302394256.connecticut Insurance Journal - Good

University of Connecticut
OpenCommons@UConn
Connecticut Insurance Law Journal School of Law
2009
Whither the Duty of Good Faith in UK Insurance Contracts?

John Lowry
Follow this and additional works at: https://opencommons.uconn.edu/cilj
Recommended Citation
Lowry, John, "Whither the Duty of Good Faith in UK Insurance Contracts?" (2009). Connecticut Insurance
Law Journal. 59.
https://opencommons.uconn.edu/cilj/59
CONNECTICUT
INSURANCE LAW
JOURNAL
Volume 16, Number 1

Fall 2009
University of Connecticut School of Law

Hartford, Connecticut
Connecticut Insurance Law Journal (ISSN 1081-9436) is published at least twice a year by
the Connecticut Insurance Law Journal Association at the University of Connecticut School
of Law. Periodicals postage paid at Hartford, Connecticut. Known office of publication:
55 Elizabeth Street, Hartford, Connecticut 06105-2209. Printing location: Western
Newspaper Publishing Company, 537 East Ohio Street, Indianapolis, Indiana 46204.
Please visit our website at http://www.insurancejournal.org or see the final page of this
issue for subscription and back issue ordering information.
Postmaster: Send address changes to Connecticut Insurance Law Journal, 55 Elizabeth

Street, Hartford, Connecticut 06105-2209.
The Journal welcomes the submission of articles and book reviews. Both text and notes
should be double or triple-spaced. Submissions in electronic form are encouraged, and
should be in Microsoft™ Word™ version 97 format or higher. Citations should conform to
the most recent edition of A UNIFORM SYSTEM OF CITATION, published by the Harvard Law
Review Association.
It is the policy of the University of Connecticut to prohibit discrimination in education,

employment, and in the provision of services on the basis of race, religion, sex, age, marital
status, national origin, ancestry, sexual preference, status as a disabled veteran or veteran
of the Vietnam Era, physical or mental disability, or record of such impairments, or mental
retardation. University policy also prohibits discrimination in employment on the basis of a
criminal record that is not related to the position being sought; and supports all state and
federal civil rights statutes whether or not specifically cited within this statement.
Copyright © 2009 by the Connecticut Insurance Law Journal Association.
Cite as CONN. INS. L.J.

CONNECTICUT
INSURANCE LAW
JOURNAL
VOLUME 16 2009-2010 NUMBER 1
EDITORIAL BOARD 2009-2010
Editor-in-Chief Managing Editor

MARK KOCHANOWICZ AGRON ETEMI
Administrative Editor Assistant Managing Editor

GWAINA WAULDON LINDSAY CORLISS
Lead Articles Editors Notes & Comments Editors Executive Editors

ALANA BARTLEY SCOTT SCHAEFFER KELLY BARRETT
SARA NADIM CAROLINE VASQUEZ STEVEN COOPER
JESSA YANNI-LAZARUS RACHEL GINSBURG
SUSAN KIRKEBY
MELISSA WONG
Legal Abstracts Editor Research Editor Associate Editors

ANNA HEHENBERGER LOUIS CRUZ MARGARET GOODELL
DOUGLAS SAUVE
Symposium Editor Technology Editor DOMINIC SPINELLI
ERIN O’LEARY TIM NAST
MEMBERS
ADAM J. ALLEGRO NATHANIEL GENTILE SHETERA PATTERSON

KENSLEY BARRETT JONATHAN HARDING JONATHAN PORZUC
OLIVER BOWERS K. RUSSELL HASAN GREGORY RADA
KENNEX CHAN RYAN LORUSSO SANA SHEIKH
DANIEL CSUKA WILLIAM MAUKE MICHAEL YEDINAK
RAPHAEL DEUTSH T. RYAN MCNEIL STEVEN ZAKRZEWSKI
SHANIQUE FENLATOR JAMES MOHER ZHIJUN ZHANG
CRYSTAL FRASER JOHN NEUMON
FACULTY ADVISOR
PATRICIA MCCOY
UNIVERSITY OF CONNECTICUT
SCHOOL OF LAW
FACULTY AND OFFICERS OF ADMINISTRATION

FOR THE ACADEMIC YEAR 2009-2010
OFFICERS OF ADMINISTRATION
Michael J. Hogan, M.A., Ph.D., President, University of Connecticut
Peter J. Nicholls , Ph.D., Provost and Executive Vice President for Academic Affairs
Jeremy Paul, J.D., Dean, School of Law
Anne Dailey, J.D., Associate Dean for Academic Affairs
Michael Fischl, J.D., Associate Dean for Research and Faculty Development
Darcy Kirk, J.D., Associate Dean for Library and Technology
Ann Crawford, M.Ed., Acting Associate Dean for Finance and Administration
Ellen K. Rutt, J.D., Associate Dean for Admissions, Student Finance and Career Services
Karen L. Demeola, J.D., Assistant Dean for Admissions and Student Finance
FACULTY EMERITI
Robert L. Bard, B.A., M.A., LL.B., J.S.D., Professor of Law, Emeritus
Phillip I. Blumberg, A.B., J.D., LL.D. (Hon.), Dean and Professor of Law and Business,
Emeritus
John C. Brittain, B.A., J.D., Professor of Law, Emeritus
Clifford Davis, S.B., LL.B., Professor of Law, Emeritus
Nell J. Newton, B.A., J.D., Dean and Professor of Law, Emeritus
Howard Sacks, A.B., LL.B., Dean and Professor of Law, Emeritus
George Schatzki, A.B., LL.B., LL.M., Dean and Professor of Law, Emeritus
Craig Shea, A.B., LL.B., LL.M., Professor of Law, Emeritus
Phillip Shuchman, B.A., M.A., LL.B., Professor of Law, Emeritus
Eileen Silverstein, A.D., J.D., Professor of Law Emeritus
Lester B. Snyder, B.S., LL.B., LL.M., Professor of Law, Emeritus
Terry Tondro, A.B., LL.B., M.Phil., Professor of Law, Emeritus
Nicholas Wolfson, A.B., J.D., George and Helen England Professor of Law, Emeritus
FACULTY OF LAW
Jill Anderson, B.A., University of Washington; J.D., Columbia University; Visiting
Assistant Professor of Law
Paul Bader, B.A., Duke University; J.D., Mercer University Walter F. George School of
Law; Assistant Clinical Professor of Law
Robin D. Barnes, B.A., J.D., State University of New York at Buffalo; LL.M., University of
Wisconsin; Professor of Law
Jon Bauer, A.B., Cornell University; J.D., Yale University; Clinical Professor of Law
Loftus E. Becker, Jr., A.B., Harvard College; LL.B., University of Pennsylvania; Professor
of Law
Bethany Berger, B.A., Wesleyan University; J.D., Yale University; Professor of Law
Robert Birmingham, A.B., J.D., Ph.D. (Econ.), Ph.D. (Phil.), University of Pittsburgh;
LL.M., Harvard University; Professor of Law
William Breetz, B.A., Dartmouth College; LL.B., University of Virginia; Executive
Director, Connecticut Urban Legal Initiative, Inc.
Sara Bronin, B.A., University of Texas; M.Sc., University of Oxford (Magdalen College);
J.D., Yale University; Associate Professor of Law
Deborah A. Calloway, B.A., Middlebury College; J.D., Georgetown University; Professor
of Law
Marcia Canavan, B.A., University of California, Los Angeles; MPH., University of
California, Los Angeles; J.D., Golden Gate University School of Law and
University of Colorado School of Law; Assistant Clinical Professor of Law
Paul Chill, B.A., Wesleyan University; J.D., University of Connecticut; Clinical Professor
of Law
Anne C. Dailey, B.A., Yale University; J.D., Harvard University; Evangeline Starr
Professor of Law and Asssociate Dean for Academic Affairs
Steven M. Davidoff, B.A., University of Pennsylvania; M.A., London Business School;
J.D., Columbia University; Associate Professor of Law
Geoffrey Dellenbaugh, A.B., Princeton University; M.A., Stanford University; Ph.D.,
University of Pittsburgh; J.D., University of Chicago; Assistant Clinical Professor
of Law
Timothy H. Everett, B.A., M.A., Clark University; J.D., University of Connecticut; Clinical
Professor of Law
Todd D. Fernow, B.A., Cornell University; J.D., University of Connecticut; Professor of
Law and Director, Clinical Programs
Richard Michael Fischl, B.A., University of Illinois; J.D., Harvard University; Professor of
Law and Associate Dean for Research and Faculty Development
Hillary Greene, B.A., J.D., Yale University; Associate Professor of Law and Director,
Intellectual Property Clinic
Kaaryn Gustafson, A.B., Harvard University; Ph.D., J.D., University of California,
Berkeley; Associate Professor of Law
Mark W. Janis, A.B., Princeton University; B.A., M.A., Oxford University; J.D., Harvard
University; William F. Starr Professor of Law
Richard S. Kay, A.B., Brandeis University; M.A., Yale University; J.D., Harvard
University; Wallace Stevens Professor of Law
Darcy Kirk, A.B., Vassar College; M.S., M.B.A., Simmons College; J.D., Boston College;
Professor of Law and Associate Dean for Library & Technology
Peter R. Kochenburger, A.B., Yale University; J.D., Harvard University; Park Associate
Clinical Professor and Director of Graduate Programs
Lewis S. Kurlantzick, B.A., Wesleyan University; LL.B., Harvard University; Zephaniah
Swift Professor of Law
Alexandra Lahav, B.A., Brown University; J.D., Harvard University; Professor of Law
Leslie C. Levin, B.S.J., Northwestern University; J.D., Columbia University; Professor of
Law
Diana L. Leyden, B.A., Union College; J.D., University of Connecticut; LL.M., Georgetown
University; Associate Clinical Professor of Law and Director, Tax Clinic
Peter L. Lindseth, B.A., J.D., Cornell University; M.A., M. Phil, Ph.D, Columbia
University; Professor of Law
Justin Long, A.B., Harvard University; J.D., University of Pennsylvania; Visiting Assistant
Professor of Law
Hugh C. Macgill, B.A., Yale University; LL.B., University of Virginia; Oliver Ellsworth
Research Professor of Law and Dean Emeritus
Neysun Mahboubi, A.B., Princeton University; J.D., Columbia University; Visiting
Assistant Professor of Law
Jennifer Brown Mailly, B.A., Brown University; J.D., Ohio State University; Assistant
Clinical Professor of Law
Margaret Martin, B.A., Boston University; J.D., Columbia University; William R. Davis
Clinical Teaching Fellow
Ruth Mason, B.A., Columbia University; J.D., Harvard University; Associate Professor of
Law
Patricia A. McCoy, B.A., Oberlin College; J.D., University of California, Berkeley;
Director of Insurance Law Center and George J. and Helen M. England Professor
of Law
Barbara McGrath, B.A., Yale University; J.D., University of Connecticut; Assistant
Director, Connecticut Urban Legal Initiative, Inc.
Willajeanne F. McLean, B.A., Wellesley College; B.S., University of Massachusetts; J.D.,
Fordham University; LL.M., Free University of Brussels; Professor of Law
Thomas H. Morawetz, A.B., Harvard College; J.D., M.Phil., Ph.D., Yale University;
Tapping Reeve Professor of Law and Ethics
Lily Neff, B.S.E.E, City College of New York; J.D., Pace University; Assistant Clinical
Professor of Law
R. Kent Newmyer, B.A., Doane College; Ph.D., University of Nebraska; Professor of Law
and History
Angel R. Oquendo, A.B., M.A., Ph.D., Harvard University; J.D., Yale University; Olimpiad
S. Ioffe Professor of Law
Leonard Orland, B.A., Rutgers University; LL.B., University of Pennsylvania; Oliver
Ellsworth Research Professor of Law
Sachin Pandya, B.A., University of California, Berkeley; M.A., Columbia University; J.D.,
Yale University; Associate Professor of Law
Richard W. Parker, A.B., Princeton University; J.D., Yale University; D.Phil., Oxford
Jeremy Paul, A.B., Princeton University; J.D., Harvard University; Dean and Thomas F.
Gallivan, Jr. Professor of Real Property Law
Ellen Ash Peters, B.A., Swarthmore College; LL.B., Yale University; LL.D., Yale
University; University of Connecticut; et al.; Visiting Professor of Law
Richard D. Pomp, B.S., University of Michigan; J.D., Harvard University; Alva P. Loiselle
Professor of Law
Jessica S. Rubin, B.S., J.D., Cornell University; Assistant Clinical Professor of Law
Susan R. Schmeiser, A.B., Princeton University; J.D., Yale University; Ph.D., Brown
Peter Siegelman, B.A., Swarthmore College; M.S.L., Ph.D., Yale University; Roger
Sherman Professor of Law
James H. Stark, A.B., Cornell University; J.D., Columbia University; Professor of Law and
Director, Mediation Clinic
Martha Stone, B.A., Wheaton College; J.D., LL.M., Georgetown University; Director,
Center for Children’s Advocacy
Kurt A. Strasser, B.A., J.D., Vanderbilt University; LL.M., J.S.D., Columbia University;
Phillip I. Blumberg Professor of Law
Colin C. Tait, A.B., Cornell University; LL.B., Yale University; Oliver Ellsworth Research
Professor of Law
Stephen G. Utz, B.A., Louisiana State University; J.D., University of Texas; Ph.D.,
Cambridge University; Professor of Law
Carol Ann Weisbrod, J.D., Columbia University; Ellen Ash Peters Professor of Law
Robert Whitman, B.B.A., City College of New York; J.D., Columbia University; LL.M.,
New York University; Professor of Law
Steven Wilf, B.S., Arizona State University; Ph.D., J.D., Yale University; Professor of Law
Richard A. Wilson, BSc., Ph.D., London School of Economics and Political Science;
Professor of Law
ADJUNCT FACULTY OF LAW

Peter G. Austin, B.S., University of Connecticut; M.S., University of Hartford; Adjunct
Professor of Law
Morris W. Banks, A.B., Dartmouth College; LL.B., Columbia University; LL.M., New
York University; Adjunct Professor of Law
Anne Davis Barry, B.S., University of Connecticut; M.S., Union College; J.D., University
of Connecticut; Lecturer in Law
Richard Baxter, B.A., Bucknell University; M.A., Yale University; J.D., Duke University;
Lecturer in Law
James W. Bergenn, B.A., Catholic University; J.D., Columbia University; Adjunct Professor
of Law
Hon. David M. Borden, B.A., Amherst College; LL.B., Harvard University; Adjunct
Professor of Law
John Buchanan, A.B., Princeton, B.A., Oxford; J.D., Harvard; Lecturer in Law
Michael A. Cantor, B.S., J.D., University of Connecticut; Adjunct Professor of Law
Christine Chabot, B.A., Northwestern University; J.D., Notre Dame Law School; Lecturer
in Law
James N. Cohen, B.A., J.D., Harvard; Lecturer in Law
Gary Collins, B.A., State University of New York, Buffalo; J.D., Vanderbilt University;
Lecturer in Law
David W. Cooney, B.A., Rutgers University; M.A., Wesleyan University; J.D., University
Dean M. Cordiano, B.A., State University of New York at Binghamton; J.D., Duke
University; Lecturer in Law
Michelle Cruz, B.A., M.A., Mount Holyoke College; J.D., University of Connecticut;
Lecturer in Law
Renee M. Dailey, B.A., J.D., University of Connecticut; Lecturer in Law
John G. Day, A.B., Oberlin College; J.D., Case Western Reserve University; Professor in
Residence
Raymond DeMeo, B.A., Williams College; J.D., University of Connecticut; Lecturer in Law
Mark Dubois, B.A., College of the Holy Cross; J.D., University of Connecticut; Lecturer in
Law
Gerald P. Dwyer, A.B., Colgate University; J.D., Gonzaga University; LL.M., Georgetown
Paul Eddy, B.A., Ohio Wesleyan University; J.D., University of Connecticut; Lecturer in
Law
Thomas Farrish, B.A., J.D., University of Connecticut; Lecturer in Law
Rebecca Flanagan, B.A., M.A., University of Connecticut; J.D., University of North
Carolina; Lecturer in Law
Evan D. Flaschen, B.A., Wesleyan University; J.D., University of Connecticut; Adjunct
Professor of Law
Charles W. Fortune, B.A., University of Rochester; J.D., University of Connecticut;
Lecturer in Law
Michael A. Gailor, B.A., Cornell University; J.D., University of Connecticut; Lecturer in
Law
William Goddard, B.A., M.B.A., Dartmouth; J.D., University of Connecticut; Lecturer in
Law
Ira H. Goldman, B.A., Cornell University; J.D., Yale University; Adjunct Professor of Law
Daniel L. Gottfried, B.A., J.D., University of Connecticut, Lecturer in Law
Andrew S. Groher, B.A., University of Virginia; J.D., University of Connecticut; Adjunct
Professor of Law
Albert B. Harper, B.A., University of Texas; J.D., Ph.D., University of Connecticut; Adjunct
Professor of Law
John Harris, B.S.E., Princeton; J.D., University of Wisconsin; Adjunct Professor of Law
Wesley Horton, B.A., Harvard University; J.D., University of Connecticut; Adjunct
Professor of Law
John Houlihan, B.A., Providence College; J.D., St. John’s University; Lecturer in Law
Daniel Klau, B.A., University of California, San Diego; J.D., Boston University; Lecturer in
Law
Charles H. Klippel, B.A., M.Ed., M.P.H., J.D., Harvard University; Lecturer in Law
Barry Kramer, B.Ch.E., Rensselaer Polytechnic Institute; J.D., New York University;
Lecturer in Law
John H. Lawrence, Jr., B.S., Washington and Lee University; J.D., University of Virginia;
Lecturer in Law
Henry C. Lee, B.S., John Jay College of Criminal Justice; M.S., Ph.D., New York
University; Dr.Sci. (Hon.), University of New Haven; Dr.Hum. (Hon.), St. Joseph
College; Adjunct Professor of Law
Erik T. Lohr, B.S., Thomas A. Edison State College; J.D., University of Connecticut;
Lecturer in Law
James Lotstein, B.S., Northwestern University; J.D., University of Connecticut, Lecturer in
Law
Joseph A. MacDougald, B.A., Brown University; M.B.A., New York University; J.D.,
University of Connecticut; M.E.M., Yale University; Lecturer in Law
Robert G. Madden, B.S.W., Providence College; M.S.S.W., Columbia University School of
Social Work; J.D., University of Connecticut, Lecturer in Law
Thomas S. Marrion, A.B., College of the Holy Cross; J.D., University of Connecticut;
Adjunct Professor of Law
James F. Meehan, B.A., University of Arizona; J.D., University of Connecticut; Lecturer in
Law
Dwight Merriam, B.A., University of Massachusetts, M.R.P., University of North Carolina,
J.D., Yale University; Lecturer in Law
Jamie L. Mills, B.A., Eastern Connecticut State University; J.D., University of Connecticut;
Lecturer in Law
Joseph Mirrione, B.A., Marist College; J.D., Vermont Law School; Lecturer in Law
Derek L. Mogck, B.A., Gordon College; M.S., University of Massachusetts; J.D., University
Umar F. Moghul, B.A., M.A., University of Pennsylvania; J.D., Temple University;
Lecturer in Law
Thomas B. Mooney, B.A., Yale University; J.D., Harvard University; Adjunct Professor of
Law
Ronald T. Murphy, B.S., Springfield College; J.D., University of Connecticut; Lecturer in
Law
Henry F. Murray, B.A., Yale University; J.D. University of Connecticut; Lecturer in Law
Kevin O’Connor, B.A., University of Notre Dame; J.D., University of Connecticut; Lecturer
in Law
Andrew J. O’Keefe, B.S., College of the Holy Cross; J.D., University of Connecticut;
Cornelius O’Leary, B.A., Williams College; M.A., Trinity College; J.D., University of
Connecticut; Adjunct Professor of Law and Mark A. Weinstein Clinical Teaching
Fellow
John Pagini, B.A., Central Connecticut State University, M.P.A., University of Hartford;
Lecturer in Law
Rosemarie Paine, B.S., Southern Connecticut State University; J.D., University of
Connecticut; Lecturer in Law
Humbert J. Polito, Jr., A.B., College of the Holy Cross; J.D., University of Connecticut;
Elliott B. Pollock, A.B., Columbia College; LL.B., Columbia Law School; Lecturer in Law
Hon. Elliot D. Prescott, B.A., University of Massachusetts; J.D., University of Connecticut;
Lecturer in Law
Renee C. Redman, B.A., Michigan State University; J.D., Brooklyn Law School; Lecturer
in Law
Leah Reimer, B.S., J.D., Ph.D., University of Connecticut; Lecturer in Law
Roger Reynolds, B.A., Macalester College; J.D., New York University; Lecturer in Law
Barbara Rezner, B.A., Gettysburg College; J.D., Washington and Lee University; Lecturer
in Law
Louis Ricciuti, B.A., Boston College; J.D., University of Connecticut; Lecturer in Law
Tracy Rich, B.A., Union College; J.D., New York University School of Law; L.L.M.,
Boston University School of Law; Lecturer in Law
James K. Robertson, Jr., B.A., Yale University; M.A., Hartford Seminary; J.D., University
Morgan Paul Rueckert, B.A., J.D., University of Connecticut; Lecturer in Law and Mark A.
Weinstein Clinical Teaching Fellow
Patrick J. Salve, B.S., J.D., University of Pennsylvania; Adjunct Professor of Law
Hon. Michael R. Sheldon, A.B., Princeton University; J.D., Yale University; Adjunct
Professor of Law
Sandra L. Sherlock-White, B.A., Central Connecticut State College; J.D., Western New
England College; Lecturer in Law
Jay. E. Sicklick, B.A., Colgate University; J.D., Boston College; Adjunct Professor of Law
Douglas Simpson, A.B., Dartmouth College; J.D., University of Connecticut; Lecturer in
Law
Mark H. Taylor, B.A., Drew University; J.D., University of Connecticut; Lecturer in Law
James R. Turcotte, B.A., Boston University; J.D., Western New England College; Lecturer
in Law
Stefan Underhill, B.A., University of Virginia; B.A., Oxford University; J.D., Yale
Arthur Webster, B.A., Bradley University; J.D., Georgetown University; Adjunct Professor
of Law
William Wilcox, B.A., University of Connecticut; J.D., Suffolk University; Lecturer in Law
Michael Wilder, B.A., Yale University; LL.B., Harvard University; Lecturer in Law
Paul B. Zolan, B.A., Trinity College, J.D., University of Connecticut; Lecturer in Law
CONNECTICUT
INSURANCE LAW
JOURNAL
VOLUME 16 2009-2010 NUMBER 1
CONTENTS
ESSAYS
CREDIT DERIVATIVES ARE NOT “INSURANCE” M. Todd Henderson 1
THE ROAD FROM “TWIN PEAKS” –

AND THE WAY BACK Michael W. Taylor 61
ARTICLES
WHITHER THE DUTY OF GOOD FAITH

IN UK INSURANCE CONTRACTS? John Lowry 97
RISK DATA IN INSURANCE INTERPRETATION Michelle Boardman 157
THE LAW AND ECONOMICS OF FIRST-PARTY Sharon Tennyson

INSURANCE BAD FAITH LIABILITY William J. Warfel 203
REGULATION OF LARGE FINANCIAL

INSTITUTIONS: LESSONS FROM John P. Harding
CORPORATE FINANCE THEORY Stephen L. Ross 243
NOTES AND COMMENTARIES
PREDATORY LENDING AND ITS

INSURANCE CONSEQUENCES Erin O’Leary 261
THE 2008 MENTAL HEALTH PARITY AND ADDICTION EQUITY ACT:
AN OVERVIEW OF THE NEW LEGISLATION AND
WHY AN AMENDMENT SHOULD BE PASSED TO
SPECIFICALLY DEFINE MENTAL ILLNESS AND
SUBSTANCE USE DISORDERS Sara Nadim 297
EXAMINING CURRENT PROPOSALS FOR

INCREASING THE FEDERAL ROLE IN
DEALING WITH COASTAL HURRICANE RISK Louis Cruz 323
CREDIT DERIVATIVES ARE NOT “INSURANCE”
M. Todd Henderson∗
***
This article explores whether credit derivatives should be regulated as

insurance and offers an alternative form of regulation for these financial
instruments. The largely unregulated credit derivates market has been
cited as a cause of the recent collapse of the housing market and resulting
credit crunch. The article explores the possibility that the credit
derivatives market should be regulated as insurance. It shows that the
argument that some credit derivatives help banks and other providers of
debt share risk with other investors is not sufficient for credit derivative
contracts in general to be deemed “insurance.” It concludes that
insurance regulation is not suitable for the credit derivatives market, while
conceding that some sort of regulation may be necessary. The first section
provides an overview of the basics of credit derivatives. The second
section presents the argument for regulating credit derivatives as
insurance. Section III describes why, although credit derivatives contracts
can result in risk sharing or transfer, they are not within insurance law.
The final section describes what one form of regulation of credit
derivatives could look like and contrasts this with insurance regulation.
***
I. INTRODUCTION
The collapse of the housing bubble and the resulting credit crunch
has caused untold harm to the economy and the lives of millions by
destroying trillions of dollars in global wealth. The search for causes and
remedies has begun in earnest, and chief among these is the largely
unregulated credit derivatives market. Regulation of one form or another is
the proposed solution in many quarters, and one of the prominent proposals
is insurance regulation. At the very least, the analogy between credit
derivatives and insurance is often made, and this faulty comparison may
lead regulators astray, regardless of the mode of regulation ultimately
∗
University of Chicago Law School. Thanks to the George J. Phocas Fund for
research support.
2 CONNECTICUT INSURANCE LAW JOURNAL [Vol. 16:1
chosen. This Essay explores the suitability of insurance regulation to the

credit derivatives market, concluding that it is a bad fit along many
dimensions. Regulation of some sort may indeed be needed to remedy
some fairly obvious market failures,1 but insurance regulation and
regulators have little if any role to play in any new regulatory regime.
The most basic form of credit derivative, known as a “credit
default swap” (CDS), is simply a contract through which a lender can
protect against the risk of default by paying premiums to a third party who
agrees to make the lender whole in the event of default by the underlying
borrower. The surficial similarity to typical insurance products, like
property or life insurance, has caused some politicians and pundits to argue
that credit derivatives are a form of insurance and should be regulated as
such.2 The former director of the Commodities Futures Trading
Commission (CFTC), which regulates most derivative products, declared:
“A credit default swap . . . is an insurance contract, but [the industry has]
been very careful not to call it that because if it were insurance, it would be
1
One market failure was the lack of a centralized clearinghouse to manage
and reduce counterparty risk in credit derivative transactions. The Federal Reserve
and Treasury are encouraging exchanges, like the Chicago Mercantile Exchange,
to handle these transactions. See USA Exchanges: Geithner Pushes for Derivatives
Shake-Up, FINREG21, July 11, 2009, http://www.finreg21.com/news/usa-
exchanges-geithner-pushes-derivatives-shake-up.
2
As shown in supra note 1, numerous politicians and observers have made
the linkage. It has also crept casually into numerous media accounts. For example,
in an account of the AIG catastrophe, an author for The New Republic calls credit
derivatives insurance: “Between March, when Greenberg left AIG, and the end of
2005, Cassano's division issued more than $40 billion in credit-default swaps
(essentially insurance) for portfolios of securities backed by subprime mortgages.
This was more than half of all the insurance of this type the company had on its
books.” Noam Scheiber, A New Theory of the AIG Catastrophe, THE NEW
REPUBLIC, Apr. 15, 2009, at 10, 11. Legal scholars believe this too. See Robert F.
Schwartz, Risk Distribution in the Capital Markets: Credit Default Swaps,
Insurance and a Theory of Demarcation, 12 FORDAM J. CORP. & FIN. L. 167, 181
(2007) (arguing that certain credit derivative contracts have “general form and
function reflect[ing] many basic insurance arrangements.”); William K. Sjostrom,
Jr., The AIG Bailout, 66 WASH. & LEE L. REV. (forthcoming 2009) (“A CDS
certainly appears to fall within this definition given that the protection seller
contractually agrees to compensate the protection buyer following the occurrence
of a credit event.”), available at http://papers.ssrn.com/sol3/papers.cfm?
abstract_id=1346552.
2009] CREDIT DERIVATIVES ARE NOT “INSURANCE” 3
regulated.”3 New York State went even further. On September 22, 2008,
Governor David Patterson announced plans to regulate credit derivatives as
insurance under the auspices of the state’s insurance department.4 New
York State Insurance Commissioner Eric Dinallo then testified before a
House Committee investigating credit derivatives: “the insurance regulator
for New York is a relevant authority on credit default swaps,” because
“[w]e believe . . . [they are] insurance.”5 Although New York has delayed
its regulatory plans pending a federal review of credit derivative
regulation,6 the question of whether credit derivatives are insurance
remains an open and much bandied about one that needs to be analyzed.
3
60 Minutes: Wall Street’s Shadow Market; Credit default swaps (CBS
television broadcast Oct. 5, 2008), available at http://www.cbsnews.com
/stories/2008/10/05/60minutes/main4502454_page1.shtml. Dr. Greenberger
argued that the industry was able to avoid regulation by simply using the word
“swap” instead of “insurance” to describe the transaction. See id. (“So they use a
magic substitute word called a 'swap,' which by federal law is deregulated.”).
Swaps were specifically excluded from regulation by the CFTC by the
Commodities Futures Modernization Act of 2000.
It is true that a typical CDS transaction does not involve a “swap” in the same
way that an interest rate swap does. In the latter case, two parties agree to swap
risks: the holder of a fixed-rate note agrees to swap that income stream with the
holder of a variable-rate note. But while the term swap is not technically accurate,
it is difficult to imagine that insurance or other regulators would be fooled by the
label.
4
Press Release, Governor Patterson Announces Plan to Limit Harm to
Markets from Credit Derivatives (Sept. 22, 2008) (on file with Errol Cockfield)
available at http://www.ins.state.ny.us/press/2008/p0809224.pdf. See also Danny
Hakim, New York to Regulate Financial Tool Behind the Credit Crisis, N.Y.,” NY
TIMES, Sept. 23, 2008, at C10, available at http://www.nytimes.com/2008/09
/23/business/23swap.html?ref=business (“The governor said the state’s insurance
department would begin regulating credit-default swaps as insurance products in
cases where the buyer of the swap also owns the underlying bond it is meant to
back.”).
5
Hearing to Review the Role of Credit Derivatives in the U.S. Economy:
Hearing before the H. Comm. on Agriculture Committee, 110th Cong. (2008)
[hereinafter Hearing] (testimony of Eric Dinallo, Ins. Comm’r, N.Y. State),
available at http://agriculture.house.gov/testimony/110/h91120/Dinallo.pdf.
6
See id. (“Based on the developments reported on by the President’s Working
Group, it is clear they are committed to comprehensive and effective federal
oversight of credit default swaps. . . . . Accordingly, New York will delay
indefinitely our plan to regulate part of this market.”). It is clear from Dinallo’s
testimony that New York is using the threat of insurance regulation as a weapon to
This Essay argues that it makes little or no sense to regulate credit

derivatives as or like “insurance,” regardless of whether they are used as
insurance, that is, to reduce risk for one party. The instinct to call credit
derivatives “insurance” is sensible enough – the lender buying credit
protection looks much like an insured and the party selling credit protection
looks much like an insurer, at least where the protection seller is in privity
with holders of notes of indebtedness.7 The analogy is obvious: in a plain-
vanilla CDS, the bank making an original loan pays a premium to a third
party that in turn agrees to make the bank whole in the event of a future
liability, that is, a default on the underlying loan or bond. This transaction
resembles a typical insurance contract, where the insured pays a premium
to a third party (an insurance company) in return for a promise to make the
insured whole in the event of a loss.
But observing that something resembles or provides insurance
against loss is not enough to warrant regulating it as “insurance.” Many
contracts that are not called insurance or regulated as insurance imbed
some component of insurance or risk sharing. For instance, when a farmer
enters into a contract that allows the farmer to sell wheat at a fixed price in
the future – a forward contract called a put option – the farmer is in effect
insuring against a drop in the price of wheat. On the other side of this
transaction, there may be a baker who enters into a forward contract that
allows the baker to insure against an increase in the price of wheat. Both
parties are buying price insurance from each other, likely with a
middleman, known as a market maker, standing between and reducing the
counterparty risk in the transaction. But these contracts, and all similar
hedging contracts entered into by regular consumers and sophisticated
financial entities, are not regulated as insurance contracts.8 The point can
encourage more comprehensive federal regulation. (“Based on the developments

reported on by the President’s Working Group, it is clear they are committed to
comprehensive and effective federal oversight of credit default swaps. . . .
Accordingly, New York will delay indefinitely our plan to regulate part of this
market.”).
7
New York State has proposed regulating only these credit derivatives –
about one-fifth of the total market – because the argument that the parties are
engaged in an insurance transaction is more difficult in cases where they are
simply wagering on the default without an actual interest in the underlying debt
instrument. See id.
8
Option contracts generally trade on exchanges, like the Chicago Mercantile
Exchange, and as such are regulated by the Commodities Futures Trading
Commission (CFTC).
be made more bluntly: it would be fanciful to argue that every contract in

which a party could be said to be reducing its risk and another party was
willing to take on some of that risk is or should be called insurance. If this
were the case, state insurance regulators would be involved in regulating
hedge funds, commodities, options, swaps, and countless other contracts
entered into by consumers and firms. In fact, every contract assigns, shares,
and apportions some sort of risk. No one seriously advocates this scope for
insurance regulation. Simply providing some risk sharing is not enough to
be regulated by state insurance commissioners.
The reason insurance regulation does not extend to every contract
that involves some element of insuring risk has to do with the purpose of
insurance regulation, as opposed to other types of regulation. There are
broadly two justifications for a special law of insurance: first, the peculiar
governance problems associated with insurance firms; and second, worries
about unsophisticated consumers being duped by complicated and essential
products. This Essay will show that neither of these justifications obtains or
makes sense for the regulation of credit derivatives.
Governance problems arise because insurance companies have an
inverted production cycle and do not generally have concentrated creditors
like non-insurance firms. This means that two crucial constraints on the
potential misuse of investment resources are missing: the feedback to the
firm provided by product and other markets is missing given the fact that
the insurance company produces its product (that is, payment of claims)
many years after the consumers pay for it; and when things go badly for the
insurance company, there is no concentrated interest to keep the firm from
adopting an excessively risky strategy (from the perspective of creditors
[that is, policy holders]).
Insurance law is designed to prevent the risk that insurers
competing for policyholders, but unconstrained by normal forces, will
charge too little for their products. This happens because of the continuous
nature of insurance company inflows and outflows, coupled with a
delinkage between the time of pricing a risk and the time of paying out the
loss from the risk. In other words, insurance can look a bit like a Ponzi
scheme, where new creditors of the firm are paying off the liabilities to old
creditors. And, just as in a Ponzi scheme, when things go badly for the firm
(that is, when actuarial estimates of liability turn out to be wrong), there is
a natural tendency to offer new investors an attractive return to increase
cash flows to pay for higher-than-estimated outflows.
The second part of the governance problem – the lack of
concentrated creditors – exacerbates this problem, since there is no
sophisticated entity with bargaining power that can keep the firm from
adopting a shareholder-friendly, go-to-Vegas strategy in the event liability

estimates were erroneous. Without these governance constraints, initial
misestimates and mistakes can fester and lead to large losses. This Essay
shows how the counterparties in credit derivative contracts do not have this
continuous investment problem or these governance problems, unless, of
course, they are insurance companies, and how insurance regulation would
be futile in any event.
Consumer problems arise because the consumers of insurance
company products (and as such creditors of the insurance company) are
average individuals without the expertise or sophisticated judgment to
assess what they are buying in insurance products. The consumer-centric
element of insurance regulation consists of three commonly recited
justifications: to make sure insurers don’t charge too much; to regulate the
substance and terms of policies; and to regulate service and coverage
issues. This basis for regulation is, to be sure, driven by a rather dim view
of the philosophy of caveat emptor, the wisdom and skill of the average
consumer, and the power of a small number of informed individuals to set
market prices for others. This Essay does not take on the soundness of
these consumer issues for insurance products, but simply compares their
rationale with what is known about the participants in credit derivative
contracts. Unlike the average consumer of insurance, the average
participant in credit derivative markets is large, sophisticated, and capable
of bearing losses. There is simply no basis for transferring the paternalistic
impulses of insurance to this market.
This Essay shows that neither the governance problem nor the
consumer abuse problem obtain in significant ways in the context of the
credit derivatives market. Section I introduces the basics of credit
derivatives. Section II presents the argument for regulating credit
derivatives as insurance based on the rough analogy describe above.
Section III then shows why the simple fact that credit derivatives
sometimes result in risk sharing or transfer does not justify bringing these
contracts and the parties to them within the ambit of insurance law. Section
IV concludes by briefly sketching out what a sensible regulation of credit
derivatives might look like, contrasting this with the approach of insurance
regulation.
II. A PRIMER ON CREDIT DERIVATIVES
Credit derivatives exist in many forms and flavors, but the essence
is simple: it may be more efficient for different entities to handle the
various aspects of lending. A typical loan has many parts, including:
origination, servicing, monitoring, and funding or risk bearing. In a world

without risk-sharing mechanisms, all of these are contained within one
entity, that is, a bank. The bank has the relationships (origination), scale in
the back office (servicing), experience (monitoring), and cash from
depositors (funding). But the bank might not want to do all these things. It
might want to become an arranger rather than a lender. One reason is
because federal regulations designed to protect depositors require the bank
to hold cash on hand to offset risk in loans.9
Another reason is that other potential lenders may be shut out of
the corporate lending market, say because they do not have relationships
with borrowers, but would provide a cheaper source of funding or be more
efficient holders of particular aspects of corporate borrowing risk. Smaller
regional banks and insurance companies come to mind here.
A final reason is that the bank may not be the most efficient
monitor of firm conduct. The bank has experience with monitoring in
general and (likely) with the specific borrower, but these advantages come
with costs too. The relationships that led to the loan may corrupt the
monitoring function—a sort of monitor capture by the borrower. Fees
earned by banks for workouts and new loans may also distort incentives. So
too might the fact that the workout group for a loan may be comprised of
only a few individuals, who are subject to biases and shortcomings that a
larger, market-based monitoring mechanism might be able to overcome. In
short, there are many reasons why banks might prefer to decouple the
bundle of loan features, but cannot without financial contracts that allow
default risk to be shared. Credit derivatives, although much maligned as a
result of current events, can help investors of various sorts allocate the
lending market to its most efficient participants.
There are many variations of credit derivatives, but to answer the
threshold question of whether credit derivatives in general can be
considered “insurance”, it makes sense to consider the two most generic
versions: the credit default swap (CDS) and the collateralized debt
obligation (CDO).
A. CREDIT DEFAULT SWAPS
A CDS is a contract in which credit risk (that is, the expected

losses arising from defaults) is transferred from one party to another. A
9
Capital reserves required by the Basle Accords are non-productive, and
therefore reduce a bank’s return.
bank makes a loan to a borrower. There is some risk that the borrower will
default on the loan, causing losses to the bank.10 Naturally, the bank wants
to minimize these losses. The bank can do this through ex ante selection
(that is, due diligence during the underwriting process), through ongoing
monitoring of the borrower, and through effective ex post workout
procedures. The bank can also contract with a third party to make the bank
whole in the event the borrower defaults.
Consider a simple example: Bank holds on its balance sheet a $100
note for a loan made to Borrower. Bank may want to shift some of the risk
that Borrower will not repay the loan, say because of costly federal capital
adequacy requirements that require Bank to hold some percentage of the
loan’s outstanding balance in cash reserves.11 For a period of (normally)
five years, Investor, who wants to hold risk of Borrower, agrees to make
Bank whole in the event of default, thus assuming the risk of default,12 in
return for a stream of periodic payments from Bank. Voilà, the risk of the
loan to Borrower has been swapped from Bank to Investor. The premium
paid by the Bank is expressed as a risk spread in basis points, say 100 basis
points or 1 percent. For a $100 loan, this would mean the bank would make
quarterly payments of $0.25 to buy protection on the note. (The spread,
which expresses the risk of default during the five-year term of protection,
varies over time, allowing information about the quality of the debtor to be
revealed and allowing investors unrelated to the loan contract to speculate
on changing credit quality for profit.) In the event of “default,”13 Bank
delivers the underlying credit instrument, in this case the loan, to Investor,
and Investor makes a payment to Bank that puts Bank in the position it
would have been in if Borrower had not defaulted.
10
The default risk is only one of many risks embedded in a loan. Lenders
(and borrowers) face interest rate risks, volatility risk, currency risks, and so on.
The significance of credit derivatives (CDSs) is the ability to unpack and isolate
credit risk, and allowed it to be transferred to others who may be more efficient
holders of it.
11
In general, US banks are subjected to the capital adequacy requirements of
the so-called Basle Accords, implemented by the Bank of International
Settlements. See Basle Committee on Banking Supervision, International
Convergence of Capital Measurement and Capital Standards, July 1998, available
at http://www.bis.org/publ/bcbs04a.htm.
12
Default risk is only one of many types of risk. Others include: interest rate
risk, counterparty risk, currency risk, and so on.
13
As described below, see infra p. 7, the issue of when a credit derivative
contract triggers payment is a complicated and tricky issue.
These contracts constitute the primary market for credit derivatives

since the parties to the transaction interact with the underlying debt
instrument: the lender writes and initially holds the instrument, while the
counterparty ends up holding the instrument in the event of default, and, in
any event, one of the two parties to the contract will hold the underlying
debt at the termination of the CDS contract. The participants in this market
are large commercial banks, as risk sellers, and insurance companies, hedge
funds, pension funds, mutual funds, and a mix of investment banks,
commercial banks, and smaller regional banks, as risk buyers.14 In this
way, CDS contracts resemble other risk-sharing arrangements, like the
syndication of credit or the sale of loans by banks. Most large loans are
shared between a lead lender and other banks with which it contracts to
share the risk of default, and there is a large and robust market for the sale
of all or parts of loans to other banks. (Although risk sharing contracts,
these are not considered or regulated as insurance.)
Returning to the example above, Bank will want to reduce the $100
risk by getting other investors to participate, both to reduce its own risk and
also to comply with capital adequacy rules. Bank could sell the loan, but
this might mean giving up its relationship with Borrower, something
neither party might want.15 Shifting the risk using a CDS preserves this
relationship – in fact, Borrower may not even know the risk has been
shifted – while also allowing conservative investors, like insurance
companies and pension funds, to participate in the credit market.16 A small
regional bank in Germany, an insurance company in Indonesia, and a
pension fund in California are thus able to achieve desired risk-return
investments in new ways.
14
See U.S. GEN. ACCOUNTING OFFICE, CREDIT DERIVATIVES: CONFIRMATION
BACKLOGS INCREASED DEALERS’ OPERATIONAL RISK, BUT WERE SUCCESSFULLY
ADDRESSED AFTER JOINT REGULATORY ACTION, GAO-07-716, p.6 n.8 (2007),
available at http://www.gao.gov/new.items/d07716.pdf (“The top five end-users of
credit derivatives are banks and broker-dealers (44 percent), hedge funds (32
percent), insurers (17 percent), pension funds (4 percent), and mutual funds (3
percent).”).
15
Borrower might not want Bank to sell the loan, since this may signal
something bad about Borrower. The positive signal derived from having Bank be a
creditor and monitor of Borrower may be quite valuable and for this reason, loan
agreements often include no-sale clauses.
16
Conservatism here may derive internally, that is, from managers and
shareholders, or from regulation.
The risk is not just swapped between two parties. In a typical CDS
contract, the risk can be swapped many times, so that there are numerous
parties in the risk-sharing chain. For instance, A, a manufacturing firm may
borrow from B, a commercial bank. B, not wanting to hold the risk, may
pay a premium to C, a dealer in CDS contracts, in return from a promise to
be made whole if A defaults. (C, like all protection sellers, will be required
to post some collateral with B to ensure payment on this obligation.) Since
C is a dealer, it will look for a buyer willing to provide the ultimate risk-
bearing function. D, an insurance company, agrees to make C whole in the
event of a default by A in return for premium payments by C. Then E, a
different commercial bank, wants exposure to A’s credit risk, so it may
agree to make D whole in the event of a default by A, in return for the
payment of premiums by D. And on and on and on. There is no limit on the
number of links in the risk-sharing chain, and, in practice, credit risk is
often transferred dozens of times after its original creation. A typical credit
derivative contract has hundreds of investors selling protection for
hundreds of lenders and even more underlying borrowers. In the mortgage
securitization market, for example, one of the problems in the collapse of
US house prices was figuring out who actually held the risk of mortgage
default so that workouts or foreclosures could happen efficiently.17
As discussed below, although C, D, E, and parties on down the
chain could be said to be providing risk-sharing contracts in this example, it
would be a dramatic expansion of the concept of insurance regulation to
call them insurance companies. These entities might be individuals, banks,
hedge funds, university endowments, or any other pool of investment
money looking for return. In addition, the kind of insurance they are
providing is not dissimilar from the insurance provided by nearly every
contract that involves risk sharing (that is, every contract), and therefore
raises irresolvable line drawing problems. The closest entity to an insurance
company is, C the original CDS dealer. But, as discussed below, these are
brokers who are regulated by numerous securities and banking laws, and
subject to the oversight of numerous federal regulators.
CDS contracts do have characteristics similar to typical insurance
contracts. Specifically, risk sharing and information asymmetries inevitably
give rise to problems of moral hazard and adverse selection. Bank knows
more about default risk of Borrower than the counterparties, and therefore
17
See, e.g., Mike McIntire, “Tracking Loans Through a Firm That Holds
Millions,” N.Y. TIMES, Apr. 23, 2009, available at http://www.nytimes.com/
2009/04/24 /business/ 24mers.html.
the latter may be unwilling to sell protection on the notes Bank brings to
the market, since the counterparties may believe these are the debtors most
likely to default. Another problem is that the existence of credit protection
may make Bank less diligent in its monitoring role, thereby increasing the
risk of default as a result of the risk-sharing contract. In theory and
practice, there are steps that can be taken to mitigate these risks. Bank can
hold back a portion of the risk of default, perhaps the first-loss position,
thereby giving it incentives to monitor. This is analogous to a deductible in
insurance contracts, and it can address both the moral hazard and adverse
selection problem. (As it turns out, however, the nature of the securitization
process made these first-loss tranches more valuable, on a risk-adjusted
basis, than their price, while more senior tranches were less valuable.)
These problems and the steps taken to mitigate them are discussed below.
There is also a rich secondary market in which the risk of default of
a particular borrower (known as a “reference entity”) is traded among
parties that have no contact with or affiliation with either the borrower or
the lender. For instance, auto parts maker Delphi had $2 billion in bonds
outstanding at the time it declared bankruptcy, but there were over $25
billion in credit derivative bets outstanding on whether or not Delphi would
default on those bonds.18 The term “bet” is chosen deliberately, since these
contracts are nothing more than wagers on whether Delphi would default.
(As a side note, we do not regulate these bets as gambling for the same
reason that the secondary market in stocks, that is, the New York Stock
Exchange, is not regulated as gambling, even though it is. The reason is
that the gambling is socially useful.) This large ratio of secondary to
primary market is common across companies used as reference entities.
After all, there is nothing (other than perhaps gambling law) that prevents
two parties from writing a contract that replicates the payoffs from the
payment or default of any debt instrument entered into anywhere. These
contracts are called “synthetic,” since they do not involve any physical
obligations to deliver on the underlying debt instrument.
The proposals to date to regulate credit derivatives have focused
entirely on the primary market, specifically disclaiming any authority over
the secondary market. As discussed below, this has something to do with
what insurance experts call “insurable interest,” which is a requirement that
the party allegedly doing the insuring has to pay only when the party that is
18
The Ballooning Credit Derivatives Market: Easing Risk or Making It
Worse?, KNOWLEDGE@WHARTON, Nov. 2, 2005, http://knowledge.wharton.
upenn.edu/article.cfm?articleid=1303.
allegedly insured actually suffers a harm unrelated to the insurance

contract. But once regulators limit control over one part of the market, the
fungibility of financial products will allow investors to move to other
unregulated products that give them the same mix of risk and return. This is
discussed further below.
B. COLLATERALIZED DEBT OBLIGATIONS
The other type of basic credit derivative is a collateralized debt

obligation (CDO). A CDO is, at its core, the same as a CDS contract.19 As
in a CDS contract, the parties to a CDO contract are shifting the risk of an
underlying debt instrument from lender to investor, but instead of doing so
for a corporate loan or bond issuance from a single borrower they do so for
a series of loans or bonds from many borrowers. In this way, some
portfolio theory-based diversification is achieved, since the risk for any
investor of any one buyer defaulting is absorbed by gains on other debtors
that do not default.
Here is a cartoon of how the basic, plain-vanilla CDO is formed. A
CDO manager, usually an investor specializing in these products from a
large investment house like Goldman Sachs, forms a special purpose
vehicle (SPV), basically a stand-alone, bankruptcy-remote firm, and then
chooses loans or bonds or mortgages from many borrowers to put into the
SPV. The SPV then sells interests in the cash flows it will generate from
these debt contracts to numerous investors. The SPV generates cash from
the instruments it holds as the borrowers pay back the debts. This cash is
then distributed to the investors according to the terms of their investment.
So far, the SPV creating the CDO looks like any firm selling a service or
product. The SPV raises money from investors, uses this money to invest in
assets (in this case, debt instruments), manages the assets, and then
distributes the profits it earns to the investors.
SPVs investing in credit derivatives have two somewhat unique
features that enable them to be attractive risk-sharing mechanisms:
tranching and securitization. The concepts are quite simple.
In a normal debt investment, a group of investors share a vertical
slice of the expected payouts from a debtor. Three investors funding a $100
loan to a firm each bear exactly the same risk if the borrower defaults – as
the recovery on the loan falls from $100 to, say, $80, each investor suffers
19
In the nomenclature, if the underlying is a bond, the instrument is called a
CDO, while if it is a loan, it is called a CLO.
a 20 percent loss. In a securitized debt investment, by contrast, the same

three investors can slice the risk horizontally, instead of vertically, allowing
them to assign different payouts, called “tranches,” based on the amount of
recovery. For example, the investments could be structured such that the
first investor bears the first $10 in losses, the second the next $10, and the
third investor the remaining $70. In that case, if the $100 loan falls in value
to $80, the first two investors would suffer complete losses, while the third
investor would suffer none (although its risk would increase, since any
additional diminution in value would impair its position). This approach
can dramatically reduce the probability of default for a particular tranche,
and thus make even risky debt investment attractive for conservative
investors. For example, the third investor reduces the riskiness of its
investment by investing in the second type of vehicle. Insurance
companies, which traditionally invest in only relatively safe instruments,
used this approach to expand the types of investments they made.
The sharing and recategorization of risk can be enhanced by
pooling together many risks through a process called securitization. This
can be seen through a simple example. Consider two banks loaning to two
companies in different and uncorrelated industries. The loans both pay
$100 in the good state of the world and $0 in the bad state of the world,
with a probability of default of 10 percent. If the banks take vertical
positions, the expected value of the loans for each bank is $90. Each faces
the identical risks.
If instead, the cash flows from the two loans are pooled and
tranched in a CDO, the 10 percent risk of default can be reduced for one of
the investors. If one of the banks bears the first $100 in losses and the
second loses only if both borrowers default (and assuming the defaults are
not correlated), the risk of default for the senior bank falls from 10 percent
to 1 percent.20 The expected value for the two banks is thus $90 for the
junior bank and $99 for the senior bank. The process can be extended
indefinitely, with each additional risk added to the pool further reducing the
risk up the tranching scheme. For instance, adding a third investor and a
third uncorrelated loan to the pool reduces the risk for the most senior
bank, which suffers only if all three firms fail, to 0.1 percent.21
20
The loss for the senior bank is the probability of firm one defaulting (10
percent) times the probability of firm two defaulting (10 percent).
21
Correlation of risks is obviously the key assumption in the creation of a
CDO. If the risks in the three-bond case are perfectly correlated (that is, the failure
of one firm means all three firms will fail), then the probability of loss for all three
Even when considering the role of tranching and securitization, the

structure of a CDO is roughly the same as any firm in any business – they
are nothing new under the sun. Tranching is just a fancy way of saying that
the SPV replicates the priority of liquidation claims created by bankruptcy
law and contracts for other firms. When a firm liquidates, equity interests
lose first, and because of the absolute priority rule in bankruptcy, more
junior interests, like unsecured creditors, lose only after equity interests are
wiped out. This is true whether the firm is a donut maker or SPV holding
debt instruments in a CDO structure. In addition, as discussed below, the
shareholders investing in a traditional firm are selling insurance to the
firm’s debt holders, managers, and other stakeholders in the same way that
the protection sellers are for the original bank in a credit derivative
contract. Equity provides a downside cushion, since no payments are
legally due equity holders, and thus provides risk sharing on favorable
terms for holders of fixed claims on a firm’s balance sheet.
Unlike regular firms, however, SPVs holding credit derivatives
generally have only a single investment period. Whereas an insurance
company is constantly adding new policy holders, SPVs are typically
formed, buy debt instruments, raise money to fund the risk of these
instruments, and then make payouts according to the terms of the credit
derivative contracts. As mentioned above and discussed more fully below,
this distinction is a crucial factor in the appropriateness of insurance
regulation.
Before moving to the merits of the arguments for and against
regulating credit derivatives as insurance, it is important to point out a few
other features of credit derivative markets. First, the CDO market is at least
two times larger than the CDS market. At the height of these markets in
2007, the single-name CDS market (that is, an entity selling protection to a
bank for a loan to a single company) had a notional value of about $20
trillion, while the total credit derivatives market was about $60 trillion in
notional value.22 This means that the CDO market, which makes up the rest
of the market, was about $40 trillion, or twice as big as the single-name
CDS market.
Second, there are numerous index products and more complicated
CDO products (such as the CDO2) that allow individual investors to buy
investors is the same 10 percent. Thus, there are no credit reduction benefits from
securitization.
22
See Posting of Tyler Durden to, Some More Facts about How CDS Market,
http://zerohedge.blogspot.com/2009/02/ (Feb. 1, 2009, 10:31 EST).
exposure to a portfolio of credit derivative investments. For instance, in the

single-name CDS world, there are several indexes, like the Dow Jones
“Investment Grade CDX” and the “High Yield CDX,” that consist of over
100 borrowing firms of different credit quality. Investors can buy securities
that track these indices in the same way they can invest in the S&P 500 or
Wilshire 5000 equity indices. (As discussed below, investments in these
indices are no more providing insurance to the underlying participants in
the borrowing transaction than a regular firm selling equity, since both
provide mechanisms of risk sharing.) Moreover, firms or investors seeking
exposure to these credit default risks often hold a portfolio of risks or an
index product for a few months or less, rolling the investments on a fairly
constant basis to meet the investors’ or firms’ balance sheet needs.23 As
such, the investors in credit derivative indexes are not generally exposed to
the possibility of actually having to pay for any losses on the original debt,
but rather are susceptible to the change in the price of the indexed securities
depending on the changing nature of the credit quality of the underlying
borrower.
A CDO2 (and more exotic credit derivative products) basically
achieves the same result for portfolio products. A CDO2 is simply a two-
pool portfolio of tranched and securitized loans in which investors face
exposure to both pools. If the credit risks are not perfectly correlated, this
structure allows investors to lower their overall risk to something less than
they would have from investing in both pools separately. These products
also allow investors to invest in funds of CDO products, in which the
exposure becomes more and more attenuated from any individual
underlying borrower and starts to look more like generic risk exposure to
the debt markets or even the market in general. As the case of CDS
indexes, the investors in these products can now be thought of as merely
identifying a unique risk-return investment as opposed to making bets
about the credit quality of individual borrowers or pools of borrowers.
III. THE ARGUMENT FOR REGULATING CREDIT

DERIVATIVES AS INSURANCE
At first blush, the similarity between property (or other) insurance

and credit derivative contracts makes the call for insurance regulation of
the latter seem reasonable. But a better case than that must be made,
23
Interview with executive at insurance company responsible for credit
derivative transactions, on Mar. 21, 2009.
because deeming something “insurance” brings with it a special set of laws

and rules administered by state insurance departments. This would mean
significant increases in the cost of using credit derivatives, and, on the
margin, less use of them. This might be fine, but we need to believe that
insurance regulation brings us something worth the price. The benefits of
the regulation, therefore, must be well calibrated to the particular risks
involved, lest otherwise socially beneficial transactions be deterred. This
Part fleshes out the analogy between insurance and credit derivatives, while
the next section shows how this analysis is highly misleading by looking
behind the analogy to the purposes and justifications for calling something
“insurance.”
The standard definition of “insurance” is an “agreement in which
one party (the insurer), in exchange for consideration provide by the other
party (the insured), assumes the other party’s risk and distributes it across a
group of similarly situated persons, each of whose risk has been assumed in
a similar transaction.”24 There are two parts of this definition – (1) risk
transfer; and (2) risk pooling. The insurer assumes not only the risk of loss,
but distributes the risk across many other similarly situated individuals or
entities, so as to reduce unpredictable events into a more predictable cash
flow stream. In technical jargon, insurance companies try to pool risk by
attracting a sufficiently large number of diverse policyholders such that the
law of large numbers will reduce the aggregate variance of claims.
The credit derivative contracts discussed above have many
characteristics that seem to fit well within the scope of at least the first part
of this definition. As in a typical insurance contract, a CDS contract
involves a party with an asset (the loan)25 with a risk of loss (default by the
borrower), paying a reoccurring premium to a counterparty, which in turn
agrees to make the first party whole in the event there is a loss. To
analogize, just as a homeowner that pays a monthly premium to an
insurance carrier in return for a promise to make the homeowner whole in
the event of a loss related to the home, so too does the lender pay a monthly
24
ROBERT H. JERRY, II & DOUGLAS R. RICHMOND, UNDERSTANDING
INSURANCE LAW § 10(d) (4th ed. 2007). But see THE NEW OXFORD AMERICAN
DICTIONARY 881 (Elizabeth J. Jewell & Frank Abate eds., 2001) (defines
“insurance” is as “a practice or arrangement by which a company . . . provides a
guarantee of compensation for specified loss . . . in return for payment of a
premium.”). This definition misses a key component of insurance – the pooling of
risk.
25
The underlying credit instrument need not be a loan, but could be any debt
instrument, such as a mortgage, bond, note or any other form of indebtedness.
premium to a third party in return for a promise to make the bank whole in
the event of a loss related to the loan. If this analogy holds, the lender is the
“insured” and the counterparty is the “insurer.” And, the insurance law of
the fifty states would then regulate the insurer and the content of contracts
it enters into with insureds.
Building on the apparent similarity between typical insurance
contracts (regulated by state insurance agencies) and credit derivatives,
New York State recently proposed deeming credit derivatives “insurance.”
The chief state regulator, Eric Dinallo, offered the rationale during
testimony before a congressional committee: “With [plain-vanilla CDS
contracts], if the issuer of a bond defaults, then the owner of the bond has
suffered a loss and the [CDS] provides some recovery for that loss.”26
Dinallo limited the reach of the proposed regulations of credit derivatives,
however, by asserting the state’s jurisdiction covers only cases where the
credit derivative contract is between an original lender and a third-party
investor, that is, single-name CDS contracts in which an individual or
entity sells protection to an originating bank.27 These are so-called
“covered” transactions (as opposed to “naked” ones), since there is privity
between the insured and the underlying debt instrument. The reason for this
limited scope for insurance regulation is based on a generally accepted
argument that the party being insured has an “insurable interest” in the
underlying amount at risk under the contract. In other words, a contract is
“insurance” only if the insuring party pays when the insured party actually
suffers a harm unrelated to the insurance contract.
This concept can be illustrated by comparing the primary and
secondary credit derivative markets. Where a bank issues a loan and then
buys credit protection on that loan that pays off if the loan defaults, the
argument is that the buyer of credit protection has an insurable interest in
the loan, and that the protection acts as insurance against this loss. In the
secondary market, by contrast, two parties unrelated to the issuance of the
loan (and without the knowledge of the bank making the loan, the borrower
26
Hearing, supra note 5, at 3 (testimony of Eric Dinallo, Ins. Comm’r, N.Y.
State).
27
Id. (“We believe that the first type of swap, let’s call it the covered swap, is
insurance. The essence of an insurance contract is that the buyer has to have a
material interest in the asset or obligation that is the subject of the contract. That
means the buyer owns property or a security and can suffer a loss from damage to
or the loss of value of that property. With insurance, the buyer only has a claim
after actually suffering a loss.”).
taking out the loan, and any parties contracting with either of them), bet on
whether or not the borrower will repay the loan. According to the
conventional wisdom of insurance regulation, this bet is not insurance. The
reason for this concession is that otherwise “insurance” would include far
too much and things far beyond the ken, expertise, or conceivable reach of
state insurance regulators.
The concession is manifest in the findings of insurance regulators.
For instance, in 2000 after Congress exempted “swaps” and other
derivatives from certain regulation,28 the New York Insurance Department
was asked whether credit derivatives were in fact insurance, which would
be subject to state regulation. The question it was asked by federal
regulators was: “Does a credit default swap transaction, wherein the seller
will make payment to the buyer upon the happening of a negative credit
event and such payment is not dependent upon the buyer having suffered a
loss, constitute a contract of insurance under the insurance law?”29 This
question is aimed at the secondary market, and was answered in the
negative for reasons of a lack of privity with the loss on the part of the
entities engaging in the derivative transaction. In his testimony before the
House, Commissioner Dinallo distinguished this prior finding of the New
York State Insurance Commission that credit derivatives were not
insurance, by pointing out that the question asked was focused only on non-
privity cases or “naked” credit derivatives.30 From this, Dinallo concluded
that a different result could obtain in the privity case (that is, CDS
contracts), since the protection seller was insuring a real loss outside of the
context of the contract. The analogy described above was thus sufficient for
him to conclude that, with privity and a real potential loss, credit
derivatives of the plain-vanilla CDS variety are insurance products.
The argument is not preposterous on its face. Insurance is about
risk sharing, and in that sense credit derivatives, which are fundamentally
risk-sharing contracts, are akin to insurance. But, as shown below, the fact
that credit derivative contracts are providing an insuring or risk-hedging
28
See Commodity Futures Modernization Act of 2000 or [CFMA], H.R.
5660, 106th Cong. §§ 103, 105, 407 (2000); CFMA of 2000, S. 3283, 106th Cong.
§§ 103, 105, 407 (2000); see also Gramm-Leach-Bliley Act [GLBA] (Financial
Services Modernization Act of 1999), Pub. L. No. 106-102, 113 Stat. 1338, 1393-
94 (1999).
29
Hearing, supra note 5, at 4-5 (testimony of Eric Dinallo, Ins. Comm’r,
N.Y. State).
30
Id. (“So at the same time, in 2000, the New York Insurance Department
was asked a very carefully crafted question.”).
function does not mean that it is sensible or efficient to regulate them as

insurance. There are lots of contracts in which one party is effectively
offering insurance as part of the deal, and yet none of these are regulated by
state insurance commissioners. The reason is the underlying policy
justifications for having a separate body of insurance law do not reach
these contracts, and thus applying that law would raise the costs of the
contracts without any likely benefit. Another reason is that the second part
of the definition of insurance – risk pooling – is absent in many of these
transactions, as it is in most CDS contracts. In fact, if there is any risk
pooling by protection sellers (the alleged insurers), it occurs in CDO
contracts or secondary-market CDSs, exactly the place where New York
claims its regulatory reach does not extend. These arguments are made in
the next Part.
A few other features shared by insurance and credit derivatives
provide some support for the analogy to insurance. The first similarity
between insurance and credit derivatives is the incomplete nature of the
risk transfer. The insured (either the homeowner or the lender) swaps the
risk of loss with respect to the underlying asset (either the home or the
loan) for the risk that the insurer will not be able to make the insured
whole. This latter risk is called “counterparty risk,” and it is a central
justification for insurance regulation. An individual who takes out an
automobile insurance policy is swapping the risk of loss from an auto
accident for the risk of loss that the insurance company will not be around
to pay the claim. Capital adequacy rules, investment restrictions, and other
aspects of insurance regulation exist to decrease this counterparty risk.
Although CDSs and other credit derivatives share this similar feature, as
discussed below, this alone does not justify regulating them as insurance as
there are many other ways of reducing the counterparty risk problem that
do not involve the full panoply of insurance regulations.
The second similarity between insurance and credit derivatives is
the presence of moral hazard. Whenever risk is transferred, there is the
possibility of misbehavior on the part of the transferor or the transferee. If
the transferor (that is, the bank) has an obligation to prevent the loss from
occurring, say by monitoring the conduct of the borrower, the transfer of
risk reduces the incentive to do this on the margin. In addition to shirking,
protection buyers may act deliberately to force the debtor into bankruptcy,
say by withholding lending that would otherwise be efficient or by
invoking covenants outside the normal usage in the industry. These
examples of destroying value to simply collect on a CDS contract can
obtain in both the primary and secondary markets – nothing prevents the
holder of synthetic protection, say a hedge fund, from taking steps to harm
the borrower in order to collect on its bet. Although plausible, this facial
similarity does not justify treating credit derivatives as insurance. As
discussed below, this argument proves too much. Numerous opportunities
exist for similar moral hazard problems outside of the insurance context,
there already exist mechanisms (both market and from industry trade
groups) to ameliorate any moral hazard, and there is nothing about
insurance law that makes it a good fit for further reducing these potential
harms, if they are substantial.
An additional argument for regulating credit derivatives as
insurance is the absence of any existing regulation by other federal or state
agencies, especially of certain players in the market, like hedge funds and
other private pools of money. Many experts and pundits blame the lack of
regulation of the credit derivatives market as contributing to the credit
crisis. The argument goes like this: credit derivatives are not traded on an
exchange, but rather through individualized contracts, known as the over-
the-counter market, and the lack of regulation, either directly or indirectly
through regulation of the exchanges on which securities trade, allowed
private parties to externalize systemic risk costs onto society. The lack of
regulation thus generated an inefficient number or type of these
transactions from a social welfare standpoint.
There may be something to the premise of this argument, that is,
that the lack of regulation exacerbated the risk that private parties would
act in ways that would be privately optimal but increase the risk of a global
financial meltdown. The premise is debatable, but even if it is true, this
Essay shows that insurance regulation is not the only way in which these
systemic costs can be internalized by firms. Most obviously, direct
regulation of the credit derivatives market by existing federal departments
responsible for derivatives and markets, such as requiring derivatives to be
traded on an exchange, is possible under current law.
In fact, it seems from the public statements of New York officials
that the purpose of the characterization of credit derivatives as insurance is
intended to stoke federal regulators to act, more than a firm belief that
credit derivatives are insurance. After all, if they are insurance, then there
should in fact be no need for or call for federal regulation. In testimony
before Congress and other public comments, New York State’s insurance
officials “stopped short of endorsing comprehensive state-level regulation
of this privately negotiated market” and agreed to delay its plan to regulate
credit derivatives based on the indication that federal regulators are
“committed to comprehensive and effective federal oversight of credit
default swaps.”31 The fact that Commissioner Dinallo’s testimony outlined

a proposed regulatory agenda for federal agencies also supports the threat-
of-regulation-as-leverage claim.32
A final argument for state-based insurance regulation is the fact
that numerous insurance companies were involved in the credit derivatives
markets as buyers and sellers of protection, as well as acting as brokers and
speculators in secondary markets. According to one estimate, insurance
companies represented about 20 percent of end users of credit derivatives.33
For instance, insurance giant AIG, invested heavily in credit derivatives of
various kinds – its portfolio of CDSs reached $526 billion at its height.34
And it is widely viewed that the losses on these credit derivatives – over
$30 billion in 2007 and 2008 alone – were the cause of the failure of AIG
and the need for the massive government bailout.35 The logic of regulation
would thus be that these products were misused by insurance companies,
among others, and this justifies regulating them as insurance. President
Obama seemed to endorse this view when he described the situation as
follows: "You've got a company, AIG, which used to be just a regular old
insurance company. ... Then they decided--some smart person decided--
31
Id. at 7.
32
Dinallo testified that :
Effective regulation of credit default swaps should include the
following provisions: All sellers must maintain adequate capital
and post sufficient trading margins to minimize counterparty
risk; A guaranty fund should be created that ensures that a failure
of one seller will not create a cascade of failures in the market;
There must be clear and inclusive dispute resolution
mechanisms; To ensure transparency and permit monitoring,
comprehensive market data should be collected and made
available to regulatory authorities; The market must have
comprehensive regulatory oversight, and regulation cannot be
voluntary. Id.
33
U.S. GEN. ACCOUNTING OFFICE, supra note 14, at 6 n.8 (“The top five end-
users of credit derivatives are banks and broker-dealers (44 percent), hedge funds
(32 percent), insurers (17 percent), pension funds (4 percent), and mutual funds (3
percent).”)..
34
Sjostrom, supra note 2, at 40 (“A CDS certainly appears to fall within this
definition given that the protection seller contractually agrees to compensate the
protection buyer following the occurrence of a credit event.”).
35
Id. at 26.
let's put a hedge fund on top of the insurance company and let's sell these
derivative products to banks all around the world."36
As discussed below, this argument proves too much, as many other
entities, like banks, hedge funds, pension funds, and so on, used credit
derivatives too, often disastrously but just as often fantastically, and
therefore there is nothing special about the end users that justifies treating
them as insurance. If anything, the fact that some insurance companies
were harmed by them justifies different regulation on insurance companies.
IV. THE ARGUMENT AGAINST REGULATING CREDIT

DERIVATIVES AS INSURANCE
This Part presents several conceptual and practical reasons why

credit derivatives should not be regulated as insurance products or why
sellers of credit protection should not be regulated as insurance companies.
Notwithstanding the superficial appeal of the analogy between insurance
contracts and credit derivative contracts, the policy justifications for special
rules regulating insurance carriers and contracts do not obtain in the credit
derivative context. Examining the rationale for insurance law and the
important differences with credit derivatives will show this.
The two primary reasons for having a separate body of insurance
law are the particular governance problems associated with insurance
companies and the fact that insureds are typically unsophisticated
individuals for whom insurance is essential and may be difficult to obtain
in the event of certain individual characteristics. (These two justifications
correspond with the two major features of insurance law – the regulation of
insurance company investments and the regulation of sales to individuals.)
Neither of these reasons justifies applying insurance law to credit
derivatives.
Before addressing these reasons for insurance law, this Part
addresses the reasons why policy arguments are needed in the first place.
The primary reason insurance contracts are treated differently than other
contracts (and “insurance law” is a separate body of law) is not because of
their nature as “insurance” but rather because they are issued by insurance
companies. This conclusion is evident from numerous problems that would
36
See “President Barack Obama on ‘The Tonight Show with Jay Leno,” N.Y.
TIMES, Mar. 19, 2009, available at http://www.nytimes.com/2009/03/20/us
/politics/ 20obama.text.html.
arise in extending the reach of insurance regulation to all contracts

providing some kind of insurance.
A. PROVIDING INSURANCE IS NOT ENOUGH
1. Line-Drawing Problems
It could not be enough for a contract that insures against risk to be

regulated as insurance or to bring the seller of that insurance within the
ambit of (state-based) insurance regulation. Every contract has a degree of
insurance embedded in it, and options and derivatives of all sorts, which
are not considered insurance or regulated as such, are mostly about insuring
or hedging against risk. If insurance law covered all contracts that are
partially or completely about insurance, the line drawing problems about
what is insurance would likely broaden the scope of insurance law to cover
vastly more than it currently does. The result would be to add regulatory
costs and uncertainty to a vast swath of the economy, with little or no
expected benefit.
Consider a simple options contract, known as a forward contract:
Farmer agrees to sell wheat in six months at a given price (a put option)
and Baker agrees to buy wheat in six months at a given price (a call
option). Both Farmer and Baker are purchasing price insurance from each
other – insurance against a price rise (for Baker) and against a price drop
(for Farmer) – by locking in a set price in advance. The contract is not
regulated as insurance, and neither Farmer nor Baker are currently
regulated as insurance companies, even though each may be providing
insurance to the other. There are several reasons for this result. Both parties
are presumably somewhat sophisticated, since they went in search of
derivative hedging tools, or are intermediated by market professionals.
These gatekeepers compete in competitive markets and are regulated by
other laws and exchange-based rules to ensure fidelity to their clients’
interests and a suitability between client needs and products sold to them.
In addition, the derivative contracts are likely made either on a competitive
derivatives exchange or as the result of arms’-length negotiation.37
37
Nor is the result different if a third party sells the insurance to Farmer or
Baker. For instance, an individual unconnected with the farming or baking
business may believe that wheat prices will rise/fall in six months based on
predictions about weather, changes in supply or demand, or other factors. This
individual can enter into a forward contract with Farmer or Baker either directly or
There are innumerable contracts that provide the same type of

insurance as a forward contract does and that are not regulated as
insurance. Any hedging contract has an element of insurance. For instance,
an investor who is long a particular security, commodity, or investment
might want to reduce the risk of the position by entering into another
investment or contract with a third party that moves the other way. The
third party might be thought of as providing some insurance for the
investor, but this is not how the contract is thought of or regulated. As
discussed below, contracts like these have many purposes, and trying to
sort contracts along this dimension is likely to be costly and highly
imperfect, especially if done in ex post litigation, where litigant
opportunism and hindsight bias will be problematic, or by regulators, who
will face inevitable public choice problems in their definitional exercises.
One can imagine trying to sort between these two categories of
contracts by inquiring into the minds of the contracting parties to see
whether the contract was about insurance or something else. As noted
above, this would require the mind of the parties to every contract to be
examined to determine whether they are providing or seeking “insurance.”
Regulators would need to know whether the investor was entering into the
contract for insurance or hedging purposes. This is not generally the
inquiry regulators make, perhaps because the question of knowledge is
malleable and costly to enforce, especially given imperfect courts and a
costly litigation system. Another dividing line could be the intent of the
investor, but this too is an unhelpful and costly line to draw. It may be
significantly over-inclusive, and it is susceptible to similar proof problems
as knowledge. There may be mixed motives for all investments – return,
hedging, speculation, and so on – that will be difficult to unpack accurately
and without being subject to ex post bias, power grabs by regulators, and
rent seeking by stakeholders of the firms in question.
Rather, some contracts tend to be primarily about insurance, while
others have multiple functions, some of which might be about risk sharing.
The former might fall within the ambit of insurance regulation, while the
latter never do. But where is this line? Consider, for instance, equity
investments in run-of-the-mill firms. As discussed above, any equity
investment in any firm could be thought of as insurance in the same way
through an options exchange, such as the Chicago Mercantile Exchange. These

exchanges are populated with relatively sophisticated parties and are covered by
alternative regulatory regimes, including licensing requirements for brokers and
dealers.
that credit derivatives are. After all, when a firm issues equity interests to
investors, it is entering into a risk-sharing contract (on behalf of managers,
creditors, and other stakeholders) with these investors. Equity holders,
unlike debt holders, have no fixed claim on a firm’s assets, and therefore
provide a source of funding that is less sensitive to downturns in
performance than debt. A firm that has more equity on its balance sheet is,
all else being equal, less risky than a firm that has less. So we could re-
characterize a firm’s decision to issue equity (to lower its debt to equity
ratio) as buying insurance (against a downturn in the firm’s affairs) and the
investors buying the equity as selling insurance to the firm. Of course, no
one thinks of equity in this way. But equity is as much about insurance as
credit derivatives are.
One reason securities are not regulated as insurance is the fact that
equity investments are regulated by a separate body of law – securities law
– specifically designed to address the policy challenges of issuing and
investing in securities. When Congress passed the securities laws in the
1930s, it could have simply called equity investments insurance and
delegated regulation to state insurance law under the same theories as those
calling for this treatment of credit derivatives. But this would have been a
reach – although arguably insurance, equity securities are sufficiently
different along numerous dimensions to justify a separate body of
regulatory law.
Another reason equity might not be regulated as insurance is
because of the particular characteristics of the contracts in question. As
noted above, typical credit derivative contracts look like typical insurance
contracts: one party makes periodic payments to another in return for a
make-whole promise in the event of a future occurrence. This similarity is
only a surficial one, however, since there are many other aspects of credit
derivative contracts that are quite different. For instance, payments may not
turn on actual losses, there may be no pooling of risk, the make-whole
promise may be purely synthetic, and so on. In addition, it is hard to
imagine regulatory treatment turning solely on the question of whether
risk-sharing payments are made on a periodic basis (as in insurance
contracts) or a lump sum basis (as in equity investments, forward contracts,
and so on). This would elevate form over substance in an arbitrary way not
anticipated by the parties, and would provide an easy roadmap to avoiding
any regulation.
Finally, few if any investors making an equity investment think
they are providing insurance. Rather, the investment is made for a whole
host of reasons, including pure investment, speculation, hedging, and so
forth. Credit derivatives are used for these multiple reasons too, and this
cuts strongly against trying to narrowly pigeonhole them as insurance.
2. Credit Derivatives Are Not Just About Reducing Risk
Another problem with regulating credit derivatives as insurance is

that they are not just (or, even, primarily) used for “insurance” purposes.
As noted above, a common use, but only about one-fifth of the current
market, is the buying and selling of credit protection on loans, bonds, and
other sources of indebtedness. It is doubtful that this use fits squarely
within the regulatory definition of insurance or that insurance regulation
would be beneficial to these markets, but importantly for definitional
purposes swapping credit risk is only one of the many uses of these
financial products.
Credit derivatives can be used to hedge risks independent of and
unrelated to the original loan or bond being used as a reference entity. For
example, a hedge fund that wants to reduce its exposure on, say, Russian
wheat futures, may find a corporate bond risk whose risk offsets its
commodity position favorably, and thus enter into a contract with a third
party, who might be hedging Texas oil prices, who is willing to pay in the
event the bond defaults. This transaction has nothing to do with the
underlying bonds, since it only uses them as a reference for calculating a
stream of payments. The transaction is akin to two individuals in China
betting on whether I will crash my car. Neither of them is insuring me, but
rather they are simply using the probability of me crashing my car (and the
amount of damage that will result) as a reference for assigning risk among
them. (The original debtor in credit derivative contracts is called a
“reference entity,” a description that well captures this concept.) These bets
are not considered insurance, because there is no privity with party
suffering a loss (that is, the lenders in the case of a default by the borrower
on the bond or me in the case I crash my car) and furthermore no proof
required that an actual loss be suffered. Even if these bets were considered
insurance, it would be impossible to regulate all of them in this way.
Detecting them would be difficult and costly, and, even if possible, would
simply direct the parties into transactions of similar risk-return
combinations but other designs. In other words, if regulators deem one
class of credit derivatives “insurance,” and thereby impose increased
regulatory costs on that class, and deem another class of credit derivative
contracts as “not insurance,” parties will naturally structure their
transactions as the not-insurance kind. More on this later.
Credit derivatives are also used for speculation about credit risk.
Investors can use credit derivative contracts to speculate on the default
probability of a particular borrower. In general, there are no insurance
contracts like this. Participants in insurance markets do not wager on
whether an individual’s probability of dying or crashing a car is rising or
falling on a daily basis, as they do with firm risk in credit derivative
markets. Insurance is based on probabilities at the macro level and across
large numbers based on actuarial science, and, when it is based on
individuals characteristics, it is done only at the point of origination and not
for speculation purposes. But this is exactly what is done with credit
derivative contracts. For example, an investor who believes that General
Electric’s credit quality is likely to worsen over the next few years can buy
protection against default by GE on its debt. If the credit quality does
deteriorate, the cost of protection will rise, and the investor will earn a
profit. Similarly, if an investor believes that GE’s credit quality is likely to
improve, it can sell protection against default by GE. If the credit quality
does improve, the cost of protection will fall, and the investor will earn a
profit. Before credit derivatives this kind of speculation was extremely
difficult, as it is practically impossible to short bonds or loans. The credit
derivative market thus allows for information about debt quality to be
processed in a market, perhaps with large gains to capital allocation
efficiency. Like the hedging examples above, these transactions are not
insurance in any meaningful sense. Nothing about the speculation contract
requires that it be held for any period of time. An investor can buy or sell
protection and hold it for an hour, a day, a year, or five years (the typical
maximum length), depending on the profit that can be made from buying or
selling at a particular time. The contract does nothing more than offer an
opportunity to buy or sell later at a higher or lower price. In this way, credit
derivatives can be, and are largely, about investment, not insurance.38 In
fact, they resemble secondary market transactions in equities, since they
involve market-based trades about the fundamental value of a third party
unrelated to the transaction in question.
A final (non-insurance) use of credit derivatives is arbitrage, of
either the pure or regulatory variety. Pure arbitrage possibilities arise when
there is temporary mispricing in markets that allows investors to engage in
38
Life insurance could be used for investment in a way, but this is not its
primary purpose or the way it is typically used. Moreover, state insurance
regulators are not really concerned about regulating investment decisions by
sophisticated parties.
paired transactions that earn sure profits, and thus eliminate the mispricing.
Mispricing in debt securities might arise because the primary and
secondary markets are traded on different markets: a firm’s bonds might be
trading on a bond market, while protection on those bonds might be trading
on over-the-counter markets that are not perfectly correlated with the
public bond markets. This difference might generate opportunities for an
investor to buy/sell the underlying security in one market, while
simultaneously buying/selling protection on it in the over-the-counter
market using a credit derivative in a way that allows it to earn a return that
is independent of the credit risk of the borrower. This kind of arbitrage
opportunity has no insurance-like characteristics, and it is potentially quite
useful in accurately pricing credit risk by removing temporary market
inefficiencies. Participants in credit derivative markets, at hedge funds,
insurance companies, and other large financial entities, describe this as a
major driver of their transactions in credit derivative markets.39
The other type of arbitrage – regulatory arbitrage – is related to
insurance, since it is one of the main reasons insurance companies are
involved in the credit derivatives markets. But the lesson here runs counter
to any regulatory story – in fact, it is a product of regulation itself.
Regulatory arbitrage works like this. Banks are often the most
efficient originators of loans, since they have relationships with lenders and
the back office to underwrite and process loans. However, they are not
always the most efficient holders of loans because of regulations that make
holding risk more costly for them and regulations that make investing in
credit risk difficult for insurance companies and other risk-sensitive
investors. In other words, smaller banks, individual investors, insurance
companies, pension funds, university endowments, foreign governments,
and a whole host of other investors would like to participate in corporate
debt markets, but cannot do so in the absence of financial instruments that
allow large commercial banks to sell the risk, especially in ways, like
securitization discussed above, that recharacterize the risk in ways that
make individual investments in it appear less risky. If insurance companies,
pension funds, or endowments can only invest in corporate debt rated
AAA, banks, who are required to hold cash reserves on corporate debt
rated below AAA will find a way of repackaging the debt so that some of
the sub-AAA debt can become AAA debt – this is the securitization and
tranching process described above.
39
Interview with executive at insurance company, supra note 23.
The first set of regulations – capital adequacy rules in the Basle

banking accords – require banks to hold a certain amount of cash in reserve
for each dollar lent out. In short, banks have to hold cash, while other
investors do not, which makes them inefficient holders of risk. The reason
for this rule is because banks have average individuals as creditors through
deposit taking, and given the moral hazard problems created by federal
deposit insurance, banks would otherwise engage in socially inefficient risk
taking. This cash, historically about 8 percent of the total value at risk, is
not productive from the standpoint of the bank’s investors, so it would be
more efficient for them to loan out the money, earn the fees on the
origination, and then offload the risk, in whole or part, to other investors so
that more of the bank’s cash can be put to use for its shareholders.
The reason insurance companies were involved heavily in these
markets (primarily as protection sellers to banks that had originated loans),
is because state law insurance regulations limit the kinds of investments
that insurance companies can make. For instance, insurance companies are
often restricted to investing in credits rated AAA by credit rating agencies.
These credit rating agencies were in turn paid by the managers of credit
derivative SPVs to rate the risks of investments in those SPVs, often to get
a slice of them to be rated AAA to attract the monies held by insurance
companies. As such, insurance companies became one of the largest
investors in credit derivatives. For example, AIG (through its financial
products business) invested nearly $400 billion in providing various
European financial institutions with “regulatory capital relief” through
credit derivatives.
Credit derivatives help complete these markets by allowing the
bank to offload the risk to investors who can more efficiently bear it, while
still having the ability to earn fees from origination. A bank that makes a
loan with a customer can now package the credit risk of that loan in a new
entity, which then uses securitization to create risk slices that will be
attractive to new classes of investors, and then sells off interests in the new
entity. To be sure, the original bank could be thought of as buying
insurance, since it is offloading to or sharing risk with others. But that
description of the activity is a cartoon representation of the transaction. The
bank is really engaging in regulatory arbitrage, but this redefinition is only
superficial. The important point is that looking at what the bank is doing is
only part of the story about whether regulation makes sense.
In addition, as discussed above, those buying interests in the bank’s
credit risk are no different than investors in any firm. A CDO is just a
business plan in which the proceeds from hundreds of credit risks from
various lenders are pooled together to generate a series of cash flows. The
firm (or SPV) holding the interests in these future cash flows is not
conceptually different from a firm that sells anything else, be it iPods,
consulting services, or what have you, and then sells interests in the cash
flows these sales generate. When a firm raises money from shareholders, it
is buying insurance in the same way that a bank transfers some risk through
a credit derivative contract. Although equity holders are not liable to make
the seller of the risk whole in the event of some specific default, the equity
investors are providing the firm with an opportunity to reduce its risk.
Equity, like insurance, provides a cushion against a downturn. Of course,
no one thinks of regulating securities as insurance despite the similarity
along this dimension.
There are at least two important differences between these two
types of risk-sharing mechanisms. First, on average the sellers of equity to
regular firms are much less sophisticated than the sellers of risk protection
to lenders. This, of course, cuts the other way from regulating credit
derivatives as insurance. Second, the structure of the standard insurance
contract (and the typical credit derivative contract) is different than the
shareholder contract. Whereas in a credit derivative or insurance contract
the party assuming the risk receives periodic payments in return for a
promise to make the party selling the risk whole, in the shareholder
contract, the sequence of payments is reversed: the party assuming the risk
of default pays the money up front, while agreeing to receive future cash
flows in the form of dividends, capital appreciation, or liquidation value at
some time in the future. This alternative structure has important
implications, which are discussed below, but it does not necessarily
undermine the attempted analogy to insurance. After all, if insurance is
defined as a contract in which risk is moved from one party to the other, the
structure and terms of the contract are, all else being equal, irrelevant to
whether risk is in fact being swapped.
The lesson to be learned from this use, which is also only
superficially similar to classic insurance, is that any regulation of insurance
company participation in credit derivative markets should focus on how
insurance companies invest in credit derivatives. This is especially true
since insurance companies are only a small fraction of the entire market in
credit derivatives.
3. The Pooling Mismatch
Another reason insurance regulation is a bad fit for credit

derivatives is that there is a conceptual difference in the function of
insurance and that of credit derivatives. The premise of insurance is risk
pooling. Insurance companies try to spread or pool risk by attracting a

sufficiently large number of diverse policyholders such that the law of large
numbers will reduce the aggregate variance of claims. In this way, the total
amount of risk can be shared by many and thus reducing its impact on any
individual in the pool. Counterparties to derivative contracts do not usually
do this.
In a single-name CDS contract, there are only two parties, so there
is no pooling of risk. When a hedge fund sells protection to bank, it does
not act like an insurance company that sells protection to an individual
property owner. While the insurance company puts together a diversified
portfolio of property owners to generate an actuarially predictable stream
of liabilities, the hedge fund does not do this. Hedge funds may try to offset
the risk of a particular CDS with other assets and liabilities in their
portfolio, but they do not pool risk by writing protection on hundreds or
thousands of firms based on predictions about default risk and correlation
of risks. Or, to be more precise, they do not always and necessarily do this.
These counterparties may be hedging risks and trying to reduce their
overall risk exposure, but they are not doing so by pooling a lot of
independent risks. Thus, the insurance component of the transaction looks
more like simple hedging, which is not regulated as insurance.
The lack of pooling is a conceptual difference, but it may have a
practical consequence. Insurance regulation requires insurance companies
to hold significant capital reserves in part because if one insurance
company fails, a significant amount of beneficiaries will lose. (Importantly,
many of these beneficiaries will be average and unsophisticated citizens
who are unable to bear the losses. This is the consumer protection angle of
insurance regulation discussed below.) The same problem does not exist for
credit derivatives generally, unless a single entity, like AIG or
Countrywide, makes a multitude of credit derivative bets (that do not
cancel or net out the risk of the sum) and the bets made are so large that it
threatens the entity and its policy holders or depositors. Note, however, that
in the rare cases in which this did or is likely to happen, the independent
regulation of the insurance company or bank making the bets exists to
ensure that the risks taken by the entity are not excessive. In other words, if
the problem is that an insurance company, like AIG, took on excessive
risks in credit derivative contracts, then the rules about what investments
insurance companies can make should be reformed.
In more complicated credit derivative transactions, such as CDOs,
there are multiple parties, and arguably more risk pooling. As discussed
above, in a CDO, a new firm (an SPV) is created to sell protection to
multiple lending banks, and numerous investors own shares in the newly
created SPV. In this case, the analogy to insurance pooling is more apt.
One could view the multiple lending banks whose notes are pooled together
in the SPV as the policyholders, while the SPV and its investors are the
“insurance company.” While this analogy has more surficial appeal than
the case of plain-vanilla CDS’, as discussed below, the policy arguments
for insurance-like regulation do not obtain. So even in the case where there
is risk pooling – a necessary conceptual component of insurance – there is
no policy justification for insurance regulation. This is discussed below.
Even insurance commissioners admit that CDOs are not insurance
for this reason.40 There is a deep irony here. There is generally no risk
pooling – an essential component of “insurance” – in CDS contracts, but
these are the contracts that state insurance regulators and pundits consider
insurance. In contrast, there is at least some risk pooling in more complex
CDO contracts, but there is often no insurable interest in these transactions,
so insurance regulators disavow any regulatory oversight of them. The
reason for the line drawn by insurance regulators has to do with experience
and thus expertise. Insurance regulators are used to dealing with entities
that pool risk, are responsible for ensuring an adequate income stream to
pay for future liabilities, and are contracting with every-day consumers
who rely on the insurance company to make them whole in the event of
large personal losses. This experience is obviously not transferable to a
market in which none of these traditional aspects of insurance exists, nor
are the key regulatory questions. This is explored in greater detail below.
4. Limits on the Reach of Regulation
The artificial distinction drawn by regulators between plain-vanilla

