Article 11 PDF

DATE DOWNLOADED: Tue Nov 15 21:07:56 2022
SOURCE: Content Downloaded from HeinOnline
Citations:
Bluebook 21st ed.

Mark L. Mitchell & Jeffry M. Netter, The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and Exchange Commission, 49 Bus. LAW. 545
(1993).
ALWD 7th ed.

Mark L. Mitchell & Jeffry M. Netter, The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and Exchange Commission, 49 Bus. Law. 545
(1993).
APA 7th ed.

Mitchell, M. L., & Netter, J. M. (1993). The role of financial economics in
securities fraud cases: applications at the securities and exchange commission.
Business Lawyer (ABA), 49(2), 545-590.
Chicago 17th ed.

Mark L. Mitchell; Jeffry M. Netter, "The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and Exchange Commission," Business Lawyer
(ABA) 49, no. 2 (1993-1994): 545-590
McGill Guide 9th ed.

Mark L. Mitchell & Jeffry M. Netter, "The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and Exchange Commission" (1993) 49:2 Bus
Law 545.
AGLC 4th ed.

Mark L. Mitchell and Jeffry M. Netter, 'The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and Exchange Commission' (1993) 49(2)
Business Lawyer (ABA) 545
MLA 9th ed.

Mitchell, Mark L., and Jeffry M. Netter. "The Role of Financial Economics in
Securities Fraud Cases: Applications at the Securities and Exchange Commission."
Business Lawyer (ABA), vol. 49, no. 2, 1993-1994, pp. 545-590. HeinOnline.
OSCOLA 4th ed.

Mark L. Mitchell & Jeffry M. Netter, 'The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and Exchange Commission' (1993) 49 Bus
Law 545
-- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and
Conditions of the license agreement available at
https://heinonline.org/HOL/License
-- The search text of this PDF is generated from uncorrected OCR text.
-- To obtain permission to use this article beyond the scope of your license, please use:
Copyright Information
The Role of Financial Economics in Securities
Fraud Cases: Applications at the Securities and
Exchange Commission
By Mark L. Mitchell and Jeffry M. Netter*
INTRODUCTION
Litigants, including the Securities and Exchange Commission (SEC),
increasingly have applied modem financial economics in securities fraud
cases. One of the most important applications of financial economics for
securities law comes from the efficient markets hypothesis. This Article
presents an overview of areas where securities fraud law has adopted some
of the reasonings and applications of the efficient markets hypothesis and
provides examples of the use of financial economics in SEC enforcement
actions. Specifically, this Article discusses how techniques developed by
financial economists can be used to establish the materiality of information
allegedly used in securities fraud, and to compute profits (or losses avoided)
resulting from fraudulent actions. It then shows how the methodology was
applied in recent SEC enforcement cases.
A leading expert on the efficient markets hypothesis, Professor Eugene
F. Fama of the University of Chicago's Graduate School of Business, re-
cently reviewed the empirical evidence on the efficient markets hypothesis
and defined market efficiency with the simple statement that security prices
fully reflect all available information.1 Fama noted that, while no market
is perfectly efficient, the idea that prices quickly adjust to the release of
new information is a useful tool to analyze many situations, especially when
information and transactions costs are low, as in the United States stock
market.
An event study, a technique developed and refined by financial econ-
omists, can be very useful in securities fraud cases. An event study relates
changes in stock prices to the release of new information. Researchers
have applied event studies to all types of events ranging from mergers to
regulatory actions. In securities fraud law, event studies are particularly
*Both authors formerly worked at the United States Securities and Exchange Commission.
The views expressed here are those of the authors and do not necessarily reflect the views
of the Commission.
1. Eugene F. Fama, Efficient Capital Markets: 11, 46 J. FINANCE 1575 (1991).
546 The Business Lawyer; Vol. 49, February 1994
beneficial because they allow the investigator to discern whether infor-