CDS’ and more complicated credit derivative contracts points out a bigger
problem with any attempt to regulate credit derivatives using an insurance
framework. If (insurance) regulation is limited to cases where there is an
insurable interest, the contract is not one of simple hedging, arbitrage, or
speculation, and there is risk pooling, then this class of cases is like an
empty set. If the set of regulated cases is limited, as regulators assert, to
cases in which there is privity, there is no risk pooling. Conversely, if the
set of cases is limited to where there is risk pooling, there is no privity, and
thus the significant line-drawing problems discussed above arise.
More importantly, from a welfare and efficiency perspective, any
regulation of one part of the market that does not cover the entire market
40
See infra note 47 and accompanying text.
will simply redirect market activity to the unregulated market. Regulation

means additional costs, and investors will try to avoid these costs if they
can do so while achieving the same returns. The fact that credit derivative
contracts are simply ways of creating a specific risk-return profile means
that the same risk-return profile can be achieved in numerous ways that fall
outside of any product (as opposed to institution-based) regulation. For
example, if new insurance regulations cover only plain-vanilla CDS
contracts, as proposed, investors can replicate the returns they would have
achieved with a single-name, real-interest CDS by using a synthetic credit
derivative contract that is by the regulators admission, unregulated. In a
synthetic CDS contract, the parties do not actually interact with the
borrower or lender and do not use the underlying debt instrument as
anything more than a probability machine that determines future payoffs
between the parties. There is simply no difference between a real and a
synthetic CDS contract from the standpoint of these investors, and
increased cost on the former will simply mean more of the latter. This fact
poses a significant problem for regulators, since there are literally an
infinite number of potential contracts and contract forms that can be used
by investors to share and transfer credit risk.
Once one form of credit derivative is regulated, other forms will
sprout up that will match exactly the same risk-return profile, but these new
forms will be unregulated for one reason or another. As discussed below, a
more sensible regulatory approach is to: (1) identify investors who are
likely to make bad investment decisions on average for one reason or
another, (2) ban them from particular forms of investment, (3) and require
them to receive special disclosures or protections, or other paternalistic
regulation. Lack of sophistication, for instance, provides a central
justification for securities regulation, while market failures that may arise
out of governance concerns provides the justification for insurance
regulation. Both of these, however, are focused largely on the impact on
particular investors, as opposed to the nature of the products being sold.
There is a case where synthetic derivative contracts may be used to
reduce real risk, and therefore are more like insurance. If the original
lending bank enters into a synthetic contract to hedge its risk, it is the same
as if the bank enters into a standard credit derivative contract with a
protection seller, since the bank is reducing its risk of loss on the default of
the original debt. The bank in both cases is seeking regulatory relief from
its capital adequacy requirements. But although this type of contract
resembles a case where there is an insurable interest more closely, the party
on the other side, (the one betting that the borrower will repay the loan)
may not know that it is providing insurance of a sort; and even if it did, it
would still make more sense to regulate this contract as a matter of banking
law. After all, the bank is engaging in regulatory arbitrage, which may be
normatively good or bad from the perspective of banking policy. Banks
hedging risk reduces the probability that they will default, and thus
jeopardize the claims of depositors. On the other hand, it is possible that
some banks used credit derivative contracts not to hedge risks but to
increase profits by repackaging loans, moving them off of the bank’s
balance sheet, receiving regulatory relief, and then bringing the risk back
onto the bank’s balance sheet through mechanisms that were not
transparent to regulators. Again, these issues are largely about banking law
– that is, capital adequacy requirements, rules about relief from these
requirements, banking oversight, compensation of bank executives, and so
on.
5. Moral Hazard Problems and Solutions
One argument in favor of insurance regulation for credit

derivatives is based on the fact that both insurance and credit derivative
contracts are subject to moral hazard concerns. But the mere existence of
moral hazard problems does not justify insurance regulation per se. Moral
hazard arises in many contracts and situations that are not deemed
insurance. In addition, there are alternative ways of reducing moral hazard
short of full insurance regulation. For instance, regulation by other
administrative agencies or resolving issues between parties by acting
collectively through trade associations. As it turns out, contractual
innovation and self-regulatory norms are already being deployed by the
International Swap Dealers Association (ISDA) to remedy some of the
moral hazard problems inherent in credit derivative markets. As discussed
below, there remain some market failures, but none of them are especially
redressable by insurance regulation alone.
Moral hazard problems arise whenever any risk is intermediated.
Just as one is less likely to take care while driving if one has good
insurance (especially with a low deductible), so too is a bank less likely to
commit to an efficient level of due diligence or otherwise monitor a
borrower if it is going to sell the risk to someone else (and not retain a first-
loss position). In equilibrium, investors in the borrowers’ credit risk have
an incentive to price this potential shirking, and therefore the arrangers of
the SPV would have an incentive to choose the credit risks to put in it
wisely, lest they be required to offer greater returns to investors. In a
frictionless world, in other words, the amount of due diligence would be
priced by the market. The credit crunch revealed significant mispricing in
credit markets, but nothing that calls this equilibrium solution into doubt.
With learning from the recent collapse, it is likely that this discipline will
return to the market.
Even if it does not, there is not much that insurance regulation is
likely to add to solve the problem. No regulators or private actors were
aware of the mispricing problem, despite the fact that there were numerous
regulators, including insurance regulators (AIG is an insurance company
after all!) monitoring these markets closely, and despite the fact that
investors were betting billions of dollars of their own money on these
instruments. To simply declare that more regulation, and in particular more
insurance regulation, is needed, is to simply declare the debate over.
Another type of moral hazard is the potential that the parties to
credit derivative contracts might act in ways that destroy social value but
increase the private value to the party. For instance, a buyer of protection,
like a bank, might have incentives to force a borrower to default on a debt
in order to collect on the credit derivative contract; even if it is not
otherwise efficient for it to do so. This problem arises only because there
may be technical defaults that would otherwise not lead to bankruptcy, but
the bank buying protection could insist upon enforcing covenants against
them now that its downside is limited by its purchase of protection. (The
analogy to insurance here is that the buyer of insurance might willingly
destroy an otherwise valuable asset to collect on an insurance premium in
cases in which the asset has value but this value is less than the value of the
insurance policy.) While this is possible, there are at least three things that
limit its practical effect.
The first of these is the fact that private contracts take this problem
into account, without the need for regulatory mandate. This is not to say
that there are not market failures, but simply that this particular problem is
not unknown or unremedied in credit derivative markets. As in insurance,
where the problem exists too, buyers of protection voluntarily reduce the
risk they will shirk because of the moral hazard problem by agreeing to
bear some of the first losses that may arise from a default by the original
borrower. In insurance, this is called a deductible, and the theory is that it
reduces on the margin the incentive of the insured to engage in socially
destructive behavior. Credit derivative contracts try to reduce this conduct
too – the buyers of credit derivative protection routinely hold the first-loss
position so as to signal to sellers of protection that the bank buying
protection has some skin the game and will not engage in this kind of
destructive behavior. A deductible, being less than 100 percent of the risk,
however, can never fully offset this risk, so there remains some moral
hazard problem. Part of this can be priced by the market, but industry
norms can try to reduce it too. This is the next solution.
The second way moral hazard is reduced is through the fact that
self-regulatory bodies, like the ISDA, are already aware of this possibility
and are structuring industry norms and boilerplate contracts to mitigate
these potential problems. The moral hazard issue arose for the first time in
the case of credit derivative contracts written on the financial services and
insurance firm Conseco. Credit derivative contracts at that time required a
payment from the seller to the buyer of protection in the event that the
underlying reference entity—in this case, Conseco—suffered a “credit
event,” which included a restructuring of the reference entity’s bank loans.
In 2000, Conseco’s credit quality deteriorated and began to suffer liquidity
problems, so it went to its borrowers in search of a restructuring agreement.
The lenders agreed, including an extension of maturity, increased interest
rates, and new covenants. The restructuring triggered payment under the
existing credit derivative contracts.
This fact created a serious moral hazard problem. The original
lenders to Conseco, who had purchased protection against a credit event,
were the ones who got to decide whether to restructure Conseco’s debt, and
thus whether a restructuring event transpired.41 The lenders could trigger
payment simply by agreeing to extend the maturity of the loan or make
other trivial changes to the loan that would cost them little (and would be
readily agreed to by the borrower) and yet trigger potentially large
payments from the sellers of protection. In fact, the situation under the
then-prevailing boilerplate terms was much worse than that. Under the
ISDA’s 1999 version of the boilerplate terms (called the “Definitions”), the
buyers of protection could deliver any debt instrument of the same kind as
that on which the lender or other party bought protection. Since Conseco
had a number of outstanding debt instruments of varying maturity, the bank
triggering a restructuring credit event could choose the cheapest of these
outstanding debt instruments, thereby making large profits on its self-
triggering claim. Specifically, Conseco had short-term bonds that were
trading at about 90 percent of face value, while its long-term bonds were
trading at about 60 percent of face value. This meant the original lender
could declare default, and then buy long-term bonds at 60 cents to settle out
its much more expensive short-term bonds. This exposed the sellers of
41
SATYAJIT DAS, CREDIT DERIVATIVES: CDOS AND STRUCTURED CREDIT
PRODUCTS 101-103 (3d ed. 2005).
protection to large losses. The same result obtained in the restructuring of

Xerox Corporation in 2002.42
There were two reactions to the moral hazard issues raised by the
Conseco and Xerox cases, both private. The market price of credit
derivative contracts quickly adjusted to account for the increased risk of
contracts including restructuring events as credit events. The contracts
“including restructuring as a credit event was 10 to 20 [basis points per
year] higher than for credit default swaps without the restructuring credit
event.”43
The other reaction was a modification of the boilerplate credit
default swap documentation by ISDA. One possibility was that ISDA could
simply eliminate restructuring as a credit event, but this was foreclosed by
a decision of the Federal Reserve that protection from the risk of
restructuring was essential to the transfer of credit risk essential to
receiving regulatory relief under the Basle accords.44 ISDA therefore issued
a “Restructuring Supplement” that provided restructuring would not be a
credit event in cases where there were fewer than four holders of the debt
in question or where less than a super-majority of unaffiliated holders
approved the restructuring.45 In addition, the amendments to the boilerplate
tried to reduce the potential arbitrage inherent in delivery options for debts
with different maturities by requiring any settlement of credit derivative
contracts to be made with debt contracts within 30 months of the
restructured facility. Further changes to the boilerplate were made in 2003
to address market developments.46 The idea with these changes was to
reduce the moral hazard problems by contract.
The third non-regulatory way moral hazard is reduced is the
presence of countervailing interests on the other side of the transaction that
generate behaviors that may cancel out any possibility of abuse. Just as the
buyer of protection has incentives to act in a socially inefficient way by
42
See Viktor Hjort, The Xerox Debt Restructuring – A Moral Hazard Issue?,
Morgan Stanley, Fixed Income Research (July 19, 2002).
43
DAS, supra note 41, at 103.
44
See William Rhode, Fed Says No To Credit Restructuring, DERIVATIVES
STRATEGY, Dec. 2000, http://www.derivativesstrategy.com/magazine
/archive/2001/1200shrt.asp.
45
See Restructuring Supplement to the 1999 ISDA Credit Derivatives
Definitions, ISDA (2001); see also Donald A. Bendernagel & Oussama Nasr,
Legal Documentation and the Restructuring Debate, CREDIT DERIVATIVES ISSUES
AND OPPORTUNITIES, 2001, at 21.
46
DAS, supra note 41, at 105.
destroying value to cash in on a credit-derivative contract, the seller of

protection has incentives to act in exactly the opposite way. More
specifically, if the original lender that bought protection refuses to make a
loan that would increase the value of the debtor in order to cash in on a
credit protection contract, the hedge fund that sold protection has the
incentive to make the loan so not to have pay on the contract. These
offsetting incentives only work under two conditions: both parties have
sufficient access to capital to provide the loans necessary, and there are
relatively symmetric incentives to act; otherwise there may be a socially
inefficient level of lending (either too much or too little). Even if the other
party to the transaction does not act to deliberately counteract the action, it
will be well positioned to detect it and report any misconduct to the market
(to impose reputational penalties) or the government (to impose civil or
perhaps criminal penalties). (As a side note, insofar as antitrust-based
uncertainties preclude collective action on the part of numerous sellers of
protection, these rules should be rethought in this light.)
The more generic version of this moral hazard concern is the risk
of sabotage. Individuals and entities that are not insurance companies
(meaning: not regulated by insurance law) are not permitted to write certain
insurance contracts, say, on an individual’s life, out of concern that one of
the parties will try to sabotage the contract. Or, to look at another way,
where there is no symmetry or where the attack and defense would simply
result in an arm’s race of dead weight costs, the risk created by insurance
contracts outweighs any gains. This is especially true when the value of the
asset that is the subject of the contract is particularly valuable or difficult to
value.
Although this logic might make sense for individuals and contracts
like life insurance, the risk of sabotage is overstated in the world of credit
derivatives. First, in this $60 trillion industry, there has never been a
reported case in which one party to a contract acted to deliberately sabotage
an underlying borrower in order to cash in on a credit derivative contract.
Second, the gains from sabotage are as great or greater in equity markets,
currency markets, and a whole host of other markets where third parties are
able to make large bets on economic outcomes. For instance, a malicious
investor could take a large short position in Firm X, and then destroy an
asset of Firm X, say by not loaning it money, blowing it up, spreading
rumors about it, or any number of activities. This risk is real, but it is
uncommon because other laws (for example, criminal law, tort law, and
securities laws banning market manipulation) and norms restrain
individuals from making socially destructive (but privately beneficial)
decisions. However, market participants, observers, and regulators should
always keep a look out for the possibility of sabotage, since reporting it to
the market and prosecutors is likely to provide sufficient deterrence.
***
As this section has showed, regulators are likely to find it very
difficult to draw sensible lines or regulate financial products in a coherent
and efficient manner under the rubric of insurance regulation. This means
that regulation might more sensibly focus on investors instead of
investments. Insurance law is based less on regulating insurance than on
regulating insurance companies. In other words, we do not regulate
insurance companies because they sell insurance, but rather regulate
insurance contracts because they are sold by insurance companies. The
right question to ask is not whether credit derivatives are “insurance,” but
rather if they are sold or issued by “insurance companies.” This, of course,
begs the question of what should count as an insurance company.
The next Part tries to answer this question by looking at the policy
reasons for having a separate body of insurance law to regulate insurance
companies. The policy reasons are uniquely applicable to insurance
companies, not all firms that participate in credit derivative markets, and
thus there is no good policy reason for applying insurance regulation.
B. THE POLICY REASONS FOR INSURANCE LAW DO NOT OBTAIN
Given the problems of defining what “insurance” is, it must be the

case that the scope of insurance law is either quite arbitrary or based on
other considerations. In fact, insurance regulation exists not to regulate
insurance contracts per se, but rather to regulate contracts issued by
insurance companies. For sophisticated or semi-sophisticate parties, the
insurance companies are the problem, not the insurance.47 Insurance
companies are regulated differently than companies producing other
products because of the unique governance problems associated with their
production cycle and unique governance structure. Let us consider these in
turn.
47
There is another justification for insurance regulation that has to do with the
consumer-facing nature of some insurance contracts. See infra p. 32.
1. Inverted Production Cycle
The first unique feature of insurance companies is the fact that they
sell their products long before they are delivered to customers. This means
the normal production cycle is inverted. The typical (that is, non-insurance)
firm produces products and then sells them in return for cash. Payment and
delivery are linked closely in time and there is an immediate feedback from
customers. Insurance companies, however, have a much different
production cycle that causes unique governance problems. An insurance
company’s customers are policyholders, who pay (in the form of premium
payments) in advance for products (payments on claims) that come many
years later, if at all.48 This is important because the discipline on how cash
can be spent that comes with having to sell valuable products or services in
the market is missing or attenuated. Payments are made based on promises
alone, and there is thus the risk that the cash reserves given in advance to
the insurance company will be squandered on risky investments, and thus
unavailable to pay off claims when they come due. This is the Ponzi-
scheme problem discussed above. When there is continuous solicitation of
investment by outsiders and a mismatched payment scheme (current
investments pay liabilities of previous investors), there is a risk that
managers will engage in too much risk when the liabilities that arise are
greater than predicted. In these bad states of the world, insurance company
managers have incentives to attract more capital on irrational terms to pay
current liabilities owed to prior investors.
The inverted production cycle of insurance companies has another
problem where there are competitive markets for insurance services. In an
unregulated market, insurance companies are bound to compete heavily on
price, and this may lead to under reserves such that future liabilities will
not be covered by sufficient assets. There are two parts of this claim, so it is
worth unpacking it.
First, competition among insurance companies is likely to focus
primarily and perhaps excessively, on price. This is because the quality of
the products insurance companies are selling (the other thing on which they
compete) is identical or unobservable. The repayment of losses less the
deductible is the same regardless of the insurer. There is some risk that the
insurer will fail and be unable to repay the liabilities, but this is something
that is, by its very nature, unobservable by the insureds. Reputation and
longevity may be correlated with this risk, but these factors coupled with
48
The payout for life insurance policies may obviously be decades away.
the governance problems noted above may simply lead to more risky
investment choices by the insurer in the next period. Another unobservable
component of potential competition is service quality – that is, the
timeliness of payments, the cost of the claims process, the quality of
customer service, and so on. Like the risk of default, these will be
unobservable by the insureds in the period when they make their
investment decision, since they happen only many years later and after
premiums have already been collected. For these reasons, price is likely the
primary way in which insurance companies would compete in an
unregulated market.
Second, price competition for insurance products is different from
price competition for non-insurance products, and, if unregulated, may lead
to pricing at below marginal cost. Non-insurance firms have no incentive to
price below marginal cost, since, as a consequence, every sale would lose
the firm money. Insurance companies, however, have inverted production
cycles, which translates that the costs of the product being sold are felt long
after the cash is collected by the firm for the sale. This means myopic
managers, hubristic managers, over-confident managers, or desperate
managers may charge too little for new insurance premiums. Insurance
involves extensive long-range forecasting and the potential for costs, which
are realized only after sales, to be much higher than expected. Absent the
immediate feedback loop of typical production cycles, the possibility of
competition leading to destructive price wars is greater than for normal
firms.
It is true that credit derivative contracts are somewhat based on
future results and forecasting problems may arise. But this is concern is
ameliorated by several factors. Credit derivative contracts are generally
much shorter in term than insurance contracts, lasting a maximum of five
years, and very often held for much shorter than that. So although make-
whole payments under the contract may occur in the future, the potential
for error is reduced by the fact that forecasting need be made over a much
shorter period. The risk is also priced much more frequently, since
payments made by protection buyers are due quarterly. For most credit
derivative contracts, the prices of buying and selling production are
adjusted quarterly depending on the financial condition of the underlying
borrower, and this generates the kind of frequent pricing data that is
common in regular product markets. Even where it is not, the continuous
pricing of the same debt in the market allows holders of risk to engage in
pairwise transactions that allow them to rebalance their portfolio on short
time horizons. In addition, the parties on both sides are highly sophisticated
financial institutions (and their investors), and it is unlikely that any
outsider, like a regulator, could do a better job at estimating the future cash
flows from particular debt instruments so as to improve pricing.
The pricing problem for insurance companies is exacerbated by the
weak governance structure described below, which may encourage overly
risky actions when policy payouts exceed expectations. In short, the
managers of an insurance firm that charges too little, for whatever reason,
and finds itself unable to meet claims as they come due, may be less
constrained by creditors in the kind of response it will take. The managers
may simply try to sell more policies to pay off existing claims from other
policyholders with the hope of someday righting the ship. This potential
that arises from price competition may turn an insurance company into a
sort of Ponzi scheme.
This super risk preference situation is unlikely to arise in the case
of non-insurance firms because of the discipline of product markets and
because of the discipline of creditors when times are bad. For credit
derivative firms (that is, SPVs holding the rights to the cash flows from
various debt instruments), the probability of this arising is even lower. This
is because investment by the SPV managers happens before any investment
is made by shareholders in the SPV. A pot of cash is created and then sold,
with an implicit promise that no more assets will be added that that
particular pot. The possibility of super risk preferring managers is thus
extremely unlikely.
2. Weak Governance Structure
The second unique feature of insurance companies is the weak

corporate governance structure that is inherent in the insurance company
model. Non-insurance firms are generally funded by a large number of
diffuse shareholders and a small number of concentrated creditors,
typically banks or other lenders. In this governance model, the shareholders
are the residual claimants of firm value, and in good times it is in their
interest that the managers operate the firm.49 The diffuse nature of the
holdings of equity, coupled with the business judgment rule, means that
firms have a lot of slack in the risk they take during good times. When
things turn for the worse, however, the interests of the shareholders are set
aside and the concentrated interests of banks and other lenders take over
49
Assuming, of course, the managers aren’t acting in their own interest.
the decision-making process.50 Loan covenants are enforced, contracts are

rewritten by the lenders, managers are replaced, and a restructuring of the
firm is undertaken. The reason creditors take effective control long before
bankruptcy is because in the absence of these control rights, shareholders
(and their agents, the managers) would have incentives to act ever more
risky as the value of shares drops. Consolidated creditors can help preserve
going concern value and reduce dangerous risk taking in near final periods.
(It is, of course, irrelevant whether this concentration of creditors takes
place when debt is issued (when times are good) or when it is consolidated
(when times are bad), since concentration is less likely in insurance
companies in either case because the creditors of the firm are its policy
holders.)
Unlike non-insurance firms (but like banks), insurance firms are
structured with both weak equity holders and weak creditors. Insurance
firms (and banks) have shareholders that are indistinguishable from other
firms, but their creditors are as diffuse and disinterested as their
shareholders. An insurance company’s creditors are its policyholders. Their
large number makes coordination difficult, either for monitoring or action,
and information costly and very unlikely to be obtained. In addition, policy
holders are not investors (like many shareholders are), and therefore likely
to be unaware of and unsophisticated about matters of corporate
governance and finance. And unlike diffuse creditors of non-insurance
firms (e.g., bondholders), the claims of policyholders cannot be and are not
consolidated or concentrated in periods of distress. This means an increased
threat of excessively risky decision making in bad times because the
insurance company’s creditors are diffuse instead of concentrated. Both in
insurance and banking, where depositors are substituted for policy holders,
this suggests the need for a prudential regulator to effectively consolidate
the diffuse policyholders into a bank-like consolidated creditor to deal with
the insurance company in bad times.
This governance problem is not present consistently, if ever, in
credit derivative transactions. To see this, consider the simplest case of a
plain-vanilla CDS. Remembering the analogy with insurance set forth
above, the bank that lent the money to the borrower is the insured, and thus
analogous to the policyholder in an insurance contract. In a single-name
CDS contract, there is no pooling and therefore no diffusion of interest
50
See, e.g., Douglas G. Baird and Robert K. Rasmussen, Private Debt and the
Missing Lever of Corporate Governance, 154 U. PA. L. REV. 1209, 1209-11
(2006).
among the alleged insurer’s “policy holders.” In addition, banks that buy
credit protection are nothing like individuals that buy auto or life insurance
policies. Banks are sophisticated, repeat players, represented by counsel,
capable of processing information about the riskiness of their counterparty,
and have tremendous bargaining power.
It is possible for a protection seller to write many CDS contracts,
and thus put any individual buyer of credit protection into the position of
holding a small claim against the firm, say insurance company or hedge
fund. Although this would not change the nature of the protection buyer
and the ability of it to fend for itself, it does raise potential concerns about
the management of the protection seller. But this just then puts the inquiry
about governance back at that level, and tied to the nature of the seller of
protection. If the protection seller has a weak governance model, like that
of a bank or an insurance company, then it may be susceptible to this
problem, but if it is an entity with strong governance in bad times, then the
concern about too much risk on the part of managers (on behalf of
shareholders) is much lower. Hedge funds, for instance, must return to the
market frequently for capital (that is, they do not have capital lock in) and
are funded by extremely sophisticated investors. They are decidedly not
subject to this concern. In short, insurance companies with weak
governance should be subject to regulation to avoid the social inefficiency
that might arise from their governance structure, while non-insurance
companies, with strong governance, are less worrisome.
The same result obtains even when we consider a more
complicated credit derivative contract. The parties buying protection that
have their default risk pooled into a CDO structure are large financial
entities with much greater sophistication and risk-bearing ability than
individuals buying typical insurance products. The risk that the sellers of
protection will “pull a fast one” on them is much lower given this
sophistication. In addition, the investors in the SPV holding the default risk
(the analogous insurance company) are likewise large financial entities
capable of making risk assessments, demanding and processing
information, pricing risk, and wielding their bargaining power in the event
a bad future state arises.
Moreover, the nature of the typical CDO structure is effectively a
one-time game, in which credit risks are pooled and the cash flows sold off
to investors. The sponsor and manager of the SPV does not continue to sell
protection based on a pool of funds provided by investors (as in an
insurance company), but rather makes the investments first (by choosing
risks to pool), then goes to the market to sell cash flow rights to investors.
This means that managers of the SPV do not really do much or can do
much in terms of risk alteration of the SPV once it has raised money from
investors. The future course of the SPV is set, and the payouts are what
they are. No future investments are made, no additional investors are
brought in, and there can be no change in strategy for the firm. Therefore,
there is less chance for abuse in the event the SPV payouts are less than
expected. Governance quality is largely irrelevant in this model firm.
Applying this governance model to the insurance company model,
it is as if the insurance company wrote all of its policies before raising
money in the market. In that case, investors would worry less about the
governance of the insurance company, since its job would simply be to
process claims from the policies it had written – it would not take on new
policies (and a new source of cash) on terms likely to be unfavorable to
existing investors. There would still be some governance risk, however,
since the decisions on what policies to pay out on, how much to
compensate executives, and other firm costs still have to be made. In some
of these, managerial interests may be aligned with those of investors, while
in others they may diverge. Importantly, however, this residual governance
risk is not present in the CDO case, since all of these decisions are made
before the investment in the firm (for example, management fees) or are
automated (for example, the amount of payouts). In short, any governance
problems simply do not obtain in the typical structure of credit derivative
contracts.
3. Consumer Protection
The third policy reason for a separate body of insurance law is the
need for strong consumer protection. While the concern with the inverted-
production-cycle and governance problems was basically insurance firms
not charging insureds enough, the consumer protection concern is that
insurance firms will charge too much. As mentioned above, the concern is
based on the following syllogism: insurance is a critical product for most
individuals; individuals are not sophisticated about insurance products or
contracts; and therefore insurance companies will take advantage of
customers by overcharging them. Accordingly, (the bulk of) state insurance
law regulates the substance and terms of insurance policies (to make them
simpler to understand and compare across firms), as well as regulating
service and coverage issues (to make sure insurance firms do not back
away from promises to pay). In other words, insurance is sometimes
regulated as a specialty consumer product in which informational and
bargaining power asymmetries are sufficiently large that social losses may
be generated from an unregulated market.
The credit derivatives market described above has none of these

characteristics or concerns. The parties to credit derivative transactions are
all large financial institutions or other sophisticated investors with access to
information, the ability to understand and process the information,
bargaining power, and the ability to bear losses. This is in sharp contrast
with insurance contracts entered into by average consumers, who have
none of these attributes. It may be sensible for insurance regulators to try to
reduce informational and bargaining power asymmetries between insurance
companies and consumers, to provide oversight of claims management and
customer service, to provide standardized contract terms that allow
comparison shopping, and to even regulate rates, but these policies are
unnecessary where the buyers and sellers of “insurance” are large financial
institutions. In fact, if anything, the sellers of protection (the alleged
insurance companies) may often be less sophisticated than the buyers of
protection (the alleged insureds). For example, a small hedge fund run by a
few investors may enter into a contract to sell protection to a large
commercial bank. In this case, it is not at all clear where insurance-law-like
consumer protection duties should run. After all, existing law will treat
both the hedge fund and the bank as not needing the protection of the
securities laws or other regulations.
In addition, as discussed above, standard-setting groups, like the
International Swap Dealers Association, are already doing much of the
work for credit derivative markets that insurance regulators do to protect
average consumers. ISDA provides, among other things, standard form
contracts (which innovate to respond to changes in the market), dispute
resolution mechanisms and guidance, information, educational services,
and so on.
C. INSURANCE LAW DOESN’T WORK WELL AND WON’T

UNIQUELY ADD MUCH
A final argument against treating credit derivative contracts as

insurance is a practical one having to do with the value that insurance
regulation, as currently constituted, might add to the market. In short,
insurance law and its generation and enforcement regime is generally
considered to be inefficient and in need of dramatic reforms, and is
therefore not the most appropriate locus of authority for a regulation of a
new financial innovation that spans numerous types of institutions and
serves innumerable purposes, most of which have nothing to do with
insuring against risk as it is commonly understood.
1. Jurisdictional Issues
As discussed above, insurance regulation is state based, meaning

there are at least fifty different regulators and models of regulation. The
state-based model is largely premised on the consumer-protection function
of insurance regulation, since it is the state police power to protect citizens
from abuse that justifies a local approach. While the merits of this model
are debatable when it comes to providing efficient insurance services for
health care or automobile risk, the global nature of modern financial
markets makes the local argument much more difficult to make for credit
derivatives. Do citizens of New Jersey need different protection in credit
derivative markets than citizens of New York? In fact, insurance companies
seemed to exploit this regulatory fracture by dividing up their businesses
into discrete components that were regulated piecemeal by various state
regulators. Given the ease of capital flows, the ability of firms to
incorporate anywhere around the globe, and the fact that even transactions
in the Cayman Islands can impact investors around the world, the idea of
insurance regulators in a particular state controlling the global market
seems fanciful. New York regulators, for instance, had authority over less
than 10 percent of AIG’s operations, because of the corporate structure of
AIG.51 As a result, one of world’s biggest insurers was, under the current
system, able to be largely below the radar screen of its primary insurance
regulators. Applying this dysfunctional model (which is based primarily on
consumer issues, also which are the least applicable to this market) to new
financial products makes little sense.
The choice of regulator, be it a question of a particular entity or a
general jurisdictional choice (like federal or state), is based on regulatory
expertise, incentives, and the expected consequences of the regulatory
model on the ability to minimize the decision costs and error costs of
regulation. State-based regulation might make sense if jurisdictional
competition is likely to lead to the efficient regulation (that is, the race-to-
the-top theory of state-based corporate law). But this is not the basis for
state-based insurance law. State insurance law is not based solely on the
state of incorporation of the insurance company, but rather the locus of
policy holders. In any event, this model will probably not work for a
market like credit derivatives. Financial markets are generally regulated by
51
See This American Life #382: The Watchmen (Chicago Public Radio
broadcast Jun. 5, 2009), available at http://www.thislife.org/extras/radio/
382_transcript.pdf.
federal agencies (or perhaps in the future by international ones) because

financial products are sold in a global market, and any state-based regime
could be avoided by simply opting into offshore regulatory regimes. If the
goal of regulation is to encourage credit derivatives to be traded on
exchanges as opposed to in over-the-counter markets, as argued below,
having a single regulator to choose from is crucial to creating
commodifiable products. If credit derivative-holding SPVs could opt into
one set of regulations through choice, this might provide some federalism-
esque benefits, but this is not the way insurance law operates.
Another factor influencing the choice of regulator is expertise.
Here too, there is nothing about state insurance regulators that seems
special or unique. State insurance regulators are used to dealing with
insurance companies and insurance contracts, which, as described above,
deal with issues of risk pooling, governance problems, consumer-facing
contract issues, and the like. None of these obtain in credit derivative
markets. Insurance commissioners are also generally concerned with
counterparty risk—a real concern in credit derivative markets—but this is
something bank regulators (like the FDIC, Federal Reserve, and Treasury),
derivative regulators (like the CFTC), and securities regulators (like the
SEC) are also especially concerned with. In addition, these latter regulators
do not have the state overlap problem described above.
2. Substantive Law
Deeming credit derivatives to be “insurance” (or credit protection

sellers “insurance companies”) would, under current law, have several
consequences, none of which is likely to improve the efficiency of credit
derivative markets.
i. Licensing
First, entities could not sell protection unless the seller was a
licensed insurance company. All fifty states require a state-issued license
before a firm may issue an insurance policy.52 Such a pre-screening
requirement might make some sense as part of trading on a credit
52
See N.Y. INS. LAW. § 1102(a) (2006). “No person, firm, association,
corporation or joint-stock company shall do an insurance business in this state
unless authorized by a license in force pursuant to the provisions of this chapter, or
exempted by the provisions of this chapter from such requirement.” Id.
derivatives exchange, say by requiring those buying or selling on the

exchange to meet certain criteria, like margin requirements. However,
certifications based on other factors unrelated to the exchange’s risk of
default would add costs without any offsetting benefits, and would merely
open up the possibility that regulators could extract rents from firms
wanting to participate in these markets. As discussed below in the context
of capital requirements, the licensing scheme for insurance companies is
doable in part because there are only a handful of firms providing insurance
in each state. In contrast, there are literally tens of thousands of investment
funds that have sold or could sell credit protection in credit derivative
markets, and this would make any licensing scheme prohibitive or
meaningless for state regulators. It would also impose potentially large
costs on funds who do not sell protection as a normal part of their
investment strategy, but might find it efficient and sensible to do so in
limited cases. Regulatory costs would therefore deter these funds from
participating in the market, without any proof that the funds have imposed
any costs on others.
ii. Duties
Second, the buyer and seller of protection would be subject to a

duty to act with the utmost good faith, that is, something beyond the
“morals of the marketplace.” This might make some sense for markets in
which buyers and sellers are of widely differing sophistication, have access
to different information, and have different bargaining power, but it makes
much less sense when the parties on both sides of a transaction are similar
giant financial institutions. In fact, the trend in the market is for large
investors to opt out of these kinds of disclosure requirements and the like
using waivers known as “Big Boy” letters. To impose fiduciary duties or
other litigation-generating obligations on parties without the potential for
opt out will increase uncertainty and costs without any obvious benefit
from ex post judicial determinations of what were and were not good deals.
iii. Capital Reserves
Third, protection sellers would be required to maintain a certain

amount of capital based on the risk inherent in its “insurance-based”
business.53 For instance, state insurance regulation requires every insurer to
53
See N.Y. INS. LAW. §§ 1322, 1324 (2006 & 2009).
maintain certain specified amounts of capital on hand to reduce

counterparty risk and to submit its risk-based capital levels to regulators on
an annual basis.54 The amounts of capital required vary by jurisdiction and
entity, depending on the riskiness of the insurance company.55
In the abstract, capital reserves are unobjectionable, since they are
about reducing counterparty risk, and therefore about increasing the
number of socially beneficial transactions. After all, these requirements are
a common element of banking law and other areas where counterparty risk
and the problem of runs and systemic risk are present. The question is then
how much capital should be required, what is the best way to reduce
counterparty risk, and which regulator has the incentive to figure these
things out. There are several reasons why insurance regulators are not
obviously the best candidate to fulfill this mission, and why the solution of
requiring credit derivatives to be exchange-traded is a more elegant
solution, albeit one fraught with problems too.
There are several problems with insurance regulators imposing this
requirement on the credit derivative markets. For one, the number of
entities and individuals writing protection on indebtedness is enormous,
making any pre-screening regulation extremely costly. For example, there
are over 15,000 hedge funds in the United States alone, each of which
could be a participant in these markets.56 The magnitude and complexity of
the job of simply tracking each of these hedge funds—only one type of
protection seller—would be beyond the capacity of any existing state
54
See N.Y. INS. LAW. § 1402(a) (2006). “Before investing its funds in any
other investments, every domestic insurer shall invest and maintain an amount
equal to the greater of the minimum capital required by law or the minimum
surplus to policyholders required to be maintained by law for a domestic stock
corporation authorized to transact the same kinds of insurance, only in investments
of the types specified in this section which are not in default as to principal or
interest.” Id.
55
The Model Insurance Act, for instance, provides for three “risk-based
capital” levels: (i) mandatory control level risk-based capital (measured at .7 times
authorized control level risk-based capital), (ii) regulatory action level risk-based
capital (measured at 1.5 times authorized control level risk-based capital), and (iii)
company action level risk-based capital (measured at 2.0 times authorized control
level risk-based capital). NAT’L ASS’N OF INS. COMM’RS, MODEL LAWS,
REGULATIONS, AND GUIDELINES vol. III (2009).
56
See Number of Hedge Funds up by Two Thirds in Two Years, Concludes
PerTrac Study, THE TRADE NEWS, Mar. 5, 2007, available at
http://www.thetradenews.com/hedge-funds/prime-brokerage/613.
regulators. The job would be made even more overwhelming since selling
protection is not necessarily a full-time job. Hedge funds and other sellers
may hold the default risk for a very short time (a few months or less) and
may engage in transactions only periodically or on a one-off basis. Unlike
insurance companies that exist to provide risk sharing services, protection
sellers are not necessarily in the business of holding debt risk. The fluid
nature of market participants would make any licensing or ex-ante
regulatory regime incredibly costly and drive many participants out of the
market.
In addition, capital requirements did not work well if at all in
preventing insurance companies, such as AIG, from investing aggressively
and, as it turns out, dangerously in credit derivative markets. The state-
based model was manipulated by AIG and others, and this possibility could
only be expected to be worse if every credit derivative protection seller
becomes a ward of insurance regulators. In other words, the job of
regulation would get much more difficult without any obvious way of
increasing the capabilities of regulators. This point is made even clearer by
reiterating the point made above about how insurance regulators are not
experts in financial markets in which most protection sellers participate. If
insurance companies can avoid insurance regulation, it is very likely that
hedge funds and other sophisticated and fast-moving private money funds
will also be able to do so.
Moreover, capital adequacy requirements imposed by regulators
(as opposed to margin requirements required by exchanges) generated the
incentive for regulatory arbitrage described above. Firms subjected to these
requirements had incentives to hold higher quality debt risk, which
received lower capital charges, and to move debt risk off of their balance
sheets and into bankruptcy-remote SPVs. Although this type of arbitrage is
likely inevitable at some level, the current regulatory model for insurance
proved ineffectual at preventing arbitrage that imposed systemic risk
externalities on society.
Finally, insurance regulators are not experts about the amount,
type, and structuring of capital requirements to reduce counterparty risk in
non-insurance financial transactions. It is arguable that insurance
regulators, representing the state, have incentives to determine the amount
of social cost from the failure of an insurance company, since many of the
social harms that would result would be paid for by a state-funded social
safety net or would otherwise result in state-based harms. But the failure of
a hedge fund or foreign bank or other protection seller may generate no
social losses, because gains from bets on one side cancel out losses from
bets on the other side, or are ones that are not clearly within the purview or
concern of state regulators. Systemic risk is something not felt completely

or even directly by one state, and therefore a collective action problem may
generate insufficient incentives to get the regulation to the efficient level.
It is important to note that, as proposed below, trading credit
derivatives on an exchange would likely require some financial assurances
akin to capital adequacy requirements on market participants, through
margin requirements, and on the exchange, which would be the ultimate
bearer of counterparty risk. For the reasons discussed below, the concerns
here are much less than through regulatory capital reserve requirements.
For one, exchanges, which act as a centralized counterparty, bear the entire
risk of loss if a trading party defaults, and therefore have the best incentives
in terms of setting up rules to ensure that traders are likely to pay for their
losses.
iv. Disclosure
Fourth, being an insurance company would trigger a detailed

disclosure requirement of any insurance business to state regulators. The
state-based requirements track roughly those firms with publicly traded
securities. Audited reports of the insurance company’s financial and
accounting situation must be made quarterly and annually.57 These include
disclosure of routine data, like the firm’s balance sheet, income statement,
and statement of cash flows, as well as more detailed information than
generally required by securities laws, like a list of every asset owned by the
firm, every asset acquired or sold during the relevant period, a report of all
derivative transactions, and so on.
Although disclosure of the assets and risks of hedge funds and
other private wealth pools may indeed be a socially valuable regulation,
there is no obvious reason why this should be tied to a regulatory apparatus
that is about only a very small part of what a hedge fund may be doing or
may have done. As noted above, there are potentially tens of thousands of
separate legal entities participating in credit derivative markets at any time,
and requiring each of them to make disclosures to insurance regulators
upon engaging in such activity is highly problematic.
As a basic principle, disclosure regulation should be implemented
and monitored by regulators that cover the primary activity of the regulated
entity. It is for this reason that the SEC is the agency responsible for the
57
See N.Y. INS. LAW §§ 307(a), 308 (2006) (requirements for filing an annual
financial statement and a quarterly financial statement).
disclosure by firms issuing securities and by brokers and dealers that buy
and sell securities, and the reason that insurance regulators are responsible
for the disclosure by insurance firms. To require an investment fund to
make disclosures to insurance regulators solely because it engaged in a
credit derivative transaction will impose potentially large costs on funds
based on potentially a single or small number of transactions. If the
disclosure rules are tied instead to how often a fund trades or how many
trades it makes in these markets, the funds will inevitably try to avoid these
costs by making the decisions on whether to sell protection based in part on
the arbitrary triggers. For example, if ten incidents of protection selling
trigger a disclosure obligation, we shouldn’t be surprised to see funds
selling protection nine times.
If the reason for insurance-based disclosure rules is because of the
lack of disclosure to other regulators – hedge funds have little or no
disclosure obligations under the securities laws – this is not an argument
for disclosure to insurance regulators, it is an argument for a securities law
disclosure regime. The regulator that can best calibrate what kinds of
disclosure are cost justified, what form disclosures should be made in, and
what to do with the disclosed information, if anything, is the regulator that
should require and monitor disclosure. For one, it is not clear what
insurance regulators would do with the disclosures, especially if the bulk or
almost all of it is about activities that are unquestionably not insurance.
v. Contract Regulation
Fifth, state law generally requires insurance companies to submit

insurance contracts, known as “policies,” to state regulators for pre-
approval before any policies can be sold using the contract. For example, in
New York, contracts for life, accident, and health insurance are subject to
prior regulatory approval.58 This requirement would layer possible fifty
different state law requirements on top of existing private contracting in the
over-the-counter credit derivative markets. There are several problems with
such a requirement.
Most obviously, as noted above, there is already a quasi-regulator,
the ISDA, that provides industry-wide boilerplate contracts for credit
derivative transactions. As the Conseco and Xerox examples above
58
Under New York law, life, accident and health and annuity policy forms are
subject to prior regulatory approval. Compare N.Y. INS. LAW. § 1102(a), with N.Y.
INS. LAW § 1108(a) (2006).
illustrate, ISDA is already incorporating best practices into these standard

contracts, as well as modifying terms that cause problems regulators would
be concerned about, such as manipulation, externalities, and contractual
unfairness that may have arisen from any bargaining power asymmetries,
mistakes, or the like. As such, contract regulation is likely redundant, and
would in any event be replacing a highly knowledgeable set of regulators
with one without any experience with credit derivatives.
Paternalistic contract regulation to protect one party or the other is
also unnecessary because of the sophistication of the parties to these
contracts. There is no obvious systematic bias in favor of one party or the
other in these contracts, and the typical arguments that may justify contract
form and substance regulation – for example, information or bargaining
power asymmetries – do not obtain or point always in one direction.
A final point has to do with the fact that the parties to credit
derivative contracts are not tied to physical locations in the way that
insureds are, and therefore any state-based regime will inevitably invite
avoidance through incorporation choice or choice of law provisions. This
may be viewed as normatively good or bad (the old race to the top versus
race to the bottom debate), but even where it might be thought of as
generating efficient contract forms that private parties would choose in any
event, it would take us simply to the current ISDA model. After all, if there
were a more efficient set of contracts that could be written – that is, the one
that parties freely choosing would choose anyway – it would exist or will
exist under the current quasi-regulatory regime.
vi. Price Control
Finally, states impose substantive restrictions on the prices that can

be charged by insurance companies. Regulation of prices varies widely by
state and by the type of insurance, but a few common themes are apparent.
There are generally three types of regulation: pre-approval, “file and use,”
and “use and file,” with the strictness of the regulation decreasing
accordingly. For example, New York law requires prices for workers’
compensation and automobile insurance to be approved in advance by
regulators, while rates for property and casualty insurance are subject only
to a pre-issuance filing policy.59 The general regulatory touchstone is that
rates shall not be too high, too low, discriminatory, or anticompetitive.60
59
Compare N.Y. INS. LAW §§ 2305(b), 2310(a), 2344 (2006), and N.Y.
COMP. CODES R. & REGS. tit. 11 §§ 161.1-161.12 (2009), and N.Y. INS. LAW §§
Whatever sense price regulation makes for contracts written by

insurance companies, it makes much less sense for credit derivative
products. Price regulation of insurance contracts is premised on the pooling
of large numbers of individuals and on the governance problems described
above. In contrast, credit derivative prices are based on the idiosyncratic
risk associated with particular firms. This is something that is difficult to
price in the abstract or to know when prices are too high, too low,
discriminatory, or the like. In addition, because the price of risk is traded in
markets, the idea of using regulators, especially ones without any expertise
or experience in this area, to set prices is nonsensical.
There is one area where the pricing of credit derivatives was
erroneous. As shown in recent research, price models used ubiquitously by
buyers, sellers, credit rating agencies, and other participants in the markets
systematically mispriced various tranches of risk.61 It turns out that highly
rated tranches were underpriced, meaning they were riskier than buyers and
sellers thought, and unrated tranches were underpriced, meaning they were
less risky than thought.62 Importantly, however, no one was aware of this
problem, even though everyone had strong incentives to be so. In addition,
this kind of error is now known, and parties to these contracts do not need
regulators to inform them about these errors. Other pricing issues might
arise in the future, but market participants have incentives to identify such
issues. The problem was not that the market for setting prices was biased in
one way or the other, but rather simply a mistake in assumptions.
Regulators are not well positioned to remedy these kinds of problems
absent a crystal ball that no one believes they possess.
Moreover, if various states are competing with each other to offer
market participants pricing regulations that fit their needs, the jurisdictional
choice point made above will obtain – contracts will migrate to those states
that offer the pricing rules that the parties would have come to anyway.
2303, 3231(d) (2006).

60
Compare N.Y. INS. LAW §§ 2305(b), 2310(a), 2344 (2006), and N.Y.
COMP. CODES R. & REGS. tit. 11 §§ 161.1-161.12 (2009), and N.Y. INS. LAW §§
2303, 3231(d).
61
See Joshua D. Coval, et al., Economic Catastrophe Bonds, 99 AM. ECON.
REV. 628 (2009) (showing how AAA-rated tranches contained very little to no
idiosyncratic risk, but large and underappreciated amounts of systematic risk).
62
See id.
V. AN ALTERNATIVE TO INSURANCE REGULATION
It is understandable why people mistakenly analogize credit

derivatives to insurance: insurance is about risk sharing and diversification,
and this is what credit derivatives are about as well. Insurance companies
were also big players in credit derivative markets. But other contracts are
about these things as well, and there were many other types of entities that
participated in these markets. In addition, credit derivatives are about many
other things than risk sharing. In fact, as shown above, credit derivatives
may have started as a risk-sharing or risk-transferring mechanism, but their
primary use was and is speculation, hedging, and other non-insurance-like
functions. Moreover, even where the insurance analogy is most apt, it does
not follow that the current insurance regulatory regime is the best available
for credit derivatives, assuming additional regulation is needed.
There may be a case for more regulation, premised on the failure of
the market to adequately address counterparty risk issues, but insurance law
has little to add. A simple rule requiring derivative contracts to be traded on
an exchange in most cases will do most of the work required.
As noted above, a credit derivative does not eliminate risk for the
original bearer of it, but simply trades default risk for counterparty risk. In
other words, the risk in a loan that the borrower will not repay is traded for
the risk that the seller of default protection will not pay in the event the
borrower does not. This counterparty risk was bigger than anyone thought;
firms no one thought would fail, like AIG, failed by taking on too much
risk. This led to a cascade of failures of brokers, like Lehman Brothers, and
other intermediaries, which in turn led to huge collateral calls and a general
constriction of credit flows. Quite simply, the mispricing of and realization
of counterparty risk caused the credit crunch.63
Fortunately there is a somewhat simple solution to reducing
counterparty risk – an exchange. Using a centralized exchange, like the
Chicago Mercantile Exchange, eliminates the counterparty risk, replacing it
with the risk of default of the exchange. If A and B have a contract that
exposes A to a net of $100 in risk to B, this risk can be eliminated if A and
B both trade through a centralized clearinghouse or exchange. A will now
have a $100 liability to the clearinghouse, while B will have a $100 credit
63
See JOHN B. TAYLOR, GETTING OFF TRACK: HOW GOVERNMENT ACTIONS
AND INTERVENTIONS CAUSED, PROLONGED, AND WORSENED THE FINANCIAL
CRISIS (Hoover Inst. Press 2009) (showing how the credit crunch was not caused
by a liquidity shortage but by an increase in counterparty risk).
with the same. If A defaults on its ability to pay B, B can still be paid by
the clearinghouse. As long as the clearinghouse is solvent, the counterparty
risk for B is eliminated.
The risk-reducing qualities of an exchange can be seen more
clearly when the number of trading parties increases. Consider the case
where A is owed $100 by B, B is owed $90 by C, and C is owed $80 by A.
In this case, A has a net risk exposure of $180 to B and C, since if they
both default, A is owed $100 from B and owes $80 to C. If these three
liabilities are managed through an exchange, however, A’s risk exposure to
B and C is reduced to zero. In this scenario, A is owed $20 from the
exchange, and B and C each owe the exchange $10. Thus, A’s risk to B and
C has been eliminated, and the netting of liabilities has reduced the
magnitude of the overall amounts owed to much more manageable sums.
So long as the exchange can ensure, say through margin requirements, that
B and C can make good on their $10 (about 10% of the total bets), the
market stays together.
It is for this reason that the clearinghouse model is used for other
derivative markets, like commodities markets, futures markets, and
currency markets. Of course, the clearinghouse must be solvent and for this
it needs sufficient scale of operations and the ability to impose rules on
trading parties that help reduce the risk that they will not be able to make
due on their contracts. This last point is precisely about the locus of
regulatory authority – who knows better how to regulate the leverage or
other characteristics of market participants? Since the clearinghouse,
typically a for-profit enterprise, stands to lose personally and dramatically
in the event of a failure, it has arguably better incentives along this
dimension than government regulators, who are not betting their own
money and who, perversely and ironically, may see increased funding from
any failures.
Given these benefits, the question is why the exchange did not
arise as a natural part of the market. One answer might be that an exchange
has some elements of a public good, since it reduces the potential for
systemic risk by decreasing the likelihood of a credit crunch from the
failure of a single firm, and public goods are chronically under supplied by
the market. But the story here is more complicated, because the analysis
above suggests that it is in each individual firm’s interest to reduce risk in
this way. Moreover, the collective action problems that typically cause the
market to under supply public goods do not obtain in this context, since
there were only about eight major intermediary market makers, and they
were all located in New York City.
There must be, therefore, some benefit to writing contracts off

exchange that exceeded any benefits from counterparty risk reduction, at
least in expectation before the actual counterparty risks, like the failure of
AIG, were known. One possibility is that the brokerage houses arranging
over-the-counter credit default contracts are able to earn higher profits for
writing specialized contracts than they could for simply dealing in
standardized contracts, as typically required for exchange-traded contracts.
There is less pricing transparency in over-the-counter contracts, since they
are written for a specific buyer and seller in a one-off fashion, and there is
also more work that arguably goes into writing these contracts. And,
private firms do not bear the full costs of the over-the-counter system, since
bankruptcy law limits the downside risk to any investor to what they
invested. Under this view, brokerage firms are able to capture the private
benefits of idiosyncratic, over-the-counter contracts, while externalizing the
risks of systemic meltdown of the entire system.
In this way, the government’s initial efforts to encourage the
trading of credit derivatives on an exchange is a sensible reform. Firms
have resisted this to date,64 because nothing has changed the private
incentives with respect to systemic risk – in fact, the rash of bailouts of
private firms have arguably exacerbated the problem. In addition, there are
multiple competing exchanges, including the CME and ICE exchanges, and
academics have shown that exchanges need a great amount of scale to be
able to adequately reduce counterparty risk.65 The government may be
rightfully worried about choosing one exchange as the preferred or
exclusive exchange, but the need for scale may force some collective
choice to be made.
As noted above, the virtue of the exchange model is that it bakes
into a private-ordering system many of the laudable aspects of the
insurance law regime. Specifically, capital requirements, disclosure, pricing
transparency, and general oversight of risk are all functions that exchanges
provide, since exchanges are on the hook for losses arising from the failure
64
See Jacob Bunge, NY Fed Examines Slow Progress in CDS Clearing,
WALL ST. J. MARKETWATCH, Apr. 1, 2009, available at
http://www.marketwatch.com/story/ny-fed-examines-slow-progress-cds.
65
See Darrell Duffie & Haoxiang Zhu, Does a Central Clearing Counterparty
Reduce Counterparty Risk? (Rock Center for Corporate Governance, Working
Paper No. 46; Stanford Graduate Sch. of Bus. Research, Paper No. 2022, 2009),
available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1348343 (showing
how reductions in counterparty risk is highly correlated with scale and the ability
to net with other derivative contracts).
of any of the market participants. Moreover, insofar as there are multiple

exchanges competing to act as a clearinghouse, there will be competition in
law making, which will increase the chances of efficient rules being
created. In the private model, there is also less chance of regulatory capture
or a public choice distortion, because rival exchanges can always arise to
offer market participants alternatives. This assumes, however, that entry is
relatively unrestricted, something that is not necessarily true in a world
where scale is so important and perhaps difficult to achieve quickly. Insofar
as this is true, some oversight of the exchange(s) may be required to simply
ensure that they are not subject to these shortcomings. A first guess at a
sensible regulator of the exchange(s) would be one of the existing
regulatory bodies that deals with exchanges (e.g., the SEC or CFTC) or the
regulators that deal with banks and systemic risk (e.g., the Federal Reserve
or the Treasury Department).
VI. CONCLUSION
This Essay has shown that the simple argument that some credit
derivatives help banks and other providers of debt share risk with other
investors is not sufficient for credit derivative contracts in general to be
deemed “insurance.” A separate body of insurance law exists not because
the underlying contracts are insurance, but rather because typical insurance
contracts are sold by insurance companies. It has also shown that the policy
justifications for regulation of insurance companies—an inverted
production cycle, weak corporate governance in bad times, and
unsophisticated insureds—do not obtain in the context of credit derivative
markets or apply to parties to credit derivative contracts. Finally, it has
shown how an exchange for credit derivative contracts can provide most if
not all of the substantive regulation insurance regulators can provide, at
lower cost and in a more efficient manner. There remain unsolved
problems with the exchange solution, including issues of scale and bilateral
netting, but this is a subject for another day.
THE ROAD FROM “TWIN PEAKS” –
AND THE WAY BACK
Michael W. Taylor *
***
This article explores the fragmented regulatory structure of financial

markets in the United States in light of the current financial crisis. Two
approaches for regulatory reform that originated in the United Kingdom
are presented. The first approach is the creation of a unified regulatory
agency responsible for regulating all the main segments of the financial
services industry. The second, also known as the “Twin Peaks” approach,
is to structure regulation around two agencies, one responsible for the
safety and soundness of all financial firms and the other for regulating
their sales practices. This article describes the debate in the UK prior to
the creation of one unified regulatory agency, the Financial Services
Authority (FSA). Next, it explores justifications for a single regulator, such
as the FSA, followed by a discussion of the rejection of the “Twin Peaks”
approach in the UK. Subsequently, the debate regarding the role of a
central bank, like the Bank of England in the UK, is discussed. Then US
regulatory reform is reviewed in terms of the lessons of the British
experience of creating a single regulatory agency. Finally, the concluding
section describes how some variation of the “Twin Peaks” alternative
would prove to be more successful than the single regulator approach.
***
I. INTRODUCTION
The Global Financial Crisis has put the spotlight on the United
States’ complex and fragmented regulatory structure as an issue of global
systemic importance. The failure of large investment banks like Bear
Stearns and Lehman Brothers has put into question the adequacy of the
regulation of large non-bank financial intermediaries. The lack of
consolidated supervision of the AIG group, with its Financial Products
*
Adviser to the Governor, Central Bank of Bahrain; formerly Head of
Banking Policy, Hong Kong Monetary Authority; Senior Economist, International
Monetary Fund; Reader in Financial Regulation, ICMA Centre, University of
Reading.
Division falling under the authority of the Office of Thrift Supervision

(OTS) while the insurance company was regulated at State level, further
illustrates the systemic problems created by regulatory fragmentation.
Finally, the Commodities Futures Modernization Act, which deliberately
excluded the regulatory authority of both the SEC and the CFTC from the
Credit Default Swaps market, resulted in a failure to ensure adequate
regulation of that market with implications for the global financial system.
In a message clearly directed to US policy-makers, the Group of
Thirty, a think tank comprising some of the most distinguished figures
from international finance, has recommended in its report on Financial
Reform: A Framework for Financial Stability: “Countries should reevaluate
their regulatory structures with a view to eliminating unnecessary overlaps
and gaps in coverage and complexity, removing the potential for regulatory
arbitrage, and improving regulatory coordination.”1 This reevaluation has
now begun, with the structure of US regulation being seriously re-
examined for the first time in over a generation. Although the 1999
Gramm-Leach-Bliley Act dismantled the structural barriers between
commercial and investment banking and between banking and insurance, it
did not result in significant structural change to the complex and over-
lapping authorities of US regulatory agencies.2 However, in March 2008,
the Bush administration unveiled a plan for a major structural reform of
regulation,3 while more recently the Obama administration has proposed a
similar, but less radical reform, to Congress.4
The U.S. debate on regulatory structure has lagged behind in other
countries of the Organisation for Economic Cooperation and Development
(OECD) by over a decade.5 By the end of the last century many of these
1
GROUP OF THIRTY, FINANCIAL REFORM: A FRAMEWORK FOR FINANCIAL
STABILITY 10 (2009), available at http://www.group30.org/pubs/recommend
ations.pdf.
2
Heidi Mandanis Schooner & Michael Taylor, United Kingdom and United
States Responses to the Regulatory Challenges of Modern Financial Markets, 38
TEX. INT’L L.J. 317, 327-29 (2003).
3
U.S. DEP’T OF THE TREASURY, THE DEPARTMENT OF THE TREASURY
BLUEPRINT FOR A MODERNIZED FINANCIAL REGULATORY STRUCTURE (2008),
available at http://www.treas.gov/press/releases/reports/Blueprint.pdf.
4
U.S. DEP’T OF THE TREASURY, FINANCIAL REGULATORY REFORM, A NEW
FOUNDATION: REBUILDING FINANCIAL SUPERVISION AND REGULATION (2009)
available at http://www.financialstability.gov/docs/regs/FinalReport_web.pdf.
5
Schooner & Taylor, supra note 2, at 320. For a discussion of reform
elsewhere in the OECD, see id. at 340-44.
2009] ROAD FROM “TWIN PEAKS” 63
countries had already embarked on major reorganizations of their

institutional structures of financial regulation.6 These reform initiatives
were presented as a response to the challenge of regulating today’s
increasingly integrated financial markets in which the traditional
distinctions between banking, securities, and insurance had become
blurred.7 Moreover, with the dismantling of the structural regulations that
had previously segmented the financial industry, diversified financial
conglomerates had emerged, necessitating a group-wide perspective to
ensure their effective regulation.8
Two broad approaches emerged in response to these challenges.9
The first, and most high profile, was the approach adopted in the United
Kingdom that created a unified regulatory agency responsible for
regulating all three of the main segments of the financial services industry
for both financial soundness and consumer protection purposes.10 The
alternative approach, which had originated in the U.K. but was not adopted
there, was to structure regulation around two agencies, one responsible for
the safety and soundness of all financial firms and the other for regulating
their sales practices.11 This “Twin Peaks” approach was adopted first in
Australia and later in the Netherlands.12 Variations of it are also to be
found in Spain, France and Canada.
This essay attempts to distil some lessons for the United States
from the U.K.’s reforms and especially the factors that led to the creation of
a single, unified regulatory agency, the Financial Services Authority
6
Id. at 340-44.
7
Id. at 340.
8
Id. at 323.
9
See Richard K. Abrams & Michael W. Taylor, Issues in the Unification of
Financial Sector Supervision 22-23 (Int’l Monetary Fund, Working Paper No.
00/213, 2000), available at http://www.imf.org/external/pubs/ft/wp/2000/
wp00213.pdf. Within these two broad forms there is also scope for substantial
variation. See id. at 21-24.
10
For a general review of the background to the U.K.’s reforms, see Eilìs
Ferran, Examining the United Kingdom’s Experience in Adopting the Single
Regulator Model, 28 BROOK. J. INT’L L. 257 (2003).
11
MICHAEL TAYLOR, “TWIN PEAKS”: A REGULATORY STRUCTURE FOR THE
NEW CENTURY 10-11 (Ctr. for the Study of Financial Innovation) (1995).
12
SELECT COMMITTEE ON ECONOMIC AFFAIRS, BANKING SUPERVISION AND
REGULATION, 2008-9, H.L. 101-I, at 34.
(FSA).13 The radicalism of the U.K.’s approach should not be

underestimated. Not only did it involve the merger of nine pre-existing
regulatory agencies14 into one but, most controversially, it involved the
decision to remove the responsibility for bank regulation from the Bank of
England, the U.K.’s central bank, and to transfer it to the FSA.15 Although
unified regulators had been previously created elsewhere, most notably in
Scandinavia, none had involved the removal of bank regulation authority
from the central bank.16
Critics of the U.K.’s arrangements at the time of the FSA’s creation
charged that the separation of bank regulation from the central bank’s
lender of last resort (LoLR) responsibilities was highly risky. It was argued
that without the detailed institutional knowledge that derives from bank
regulatory authority, the Bank of England would be unable to perform its
LoLR responsibilities adequately. The subsequent experience of the run on
the British mortgage bank Northern Rock in September 2007 seemed to
confirm these critics. However, as this essay will argue, this conclusion
overlooks the range of possible alternatives to the U.K.’s reforms, and
particularly the Twin Peaks model. A “Twin Peaks” separation of
prudential (safety and soundness) and consumer protection regulation
would have offered a number of advantages over the FSA, including in
relation to crisis management arrangements. The essay concludes by
13
For an assessment of the FSA more generally, see Howell E. Jackson, An
American Perspective on the U.K. Financial Services Authority: Politics, Goals &
Regulatory Intensity (Harvard, John M. Olin Ctr. for Law, Econ. & Bus.,
Discussion Paper No. 522, 2005), available at http://www.law.harvard.edu/
programs/olin_center/papers/pdf/ Jackson_522.pdf.
14
Clive Briault, The Rationale for a Single National Financial Services
Regulator 6 (Fin. Servs. Auth., Occasional Paper No. 2, 1999), available at
http://www.fsa.gov.uk/pubs/ occpapers/OP02.pdf. These agencies include the
Securities and Investments Board, the Personal Investment Authority, the
Investment Management Regulatory Organisation, the Securities and Futures
Authority, the Supervision and Surveillance Division of the Bank of England, the
Building Societies Commission, the Insurance Directorate of the Department of
Trade and Industry, the Friendly Societies Commission, and the Registrar of
Friendly Societies. Id. at 6 n.1.
15
Id. at 7.
16
See Michael Taylor & Alex Fleming, Integrated Financial Supervision:
Lessons from Northern European Experience 17 (World Bank, Working Paper
2223, 1999), available at http://info.worldbank.org/etools/docs/library/
50180/TaylorFleming_1999.pdf.
drawing some conclusions from the British experience that might be

considered by policy-makers in the U.S.17
II. THE DEBATE IN THE UK PRIOR TO THE FSA
What is striking about the policy debate within the U.K. prior to
the formation of what became the FSA, is just how little attention was
given to the possibility of creating a single integrated financial regulator.
For several years prior to the election of a new Labour government in May
1997, there had been discussion of the need to reform the U.K.’s regulatory
system, but the ideas being debated stopped short of proposing to create a
single regulatory agency with a mandate that covered the entire banking,
insurance and investment industries.18 The concept only came to
prominence on May 20, 1997 with an announcement to the House of
Commons by the new Chancellor of the Exchequer that the government
intended to create a single regulatory authority for the banking and
securities industries. The announcement itself came as a surprise to many
observers and showed signs of having been rapidly prepared. This
impression arose not only because the statement was vague concerning
matters of detail, but also because it did not address some more
fundamental issues, such as whether the prudential regulation of insurance
companies would be included in the scope of the new regulator.19
Prior to this announcement, the British regulatory system combined
institutional and functional regulation, similar to the system created by the
17
For further discussion regarding the Twin Peaks model in a U.S. context,
see Heidi Mandanis Schooner, Regulating Risk Not Function, 66 U. CIN. L. REV.
441 (1998); Howell E. Jackson, A Pragmatic Approach to the Phased
Consolidation of Financial Regulation in the United States, (Harvard Law Sch.,
Working Paper No. 09-19, 2008), available at http://ssrn.com/abstract=1300431;
Cynthia Crawford Lichtenstein, The Fed’s New Model of Supervision for “Large
Complex Banking Organizations”: Coordinated Risk-Based Supervision of
Financial Multinationals for International Financial Stability (Boston College
Law Sch. Research Paper No. 89, 2006), available at http://ssrn.com/
abstract=882474.
18
Schooner & Taylor, supra note 2, at 320.
19
Ensuring the solvency of insurance companies had been the responsibility
of the Department of Trade and Industry although it was briefly transferred to HM
Treasury before the FSA was established. In July 1997, i.e. two months after the
original announcement, the government confirmed that this function would also
form part of the responsibilities of the new agency.
Gramm-Leach-Bliley Act in the United States.20 Banks were regulated by

the Bank of England under the Banking Act 198721 with respect to their
safety and soundness, while insurance companies were subject to solvency
regulation under the Insurance Companies Act of 198222 by a department
of the Treasury (a function which was previously discharged by the
Department of Trade and Industry). Sales practice (“conduct of business”)
regulation was in the hands of a network of self-regulating organisations
(SROs) which were also responsible for the safety and soundness
regulation of non-bank financial intermediaries such as securities brokers
and dealers and investment managers.
The SRO system was established by the Financial Services Act
23
1986 which had been described as “self-regulation within a statutory
framework” both by its chief architect24 and the Conservative government
that enacted it. This system had been designed to provide an all-
encompassing investor protection regime for securities, mutual funds, and
other forms of collective investment through a number of “Self-Regulating
Organizations” overseen by a quasi-governmental body, the Securities and
Investments Board (SIB).25 The SROs administered the sales practice
regime and were responsible for ensuring that the users of financial
services (generally speaking, securities brokering and dealing; futures
brokering and dealing; investment management; financial advice; and sales
practices relating to collective investment schemes like personal pensions
and life insurance) were subject to appropriate levels of consumer
protection. This system applied a functional approach to the regulation of
investment services, products, and advice. If a service or product was being
offered, it needed to be regulated by the relevant SRO, no matter what the
nature of the firm offering the service.
The Financial Services Act was initially administered by no fewer
than five separate SROs: The Securities Association (TSA) for Stock
20
Schooner & Taylor, supra note 2, at 324-25.
21
Banking Act, 1987, c. 22, § 1 (Eng.) (repealed 2001).
22
Insurance Companies Act, 1982, c. 50, § 3 (Eng.) (repealed 2001).
23
Financial Services Act, 1986, c. 60, § 8 (Eng.) (repealed 2001).
24
L.C.B. Gower, “Big Bang” and City Regulation, 51 MOD. L. REV. 1, 11
(1988).
25
The SIB exercised powers that were transferred to it under the Financial
Services Act by the Secretary of State for Trade and Industry (a government
minister). However, the SIB itself was in the unusual position of being a company
limited by guarantee and not a department of government. A similar structure was
subsequently adopted for the Financial Services Authority.
Exchange brokers and dealers; the Association of Futures Brokers and

Dealers (AFBD) for dealers in futures and options; the Investment
Management Regulatory Organisation (IMRO) for asset management and
mutual funds; the Life Assurance and Unit Trust Regulatory Organisation
(LAUTRO) for collective investment schemes marketed by insurance
companies; and the Financial Intermediaries, Managers and Brokers
Regulatory Association (FIMBRA) for independent financial advisers,
many of whom acted as agents of the insurance companies.26 During the
later years of the self-regulatory system’s existence some streamlining took
place: the TSA and AFBD merged, as did LAUTRO and FIMBRA, thus
reducing the number of SROs to three. Nonetheless the system was
criticized for its complexity and opacity to the consumer, especially as it
gave rise to what was described as an “ ‘alphabet soup’ of regulatory
agencies.”27 At the same time, the financial services industry criticised the
system for not being genuinely self-regulatory, and for imposing an
inappropriate regulatory burden on the interprofessional (“wholesale”)
money and capital markets. The SIB developed its own rulebook and
required the SROs to adopt “equivalent” standards. This resulted in a
lesser role for practitioner input and greater uniformity in the SRO
rulebooks than had originally been intended.28
However, the SRO system was most thoroughly discredited in the
eyes of opposition lawmakers by what became known as the “pensions
mis-selling scandal.”29 It had been the policy of the Conservative
government in the mid-1980s to encourage more personal provision for
retirement, rather than relying on occupational or state-provided schemes.
Approximately eight million personal pensions were sold in the UK
between 1988 and 1995.30 The SRO system was intended in part to provide
protection for individuals who entered into one of these personal savings
schemes; in the words of John Major (then a junior minister but later
26
See the account given by DAVID F. LOMAX, LONDON MARKETS AFTER THE
FINANCIAL SERVICES ACT 78 (1987).
27
Taylor, supra note 11, at 7.
28
These criticisms were recognized in a report issued by Andrew Large when
he assumed the Chairmanship of the SIB in 1993. ANDREW LARGE, FINANCIAL
SERVICES REGULATION: MAKING THE TWO TIER SYSTEM WORK (London:
Securities and Investments Board, 1993).
29
PETER CARTWRIGHT, CONSUMER PROTECTION IN FINANCIAL SERVICES 152
(1999).
30
Michael Taylor, The Policy Background in BLACKSTONE’S GUIDE TO THE
FINANCIAL SERVICES. & MARKETS ACT OF 2000 14 (Michael Blair ed., 2000).
Prime Minister) the Financial Services Act would “safeguard people

against the unscrupulous overselling of personal pensions.”31
Personal pension plans were mainly offered by insurance
companies (regulated by LAUTRO) which employed a sales staff with a
large commission element in their remuneration. By 1993 it had emerged
that a significant number of public sector employees, including teachers,
nurses and the employees of former state-owned industries such as coal
mining, had been encouraged by these salespeople to switch from their
occupational schemes to personal pension plans. As the employer-
provided plans were defined benefit, whereas the personal plans were
defined contribution, this arguably placed these individuals at a potentially
serious financial disadvantage.32 A report commissioned by the SIB
suggested that as many as 1.5 million pensions had been mis-sold with
compensation costs amounting to some ₤4 billion. Many of those affected
were a core constituency of the Labour party – public sector workers – and
hence the issue became highly politicised with the opposition party using it
as a stick with which to beat the government.33
Before winning the 1997 General Election Labour, party
spokesmen had committed the party to end what they termed “City self-
regulation.”34 One of the few definite policy commitments to emerge from
their pledge was the intention, once in government, to abolish the two-tier
system of SIB and SROs.35 In its place they undertook to establish a single,
statutory regulatory agency for securities and investments. Thus the
commitment to end City self-regulation might be narrowly construed as the
commitment to replace the system created by the Financial Services Act.
At the same time, however, there were indications that the Labour
party also considered the Bank of England to be part of the City’s “self-
regulatory” system, even though its powers to regulate banks derived from
31
Michael Taylor, Fin. Svcs..& Mkts. Act: The Policy Background-II, 31
AMICUS CURIAE 4, 6 (2000).
32
Whether the individuals were disadvantaged and to what extent depended
on a number of actuarial assumptions and assumptions about investment returns.
The intricacies of these issues were, however, drowned out in the subsequent
political debate.
33
See, e.g., 318, PARL. DEB., H.C. (6th ser.) (1998) 716, 718. (It continued to
be used by Labour ministers against their Conservative opposite numbers even
after the change of government).
34
Mike O’Brien, Labour's Proposals for Regulation into the 21st Century, 5
J. FIN. REG. & COMPLIANCE 115, 115-17 (1997).
35
LABOUR PARTY, LABOUR’S BUSINESS MANIFESTO (1997).
a separate statute (the Banking Act 1987) and even though, unlike the
SROs, it was a government agency. Labour suspicion of Britain’s central
bank ran deep, with some commentators suggesting that it can be traced to
the Bank’s role in the sterling crisis of 1931 that had helped to bring down
a minority Labour administration headed by Ramsay Macdonald. 36 This
fuelled Labour suspicions that the Bank of England was too closely aligned
with the Conservative party, in which the financial interests of the City of
London had a major influence. Thus when the British government
considered the introduction of statute-based bank regulation in the mid-
1970s some members of the governing Labour party proposed establishing
a banking commission independent of the Bank of England to exercise
regulatory powers. These proposals were rejected by the Cabinet after the
then Governor of the Bank of England fought a rearguard action to ensure
that it became the bank regulator.37 Nonetheless, in subsequent years the
Bank was to show itself a reluctant regulator which above all wished to
maintain its traditional, informal relationship with the leading financial
institutions in the City. Against this background it was possible to present
it as part of the City’s “self-regulating” system and as merely the chief
spokesman for a “cosy club.”
Nonetheless, it is doubtful whether the Bank of England’s
responsibility for regulating the banking sector would have come under
renewed scrutiny had it not been for two incidents in the first half of the
1990s. The first was the failure of the Bank of Credit and Commerce
International (BCCI), which went into insolvent liquidation once it became
clear that it had been a vehicle for a massive fraud.38 Although the bank
only had branches in the U.K. (its holding company was registered in
Luxembourg), the group’s “mind and management” had been in London
and hence there was a case for the Bank of England having taken the lead
in ensuring that the group as a whole was subject to consolidated
supervision. In a subsequent investigation conducted by Sir Thomas (later
Lord Justice) Bingham, a senior judge, the Bank was found to have adopted
36
MICHAEL MORAN, THE POLITICS OF BANKING: THE STRANGE CASE OF
COMPETITION & CREDIT CONTROL 120 (2d ed., 1986).
37
Id.
38
H.M. STATIONARY OFFICE, INQUIRY INTO THE SUPERVISION OF THE BANK
OF CREDIT & COMMERCE INT’L. (1992).
an excessively narrow interpretation of its powers and, partly for that

reason, it had not been sufficiently proactive in regulating BCCI.39
In BCCI’s case, the Bank of England could argue in its defence
that it was a bank regulator not a fraud investigator. No such defence was
available in relation to the second episode – the failure of Barings merchant
bank in early 1995.40 Barings had been part of the City of London’s
“aristocracy,” a centuries old merchant bank that had for generations been
at the heart of the City’s establishment to the extent of providing several
Governors of the Bank of England. Barings had failed once before, in
1890, as the result of speculation in railroad construction in South
America.41 It had been then bailed out by the Bank of England, at that
time still a privately owned corporation.42 One hundred and five years
later, Barings failed again, this time due to the poorly controlled activities
of a futures trader based in Singapore who took large unhedged positions in
the Singapore and Osaka futures exchanges.43 This was the first of several
episodes involving what came to be called “rogue traders” in the years that
followed.44 The episode was damaging to the Bank of England since it
appeared that Barings had enjoyed a relatively light touch regulatory
regime and thus provided an illustration of the operation of a so-called
“self-regulatory system,” at least as far as it applied to members of the
City’s establishment.45
39
212 PARL. DEB., H.C. (6th ser.) (1992) 575. According to a statement
given to the House of Commons by the then Chancellor of the Exchequer Norman
Lamont, the Bingham report “argues that the Bank was slow to impose on BCCI
an appropriate supervisory regime, and concludes that the Bank continued for too
long to rely on the Luxembourg authorities to play the leading role.” Id.
40
H.M. STATIONARY OFFICE, REPORT OF THE BOARD OF BANKING
SUPERVISION INQUIRY INTO THE CIRCUMSTANCES OF THE COLLAPSE OF BARINGS
(1995).
41
See JOHN GAPPER & NICHOLAS DENTON, ALL THAT GLITTERS: THE FALL
OF BARINGS 2 (1996).
42
Id.
43
Id. at 28-29.
44
A phrase that was originally coined to describe the Barings trader, Nick
Leeson, which he used as the title of his subsequent book: NICK LEESON &
EDWARD WHITLEY, ROGUE TRADER: HOW I BROUGHT DOWN BARINGS BANK AND
SHOOK THE FINANCIAL WORLD (1996).
45
See Gordon Brown, Ch. of the Exch., Statement to the H.C. on the Bank of
Eng. (May 20, 1997), (“SIB will become the single regulator underpinned by
statute. The current system of self-regulation will be replaced by a new and fully
statutory system, which will put the public interest first, and increase public
Despite these episodes, Labour entered office in 1997 without a

clear commitment to removing the Bank of England’s responsibility for
bank regulation. Nor was there any indication of the possibility that a
single financial regulator was on the policy agenda. What changed this
situation was the new government’s announcement in its first few days in
office that it would grant the Bank of England independence to set interest
rates. Although this policy was not featured in the Labour party’s
manifesto, central bank independence had been debated extensively in
Britain since the early part of the decade.46 On occasion in this debate the
question of the central bank’s regulatory powers had arisen without,
however, any definitive conclusion being reached. Nonetheless, once the
decision was taken to create an independent central bank, a new Bank of
England Act was required and this seems to have provided the pretext for a
re-examination of the Bank’s role as bank supervisor.47
The decision to remove banking supervision from the Bank of
England appears to have been taken opportunistically. Before the start of
each parliamentary year in Britain, each government department must put
in “bids” for parliamentary time for the passage of legislation that it
considers essential. The successful bids are then included in the
government’s annual legislative program announced to parliament in the
“Queen’s Speech.” In 1997 Treasury ministers wished to introduce two
major bills – one to grant the Bank of England its independence, the other
to abolish the “two tier” system of SIB and SROs created by the Financial
Services Act. However, the new government had an ambitious policy
agenda and a crowded legislative timetable, resulting in the Treasury being
granted the time for only one major bill. According to the director of the
Association of British Insurers, speaking the year after the event:
The Treasury team had failed to secure in the first Queen’s

Speech legislation to abolish the two tier system under the
Financial Services Act and Markets Act. However, a
separate decision had been taken to give the Bank of
England independence in respect of conducting monetary
confidence in the system.” (From the context it appeared that he considered the
Bank of England to be part of the self-regulatory system). Id.
46
See Michael Taylor, Central Bank Independence.: The Policy Background,
in BLACKSTONE’S GUIDE TO THE FINANCIAL SERVICES AND MARKETS ACT OF 2000
at 10 (Michael Blair ed., 1998).
47
See Ferran, supra note 10, at 271-72.
policy and this did require legislation. It seems that an

opportunist decision was taken at this stage to move
towards a single regulator because the legislation to give
the Bank of England independence in respect of monetary
policy could be used for any other purpose relevant to the
Bank of England.48
One of these “other purposes” was the transfer of responsibility for

bank regulation from the Bank of England to the SIB, which then became
the nucleus of the FSA. In other words, the parliamentary timetable rather
than a reasoned policy debate seems to have triggered the decision to move
to a single regulator. This also would have been consistent with an
apparent about-turn in government policy after the Governor of the Bank of
England apparently had been assured there were no immediate plans to
strip the Bank of its bank regulatory function.
There have also been allegations that the concept of a single
financial regulator had been developed within the Treasury before the
change of government and had been inspired as much by Treasury rivalry
with the Bank as by any policy considerations.49 It was certainly the case
that government ministers saw a single financial services regulator as an
alternative centre of power to the Bank and hoped that the FSA would
assume the Bank’s role as overseer of the City’s interests. Since the
government was committed to establishing Bank of England independence
in respect of monetary policy, it is also possible that removing its banking
supervision function was seen as a way of preventing it from becoming “an
over-mighty subject.” Whatever the exact motivation, it is clear that the
momentous implications of an opportunistic and essentially political
decision may not have been fully appreciated by government ministers who
were still new to power after their party’s unusually long period out of
office.50
48
See Ferran, supra note 10, at 271.
49
See SIR MARTIN JACOMB, RE-EMPOWER THE BANK OF ENG. 2-4 (Centre for
Policy Stud.) (2009) available at http://www.cps.org.uk/cps_catalog/Re-
empower%20the %20Bank%20of%20England.pdf.
50
It is important in this regard that the Labour party had been out of power
for 18 years and few of its lawmakers had experience of government.
III. JUSTIFYING THE SINGLE REGULATOR
The decision to create a single financial regulator had to be

justified after the fact. Government ministers and the FSA itself put
forward a series of justifications for the creation of a single regulator. They
fell into two broad categories: those relating to market developments and
those relating to the purported effectiveness and efficiency of a single
regulatory agency.
The argument that market developments justified a single financial
regulator became known as the “blurring the boundaries” argument. In his
statement to the House of Commons on May 20, 1997, Britain’s Chancellor
of the Exchequer argued that:
At the same time, it is clear that the distinctions

between different types of financial institution--banks,
securities firms and insurance companies--are becoming
increasingly blurred. Many of today's financial institutions
are regulated by a plethora of different supervisors. This
increases the cost and reduces the effectiveness of the
supervision.
There is therefore a strong case in principle for
bringing the regulation of banking, securities and insurance
together under one roof. Firms now organise and manage
their businesses on a group-wide basis. Regulators need to
look at them in a consistent way. That would bring the
regulatory structure closer into line with today's
increasingly integrated financial markets. It would deliver
more effective and efficient supervision, giving both firms
and customers better value for money, and would improve
the competitiveness of the sector and create a regulatory
regime to genuinely meet the challenges of the 21st
century.51
The argument was further developed in a document issued by the

Treasury the following year:
The existing arrangements for financial regulation involve

a large number of regulators, each responsible for different
51
294 PARL. DEB., H.C. (6th ser.) (1997) 510.
parts of the industry. In recent years there has been a

blurring of the distinctions between different kinds of
financial services business: banks, building societies,
investment firms, insurance companies and others. This
has added further to the complexity of financial regulation.
The Government believes the current system is costly,
inefficient and confusing for both regulated firms and their
customers. It is not delivering a standard of supervision
and investor protection that the public has a right to expect.
We are therefore establishing a single, statutory regulator
for the UK financial services industry with clearly defined
regulatory objectives and a single set of coherent functions
and powers.52
However, it was left to the FSA itself to provide the most extensive
justification for its own existence. While the FSA was still under
construction, it published a paper written by one of its own officials, Clive
Briault, who set out to defend the single regulator concept.53 He did so by
first invoking the “blurring of boundaries” argument:
The securitisation of traditional forms of credit

(including mortgages, credit card outstandings and
commercial loans) and, with the growth of options,
increasingly elaborate ways of unbundling, repackaging
and trading risks, have weakened the distinction between
equity, debt and loans, and even between banking and
insurance business (where, for example, credit derivatives
bear many of the characteristics of an insurance product
and insurance companies offer short-term deposit-like
products).54
This development, Briault explained, had an important consumer protection

dimension in that the disappearance of a neat conjunction between a
particular type of firm and a limited range of products being supplied by
that firm means that it is difficult to regulate on a functional basis, since the
52
H.M. STATIONARY OFFICE, FIN. SVCS. & MKTS. BILL: A CONSULTATION
DOCUMENT, pt. 1 (1998).
53
See Briault, supra note 14.
54
Id. at 13-14.
traditional functional approach no longer matches the structure of either

firms or markets.55
Accordingly, a single financial regulator was essential to provide
adequate consumer protection when financial products could no longer be
neatly slotted into the traditional contractual forms which have underpinned
the functional approach to regulation.56 Trying to regulate the sale and
marketing of products on a functional basis would result in inadequate
consumer protection, either because similar products would become subject
to different levels of consumer protection or the regulatory agencies
disputed jurisdiction over certain types of product.57
The blurring the boundaries argument also related to the formation
of financial conglomerate groups. The emergence of financial
conglomerates (usually defined as a group which undertakes at least two of
the activities of banking, securities or insurance) resulted from mergers and
acquisitions that occurred most frequently between banks and securities
firms and between banks and insurance companies.58 In some cases they
also involved the purchase of fund managers by banks and by insurance
companies.59 These combinations were permitted as the result of the
dismantling of structural barriers – which in the U.K. had been mainly
informal and non-statute based – in the course of the 1980s.60 In response
to these and similar developments elsewhere in the G10, the Tripartite
Group of banking, securities, and insurance supervisors argued in a 1995
report that a “group-wide” perspective was required to obtain an adequate
supervisory overview of these financial conglomerates.61 Nonetheless, as
long as regulation remained structured along traditional
institutional/functional lines, obtaining such a group-wide perspective
would be difficult.
The British solution was to adopt the lead regulator concept.62 The
lead regulator would be responsible for taking a consolidated view of the
55
Id. at 14.
56
See id.
57
See id.
58
Id. at 12-13.
59
Briault, supra, note 14, at 13.
60
TRIPARTITE GROUP OF BANK SEC. & INS. REG., THE SUPERVISION OF FIN.
CONGLOMERATES at i (1995). Subsequently the Tripartite Group was renamed the
Joint Forum.
61
Id. at i-ii.
62
See GEORGE ALEXANDER WALKER, INT’L BANKING REGULATION: LAW,
POLICY AND PRACTICE 258 (Kluwer Law Int’l) (2001).
capital adequacy and liquidity of the consolidated group; taking a similarly

group-wide view of more qualitative factors such as the calibre of senior
management and the high-level systems and controls of the financial
conglomerate; and co-ordinating and encouraging the exchange of
information among the relevant regulatory bodies, both routinely and in the
event of an emergency. Typically, since most such groups were headed by
a bank, the Bank of England usually assumed this responsibility ,which was
similar to the Fed’s umbrella supervisor role created by Gramm-Leach-
Bliley.63 In contrast to the U.S. arrangements, however, the Bank of
England’s role was largely extra-statutory and was the result of a
framework of Memorandums of Understanding (MoUs) between the Bank
of England and the functional regulators.
Although Briault claimed that the lead regulator concept had
worked well, he nonetheless stressed that countries that had moved towards
a single regulator had “done so in part because, with the growth in the
number of multiple-function firms, the need for communication, co-
ordination, co-operation and consistency across specialist regulatory bodies
had become increasingly acute and increasingly difficult to manage
efficiently.”64 If such firms were the rule rather than the exception (in
contrast to the situation in the past) then new institutional arrangements
were required to ensure that that they were subject to more efficient
oversight. Briault cited statistics to show that many firms were now subject
to multiple regulators: eight firms (including HSBC, Halifax, Abbey
National and the Royal Bank of Scotland) were authorised to conduct all
five of the main regulated activities (“deposit-taking, insurance, securities
and corporate finance, fund management, and advising on or selling
investment products to retail customers”).65 A further 13 firms were
authorised to conduct four of these activities, and more than 50 other firms
were authorised for three of these five functions.66
The efficient supervision of financial conglomerates was only one
dimension of the superior efficiency claimed for the single regulator. It
was also argued that it would allow scarce supervisory resources to be
deployed more effectively; an example concerned the development of
specialist teams to review firms’ internal risk management models that had
become an integral part of regulation during the 1990s. In the pre-FSA
63
Schooner & Taylor, supra note 2.
64
Briault, supra note 14, at 15.
65
66
Id.
system, several different regulators had needed to build their own specialist
model review teams, but individuals with the requisite skills were in high
demand which made it difficult for regulatory agencies to recruit them in
sufficient numbers.67 By centralizing the available resources, a single
regulator seemed to offer a way out of this impasse. Similarly, it was also
argued that the creation of a single support infrastructure (e.g. IT system)
would lead to significant cost savings as the duplication and overlap
resulting from the nine pre-existing regulators was eliminated. The
argument that a single regulator would be more cost effective was vital in
selling the concept to the financial industry. It was therefore not surprising
that Briault made much of this argument:
Economies of scale and scope should arise because a single

regulator can take advantage of a single set of central
support services (human resources, information services,
financial control, premises etc); introduce a unified
statistical reporting system for regulated firms; operate a
single database for the authorisation of firms and the
approval/registration of individuals; avoid unnecessary
duplication or underlap across multiple specialised
regulators; introduce a consolidated set of rules and
guidance; tackle problems of co-ordination, co-operation
and communication more effectively within a single entity
and under a unified management structure than might be
possible across separate specialist entities; offer a single
point of contact to both regulated firms and to consumers
(through a single complaints handling regime and a single
compensation scheme); and adopt a more effective and
focused approach to areas of common interest to most
regulated financial activities (for example, handling Year
2000 issues and turbulence in international financial
markets).68
These arguments – consumer protection arrangements that were

better suited to the characteristics of new financial instruments, improved
oversight of financial conglomerate groups, and cost savings and
efficiencies from a common regulatory platform – were at the heart of the
67
See Taylor, supra note 11, at 6.
68
case constructed for the single financial regulator. The difficulty was that
exactly these same arguments had been made in favor of an alternative
regulatory structure – the so-called Twin Peaks model. It was therefore
also necessary for the defenders of the single regulator to explain why this
structure would be superior to the Twin Peaks alternative.
IV. THE REJECTION OF THE TWIN PEAKS ALTERNATIVE
Unlike the single regulator, a Twin Peaks structure had been

actively debated in the U.K. prior to the 1997 reform, and it had attracted
support from a number of influential figures both in the industry and in
regulation.69 It was, however, strongly opposed by the Bank of England
which regarded the proposals as primarily an attempt to divest it of its
regulatory responsibilities.
Twin Peaks proposed that, instead of being structured around the
traditional tripartite distinction of banking, securities and insurance, the
institutional structure of regulation should in future comprise two
regulatory agencies, a Financial Stability Commission and a Consumer
Protection Commission.70 The first would be responsible for ensuring the
stability of the financial system as a whole, mainly through the application
of prudential regulations.71 The second would be charged with ensuring
that firms deal with their (retail) customers in a fair and transparent
manner.72 The two Commissions would be responsible for discharging
their mandate irrespective of the legal form of the firms that they
regulated.73
The source of the “blurring the boundaries” argument can be traced
to the Twin Peaks proposals, which also placed heavy emphasis on the
need to ensure proper group-wide supervision of financial conglomerates.74
69
See Jill Treanor, Regulators Back Taylor’s Twin Peaks Theory, THE INDEP.,
Oct. 29, 1996, available at http://www.independent.co.uk/news/business/
regulators-back-taylors-twinpeaks-theory-1360780.html.
70
71
Id.
72
Id.
73
See id.
74
Id. at 4. However, the earliest occurrence of this argument is to be found in
a working paper published the year before Twin Peaks. See Claudio E.V. Borrio &
Renata Filosa, The Changing Borders of Banking: Trends and Implications, 3, 16
(Bank for Int’l Settlements, Working Paper No. 23, 1994), available at
http://www.bis.org/publ/ work23.pdf?noframes=1.
The case for Twin Peaks also invoked the economies of scale that would
result from (the admittedly more limited) regulatory consolidation that it
also involved. Thus, because the arguments for a single regulator and for
Twin Peaks were almost identical, it was necessary for the FSA’s defenders
to show that theirs was the superior solution. The crux of the argument
concerned the separation of prudential and conduct of business regulation
that was the main feature of the Twin Peaks model; the defenders of a
single regulator argued that the separation was not so clear cut as the Twin
Peaks model presupposed.75
The first strand of this argument was to contest the claim, central to
the Twin Peaks analysis, that there were two relatively, clearly
distinguishable regulatory objectives – financial stability on the one hand
and consumer protection on the other. This case for treating these two
objectives as interlinked is well summarized by Davies and Green:
The ultimate argument for financially sound and

prudentially well regulated financial institutions is that they
are then able to provide financial services and investment
opportunities to consumer and businesses which those
customers may use with confidence. A breakdown in
consumer protection, whether in banking, investment or
insurance products, may itself precipitate a wider loss of
confidence in types of product or firms. There is therefore
no necessary conflict between the two aims of regulation. In
the long run they are aligned.76
Closely related to this was the claim further claim that, in practice,
prudential and conduct of business (sales practice) regulation required
examination of very similar issues, and therefore that there would be
significant overlap between the Twin Peaks agencies.77 Briault put the
point with characteristic clarity:
[T]here is a considerable overlap – both

conceptually and in practice – between prudential and
conduct of business regulation. Both have a close and
75
See Briault, supra note 14, at 25.
76
HOWARD DAVIES & DAVID GREEN, GLOBAL FIN. REGULATION: THE
ESSENTIAL GUIDE 192 (2008).
77
legitimate interest in the senior management of any

financial institution subject to both of these types of
regulation, in particular because of the crucial roles of
senior management in setting the “compliance culture” of a
firm, in ensuring that management responsibilities are
properly allocated and cover comprehensively the business
of the firm, and in ensuring that other internal systems and
controls are in place. The detail of some of these systems
and controls may indeed be specific to either prudential or
conduct of business considerations, but many of them will
be more general.78
In short, a single regulator was superior to a Twin Peaks structure

because many of the same supervisory judgments would arise in
considering prudential and sales practice regulation. There seemed little
point in having two regulators reaching essentially duplicate judgments of
broadly similar matters. Since there is substantial overlap between the two
regulatory objectives and, in practice, prudential and conduct of business
regulation will focus on the same fundamental issues, they were best
administered by a single regulatory agency.
The Global Financial Crisis has created a very different perspective
on this argument. The British Government’s own White Paper on
regulatory reform after the crisis has concluded that the system places too
much weight on “ensuring that systems and processes were correctly
defined rather than on challenging business models and strategies” and on
“conduct of business regulation of the banking sector rather than prudential
regulation of banking institutions.”79 Even the FSA’s senior management
has acknowledged that the agency neglected prudential supervision.80 In
the words of the report on the banking crisis produced by the FSA’s current
chairman, Lord Turner, the agency’s regulatory practices resulted in “[a]
balance between conduct of business and prudential regulation which, with
the benefit of hindsight, now appears biased towards the former.”81 Turner
repeated this admission to a committee of the British House of Lords which
78
Id.
79
H.M. TREASURY, REFORMING FINANCIAL MARKETS, 2009, Cm. 7667, at 56.
80
See FIN. SERV. AUTH., THE TURNER REVIEW: A REGULATORY RESPONSE
TO THE GLOBAL BANKING CRISIS 87 (2009), available at
http://www.fsa.gov.uk/pubs/ other/ turner_review.pdf.
81
Id.
referred in its final report to the “widely held perception that, in recent
years, the FSA has emphasized conduct-of-business supervision at the
expense of prudential supervision.”82
This situation was especially apparent in the FSA’s supervision of
the mortgage bank Northern Rock which was the first British casualty of
the crisis.83 The bank had received numerous contacts from the FSA
concerning a consumer protection initiative (“treating customers fairly”),
but supervision of capital and liquidity had been deficient and the bank had
been placed in a category that subjected it to one major prudential meeting
once every three years. The FSA’s own report on Northern Rock stated
that “some of the fundamentals of work on assessing risks in firms (notably
some of the core elements related to prudential supervision, such as
liquidity) have been squeezed out.”84
The House of Lords Committee thought it could identify the reason
why the FSA had failed to give sufficient attention to prudential regulation:
Conduct-of-business is important and politically sensitive,

and its results are easy to measure. In contrast, prudential
supervision, while arguably more important, is conducted
privately; its success is less easily measured, and, most of
the time, it has a lower political impact than conduct-of-
business supervision though in times of crisis such as the
present its political impact, its effect on businesses,
individuals and the economy, is very much greater than
conduct-of-business supervision. It is natural and rational
for a supervisor with responsibility for both activities to
concentrate on the one with the greater immediate political
sensitivity.85
In other words, the argument that there were synergies between prudential
and conduct of business regulation overlooked the distinct possibility that
82
83
FINANCIAL SERVICES AUTHORITY, THE SUPERVISION OF NORTHERN ROCK:
A LESSONS LEARNED REVIEW ¶¶ 8, 27 (2008), available at http://www.fsa.gov.uk/
pubs/other/ nr_report.pdf.
84
Id. at ¶ 36.
85
one type of regulation would come to dominate within a single regulator

and that this would likely to be consumer protection given the realities of
the political process.86 Twin Peaks had predicted that this outcome was
likely, and used it as one of the arguments against creating a single
regulatory agency.87
Thus the argument that there was a natural synergy between
prudential and consumer protection regulation has been discredited by
events both before and during the crisis. If the purported synergies were
really as strong as was claimed, then the multiple reviews of Northern
Rock’s systems for handling consumer issues should have thrown out
evidence that the bank’s business strategy was dangerously flawed. They
did not. Nor is this outcome really surprising. Although there may be
some overlap of the relevant judgments at the margin, they ultimately
involve quite fundamentally different matters. Weaknesses of internal
control systems for dealing with consumer issues may be indicative of
more general weaknesses in internal control within the institution as a
whole, and this could indeed raise matters of prudential concern. But it is
doubtful that these kinds of findings will demonstrate that the bank’s
management is following a deeply flawed and highly risky business
strategy which is likely to end in failure. To reach this conclusion it is
necessary to ask different questions to those a consumer protection
regulator might ask.
V. THE ROLE OF THE CENTRAL BANK
One dimension of the Twin Peaks structure that had been actively
debated before the decision to create the FSA was the role of the central
bank. In a number of speeches and articles, the Bank of England’s senior
management defended the Bank’s role as a bank regulator against the
proposed Twin Peaks structure.88 The central bank, it was argued, needed
to be concerned with the financial condition of the banking system, as this
was the conduit through which its monetary policy was transmitted to the
wider economy. As Governor Eddie George argued in a speech given in
1994, before the Twin Peaks debate began, the soundness of banks and the
central bank’s ability to conduct monetary policy were intimately related:
86
See id.
87
88
Howard Davies, Financial Regulation: Why, How and By Whom, in BANK
OF ENG. Q’LY BULL. 107, 111 (1997).
Monetary and financial stability are inter-related. It

is inconceivable that the monetary authorities could quietly
pursue their stability-oriented monetary policy objectives if
the financial system through which policy is carried on –
and which provides the link with the real economy – were
collapsing around their ears. The liabilities of banks in
particular are money, and you cannot be concerned with
preserving the value of money without being concerned
also with preserving public confidence in money in this
broader sense. Equally though, the financial system is
much less likely to be collapsing around the ears of the
monetary authorities in an environment of macro-economic
stability than in one of exaggerated boom and bust and
volatile asset values. This inter-relationship means that,
whatever the precise institutional arrangements for
financial regulation and supervision, central banks
necessarily have a vital interest in the soundness of the
financial system.89
Moreover, banks were a “special” type of financial intermediary:

as Sir Howard Davies, at that time still the Deputy Governor of the Bank of
England,90 said in early 1997 “in our view, there is still a reasonably clear
distinction to be made between banks and other financial institutions, and
their prudential soundness, or lack of it, can have rather different
implications for the rest of the market.” As a result, he continued,
Of course it may be argued that the distinctive

characteristics of banks, and their potential to create
systemic risk—which central banks can counteract—does
not necessarily mean that the central bank should act as
their regulator. I agree. But there are significant synergies
to be had from maintaining an institutional link between
the two functions, and the burden of proof rests, I think,
89
E. A. J. George, Governor, Bank of England, The Bank of England –
Objectives and Activities, The Capital Market Research Institute, Frankfurt,
Germany (Dec. 5, 1994).
90
Shortly afterwards he was appointed the first Chairman and Chief
Executive of the FSA. His views on the specialness of banks underwent a
subsequent change.
with those who wish to make the case for disturbing that
relationship.91
The main “synergy” that arose from retaining banking supervision within
the central bank was with the Bank’s role as lender of last resort (LoLR). It
was argued that the information acquired in the capacity of the bank
supervisor was essential to the central bank performing the lender of last
resort function, and that therefore the best arrangement was for LoLR and
banking supervision to be located in the same institution. Following the
Northern Rock experience, a number of commentators have reached the
conclusion that this argument was correct. As Professor Willem Buiter
argued in evidence to the House of Commons Treasury Select Committee:
The notion that the institution that has the

knowledge of the individual banks that may or may not be
in trouble would be a different institution from the one that
has the money, the resources, to act upon the observation
that a particular bank needs lender of last resort support is
risky. It is possible, if you are lucky, to manage it, but it is
an invitation to disaster, to delay, and to wrong decisions.
The key implication of that is that the same institution—it
could be the FSA or it could be the Bank of England—
should have both the individual, specific information and
the money to do something about it.92
Against these arguments, proponents of separation argued that

theoretical considerations and empirical evidence indicated that central
banks with banking supervisory responsibilities tended to err on the side of
laxity in monetary policy; as Goodhart and Schoenmaker argued in a
widely cited paper, monetary policy aimed to be countercyclical, whereas
regulatory policy was pro-cyclical.93 Concerns were also expressed that
banking supervision “failures” – which it was generally accepted were
91
DAVIES, supra note 88, at 110.
92
Id.
93
C. A. E. Goodhart & D. Schoenmaker, The Institutional Separation
Between Supervisory and Monetary Agencies, in THE CENTRAL BANK AND THE
FINANCIAL SYSTEM 341 (1995)..
almost inevitable – would damage the reputation and credibility of the

central bank as a monetary policy institution.94
Twin Peaks further argued that “[a]s the once-clear demarcation
lines between types of financial markets and institutions are broken down,
the Bank’s role appears increasingly anomalous.”95 In other words, owing
to the changing nature of the financial system, banks could no longer be
considered the unique source of systemic risk that traditionalists insisted
that they remain. In consequence of these developments, it became
increasingly difficult to argue that banks were “special” in the sense that
they were uniquely, systemically important.96 Many large non-banks were
now “too interconnected to fail,” a phrase that was coined when Bear
Stearns teetered on the brink of failure in March 2008.97 On the traditional
view, Bear Stearns would not have been considered systemically important;
however, the episode confirmed the argument of Twin Peaks that “the rise
of the OTC markets means that we must extend our concept of what
constitutes a systemically important firm.”98
Yet if the concept of a systemically important firm was extended in
this way, it was by no means obvious that the central bank was the right
institution to regulate these firms. Twin Peaks acknowledged the
possibility that the functions of the FSC could be performed by the central
bank and that LoLR was an important issue.99 However, on balance it
rejected the case for the central bank also performing the role of prudential
regulator of the new, broader category of systemically important firms.100
In the first place, a broader concept of systemically important firms meant
that the central bank would need to interact with various institutions that
were not its traditional counterparties (a prediction that has come to pass
following the Federal Reserve’s expansion of its facilities in the wake of
the financial crisis).101 Secondly, the expertise necessary to regulate
investment banks and insurance companies does not naturally reside in
central banks.102 As Twin Peaks identified, a major problem for central
banks is in finding a place for such regulatory specialists in organizations
94
Id. at 341
95
Taylor, supra note 11, at 13-14.
96
Id. at 4.
97
Id.
98
Id. at 5.
99
Id. at 14.
100
Id. at 13-14.
101
102
Id. at 6, 12.
where they will have few opportunities for career progression.103

Nonetheless, Twin Peaks also recognized that close links would be needed
between the central bank and the FSC.104 Although it was comparatively
sketchy about the nature of those links, apart from proposing overlapping
board membership, the need for close coordination between the central
bank and the prudential regulator was an important component of the Twin
Peaks structure.105
The FSA’s relationship with the Bank of England was, in theory,
also to be a close one.106 Yet when the FSA was established, very little
attention was given to the need for institutional linkages between the
regulator and the central bank.107 Instead, given the prominent role played
by ex-Bank of England staff in the FSA, the professional relationships
between former colleagues were supposed to guarantee cooperation
between the two institutions. However, once this generation of officials
had either retired or left the FSA, there was no institutional mechanism to
ensure close collaboration between the two institutions. More recently, the
British government has announced the formation of a Financial Stability
Council which can be seen as a belated attempt to build the stronger
institutional linkages between the Bank and the FSA that were required
from the outset.
Briault acknowledged that in the “multi-faceted” relationship
between Bank and FSA, close cooperation and regular information flows
would be essential. These would need to occur both routinely for those
aspects of financial stability in which the central bank has an interest for
the setting of monetary policy, and in exceptional circumstances for more
specific and detailed information relating to the position of financial
institutions for whom support operations are being considered (where the
fiscal authority is also likely to have a close interest). “The UK
Memorandum of Understanding… provides an important underpinning to
the necessary exchange of such information under the new arrangements in
the UK.”108
The Memorandum of Understanding to which Briault refers was
between the Treasury, Bank of England, and FSA and it supposedly created
103
Id. at 12.
104
Id. at 14.
105
Id.
106
Id. at 13-14.
107
108
the framework for both information exchange and for crisis management.109
These are referred to under the MoU as the “Tripartite Authorities” and the
Bank of England’s responsibilities are summarised as contributing “to the
maintenance of the stability of the financial system as a whole.”110 The
FSA has the responsibility of authorising and supervising individual
banks.111 HM Treasury has responsibility for the institutional structure of
the financial regulatory system, and the legislation behind it.112 In a crisis,
the Financial Services Authority would, according to the Memorandum of
Understanding, be responsible for monitoring “the health of institutions
that fall within its regulatory remit” and for ensuring, “as far as is
appropriate in the circumstances, continuing compliance with regulatory
standards.”113 However, the Bank of England would remain in charge of
“official financial operations … in order to limit the risk of problems in or
affecting particular institutions spreading to other parts of the financial
system”.114
The MoU also established a Joint Crisis Management Committee,
chaired by the Chancellor, for dealing with what the MoU referred to
generically as “support operations.”115 It did not, however, clearly
distinguish between those operations that relate to emergency liquidity
assistance and those that would involve solvency support.116 In both cases
the Treasury sat at the apex of a pyramid with both the Bank and FSA in
subordinate roles.117 This contrasts with the practice of most other
countries in crisis management, which is to ensure that as long as the issue
remains one of liquidity the central bank will be in the lead.118 It alone has
(or should have) the information and the ability to react sufficiently
109
GORDON BROWN ET AL., MEMORANDUM OF UNDERSTANDING BETWEEN
H.M. TREASURY, THE BANK OF ENGLAND AND THE FINANCIAL SERVICES
AUTHORITY 4-5 (2009), available at http://www.bankofengland.co.uk/
financialstability/mou.pdf [hereinafter MoU].
110
Id. ¶ 2.
111
Id. ¶ 3(i).
112
Id. ¶ 4(i).
113
Id. ¶ 17(iii).
114
Id. ¶ 2(iv).
115
GORDON BROWN ET AL., supra note 109, at ¶ 14.
116
Id. ¶ 17(iii).
117
Id. at ¶ 4, 10, 13.
118
Dong He, “Emergency Liquidity Support Facilities”, Appendix in
CHARLES ENOCH et al (eds) BUILDING STRONG BANKS THROUGH
SURVEILLANCE AND RESOLUTION (IMF 2002), p.134
promptly to emerging problems.119 In this case the FSA’s role would be

clearly established as that of a handmaiden to the Bank, under an explicit
obligation to provide it with any and all information required by for the
discharge of its duties.120 Only in the event that the issue becomes one of
providing solvency support should the Treasury have taken the lead, with
both the Bank and the FSA in supporting roles.121 The subordinate role to
which the Bank was assigned in the MoU provides support to those who
argue that the post-1997 arrangements were designed to reduce the Bank’s
status.
In practice, however, the arrangements envisaged by the MoU were
rarely tried in practice and the Joint Crisis Management Committee rarely
met. The House of Commons Treasury Select Committee, in reviewing the
Northern Rock experience, concluded that “in terms of information
exchange between the Tripartite authorities, the system might have ensured
that all the Tripartite authorities were fully informed. However, for a run on
a bank to have occurred in the United Kingdom is unacceptable, and
represents a significant failure of the Tripartite system.”122
VI. LESSONS OF THE BRITISH EXPERIENCE FOR UNITED

STATES REGULATORY REFORM
Reviewing the lessons of the British experience, and considering

the current regulatory reform debate in the United States, I offer the
following conclusions:
A. THE CONCEPT OF A SINGLE REGULATOR HAS NOT BEEN

DISCREDITED, BUT ITS LIMITATIONS HAVE BEEN EXPOSED
The U.K. was not the first country to establish a unified regulatory
agency outside the central bank: that honor belongs to the Scandinavian
countries, with Norway (1986) as the pioneer followed by Denmark (1988)
and Sweden (1991).123 In these countries an important consideration was
119
Id. at ¶ 2, 4.
120
Id. at ¶ 6.
121
GORDON BROWN ET AL., supra note 109, at ¶ 14.
122
HOUSE OF COMMONS TREASURY COMMITTEE, THE RUN ON THE ROCK,
2007-08, H.C. 56-1, at 107.
123
Taylor & Fleming, supra note 16, at 6-7.
the “economies of scale” argument.124 As relatively small countries with

financial systems dominated by a small number of financial conglomerate
groups, combining all regulatory functions within a single agency appeared
to offer numerous efficiency benefits.125 Moreover, since the central bank
was not involved in banking supervision in any of these countries, the
powerful – and sometimes emotive – issue of the central bank’s powers did
not arise.126
There continues to be a case for single regulatory agencies in
comparatively small countries where the economies of scale gains are
significant.127 It is expensive to establish regulatory agencies with their
associated support services and infrastructure, and therefore minimizing
overhead costs is a worthwhile ambition. However, in larger countries,
especially those with a large and complex financial system, any potential
efficiency gains are far outweighed by the inefficiencies of combining too
many regulatory functions in a single agency. As noted above, the FSA has
struggled with the combination of prudential and conduct of business
regulation and its past performance suggests that it was simply tasked with
too many functions to perform all of them adequately.
B. PRUDENTIAL AND CONDUCT OF BUSINESS REGULATION DON’T

MIX
Despite the claims of the FSA’s supporters that there are

substantial synergies between prudential and conduct of business
regulation, the crisis has shown the limits of these synergies. While some
of the relevant supervisory judgements do overlap, especially on such
matters as internal controls and the probity of management, prudential
regulation needs a different focus. The factors influencing the financial
soundness of an institution and the likelihood that it might fail go far
beyond those of concern to a consumer protection regulator. Moreover, as
the House of Lords Select Committee on Economic Affairs observed,
There is also a cultural difference between conduct-of-

business and prudential supervision. Conduct-of business
supervision is often performed by lawyers. Prudential
124
Id. at 25.
125
See id.
126
Id. at 17.
127
Id. at 31.
supervision is largely an economic activity, particularly at

the macro level. It seems likely that either a lawyerly or an
economic approach would dominate in a supervisory body
that performed both prudential and conduct of business
supervision, and that this dominance would reduce the
effectiveness of the dominated half of the organisation.128
The function that receives the greatest emphasis will be that having
the greatest political saliency: this means that in normal times, when bank
failures are rare, consumer protection regulation is likely to be the main
focus of agency attention. Although the FSA has now increased the
resources it devotes to prudential regulation,129 the above analysis suggests
that this is likely to be a relatively short term development, remaining in
place only as long as political attention is focused on the fall-out from the
crisis.
C. IT IS ESSENTIAL TO ACHIEVE A BALANCE BETWEEN THE

FINANCIAL STABILITY AND CONSUMER PROTECTION
OBJECTIVES
Because of the circumstances in which the U.K.’s regulatory

reforms took place – as a reaction to perceived regulatory failures in
consumer protection – it was perhaps inevitable that this aspect of
regulation should have been their main focus. The overarching desire on
the part of the FSA’s architects was to establish a strong consumer
protection regulator that would be independent of the industry.130
However, one consequence of the consumer protection focus was that the
financial stability objective did not receive the attention that it either
warranted or deserved.131 Fortunately, the Northern Rock episode has
provided the impetus to restore some balance to the post-1997
arrangements. The Bank of England has now been given both formal
statutory responsibility for financial stability and for handling bank
resolutions under a new legislative framework, the Banking Act 2009.132
128
129
Id. at 32.
130
Id. at 30, 32, 52-53.
131
Id. at 31-32.
132
Id. at 30-31.
However, as noted earlier, there are still a number of aspects of the

“Tripartite system” where reforms are still needed. In addition, although
charged with the formal statutory responsibility for maintaining financial
stability the Bank of England lacks most of the policy tools it needs for this
task. Hence, even now, the rebalancing is only partly finished.
A second dimension of the financial stability focus concerns what
is now termed “macroprudential” regulation. When the U.K.’s
arrangements were put in place, prudential regulation was conceptualized
in terms of ensuring the soundness of individual institutions. As the
financial crisis has made clear, however, ensuring the soundness of
individual firms is a necessary, but not sufficient condition for ensuring
financial stability. While in one respect the comparative neglect of
financial stability issues under the U.K.’s post-1997 arrangements was due
to an oversight, it also reflected the fact that what is now called the macro-
prudential perspective had not at the time gained the prominence that it
now enjoys. As noted above, an unfinished aspect of the U.K’s attempt to
re-balance its regulatory system concerns the additional macroprudential
powers that should be assigned to the Bank of England.
D. POLITICALLY MOTIVATED REFORMS OR THOSE MOTIVATED BY

A DESIRE TO “PUNISH” THE CENTRAL BANK ARE
COUNTERPRODUCTIVE
There is at least some circumstantial evidence for concluding that

part of the motivation for the U.K.’s reforms was to “punish” the central
bank or to “cut it down to size.”133 However, as the subsequent British
experience shows, there is no plausible alternative to having a central bank
with an extensive mandate and the ability to intervene to mitigate a crisis.
The FSA’s architects appear to have believed that it would be possible to
create a rival center of power to the Bank, without realizing the reality that
without significant financial muscle of its own, the FSA was destined to
play a subsidiary role in any crisis. Only the central bank has the ability to
play the role of LoLR and this fact means that it must play a unique role in
any financial safety net arrangement. The members of the U.S. Congress
who have recently criticized the Federal Reserve for its actions in
stemming the crisis need to reflect on whether there are any viable
alternatives. The British experience suggests that there are not.
133
Jacomb, supra note 49, at 3-4.
E. SOME “OVERLAP” AND “DUPLICATION” OF REGULATORY

FUNCTIONS IS UNAVOIDABLE
As should now be apparent, the U.K.’s regulatory reforms were

inspired, to a very large extent, by the desire to eliminate the perceived
duplication and overlap of regulatory authority resulting from the Financial
Services Act system in particular. While the Act had indeed created a
system that was excessively complex – especially from the point of view of
the individual consumer – this factor arguably received too much attention
in the resulting reforms.
A particularly clear example was the decision not to give the Bank
of England its own powers to gather information from the financial sector
(banks in particular). It was therefore reliant on the FSA to provide it with
the data it required to perform its “financial stability” function.134 The
thinking appeared to be that if the FSA was to be the banking supervisor,
the Bank of England should have only a general role in relation to overall
financial stability, and did not require the ability to gather institution-
specific information. 135 Because one stated objective of the 1997 reforms
was to reduce regulatory duplication and overlap – a major selling point
with the industry – only the FSA was given information-gathering powers.
This decision ignored the experience of many other countries
where the central bank was not itself the prudential regulator, and indeed
the Bank of England’s own history before it assumed the statutory
responsibility for bank regulation in 1979.136 In its role as lender of last
resort it had been able to exert significant moral suasion over the banking
sector, and the Discount Office was able to obtain information from banks
on a purely informal basis.137 Other central banks also enjoy substantial
information gathering powers of their own.138 For example, the Bank of
Japan’s information-gathering ability includes the power to conduct bank
examinations, notwithstanding that this duplicates the function of the
Financial Services Agency. These precedents should have shown that even
without the formal statutory responsibility for banking supervision, the
134
BROWN, supra note 109, at ¶ 8.
135
Id. at ¶ 6.
136
See Schooner & Taylor, supra note 2, at 629-32.
137
Id. at 614-15.
138
INSTITUTE FOR MONETARY AND ECONOMIC STUDIES, FUNCTIONS AND
OPERATIONS OF THE BANK OF JAPAN 80 (2004), available at
www.boj.or.jp/en/type/exp/about/data/ foboj01.pdf
central bank still needed to have access to substantial amounts of

institution-specific information and ideally its own capacity to go about
gathering that information.
F. CRISIS MANAGEMENT PREPAREDNESS MATTERS
Finally, insufficient attention was given to crisis management

arrangements. For at least two decades prior to the formation of the FSA,
the U.K. had not experienced any episodes of serious financial distress.
This may have bred a certain degree of complacency about the need for
adequacy crisis management preparedness and planning. Although the
Memorandum of Understanding was drawn up between the Treasury, Bank
of England, and FSA, the arrangements envisaged were rarely tried in
practice and the Joint Crisis Management Committee rarely met. The
arrangements also assumed that the Treasury would be the glue that held
this system together, thus involving it in the minutiae of crisis management
decision-making – a role that it was ill-equipped to perform and one that
hampered the ability to reach quick decisions in an environment where time
was of the essence.139 It is therefore necessary to ensure that the central
bank’s freedom of manoeuvre is not excessively constrained by any
arrangements that are put in place.
VII. CONCLUSION: THE RETURN TO “TWIN PEAKS”
In the aftermath of the Global Financial Crisis there has been a

revival of interest in the Twin Peaks model. The experience of the U.K.
during the financial crisis has strengthened the arguments of the FSA’s
critics who challenged the viability of a single regulatory agency in a
financial centre as large and diverse as the U.K. The British Conservative
party, which at the time of writing is still in opposition but is expected to
win the election due in 2010, has now adopted the policy of abolishing the
FSA and introducing a division between prudential and conduct of business
regulation with the former being returned to the Bank of England.140
Similarly, in the U.S., the Twin Peaks concept has received attention in
139
Peter Hayward “The Financial Sector – The Responsibilities of Public
Agencies” in CHARLES ENOCH et al, supra note 118.
140
George Osborne, Foreword, in FROM CRISIS TO CONFIDENCE: PLAN FOR
SOUND BANKING (2009).
evidence given to Congressional committees141 and as a major source of the

Bush administration’s proposals of 2008.
As can be seen from the above analysis, when Britain adopted its
single regulator structure in 1997, it did not do so due to a conscious
rejection of the Twin Peaks alternative. Rather, the British government’s
decision seems to have owed more to the legislative timetable and the
apparent simplicity of the single regulator in avoiding some of the
complexity, duplication and opacity which had been a focus of the
criticisms of the previous system. The single regulator’s very simplicity
may well have been a factor in its favour; but the apparent simplicity of the
structure was deceptive as it resulted in some of the complexities of
financial regulation and crisis management being neglected.
The Twin Peaks alternative might, arguably, have avoided some of
the design flaws of the post-1997 arrangements. In particular, it would
have avoided trying to set up a rival center of power to the Bank of
England, thereby creating crisis management arrangements that were far
too unwieldy. Because the Bank and the FSA were assigned equal status in
the Tripartite arrangements, the active role of the Treasury was essential to
hold the ring and to ensure a cooperative relationship between the two
agencies. By contrast, a specialist prudential regulator might have been
established more clearly under the Bank of England’s wing, and as a result
could have enjoyed much closer links with the central bank than did the
FSA. There are a variety of precedents for this possible arrangement: the
relationship between the Bank of France and the Commission Bancaire, or
between the Finnish Central Bank and that country’s Financial Supervision
Agency could have been potential models.142 In these structures, although
the regulatory agencies are governed by boards separate from those of the
central bank, their staff are central bank employees and extensive use is
made of shared facilities, information technology platforms and databases.
Nonetheless, although Twin Peaks has its attractions, it is
necessary to be cautious about trying to introduce too much neatness and
tidiness into regulatory structures. The objectives of financial regulation
can be neatly packaged into two, but the range of regulatory functions is far
more diverse. At least six (or possibly seven) regulatory functions can be
141
See Enhancing Investor Protection and the Regulation Of Securities
Markets: Hearing Before S. Comm. on Banking, Housing and Urban Affairs, 11th
Cong. 36-39 (2009) (testimony of Mr. John Coffee, Professor, Columbia Law
School).
142
See Abrams & Taylor, supra note 9, at 23-24.
identified: financial system stability; crisis management; the prudential

regulation of systemically important firms; the prudential regulation of
firms that are not systemically important; sales practice regulation; and
market conduct regulation.143 (Competition policy is a possible seventh
regulatory function although it applies in many sectors other than financial
services.) At its most basic, the problem of designing a regulatory structure
is one of deciding which of these functions belong together in the same
agency. The single regulator concept tried to combine most of these
functions within one agency. That has been shown to be a step too far. But
there are many possible configurations between this option and the current
highly fragemented regulatory system in the United States.
143
See C. A. E. GOODHART ET AL., FINANCIAL REGULATION: HOW, WHO AND
WHERE NOW? (1998).
WHITHER THE DUTY OF GOOD FAITH IN
UK INSURANCE CONTRACTS
John Lowry∗
***
This article explores the current state of the law in the United Kingdom
concerning the duty of good faith in insurance contracts. Recent case law
provides that the duty of disclosure by insureds is constantly being refined.
It argues that due to the current fragmentation of the law, future reform
should be focused on creating a consistent regime for insurance contracts.
Such regime should be flexible enough to encompass both consumer and
commercial insurance, while demonstrating certain and clear objectives.
The first part examines the duty of disclosure by an insured as formulated
by Lord Mansfield CJ. The second part analyzes the case law that followed
Carter v. Bohem, which developed the notion of good faith and expanded it
into a duty of utmost good faith. Third, the discomfort of the UK courts
and UK law reform agencies over the severity of the insured’s duty along
with the injustices that result when insurers avoid a policy for non-
disclosure is explored. Fourth, recent judicial opinions that attempt to
alleviate the position of the insured are assessed. The fifth and concluding
part of this article briefly examines the 2009 [UK] Consumer Insurance
(Disclosure and Representations) Bill published by the English and
Scottish Law Commissions in December 2009. It constructs an alternative
model which takes account of recent developments in Australian law. It is
argued that the focus should be on balancing the economic costs of reform
with the benefits of a more balanced regime which does not create a
distinction between consumer and business insureds.
***
∗
Professor of Law and Vice Dean of the Faculty of Laws, UCL. I owe a debt
of gratitude to the anonymous referees for their helpful comments. Liability for
any errors, however, is mine alone. I would also like to thank the editors of the
Journal for affording me the opportunity to amend the article immediately prior to
its publication so as to include reference to UK reform proposals published in
December 2009.
Insurance contracts are highly unusual in that they are founded

upon the doctrine of uberrimae fidei.1 The consequence is that the general
contractual duty borne by parties to avoid misrepresentation is extended
and reinforced by the additional obligation to disclose all material facts that
would induce the insurers to underwrite the risk. This was first laid down
by Lord Mansfield C.J. in Carter v. Boehm,2 and his formulation of the
disclosure duty is partially codified in the Marine Insurance Act 1906.3
Lord Mansfield was at the time attempting to import into English
commercial law the civil law notion of good faith, but this ultimately
proved unsuccessful and only survived for a very limited class of
transactions, including insurance.4
The effect of non-disclosure by the insured entitles the insurer to
avoid the contract ab initio, notwithstanding the absence of any fraudulent
intent. The economic consequences are severe and disproportionately
harsh. The policy becomes valueless so that the insured loses the financial
safeguard that the policy was designed to provide should the losses caused
by the insured risks come to pass.5 This is not to say that the rationale
underlying the disclosure duty and the remedy for its breach is in any way
obscure. The role of underwriters is to assess risk and if there are material
factors known only to the insured, then the insured must disclose them.
The reason for the duty is clear where the underwriter is not in a position to
1
For the common law position governing the general law of contracts see,
e.g., Keates v. Cadogan, (1851) 138 Eng. Rep. 234. For a review of the policy
considerations underlying the general contractual position see J. BEATSON,
ANSON’S LAW OF CONTRACT 263-64 (28th ed. 2002).
2
(1766) 97 Eng. Rep. 1162, 1165.
3
See Marine Insurance Act, 1906, 6 Edw. 7, c. 41, §§ 17-20 (Eng.).
4
See Lord Mustill’s speech in Pan Atl. Ins. Co. Ltd. v. Pine Top Ins. Co.,
[1994] 1 A.C. 501 (H.L.), and Lord Hobhouse’s speech in Manifest Shipping Co.
Ltd. v. Uni-Polaris Shipping Co. Ltd., [2003] 1 A.C. 469. See also Potter L.J.’s
observations in James Spencer & Co. Ltd. v. Tame Valey Padding Co. Ltd.,
QBENI 97/1118 CMS1 (A.C. April 8, 1998) (Smith Bernal Reporting Ltd. for
Lawtel). It is noteworthy that Lord Mansfield was familiar with the civilian
tradition (see Sir WILLIAM HOLDSWORTH, 5 A HISTORY OF ENGLISH LAW 147
(1966)). This is not to suggest that there is a universal view among civilian
jurisdictions on the meaning and scope of good faith. See R. ZIMMERMANN & S.
WHITTAKER, GOOD FAITH IN EUROPEAN CONTRACT LAW (Cambridge, CUP,
2000). See also M. Bridge, Doubting Good Faith, NEW ZEALAND BUSINESS LAW
QUARTERLY, 2005, at 426.
5
HIH Cas. & Gen. Ins. Co. v. Chase Manhattan Bank [2003] UKHL 6
(statement of Lord Hobhouse).
2009] GOOD FAITH IN UK INSURANCE CONTRACTS 99
discover for himself or herself the circumstances that may impact upon the
risk. For example, one case would be where the insured seeks to take out a
life policy having received a death threat. Or a vessel the underwriter is
asked to insure may be on the high seas and, therefore, not available for
inspection.6 In such circumstances there can be no quarrel with the fairness
of the insured’s duty of good faith. But in the case law subsequent to
Carter v. Boehm, the boundaries of the insured’s duty were expanded
beyond, it is suggested, what Lord Mansfield originally envisaged. For
example, it has been held that the failure of an insured to disclose criminal
convictions and his precarious financial position when applying for fire
insurance will enable the insurer to avoid the policy following a fire caused
by an electrical fault.7
In practical terms, the issue which pervades the duty of good faith
can be reduced to the following question: how can the ordinary insured,
whether acting in a private or commercial capacity, untutored in the
niceties of insurance law, be expected to know what particular
circumstances are material and would, therefore, influence the prudent
underwriter?8 The sheer breadth of the insured’s duty together with the all-
or-nothing consequence of avoidance, therefore, rightly gives rise to
legitimate concern. In a series of recent cases, decided over the last ten
years or so, the English courts have been steadily refining the disclosure
duty while, at the same time, laying considerable emphasis on the mutuality
of the requirement of good faith by giving content to that borne by insurers.
This process is a rebalancing exercise.9 As such, it involves a tacit
6
See NEW ZEALAND LAW COMMISSION, No. 46: SOME INS. LAW PROBLEMS 2
(1998).
7
Quinby Enter. Ltd. v. Gen. Accident Ltd. [1995] 1 N.Z.L.R. 736. But cf.
Waller L.J.’s reasoning in North Star Shipping Ltd. v. Sphere Drake Ins. plc [2006]
EWCA (Civ) 378; see also Doheny v. New India Assurance Co. [2004] EWCA
Civ. 1705; O’Kane v. Jones [2003] EWHC 3470 (Comm.); James v. CGU Ins. plc,
[2002] Lloyd’s Rep. I.R. 206; March Cabaret Club & Casino Ltd. v. London
Assurance, [1975] 1 Lloyd’s Rep. 169.
8
See Lord Mansfield’s formulation of the duty and the Marine Insurance Act,
1906, 6 Edw. 7, c.41, § 18.
9
Indeed, this is in line with various calls for reform which have long gone
unheeded by the legislature. See infra note 15. That said, the UK Financial
Ombudsman Service (FOS) has sought to mitigate the harshness of the duty, at
least in so far as it applies to consumer insureds and small businesses with a group
annual turnover of less than £1 million. For commercial insureds, however, who
unraveling of case law spanning over a century in which the courts had
adopted an unequivocal stance in permitting avoidance for non-disclosure
across the range of insurance, both consumer and commercial, without
regard to notions of fairness, proportionality, or whether there was actual
inducement.10 Admittedly the anxiety of modern judges has generally been
directed towards relieving the position of the consumer or private insured,
but the limits of this shift of emphasis are not entirely clear. It begs the
question whether commercial insurance is also being targeted. Such doubt
carries the danger of undermining the very certainty that should represent
the cornerstone of commercial law in this respect and the economic
implications are potentially significant.
Taken in the round, it is possible to distill several strands of
reasoning from the modern case law. First, focusing solely upon the
content of the insured’s pre-contractual duty of disclosure, the courts have
sought to limit its scope by refining the conditions, most notably the
requirement of inducement, that must be met before the insurer may
justifiably avoid the policy for non-disclosure. Further, in relation to
avoidance, it is noteworthy that in recent times some judges have had
recourse to notions of good faith, conscience, and fairness when assessing
whether insurers may exercise the remedy. But judicial thinking in this
respect is not entirely consistent for it has been suggested that, as with the
remedy of rescission for misrepresentation, the rights of insurers are
unfettered by such considerations.11 Another strand of reasoning that has
emerged has been directed towards the contours of waiver. An insurer has
every opportunity to ask specific questions of the applicant for insurance in
the proposal form. Typically, those questions will be directed towards
claims history or health where the application relates to life or sickness
insurance. Nonetheless, even where such questions are raised, the insured
is not relieved from his or her duty to volunteer any further material
circumstances that fall outside the scope of them. Any defense that an
insured might seek to raise based on waiver is, in the orthodox view,
doomed to failure.12 However, this has been challenged recently in a
are outwith the FOS jurisdiction, it has been left to the courts to alleviate their
position.
10
See, e.g., the judgments in Lambert v. Co-operative Ins. Soc’y, [1975] 2
Lloyd’s Rep. 485. See also infra notes 82-91.
11
See the judgment of Mance L.J. in Brotherton v. Aseguradora Colseguros
S.A. [2003] EWCA Civ. 705. See infra note 197.
12
See, e.g., Roselodge, Ltd. v. Castle, [1966] 2 Lloyd’s Rep. 113, 132.
powerful dissent delivered by Rix L.J., whose reasoning may portend

future developments.13 In this regard, considerable emphasis is now being
paid to the importance of the presentation of risk. The modern view is that
insurers should not be content to play a passive role during the disclosure
process but should be prepared to make necessary enquiries about the risk
to be underwritten.14 Finally, the question of whether or not the insured’s
duty of good faith continues post-contractually and again triggers at the
claims stage has also attracted considerable judicial scrutiny of late.
This article is in five parts. It first examines the scope of the
insured’s duty of disclosure originally formulated by Lord Mansfield C.J..
Secondly, it considers the case law subsequent to Carter v. Boehm in which
the notion of good faith was developed and expanded into a duty of so-
called utmost good faith. The third part will outline the unease expressed
by the courts and by the law reform agencies, over the harshness of the
insured’s duty and the injustices that result when insurers avoid a policy for
non-disclosure.15 The fourth part will assess recent judicial inroads into the
orthodox position that appear to be aimed at alleviating the position of the
insured. In this respect, the starting point will be Pan Atlantic Insurance
Co. v. Pine Top Insurance Co.16 Although the House of Lords took the
opportunity to settle the legal position relating to the insured’s duty of
disclosure, it did not quell the debate surrounding the perceived iniquities
of the insurer’s remedy. Finally, against the backdrop of modern English
case law, together with key developments in Australia, a model will be
13
See WISE Ltd. v. Grupo Nacional Provincial S.A. [2004] EWCA Civ. 962.
See infra note 168-87.
14
This is not a novel development but reflects the view expressed by Lord
Mansfield C.J. in Carter v. Boehm, (1766) 97 Eng. Rep. 1162. However, by the
mid-nineteenth century, the point seems have faded from judicial thinking when
addressing the scope of the disclosure duty.
15
See, e.g., LAW REFORM COMMITTEE, CONDITIONS AND EXCEPTIONS IN
INSURANCE POLICIES, 1957, Cmnd. 62; THE LAW COMMISSION, INSURANCE LAW –
NON-DISCLOSURE AND BREACH OF WARRANTY, 1980, Cmnd. 8064; JOHN BIRDS,
NAT’L CONSUMER COUNCIL, INSURANCE LAW REFORM: THE CONSUMER CASE FOR
A REVIEW OF INSURANCE LAW (1997). Although the UK Department of Trade and
Industry (now renamed the Department for Business, Innovation and Skills
(DBIS)) issued a draft Insurance Bill following the Law Commission’s report in
1980, nothing came of it, the government of the day being satisfied that the
industry’s response via the Association of British Insurers’ Statement of Practice
struck the appropriate balance for the consumer-insured.
16
[1995] 1 A.C. 501 (H.L.).
proposed that might serve to inform the English and Scottish Law
Commissions’ current re-examinations of insurance law that identifies non-
disclosure as a key issue.17 It will be seen that the current state of the law is
fragmented and complicated and that future reform should be directed
towards constructing a coherent regime for insurance contracts that meets
the objectives of certainty and clarity, while being sufficiently flexible to
encompass both consumer and commercial insurance.
I. THE ORIGINS OF THE DISCLOSURE DUTY AND ITS

STATUTORY CODIFICATION
A. CARTER V. BOEHM
An enduring legacy of the Seven Years War is that it left us with a

landmark decision which contains the most quoted passage in U.K.
insurance law. In Carter v. Boehm, the issue of non-disclosure came to
court as a result of the Governor of Sumatra, George Carter, effecting a
policy of insurance on Fort Marlborough, a trading fort, against the
likelihood of a French attack.18 His decision to insure was vindicated
when, in April 1760, the fort was seized by the French.19 The Governor’s
claim was disputed by the underwriter and in 1766, Lord Mansfield C.J.,
17
The English and Scottish Law Commissions are statutory independent
bodies created by the [UK] Law Commissions Act 1965 c. 22, to keep the law
under review and to recommend reform where it is needed. The insurance contract
law reform project was announced on 14 October 2005, the first “issues paper” on
misrepresentation and non-disclosure was published at the end of September 2006.
See The Law Commission, Insurance Contract Law, Misrepresentation and Non-
Disclosure (Sept. 2006), available at http://www.lawcom.gov.uk/docs/insurance_
contact_law_issues_paper_1. See infra notes 305-11. On December 15 2009 the
Law Commissions published their joint report and draft Bill to reform the law on
what a consumer-insured must disclose to the insurers prior to the conclusion of
the policy; see CONSUMER INSURANCE LAW: PRE-CONTRACT DISCLOSURE AND
REPRESENTATION, 2009, Cm 7758, discussed infra note 307 et seq..
18
Carter, 97 Eng. Rep. at 1163. Park notes that Carter is a seminal case
(“. . . . from it may be collected all the general principles which the doctrine of
concealments, in matters of insurance, is founded, as well as all the exceptions. . . .
”). JAMES ALLAN PARK, A SYSTEM OF THE LAW OF MARINE INSURANCES 193
(Thomas & Andrews 1800) (1787).
19
Carter, 97 Eng. Rep. at 1163.
presiding at Guildhall, heard the consequent action.20 The underwriters had

sought to avoid the contract on the basis that the Governor had failed to
disclose the fort’s weakness and its vulnerability to an attack by the
French.21 Their defence failed, but Lord Mansfield took the opportunity to
formulate the duty of good faith which has come to represent a cornerstone
of English insurance law:
The special facts, upon which the contingent

chance is to be computed, lie most commonly in the
knowledge of the insured only; the under-writer trusts to
his representation, and proceeds upon confidence that he
does not keep back any circumstance in his knowledge, to
mislead the under-writer into a belief that the circumstance
does not exist, and to induce him to estimate the risqué, as
if it did not exist.
The keeping back such circumstance is a fraud,
and therefore the policy is void. Although the suppression
should happen through mistake, without any fraudulent
intention; yet still the under-writer is deceived, and the
policy is void; because the risque run is really different
from the risque understood and intended to be run at the
time of the agreement.
....
Good faith forbids either party by concealing what
he privately knows, to draw the other into a bargain, from
his ignorance of that fact, and his believing the contrary.
But either party may be innocently silent, as to
grounds open to both, to exercise their judgment upon.
Aliud est celare; aliud, tacere; neque enim id est celare
quicquid reticeas; sed cum quod tuscias, id ignorare
emolumenti tui causa velis eos, quorum intersit id scire.22
This formulation begs the question as to why insurance contracts

are exceptional in requiring a positive duty of disclosure. Although the
disparity of knowledge between the parties has been proffered as the
explanation, this is hardly a satisfactory rationale in itself given that in
20
Id. at 1162-63.
21
Id. at 1163.
22
other common contracting situations, the parties similarly lack equality of

information. The explanation for this exceptional feature of insurance
must, therefore, lie elsewhere. The investigation begins by considering
Lord Mansfield’s reasoning. It then goes on to review the jurisprudence
surrounding the good faith requirement and how it evolved into a duty of
so-called utmost good faith.
B. THE RATIONALE
Lord Mansfield explained that the policy considerations underlying

the duty are the prevention of fraud and the furtherance of good faith: it
therefore fulfils a prophylactic role.23 He based it upon the concept of
“concealment,” but over time this developed beyond deliberate
concealment so as to encompass all non-disclosure, however innocent, of a
material fact. In Carter v. Boehm, the underwriter had argued that the
insured had been fraudulent in failing to disclose the fort’s vulnerability to
attack.24 This contention was unsuccessful, it being held that the
underwriter must be taken to have realised that the Governor, by insuring,
obviously apprehended the possibility of an attack.25 By underwriting the
risk, the insurer thereby assumed knowledge of the state of the fort.26 It
was stressed that the underwriter, sitting in London, was in a better position
than the Governor to stay informed about the fortunes of the war and so it
was not a matter within the private knowledge of the Governor only, but
was, in fact, in the public domain.27 Lord Mansfield concluded that a
verdict in favour of the underwriters would have had the effect of turning a
rule against fraud into an instrument of fraud.28 He proceeded on the basis
that good faith was a mutual duty, not an obligation borne solely by
23
This is a narrower view than that expressed over a century later by
Channell J in Re Yager & Guardian Assurance Co., [1912] 108 L.T. 38 (K.B.), to
the effect that the rationale underlying the disclosure duty is not the need to
prevent harm to the insurer as such, but the need for a true and fair agreement
whereby risk is transferred. Id. at 44-45.
24
25
Id. at 1167.
26
Id.
27
Id.
28
Id. at 1169.
insureds, and being mutual the problem of unfairness was shared between
the parties.29
Lord Mansfield also placed emphasis on the need for insurers to
demonstrate reliance. He explained that the underwriter “proceeds upon
confidence that [the insured] does not keep back any circumstance in his
knowledge, to mislead the under-writer into a belief that the circumstance
does not exist, and to induce him to estimate the risque, as if it did not
exist.”30 Herein we see the assimilation of non-disclosure with
misrepresentation. The good faith duty converts non-disclosure into
misrepresentation because an insured who fails to disclose a material fact is
effectively misrepresenting the true state of affairs. Lord Mansfield’s
choice of language is critical: it traverses the two vitiating factors and
reliance and inducement lies at the heart of both. For misrepresentation,
the consequence is therefore the same as with pure non-disclosure, namely
avoidance of the contract ab initio.31
Both in Carter v. Boehm, and in subsequent cases, Lord Mansfield
sought to limit the scope of the insured’s duty by, for example, stressing the
need for underwriters to be proactive in ascertaining facts material to the
29
Indeed, Lord Mansfield was scathing in his condemnation of the
underwriter’s defense:
The underwriter, here, knowing the governor to be

acquainted with the state of the place; knowing that he
apprehended danger, and must have some ground for his
apprehension; being told nothing of neither; signed this policy,
without asking a question. If the objection ‘that he was not told’
is sufficient to vacate it, he took the premium, knowing the
policy to be void; in order to gain, if the alternative turned out
one way; and to make no satisfaction, if it turned out the other:
he drew the governor into a false confidence . . .
. . . If he thought that omission an objection at the time,

he ought not to have signed the policy with a secret reserve in his
own mind to make it void; if he dispensed with the information,
and did not think this silence an objection then; he cannot take it
up now, after the event.
Id. at 1169.
30
31
The insurers remedy in this respect is codified by the Marine Insurance Act,
1906, 6 Edw. 7, c. 41, § 20 (Eng.).
risk. In Noble v. Kennoway, the insured vessel had arrived safely at

Labrador but prior to being unloaded, it was used for fishing.32
Subsequently, the vessel was taken by privateers while it was unmanned.33
The owners’ claim for the value of the cargo was met with the defence by
the underwriters that they were not liable because of the delay in
unloading.34 The insured argued that this was a trade usage in this
particular port because of the lack of warehousing.35 Lord Mansfield,
finding for the insured, reasoned that every underwriter was presumed to
know the practices of the trade he insures and if he does not know then it is
his duty to inform himself of it.36 He returned to the point in Mayne v.
Walter, where the insured’s claim for the loss of supercargo seized by the
French was met with the defence that he should have disclosed the
existence of a French ordinance prohibiting Dutch ships carrying the
supercargo of any country at war with France on pain of it being taken as
prize.37 Lord Mansfield said that if both parties were ignorant of the
relevant fact, “the underwriter must run all risks: and if the [underwriter]
knew of such an edict, it was his duty to inquire, if such supercargo were
on board.”38 He went on to note that “[i]t must be a fraudulent
concealment of circumstances, that will vitiate a policy.”39 This has been
termed the narrow Mansfield rule.40 Reflecting upon his original
formulation, Lord Mansfield appears to have come around to the view that
the duty is limited insofar as it must strike a balance between the parties so
as to achieve some symmetry between them. Indeed, by the early
nineteenth century, emphasis was being placed on Lord Mansfield’s clear
admonition that underwriters have a distinct investigative role to play in the
disclosure process. For example, in Friere v. Woodhouse, a marine
insurance case, Burrough J. said, “what is exclusively known to the assured
ought to be communicated; but what the underwriter, by fair inquiry and
due diligence, may learn from ordinary sources of information need not be
32
99 Eng. Rep. 326, 326-27.
33
Id. at 326.
34
Id.
35
Id.
36
Id. at 327.
37
Mayne v. Walter, (1782) 99 Eng. Rep. 548, 548-49. PARK, supra note 18,
at 196.
38
PARK, supra note 18, at 196.
39
Id.
40
See R. A. Hasson, The Doctrine of Uberrima Fides in Insurance Law - A
Critical Evaluation, 32 M.L.R. 615, 618 (1969).
disclosed.”41 The material information could have been discovered by the

underwriter from Lloyd’s List.42
It is striking that throughout his judgments on the issue of non-
disclosure Lord Mansfield avoided the terminology of “utmost” good faith.
Yet section 17 of the Marine Insurance Act 1906, the preamble of which
declares it to be a codifying statute, states that insurance is uberrimae
fidei.43 It goes on to provide that a contract of insurance is a contract based
upon the duty of utmost good faith which, if broken, entitles the other party
to avoid the contract.44 Section 17 does not, therefore, precisely mirror the
language of Lord Mansfield’s formulation which, as seen above, draws the
distinction between deliberate concealment and misrepresentation (bad
faith) and innocent (good faith) mistaken non-disclosure. The provision
must, therefore, be seen as synthesising not Lord Mansfield’s views, but
rather the dominant view emerging from the case law decided during the
latter half of the nineteenth century to the effect that the underwriter is a
passive recipient of information supplied by the insured when presenting
the risk. As such, this loses sight of the more restrictive views expressed
not only in Carter v. Boehm, but which was also in the judgments found in
Noble v. Kennoway, Mayne v. Walter, and Friere v. Woodhouse.45 It was
certainly not within the mandate of Sir Mackenzie Chalmers, who drafted
the Digest upon which the 1906 Act was based, to correct the significant
body of case law he sought to codify.46 The question that arises, therefore,
41
Friere v. Woodhouse, (1817) 171 Eng. Rep. 345, 345. See also Gandy v.
The Adelaide Marine Ins. Co., 6 Eng. Rep. 746, 757 (Q.B.).
42
Friere, 171 Eng. Rep. at 345.
43
It has long been settled that the Marine Insurance Act, 1906, 6 Edw. 7, c.
41, §§ 17-18 (Eng.) are of general application in insurance law. See, e.g., P.C.W.
Syndicates v. P.C.W. Reinsurers, (1996) 1 W.L.R. 1136, 1140 (A.C.); Australia
and New Zealand Bank, Ltd. v. Colonial and Eagle Wharves Ltd., (1960) 2
Lloyd’s L. Rep. 241, 251 (Q.B.D ); Cantiere Meccanico Brindisino v. Janson,
(1912) 3 K.B. 452, 467. More recent cases have also proceeded on the assumption
that the provisions apply to all types of insurance. See Pan Atlantic Ins. Co. Ltd. v.
Pine Top Ins. Co. Ltd., [1994] 1 A.C. 501, 518 (H.L.); Lambert v. Co-operative
Ins. Soc’y, (1975) 2 Lloyd’s Rep 485, 487.
44
Marine Insurance Act, 1906, 6 Edw. 7, c. 41, § 17 (emphasis supplied).
45
Friere v. Woodhouse, (1817) 171 Eng. Rep. 345; Mayne v. Walter, (1782)
99 Eng. Rep. 548; Noble v. Kennoway, (1780) 99 Eng. Rep. 326. See also Lord
Ellenborough’s closely reasoned judgment delivered in Haywood v. Rodgers,
(1804) 4 East 590.
46
See note 45 infra for conflicting case law.
is when did the disclosure duty metamorphose into something requiring

utmost good faith?
II. THE EVOLUTION OF UBERRIMAE FIDEI
In broad terms, a hallmark of much of the case law subsequent to

Carter v. Boehm is the expansive approach that was taken towards Lord
Mansfield’s formulation of the disclosure duty. It is not proposed to
examine the merits of this case law in terms of whether the scope of Lord
Mansfield’s judgments were misconstrued,47 but, as commented above, it
was such that by the end of the nineteenth century, synthesising it required
section 17 of the 1906 Act to declare insurance contracts to be of “utmost”
good faith. Perhaps surprisingly, the suggestion that an insured must be of
utmost honesty (as if there may be lower degrees of honesty) as represented
in this statutory declaration was not seen as being particularly controversial
or novel. As Lord Herschell, who originally took charge of the Bill when it
was introduced in the House of Lords in 1894, explained, its purpose was
to reproduce as exactly as possible the state of the existing law.48
Tracing the antecedents of “utmost” good faith is an intriguing
exercise, for it has no equivalent in the civil law.49 Indeed, in Mutual and
Federal Insurance Co. Ltd. v. Oudtshoorn Municipality,50 the Supreme
Court of Appeal of South Africa, expressing the view that the effect of the
Pre-Union Statute Revision Act 43 of 1977 was to make South African
insurance law governed by Roman-Dutch law, was moved to observe that
“uberrimare fides is an alien, vague, useless expression without any
meaning in law…our law of insurance has no need for uberrimae fides and
the time has come to jettison it.”51 Its origins can, however, be discerned in
U.K. case law decided during the latter half of the nineteenth century.52
For example, in Bates v. Hewitt, the court paid little heed to Lord
47
See Hasson, supra note 40.
48
Durant v. Durant, I Haggard Eccl. Rep. 733.
49
See M.A. Millner, Fraudulent Non-disclosure 74 S.A.L.J. 177, 188 (1957).
50
Mut. and Fed. Ins. Co Ltd. v. Oudtshoorn Mun. 1985 (1) SA 419 (AD).
51
Id. at 433F.
52
See A. D. M. FORTE, Good Faith and Utmost Good Faith in A.D. M. Forte
(ed), GOOD FAITH IN CONTRACT AND PROPERTY LAW, (Hart Publishing 1999).
Joubert J.A. cites Dalglish v. Jarvis, (1850) 2 Mac. & Gord. R. 231, as the decision
in which the term uberrimae fides is first used in Mut. and Fed. Ins., (1) SA at
431I.
Mansfield’s views expressed over a century earlier.53 The claimant had not
informed the insurer that the insured vessel, the Georgia, had been a
Confederate cruiser.54 The Georgia was well known to the British public,
and when the ship came to Liverpool for breaking she attracted
considerable interest in both the press and in the House of Commons.55
The defendant underwriter admitted he had known of the ship’s history but
that at the time of underwriting it was not present in his mind.56 The jury
found that the underwriter was ignorant of the vessel’s notoriety at that
particular time, although they did go on to express the view that when the
risk was presented, he did have the means available for identifying the
ship.57 The court held that the claimant was in breach of his duty of
disclosure.58 Curiously, the judges in the case went to considerable lengths
to explain that they were merely applying a long established principle.
Lord Cockburn CJ stated that a proposer of insurance “is bound to
communicate to the insurer all matters which will enable him to determine
the extent of the risk against which he undertakes to guarantee the
assured.”59 Shee J., while admitting that the underwriter might through his
own investigations have discovered the material fact about the Georgia’s
history, concluded, however, that he was under no duty to make such
enquiries.60 This fails to sit with Lord Mansfield’s notion of an insurer’s
constructive knowledge - a critical factor in his finding in Carter v. Boehm.
It also fails to sit with Friere v. Woodhouse, in which it will be recalled, the
court, applying Lord Mansfield’s formulation of the duty, had no hesitation
in finding that underwriters had a pro-active role to play during the
disclosure process.61 Nonetheless, towards the close of the nineteenth
century, the consensus of judicial opinion was such that determining if the
duty of disclosure has been discharged requires something more than
merely exacting a duty of honesty from the insured. This came to the fore
in Life Association of Scotland v. Foster, in which the term “utmost good
faith” is adopted by Lord President Inglis: “Contracts of Insurance are in
53
Bates v. Hewitt, (1867) 2 L.R.Q.B. 595; 1867 WL 9866.
54
Id. at 604.
55
Id. at 595.
56
Id. at 604.
57
Id.
58
Id. at 599.
59
Bates, 2 L.R.Q.B. at 604-05.
60
Id. at 611.
61
Friere v. Woodhouse, (1817) 171 Eng. Rep. 345. See also Noble v.
Kennoway, (1780) 99 Eng. Rep. 326.
this, among other particulars, exceptional, in that they require on both sides
uberrima fides. Hence, without fraudulent intent, and even bona fides, the
insured may fail in the duty of disclosure.”62
This decision is followed soon after by Ionides v. Pender, in which
Blackburn J. was moved to assimilate the prevailing view into the so-called
“prudent insurer” test.63 While the judge accepted that “it would be too
much to put on the assured the duty of disclosing everything which might
influence the mind of an underwriter,” he nevertheless conceded that “a
concealment of a material fact, though made without any fraudulent
intention, vitiates the policy.”64 Blackburn J. concluded by stating that “all
should be disclosed which would affect the judgment of a rational
underwriter governing himself by the principles and calculations on which
underwriters do in practice act.”65
62
(1873) 11 M. 351, 359. It might be the case that this generation of judges
was perhaps placing too much reliance on Park’s Law of Marine Insurance
published in 1787, rather than tracing Lord Mansfield’s reasoning first hand. For
example, chapter 10 of Park’s treatise states that “the learned judges of our courts
of law, feeling that the very essence of insurance consists in a rigid attention to the
purest good faith, and the strictest integrity, have constantly held that it is vacated
and annulled by any the least shadow of fraud or undue concealment.” PARK,
supra note 18, at 174 (emphasis supplied).
63
(1874) 9 L.R.Q.B. 531. This was incorporated by Chalmers into § 18(2) of
the Marine Insurance Act, 1906, 6 Edw. 7, c.41 (Eng.), which provides, “[e]very
circumstance is material which would influence the judgment of a prudent insurer
in fixing the premium, or determining whether he will take the risk.”
64
Ionides, 9 L.R.Q.B. at 537, 539.
65
Id. at 539. A further opportunity to put forward his view on the scope of
the duty of disclosure was taken by him, now Lord Blackburn, in Brownlie v.
Campbell, (1880) 5 App. Cas. 925, in which he noted:
… [i]n policies of insurance, whether marine insurance or life

insurance, there is an understanding that the contract is uberrima
fides, that if you know any circumstance at all that may influence
the underwriter's opinion as to the risk he is incurring, and
consequently as to whether he will take it, or what premium he
will charge if he does take it, you will state what you know.
There is an obligation there to disclose what you know; and the
concealment of a material circumstance known to you, whether
you thought it material or not, avoids the policy.
Id. at 954. Material facts are typically categorised as either those relating to
Blackburn J.’s formulation is encapsulated in sections 17 and 18(1)

and (2) of the 1906 Marine Insurance Act. More particularly, section 18(1)
lays down the overriding pre-contractual duty of disclosure while section
18(2), which provides that “every circumstance is material which would
influence the judgment of a prudent insurer in fixing the premium, or
determining whether he will take the risk,” gives content to the governing
principle of utmost good faith declared by section 17.66 The view that
insurance required nothing less than utmost good faith, or the idea of
comparative degrees of honesty, thus became firmly entrenched in English
insurance law.
Shortly after the 1906 Act received Royal Assent, the opportunity
to explore the insured’s duty of disclosure came before the Court of Appeal
in Joel v. Law Union and Crown Insurance Company.67 As a means of
overcoming the practical difficulties of proof which a duty based solely
upon utmost good faith could give rise to,68 Fletcher Moulton L.J.
superimposed a requirement of reasonableness. The judge explained that:
There is, therefore, something more than an obligation to

treat the insurer honestly and frankly… There is the further
duty that he should do it to the extent that a reasonable
man would have done it; and, if he has fallen short of that
by reason of his bona fide considering the matter not
material, whereas the jury, as representing what a
reasonable man would think, hold that it was material, he
has failed in his duty, and the policy is avoided. This
further duty is analogous to a duty to do an act which you
undertake with reasonable care and skill, a failure to do
which amounts to negligence, which is not atoned for by
any amount of honesty or good intention. The disclosure
must be of all you ought to have realized to be material,
not of that only which you did in fact realize to be so.69
physical hazard or those relating to moral hazard. See JOHN LOWRY & PHILIP
RAWLINGS, INSURANCE LAW: DOCTRINES AND PRINCIPLES 93-99 (2d ed. 2005).
66
Marine Insurance Act, 1906, 6 Edw. 7, c. 41, § 18 (Eng.).
67
[1908] 2 K.B. 863.
68
See PETER MACDONALD EGGERS ET AL., GOOD FAITH AND INSURANCE
CONTRACTS ¶ 3.10 (2004).
69
Joel, [1908] 2 K.B. at 883-84 (emphasis added). This passage occurs in a
reserved judgment, and there is no indication that Vaughan Williams or Buckley
The insured is thus under a duty to disclose material facts, irrespective of

whether he or she appreciated their materiality.70 This combined test can,
and does give rise to unjust results. For example, in Horne v. Poland, the
insured’s policy was voidable due to his failure to disclose that he was an
alien having come to this country at the age of twelve, and that he had
changed his name from that of Euda Gedale to Harry Horne.71 Lush J.,
having noted that the applicable principle of law had been stated by
Fletcher Moulton L.J. in Joel, added:
If a reasonable person would know that underwriters would

naturally be influenced, in deciding whether to accept the
risk and what premiums to charge, by those circumstances
[i.e., that he came from a country where his countrymen
were not as careful and trustworthy as Englishmen], the
fact that they were kept in ignorance of them and indeed
were misled, is fatal to the plaintiff's claim.72
Similarly, in Becker v. Marshall, Salter J., also applying the test laid down
by Fletcher Moulton L.J., held on an issue of concealment as to foreign
origin and change of name, that while the claimant:
… in good faith did not realise that these were things

material to be disclosed…the average business man, the
average reasonable man, would not have taken that view,
and…that a reasonable man, the average reasonable man,
would have disclosed and would have known that it was
necessary to disclose.73
L.J.J. dissented, although they base their concurring judgments on different

grounds.
70
See EGGERS ET AL., supra note 68, ¶¶ 3.10-3.11.
71
[1922] 2 K.B. 364, 364.
72
Id. at 367. The decision must now be viewed as running counter to the
Race Relations Act, 1976, c.74 (Eng.).
73
(1922) 11 Lloyd’s List L.R. 114, 119 (K.B.). Fletcher Moulton L.J.’s
reasoning was further explained by Roskill J in Godfrey v. Britannic Assurance
Co. Ltd., [1963] 2 Lloyd’s Rep 515, 529 (Q.B.D.):
Curiously, for McNair J. in Roselodge v. Castle,74 Fletcher

Moulton L.J.’s formulation represents nothing less than a pure application
of Lord Mansfield’s rule. The dispute arose out of the rejection by the
insurers of the insured’s claim who, as diamond merchants, had insured
diamonds against all risks.75 The insurers’ defence was founded upon the
non-disclosure of two alleged material facts. First, that the principal
director of the insured company had been convicted of bribing a police
officer in 1946 and second, that the insured’s sales manager had been
convicted of smuggling diamonds into the United States in 1956.76
According to one of the expert witnesses called by the insurer, a person
who stole apples when aged 17 is much more likely to steal diamonds at
the age of 67 even if he had led a blameless life for 50 years, than someone
[T]here is a point here which often is not sufficiently kept in

mind. The duty is a duty to disclose, and you cannot disclose
what you do not know. The obligation to disclose, therefore,
necessarily depends on the knowledge you possess. I must not be
misunderstood. Your opinion of the materiality of that
knowledge is of no moment. If a reasonable man would have
recognized that it was material to disclose the knowledge in
question, it is no excuse that you did not recognize it to be so.
But the question always is, Was the knowledge you possessed
such that you ought to have disclosed it?
Such was the momentum of this approach that the 2nd (Hailsham) edition
of the Laws of England, Volume 18, prepared by Scott L.J., stated that:
Materiality is a question of fact, not of belief or opinion. The

assured does not therefore discharge his duty by a full and frank
disclosure of what he honestly thinks to be material; he must go
further and disclose every fact which a reasonable man would
have thought material… If, however, the fact, though material, is
one which he did not and could not in the particular
circumstances have been expected to know, or if its materiality
would not have been apparent to a reasonable man, his failure to
disclose it is not a breach of duty.
Id. at 586(3). This passage is repeated in the 3rd (Simonds) ed., vol. 22,
360.
74
[1966] 2 Lloyd’s List L.R. 113, 131-32 (QBD).
75
Id. at 113.
76
Id.
who had led a totally blameless life.77 This did not convince McNair J.
who held that the 1946 conviction was not a material fact, having “no
direct relation to trading as a diamond merchant.”78 Having examined the
authorities, with particular emphasis being given to Horne v. Poland, the
judge concluded:
In my judgment, on this review of the authorities the

judgment of Lord Justice Fletcher Moulton in Joel's case
contains, if I may respectfully say so, a correct statement of
the law on the topic. It has the merit…of emphasizing that
even under the present practice of admitting expert
evidence from underwriters as to materiality, the issue as to
disclosability is one which has to be determined as it was
in Lord Mansfield's day by the view of the Jury of
reasonable men.79
As conceded by Fletcher Moulton L.J., the duty does not require the
insured to disclose that of which he or she is ignorant, unless the insured
ought to have known of such circumstances in the ordinary course of
business.80 But, nevertheless, from the insured’s perspective the disclosure
duty laid down by the Marine Insurance Act of 1906, as explained by the
subsequent case law, is particularly harsh and, it is suggested, represents an
overly expanded view of Lord Mansfield’s original formulation which was
premised upon the notion of “concealment.” From a contemporary
standpoint, it is hardly surprising that by the second half of the twentieth
century both the courts and the law reform bodies were questioning
whether such a strict approach was necessarily appropriate for all classes of
insurance.
77
Id. at 132.
78
Id.
79
Id. at 131.
80
See also Lambert v. Co-operative Ins. Soc’y Ltd., [1975] 2 Lloyd’s Rep.
485, 490 (A.C.). See infra notes 63-66. The private insured will not, of course, be
deemed to possess constructive knowledge. See Economides v. Commercial
Union Assurance Co. plc, [1998] Q.B. 587, 601-03 (Simon Brown, L.J.).
III. CRITICISMS OF THE DUTY OF DISCLOSURE AND THE

REMEDY FOR BREACH
Although modern judges inevitably follow the substantial line of

case law on the Marine Insurance Act of 1906 §§ 17 and 18, they have not
been timid in expressing their unease over the rigours of the disclosure
duty. For example, in Anglo-African Merchants Ltd. v. Bayley, Megaw J.
queried whether the insured should be bound to disclose that which he does
not appreciate to be material.81 Further, in Lambert v. Co-operative
Insurance Society, Ltd., all three judges in the Court of Appeal took the
opportunity to criticise the prudent insurer test.82 Lawton and Cairns L.J.J.
went so far as to call for Parliamentary intervention to address the
injustices caused by the harshness of the duty.83 Briefly, the facts
concerned an insured, Mrs. Lambert, who claimed under a household “all
risks” policy that she and her husband had held for some nine years.84
Neither at the commencement of the policy nor on its subsequent renewals
had the insurers asked whether they had any criminal convictions.85 When
a claim was made for £311.00, representing the value of items of jewellery
that had been lost, the insurers avoided liability on the basis that the
criminal convictions of Mr. Lambert for, amongst other things, handling
stolen cigarettes and stealing shirts, had not been disclosed.86 In fact he
was in prison at the time of the claim and could not, therefore, have been
responsible for the loss.87
In the event, Mrs. Lambert’s appeal against the decision of the trial
judge who found in favour of the insurers failed.88 MacKenna J.,
delivering the leading judgment in the Court of Appeal, took the law to be
that stated by the Law Reform Committee in its 1957 report,89 namely that
the “question in every case is whether the fact not disclosed was material to
the risk, and not whether the insured, whether reasonably or otherwise,
81
[1970] 1 Q.B. 311, 319.
82
[1975] 2 Lloyd’s Rep. 485, 491-93 (A.C.).
83
Id. at 492-93.
84
Id. at 486.
85
Id.
86
Id.
87
Id.
88
Lambert, [1975] 2 Lloyd’s Rep. at 491.
89
LAW REFORM COMMITTEE, CONDITIONS AND EXCEPTIONS IN INSURANCE
POLICIES, supra note 15.
believed or understood it to be so.”90 Nevertheless, the judge went on to

express considerable sympathy for Mrs. Lambert and stated that he hoped
the insurers “would act decently if, having established the point of
principle, they were to pay her. It might be thought a heartless thing if they
did not, but that is their business, not mine.”91 As a matter of principle, the
decision in Lambert is, of course, correct, although nowadays a consumer-
insured would be able to refer the issue to the UK Financial Ombudsman
Service rather than the courts, which, as will be seen, does not follow the
strict law.
Considerable anxiety has also been expressed over the need for an
insured to disclose allegations of dishonesty which, in fact, are false. The
conundrum which arises here was identified by Forbes J. in Reynolds and
Anderson v. Phoenix Assurance Co. Ltd., who explained that the rule
applied only to unfounded allegations.92 If the allegation was true, the
insured was bound to disclose that he had committed the fraud and
disclosure of the allegation added nothing.93 Forbes J. noted that “the only
occasion on which the allegation as an allegation must be disclosed is when
it is not true. This appears to me to be a conclusion so devoid of any merit
that I do not consider that a responsible insurer would adopt it and nor do
I.”94
However, against this, the view of Colman J. in Strive Shipping
Corp. v. Hellenic Mutual War Risks Association (Bermuda) Ltd.,95 reflects
the orthodox approach taken towards the disclosure duty:
If an allegation of criminal conduct has been made against

an assured but is as yet unresolved at the time of placing
the risk and the evidence is that the allegation would have
influenced the judgment of a prudent insurer, the fact the
90
Lambert, [1975] 2 Lloyd’s Rep. at 489. MacKenna J. was particularly
influenced by the opinion of the Privy Council in Mutual Life Insurance Co. of
New York v. Ontario Metal Products Co. Ltd., [1925] A.C. 344, 351-52, to the
effect that the test, as laid down in the Marine Insurance Act, 1906, 6 Edw. 7, c.
41, § 18, is whether the non-disclosed fact would have influenced a reasonable
insurer to decline the risk or to have stipulated for a higher premium.
91
Lambert, [1975] 2 Lloyd’s Rep. at 491.
92
[1978] 2 Lloyd’s Rep. 440, 460 (Q.B.D.).
93
Id.
94
Id.
95
[2002] EWHC 203.
allegation is unfounded cannot divest the circumstances of

the allegation of the attribute of materiality.96
But, having held that the allegation was material, Colman J. nevertheless
mitigated his finding by holding that for the insurers to persist at trial in
taking the point, in the face of evidence that pointed to the suggested facts
being totally false, would be contrary to their obligation of good faith.97 It
is noteworthy that recently, the Court of Appeal in North Star Shipping Ltd.
v. Sphere Drake Ins. plc expressed sympathy for Forbes J.’s views and
urged the insured to argue that allegations of dishonesty which were
unrelated to the risk were immaterial.98 Ultimately though, it felt
constrained by authority to reject the contention which Waller L.J. stated he
otherwise “might be tempted to follow.”99 However, on the issue of
whether the impecuniosity of the insured was a material fact, Waller L.J.
stated that, “the non-payment of premium is either material on its own or
not, and since it seems to go to the owner’s credit risk, and not to the risk
insured, I would have thought it was not material.”100 In so finding, Waller
L.J. admitted that he was placing a significant limitation on section 18(2) of
the 1906 Marine Insurance Act given that this was plainly a material fact
which went to the decision of a prudent underwriter whether or not to
underwrite the risk.101
With respect to the insurers right of avoidance, the judiciary has
also displayed considerable tenacity in its condemnation of the results
which necessarily flow from the exercise of the remedy. To take just one
recent example, in Kausar v. Eagle Star Ins. Co. Ltd., Staughton L.J.
stated:
96
Id. In North Star Shipping Ltd. v. Sphere Drake Insurance plc, [2005]
EWHC 665, Colman J. again took this view. See also Brotherton v. Aseguradora
Colseguros S.A. (No. 2) [2003] EWCA Civ. 705 (Mance, L.J.) (discussed infra
notes 105 and 132); The Dora, [1989] 1 Lloyd’s Rep. 69, 93-94 (Q.B.D.) (Phillips,
J.); March Cabaret Club & Casino Ltd. v. London Assurance, [1975] 1 Lloyd’s
Rep. 169, 175-77 (Q.B.D.). Cf. Norwich Union Ins. Ltd. v. Meisels [2006] EWHC
2811 (Q.B.) (taking a narrower view of materiality).
97
Strive Shipping Corp. [2002] EWHC 203. See infra note 193.
98
[2006] EWCA (Civ.) 378.
99
Id.
100
Id.
101
Id.
Avoidance for non-disclosure is a drastic remedy. It

enables the insurer to disclaim liability after, and not
before, he has discovered that the risk turns out to be a bad
one; it leaves the insured without the protection which he
thought he had contracted and paid for...I do consider there
should be some restraint in the operation of the doctrine.
Avoidance for honest non-disclosure should be confined to
plain cases.102
However, the weight of the case law and the force of the 1906 Marine
Insurance Act inevitably present considerable hurdles to judicial
intervention. Nevertheless, the subject of non-disclosure and the insurers’
remedy has not escaped the attention of law reform agencies.
A. LEGISLATIVE REFORM: A FALSE DAWN
In 1978, the Law Commission was given the opportunity to review

non-disclosure.103 This was to be carried out in the light of a proposed EEC
Directive on the Co-ordination of Legislative, Statutory and Administrative
Provisions relating to Insurance Contracts, the object of which was to
harmonise the law in the Community.104 Of particular concern to the U.K.
was the recommendation that the proportionality principle should be
adopted. Under French law, for example, an insurer is obliged to pay the
proportion of the claim which the actual premium paid bears to the
premium which would have been payable if the material facts had been
disclosed.105 In this way, any additional risk and the loss attributable to that
additional risk is, in effect, borne by the insured. A more complex set of
provisions was adopted in the proposed EEC Directive. Article 3.3(c) of
the Proposed Directive dealt with the insurer's right in respect of innocent
non-disclosure.106 It adopted the principle of proportionality only where
non-disclosure is due to fault (short of fraud) on the part of the insured.107
102
[1997] C.L.C. 129, 132-33.
103
See THE LAW COMMISSION, supra note 15, § 1.1.
104
Id. § 1.2. In fact, the Directive did not materialise.
105
CODE DES ASSURANCES, art. L113-9.
106
See THE LAW COMMISSION, supra note 15, § 4.2.
107
Id. § 4.3.
In a case where the non-disclosure is not due to fault, the insurer would
remain liable for any loss.108
The Law Commission concluded that proportionality was
unworkable in England and Wales.109 It stated that the principle gives no
guidance as to how the insured’s entitlement is to be computed if the
insurer would have either declined the risk, imposed additional terms on
the insured, narrowed the risk via exclusion clauses or imposed or
increased “an excess.”110 Further, the Commission found that
proportionality provides no solution where knowledge of the undisclosed
facts would have led the insurer to decline the risk altogether.111 Whatever
the merits of this strident dismissal, such hurdles have not prevented the
Ombudsman importing the principle into the range of his remedies.112 That
aside, the Law Commission did endorse the views expressed by way of
obiter in Lambert to the effect that the law should be changed.113 It found
that the insured’s duty of disclosure can give rise to grave injustice and
there was, notwithstanding the protestations of the insurance industry, “a
formidable case for reform.”114 In essence, the Commission recommended
a substantially revised duty of disclosure that, had it been implemented,
would have resulted in shifting the focus away from the “prudent insurer”
as the determinative test of materiality.115 It proposed a modified duty of
disclosure for both consumers and businesses whereby an insured would be
required to disclose those facts that a reasonable person in the position of
the applicant would disclose.116 However, an insured’s individual personal
characteristics would not be taken into account.117
Despite early optimism that legislative reform would follow the
Law Commission’s recommendations, this soon petered out. While there
was an initial flurry of activity by the DTI (now DBIS), the impetus for
108
Id.
109
Id.
110
Id.
111
Id. § 4.5-4.6.
112
The Ombudsman has adopted proportionality for cases of unintentional
non-disclosure and misrepresentation. See the Ombudsman Report for 1989, para
2.16-7) and the Annual Report for 1994, para 2.10. See infra text accompanying
note 121.
113
THE LAW COMMISSION, supra note 15, § 4.44.
114
See id. § 3.23.
115
Id. § 6.2.3.
116
Id. § 4.47
117
Id.
reform ground to a halt no doubt because of the intense lobbying by the

insurance industry. Some six years after the Report was laid before
Parliament, the Secretary of State for Trade and Industry, Mr. Channon,
said, by way of a written response to a question asking for a progress report
on the reforms, that he was convinced that self-regulation, through the
Statements of Practice issued by the Association of British Insurers, would
meet the need of protecting private insureds from the harsher aspects of the
disclosure duty.118
118
Paul Channon, Secretary of State, Written Answers (Commons) of 21
Feburary 1986, Insurance Contracts (Feb. 21, 1986) (transcript available at
http://hansard.millbanksystems.com/written_answers/1986/feb/21/insurance-
contracts). The Secretary of State said,
The insurers have informed me that they are willing to

strengthen the non-life and long-term statements of insurance
practice on certain aspects proposed by the Department. These
concern the limitation of the duty of disclosure, warranties,
disputes procedures and, in the case of the long-term statements,
the payment of interest on life insurance claims. The statements
apply to insurance taken out by private consumers…These
changes are in the right direction. I am well aware of the
arguments, advanced amongst others by the representatives of
consumers, in favour of legislation on non-disclosure and breach
of warranty. But I consider that on balance the case for
legislation is out-weighed by the advantages of self-regulation so
long as this is effective. I look to all insurers, whether or not
they belong to the Association of British Insurers which has
promulgated the statements, to observe both their spirit and their
letter. In the light of the insurers' undertakings I do not consider
there is any need for the moment to proceed with earlier
proposals for a change in the law….
The Statements of Practice, first issued in 1977 (revised in 1986), covered

General and Long Term Insurance. The General Statement came to an end in
January 2005 when it was incorporated into the Code of Business Conduct Rules
(ICOB) by the Financial Services Authority. The ICOB adopts the language of the
ABI’s Statements. In essence, rule 7.3.6 provides that except where there is
evidence of fraud, an insurer should not avoid a claim by a retail customer on the
ground of non-disclosure of a fact material to the risk that the customer could not
reasonably be expected to have disclosed. It is also noteworthy that in the field of
motor insurance the right of insurers to avoid liability to a third party is
substantially restricted by section 152 of the Road Traffic Act 1988. The Long
The move towards self-regulation was reinforced by the industry

establishing and financing the Insurance Ombudsman in 1981. Rather than
following the common law rules relating to avoidance, the Ombudsman
seeks to reach a decision that he considers to be a “fair and reasonable”
solution to a dispute.119 In reaching this objective, the criteria taken into
account include whether the non-disclosure was deliberate or innocent.120
He has also sought to mitigate the draconian consequences of “inadvertent”
non-disclosure by, for example, requiring insurers to pay a proportion of a
claim that the premium actually fixed bears to the premium that would
have been charged had the fact been disclosed.121 The question remains,
however, whether the general law should adopt the same approach, and if
so, should the position of the consumer be separated out from that of the
commercial insured.
Subsequent investigations have arrived at conclusions very similar
to those of the Law Commission. For example, in 1997 the National
Consumer Council embarked on a thorough review of the disclosure duty
among other areas of insurance law. Its report, Insurance Law Reform:
The Consumer Case for a Review of Insurance Law, written by Professor
Birds, recommended that the consumer-insured’s duty of disclosure should
be restricted to facts within his or her knowledge which either he or she
knows to be relevant to the insurer's decision or which a reasonable person
Term Statement remains in place. Long term insurance is governed by COB which
has no equivalent provision to ICOB 7.3.6.
119
See, e.g., INSURANCE OMBUDSMAN, ANNUAL REVIEW 1992/93, ¶¶ 6.48-
6.55 (1993). The Insurance Ombudsman is now part of the Financial Ombudsman
Service, regulated by Part XVI of the Financial Markets and Services Act, 2000, c.
8, §§ 225-34. The Association of British Insurers Statements will continue to
inform the approach of the Ombudsman. See Financial Ombudsman Service, Non-
Disclosure in Insurance Cases, OMBUDSMAN NEWS, May-June 2005, at 8,
available at http://www.financial-ombudsman.org.uk/publications/ombudsman-
news/46/46.pdf.
120
Financial Ombudsman Service, supra note 119, at 8.
121
See INSURANCE OMBUDSMAN, ANNUAL REVIEW 1988/89, ¶ 2.17. See also
Financial Ombudsman Service, Insurance Complaints Involving Non-Disclosure,
OMBUDSMAN NEWS, Apr. 2003, at 9, available at http://www.financial-
ombudsman.org.uk/publications/ombudsman-news/27/27.pdf. See also JOHN
LOWRY & PHILIP RAWLINGS, INSURANCE LAW: CASES AND MATERIALS 129-230
(Hart Publishing 2004). See also P.M. North, Law Reform: Processes and
Problems, 101 L.Q.R. 338, 349 (July 1985); John Birds, The Reform of Insurance
Law, 1982 J. BUS. L. 449, 450, 454 (Nov. 1982).
in the circumstances could be expected to know to be relevant.122 More

recently, in January 2001, the British Insurance Law Association
established a sub-committee to examine contentious areas of insurance law
and to make recommendations to the Law Commission “as to the
desirability of drafting a new Insurance Contracts Act.”123 Its report,
Insurance Contract Law Reform, published in September 2002, also
endorsed the reasonable insured test, for example, whether a reasonable
insured would have considered the undisclosed matter to be material to a
prudent insurer.124 As will be seen, the most recent report by the English
and Scottish Law Commissions broadly follows this proposal.125
Although the legislature has not responded to these calls for
reform, the issue has not escaped the attention of the courts. Recent case
law suggests that there is a distinct shift in the judicial focus and that the
attention of the judges is being channelled along several lines of
investigation. For example, particular attention is being directed towards
the requirement of inducement as a determinant of non-disclosure together
with a wider-visioned approach being adopted towards the role of the
insurer during the disclosure process.126 As commented above, this may be
seen as adding content to the insurers’ duty of good faith and in this regard,
attention is now also being directed towards the exercise of the remedy of
avoidance.
IV. JUDICIAL INTERVENTION: REDRESSING THE BALANCE
The opportunity for an authoritative review of the insured’s duty of

disclosure came before the House of Lords just over ten years ago in Pan
Atlantic Insurance Co. Ltd. v. Pine Top Insurance Co.127 The defendant
122
John Birds Insurance Law Reform: the Consumer Case for a Review of
Insurance Law, National Consumer Council, 1997.
123
Id.
124
British Insurance Law Association, Insurance Contract Law Reform (Sept.
2002).
125
See THE LAW COMMISSION, supra note 17.
126
[1995] 1 AC 501.
127
Pan Atl. Ins. Co. Ltd. v. Pine Top Ins. Co. Ltd., [1995] 1 A.C. 501, 505.
Both parties were insolvent by the time the case reached the appellate courts. Id.
However, it proceeded as a friendly action because of the perceived loss of
business being suffered by the UK, and London in particular, as a result of the
overly insurer-friendly approach being adopted in relation to non-disclosure. Id. It
was hoped that the House of Lords would redress the balance. Id.
reinsurers had written excess of loss policies for three years.128 Their
defence to a claim arising out of losses suffered in the third year was based
on the inadvertent failure to disclose the extent of losses occurring in the
first two years.129 The issues were first, should materiality be measured by
reference to whether its ‘influence’ on the prudent insurer's judgment was
‘decisive’, or should some lesser degree of impact be sufficient?130 Second,
where there has been non-disclosure of a material fact, must it induce the
actual insurer to enter into the contract?131
With respect to the first issue Lord Mustill, with whom Lords Goff
and Slynn concurred, could see no good reason for departing from the
principle which had guided insurance law for more than 200 years.132 Lord
Mustill stated that disclosure was not limited to matters which would have
caused the prudent insurer to decline the risk or increase the premium but
rather the insured’s duty to disclose “all matters which would have been
taken into account by the underwriter when assessing the risk . . . which he
was consenting to assume.”133 On the question of statutory interpretation,
the majority view was that since Parliament had left the word “influence”
in section 18(2) unadorned by phrases such as “decisively” or
“conclusively,” it must bear its ordinary meaning.134 His Lordship stated
that “. . . this expression clearly denotes an effect on the thought processes
of the insurer in weighing up the risk, quite different from words which
might have been used but were not, such as ‘influencing the insurer to take
the risk.’”135 The majority decision therefore was to reject the “decisive
influence” test. The position remains that a circumstance is material and
must be disclosed even though the prudent insurer, had he known of the
fact, would have insured the risk on the same terms.136
128
Id. at 519.
129
Id. at 520.
130
Id. at 516-17
131
Id. at 517-18.
132
Id. at 536.
133
Pan Atl. Ins. Co. Ltd., [1995] 1 A.C. at 538. Lord Mustill thus rejected a
test based upon the decisive influence of the non-disclosed/misrepresented fact: “I
can see no room within [the principle] for a more lenient test expressed solely by
reference to the decisive effect which the circumstance would have on the mind of
the prudent underwriter.” Id. at 536.
134
Id. at 531.
135
Id. See also id. at 517 (speech of Lord Goff).
136
It is noteworthy that Lord Lloyd, in a powerful dissent, agreed with the
appellants' submission that there should be a twofold test under which the insurer
In relation to the second issue, however, the House of Lords

unanimously held that in Pan Atlantic Insurance Company v. Pine Top
Limited, the non-disclosure of a material fact, as with misrepresentation,
must induce the particular insurer to enter into the contract.137 In reaching
this conclusion, their Lordships were clearly influenced by the argument
that the 1906 Act codified the common law, and given that inducement was
a requirement under the general law which provides for rescission of a
contract, the Act must be taken as having the same effect.138 In language
that resonates with that of Lord Mansfield’s in so far as it traverses the
terrain of misrepresentation and non-disclosure (and in so doing aligns the
requirement of inducement with both vitiating factors), Lord Mustill stated
that:
I conclude that there is to be implied in the Act of 1906 a

qualification that a material misrepresentation will not
entitle the underwriter to avoid the policy unless the
misrepresentation induced the making of the contract,
using “induced” in the sense in which it is used in the
general law of contract.139
must show that a prudent insurer, if aware of the undisclosed fact, would either
have declined the risk or charged a higher premium and that the actual insurer
would have declined the risk or required a higher premium. Id. at 554. See also
John Birds & Norma J. Hird, Misrepresentation and Non-Disclosure in Insurance
Law - Identical Twins or Separate Issues, 59 M.L.R. 285, 285 (1996).
137
Pan Atl. Ins. Co., [1995] 1 A.C. at 551. In essence, the House of Lords
were injecting into the law on non-disclosure a requirement of causation analogous
to the “but for” test familiar to tort lawyers. See id. at 551 (Lord Mustill’s
reference to causative effect). In his reasoning, Lord Mustill gave prominence to
the decision of Kerr J. in Berger v. Pollock, [1973] 2 Lloyd’s Rep. 442, in which
the judge stated the principles in a way that suggested that the insurer could avoid
the policy only if he had in fact been influenced by the non-disclosure. Id. at 463.
138
Pan Atl. Ins. Co., [1995] 1 A.C. at 549.
139
Id. Where there is a material misrepresentation, there is a rebuttable
presumption of inducement. See Redgrave v. Hurd, (1881) 20 Ch.D. 1, 21; Smith
v. Chadwick, (1884) 9 App. Cas. 187, 196. Lord Mustill went on to add that, “As
a matter of common sense however even where the underwriter is shown to have
been careless in other respects the assured will have an uphill task in persuading
the court that the witholding or mistatement of circumstances satifying the test of
materiality has made no difference.” Pan Atl. Ins. Co., [1995] 1 A.C. at 551. See
also Svenska Handelsbanken v. Sun Alliance & London Ins. plc, [1996] 1 Lloyd’s
Rep. 519.
Lord Goff, concurring, thought that the need to show inducement

on the part of the actual insurer addresses the criticisms directed against the
harshness of the duty.140 He reasoned that it was the absence of this
requirement that prompted the call for the test of materiality to “be
hardened into the decisive influence test.”141 However, this concession
must be measured against the view of Lord Mustill, vigorously opposed by
Lord Lloyd, that there should be a presumption of inducement.142 In
essence once objective materiality is established a presumption that the
actual insurer was induced triggers. On the facts, the House of Lords held
that the non-disclosed losses were so obviously material that inducement
could be inferred.143
The presumption of inducement has had a chequered reception in
the case law following Pan Atlantic.144 From the perspective of the
insured, modern case law has sought to preserve the benefit of the
requirement by limiting the scope of its presumption to exceptional cases
only. For example, in Marc Rich & Co. A.G. v. Portman, Longmore J.
suggested that unless there was good reason for the underwriter not to give
evidence, the presumption would simply not arise.145 The judge stressed
140
Pan Atl. Ins. Co., [1995] 1 A.C. at 516-18.
141
Id. at 518.
142
Id. at 542, 571.
143
Id. at 562.
144
The only cases in which the presumption of inducement has been applied
are those involving market subscriptions when one member of the following
market has been unavailable to give evidence of his own state of mind. See, e.g.,
Talbot Underwriting Ltd. v. Nausch, Hogan & Murray Inc. [2006] EWCA Civ.
889, [2006] 2 Lloyd’s Rep. 195; Toomey v. Banco Vitalicio de Espana SA de
Seguros y Reasseguros [2003] EWHC 1102, [2004] Lloyd’s Rep. I.R. 354; St. Paul
Fire & Marine Co. (U.K.) Ltd. v. McConnell Dowell Constructors Ltd., [1996] 1
All E.R. 96, [1995] 2 Lloyd’s Rep. 116. See also Sirius Int’l Ins. Group Corp. v.
Oriental Assurance Corp., [1999] 1 All E.R. (Comm.) 699, [1999] Lloyd’s Rep.
I.R. 343; Ins. Corp. of the Channel Islands v. Royal Hotel Ltd., [1998] Lloyd’s
Rep. I.R. 151.
145
[1996] 1 Lloyd’s Rep. 430. See also Sirius Int’l Ins. Group Corp., [1999]
1 All E.R. (Comm.) 699, where, in relation to misrepresentation, Longmore J. also
stressed that it is for the insurer to prove inducement. The judge did, however,
recognize that the onus of proof is difficult to discharge. Id. In his Pat Saxton
Memorial Lecture, “An Insurance Contracts Act for a New Century”, delivered on
5 March 2001 to the British Insurance Law Association, Longmore J gave death as
an example of a good reason for failing to give evidence. Sir Andrew Longmore,
Pat Saxton Memorial Lecture at the British Insurance Law Association: An
that in cases where the court is in doubt, the defence of non-disclosure

should fail because “[a]t the end of the day it is for the insurer to prove that
the non-disclosure did induce the writing of the risk....”146 Further, in
Assicurazioni Generali SpA v. Arab Ins. Group, the Court of Appeal took
the view that although the non-disclosed (or misrepresented) fact need not
be the sole inducement operating on the insurer, it must cause the actual
insurer to enter into the contract.147 Significantly, the majority of the court
followed earlier decisions to the effect that the insurer must give evidence
as to his state of mind.148 This, therefore, gives the insured the opportunity
to cross-examine the insurer with a view to demonstrating that he was not
induced by the non-disclosed fact but would have entered into the contract
on the same terms had there had been full disclosure of all material facts.149
If the underwriter fails to give evidence, without "good reason",
inducement will not be made out.150 Clarke L.J. summarised the position as
follows:
1. In order to be entitled to avoid a contract of insurance or

reinsurance, an insurer… must prove on the balance of
probabilities that he was induced to enter into the contract
by a material non-disclosure or by a material
misrepresentation.
2. There is no presumption of law that an insurer… is

induced to enter in the contract by a material non-
disclosure or misrepresentation.
3. The facts may, however, be such that it is to be inferred

that the particular insurer… was so induced even in the
absence from evidence from him.
Insurance Contracts Act for a New Century? (Mar. 5, 2001), available at

http://www.bila.org.uk/lecture_scripts/lectsaxt.asp.
146
Marc Rich & Co., [1996] 1 Lloyd’s Rep. at 442.
147
[2002] EWCA Civ. 1642, [2003] 1 W.L.R. 577. Following the decision in
St. Paul’s Fire & Marine Ins. Co. v. McConnell Dowell Constructors Ltd., [1995]
1 All E.R. 96, 104.
148
Assicurazioni Generali SpA [2002] EWCA Civ. 1642.
149
Id.
150
Marc Rich & Co. AG, 1 Lloyd’s Rep. at 442.
4. In order to prove inducement the insurer or reinsurer

must show that the non-disclosure or misrepresentation
was an effective cause of his entering into the contract on
the terms on which he did. He must therefore show at least
that, but for the relevant non-disclosure or
misrepresentation, he would not have entered into the
contract on those terms. On the other hand, he does not
have to show that it was the sole effective cause of his
doing so.151
Insurers must, therefore, establish that the non-disclosed or misrepresented

fact was an effective cause, although not necessarily the only cause, of their
agreement to underwrite the risk. The link between materiality and
inducement has thus been severed. Where the underwriter does give
evidence, he will need to demonstrate a causal link between the
presentation of the risk and its acceptance.
Further inroads have recently been made into the notion of utmost
good faith. This has been done by first adopting a narrow, insured friendly
approach towards the requirement of inducement, one incident of which
has led to the courts to re-examine the defence of waiver; and second, by
aligning the exercise of the remedy of avoidance for non-disclosure with
the insurers’ duty of good faith. In tandem with this process, the courts
have also been examining the role of insurers as recipients of information
during the disclosure process. The effect is that the burden imposed by the
duty is being recalibrated so as to strike some balance between the
respective obligations of the parties especially at the stage when the risk is
being presented for underwriting. A further and significant recent
development relates to the vexed question of whether the insured’s duty of
good faith continues after the insurance contract has been concluded so that
it again triggers when a claim is made under the policy.
A. REFINING INDUCEMENT
The emphasis now being placed on the need to demonstrate

inducement has been bolstered by the radical step taken by the court in
Drake Insurance plc v. Provident Insurance plc, to the effect that, in
deciding whether the non-disclosed fact had induced the insurer to enter the
contract, it is necessary to examine what would have happened had full
151
[2003] 1 W.L.R. 577.
disclosure been made.152 In this case, the insurer sought to avoid the motor
policy on the basis that it would have charged a higher premium had a
speeding conviction been disclosed.153 The insured had disclosed an earlier
fault accident, but failed to disclose before the present policy was
concluded that it had been reclassified as a no-fault accident.154 The
majority of the Court of Appeal held that, even if the conviction had been
disclosed, information would have come to light that the earlier accident
had not been the insured’s fault and this would have resulted in the
proposal being accepted at a normal rate of premium.155
Rix L.J., delivering the leading judgment,156 stated that the issue is
not what actually happened, but what would have happened had the
speeding conviction been declared. To prove inducement, the insurer,
Provident, would need to show that a higher premium would have
resulted.157 This it could not do because it was common ground that it
would not have increased the premium if the earlier accident had been no-
fault: “So the question resolves itself into this: if the conviction had been
mentioned, would the question of the status of the accident have been
discussed? It seems to me to be very likely that it would have been… .”158
It is noteworthy that Rix L.J. went on to express the view that he
could see no reason in principle why an insured should not be able to rely
on facts which would have been material in his favour had they been
disclosed.159 This, after all, is the case with insurers and the logic is,
therefore, compelling. Further, this reasoning marks a clear departure from
the view expressed by Mance L.J. in Brotherton v. Aseguradora
Colseguros SA,160 to the effect that an insured is not entitled to prove what
the true position was at the time the contract was concluded as a means of
proving that a particular fact was immaterial.161
152
[2003] EWCA Civ. 1834; [2004] Q.B. 601.
153
Id.
154
Id.
155
Id.
156
Id. Clarke L.J. agreeing and Pill L.J. dissenting.
157
Id.
158
Drake Ins. [2003] EWCA Civ. 1834.
159
Id.
160
[2003] EWCA Civ. 705. See also Malcolm Clarke, Non-disclosure and
Avoidance: Lies, Damned Lies, and ”Intelligence” [2004] L.M.C.L.Q. 1.
161
Brotherton [2003] EWCA Civ. 705 (rejecting the view of Colman J. in
Strive Shipping Corp. v. Hellenic Mut. War Risks Ass’n (Bermuda) Ltd., (The
Grecia Express), [2002] EWHC 203, 213, that “an insured is, if necessary, entitled
Overall, the approach taken towards inducement by Rix L.J. has

the effect of the court placing itself in the position of the underwriter. The
result is that the court is prepared to reopen the negotiations between the
parties, certainly in relation to disclosure, and to speculate on their likely
responses.162 In effect, the court is assessing what the underwriter’s most
likely course of action would have been with full disclosure. The onus is
on the underwriter to prove that it would not have accepted the risk either
at all or on the premium actually charged. Given this proactive position
being taken towards the issue, there seems no reason in principle why the
judges should not also be able to apply the proportionality doctrine, long
harnessed by the Ombudsman, rather than continue with the all or nothing
approach of avoidance. The approach adopted by Rix L.J. towards the
determination of the particular risk in question thus renders the Law
Commission’s reasoning, in its 1980 report rejecting the proportionality
doctrine as "unworkable," less than compelling.
B. THE PRESENTATION OF THE RISK: A PRO-ACTIVE ROLE FOR