mation that is used in an allegedly fraudulent action is important to inves-
tors and to determine the value of the information. This Article illustrates
the application of this event study methodology in SEC enforcement cases.
The application, however, can be used easily in private suits as well.
While financial economics is becoming increasingly important in secu-
rities fraud law, both indirectly through the courts' approach to certain
issues and directly through its application in specific cases, many litigants
have not been exposed to this application or to a detailed description of
the methodology. Recognition of the important relationship between eco-
nomics and the law by the legal community, however, is not a new devel-
opment. In 1897,Justice Oliver Wendall Holmes, Jr. said: "For the rational
study of the law the black-letter man may be the man of the present, but
' 2
the man of the future is the man of statistics and the master of economics.
Over the intervening years, the legal profession adopted many economic
theories and reasonings. The application of financial economics in litiga-
tion has lagged behind other fields in economics largely because many of
the accepted theories in finance are relatively new.' Today, the courts are
more receptive to the use of financial economic analysis. For instance, in
1988 when the United States Supreme Court adopted the fraud-on-the-
market theory in Basic, Inc. v. Levinson, 4 it stated: "Recent empirical studies
have tended to confirm Congress' premise that the market price of shares
traded on well-developed markets reflects all publicly available informa-
tion, and, hence, any material misrepresentations. ' 5 This type of reasoning
suggests the manner in which financial economics is used in securities fraud
litigation and will be analyzed in this Article. While litigants in private and
public suits have used financial economics in connection with a variety of
securities issues, this Article will focus on the recent application in SEC
enforcement cases.
THE BASIS FOR FINANCIAL ECONOMICS ANALYSIS IN

SECURITIES FRAUD CASES
In an influential article published in The Business Lawyer in 1982, Pro-
fessor Daniel R. Fischel argued that in a rule 1Ob-5 suit the court should
determine whether the stock price was "artificially affected by false infor-
mation" instead of separate determinations into materiality, reliance, caus-
2. Oliver Wendall Holmes, Jr., The Path of the Law, in COLLECTED LEGAL PAPERS 167, 187
(1920).
3. For example, it was not until 1990 that the Nobel Prize Committee recognized finance
as a legitimate scientific area of study within the field of economics and awarded the Nobel
Prize in Economics to three researchers, Harry Markowitz, Merton Miller and William Sharpe,
for their seminal contributions to the field of finance.
4. 485 U.S. 224 (1988).
5. Id. at 246.
Securities Fraud: Financial Economics 547
ation, and damages. 6 While courts have not followed Fischel's suggestion
unanimously, they have relied directly on the efficient markets hypothesis
in recent years with the use of evidence provided by expert testimony of
financial economists and indirectly through the acceptance of many of the
implications flowing from the hypothesis.
FRAUD-ON-THE-MARKET THEORY
The seminal adoption of financial economics in securities fraud litigation
is in the fraud-on-the-market theory, which enables a plaintiff who has not
actually seen a misleading statement to satisfy, nevertheless, the reliance
requirement in a fraud suit. Fischel noted that the fraud-on-the-market 7
theory originated to ease the proof of reliance in large class action suits.
A derivative of the efficient markets hypothesis, the fraud-on-the-market
theory assumes that investors rely on the market price of a security as a
reflection of its value. 8 Thus, a misleading statement that distorts securities
prices is fraudulent even if the average securityholder has no knowledge
of the statement. 9 Applying the fraud-on-the-market theory, a court can
presume the plaintiffs relied on the integrity of the market price for the
securities they bought or sold, therefore dispensing with the traditional
reliance requirement that the plaintiff relied on the fraudulent statement
in making his or her investment decisions)10
The acceptance of the fraud-on-the-market theory varied among lower
courts until Basic, Inc. v. Levinson. IIn this case, corporate officers of Basic,
Inc. falsely denied the existence of on-going merger negotiations with
Combustion Engineering during 1977-1978. In fact, "[n]ot only was Basic
[Inc.] involved in negotiations, but on December 20, 1978, Basic [Inc.]
announced that its Board of Directors had approved a tender offer by
Combustion Engineering."'12 Stockholders, who sold stock after officers in
Basic, Inc. first denied merger negotiations and before the merger an-
nouncement, sued claiming a violation of rule 10b-5.13 The United States
Supreme Court held that plaintiffs may use the fraud-on-the-market theory
to presume reliance so long as plaintiffs can show the affected shares traded
4
in an "efficient" market.'
6. Daniel R. Fischel, Use of Modern FinanceTheory in Securities FraudCases Involving Actively