INSURERS?
This renewed focus on the requirement of inducement can be seen

as part of the overarching anxiety that the presentation of the risk should be
fair. Here, the mutuality of the good faith duty has come to the fore and the
judges have been directing their attention towards the content of the
insurers’ obligation. It will be recalled that in Carter v. Boehm, Lord
Mansfield laid particular emphasis on both the need for a fair presentation
of risk and the limits of the insured’s disclosure duty.163 In this respect, he
excluded from the realms of the duty those facts which the insurer “waives
being informed of” together with facts the insurer is presumed to know.164
to litigate the issue of the truth or falsity of known but undisclosed intelligence, in
order to argue that, if it is shown to be incorrect, the insurer would be acting in bad
faith or unconscionably in avoiding.")
162
See R. Merkin, 16 Insurance Law Monthly (2004). See also Bonner v. Cox
Dedicated Corporate Member [2004] EWHC 2963 (Comm).
163
Carter v. Boehm, (1766) 97 Eng. Rep. 1162, 1164-65.
164
Id. at 1165. Lord Mansfield explained that the insured need not disclose
… what the underwriter knows ...what way soever he came

to the knowledge. The insured need not mention what the under-
writer ought to know; what he takes upon himself the knowledge
of; or what he waives being informed of. The underwriter needs
It will be recalled that Lord Mansfield returned to the role expected of

underwriters during the disclosure process in Noble v. Kennoway, in which
he held that the insurer was under a duty to inform himself of the practices
of the trade he insures.165 Further, in Court v. Martineau, he was prepared
to draw the inference that the insurer had waived the disclosure of certain
facts by the large premium he charged for underwriting the risk in
question.166
Opportunities to consider the insurers’ duty of utmost good faith
have been rare in modern times. However, in Banque Keyser Ullman S.A.
v. Skandia (U.K.) Insurance (C.A.), Slade L.J. said that the insurers’ duty of
disclosure should
… extend to disclosing facts known to him which are

material either to the nature of the risk sought to be
covered or the recoverability of a claim under the policy
which a prudent insured would take into account in
deciding whether to place the risk for which he seeks cover
with that insurer.167
The House of Lords approved Slade L.J.’s reasoning in this respect. The
only remedy available to the insured where the insurer is in breach of duty
is avoidance ab initio. In practice, this affords little or no benefit to
insureds. An insurer’s breach will come to light when the loss has been
suffered – a time when an insured will want full recovery rather than a
return of the premium.
More recently, however, the issue has come to the fore in the
context of determining the insurer’s role during the disclosure process
not be told what lessens the risque agreed and understood to be

run by the express terms of the policy. He needs not to be told
general topics of speculation: as for instance - the under-writer is
bound to know every cause which may occasion natural perils….
Id. This particular element of the judgment was later codified, virtually verbatim,
in the Marine Insurance Act, 1906, 6 Edw. 7, c. 41, § 18(3) (Eng.).
165
(1780) 99 Eng. Rep. 326. See also Mayne v. Walter, (1782) 99 Eng. Rep.
548.
166
(1782) 99 Eng. Rep. 591. See also Drake Ins. plc v. Provident Ins. plc,
[2004] Q.B. 601 (reasoning of Rix L.J.).
167
[1990] 1 Q.B. 665 (A.C.) at 772. See also Aldrich v. Norwich Union Life
Ins. Co. Ltd., [1999] 2 All E.R. (Comm) 707.
where the insured has raised the defence of waiver. Of significance in this
respect are the views expressed by the Court of Appeal in WISE Ltd. v.
Grupo Nacional Provincial SA.168 The issue arose in the context of
commercial insurance. The defendant, Mexican insurer GNP, appealed to
the Court of Appeal against a decision of Simon J. that the claimants-
reinsurers WISE were entitled to avoid a reinsurance contract on the basis
that the presence of high-value Rolex watches in the insured consignment
of goods was not disclosed.169 This occurred as a result of a translation
error in which the watches were described as clocks.170 It was held, by a
majority, that GNP was entitled to recover.171 Although it was
unanimously held that WISE had been induced by the presentation of the
risk, Rix and Peter Gibson L.J.J. held that the reinsurers had affirmed the
policy, notwithstanding the breach of the duty of disclosure, by giving
notice of its cancellation.172 Such notice was inconsistent with any claim to
avoid the policy ab initio.173 Both judges took the view that the trial judge
had overlooked a vital email which showed that WISE were unequivocal in
cancelling the policy.174
With respect to the issue of waiver, the parties agreed that the law
was accurately set out in MacGillivray on Insurance Law,175 which, citing
CTI v. Oceanus Mutual Underwriting Association (Bermuda) Ltd.,176 states:
The assured must perform his duty of disclosure properly

by making a fair presentation of the risk proposed for
insurance. If the insurers thereby receive information from
the assured or his agent which, taken into conjunction with
other facts known to them or which they are presumed to
know, would naturally prompt a reasonably careful insurer
to make further inquiries, then, if they omit to make the
appropriate check or inquiry, assuming it can be made
168
[2004] EWCA Civ. 962.
169
Id.
170
Id.
171
Id.
172
Id.
173
Id.
174
WISE Ltd. [2004] EWCA Civ. 962. Longmore L.J. dissented on the basis
that the judge’s findings of fact could not be reversed. Id.
175
NICHOLAS LEIGH-JONES ET AL., MACGILLIVRAY ON INSURANCE LAW (10th
ed. 2003).
176
[1984] 1 Lloyd’s Rep. 476.
simply, they will be held to have waived disclosure of the

material fact which that inquiry would necessarily have
revealed. Waiver is not established by showing merely that
the insurers were aware of the possibility of the existence
of other material facts; they must be put fairly on inquiry
about them.177
Longmore L.J., with whom Peter Gibson L.J. agreed, took the view that
since the carriage of Rolex watches was a material fact which was not
disclosed, the presentation of the risk was unfair.178 The issue was,
therefore, whether the facts that were disclosed would prompt a reasonably
careful insurer to enquire whether watches were included in the shipment.
As Longmore LJ explained, the issue came down to whether or not the
insurer was “put on inquiry by the disclosure of facts which would raise in
the mind of the reasonable insurer at least the suspicion that there were
other circumstances which would or might vitiate the presentation.”179 On
the facts he held that there was nothing in the presentation of the risk that
could be said to have raised the suspicion that Rolex watches were to be
included in the consignment.180
Of particular interest for present purposes is Rix L.J.’s dissenting
judgment on this issue. In finding that there had been waiver, he placed
particular emphasis on the mutuality of the duty of utmost good faith, and
stated that the only relevant question was whether the presentation was
fair.181 This could not be judged in isolation, although an obviously unfair
presentation would rarely leave room for waiver to operate.182 The
insurers’ reaction and the issue of possible waiver had to be taken into
account.183 The question is not whether an “unfair” presentation had been
waived but whether, taking both sides into account, the presentation was
unfair or, alternatively, it would be unfair of the insurers to avoid the
contract on a ground on which they were put on inquiry and should have
satisfied themselves by making appropriate enquiries.184
177
LEIGH-JONES ET AL., supra note 175, ¶¶ 17-83.
178
WISE Ltd. [2004] EWCA Civ. 962.
179
Id.
180
Id.
181
Id.
182
Id.
183
Id.
184
WISE Ltd. [2004] EWCA Civ. 962.
Ultimately, it seems, the question is: Has the insurer been

put fairly on inquiry about the existence of other material
facts, which such inquiry would necessarily have revealed?
The test has to be applied by reference to a reasonably
careful insurer rather than the actual insurer, and not
merely by reference to what such an insurer is told in the
assured's actual presentation but also by reference to what
he knows or ought to know, ie. his s 18(3)(b) [of the 1906
Act] knowledge… Overriding all, however, is the notion of
fairness, and that applies mutually to both parties, even if
the presentation starts with the would-be assured.185
Rix L.J. therefore concluded that a reasonably careful insurer would have
been fairly put on inquiry given what he knew from GNP's presentation and
his general, presumed knowledge.186 The question as to what types of
clocks were being transported was one that should have been asked by the
reinsurers. He went on to state that:
If the question is instead the overriding question: Is the

ultimate assessment of GNP's presentation that it is unfair,
or would it be unfair to allow the reinsurers a remedy of
avoidance in such a case? I would answer that the
presentation was fair, and that it would be unfair to allow
reinsurers to take advantage of an error of translation in a
case where, on the evidence, an exclusion of watches
would seem to have been the obvious solution.187
For Rix L.J., the duty of utmost good faith and, more particularly, its
content insofar as it applies to insurers or reinsurers, requires them to play a
pro-active role in the disclosure process rather than relying solely upon the
insured’s presentation.188 In this respect, his approach resonates with that
taken by Lord Mansfield in Noble v. Kennoway,189 and Court v.
185
Id.
186
Id.
187
Id.
188
See Merkin, supra note 162.
189
(1780) 99 Eng. Rep. 326.
Martineau.190 Further, the clear implication seems to be that given that the
insurers or reinsurers draft the policy terms; there is more than adequate
opportunity for them to take the necessary steps to protect themselves in
relation to the risks to be underwritten.
C. ALIGNING THE INSURER’S REMEDY OF AVOIDANCE WITH THE

DUTY OF UTMOST GOOD FAITH
We have seen that the consequence of non-disclosure is to render

the insurance contract voidable, thereby entitling the insurer to avoid it ab
initio.191 Any premium paid is returnable to the insured except in cases of
fraud (unless the policy otherwise provides).192 Not surprisingly, as with
the requirements of inducement and waiver, the conditions governing the
exercise of the avoidance remedy have also attracted considerable judicial
attention of late. Again, the views expressed have not been entirely
consistent. In The Grecia Express, Colman J. suggested that the right to
avoid is conditional upon the insurer acting with “duty of the utmost good
faith.”193 He also reasoned, as commented above, that an insured is entitled
to litigate the issue of the truth or falsity of material circumstances in order
to argue that, if it is shown to be incorrect, the insurer would be acting in
bad faith or unconscionably in avoiding the policy.194 In a similar vein, in
Manifest Shipping Co. Ltd. v. Uni-Polaris Shipping Co. Ltd.,195 Lord
Hobhouse, delivering the leading speech, identifies the need for some
balance to be struck between the parties in the post-contract situation and
suggests, as did Lord Mansfield in Carter v. Boehm, that the courts should
guard against the danger of the good faith duty being turned into an
190
(1782) 99 Eng. Rep. 591.
191
Abram Steamship Co. v. Westville Shipping Co., [1923] A.C. 773, 781.
See also Glasgow Assurance Corp. v. Symondson & Co., (1911) 16 Com. Cas.
109, 121 (Scrutton J., suggesting that the only remedy available for non-disclosure
is avoidance of the contract).
192
PARK, supra note 18, at 218. The Marine Insurance Act, 1906, 6 Edw. 7,
c. 41, § 83(3)(a) (Eng.) provides, “Where the policy is void, or is avoided by the
insurer, as from the commencement of the risk, the premium is returnable,
provided that there has been no fraud or illegality on the part of the assured . . . .”
193
Strive Shipping Corp. v. Hellenic Mut. War Risks Ass’n (Bermuda) Ltd.
(The Grecia Express) [2002] EWHC 203.
194
Id.
195
[2003] 1 A.C. 469, 496-97.
instrument permitting unconscionable behaviour on the part of

underwriters. He wrote:
The courts have consistently set their face against allowing

the assured's duty of good faith to be used by the insurer as
an instrument for enabling the insurer himself to act in bad
faith. An inevitable consequence in the post-contract
situation is that the remedy of avoidance of the contract is
in practical terms wholly one-sided. It is a remedy of value
to the insurer and, if the defendants’ argument is accepted,
of disproportionate benefit to him; it enables him to escape
retrospectively the liability to indemnify which he has
previously and (on this hypothesis) validly undertaken.196
Against this, a rather more rigid view was taken towards the exercise of the
remedy in Brotherton v. Aseguradora Colseguros SA (No. 2).197 Mance
L.J. explained that the right to avoid is a self-help remedy that is exercised
without the court’s authorisation.198 He stated that avoidance for non-
disclosure is to be treated in the same way as rescission for
misrepresentation under the general law of contract, which is “by act of the
innocent party operating independently of the court.”199 In short, the court
at trial cannot reverse a valid avoidance. This affirms the orthodoxy that
holds that an insurer has an unfettered discretion to avoid the contract in
cases where there has been a breach of the duty of disclosure even where
the facts relied on, which in this case concerned allegations going as to
moral hazard, turn out to be unfounded.
Notwithstanding this strict stance, the courts have continued to
subject the conditions governing the right of avoidance to scrutiny and have
suggested that the good faith duty triggers whenever underwriters seek to
exercise the remedy. In his far reaching judgment delivered in Drake
Insurance plc v. Provident Insurance plc, Rix L.J. observed that the
doctrine of good faith should be capable of limiting the insurer's right to
196
Id. at 497.
197
[2003] EWCA Civ. 705.
198
Id.
199
Id. Note the criticism of this reasoning by Malcolm Clarke, Rescission:
Inducement and Good Faith, CAMBRIDGE L.J. 286, 287 (2004). See generally
Peter Macdonald Eggers, Remedies for the Failure to Observe the Utmost Good
Faith, L.M.C.L.Q. 249, 262-71 (2003).
avoid in circumstances where that remedy, “which has been described in

recent years as draconian,” would operate unfairly.200 He went on to note
that in recent years, there has been a realisation that in certain respects
English insurance law has developed too stringently.201 Citing Pan
Atlantic, Rix L.J. stated that leading modern cases show that the courts are
prepared to introduce safeguards and flexibility.202 Importantly, he said
that it would not be in good faith to avoid a policy without first allowing
the insured an opportunity to address the reason for the avoidance.203 He
concluded by stressing that not all insurance contracts are made by those
engaged in commerce and the widespread nature of consumer insurance
presented new problems.204 “It may be necessary to give wider effect to the
doctrine of good faith and recognize that its impact may demand that
ultimately regard must be had to a concept of proportionality implicit in
fair dealing.”205
Turning to the mutuality of the duty of utmost good faith, the Court
of Appeal sought to refine the insurer’s duty further. Rix and Clarke L.J.J.
took the view that if the insurer had actual knowledge or blind-eye
knowledge of the fact that the accident was “no-fault,” it would have been
a matter of bad faith had the insurer avoided the policy.206 Rix L.J. left the
question open whether “something less than such knowledge would have
been enough to qualify an unrestricted right to avoid.”207 Pill L.J.,
however, discusses blind-eye knowledge and points out that there must be a
suspicion that relevant facts exist and a deliberate decision not to make an
200
[2003] EWCA Civ. 1834, [2004] Q.B. 601, 628. For criticism of the
court’s finding that the insurer’s right of avoidance was subject to good faith, see
Neil Campbell, Good Faith: Lessons from Insurance Law, 11 N.Z. BUS. L.Q. 479
(2005). It was, of course, inevitable that the Court of Appeal in North Star
Shipping Ltd. v. Sphere Drake Ins. plc, [2006] EWCA Civ. 378, did not permit the
amendment to the notice of appeal and, therefore, did not have the opportunity to
comment on this aspect of Drake.
201
Drake Ins. [2003] EWCA Civ. 1834.
202
Id.
203
Id., [2004] Q.B. at 628, 630.
204
Id. at 629.
205
Id.
206
Id.
207
Drake Ins., [2004] Q.B. at 630. Clarke L.J., denied the existence of “a
general principle that insurers must always give the insured an opportunity to
address the reason why they are considering avoidance.” Id. at 642.
enquiry.208 He goes on to state, “failure to make any enquiry of the insured

before taking the drastic step of avoiding the policy was in my judgment a
breach by the insurer of the duty of good faith.”209 While he concluded
that they did not have blind-eye knowledge, nevertheless, he took the view
that they were under a duty of good faith to inform the insured of their
intention to avoid the policy and to give him an opportunity to update them
with respect to the accident.210
It is noteworthy that the issue of the (re)insurer’s good faith duty
has recently been considered in relation to express terms contained in the
insurance policy. In Gan Insurance Co. Ltd. v. Tai Ping Insurance Co.
Ltd., it was held that claims co-operation clauses are subject to a rationality
test which owes its origins to the insurers’ duty of good faith.211 Although
there was no implied term that approval of a settlement could not be
unreasonably withheld, the right to withold approval was not unqualified.212
It must be exercised in good faith.213 Thus, (re)insurers are under a duty of
good faith in exercising their rights under a claims co-operation clause, and
must not, therefore, arbitrarily refuse to approve a settlement.
IV. REINING IN THE NOTION OF THE INSURED’S POST-

CONTRACTUAL GOOD FAITH DUTY
The move away from the position permitting an unfettered right of

avoidance must also be viewed against the wider landscape in which the
nature and scope of the insured’s post-contractual duty of good faith has
similarly been tested by the courts. Until recently, the judicial consensus
was that the insured’s duty of good faith revived in appropriate
circumstances during the currency of the contract.214 In this regard the
208
Id. at 649.
209
Id.
210
Id.
211
[2001] EWCA Civ. 1047 (Mance L.J.). See also Eagle Star Ins. Co. Ltd.
v. Cresswell [2004] EWCA Civ. 602.
212
Gan Ins. Co. Ltd. [2001] EWCA 1047.
213
Id.
214
Manifest Shipping Co. Ltd. v. Uni-Polaris Shipping. Co. Ltd. [2003] 1
A.C. 469 (Leggatt L.J.) (approving the judgment of Rix J. in Royal Boskalis
Westminster NV v. Mountain, [1999] Q.B. 674, in which the judge cited the
following passage from MALCOLM A. CLARKE, THE LAW OF INSURANCE
CONTRACTS 708 (2d ed. 1994):
position was stated Mathew L.J. in Boulton v. Houlder Bros. & Co., that it
“is an essential condition of the policy of insurance that the underwriters
shall be treated with good faith, not merely in reference to the inception of
the risk, but in the steps taken to carry out the contract.”215 The underlying
rationale for this view was explained by Hoffmann L.J. in Orakpo v.
Barclays Ins. Services Co. Ltd.216
I do not see why the duty of good faith on the part of the
assured should expire when the contract has been made.
The reasons for requiring good faith continue to exist. Just
as the nature of the risk will usually be within the peculiar
knowledge of the insured, so will the circumstances of the
casualty; it will rarely be within the knowledge of the
insurance company. I think that the insurance company
should be able to trust the assured to put forward a claim in
good faith.217
Sir Roger Parker agreed with Hoffmann L.J..218 There the Court of Appeal
held that a claim which is fraudulent entitles the insurer to avoid the contact
ab initio irrespective of whether there is a term in the policy to that
effect.219 However, Staughton L.J. differed.220 While he thought this
As regards insurance contracts, the duty of good faith

continues throughout the contractual relationship at a level
appropriate to the moment. In particular, the duty of disclosure,
most prominent prior to contract formation, revives whenever the
insured has an express or implied duty to supply information to
enable the insurer to make a decision. Hence, it applies if cover
is extended or renewed. It also applies when the insured claims
insurance money: he must make ‘full disclosure of the
circumstances of the case’ [citing Shepherd v. Chewter, (1808) 1
Camp. 274, 275 (Lord Ellenborough)].
215
[1904] 1 KB 784, 791-92. See also Shepherd v. Chewter, (1808) 170 Eng.
Rep. 955, 957.
216
[1994] C.L.C. 373 (A.C.).
217
Id. at 383.
218
Id. at 384.
219
See generally Id.
220
Id.
should certainly be the case where the policy so provided,221 he was not
convinced this should necessarily be the case in the absence of such term:
I do not know of any other corner of the law where the

plaintiff who has made a fraudulent claim is deprived even
of that which he is lawfully entitled to… True, there is
distinguished support for such a doctrine… But we were
not told of any authority which binds us to teach that
conclusion.222
It is settled that if the insured makes a fraudulent claim, he or she

will not be able to recover.223 The consequence is that the insured will
forfeit all rights under the policy.224 However, the question whether the
policy can be avoided ab initio so that the insurer can recover any
payments made with respect to an earlier loss, or whether the insurer
should be restricted to recovering only from the date of the fraudulent
claim, has received inconsistent responses by the courts.225 For example, in
Black King Shipping Corp. v. Massie (The Litsion Pride), it was held that a
fraudulent claim could amount to breach of section 17 of the 1906 Act,
thereby entitling the insurer to avoid the contract ab initio.226 However, the
courts have recently been retreating from this position by placing limits on
the insureds’ post-contractual good faith duty. In Orakpo, the majority of
the Court of Appeal was of the view that where an insured’s claim is
fraudulent to a “substantial extent,” it must fail.227 The meaning of
“substantial” was considered by the Court of Appeal in Galloway v.
221
Albeit, subject to the Unfair Terms in Consumer Contracts Regulations.
222
Orakpo, [1994] C.L.C. at 382-83 (citing, inter alia, Britton v. Royal Ins.
Co., (1866) 176 Eng. Rep. 843; Black King Shipping Corp. v. Massie (The Litsion
Pride), [1985] 1 Lloyd’s Rep. 437).
223
Britton v. Royal Ins. Co., (1866) 176 Eng. Rep. 843, 844.
224
Id.
225
See, e.g., Gore Mut. Ins. Co. v. Bifford, [1987] 45 D.L.R (Ltd.) 763; Reid
& Co., Ltd. v. Employers’ Accident & Livestock Ins. Co. Ltd., (1899) 1 F. 1031;
The Litsion Pride, 1 Lloyd’s Rep. 437. See also MALCOLM A. CLARKE, THE LAW
OF INSURANCE CONTRACTS, supra note 214, at 859; Malcolm A. Clarke, Lies,
Damned Lies, and Insurance Claims: the Elements and Effect of Fraud, [2000]
N.Z.L.R. 233, 251.
226
The Litsion Pride, 1 Lloyd’s Rep. at 438.
227
Orakpo, [1994] C.L.C. at 385 (Hoffman L.J., Sir Roger Parker). This is
merely the application of the de minimis rule.
Guardian Royal Exchange (U.K.) Ltd.228 The claimant’s premises were

burgled and he claimed under a home contents policy some £16,133.94 (the
probable true value of the loss) and an additional £2,000 for a computer.229
In fact, there had been no loss of a computer and a receipt which the
claimant produced as evidence of purchase was a forgery.230 Further, when
completing the proposal form for this insurance some five months prior to
the claim, he had failed to disclose a conviction for obtaining property by
deception.231 Lord Woolf M.R., stressing that the policy of the law must be
to deter the making of fraudulent claims, stated that the phrase
“substantial:”
is to be understood as indicating that, if there is some

immaterial non-disclosure, then of course, even though that
material non-disclosure was fraudulent dire consequences
do not follow in relation to the claim as a whole; but if the
fraud is material, it does have the effect that it taints the
whole.232
For Lord Woolf M.R., the whole of the claim must be looked at in
order to determine whether the fraud is material.233 On the facts of the
case, the claim for £2,000 amounted to some 10 per cent of the whole.234
This was an amount that was thought substantial and it therefore tainted the
whole claim.235
228
[1999] Lloyd’s Rep. I.R. 209, 213.
229
Id. at 210.
230
Id.
231
Id.
232
Id. at 213.
233
Id.
234
Galloway, [1999] Lloyd’s Rep. I.R. at 213-14.
235
Id. Millett L.J., however, disagreed with this reasoning. He said that the
determination of whether or not a claim is “substantially” fraudulent should not be
tested by reference to the proportion it bears to the entire claim. Id. at 214. To do
so “would lead to the absurd conclusion that the greater the genuine loss, the larger
the fraudulent claim which may be made at the same time without penalty.” Id. In
Millett L.J.’s view, the size of the genuine claim should not be taken into account.
Id. All that matters is that the insured is in breach of the duty of good faith which
leaves him without cover. As a matter of policy, he added that he would not
support any dilution of the insured’s duty of good faith. Id.
The need for certainty was finally addressed by the House of Lords
decision in Manifest Shipping Co. Ltd. v. Uni-Polaris Ins. Co. Ltd. (The
Star Sea).236 While the trial judge had doubted the independent application
of utmost good faith to the claims process, the Court of Appeal, took the
view that the duty of good faith binds both the insured and the insurer when
a claim is made.237 Leggatt L.J. observed that “[i]t is less clear from the
cases whether there is a duty to disclose co-extensive with that which exists
before the contract of insurance is entered into, as opposed to a rather
different obligation to make full disclosure of the circumstances of the
claim. But that distinction matters not.”238 Leggatt L.J. went on to state
that that the insured’s duty of good faith requires that the claim should not
be made fraudulently and that the duty “is coincident with the term to be
implied by law, as forming part of a contract of insurance, that where fraud
is proved in the making of a claim the insurer is discharged from all
liability.”239 In conclusion, the judge stressed that given the draconian
remedy available to insurers where a claim is made fraudulently, there
should be no enlargement of the insured’s duty so as to encompass claims
made “culpably.”240
236
[2001] UKHL 1, [2003] 1 A.C. 469, 508-09. See also Sir Andrew
Longmore, Good Faith and Breach of Warranty: Are We Moving Forwards or
Backwards?, 2004 L.M.C.L.Q. 158, 166-71.
237
Manifest Shipping Co. Ltd. v. Uni-Polaris Ins. Co. Ltd., [1995] 1 Lloyd’s
Rep. 651, 667; aff’d, [1997] 1 Lloyd’s Rep. 360. The judge held that even if it did
operate there had to be at the very least recklessness by the insured and that the
duty came to an end once legal proceedings had been commenced as after that date
false statements were to be dealt with as part of the court’s processes rather than as
part of the claim. Id.
238
Manifest Shipping Co., [1997] 1 Lloyd’s Rep. at 371.
239
Id. See also Orakpo v. Barclays Ins. Servs., [1999] C.L.C. 373, 383
(Hoffmann L.J. stated “[a]ny fraud in making the claim goes to the root of the
contract and entitles the insurer to be discharged.”) As has been seen in Galloway,
the Court of Appeal held that the absence of an express condition providing that
where there was a fraudulent claim the policy would be void made no difference
for the duty of good faith continued long after the policy was effected and applied
to the claims process. Galloway, [1999] Lloyd’s Rep. I. R. at 211.
240
Manifest Shipping Co., [1997] 1 Lloyd’s Rep. at 372. See also Diggens v.
Sun Alliance and London Ins. plc, [1994] C.L.C. 1146 (the duty is not broken by
an innocent or negligent non-disclosure).
The House of Lords, doubting the reasoning of Hirst J. in The

Litsion Pride,241 accepted that the duty of good faith continued to apply
after the conclusion of the insurance contract but held that the claim of
fraud had not been proved. As seen above, Lord Hobhouse, noting that the
right to avoid under the Marine Insurance Act 1906, section 17 entitles the
insurer to rescind the contract ab initio, thought that were this remedy to
apply where the breach of duty occurs post-contractually, the effect would
be effectively penal.242 In his reasoning in this regard, Lord Hobhouse
could find no authority to support the notion that the duty of utmost good
faith declared by section 17 continued to bind the insured post-
contractually:
[The] authorities show that there is a clear distinction to be

made between the pre-contract duty of disclosure and any
duty of disclosure which may exist after the contract has
been made. It is not right to reason, . . . from the existence
of an extensive duty pre-contract positively to disclose all
material facts to the conclusion that post-contract there is a
similarly extensive obligation to disclose all facts which
the insurer has an interest in knowing and which might
affect his conduct.243
With respect to the majority view in Orakpo, Lord Hobhouse observed that
the decision “cannot be treated as fully authoritative in view of the
contractual analysis there adopted” with respect to the duty of good faith.244
His Lordship, stressing that the duty of utmost good faith applies only up
until the conclusion of the contract, noted that a duty to disclose
information can nevertheless arise later, during the currency of the policy,
as a result of an express or implied term.245
Recently the issue again arose in K/S Merc-Scandia XXXXII v.
Lloyd's Underwriters (The Mercandian Continent).246 The insured
241
Black King Shipping Corp. v. Massie (The Litsion Pride), [1985] 1
Lloyd’s Rep. 437, 437. See also Howard N. Bennett, Mapping the Doctrine of
Utmost Good Faith in Insurance Contract Law, 1999 L.M.C.L.Q. 165.
242
Manifest Shipping Co. Ltd. v. Uni-Polaris Ins. Co. Ltd. [2001] UKHL 1,
[2003] 1 A.C. 469, 494.
243
Id., [2003] 1 A.C. at 496-97.
244
Id. at 501.
245
Id. at 495.
246
[2001] EWCA Civ. 1275, [2001] 2 Lloyd’s Rep. 563.
submitted a forged letter to his liability insurers to assist them in defending

a claim that had been brought against the insured by a third party.247 The
purpose was to show that the insured had not entered into a contract with
the claimant third party conferring exclusive jurisdiction on the English
courts.248 The letter was found to be immaterial and the insurers were
therefore held liable.249 The Court of Appeal, aligning the duty of
disclosure during the claims process with its pre-contract counterpart, took
the view that the non-disclosed or misrepresented fact must be material and
it must induce the insurer to pay the claim.250 With respect to the remedy
available to the innocent party, Longmore L.J. explained that the right to
avoid the contract with retrospective effect is only exercisable in
circumstances where the innocent party would, in any event, be entitled to
terminate the contract for breach.251 He went on to state that:
[T]he giving of information, pursuant to an express or

implied obligation to do so in the contract of insurance, is
an occasion when good faith should be exercised. Since, . .
. the giving of information is essentially an obligation
stemming from contract, the remedy for the insured
fraudulently misinforming the insurer must be
commensurate with the insurer's remedies for breach of
contract. The insurer will not, therefore, be able to avoid
the contract of insurance with retrospective effect unless he
can show that the fraud was relevant to his ultimate
liability under the policy and was such as would entitle
him to terminate the insurance contract.252
Not surprisingingly, the issue continued to be litigated and the

Court of Appeal was soon afforded another opportunity to settle the point,
at least with some measure of clarity. In Agapitos v. Agnew (The Aegeon),
the question which Mance L.J. focused upon was whether a genuine claim
could become fraudulent because it was made fraudulently and whether, in
247
Id., [2001] 2 Lloyd’s Rep. at 567.
248
Id. at 566-67.
249
Id. at 576.
250
Id. at 573.
251
Id. at 575.
252
The Mercandian Continent, [2001] 2 Lloyd’s Rep. at 576.
consequence, the duty of utmost good faith was broken.253 Holding that the
duty did not apply to fraudulent claims so that the policy could not be
avoided ab initio, Mance L.J. went on to state the position with respect to
fraudulent devices.254 He thought that an acceptable solution would be to
"treat the use of a fraudulent device as a sub-species of making a fraudulent
claim" and to treat as relevant for this purpose
any lie, directly related to the claim to which the fraudulent

device relates, which is intended to improve the insured's
prospects of obtaining a settlement or winning the case,
and which would, if believed, tend, objectively, prior to
any final determination at trial of the parties' rights, to
yield a not insignificant improvement in the insured's
prospects - whether they be prospects of obtaining a
settlement, or a better settlement, or of winning at trial.255
The insurer is therefore discharged from liability in respect of such a

claim.256 Concluding the point, it was held that the common law rules
governing the making of a fraudulent claim (including the use of fraudulent
device) fell outside the scope of section 17 of the 1906 Act.257 Further, the
253
[2003] Q.B. 556. See also, H. Y. Yeo, Post-contractual good faith –
change in judicial attitude?, 66 M.L.R. 425 (2003).
254
Agapitos, [2003] Q.B. at 574-75.
255
Id. at 575. In Stemson v. AMP Gen. Ins. (NZ) Ltd. [2006] UKPC 30, the
Privy Council endorsed this approach.
256
The reasoning of Mance L.J. has recently been applied by Simon J. in
Eagle Star Ins. Co. Ltd. v. Games Video Co. SA (The Game Boy), [2004] EWHC
15. See also Interpart Comerciao e Gestao SA v. Lexington Ins. Co., [2004]
Lloyd’s Rep I.R. 690.
257
See also Goshawk Dedicated Ltd. v. Tyser & Co. [2006] EWCA Civ. 54,
[2006] 1 All E.R. (Comm.) 501, which held that any notion that the insured’s duty
of good faith continues post-contractually cannot be divorced from the terms of the
policy. The way in which such a continuing duty can arise is by implying a term
into the contract, on the basis that it is necessary for business efficacy, which
requires the insured to provide information in appropriate circumstances. It
therefore follows that there is no independent post-contractual good faith duty. All
post-contract issues are to be determined according to the terms of the policy and
in this respect, the decision appears to accord with Staughton L.J.’s minority view
in Orakpo v. Barclays Ins. Serv.’s, [1994] C.L.C. 373 (A.C.). See also Bonner v.
Cox Dedicated Corporate Member Ltd. [2004] EWHC 2963 (Comm.) (Morison
J.). Further, following Friends Provident Life & Pensions Ltd. v. Sirius Int’l. Ins.
Court of Appeal also went on to hold that once litigation between the
insurers and the insured has commenced, the consequences of making a
fraudulent claim or promoting a claim with fraudulent devices are
superseded by the procedural rules governing civil litigation.258
Mance L.J. was again given the opportunity to revisit the issue in
AXA General Insurance Ltd. v. Gottlieb.259 The issue was whether under
the common law rule relating to fraudulent claims, an insurer could recover
interim payments made prior to any fraud in respect of genuine losses
incurred on the claim to which the subsequent fraud related.260 The judge
rejected the submission of the insureds’ counsel to the effect that where a
genuine right to indemnity has both arisen and been subject of a payment
made prior to any fraud committed in respect of the same claim, there can
be no conceptual basis for requiring the insured to repay the sums
received.261 Mance L.J. stated that:
If a later fraud forfeits a genuine claim which has already

accrued but not been paid, the obvious conceptual basis is
that the whole claim is forfeit… If the whole claim is
forfeit, then the fact that sums have been advanced towards
it is of itself no answer to their recovery.262
The effect of counsel’s argument would be to result in the anomaly that

forfeiture of the whole claim should be restricted to the whole of the
outstanding claim only; in other words, to any part that remains unpaid as
of the date of the fraud. Mance L.J. explained that the rationale of the rule
relating to fraudulent claims is that an insured should not expect that,
Co. [2005] EWCA Civ. 601, it will probably never be the case that breach of such
a term is repudiatory of the policy as a whole. Waller L.J. expressed the
unconvincing view that perhaps a series of breaches might be repudiatory so that,
in effect, the notion of a continuing duty is dead and buried. Id.
258
See also Manifest Shipping Co. Ltd. v. Uni-Polaris Shipping Co. Ltd.,
[2001] UKHL 1, 481. But see Eagle Star Ins. Co. Ltd. [2004] EWHC 15. Simon J.
explained that this could give rise to anomalous consequences: “After litigation has
commenced an insured may advance false documentation and lie without the
drastic consequences which follow if the deployment of false documentation and
lies are less well timed.” Id.
259
[2005] EWCA Civ. 112 (Keene and Pill L.J.J., concurring).
260
Id.
261
Id.
262
Id.
should the fraud fail, he or she will lose nothing.263 The court should not,
therefore, undermine the prophylactic policy of the common law rule by
holding that forfeiture should not apply to a part of a claim that is otherwise
honest.264 Accordingly, it was held that the effect of the common law is to
forfeit the whole of the fraudulent claim so that the consideration for any
interim payments made on that claim fails.265 Such sums are thus
recoverable by the insurers irrespective of whether they were paid prior to
the fraud.266
The issue of fraudulent claims again came to the fore in Danepoint
Ltd. v. Allied Underwriting Insurance Ltd., in which a block of some
thirteen flats was damaged by a fire. 267 The insured lodged a number of
exaggerated claims together with a fraudulent claim relating to loss of
rent.268 Coulson J. subjected the authorities to thorough review. He
concluded that the duty of utmost good faith declared by section 17 of the
1906 Act does not trigger during the claims process.269 An insurer cannot,
therefore, avoid the policy ab initio on the ground of fraud.270 Where all or
part of the claim is fraudulent, or where fraudulent devices are enlisted to
promote a genuine claim, the insured will not be permitted to recover in
respect of any part of the claim.271 Mere exaggeration will not, in itself,
suffice to substantiate an allegation of fraud.272 But if the exaggeration is
wilful, or is allied to misrepresentation or concealment, it will, in the
judge’s view, probably be held to be fraudulent. In this regard, an
exaggeration is more excusable where the value of the particular claim or
head of loss in question is unclear or is a matter of opinion.273
263
Id.
264
Id.
265
AXA Gen. Ins. Ltd. [2005] EWCA Civ. 112.
266
Id.
267
[2005] EWHC 2318 (TCC). See also J. Lowry and P. Rawlings,
Fraudulent claims: framing the appropriate remedy, [2006] J. BUS. L. 339. Cf.
Stemson v. AMP Gen. Ins. (NZ) Ltd., [2006] UKPC 30.
268
Danepoint Ltd. [2005] EWHC 2318.
269
Id.
270
Id.
271
Id. at 432.
272
Id. at 438.
273
See, e.g., Orakpo v. Barclays Ins. Serv., [1994] C.L.C. 373 (A.C.).
V. THE FUTURE
The reasoning expressed in the modern cases demonstrates a

significant shift in the way the courts approach the good faith duty. The
process of recalibrating the insured’s pre-contractual duty of good faith
seen in the case law over the last decade or so is not being done in isloation
from other aspects of the insurance contract for the courts are also adding
content to the duty of good faith which the insurer owes to the insured at
the time the risk is presented, at the time when the remedy of avoidance is
exercised and at the time when insurers assert the benefit of claims clauses.
With respect to consumer insurance there is a considerable body of
Ombudsman jurisprudence to be added to the burgeoning case law. The
result is that we now have two parallel regimes governing insurance
contracts: one relating to commercial insurance and one relating to
consumer insurance. In terms of coherence, this is not satisfactory.
However, the Scottish and English Law Commissions current re-
examination of insurance offers the potential for a thorough overhaul of the
law, and in this regard it will be recalled that non-disclosure is identified as
a key issue. It was to be hoped that the exercise would seek to assimilate
the developments seen in the modern decisions into a single scheme for
both consumer and commercial insurance. Admittedly, for other types of
contracts the legislature has seen fit to distinguish between consumer and
commercial transactions,274 but, as is pointed out by Professor Clarke, for
insurance the distinction necessarily results in the adoption of arbitrary
tests based on turnover.275 In any case, English judges have shown some
reluctance in recognising such a dichotomy. For example, in Cook v.
Financial Insurance Co.,276 Lord Lloyd,277 considering the effect of a term
contained in a policy for disability insurance effected by a self-employed
builder, took the view that it “must be construed in the sense in which it
would have been reasonably understood by him as the consumer….”278
In framing a suitable model for the duty of good faith, an
appropriate balance needs to be struck between the economic costs of
reform and the benefits, including social benefits, which a more balanced
regime will bring. Such considerations must also be weighed against the
274
For example, in the realms of sales law and credit transactions.
275
Clarke, supra note 199, at 288. See supra text accompanying note 9.
276
[1998] 1 W.L.R. 1765.
277
Both Lord Steyn and Lord Hope agreed.
278
Cook, [1998] 1 W.L.R. at 1768.
objective of improving the competitiveness of the insurance market.279

This is not to under-estimate the difficulties of framing a solution. Any
such reform needs to avoid excessive interference with commercial
practices and avoid introducing uncertainty into the law.280 Over twenty
years ago, the Australian Law Reform Commission (hereafter, the ALRC)
published its wide-ranging report on insurance law,281 which unlike the
1980 report of its English counterpart, led to statutory reform by way of the
Australian Insurance Contracts Act 1984 (hereafter, ICA 1984).282 The
ALRC recommended a new test for the determination of the insured’s duty
of disclosure, namely that the duty should be tested by what the insured
knew, or what a reasonable person in the insured’s circumstances would
have known, to be relevant to the assessment of the risk.283 The ALRC
considered that this formulation of the duty was more consistent with the
limits of the insured’s duty to exercise utmost good faith.284 It also thought
that the formulation would achieve a fairer balance between insured and
insurer than would the more objective test recommended by the English
Law Commission in 1980.285
In reaching its conclusions, the ALRC took into account a number
of factors which offer important lessons for the investigation now
underway in the UK. It found that fairness to the insured can only be
achieved if both insurers and the law which regulates the insurance
relationship are sensitive to the literacy, knowledge, experience, and
279
This objective informed the deliberations and recommendations put
forward by the Australian Law Reform Commission. AUSTRALIAN LAW REFORM
COMMISSION, INSURANCE CONTRACTS, Rep. No. 20, at xxi (1982).
280
See report of the NEW ZEALAND LAW COMMISSION, supra note 6, ¶ 10.
281
See AUSTRALIAN LAW REFORM COMMISSION, supra note 279. The Report
is regarded as authoritative in the interpretation of the Insurance Contracts Act,
1984 (Austl.). See Ferrcom Pty. Ltd. v. Commercial Union Assurance Co. of
Austl. Ltd., (1993) 176 C.L.R. 332, 340. By virtue of §§ 13-14 of the 1984 Act,
utmost good faith is an implied term that applies to both parties to the contract.
Thus, breach of the duty is a breach of contract giving rise to damages or to an
estoppel and not to avoidance ab initio.
282
Most of the 1984 Act’s provisions came into operation on 1 January 1986.
283
See AUSTRALIAN LAW REFORM COMMISSION, supra note 279, ¶ 24, at
xxix. See also ICA 1984, supra note 281, §§ 21 and 21A (as amended).
284
AUSTRALIAN LAW REFORM COMMISSION, supra note 279, ¶ 328, at 202.
285
Id. ¶ 182, at 110.
cultural background of insureds.286 More particularly, it was emphasised

that the law should recognise the modern conditions in which insurance is
marketed.287 Nowadays, insurance contracts are concluded with a
minimum of formality and so, subject only to the principle of good faith,
insurers should take individual members of the relevant market as they find
them. The ALRC found that the existing duty of disclosure imposes
obligations which many prospective insureds, acting in the utmost good
faith, are unable to discharge.288 Indeed, in the current market place
marketing methods are adopted which increase the risk of non-disclosure,
and where intermediaries are not involved, there is no one to bring to the
insured's attention the breadth of the disclosure obligation. For reasons of
cost and competition, proposal forms are often kept to a minimum,
especially so where direct marketing of insurance products is used whereby
policies are purchased by means of computer-based communications
systems. Taken in the round, these developments increase the risk of
innocent non-disclosure. A modern regime should therefore take account
not only of the various subjective factors affecting insureds, but also of the
diverse methods enlisted by insurers to transact with their prospective
customers. A test of disclosure based on the twin attributes of the actual
insured together with the reasonable insured strikes the optimum balance in
maintaining a single test, albeit dual-limbed, for both consumer and
286
Perhaps surprisingly, these factors were specifically excluded by the Law
Commission’s final report in 1980. That said, it should be noted that the ICA 1984
differs from the ALRC’s formulation of the duty. Section 21(1)(b) refers to
matters that “a reasonable person in the circumstances” could be expected to know
(emphasis added). Notwithstanding the pure objectivity of this statutory
formulation, it has received positive endorsement beyond the shores of Australia.
See, e.g., INSURANCE LAW REFORM: THE CONSUMER CASE FOR A REVIEW OF
INSURANCE LAW, supra note 15; BRITISH INSURANCE LAW ASSOCIATION, REFORM
OF INSURANCE CONTRACT LAW (2006); and more recently it has received broad
support from the English and Scottish Law Commissions, supra note 17. It has
also been welcomed by the New Zealand courts. For example, in State Insurance
v. McHale, [1992] 2 N.Z.L.R. 399, 415, Richardson and Hardie Boys J.J.
concluded that: “[t]he law in New Zealand as to materiality and the duty of
disclosure is not satisfactory. It can lead to uncertainty and injustice…. The test of
the reasonable assured has much to commend it. The Australian legislation
adopting that test … could well be followed in this country.” See also the remarks
of Cooke P, [1992] 2 N.Z.L.R. at 404. See also Quinby Enter. Ltd. v. Gen.
Accident Ltd., [1995] 1 N.Z.L.R. 736, 740.
287
AUSTRALIAN LAW REFORM COMMISSION, supra note 279, ¶ 183, at 111.
288
Id.
commercial insurance.289 Although insurers often have exclusive recourse

to data relevant to particular types of risks, they do not possess superior
knowledge in relation to factors peculiar to the particular risk sought to be
insured for this generally lies within the province of the insured.
Therefore, an insured under this test would be required to prove the
existence of any circumstances which he or she relies on to reduce the
scope of the duty of disclosure.
As seen above, a particular feature of the modern English case law
is the emphasis now being placed on the way in which insurers seek to
exercise the remedy of avoidance. It will be recalled that while insurers are
entitled to avoid the contract ab initio, the judges have expressed
considerable unease over the draconian consequences suffered by insureds.
In Drake Insurance, Rix L.J. addressed the issue in wider terms than most
in calling for regard to be had to the concept of proportionality.290 As noted
above, the Law Commission’s 1980 report expressly rejected the
proportionality doctrine on the basis that it was unworkable.291 It also
rejected a “nexus test” whereby the insurers would be required to
demonstrate that the undisclosed fact is in some way connected to the
loss.292 In reaching this conclusion the Law Commission reasoned that:
all considerations relating to non-disclosure must focus on

the moment when a proposal for insurance is put forward
and either accepted on certain terms or rejected, in either
event by reference to what the insurer judges to be the
quality of the risk. The technique - one might almost say
the art - of good underwriting is to judge all the factors
affecting an offered risk at this moment, when the
underwriter must then and there assess its quality on the
basis of his experience, as though he were considering the
289
Such a subjective/objective form is now the accepted test for determining
the appropriate standard of care for directors. See, e.g., Norman & Anor. v.
Theodore Goddard & Ors., [1992] B.C.C. 14. This appears in the statutory
statement of directors’ duties contained in the Companies Act, 2006, c.46, § 174
(Eng.).
290
[2003] EWCA Civ. 1834, [2004] Q.B. at 628, 629.
291
For the Law Commission’s reasoning in this regard, see supra text
accompanying note 15.
292
See THE LAW COMMISSION, supra note 15, ¶¶ 4.91-4.97.
overall impression given by a “still photograph” of the risk

at this point.293
While the ALRC had some sympathy for the misgivings expressed
by the Law Commission over the difficulties of proof in relation to
proportionality and causation, it did not think these were insurmountable.
The ALRC saw no reason why in most cases insurers would not be able to
establish, “whether from rating guides, from its instructions to its agents or
staff or from its prior conduct, the nature and extent of the loss which it had
suffered.”294 While conceding that it would sometimes be difficult to
establish how it would have reacted to additional moral, as distinct from
statistical, risks the ALRC concluded, in a robust statement of principle,
that:
difficulties of proof cannot be avoided if a proper balance

is to be reached between the interests of the insurer and
those of the insured. It is quite plainly contrary to the true
principle of uberrima fides to impose on the insured a
burden which far exceeds the harm which he has done.
The insurer should not be entitled to any redress which
exceeds the loss which it has in fact suffered. That is the
basic principle which lies behind the law of damages, both
in contract and in tort.295
It therefore recommended that the nature and extent of the insurer’s redress
should depend on the nature and extent of the loss which it has suffered as
a result of the insured’s conduct and that it should no longer be entitled to
avoid a contract, and a heavy claim under that contract, merely because it
has suffered a small loss as a result of non-disclosure.296 This certainly
293
Id. ¶ 4.94.
294
AUSTRALIAN LAW REFORM COMMISSION, supra note 279, ¶ 194.
295
Id.
296
Id. See Insurance Contracts Act, 1984, §§ 28–31 (Austl.) (as amended).
See Advance (NSW) Ins. Agencies Pty. Ltd. v. Matthews, (1987) 4 ANZ Ins. Cas.
60–813 (Young J.) (New South Wales Sup. Ct.); cf. Lindsay v. CIC Ins. Ltd.,
(1989) 16 N.S.W.L.R. 673; Ferrcom Pty. Ltd. v. Commercial Union Assurance Co.
of Austl. Ltd., (1989) 5 ANZ Ins. Cas. 60-907 (Giles J.) (New South Wales Sup.
Ct.). See also Alexander Stenhouse Ltd. v. Austcan Inv. Pty. Ltd., (1993) 112
A.L.R. 353 (Mason C.J., Deane, Dawson, Toohey & McHugh J.J.). Section 31 of
the ICA provides that a court may disregard an avoidance and order the insurers to
aligns the position in insurance with fundamental principles underpinning

the law of damages both in contract and tort. More particularly, it also
reflects the approach taken in the general law of contract towards
restricting the right to rescind a contract for innocent misrepresentation.297
Accordingly, the ALRC took the view that the right of insurers to avoid a
contract from its inception should be abolished except for cases of
fraudulent non-disclosure on the basis that avoidance ab initio was wholly
disproportionate to the harm caused by an insured’s non-fraudulent non-
disclosure.298 Rather, the insurer should be able to cancel the contract
prospectively and be entitled to adjust a claim to take into account the loss
actually suffered by it as a result of the insured’s breach of the disclosure
duty.299 As to the question of assessing damages, the ALRC favoured the
approach taken by the common law in claims for misrepresentation
whereby damages for a breach of duty would depend upon what the insurer
would have done had it known of the true facts.300 Any other remedies
available to the insurer would depend on the response it would have made
if it had known of the undisclosed material facts. For example, if it would
have declined the risk outright, the insurer’s loss is equivalent to the
amount claimed against it. If it would have accepted the risk albeit at a
higher premium, its loss is the difference between the actual and notional
premiums. If it would have accepted the risk but on different terms,
whether at the same premium or not, its loss is the difference between its
liabilities under the actual and notional contracts.
Not surprisingly, modern English decisions such as Drake
Insurance and Rix L.J.’s dissent in WISE Ltd. illustrate the anxiety of the
modern judges to address unfair dealings, certainly with respect to the
pay some or all of a claim consistently with what is just and equitable in the
circumstances. It should be noted that the case law on sections 28-31 of the 1984
Act is confusing. The courts have developed their own principles on the
assessment of damages.
297
See Misrepresentation Act, 1967, c. 7, § 2(2) (Eng.). It should be noted,
however, that the Australian courts have frequently questioned whether the
analogy between damages for breach of contract and damages for
misrepresentation/non-disclosure is strictly correct.
298
AUSTRALIAN LAW REFORM COMMISSION, supra note 279, § 194.
299
Id. The Australian Law Reform Commission recommended reforms along
similar lines for the law relating to misrepresentation.
300
Id.
exercise of avoidance by insurers.301 From the pre-contractual standpoint,

we have also seen the courts aligning insurance law with general contract
law in terms of the assimilation of non-disclosure with misrepresentation
whereby the insurers must have been induced by the non-disclosed fact.302
This process, an objective of which appears to be a rebalancing of the
respective rights and obligations of the parties, is not restricted to the
modern case law. It should be viewed in tandem with the regulatory
approach now being adopted by the UK’s Financial Services Authority
through its Insurance: Conduct of Business (ICOB). The ICOB, while less
prescriptive in relation to commercial policies than it is for consumer
contracts, nevertheless requires insurers to treat commercial customers
“fairly” and not to unreasonably reject claims.303
From the European perspective it is of interest to view these
developments against the background of the Directive on Unfair
Commercial Practices which was approved on 11 May 2005.304 The
Directive is aimed at providing a uniform and comprehensive standard for
prohibiting unfair commercial practices. Although it is aimed at the
position of consumers, including those affected by unfair commercial
practices of insurers, there seems no reason in principle why the objective
of the Directive (taken with the emphasis placed on ‘fairness’ in the ICOB)
together with the current judicial thinking on the content of the duty of
good faith borne by both parties, should not underpin the Law
Commissions’ current review of insurance law. Indeed, in this respect, a
degree of optimism is warranted. In September 2006, the English and
Scottish Law Commissions published an Issues Paper on Misrepresentation
and Non-Disclosure which was intended to promote discussion and
301
Drake Ins. plc v. Provident Ins. plc [2003] EWCA Civ. 1834, [2004] Q.B.
601; WISE Ltd. v. Grupo Nacional Provincial SA [2004] EWCA Civ. 962.
302
Whether this is particularly novel is another question. Certainly, in Carter
v. Boehm, (1766) 97 Eng. Rep. 1162, 1165, Lord Mansfield did not draw any sharp
distinction between them. As Professor Clarke has observed, “If I describe the
shandy that I have just bought you as lemonade, is that non-disclosure of part, the
beer, or misrepresentation of the whole?” Clarke, supra note 199, at 288.
303
Insurance: Conduct of Business Sourcebook, Financial Services Authority
Handbook, Release No. 094, § 8.1.2 (Oct. 2009).
304
Unfair Commercial Practices Directive, Council Directive 2005/29, 2005
O.J. (L 149) 22 (EC). See Hugh Collins, The Unfair Commercial Practices
Directive, 1(4) EUR. REV. CONT. L. 417 (2005); and Hugh Collins, Harmonisation
by Example: European Laws against Unfair Commercial Practices 17 M.L.R. 89
(2010).
feedback.305 The current system is criticised as “incoherent and flawed” on

the basis that insurers can avoid policies inappropriately; that consumers
are “deprived of a genuine choice between the FOS and the courts;” and
that it “requires the FOS to exercise undue discretion.”306
The initial recommendation was that the duty of disclosure in
consumer insurance should be abolished. This proposal survived the
various consultation exercises carried out by the Law Commissions and
now forms the central plank of their proposals and draft Bill which was
published in December 2009.307 Clause 2 of the Bill replaces the
consumer-insured’s duty of disclosure with the duty “to take reasonable
care not to make a misrepresentation.” This therefore removes the
consumer’s duty to volunteer information to the insurer. Instead,
consumers will be required to answer insurers’ questions honestly and to
take reasonable care that their replies are accurate and complete. If
consumers do, however, provide insurers with information which was not
asked for, they must also do so honestly and carefully. The thinking here is
that abolition of the disclosure duty would force insurers to ask the right
questions in proposal forms. The draft Bill does not require the insurer to
ask specific questions. However, clause 3(2) provides that in assessing the
reasonableness of the consumer’s answer to a question, the court (or
ombudsman) will take account “how clear, and how specific, the insurer’s
questions were.” Clause 10, amongst other things, goes on to prevent
insurers from contracting out of the provisions of the Bill. Thus, a policy
term, or a term in any other contract, is rendered void to the extent that it
would put the consumer in a worse position than under the draft Bill.
The prudent underwriter test is thus replaced with a reasonable
insured test. Schedule 1 of the draft Bill goes on to lay down the insurers’
remedies for misrepresentation. The applicable remedy should depend on
the insured’s state of mind. Where a consumer acts honestly and
reasonably the insurer will be required to pay the claim. In cases of fraud
(termed a “deliberate or reckless” misrepresentation), the insurer will be
entitled to refuse to pay the claim. In such a case the insurer will need to
prove on the balance of probabilities that the consumer knew (a) that the
statement was untrue or misleading, or did not care whether it was or not,
and (b) that the matter was relevant to the insurer, or did not care whether it
305
See The Law Commission, supra note 17.
306
Id. § 5.24.
307
CONSUMER INSURANCE LAW: PRE-CONTRACT DISCLOSURE AND
REPRESENTATION, supra note 17.
was or not. However, with respect to negligent misrepresentation, a

scheme of apportionment will apply in that both parties should be put into
the position they would have been in had the insurer known the true facts.
For example, if the insurers would have charged a higher premium, the
insured will be able to recover that proportion of his loss which
corresponds to the proportion of the proper premium actually paid.308 For
innocent misrepresentation, i.e. where the insured had reasonable grounds
for believing the truth of what is stated, the insurer will have no remedy.
The Law Commissions thus draw a distinction between consumer
and non-consumer insurance contracts. The received wisdom is that
businesses require less protection in part because businesses use expert
brokers, but also because the market for commercial insurance is
competitive and businesses can generally negotiate with insurers in a way
not available to consumers. Overall, it is provisionally recommended that
the duty of disclosure should continue to apply to non-consumer insurance
but subject to a “reasonable insured” test for materiality. This would also
apply to misrepresentation.309 It is also proposed that the same remedies as
recommended for consumer policies should be available for fraudulent,
negligent and innocent misrepresentation, although a range of questions
concerning negligent misrepresentation are put forward for discussion. A
policy statement on pre-contract disclosure and misrepresentation in
business insurance is expected to be issued in 2010.
It would be churlish not to welcome the Law Commissions’
proposals for reforming insurance contracts. More particularly, a positive
feature is that a number of their proposals resonate with those originally put
forward by the ALRC in 1982. However, perpetrating the distinction
between consumer and business insureds is open to question.310
Maintaining the division between the two does little to further the objective
308
For example, if an insurer only charged a premium of £1,000 but should
have charged £1,500, the consumer will receive two thirds of the claim. See the
Explanatory Notes to the draft Bill, A.17(3), CONSUMER INSURANCE LAW: PRE-
CONTRACT DISCLOSURE AND REPRESENTATION, supra note 17.
309
Such rules will be compulsory for business policies. See Law
Commission, Insurance Contract Law, Warranties (Nov. 2006), available at
http://www.lawcom.gov.uk/docs/Insurance_Contract_Law_Issues_Paper_2.pdf.
See also, the Law Commissions’ Consultation Paper No 182; Discussion Paper No
134, (July, 2007), Insurance Contract Law: Misrepresentation, Non-Disclosure
and Breach of Warranty by the Insured.
310
Regrettably, the Australian ICA 1984, § 21A also draws the distinction
between private and business insureds.
of constructing a coherent regime for insurance. On a more positive note,

the recommendations relating to remedies are to be welcomed especially in
so far as the Law Commissions now embrace a form of apportionment.
Other issues still remain to be addressed, not least the question of whether
or not the good faith duty continues to operate post-contractually. As seen
above, the case law has at last settled the point and no doubt the Law
Commissions will assimilate this position in their final proposals. Their
work on this is expected to begin in 2010.311 Finally, in relation to claims
there remains a significant question which the authorities have thus far
failed to address. Will a fraudulent claim bring the insurance contract to an
end so that insurers can refuse to pay a legitimate claim that is made
subsequent to that which is held to be fraudulent? Although the issue has
so far evaded judicial determination, the English and Scottish Law
Commissions now have the opportunity to grasp the nettle.
311
See http://www.lawcom.gov.uk/insurance_contract.htm (accessed 11
January 2010).
RISK DATA IN INSURANCE INTERPRETATION
Michelle Boardman∗
***
Insurance companies use facts about past risks—actuarial data—in

essential ways. The data are at the crux of creating the product and
clearing regulatory hurdles to selling the product. Given the centrality of
the data in the drafting, pricing, and legitimizing of insurance policies, it is
peculiar that courts, insurers, and policyholders tend to ignore it when the
time comes to interpret and apply a policy in court. This article imagines
the shape its use would take and considers casual empirics on why it is not
used more now.
There are three ways for actuarial data to advance interpretation and
construction. The first is in proving or disproving insurer good faith.
Actuarial data can show an insurer’s bona fides—countering the universal
underlying assumption of the swindling insurer. Comparing money taken
in (the premium calculation) with money paid out (the risks covered) can
confirm or deny a bait and switch scheme. Second, the data can prove an
otherwise abstract claim of actuarial purpose, providing the context that
resolves a nascent ambiguity. This is important because a finding of
ambiguity is four-fifths of a finding that the policyholder wins. Third,
actuarial data can reveal insurer intent—not simply a lack of bad faith but
a particular intent. Recognizing this intent brings some surprising benefits
to both consumers and insurers.
In insurance, courts are often engaged in a project that is both more than
and less than interpretation. Courts are engaged in regulation of the
insurance policy directly, dictating the clauses insurers can and cannot
enforce. Actuarial function provides the court intent on regulating the
insurance field with the policy implications of a particular ruling. The data
∗
Assistant Professor of Law, George Mason University School of Law.
Brown University, B.A. 1994, University of Chicago Law School, J.D. 1998.
Thanks to Tom Baker, Lloyd Cohen, Kimberly Moore, Tony Sebok, James
Siewert, Nicholas Quinn Rosenkranz, Alice Wellford, Hill Wellford, and
participants in the Vanderbilt University Law School faculty workshop. Thank
you to Adina Horvath and Cody Williams for excellent research support.
will allow a court engaging in regulation to consider policyholders other

than the one momentarily before it.
***
Insurance companies use facts about past risks—actuarial data—in

essential ways. The data are at the crux of creating the product (what risks
to cover, at what cost) and clearing regulatory hurdles to selling the product
(proving that they are charging enough but not too much).1 Given the
centrality of the data in the drafting, pricing, and legitimizing of insurance
policies, it is peculiar that courts, insurers, and policyholders tend to ignore
it when the time comes to interpret and apply a policy in court. This article
imagines the shape its use would take and considers why it is not used
more than sparingly now.
Behind each insurance contract there lies a city of statistics. Here
you will find answers to when a risk becomes a loss, how many people will
lose each year, and how much will be lost. Following the hundreds of
intertwining streets will lead you to the insurer’s ultimate question: how
much one would have to be paid to take on the risk of all the losses
together.
These city streets are paved with data: raw data of yesterday’s
losses and calculations of tomorrow’s. (Those craving a more elaborate
description of actuarial data can dash ahead to section I.) From these data,
insurers decide which risks to insure and which to omit. An insurance
policy can be thought of as the insurer’s attempt to explain this, to explain
in words what the equation includes and excludes. 2 But just as describing
a dance is not the dance itself, the contract language is not the underlying
truth of what risks insurers have included and excluded from their premium
and expected risk calculations.
The preliminary question is why courts engaged in interpretation
would care about this hidden city. Its secrets may seem no more relevant to
1
See 1 LEE R. RUSS & THOMAS F. SEGALLA, COUCH ON INS. §§ 2:27, 2:29,
2:31 (3d ed. 2009). See generally, Susan Randall, Freedom of Contract in
Insurance, 14 CONN. INS. L. J. 107, 124-34 (2007).
2
To whom the insurer offers its explanation is not an easy question. From a
contractual standpoint, the policyholder should be the audience. For my
explanation of why insurers might be more interested in communicating with the
courts, see Michelle Boardman, Contra Proferentem: The Allure of Ambiguous
Boilerplate, 104 MICH. L. REV. 1105 (2006).
2009] RISK DATA IN INSURANCE INTERPRETATION 159
the act of interpretation than the fact that the policy was issued on a
Tuesday. In insurance, however, courts are often engaged in a project that
is both more than and less than interpretation. Courts are engaged in
regulation of the insurance policy directly, dictating the clauses insurers
can and cannot enforce. Actuarial data can directly address the
construction and judicial regulation of insurance contracts.
This article offers three main ways in which actuarial data can
advance interpretation and construction. The first is in proving or
disproving insurer good faith, an area of heightened relevance in insurance
because insurers have a near-fiduciary duty toward policyholders.
Actuarial data can show an insurer’s bona fides—countering the universal
underlying assumption of the swindling insurer. Is an insurer arguing for a
particular reading of a provision, not because the insurer “means it” in
some sense, but because the insurer seeks to avoid paying for any loss, at
any time, under any theory? At times, actuarial data can answer this
question. Comparing money taken in (the premium calculation) with
money paid out (the risks covered) can confirm or deny a bait and switch.
Ultimately, the threat of this comparison will shrink the number of genuine
swindlers by bringing the con to light.
Second, the data use contains within it its own purpose. It can
prove an otherwise abstract claim of actuarial purpose, such as avoiding
synchronized losses across many people or avoiding moral hazard.3
Actuarial purpose supports a “reasonable reading” of a clause, perhaps
providing the context that resolves a nascent ambiguity. This is important
because a finding of ambiguity is four-fifths of a finding that the
policyholder wins.4 Actuarial purpose also provides the court intent on
regulating the insurance field with the policy implications of a particular
ruling. Courts are accustomed to looking beyond (or over and around) the
language of insurance policies to determine not just their written meaning,
but the meaning the court is willing to enforce. Kenneth Abraham’s central
3
Black’s Law Dictionary defines “moral hazard” as a “hazard that has its
inception in mental attitudes,” such as the “risk that an insured will destroy
property or allow it to be destroyed (usually by burning) in order to collect the
insurance proceeds is a moral hazard.” BLACK’S LAW DICTIONARY 786 (9th ed.
2009). In modern insurance discussions, however, it is used to refer to the danger
that a person may take less care in avoiding a hazard, knowing that insurance will
cover part of the loss. See generally, Tom Baker, On the Genealogy of Moral
Hazard, 75 TEX. L. REV. 237 (1996).
4
See TOM BAKER, INSURANCE LAW AND POLICY: CASES AND MATERIALS 466
(2003).
“A Theory of Insurance Policy Interpretation,” explores how several of the

fundamental principles of insurance law “direct that under specific
circumstances the meaning of even clear policy language must be
disregarded, not interpreted.”5 Actuarial data will allow a court engaging
in regulation to consider policyholders other than the one before
momentarily before it.6
Third, actuarial data can reveal insurer intent—not simply a lack of
bad faith but a particular intent. Courts do not ordinarily treasure the secret
meaning a seller harbors in his heart when making an offer to a buyer. But
left with complex contract language and an amorphous buyer’s intent that
does not mirror the seller’s, judges are hard pressed to turn away guidance.
Intent has a few specific uses. First, hewing more closely to
insurer intent should decrease the cost of insurance. Uncertainty about how
courts will rule increases “contract risk,” which increases premiums.7
Surveys “illustrate that uncertainty about losses and ambiguity about
probability lead to higher prices.”8 Second, knowing the intent behind a
5
Kenneth S. Abraham, A Theory of Insurance Policy Interpretation, 95
MICH. L. REV. 531, 532 (1996) (Some of the doctrines “have nothing to do with
‘interpretation’ as it is normally understood.”).
6
For a modern defense of (limited) judicial regulation, see Daniel Schwarcz,
A Products Liability Theory for the Judicial Regulation of Insurance Policies, 48
WM. & MARY L. REV. 1389 (2007). Schwarcz does not advocate the introduction
of specific actuarial data but his framework requires courts to consider whether the
policy causes “insurance harm” and to “ask whether the insurer has any legitimate
underwriting purpose for not insuring against the specific loss that befell the
insured.” Id. at 1448 (emphasis added).
7
See Tom Baker, Insuring Liability Risks, 29 GENEVA PAPERS ON RISK & INS.
128, 139-40 (2004).
8
Howard Kunreuther, Robin Hogarth & Jacqueline Meszaros, Insurer
Ambiguity and Market Failure, 7 J. OF RISK AND UNCERTAINTY 71, 79 (1993).
The 1993 Kunreuther et al. studies of insurer ambiguity presented the insurance
actors with a set probability on which “all experts agree” and contrasted that with
an ambiguous probability, defined as a “wide disagreement about the estimate of p
[the probability of loss] and a high degree of uncertainty among the experts.” Id. at
72. See also Howard Kunreuther et al., Ambiguity and Underwriter Decision
Processes, 26 J. ECON. BEHAV. & ORG 337, 342-44 (1995). Instead of a wide
range in the probability that a loss will occur, here the issue is a wide range that a
policy will be read to cover the loss. The surveys also found some evidence that
the risk premium could be double-charged: “To the extent that primary
underwriters do not recognize that the prices of actuaries may already include
clause makes it possible to assess whether the clause could have been
written more clearly. Where the intent, while legitimate, is too complex to
be conveyed well to consumers, courts must decide whether to effectuate
the intent or forbid any similar clause.
Courts adopt a regulatory view toward the health of the insurance
industry and the interests of those policyholders not before the court. If the
purpose of the clause is insurer solvency, for example, a regulatory court
may prefer to protect solvency over literal language interpretation or other
values. Courts most commonly regulate by mandating coverage and
forbidding exclusions to coverage. But regulating the substance of
insurance clauses without access to the actuarial function of those clauses is
looking left and leaping right. Insurer purpose and intent do not need to
control the outcome of a court’s decision to improve the outcome of its
regulation.
Three parties are in this game—policyholders, insurers, and courts.

Each has incentives, the pursuit of which creates externalities. Each has
motives that can be described generously or with distrust; there is benefit to
doing both. There is a fourth major player, of course, in the public, but the
public’s interest is ubiquitous. The public’s needs are partially represented
by courts (although their power to pursue policy aims is cabined), and
partially represented by policyholders, in that most individuals are
policyholders. But a policyholder after a loss may pursue his individual
compensation over a healthy insurance market and insurers can represent
the interest of the many policyholders against the few (or the future
policyholders against the demands of the present).9
adjustments for ambiguity and uncertainty, they may recommend a premium that
reflects their concerns with these factors.” Kunreuther, Hogarth & Meszaros at 75
(emphasis added).
9
“First, insurance looks at groups, at the socialization of risk through
standard contracts sold to large numbers of similarly situated persons who face an
uncertain risk. What is good for the group, as a whole, in face of uncertainty, may
not be what is good for any individual when sued.” Kent D. Syverud, What
Professional Responsibility Scholars Should Know About Insurance, 4 CONN. INS.
L.J. 17, 19 (1998). “The needs of the many outweigh the needs of the few or the
one.” Spock’s dying words, with help from Captain Kirk. STAR TREK II: THE
WRATH OF KHAN (Paramount Pictures 1982).
In addition to being policyholders, the public’s interest in insurance

can hardly be exaggerated. (The public’s intellectual interest in insurance
can, alas, readily be exaggerated.) We are all potential tort victims, to be
affected by a tortfeasor’s insurance coverage, and we are all safer or less
safe as insurers create incentives to prevent their insureds from harming us
as consumers. We are also taxpayers, whose contribution to the public fisc
may increase if government pays what insurance does not. Who would
benefit from the introduction of actuarial data in court? The conflicting
needs of the public will come into play in the normative question but the
incentives of the three direct players will help answer the descriptive
question first.
Section one is optional reading; it presents a primer on actuarial
data and premium calculation. The core of the article lies in section two.
Section three analyzes the limited ways courts already use actuarial data, in
and outside of insurance. The concluding section presents some casual
empirics on why insurers—those with direct access to the data—do not
bring it to court.
The difficulty for both insurers and policyholders is that unless the
courts are willing to adopt a rule that benefits one over the other, whether
increased use of actuarial data will harm or hurt individual players is a
factual question that can only be answered with certainty if the experiment
takes place. This article invites the experiment.
I. ACTUARIAL DATA AND PREMIUM CALCULATION

(OPTIONAL READING)
Two basic types of data can be submitted to support actuarial

claims. First, the underlying statistical data used to price a category of risk,
which would show what was being “counted” in a particular risk. This is
the actuarial data proper. The second type is the application of the
actuarial data to evaluation of a particular policyholder or risk, which is
performed by an underwriter. Either would need to be tied to the premium
charged the policyholder (or the decision to offer the coverage at all) and
the language in the policy. For example, the additional premium charged
for a hurricane endorsement would be linked to the data under the
hurricane risk, the language of the endorsement, and highlighted by the fact
that those without the endorsement pay less.
Actuaries and underwriters create and analyze the relevant data.
An actuary is a “statistician who determines the present effects of future
contingent events, esp[ecially], one who calculates insurance and pension

rates on the basis of empirically based tables.”10 “Underwriter” and
“underwriting” have their etymology in Lloyd’s Coffeehouse, where those
willing to take part in the insuring of a ship’s outgoing cargo would write
their names under a description of the ship, journey, and goods.11 Today,
underwriters use the data compiled and created by actuaries to evaluate a
specific risk seeking insurance.
Taking the simple case of life insurance, “the actuary develops the
mathematical models to be used to analyze the data, and the underwriter
applies the findings of the” medical professionals and actuaries “into the
underwriting decision made for a single proposed insured.”12 Doctors and
medical researchers report the initial facts of death by cause and age.
Actuaries collect and interpret that data, creating tables into which people
of various ages with various traits and medical histories fall. Underwriters
then apply those tables, with judgment, to an individual applying for life
insurance and recommend a certain place on the table or recommend
rejection. The price that corresponds to a place on the table will depend on
the insurer—their costs, profit expectations, and more.
How much to charge for an insurance policy is not simply a
calculation of the expected risk of loss. The premium charged includes the
actuarial premium and the non-actuarial premium.13 Underwriters
calculate the actuarial portion for a particular risk by applying “statistical
data and judgment. Probably in no case is either the sole basis for a rate;
every conceivable combination of the two is found.”14
10
BLACK’S LAW DICTIONARY 41 (9th ed. 2009).
11
Hence the individual members of the Lloyds insurance market are called
Names. See PETER L. BERNSTEIN, AGAINST THE GODS: THE REMARKABLE STORY
OF RISK 90-91 (1998).
12
F. Daniel Perkins, Can “Sound Actuarial Principles” Be Found in Life
Insurance Underwriting?, 38 TORT TRIAL & INS. PRAC. L.J. 125, 138 (2002).
13
The actuarial portion may include “feature rating,” where “[d]ata collected
over the years and intuitive hunches by insurers suggest which features are
correlated with loss rates.” “Experience rating,” which is coupled with feature
rating, “uses the loss experience of the insured during one period to help set the
premiums charged” in the next. KENNETH S. ABRAHAM, DISTRIBUTING RISK:
INSURANCE, LEGAL THEORY, AND PUBLIC POLICY 72 (1986).
14
Ralph H. Blanchard, The Basis of Premium Rates, J. OF AM. INS. (Feb.
1928), reprinted in RALPH H. BLANCHARD, RISK AND INSURANCE AND OTHER
PAPERS 159, 161 (1965). What was true in 1928 is true today.
The non-actuarial premium covers costs and profits, broadly. It

may be calculated as a percentage of the actuarial premium, as a lump sum,
or some combination of the two.15 This bifurcation does not affect the
applicability of actuarial data in court; the question remains whether the
insurer has made a calculation of the relevant loss or made a calculation
that clearly excludes the loss.
Imagine a 1% chance that your house will burn down in the next
year, leading to a $100,000 insured loss. The expected loss is $1,000 =
($100,000)(.01). For any one policyholder we can assume the insurer will
not need $1000, but either $0 or $100,000. An insurer worthy of the name
will have spread the risk across a large pool of similarly situated
policyholders and will need to be able to pay out $1000 per policyholder by
the end of the policy year. The risk being carried is therefore a $1000 risk.
A premium calculation will (usually) include this computation, but
what insurers want to determine is how much money it would take to meet
the expected cost of the risk, allowing for administrative costs and profit.
Even without cost and profit, however, the answer is not $1000. The time
value of money should allow an insurer to collect less than $1000 in
January and have $1000 or more in December.16 Fluctuation in interest
rates and the insurer’s access to investments must be predicted in order to
calculate how much needs to be collected from the policyholder in
January.17
In short, insurers would not want to report, and courts would not
want to hear, the entire premium calculation for a policy. Nonetheless, the
data for a particular clause or the actuarial function behind the structuring
of a policy could be explained without excessive fuss.18
15
See NEIL A. DOHERTY, INSURANCE PRICING AND LOSS PREVENTION 16-21
(1976).
16
Come the start of 2009, some are laughing at this statement.
17
The process gets more complex, of course. On the surface, the timeline for
loss payments in the Commercial General Liability context is often years if not
decades after the collection of the original premium. This puts a strain on even the
most careful predictions of interest rates, taxes, loss amounts, and reserves for
payment.
18
This is not to say there are not questions the data will not answer. There
are many. See infra text accompanying notes 73-74.
II. USING ACTUARIAL DATA IN INTERPRETATION
Opinions in insurance cases, read collectively, have been compared

to “a chapter out of Alice in Wonderland.”19 In some ways this is true; the
policyholder does not have to read the policy, yet it is still the basis of the
contract. The policy language gets a lay reading except when it doesn’t.
Worse, the individual words get a lay reading where there is no lay
meaning to the string of the words put together. The parties have no
mutual intent in this contract of adhesion but their individual unshared
intent might define the contract. Indeed, the policyholder’s “reasonable
expectation” of the policy might define coverage even if the expectations
come from nowhere and would be dashed by a reading of the policy. But
the policy is incomprehensible and thus does not have to be read.20 “It
takes all the running you can do, to keep in the same place.”21
Still, there is reason in insurance policy interpretation and
construction.22 Actuarial data may inject additional reason or at least cut
off some of the whims of the Red Queen23. There are holes in the existing
paradigm of insurance policy interpretation into which the data fit. Three
main avenues for the useful introduction of actuarial data are explored here:
faith, purpose, and intent.
First, the data can show insurer faith—good faith or bad faith and
potentially fraud. In particular, it could offer evidence instead of
conjecture in the otherwise presumed bait and switch scheme where the
insurer attempts to sell big and deliver small. The value here is to prove (or
19
EMERIC FISCHER, PETER NASH SWISHER, & JEFFREY W. STEMPEL,
PRINCIPLES OF INSURANCE LAW 90 (rev. 3rd ed. 2006).
20
See, e.g., Prudential Ins. Co. v. Lamme, 425 P.2d 346, 347 (Nev. 1967)
(policies are “complex instrument[s], unilaterally prepared [by the insurer], and
seldom understood by the assured.”).
21
So says The Red Queen. LEWIS CARROLL, THE ANNOTATED ALICE:
ALICE’S ADVENTURES IN WONDERLAND AND THROUGH THE LOOKING-GLASS 210
(New American Library ed. 1960).
22
“The distinction [between interpretation and construction] is, for the most
part, not dwelled upon by the courts, with the result that it is difficult to tell which
process is being employed.” JOHN D. CALAMARI & JOSEPH M. PERILLO, THE LAW
OF CONTRACTS 165 (3rd ed. 1987). The distinction will be partially honored here
because the two approaches call for distinct uses of actuarial data.
23
“Alice . . . explained, as well as she could, that she had lost her way. ‘I
don't know what you mean by your way,’ said the Queen: ‘all the ways about here
belong to me.’” CARROLL, supra note 21, at 206.
disprove) consistent behavior, consistent from drafting the policy to

adjusting the claim and, finally, to taking a litigation position. Inconsistent
behavior, either sloppy or greedy, should decline as insurers realize courts
now have insight into their inner workings.
Second, actuarial data can reveal, astoundingly, the actuarial
purpose of the structure of coverage or the actuarial pressure behind an
exclusion. Under several doctrines, courts reconstruct, misread, or refuse
to enforce a clause because the court can either discern no meaning, no
“reasonable reading,” or can discern only an illusory or devious meaning.
In these cases, the court may be blameless, left adrift by poor insurer
counsel, but the outcome is no less mistaken.24 A court with a good handle
on the purpose of a clause may still reject it, of course, but often the court
will find a purpose worth protecting. Between an insurer and a lone
policyholder in court, the lone policyholder’s needs cry out sharply.
Actuarial purpose can show where those needs are misaligned with the
needs of other policyholders.
Third, the data can show substantive insurer intent, not just a
consistent but a specific intent. It might appear from a casual review of
insurance cases that insurer intent has no relevance to courts but this is not
so. It can, should, and (sometimes) does matter to courts, regulators, and
policyholders that an insurer has a specific intent for taking a particular
action. Where the courts have little to go on, insurer intent remains, even
disfavored, one piece of the puzzle.
This value of this piece is explored below to address the question
whether “tangible property” includes information stored on computers.25
Before the computer age, commercial general liability policies covered
damage to tangible property but not intangible property. What should
courts have done in the first days of damage to electronically stored
information? It was an indeterminate question that could have had one
clear input: insurers had not calculated or charged for the risk.
Insurer intent and actuarial purpose can blend together at the edges
but they are distinct. To start, the proof may differ. To show actual intent,
in lieu of ex post excuses, courts could require an insurer to provide its own
underwriting data—proof that the insurer did in fact apply the actuarial data
24
Of course, it is foolish to expect the sins of inadequate counsel to be
remedied by asking that same counsel to present actuarial data. Unless the insurer,
the policyholder, or the court invokes the data, the lawyer who fails to explain the
function of a clause will also fail to introduce the data behind it.
25
See infra notes 71-74 and accompanying text.
available. In some circumstances, an insurer will not be able to prove use.

In others, an insurer will not want to submit its own underwriting data,
thereby forgoing the intent argument. Both could still submit general data
on the actuarial purpose of a clause, thus giving the clause another
plausible, legitimate meaning from which the court can choose.
Actuarial purpose at times will be more appropriate than intent. In
the wake of September 11, insurers collectively stated that they would not
seek to exclude coverage under their various war exclusions.26 If the
attacks had been more widespread and even more destructive, insurers
might not have taken that position. But an insurer would have a hard time
proving that its specific intent in drafting or including the war exclusion
was to avoid losses from a large scale attack by a nongovernmental
organization with whom the United States was not at war, declared or
undeclared. Where intent would fail, however, actuarial purpose might
convince a court that the social purpose of war exclusions—insurer
solvency—applied as equally to large scale terrorist attacks as to
conventional war.
Finally, those courts that see no slot in the interpretative equation
for the input of insurer intent may still be open to the purpose of a clause.
Courts adopt a regulatory view toward the health of the insurance industry
and the interests of those policyholders not before the court. If the purpose
of the clause is insurer solvency, for example, a regulatory court may prefer
to protect solvency over literal language interpretation or other values.
But to judicially regulate an insurance clause without access to the
actuarial function of the clause is to read a map while ignoring the road. It
only makes sense for a court to create coverage if it can accurately identify
desired coverage that insurers have failed to perceive. In litigation, a court
can be confident that the policyholder before the court wants coverage—of
flood damage from Katrina, for example. And so do other similarly
situated policyholders, at least ex post. But non-losing current and future
policyholders of the state might prefer a ruling that does not cause insurers
26
Under the ISO war exclusion, the insurer “will not pay for loss or damage
caused directly or indirectly by . . . (1) war, including undeclared or civil war; (2)
warlike action by a military force, including action in hindering or defending
against an actual or expected attack, by any government, sovereign or other
authority using military personnel or other agents; or (3) insurrection, rebellion,
revolution, usurped power, or action taken by governmental authority in hindering
or defending any of these.” Insurance Services Office, Cause of Loss Special
Form CP1030, available at http://www.vwcos.com/documents/ forms/CP1030-
0402.pdf.
to flee or prices to rise.27 The court faces a tradeoff, although it may not
know or care to know.
Let us assume that insurers fail to provide the best coverage the
market will bear. Policyholders as a group want and are willing to pay for
a particular coverage, call it the happy clause. Should not judges simply
award that coverage to the policyholder seeking redress in court? Courts
could read the happy clause into policies, perhaps under the reasonable
expectations doctrine because policyholders reasonably expect it. 28 After
all, if judges regularly award happy coverage, in theory insurers will soon
sell what they are being forced to provide. Sure, the premium will rise to
reflect the cost of the new coverage but recall that in this example
policyholders as a group are willing to pay.
Still, policyholders in general may ask the court not to do them any
favors. The specific preference of a given policyholder after a loss can
easily be at odds with the preference of policyholders generally. Moreover,
policyholders’ preferences are not simply cumulative. Case by case,
policyholders might approve of each of one thousand subtypes of coverage
at various prices. There are therefore one thousand “units” of coverage that
Every Policyholder would be willing to buy at its market price. However,
the Policyholder is not willing to buy all thousand because he has a limit on
how much he is willing to devote to insurance, say $1000 a year.
If he can rank the units of coverage, his ideal policy will include
unit 1 to unit 100, or wherever the premium reaches $1000. If a third party
(the courts, legislature, or regulators) instead requires that unit 20 be
expanded or that unit 500 be added, he either will have to let unit 100 go or
pay more than his maximum price. Of course, the policyholder cannot go
to the insurer in response to a court case and request the substitution; the
insurer will decide for him. If the insurer removes nothing, but adds in the
court’s mandated coverage, the insurer will eventually charge more.
Whatever insurers are selling today, we can be sure that Every
Policyholder is not in fact receiving his perfect policy at his perfect price.
The question is whether the terms that courts “add” are likely to make the
policy better or worse. To the extent courts are regulating insurance—
mandating coverage because policyholders deserve it or intractably expect
27
On fleeing insurers, see Daniel Hays, State Farm to Leave Florida
Homeowners Insurance Market, NAT’L UNDERWRITER PROP. & CASUALTY, Feb. 2,
2009, at 7. See also Richard A. Epstein, Exit Rights and Insurance Regulation:
From Federalism to Takings, 7 GEO. MASON L. REV. 293 (1999).
28
See infra notes 32-34 and accompanying text.
it—that regulatory decision is poorly made if the court does not consider
(or makes erroneous assumptions about) the actuarial facts behind
coverage.
A. INSURANCE INTERPRETATION WITHOUT ACTUARIAL DATA
A full discussion of how a judge ought to approach the

interpretation of an insurance policy is beyond the scope of this article.
However, we can loosely assume that a judge will approach the policy as a
contract and consider, in order: text, intent, purpose, and public policy.29
Differences across states and a general lack of a clear order for
interpretive inputs makes a linear statement about the principles of
insurance interpretation mostly fictional.30
The components can be identified, however, even as the order and
interaction of the components vary. The components fall into (at least)
three categories: basic rules, overrides, and penchants. These lists are
illustrative, not comprehensive. Do not read the lists horizontally; public
policy is a general override, for example, not one specific to plain
language.
Basic rules Overrides Penchants

Plain language Public policy Pro-coverage
Mutual intent Bad faith Ambiguity Probable
Purpose Industry health Pay tort victim
Contra proferentem Avoid moral hazard
Reasonable expectations Find adhesion
29
The text will be weighed in context with intent and purpose potentially
providing part of the context. It may be better to consider the policy’s purpose
before the specific intent of the parties, as evidence of intent will be extrinsic and
potentially less reliable than the self-evident purpose of a policy. On the other
hand, undue confidence in the “self-evident” purpose of particular clauses is one of
the ripest areas for the introduction of actuarial data.
30
Some of the best works on insurance interpretation are: ABRAHAM, supra
note 13, at 101; Abraham, supra note 5; James M. Fischer, Why Are Insurance
Contracts Subject to Special Rules of Interpretation?: Text Versus Context, 24
ARIZ. ST. L.J. 995 (1992); Michael B. Rappaport, The Ambiguity Rule and
Insurance Law: Why Insurance Contracts Should Not be Construed Against the
Drafter, 30 GA. L. REV. 171 (1995); Peter Nash Swisher, Judicial Interpretations
of Insurance Contract Disputes: Toward a Realistic Middle Ground Approach, 57
OHIO ST. L.J. 543 (1996).
Most of these concepts are familiar from ordinary contract

interpretation. Two basic rules bear elaboration. As early as 1923, the
Supreme Court applied contra proferentem in the insurance context: “The
rule is settled that in case of ambiguity that construction of the policy will
be adopted which is most favorable to the insured.”31 The rule takes on
subtleties32 but this definition will do here.
Next, the strong form of the reasonable expectations doctrine gives
as “the rule that the reasonable expectations of the insured should be
honored even if those expectations are unambiguously contradicted by fine-
print provisions in the policy.”33 After first being recognized by Robert
Keeton in 1970, the doctrine has enjoyed a sharp upswing followed by a
gradual relaxing of its application.34 Many courts trend toward the
doctrine’s weak form, ruling that a policyholder’s intent cannot be
reasonable if an ordinary reading of the policy would have corrected it.
For one view of how these pieces work together, consider the
“textual approach” and “modern contract theory” models proposed by the
authors of “Insurance Coverage Litigation.”35
The textual model starts with

(1) the plain and ordinary meaning of the language,
which if (2) patently ambiguous,
triggers (3) contra proferentem,
and/or (4) reasonable expectations,
and/or (5) the admission of extrinsic evidence.36
Note that this model sensibly does not attempt to claim a hierarchy for the
treatments of ambiguity.
If the language is not ambiguous on its face, the textual court may
admit extrinsic evidence to determine (a) a latent ambiguity; (b) the intent
31
Mut. Life Ins. Co. of N.Y. v. Hurni Packing Co., 263 U.S. 167, 174 (1923).
32
See Abraham, supra note 5, at 533.
33
Abraham, supra note 5, at 532 (citing ROBERT E. KEETON & ALAN I.
WIDISS, INSURANCE LAW §§ 6.1(a)-(b), at 614-21 (Practitioner’s ed. 1988)).
34
See infra Section II.C. Keeton “introduced” the doctrine in Robert E.
Keeton, Insurance Law at Variance with Policy Provisions, 83 HARV. L. REV. 961
(1970).
35
ANDERSON ET AL., INSURANCE COVERAGE LITIGATION § 2.01 (2nd ed. 2009
& Supp. 2010).
36
Id.
of the drafters; (c) any accepted usage in the industry; (d) if the
policyholder’s interpretation is reasonable in light of the insurer’s; and (e)
insurer bad faith.37 According to Anderson et al., if after this investigation
the court finds the language ambiguous, it will construe it in favor of
coverage.38 This is accurate as long as one understands that a finding of
ambiguity after the total analysis is a legal conclusion that is the same as
finding that the court will construe in favor of coverage.
A formalist judge will start with the policy text and veer off-
document with reluctance. Actuarial data could be viewed as extrinsic
evidence but it can also be viewed as a contextual fact about the purpose of
the document. Purpose, text, and logic will help resolve a budding
ambiguity or surface tension between two clauses.
As with non-insurance contracts, the “modern contract theory”
model does not completely supplant this textual approach. The tools used
are the same, but with a greater focus on reasonable expectations and the
intended use and purpose of the policy.39 On the one hand, the contextual,
functionalist approach should be less open to actuarial data because it
rejects the idea of subjective mutual intent in the standardized adhesive
contract.40 On the other hand, it also rejects a full obligation on the part of
the policyholder to read the policy; as the policy language loses its use in
determining the meaning of the contract, other sources, including insurer
intent and actuarial purpose, become more useful. Moreover, the actuarial
purpose discussed below may be of greatest use to those courts taking a
regulatory approach to policy language. Regulating courts are more prone
to be in the “modern” interpretative camp.
That said, all judges behave somewhat curiously in the insurance
realm.41 A functionalist or realist judge views a policy through its purpose
but will focus on the public purpose; this judge will be more open to
evidence about actuarial purpose than to an insurer’s specific intent. He
37
Id.
38
Id. §2.02
39
Id. § 2.01.
40
See, e.g., E. Allan Farnsworth, “Meaning” in the Law of Contracts, 76
YALE L.J. 939, 943 (1967).
41
Associate Justice of the Supreme Court, Sonia Sotomayor provides one
example of unexpected behavior in insurance cases. On the lower bench, then
Judge Sotomayor ruled “consistently, across the board in favor of insurers.”
Andrea Ortega-Wells, Supreme Court Nominee Sotomayor Shows Record of
Favoring Insurers, INSURANCE JOURNAL, June 1. 2009,
http://www.insurancejournal.com/news/national/2009/06/01/101001.htm.
may also be more inclined to construct than interpret a policy, reforming

inequitable clauses or refusing to enforce “bait and switch” clauses.
Both functionalist and formalist judges will be open to evidence of
insurer faith, good or bad. Formalist judges will “tend to focus less on the
meaning of particular words and more on the organization of the policy, its
history, and insuring intent and purpose—tactics that are more normally
employed by functionalist judges.”42 Actuarial data address all of this.
B. CONSISTENT INTENT
To generalize, the bad or good faith of an insurer is the backdrop to

many interpretive questions. To over generalize, courts tend to assume
insurer bad faith. Although faith does not always fit neatly into the
interpretation decision tree, actuarial data might make the most strides in
demonstrating an insurer’s bona fides. A leading treatise on insurance
coverage litigation states the common view that “the insurance coverage
that was promised at the time of purchase often disappears down the road
when a policyholder submits a claim.”43 Sorting insurers whom this
accurately describes from the rest will help the policyholders in court, those
not in court, and honest insurers.
The ability to show consistent intent is more important in insurance
than elsewhere. While the disappointed party to a contract may be able to
claim bad faith, insurers have fiduciaryesque responsibilities toward their
42
FISCHER ET AL., supra note 19, at 95 n.10.
43
ANDERSON ET AL., supra note 35 at xx. See also Jarrett v. E.L. Harper &
Son, 235 S.E.2d 362, 366 (W.Va. 1977). The West Virginia Supreme Court wrote:
Insurance is different from any other business. If a man goes
into a butcher shop, asks for two pounds of ground meat, and
tenders $2.89 in payment, he will expect his meat to be
forthcoming from the grinder. Imagine the scene were the
customer to ask for his meat, and be answered that the butcher
has no intention to deliver the same. ‘Where is my meat?’ the
customer would reply, possibly in other than dulcet tones. ‘I
won't give you any meat,’ replies the butcher firmly. ‘Then give
me back my $2.89 and I shall go elsewhere,’ says the customer.
‘I won't give you the $2.89 either,’ replies the butcher, ‘for you
must bring a law suit to get it from me.’ Sock! Pow! Blam!
And much property damage of a different sort. Id.
policyholders.44 A few courts go further, describing the insurer’s

obligation as fully fiduciary.45 In general, however, an insurer has some
level of duty to its policyholders, one that may “preclude the insurer from
taking positions on the meaning of a contract term that would be available
in an ordinary commercial setting.”46 Even in those jurisdictions where
insurance policies are interpreted using ordinary contract principles, there
is at least one “proviso”, that “the contract raises quasi-fiduciary
obligations owed by the insurer to the insured […] [and a]s a result, the
insurer has a common-law duty ‘not to unreasonably withhold payment of
benefits it is obligated to make under the insurance contract.’”47
The question of bad faith can go beyond contract interpretation.
Numerous states allow a claim for bad faith breach, including for the denial
of a claim, as an independent tort worthy of noneconomic and punitive
damages.48 This can be true where the law does not allow the same claim
to sound in tort for other contracts because “[i]nsurance is different. Once
an insured files a claim, the insurer has a strong incentive to conserve its
financial resources balanced against the effect on its reputation.”49 If a
jurisdiction does not allow the tort claim in insurance, policyholders may
still be able to seek additional damages under a bad faith contract claim not
permitted outside of the insurance context.50
44
See JEFFREY W. STEMPEL, LAW OF INSURANCE CONTRACT DISPUTES, §
10.01 (2d ed. 1999 & Supp. 2005); see also William T. Barker, Paul E.B. Glad,
and Steven M. Levy, Is an Insurer a Fiduciary to Its Insureds?, 25 TORT & INS.
L.J. 1 (1989).
45
See, e.g., Tank v. State Farm Fire & Cas. Co., 715 P.2d 1133, 1137 (Wash.
1986); Short v. Dairyland Ins. Co., 334 N.W.2d 384, 387 (Minn. 1983).
46
FISCHER ET AL., supra note 19, at 93. Fischer et al. describe the “tension of
sorts between [an insurer’s] duty to its shareholders or other investors (to make
money) and duty to its policyholders (to pay money).” Id. One could add the
tension with the duty to other policyholders to pay other money or money in the
future.
47
State Farm Mut. Auto. Ins. Co. v. Kastner, 77 P.3d 1256, 1259-60 (Colo.
2003) (quoting Farmers Group, Inc. v. Williams, 805 P.2d 419, 423 (Colo. 1991)).
48
FISCHER ET AL., supra note 19, at 88.
49
E.I. DePont de Nemours & Co. v. Pressman, 679 A.2d 436, 447 (Del.
1996).
50
See, e.g., Tibbs v. Great Am. Ins. Co., 755 F.2d 1370, 1375 (9th Cir. 1985)
(“[T]here is sufficient evidence to support a finding that Great American refused to
defend Tibbs in bad faith and is guilty of oppression, fraud, or malice,” triggering
punitive damages).
In the right case, evidence of “faith” can be readily found in

actuarial data. Take an insurer’s premium calculation based on risks A + B
but not risk C. If the policyholder claims or the court suspects that the
insurer sold C and is now trying to avoid paying for it, actuarial data can
show whether or not C was indeed “sold”. More concretely, if a premium
calculation were based on wind damage + hurricane damage but not flood
damage (and not flood damage during a hurricane), the insurer should be
able to prove this to the court.
This next example assumes the insurer can prove a premium
calculation based on risks A + B but not risk C; this assumption sets up the
question whether courts can ever find meaning in an insurer’s proof of a
particular actuarial calculation. First, imagine a duplicitous insurer.51 This
first insurer will be the Tricky Insurer, to be compared to the Breaching
Insurer down the road. The Tricky Insurer intends to mislead consumers
into believing they are purchasing A + B + C but when loss C occurs, it
does not intend to provide coverage. For some percentage of
policyholders, when the C claim is denied, the policyholder will take no
action and the insurer will have what they want: purchases based on the
illusion of C without having to pay out C losses.
But some (smaller) percentage of policyholders will sue for
coverage C. Assume one half wins and one half loses. The insurer has to
pay the cost of defense against all of these policyholders. It also must pay
coverage C for the winning half. Even if the percentage of policyholders
who sue is ten percent—a high number—the insurer has collected a
premium based on A + B but not C and now has to pay litigation costs and
coverage C for some percentage of policyholders.52 The insurer puts itself
in this situation:
Insurer pretends to cover: A + B + C

Insurer takes in premium charges: A + B
Insurer plans to cover: A + B
Insurer required to pay: A + B + (.05)(C) + litigation costs of C
51
Most people report being able to do this with ease.
52
Some speculate that insurers care not for litigation costs because the time
value of the money being retained while the litigation progresses pays for itself,
but this cannot be true in small cases. See Alan O. Sykes, “Bad Faith” Breach of
Contract by First-Party Insurers, 25 J. LEGAL STUD. 405, 413 (1996).
This arrangement should be profitable for few insurers, far fewer than
courts assume can profit from the pretense of coverage.
The sole potential benefit to the tricky type of duplicitous insurer is
in an increase in clients. It hasn’t increased its rates, after all. To be a
profitable ruse, the increase in sales from pretending to provide coverage C
would have to outweigh the costs of litigating and the cost of covering
some percentage of C risks. Recall, no premium was charged for C.
How does the insurer lure in new clients using risk C? Given that
most policyholders do not fully read or understand their policies, the
creation of the expectation of coverage C would either have to come from
(a) fraudulent advertising or (b) taking advantage of a pre-existing
expectation of coverage. Fraudulent advertising is a dicey proposition; it
will invite the scrutiny of state insurance commissioners and the attention
of class action lawyers. On an individual basis, an ad that promises a
particular type of coverage will solidly support the consumer’s claim of
“reasonable expectations” in court. The percentages of those who sue on C
and win will rise, leading to a rise in the percentage of consumers who sue.
This may explain why most insurance ads make vague promises of “good
neighbors” with “good hands” in the “company you keep.”53
In the second, more likely scenario, the insurer does not advertise
but relies on expectations to ensure sales. But an insurer will be hard-
pressed to increase sales based on the expectations of coverage C because
the policyholder’s pre-existing expectations of coverage are not insurer
specific—the policyholder will assume all policies include coverage for
risk C. Thus, a Tricky Insurer who does not charge more but only hopes to
increase sales based on a universal assumption is not so tricky. In other
words, this scenario seems an unlikely one for swindling insurers, but this
is an empirical question courts can investigate on their own using actuarial
data.
A more likely model for the swindling insurer is one who
calculates and charges a premium based on risks A + B + C but, when C
loss occurs, does not intend to provide coverage. (This differs from the
insurer above because the premium actually includes risk C here.)
53
State Farm Mutual Automobile Insurance Company: “Like a good
neighbor, State Farm is there.”® Allstate Insurance Company: “You’re in Good
Hands with Allstate.”® New York Life Insurance Company: “New York Life.
The Company You Keep.”®
Insurer pretends to cover: A + B + C

Insurer takes in premium charges: A + B + C
Insurer plans to cover: A + B
The benefit here is obvious: collect money for C and keep it. This
Breaching Insurer may be similarly happy where policyholders either fail
to claim the coverage or do not sue when denied. If the pretense of
providing for C is in the ether, not an ad or the policy language, the insurer
can hope that many will not sue after a loss. But if courts look to the
actuarial data when the policyholder does sue, the court should be able to
readily grant the C coverage paid for (assuming the policyholder has
suffered C).
This will be most straightforward where risk C falls within the
general grant of coverage, say to a small guesthouse near the insured home,
but is removed by an exclusion, such as if the guesthouse is used as an
office part-time.54 Many homeowners policies “cover other structures on
the ‘residence premises’ set apart from the dwelling by clear space,” but
not “[o]ther structures from which any ‘business’ is conducted,” where
“business” is defined as “[a] trade, profession or occupation engaged in on
a full-time, part-time or occasional basis.”55 If the premium charged takes
into account the loss data for all “other structures” but fails to exclude those
losses where “business is conducted,” the consumer has been charged for
and deserves compensation.
Of course, the set-up need not be so formalized or consistently
applied. Many people and judges seem to believe that insurers randomly
deny payment for losses that are covered, for which the policyholder has
paid a premium. Actuarial data should be able to give a clearer picture of
how often this happens. If courts regularly examine actuarial data, as time
goes by insurers should become less and less willing to charge for coverage
and then deny the claim.
In addition to tricks and breach, there is a middle way. An insurer
may have created an expectation of coverage from past sales; all prior
policies did cover risk C, at price P. When a new day dawns, the insurer
54
INSURANCE SERVICES OFFICE, INC., STANDARD HOMEOWNERS POLICY,
HOMEOWNERS 3 – SPECIAL FORM HO 00 03 10 00 1, 3 (1999), quoting p. 3,
Section I – Property Coverages, B.1. & B.2.c. and p. 1, Definitions, B.3.a,
available at http://www.mypolicyforms.com/ho3/default2.aspx. The definition of
“business” contains other provisions not relevant here.
55
Id. at 1, 3.
continues to charge P (or even slightly less than P) but removes risk C from
the policy language and stops compensating for C loss.56 Behold the
Tricky Breaching Insurer. The marketing of the policy is unchanged, so
applicants assume C coverage continues to exist. If the coverage is sold at
slightly less than P, it appears to the unsophisticated buyer to be a great
bargain.
What will the actuarial data show here? The analysis parallels that
of the Breaching Insurer if the premium fully includes the risk of C. If the
premium is the original P, discounted slightly, the analysis may be
identical. It depends on how the discount is taken. If the actuarial data
continues to include risk C but a slight discount is taken from the total
premium calculation at the end, the court still easily sees that C has been
charged for but denied. Breach. If the C component of the premium is
itself reduced by some percentage, as long as the basis is still C, breach is
found again.
This is not to say that the calculation cannot be too convoluted to
prove breach. Indeed, if actuarial data becomes widely used in court, some
swindling insurers can be counted upon to re-master the premium
calculations to increase convolution. On the other hand, honest insurers
who currently operate under a cloud of doubt may work to ensure their
calculations show their honesty to its best advantage.
Let us return for a moment to methods of misleading the consumer
at the front end. We can imagine either that the insurer actively attempts to
provide the illusion of coverage C, as in the first example, or that the
insurer simply takes advantage of a pre-existing consumer misconception
about that type of coverage. It would be easy for insurers to identify the
many misconceptions policyholders have about their insurance coverage.
A 2007 phone survey of policyholders found that 71% of Americans with
homeowners insurance believed they had full coverage to rebuild after a
natural disaster or fire and 72% believed their personal belongings were
covered at the cost of replacement.57 Their actual coverage had caps below
56
This is a common marketing move in the grocery aisle. You may
sporadically find your cereal box is slightly thinner or has fewer flakes, although
the price per box does not decrease. See, e.g., Jessica Dickler, The Incredible
Shrinking Cereal Box: The packaging may look the same but the amount inside has
gone down, that’s how companies try to pass on food inflation, CNNMONEY.COM,
Sept. 10, 2008, http://money.cnn.com/2008/09/09/pf/food_downsizing/index.htm.
57
Press Release, Metlife Auto & Home, Insurance Surprises: Survey Finds
Many Americans Dramatically Overestimate the Level of Insurance Protection
full coverage and would cover the depreciated value of personal items, well
below replacement cost.
Do insurers take advantage of mistakes like this? Oren Bar-Gill
and Richard Epstein recently asked the more general question “do
sophisticated sellers respond strategically to consumer misperception? In
particular, do sellers design their products, contracts, and pricing schemes
in response to consumer misperception?” 58 They consider the answer in a
debate between behavioral and neoclassical law and economics; Bar-Gill
emphasizes that the answer is ultimately empirical.59 Contrary to courts’
assumptions, both conclude that it “is probably correct” that “mistakes
about a standardized product are not sustainable.”60
Insurance may be the exception. Under a reasonable conception of
standardization, homeowners insurance policies are a standardized product
and consumer mistakes certainly persist. Whether insurers are strategically
taking advantage of these mistakes can be answered by actuarial data. And
while Bar-Gill is correct that the question is ultimately empirical, the
analysis of premium calculation presented here can provide some initial
answers even before the actuarial data is reviewed by courts.
The point here is that if the insurer attempts to swindle
policyholders by the second breaching method—charging for A+B+C but
not covering C—actuarial data should reveal the C charge and confirm that
policyholders are due C coverage. The assumption here is not that there
are only honest insurers, but that actuarial data can either keep insurers
honest or at least allow courts to identify the dishonest ones. 61
They Have (July 10, 2007) (available at http://www.metlife.com/about/press-

room/us-press-releases/2007/index.html?compID=518).
58
Oren Bar-Gill, The Behavioral Economics of Consumer Contracts, 92
MINN. L. REV. 749, 749 (2008); Richard A. Epstein, The Neoclassical Economics
of Consumer Contracts, 92 MINN. L. REV. 803, 807 (2008).
59
Bar-Gill, supra note 58, at 751.
60
Id. at 750.
61
On the flip side, when a policyholder is dishonest and commits fraud in
applying for a policy, state law often requires the insurer to prove that the fraud
was material; the insurer must be able to show that it would not have issued the
policy or would have required different terms in the absence of the policyholder’s
misrepresentation or concealment. Actuarial data can inform this question too.
See, e.g., Hill v. Allstate Ins. Co., No. 04-CV-0865-REB-CBS, 2006 WL 173693,
at *2 (D. Colo. Jan. 24, 2006) (Insurer “charged a premium based on the actuarial
risk associated with a one-household [auto] policy,” and “[a]bsent the fraudulent
misrepresentations and concealment” of the policyholders “it would not have
C. SPECIFIC INTENT
Policyholder intent has a settled place in insurance interpretation.

The traditional form of the reasonable expectations doctrine gives as “the
rule that the reasonable expectations of the insured should be honored even
if those expectations are unambiguously contradicted by fine-print
provisions in the policy.”62 Not all courts go so far, however. Many courts
are more likely to hold that a policyholder’s intent cannot be reasonable if a
reading of the policy would have corrected it.
Courts using the milder version of reasonable expectations
recognize the “rule that the policy must receive a reasonable interpretation
consistent with the parties’ object and intent.”63 Courts that are interested
in insurer intent (and expectations64) will consider the intersection of
policyholder and insurer intent, but insurers have two ongoing problems in
proving intent.
First, given the presumption of poor faith, courts are simply less
likely to believe an insurer. An insurer may be taking a litigation position
or the insurer’s lawyer may simply be cleverly supporting whatever
interpretation supports the insurer’s cause in this case. By contrast, if a
policyholder states that it expected or intended a particular coverage, or
that it would have expected coverage had it read the policy language, courts
are inclined to believe the policyholder if the claim is at all reasonable.
Second, insurers tend to point to the language of the policy for
proof of their intent. This is a textbook contractual approach but it severely
limits the insurer where the court finds the language confusing and
inaccessible to the policyholder. Moreover, it conflates what could be two
separate thrusts for the insurer: one about the natural reading of the
language and one about the value of its own intent. Both of these problems
could be solved with the introduction of actuarial data. The data are a
separate and different input for proof of the insurer’s intent. More
issued and renewed the policy as it did.”).

62
See Abraham, supra note 5, at 532 (citing ROBERT E. KEETON & ALAN I.
WIDISS, INSURANCE LAW §§ 6.1(a)-(b), at 614-21 (Practitioner’s ed. 1988)).
63
Brown v. Ind. Ins. Co., 184 S.W.3d 528, 540 (Ky. 2005) (quoting St. Paul
Fire & Marine Ins. Co. v. Powell-Walton-Milward, Inc., 870 S.W.2d 223, 226-27
(Ky. 1994)).
64
“Party expectations are a cousin of party intent.” FISCHER ET AL., supra
note 19, at 97.
importantly, the data are actual proof of intent; insurers would no longer
have to rely on their mere claims of intent in the face of skeptical courts.
If courts could be made more comfortable that an insurer’s
representation of its intent was its actual intent, the question becomes what
to do with the knowledge. On the day the contract was formed, the insurer
had a definite intent as to what the words meant; the (consumer or
unsophisticated) policyholder had a general intent—cover my losses if my
house is partially or totally destroyed. Karl Llewellyn called a consumer’s
intent in a standardized contract the “blanket assent . . . to any not
unreasonable or indecent terms the seller may have on its form, which do
not alter or eviscerate the reasonable meaning of the dickered terms.”65 In
keeping with this, not all reasonable expectations options require the
policyholder to have had an actual expectation before the loss occurred;
many courts are content to ask what the policyholder would have expected
had the question come to mind. Indeed, the policyholder is rarely asked to
explain, let alone prove, the source of his expectation.
But what if the policyholder would not have had a particular
expectation even had the question come to mind? For example, does your
homeowners policy provide coverage if the fence around your yard is
damaged?66 Does it provide coverage for limited personal effects lost in
hotel rooms while traveling? (Yes, to some extent).67 If your spouse
intentionally burns down your house, are you still covered for your half of
the loss? (It depends). 68 Does the pollution exclusion exclude coverage
65
KARL N. LLEWELLYN, THE COMMON LAW TRADITION—DECIDING APPEALS
370 (1960) (emphasis added); see also Swisher, supra note 30, at 570 n.77.
66
Your homeowners policy may cover your neighbor’s fence—if you
negligently set a fire in your backyard. See Prather v. Audubon Ins. Co., 488 So.
2d 383, 384-85 (La. Ct. App. 1986). A homeowner was recently paid $25,000 for
damage from Hurricane Katrina to a den and the fencing around his home.
Although the plaintiff was not happy with the sum he received, a portion was
attributed to the fence. See Gustings v. Travelers & Standard Fire Ins. Co., No. 07-
4443, 2008 WL 4948837 at *1 (E.D. La. Nov. 18, 2008).
67
INSURANCE SERVICES OFFICE, supra note 54, at 3 (“We cover personal
property owned or used by an ‘insured’ while it is anywhere in the world.”).
68
See Rachel R. Watkins Schoenig, Note, Property Insurance and the
Innocent Co-Insured: Was It All Pay and No Gain for the Innocent Co-Insured? 43
DRAKE L. REV. 893, 895-96 (1995). If the policy is jointly held between husband
and wife, the non-arsonist has suffered an unintentional loss; whether that loss is
covered depends on the contract and the state. See Randall, supra note 1, at 144-
45. “Whether the intentional acts of a co-insured will defeat coverage for an
for the injury to people who inhaled fumes from floor cleaner? (Depends
on the jurisdiction).69
If the policyholder would not have had a particular view, it could
not be disappointed in a lack of coverage. If the insurer has a particular
intent, one acted upon in premium calculation, it has something to lose
from an adverse interpretation. There is always a decent chance that
nonetheless a court will apply contra proferentem, construing the language
in favor of coverage. The undercurrent of contra proferentem is protection
of the policyholder against a scheming insurer and “encouragement” to the
insurer to draft more clearly. In circumstances where neither of these
motivations applies, courts that give lip service to valuing the insurer’s
intent could pay up, with the aid of actuarial data.70
For example, commercial general liability policies provide
coverage for liability arising from damage to “tangible property.”71 Data
stored on computers came along, quickly followed by possible liability
arising from its destruction. When policyholders first started seeking
innocent co-insured turns on the exclusionary language used in the policy. A

policy excluding losses caused by intentional acts of ‘any insured’ or ‘an insured’
creates a joint obligation among co-insureds and bars coverage for both the
malefactor and innocent co-insureds. Where the policy uses the words, ‘the
insured’, the obligation is several, and the exclusion applies only to the insured
who intended the act and caused injury, not an innocent co-insured.” Id. See also
N.J. Mfrs. Ins. Co. v. Carney, No. 3:04-CV-2468, 2006 WL 2092571 at *3-4
(M.D. Pa. July 26, 2006) (holding the intentional act of one insured excludes
coverage for the innocent co-insured under the language “an insured” or “any
insured”; but a wife’s arson does not stop her husband’s recovery when he is the
sole owner).
69
Compare Nav-Its, Inc. v. Selective Ins. Co. of Am., 869 A.2d 929 (N.J.
2005) (holding that the pollution exclusion in a Commercial General Liability
policy applies only to traditional environmental pollution, not to indoor chemical
use) with Fireman’s Ins. Co. v. Kline & Son Cement Repair, Inc., 474 F. Supp. 2d
799 (E.D. Va. 2007) (holding that fumes from floor sealant were “pollution” within
the definition of the pollution exclusion).
70
Some will say that the two animating factors behind contra proferentem
always apply. At a minimum, neither is at stake when older policy language is
applied to emergent and new risks.
71
The policyholder is covered for “property damage,” defined as “physical
injury to tangible property . . . or loss of use of tangible property that is not
physically injured.” INSURANCE SERVICES OFFICE, INC., COMMERCIAL GENERAL
LIABILITY COVERAGE FORM CG 00 01 10 01 15 (2000) available at http://www.
certifiedriskmanagers.com/NewISOforms.htm.
coverage for damage to computer files, courts were at a loss.72 Consider

how a court could attempt to answer the question.
First, the language is not ambiguous as much as indeterminate.
Some courts found ambiguity but most found the language unambiguously
supported the judge’s individual sense of whether electronic data was
tangible property; opinions were widely split, yea or nay. Some courts had
a better understanding than others of the physical space (however
miniscule) electronic data occupy, and these courts found the data to be
tangible property. While reading the words and applying them accurately
to a new context has some basis for support, it does require “tangible” to do
work it was not selected to do; “tangible” property was meant to set apart
intangible property, such as intellectual property. Like threats to
intellectual property, threats to electronic data come from different sources.
These threats require different risk calculations than threats to tangible
property.73
Second, the courts could have turned to the reasonable expectations
doctrine. But, in the earlier years, the policyholders either did not have any
expectation or would not have expected coverage. (The earlier years were
when the question was relevant; insurers eventually addressed the split
between tangible property and electronic data with policy options.) The
actuarial data from these years would not have included the risk of loss to
electronically stored data—an example of the lack of data providing
definite information.
Third, those courts finding the language ambiguous could have
construed in favor of the policyholder, including electronic data in tangible
property. This is not necessarily a good policyholder outcome because the
newly found coverage will cost future policyholders. Again, however,
many courts were not willing to find the language ambiguous.74
72
See, e.g., Computer Corner Inc. v. Fireman’s Fund Ins., 46 P.3d 1264, 1266
(N.M. Ct. App. 2002) (electronically stored data is “tangible property”); Am.
Online, Inc. v. St. Paul Mercury Ins. Co., 207 F. Supp. 2d 459 (E.D. Va. 2002)
(electronically stored data not “tangible property”).
73
Of course, threats to physical property also threaten the physical storage of
electronic data. When fire destroys a computer, it destroys any data stored on that
computer but electronic data are subject mainly to electronic threats.
74
Try this experiment. Show the policy language on tangible property to
several friends and ask each one (a) if the clause covers electronically stored data
and (b) if the question is close, i.e. if the language is ambiguous as applied to
electronic data. You will find that people vary in their answer to (a) but do not
believe (b).
Fourth, courts could have considered the insurer’s intent, not as

asserted but as shown by the actuarial data. Insurers had not been counting
electronic data in the tangible property loss statistics. They were not
charging for the risk of loss to electronically stored materials. In other
words, one of the parties had a specific intent about “tangible property” and
relied on that intent in its initial performance under the contract. Barring a
public policy conclusion that insurers should have been on the loss for not
addressing electronic data in the policy language at all, the evidence of this
specific insurer intent would have provided useful, perhaps definitive,
interpretive guidance.
Insurer intent thus could have given courts a tool for solving the
“tangible property” dilemma without resorting to individual judge’s
happenstance first impressions. But there is another benefit to construing
language in keeping with insurer intent, at least some of the time. Bringing
the interpretation of a clause closer to the coverage intended when the
policy was issued decreases what Tom Baker has called the “contract risk”.
75
The contract risk is part of the insurers’ risk equation. Baker defines it
as “the risk relating to the drafting and interpretation of insurance
policies.”76 Baker partitions the risk liability insurers take on into:
(1) the baseline risk,

(2) the developments risk (“relating to developments that change
the rate or cost of loss”),
(3) the contract risk, and
(4) the financing risk.77
As with the other three risks, the higher the contract risk, the more
expensive it will be to cover and the more policyholders will have to pay.
Another way to think about contract risk is as a species of
ambiguity aversion for which insurers will charge a risk premium. Surveys
of actuaries, underwriters, and reinsurance underwriters suggest strong
aversion to loss ambiguity. These surveys “illustrate that uncertainty about
75
See Baker, supra note 7, at 139-40.
76
Id. at 128, 130. Kenneth Abraham calls this an “uncertainty tax.” Kenneth
S. Abraham, The Insurance Effects of Regulation by Litigation, in REGULATION
THROUGH LITIGATION 212, 222-23 (W. Kip Viscusi ed. 2002).
77
Baker, supra note 7, at 128-30.
losses and ambiguity about probability lead to higher prices.” 78 The

research suggests that the risk premia for ambiguity is inefficient and
results in higher prices to policyholders.
There can be good reason for knowingly increasing contract risk.
If a clause is avoidably unclear or misleading, for example, an insurer
should bear the risk its confusion causes.79 Courts want to motivate
insurers to decrease the level of contract risk the insurers cause
themselves.80 Even clear language does not eviscerate contract risk; a
clearly written clause may go unenforced because it is against public policy
as applied to the particular facts at hand. In other words, the optimum
contract risk is not zero. If the risk can be decreased appropriately,
however, it will benefit all parties by lowering the price of insurance and
decreasing unnecessary litigation.81
D. ACTUARIAL PURPOSE
A primary and simple reason to show actuarial purpose is to

explain why an insurer includes a particular exclusion. Courts often
behave as though exclusions are mere traps for unwary policyholders, not a
decision by an insurer (in keeping with other insurers) that a particular risk
is uninsurable. Actuarial data would be able to demonstrate the legitimate
business purpose behind these “natural” exclusions.
Other exclusions often fall into the categories of public policy
exclusions or exclusions used by insurers to fight moral hazard. Intentional
act exclusions fall into both. These exclusions bar coverage for harm that
results from an intentional act of the policyholder, although whether both
the act and its result must be intentioned differs among jurisdictions.
In Arizona, for example, there is a wrinkle in the application of
intentional act exclusions. The Arizona courts have “abandoned” the
contra proferentem approach, believing that “a finding of ambiguity is the
easy way out since it permits the court to create its own version of the
78
Kunreuther, Hogarth & Meszaros, supra note 8, at 79. See also sources
cited supra note 8.
79
Inflicting a high contract risk for unavoidably complex language is another
story. See Michelle Boardman, Insuring Understanding: The Tested Language
Defense, 95 IOWA L. REV. (forthcoming 2010).
80
But see Boardman, supra note 2, at 1112-17.
81
As with any increase in litigation certainty, decreasing the contract risk
should also increase the resolution of claims out of court.
contract and to find, or fail to find, ambiguity in order to justify an almost

predetermined result.”82 Ambiguity cannot be willed away, however, so
while the Arizona Supreme Court has cautioned against hunting for
ambiguity, the courts need an approach to truly ambiguous clauses: the
Arizona courts “determine the meaning of the clause . . . by examining the
purpose of the exclusion [or clause] in question, the public policy
considerations involved and the transaction as a whole.”83
In Transamerica Insurance Group v. Meere, the policyholder
injured a person when acting in self defense.84 The policy, like most, had
an unambiguous exclusion for intentional acts; the act of self defense was
intentional. Should coverage be found to be excluded, giving policyholders
some incentive to refrain from self defense? After all, the moral hazard
concern behind intentional act exclusions (and their common law parallel)
applies weakly, if at all, to self defense.
Consider the approaches a court could take:
1. Apply the clause as written. The meaning is clear.

→ no coverage for policyholder acting in self defense
2. Find the unambiguous clause ambiguous.
82
Ohio Cas. Ins. Co. v. Henderson, 939 P.2d 1337, 1339 (Ariz. 1997)
(quoting Transamerica Ins. Group v. Meere, 694 P.2d 181, 185 (Ariz. 1984)).
83
Id. at 1339 (emphasis added). See also Cal. Cas. Ins. v. Am. Family Mut.
Ins., 94 P.3d 616, 618 (Ariz. App. 2004).
84
Transamerica Ins. Group v. Meere, 694 P.2d 181 (Ariz. 1984).
On March 27, 1980, at about 12:30 a.m., Meere and a friend,
Leon Ivey, were leaving Lindy’s, a bar in Florence, Arizona.
Outside Lindy’s, Meere and Ivey were confronted by several off-
duty employees of the Arizona State Prison. Meere alleges that
he was quite apprehensive because he had been informed by a
captain at Arizona State Prison that a rumor was circulating
among the guards that Ivey and Meere, both ex-police officers,
were undercover investigators of narcotics flow into the prison.
One of the guards, Dennis Pruitt, approached Meere. Meere and
Pruitt exchanged words. Pruitt then struck Meere, knocking him
to the ground; Meere put up his hands, said ‘I don’t want to
fight,’ and was struck again by Pruitt. The two then exchanged
blows. The fight ended when Meere knocked Pruitt to the
ground and kicked Pruitt as he attempted to get up and come at
Meere again. Pruitt lost partial use of an eye as a result of this
fight. Id. at 183.
→ coverage without honor

→ litigation increases because policyholders and their counsel
conclude there is always an argument for ambiguity
3. Review the actuarial data and actuarial purpose of the clause.

→ the actuarial purpose of the clause, which is to avoid
creating a moral hazard of lowering a policyholder’s
disincentive to cause harm, does not apply to true self
defense
→ the actuarial data would reflect this fact
→ policyholder covered, in keeping with the insurer’s intent
4. Meere court: determine purpose of clause from case law, a

treatise, and public policy
→ coverage, in keeping with public policy
→ route most courts haven’t taken
Choices three and four may be equally good outcomes.

Unfortunately, a number of courts to look at this issue have taken choices
one and two, unnecessarily.85 Perhaps these courts felt uncomfortable
making public policy openly dispositive86 and were unwilling to deny the
policyholder at hand coverage while futilely recommending a future change
to the state insurance commission.
While the Meere court’s choice seems sufficient, the approach was
not attractive enough to other courts. Use of actuarial data would have
provided another avenue, a line of reasoning that could be substantiated,
and that was fact-based and provable. Moreover, an insurer aware that a
court might consider the actuarial data and purpose of the clause would
have been less likely to deny and litigate the claim in the first place.
This case strikes some as silly—“it is not a serious question
whether the policyholder should be covered for liability stemming from
self defense.” But it was (and is) a question in Arizona and elsewhere. “In
a majority of cases, courts have held that injury inflicted in self-defense is
85
See, e.g., Lockhart v. Allstate Ins. Co., 579 P.2d 1120, 1122-23 (Ariz. Ct.
App. 1978); Auto-Owners Ins. Co. v. Harrington, 538 N.W.2d 106, 108-10 (Mich.
Ct. App. 1995); Grange Ins. Co. v. Brosseau, 776 P.2d 123, 126-27 (Wash. 1989).
86
The dissent in Meere characterized the decision by the majority as one
“based on policy to distribute the consequences of the loss on an insurance
company.” Meere, 694 P.2d at 190.
expected or intended under the intentional injury exclusion clause.”87

Insurers fight these claims. One benefit of even the potential use of
actuarial data is that swindling or sloppy insurers should more readily pay
claims for risks the insurer has calculated and charged. Note that for this
benefit, the policyholders do not need to be able to determine where the
insurer is being inconsistent; the insurer’s self-knowledge is enough.
Actuarial data should also be able to prove or disprove the
existence of a “natural” exclusion. In a typical Hurricane Katrina case,
Tuepker v. State Farm, the court concluded that a combination of clauses
was ambiguous and therefore to be construed against the insurer. 88 As in
most homeowners policies, wind and rain damage were covered and water
damage (flood, inundation by water) was excluded. This policy also
included a “hurricane deductible endorsement,” meaning that the
policyholder had paid an additional premium for a type of hurricane
coverage (with a deductible).89
Finally, the policy excluded losses that were caused in part by a
covered cause but would not have occurred without an excluded cause.90
87
ROBERT H. JERRY, II & DOUGLAS R. RICHMOND, UNDERSTANDING
INSURANCE LAW § 63C(c) (4th ed. 2007). See, e.g., Stout v. Grain Dealers Mut.
Ins. Co., 201 F. Supp. 647 (M.D.N.C. 1962) (coverage for intentional shooting and
unintentional killing of persistent prowler excluded). What is at stake in these
cases is the insurer’s duty to defend, or pay for the defense, of the policyholder. A
policyholder who is found in a civil action to have acted properly in self-defense is
not liable to his foe. A policyholder whose claim of self-defense is rejected is
liable to his foe but his intentional act and harm clearly falls within the intentional
act exclusion. See also John Dwight Ingram, The Expected or Intended Exclusion
in Liability Insurance: What About Self-Defense?, 42 CREIGHTON L. REV. 123
(2009).
88
Tuepker v. State Farm Fire & Cas. Co., No. 1:05CV559 LTS-JMR, 2006
WL 1442489 at *4-*5 (S.D. Miss., May 24, 2006).
89
Id. at *4.
90
STATE FARM HOME OWNERS POLICY 24100401-1 (2005), reprinted in
Tuepker 2006 WL 1442489 at *2.
We do not insure under any coverage for any loss which would
not have occurred in the absence of one or more of the following
events. We do not insure for such loss regardless of: (a) the cause
of the excluded event; or (b) other causes of the excluded event;
or (c) whether other causes acted concurrently or in any sequence
with the excluded event to produce the loss; or (d) whether the
event occurs suddenly or gradually, involves isolated or
widespread damage, arises from natural or external forces, or
For these purposes, the policy would therefore exclude wind and rain
damage (otherwise covered) if it would not have occurred but for the storm
surge (excluded water damage). If a house withstood the wind up until the
storm surge knocked down a wall, for example, leading wind and rain to
damage the inside of the house, all would be excluded. On the other hand,
if wind tore a hole in the roof, letting rain in to damage the second story,
both would be covered whether or not the first floor was also flooded.
This is a simple explanation of what is admittedly dense policy
language. Given that the court had to find some meaning in the language,
however, it is a reasonable reading that does not torture or ignore any of the
policy language. However, the judge in Tuepker—a capable judge who
handled many Katrina cases with some skill—found the addition of the
hurricane endorsement to the rest of the policy (wind covered, flood
excluded) ambiguous.91 The judge seemed to conclude that since the
hurricane coverage must have meant something, and the policy already
covered wind and rain damage, it must have meant that the “combined
cause” exclusion did not apply during a hurricane.92
This reading may be less reasonable than the one offered above or
it may be an improvement; the point is that the actuarial data should have
been able to answer the question. The policyholder paid an additional
premium for the hurricane endorsement. Which risks did the insurer enter
in calculating that premium? Which risks are excluded? In particular, the
occurs as a result of any combination [enumerated excluded

causes]. Id.
91
On the other hand, Former Senator Trent Lott’s home was destroyed during
Hurricane Katrina, leaving nothing but a slab. State Farm held that a “storm
surge,” and not hurricane winds, caused the damage. This meant that Lott’s
insurance policy did not cover his damage. See Bob Kemper, Senator who Lost his
Home Sues Insurer, ATLANTA JOURNAL-CONSTITUTION, Dec. 16, 2005, at C3.
92
The court took this reading to avoid a finding that the hurricane coverage
was illusory. Insurance policies or clauses that are “illusory”—that would provide
coverage under no circumstances—are fraudulent and unenforceable.
Policyholders bringing suit against an illusory clause should be able to demonstrate
that the insurer knew the clause to be illusory from the insurer’s calculations of
expected loss. Conversely, an insurer defending a legitimate clause from attack
might be able to provide proof that it expected loss under certain circumstances
and planned for it. See, e.g., Frye v. S. Farm Bureau Cas. Ins., 915 So. 2d 486, 491
(Miss. Ct. App. 2005) (Policyholders seeking to prove “phantom” coverage
attempted to discover “information relating to the historical make-up and design of
the policy, as well as information relating to actuarial composition, loss reserves,
claims experience of [the insurer’s] agents, and pure profit.”).
correlated losses from inundation by water, caused by hurricane or any

other source, may be too high for insurers to offer solvent coverage. If so,
the court’s reading of the hurricane endorsement is flawed.
Here, actuarial data could have made a clause sensible to a court; it
could have shown not just a lack of fraud or decreased chance of deceit by
the insurer but also an understanding of what the insurer was getting at with
its language. A circumlocutory or technical clause will not be transformed
by data into a thing of beauty, but the court will no longer suspect the
insurer of careless randomness. Moreover, the court may see the public
policy value behind the enforcement of a “natural” exclusion.
Of course, judges sometimes mandate coverage for public policy
reasons and some judges involved in the many Katrina cases clearly saw
themselves in that role; thousands of people lost their homes and stingy
insurers could be made to pay. “Judges in insurance cases not only make
insurance law; sometimes they also make insurance.” 93 A full explanation
of why or how judges award uncontracted-for coverage is beyond this
discussion. Courts do it and will continue to do it; the point here is that
courts engaging in the regulation of the insurance industry should not do so
blindly. They should be armed with a full understanding of the actuarial
purpose behind the clauses to be regulated and the likely outcome of any
regulation. 94
93
ABRAHAM, supra note 13 at 101. Abraham is right that while “judicial
techniques of interpretation frequently create insurance coverage where policies do
not provide for it,” the total interpretation “practice turns out to be considerably
more complicated” than courts simply handing money from wealthy insurers to
policyholders. Id. at 101-02.
94
For example, state regulation can have disastrous results. State Farm is
currently withdrawing all homeowners insurance from the state of Florida. The
last straw for State Farm was not court action but the denial by the state Insurance
Commissioner of State Farm’s request to raise rates 47.1 percent. State Farm Can
Go? But Not on Its own Terms, SUWANNEE DEMOCRAT, Feb. 19,2009 available at
http://www.suwanneedemocrat.com/archivesearch/local_story_044133837.html.
State Farm claims that Florida’s policies, such as required discounts to its
customers “have further reduced needed revenues. During the first three quarters
of 2008 (a year with relatively modest catastrophe impact and no major hurricane),
State Farm Florida saw its surplus reduced by $201 million.” Hays, supra note 27,
at 7.
1. Ambiguity
A side note about ambiguity is called for here. More than half of
insurance disputes involve some claim of ambiguity. 95 If a court concludes
that a clause (or collection of clauses) is ambiguous, it will be construed
against the insurer and in favor of coverage under the doctrine of contra
proferentem. Actuarial data can help here too, although its role is more
amorphous.
In an opinion that is a favorite of insurance textbooks, a court
found the phrase “occupied as Janitor’s residence” ambiguous.96 The
question was whether a floor, a portion of which was used as a massage
parlor, was “occupied as a janitor’s residence” because a janitor slept there
on occasion. The court attempted to consider the purpose of the clause
from the insurer’s perspective but did so by guessing. The court assumed
the value of the janitor’s residence was to monitor for fire or trouble, and
this may have been so. On the other hand, the court also speculates that the
insurer may have wanted the janitor to exclusively occupy the floor,
keeping out more dangerous uses.97
Actuarial data should have been provided or solicited to prove one
of these purposes. The kink here is that the data would not directly have
addressed the question of whether the clause was ambiguous; either the
clause is susceptible of two plausible interpretations, in context, or it isn’t.
Appropriately or no, if one reasonable meaning of a clause is proven and
sensible, courts are less likely to find ambiguity. First, while courts often
admonish themselves not to “seek” or “create” ambiguity for policy
reasons, more factors go into the decision than the sheer ambiguity of the
word at hand. Second, once a plausible meaning is available, backed with
proof of consistent intent, the mind is less willing to entertain a weaker
alternative as proof of legitimate ambiguity.
95
See Rappaport, supra note 30, at 173. (“The ambiguity rule is probably the
most important rule in insurance law.”).
96
Vlastos v. Sumitomo Marine & Fire Ins. Co., 707 F.2d 775, 776 (3d Cir.
1983) (warranty as to state of building on day policy issued).
97
Id. at 779. The court hypothesizes why having a janitor occupying the floor
would be in the best interest of the insurance company, and states “[a] full-time
resident janitor might also deter prowlers and vandals from entering the building,”
among other reasons. Id.
III. ACTUARIAL DATA ALREADY IN COURT
This section addresses the concern that courts are ill-equipped to

handle actuarial data. Bottom line: courts have no choice. Actuarial data
are ubiquitous in modern decision making, too ubiquitous for courts to
duck entirely.98 The Supreme Court has dealt with the data directly on
several occasions, such as evaluating the use of gendered statistics in
employee benefits.99 Despite controversy, its use is increasing in civil
commitment and criminal sentencing.100 These cases often require expert
analysis of the data but, as would be the case in interpretation, they do not
require courts to undertake calculations or learn advanced statistics.
Courts accept, on occasion demand, and use actuarial or
underwriting data in limited insurance cases now. While these
circumstances do not usually involve the interpretation of insurance
policies head on, they do indicate that courts have some facility with the
data. For example, courts expect insurers to use actuarial or underwriting
data to prove materiality in a claim against a policyholder for
misrepresentation. A recent California opinion is a typical case:
The insurer’s “senior underwriter testified to the
misrepresentation’s materiality. She explained joint ventures pose
98
See Jonathan Simon, The Ideological Effects of Actuarial Practices, 22
LAW & SOC’Y. REV. 771 (1988).
Actuarial techniques play a central role in a proliferating set
of social practices. They are at the same time a regime of truth, a
way of exercising power, and a method of ordering social life.
Actuarial practices have not seemed very important nor attracted
much interest from social observers in part because they are
already so familiar, and in part because they fit so unobtrusively
into various substantive projects (e.g., educating, hiring,
premium setting) in which they are subordinated as a means to
an end. Id. at 772.
99
See Los Angeles Dep’t of Water & Power v. Manhart, 435 U.S. 702 (1978)
(holding that requiring larger pension contributions from women than men violated
Title VII of the 1964 Civil Rights Act, actuarial evidence of longer female life
expectancy notwithstanding).
100
See, e.g., Thomas R. Litwack, Actuarial Versus Clinical Assessments of
Dangerousness, 7 PSYCHOL. PUB. POL’Y & L. 409 (2001); Christopher Slobogin,
Dangerousness and Expertise, 133 U. PA. L. REV. 97 (1984); John W. Bagby,
Book Review, 4 J.L. ECON. & POL’Y 213 (2006) (reviewing BERNARD E.
HARCOURT, AGAINST PREDICTION: PROFILING, POLICING, AND PUNISHING IN AN
ACTUARIAL AGE (2006)).
increased risks, require additional underwriting, and warrant charging ‘an

additional premium’ before [the insurer] will cover them. Thus, the
misrepresentation is also material because it affected [the insurer’s]
evaluation of risk and the amount of the premium charged.’” 101
In a second example, if a policyholder is turned down for
insurance, the policyholder may be able to seek redress for a violation of
state law or regulation.102 The same is true when an existing policy is
cancelled or not renewed.103 Courts in these cases evaluate the
underwriting process but not for the purpose of interpreting policy
language.104 Similarly, insurers can challenge a state insurance
commissioner’s rejection of their language or rates, a dispute that will
involve the insurer’s premium data.105
Third, parties can point to actuarial theory or abstract fact without
presenting numbers or calculations.106 For example, insurers can point to
101
LA Sound USA, Inc. v. St. Paul Fire & Marine Ins. Co., 67 Cal. Rptr. 3d
917, 924 (Cal. Ct. App. 2007). See, e.g., Hill v. Allstate Ins. Co., No. 04-CV-
0865-REB-CBS, 2006 WL 173693, at *2 (D. Colo. 2006) (Insurer “charged a
premium based on the actuarial risk associated with a one-household [auto]
policy,” and “[a]bsent the fraudulent misrepresentations and concealment” of the
policyholders “it would not have issued and renewed the policy as it did.”). In
L.A. Sound, the court held that the insurer did not need to produce the specific
underwriter who had processed the policyholder’s application. L.A. Sound, 67 Cal.
Rptr. 3d at 924.
102
See CAL. INS. CODE § 791.10 (2006) (Adverse underwriting decisions;
declination, cancellation or nonrenewal of enumerated policies; specific reasons
for decision). Discriminatory underwriting was treated under the McBride-
Grunsky Insurance Regulatory Act of 1947, but repealed by Prop 103. CAL. INS.
CODE § 1861.05 (2009).
103
CAL. INS. CODE § 791.10 (2006).
104
See, e.g., Gonzales v. Pac. Specialty Ins. Co., No. B193051, 2007 WL
2005059 (Cal. Ct. App. July 12, 2007).
105
See, e.g., Mass. Auto. Rating & Accident Prevention Bureau v. Comm’ r,
453 N.E.2d 381 (Mass. 1983) (Action by insurers challenging the Massachusetts
Insurance Commissioner’s establishment of automobile insurance rates).
106
As with most forms of evidence or logical support, courts at times include
a reference to the actuarial function of a clause or policy where the point is not
outcome determinative but merely serves to buttress the court’s conclusion as
doubly right. The non-essential support for a legal outcome is not necessarily
dictum but nor is it always as well-considered as the evidence upon which the
court actively relies. Current judicial use of abstract actuarial principles is more
the fact that a particular type of coverage can be had for an additional
premium as proof that, in the absence of that premium, the loss is not
covered. Courts may but often do not require the insurer to demonstrate
that the policyholder was aware of the additional coverage for sale. Thus
courts are at times willing to use “inside” insurer information to interpret a
clause despite the information asymmetry between the parties. (Focusing
on a policyholder’s reasonable expectations, ignoring any insurer role in
those expectations, is a more common example of courts using
asymmetrical information, but to the policyholder’s benefit.)
Some courts seem to be open to a more sophisticated use of
actuarial data. If deeds speak louder than words, however, it should be
noted that encouraging language such as this often appears in an opinion
that does not rely upon the data:
If the primary goal is to fulfill the reasonable expectations of the
insured, then there is no need to look at anything beyond the language of
the policy itself. If, on the other hand, the primary goal is to give insureds
what they pay for, then we should, at the very least, be concerned with the
actuarial methods used to arrive at the premium and should look behind the
policy language itself.107
It is an open question whether courts would ever use the same data
to see that an insurer gives no more than he sells. 108
Fourth, in the specific area of uninsured motorist coverage, some
courts have been considering how premiums are calculated in policy
interpretation. In these cases courts look at the structure of the premiums
charged rather than at specific actuarial calculations. The question arises in
likely to fall into this category because a court need not expend much effort to
evaluate these forms of support.
107
Montano v. Allstate Indem. Co., 92 P.3d 1255, 1261-62 (N.M. 2004). The
court went on to “conclude that we need not resolve which rationale to give
primary effect.” Id. at 1262. Other courts have agreed that the “actuarial methods
used to arrive at the premiums [can be] considered to determine whether the
insured gets what he pays for.” Rehders v. Allstate Ins. Co., 135 P.3d 237, 248
(N.M. Ct. App. 2006).
108
In 1934, the Supreme Court expressed a similar sentiment: “While it is
highly important that ambiguous clauses should not be permitted to serve as traps
for policyholders, it is equally important, to the insured as to the insurer, that the
provisions of insurance policies which are clearly and definitely set forth in
appropriate language, and upon which the calculations of the company are based,
should be maintained unimpaired by loose and ill-considered interpretations.”
Williams v. Union Cent. Life Ins. Co., 291 U.S. 170, 180 (1934) (emphasis added).
underinsured and uninsured motorist (UM) coverage, where one’s own

auto policy provides coverage for losses from an accident caused by
another driver who has insufficient or no insurance. The premium for UM
coverage is usually broken out, so that the driver can see how much is
being charged for the UM portion of the policy.
Imagine a judge faced with an auto policy that has a UM coverage
limit of $50,000 and charges a UM premium per car. For example, $20 per
car, so that a one-car policy has a $20 UM premium and a three-car policy
has a $60 UM premium.109 If the three-car policy owner gets into an
accident with one of the cars, does he have $50,000 in coverage or
$150,000? Where policy language was ambiguous, courts tended to grant
the greater coverage.110
Insurers apparently had not meant to allow this “stacking” of
limits; insurers rewrote the policies to charge a unified UM premium. It
might appear that this was an insurer sleight of hand. The driver with three
cars is financially indifferent to paying $60 once or $20 three times. But
most courts reviewing the change in premium calculation have concluded
that with the new single premium, policyholders have “no reasonable
expectation of aggregate coverage.”111
One court referred to the new single premium as “actuarial and not
based on the number of vehicles covered.”112 This is twice right and twice
wrong: both calculations are actuarial ones and both take account of the
number of cars. The new single premium is probably more actuarially
accurate (or specific) than taking the premium for one car and multiplying
it, but it should still include the increased risk that comes from owning
additional cars. Assume a uni-car family with two drivers carries the risk
of needing UM coverage, R. Adding a second car to the family might
109
See Adkins v. Ky. Nat’l Ins. Co., 220 S.W.3d 296, 297 (Ky. Ct. App.
2007).
110
The endorsement page of the policy in Adkins read: “The limit of liability
shown in the Declarations for each person for Uninsured Motorists Coverage is our
maximum limit of liability for all damages ... sustained by any one person in any
one accident.... This is the most we will pay regardless of the number of ...
[v]ehicles shown in the Declarations....” Adkins, 220 S.W.3d at 299. See also
Sturdy v. Allied Mut. Ins. Co., 457 P.2d 34, 42 (Kan. Sup. Ct. 1969) (“When we
pay a double premium we expect double coverage. This is certainly not
unreasonable but, to the contrary, is in accord with general principles of indemnity
that amounts of premiums are based on amounts of liability.”).
111
Adkins, 220 S.W.3d at 300.
112
Id. at 299.
double that risk to 2R because now both drivers can be on the road at the
same time, although one would have to see the actual data to know.
Adding a third car seems unlikely to raise the risk to 3R, however, even if it
is greater than 2R. (The third car might be correlated with being a higher
risk rather than directly causing the risk to increase, for example if those
who own more than one car per person tend to drive more frequently or
hastily.)
Moreover, the reverse implication of the court’s description is that
the initial per-car premium was not “actuarial.” No doubt the initial
premium charged for one car did reflect the amount necessary for UM
coverage. The shorthand of multiplying that number by additional cars is
not as actuarially precise but nor was it pulled from thin air. Insurers
presumably found that any more elaborate calculation of the risk per added
car was not worth the candle.
This “stacking” question is the one most consistently being
answered with reference to actuarial data. In two exceptionally rare
circumstances, insurance policies refer directly to actuarial data. First,
courts have no choice but to accept actuarial data if the policy is one with a
retrospective premium. “A retrospective premium has two components: a
basic premium and a conversion loss factor to adjust the premium by
consideration of the insured's actual losses during the policy period.”113
Second, in unusual circumstances, actuarial data is incorporated
into the contract.114 Even then, courts will not allow themselves to be
forced to consider data in a policy; courts have shown their willingness to
ignore portions or entire policies that they conclude policyholders could not
113
Hartford Fire Ins. Co. v. Terra Ins. Co., No. CIV.A.01-5961, 2004 WL
1770298 at *2 (E.D. Pa. Aug. 2. 2004). “An insurance policy with retrospectively-
rated premium is sometimes referred to as a form of ‘self-insurance’ because the
policy covers only claims exceeding the maximum premium under the policy.” Id.
See also Douglas R. Richmond, Issues and Problems in “Other Insurance,”
Multiple Insurance, and Self-Insurance, 22 PEPP. L. REV. 1373, 1450 (2002).
114
See, e.g., Conrad v. Ace Property & Cas. Ins. Co., No. CV-05-5117-FVS,
2006 WL 1582376, at *1 (E.D. Wash. 2006) (“The Policy contains the following
definition of ‘Policy’: The agreement between you and us consisting of the
accepted application, these provisions, Special Provisions, actuarial documents,
and the applicable regulations published in 7 C.F.R. chapter IV.”) (emphasis
added)).
understand. But these examples suggest that judges are able to handle
actuarial data.115
As one of the largest insurance markets in the country (and the
world), California may be a good place to begin the actuarial data
experiment.116 California’s courts often provide a prototype of legal
evolution for other jurisdictions. “Other states may or may not choose to
follow California’s example on a particular issue or principle, but they
certainly note and examine what California does.”117
On the actuarial data front, the California Supreme Court has
repeatedly stated that it will look to “the reasonable expectations of the
insurer and the insured . . . as manifested in the distribution of risks, the
proportionate premiums charged and the coverage for all risks except those
specifically excluded.”118 In a recent decision, a court of appeals held:
No rational insurer would wish to undertake such an insuring
obligation. It would be literally impossible, from an actuarial standpoint,
to set appropriate premiums to guard against the risk that an association
would enter into multimillion-dollar construction contracts, and then not
pay for the construction work. That type of risk would be virtually
impossible to underwrite.119
115
These examples show courts have the ability to handle the data on some
level but in these last three examples, the data are represented in part in the policy
itself. Therefore, these examples are not meant to strongly show that courts are
necessarily inclined to use data more broadly.
116
In California, a search to capture every case that includes the word
“insurance” and some form or either “underwriting” or “actuarial” resulted in close
to 2000 cases. (Westlaw search for “insurance & (underwr! actuar!)”.) More
specific searches were used as well but one extremely broad search proved useful
in locating categories of use.) The majority of these were not relevant to the
question of whether or how California courts currently use the data. Many were
not insurance cases and many simply had citations to common party names that
included the word “underwriters,” such as “Universal Underwriters Ins. Co.” or
“Certain Underwriters of Lloyd’s London.” Some involve the underwriting of
pension plans or other funds that do not involve application of an insurance policy.
117
H. Walter Croskey, The Doctrine of Reasonable Expectations in
California: A Judge’s View, 5 CONN. INS. L.J. 451, 452 n.1 (1998) (“California
insurance jurisprudence has considerable influence on that of other jurisdictions.”).
118
Garvey v. State Farm Fire & Cas. Co., 770 P.2d 704, 711 (Cal. 1989)
(emphasis added).
119
Oak Park Calabasas Condominium. v. State Farm Fire & Cas. Co., 40 Cal.
Rptr. 3d 263, 268 (Cal. Ct. App. 2006) (emphasis added). This approach was
Although the California courts do not now entertain the data in the
way encouraged here, its use is compatible with the general approach taken
because “[t]he goal is to give effect to the reasonable expectations of both
the insured and the insurer.”120
IV. CONCLUSION: RESISTANCE AND REAL WORLD

OBSERVATIONS
This article introduces the possibility of improving the

interpretation and construction of insurance contracts through actuarial
data. As with any innovation, one must ask if there is a barrier beyond
simple lack of creativity or inertia to explain its current absence. Each of
the main players offers their own potential resistance.
A. INSURERS
Insurers have voted with their briefs, so to speak, in that they have
access to data and rarely seek to introduce it in court. As I have argued
elsewhere, there is some reason to believe that insurers are resistant to
change.121 Insurers also may assume that courts would not welcome this
development or, and only some will find this plausible, it may be that it has
simply not occurred to insurers to routinely use this type of data in court.122
Based on casual interviews with insurer counsel, I have collected
explanations for why insurers resist the advice of their own counsel that the
use of actuarial data would support an important position. In interviewing
outside counsel I was most interested in cases where the firm lawyer
instigated the idea of using actuarial data and was rebuffed by inside
counsel.123
specifically followed in August Entm’t, Inc. v. Philadelphia Indem. Ins Co., 52 Cal.
Rptr. 3d 908, 914-16 (Cal. Ct. App. 2007).
120
August Entm’t, Inc., 52 Cal. Rptr. 3d at 913 (emphasis added). As with
almost all jurisdictions, the opening move in California is to determine the intent
from the language if possible. California courts are not reticent to move on to
additional inputs, however.
121
See Boardman, supra note 2, at 1116-17.
122
As a lawyer, I would occasionally suggest to in-house insurer counsel that
actuarial data be used to prove a particular point. The resistance I encountered
rested upon some of the reasons discussed below.
123
In a future project I may ask insurers directly, although I would expect to
receive vetted public relations answers.
Most of the reasons given are not general objections but reasons to
resist the use of actuarial data in certain cases. The approach appears to be
piecemeal rather than a considered policy of always resisting actuarial data.
An insurer who has concluded, on net, that the general use of actuarial data
is not in their interest may of course object to all policyholder attempts to
discover or use it, even in cases where the data could be beneficial.
As expected, sometimes there is no data on point. An outside
counsel suggested backing up an insurer’s claim about an “additional
insured” provision with reference to the data. The insurer response: “There
is no data. We sometimes add ‘additional insureds’ without changing the
premium or putting aside additional reserves;” a fact the insurer did not
want to advertise.124 This does not seem to be an uncommon practice with
additional insured provisions.125
Similarly, there are circumstances where the data would not be
useful because the premium charged does not directly reflect the cost of the
risk but instead reflects interest rates, price competition, (past expected but
not yet incurred losses from past calculations that turned out to be
insufficient), etc. Of course, there are other times when the available
actuarial data just does not speak to the question at hand. In these cases
insurers may not be resistant to the court reading it, if its production were
costless, but they would not seek to introduce it themselves. Given that
production is never costless, an insurer may still object to discovery.
Another common response: “We just don’t refer to actuarial data in
our pleadings and briefs. We never have and nor do our competitors.”126 A
taller objection than inertia: Courts will reject the data but we will be worse
off for having offered it; it creates the impression that in order to
understand our policies a policyholder would need to be an actuary.
Finally, certain applications of actuarial data in the underwriting
process are proprietary. Insurers have moved to quash subpoenas on the
grounds that “materials contain[ing] reserving information and actuarial
124
As with all of these illustrations, this is a paraphrase of the insurer
counsel’s response, not a direct quotation.
125
“In practice, [additional insured] endorsements that are issued
automatically or without charge are usually limited to vicarious liability by express
statement.” James E. Joseph, Indemnification and Insurance: The Risk Shifting
Tools (Part II), PA. B. ASS’N Q. 1, 16 (2009). For a collection of cases in which
courts note that no or little additional premium had been charged for additional
insured coverage, see Note, Recognizing the Unique Status of Additional Names
Insured, 53 FORDHAM L. REV. 117, 120 and n.12 (1984).
126
Again, this is a paraphrase.
formulas or analysis particular to [that insurer]” are proprietary, in part

because in some segments “it is industry standard for each company to
develop its own product forms and underwriting systems.”127 Similarly,
insurers may fear that opening their books will open them to charges of
misbehavior, such as price-fixing or redlining by racial data.
B. POLICYHOLDERS
In many cases policyholders benefit from moving away from the

“contract context,” and away from what the insurer expected, so that
actuarial data might be viewed as a step in the wrong direction.
Policyholders have discovered or sought to discover actuarial data in
limited cases.128 Perhaps the data would open a Pandora’s box that
policyholders and their counsel are afraid to open because they are unsure
what waits inside.
The small sample of cases shows that policyholders are more likely
to seek to discover the data in class actions. The starting assumption is that
policyholders seek out the data when they predict or know (from others’
litigation or regulatory history) that it will support their preferred
interpretation and seek to exclude it when it will not. Another simplifying
but plausible assumption is that policyholders in litigation are pursuing an
interpretation of a policy or a clause that provides or extends coverage and
the insurer is pursuing an interpretation that excludes or limits coverage.129
127
Richter v. Mut. of Omaha Ins. Co., No. 06-Misc.-011, No. CV 05-498
ABC (PJWX), 2006 WL 1277906 at *1 (E.D. Wis. May 5, 2006) (assertion of
privilege deemed waived).
128
See, e.g., Vos v. Farm Bureau Life Ins. Co., 667 N.W.2d 36, 41 (Iowa
2003) (discovery in class action against life insurer for breach of contract and
various fraud claims included “actuarial material regarding the policies in
question”); Frye v. S. Farm Bureau Cas. Ins. Co., 915 So. 2d 486, 490-91 (Miss.
Ct. App. 2005) (denying a policyholder’s motion for discovery of actuarial
information); Robinson v. S. Farm Bureau Cas. Co., 915 So. 2d 516, 520 (Miss.
Ct. App. 2005) (denying a claim by policyholder plaintiffs that “alleged in their
motion for continuance that [the insurer] Southern Farm failed to respond to
written discovery concerning the historical makeup and actuarial composition of
the disability income coverage.”); Beller v. William Penn Life Ins. Co. of N.Y.,
828 N.Y.S.2d 869, 875 (N.Y. Sup. Ct. 2007) (actuarial data about life insurance
plan sought and obtained in class action).
129
Note that in considering the benign or evil motives of each player, this last
does not require a dark view. There are policyholders who would prefer their
Apart from wanting to use specific data that support their view and
exclude data that do not, the increased use of actuarial data should help
policyholders in general across time. Start with the arguable but
contentious assumption that insurers tend to offer the coverage the market
demands. This does not mean that insurance coverage is a wonderland for
consumers. Some coverage that policyholders would like to purchase is
not sold because policyholders are not willing to pay the minimum price
insurers demand. Some coverage is not sold because insurers are not
willing to insure the type of event. To maintain risk distribution in the risk
pool, insurers do not want to sell coverage for correlated large losses, such
as flood coverage in a flood plain. For reasons of public policy, moral
hazard, and adverse selection, insurers limit other options that individual
policyholders desire.
If there are sound actuarial reasons to limit coverage, the limitation
is generally to the benefit of insurers and policyholders. Courts can make a
mistake when they interfere with these types of insurer choices. They do
make a mistake when they choose the needs of one plaintiff policyholder
over all other policyholders. Actuarial data can improve the decision-
making of courts inclined to regulate.
C. COURTS
Are courts willing to entertain actuarial data more regularly? The

recitations of canons of insurance interpretation supply conflicting
statements. These are sometimes one on top of another, as though the
proximity of delivery will blind the reader to the inconsistency. Take a
representative summary from the Wisconsin Supreme Court:
The goal of construction is to ascertain the true intentions of the
parties to an insurance contract. In the case of an insurance contract, the
words are to be construed in accordance with the principle that the test is
not what the insurer intended the words to mean but what a reasonable
insurance to cover a loss but who read the policy to exclude it and who thus do not
sue. A particularly rosy view would be required to assume that every policyholder
in court firmly believes in the accuracy of their interpretation, but many no doubt
do. Likewise, insurers pursuing no or low coverage positions in court is not proof
that all insurers deny all claims all the time; if the insurer pays what the
policyholder expects, or close to it, there is no need for court.
person in the position of an insured would have understood the words to

mean.130
Translation of the interpretive goal:
(1) to find the “true,” i.e. subjective, intent of both the

policyholder and the insurer
(2) to not find the intent of the insurer
(3) to find the objective understanding of a reasonable
policyholder, which may or may not equate with
(a) the intent of this policyholder, or
(b) the understanding of this policyholder
Given these goals, actuarial data have a role to play. If,

contradiction accepted, a court is interested in knowing the insurer’s intent,
the court would be interested in actuarial intent. If a court is interested
neither in the policyholder’s intent (as it may not have one) nor the
insurer’s intent, it will pursue goal three.
Under this goal, the court will pursue the policyholder’s original
understanding, not the original intent. The goal is to decipher what the
policyholder would have understood the words to mean. With this
aspiration, often there is no answer to the question or there are several
answers; the language is ambiguous. Courts are then open to explanations
of:
(1) why one of several interpretations is reasonable (actuarial

purpose),
(2) what one party thought or did (actuarial intent),131
(3) whether the insurer acted in bad faith (consistent intent), to
decide how far to construe against the drafter, and
(4) the future drafting and premium consequences of reading the
language a particular way (actuarial purpose).132
130
Wood v. Am. Fam. Mut. Ins. Co., 436 N.W.2d 594, 599 (Wis. 1989)
(internal citations omitted) (emphasis added) (partially overruled on grounds not
relevant here).
131
Wisconsin courts have taken the position that “the policy should not be
rewritten by construction to cover matters not contemplated by the insurer nor paid
for by the insured.” Vidmar v. Am. Fam. Mut. Ins. Co., 312 N.W.2d 129, 131
(Wis. 1981).
132
“[W]ere insurers not able to enforce reasonable conditions upon their
liability, in accordance with actuarial standards and projections, their industry
As with other types of evidence, there will be debate about whether

most judges are qualified to evaluate actuarial data. Aspects of actuarial
calculation obviously require higher math skills and training. But one does
not need to be an actuary to understand the pricing mechanism or to see
that a certain risk was included in a calculation and therefore that the risk
should be covered. The relevant “pricing principles were in operational use
14 years before the publication of Adam Smith’s The Wealth of Nations
and considerably before the nineteenth century classical economists had
turned their attention to a general theory of prices.”133
Of course, the ability to use data is not sufficient; courts have to be
willing to use it. The limited judicial use of the data now shows both
ability and willingness to handle the evidence when it is part of set
interpretative routine.134 Those places where it has arisen in recent years
suggest that courts use the data when a framework for its use is obvious or
available.
Whether particular players in the insurance interpretation game

will support the introduction of actuarial data into the process is an
empirical question. To the extent the data can accurately reveal consistent
(or inconsistent) pricing and paying behavior, it will supplant what is now
mere assumption. This will decrease swindling insurer behavior by
increasing the chance of getting caught. It will also free honest insurers
from excessive court interference.
Where the data reveal insurer intent, hewing closer to that intent
will increase contract certainty, decrease cost, and reduce litigation. With
an understanding of actuarial purpose, the many courts engaged in the
regulation of insurance contracts can avoid current mistakes that harm
insurers and policyholders. Rather than regulate blindly, based on the
current needs of the one policyholder before it, proof of actuarial purpose
will allow a court take account of all the absence policyholders and the
health of the industry.
would be hindered in its ability to serve the important function it does in our
society.” Hartzo v. American Nat. Property. & Cas. Ins. Co., 951 So. 2d 1120,
1124 (La. Ct. App. 2006).
133
James C. Hickman & Robert B. Miller, Insurance Premiums and Decision
Analysis, 37 J. OF RISK AND INS. 567, 568 (1970).
134
See supra Section III.
THE LAW AND ECONOMICS OF
FIRST-PARTY INSURANCE BAD FAITH LIABILITY
Sharon Tennyson *
William J. Warfel**
***
States differ in the legal avenues available to policyholders to pursue

actions against their insurers for bad faith in claims settlement. This
article discusses the various approaches to first-party insurance bad faith
law that have been taken by the states, and discusses the potential benefits
and costs of different approaches. Regimes that are likely to grant large
damages awards to aggrieved policyholders provide the greatest deterrent
to insurer bad faith; but such regimes may also create incentives for
fraudulent insurance claiming and disincentives for rigorous claims
investigations by insurers. This article evaluates the empirical relevance of
these potential incentive distortions through an analysis of automobile
insurance claim settlement data in states with different bad faith regimes.
The data show that claim characteristics and claim investigations differ
significantly in states which permit tort-based bad faith from those in other
states, in ways consistent with the hypothesized effects.
***
*
Sharon Tennyson is Associate Professor in the Department of Policy
Analysis and Management at Cornell University. She holds a Ph.D. in economics
from Northwestern University with a specialization in industrial organization and
regulation. Her professional interest centers on economic and policy analysis of
insurance markets, and she has published widely on these topics. Sharon is a noted
expert on insurance rate regulation and insurance fraud, and is a frequent speaker
on these issues.
**
William J. Warfel, Ph.D., CPCU, CLU, is Professor of Insurance and Risk
Management at Indiana State University. He received his doctorate from Indiana
University in 1990. His research focuses largely on the interface of law and
insurance; he has published extensively in the CPCU eJournal, The John Liner
Review, Risk Management Magazine, and various legal publications. Also, to
date, he has been retained as a testifying and consulting expert witness in about 45
cases; his specialty includes breach of contract, bad faith, and agent/broker liability
issues. He can be reached at [email protected].
I. INTRODUCTION
The idea that insurers should be penalized for unfair claim

settlement practices involving first-party insurance coverage is a relatively
recent development in the long history of insurance law. Historically,
insurers were not penalized for unfair claim settlement practices involving
first-party insurance coverage. Pursuant to the nineteenth-century English
common law rule articulated in Hadley v. Baxendale,1 the policyholder was
allowed to recover only those damages that were in the contemplation of
the parties to the contract at the time the policy was purchased.2 This rule
meant that damage awards could not exceed the amount specified in the
insurance policy, even if the breach of contract was intentional on the part
of the insurer.
Beginning in the early 20th century, this rule was modified through
enactment of a progression of state statutes including model legislation on
unfair trade practices developed by the National Association of Insurance
Commissioners (NAIC) in 1959 and amended in 1972.3 The model Unfair
Trade Practices Act prohibits specified acts by an insurer when “committed
flagrantly and in conscious disregard of” the statute, or “with such
frequency as to indicate a general business practice.”4 The 1971 model
legislation and many of the “statutes originally adopted by the states” were
silent as to whether it created a private cause of action.5 This silence meant
the insured’s only recourse was to file a complaint with the state insurance
department.6
1
See Hadley v. Baxendale, (1854) 156 Eng. Rep. 145 (Exch. Div.).
2
See id.
3
4 NAT’L ASS’N OF INS. COMM’RS, MODEL LAWS, REGULATIONS, AND
GUIDELINES: UNFAIR TRADE PRACTICES ACT (2008).
4
Id. See also Laureen Regan & Paul M. Rettinger, Private Rights of Action
Under State Unfair Claims Settlement Practices Acts: A Review, J. INS. REG.
(1998) (identifies all 14 of the prohibited acts). This model legislation, or some
variant of it, has been adopted by all U.S. states. Efforts to expand first-party bad
faith liability continue today. In Connecticut, for example, a proposal has been
made to delete the requirement that violations occur with such frequency as to
indicate a general business practice. If this proposal is adopted, a single violation
would be sufficient to constitute bad faith.
5
Regan & Rettinger, supra note 4, at 298.
6
See Allen v. State Farm Fire & Cas. Co., 59 F. Supp. 2d 1217, 1227 (S.D.
Ala. 1999) (“[P]laintiffs should follow the procedure for review required by the
2009] FIRST-PARTY INSURANCE BAD FAITH LIABILITY 205
In the landmark case Gruenberg v. Aetna Insurance Company7 the

California Supreme Court extended the tort of bad faith to include first-
party insurance coverage disputes.8 In the wake of this decision, courts and
state legislatures across the country began to recognize the right to file
private causes of action against insurers alleging unfair claim settlement
practices in first-party insurance coverage disputes.9 Three different
avenues have been taken by state courts and legislatures in recognizing this
right to file a private cause of action: tort action based solely on bad faith;
contract action with broad definition of damages; and statutes.10
Tort Action Based Solely on Bad Faith: Tort action based solely on
bad faith relies exclusively on breach of the implied covenant of utmost
good faith.11 “Policyholders are not required to allege an independent tort
such as fraud or intentional infliction of emotional distress in order to
recover under the tort laws.”12 The general rule of damages in tort is that
the injured party may recover for all harm or injuries sustained (including
legal expenses, and damages for economic loss and mental distress),
regardless of whether these damages could have been anticipated.13
Punitive damages may be awarded if the conduct giving rise to liability was
particularly egregious.14
Contract Action with Broad Definition of Damages: A contract
action with broad definition of damages involves a good faith and bad faith
inquiry confined to the realm of contract15 where damages are broadly
defined to include both general damages (i.e., those following naturally
from the breach)16 and consequential, or incidental, damages (i.e., those
reasonably within the contemplation of, or reasonably foreseeable by, the
insurance code and first seek relief from the insurance department and the
insurance commissioner.”).
7
510 P.2d 1032 (Cal. 1973).
8
See id.
9
See SHARON TENNYSON & WILLIAM J. WARFEL, NAT’L ASS’N OF MUT. INS.
COS., ISSUE ANALYSIS: FIRST-PARTY INSURANCE BAD FAITH LIABILITY: LAW,
THEORY, AND ECONOMIC CONSEQUENCES 3 (Sept. 2008) available at
www.namic.org/insbriefs/080926BadFaith.pdf.
10
Id. at 3.
11
Id.
12
Id.
13
RESTATEMENT (SECOND) OF TORTS §§ 901, 903 (1979).
14
Id. § 908.
15
See Beck v. Farmers Ins. Exch., 701 P.2d 795, 801-02 (Utah 1985).
16
BLACK’S LAW DICTIONARY 416 (8th ed. 2004).
parties at the time the contract was made).17 “Consequential damages may
reach beyond the strict contract terms and include prejudgment interest and
legal expenses, and damages for economic loss and mental distress.”18
However, “an independent tort such as fraud or intentional infliction of
emotional distress must be alleged in order to make a claim for punitive
damages.”19
Statute: The right to file a private cause of action alleging bad faith
is based on statute and judicial recognition of an implied, private cause of
action under an Unfair Trade Practices Act that includes an unfair claim
settlement practices provision.20 Damages may include prejudgment
interest and legal expenses, consequential, or incidental, damages for
economic loss and mental distress. With a few notable exceptions,21 in
states that have adopted a statutory approach to first-party insurance bad
faith, punitive damages are not permitted.22 Or, if punitive damages are
permitted, a cap is placed on such damages and/or the standard of conduct
for awarding such damages is very stringent and thus the exposure to
punitive damages is minimal.23
A majority of states that recognize first-party insurance bad faith
allow actions under tort law. The tort of bad faith is a unique application of
tort law because it applies despite the existence of a contract, and does not
require the policyholder to allege a traditional tort such as fraud or
intentional infliction of emotional distress in order to recover punitive
damages. Relative to other bad faith liability regimes, the tort of bad faith
increases both the potential damages and the uncertainty of judgments for
insurance companies. Thus, the legal basis for a first-party insurance bad
faith allegation determines the realistic potential for a punitive damages
17
Id. at 417.
18
Tennyson & Warfel, supra note 9, at 3.
19
Id.
20
See CONN. GEN. STAT. ANN. §§ 38a-816, 42-110q (West 1995).
21
Pennsylvania statutory law recognizes private actions for first-party
insurance bad faith, and authorizes punitive damages. 42 PA. CONS. STAT. ANN §
8371 (West 2006). The state of Washington recently adopted a bad faith statute that
allows punitive damages awards under a standard similar to that in many tort-based
regimes. WASH. REV. CODE ANN. § 48.30.010 (West 1998).
22
See generally Mark J. Browne, et. al., The Effect of Bad Faith Laws on
First-Party Insurance Claims Decisions, 33 J. LEGAL STUDIES 355, 355-390
(2004).
23
Id.
award, dramatically altering the “stakes” of first-party insurance bad faith

litigation.
In theory, allowing policyholders to easily recover damages in
excess of the insurance benefit owed will provide insurers with added
incentives to engage in fair and efficient claims settlement practices.
Policyholders are provided assurance that an insurer will not unreasonably
withhold payment of a rightful policy benefit, or otherwise engage in
conduct that is designed to withhold such payment. However, tort liability
may also affect the claim settlement process in ways that are not socially
beneficial. For example, high and uncertain penalties for insurer bad faith
may create potential gains to policyholders from initiating bad faith actions
based on questionable or fraudulent claims. High penalties may also
reduce the willingness of insurers to vigorously challenge questionable
claims.
This article provides a discussion and analysis of first-party
insurance bad faith liability. It traces the evolution of first-party insurance
bad faith law, and identifies and discusses the various approaches that have
been taken by the courts and state legislatures. The economic rationale for
allowing bad faith actions in first-party insurance cases is developed, and
the potential gains to insurance market participants are considered. This
article also considers potential adverse effects of excessive or uncertain
first-party bad faith liability for the insurance claim settlement process, in
terms of creating incentives for policyholders to file fraudulent claims and
creating disincentives for insurers to investigate potentially fraudulent
claims. Automobile insurance claims data are analyzed to investigate the
empirical importance of these effects.
II. LEGAL PERSPECTIVES
A. LEGAL DEVELOPMENT OF TORT ACTION BASED SOLELY ON

BAD FAITH
Among jurisdictions that permit a tort action based solely on bad

faith, a large minority have adopted a “negligence” standard for
determining whether an insurer has acted in bad faith;24 the most common
24
At least eleven states have embraced the negligence standard for first-party
bad faith claims. These states include Alaska (State Farm Fire & Cas. Co. v.
Nicholson, 777 P.2d 1152 (Alaska 1989)); California (Gruenberg v. Aetna Ins.
Co., 510 P.2d 1032 (Cal. 1973)); Connecticut (Grand Sheet Metal Prods. Co. v.
standard among tort jurisdictions is an “intentional tort” standard;25 and one

jurisdiction has adopted a “quasi-criminal” standard.26 We discuss each of
these standards as they relate to first-party bad faith claims.
1. Negligence Standard
The “negligence” standard was first adopted in a third-party

liability insurance case in California.27 Courts following this approach in
Protective Mut. Ins. Co., 375 A.2d 428 (Conn. Super. Ct. 1977)); Hawaii (Best
Place, Inc. v. Penn Am. Ins. Co., 920 P.2d 334 (Haw. 1996)); Nevada (Hart v.
Prudential Prop. & Cas. Ins. Co., 848 F. Supp. 900 (D. Nev. 1994)); North Dakota
(Seifert v. Farmers Union Mut. Ins. Co., 497 N.W.2d 694 (N.D. 1993)); Ohio
(Hoskins v. Aetna Life Ins. Co., 452 N.E.2d 1315 (Ohio 1983)); Oklahoma
(Christian v. Am. Home Assur. Co., 577 P.2d 899 (Okla. 1977)); South Carolina
(Nichols v. State Farm Mut. Auto. Ins. Co., 306 S.E.2d 616 (S.C. 1983)); Texas
(Arando v. Ins. Co. of N. Am., 748 S.W.2d 210 (Tex. 1988)); and Washington
(Griffin v. Allstate Ins. Co., 29 P.3d 777 (Wash. Ct. App. 2001), review denied, 45
P.3d 551 (Wash. 2002)).
25
At least fourteen states have embraced the intentional tort standard for
first-party bad faith claims. These states include Alabama (Chavers v. Nat'l Sec.
Fire & Cas. Co., 405 So. 2d 1 (Ala. 1981)); Colorado (Herod v. Colo. Farm Bureau
Mut. Ins. Co., 928 P.2d 834 (Colo. App. 1996)); Idaho (Robinson v. State Farm
Mut. Auto. Ins. Co., 45 P.3d 829 (Idaho 2002)); Indiana (Erie Ins. Co. v. Hickman,
622 N.E.2d 515 (Ind. 1993)); Iowa (Dolan v. Aid Ins. Co., 431 N.W.2d 790 (Iowa
1993)); Kentucky (Wittmer v. Jones, 864 S.W.2d 885 (Ky. 1993)); Mississippi
(Universal Life Ins. Co. v. Veasley, 610 So. 2d 290 (Miss. 1992)); Nebraska
(Braesch v. Union Ins. Co., 464 N.W.2d 769 (Neb. 1991)); New Mexico (Chavez
v. Chenoweth, 553 P.2d 703 (N.M. Ct. App. 1976)); Rhode Island (Zarrella v.
Minn. Mut. Life Ins. Co., 824 A.2d 1249 (R.I. 2003)); South Dakota (Stene v.
State Farm Mut. Auto. Ins. Co., 583 N.W.2d 399 (S.D. 1998)); Vermont (Bushey
v. Allstate Ins. Co., 670 A.2d 807 (Vt. 1995)); Wisconsin (Anderson v. Cont'l Ins.
Co., 271 N.W.2d 368 (Wis. 1978)); and Wyoming (Hulse v. First Am. Title Co.,
33 P.3d 122 (Wyo. 2001)).
26
Aetna Cas. & Sur. v. Broadway Arms Corp., 664 S.W.2d 463 (Ark. 1984).
See JEFFREY W. STEMPEL, STEMPEL ON INS. CONTRACTS 10-87 (3d ed. 2009)
(1994). According to Stempel, “[t]o appreciate the differences across the states
concerning insurer bad faith, one must pay attention to the nuances of precedent
and doctrine, which tend to be glossed over in any classification….” Id. For this
reason, legal scholars often disagree on occasion over whether a particular state has
adopted the “negligence” standard or the “intentional tort” standard in first-party
cases.
27
See Comunale v. Traders & Gen. Ins. Co., 328 P.2d 198 (Cal. 1958).
third-party cases have reasoned that insurers must be held to a stringent

standard because of their disproportionate ability to influence the
acceptance or rejection of a settlement offer made by a claimant.28 In
particular, the standard demands that an insurer consider the insured’s
interest in addition to its own in deciding whether to accept or reject the
settlement offer.29
Claim-handling practices that are arguably unreasonable can
extend beyond third-party claims to include first-party claims; therefore,
plaintiffs’ attorneys soon asserted that first-party insureds also should be
permitted to file a tort action based solely on bad faith.30 Insurers
countered that breach of contract should be the exclusive cause of action
for first-party insurance bad faith actions because the relationship between
an insurer and a policyholder in a first-party context differs from that in a
third-party context.31 In a first-party context, the relationship might lead to
a dispute that could be characterized as “adversarial” (i.e., first-party cases
involve disputes over the terms of coverage, whether a loss occurred, or the
value of the loss).32 The relationship between an insurer and a policyholder
in a third-party context could be characterized as “fiduciary” (i.e., the
policy agreement transfers from the insured to the insurer the authority to
accept or reject on behalf of the insured a settlement offer presented by a
claimant).33
In the landmark Gruenberg v. Aetna Insurance Company
decision,34 the California Supreme Court rejected an insurer’s argument
that third-party cases are different from first-party cases, extending the tort
of bad faith to include first-party insurance coverage disputes.35 In
Gruenberg, the policyholder’s business was destroyed in a fire.36 The
claim representative informed the fire department investigator that the
28
See id.
29
Id. at 200-201 (citing Ivy v. Pac. Auto. Ins. Co., 320 P.2d 140 (Cal.
1958)).
30
See Gruenberg, 510 P.2d at 1032; Margaret Cronin Fisk, Looking for a
New Cause of Action?, NAT'L L. J., May 19, 1997, at A1.
31
See Brief for Am. Ins. Ass'n et al. as Amici Curiae Supporting Appellant,
St. Paul Fire & Marine Ins. Co. v. Onvia, Inc., 196 P.3d 664 (Wash. 2008) (No.
80359-5), 2009 WL 907292.
32
Tennyson & Warfel, supra note 9, at 4.
33
See id.; see STEMPEL, supra note 26, at 10-38.
34
510 P.2d 1032 (Cal. 1973).
35
Id. at 1038.
36
Id. at 1034.
policy limit in place was excessive, suggesting that the policyholder

intentionally caused the loss.37 Shortly thereafter, the policyholder was
charged with arson.38 Based on the advice of criminal defense counsel, the
policyholder initially declined to submit to an examination under oath,
which was requested by the insurer shortly after the fire pursuant to the
“Your Duties After Loss” provision contained in the policy.39 The charges
were subsequently dismissed at a preliminary hearing for lack of probable
cause.40 Shortly after disposal of the criminal matter, the policyholder
informed the insurer that he was now prepared to submit to an examination
under oath.41 The insurer declined to depose the policyholder based on its
contention that the coverage was void because the policyholder had
previously breached a condition in the policy requiring the insured to
submit to an examination under oath at the insurer’s request. 42
Arguing that the insurer had unreasonably suggested that he
intentionally caused the loss, the policyholder sought both compensatory
and punitive damages.43 In adopting the negligence standard in this first-
party case, the California Supreme Court reasoned that the third-party
context cannot be distinguished from the first-party context.44 In third-
party claims, the insurer has a “duty to accept reasonable settlements,”
whereas in a first-party claim, the insurer has a “duty not to withhold
unreasonably payments due under a policy.” 45 The court observed that
“these are merely two different aspects of the same duty.”46 When an
insurer “[refuses], without proper cause, to compensate its insured for a
loss covered by the policy, such conduct may give rise to a cause of action
in tort for breach of an implied covenant of good faith and fair dealing.”47
37
Id.
38
Id.
39
Id. at 1035.
40
See Gruenberg, 510 P.2d at 1035.
41
Id.
42
Id.
43
Id.
44
Id. at 1036.
45
Id. at 1037.
46
Gruenberg, 510 P.2d at 1037.
47
Id. The California court did address the issue pertaining to recovery for
mental distress. Given that the policyholder alleged substantial economic losses
(e.g., loss of earnings, loss associated with bankruptcy) apart from damages for
mental distress, the policyholder was entitled to make a claim for mental distress.
Generally, when a policyholder substantially prevails in a first-party claim against
2. Intentional Tort Standard
An “intentional tort” standard was first adopted in Anderson v.

Continental Insurance Co., a first-party homeowner’s insurance case in
Wisconsin in 1978.48 Like the California Supreme Court in Gruenberg, the
Supreme Court of Wisconsin ruled that the theoretical underpinnings of the
bad faith tort in the third-party claim context apply equally in the first-party
claim context.49 Importantly, however, the Wisconsin Supreme Court
departed from the California legal precedent, ruling that “the tort of bad
faith is not a tortious breach of contract. It is a separate intentional wrong,
which results from a breach of duty imposed as a consequence of the
relationship established by contract.”50
This dichotomy is the foundation of the intentional tort standard;
the denial of a claim may constitute a breach of contract, but it does not
constitute bad faith. In other words, an insurer is entitled to contest a claim
so long as it has a reasonable basis grounded in law or fact. Whether the
insurer ultimately is correct in its position is of no consequence in resolving
the bad faith issue. Denying a claim whose validity is “fairly debatable”
does not necessarily constitute bad faith, even if the insurer ultimately is
incorrect in its position.51 Rather, the issue is first, whether the insurer
undertook a proper investigation, and second, whether the results of the
investigation were subjected to a reasonable evaluation and review.52 If
these conditions are met, the insurer will have established that its denial of
the claim was reasonably grounded in law or fact.
Because the intentional tort standard is more stringent than the
negligence standard, insurers are more likely to be successful in pretrial
pleadings. Judges are more likely to dismiss as a matter of law an
allegation of bad faith that involves nothing more than an insurance
coverage dispute. Specifically, the Wisconsin court ruled that “there must
be a showing of an evil intent deserving of punishment or of something in
the nature of special ill-will or wanton disregard of duty or gross or
an insurer, the policyholder is entitled to damages for aggravation and

inconvenience. See, e.g., Hayseeds, Inc. v. State Farm Fire & Cas. Ins. Co., 352
S.E.2d 73 (W. Va. 1986).
48
See Anderson v. Cont’l Ins. Co., 271 N.W.2d 368 (Wis. 1978).
49
Id. at 374.
50
Id.
51
Id. at 375-76.
52
Id. at 377.
outrageous conduct” in order to recover punitive damages.53 The court

added, “[An insurer] must not only intentionally have breached [its] duty of
good faith, but in addition must have been guilty of oppression, fraud, or
malice ….”54 This heightened standard means that only a small subset of
bad faith claims will warrant punitive damages. Direct proof must be
presented establishing either that the misconduct was extreme, or that the
misconduct was the result of a deliberate, company-wide practice of
underpaying claims.
3. Quasi-Criminal Standard
In 1984 the Arkansas Supreme Court overturned a jury verdict that,

in the view of the Court, did not require sufficiently stringent standards of
conduct and proof to support an award of punitive damages in a first-party
bad faith case.55 In Aetna Casualty and Surety v. Broadway Arms, the
policyholder alleged bad faith in the handling of a fire insurance claim.56
The evidence presented to the jury to support a finding of intentional
oppressive conduct was the claim representative’s statement to the
policyholder that he might be asked by the Internal Revenue Service (IRS)
to explain why the insurance carrier would pay $75,000 for loss of
inventory when the policyholder’s financial statement showed an inventory
valued at only $23,000.57 Apparently convinced that the claim
representative had made a thinly veiled threat to report the policyholder to
the IRS if the policyholder refused a reduced settlement offer, the jury
awarded the policyholder $5 million in punitive damages.58 The judgment
was reversed on appeal, and the case was remanded for a new trial based on
a “quasi-criminal” standard of conduct.59
53
Id. at 379.
54
Anderson, 271 N.W.2d at 379.
55
See Aetna Cas. & Sur. v. Broadway Arms Corp., 664 S.W.2d 463 (Ark.
1984). Arkansas is the only state that has embraced the quasi-criminal standard.
This standard was upheld in Columbia Nat’l Ins. Co. v. Freeman, 64 S.W.3d 720
(Ark. 2002). In this particular case, which involved a property insurer, the court
held that the conduct of the insurer must be carried out with a state of mind
characterized by hatred, ill will, or a spirit of revenge.
56
Broadway Arms, 664 S.W.2d at 464.
57
Id. at 469.
58
Id. at 465-66.
59
Id. at 470.
In adopting this standard, the court declared that “evidence of bad

faith must be sufficient to show affirmative misconduct of a nature which is
malicious, dishonest, or oppressive.”60 As articulated by the court, the
quasi-criminal standard has three elements. First, the court noted that a
single violation of the Arkansas Trade Practices Act does not necessarily
constitute bad faith.61 At minimum, there must be multiple violations in the
handling of the claim. Alternatively, a pattern of institutional misconduct
(e.g., a company-wide practice of deliberately underpaying claims) would
constitute bad faith.62 Assuming multiple violations in the handling of the
claim, or institutional misconduct, an inference can be made that the
evidence is “sufficient to show affirmative misconduct of a nature which is
malicious, dishonest, or oppressive.”63 Second, the court ruled that the
purpose of the tort of bad faith is not to address the situation where the
insurance carrier simply refuses or fails, through nonfeasance, to pay an
insurance claim.64 In cases of this sort, an adequate remedy already exists
under Arkansas law (See Arkansas Stat. Ann. Section 66-3001 et seq.
(Repl. 1980)).65 Third, the court reasoned that the public interest demands
that the tort of bad faith, which includes a substantial punitive damages
exposure, be carefully confined to extreme cases of misconduct.66
Otherwise, insurers will be inappropriately discouraged from questioning
false, suspicious, or inflated claims – a result that will increase insurers’
claim costs and raise policyholders’ premiums. The court suggested that
alternative remedies should be used to assure that policyholders are
appropriately compensated in those cases where an insurer simply refuses
or fails to pay a valid insurance claim.67
60
Id. at 467.
61
Id. at 466.
62
Broadway Arms, 664 S.W.2d at 466 (citing ARK. STAT. ANN. § 66-3005(9)
(current version at ARK. CODE ANN. § 23-66-206(13) (2009))).
63
Id. at 467.
64
Id. at 468.
65
ARK. STAT. ANN. §§ 66-3001–3014 (current version at ARK. CODE ANN.
§§ 23-66-201–215 (2009)).
66
Broadway Arms, 664 S.W.2d at 469.
67
Id. at 468.
B. LEGAL DEVELOPMENT OF CONTRACT ACTION WITH A BROAD

DEFINITION OF DAMAGES
The contract standard was embraced in the landmark Beck v.

Farmers Insurance Exchange decision,68 a case in Utah in which the
insurer rejected the insured’s claim for uninsured motorist benefits without
explanation and without conducting an investigation to determine the
validity of the insured’s claim.69 In this particular case, the court reasoned
that “[a]lthough the policy limits define the amount for which the insurer
may be held responsible in performing the contract, they do not define the
amount for which it may be liable upon a breach.”70 Confinement of the
good faith/bad faith inquiry to the realm of contract assures compensation
in the situation where the insurer fails to (1) diligently investigate the facts
to enable it to determine whether a claim is valid, (2) fairly evaluate the
claim, or (3) act promptly and reasonably in rejecting or settling the
claim.71 However, the contract standard forecloses the possibility of a
punitive damages award in the absence of proof that an independent tort
such as fraud or intentional infliction of emotional distress occurred.72
Indeed, in Beck v. Farmers Insurance Exchange, the court reasoned that
“the practical end of providing a strong incentive for insurers to fulfill their
contractual obligations can be accomplished … through a contract cause of
action, without the analytical straining necessitated by the tort approach
and with far less potential for unforeseen consequences to the law of
contracts.”73
68
Beck v. Farmers Ins. Exch., 701 P.2d 795 (Utah 1985). At least six states
have embraced the contract standard for first-party bad faith claims. These states
include Maine (Marquis v. Farm Family Mut. Ins. Co., 628 A.2d 644 (Me. 1993));
Maryland (Johnson v. Fed. Kemper Ins. Co., 536 A.2d 1211 (Md. Ct. Spec. App.
1987)); New York (Bi-Econ. Mkt., Inc. v. Harleysville Ins. Co., 886 N.E.2d 127
(N.Y. 2008)); Oregon (Nw. Pump & Equip. Co. v. Am. States Ins. Co., 925 P.2d
1241 (Or. 1996)) and Virginia (A & E Supply Co. v. Nationwide Mut. Fire Ins.
Co., 798 F.2d 669 (4th Cir. 1986) (interpreting Virginia law), cert. denied, 479
U.S. 1091 (1987)).
69
Beck, 701 P.2d at 796-97.
70
Id. at 801.
71
Id.
72
Id. at 800-801.
73
Id. at 799.
C. LEGAL DEVELOPMENT OF PRIVATE CAUSE OF ACTION BASED

ON STATUTE
In a majority of states, a private cause of action is not statutorily or

judicially permitted under the state’s Unfair Trade Practices Act. In a small
number of states, however, either the state legislature has written the law to
permit a private cause of action or a court has recognized an implied
private cause of action under the law. For example, the Connecticut statute
identifies specific types of conduct that constitute bad faith, sets forth the
burden of proof, and specifies the damages that can be recovered.74
Furthermore, in many states where the courts have failed to recognize a
common law cause of action for first-party bad faith, the state legislatures
have responded by enacting a statute that permits a private cause of action
for the first-party bad faith. Typically, these statutes identify the standard
of conduct, the burden of proof, and the damages that can be recovered in a
first-party bad faith action.75
There is considerable variation among state bad-faith statutes with
respect to the standard of conduct, burden of proof, and damages that can
be recovered. Some statutes, for example, only allow for limited recovery
of damages (e.g., prejudgment interest and attorney fees).76 Other statutes
contain language that has been broadly construed by at least one court to
permit unlimited punitive damages in those cases where the insurer has
engaged in more than one listed prohibited practice with respect to the
processing of a single claim.77
The statutory basis for first party insurance bad faith is still
evolving. In recent years, a number of states have enacted new legislation
creating or modifying the first-party bad faith liability exposure for
insurers. For example, Minnesota passed legislation in 200878 that creates
a new private cause of action for first-party insurance bad faith where one
previously did not exist. The statute codifies the intentional tort standard,
providing for damages if the insured can show (1) the absence of a
reasonable basis for denying the benefits of the insurance policy, and (2)
that the insurer knew or acted in reckless disregard of the lack of a
74
See CONN. GEN. STAT. §§ 38a-816, 42-110q (2009).
75
See, e.g., FLA. STAT. ANN. § 624.155(1)(b)(1) (West 2004).
76
See, e.g., MICH. COMP. LAWS SERV. § 500.3148(1) (LexisNexis 2001).
77
See, e.g., Maher v. Cont’l Cas. Co., 76 F.3d 535 (4th Cir. 1999) (applying
West Virginia law).
78
MINN. STAT. ANN. § 604.18 (West 2008).
reasonable basis for denying the benefits of the insurance policy.79 The law
allows policyholders to be awarded up to $250,000 in “taxable costs” if an
insurer is found to be acting in bad faith and up to $100,000 in attorney's
fees, but specifically precludes punitive damages in the absence of an
independent tort such as fraud or intentional infliction of emotional
distress.80
Recent Colorado legislation lowered the legal standard for
asserting a first-party bad faith claim and increased the penalties levied
against an insurer, relative to existing common law.81 The new legislation
adopts the negligence standard, whereas the intentional tort standard
applies under common law.82 In addition, under common law,
consequential, or incidental, damages for economic loss and mental distress
can be recovered, but the cost of litigation cannot be recovered.83 The new
legislation allows for the recovery of the cost of litigation and caps the
damages award at two times the policy benefit that was unreasonably
denied.84
Recent first-party insurance bad-faith legislation in Maryland85
applies exclusively to property/casualty insurance policies and allows
policyholders to initiate bad-faith claims through the Maryland Insurance
Administration (MIA), the state agency responsible for enforcing
Maryland’s insurance laws.86 The new law adopts the negligence standard
and caps damages the insured can recover at the policy limit.87 In addition,
it provides for recovery of pre-judgment interest and allows recovery of
attorney’s fees, but limits the recoverable amount to one third of the actual
79
§ 604.18(2)-(3).
80
§ 604.18(3)
81
COLO. REV. STAT. § 10-3-1113 (2008).
82
Id.
83
See Am. Family. Mut. Ins. Co. v. Allen, 102 P.3d 333, 342 (Colo. 2004)
(applying Colorado common law).
84
COLO. REV. STAT. §§ 10-3-1116 & 10-16-106.5. Moreover, the new
legislation imposes a special penalty on health insurers that unreasonably delay the
payment of the policy benefit (i.e., the penalty is 20 percent of the policy benefit,
the payment of which was delayed 90 days or longer past the submission of the
claim).
85
MD. CODE ANN., [INS.] § 27-1001(e)(2)(i) – (ii) (West 2006).
86
The law does not apply to claims that fall under the small-claims
jurisdiction of district courts or to commercial insurance policies with policy limits
exceeding $1 million.
87
§§ 27-303(9) & 27-305(c)(2).
damages.88 Previously, an insured could recover only the amount of actual

damages, subject to the policy’s limits. Such actions could be pursued
either through the MIA or as a civil action for breach of contract, but
neither avenue allowed recovery of litigation expenses.89
Legislation adopted in Washington90 expands the definition of first-
party insurance bad faith and increases the damages awards available to
policyholders in cases alleging insurer bad faith. The remedies specified in
the act are separate and distinct from the remedies provided under common
law as well as those prescribed in the state’s Consumer Protection Act.91
The new legislation provides for a private cause of action in the event an
insurer “unreasonably” denies or delays payment of a policy benefit or
commits a specified unfair claims settlement practice, recovery of “actual
damages sustained,” recovery of the cost of reasonable attorney’s fees, and
treble actual damages sustained, at the discretion of the trial judge.92 This
legislation represents a significant departure from most other states’
statutory approaches to first-party insurance bad faith, because it permits
both unlimited punitive damages and does not contain a stringent standard
of conduct for the awarding of such damages.
III. ECONOMIC PERSPECTIVES
A. THE ECONOMIC RATIONALE FOR TORT LIABILITY
Applying economic analysis to first-party bad faith insurance law

leads to important insights regarding the purpose of first-party bad faith
actions. A key concept is that allowing the courts to impose extra
contractual liability on insurers in cases of intentional or unintentional bad
faith denial of claims serves the obvious purpose of compensating
88
§ 27-305(4).
89
§ 27-305(3)(1)-(2).
90
WASH. REV. CODE ANN. § 48.30.010.
91
§ 19.86.090 Washington common law provides for the tort of bad faith
with a negligence standard, and the Consumer Protection Act provides for recovery
of actual damages sustained, the cost of litigation, and treble damages, subject to a
cap of $10,000 in the event the insurer violates a claims handling regulation.
92
Id. The specific unfair claims settlement practices covered by the
legislation include misrepresentation of policy provisions, failure to acknowledge
pertinent communications, failure to meet standards for prompt investigation of
claims, and failure to meet standards for prompt, fair, and equitable settlements
applicable to all insurers.
policyholders for their unwarranted losses; but it may also serve the
broader economic purpose of enhancing the efficiency of insurance
contracting. In competitive insurance markets, systematic bad faith in
claim settlement practices will reduce demand for insurance from a
company that engages in such practices. However, reputation penalties
may not be sufficient to guarantee that an insurer will never have an
incentive to engage in intentional bad faith claim settlement practices. Nor
can reputation penalties guarantee that an insurer will never engage in
behaviors that lead to unintentional bad faith denial of a claim settlement.93
An isolated example of intentional bad faith would include, for
example, if an insurer strategically denies or delays the settlement of a
particularly large insurance claim for the purpose of coercing the
policyholder to accept a reduced claim settlement.94 Market sanctions
alone may not deter this kind of behavior, because the potential cost
savings on the claim could outweigh the cost of reputation penalties meted
out in the market in the form of reduced demand for insurance. In such
cases, the potential for tort litigation creates an incentive for an insurer to
avoid unwarranted strategic denial or delay of a claim settlement, by
imposing a potentially large financial penalty for such conduct.95 Indeed,
the mere threat of substantial extra contractual liability will reduce the
incentive for an insurer to strategically deny or delay a claim settlement.
Most importantly, because a practical mechanism does not exist for an
93
See Joseph M. Belth, Two Recent Court Decisions Critical of UNUM’s
Disability Insurance Claims Practices, THE INS. FORUM, Mar. 2009, at 161.
Perhaps the classic case that demonstrates this point concerns disability insurance
claim settlement practices implemented by several operating companies of the
UNUM Group. These companies implemented a “claims management
philosophy” in 1993; previously these companies had a “claims payment
philosophy.” Various claim settlement practices used to deny, terminate, and settle
disability insurance claims resulted in numerous lawsuits alleging bad faith. The
volume of lawsuits was such that they resulted in widespread media attention, two
far-reaching regulatory probes, and eventual settlements with regulatory authorities
under which UNUM agreed to reassess some denied claims and improve claim
procedures. While the settlements did not include an admission to the effect that a
statute or regulation had been violated, UNUM did agree to pay a $15 million fine
divided among 49 jurisdictions that had signed on to the settlements. Id.
94
Alan O. Sykes, “Bad Faith” Breach of Contract by First-Party Insurers,
25 J. OF LEGAL STUD. 405, 412-13 (1996).
95
Id. See also KENNETH S. ABRAHAM, DISTRIBUTING RISK: INSURANCE,
LEGAL THEORY, AND PUBLIC POLICY 183-85 (1986).
insurer to provide a credible and binding contractual commitment not to

deny or delay payment of a legitimate claim, the threat of substantial extra
contractual liability for intentional bad faith may reassure policyholders
that a valid claim will be paid in a timely manner. This assurance may
improve the insurance contracting environment, thereby benefiting both
policyholders and insurers.
Because insurance contracts do not contain a binding contractual
commitment not to deny payment of a valid insurance claim, and because
insurance claims may be complex and policy language cannot fully
anticipate all of the details or nuances of a loss, coverage disputes are
inevitable.96 Some coverage disputes involve differing interpretations of the
policy language by the insurer and the policyholder, leading to different
conclusions about whether a loss is covered. For example, under a
builder’s risk policy, the reporting form may instruct the policyholder to
deduct the value of land when reporting the estimated completed value of
structures. The insurer may interpret “land” to include just the value of the
land itself; the policyholder may interpret “land” to include not only the
value of the land but also the value of land improvements including paving,
gutters, and curbs, for example. Assuming a loss caused by the collapse of
paving, gutters or curbs,97 the insurer may deny the claim because the
policyholder did not report the completed value of these items and thus did
not pay a premium for coverage on them. The policyholder may insist that
coverage exists for these items based on the instruction contained in the
reporting form to the effect that land should be deducted from total
estimated completed value of structures, which presumably would include
land improvements as well as the value of land itself.
A particularly interesting special case is the potentially fraudulent
claim, in which the insurer believes that the policyholder may have
“manufactured” false information about the loss event or the amount of
loss.98 One such case is when an insurer may incorrectly but reasonably
believe that the policyholder intentionally caused a loss.99 For example, if
a summary judgment concerning a mortgage default was issued by a court
the day preceding a fire, an insurer may have reasonably believed that the
96
Sykes, supra note 94, at 429.
97
For example, the soil was improperly compacted when the sewer lines
were laid.
98
Sykes, supra note 94, at 425-26.
99
Id. at 426.
policyholder conspired with a friend to set the fire, notwithstanding the fact
that a jury ultimately ruled that the friend acted alone in setting the fire.
With respect to some coverage disputes, further claim investigation
can be helpful in enabling the insurer to reach the correct conclusion
concerning the existence or non-existence of coverage. With respect to
other coverage disputes, however, further claim investigation is not viable
because of cost considerations or factual issues that simply cannot be
resolved.100 In these cases, the discovery process attendant to litigation
may be the best mechanism for bringing a claim to resolution.101 For
example, if an insurer reaches the conclusion that further claim
investigation is not viable, the insurer may offer a nuisance settlement even
though it believes that coverage does not exist.
In the case of suspected fraud, if we assume that it is more costly
for a policyholder to pursue a fraudulent claim than a valid claim, a
policyholder with a fraudulent claim will be less likely to persist in the face
of a claim denial than a policyholder with a valid claim.102 This fact implies
that costly litigation may be used as a screening device by insurers for the
purpose of sorting valid claims from fraudulent ones. For this reason, even
though the denial of a claim may lead to litigation, an insurer may find that
the denial of some fraction of suspicious claims is efficient.103 The benefit
of this approach for the insurer is that a policyholder pursuing a fraudulent
claim may drop it rather than engage in costly litigation; policyholders with
legitimate claims will be more likely to pursue the litigation.104 While
some policyholders needlessly suffer because they are forced to litigate
legitimate claims, policyholders as a group benefit from this approach
because it minimizes unwarranted claim costs and results in reduced
insurance premiums.105 In specifying a claims denial fraction, the insurer
will balance the expected reduction in claims fraud with the expected costs
of litigating denied claims.
A similar claims settlement dynamic may occur if we consider the
insurer’s decision regarding the amount of payment, rather than the
decision regarding claim denial. Again, in cases where establishing the
complete truth through investigation is not practical, and assuming that the
100
Id.
101
Id. at 425-29.
102
Id. at 428.
103
Id.
104
105
Id. at 426-27.
“manufacture” of documentation to support an overstated claim is costly

for a policyholder, underpayment of suspicious claims may be an optimal
fraud-deterrent strategy for an insurer.106 The benefit of this approach for
the insurer is that underpaying suspicious claims reduces policyholders’
incentives to exaggerate the claimed amount.107 While some policyholders
are forced to litigate in order to collect the full amount of a valid claim, the
benefit of this approach for policyholders as a group is reduced fraud and
reduced insurance premiums. The cost to the insurer is the potential for
litigation due to underpayment. Thus, the insurer must balance the amount
of claim underpayment against the expected costs of dispute resolution and
litigation when specifying a claims payment strategy for suspicious claims.
These perspectives on insurers’ claims payment strategies for
responding to suspicious claims provide an economic rationale for
permitting first-party private actions for insurer bad faith failure to settle,
and for allowing extra contractual damages in these cases. First, the threat
of bad faith litigation serves to mitigate insurers’ incentives to strategically
deny or delay the payment of valid claims.108 Second, the threat of bad
faith litigation places an appropriate limit on insurers’ use of claim denial
or underpayment as a fraud-deterrent strategy.109 In the absence of
potential extra contractual liability, an insurer will consider only the
benefits of this strategy to its claims operation, ignoring the costs imposed
on policyholders who have legitimate claims that were denied or underpaid.
If an insurer faces the possibility of a damage award in excess of the
benefit specified in the policy, the insurer is given an incentive to take into
account the costs imposed on a policyholder when a legitimate claim is
denied, delayed or underpaid. The result is an efficient balance (of costs
and benefits) – i.e., between the costs associated with unwarranted claims
denial, delay or underpayment,110 and the benefits of reduced insurance
claim fraud, and thus the avoidance of excessive costs in the insurance
system.111
106
Keith J. Crocker & Sharon Tennyson, Insurance Fraud and Optimal
Claims Settlement Strategies, 45 J.L. & ECON. 469, 470 (2002).
107
Id. at 504.
108
Id. at 472-73.
109
Id. at 475.
110
Similar arguments can be made regarding insurers’ claim settlement
strategies for claims disputes centering on contractual language or factual disputes
not involving suspected fraud.
111
Crocker & Tennyson, supra note 106, at 504-505.
B. POTENTIAL UNINTENDED EFFECTS OF TORT LIABILITY
Although allowing tort actions for the purpose of addressing

insurer bad faith in claims settlement may be efficient in theory, practical
considerations concerning implementation of tort law have important
implications for whether the tort system is in fact efficiency-enhancing. If
the standard applied by the court for a finding of insurer bad faith is too
lax, and/or if the damages award is too high relative to the actual damages
sustained by a policyholder whose claim was denied or underpaid,
substantial incentive distortions may arise. If first-party bad faith laws
create substantial incentive distortions, the benefits of these laws will be
lessened because of increases in insurance costs due to fraud. Because it is
the possibility (more specifically, the expected value) of damages from bad
faith actions that affects insurer and policyholder decisions, interactions
between insurers and policyholders may be distorted irrespective of
whether an injured policyholder actually files a suit.112
A major concern is the increased pressure on insurers to pay
reasonably disputable claims.113 Insurers balance the benefits of reduced
fraud costs with the expected costs of litigation.114 If the expected costs of
litigation to insurers are sufficiently high that they exceed the expected
cost-savings from reduced fraud costs, insurers will have less incentive to
employ fraud reduction strategies.115 Specifically, claim investigations
may lead to insurer actions that bring accusations of bad faith, and thus an
excessive threat of bad faith liability may reduce the number and scope of
claim investigations below what they should be.116
This effect on insurer incentives will raise the costs of fraud in both
the immediate term because fewer fraudulent claims will be detected, and
over the longer term because of reduced fraud deterrence. This latter point
is particularly important. A key insight gleaned from economic theory is
that the largest savings to an insurer from investigating claims fraud may
stem from the deterrence of fraudulent claiming, rather than from cost
112
See Sykes, supra note 94.
113
ABRAHAM, supra note 95, at 184.
114
Id.
115
Id.
116
Id.
savings associated with the detection of fraudulent claims.117 By reducing

insurer resistance to potentially fraudulent claims and thus increasing the
expected payoff to policyholders from filing fraudulent claims, excessive
liability for insurer bad faith may increase policyholders’ incentives to
engage in claims fraud or exaggeration.118
IV. EMPIRICAL EVIDENCE
A. RELATIONSHIP TO PREVIOUS STUDIES
While the impact of first-party bad faith law on insurer claims

settlement behavior has not been extensively studied, anecdotal evidence
exists from case law to the effect that in some cases tort liability standards
have been too lax and/or damages awards have been too high. Based on an
examination of a variety of cases involving application of first-party bad
faith law, Sykes reached the conclusion that the courts have made
substantial errors in applying the law.119 Specifically, in some of these
cases, the court found insurer bad faith even though the dispute arose as a
result of the insurer’s reasonable suspicion of claims fraud;120 in other
cases, the intentional tort standard was misapplied by the court, with the
result being a finding of insurer bad faith even though the claim was
reasonably debatable;121 in still other cases, the size of a punitive damages
award appeared to be disproportionately high in comparison to the offense
of the insurer.122 Of all the cases examined by Sykes, perhaps the most
perplexing were those cases in which the court found bad faith based on an
insurers’ strict reading of the policy language.123 Sykes concluded that “the
remedy may be worse than the problem, as the courts seem to find bad faith
on the part of insurers who have genuine and reasonable disputes with their
117
Picard provides an excellent discussion of the theoretical literature on
insurance fraud. See Pierre Picard, Economic Analysis of Insurance Fraud, in
HANDBOOK OF INSURANCE 337, 339 (Georges Dionne ed., 2000).
118
See id. at 337.
119
Sykes, supra note 94, at 407-408.
120
See T.D.S. Inc. v. Shelby Mut. Ins. Co., 760 F.2d 1520, 1527-29 (11th
Cir. 1985); Capstick v. Allstate Ins. Co., 998 F.2d 810, 815-16 (10th Cir. 1993).
121
See Aetna Life Ins. Co. v. Lavioe, 475 U.S. 813, 823 (1986).
122
See id. and Nationwide Mut. Ins. Co. v. Clay, 525 So. 2d 1339, 1342,
1344 (Ala. 1989).
123
See Silberg v. Cal. Life Ins. Co., 521 P.2d 1103, 1112 (Cal. 1974); Sparks
v. Republic Nat’l Life Ins. Co., 647 P.2d 1127, 1137 (Ariz. 1982).
policyholders,”124 and that “the ability of the courts to identify

opportunistic behavior…is very much in doubt.”125
In addition, empirical analysis of insurance claims data reveals that
tort liability for first-party insurer bad faith is associated with higher claim
payments. Browne, Pryor and Puelz analyze a large dataset of first-party
automobile insurance claims settled in 38 different states in 1992.126 They
find that, even after controlling for a wide array of claim characteristics and
for other features of states’ legal and claims environments, claim payments
are significantly higher in states that allow tort actions for insurer bad faith
compared to states that do not.127
A Rand Institute study of the impact of the Royal Globe case (the
case that allowed third-party bad faith tort liability claims in California) on
insurance claim payments found similar effects.128 This study found that
when third-party bad faith tort liability claims were allowed, claim
payments for automobile bodily injury liability (BIL) claims in California
were 25 percent higher than similar claims in other states and that this trend
was reversed after the Royal Globe ruling was overturned.129 The Rand
study also found that the number of BIL claims was higher in California
when third-party bad faith tort liability claims were allowed, and this
frequency declined when the Royal Globe ruling was overturned.130
Of course, higher claim payments or claim filing rates should not
be construed negatively if, in the absence of bad faith liability, a tendency
124
125
Id. at 443.
126
Mark J. Browne et al., The Effect of Bad-Faith Laws on First-Party
Insurance Claims Decisions, 33 J. LEGAL STUD. 355, 357 (2004). These authors
study uninsured and underinsured motorist claims using data compiled by the
Insurance Research Council from a survey of closed claims obtained from
insurance companies.
127
Id. at 386.
128
ANGELA HAWKEN ET AL., RAND INST. FOR CIVIL JUSTICE, PUB. NO. MR-
1199-ICJ, THE EFFECTS OF THIRD-PARTY, BAD FAITH DOCTRINE ON AUTOMOBILE
INSURANCE COSTS AND COMPENSATION 49-53 (2001) (analyzing the impact of
Royal Globe Ins. Co. v. Super. Ct., 592 P.2d 329 (Cal. 1979)). In this landmark
case, the California Supreme Court held that third-party claimants may sue insurers
for violating the Unfair Claim Settlement Practices Act. Id. Ultimately, this
decision was overruled in Moradi-Shalal v. Fireman’s Fund Ins. Co., 758 P.2d 58
(Cal. 1988).
129
HAWKEN, supra note 128, at 49-50.
130
Id. at 49.
exists for insurers to underpay or to wrongfully deny claims. Indeed, in

this circumstance, bad faith liability would be performing its intended
function. However, if higher payments or claim filings are occurring
because insurers are less inclined to investigate potential fraud (that is,
unwarranted amounts are paid in order to avoid potential bad faith
liability), this phenomenon should be a source of concern for policymakers.
In order to explore the relevance of this concern, we undertake an
analysis of the relationship between a state’s first-party bad faith regime
and the settlement of insurance claims. We focus on the legal regime
because this affects the expected value of potential damages faced by an
insurer, by affecting both the likelihood of a finding of bad faith and the
damages awarded in the event of such a finding. Tort-based bad faith –
especially using the California rule – generally leads to higher expected
damages for an insurer than contract-based or statutory-based laws, due to
the expanded possibility for a punitive damages award. States in which bad
faith actions are not permitted, or states in which the law is silent on such
actions, will impose lower expected penalties on insurers than other states.
We examine two aspects of insurance claims that may be affected
by bad faith liability: the characteristics of claims (specifically the accident,
injury and medical treatments), and the claim settlement behavior of
insurers (specifically the claim investigations). Consistent with the
economic analysis above, we hypothesize that claims will be more likely to
exhibit characteristics consistent with (possible) fraud in states where
insurers face greater potential liability for bad faith in claims settlement.
We further hypothesize that insurers will be less likely to engage in
vigorous investigation of claims in those states.
We utilize a large database of paid automobile insurance claims to
test these hypotheses. Because the database that we use includes only
claims that are closed with some payment by the insurer, a higher
prevalence of fraud suspicion indicators and a lower prevalence of insurer
investigations among the claims may indicate that insurer claim settlement
behavior is different in states that impose liability for bad faith as compared
to other states.
B. THE DATA
1. Insurance Claims Data
Individual claim data pertaining to uninsured motorist (UM) claims

collected by the Insurance Research Council (IRC) are analyzed.131 UM
coverage is an element of the private passenger automobile insurance
policy, and it provides coverage for bodily injury to the policyholder with
respect to an accident in which the other driver was at fault, but the other
driver was the owner or operator of an “uninsured motor vehicle” (as
defined in the policy). Under this scenario, the injured policyholder files a
UM claim with his own insurer and receives compensation for both
economic and non-economic damages. UM insurance is considered a first-
party insurance contract and, consequently, courts in a number of states
have specifically upheld the applicability of first-party bad faith remedies
in the UM context.132
The data are obtained from a national sampling of claims from a
large number of insurance companies. The original dataset includes nearly
6,000 UM claims closed in 1997 (the latest year data are available to us),
from accidents occurring throughout the entire United States. Most claims
arise from accidents occurring in 1996 or 1997, but accident dates extend
back to 1986.133 The closed claim survey provides a wealth of information
for each claim, including claim characteristics, insurer investigations of the
claim, the claimed amount and the paid amount.
131
See Insurance Research Council Home Page, http://www.ircweb.org. The
IRC is an independent, not-for-profit organization founded in 1977, and it is
supported by leading property-casualty insurance organizations, including
property-casualty insurance carriers and trade associations that represent property-
casualty insurance carriers. Its purpose is to provide timely and reliable
information that examines public policy issues affecting property-casualty insurers,
their customers and the general public. The IRC is devoted solely to research and
the communication of its research findings to interested parties; it does not
advocate for property-casualty insurers on public policy issues as such.
132
Browne, supra note 126, at 360-61.
133
Overall, 76.3 percent of accidents in the dataset occur in 1996-1997, 21.9
percent of accidents occur in 1993-1995 and 1.8 percent of accidents in the dataset
occur prior to 1993.
2. Claim Characteristics
Studies of automobile insurance fraud have developed a catalog of

fraud suspicion indicators or “red flags” that most claim professionals find
to indicate potential fraud.134 The claim characteristics identified as
suspicion indicators encompass a variety of characteristics of the insured,
the accident, the injury and the injury treatment. One fraud suspicion
indicator is the lack of a police report for the accident that produced the
claim.135 The rationale is that in the normal course of an accident, the
police will be called and a report will be filed. If there is no police report,
it is more likely that the accident (and hence the injury) is fictitious.
Another fraud suspicion indicator is the lack of a visible injury at the scene
of the accident.136 While it is possible that a policyholder could realize his
or her injuries only with some delay, if no injury was apparent at the scene
of the accident, the likelihood that the injury is fictitious or exaggerated is
enhanced.
Soft tissue injuries such as sprains and strains are difficult to
medically verify and, therefore, fall into the category of claims that may
not lend themselves to discovery through investigation.137 As a result, this
sort of injury is notorious for being prone to falsification and exaggeration,
and a claim involving only or primarily a sprain injury is a fraud suspicion
indicator for insurers.138 Appropriate treatment of sprain injuries is also
difficult to determine, providing an additional avenue for a policyholder to
falsify the treatment or to exaggerate the amount of treatment. Thus, a
large number of visits to a chiropractor for treatment of injuries allegedly
sustained in an accident is another fraud suspicion indicator or “red
flag.”139
134
See Herbert I. Weisberg & Richard A. Derrig, Fraud and Automobile
Insurance: A Report on Bodily Injury Claims in Massachusetts, 9 J. INS. REG. 523,
534 (1991); Herbert I. Weisberg & Richard A. Derrig, Detection de la Fraude:
Methodes Quantitatives, 35 RISQUES 75-99 (1998) (in English translation).
135
See, e.g., Weisberg & Derrig, supra note 134, at 523.
136
Id. at 534.
137
See Georges Dionne & Pierre St-Michel, Workers' Compensation and
Moral Hazard, 73 REV. ECON. & STAT. 236, 238-39 (1991).
138
See, e.g., Weisberg & Derrig, supra note 134, at 534, 537.
139
Id. at 534, 536.
3. Claim Investigations
Insurers have several methods at their disposal to investigate the

validity of medical claims. One method is a medical audit, which entails
having a medical professional (usually a nurse) review the medical
treatment and bills submitted.140 This review will provide information from
a medical perspective on whether the treatment and billed amounts are
appropriate. Another, more costly and detailed, investigative method is an
independent medical exam (IME).141 An IME is an examination of the
injured policyholder by a medical professional (usually a doctor) chosen by
the insurance company. An IME provides a second medical opinion
concerning the nature and severity of the injuries to the policyholder. An
IME is more expensive than a medical audit and necessitates the
cooperation and involvement of the policyholder.
1. State Bad Faith Liability Regimes
We combine the data on UM claims with data on each state’s legal

regime for first-party insurance bad faith to facilitate a comparison of
outcomes across states with different bad faith regimes. For each claim, we
identify the bad faith regime in effect in the state and year that the accident
occurred. After omitting claims for which the bad faith legal regime cannot
be determined, our sample for analysis contains 5,338 claims from 48 states
and the District of Columbia.142
State bad faith laws are compiled from Jeffrey W. Stempel,
Stempel on Ins. Contracts 10-87 (Aspen Publ’s 3d ed. 2009) (1994).143 The
Appendix table displays the bad faith regimes in effect in each state during
the sample period for the study. Twenty-four states permitted tort-based
bad faith actions during the entire sample period,144 and 4 states permitted
140
Weisberg & Derrig, supra note 134, at III.1.
141
Id.
142
We omit claims for which the accident state or the accident date is
missing. We also omit claims that arise in Pennsylvania, Montana, Puerto Rico and
the Virgin Islands because of ambiguities surrounding the treatment of first part
bad faith claims in these states and territories.
143
STEMPEL, supra note 26, at 10-3 to 10-20; BARRY R. OSTRAGER &
THOMAS R. NEWMAN, HANDBOOK ON INSURANCE COVERAGE DISPUTES 990-1012
(14th ed. 2008) (1988); GENRE & EDWARDS ANGELL PALMER & DODGE LLP, BAD
FAITH LAWS FOR PROPERTY/CASUALTY CLAIMS (2008).
144
Id.
tort-based bad faith actions for a portion of the sample period.145 Seven
states permitted contract-based bad faith actions during the entire sample
period,146 and 2 states permitted contract-based bad faith actions for a
portion of the sample period.147 First-party bad faith actions were
permitted by statute in 5 states during the entire sample period;148 first-
party bad faith actions were not permitted in 4 states during the entire
sample period;149 and first-party bad faith actions were not authorized by
either statute or legal precedent in 4 states during the entire sample
period.150
Table 1 displayed below indicates the number of claims in the
dataset that were filed under each of the bad faith regimes. The majority of
claims in the dataset stem from accidents in states that permit tort actions
for first-party bad faith.
Table 1: Distribution of Claims by Bad Faith Regime

Number of Percent of
Law Regime Claims Claims
No Private Actions Allowed 182 3.4%

No Private Actions Defined 164 3.1%
Contract Law Actions 592 11.1%
Statutory Actions 930 17.4%
Tort Actions, Intentional Standard 832 15.6%
Tort Actions, Negligence Standard 2,638 49.4%
Total Claims 5,338 100.0%
Source: Authors' calculations from IR C survey data.

N ote: Arkansas is included in the intentional tort category.
145
Id.
146
Id.
147
Id.
148
Id.
149
STEMPEL, supra note 26, at 10-3 to 10-20; OSTRAGER & NEWMAN, supra
note 143, at 990-1012; GENRE & EDWARDS ANGELL PALMER & DODGE LLP, supra
note 143.
150
Id.
Specifically, 3,470 claims (65.0% of the sample) arise in states that

permit tort-based bad faith actions, under either a negligence or an
intentional tort standard. A large minority of claims arise in states that
permit contract-based bad faith actions (592 claims or 11.1% of the
sample) or statute-based bad faith actions (930 claims or 17.4% of the
sample). A small minority of claims arise in states that either specifically
do not permit bad faith actions (182 claims or 3.4% of the sample) or are
silent with respect to whether a bad faith action is permitted (164 claims or
3.1% of the sample).
C. ANALYSIS OF CLAIM CHARACTERISTICS AND CLAIM

INVESTIGATIONS
We investigate the effects of first-party bad faith liability on claim

characteristics and claim investigations by conducting t-tests of differences
in mean values of relevant variables across states with different bad faith
regimes. We first compare the states that recognize bad faith actions
(through tort, contract or statute) to the states, which do not recognize
(either do not permit or have not specifically authorized) bad faith actions.
The results of the comparisons are reported in Table 2.
Table 2: Comparison of Bad Faith versus No Bad Faith

M eans T -T est Statistics
A B A vs B
S tates that States that d o
A llow B ad n ot A llow B ad B ad F aith vs N o
F aith A ction s F aith A ctions B ad F aith
Police at the scene 0.833 0.895 -2.98***

A ny police report 0.879 0.958 -4.32***
O n scene police report 0.809 0.855 -2.14**
N o visible injury at scene 0.681 0.556 4.79***

A ny sprain injury 0.833 0.769 3.07***
W orst injury is sprain 0.666 0.586 3.04***
A ny chiropractor visits 0.355 0.364 -0.34

N umber chiropractor visits 25.440 26.300 -0.42
C hiropractor cost/total cost 0.235 0.160 3.22***
A ny medical audit 0.367 0.261 3.95***

E xternal medical audit 0.064 0.086 -1.61
Independent medical exam 0.066 0.263 -13.01***
So urce: Autho rs' calculatio ns fro m IR C survey d ata.

N o te: *** ind icates statistically d ifferent fro m zero at the 1 p ercent co nfid ence level; ** ind icates
statistically d ifferent fro m zero at the 5 p ercent co nfid ence level; * ind icates statistically d ifferent fro m
zero at the 1 0 p ercent co nfid ence level. All are two -sid ed tests. The num b er o f o b servatio ns with no n-
m issing d ata d iffers b y variab le.
The first two columns of the table report mean values of variables
in states that permit bad-faith actions and states that do not, respectively.
The third column of the table reports t-statistics and significance levels for
comparisons of the means across the two sets of states. The t-statistics
provide a test of whether the differences in means across these groups of
states are significantly different from zero at the one percent (***), five
percent (**) or ten percent (*) confidence level.
The comparisons reveal significant differences in claim
characteristics in states that permit bad faith actions. We observe that
police verification of accidents is less prevalent for claims in states that
permit bad faith actions; police are less likely to be at the scene of the
accident, and accidents are less likely to be the subject of a police report, as
compared to other states. Specifically, 80.9 percent of claims paid have a
police report from the scene of the accident in states that permit bad faith
while this is true for 85.5 percent of claims paid in states that do not permit
bad faith actions. The right hand columns of the table confirm that the
differences are statistically significant.
Claims that involve no visible injury at the scene of the accident
are more prevalent in bad-faith states (68.1 percent) than in states with no
bad faith (55.6 percent). Claims in bad-faith states are more likely to
involve a sprain injury (by a margin of 83.3 percent to 76.9 percent).
Similarly, sprain injuries are more likely to be the most severe injury
experienced in bad-faith states (66.6 percent in bad-faith states compared to
58.6 percent in states with no bad faith). All of these differences are
statistically significant, as indicated in the right hand column of the table.
The use of chiropractors by injured policyholders is about the same
across the two sets of states, as is the number of chiropractor visits, and the
differences are not statistically significant. However, the proportion of the
total claimed amount that arises from chiropractor care is larger in bad-faith
states (23.5 percent) compared to states with no bad faith (16.0 percent),
and this difference is statistically significant.
The data also suggest differences in insurer investments in claim
investigation in states that permit bad-faith actions relative to investigations
in other states. Insurers faced with potential bad-faith actions are more
likely to conduct a medical audit (36.7 percent of claims versus 26.1
percent of claims in states with no bad faith actions), and this difference is
statistically significant. However, this result is entirely due to a greater
propensity to conduct in-house medical audits. Insurers in bad-faith states
are slightly less likely to invest in external medical audits than in other
states (although the difference is not statistically significant). This greater
use of in-house medical audits may indicate greater investments in claim
processing bureaucracy so that a defense can be mounted in the event of a
bad-faith lawsuit.
The above interpretation is reinforced by the fact that the potential
for a bad faith claim has the opposite effect on insurers’ IME use; the
proportion of claims for which insurers request an IME is only 6.6 percent
in states that permit bad faith actions, while it is 26.3 percent in states with
no bad faith. This difference is both large and statistically significant.
Because an IME requires the notification and cooperation of the
policyholder, insurers may be particularly reluctant to undertake this type
of investigation when faced with the potential for bad-faith lawsuits. Such
is the case because an IME may lead to an allegation that the insurer
unnecessarily engaged in delay tactics with respect to resolution of the
claim so as to coerce a reduced claim settlement.
Because there are important differences in legal standards and
potential damage awards across different legal regimes for insurance bad
faith, we also separately compare claims filed under tort-based bad faith
regimes to those filed under non-tort-based regimes (contract-based or
statute-based), and to those filed in states that do not recognize first-party
bad faith. Table 3 presents the results of these comparisons.
Table 3: Comparison of Bad Faith Regimes

Means T-Test Statistics
A B C A vs B A vs C B vs C
States with States with States with Tort vs Contract or

Tort-based Contract or No Private Contract Tort vs No Statute vs
Bad Faith Statute Actions or Statute Actions No Actions
Police at the scene 0.814 0.875 0.895 -5.31*** -3.69*** -1.01

Any police report 0.865 0.911 0.958 -4.53*** -4.84*** -2.82***
On scene police report 0.794 0.843 0.855 -4.12*** -2.73*** -0.56
No visible injury at scene 0.699 0.641 0.556 4.08*** 5.48*** 2.94***

Any sprain injury 0.846 0.804 0.769 3.61*** 3.71*** 1.49
Worst injury is sprain 0.690 0.612 0.586 5.43*** 3.96*** 0.89
Any chiropractor visits 0.359 0.346 0.364 0.87 -0.19 -0.63

Number chiropractor visits 23.390 30.230 26.300 -6.59*** -1.59 -1.56
Chiropractor cost/total cost 0.246 0.211 0.160 2.66*** 3.87*** 1.89*
Any medical audit 0.395 0.305 0.261 6.06*** 4.85*** 1.59

External medical audit 0.064 0.063 0.086 0.16 -1.55 -1.54
Independent medical exam 0.043 0.120 0.263 -10.04*** -16.33*** -6.72***
Source: Authors' calculations from IRC survey data.

Note: *** indicates statistically different from zero at the 1 percent confidence level; ** indicates statistically different from zero
at the 5 percent confidence level; * indicates statistically different from zero at the 10 percent confidence level. All are two-sided
tests. The number of observations with non-missing data differs by variable.
The three left-hand columns of the table report mean values of

variables in states that permit tort-based bad-faith actions, states that permit
contract-based or statute-based actions, and states that do not recognize bad
faith actions, respectively. The three right-hand side columns of the table
report t-statistics and significance levels for comparisons of mean values
across these sets of states. As in the previous table, the t-statistics provide
a test of whether the differences in means across these groups of states are
significantly different from zero at the one percent (***), five percent (**)
or ten percent (*) confidence level.
The comparisons are striking, and are consistent with the
hypothesis that the impact of permitting tort-based bad faith actions is
greatest. Claims in tort-based states are notably different from claims in
the other two sets of states. These differences are in the hypothesized
directions and are statistically significant.
Relative to states that permit contract or statute-based actions,

insurance claims in states that permit tort-based bad-faith actions are
significantly more likely to include fraud suspicion indicators (i.e. “red
flags”), including, less prevalence of police verification of accidents,
greater prevalence of claims that involve no visible injury at the scene of
the accident, greater prevalence of claims that involve only sprain injuries,
and greater proportion of treatment costs from chiropractors. The second
of the right hand columns of the table confirms that these differences
between tort-based states and contract or statute-based states are
statistically significant at better than the 1 percent confidence level. Insurer
investigation patterns in states that permit tort-based faith actions also
differ from those in states with contract or statute-based actions. Insurers
faced with potential tort-based bad faith are more likely to conduct medical
audits but are less likely to conduct IME’s than insurers in states with
contract or statutory bad faith regimes, and these differences are
statistically significant. Insurers in tort-based bad-faith states are no more
likely to undertake external medical audits than insurers in states with
contract or statute-based bad faith.
The differences between states that allow tort-based bad faith
actions and states that allow contract or statute-based actions are generally
smaller than those between tort-based states and states that do not
recognize bad faith actions. For example, claims that involve no visible
injury at the scene of the accident are more prevalent in tort-based states
(69.9 percent) than in contract-based or statute-based states (64.1 percent)
or states with no bad faith (55.6 percent). Similarly, sprain injuries are
more likely to be the most severe injury experienced in tort-based states
(69.0 percent in tort-based states, 61.2 percent in contract-based or statute-
based states and 58.6 percent in states with no bad faith). The proportion
of the total claimed amount that arises from chiropractor care is
significantly larger in tort-based states (24.6 percent) compared to contract-
based or statute-based states (21.1 percent) or states with no bad faith (16.0
percent). There is also a large difference in the propensity of insurers to
use IME’s across the different bad faith regimes. The proportion of claims
for which insurers request an IME is only 4.3 percent in states that permit
tort-based bad faith actions, while it is 12.0 percent in states that permit
contract or statute-based claims and 26.3 percent in states with no bad faith.
Furthermore, there are fewer significant differences between states
that permit contract-based or statute-based bad faith and states that do not
recognize bad faith actions. These patterns are consistent with the
hypothesis that the expected value of potential damages faced by an insurer
is an important influence in the claim settlement process, and that the
prospect of punitive damages is of particular importance. In states where

insurers face a greater potential for punitive damages in bad faith actions,
“red flags” are more prevalent in paid claims and insurers are less likely to
employ certain claim investigation techniques.
To further investigate the role of potential punitive damages, we
compare UM claims in states that permit tort-based bad faith actions under
a “negligence” standard to states that permit bad faith actions but require
more stringent standards of proof for punitive damages (i.e., intentional
tort, contract law or statutory law). The results of this comparison,
displayed in Table 4, yield conclusions that are identical to those obtained
from the previous comparisons. This suggests that differences in the
standards required for a finding of bad faith, and the attendant penalties
arising from such a finding, play a significant role in the claim settlement
process.
1. Robustness Checks
One caveat to the previous analysis is that characteristics of

automobile accidents may differ across the states. If accidents tend to be
less severe in states that permit tort-based bad faith actions, this factor
could partially explain the differences in claim characteristics and insurers’
use of investigative techniques. For example, less severe accidents may be
more likely to result in sprain injuries (or only sprain injuries). Or, with
respect to a small claim, an investigation may be cost-prohibitive even if
the claim has “red flags.” Having acknowledged this, observing a larger
proportion of small claims (as opposed to large claims) filed and paid in
states that permit bad faith actions may itself constitute evidence of
incentive distortions induced by bad faith laws. That is, policyholders may
file small, illegitimate claims knowing that the insurer will not fight the
claim because of the exposure to a bad faith lawsuit. On the other hand, in
states that permit contract-based bad faith actions, or no bad faith actions,
insurers may “nickel and dime” small claims because they know that an
insufficient incentive exists for a policyholder to sue for bad faith.
However, these hypotheses cannot be tested because our data do not report
accident or claim frequency by state.
Table 4: Comparison of Negligence Standard and Other Standards

M eans T -T est Statistics
A B A vs B
States w ith States w ith
N egligence- O ther B asis for N egligence T ort
based T ort B ad F aith vs O ther B ad
A ctions A ctions F aith
Police at the scene 0.779 0.894 -10.93***

A ny police report 0.836 0.927 -9.82***
O n scene police report 0.757 0.867 -9.93***
N o visible injury at scene 0.720 0.638 6.22***

A ny sprain injury 0.866 0.796 6.62***
W orst injury is sprain 0.712 0.614 7.36***
A ny chiropractor visits 0.397 0.309 6.49***

N umber chiropractor visits 23.553 28.183 -4.70***
C hiropractor cost/total cost 0.274 0.191 6.92***
A ny medical audit 0.416 0.313 7.55***

External medical audit 0.068 0.058 1.42
Independent medical exam 0.045 0.091 -6.33***
So urc e: Autho rs' calc ula tio ns fro m IR C surve y d a ta.

N o te: *** ind ica tes statistic ally d ifferent fro m zero at the 1 p e rce nt co nfid enc e le vel; ** ind ic ates
statistically d iffere nt fro m z ero at the 5 p erce nt co nfid enc e leve l; * ind ic ates statistically d ifferent fro m
ze ro at the 1 0 p ercent c o nfid ence level. All are two -sid ed tests. T he num b e r o f o b servatio ns with no n-
m issing d ata d iffers a cro ss varia b les.
We are nonetheless able to examine whether the differences in

claim characteristics and insurer investigations remain statistically
significant for claims that are of equivalent size in the different states. We
compare claim characteristics and insurer investigations for claims in the
third quarter (the quarter in which claim amounts fall above the median
claim amount) and the second quarter (the quarter in which claim amounts
fall below the median claim amount) of the distribution of claimed amounts
in our database, respectively. The benefit of this approach is that it controls
for claimed loss amount and focuses attention on the “typical” claim by
removing both very small claims and very large claims from the
comparisons. One drawback of this approach is that it reduces the sample
sizes for our statistical tests.
Table 5 reports the results of the comparison of claims of similar
size in states with different bad faith regimes. Because all of our previous
results are consistent across the various alternative methods of categorizing

states by bad faith regime, we report only the comparison of states with
tort-based bad faith regimes to all other states. The left-hand columns of
the table report mean values and t-test statistics for characteristics of claims
in the second quarter of the claim size distribution, and the right-hand
columns report means and t-test statistics for claims in the third quarter.
Table 5: Comparison of Claims by Quarter

Claims in 2nd Quarter Claims in 3rd Quarter
Means T-tests Means T-tests
States with States with
Tort-based All Other Tort vs All Tort-based All Other Tort vs All
Bad Faith States Other Bad Faith States Other
Police at the scene 0.805 0.880 -3.50*** 0.745 0.874 -4.82***

Any police report 0.850 0.932 -3.82*** 0.808 0.912 -4.35***
On scene police report 0.783 0.868 -3.36*** 0.727 0.84 -4.12***
No visible injury at scene 0.723 0.679 1.50 0.739 0.659 2.74***

Any sprain injury 0.908 0.885 1.20 0.949 0.915 2.14**
W orst injury is sprain 0.758 0.764 0.21 0.797 0.706 3.38***
Any chiropractor visits 0.456 0.399 1.73* 0.616 0.519 3.06***

Num ber chiropractor visits 15.657 15.539 0.13 27.468 27.349 0.09
Chiropractor cost/total cost 0.356 0.306 1.75* 0.470 0.336 4.80***
Any m edical audit 0.445 0.317 4.04*** 0.575 0.361 6.67***

External m edical audit 0.069 0.067 0.11 0.104 0.084 1.03
Independent m edical exam 0.019 0.073 -4.60*** 0.043 0.117 -4.57***
S o ur c e : A utho r s' c a lc ula tio ns fro m I R C sur ve y d a ta .

N o te : *** ind ic a te s sta tistic a lly d iffe r e nt fr o m z e r o a t the 1 p e r c e nt c o nfid e nc e le ve l; ** ind ic a te s sta tistic a lly d iffe r e nt fr o m z e r o a t
the 5 p e r c e nt c o nfid e nc e le ve l; * ind ic a te s sta tistic a lly d iffe r e nt fr o m z e r o a t the 1 0 p e r c e nt c o nfid e nc e le ve l. A ll a r e tw o -sid e d te sts.
T he num b e r o f o b se r va tio ns w ith no n-m issing d a ta d iffe r s b y va r ia b le .
The table shows the same general results as the comparisons based
on all claims in the database. Indeed, the comparisons of claims in the
third quarter follow the exact patterns of signs and statistical significance as
those for the full sample of claims. For the smaller claims (those in the
second quarter) there are some differences from the full sample. Most
notable is that there are no significant differences in injury characteristics
(visible injuries, prevalence of sprain) across states with tort-based bad
faith and other states. This suggests that the relevant difference in claim
characteristics is the greater prevalence of large-valued claims that involve
sprain injuries in states with tort-based bad faith, consistent with the
hypothesis that claim exaggeration may be more prevalent in those states.
Overall, it appears from the table that differences in claim size across states
with different bad faith regimes is not the determining factor in explaining
differences in claim characteristics and claim investigations, although

results for claim characteristics may be somewhat less robust.
A second caveat to the preceding analysis is that results are based
on the pooling of claims from different states into categories based solely
on first-party insurance bad faith regime.151 One concern is that there may
be other important sources of heterogeneity across states; a second concern
is the potential for a large state to dominate the comparisons. A specific
issue that has been considered in other studies is that the compensation
system for automobile accident claims may affect the claim settlement
process.152 To explore whether claim and investigation characteristics
differ across bad faith legal regimes when the accident compensation
regime is held constant, we modify our sample to exclude claims that were
settled under an automobile no-fault compensation system. Additionally,
we investigate the influence of a large state on our previous results by
excluding claims from California (a state with tort-based bad faith) from
the sample. This is done because claims from this state make up a nearly
one-quarter of the sample and thus may be influential on the results.
The results of comparisons of claim characteristics and claim
investigations in states with tort-based bad faith to all other states, for both
of these alternative samples of claims, are displayed in Table 6. The left-
hand columns display means and t-test statistics for the sample that omits
claims settled under no-fault insurance; the right-hand columns display
means and t-test statistics for the sample that omits California.
151
To more fully account for this issue, additional investigations were
undertaken using a logistic regression approach that can account for the clustering
of claims by state. Like the t-tests, this approach analyzes whether different bad
faith regimes are associated with significantly different mean values of each
relevant claim characteristic, but takes into account that claims occur within a state
and the bad faith regime varies only by state and not for each claim. Thus,
clustering by state reduces the likelihood that differences across bad faith regimes
are statistically significant, by allowing for the possibility that variable values are
correlated across claims within each state. Nonetheless, the results of this analysis
remain generally consistent with the t-test results, and are most similar to the t-test
results for the sample omitting the state of California. Claim injury characteristics
remain statistically significant and higher in tort-based states, and insurer
investigations also remain statistically significant and of the same signs as in the t-
test analysis. Police presence and police reports and use of chiropractic treatments
become statistically insignificant in the logistic regression analysis.
152
Browne, supra note 126, at 360-61; HAWKEN, supra note 128.
The results for the sample which omits claims settled under a no-
fault regime are extremely similar to results obtained from the full sample
of states: the signs and statistical significance of all variables remain the
same. This confirms that the automobile accident compensation regime is
not the determining factor in explaining differences in claim characteristics
and claim investigations.
Table 6: Comparison of Alternative Samples

Sample omitting Nofault States Sample omitting California
Means T-tests Means T-tests
States with States with
Tort-based All O ther Tort vs All Tort-based All O ther Tort vs All
Bad Faith States Other Bad Faith States O ther
Police at the scene 0.813 0.868 -4.55*** 0.901 0.879 2.01**

Any police report 0.864 0.908 -4.122*** 0.942 0.919 2.72***
On scene police report 0.793 0.833 -3.20*** 0.879 0.846 3.04***
No visible injury at scene 0.701 0.633 4.54*** 0.664 0.625 2.53**

Any sprain injury 0.847 0.799 3.96*** 0.802 0.798 0.32
W orst injury is sprain 0.693 0.645 3.15*** 0.641 0.607 2.15**
Any chiropractor visits 0.360 0.300 3.94*** 0.275 0.35 -5.05***

Num ber chiropractor visits 23.412 23.272 0.13 22.318 29.636 -5.58***
Chiropractor cost/total cost 0.247 0.192 3.96*** 0.174 0.201 -2.08**
Any m edical audit 0.396 0.280 7.46*** 0.343 0.27 3.02***

E xternal m edical audit 0.063 0.054 1.20 0.059 0.067 1.08
Independent m edical exam 0.040 0.069 -4.10*** 0.048 0.146 -10.31***
S o ur c e : A utho r s' c a lc ula tio ns fr o m I R C sur ve y d a ta .

N o te : *** ind ic a te s sta tistic a lly d iffe r e nt fr o m z e r o a t the 1 p e r c e nt c o nfid e nc e le ve l; ** ind ic a te s sta tistic a lly d iffe r e nt fr o m z e r o a t
the 5 p e r c e nt c o nfid e nc e le ve l; * ind ic a te s sta tistic a lly d iffe r e nt fr o m z e r o a t the 1 0 p e r c e nt c o nfid e nc e le ve l. A ll a r e tw o -sid e d te sts.
T he num b e r o f o b se r va tio ns w ith no n-m issing d a ta d iffe r s b y va r ia b le .
In contrast, results for the sample from which California claims are
omitted differ from previous comparisons in several ways. Most notable
are the changes in sign for the police report and chiropractor use variables.
In the sample without California, police are more likely to be at the
accident scene and to submit a report of the accident in states with tort-
based bad faith than in other states. The percent of claim costs stemming
from chiropractor use is also smaller in tort-based bad faith states than in
other states, once California is omitted from the sample. Thus it appears
that these claim characteristics may be prevalent in California rather than in
states with tort-based bad faith more generally. On the other hand, several
of the comparisons without California remain statistically significant and
consistent with the findings for the full sample of states. Claims with no
visible injury at the accident scene and claims in which the worst injury is a
sprain remain significantly more prevalent in states with tort-based bad
faith than in other states. Moreover, the pattern of insurer claim

investigation activities remains the same in this sample as in the full
sample; insurers in states with tort-based bad faith are more likely to invest
in internal routine medical audits and less likely to invest in independent
medical examinations. Thus, there remain important differences in claim
“red flags” and insurer investigations between states with tort-based bad
faith and other states, even after the influence of California is removed
from the sample.
V. CONCLUSION
This article has examined first-party insurance bad faith remedies

under common law and the recent legislative expansion of such remedies.
Theory predicts that allowing policyholders to recover damages over and
above the value of the insurance benefit owed will provide insurers with
added incentives to engage in fair claim settlement practices, with the result
being an enhancement in the efficiency of insurance market contracting.
However, theory also predicts that uncertain bad faith standards for insurers
and excessive punitive damages awards for policyholders will undermine
the benefits of the bad faith remedy by distorting insurers’ claim settlement
practices and policyholders’ claim filing incentives, in ways that will lead
to more borderline (or even fraudulent and exaggerated) claims and
unwarranted increases in insurance premiums.
Previous empirical studies have found that tort-based standards for
insurer bad faith are associated with higher insurance claim payments.
This article notes that higher claim payments may be evidence of beneficial
effects of bad faith liability, if in its absence insurers would underpay
claims. A more pertinent concern is whether tort liability for insurer bad
faith deters insurers from appropriately challenging potentially fraudulent
or otherwise invalid claims, leading to greater amounts of fraud and to
unwarranted costs and higher insurance premiums. The article provides
new evidence that tort liability for first-party bad faith may reduce insurers’
incentives to monitor for claim fraud, leading to less intensive use of
investigative techniques and to more paid claims that contain
characteristics often associated with fraud. Although constructing a
baseline comparison for determining the appropriateness of claim
investigations is difficult, these findings are consistent with the predictions
of theory when bad faith liability is uncertain and/or excessive. This raises
questions about whether tort liability facilitates efficient claim settlement
practices in insurance markets. Additional empirical study of the
relationship between bad faith liability standards and the insurance claim
settlement process would be useful.
Appendix Table
State First-Party Bad Faith Regimes for Sample Period of Claims Data
State First Party Bad Faith Law State First Party Bad Faith Law
Alaska Tort Actions Mississippi Tort Actions
Alabama Tort Actions Nebraska Tort Actions
Arkansas Tort Actions New Hampshire Contract Law Actions
Arizona Tort Actions New Jersey Tort Actions (until 1993)
California Tort Actions Contract Law Actions (since 1993)
Colorado Tort Actions New Mexico Tort Actions
Connecticut Tort Actions Nevada Tort Actions
District of Columbia No Private Actions Defined New York No Private Actions Allowed
Delaware Tort Actions (until 1995); North Carolina Tort Actions
Contract Law Actions (since 1995) North Dakota Tort Actions
Florida Statutory Actions Ohio Tort Actions
Georgia Statutory Actions Oklahoma Tort Actions
Hawaii No Private Actions Defined Oregon Contract Law Actions
Iowa Tort Actions Rhode Island Tort Actions
Idaho Tort Actions South Carolina Tort Actions
Illinois Tort Actions South Dakota No Private Actions Defined
Indiana Tort Actions (since 1993) Tennessee Statutory Actions
Kansas No Private Actions Allowed Texas Tort Actions
Kentucky Tort Actions Utah Contract Law Actions
Louisiana Statutory Actions Virginia Contract Law Actions
Massachusetts Statutory Actions Vermont Tort Actions
Maryland Contract Law Actions Washington Tort Actions (since 1992)
Maine Contract Law Actions Wisconsin Tort Actions
Michigan No Private Actions Allowed West Virginia Contract Law Actions
Minnesota No Private Actions Allowed Wyoming Tort Actions (since 1990)
Missouri No Private Actions Defined
Source: Authors' calculations from GenRe (2008), Stempel (2006) and Ostrager and Newman (2008).
Note: The sample period includes years 1986 through 1997.
REGULATION OF LARGE FINANCIAL INSTITUTIONS:
LESSONS FROM CORPORATE FINANCE THEORY
John P. Harding*
Stephen L. Ross **
***
This article applies a model of firm capital structure to the current
financial crisis and summarizes the insights the model provides regarding
the regulation of large financial institutions in a post-crisis world. Firm
capital structure is evaluated by studying how firms finance their activities
using debt and equity, which in turn captures an important element of firm
risk taking. First, the simple model is briefly summarized. Second, the
model’s results are used in order to interpret the evolution of the current
financial crisis and to put it into perspective. Finally, the article presents
forward-looking observations and suggestions for future regulation. The
article concludes that an essential element of a new regulatory framework
must include an effective method for dealing with the extension of the “Too
Big to Fail” umbrella, which has extended the risk of moral hazard beyond
depository institutions. It asserts that a successful new framework must
include stringent capital standards for financial institutions combined with
regulators that have the authority and commitment to enforce those
standards putting owners and managers at risk when capital standards are
violated, even during financial crises when there are strong incentives for
regulator accommodation to preserve asset value. The new framework
must be flexible in order to adapt to changing financial conditions
especially developments that affect franchise value, and contain provisions
that expose uninsured debtors to risk when capital standards are violated
so that debt holders have an incentive to monitor the activities of very large
financial firms.
***
*
John P. Harding, Professor, Finance Department, University of Connecticut
**
Stephen L Ross, Professor, Economics Department, University of
Connecticut. The authors benefited from conversations with Dwight Jaffe, Ed
Kane, Pat McCoy and seminar participants at the Federal Reserve Bank of Boston.
I. INTRODUCTION
Equity capital is the shock absorber for our financial system and
the current financial crisis, like a bumpy road for an auto designer, provides
a unique opportunity for financial regulators to evaluate the predictions of
theory and improve the design of the regulatory system. The purpose of
this paper is to apply a simple model of firm capital structure to the current
situation and summarize the insights it provides regarding the regulation of
large financial institutions in a post-crisis world. The paper begins with a
brief summary of the model and uses the results of that model to place the
evolution of the current crisis into perspective. The paper concludes with
forward-looking observations and suggestions for future regulation.
II. A SIMPLE MODEL OF FIRM CAPITAL STRUCTURE
The study of firm capital structure is basically the analysis of how

firms finance their economic activities – in particular, the allocation of
funding between equity and debt or firm leverage (the ratio of debt to
equity). In much of the literature, the fundamental risk of the firm’s assets
is taken as given by economic forces outside the control of the firm and the
choice of leverage can be viewed as allocating that risk between different
providers of capital. In general, the use of debt financing results in creating
a lower risk/lower return investment opportunity for debt providers and a
higher risk/higher return investment for equity providers.
Leland develops a simple model of the choice of firm capital
structure that has proven to provide valuable insights into important
questions.1 In Leland’s model, a private firm in a given industry must fund
its initial investment in productive assets with a combination of debt and
equity. The return on the firm’s assets evolves stochastically, but the firm
has limited ability to adjust its capital structure in response to the actual
asset outcomes.2 As a result, the firm’s managers must choose their initial
leverage to maximize the expected present value of the firm’s future
operations in the presence of taxes, bankruptcy costs and tax-deductible
1
Hayne E. Leland, Corporate Debt Value, Bond Covenants, and Optimal
Capital Structure, 49 J. OF FIN. 1213 (1994).
2
Id. at 1248. While such a restriction may seem unrealistic, the recent crisis
has demonstrated that firms experiencing financial difficulties often have no access
to additional equity capital at the time of distress. This fact places more
importance on the forward-looking choice of initial leverage. Id.
2009] REGULATION OF FINANCIAL INSTITUTIONS 245
interest payments on the debt. In this model, firms rationally choose to

fund themselves with some debt and some equity as they trade-off expected
bankruptcy costs that could be avoided if the firm chose all equity and the
tax advantage of debt, which is maximized by choosing all debt.
A key lesson of this paper arises from the author’s consideration of
the firm’s choice of a bankruptcy threshold – the level of asset value at
which the firm’s owners voluntarily turn over the assets to the debt holders
for liquidation. In financial terms, the firm’s owners have a call option on
the assets with a strike price equal to the debt payoff. Because an option
can never have a negative value, ignoring the periodic debt service
payment, firm equity could never be negative and the owners would never
choose bankruptcy.3 However, to keep their option alive, the owners must
make the periodic debt service payments and thus choose bankruptcy when
the cost of the required debt service payment exceeds the current call
option value. Two important insights from Leland’s analysis are that for
reasonable parameter assumptions owners will choose to continue to
operate the firm (maintain their option value) until the market value of
assets falls well below the face value of debt and the key factor that limits
this behavior of owners is the required debt service to bond holders.
Because equity holders can choose the best time (from their
perspective) to turn over the firm to debt holders, the bond holders are
exposed to greater risk by the continued operation of an insolvent firm.
Once declines in asset values have made a firm insolvent, the equity
holder’s decision whether to continue operating the firm only depends upon
the subset of outcomes where asset values recover sufficiently for the firm
value to exceed debt obligations. The return to equity holders is zero in all
other scenarios, no matter how bad the outcome. However, debt holders
care about the expected future value of assets over all states of the world,
but are especially concerned about situations where asset values continue to
erode because they will be the owner of the assets in those states of the
world. Furthermore, insolvency is likely to affect the equity holders’
choice of investment risk because, from their perspective, they still benefit
from exceptionally good outcomes and do not bear increased losses if the
3
A call option provides the holder the right, but not the obligation, to
purchase a security or asset at a fixed price. Such an option always has a positive
value regardless of the current value of the asset as long as there is some chance
that the value of the asset may rise above the strike price. For a complete
introduction to options and their valuation see JOHN C. HULL, OPTIONS, FUTURES,
AND OTHER DERIVATIVES (Prentice-Hall) (2008).
bad outcomes become worse. Such increased investment risk exposes debt
holders to substantially more risk than the likely investment decisions of a
solvent firm.4
The providers of debt capital can anticipate this behavior on the
part of owners and demand fair compensation for bearing the risk of
significant losses from asset declines by increasing the required coupon
rate on the debt. However, this creates “risky” debt and limits the ability to
allocate risk that was one of the original motivations for leverage.5 A
common solution to this problem is for equity and debt providers to
negotiate bond covenants whereby owners contract to transfer control of
the firm to debt holders at a higher threshold value of assets. For example,
a positive net worth covenant would transfer control when the market value
of assets first falls below the face value of debt.6
A final important result from Leland’s model is that whether or not
the debt is protected by bond covenants, the optimal choice of initial
leverage provides a substantial equity “cushion” above the bankruptcy
threshold. Leverage levels near the bankruptcy threshold raise the risk of
incurring the transactions costs associated with bankruptcy. Since debt
holders will bear those costs in bankruptcy, firm owners must pay for those
costs with higher interest rates and so have an incentive to hold additional
equity in order to lower the interest rate on debt.
4
This behavior arises from the limited liability exposure of equity holders, so
that owners of an insolvent firm face no downside risk and benefit from large risks
that have a low probability of success, but, if successful, create the potential for
future positive net worth of the firm. FRANK H. EASTERBROOK & DANIEL R.
FISCHEL, THE ECONOMIC STRUCTURE OF CORPORATE LAW 49-50 (1996).
5
To see the intuition underlying this claim, envision a continuum of risk
measured along a straight line where the firm’s assets are located in the middle of
the line. Adding leverage to the firm creates one security with less risk (debt) and
another with more risk (leveraged equity) – splitting the single center point into
two, one to the safer side of center and one to the riskier side. To the extent that
more of the risk of future outcomes falls on the debt holders, the two new points
must remain closer to the original center point.
6
There are many possible variations of this basic idea. For example, some
covenants might limit the managers’ ability to finance new growth or sell assets in
lieu of transferring total control to debt holders. Another variation might provide
debt holders representation on the board when certain financial thresholds are
crossed. Clifford W. Smith & Jerold B. Warner, On Financial Contracting: An
Analysis of Bond Covenants, 7 J. OF FIN. ECON. 117 (1979).
Regulated depository institutions have two significant differences

from the firms Leland studies: they are able to issue debt at the riskless
rate without regard to their individual financial condition or the selected
leverage, and they are subject to binding capital regulation. These capital
regulations can be viewed in Leland’s terms as an exogenously determined
bankruptcy threshold at which level the debt holders are paid in full and the
assets turned over to the regulator for liquidation. As a result of the
reduced market discipline via debt cost, many scholars have discussed the
concern that deposit insurance creates a moral hazard that encourages
owners of insured depositories to select very high levels of leverage, i.e.,
hold the minimum capital allowed by the capital regulations.7 On the other
hand, Buser, Chen and Kane (1981) point out that banks incur significant
costs that are not explicitly priced attributable to regulations, investment
restrictions and monitoring.8 Merton (1978) and Markus (1984) argue that
these costs increase when banks have weak capital positions creating an
equity favoring factor that offsets the incentives created by deposit
insurance and could lead to banks choosing to hold high levels of equity
and so provide very limited deposit services.9
Ideally, regulation should preserve incentives for owners to operate
as deposit issuing banks, while selecting a prudent level of leverage
tailored to the riskiness of their asset strategy – voluntarily holding capital
in excess of the specified minimum. It is unreasonable and impractical to
expect regulators to be able to establish capital standards tailored to an
individual bank’s operating strategy and asset risk. Similarly, regulators
cannot monitor the portfolios of all financial institutions closely enough to
accurately classify banks by the riskinesss of their assets. Consequently, it
is important that the capital regulation framework provides owners and
managers sufficient incentives to provide a capital cushion that reflects the
7
See Jean-Pierre Gueyie & Van Son Lai, Bank Moral Hazard and the
Introduction of Official Deposit Insurance in Canada, 12 INT’L REV. OF ECON. &
FIN. 247 (2003); Michael C. Keeley, Deposit Insurance, Risk, and Market Power
in Banking, 80 THE AM. ECON. REV. 1183 (1990); and David A. Marshall &
Edward S. Prescott, Bank Capital Regulation With and Without State-Contingent
Penalties (Federal Reserve Bank of Chicago, Working Paper No. 2000-10, 2000).
8
Stephen A. Buser, Andrew H. Chen & Edward J. Kane, Federal Deposit
Insurance, Regulatory Policy and Optimal Bank Capital, 36 J. OF FIN. 51, 56
(1981).
9
Alan J. Marcus, Deregulation and Bank Financial Policy, 8 J. OF BANKING
AND FIN. 51, 59 (1984); Robert C. Merton, On the Cost of Deposit Insurance When
There are Surveillance Costs, 51 J. OF BUS. 439, 448 (1984).
operating strategy of the bank and appropriately reduces the risk of

violating the standards in adverse times.
Harding, Liang, and Ross (2008) (HLR) extend Leland’s model to
incorporate the additional factors influencing bank capital choices.10 They
find that capital regulation alone is not effective for establishing an
incentive for banks to hold capital reserves in excess of the minimum
capital standards. Bankruptcy costs and insurance benefits both vary with
capital structure in such a way that banks would rationally choose to
operate with the minimum capital allowed or forego being a bank entirely
(i.e., choose all equity financing), depending upon the strictness of capital
regulation. A bank will only choose to hold capital reserves in excess of
minimum capital requirements if there is some additional firm franchise
value that is placed at risk by the capital regulation.
HLR find that tax-advantaged debt as parameterized by Leland
(1994), or any franchise value that depends upon the total amount of an
institution’s deposits, creates such a franchise value and, when this
franchise value is considered, banks will hold excess capital without
shifting to the extreme of foregoing issuing deposits. Counter-intuitively,
the tax benefits of debt or the franchise benefits of the deposit base lead to
lower leverage which is quite striking given that tax-advantaged debt
traditionally is viewed as increasing the incentive for firms to take on more
debt. The result is explained by the fact that a firm’s franchise value is
placed at risk by capital standards causing firms to choose lower levels of
leverage in order to protect this value for equity holders. Overall, the
combination of capital standards with the power of the regulator to wipe
out equity when the standard is violated acts like the bond covenants
discussed in Leland encouraging managers to operate with a capital
cushion above the threshold that would trigger takeover.
HLR demonstrate that it is the regulator’s ability to move swiftly to
place a firm that violates capital standards into receivership that leads firms
to hold excess capital. Alternative structures such as warning thresholds
that trigger additional regulatory burdens have little effect on leverage. It is
the ability to liquidate the firm, placing the franchise value at risk, that
leads to more conservative choices of leverage. Consistent with these
predictions, the authors note that commercial banks, that are subject to
dissolution if capital standards are violated, hold a substantial capital
10
John P. Harding, Xiaozhong Liang & Stephen L. Ross, Bank Capital
Requirements and Capital Structure, 4-6 (Univ. of Conn., Dept. of Econ, Working
Paper No. 2009-09, 2008).
cushion above the minimum capital requirement while the mortgage GSEs,
Freddie Mac and Fannie Mae, who prior to the summer of 2008 were free
from the risk of receivership, held little or no capital cushion through the
same period.11
III. UNDERSTANDING THE CURRENT CRISIS
The origins of the current financial crisis are complex and no single
factor can be singled out as the primary cause. However, most observers
believe that increasing use of leverage, broadly defined, was a contributing
factor. The U.S. Government Accountability Office (GAO) report to
Congress on the origins of the crisis shows that total debt in the U.S.
economy rose significantly in the years preceding the crisis. Measured as a
ratio to nominal GDP, total debt rose from roughly two times GDP in 2002
to 2.25 times GDP in 2007.12
11
Fannie Mae and Freddie Mac first began operating under the statutory
minimum capital requirement in 1993, and in that year they held excess capital of
roughly $1 billion and $0.7 billion, respectively. These amounts expressed as a
percentage of assets plus Mortgage Backed Securities (MBS) outstanding were
.14% and .13%, respectively. In most of the eight subsequent years, Fannie Mae
held excess capital well under $1 billion and in 1998 and 1999 its excess capital
was 1/100th of a percent of the assets and MBS. Freddie Mac’s excess capital,
while slightly higher when measured as a percentage of assets and MBS, was
smaller when measured in dollars and is also consistent with the claim that the firm
intended to meet, but not exceed its capital standard. On average over the period
from 1993 through 2001, Fannie Mae and Freddie Mac held less than 1/10th of a
percent of excess capital. The period from 2002 through 2007 is distorted by the
effects of financial restatements arising from accounting problems experienced in
the period from 2003 through 2005. However, the numbers for 2007 most likely
reflect the firms’ contemporaneous intentions, and they still suggest that the firms
were not holding precautionary excess capital. See FHFA AND OFHEO ANNUAL
REPORTS TO CONGRESS 1994 - 2008 available at http://www.fhfa.gov/
Default.aspx?Page=240. From 2001 through 2008, commercial banks held total
capital (Tier I plus Tier II) of 12.3% of risk-based assets. During this period, a
bank was deemed to be well-capitalized with a total capital ratio of 10% of risk-
based assets and adequately capitalized with a ratio of 8%. Data on commercial
bank capital ratios is available at http://www2.fdic.gov/SDI.
12
See, e.g., U.S. GOV’T. ACCOUNTABILITY OFFICE, FINANCIAL MARKETS
REGULATION: FINANCIAL CRISIS HIGHLIGHTS NEED TO IMPROVE OVERSIGHT OF
LEVERAGE AT FINANCIAL INSTITUTIONS AND ACROSS SYSTEM, GAO-09-739, p.13,
By their nature, financial institutions have always been some of the

most highly leveraged firms. The five largest U.S. investment banks
together had an average leverage ratio of about 30 to 1 during the Asian
Financial Crisis in 1997. While this ratio declined to 22 to 1 in the period
following 1997, it had risen back to 30 to 1 by 2007. The five largest bank
holding companies had an average leverage ratio of about 13 to 1
throughout the same period.13
However, the effective growth of leverage at financial institutions
in recent years is difficult to measure precisely because recent
developments in financial assets and derivatives allow institutions
numerous opportunities to effectively leverage their risk “off balance
sheet” while still maintaining excess capital as measured by traditional
measures of on-balance sheet risks. For example, before the development
of credit default swaps, an institution willing to bear the credit risk of an
industrial firm would make a loan or purchase the debt of that company and
report the loan as an asset, thereby increasing its required capital. In recent
years, institutions could take on that same credit risk without making a loan
by writing an insurance policy through a credit default swap.
Leverage in the economy also increased with greater use of hedge
funds, Special Purpose Entities (SPE) as part of holding company
structures and the ability to issue structured debt securities. One sector of
the economy that made prominent use of structured debt to expand credit
was the housing sector where subprime mortgages were used to expand the
population of mortgagors. The resulting pools of subprime mortgages were
structured in ways to attract investment from a wide array of non-
traditional investors. While most large investment banks and bank holding
company subsidiaries that originated subprime mortgages operated with the
intent to pool and sell mortgage-backed securities as soon as a sufficient
number of loans had been originated, at any given time, they nevertheless
had significant exposure to subprime loans because they were holding
mortgages as inventory awaiting future sales or holding securities as part of
their underwriting and trading operations. In some cases these institutions
Fig.3 (2009) (hereinafter GAO REPORT) available at www.gao.gov/

new.items/d09739.pdf.
13
Id. at 19. At the time, the five largest investment banks or broker-dealer
holding companies were Bear Stearns, Goldman Sachs, Lehman Brothers, Merrill
Lynch, and Morgan Stanley. The five largest bank holding companies were Bank
of America, Citigroup, JPMorgan Chase, Wachovia, and Wells Fargo. Id. at 18-19
figs.4 & 5.
also bought mortgage related securities as investments through special

purpose entities or subsidiaries.14
The role of SPE’s is especially instructive. If a regulated financial
institution committed to provide contingent funding for a SPE, such as
liquidity facilities or credit enhancements, the institution would have been
required to hold capital against that commitment. In many cases, however,
the institutions establishing the SPE’s provided no such guarantees and so
the business activities of the SPE were not considered in calculating capital
requirements. Nevertheless, these institutions still faced both reputation and
legal risks associated with the possibility of an SPE failure. Many of these
SPE’s increased returns by investing in long-term assets like mortgage-
backed securities and financed these investments with short term
commercial paper. As the financial crisis began in 2007, many of these
SPE’s could not renew their debt financing and the regulated financial
institutions either extended financing themselves or directly brought the
SPE’s onto their books in order to avoid the reputation damage associated
with SPE failure.15
The first signs of the current crisis arose in the subprime mortgage
market where delinquencies and defaults began to increase in the first half
of 2007. In June, 2007, Standard and Poor’s and Moody’s began to
downgrade structured debt backed by subprime mortgage securities. This,
in turn, led Bear Stearns to suspend redemptions in certain subprime
investment funds it was managing, and later in July to liquidate two hedge
funds that invested in subprime mortgage-backed securities. By March
2008, Bear Stearns was taken over by JPMorgan Chase in a deal brokered
by the U.S. government.16 Throughout this period, subprime mortgage
securities market values fell sharply as uncertainty increased about the
ability of the structures to withstand higher levels of delinquency and
default.17 These concerns were exacerbated when home prices began to fall
in several major markets.18
14
Id. at 19-20.
15
Id. at 56-58.
16
For a detailed timeline of events, see The Federal Reserve Bank of St.
Louis, The Financial Crisis: A Timeline of Events and Policy Actions,
http://timeline.stlouisfed.org/index.cfm?p=timeline.
17
See, e.g., Roddy Boyd, How the Bear Stearns Deal Got Done, CNN
MONEY.COM, Mar. 17, 2008, http://money.cnn.com/2008/03/17/news/companies
/boyd_bear.fortune/index.htm?postversion=2008031717. See also, Economic
Outlook Before the S. Comm. On Banking, Housing, and Urban Affairs, 110th
Cong. (2008) (statement of Federal Reserve Chairman Ben S. Bernanke)
While much of the public attention was focused on “toxic assets”,

the combination of declining asset values with high leverage and increasing
uncertainty and risk aversion resulted in a rapid spread of the crisis
throughout the global financial system. Asset value declines forced
institutions to deleverage initially by raising capital,19 but as raising capital
became more difficult institutions were forced to deleverage by selling
assets into markets dominated by sellers.20 Brunnermeier and Kashyap,
Rajan, and Stein21 emphasize the role of deleveraging in asset price
declines suggesting that deleveraging could trigger downward spirals as
asset sales depress asset prices requiring further deleveraging. Further,
Mark-to-Market22 rules could exacerbate a deleveraging based price spiral
as financial institutions are forced to re-value assets in the face of rapid
declines in asset prices and then due to Mark-to-Market rules must respond
to those lower values with further deleveraging. However, it is important
to recognize that the falling prices of assets are not simply the result of
available at http://banking.senate.gov/public/index.cfm?FuseAction=Files.
View&FileStore_id=5a34e5df-2de7-4c9b-a5a7-4f0893acd79e.
18
See, e.g., Lee Christie, Home Prices See Another Record Plunge, CNN
MONEY.COM, Oct. 28, 2008, http://money.cnn.com/2008/10/28/real_estate/
August_Case_Shiller/index.htm. See also Lee Christie, Home Prices in Record
Decline, CNN MONEY.COM, Nov. 25, 2008, http://money.cnn.com/2008/11/25/
real_estate/third_quarter_case_shiller/index.htm.
19
GAO REPORT, supra note 13, at 20.
20
If a firm is levered 30:1 and experiences a 10% decline in value on 10% of
its assets, its equity base declines by 30% and its leverage increases to more than
40:1. In order to return leverage to the original ratio of 30:1, it must sell close to
one third of its assets.
21
Markus K. Brunnermeier, Deciphering the 2007-08 Liquidity and Credit
Crunch, 1 J. OF ECON. PERSPECTIVES 23, 77-100 (2009); Anil K., Kashyap,
Raghuram G. Rajan, & Jeremy C. Stein, Rethinking Capital Regulation, Jackson
Hole Symposium: “Maintaining Stability in a Changing Financial System” (2008),
available at http://kansascityfed.org/publicat/sympos/2008/KashyapRajanStein.
03.12.09.pdf.
22
Mark-to-Market is a way to measure assets and liabilities that appear on a
company’s balance sheet and income statement that involves an attempt to measure
companies’ assets and liabilities at fair or market value. For more detailed
information, see SEC. AND EXCH. COMM’N, OFFICE OF CHIEF ACCOUNTANT & DIV.
OF CORP. FIN., REPORT AND RECOMMENDATIONS PURSUANT TO SECTION 133 OF
THE EMERGENCY ECONOMIC STABILIZATION ACT OF 2008: STUDY ON MARK-TO-
MARKET ACCOUNTING (2008) available at www.sec.gov/news/studies/2008/
marktomarket123008.pdf.
psychology and forced sales, but also reflect declines in value due to
greater economic uncertainty in general and/or a higher likelihood of
extreme or tail events in the fundamental markets on which those assets
draw their value. Further, even without a deleveraging price spiral, real
declines in complex asset values such as subprime mortgage-backed
securities could be exacerbated by a lemons problem where firms have
private information on the quality of securities and sell the lowest quality
assets as they deleverage.23
The initial regulatory reaction focused on the immediate symptom
– the freezing up of markets – by providing liquidity in the hope of
stabilizing the markets by stimulating buyers of assets. In August 2007, the
Federal Reserve publicly emphasized its intention to provide reserves as
necessary to meet the needs of depository institutions,24 and then lowered
the federal funds rate throughout the fall of 2007. In December 2007, the
Federal Reserve Board announced the creation of a Term Auction Facility
(TAF) that would auction fixed amounts of term funds to depository
institutions allowing those institutions to use a wide variety of assets
including mortgage-backed securities as collateral. This action was
followed by related efforts throughout the spring and summer of 2008
including legislative authority to extend credit to the Government
Sponsored Enterprises (GSEs) Fannie Mae and Freddie Mac.25 These
efforts proved to be ineffective in containing the crisis and repeated waves
23
See GAO REPORT, supra note 13, at 15-23 for a general discussion of the
deleveraging of financial institutions during this period.
24
The Federal Reserve is often viewed as limiting the use of its discount
window to meet reserve requirements by requiring detailed explanation for
substantial requests. ROBERT E. HALL & JOHN B. TAYLOR, MACROECONOMICS:
THEORY, PERFORMANCE, AND POLICY, 327 (2nd ed., 1988). Banks usually meet
this requirement via inter-bank lending. See OECD Glossary of Statistical Terms,
http://stats.oecd.org/glossary/detail.asp?ID=1385. Accordingly, this statement
addressed the increasing illiquidity in inter-bank lending markets as indicated by
an increasing London Inter-Bank Offer Rate (LIBOR). Asani Sarkar, Liquidity
Risk, Credit Risk, and the Federal Reserve’s Response to the Crisis. Federal
Reserve Bank of New York Staff Report, available at http://www.newyorkfed.org
/research/staff_reports/sr389.pdf.
25
Credit was not granted prior to the GSEs being placed in receivership in
September. However, the recovery act passed in the summer of 2008 did authorize
the U.S. Treasury Department to extend credit to the GSEs if necessary and was
intended to improve the GSEs borrowing ability by increasing investor confidence.
See Timeline, supra note 17.
of deleveraging asset sales simply led to further asset value declines and
the need for additional deleveraging. By the end of the summer of 2008,
U.S. regulators shifted focus from providing liquidity and shoring up
specific markets to working to preserve the solvency of financial
institutions through the infusion of capital. Most notably, this shift was
signaled by the U.S. Treasury Department proposed legislation to purchase
“troubled assets” of financial institutions that was eventually passed by
congress in October as the Troubled Asset Relief Program (TARP). Soon
after passage, the plan to purchase “troubled assets” was abandoned and
replaced by efforts to directly infuse capital into financial institutions
through the purchase of preferred shares of stock.26
Clearly leverage was a key element of the financial crisis and
understanding the drivers that led to the sharp increase in leverage is
critical for planning the future regulatory regime.27 HLR’s model of bank
capital regulation provides useful insights. First, consider entities such as
investment banks, the two large mortgage GSEs and other non-depository
financial institutions. Most of the larger firms in this category had
significant benefits from relationships with the federal government and
arguably were protected by the “Too Big to Fail” principle. Such firms
differ significantly from the purely private firms studied by Leland in that
they benefit from an implicit government guarantee of liabilities due to the
risk their failure would pose to the national economy. This implicit
guarantee meant that these institutions were similar in some respects to
depository institutions in that they could issue debt at lower cost and with
less market scrutiny than private firms. Significantly, however, there was
26
See id.
27
As noted earlier, leverage cannot be singled out as the sole or even primary
contributing factor. Like a “perfect storm” several factors came together to create
the financial crisis. Low interest rates encouraged institutions and investors to seek
out higher risk/higher return investment opportunities. Technology enabled
investment banks to quickly create and sell structured securities. The global trade
imbalances led to rapid growth of dollar balances in the portfolios of developing
nations and as those investors sought out higher return investments they financed
increased leverage by consumers and institutions in developed countries. Ethan
Cohen-Cole et al., Looking Behind the Aggregates: A Reply to “Facts and Myths
about the Financial Crisis of 2008” (Fed. Reserve Bank of Boston, Working Paper
No. QAU08-5, 2008) available at http://www.bos.frb.org/bankinfo/qau/wp/
2008/qau0805.pdf; V.V. Chari et al., Facts and Myths about the Financial Crisis
of 2008 (Fed. Reserve Bank of Minneapolis, Working Paper No. 666, 2008)
available at http://www.minneapolisfed.org/research/WP/WP666.pdf.
no direct mechanism for regulatory authorities to take control of these firms

and wipe-out the equity holders’ claims if the firms’ operations became too
risky, and therefore these firms did not face the direct threat to franchise
value that HLR emphasize is essential to motivate maintaining a capital
cushion.
As discussed earlier, the two mortgage GSEs operated throughout
the 2002-2007 period with essentially the minimum capital required by
their regulator and returned excess capital to equity holders through
dividends and share repurchases. The largest commercial banks (those with
assets in excess of $10 billion) maintained an average total capital ratio of
11.8% from 2001 through 2008 compared to the 12.3% average for all
commercial banks, which is a substantial difference when compared to the
10% standard for well-capitalized banks. Many of the banks in this
category might have viewed themselves as being too big to fail and were
therefore willing to carry a smaller capital cushion.28 HLR points out that
this behavior is entirely rational and predictable for owners of firms in their
position. The debt-favoring factors (tax-benefits and implicit insurance
benefits) significantly outweigh the expected cost of lost franchise value
when early intervention by a regulator is unlikely. Thus, the optimal
leverage for these firms lies above that implied by the capital standard and
they rationally choose to hold no more capital than required. In addition,
these institutions were able to further circumvent regulation by using
derivatives and special purpose subsidiaries of holding companies to
undertake additional financial risk in ways that entailed less risk of early
regulatory takeover and loss of franchise value. Further, while subject to
capital regulations, neither bank holding companies nor investment banks
are subject to statutory threat of receivership for violating capital standards
under Prompt Corrective Action (PCA).29
During a crisis that involves substantial deleveraging that is
destabilizing markets, regulators may have an incentive not to enforce
capital standards. Enforcing capital standards during a period of rapid asset
price decline and thin securities markets will lead to additional asset sales
exacerbating asset price declines. As noted by HLR, it is the threat of
receivership that causes financial institutions to hold a capital buffer, and
those institutions will not hold a sufficient buffer if they know that
28
For data on commercial bank capital ratios, see Federal Deposit Insurance
Corporation, Statistics on Depository Institutions, http://www2.fdic.gov/sdi/
index.asp (last visited Nov. 3, 2009).
29
GAO REPORT, supra note 13, at 28-42.
regulators are unlikely to enforce capital standards during a financial

crisis.30 PCA might be viewed as a reasonable policy to address this
problem in that the government has committed to act when banks violate
capital standards by placing a legislative mandate on bank regulators.31
Finally, the HLR model suggests that an additional factor
contributing to the trend toward higher leverage was diminished franchise
value. Franchise value has many different sources, but a common source is
the restriction of competition. In the last two decades, there has been a
strong trend toward deregulation and the elimination of barriers to
competition. Before 1999, in the U.S. the Glass-Steagall Act limited
competition between commercial banks and investment banks for the
provision of certain financial services. In 1999, the Gramm-Leach-Bliley
Act reduced those barriers. While increased competition may well lead to
the elimination of excess profits and more competitive pricing of services,
it also eliminates a factor that contributes to franchise value.32 In saying
this, we are not attributing blame to the Gramm-Leach-Bliley Act, but
rather simply pointing out that regulators need to monitor all sources of
franchise value and react appropriately. The loss of franchise value due to
increased competition domestically and globally might help explain the
increasing levels of leverage and other risk taking behavior of well
established financial institutions, such as the major investment banks and
the mortgage GSEs, over the past decade as global competition, in the case
30
In game theory, this situation is known as a non-credible threat where an
individual cannot commit to an action in the future (a priori) because all players
know that the action will be irrational in the future (ex post). A standard solution in
such situations is a commitment device that the player imposes on themselves
requiring the action in the future even though irrational ex post. PRAJIT K.
DUTTA, STRATEGIES AND GAMES: THEORY AND PRACTICE (M.I.T. Press
1999).
31
While a minority view, some have argued that the federal government is
not enforcing PCA in violation of federal law during this crisis. See Bill Moyer’s
Journal, William K. Black on the Prompt Corrective Action Law,
http://www.pbs.org/moyers/journal/blog/2009/04/william_k_black_on_the_prompt
.html (April 6, 2009 8:28 EDT).
32
For a discussion of how increased competition can reduce franchise value
and increase risk taking by financial institutions, see Thomas F. Hellmann et al.,
Liberalization, Moral Hazard in Banking, and Prudential Regulation: Are Capital
Requirements Enough?, 90 AM. ECON. REV. 147 (2000).
of investment banks, and an expanding subprime market, in the case of the

GSEs, eroded the franchise values of those firms.33
IV. IMPLICATIONS FOR REGULATORY REFORM
Recent actions to prevent the failure of many large financial

institutions have significantly extended the “Too Big to Fail” umbrella and
suggest an extension of the moral hazard problem well beyond the realm of
depository institutions. We believe an essential element of the new
regulatory framework must include an effective method for dealing with
this extension. Based on our work, we believe a critical element of the new
regulatory framework must be the implementation of substantive capital
regulation for all large highly-levered financial institutions (i.e., all
institutions “Too Big to Fail”). The U.S. Treasury Department’s recent
legislative proposal to create a single regulator responsible for the capital
regulation of systematically important firms provides a fairly simple and
straightforward mechanism for accomplishing this goal.34
Moreover, to achieve the appropriate incentive structure, regulators
must have the clear authority to move quickly to takeover firms that violate
those standards and wipe-out the shareholder equity claims. Only by
putting significant franchise value at risk will the capital standards provide
the incentive for owners and managers to maintain excess capital
appropriate to the unique risks of their firms. As discussed earlier, we
believe that it is not feasible to expect regulators to have as much detailed
knowledge of a financial institution's investment strategies and risk
exposure as the institution’s employees, and so it is important to have a
regulatory system that causes the institutions themselves to maintain a
capital cushion out of self-interest based their own mix of financial assets
and risk exposure.
A credible threat of receivership requires that regulators be
committed to enforcing capital standards for large financial institutions in
33
See Bharat N. Anand et al., Does Competition Kill Relationships? Inside
Investment Banking, December 2001, http://www.webmanager.cl/prontus_
cea/cea_2002/site/asocfile/ASOCFILE120030327160830.pdf.
34
For a general outline of this proposal see Press Release, U.S. Dep’t of the
Treasury, Treasury Outlines Framework for Regulatory Reform (Mar. 26, 2009)
available at http://www.treasury.gov/press/releases/tg72.htm. Also, for a
discussion of non-legislative actions to tighten regulation of financial institutions,
see Binyamin Appelbaum, Obama Administration Pushing for Regulatory Reform
on Many Fronts, WASH. POST, Sept. 23, 2009, at A25.
the midst of serious financial turmoil. If large institutions believe that

regulators will be unwilling to enforce capital regulations during major
economic downturns they will not view the threat of receivership as
credible. Legislative requirements of Prompt Corrective Action or PCA is
one possible commitment tool that might be imposed on “Too Big to Fail”
institutions in order to create the proper incentives for these large financial
institutions.
At the same time, regulators must consider the implications of
enforcing capital standards during economic crises on deleveraging and
asset prices when designing policies for capital regulation. During the
crisis, financial institutions may not be able to raise capital and might be
forced to sell assets further depressing asset values if capital regulation is
strictly enforced. One solution to deleveraging is to actually take over the
failed institution allowing regulators to hold onto assets until the market
has stabilized much like the Resolution Trust Corporation did during the
savings and loan crisis.35 Further, it may be very difficult to value assets in
order to “Mark-to-Market” during these periods as a result of the lemons
problem and/or very few trades in the asset markets. There are no clear
solutions to the problem of marking assets to market during periods of
financial turmoil, but at a minimum we believe that “Mark-to-Market”
should be based on Generally Accepted Accounting Principles (GAAP).36
Further, the new regulatory framework must be comprehensive and
flexible enough to adapt to changing financial instruments. For example,
we must avoid the problem of basing capital standards on balance sheet
assets and liabilities while institutions use off balance sheet derivatives to
take on equivalent risks. Compartmentalizing regulatory authority based
on type of instrument will provide institutions tempting opportunities for
circumventing the intent of the regulation. However, recognizing that no
regulatory scheme can keep ahead of the market in terms of security design
just re-emphasizes the importance of a regulatory capital scheme that
35
The Resolution Trust Corporation was created by the federal government
during the savings and loan crisis in order to hold and dispose of assets of
insolvent thrifts. Lee Davidson, Politics and Policy: The Creation of the
Resolution Trust Corporation, 17(2) FDIC BANKING REV. 17 (2005).
36
During the savings and loan crisis, S&L’s were subject to Regulatory
Accounting Principles rather than GAAP. In May 1987, the Federal Home Loan
Bank Board began phasing in the use of GAAP in response to concerns that
regulatory accounting principles had contributed to the crisis. See Federal
Department Insurance Corporation, The S&L Crisis: A Chrono-Bibliography,
http://www.fdic.gov/bank/Historical/s&l/ (last visited Nov. 3, 2009).
provides incentives for firm managers themselves to assess the risk of new
securities and provide a sufficient capital cushion against future shocks.
Finally, the new framework should provide that when capital
standards are violated, unsecured debt holders are not necessarily fully
protected as they have been in the current crisis. Properly motivated,
unsecured, debt holders can provide significant market discipline to an
institution. Throughout the public debate over “bailing out” various firms,
most attention has been given to the fact that in some cases the owners
have retained a valuable equity stake (e.g., Bear Stearns, Merrill Lynch,
Countrywide and AIG) and managers have received significant bonuses
(famously, AIG).37 However, little attention has been placed on the fact
that in most cases, debt holders were the major beneficiaries of the bailouts.
The government intervention and apparent willingness to fund future
operating losses via further infusions of capital provides assurance of
payment in full to debt holders.38 Regulators should have powers similar to
those of a bankruptcy court to impose modifications to unsecured debt
contracts to avoid such windfalls and in fact to impose a share of the loss
on unsecured debt holders so that debt holders will have an incentive to
monitor and assess the investment and financial decisions of large, highly
levered financial institutions.
Finally, because franchise value plays such an important role in
encouraging prudent managerial choices, the new regulatory framework
must be prepared to react to market developments that affect franchise
value. Global competition, domestic and international tax policies,
technological changes and regulatory changes all have the potential to
erode franchise value. Clearly, concerns about financial institution
franchise value cannot and should not drive policy in these areas.
However, regulators need to be cognizant that financial institution
managers will rationally react to declines in franchise value by taking on
additional risk. Capital standards and regulatory monitoring activities need
to reflect this reality.
37
David Goldman, AIG Bonuses: $235 Million To Go, CNN MONEY.COM,
July 10, 2009, http://money.cnn.com/2009/07/10/news/companies/aig_bonuses/
index.htm.
38
For example, the first business day after the mortgage GSEs were placed in
receivership, the market value of GSE debt and MBS securities jumped
significantly providing a windfall profit to recent purchasers.
PREDATORY LENDING AND ITS INSURANCE
CONSEQUENCES
Erin O’Leary *
***
This note distinguishes predatory from subprime lending, while focusing on

the insurance consequences of predatory lending. It considers how single
premium credit insurance (SPCI) and private mortgage insurance (PMI),
two mortgage-related insurance products, have affected the current
predatory lending crisis. This note argues for reform that eliminates SPCI
and makes PMI a more feasible option for insureds. Such reform would
allow subprime lenders to offer mortgages to qualified borrowers, while
reducing the amount of predatory lending and foreclosures. The
introduction of this note presents some background information regarding
subprime lending and predatory lending. The second part examines
several issues concerning the role of insurance in the subprime mortgage
market. Third, reform measures necessary to alleviate the issues with
mortgage insurance are discussed. Finally, the fourth section studies
recent actions by the Federal Reserve Board and analyzes whether they
can be expected to bring meaningful change. It concludes that, although
the Fed’s new regulations are a step in the right direction, there needs to
be an outright ban of SPCI and predatory must be stopped completely.
***
I. INTRODUCTION
There is no question that a crisis is gripping the subprime market.

As of January 2009, 1.5 million homes had been lost to subprime
foreclosure.1 Another two million subprime mortgage holders are currently
*
University of Connecticut School of Law, Juris Doctor Candidate for the
Class of 2010. The author wishes to thank John McGrath for his assistance and
topical expertise.
1
SONIA GARRISON ET AL., CTR. FOR RESPONSIBLE LENDING, CONTINUED
DECAY AND SHAKY REPAIRS: THE STATE OF SUBPRIME LOANS TODAY 2–3 (2009),
available at http://www.responsiblelending.org/mortgage-lending/research-
analysis/continued_decay_and_shaky_repairs.pdf.
delinquent.2 The fallout from the crumbling subprime market affects not
only subprime borrowers but also their communities.3 The crisis has turned
“subprime” into a dirty word. Many Americans have heard enough on the
news to know that this type of lending is largely to blame for the credit
crisis. However, the subprime market has valid and socially valuable
applications; thus, before one can identify the most troubled areas of the
mortgage industry in order to aim reform measures at these problem spots,
the critical distinction one must draw is between subprime and predatory
lending, which represents only a narrow but dangerous sliver of the
subprime market.
There is a legitimate need for subprime lending. Not all borrowers
can qualify for prime loans, typically because of a negative credit event in
their history or because they lack the cash for a down payment.4 Loan
qualification is typically based on credit scores, as calculated by the Fair
Isaac Credit Organization (FICO). Some examples of negative credit
events that would disqualify a borrower from prime-rate loans include a
history of default, bankruptcy, or low or no credit. Such borrowers are
considered to present too great a risk of delinquency, default, or foreclosure
to justify the prime rate. Nevertheless, there is a strong policy argument in
the United States for enabling as many people as possible to become
homeowners. Therefore, in order for these somewhat risky borrowers to
become homeowners, they may need to resort to other types of loans.
These loans are usually considered subprime. There is no one accepted
2
Id.
3
The Center for Responsible Lending (CRL) estimates that because of what
is known as the “spillover effect,” subprime foreclosures will drain properties
surrounding foreclosed homes of nearly $352 billion in value from 40 million
nearby families, averaging out to almost $9,000 per family. CTR. FOR RESPONSIBLE
LENDING, UPDATED PROJECTIONS OF SUBPRIME FORECLOSURES IN THE UNITED
STATES AND THEIR IMPACT ON HOME VALUES AND COMMUNITIES 1 (2008),
available at http://www.responsiblelending.org/mortgage-lending/research-
analysis /updated-foreclosure-and-spillover-brief-8-18.pdf.. These estimates were
made in August 2008 and update an earlier study made in January 2008 that was
considered “wildly pessimistic” in its estimate of a $202 billion loss in neighboring
home values and an average loss of $5,000 for 40 million nearby families. See
CTR. FOR RESPONSIBLE LENDING, SUBPRIME SPILLOVER: FORECLOSURES COST
NEIGHBORS $202 BILLION; 40.6 MILLION HOMES LOSE $5,000 ON AVERAGE 1
(2008) available at http://www.responsiblelending.org/mortgage-lending/research-
analysis/subprime-spillover.pdf
4
David Anderson, The Subprime Lending Crisis, 71 TEX. B.J. 20 (2008).
2009] PREDATORY LENDING 263
definition of subprime, but it is generally thought to encompass this

category of loans made to borrowers whose backgrounds make them a
riskier investment for lenders.
Because these borrowers present an increased risk to the lender, the
terms of the loans they receive are generally less favorable than those
offered to prime borrowers.5 For example, 70% of subprime loans come
with a prepayment penalty, while only 2% of prime loans do.6 Subprime
loans also tend to have higher fees than their prime counterparts. Most
subprime loans have interest rates that change, usually with little
predictability as to the direction or magnitude of the change. Additionally,
prime loans tend to be similar from lender to lender, facilitating
comparison. Subprime loans are considered by Fannie Mae and Freddie
Mac to be non-conforming loans and tend to vary widely among
originators. A higher annual percentage rate (APR) is the price that
subprime borrowers must pay in order for the lender to take on the risk of
the borrower defaulting, going into delinquency, or being forced to
surrender his home to foreclosure.
Subprime loans have been increasing in popularity at a rapid pace.
In 2003, just 8% of all mortgage originations were subprime loans. By
2006, that figure had jumped to 28%.7 Several factors contributed to this
rise. One important example is the recent housing bubble. Soaring home
5
See Kathleen C. Engel & Patricia A. McCoy, A Tale of Three Markets: The
Law and Economics of Predatory Lending, 80 TEX. L. REV. 1255, 1265–66 (2002).
(explaining the costs that justify higher rates and fees for subprime loans). It is
important to note that many of the costs that justify higher costs and fees, such as a
more thorough review of a customer’s income and credit, lower loan principal
amounts that lead to higher origination costs as a percentage of the total loan, and
higher incidence of prepayment, are much lower in predatory loans, if they exist at
all. For example, predatory lenders often fail to verify a customer’s income, debt
obligations, and assets before making a loan, thereby erasing most of the cost of
income verification. See id.
6
CRL, A SNAPSHOT OF THE SUBPRIME MARKET 1–3 (2007), available at
http://www.responsiblelending.org/mortgage-lending/tools-resources/snapshot-of-
the-subprime-market.pdf.. That number is probably low; Professors Engel and
McCoy estimate that at least 98% of subprime home loans contain “substantial
prepayment penalties.” Engel & McCoy, supra note 5, at 1285.
7
CTR. FOR RESPONSIBLE LENDING, supra note 6, at 2.
values reassured lenders that if the properties for which they were issuing
mortgages would at the very least retain their value.8
Predatory lending, on the other hand, has no legitimate basis. The
vast majority of predatory loans occur in mortgages that would otherwise
be considered subprime. While subprime lenders have a legitimate
argument that they must increase the fees and rates associated with their
products in order to hedge against the increased risk of default presented by
borrowers with less than ideal credit, predatory lending practices have no
place in the world of legitimate loans. They are characterized by practices
that strip homeowners of the equity they have in their properties extract as
much in fees, rates, and other charges as possible from the homeowner.
The subprime lending crisis has had obvious and devastating
ramifications for millions of subprime borrowers. There have also been
less apparent consequences for the insurance industry. This industry has
become closely intertwined with the mortgage industry. As the problems
in the subprime market have become exacerbated, so too have the problems
in mortgage-related insurance products. This paper considers how two
mortgage-related insurance products, namely, single premium credit
insurance (SPCI) and private mortgage insurance (PMI), have affected the
predatory lending crisis. I contend that through reform measures that
eliminate SPCI and make PMI a more feasible option, the subprime lending
market can continue to offer mortgages to qualified subprime borrowers
while simultaneously reducing the tide of predatory lending and
foreclosures.9
In Part II of this Note, I examine two issues concerning the role of
insurance in the subprime mortgage market. The first is the declining
centrality of the role played by providers of private mortgage insurance
(PMI) to insure many subprime mortgages. I will explain the effect that
predatory lending has had on legitimate PMI policy providers. The second
issue is the abuse of credit insurance in subprime mortgage loans. In
particular, I will focus on how single premium credit insurance functions as
one of the more abusive terms that predatory lenders employ. I will also
explore the related nature of these two problems. Moreover, I will provide
a brief synopsis of some of the most common abusive tactics used by
8
See Christopher Barlow, Developments in Banking & Financial Law, 2006-
2007: The Subprime Mortgage Crisis: Macroeconomic Implications of the
Subprime Lending Crisis, 27 REV. BANKING & FIN. L. 57, 57–58 (2008).
9
This paper considers primarily legislation-based reform rather than
litigation.
predatory lenders that have led to the complications that the mortgage
industry is facing today.
In Part III, I discuss the reform measures that will be necessary to
alleviate the problems related to mortgage insurance. I will also focus on
two reform measures: first, the Federal Reserve Board’s newly finalized
amendments to Regulation Z, which implements the Truth in Lending Act
(TILA) and the Home Ownership and Equity Protection Act (HOEPA); and
second, H.R. 3915, or the Mortgage Reform and Anti-Predatory Lending
Act of 2007, which was approved by the House but was never voted on by
the Senate. Additionally, I suggest that PMI is an important product that
can be used to stem the tide of foreclosures because of the mutual interests
shared by borrowers and insurers.
Part IV examines whether the Fed’s new amendments to
Regulation Z can be expected to bring meaningful change and explores
what other measures must be taken in order to stop predatory lending for
good, including the effect that H.R. 3915 would have if re-introduced in the
next session of Congress.
II. MORTGAGE-RELATED INSURANCE PROBLEMS IN THE

SUBPRIME AND PREDATORY LENDING CONTEXT
A. PRIVATE MORTGAGE INSURANCE10
1. Overview
A private mortgage insurance (PMI) policy is one that protects the

lender by paying the costs of foreclosing and guaranteeing a certain portion
of the debt; in most cases, that portion is 20 per cent11. PMI is the only
10
There are multiple types of PMI, but the type that is the focus of this article
is primary PMI, which protects against mortgage default. PMI in this article is
used to refer to primary PMI. Other kinds of PMI not discussed in this article
include pool insurance, which insures groups of individual mortgages and provides
100 per cent coverage for any default losses on mortgages in the pool, subject to a
total loss limit on all mortgages in the pool, and PMI reinsurance. Quinton
Johnstone, Private Mortgage Insurance, 39 WAKE FOREST L. REV. 783, 783
(2004).
11
OFFICE OF THE COMPTROLLER OF THE CURRENCY, Answers About Credit
Insurance, http://www.helpwithmybank.gov/faqs/insurance_credit_life.html.
Policies that pay 20-30% of the amount of the outstanding debt typically allow
lenders to retain title to the property after foreclosure. Mortgage INSURANCE
type of mortgage insurance that lenders can require before making a loan.12
Often, a potential borrower cannot afford a 20 per cent down payment on a
home mortgage.13 On the other hand, most lenders are hesitant to make
such a loan because these borrowers tend to have a higher risk of
defaulting. PMI has served a conciliatory function and permitted these
individuals to become homeowners, while alleviating the fears of lenders.14
Not only is the PMI industry regulated at the state level by laws
that vary from state to state15, but also in part by the federal Homeowner
Protection Act of 1998, which took effect on July 29, 1999.16 If PMI is
required as a condition for a mortgage, this statute protects homeowners by
requiring both automatic cancellation of the PMI and notice of cancellation
rights with respect to PMI.17 For borrowers with good payment history
who are current on their mortgage payments and who can show that the
property has not declined below its original value, the PMI can be
cancelled on a predetermined cancellation date.18 This date typically
occurs when the balance of the mortgage reaches 80 per cent, which occurs
by the borrower making sufficient payments of principal on the mortgage,
through appreciation in the value of the property value, or a combination of
both.19 High risk loans are exceptions, however: lenders must give
COMPANIES OF AMERICA, FACT BOOK (2008-09), available at

http://www.privatemi.com/news/factsheets/2008-2009.pdf.
12
Id.
13
SAN FRANCISCO FEDERAL RESERVE BANK, PRIVATE MORTGAGE
INSURANCE (2008).
14
Id.
15
Johnstone, supra note 11, at 802-03. Although all states impose their own
financial requirements on PMI companies doing business within their borders,
there are several similar requirements: all states have substantial reserve
requirements, including contingency reserves, loss reserves, and unearned
premium reserves. Id. at 813-14. Many have paid-in capital and paid-in surplus
requirements. Id. at 814. Some require that the risk to capital ratio does not
exceed a certain threshold; for example, it must not be greater than 25:1. Id. at
815. Most also restrict the kinds of investments that PMI companies can make and
require that master policies of PMI companies be filed for approval with the state’s
regulatory authority. Id. at 815-16.
16
12 U.S.C.A. §§ 4902-4903.
17
47 C.J.S. Interest & Usury § 515 Homeowners Protection Act (2008).
18
Id.
19
WILLIAM F. GALVIN, SECRETARY OF THE COMMONWEALTH OF
MASSACHUSETTS, QUESTIONS AND ANSWERS ON PRIVATE MORTGAGE INSURANCE
borrowers notice of the automatic cancellation provisions, which state that

while PMI may be cancelled when the borrower has reached 20 per cent
equity, it must be cancelled by the lender when the loan reaches 22 per cent
equity or 78 per cent of the loan outstanding.20 The statute also requires
that, at the time the home is purchased, the lender must give written notice
of when the borrower may cancel PMI.21
The impact of the subprime crisis is becoming more apparent, as
well as the effect on PMI providers. Most prime lenders require PMI when
the loan-to-value ratio is 80 per cent.22 This helps to ensure that if the
borrower is forced into foreclosure, the lender will not lose the difference
between the selling price and the balance on the loan as well as the
foreclosure fees.23 If the borrower does default, the insurance policy pays
out 20 per cent of the loan amount and the bank can recoup the rest through
the foreclosure sale. PMI insurers are left holding the bag if borrowers
default early into their mortgages.
The subprime crisis has caused two central problems for PMI
providers: loss of market share and significant financial losses. First, at the
height of subprime lending in 2005 and 2006, PMI providers lost
significant market share. With the relatively new risk-spreading maneuver
known as the 80-10-10 or "piggyback loan", a buyer could finance enough
of the purchase price such that the first mortgage holder would not require
PMI, discussed infra. 24 Borrowers sought to avoid PMI for a variety of
reasons. Until recently, PMI could not be used as a tax writeoff, but the
AND THE FEDERAL HOMEOWNER PROTECTION ACT OF 1998, available at

http://www.sec.state.ma.us/cis/cispmi/pmiidx.htm.
20
Id.
21
12 U.S.C.A. § 4902-4903.
22
SAN FRANCISCO FEDERAL RESERVE BOARD, supra note 14.
23
NATIONAL CONSUMER LAW CENTER, COMMENTS TO THE FEDERAL
RESERVE BOARD’S PROPOSED REVISIONS TO REGULATION Z TRUTH IN LENDING
REGARDING PROPOSALS TO ADDRESS PREDATORY MORTGAGE LENDING (Mar. 9,
2001).
24
Patrick Rucker, Subprime crisis will boost mortgage insurers, REUTERS,
Aug. 23, 2007, available at http://www.reuters.com/article/gc06/idUSN226325
9520070823 (acknowledging the loss in market share experienced by PMI
companies, but predicting that those mortgage insurers who can “weather” the
current financial crisis will be in a good position to regain market share).
piggyback loan did.25Also, PMI is paid for and maintained by the borrower,
thus adding an obvious cost to the loan.26 For many borrowers, this cost
can be especially difficult to swallow because PMI is not first party
insurance; rather, it is in place to protect the lender. By contrast, the costs
of piggyback lending are more subtle, such as higher interest rates.
Additionally, underwriting standards employed by PMI companies tend to
be much stricter than those used by piggyback lenders.27 As I will explain,
this is due in part becausePMI companies and borrowers share a unity of
interest that simply does not exist between brokers of piggyback loans and
their borrowers.
An 80-10-10 loan is actually two loans: one is a first-lien mortgage
on the property that covers 80 per cent of the purchase price. The second
loan is, in most cases, a ten per cent loan that covers part of the remaining
20 per cent so that the borrower does not have to purchase PMI.28 Thus,
with the increase in predatory loans in the past five years, the mortgage
industry has seen a concurrent increase in 80-10-10 loans. Because the 80-
10-10 loan is a tool invented in part to avoid PMI, it is easy to see how a
sharp increase in its usage would translate into decreased profits for PMI
providers.29 As an illustration of this point, piggyback loans were virtually
nonexistent in 2000, but by 2006 about 22% of owner-occupied houses had
piggyback subordinate lien mortgages, and the number and dollar volume
25
See Robert B. Avery, Kenneth P. Brevoort & Glenn B. Canner, The 2006
HMDA Data, 93 Fed. Reserve Bulletin A73, A84-A85 (2007), available at
http://www.federalreserve.gov/pubs/bulletin/2007/pdf/hmda06final.pdf.
26
Todd J. Zywicki & Joseph D. Adamson, The Law & Economics of
Subprime Lending, 80 U. COLO. L. REV. 1, 42 (2009).
27
Id.
28
While most prime loans require PMI if the borrower puts less than 20 per
cent down, many subprime lenders, although they are willing to lend to customers
who cannot put very much down, will demand either PMI or some other type of
credit insurance if the borrower is not able to put down 25 or even up to 30 per
cent. NATIONAL CONSUMER LAW CENTER, INC., supra note 24.
29
Id. Private mortgage insurers have recognized the threat to their product
that is posed by this 80-10-10 or “piggyback” loan. On the website of the
Mortgage Insurance Companies of America (MICA), an industry organization,
there is an entire section dedicated to warning consumers of the dangers of this
type of loan. See MICA, QUICK FACTS ABOUT 80-10-10 LOANS (2009),
http://www.privatemi.com/news/factsheets/quickfacts.cfm; see also MICA,
MYTHS AND FACTS ABOUT PIGGYBACK LOANS (2009),
http://www.privatemi.com/news/factsheets/myths.cfm.
of piggyback loans had risen dramatically during that period.30 By

contrast, the number of homes purchased backed by PMI declined about
6% from 2005 to 2006 alone.31
The second major problem that predatory lending has created for
PMI companies is an unprecedented financial loss.32 Traditionally, PMI
has been relatively inexpensive,33 for the simple reason that people tend to
have tremendous incentives and motivation to pay off their home mortgage
loans.34 However, as predatory lending creates an increasing number of
dangerous loans whose terms quickly become unaffordable, mortgages are
entering foreclosure at unprecedented rates.
It became clear that many subprime loans were dangerous
investments and carried with them a high risk of borrower default, and
providers shied away from them. Insurers who had issued policies for any
of these predatory loans suddenly found themselves paying claims on the
policies at rapidly increasing rate.
Most PMI companies did not want to be associated with predatory
lending, particularly as it became an increasingly well-known phenomenon.
Moreover, because predatory lending often leaves a borrower with a loan
that he or she cannot afford, from a business standpoint, these loans were
much more likely to result in default and to trigger a payout. Now, several
insurers have publicly refused to insure loans with certain characteristics.35
Many of these are characteristics that often lead commentators to label
loans as “predatory.”36
An examination into the evolution of underwriting standards of
PMI companies reveals their scramble to limit their losses and regain
financial stability. For example, Genworth Mortgage Insurance has issued
statements explaining that, as a response to predatory lending, it will not
insure loans that have excessive fees and costs, prepayment penalties
without a correlating borrower benefit, a history of repetitive financing
30
Zywicki & Adamson, supra note 27.
31
Id.
32
Vikas Bajaj, More Lenders Feeling Pain from Defaults, N.Y. TIMES, July
31, 2007.
33
Bankrate.com, The Basics of Private Mortgage Insurance (June 1, 2001),
http://www.bankrate.com /brm/news/mtg/20010601b.asp (last visited July 19,
2009).
34
Brian M. Heat, Note, Hoosier Inhospitality: Examining Excessive
Foreclosure Rates in Indiana, 39 IND. L. REV. 87, 100-01 (2005).
35
Id.
36
See § 2.a.ii, infra; see also Engel & McCoy, supra note 6, at 1260.
(“flipping”), or SPCI attached to the loan.37 The policy in place at the

Mortgage Guaranty Insurance Corporation (“MGIC”) is very similar to
Genworth’s.38 A series of updates and changes to the underwriting policies
of the Radian Guaranty, Inc. reflect the company’s progression toward
increasingly strict standards. The first major change to Radian’s guidelines
became effective on September 17, 2007 and eliminated some of the more
obviously dangerous practices, such as the Radian 103% LTV program,
which presumably insured loans that were made for up to 103% of the
value of the property.39 Effective February 1, 2008, the maximum LTV
ratio for subprime loans became 95%.40 Effective March 31, 2008, the
minimum credit score for a subprime loan was raised to 660, and interest
only loans as well as cash-out refinances became ineligible for insurance.41
Radian’s increasingly strict guidelines provides an effective illustration of
how the fallout from the subprime crisis became slowly apparent to PMI
companies over a period of time.
Mortgage-related insurance products and predatory lending are
inextricably intertwined in that as long as predatory lenders continue to
make extremely high-cost loans that are likely to cause borrower default,
PMI providers will continue to be reluctant to insure these policies because
of the likelihood that they will have to pay out the policy. PMI providers
have identified several practices that are associated with predatory lenders
and, for reasons both practical and ethical, they often refuse to insure any
loans with some of the following properties.
37
See GENWORTH MORTGAGE INSURANCE, POLICY STATEMENT ON
“PREDATORY” LENDING (2008), http://www.mortgageinsurance.genworth.com/
Legal/PredatoryLendingGuidelines.aspx (discussing practices that will cause the
company to refuse to insure the loan in which the practices are found).
38
MORTGAGE GUARANTY INSURANCE CORPORATION, ANTI-PREDATORY
LENDING POLICY (2008), http://www.mgic.com/emergingmkts/antipredatory
lending.html.
39
eBulletin from Radian Guaranty, Inc. (Aug. 28, 2007),
http://www.radian.biz/pdf/EBULLETIN%202007-02%20GUIDELINE%20
CHANGES%20EFFECTIVE%2009-17-07%20Final.pdf. Because the program is
no longer available, Radian does not thoroughly explain it anywhere.
40
eBulletin from Radian Guaranty, Inc. 2 (Dec. 17, 2007),
http://www.radian.biz/pdf/ Radian%20eBulletin%202007-03.pdf.
41
eBulletin from Radian Guaranty, Inc. 2 (Mar. 7, 2008),
http://www.radian.biz/pdf/ Radian%20eBulletin%202008-02.pdf.
2. Characteristics of Predatory Loans
a. Lack of Verification
No-income, no-asset (NINA) loans have gained notoriety as a

significant cause of the subprime crisis, and rightfully so. Brokers were so
eager to generate loans that they offered mortgages while requiring little to
no documentation of a borrower’s income, job, assets, or other
obligations,42 and they did so at a startling rate: it is estimated that as of
2007, up to half of all subprime mortgages made between 2004 and 2006
had been made without fully documented income.43 Thus, even if they
wanted to, brokers could not ensure that these borrowers would be able to
meet their monthly payments, particularly when the initial low “teaser” rate
reset to a much higher rate.44
These practices are all harmful to borrowers but often have almost
no cost to the brokers and, to a lesser degree, to the lenders. Once the
broker sells the mortgage on behalf of the lender, often earning a sizeable
kickback for doing so, he is no longer responsible for servicing the loan, or
collecting on the payments each month. Therefore, a broker has no interest
in ensuring that a borrower can actually make the loan payments. Rather,
the broker’s main interest is in maximizing his own profit, which is best
accomplished by selling loans at higher rates than the borrower qualifies
for and by including prepayment penalties. These are both actions that loan
originators reward with a kickback.45 Sometimes, brokers are even
rewarded with a commission if the borrower refinances and incurs the
prepayment penalty.
The loan originators themselves often have no more incentive to
write affordable loans than do the brokers. Most lenders bundle these loans
into mortgage-backed securities and sell them on the secondary market.46
Some retain servicing rights, but others sell these rights, leaving them with
virtually no exposure if the borrower defaults. This resale of the mortgage
42
AMERICAN ASSOCIATION OF RETIRED PERSONS, AARP COMMENTS TO
FEDERAL RESERVE BOARD ON THE HOME OWNERSHIP AND EQUITY PROTECTION
ACT (Aug. 15, 2007), available at http://www.federalreserve.gov/SECRS
/2007/August/20070816/OP-1288/OP-1288_51_1.pdf.
43
Center for Responsible Lending, supra note 7.
44
See infra § II.a.ii.2.
45
See infra § II.a.ii.7
46
See Engel & McCoy, supra note 6, at 1273-74.
bundles serves two crucial functions: it gets the mortgages off of the
originator’s books and also provides an inflow of capital so the originator
can continue to make loans.47 Secondary market purchasers of these
securities were unable to inspect the individual loans underlying the
securities they were purchasing. The securities, however, were rated
investment-grade for the most part by the ratings agencies and were
thought to be safe investments because they were diverse and because it
was thought that with their homes at stake, borrowers would generally not
default. Even if one or two failed, investors reasoned, the rest of the
portfolio would be sound. When they failed in large numbers, the so-called
“subprime crisis” ensued.
However, unlike most brokers or lenders, PMI companies have a
strong interest in their insured’s ability to repay his or her mortgage. The
company will have to pay a claim on the policy if the borrower does not
repay; thus, it is in the company’s best interest to fully investigate a
borrower's ability to repay the loan. The requirement that a customer
document his or her ability to repay is now well-reflected in the
underwriting standards of PMI companies.48
b. “Exploding” Adjustable Rate Mortgages
Many predatory loans are advertised at a very low introductory or

“teaser” rate.49 Frequently, advertisements claim that the loan is at a low
fixed rate. What these advertisements do not always disclose, however, is
that once the introductory period, which typically lasts no more than two to
three years, is over, an estimated 90% of these mortgages suddenly switch
to some type of adjustable rate mortgage.50 On a 2/28, which is a 30-year
mortgage whose rate is fixed only for the first two years and then
adjustable for the next 28, a borrower’s monthly payments will increase by,
on average, 30 to 50 percent in the first month of the third year, hence the
47
See id.
48
See, e.g., RADIAN GUARANTY, INC., UNDERWRITING GUIDELINES, 6-7, 10-
11 (2009), available at http://www.radian.biz/pdf/Radian_Standard_Guidelines_
08_17_09.pdf.
49
CENTER FOR RESPONSIBLE LENDING, supra note 7, at 1-2, 4.
50
Id. at 1. These adjustable rate mortgages typically float in relation to an
index like the London Interbank Offered Rate (LIBOR), but subprime borrowers
often pay a premium over the rate that prime borrowers would pay for a similar
product. Anderson, supra note 5, at 20.
term “exploding.”51 Sometimes, a loan is structured as an optional

adjustable rate mortgage, which gives a borrower flexibility in the amount
he wants to pay on his mortgage each month. These can end up being
devastating for the borrower, however: at times, a borrower may be
permitted to pay less than the monthly interest rate each month, thereby
actually adding to his principal. This is a process known as negative
amortization.52
c. Excessive Fees
Predatory loans virtually always come with numerous fees, deftly

packaged into the loan such that a typical borrower would have a difficult
time spotting them.53 Up until now HOEPA has structured its definition of
“high cost mortgages” such that lenders could avoid falling within
HOEPA’s regulations simply by hiding large fees in their mortgages rather
than simply increasing the interest rate.54 One of the most onerous and
frequently used of these fees is the prepayment penalty.
d. Prepayment Penalties
A prepayment penalty is a provision of the mortgage contract that

states that if the borrower pays off the loan entirely or in part, either
through payments against the principal or, as is more common, through
refinancing, the borrower will incur a penalty. Penalties are usually
expressed as a percentage of the mortgage balance at the time of
prepayment or alternatively as a set number of months’ worth of interest,
51
CENTER FOR RESPONSIBLE LENDING, supra note 7, at 2. While the average
increase in rate is only 30 to 50 per cent, a loan resetting from 7 to 12 per cent
would cause the borrower’s payments to increase by 70 per cent. See Id.
52
Engel & McCoy, supra note 6, at 1263. See also Mincey v. World Sav.
Bank, FSB, 614 F. Supp. 2d 610, 635-638 (D. S.C. 2008) (holding that where a
lender violated TILA where he failed to disclose that if a borrower chose to pay off
his mortgage as an option-ARM, it would cause negative amortization).
53
These fees greatly exceed the amounts justified by the costs of the services
provided and the credit and interest rate risks involved. David Reiss, Subprime
Standardization: How Rating Agencies Allow Predatory Lending to Flourish in the
Secondary Mortgage Market, 33 FLA. ST. U. L. REV. 985, 999 (2006).
54
See CENTER FOR RESPONSIBLE LENDING, supra note 7, at 1-3.
and can easily total five percent of the principal balance of the loan.55 The
penalties usually can only be invoked in the first part of a loan, i.e. within
the first four or five years. Brokers defend these particular penalties with
the argument that by applying a prepayment penalties, lenders are willing
to lower the interest rate because they are more confident that the borrower
will not prepay, such as through a cash-out refinance. Thus, contend
brokers, because of the lower interest rate, there is a net benefit to
borrowers. However, research indicates that even where lenders do lower
the interest rate on mortgage with prepayment penalties, the difference is
made up in added fees, evidencing a lack of correlation between the penalty
and the purported savings to the consumer. Instead, the cost of the penalty
to the average borrower is three to four times the average savings in interest
payments.56
Prepayment penalties are particularly troublesome for borrowers in
two situations. First, if a borrower has successfully improved his credit
score such that he now qualifies for a better loan at a prime rate, he invokes
a prepayment penalty. Those borrowers who cannot afford to pay the
penalty- and there are many- feel they have no choice but to stay with their
current mortgages.57 Many subprime lenders market their product as a loan
that a borrower can use while he or she tries to improve his credit rating
such that he can qualify for a prime loan. Given the fact that prepayment
penalties lock a consumer into the loan, this claim seems ironic at best and
downright fraudulent at worst.58 If the borrower chooses to refinance, he
incurs a large penalty for doing so. In the words of the Center for
Responsible Lending, the borrower incurs “punishment for obtaining a
better loan.”59
The second situation in which prepayment penalties become
onerous for borrowers occurs when the mortgage “explodes” at the end of
the introductory period. The borrower abruptly finds himself owing on
55
A typical prepayment penalty would be six month’s interest, or about 4.5
per cent of the initial loan balance. Supra note 45, at 5. This translates into
thousands of additional dollars required for refinancing. Id.
56
Id.
57
DEBBIE GOLDSTEIN & STACY STROHAUER SON, CRL, WHY PREPAYMENT
PENALTIES ARE ABUSIVE IN SUBPRIME HOME LOANS 1, 4 (2003), available at
http://www.responsiblelending.org/mortgage-lending/research-
analysis/PPP_Policy_Paper2.pdf.
58
See id.
59
Id.
average 30 to 50 percent more each month than he did previously.60 Faced

with this financial burden, many borrowers attempt to refinance their loan.
However, in doing so, they incur the prepayment penalty. Many can’t
afford to pay the penalty and so are forced to remain in the mortgage. For
others, the original lender offers to reduce the prepayment penalty in
exchange for an agreement to refinance with the same lender. The
prepayment penalty is then added to the principal, along with other fees
charged by the lender for refinancing, and the borrower’s principal
increases.
This tactic is used almost exclusively in subprime loans: 70% of
subprime loans have prepayment penalties, while less than two per cent of
prime loans have them.61 This is partially fueled by lenders who pay a
premium to brokers who can pad a loan with a high prepayment penalty.62
e. Equity Stripping & Loan Flipping
One of the most prevalent practices among predatory lenders is

equity stripping. Equity stripping can take many forms, but in the context
of mortgage lending the method of choice for accomplishing this is known
as loan flipping.63 When one of these mortgages becomes onerous or
unaffordable for the borrower, the lender is quick to offer the opportunity
to refinance repeatedly and at short intervals.64 The borrower incurs a large
prepayment penalty, described below, as well as other fees in addition to
the penalty. These costs usually get added to the principal and the
borrower begins paying interest on the fees themselves.65 Some lenders
increase the fees every time they refinance. With each refinancing, the
lenders pocket more in fees and the borrower’s equity dwindles.
60
See Ctr. for Responsible Lending, supra note 6, at 2.
61
Id. Prepayment penalties have been limited in the prime market because of
better competition among lenders. Id. But see Engel & McCoy, supra note 6, at
1285 (estimating that at least 98% of subprime loans contain substantial
prepayment penalties).
62
See infra § II.a.6; William F. Bennett, Mortgage Brokers Get Fatter
Payoffs for Selling Riskier Loans, NORTH COUNTY TIMES, May 5, 2007
(considering the case of California mortgage brokers who are legally obligated to
act in the best interest of the borrowers, but who often do not because there is no
enforcement mechanism).
63
See Engel & McCoy, supra note 5, at 1263.
64
Id.
65
Id.
Sometimes the prepayment penalty added to principal borrower pays for

every refinancing is less than the cash recieved during that transaction.
Ultimately, having leveraged it all, homeowners are left with little to no
equity and owe enormous sums because fees have been folded into the
principal amount of their loan.66 Thus, the equity has been stripped.
Most of the wealth of the average American is found in the equity
of his home.67 To unscrupulous lenders, this represents a veritable jackpot,
and they attack it as such with aggressive advertising practices, as
described below.
f. Aggressive Advertising
Some scholars characterize the tactics employed by predatory

brokers and lenders as one calculated to overcome free will.68 Whereas
borrowers of prime loans tend to seek out a lender, the opposite is true with
predatory loans. Predatory lenders tend to use mass mailings, cold calls,
and other aggressive advertising techniques to seek out people who often
are not even in the market for a loan to begin with.69 When seeking out
these borrowers, brokers and lenders focus on the most vulnerable
populations.70 They use public records to identify which individuals or
households may owe back taxes. Many prey on elderly homeowners who
have built up significant amounts of equity in their homes.71 The CRL
estimates that borrowers age 65 and older have five times the odds of
receiving a subprime loan than borrowers younger than 35.72 Others use
census data to identify low to moderate-income neighborhoods, typically
targeting minority groups within those neighborhoods. In 2006, over 50%
of loans made to African-Americans were subprime, 40% of loans made to
people of Hispanic descent were subprime, but only 22% of loans made to
Caucasians were subprime.73 These populations often feel that the prime
66
See Goldstein & Son, supra note 57, at 3.
67
Id at 12..
68
See Engel & McCoy, supra note 6, at 1346
69
See id. at 1296.
70
See id.
71
AARP, supra note 42.
72
CTR. FOR RESPONSIBLE LENDING, FACT SHEET: PREDATORY MORTGAGE
LENDING ROBS HOMEOWNERS & DEVASTATES COMMUNITIES (2006), available at
http://www.responsiblelending.org/mortgage-lending/research-analysis/2b003-
mortgage2005.pdf.
73
Ctr. for Responsible Lending, supra note 6, at 2
market is not available to them because of poor credit history or other

major credit event. Of course, as discussed above, up to 50% of these
borrowers may in fact qualify for a prime loan.
Brokers and lenders rely on high-pressure marketing tactics, and
may go door to door or cold-call.74 Their target borrowers are often not
even seeking a loan, but marketers tend to be charming and friendly,
convincing the target that they are going to help them. Sometimes, would-
be borrowers, especially elderly ones, have little to no debt on their home.
For these groups, brokers focus on convincing them that they can “help”
them by giving them some extra cash. Of course, the customer is then
subject to prepayment penalties, exorbitant fees that may get folded into the
principal itself, and other abusive practices. For these groups, brokers
focus on convincing them that they can “help” them by giving them some
extra cash. Of course, the customer is then subject to prepayment penalties,
exorbitant fees that may get folded into the principal itself, and other
abusive practices. In the worst scenarios, people who once owned their
homes free and clear (or close to it) end up the victims of foreclosure.75
Predatory lenders act quickly once they have identified their
targets,. They restrict the amount of information they give out and what
information borrowers do receive is confusing and hard to read. Closings
are often rushed and accompanied by stacks of paperwork that are
insurmountable to the average borrower. Sometimes, brokers misrepresent
what the borrower must do in order to close. For example, the broker may
tell the borrower that he has to purchase life insurance in conjunction with
his loan when that is not the case at all.76
g. Yield-Spread Premiums
Yield-spread premiums are kickbacks paid by lenders or

originators to brokers if the broker “upsells” a loan. By steering a customer
who qualifies for a lower cost subprime or even a prime rate into a much
costlier loan, originators can increase the rates and fees on their loans.
However, this means that brokers are trying to sell borrowers loans at rates
much higher than the customer would otherwise qualify for since these
loans are more suitable for customers with poorer credit.77 Usually, Fannie
74
Engel & McCoy, supra note 5, at 1295.
75
See AARP, supra note 42.
76
See Engel & McCoy, supra note 5, at 1267.
77
See id. at 1264-66.
Mae and Freddie Mac studies have found that up to about 50% of all
borrowers steered into subprime loans could have qualified for a prime
rate.78 This figure is partially a result of brokers working in concert with
lenders to drive up mortgage rates. In addition, many lenders offer these
kickbacks if a brokers can pad the loans with large prepayment penalties.79
Lenders are willing to pay a premium for a prepayment penalty that
functions as a guarantee that the borrower will either stay in the overpriced
predatory loan or that the lender will make money when the borrower is
forced to pay the penalty in order to get out of his expensive loan.80
These kickbacks are dangerous for another reason. Many
unscrupulous brokers employ tactics designed to convince the potential
borrower that they, the brokers, are working on behalf of the borrower to
obtain the best rate possible. Borrowers are lulled into a false sense of
comfort and a belief that the broker is laboring under some sort of fiduciary
duty. In reality, exactly the opposite is true: brokers are motivated to sell
the borrower the most expensive mortgage possible.81
h. Over-appraisal and Overselling
Lenders and brokers often work hand in hand with appraisers:

when appraisers return favorable, overstated estimates of home values, the
lenders and brokers continue to use their services.82 Because the cost of
their home has been overstated, borrowers end up with loans whose
principal is greater than the true value of their home. 83 This benefits the
lenders, who make more money from interest, fees, and charges if the
underlying principal is higher. Many of these loans were made during the
housing bubble, when people expected home prices to continue to rise.
78
See Goldstein and Son, supra note 57, at 1 n.3.
79
See id. at 1.
80
For example, a ContiMortgage Corporation rate sheet shows that for loans
with prepayment penalties, the maximum yield-spread premium is 2.5 per cent.
For mortgages with prepayment penalties, the maximum premium jumps to 4.25
per cent. AARP, supra note 42, at 5-6.
81
Bennett, supra note 62.
82
See 73 Fed. Reg. 44522, 44566 (proposed Jul. 30, 2008) (to be codified at
12 C.F.R. pt. 226).
83
See Christopher L. Peterson, Predatory Structured Finance, 28 CARDOZO
L. REV. 2185, 2223 (2007) (describing a predatory lender who used overstated
appraisals to justify loans packed with excessive fees and charges).
Instead, in 2007, the bubble burst and home prices fell.84 At the same time,
many rates on subprime loans reset from the introductory “low fixed rates”
to a higher rate. Borrowers who relied on increasing home values were
simultaneously faced with increased mortgage payments, , leading to
increased foreclosures.
B. THE ABUSE OF CREDIT INSURANCE
Among the many tools in a predatory lender’s arsenal is the use of

insurance to extract even more money out of subprime borrowers.85 Credit
insurance is linked to a specific debt or loan and will pay off that debt if the
borrower becomes unable to do so.86 Although such a policy protects the
lender, it is paid for by the borrower. There are various types of insurance
against the borrower’s failure to pay. Among others, some examples
include credit health insurance, which protects the lender in case the
borrower becomes ill and can no longer make payments on the loan; credit
life insurance to protect against the risk of the borrower’s death prior to
repaying the loan; and credit unemployment insurance protects against the
borrower’s inability to make payments due to job loss.87 Whereas it is not
uncommon for prime lenders to require private mortgage insurance for
loans in which the borrower pays less than twenty per cent of the home’s
value as a down payment, these types of credit insurance are rarely found
in prime loans, but have been aggressively marketed in the subprime
market.
Single premium credit insurance (SPCI) is particularly abusive.
With this type of insurance, borrowers are forced to pay a one-time
premium that is very high, tending to cost four to five times as much as
credit insurance whose premiums are paid on a monthly basis.88 This
premium is nearly always financed with the mortgage leaving the borrower
paying a hefty interest rate in addition to an already enlarged insurance
premium.89 For these reasons, the Consumer Federation of America has
84
Anderson, supra note 4, at 20.
85
For a description of other abusive practices found in predatory loans, see
supra Part II.a.
86
ASSOCIATION OF COMMUNITY ORGANIZERS FOR REFORM NOW,
PREDATORY LENDING PRACTICES, ¶ 15 (2004), http://www.acorn.org/
index.php?id=754.
87
Id.
88
Id. at ¶ 16.
89
Id. at ¶ 17.
called SPCI on mortgages “the worst insurance rip-off in the nation,” and
Fannie Mae and Freddie Mac refuse to purchase loans that included
financed credit insurance.90
While onerous for the borrowers who must pay the premiums,
mortgage credit insurance is lucrative for the lenders. Credit insurance
companies typically sell group insurance products to home mortgage and
other lenders.91 One author suggests that some lenders make up to 50% of
their pre-tax income from the sale of credit insurance.92 The Consumer
Federation of America also found in a 1999 report that the credit insurance
industry had unusually low loss ratios, with the ratio of claims to premiums
often no more than 40% for credit life and disability insurance policies.93
This exceeds the National Association of Insurance
Commissioners’(NAIC) recommended 60% claims to premiums ratio.94
The implication from these numbers is that lenders selling credit insurance
are profiting at the disproportionate expense of the borrowers who are often
deceived into purchasing the policies.95
In theory, these products are voluntary. The only form of
insurance that a lender may legally require in order to obtain a loan is
private mortgage insurance, but once a borrower does purchase credit
insurance, it becomes part of his contract.96 In many cases, the relatively
meager consumer benefits provided by credit insurance last only a few
years, while the borrower continues to pay a high interest fee on a large
premium for years after the benefits have stopped.97 In previous
90
Id. See also Broderick Perkins, California Enforcing New Predatory Loan
Rules, REALTY TIMES, July 10, 2002, available at http://realtytimes.com
/rtpages/20020710_predatoryloans.htm
91
Ronald H. Silverman, Toward Curing Predatory Lending, 122 BANKING
L.J. 483, 504 (2002).
92
Id. at 504-505
93
Id. at 505.
94
Id.
95
See id.
96
NATIONAL ASSOCIATION OF INSURANCE COMMISSIONERS, CREDIT
INSURANCE: SAFETY NET OR NO NET GAIN? (Nov. 2006),
http://www.naic.org/documents/consumer_alert_credit_insurance.htm; see also
Office of the Comptroller of the Currency, supra note 11.
97
Edward M. Gramlich, Governor, Federal Reserve Board, Remarks at the
Housing Bureau for Seniors Conference (Jan. 18, 2002).
amendments made to HOEPA, the Federal Reserve Board has attempted to

regulate SPCI without prohibiting it altogether, but it persists.98
At least one appellate-level court has agreed with the proposition
that SPCI is inherently unfair. In Richardson v. Bank of America, the
North Carolina Court of Appeals held that the lender’s illegal sale of SPCI
was an unlawful and deceptive trade practice under North Carolina law,
that it was not made in good faith and did not represent fair dealing, and
that it constituted willful and wanton tortuous activity sufficient to support
the imposition of punitive damages.99
III. REGULATIONS AND REFORM MEASURES
A. OBJECTIVES OF REFORM
1. Competing Interests
Any reform measures aimed at the insurance industry in the

context of predatory lending, whether directly or indirectly, must seek to
further some important goals. Many of these goals in fact consist of
striking a balance between two equally compelling but competing interests.
The first and perhaps the most frequently cited set of competing interests
involves the conflict between providing consumers with adequate
protection against predatory loans and promoting the availability of credit.
In December, 2007, Ben Bernanke, Chairman of the Federal Reserve
Board, referenced the need to balance these interests when introducing a
proposal to amend existing regulations, stating that the proposed rules were
crafted with an eye toward deterring improper lending and advertising
practices without unduly restricting mortgage credit availability.100
A broad scope of regulations, whether at the federal or state level,
is appealing to consumer advocacy groups but carries a greater risk of
encompassing loans and lenders that the drafters of the regulations never
contemplated. Stricter regulations tend to cause legitimate subprime
lenders to restrict credit for fear of falling into regulated categories. At a
time when the country desperately needs credit, any reform measures must
98
See id.
99
643 S.E.2d 410, 424-425 (N.C. Ct. App. 2007).
100
Ben S. Bernake, Chairman, Federal Reserve, Statement of Jul. 14, 2008,
available at http://www.federalreserve.gov/newsevents/press/bcreg/bernankeregz
20080714.htm
be just as careful to preserve lenders’ willingness to lend as to protect

consumers. Of course, on the other hand, most regulatory reform measures
are motivated by a desire to stop the abuses inflicted upon subprime
borrowers by predatory lenders. To that end, any measures that did not
increase protections for consumers would simply be ineffective.
There is also a conflict between forcing predatory brokers and
lenders to internalize the harm they cause and requiring borrowers to accept
greater personal financial responsibility. On the one hand, when lenders
make irresponsible loans with abusive terms that cause a consumer to
default and frequently to lose his home to foreclosure, society often must
bear the costs.101 On the other hand, the hallmark of responsible borrowing
is knowing how much one can afford. According to one author, “the social
and moral question centers around who should determine that a particular
borrower cannot afford to pay the proposed mortgage.”102
Especially in today’s real estate situation, where it is not
uncommon for a borrower to owe more on his home than it is worth, when
a home is lost to foreclosure the borrower may be left with nothing and
suddenly become dependent on social programs such as welfare to provide
for him and his family. Strict regulations, such as stringent requirements
regarding verification of a borrower’s ability to pay or an escrow
requirement103 would be costly to implement but ultimately would be likely
to reduce foreclosures. Since the lender is in the best position to prevent
the harm caused by predatory mortgage loans and their abusive terms, one
101
See, e.g. Manny Fernandez, Helping to Keep Homelessness at Bay as
Foreclosures Hit More Families, N.Y. TIMES, Feb. 4, 2008, at B6 (discussing a
program adopted by New York charitable organizations to help homeowners who
are facing foreclosure avoid eviction by giving them up to $10,000 to pay for the
costs of moving, the first month’s rent at their new apartments, and other related
costs).
102
David Schmudde, Responding to the Subprime Mess: The New Regulatory
Landscape, 14 FORDHAM J. CORP. & FIN. L. 709, 730 (2009).
103
An escrow requirement like the one proposed in the Fed’s finalized
amendments to Regulation Z requires lenders to set up an escrow to cover the costs
of property taxes and insurance. See CENTER FOR RESPONSIBLE LENDING,
PROPOSED RULES REGARDING UNFAIR, DECEPTIVE, ABUSIVE LENDING AND
SERVICING PRACTICES PURSUANT TO HOME OWNERSHIP AND EQUITY PROTECTION
ACT at 52 (Apr. 8, 2008), available at http://www.responsiblelending.org
/mortgage-lending/policy-legislation/regulators/fed-udap-comments-final-
040808.pdf.
side of the policy argument says that any regulation should favor the
consumer in order to force lenders to internalize the harm they cause.104
The other side of that same policy argument is the side favored by
Senator Phil Gramm, among others. Gramm attributed the increase in
foreclosures, particularly in the subprime market, to “predatory
borrowers,” not predatory lenders.105 With the spectrum of available
subprime products, many borrowers were able to obtain mortgages that
they ultimately could not afford.106 People such as Senator Gramm favor
regulations that require the borrower to be responsible for ensuring that he
only borrows what he can afford. The problem with this, however, is that
predatory brokers and lenders often structure their marketing strategies
around pressure and removing free will.
2. Policy Goals
The mortgage insurance industry can be reformed by measures

aimed specifically at SPCI and PMI reform. However, it is equally if not
more important that broader reforms encompassing predatory lending
generally be passed.
The preferred method for avoiding PMI is the use of the 80-10-10
loan, which is often a favorite target of predatory lenders.107 The type of
borrower who seeks out these loans is often one who has very little cash
and may feel excluded or disenfranchised from the prime market. There
may be an element of desperation: often, borrowers want to avoid incurring
a PMI premium so badly that they accept credit from any lender willing to
offer it. Some may feel that because the loan is quite small in comparison
to the larger 80% loan, it is less risky to accept more onerous terms. The
80% loans, particularly when made to subprime borrowers, may also be
predatory. Whatever the reason, legislative efforts toward eliminating
104
See generally Lauren E. Willis, Against Financial-Literacy Education, 94
IOWA L. REV. 197 (2008) at A1 (discussing the interaction of foreclosures and
societal reaction to the same and positing that regulations are the best way to attack
foreclosures, which are costly to the borrower and to society).
105
Eric Lipton & Stephen Labaton, Deregulator Looks Back, Unswayed, N.Y.
TIMES, Nov. 17, 2008 (emphasis added)
106
See Fernandez, supra note 102, at B6 (discussing the example of a woman
who was able to obtain a $486,000 mortgage with $4,000 monthly payments
despite the fact that her monthly income was only $2,800).
107
Johnstone, supra note 10, at 787.
predatory lending are a critical first step toward improving the PMI outlook
and making mortgages safer and more affordable for subprime borrowers.
PMI providers avoid predatory loans for reasons that are both
ethical and practical. From an ethical standpoint, it is likely that these
providers do not want to be associated with a product that has gained so
much notoriety as a major cause of rising foreclosure costs and the seizing
up of the capital markets. And from a practical point of view, insurance
companies are loath to insure loans that carry with them a high likelihood
of a borrower default that would trigger payout on the policy. As discussed
above, predatory loans carry with them a higher rate of default. Therefore,
curbing predatory lending would make subprime loans safer and less
expensive for PMI providers to insure. Although there are a relatively
small number of PMI providers,108 there is intense competition among this
limited group of companies. Therefore, if the costs of PMI can be reduced
across the board, then costs of PMI policies are likely to fall, making it a
much more feasible option for borrowers. This will help drop demand for
the 80-10-10 loans. If faced with decreased demand and more of the right
regulations, along with the increased public awareness, predatory lenders
would find it much more difficult to find willing victims.
Perhaps the most appealing aspect of PMI is that it aligns the
borrower’s interest with the insurer’s interest. Restrictive or requirement-
based regulations, such as prohibitions on prepayment penalties or
mandatory disclosures, are often difficult because the brokers and lenders
targeted by these regulations realize that it is frequently not in their best
interests. Brokers in particular have little incentive to ensure that the loans
they write on behalf of their lenders are sustainable and affordable.109 If
the borrower defaults, the broker is not implicated: he or she has likely
already collected his or her yield-spread premium and is unaffected by the
foreclosure.
One author has identified the moral hazard faced by brokers and
explains that at the same time that the law permits and even encourages a
mortgage broker to bargain for the most advantageous outcome, it also asks
that broker to serve as the conduit for information that is intended to reduce
108
Members of MICA, the industry association for mortgage insurers includes
AIG United Guaranty, Genworth Mortgage Insurance Corporation, Mortgage
Insurance Guaranty Corporation, PMI Mortgage Insurance Co., Radian Guaranty,
and Republic Mortgage Insurance Company. Private MI, ABOUT MICA (2009),
available at http://www.privatemi.com/about.cfm.
109
See Bennett, supra note 62.
his advantage. Setting up such a conflict between the duties we expect a

mortgage broker to discharge and that broker’s self-interest is ill-
conceived.110
Furthermore, in the age of increasing securitization, fewer lenders
have kept the mortgage on their books, although some have kept the
servicing rights.
PMI is exactly the opposite: here, there is a unity of interest
between the borrower and the seller of the product. If the policy is still in
effect, a foreclosure triggers a payout. Indeed, some PMI providers have
even paid off an insured’s monthly mortgage debt if the insured loses his
job or is temporarily unable to pay, or have otherwise negotiated or
mediated between the insured and the lender.111 This is just one example of
a practice that is mutually beneficial to the insurer and the policyholder: the
insured avoids a devastating foreclosure while the insurer does not have to
pay out the policy’s limits.
For this reason, PMI does not need as much heavy federal
regulation that is required by loan originators and brokers; regulatory
efforts should thus remain focused on predatory lending tactics. PMI
companies are already extensively regulated as insurance at the state
levels.112 Furthermore, while the government must implement restrictions
and requirements to force brokers and lenders to conduct business in a
manner that is fair for consumers, the common desire to avoid foreclosure
shared by PMI providers and policyholders ensures that insurers are
already motivated to act in the interests of the consumer. In short, if
reforms to the predatory lending practices are effectuated such that loans
become more affordable and sustainable, private mortgage insurers can
once again insure those risks without the fears of foreclosure that have been
brought on by the subprime crisis.
110
Lloyd T. Wilson, Jr., Effecting Responsibility in the Mortgage Broker-
Borrower Relationship: A Role for Agency Principles in Predatory Lending
Regulation, 73 U. CIN. L. REV. 1471, 1500 (2005).
111
Genworth Financial has stepped in to facilitate mediations and workouts
that will keep their insureds in their homes and avoid foreclosure and claims. See
Shannon Behnken, Insurance Company Helps Homeowner Avoid Foreclosure,
TAMPA TRIBUNE, Nov. 25, 2008.
112
Johnstone, supra note 10, at 783. This is not to say that regulations are
never needed. The Homeowners Protection Act is an example of regulations that
protect consumers by requiring that PMI be cancelable after a certain loan to value
is reached. See also 47 C.J.S. Interest & Usury § 515 supra note 17.
Turning to the another problem with mortgage-related insurance

products, reform measures aimed at predatory lending overall should also
seek to eliminate SPCI rather than just to regulate it. If this is
accomplished, then lenders who require an extra measure of security before
lending to subprime borrowers are much more likely to turn to other
options. One such option is likely to be PMI: lenders could, if a borrower
was paying below 20 per cent down, require PMI. This would mean that
instead of disguising SPCI in complicated loan documents to essentially
dupe borrowers into financing a single large premium, they would need to
find other options if credit protection was truly important to them.
Whereas SPCI tends to be deceptive and is a product often offered by
predatory brokers and lenders, PMI is a highly regulated industry at both
the federal and usually the state levels. Furthermore, by reducing predatory
lending, including the use of SPCI, not only are loans more safe for
borrowers, they are safer, and by extension cheaper, for policy providers.113
B. PROPOSED PROTECTIONS
1. The Federal Reserve’s Measures
a. Proposed Measures.
Creating effective reform measures is, of course, easier said than

done. Competing interests, such as those described above,114 are indicative
of the wide range of viewpoints about reform. However, the Federal
Reserve has made an attempt at effective reform. On December 18, 2007,
the Fed proposed and asked for public comment on changes to Regulation
Z (Truth in Lending), which was to be adopted under HOEPA115. The
board used primarily objective-based triggers, but drew largely from
Federal Trade Commission notions about what is considered “unfair” or
“deceptive.” The key points of the proposed reform included:
113
See Kathleen C. Engel & Patricia A. McCoy, Turning a Blind Eye: Wall
Street Finance of Predatory Lending, 75 FORDHAM L. REV. 2039, 2041 (2007).
114
See supra § III.a.i.
115
Press Release, Board of Governors of the Federal Reserve System,
(December 18, 2007) (available at http://www.federalreserve.gov
/newsevents/press/bcreg/20071218a.htm).
• Prohibiting a lender from engaging in a pattern or practice of

lending without considering borrowers’ ability to repay the loans
from sources other than the home’s value
• Restrictions on prepayment penalties only to loans that met certain

conditions, including a requirement that the penalty expire at least
sixty days before any possible payment increase
• Requiring the lender to establish an escrow account for payment of

property taxes and homeowners’ insurance, with an opt out
provision available only after one year
• Prohibiting abusive servicing practices, including failing to credit

payments when the servicer receives it, failure to provide a payoff
statement within a reasonable period of time, and “pyramiding”
late fees.
• Prohibiting creditors or brokers from coercing or encouraging an

appraiser to misrepresent the value of a home
• Prohibiting misleading and deceptive advertising practices for

closed-end loans. For example, prohibiting use of the term “fixed”
to describe a rate that will explode to an ARM within a few years;
requiring applicable rates and payments to be disclosed in
advertisements.
• Adjustment to the time frame for disclosures to permit borrowers

to use the information to make comparisons to other mortgage
products.116
b. Public Reaction and Comments
During the first part of 2008, various groups, including industry

interest groups, consumer advocacy groups, local governments and others
submitted comments about the proposal. Two of the most significant
groups to offer comment included the consumer advocates and the industry
interest groups.
116
Id.
Consumer interest groups offering comment included the American

Association for Retired Persons (AARP), the National Consumer Law
Center (NCLC), the Center for Responsible Lending (CRL), National
Council of La Raza (NCLR), and several smaller, state-level groups.
Overall, consumer advocates agreed with most of the proposal but felt that
it did not go too far. More specifically, they wanted to see the new
protections extend to all non-traditional loans. In support of this argument,
consumer groups pointed to the narrow scope of the existing HOEPA
regulations, which were only covering about five percent of all subprime
loans.
Several of these groups, with the AARP leading the charge, called
for a return to traditional underwriting standards117 that emphasized the
three “C’s”: capacity, credit, and collateral. Capacity is a measure of
whether the borrower is able to repay; credit is a measure of whether the
borrower is likely to repay; and collateral requires an assessment of what
assets the borrower has in case he does not repay.118 Along these lines,
consumer groups favored requiring more specific standards in assessing
ability to repay, including requiring the creditor or broker to consider W2
forms, payroll receipts, and other concrete measures of income verification.
The consumer groups differed radically from industry insiders on
their treatment of prepayment penalties. Not all of the consumer groups
recommended banning them outright, but most at least favored strict
restrictions so that lenders could not penalize borrowers who wanted to
refinance their mortgages once the rates exploded or if they were able to
obtain a prime loan.
Among these groups, the escrow requirement was also popular.119
Many low- to moderate-income families may need cash more than their
moderate- to high-income counterparts. This characteristic not only makes
them more susceptible to advertising from predatory brokers and lenders,
but it makes them less likely to be able to save for expenses such as
property taxes and insurance. However, without paying these costs the
borrowers are much more likely to fall victim to foreclosure.
Consumer groups also approved of the scope of the new disclosure
requirements, although many noted that they, alone, were not sufficient. A
few saw it as a trap to bait and switch because lenders are not required to
disclose some applicable points and fees until the closing, at which time
117
See AARP, supra note 42, at 1, 4, 11.
118
Id.
119
Id. at 7-8.
many borrowers would be less likely to withdraw their application and

shop for a new lender.120
Additionally, in the proposed requirements, the Fed contemplated
an outright ban on yield-spread premiums, which was applauded by
consumer groups.121
These groups also were strongly opposed to the “pattern or
practice” requirement, which allowed a borrower to pursue legal recourse if
he could prove a pattern or practice of making loans without verifying
income or assets.122 The criticism of this requirement was that it would
essentially preclude borrowers from relief as it was difficult to establish
this pattern. Furthermore, the consumer groups argued that it was
counterintuitive to require would-be plaintiffs to prove a pattern or practice
because by the time a pattern had been established, many borrowers would
have been hurt, whereas if borrowers were allowed to sue based only on
their own injuries from the lender
Consumer groups, and the NCLC in particular, were wary of a
loophole that allowed lenders to continue to make no-documentation loans
without liability as long as the originator’s loan decision would not have
been different if the proper information had been available.123 The group
was concerned that this would serve as an incentive for originators to avoid
proper underwriting techniques.
The Fed also received comments from industry interest groups,
including the Mortgage Bankers Association, the American Securitization
Forum (ASF), the American Community Bankers, and local real estate and
lending groups. These groups, of course, viewed the proposed regulations
differently than did consumer advocacy groups. They were primarily
concerned with increased exposure to “extreme civil liability”124 and an
ensuing reluctance to extend credit to subprime borrowers.125
120
NATIONAL CONSUMER LAW CENTER, INC., THE FRB’S FINAL HOEPA
RULE: A FIRST STEP, BUT REAL REFORMS ARE STILL NEEDED 2 (2008), available at
http://www.consumerlaw.org/issues/predatory_mortgage/content/FRB-HOEPA-
Rule-NCLCquickanalysis.pdf.
121
Center For Responsible Lending, supra note 106, at 4, 22.
122
Id. at 4.
123
National Consumer Law Center, Inc., supra note 120, at 4.
124
CONSUMER BANKERS’ ASSOCIATION, FED ISSUES TOUGH FINAL HOEPA
RULES; DROPS BROKER DISCLOSURE 1 (2008), at 1. http://www.cbanet.org
/files/GRFiles/HOEpfinal.pdf.
125
Letter from the American Securitization Forum & Securities Industry &
Financial Markets Association to the Board of Governors of the Federal Reserve
Like the consumer groups, industry groups generally offered little

dissent regarding the disclosure requirements. This is likely because
disclosure poses a relatively low burden to the lender, and as such there are
very few legitimate arguments that brokers and lenders could make in
opposition to these requirements.
Another concern was that the proposed regulations were too broad
and would encompass prime as well as subprime loans, thereby harming an
already fragile mortgage market by restricting lenders’ willingness to
lend.126
Additionally, industry groups did not favor a prohibition on the
yield-spread premium. Some defended these kickbacks on the basis that
they a way for creditors and originators to compete for the best brokers to
sell their loans.
Interestingly, from a general standpoint the consumer groups were
less critical of the proposal than were the consumer advocate groups.127
c. The Fed’s Final Rule
After accepting comments, on July 30, 2008, the Fed published the
final rule amending Regulation Z implementing HOEPA and TILA.128
There were some notable changes from the proposed rule. First, at the
behest of consumer groups like the CRL, the Fed changed the definition of
“high-cost” in the context of mortgage loans.129 In the proposed rules,
whether a loan was defined as “high-cost” was determined by comparing
that loan’s rate to the yield on Treasury securities of comparable maturity.
The CRL pointed out that a mortgage-based trigger was more appropriate,
to account for economy-wide wide credit events. As a result, under the
final rules the Fed will publish an “average prime offer rate based on a
survey currently published by Freddie Mac. A loan is high-cost if it is a
first loan and is at a rate that is 1.5 percentage points or more above the
prime offer rate, or for subordinate-lien mortgages, if it is 3.5 percentage
points above the prime offer rate.130
System (Apr. 8, 2008), available at http://www.americansecuritization.com/

uploadedfiles/SIFMAASFRegZComments040808.pdf
126
Id, at 2.
127
See id.
128
Truth in Lending (Regulation Z), 73 Fed. Reg. 44522 (Jan. 9, 2008) (to be
codified at 12 C.F.R. 226).
129
Id. at 44531.
130
Id.
There were key protections passed specifically for higher-priced

mortgage loans. The first requires lenders to verify repayment ability
based on the highest scheduled payment in the first seven years,
considering income, assets, and other debt obligations in the process.
Lenders would be required to actually verify these facts as well. Because
groups like Radian, Genworth, and MGIC have all been plagued by low- or
no-documentation loans and have subsequently refused to insure them, the
requirement that lenders verify a consumer’s ability to repay is likely to
qualify more subprime borrowers for loans under the new underwriting
criteria.
Another protection for high-cost loans, representing a compromise
between consumer groups and the industry groups, generally permits
prepayment penalties, but not if the payment could change within the first
four years.131 For other high-cost loans, the prepayment penalty period
cannot last longer than two years.132 This also strikes a balance between
paternalistic measures which would ban them altogether and promoting
free choice by letting borrowers compare the rates on loans with penalties
versus the ones on loans without in that it bans the most onerous penalties
but permits others. It is however, substantially more restrictive than
originally proposed.133
Lastly, for high-cost mortgages the Fed kept the escrow
requirement from the proposed rules largely intact. However,
acknowledging the argument by the mortgage industry about the costs of
implementation, instead of becoming effective in October 2009 like the rest
of the provisions, the escrow requirement will take effect in 2010.134
Provisions were included, such as those recommended by CRL that
the regulationsgovern all mortgage loans, whether high-cost or not.. The
Fed left intact the rules as proposed prohibiting abusive service practices135
and coercion of appraisers to misstate home values.136 The requirement to
provide a good faith estimate of loan costs, including a payment schedule,
within three days of receiving a consumer’s application also remained
131
Id. at 44551.
132
Id.
133
See id.
134
Truth in Lending (Regulation Z), 73 Fed. Reg. 44522, 44523, 44595 (Jan
9, 2008) (to be codified at 12 C.F.R. 226).
135
Id. at 44598.
136
Id. at 44604.
intact.137 Consumers can only be charged fees after receiving early

disclosure. Furthermore, all mortgages are subject to more stringent
standards regarding advertising: more information is required on
advertisements, and lenders cannot advertise a “low fixed rate” if that rate
in fact may suddenly explode to an adjustable rate after two or three
years.138
Notably, the Fed withdrew for further consideration its original
proposal to prohibit brokers from receiving kickbacks in the form of yield-
spread premiums. It also withdrew for further consideration more stringent
disclosure requirements.139However, the Fed left intact a loophole allowing
for the originators of no-documentation loans to escape liability if their
decision would not have changed even with proper documentation.140
2. H.R. 3915, or the Mortgage Reform and Anti-

Predatory Lending Act of 2007
H.R. 3915, or the Mortgage Reform and Anti-Predatory Lending

Act of 2007, would have filled most of the gaps left by the Fed’s proposed
reform measures.141 However, although passed by the House, the Senate
did not pass the Act which was thus cleared from the books at the end of
110th session of Congress. If t reintroduced and passed, the Act would
represent a major step toward mortgage reform, even stronger than that
taken by the Fed.
H.R. 3915 had a goal similar to that of the Fed’s reform measures,
but added some key protections. One of the most important additional
protections was a new duty of care imposed on all mortgage originators,
including brokers and lenders.142 This would encompass the subjective
triggers that were lacking in the Fed’s reform regulations and would help
add much-needed teeth to reform measures. A bill like this, with its stricter
provisions and more adequate remedies, would help ensure that lenders
cannot avoid penalties for abusive practices by structuring the terms so as
to fall just outside a definitional trigger. Furthermore, SPCI would have
been prohibited on any residential mortgage, a move that acknowledged its
137
Id. at 44600-01.
138
See id. at 4457495.
139
Id. at 44563.
140
See id. at 44574.
141
Id. § 129A.
142
Id. at § 206(g).
inherent unfairness.143 The bill would have also prohibited yield-spread

premiums and other forms of steering incentives,144 as well as unfair
servicing practices such as refusing to credit a payment the day it was
received and then applying a late fee.145 Mortgage brokers and lenders
would have been required to be licensed and registered.146
The proposal also required creditors to verify ability to repay by
considering several factors, including credit history, current and expected
income, The scope of covered mortgages was also increased to cover open-
end loans, whereas the Fed’s regulations cover only closed-end loans. The
triggers are based on the yield for Treasury securities of comparable
maturities. There were special, additional protections for high-cost
mortgages,147 including a prohibition on balloon payments, recommending
or encouraging default, excessive late fees, financing any points or fees
abusive modification and deferral, and other abusive practices.148 The
borrower of such loans would also have been required to go through pre-
loan counseling to ensure that he can actually understand and interpret all
of the information that he would receive in required pre-closing
disclosures.149
IV. CONCLUSION: IS IT ENOUGH?
The recent subprime lending crisis finds its origins in predatory

lending. Whereas legitimate subprime lenders help countless Americans
with less-than-perfect credit get homes, predatory lending serves no
function other than to pad the pockets of unscrupulous lenders. As the
subprime crisis has spiraled out of control, two problems in mortgage
insurance have emerged: the abuse of single-premium credit insurance and
the issues plaguing the private mortgage insurance providers as a result of
143
Id. at § 123(b).
144
Id. at §§ 601-604. Steering incentives are defined as originator
compensation that varies, directly or indirectly, or is based on the terms of any loan
that is not a qualified mortgage as defined in § 129B(c)(3). Id.
145
Id. at §§ 101-113.
146
Id. at §§ 301(aa)(1)..
147
Id. at § 303. “Extremely” high-cost mortgages would be defined as those
with very high points and fees, exceeding 5% of the total loan amount, using a
comprehensive definition of points and fees that would include yield-spread
premiums, and prepayment penalties. Id. at § 301(aa)(1).
148
Id. at § 303(t).
149
Id. at § 303(t).
the crisis. Both of these phenomena have proved a boon for predatory
lenders.
Single premium credit insurance gets financed into the principal of
the loan, which translates into earnings in interest for the loan originator as
well as a large lump sum. The borrower’s benefits are not nearly as
appealing; usually, the insurance policy expires long before the premium is
paid off. Any effective reforms must ban the use of this; disclosure alone is
not enough.150 At the very least, an intermediate step would be to prohibit
financing the premium; that is, prohibit lenders from charging interest on
the large premium. This will also require lenders to turn increasingly to
PMI if they want extra security for loans on which borrowers are paying
less than 20 per cent down, thereby shifting borrowers into a the much
safer and better-regulated world of PMI.
As predatory lenders became increasingly aggressive, they set their
sights on borrowers who wanted to avoid purchasing PMI when they had
less than 20 per cent to put down on a home. Demand fell for PMI policies
and as a result, predatory loan solutions for the problem of the borrower
with little cash to put down rose. Addressing predatory lending and
restoring PMI to a legitimate, feasible option for borrowers are symbiotic
propositions. The best chance of accomplishing both of these objectives is
for the federal government to use its authority to effectuate sweeping
reforms.
Had the Senate passed H.R. 3915 and the bill had been signed into
law and enacted, it would have represented a major step against predatory
lending that encompassed nearly all abusive practices and also provided
sufficient penalties and remedies to enforce them. As it is, however, the
Fed’s regulations alone are not enough to effectuate the changes needed in
mortgage-related insurance products. What is needed are not strong
disclosure requirements, because even if disclosed, the abusive practices
tend to be complicated and hidden within an even more complicated
framework of loan terms.151 Whereas disclosure may be sufficient for
150
See generally Patricia A. McCoy, Rethinking Disclosure in a World of
Risk-Based Pricing, 44 HARV. J. ON LEGIS. 123 at 147, 154 (2007).
151
See Elizabeth Renuart & Diane E. Thompson, The Truth, the Whole Truth,
and Nothing But the Truth: Fulfilling the Promise of Truth in Lending, 25 YALE J.
ON REG. 181 (2008) (explaining that evaluating the cost of credit and comparison
shopping is very difficult even for sophisticated shoppers, but as lenders
increasingly “unbundle the costs of their loans from the interest into an array of
fees, outsource their overhead to third parties who add to consumers’ costs, and
unveil amazingly complex loan products that dazzle and confuse borrowers,” even
comparison-shopping in the prime market, disclosure is no longer sufficient

as the primary regulator of the subprime credit marketplace.152 Rather, the
abusive practices must be banned outright. This was well-illustrated in
Cetto v. LaSalle Bank National Association. debt obligations, and assets,
among others.153 Prepayment penalties would be prohibited on subprime
loans, but permitted on other mortgages provided they expired three
months before a loan resets. They would have been forbidden entirely on
any subprime loan.154 Because there is no pattern or practice requirement, a
lender could have become liable based only on one mortgage. Therefore,
there was an allowance for bona fide errors within thirty days of the loan
closing, provided the lender corrected them.
The scope of covered mortgages was also increased to cover open-
end loans, whereas the Fed’s regulations cover only closed-end loans. The
triggers are based on the yield for Treasury securities of comparable
maturities. There were special, additional protections for high-cost
mortgages,155 including a prohibition on balloon payments, recommending
or encouraging default, excessive late fees, financing any points or fees
abusive modification and deferral, and other abusive practices.156 The
borrower of such loans would also have been required to go through pre-
loan counseling to ensure that he can actually understand and interpret all
of the information that he would receive in required pre-closing
disclosures.157
sophisticated shoppers would be hard-pressed to draw any sort of meaningful

comparison).
152
Id.
153
H.R. 3915, 110th Cong. at § 129B(a) (2007).
154
Id. at 206(f).
155
Id. at § 301(aa)(1). “Extremely” high-cost mortgages would be defined as
those with very high points and fees, exceeding 5% of the total loan amount, using
a comprehensive definition of points and fees that would include yield-spread
premiums, and prepayment penalties. Id. at § 301(aa)(1).
156
Id. at § 303.
157
Id. at § 303(t).
THE 2008 MENTAL HEALTH PARITY
AND ADDICTION EQUITY ACT:
AN OVERVIEW OF THE NEW LEGISLATION AND
WHY AN AMENDMENT SHOULD BE PASSED TO
SPECIFICALLY DEFINE MENTAL ILLNESS AND
SUBSTANCE USE DISORDERS
Sara Nadim *
***
This note examines the 2008 Mental Health Parity and Addiction Equity
Act and argues that even though this Act represents a landmark
improvement in mental illness parity coverage, an amendment should be
passed to define what is specifically considered to be a mental illness or
substance use disorder. The first part explores the history of federal
mental parity law along with the efforts made to achieve parity. The
second part discusses the specific provisions of the 2008 Act, specifically
that it does not provide explicit definitions for mental health conditions or
substance use disorders. Third, state definitions of mental illness are
reviewed. Recent developments supporting the biological basis of mental
illness are presented in the fourth part. Finally, the fifth part of the note
evaluates the diminished societal costs that will come with the addition of
parity in mental illness insurance coverage. This note argues that certain
severe biologically based mental illnesses should be listed under the
definition of mental illness that at minimum insurers should be required to
cover. It supports its proposition by providing evidence that group health
plan costs for employers will not increase greatly, and that societal costs,
such as homelessness and loss of productivity in the workplace, will be
greatly reduced when mental illness and substance use disorders are
adequately treated.
***
*
Class of 2010.
I. INTRODUCTION
For decades mental health advocates have fought for insurance

coverage parity for mental illnesses. On October 3, 2008 the Paul
Wellstone and Pete Domenici Mental Health Parity and Addiction Equity
Act of 2008 was signed in to law, bringing to fruition decades of advocacy
work on behalf of the mentally ill. Over fifty-seven million, or one in four
Americans, currently suffer from a mental illness.1 In the words of the
Speaker of the House, Nancy Pelosi:
This long-overdue legislation has brought mental illness and
addiction out of the shadows and to the forefront of our work here in
Congress. By requiring that illness in the brain be treated just like illness
anywhere else in the body for insurance purposes, we are helping to end
discrimination against those who seek treatment for mental illness and
saving lives.2
While this bill provides positive advancement for mental illness
coverage, the lack of a clear definition for what is considered a mental
illness or substance use disorder will result in inequitable coverage for
many individuals. There are many reasons why mental parity proponents
were able to finally achieve the passage of this bill. The reasons have
ranged from: (1): proof that group health plan costs would not become
exorbitant; (2) a greater acceptance of and a reduction in stigma
surrounding mental illness; (3) the recent evidence of a biological basis of
mental illness; and (4) recognition of the enormous societal costs of not
treating the mentally ill.
This note will examine the new bill and argue that while the bill
provides significant advancements in mental illness parity coverage, an
amendment should be passed to define what is specifically considered to be
a mental illness or substance use disorder. Two reasons why this bill was
able to garner the support necessary for passage in both the House and
Senate was that it was proven that the cost of group health plans would not
increase, as well as evidence that many mental illnesses have a biological
1
NAT'L INST. OF MENTAL HEALTH, HEALTH & OUTREACH, STATISTICS
(2009), http://www.nimh.nih.gov/health/statistics/index.shtml.
2
Press Release, Nancy Pelosi, Speaker of the House, House of
Representatives, Pelosi: Passage of Mental Health Parity Bill Gives Hope and Help
to Millions of American Families (Sept. 23, 2008), http://speaker.house.gov
/newsroom/pressreleases?id=0831.
2009] MENTAL HEALTH PARITY AND ADDICTION 299
basis.3 This note suggests that both of these reasons also support the
advancement of an amendment that would provide a specific list of severe
mental illnesses to be covered instead of leaving it up to the states and
group health plans to determine.
While the House of Representatives and a handful of states have
advocated or supported a definition of mental illness that includes all of the
diseases listed in the Diagnostic and Statistical Manual of Mental Disorders
(DSM), published by the American Psychiatric Association, this note does
not advocate for such an approach. Rather, this note suggests that certain
severe biologically based mental illnesses, specifically those based on the
definition advocated by Paul Wellstone and Pete Domenic in 1999, should
be listed under the definition of mental illness that, at a minimum, insurers
should be required to cover. This proposition is supported by evidence that
group health plan costs for employers will not increase greatly and that
societal costs, such as homelessness and loss of productivity in the
workplace, will be greatly reduced when mental illness and substance use
disorders are adequately treated.
Part I will discuss the history of federal mental parity law and the
efforts to achieve parity; Part II will discuss the specific provisions of the
Paul Wellstone and Pete Domenici Mental Health Parity and Addiction
Equity Act of 2008 and the fact that the bill does not provide explicit
definitions for mental health conditions or substance use disorders; Part III
will discuss how the states define mental illness; Part IV will cover recent
developments supporting the biological basis of mental illness; and Part V
will discuss the diminished societal costs that will come with the addition
of parity in mental illness insurance coverage. The argument that coverage
of specific illnesses will be cost prohibitive is not sound. Certain illnesses
have been found to have a biological basis and, at a minimum, a specific
list of illnesses should be included in the parity legislation in order to
ensure that individuals throughout the country receive equal treatment and
coverage.
3 See infra Parts IV, V.

II. THE HISTORY OF FEDERAL MENTAL HEALTH PARITY

LAW
A. EARLY EFFORTS TO ACHIEVE PARITY
In the 1990’s, the majority of employer-sponsored health plans that

did include mental health services placed far greater restrictions on mental
health services than for other medical services.4 In 1998, sixty-two percent
of health plans imposed limits on inpatient treatment for mental health
services and fifty-seven percent imposed limits on outpatient treatment.5
These limits were imposed purely on mental health services and typically
were not placed on other medical services.6
In 1996, Paul Wellstone (D-MN) and Peter Domenici (R-MN)
introduced the Mental Health Parity Act (MHPA) as an amendment to the
Kassebaum-Kennedy bill for healthcare portability.7 The Health Insurance
Portability and Accountability Act of 1996 (HIPAA) established standards
for the privacy and security of health information, as well as standards for
electronic data interchange (EDI) of health information.8 The proposed
MHPA amendment was passed by the Senate; however, it was met with
objections in the House.9 The concerns raised in the House included
whether the enactment of such an amendment would result in an increase in
premiums for private health plans and, if that were the case, whether it
would be necessary to provide for an amendment which would only
4
Dana L. Kaplan, Can Legislation Alone Solve America’s Mental Health
Dilemma? Current State Legislative Schemes Cannot Achieve Mental Health
Parity, 8 QUINNIPIAC HEALTH L.J. 325, 329 (2005). “Ninety-one percent of small
firms and 99% of large firms offer mental health and substance abuse coverage in
their most used medical plans.” Id. at n.26. These medical plans restricted one or
all of the following for mental illness treatments: inpatient day limitations, office
visit limitations, annual and/or lifetime maximums, or higher deductibles and co-
payment rates. Id.
5
Id. at 329.
6
See id.
7
Vallerie Propper & Ginger L. Pomiecko, Parity for Mental Health: History
and Consequences, in PUBLIC HEALTH MANAGEMENT & POLICY (2009),
http://www.case.edu/med/epidbio/mphp439 (last visited Nov. 21, 2009).
8
Univ. of Miami, Miller Sch. of Med. Health Insurance and Portability Act
of 1996 (HIPAA), http://privacy.med.miami.edu/glossary/xd_hipaa.htm (last
visited Jan. 25, 2008).
9
142 CONG. REC. H9473-9564 (1996).
increase the coverage of mental health services under health plans without
increasing such premiums.10 In the end, Senator Kassebaum and Kennedy
decided to remove the amendment in order to pass their bill more swiftly.11
Both of these men had personal experiences with mental illness,
driving them to further advocate for this bill. Senator Domenici’s daughter
suffers from schizophrenia and Senator Wellstone’s brother suffers from
bipolar disease. Wellstone and Domenici once again attempted to gain the
passage of the MHPA when they attached the amendment to the Employee
Retirement Income Security Act of 1974 and to the Public Health Services
Act.12 This time the supporters of the amendment threatened to filibuster if
the amendment was removed, as it had been in 1996.13
B. THE MENTAL HEALTH PARITY ACT OF 1998
The MHPA amendment went into effect on January 1, 1998.14 The

amendment that was implemented reflected a compromise between the
proponents and opponents of the bill, with the understanding and
expectation that Congress would reach a more comprehensive agreement
within six years.15 The MPHA included a sunset provision stating that
“this section shall not apply to benefits for services furnished on or after
September 30, 2001” by group health plans or health insurance coverage
offered in connection with such a plan.16 However, Congress decided to
extend the sunset provision every year since 2001, and a more
comprehensive agreement was not reached until 2008 with the passage of
the Mental Health Parity and Addiction Equity Act.17
10
Id.
11
See id.
12
Mental Health Parity Act of 1996, Pub. L. No. 104-204, § 702, 110 Stat.
2944 (current versions at 42 U.S.C. § 300gg-5 (2006); 29 U.S.C. § 1185a (2006)).
13
See Propper & Pomiecko, supra note 7.
14
29 U.S.C. § 1185a (Supp. II 1996) (current version at 29 U.S.C. § 1185a
(2006)).
15
Glen Cheng, Note, Caring for New Jersey’s Children with Autism: A
Multifaceted Struggle for Parity, 60 RUTGERS L. REV. 997, 1016 (2008).
16
42 U.S.C. § 300gg-5(f) (Supp. II 1996) (repealed 2008).
17
See Cheng, supra note 15, at 1016 & n.135. “Public Law 107-116, § 701(a)
extended the sunset to December 31, 2002. Public Law 107-313, § 2(a) extended
the sunset to December 31, 2003. Public Law 108-197, § 2(a) extended the sunset
to December 31, 2004. Public Law 108-311, § 302(c) extended the sunset to
December 31, 2005.” Id. at n.135.
The original MHPA legislation mandated generous mental health

insurance benefits, although unfortunately the final legislation did not
reflect this mandate.18 This was due to the fact that in order to satisfy the
opposition groups and achieve some type of parity, several important
measure that would have ensured comparable coverage between mental and
health services and other medical services were abandoned.19 Due to the
need for such compromises, the bill was viewed as only a step toward
achieving full mental health parity.20 Senator Domenici, a sponsor of the
bill, stated that the legislation was “a compromise to begin down the path
of parity and nondiscrimination for mentally ill people in this country who
have health insurance.”21
The bill stated that if a group health plan does not include an
aggregate lifetime limit on all medical and surgical benefits, the plan
cannot impose any aggregate lifetime limit on mental health benefits.22
Additionally, if the plan does have an aggregate lifetime limit on medical
and surgical benefits it must apply the limit to mental health benefits and
not distinguish between the two.23 The bill also contained provisions
articulating the same standards for group health plans with regards to
annual limits.24 If a plan does not include an annual limit on all medical
and surgical benefits, the plan may not impose any annual limits on mental
health benefits.25 One positive provision of the bill stated that for insurance
plans that were subject to state laws, the MHPA did not preempt state laws
that require more favorable treatment of mental health benefits.26
However, this bill included many limitations that diminished the scope and
force of the positive mandates.
The bill did not require a group health plan to provide any mental
health benefits. If a group health plan did provide mental health benefits,
the provisions of the bill did not affect the terms and conditions of the
coverage such as limits on outpatient visits, in-patient day limits, days of
18
See Kaplan, supra note 4, at 330.
19
Id.
20
Id. at 342.
21
Id.
22
29 U.S.C. § 1185a(a)(1)(A) (Supp. II 1996) (current version at 29 U.S.C. §
1185a (2006)).
23
Id. § 1185a(a)(1)(B).
24
Id. § 1185a(a)(2)(B).
25
Id. § 1185a(a)(2)(A).
26
Michael J. Carroll, Note, The Mental Health Parity Act of 1996: Let It
Sunset if Real Changes Are Not Made, 52 DRAKE L. REV. 553, 557-58 (2004).
coverage, deductibles, prior authorization requirements, requirements

relating to medical necessity, or, in the case of a managed care plan, a
primary care physician's referral requirement.27 And lastly, if the
application of the bill to the plan resulted in an increase in cost of at least
one percent, the group health plan was exempt.28 This last provision was
included to alleviate fears that the bill would increase costs to a prohibitive
level.
C. SUBSTANCE USE DISORDERS
The bill also did not include coverage for substance abuse
treatment. Many found this to be illogical due to the fact that many people
who suffer from mental illnesses also experience a co-occurring substance
abuse problem.29 The co-occurrence or dual diagnosis is often known as
comorbidity.30 Treatment for substance use disorders is a critical element
in an individual’s treatment for a mental disorder; similarly, treatment for a
mental illness is a critical element in the recovery of a person with a
substance use disorder.31 Often times the two are intertwined and recovery
from either disorder is dependent upon the other.32 Treatment of each
disorder separately has proven to be ineffective and research supports
treatment that addresses both conditions.33 A successful model of such
treatment includes case management, group interventions, and assertive
outreach to bring people into treatment.34
The Journal of the American Medical Association has found that
roughly fifty percent of individuals with severe mental disorders are
affected by substance use disorders; thirty-seven percent of alcohol abusers
and fifty-three percent of drug abusers also have at least one serious mental
27
Id. at 557.
28
Id. at 561.
29
E.g., Charity Felts, Comment, Dealing with a Depressed Workforce: Are
American Employers Doing Enough to Support the Mental Health Challenges
Affecting Today’s Employees? 9 SCHOLAR 119, 129 (2006).
30
Id. at 129-30 & n. 64.
31
Id. at 129-30.
32
Id.
33
See id.
34
NAT’L ALLIANCE ON MENTAL ILLNESS, DUAL DIAGNOSIS AND INTEGRATED
TREATMENT OF MENTAL ILLNESS AND SUBSTANCE ABUSE DISORDERS (2009),
http://www.nami.org/Template.cfm?Section=By_Illness&Template=/TaggedPage/
TaggedPageDisplay.cfm&TPLID=54&ContentID=23049.
illness; and of all people diagnosed as mentally ill, twenty-nine percent

abuse either alcohol or drugs.35 An Epidemiologic Catchment Area Survey
found that individuals with severe mental disorders were at significant risk
for developing a substance use disorder during their lifetime.36 The
research specifically found that forty-seven percent of individuals with
schizophrenia had a substance use disorder, more than four times greater
than the general population, and that sixty-one percent of individuals with
bipolar disorder also had a substance use disorder, which is five times
greater than the general population.37
It is clear that in order to provide effective treatment to the
mentally ill, substance use disorders must also be treated. However,
mandated coverage for substance use disorders was not implemented until
the Paul Wellstone and Pete Domenici Mental Health Parity and Addiction
Equity Act of 2008 was signed in to law.
D. THE ROAD TO THE MENTAL HEALTH PARITY AND ADDICTION

EQUITY ACT OF 2008
Due to the compromises that had to be made in the 1998 parity

legislation, Senators Domenici and Wellstone came back in 2001 with a
new bill in an effort to achieve full mental health parity. This time they
were joined by Patrick Kennedy (R- RI) and Jim Ramstad (R-MN), both of
whom had struggled with addiction issues.38 The bill had begun to pick up
momentum when Wellstone was killed in a plane crash in 2002.39 After
this occurred, his son David began to advocate for the measure.40 The
accumulation of the effort to pass a comprehensive mental parity bill was
achieved this past year when the Mental Health Parity and Addiction
Equity Act was passed under the 2008 economic stimulus package. One
reason the bill was able to achieve passage was that just before the start of
the current Congressional session, several mental health groups won the
support of employers and health insurers by alleviating concerns over the
35
Id.
36
Id.
37
Id.
38
Julie Rovner, Mental Health Parity Approved with Bailout Bill, NAT’L PUB.
RADIO (2008), http://www.npr.org/templates/story/story.php?storyId=95435676.
39
Id.
40
Id.
cost of mental parity.41 A 2006 study in the New England Journal of

Medicine found that insurers’ costs rose less than half a percentage point
when full parity was required for federal workers starting in 2001. 42 The
Congressional Budget Office Cost Estimate also stated that if the more
generous House bill were enacted, the costs for premiums would increase
for group health insurance by an average of only about 0.4 percent.43
Two bills were originally formulated in the House and Senate.
Through a series of compromises, the two sides were able to arrive at an
agreement. “A breakthrough occurred when sponsors of the House bill
agreed to drop a provision that required insurers to cover treatment for any
condition listed in the Diagnostic and Statistical Manual of Mental
Disorders (DSM), published by the American Psychiatric Association.”44
The Senate also made concessions in the negotiation agreements and
agreed to adopt some of the language in the House bill that required parity
for out-of-network coverage.45
Other reasons that such a comprehensive mental parity bill gained
widespread support included the fact that new scientific research had
revealed a biological basis and effective medical treatments for numerous
mental illnesses, as well as evidence suggesting that providing mental
health parity coverage will not be cost prohibitive. Additionally,
employers have realized that productivity tends to decrease when workers
are not treated for mental illnesses and substance use disorders, and that if
they do receive treatment it can reduce the number of lost work days.46
Furthermore, some argue that the stigma of mental illness may have faded
as society has seen many members of the armed forces returning from the
41
Id.
42
Nancy Shute, Paying a High Price for Mental Health, U.S. NEWS, Oct. 25,
2007, available at http://health.usnews.com/articles/health/health-plans/2007/10
/25/paying-a-high-price-for-mental-health.html.
43
CONGRESSIONAL BUDGET OFFICE COST ESTIMATE (2007), available at
http://www.cbo.gov/ftpdocs/88xx/doc8837/hr1424ec.pdf.
44
Robert Pear, Bailout Provides More Mental Health Coverage, N.Y. TIMES,
Oct. 5, 2008, available at http://www.nytimes.com/2008/10/06/washington/
06mental.html.
45
Frederic J. Frommer, Wellstone Son Launches Push on Mental Health Bill,
BOSTON GLOBE, July 9, 2008, available at http://www.boston.com/news/
local/rhode_island/articles/2008/07/09/wellstone_son_launches_push_on_mental_
health_bill/.
46
See Pear, supra note 44.
Middle East with serious mental health issues.47 And lastly, the
experimentation with parity at both the state level and in the health
insurance program for federal employees, including members of Congress,
has proved workable.48
III. THE PAUL WELLSTONE AND PETE DOMENICI MENTAL

HEALTH PARITY AND ADDICTION EQUITY ACT OF 2008
The Mental Health Parity and Addiction Equity Act (MHPAEA)

was signed into law on October 3, 2008. The bill is an amendment to
Section 712 of the Employee Retirement Income Security Act of 1974 ( 29
U.S.C. 1185a).49 This bill will provide over one-third of Americans with
improved mental health insurance coverage.50 It will also eliminate a
practice that has been in place for decades, where insurers have placed
much higher co-payments and deductibles on treatment for mental illness.
The new bill also eradicates the practice of restricting the number of
inpatient hospital treatment and outpatient visits for mental health
treatment.51 Federal officials have stated that the new law will improve
coverage for 113 million people.52 This figure includes 82 million people
who are in employer-sponsored plans that are not subject to state
regulations.53 The effective date, for most health plans, will be January 1,
2010.54
A. SPECIFIC PROVISIONS OF THE MHPAEA
The new bill provides parity between medical and surgical benefits
and mental health or substance use disorders for all “deductibles,
copayments, coinsurance, and out-of-pocket expenses,” specifying that
there are to be no separate cost sharing requirements that apply only to
mental health or substance abuse disorder benefits.55 The bill also states
that there shall be no difference between treatment limitations for mental
47
Id.
48
Id.
49
H.R. 1424, 110th Cong. (2008).
50
Pear, supra note 44.
51
See infra Part II.A.
52
Pear, supra note 44.
53
Id.
54
Id.
55
42 U.S.C.A. § 300gg-5(a)(3) (West Supp. 2009).
health or substance use disorder benefits and medical and surgical

benefits.56 Treatment limitations include “limits on the frequency of
treatment, number of visits, days of coverage, or other similar limits on the
scope or duration of treatment.”57 The bill also states that if a plan or
coverage provides medical or surgical benefits by out-of-network
providers, the plan must afford the same for mental health or substance use
disorders.58
The bill also has a few exceptions, including a cost exemption.
The bill states that an employer is exempt from the parity requirements if
the overall implementation of the bill would result in an increased cost of
two percent or more during the first year after the legislation goes into
effect and one percent in the following years.59 Furthermore, employers
with fifty employees or less are exempt from the parity requirements.60
This affects the almost 113 million American employees who work at
companies with fewer than fifty employees.61
The new bill also provides specific guidelines which articulate how
an employer will qualify for the cost exemption. To qualify for the
exemption, the plan must implement the new requirements for at least six
months.62 If after six months the employer can show that the
implementation of the bill results in a cost increase of one percent, the plan
must give notification to the participants and beneficiaries of its decision to
claim the exemption to the health benefit plan.63 The exemption will not go
into effect until thirty days after the notice requirements are fulfilled.64 If
the employer fails to follow these requirements they will be subject to a tax
penalty and fined $100 per day per individual who is affected by such a
failure.65
56
Id. § 300gg-5(a)(3)(A)(ii).
57
Id. § 300gg-5(a)(3)(B)(iii).
58
Id. § 300gg-5(a)(5).
59
Id. § 300gg-5(c)(2)(A)-(B).
60
H.R. 1424, 110th Cong. (2008).
61
LONG-AWAITED BREAKTHROUGH ON FEDERAL MENTAL HEALTH PARITY
(2007), http://www.mhafc.org/pdf/FedParity2007.pdf.
62
Peter M. Panken et al., Employment and Labor Relations Law for the
Corporate Counsel and the General Practitioner, Litigating Claims of Employee
Benefits, SN020 A.L.I-A.B.A. 169, 179 (2008).
63
Id.
64
Id.
65
Id.
It is not clear at this time how many companies will qualify for this
exemption. The states that have experimented with parity have found that
costs often did not rise to the one percent exemption level and frequently
costs stayed the same or decreased.66 For example, when Texas
implemented parity for severe mental illnesses and substance use disorders,
a study found that there was a decrease of fifty percent in per-member, per-
person cost.67 Managed care was also introduced at the same time.68
Similar results were found in North Carolina, and a study on the impact of
mental health parity in California revealed that costs did not increase after
one year.69
The bill further states that insurers must publish the criteria for
medical necessity determinations.70 The insurer must also provide an
explanation for any denial of a claim made for mental health services.71
However, the bill does not provide an explicit definition for mental health
conditions or substance use disorders. The bill states that “the term ‘mental
health conditions’ and ‘substance use disorders’ are defined under the
terms of the group plans and in accordance with applicable Federal and
State law.”72
The House of Representatives had urged the inclusion of a
provision that would have required insurers to provide coverage for any
condition listed in the DSM; however this was met with strong
opposition.73 Many opposed the use of the DSM due to the fact that it
contains conditions such as caffeine intoxication, sleep disorders and jetlag.
In the end, the House dropped its requirement that all DMS-IV disorders be
covered equitably.
Without a clear definition of what the federal government
considers to be a mental health condition or substance use disorder, such
determinations will vary widely due to the fact that such definitions differ
greatly from state to state. This will result in a great variation in individual
coverage depending on where people reside. Additionally, with no clear
66
See Kate Mulligan, More Data Confirm Affordability of Parity, 37
PSYCHIATRIC NEWS 18 (2002), http://pn.psychiatryonline.org/cgi/content/full
/37/12/18.
67
Id.
68
Id.
69
Id.
70
71
Id.
72
42 U.S.C.A. § 300gg-5(e)(4) (West Supp. 2009).
73
definitions of what is considered to be a mental illness or substance use

disorder, the issue of whether something is medically necessary will
continue to come into debate. Furthermore, the lack of categorization or
definitions for mental health conditions and substance use disorders
potentially leaves the door open for insurance providers to exclude illnesses
more readily than if there were more stringent guidelines.
IV. HOW STATES DEFINE MENTAL ILLNESS
In the article An Analysis of the Definitions of Mental Illness Used

in State Parity Laws, Marcia C. Peck, M.D., M.P.H., and Richard M.
Scheffler,Ph.D. analyzed the mental health parity laws in thirty-four
states.74 The study analyzed the different definitions used by the states in
defining mental illness and the effect these variations have on the coverage
that individuals with a mental illness receive.75 They found that three
statutory terms were used to define mental illness in state parity legislation:
(1) “broad-based mental illness,” (2) “serious mental illness,” or (3)
“biologically based mental illness”.76 They further found that,
States rarely, if ever, considered disease prevalence, needs-based
studies, and clinical judgment. In our opinion the definitions that states use
result from a political and economic process involving mental health
advocates and providers, pro- and antiparity legislators, insurers, and
employers.77
Generally, the majority of states have laws specifically
enumerating what is a mental illness. These states base their mental illness
definitions on “biologically based” mental illnesses and “serious” mental
illnesses.78 These definitions are most frequently included in parity
legislation that outline the specific mental illnesses which must be covered
by insurers on an equal basis with physical illnesses. The states that use the
term “biologically based mental illness” as the statutory term used to define
mental illness in their parity legislation are: Alabama, Colorado, Iowa,
74
Marcia C. Peck, M.D., M.P.H. & Richard M. Scheffler, Ph.D., An Analysis
of the Definitions of Mental Illness Used in State Parity Laws, 53 PSYCHIATRIC
SERVICES 1089, 1089 (2002), http://psychservices.psychiatryonline.org/
cgi/content/full/53/9/1089.
75
Id.
76
Id.
77
Id. at 1091.
78
Id. at 1089.
Massachusetts, Missouri, New Hampshire, New Jersey, South Carolina,

South Dakota and Virginia.79 The theoretical basis for this type of
definition is biological psychiatry.80 The mental illnesses that these states
list have a scientifically demonstrable effect on the brain.81 The argument
for covering these illnesses is that the brain is being damaged by the mental
illness and therefore should be treated and covered like any other damaged
or injured organ.82
The states that use the term “serious mental illness” as the statutory
term used to define mental illness in their parity legislation are: California,
Kansas, West Virginia, Louisiana, Illinois, Pennsylvania, Delaware, Texas,
Montana, Maine, Nebraska, Nevada, Oklahoma, and Hawaii.83 The
theoretical basis for this type of definition is public policy, not clinical.84
The aim of this definition is to identity severe and persistent mental
illnesses based on “functional disability and duration of [the] illness.”85
While these states all define mental illness on the same theoretical basis,
either as “biologically based mental illness” or “serious mental illness,” it
is important to note that there is still a great variety among what mental
illnesses are covered.
The states that define mental illness as “biologically based mental
illness” all provide coverage for schizophrenia, major depressive disorder,
obsessive-compulsive disorder, and bipolar disorder.86 However, out of the
ten states that define mental illness in that manner, only two (Virginia and
New Jersey) define autism as a mental illness.87 Only Massachusetts,
Missouri, Virginia, and South Carolina define childhood depression as a
79
JERRY CONNOLLY & JILL STRASSER, PH.D., DEFINING “MENTAL OR
NERVOUS CONDITIONS,” 1st Sess., at 5 (Feb. 2006), http://www.cbs.state.or.us/ins/
public_meetings/sb_minutes/attachments/SB1_defining-presentation.pdf.
80
Id.
81
Id.
82
See id.
83
See Peck & Scheffler, supra note 74, at 1092 tbl. 2.
84
CONNOLY, supra note 29, at 5.
85
Id.
86
See Peck & Scheffler, supra note 74, at 1093 tbl. 3.
87
Id.; see also, IOWA CODE ANN. § 514c.22 (West 2007); N.J. STAT. ANN. §
17B:26-2.1s (West 2006); VA. CODE ANN. § 38.2-3412.1:01 (2007). On July 1,
2009, Massachusetts began to include autism in its definition of mental illness.
MASS. GEN. LAWS ANN. ch. 176B § 4A (2007).
mental illness88 and only Virginia defines attention deficient/hyperactive

disorder (ADHD) as a mental illness.89 Massachusetts, Missouri and New
Hampshire are the only states in the group that consider post-traumatic
stress disorder (PTSD) to be a mental illness.90
Similar to the variation of definitions that exist between states that
define mental illness as “biologically based,” there is great variation in the
illnesses covered in states which define it as “serious mental illness.” For
example, California’s definition includes five disorders: schizophrenia,
schizo-affective disorder, bipolar disorders and delusional depressions, and
pervasive developmental disorder.91 Maine’s definition, on the other hand,
includes fourteen disorders: psychotic disorders, including schizophrenia,
dissociative disorders, mood disorders, anxiety disorders, personality
disorders, paraphilias, attention deficit and disruptive behavior disorders,
pervasive developmental disorders, tic disorders, eating disorders,
including bulimia and anorexia, and substance abuse-related disorders.92
These examples illustrate that simply because states have drafted their
statutory definitions based on the same theoretical basis, it does not mean
that the same illnesses will be covered.
As Doctors. Peck and Scheffler noted, in addition to “biologically
based” definitions, some states use “broad-based” definitions of mental
illness in their statutes.93 “Broad-based” definitions are based upon the
88
MASS. GEN. LAWS ANN. ch. 176B, § 4A(a) (West 2009); MO. ANN. STAT. §
376.826(4)(b),(d) (West 2002); S.C. CODE ANN. § 38-71-290 (1976); VA. CODE
ANN. § 38.2-3412.1:01(E) (LexisNexis 2007).
89
VA. CODE ANN. § 38.2-3412.1:01(E) (LexisNexis 2007). See MASS. GEN.
LAWS ANN. ch. 176B, § 4A(c) (West 2009) which states that "[A]ny such
subscription certificate shall also provide benefits on a non-discriminatory basis for
children and adolescents under the age of 19 for the diagnosis and treatment of
non-biologically-based mental, behavioral or emotional disorders, as described in
the most recent edition of the DSM, which substantially interfere with or
substantially limit the functioning and social interactions of such a child or
adolescent." This would include attention deficient/hyperactive disorder; however,
it is not defined as a "biologically based" mental illness by Massachusetts.
90
MASS. GEN. LAWS ANN. ch. 176B, § 4A(a) (West 2009); MO. ANN. STAT. §
376.826(4)(c) (West 2002); N.H. Rev. Stat. Ann. § 417-E:1(III)(i) (LexisNexis
2009).
91
CAL. INS. CODE § 10123.15 (West 2005).
92
ME. REV. STAT. ANN. tit. 24-A § 2843(5-C)(A), (5-D)(A) (2000).
93
See Peck & Scheffler, supra note 74, at 1090, 1091 tbl.1.
DSM.94 The use of a “broad-based” definition of mental illness was the

position advocated by the House of Representatives during the debates over
the current parity bill. An example of states that use “broad-based”
definitions include: Connecticut, Kentucky, Utah, Rhode Island, and
Washington.95 These states undoubtedly provide greater coverage than do
states that define mental illness according to a short list of “biologically
based” or “serious” mental illnesses.
There are two major mental health advocacy groups in the United
States, the National Association for the Mentally Ill (NAMI) and the
National Mental Health Associations (NMHA). Each of these groups
define mental illness differently. NAMI promotes ending discrimination
and demands fair legislative policies for “priority populations with serious
mental illness.” Priority populations include those with schizophrenia,
schizoaffective disorder, major depressive disorder, obsessive-compulsive
disorder, panic disorder and other severe anxiety disorders, and ADHD.96
The NMHA defines mental illness broadly, addressing a person's ability to
function rather than his or her diagnosis.97
These different definitions can lead to different conclusions about
what is a mental illness and therefore what is “medically necessary.” The
MHPAEA states that the definition of mental disorders and substance use
disorders should be in accordance with state and federal laws.98 From the
survey of many state statutes, it is clear that there is not a consensus on the
state level of what is a mental illness.99 It should be noted that this
94
Id. at 1090.
95
JERRY CONNOLLY & JILL STRASSER, DEFINING “MENTAL OR
NERVOUS CONDITIONS,” Address before the Senate Advisory Committee,
(Feb. 27, 2006).
96
NAT’L ALLIANCE ON MENTAL ILLNESS, PRIORITY AND SPECIAL POPULATION
(2009), http://www.nami.org/Content/NavigationMenu/Inform_Yourself/About_
Public_Policy/NAMI_Policy_Platform/2_Priority_and_Special_Populations.htm
(last visited January 26, 2009).
97
Peck & Scheffler, supra note 74, at 1091.
98
42 U.S.C.A. § 300gg-5(e)(4) (West. Supp. 2009).
99
In addition to a lack of consensus on the state level for mental illness,
Medicaid coverage that individuals receive also varies greatly from state to state.
The states have wide latitude within the confines of the federal guidelines and
therefore the number of people covered and the amounts that each states spends on
services varies across the states. Similar to this note, this has led to a discussion of
solutions that would result in more equitable coverage across the states. See John
Holahan & David Liska, Variations in Medicaid Spending Among States Series A,
variation does not appear with regard to what the states view to be a
physical illness. In fact the states do not even deem it necessary to
statutorily define physical illness. Therefore, with such a wide variation on
the definition of mental illness in state laws, insurers will likely try to adopt
the most stringent and narrow definitions of mental illness. As evidenced
by the varying state laws, depending on the state in which one resides, the
illnesses covered vary greatly.
V. BIOLOGICAL BASIS OF MENTAL ILLNESSES
The way society views mental illness has greatly evolved over the
last two centuries.100 The most recent report from the Surgeon General
found that the stigma attached to mental illness dates back to the 19th
century separation between mental health treatment and mainstream
treatment in the United States.101 National surveys have tracked the
public’s perception of mental illness since the 1950’s.102 The Surgeon
General’s Report states that “[i]n the 1950s, the public viewed mental
illness as a stigmatized condition and displayed an unscientific
understanding of mental illness.”103 In contrast, a 1996 survey found that
people had a greater scientific understanding of mental illness; however,
there was still a large amount of social stigma and people were more
inclined to consider an individual with schizophrenia to have a mental
illness in comparison to an individual with depression.104
Mental illness is not diagnosed in the same way that physical
illnesses may be diagnosed. One cannot test for mental illness by doing a
No. A-3 URB. INST. 1, 1-3 (1997), http://www.urban.org/publications/307035.html;

Nicole Huberfeld, Bizarre Love Triangle: The Spending Clause, Section 1983, and
Medicaid Entitlements, 42 U.C. DAVIS L. REV. 413, 469-70 (2008) stating
“Though Medicaid was created to provide a statutory entitlement to states,
providers, and enrollees, it has failed to ensure that enrollees receive promised
benefits, both by lack of agency action and lack of statutory enforcement
provisions.” Id. at 469.
100
See MENTAL HEALTH: A REPORT OF THE SURGEON GENERAL 6-7 (1999),
http://www.surgeongeneral.gov/library/mentalhealth/chapter1/sec1.html#overarchi
ng.
101
Id. at 6.
102
Id. at 7.
103
Id.
104
Id.
blood test, x-rays, or throat swab.105 Mental health professionals often

meet with their patients to discuss the patients’ symptoms and to have them
describe how they have been feeling and how these feelings have affected
their daily lives.106 They may ask the patient about the length of time these
symptoms have persisted as well as their severity.107 After conducting this
type of consultation, the mental health professionals will consult the DSM-
IV.108
In addition to this traditional method of diagnosing mental illness,
there have been many advances in science that have permitted mental
health professionals to analyze and diagnose their patients. For years
scientists have been trying to investigate the biological and chemical
processes of the brain. There have been many technological advances
since the 1970’s that now permit scientists and researchers to more closely
examine and study the living brain.109 One such technological advance
includes magnetic resonance imagining, commonly known as MRI. This
technology has allowed researchers to compare normal brain functions with
those of individuals suffering from mental illness.110
Today it is clear that many mental disorders have a biological
basis.111 Mental illness is associated with changes in the brain’s structure,
chemistry and function.112 Scientists already have the knowledge of how
the brain typically functions and this new and ongoing research, which has
revealed how the biological processes change when a person has a mental
illness, has caused “scientists to minimize the distinctions between mental
illnesses and these other brain disorders.”113
105
NAT’L INST. OF HEALTH, THE SCIENCE OF MENTAL ILLNESS 23 (2005),
http://science-education.nih.gov/supplements/nih5/Mental/guide/info-mental-a.htm
(last visited March 5, 2009).
106
Id.
107
Id.
108
Id. at 23-24.
109
Okianer Christian Dark, Tort Liability and the “Unquiet Mind”: A
Proposal to Incorporate Mental Disabilities into the Standard of Care, 30 T.
MARSHALL L. REV. 169, 202 n.176 (2004).
110
Id.
111
See MENTAL HEALTH: A REPORT OF THE SURGEON GENERAL, supra note
100, at 5, 15.
112
See NAT’L INST. OF HEALTH, supra note 105, at 21.
113
Id.
The brain’s basic functional unit is the neuron.114 The neuron

possesses dendrites which receive signals and an axon that transmits the
signals to other neurons.115 The area where an axon terminal ends near a
receiving dendrite is called the synapse.116 In order for a neuron to relay
information it uses both electrical signals and chemical messages called
neurotransmission.117 People with mental illness have been found to have
brain scans that do not reflect a normal functioning brain.
In the United States over the last five years, research studies
examining the link between physical brain abnormalities and disorders like
severe depression and schizophrenia have begun to make a strong case that
the disorders are not scary tales of minds gone mad but manifestations of
actual, and often fatal, problems in brain circuitry.118
In 1990 Congress and the President declared the 1990s to be the
“Decade of the Brain,”119 and many studies over the last twenty years have
made the biological connection between mental illness and the brain which
has helped to reshape the way people look at mental illness.120 This new
research has refuted “the nineteenth century distinction between the organic
mental illnesses (dementias and toxic psychoses) and the functional mental
illnesses (including the neuroses and various affective or depressive
disorders and the schizophrenic syndromes.)”121 Physical brain
abnormalities can be viewed as biological instead of “mental” illness.”122
In addition to being able to view the brain and its neural and electrical
responses, scientists have found that certain mental illnesses also have a
genetic basis. For example, Dr. Steven E. Hyman, a former director of the
National Institute of Mental Health, has stated that “[g]enetic mutations and
114
Id. at 24.
115
Id. at 25.
116
Id.
117
Id. at 25-26.
118
Sarah Kershaw, Insure Me, Please, The Murky Politics of Mind-Body,
N.Y. TIMES, Mar. 30, 2008, at WK, available at
http://www.nytimes.com/2008/03/30 /weekinreview/30kers.html.
119
Proclamation No. 6158, 55 Fed. Reg. 29,553-54 (July 20, 1990).
120
See, e.g., Richard E. Gardner, III, Comment, Mind Over Matter?: The
Historical Search for Meaningful Parity Between Mental and Physical Health
Care Coverage, 49 EMORY L.J. 675, 682 (2000); Kershaw, supra note 118.
121
Gardner, supra note 120, at 682.
122
Id. at 682-83.
unlucky combinations of normal genes contribute to the risk of autism and

schizophrenia.”123
Additional research has proven that there is a biological basis for
schizophrenia, bipolar disorder and major depression.124 Schizophrenia is
one of the most debilitating types of mental illness.125 It can often interfere
with an individual’s ability to think clearly, manage emotions, make
decisions, interact and relate, and differentiate between reality and
fantasy.126 Research on the biological basis of schizophrenia has shown
that there is a possible genetic disposition to the illness.127
Research has shown that there are possible abnormalities in certain
genes or in certain areas of the genome at a specific point on a specific
chromosome.128 Hundreds of studies have also proven that schizophrenic
patients have less grey matter than non-schizophrenic patients in addition
to enlarged ventricles and fluid-filled spaces in the brain.129 Additionally,
electrical transmissions in schizophrenic patients have been found to be
abnormal.130 This conclusion was reached in a study where electrodes were
placed on the heads of schizophrenic patients and the electrical events were
recorded and analyzed.131 A schizophrenic’s brain’s neurons do not
function normally in the frontal lobe, revealing that a schizophrenic patient
has fewer neurons than a normal patient, that their neurons are more
randomly organized, and that their frontal lobes are also smaller.132 “The
National Institute of Mental Health likens the search for better treatments
123
124
See Gardner, supra note 120, at 683-85.
125
Id. at 683.
126
NAT’L ALLIANCE ON MENTAL ILLNESS, SCHIZOPHRENIA (2009),
http://www.nami.org/Content/ContentGroups/Helpline1/Schizophrenia_Fact_Sheet
.htm.
127
Subhagata Chattopadhyay, Tracking Genetic and Biological Basis of
Schizophrenia, 2 INTERNET J. MENTAL HEALTH (2004).
128
See generally, Richard E. Straub & Daniel R. Weinberger, Schizophrenia
Genes-Famine to Feast, 60 BIOLOGICAL PSYCHIATRY 81, 82 (2006).
129
Martha E. Shenton et al., A Review of MRI Findings in Schizophrenia, 49
SCHIZOPHRENIA RES. 1, 23, 34-35 (2001).
130
See id. at 35.
131
Jyrki Ahveninen et al., Inherited Auditory-Cortical Dysfunction in Twin
Pairs Discordant for Schizophrenia, 60 BIOLOGICAL PSYCHIATRY 612, 613-14
(2006).
132
See Gardner, supra note 120, at 683.
for those suffering from schizophrenia . . . to those involving heart disease

or diabetes.”133
Research over the last few decades has also revealed a biological
basis in depression. The National Alliance on Mental Health has stated that
scientific research “has firmly established that major depression is a
biological, medical illness.”134 Major depression is a condition which is
continually persistent and can drastically interfere with an individual’s
behavior, mood, level of activity, and physical health.135 Research has
shown that patients with major depression have a decreased level of neural
activity in a specific area of the brain.136 Additionally, major depression
responds well to biologically based therapy which suggests that there is an
organic nature to the illness.137 Among medical illnesses, depression is the
leading cause of disability in the United States and many other countries.138
In addition to schizophrenia and major depression, research has
shown that there is also a biological basis for bipolar disorder. Bipolar
disorder, which is also known as manic depression, causes extreme shifts in
one’s mood, energy and overall functioning.139 Recurring episodes of
mania and depression can last from a few days to months and usually begin
in adolescence or early adulthood.140 Studies have suggested that there
may be a genetic basis for bipolar disorder due in part to the fact that the
disorder often runs in families.141 The exact cause of bipolar disorder is
unknown, although scientists believe that it is caused by multiple factors
which produce a chemical imbalance in certain areas of the brain.142 While
schizophrenia, major depression and bipolar disorder are included in most
state definitions of mental illness, many other severe disorders “for which a
133
See Dark, supra note 109, at 202.
134
NAT’L ALLIANCE ON MENTAL ILLNESS, MAJOR DEPRESSION (2009),
http://www.nami.org/Template.cfm?Section=By_Illness&template=/ContentMana
gement/ContentDisplay.cfm&ContentID=7725 .
135
Id.
136
See Gardner, supra note 120, at 684.
137
Id. at 685.
138
Id.
139
NAT’L ALLIANCE ON MENTAL ILLNESS, BIPOLAR DISORDER (2009),
http://www.nami.org/Template.cfm?Section=By_Illness&Template=/TaggedPage/
TaggedPageDisplay.cfm&TPLID=54&ContentID=23037.
140
Id.
141
Id.
142
Id.
biological basis has emerged through recent research” are excluded from
the definition of mental illness by many states.143
The advocates of mental health parity have tried to incorporate
specific definitions of mental illness in past legislation. In 1999, Senators
Domenici and Wellstone introduced the Mental Health Equitable
Treatment Act of 1999.144 This bill provided for specific illnesses to be
covered called “severe biologically based mental illness” which included
“schizophrenia, bipolar disorder, major depression, obsessive compulsive
and panic disorders, posttraumatic stress disorder, autism, and other severe
and disabling mental disorders such as severe anorexia nervosa and
attention-deficit/hyper activity disorder.”145 This bill did not pass and three
years later they introduced the Mental Health Equitable Treatment Act of
2002.146 This bill advocated for a broader definition of mental illness and
stated that mental health benefits should include “all categories of mental
health conditions listed in the Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition, or the most recent edition.”147 This was the
position advocated by the House of Representatives during the debate over
the Paul Wellstone and Pete Domenici Mental Health Parity and Addiction
Equity Act of 2008, but it was ultimately rejected.148
With the growing acceptance of mental illnesses and their
biological basis, there is no reason not to have a specific list of severe
mental illness and substance use disorders in mental parity legislation. The
new legislation states that mental health conditions or substance use
disorders can be defined under the terms of the group health plan in
accordance with applicable federal and state law.149 Certain states do not
even define mental illness at all and leave it solely to the health plan
provider to define.150 The evidence that many mental illnesses are
143
Gardner, supra note 120, at 685. For example, many states exclude
obsessive-compulsive disorder, panic disorder, eating disorders, and post-traumatic
stress disorder. See supra Part III.
144
Mental Health Equitable Treatment Act of 1999, S. 796, 106th Cong.
(1999).
145
Id. § 2(b)(5).
146
(2001).
147
Id. § 2(e)(3).
148
149
Emergency Economic Stabilization Act of 2008, Pub. L. No. 110-343,
§512, 122 Stat. 3881-3892 (2008).
150
See Kaplan, supra note 4, at 353.
biologically based and treatable is stronger today than it ever has been.
Therefore, moving forward, there should be an attempt to pass an
amendment that would more specifically state which mental illnesses
should be covered instead of leaving it to the states and group plans, which
results in a great variation in coverage that individuals may receive.
A good starting point for defining mental illness in parity
legislation would be the definition of mental illness advocated for in 1999
that listed only severe mental illnesses. This definition was much narrower
than the position advocated in 2002 and advocated by the House of
Representatives in 2008, which was to include all illnesses in the DSM.
While many states provide coverage for schizophrenia, bipolar disorder and
major depression, many states do not cover obsessive-compulsive and
panic disorders, PTSD, autism, ADHD and severe eating disorders, all of
which were included in the list of severe mental illness in 1999. An
amendment defining mental illness under these terms151 would provide the
minimum list of mental illnesses that insurers would be required to cover.
States should of course be allowed to provide additional coverage at their
discretion.
VI. THE DIMINISHED SOCIETAL COSTS ASSOCIATED WITH

PROVIDING MENTAL HEALTH PARITY COVERAGE
One of the reasons that MHPAEA was able to gain passage is that
concerns over the cost implications of providing parity for mental health
coverage were alleviated. Many feared that the cost of health plans for
employers and insurers would skyrocket, however that was proven to be an
unwarranted concern. As previously mentioned, a 2006 study in the New
England Journal of Medicine found that insurers' costs rose less than half a
percentage point when full parity was required for federal workers starting
in 2001.152 Additionally, the Congressional Budget Office Cost Estimate,
prepared in 2007, found that if the more generous House bill were enacted,
151
The definition would include “severe biologically based mental illness”
defined as schizophrenia, bipolar disorder, major depression, obsessive-compulsive
and panic disorders, posttraumatic stress disorder, autism, the eating disorders
anorexia and bulimia, and severe attention-deficit/hyperactivity disorder.
152
See Shute, supra note 42.
the costs for premiums would increase for group health insurance by an
average of only about 0.4 percent.153
In addition to the fact that employer costs regarding health plans
will not increase, many factors point towards proving that providing mental
health parity will reduce the costs associated with treating mental illness.154
For example, many studies have shown that individuals who receive
psychiatric care incur fewer medical costs over the long term and less
morbidity than those individuals who do not receive psychiatric
treatment.155
Psychiatrist Stephen M. Stahl, Director of the Clinical
Neuroscience Research Center, has found numerous hidden costs that are
associated with not treating major depression. Examples of these hidden
costs include but are not limited to: fatal accidents resulting from impaired
concentration; patient morbidity such as suicide attempts, accidents,
resultant illnesses, lost jobs, failure to advance in career and school; and
social costs such as dysfunctional families, absenteeism, decreased
productivity, job-related illnesses, and adverse effects on quality control in
the workplace.156 Employers have also found that productivity tends to
increase after workers are treated for mental illnesses and substance abuse
problems and that such treatments can reduce the number of lost
workdays.157 The American Psychiatric Association has reported that
untreated mental illness costs employers $70 billion each year, which is
primarily due to lost productivity.158
In addition to the impact that mental illness has on work
productivity, the mentally ill account for a large portion of both the
homeless and the incarcerated. The Federal Task Force on Homelessness
and Severe Mental Illness found that approximately one-third of the
153
CONGRESSIONAL BUDGET OFFICE COST ESTIMATE (2007) at 4,
http://www.cbo.gov/ftpdocs/88xx/doc8837/hr1424ec.pdf.
154
Maria A. Morrison, Changing Perceptions of Mental Illness and the
Emergence of Expansive Mental Health Parity Legislation, 45 S.D. L. REV. 8, 22-
26 (2000).
155
Id. at 26.
156
Id.
157
See Carroll, supra note 26, at 582.
158
Sara Noel, Comment, Parity in Mental Health Coverage: The Goal of
Equal Access to Mental Health Treatment Under the Mental Health Parity Act of
1996 and the Mental Health Equitable Treatment Act of 2001, 26 HAMLINE L.
REV. 377, 398-99 (2003).
estimated 600,000 homeless people suffer from a severe mental illness.159

Additionally, “[a]mong all jail inmates, twenty-four percent reported at
least one symptom of psychotic disorder, and sixty-four percent reported
some degree of mental health problems.”160 The trend in the last several
decades has been a shift from institutionalization of the mentally ill to the
incarceration of the mentally ill.161 The reasons for this trend often have to
do with a lack of funding from both the federal and state level, disallowing
many mental hospitals and community-based mental health services the
ability to provide adequate treatment.162 This has lead to an increase,
among other things, to the already increasing incarceration rate.163 Studies
have shown that incarcerating one individual costs more than $23,000 per
year, which is undoubtedly a huge cost on society.164 While it is clear that
there have been grave concerns over the cost of implementing mental
health parity in the past,165 numerous studies and data reports have
alleviated those fears.166 Without a specific definition of mental illness,
many severe disorders will continue to go untreated and society will
continue to suffer the consequences of inadequate treatment of the mentally
ill.
VII. CONCLUSION
The Paul Wellstone and Pete Domenici Mental Health Parity and
Addiction Equity Act of 2008 is a significant advancement in providing
parity for mental illnesses and substance use disorders in that it finally
provided complete parity between medical and surgical benefits and mental
health and substance use disorders.167 Although it is a significant step up
159
Id. at 399 n.128.
160
John D. King, Candor, Zeal, and the Substitution of Judgment: Ethics and
the Mentally Ill Criminal Defendant, 58 AM. U. L. REV. 207, 211 n.12 (2008).
161
See id. at 212 n.15.
162
Id.
163
Id.
164
Maureen Carroll, Comment, Educating Expelled Students After No Child
Left Behind: Mending an Incentive Structure That Discourages Alternative
Education and Reinstatement, 55 UCLA L. REV. 1909, 1914 n.23 (2008).
165
“Cost was the second most influential factor cited by interviewees as
shaping the definition of mental illness.” Peck & Scheffler, supra note 74, at 1092.
166
See Shute, supra note 42; CONGRESSIONAL BUDGET OFFICE COST
ESTIMATE (2007), http://www.cbo.gov/ftpdocs/88xx/doc8837/hr1424ec.pdf.
167
See 42 U.S.C.A. § 300gg-5(a)(3)(A) (West Supp. 2009).
from all previous parity bills, the new bill still has a glaring deficiency
which, in order to be remedied, requires the passage of an amendment
which would aid in achieving full parity for all citizens across all states.
The bill currently does not specifically define mental illness or
substance use disorders. It states that the definition of mental illness and
substance use disorders should be in accordance with state and federal
laws.168 As evidenced by the varying state laws, the illnesses covered, and
thus how much coverage individuals are afforded, varies greatly depending
on the state in which one resides. With the growing acceptance of mental
illnesses and the evidence that mental illnesses are biologically based, it is
clear that mental illnesses are serious disorders that can be and must be
treated. Monetary costs have been a central part of the parity debates for
years and have often been a major factor in how states shape their
definitions of mental illness. Now that concerns over cost have been
alleviated, it is no longer a valid argument against a list of specific mental
illnesses to be covered. Furthermore, society as a whole will greatly
benefit from the adequate treatment of the mentally ill.169
An amendment should be passed in order to provide a specific
definition of what constitutes a mental illness or substance use disorder so
that the coverage people receive is not varied. The amendment should be
based on the definition that was advocated in the Mental Health Equitable
Treatment Act of 1999.170 That bill provided for specific illnesses to be
covered called “severe biologically-based mental illness” and included
“schizophrenia, bipolar disorder, major depression, obsessive compulsive
and panic disorders, posttraumatic stress disorder, autism” as well as
“anorexia nervosa and attention-deficit/hyper activity disorder.”171 An
amendment should be passed based on this previously proposed definition
and should identify each of the illnesses listed above. The MHPAEA has
done a great deal in providing parity between mental illness, substance use
coverage and physical illness coverage. However, in order to truly provide
parity for mental illness and substance use disorders, an amendment should
be passed so that individuals throughout the country receive the same level
of coverage and can receive treatment for the most debilitating mental
disorders.
168
Id. § 300gg-5(e)(4).
169
See supra Part V.
170
(1999).
171
Id.
EXAMINING CURRENT PROPOSALS FOR
INCREASING THE FEDERAL ROLE IN
DEALING WITH COASTAL HURRICANE RISK
Louis Cruz *
***
This note distinguishes predatory from subprime lending, while focusing on

the insurance consequences of predatory lending. It considers how single
premium credit insurance (SPCI) and private mortgage insurance (PMI),
two mortgage-related insurance products, have affected the current
predatory lending crisis. This note argues for reform that eliminates SPCI
and makes PMI a more feasible option for insureds. Such reform would
allow subprime lenders to offer mortgages to qualified borrowers, while
reducing the amount of predatory lending and foreclosures. The
introduction of this note presents some background information regarding
subprime lending and predatory lending. The second part examines
several issues concerning the role of insurance in the subprime mortgage
market. Third, reform measures necessary to alleviate the issues with
mortgage insurance are discussed. Finally, the fourth section studies
recent actions by the Federal Reserve Board and analyzes whether they
can be expected to bring meaningful change. It concludes that, although
the Fed’s new regulations are a step in the right direction, there needs to
be an outright ban of SPCI and predatory must be stopped completely.
***
I. INTRODUCTION
The recent flurry of coastal risk proposals vying for federal

adoption mirrors the sudden onslaught of a severe coastal storm. The
clamor to develop a framework to effectively manage coastal risk is
understandable. The public and private sector response to Hurricane
Katrina show that the current system for handling coastal risk is becoming
less feasible by the day. Katrina has led leading observers to note that,
*
Class of 2010.
“[i]nsurers – and society as a whole – need to reconsider the potential for

mega-catastrophes” and in turn “will need to adjust to this new reality.”1
Ideally, the proposal that emerges as successful should exhibit
three characteristics. First, the plan the federal government ultimately
adopts should expand the capacity of the existent private market to field
coastal risk so as to better deal with future storms. Second, the plan should
avoid completely crowding out the existent private sector players that
already seek to insure coastal risk. Finally, the successful plan should
minimize the degree to which non-coastal property owners subsidize the
risk undertaken by those who choose to live in storm-prone coastal areas.
It may be that no current proposal fully satisfies all three criteria.
Indeed, upon analysis, none of them seem to. However, some of the
current coastal risk plans contain aspects that show promise in light of
these three criteria. The proposals that show promise lack federal
reinsurance mechanisms, thereby not crowding out private market
reinsurers and sidestepping large scale cross-subsidization. Additionally,
the most favorable of the available plans contain coastal risk transferring
mechanisms designed to increase the use of the private capital market’s
capacity to deal with future catastrophic coastal risk. Also, plans that
contemplate homeowner catastrophe savings accounts may provide a way
to enable coastal residents to afford actuarially sound private market
insurance rates. Finally, plans that envision private insurer catastrophe
reserving may present ways to increase private market capacity to deal with
coastal catastrophe risk, thereby keeping private insurers in the business of
insuring coastal risk.
The following examination of the current proposals begins with a
brief survey of the meteorological and geographical settlement trends that
have worked to necessitate an immediate rethinking of coastal risk policy.
Next, this study will move into a review of the current coastal risk
framework. Following this, the current proposals will each be discussed.
Finally, each proposal will be evaluated against the capacity, displacement
and cross-subsidization criteria discussed above.
1
TOWERS PERRIN, HURRICANE KATRINA: ANALYSIS OF THE IMPACT ON THE
INSURANCE INDUSTRY 4 (2005).
2009] COASTAL HURRICANE RISK 325
II. BACKGROUND
A. EXTREME WEATHER TRENDS
On a world-wide scale, extreme weather events are on the rise.

The number of major weather-caused natural catastrophes has increased
from an average of 1.5 per year since the 1950s to 4.5 in recent years.2 As
part of this trend, catastrophic storms in the Atlantic and Gulf Coasts of the
U.S. have also become more prevalent.3 While climatologists disagree as
to whether the available data is adequate to determine the magnitude of this
current extreme weather event uptick,4 what cannot be debated is that the
number of storms per hurricane season in the Atlantic and Gulf coasts has
increased since as recently as 1995.5 In comparison to the period of 1970-
1994, the amount of hurricane activity on the Atlantic and Gulf coasts has
increased by over 60% in the ten years that followed.6 Further examination
of the historical extreme weather data reveals that compared to an Atlantic
basin annual average of 1.5 major storms a year in the period from 1970 to
1994, the region has experienced an average of 3.9 major storms per year
since 1995.7 When measuring only the subset of storms that actually make
2
COMITÉ EUROPÉEN DES ASSURANCES, REDUCING THE SOCIAL AND
ECONOMIC IMPACT OF CLIMATE CHANGE AND NATURAL CATASTROPHES 10
(2007).
3
Rick Jervis, Data Show U.S. Riding Out Worst Storms on Record, USA
TODAY, Oct. 22, 2008, at 3a.
4
See Eliot Kleinberg, Hurricanes May Trace to Cycles, Not Warming, PALM
BEACH POST, June 3, 2007, at 1C (noting that there are problems with researchers
basing their conclusions on a small statistical sample of scientifically significant
research); but cf. Cathy Zollo, Experts Spar Over Warming’s Impact, SARASOTA
HERALD-TRIBUNE, Apr. 15, 2006, at A1 (many climatologists argue there is
enough data, and some describe the recent uptick in storms as a natural cyclical
occurrence, while others attribute it to global warming).
5
See generally Stanley B. Goldenberg et al., The Recent Increase in Atlantic
Hurricane Activity: Causes and Implications, 293 SCIENCE 474 (2001) (discussing
the years 1995 to 2000 as experiencing the “highest level of North Atlantic
hurricane activity in the reliable record.”).
6
HOWARD KUNREUTHER, REFLECTIONS ON U.S. DISASTER INSURANCE
POLICY FOR THE 21ST CENTURY 2 (Am. Enter. Inst.-Brookings Joint Ctr. For
Regulatory Studies, 2007).
7
PHILIP J. KLOTZBACH & WILLIAM M. GRAY, EXTENDED RANGE FORECAST
OF ATLANTIC SEASONAL HURRICANE ACTIVITY AND LANDFALL STRIKE
PROBABILITY FOR 2009 26 (2009).
landfall the data is just as striking. From the 1930s to the 1960s, there was
an average of 1.8 storms per year making landfall, while over the last
twelve years that figure has risen to 2.2.8
In spite of the fact that the causes of this trend are beyond the
scope of this study, what is clear is that there has been a pronounced
increase in the number of coastal storms in the southeast Atlantic and Gulf
coasts over the past decade. While this meteorological trend is startling
when considered on its own, what makes it all the more attention worthy is
that it coincides with a trend of increasing coastal populations in the United
States.
B. COASTAL POPULATION TRENDS
At present, coastal counties comprise 17% of the total land area of

the United States, but claim 53% of the country’s total population.9 While
the country’s coastal-to-non-coastal population ratio has remained
relatively stable over the last forty years, the limited area of coastal
geography has contributed to significantly higher population densities in
these coastal areas.10 Data from the 2000 census show that while the
country at large experienced a density increase in the range of 38% from
1970 to 2000, the Southeast Atlantic coast area had an increase in
population density of nearly 66% over the same period.11 This study
focuses particularly on those coastal regions that have historically been
most susceptible to coastal storm damage: the Southeast Atlantic and Gulf
coasts.12
Currently, 9% of the nation’s coastal population resides in the
Southeast Atlantic region.13 Florida in particular has most of its total
8
Manuel Lonfat et al., Atlantic Basin, U.S. and Caribbean Landfall Activity
Rates over the 2006-2010 Period: An Insurance Industry Perspective, 59A TELLUS
499, 500 (2007).
9
KRISTEN M. CROSSETT ET AL., NAT’L OCEANIC AND ATMOSPHERIC ADMIN.,
POPULATION TRENDS ALONG THE COASTAL UNITED STATES: 1980-2008 6 (2004).
10
Id.
11
KUNREUTHER, supra note 6, at 1.
12
See Stanley A. Changnon, Characteristics of Severe Atlantic Hurricanes in
the United States: 1949-2006, 48 NAT. HAZARDS 329, 333 (2009) (noting that in
the 57 year period examined, the Southeast and South ranked first and second in
the United States for the number of losses from storm events).
13
CROSSETT ET AL., supra note 9, at 16.
population distributed throughout its coastal counties.14 Specifically

noteworthy is the fact that it had been anticipated that the southeast
Atlantic region’s coastal population would increase by 1.1 million people,
or 8%, between 2003 and 2008.15 This would be the largest percentage
increase of all U.S. coastal regions within that five year period.16
Concomitant with this population growth is an increase in insured coastal
property. More than 80% of Florida’s total insured property exposure can
be classified as coastal.17 The estimated total value of insured coastal
exposure in Florida is over $2 trillion dollars.18
Extending from the Florida Keys to southern Texas and including
the coastline of six states, the Gulf region’s coastal population of 19.1
million residents claims just over 13% of the nation’s total coastal
population.19 The majority of the population in both Louisiana and the
western coast of Florida are in Gulf coastal counties.20 A total of 23% of
the region’s total land area and 32% of the region’s population are
distributed throughout the Gulf Region’s 144 coastal counties.21 The
combined commercial and residential insured exposure in the Gulf region
is estimated to be over one trillion dollars. 22
C. IMPLICATIONS AND IMPACT
On the whole, the total insured value of property in U.S. coastal

regions is increasing at an astonishing rate. Recently it was determined
that from 2004 through 2007, the insured value of properties in coastal
areas of the United States grew at a compound annual growth rate of just
14
Id.
15
Id.
16
Id.
17
Press Release, Prop. Cas. Insurers Ass’n of Am., Coastal Insurance
Concerns Demonstrate Need for New Catastrophe Solutions (Mar. 15, 2007),
http://www.pciaa.net/ (follow "Media Center: News Releases: By Date" menu;
then scroll to Mar. 15, 2007).
18
Id.
19
CROSSETT ET AL., supra note 9, at 18.
20
Id.
21
Id.
22
See Ieva M. Augustums, Nearly $1T of Insured Property in Ike’s Path,
USA TODAY, Sept. 12, 2008, http://www.usatoday.com/money/ economy/2008-09-
11-2141571223_x.htm.
over 7%.23 This annual growth rate will lead to a doubling of the total
insured coastal value every decade.24 In light of the recent increases in
extreme weather events in the Atlantic and Gulf coasts, this insured growth
takes on a number of serious implications. Due to the increasing frequency
of coastal storms, the growing size of coastal communities and the rapid
growth rate of insured exposure concentrated in coastal areas, extreme
weather related events are becoming increasingly financially destructive.
Five of the ten most expensive storms in U.S. history have occurred since
1990.25 Of particular import, the 2004/05 hurricane season was
unprecedented. In terms of U.S. insured loss, the six hurricanes that hit the
southeast Atlantic and Gulf Coasts make up half of the list of the twelve
largest disasters in the last forty years.26 Katrina alone is the most costly
event the U.S. insurance industry has ever experienced, resulting in more
than $61 billion in insured losses and $125 billion in total losses.27
Another important implication for these current trends in extreme
weather and coastal population growth is the need for a revaluation of
current policy. The federal role in coastal disaster policy has been the
subject of recent intense debate. To date, a flurry of legislative and private
sector proposals have followed in the wake of the 2004/05 hurricane
season. Each of these proposals envisions a reoriented role for federal
government in coastal catastrophe policy in an attempt to overhaul the
current coastal risk policy framework.
III. CURRENT COASTAL RISK POLICY FRAMEWORK
In the context of coastal hurricanes, the current policy landscape is

a patchwork of private insurance, government assistance and government
subsidized insurance programs. Presently, private insurance plays a
relatively limited role. Due to the fact that many private insurers perceive
the risks associated with coastal catastrophic loss as unacceptably high,
many have curtailed writing such polices.28 Some major insurers, such as
23
AIR WORLDWIDE CORP., THE COASTLINE AT RISK: 2008 UPDATE TO THE
ESTIMATED INSURED VALUE OF U.S. COASTAL PROPERTIES 1 (2008).
24
Id.
25
ALEXANDER BOLONKIN, CONTROL OF REGIONAL AND GLOBAL WEATHER 3
(2007), http://arxiv.org/pdf/physics/0701097.
26
KUNREUTHER, supra note 6, at 2.
27
COMITÉ EUROPÉEN DES ASSURANCES, supra note 2, at 10.
28
See, e.g., Sandra Fleishman, Sea Change in Insurers’ Coastal Coverage;
Many Firms Opt to End or
Allstate, no longer write new policies in Florida, Louisiana, Mississippi,

parts of New York, and coastal Texas.29 Similarly, State Farm has stopped
offering insurance that would protect against storm damage within one mile
of the ocean and announced in 2006 that it would sell no new policies in
Mississippi.30 Moreover, where available, private coverage often protects
against wind damage but, excludes flood coverage.31 As a result, property
owners who want flood coverage must purchase it separately. This is done
through either the National Flood Insurance Program or state sponsored
Fair Access to Insurance Requirement plans.32
A. NATIONAL FLOOD INSURANCE PROGRAM
The National Flood Insurance Program was established by

Congress in 1968.33 When enacted the goals of the National Flood
Insurance Program were twofold: minimize flood damage through
floodplain management, and provide property owners with flood
insurance.34 Corresponding to these goals are the two main initiatives that
currently characterize the National Flood Insurance Program. First,
National Flood Insurance Program actively generates flood maps indicating
100-year floodplains.35 These maps anchor the National Flood Insurance
Limit New Policies, WASH. POST, Dec. 30, 2006, at F1.

29
Spencer S. Hsu, Insurers Retreat from Coasts; Katrina Losses May Force
More Costs on Taxpayers, WASH. POST, Apr. 30, 2006, at A1; see also Examining
the Terrorism Risk Insurance Program, Before S. Comm. on Banking, Housing and
Urban Affairs, 107th Cong. 24 (2007) (Statement of J. Robert Hunter, Director of
Insurance, Consumer Federation of America).
30
Fleishman, supra note 28; see also An Examination of the Availability and
Affordability of Property and Casualty Insurance in the Gulf Coast and Other
Coastal Regions: Hearing Before the S. Committee on Banking, Housing and
Urban Affairs, 109th Cong. 7 (2007) (statement of Mel Martinez, Member, S.
Comm. on Banking, Housing, and Urban Affairs).
31
JUSTIN R. PIDOT, COASTAL DISASTER INSURANCE IN THE ERA OF GLOBAL
WARMING: THE CASE FOR RELYING ON THE PRIVATE MARKET 37 (Georgetown
Envtl. Law & Policy Inst., 2007).
32
Id. at 12, 21.
33
U.S. GOV’T ACCOUNTABILITY OFFICE, GAO-07-285, CLIMATE CHANGE:
FINANCIAL RISKS TO FEDERAL AND PRIVATE INSURERS IN COMING DECADES ARE
POTENTIALLY SIGNIFICANT 46 (2007).
34
See PIDOT, supra note 31, at 12-13.
35
Id. at 13.
Program’s premium rates and mitigation requirements.36 Second, the

program offers up to $250,000 in insurance against flood damage to
homeowners in communities that have adopted floodplain regulations
meeting the minimum standards set by the program.37 Although
participation was initially voluntary the National Flood Insurance Program
now requires homeowner’s within mapped floodplains who have federally
insured mortgages to purchase and maintain flood insurance. 38
Perhaps the most important aspect of the National Flood Insurance
Program is the Write-Your-Own program. In 1983 the federal government
permitted private companies to write National Flood Insurance Program
coverage under the Write-Your-Own program.39 Under this program
private insurers wrote policies and handled claims and in exchange
received roughly a third of the premiums collected as sales commission and
sustained 3.3% of incurred losses.40 Currently, of the nearly 5.3 million
policies written by the National Flood Insurance Program, upwards of 95%
of the policies are written by private companies under the Write-Your-Own
program.41
In spite of its long history, the National Flood Insurance Program
has significant shortcomings. Due to the heavy subsidization of the
premiums by the federal government the program is often criticized as not
being actuarially sound and therefore ultimately not self-supporting.42 The
effect of subsidized premiums is two-fold. First, the program does not
generate enough reserves to protect against catastrophic loss.43 This can be
seen by the fact that of the $23 billion in claims paid out by the National
Flood Insurance Program following Hurricane Katrina much of this money
36
Id.
37
42 U.S.C. §§ 4012, 4013 (2006).
38
42 U.S.C. § 4012a (b)(1) (2006).
39
DEP’T OF HOMELAND SEC., PRIVACY IMPACT ASSESSMENT FOR THE
NATIONAL FLOOD INSURANCE PROGRAM APPEALS PROCEDURE 2, 5 (2006).
40
WILLIAM O. JENKINS, JR., U.S. GOV’T ACCOUNTABILITY OFFICE GAO-05-
532T, NATIONAL FLOOD INSURANCE PROGRAM: OVERSIGHT OF POLICY ISSUANCE
AND CLAIMS 8 (2005); Robert J. Rhee, Catastrophic Risk and Governance After
Hurricane Katrina: A Postscript to Terrorism Risk in a Post-9/11 Economy, 38
ARIZ. ST. L.J. 581, 610 (2006).
41
JENKINS, supra note 40, at 7.
42
U.S. GOV’T ACCOUNTABILITY OFFICE GAO-06-119, FEDERAL EMERGENCY
MANAGEMENT AGENCY: IMPROVEMENTS NEEDED TO ENHANCE OVERSIGHT AND
MANAGEMENT OF THE NATIONAL FLOOD INSURANCE PROGRAM 9 (2005).
43
Id.
came from loans taken directly from the federal treasury.44 Second,
because the premiums are subsidized and therefore artificially low,
property owners have no incentive to avoid moral hazard in or to invest
mitigation measures.45 This dynamic is demonstrated by the fact that
property owners that have repeatedly suffered damage from floods have
rebuilt their property in the same locations and continue to receive
subsidized rates.46 This is in stark contrast to the ideally analogous
scenario in which private insurers have the ability to cancel coverage.
B. STATE DISASTER INSURANCE PLANS
Many states have in operation, Fair Access to Insurance

Requirement or “FAIR” plans. Ostensibly, these plans provide coverage to
property owners unable to secure insurance in the private market.
Although FAIR plans were not originally designed to cope with coastal
natural catastrophe loss, these plans have expanded considerably and today
many states operate plans that primarily provide coastal storm coverage.47
FAIR plans are state-run insurance pools in which all property insurers
licensed in a state are required to participate.48 Participating insurers share
in the profits and losses of the high risk coverage.49 In 2004, a total of
$400 billion of coastal property was insured by state run FAIR plans.50
Instead of FAIR plans, other states have enacted Beach and
Windstorm Insurance Plans that provide coverage to property-owners in
coastal communities.51 Much like FAIR plans, most states that operate
Windstorm plans require participation by private property insurance
carriers in the state. Again, as in the case of FAIR plans, here the insurers
44
See RAWLE O. KING, CONG. RESEARCH SERV., NATIONAL FLOOD
INSURANCE PROGRAM: TREASURY BORROWING IN THE AFTERMATH OF HURRICANE
KATRINA 1, 3-4 (2006).
45
See PIDOT, supra note 31, at 34.
46
Id; Spencer M. Taylor, Insuring Against the Natural Catastrophe after
Hurricane Katrina, 20 NAT. RESOURCES & ENV’T. 26, 28 (2006).
47
See PIDOT, supra note 31, at 21.
48
See Ins. Info. Inst., Residual Markets, Aug. 2009, http://www.iii.org/
media/hottopics/insurance/residual/.
49
Id.
50
ROBERT P. HARTWIG & CLAIRE WILKINSON, INS. INFO. INST.,
PUBLIC/PRIVATE MECHANISMS FOR HANDLING CATASTROPHIC RISKS IN THE
UNITED STATES 33 (2005).
51
See Ins. Info. Inst., supra note 48.
share profits and losses. In 2004 upwards of $30 million of coastal

property was insured under some form of state Beach or Windstorm plan.52
Still other states have crafted even more unique programs. Florida
created the “Florida Citizens”, not-for-profit “insurer of last resort.”53 In
contrast to FAIR and Windstorm plans, Florida Citizens is less an
insurance pool, than a state run insurer.54 At present, Florida Citizens is the
largest insurer in the state with more than 1.3 million policyholders.55
Unfortunately, Florida Citizens’ premium rates are set by statute at levels
that are not actuarially sound.56 In 2007 Florida Citizens had more than
$400 billion in exposure, but received only $3 billion in yearly premiums.57
When Florida Citizens' losses exceed its claims-paying capacity in a single
year, it is required by statute to post-fund itself by imposing a statewide
assessment on every other line of insurance sold in the state.58
IV. THE PROPOSED PLANS
A. COALITION OF AMERICANS FOR SMART NATURAL

CATASTROPHE POLICY: MITIGATION
Formed in late 2005, the Coalition of Americans for Smart Natural

Catastrophe Policy is a grouping of environmentalists, academics, and
consumer rights groups.59 At present, the Coalition is one of the most high
52
HARTWIG & WILKINSON, supra note 50, at 39.
53
Id. at 32, 34.
54
Id. at 40.
55
INS. INFO. INST., AN EXCERPT FROM: RESIDUAL MARKET PROPERTY PLANS
-- FROM MARKETS OF LAST RESORT TO MARKETS OF FIRST CHOICE 2 (2007).
56
See PIDOT, supra note 31 at 18.
57
John W. Rollins, Florida Property Insurance – The “Citizens” View (June
19, 2007), http://www.casact.org/education/spring/2007/handouts/rollins.pdf.
58
INS. INFO. INST., supra note 55, at 4.
59
Presently the Coalition consists of: Consumer Federation of America,
Defenders of Wildlife, Environmental Defense, Friends of the Earth, National
Wildlife Federation, Republicans for Environmental Protection, Association of
State Floodplain Managers, Americans for Prosperity, Council for Citizens Against
Government Waste, Competitive Enterprise Institute, FreedomWorks, Taxpayers
for Common Sense, Association of Bermuda Insurers and Reinsurers, Reinsurance
Association of America, and the National Association of Professional Insurance
Agents. See Smarter Safer, About Us, http://www.smartersafer.org/about-us/ (last
visited Sept. 30, 2009)(for more information on the coalition’s expanding roster).
profile groups advocating their own national catastrophe proposal. Unlike

many of the other proposals discussed later, the Coalition’s plan focuses on
policy encouraging homeowner mitigation methods rather than federal
reinsurance or federal lending to state insurance pools. According to the
Coalition’s current mission statement they believe:
[T]he Federal government has a role in encouraging and helping
homeowners to undertake mitigation efforts to safeguard their homes
against hurricanes, [but] ... the coalition oppose proposals being considered
in Congress that would create moral hazards by providing direct or indirect
subsidies for coastal homeowners' insurance policies, thereby giving people
incentives to build homes in hurricane-prone, environmentally sensitive
areas.60
Specifically, the Coalition proposes that federal intervention be
limited to existing programs modified to increase emphasis on
preparedness and mitigation as opposed to any sort of federally aided
insurance coverage expansion.61 The full breadth of the Coalition’s
platform can be seen in its recent promotion of the Flood Insurance Reform
Modernization Act of 2007 and the Property Mitigation Assistance Act; as
well as its opposition to the Homeowners Defense Act of 2007.
The Coalition supported the version of the Flood Insurance Reform
and Modernization Act of 2007 that appeared before the Senate.62 While
described by the Coalition as modest legislation reauthorizing the National
Flood Insurance Program, the Flood Insurance Reform and Modernization
Act of 2007 would have made fairly significant changes to the program.63
First, the act would have attempted to make the National Flood Insurance
Program satisfy traditional criteria for actuarial soundness by phasing out
discounted premiums previously available for structures built prior to the
mapping and implementation of the program’s floodplain management
requirements.64 At present, these prerate map structures pay heavily
discounted rates on the first $35,000 of their structure’s insured value, and
full risk-based premium rates for the remaining insured value.65 Currently,
60
Id.
61
See id.
62
Smarter Safer, Senate Legislation, http://www.smartersafer.org/about -the-
legislation/senate-legislation, (last visited Sept. 30, 2009).
63
Id.
64
See Flood Insurance Reform and Modernization Act of 2007, S. 2284,
110th Cong. §§ 8(g), 14 (2007).
65
An Examination of the National Flood Insurance Program: Hearing Before
The United States Senate Comm. On Banking, Housing, and Urban Affairs, 110th
nearly a quarter of all policies written by the National Flood Insurance

Program are subsidized under this aspect of the program.66
Further, the Senate version of the Flood Insurance Reform and
Modernization Act of 2007 would have allowed the increase of National
Flood Insurance policy rates by 15% a year, up from the previous 10%
cap.67 The bill’s drafters arrived at this figure because the previous 10%
ceiling was shown not to be enough to ensure the program would have
sufficient funds to cover future obligations for policyholder claims,
operating expenses, and interest on debt stemming from the 2005 hurricane
season.68
In addition to supporting the Flood Insurance Reform and
Modernization Act of 2007 the Coalition also successfully opposed a
substantial amendment to the bill that would have made wind coverage
available to National Flood Insurance policyholders.69 In support of their
opposition to the wind coverage amendment, the Coalition cited estimates
that such an amendment would result in as much as $161 billion in new
taxpayer liabilities in 2009 alone if the U.S. Gulf Coast suffered a hurricane
season comparable to that of 2005.70 Further still, the Coalition urged that
the expanded program would also threaten public safety by encouraging
further development in hurricane prone coastal areas.71 In a press release
lauding the Senate’s subsequent rejection of the amended bill, the Coalition
suggested that instead of expanding the National Flood Insurance Program,
Congress should look to, “safety-oriented reform solutions that would help
homeowners better prepare for storms and reduce destruction caused by
Cong. 2 (2007) (testimony of David I. Maurstad Assistant Admin. for Mitigation

and Fed. Ins. Admin. of Fed. Emergency Mgmt. Agency).
66
Id.
67
See Flood Insurance Reform and Modernization Act of 2007, S. 2284,
110th Cong. § 6(b) (2007).
68
KING, supra note 44, at 2-3.
69
Press Release, Americans for Smart Natural Catastrophe Policy, Americans
for Smart Natural Catastrophe Policy Applauds Congress’ Extension of National
Flood Program Without Irresponsible Expansion (Sept. 29, 2008),
www.smartersafer.org/newsroom/press-releases (follow “Americans for Smart
Natural Catastrophe Policy Applauds Congress’ Extension of National Flood
Program Without Irresponsible Expansion” hyperlink).
70
Id.
71
Id.
natural catastrophes.”72 The 2007 Property Mitigation Act is an example of

a safety-oriented solution can that was supported by the Coalition.
As part of the Coalition’s overall program, the group championed
the unsuccessful 2007 Property Mitigation Assistance Act.73 The Act
sought to “authorize grants and loans to homeowners to harden their homes
against hurricanes and other disasters.”74 Pursuant to the Act, states that
met the terms in the bill would have received grants of at least $500,000 to
be distributed to residents through loan and grant programs to help
residents take such measures as adding storm shutters, hurricane clips, safe
rooms or any other activity that would mitigate the risks of future hazards
and natural disasters.75 The idea was to help at-risk homeowners protect
their homes from damage instead of expanding federal involvement in
insurance coverage or aid.76
Finally, the Coalition’s overall view of national coastal natural
catastrophe policy can be seen in their opposition to the Senate version of
the Homeowners’ Defense Act of 2007. The Coalition described the
proposed legislation as irresponsible.77 Essentially, the Homeowners’
Defense Act of 2007 consisted of three bundled programs designed to work
with existing state insurance pools. The first program would have
established an interstate federal consortium that would have attempted to
aid multiple states running coastal risk insurance pools in combining and
transferring this risk to the capital markets.78 The second program was an
insurance stabilization plan that would have allowed the Federal Treasury
to make loans to state pools in order to ensure their continued, liquidity in
the aftermath of a natural catastrophe.79 The third and final component of
the Act would have established a Federal reinsurance program designed to
sell reinsurance to state and interstate pools.80
72
Id.
73
Smarter Safer, supra note 62.
74
Id.; Property Mitigation Assistance Act of 2007, S. 2328, 110th Cong. §
2(m) (2007).
75
Property Mitigation Assistance Act of 2007, S. 2328, 110th Cong. § 2(m)
(2007).
76
See Smarter Safer, The Senate Should Support The Property Mitigation
Assistance Act of 2007, http://www.smartersafer.org/uploads/ S2328.pdf (last
visited Sep.15, 2009).
77
Smarter Safer, supra note 62.
78
Homeowners’ Defense Act of 2007, S.2310 §§ 101-108 (2007).
79
See id. § 201.
80
See id. § 301.
The Coalition’s criticism of the Senate version of the

Homeowners’ Defense Act of 2007 was that each aspect of the Act would
simply expand federal liability by introducing a federal role in state
insurance pools.81 As to the Federal Risk Consortium, the Coalition pointed
out that the States were already free to associate, and address catastrophe
risk.82 Therefore a permanent federal role in this process would be
unnecessary.83 Similarly, in their opposition to the aspects of the Senate
version of the Homeowners’ Defense Act of 2007 that provided for Federal
reinsurance and Federal loans to state programs, the Coalition adopted the
position that the Act would both displace existing private reinsurers and
promote the continuation of financially unsound state insurance
programs.84
B. ALLSTATE’S PROPOSAL: FEDERAL REINSURANCE BACKSTOP

TO STATE POOLS
Supporting a bill that would have created a federal reinsurance

backstop even prior to the monumental 2005 hurricane season, Allstate was
among the first private sector supporters of the concept of an increased
federal role in coastal natural catastrophe insurance coverage.85 As
Allstate’s was the first of the major insurer backed plans, the proposal was
vaguer than some subsequent proposals. However, a more fleshed out
version of the original Allstate proposal was soon incorporated into the
stalled Homeowners Insurance Protection Act of 2005. Allstate
86
subsequently endorsed the Act. With this in mind, the following
description of the Allstate proposal includes details from the Homeowners
Insurance Protection Act of 2005 to fill in some of the gaps that this early
private sector proposal exhibited.
81
See Smarter Safer, The Senate Should Not Adopt S. 2310 (2007),
http://www.abir.bm/downloads/SmartNatCatBriefingPapers.pdf (last visited
Sep.15, 2009).
82
See id.
83
See id.
84
See id.
85
R.J. Lehmann, A Catastrophic Battle, BEST’S REVIEW, Oct. 2006, at 38.
86
Id.; Business Wire, Allstate CEO to Call for National Catastrophe Plan at
National Press Club on Friday, Jan. 13; Devastation of 2005 Makes Reform an
Immediate National Priority, Jan. 9, 2006, available at
http://findarticles.com/p/articles/mi_m0EIN/is_2006_Jan_9/ai_n15989710/ (last
visited Sept. 30, 2009).
Allstate’s initial position was to create a federal government

sponsored catastrophe program that would provide reinsurance to a system
of state and regional catastrophe funds.87 This required states that did not
already have them to create catastrophe pools funded by, “all entities that
benefit from a robust local economy such as the banking and real estate
sectors.”88 In turn, the Department of the Treasury would make
reinsurance contracts available to any of the state programs that met
minimum requirements.89
As to rates, the text of the Homeowners Insurance Protection Act
of 2005 required that rates be risk-based.90 States would pay premiums
directly to the federal government.91 The federal reinsurance would cover
90% of losses in excess of either the capacity of each state program or the
projected losses from a 200-year event, whichever is greater.92 This
translates into the federal reinsurance coverage being triggered by events
causing roughly $50 billion or more in homeowners’ losses.93
C. TRAVELERS AND NATIONWIDE: FEDERAL REINSURANCE AND

INCREASED FEDERAL REGULATION
Two other leading property-casualty insurance carriers, Travelers

and Nationwide, soon followed Allstate in calling for an increased federal
role in coastal catastrophe policy. The Travelers/Nationwide proposal had
two significant aspects. First, the proposal sought to spread big windstorm
insurance risk across state borders by having it federally regulated.94
87
Lehmann, supra note 85.
88
Press Release, Ins. Info. Inst., Reforms Needed to Prepare for Major
Catastrophes, Insurance CEOs Tell Forum, http://www.iii.org/media/
updates/archive/press.748575/ (Jan.10, 2006).
89
Homeowners Insurance Protection Act of 2005, H.R. 4366, 109th Cong. § 4
(2005).
90
Id. § 7(b)(6)(B).
91
Id. § 7(a).
92
Id. § 7(b)(3).
93
David Dankwa, Allstate Takes Hot Seat as Industry Debates Catastrophe-
Fund Options, BESTWIRE Jan. 16, 2006, available at
http://www3.ambest.com/Frames/FrameServer.asp?AltSrc=23&Tab=1&Site=news
&refnum=80532.
94
The Travelers Cos., Insurance Agents, Travelers, Nationwide Urge Coastal
Wind Policy, INS. J., July 16, 2008, http://www.insurancejournal.com/news
/national/2008/07/16/91909.html.
Second, similar to the Allstate proposal, the Travelers/Nationwide plan

called for a federal reinsurance mechanism for extreme weather events.95
The first notable aspect of the Travelers/Nationwide plan was that
it called for a federal role in promoting a plan to spread coastal windstorm
risks across state borders.96 As opposed to setting up individual catastrophe
funds state by state, the Travelers/Nationwide plan encouraged a system of
inter-state zones: the Gulf, Florida, Southeast and Northeast.97 Pursuant to
the Travelers/Nationwide plan the federal government would not have a
financial role in maintaining these pools, but would oversee the wind
underwriting by private insurers.98 To do this, the Travelers/Nationwide
plan would require the creation of an independent federal agency to
regulate rates and set uniform rules from Texas to Maine for the wind
coverage portion of a homeowner's policy, while the states would continue
to regulate the other portions of a homeowner's policy.99
With regard to premiums, the Travelers/Nationwide proposal
requires setting “risk-based and actuarially sound rates using approved
standards and certified windstorm risk models approved by the federal
commission.”100 However, the drafters of the Travelers/Nationwide
proposal recognized that risk-based premiums may be out of reach for
many coastal residents.101 In response, the proposal provided for temporary
transitional subsidies lasting between 10 to 15 years. 102 To this end,
coastal residents who were unable to afford coverage would receive tax
credits to help them afford premiums.103 The credits would be funded by
tax surcharges imposed on those coastal residents able to afford
coverage.104
The Travelers/Nationwide federal reinsurance mechanism
functioned much like the Allstate backstop. The federal backstop required
regions set up catastrophe funds and pay premiums to a federal program in
95
Id.
96
See id.
97
THE TRAVELERS INST., TRAVELERS COASTAL WIND ZONE PLAN 5 (2009),
http://www.travelers.com/iwcm/Trv/docs/TRV_Coastal_Wind_ Zone_Web.pdf.
98
Id. at 6.
99
Id. at 5.
100
Id.
101
Jay S. Fishman, Before the Next 'Big One' Hits, WALL ST. J., Aug. 27,
2007, at A10.
102
Id.
103
THE TRAVELERS INST., supra note 97, at 6.
104
Id.
return for reinsurance that would be used to support the regional fund.105
The threshold for the Travelers/Nationwide federal reinsurance mechanism
has yet to be finalized as it is stated to be for “extreme events (such as
hurricanes causing losses several times greater than those arising out of
Hurricane Katrina).”106
D. THE HARTFORD’S PLAN: HOMEOWNER CATASTROPHE

SAVINGS ACCOUNTS, MANDATORY FLOOD INSURANCE, AND
FEDERAL BACKSTOPPING
The Hartford announced its public-private natural catastrophe plan

in the summer of 2008.107 The Hartford’s Coastal Catastrophe Partnership
Plan was “designed to deal with the looming economic crisis posed by a
major hurricane.”108 The plan was to occupy a middle ground of
government involvement. Hartford CEO, Romani Ayer, said the Coastal
Catastrophe Plan was designed to navigate between the “‘socialistic’ efforts
of some in Congress, and reinsurers who believe all catastrophe protection
should be left to the private market.”109 It was the last of the plans put forth
by a major private insurer and in many ways it was the most detailed. The
plan had three facets that merit discussion. First, it called for tax-deferred
catastrophe savings accounts.110 Second, the Hartford proposal required
105
See supra Part III.B.
106
The Travelers Cos., supra note 94.
107
See The Hartford Fin. Serv. Group Inc., The Hartford Unveils National
Coastal Home Insurance Plan, INS. J., July 31, 2008,
http://www.insurancejournal.com/news/national/2008/07/31/92393.htm (noting
that the plan will include subsidized premiums for homeowners and a federal
reinsurance program).
108
Daniel Hays, Hartford Proposes National Cat Plan Featuring Federal
Backstop, IRA-Type Fund, NAT’L UNDERWRITER – PROP. & CAS. INS., Aug. 4,
2008, available at 2008 WLNR 14506223 (quotation omitted).
109
Id.
110
See RAMANI AYER, THE HARTFORD FIN. SERV. GROUP, BUILDING A
NATURAL CATASTROPHE SOLUTION 5 (2007), http://www.coastalpartner
ship.org/documents/CCP Plan Outline.pdf (noting that the Hartford proposes “the
creation of IRA-like savings vehicles – perhaps called ‘supplemental catastrophic
security accounts’”).
mandatory homeowner’s flood insurance.111 Finally, the proposal called

for a Federal reinsurance backstop.112
Perhaps the most striking aspect of the Hartford’s plan was that it
encouraged the creation of IRA-like savings vehicles dubbed
“supplemental catastrophic security accounts.”113 The accounts would have
permitted property owners in coastal regions to establish tax-deferred
reserves to pay for insurance and other expenses related to disasters.114
Specifically, the idea is that allowing homeowners to use tax-deferred
dollars to pay for catastrophe insurance could induce more people to buy
it.115
The second aspect of the Hartford plan required coastal
homeowners to buy flood insurance.116 The Hartford’s plan would achieve
this by allowing for the alternative policy options of either requiring
certification of flood insurance coverage through current programs such as
the National Flood Insurance Program or by inserting flood coverage as
part of the standard homeowner’s policy.117
Finally, much like the Allstate and Travelers/Nationwide plans, the
Hartford proposal also included a federal reinsurance backstop for existing
state and regional funds.118 The Hartford backstop would be triggered upon
the occurrence of a 1-in-100 year loss.119 Thus, the key difference between
the previous proposed federal reinsurance backstops and the Hartford’s
plan lies in the fact that the latter seems to anticipate the most amount of
government intervention. 120
111
Id.
112
Id. at 6.
113
Hays, supra note 108.
114
Id.
115
U.S. GOV’T ACCOUNTABILITY OFFICE, GAO-08-7 PUBLIC POLICY OPTIONS
FOR CHANGING THE FEDERAL ROLE IN NATURAL CATASTROPHE INSURANCE 35
(2007).
116
Ayer, supra note 110, at 5.
117
Id.
118
Id. at 6.
119
Id.
120
Id.
E. THE HOUSE VERSION OF THE HOMEOWNERS DEFENSE ACT OF

2007: BONDS, LOANS, REINSURANCE
In response to the 2005 hurricane season U.S. House Reps. Tim

Mahoney and Ron Klein co-authored a bill designed “to bring relief to
property owners struggling with the affordability and availability of
homeowners insurance.”121 The resultant Homeowners' Defense Act of
2007 sought to, “assist state-sponsored insurance programs on covering
losses from natural disasters.”122 Essentially, the proposal offered
by the Homeowners' Defense Act of 2007 had three major features.
First, the proposal would essentially securitize coastal risk. This
aspect of the proposal would have established a federally overseen
mechanism that permitted multiple states to join together to help pay for
each others' disaster costs and then transfer those costs to the private
markets through catastrophe bonds and reinsurance contracts.123 The
program would have accomplished the private market risk transfer via the
establishment of a central National Catastrophe Risk Consortium tasked to,
work with states to create an inventory of catastrophe risk obligations held
by state reinsurance funds and issue securities linked to the catastrophe risk
insured in capital markets.124
Second, the proposal would have created a federal reinsurance
program similar to the type championed by both the Allstate and
Travelers/Nationwide plans to backstop existing state insurance pools.125
The actual reinsurance mechanism would be fiscally backed by the Federal
Natural Catastrophe Reinsurance Fund.126 This fund would be directed
towards covering contract payments for losses resulting from the federal
reinsurance program.127 The fund would be financially supported by,
amounts received annually from the sale of reinsurance contracts, amounts
earned on investments, and appropriations.128
121
See Amie Parnes, Bill Seeks to Provide Homeowners Insurance Relief,
NAPLES NEWS, Aug. 3, 2007, available at http://www.naplesnews.com/
news/2007/aug/03/bill_seeks_provide_homeowners_insurance_relief/.
122
Id.
123
See Homeowners’ Defense Act of 2007, H.R. 3355, 110th Cong. §§ 2(b),
101(d), 102 (2007).
124
See id. § 102.
125
See id. §§ 301-303.
126
Id. § 305.
127
Id.
128
H.R. 3355, § 305.
Finally, the proposal would have provided loans to state insurance

programs.129 This aspect of the proposal called for the creation of a
National Homeowners' Insurance Stabilization Program, which would have
provided low-interest federal loans directly to qualified state insurance
programs in order to, “ensure the solvency of such programs, to improve
the availability and affordability of homeowners' insurance, to incent risk
transfer to the private capital and reinsurance markets and to spread the risk
of catastrophic financial loss resulting from natural disasters.”130
F. NAIC’S COMPREHENSIVE NATIONAL CATASTROPHE PLAN:

CATASTROPHE RESERVING, ALL PERILS POLICIES AND
TENTATIVE SUPPORT FOR FEDERAL REINSURANCE
The National Association of Insurance Commissioners’

Catastrophe Insurance Working Group developed its national natural
disaster plan in 2005.131 The NAIC plan addresses natural catastrophe
coastal risk in terms of three main “layers.”132 The first layer corresponds
to private market solutions and calls for a mandatory all-perils residential
policy and insurance company catastrophe reserving.133 Dealing primarily
with state level issues, the second “layer” calls for the establishment of a
uniform system of state catastrophe funds.134 While corresponding to
federal level policy, the third layer tentatively supports further examination
of creating a federally overseen catastrophe reinsurance mechanism.135
Under the all-perils component of the NAIC plan, the idea would
be to create a homeowner insurance policy that would provide coverage
against all types of natural catastrophes.136 The NAIC posits that, private
insurers would provide coverage against all perils in a single standard
homeowner’s policy that would reflect each property owner’s risk of loss
129
See id. § 202.
130
Id. § 201.
131
NAT’L ASS’N OF INS. COMM’RS, NATURAL CATASTROPHE RISK: CREATING
A COMPREHENSIVE NATIONAL PLAN 2 (Version 15a 2009),
http://www.naic.org/documents/committees_c_090615_nat_catastrophe_paper_dra
ft.pdf.
132
See id. at 7.
133
Id. at 7-8.
134
Id. at 9.
135
Id. at 10.
136
NAT’L ASS’N OF INS. COMM’RS, supra note 131, at 8.
due to natural disaster.137 At this point coastal the NAIC plan would not
require coastal residents purchase an all perils policy, however, the
availability of such a policy would be mandatory. 138
Pursuant to the catastrophe reserving aspect of the NAIC plan,
private insurers would be permitted to establish tax-deferred reserves to
pay expenses related to natural coastal disasters in order to further expand
the financial base available for underwriting catastrophe risk.139 The idea
underlying catastrophe reserving is that with an additional means of
building cash reserves insurance companies would be more able and
willing to underwrite policies. At present, tax-exempt reserves would
necessarily require amending the US tax code.140
Within the second layer of the NAIC’s plan, each state would be
asked to decide whether its exposure to coastal natural catastrophes
warrants creating a catastrophe fund or whether the private market has the
capacity to provide adequate coverage without additional funding.141 If
established the state catastrophe funds would be responsible for creating
and managing the insurance capacity of their respective regions.142 The
funds would have the discretion to implement their own operating
structures to best fit their particular needs.143 This would entail defining
catastrophic loss thresholds, determining appropriate private insurers/state
fund retention amounts, and ensuring that premium rates are actuarially
sound.144
The third and final layer of the NAIC plan consists of cautious
support for a federal reinsurance role.145 Although the NAIC acknowledges
that a federal reinsurance program seems to be a potential solution, the
Commission also acknowledges that the debate over whether federal
reinsurance involvement is necessary is still ongoing.146 Beyond this, the
specific character of the federal reinsurance mechanism the NAIC would
support is still not clear.
137
Id.
138
Id. at 9.
139
140
Id.
141
Id. at 9.
142
Id.
143
Id.
144
Id.
145
See id. at 10-11.
146
Id. at 11.
V. ANALYSIS: CAPACITY, PRIVATE MARKET

DISPLACEMENT, CROSS-SUBSIDIZATION
The plans proposed offer a myriad of policy options. Keeping

track of them is a task in itself. One useful way of evaluating these current
coastal risk proposals is to look at how well they deal with certain of
evaluative criteria: private insurance industry capacity, private market
displacement, and cross subsidization.
Capacity means “the ability of an insurance company… to pay
claims in the event of a loss.”147 In this context, capacity specifically refers
to the private insurance industry’s ability to cover the insured costs of
property damage wrought by a coastal storm. With respect to capacity, the
analysis of each plan concerns two issues. First, the extent each plan
suggests the private market presently has adequate capacity to handle
coastal risk. Second, how each plan seeks to expand private market
capacity to handle growing future coastal risk.
Similarly, private market displacement refers to the concern that
federal government activity in coastal natural catastrophe insurance will
hinder the private market actors already involved.148 The focus of this
inquiry is on how each plan seeks to minimize the displacement of the
private coastal risk insurance markets. That is, whether any of the plans
utilize policy options or forms of government intervention that will either
encourage private market participation or, at least, not completely crowd
out private insurers from covering coastal risk.
Finally, in this context cross subsidization refers to the situation
whereby the government covers the exposure property owners who choose
to live in high risk coastal locations take on by spreading the cost of that
risk to taxpayers who do not reside in these areas.149 The issue here is
whether and to what extent each plan contains policy tools that could help
localize the costs of owning property in storm-prone coastal areas.
147
148
See Press Release, Prop. Cas. Insurers of Am., Windstorm Coverage
Widely Available, Federal Involvement Unnecessary (April 21, 2008)
(http://www.pciaa.org (follow “MediaCenter: News Releases: By Date:” then
scroll to “April 21, 2008”)).
149
See CHARLES PERROW, THE NEXT CATASTROPHE: REDUCING OUR
VULNERABILITIES TO NATURAL, INDUSTRIAL, AND TERRORIST DISASTERS 65
(Princeton 2007).
A. CAPACITY
The nation's two largest homeowners’ insurers, State Farm and

Allstate, argue that some natural disasters are too large for the private
insurance market to handle.150 State Farm spokesman Jeff McCollum
argues that “[c]ommercially available insurance is designed for localized
disasters, not ‘mega-catastrophes’.”151 He goes on to note that,
“[i]nsurance was never designed to cover a catastrophe as big or bigger
than Katrina.”152 Similarly, a spokesperson for the Allstate remarked that,
“[o]ur view is that there are some events that have the potential to be so
large as to exceed the capabilities of the insurance industry, as well as the
funding and financing capability of individual states.”153
Similarly, some state insurance commissioners recognize the
capacity issue as a potential stumbling block to completely private market
solutions to coastal risk. Florida Insurance Commissioner Kevin McCarty
told the House Financial Services Committee that in his view, "large
natural catastrophes are a national economic problem, not simply a local
insurance problem.”154 McCarty went on to argue that, “[b]ecause of the
absolute size of the economic losses that are possible due to hurricanes in
Florida, the private market, public mechanisms, and even the states
themselves simply do not have sufficient capacity to provide recovery from
a truly mega-catastrophic hurricane event.”155 As a result, McCarty
concluded that, “Congress and the states need to work together to develop a
comprehensive plan to better manage and mitigate the natural catastrophic
events of tomorrow."156
150
Larry Lipman, Catastrophe Insurance Spawns Storm of Debate, PALM
BEACH POST, April 8, 2007 at A13.
151
Id.
152
Id.
153
Joel Garreau, A Dream Blown Away: Climate Change Already Has a
Chilling Effect on Where Americans Can Build Their Homes, WASH. POST, Dec. 2,
2006 at C01.
154
Lipman, supra note 150.
155
Associated Press, Citizens Property Insurance Increases Upset Policy-
holders in Florida, INSURANCE JOURNAL, http://www.insurancejournal.com
/news/southeast/2006/05/15/68332.htm (last visited Oct. 12, 2009).
156
Id.
1. The Coalition of Americans for Smart Natural

Catastrophe Policy
Not all think private market capacity is such a problem. The

proposal put forth by the Coalition of Americans for Smart Natural
Catastrophe Policy is concentrated almost entirely on policy options aimed
at risk mitigation as opposed to augmenting private market capacity.157 At
base, the Coalition’s position consists of supporting the homeowner
mitigation grants and the gradual phasing out of subsidized National Flood
Insurance Program premiums.158 Implicit in this combination of policy
choices is the assumption that, if implemented, an increased government
role in coastal catastrophe insurance is unnecessary because the private
market has the capacity to cope with such risks.
As a result of this assumption the Coalition’s plan lacks policy
aimed at expanding private market capacity. While policy that encourages
mitigation can be viewed as an indirect means of decreasing coastal risk
relative to market capacity, it does not directly address the capacity issue.
At this point the Coalition’s plan offers no policy innovations aimed at
aiding private insurers increase their own ability to meet the needs of the
growing coastal catastrophe insurance market. This sort of position is
problematic insofar as the underlying assumption that present market
capacity is sufficient is debatable.159 Further still, with global climate
changes in mind market capacity will need to keep pace with coastal
catastrophes of increasing severity. In this way the Coalition’s plan does
not offer an adequate response to the private market capacity issue.
2. Allstate, Travelers’, and the Hartford
In contrast to the position taken in the plan put forth by the

Coalition of Americans for Smart Natural Catastrophe Policy, the stances
taken by Allstate, Travelers’, and the Hartford all suggest significantly less
confidence in private industry capacity to handle coastal catastrophic risk.
Common to each of these three private insurer proposals are federal
reinsurance programs.160 Acknowledgement that a federal reinsurance
backstop is necessary clearly serves as an avowal that the private
157
See supra notes 61-83 and accompanying text.
158
Id.
159
160
See supra notes 104-6, 118-20 and accompanying text.
reinsurance industry does not have the capacity to handle major coastal
catastrophe.
Similarly, the plans put forth by all three of these private industry
leaders signal a perceived lack of private market capacity to cover
catastrophic coastal loss in that all three share common support for a
continued role for state and regional government overseen insurance
pools.161 Although insurance pools essentially spread risk, in part, to
prevent insolvency,162 the history of government organized coastal risk
insurance pools is such that after a catastrophe has struck they are often
funded in part by the public sector.163 In this sense, it is important to note
that each of the private industry backed proposals requires the erection of
some form of a system of government backed insurance pools.
Interestingly, because of their reliance on federal reinsurance
mechanisms, the private industry response to the capacity issue, as seen by
the proposals put forth by Allstate, Travelers’, The Hartford and State
Farm, have less to do with growing private industry capacity than enlisting
the federal government as a reinsurance safety net. Instead of approaches
that would augment private industry ability to cover catastrophic risk, such
as catastrophe reserving, the plans set forth by the three leading insurers
look to the federal government as a potential backstop as a primary means
of shifting the exposure of coastal loss. This approach is undesirable
because over-reliance on federal support implicates creating large liabilities
for the federal government and, as will be discussed shortly, unfairly cross-
subsidizing coastal residents at the expense of non-coastal residents.
3. House Version of the Homeowners Defense Act of

2007
The House version of the Homeowners’ Defense Act of 2007

proposal evidences significant reservation as to the ability of private
markets to handle catastrophic coastal risk. The Act would both create a
161
See supra notes 87-88, 118 and accompanying text.
162
See generally, John Pollner, Catastrophe Risk Management: Using
Alternative Risk Financing and Insurance Pooling Mechanisms 87 (The World
Bank, Working Paper No. WPS2560, 2001), available at
http://www.worldbank.org/reference/ (follow “Documents and Reports” hyperlink;
then search “WPS2560”; then follow “Catastrophe Risk Management…”
hyperlink).
163
Id. at 85-86.
federal reinsurance presence in already established state catastrophe pools

and provide low-interest federal loans directly to existing state reinsurance
programs.164 However, through its National Catastrophe Risk Consortium
program,165 the Act provides a more complicated overall take on the private
market capacity expansion issue.
The Homeowners’ Defense Act is the first of the plans examined to
offer a policy option directly aimed at expanding the private market’s
ability to handle catastrophic coastal risk. The Consortium component of
the Act seeks to transfer coastal risk to the private financial market.166
Through the creation of financial instruments linked to catastrophe risks
insured or reinsured by members of the Consortium, this aspect of the
Homeowners’ Defense Act sought to create and encourage a securities
market linked to coastal risk.167 This approach essentially uses a federally
overseen mechanism to funnel coastal risk from state and regional pools to
private securities markets. 168 In theory, by tapping into the private capital
markets this approach has the potential to drastically increase private sector
capacity to handle coastal risk. As a result, this approach offers an
innovative policy option geared towards expanding private sector capacity.
4. The National Association of Insurance Commissioners
Due to its limited tentative support for some sort of federal

reinsurance mechanism,169 NAIC’s Comprehensive National Catastrophe
Plan is marked by a relatively limited direct federal intervention that
assumes the private markets’ have significant ability to handle coastal
catastrophic storm related loss. Moreover, the NAIC proposal contains a
policy tool in catastrophe reserving that may actually expand the private
market’s capacity to handle coastal risk.
The catastrophe reserving mechanism contained in the NAIC
proposal seeks to expand the ability of private market insurers to
underwrite coastal risk by allowing them to build limited tax-deferred pre-
164
165
166
Id.
167
Id.
168
See Homeowners’ Defense Act of 2007, H.R. 3355, 110th Cong. §102(6)
(2007).
169
catastrophe capital bases.170 This sort of option represents the most direct
means of increasing private market capacity to deal with coastal storms.
B. PRIVATE MARKET DISPLACEMENT
There is a line of reasoning with considerable traction within the

coastal insurance industry that the greatest threat to the property insurance
market, "is not the force of hurricane winds but legislation and regulations
that displace available private capital or make it economically unfeasible
for private companies to operate in coastal markets."171 Similarly, there is a
movement within the industry that would prefer to see a halt to federal
expansion into the coastal catastrophe market.172 Frank Nutter, president of
the Reinsurance Association of America argues that expanding the current
federal role "would displace a vibrant [private] reinsurance market to the
detriment and cost of the U.S. taxpayers.”173

Catastrophe Policy
Clearly, the mitigation based proposal put forth by the Coalition of

Americans for Smart Natural Catastrophe Policy would require the least
amount of private market displacement. Perhaps overestimating the
capacity of the private market, the Coalition proposal requires no federal
reinsurance program and does not require any other new federal
intervention beyond modest homeowner mitigation grants and loans.174 In
this way, the Coalition plan is at the far end of the spectrum occupied by
proposals that would cause little private market displacement.
170
See supra note 139 and accompanying text.
171
Larry Lipman, Catastrophe Insurance Spawns Storm of Debate, PALM
BEACH POST, Apr. 8, 2007, at A13.
172
See id.
173
Press Release, Reinsurance Ass’n of Am., RAA Testifies Before Senate
Banking, Housing, and Urban Affairs Committee (Apr. 11, 2007),
www.reinsurance.org/files/public/pr41107.pdf.174 See supra notes 61-83 and
accompanying text.
174
In contrast to the Coalition plan, the proposals put forth by Allstate,

Travelers’, and the Hartford would all anticipate significant private market
displacement. All three proposals require a federal reinsurance mechanism
specifically designed to support state insurance pools.175 The displacement
implications for this type of federal reinsurance mechanism are clear. If
even eligible, private reinsurers would be forced to compete with a federal
reinsurance program in providing reinsurance to the pools.176 This scenario
has the potential to cause significant private market disruption.
However, this reinsurance displacement should be viewed in
broader context. A federal reinsurance option could lead to greater
participation from private primary insurers.177 In this sense, any
displacement of private reinsurers may be offset or even surpassed by
greater primary insurer willingness to underwrite coastal risk.
3. Homeowners Defense Act of 2007
The response to the private market displacement question offered

by the proposal embodied in the House version of the Homeowners
Defense Act of 2007 is unique. Although the plan advocates government-
run insurance pools and a federal reinsurance fund,178 the role of the
Consortium in transferring risk to the private securities market complicates
the private market displacement analysis. In the narrow sense, the federal
reinsurance fund would certainly displace private reinsurers.179 However,
the net effect of the Consortium’s risk transferring scheme may potentially
offset some of the private market reinsurance displacement by facilitating
risk transfer to the private securities market.
The NAIC program, consisting of private industry catastrophe

reserving and mandatory offer of an all-perils policy180 suggests expanding
175
See supra notes 87-89, 118 and accompanying text.
176
U.S. GOV’T ACCOUNTABILITY OFFICE, supra note 115, at 34-35.
177
Id.
178
179
U.S. GOV’T ACCOUNTABILITY OFFICE, supra note 115, at 33.
180
See supra notes 136-39.
the role of private insurers as opposed to displacing them. The catastrophe

reserving policy option would potentially increase private insurer capacity
to handle large coastal risk and thereby increase its ability to underwrite
risk.181 In this way, catastrophe reserving may prevent private market
displacement by fortifying its ability to handle coastal risk. Similarly, the
all-perils portion of the NAIC plan explicitly seeks to have private insurers
provide comprehensive storm coverage in a single standard homeowners’
policy.182 Here, the effort to expand private activity in the context of
coastal risk is quite clear. An all-perils policy program such as the one
endorsed by the NAIC may have the opposite impact of actually further
entrenching the private market in fielding coastal risk.183
C. SUBSIDIZING COASTAL PROPERTY OWNERS
Federal intervention in coastal risk could result in significant public

subsidies to property owners who make the choice to construct or maintain
a home or business in a coastal area. The current National Flood Insurance
Program is a prime example of a program that facilitates this dynamic.184
Ultimately the subsidized rates that many coastal property owners pay are
made possible by tax dollars of many non-coastal residents.185 The term
that describes this occurrence is “cross-subsidization.”

Catastrophe Policy
As for current proposals that attempt to minimize such

subsidization, the Coalition of Americans for Smart Natural Catastrophe
Policy’s mitigation based plan confronts the issue in blunt terms. In effect,
the Coalition’s platform is built on the fear that if its policyholders are not
adopted citizens in regions who face little coastal storm risk, they would be
181
NAT’L ASS’N OF INS. COMM’RS, supra note 131, at 7; U.S. GOV’T
ACCOUNTABILITY OFFICE, supra note 111, at 34-35.
182
See NAT’L ASS’N OF INS. COMM’RS, supra note 131, at 8.
183
See U.S. GOV’T ACCOUNTABILITY OFFICE, supra note 115, at 34-35.
184
See The Nat’l Flood Ins. and Repetitive Loss Properties: Hearing Before the
Subcomm. on Housing and Cmty. Opportunity of the H. Comm. on Fin. Serv.,107th
Cong. 3 (2001) (statement of Rep. Bereuter, Member House Subcomm. on
Housing and Cmty. Opportunity).
185
See id. at 3-4.
subsidizing residents of areas where hurricanes are a constant threat.186

The Coalition urges no federal involvement in insuring coastal risk beyond
the existing National Flood Insurance Program, and even then would have
the subsidized rates that characterize much of the program phased out
within a matter of years.187 In addition, the Coalition successfully fought
adding windstorm coverage to the National Flood Insurance Program
because it would have the impact of transferring more coastal risk from
coastal property owners to non-coastal homeowners by way of the federal
government.188 For these reasons, the Coalition’s plan deals most
effectively with curbing cross-subsidization.
At the other end of the spectrum the Allstate,

Travelers/Nationwide, and Hartford plans would all have non-coastal
residents subsidize coastal property owners. The federal reinsurance
program that is common to three plans would likely rely on post-funding
appropriations in the event of a coastal catastrophe.189 In the context of a
federal reinsurance mechanism, post-funding appropriations would likely
be drawn from federal tax dollars and would therefore represent a
significant form of cross-subsidies.190
In theory, a federal reinsurance mechanism does not necessarily
have to lead to federal subsidies. If the proposed federal reinsurance
programs were operated according to actuarially sound principles it should
be self-sufficient and not require tax derived post-funding in the event of a
catastrophe. However, given the history of programs such as the National
Flood Insurance Program it is difficult to see a government devised
insurance scheme not relying on post-funding measures.191 If the National
Flood Insurance Program is any indication, a federal reinsurance
mechanism will not be self-sufficient and the program will likely rely on
the liberal use of cross-subsidized post-event appropriations.192 For this
186
See Smarter Safer, Statement of Principles, http://www.smartersafer.org/
about-us/statement-of-principles (last visited Oct. 8, 2009).
187
Id.; see supra notes 62-72 and accompanying text.
188
189
The history of other government managed coastal risk programs supports
this assertion. See supra notes 45, 57-58 and accompanying text.
190
See, e.g., supra note 45 and accompanying text.
191
192
reason the Allstate/Nationwide, Travelers’, and Hartford plans do not

successfully minimize cross-subsidization.
3. Homeowners Defense Act of 2007
On balance the three prongs of the program offered in the

Homeowners' Defense Act of 2007 favor significant cross-subsidization.
Clearly, the plan’s federal reinsurance backstop and related reinsurance
fund could lead to non-coastal taxpayers subsidizing their coastal
counterparts.193 Similarly, the federal stabilization program designed to
provide low-interest federal loans directly to qualified state reinsurance
programs would also lead to government facilitated cross-subsidization.194
However, on the other side of the scale, the risk transferring National
Catastrophe Risk Consortium has the potential to offset some of the other
two programs insofar as it can tap into private sector capacity to take on
coastal risk and thereby eliminate some need for utilization of the other two
programs.195 Since the potential success of the Consortium is difficult to
gauge it is ultimately hard to tell how effective it will be in curbing the
need to utilize the proposals other cross-subsidy funded programs.
The NAIC proposal contains policy measures that localize the risk
of catastrophic coastal storms, but also some that may lead to cross-
subsidization of coastal risk. By withholding unreserved support for a
federal reinsurance system, the NAIC plan, in its present form would
essentially only allow for catastrophe reserving and all-perils homeowners’
policies to be provided by private companies.196 These two remaining
aspects of the NAIC proposal would have very different impacts in terms
of cross-subsidization. While catastrophe reserving would allow private
insurers to build up capital reserves to better deal with future catastrophic
events, it would not directly involve having noncoastal tax-payers subsidize
coastal residents. However, the all-perils homeowners’ policy aspect of the
NAIC program may indirectly lead to moderate cross-subsidization. While
193
See supra notes 125-28,189-92 and accompanying text.
194
See KING, supra note 44 at 5; see also supra notes 129-31 and
accompanying text.
195
196
the NAIC would require these policies be written through private insurers,
the all-perils option could lead to government subsidies for low-income
property owners to afford such coverage.197
VI. CONCLUSION
Some estimate that the destructive potential of tropical storms in

the North Atlantic has increased dramatically since the 1970s.198 Alarming
enough on its own, this weather trend coincides with a time in which many
parts of the coastal U.S. are becoming more densely populated than ever.199
Add to this an average annual growth rate that will lead to the doubling of
the total covered coastal value every decade,200 and it becomes clear that
the coastal catastrophe policy framework needs reevaluation.
Even before the wakeup call of the monumental 2005 hurricane
season policymakers and private industry leaders were beginning to rethink
the coastal risk landscape. The proposals examined in this paper represent
a fairly wide range of thought on the federal role in insuring coastal risk.
Evaluated by way of the capacity, private market displacement, cross
subsidization criteria laid out here, none of the plans are perfect. Instead,
each proposal is flawed while a few exhibit promising aspects.
In advocating systems of federally reinsured state run catastrophe
pools, the proposals made by three of the largest property and casualty
insurance companies, could result in a considerable expansion of federal
involvement. Federal reinsurance would not directly expand private
market capacity, but would instead shift coastal risk exposure to the federal
government. Similarly, this expansion would come at the cost of private
market displacement and significant cross-subsidization.
In contrast, the plan put forth by the Coalition of Americans for
Smart Natural Catastrophe Policy is characterized almost exclusively by
risk mitigation efforts and would seemingly halt any additional federal
involvement in managing coastal risk. While the absence of a federal
reinsurance scheme is laudable insofar as it would curb private market
displacement and significant cross-subsidization, the Coalition approach is
197
U.S. GOV’T ACCOUNTABILITY OFFICE, supra note 115, at 34-35.
198
See Kerry Emanuel, Increasing Destructiveness of Tropical Cyclones Over
the Past 30 Years, 436 NATURE 686 (2005).
199
See supra notes 9-22 and accompanying text
200
See supra note 24 and accompanying text.
problematic because places unwarranted confidence in the capacity of the

private market and existing coastal risk framework to handle future storms.
Containing measures aimed at expanding private insurer capacity
to take on coastal risk and only reserved support for a federal reinsurance
mechanism, the proposal made by the National Association of Insurance
Commissioners offers a possible approach with promise. The NAIC plan’s
support of insurance company catastrophe reserving may potentially be
effective in increasing private market capacity to handle coastal catastrophe
risk. This capacity expanding upside might offset the private market
displacement and cross-subsidization the NAIC plan’s potential reinsurance
mechanism would cause.
Finally, the House version of the Homeowners’ Defense Act of
2007 would have created an innovative mechanism to transfer coastal risk
to private capital markets, but would also have provided for federal loans to
state insurance pools and erected a federal reinsurance backstop. The
innovative capacity expansion upside that the Defense Act’s risk
transferring mechanism provides is undercut by the amount of cross-
subsidization and private market displacement the proposal’s federal
reinsurance and loan programs could lead to.
In the months ahead other proposals will be made and eventually a
comprehensive program will be settled upon. It was the goal of this modest
examination to add to the discussion that will lead to an ultimate decision
by examining and comparing some of the high profile options with
particular attention to certain concerns: private market capacity to field
catastrophic coastal risk, continuing private market role in coastal
catastrophic risk, and cross-subsidization of coastal risk.
CONNECTICUT INSURANCE
LAW JOURNAL
The Connecticut Insurance Law Journal is published at least twice per year.
Domestic subscriptions are $28.00 per year, international subscriptions are
$32.00 per year. Back issues and back volumes are available (for any
volume up to and including Volume 12) for $15.00 per issue or $25.00 per
volume; international orders add $3.00 per issue or $5.00 per volume.
Absent timely notice of termination, subscriptions are renewed

automatically. To subscribe to the Insurance Journal or order back issues
online, please visit our website located at www.insurancejournal.org, or
you may fill out the form below and mail to:
Pat Carbray,
Connecticut Insurance Law Journal,
65 Elizabeth Street,
Hartford, CT 06105;
[email protected].
Please Print below and complete reverse:
Contact Name: _______________________________________
Contact Title: ________________________________________
Organization: ________________________________________
Department: _________________________________________
Address: ____________________________________________
____________________________________________________
City: _________________ State: _____ Zip Code: ________
Phone: __________________ Fax: ______________________
E-Mail: ____________________________________________
Additional Comments: ________________________________

Check all that apply:
I would like to order (number) __________ subscriptions at $28.00

per subscription ($32.00 for international orders).
I would like to order (number) __________ complete set(s) of back

issues (Volumes 1-12) at $150.00 per set ($165.00 for international
orders).
I would like to order the following back volumes or back issues at

$15.00 per issue, $25.00 per volume (international orders are
$18.00 per issue and $30.00 per volume). Please list volume, issue
and quantity for each.
Examples: 4 copies, Volume 3, Issue 2

2 copies - Volume 3
_________________________________________________________
_________________________________________________________
_________________________________________________________
_________________________________________________________

302394256.connecticut Insurance Journal - Good

Uploaded by

Copyright:

Available Formats

302394256.connecticut Insurance Journal - Good

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

302394256.connecticut Insurance Journal - Good

Uploaded by

Copyright:

Available Formats

University of Connecticut

Connecticut Insurance Law Journal School of Law

Whither the Duty of Good Faith in UK Insurance Contracts?

Follow this and additional works at: https://opencommons.uconn.edu/cilj

Volume 16, Number 1

University of Connecticut School of Law

Postmaster: Send address changes to Connecticut Insurance Law Journal, 55 Elizabeth

It is the policy of the University of Connecticut to prohibit discrimination in education,

Copyright © 2009 by the Connecticut Insurance Law Journal Association.

Cite as CONN. INS. L.J.

Editor-in-Chief Managing Editor

Administrative Editor Assistant Managing Editor

Lead Articles Editors Notes & Comments Editors Executive Editors

Legal Abstracts Editor Research Editor Associate Editors

ADAM J. ALLEGRO NATHANIEL GENTILE SHETERA PATTERSON

FACULTY AND OFFICERS OF ADMINISTRATION

ADJUNCT FACULTY OF LAW

CREDIT DERIVATIVES ARE NOT “INSURANCE” M. Todd Henderson 1

THE ROAD FROM “TWIN PEAKS” –

WHITHER THE DUTY OF GOOD FAITH

RISK DATA IN INSURANCE INTERPRETATION Michelle Boardman 157

THE LAW AND ECONOMICS OF FIRST-PARTY Sharon Tennyson

REGULATION OF LARGE FINANCIAL

NOTES AND COMMENTARIES

PREDATORY LENDING AND ITS

EXAMINING CURRENT PROPOSALS FOR

This article explores whether credit derivatives should be regulated as

chosen. This Essay explores the suitability of insurance regulation to the

This Essay argues that it makes little or no sense to regulate credit

encourage more comprehensive federal regulation. (“Based on the developments

be made more bluntly: it would be fanciful to argue that every contract in

adopting a shareholder-friendly, go-to-Vegas strategy in the event liability

II. A PRIMER ON CREDIT DERIVATIVES

origination, servicing, monitoring, and funding or risk bearing. In a world

A. CREDIT DEFAULT SWAPS

A CDS is a contract in which credit risk (that is, the expected

These contracts constitute the primary market for credit derivatives

allegedly insured actually suffers a harm unrelated to the insurance

B. COLLATERALIZED DEBT OBLIGATIONS

The other type of basic credit derivative is a collateralized debt

a 20 percent loss. In a securitized debt investment, by contrast, the same

Even when considering the role of tranching and securitization, the

exposure to a portfolio of credit derivative investments. For instance, in the

III. THE ARGUMENT FOR REGULATING CREDIT

At first blush, the similarity between property (or other) insurance

because deeming something “insurance” brings with it a special set of laws

function does not mean that it is sensible or efficient to regulate them as

default swaps.”31 The fact that Commissioner Dinallo’s testimony outlined

IV. THE ARGUMENT AGAINST REGULATING CREDIT

This Part presents several conceptual and practical reasons why

arise in extending the reach of insurance regulation to all contracts

A. PROVIDING INSURANCE IS NOT ENOUGH

It could not be enough for a contract that insures against risk to be

There are innumerable contracts that provide the same type of

through an options exchange, such as the Chicago Mercantile Exchange. These

2. Credit Derivatives Are Not Just About Reducing Risk

Another problem with regulating credit derivatives as insurance is

The first set of regulations – capital adequacy rules in the Basle

3. The Pooling Mismatch