Traded Securities, 38 Bus. LAW. 1, 13 (1982).
7. Id. at 9.
8. See generally Jonathan R. Macey et al., Lessons From Financial Economics: Materiality
Reliance, and Extending the Reach of Basic v. Levinson, 77 VA. L. REV. 1017, 1020 (1991).
9. Id.
10. Id.
11. 485 U.S. 224 (1988).
12. Macey, supra note 8, at 1019.
13. Basic, 485 U.S. at 228.
14. Id. at 245. Inquiry into whether a stock trades in an efficient market is unnecessary.
Instead, courts should address whether a misstatement caused the stock to trade at an ar-
tificially low or high price. Macey, supra note 8, at 1021.
The fraud-on-the-market theory is useful to those attempting to satisfy

the reliance requirement in private rule 10b-5 suits such as Basic. The
acceptance of the theory by lower courts and the United States Supreme
Court, however, also provided an intellectual basis for the application of
financial economics in other contexts such as the SEC's use of financial
economics in its enforcement actions. The two primary elements of rule
1Ob-5 cases that directly relate to SEC securities fraud cases are materiality
and disgorgement.
MATERIALITY
A key element of a rule 1 Ob-5 case is proof that the fraudulent or inside
information is material.' 5 The plaintiff must establish the importance of
information provided in the fraudulent statements or of information ex-
ploited in insider trading.
Standards for Materiality

The materiality requirement originated in the common law of fraud,
which is the basis for the courts' general interpretation of rule 1Ob-5.
According to the United States Supreme Court, "materiality may be char-
acterized as a mixed question of law and fact, involving as it does the
application of a legal standard to a particular set of facts."' 6 Arnold S.
Jacobs suggested there are three interrelated groups of materiality stan-
dards that courts use-reasonable investor, probability/magnitude, and
market impact.' 7
The United States Supreme Court in Basic described the reasonable
investor approach as "[a]n omitted fact is material if there is a substantial
likelihood that a reasonable shareholder would consider it important."' 8
The Court expressly adopted the reasonable investor standard of mate-
riality for all section 10(b) and rule 10b-5 cases.' 9 The probability/mag-
nitude approach concerns events that do not always take place. For ex-
ample, in Basic the SEC urged the Court to treat preliminary merger
negotiations as material under this probability/magnitude standard.20 The
SEC argued that while a merger might not actually result from negotia-
tions, the possibility of the merger occurring combined with the potential
large market impact indicates the materiality of information regarding
15. See Donald C. Langevoort, Investment Analysts and the Law of Insider Trading, 76 VA.
L. REV. 1023 (1990), for an extended discussion of materiality in SEC insider trading cases.
16. TSC Indus., Inc. v. Northway, Inc., 426 U.S. 438, 450 (1976).
17. ARNOLD S. JACOBS, LITIGATION AND PRACTICE UNDER RULE 1Ob-5 § 61.02 [b] [ii] (1993).
18. Basic, Inc. v. Levinson, 485 U.S. 224, 231 (1988) (quoting TSC Indus., 426 U.S. at
449).
19. See id. at 230.
20. Id. at 239 n.16.
preliminary merger negotiations. 21 The third standard, market impact, de-

fines materiality in terms of whether the relevant information, if released,
would have had an impact on the price of the affected securities. 22 The
United States Court of Appeals for the Second Circuit, in the influential
SEC v. Texas Gulf Sulphur Co. case,23 defined materiality under the market
impact approach as "those situations which are essentially extraordinary
in nature and which are reasonably certain to have a substantial effect on
' '24
the market price of the security.
These three standards, however, are certainly not mutually exclusive;
some courts applied more than one standard in the same case. For example,
in adopting the reasonable investor standard in Basic, the Court suggested
the probability/magnitude test could be used to analyze the importance 25
a reasonable investor would attach to preliminary merger negotiations.
Similarly, in Texas Gulf Sulphur, the Second Circuit employed all three 26
standards to determine the materiality of news about an ore discovery.
Most importantly for financial economics, courts sometimes applied a mar-
ket impact test to determine whether information met the reasonable inves-
27
tor standard.
Materiality and Financial Economics

Information allegedly used in fraudulent activity that is important
enough to affect security prices when publicly released provides compelling
evidence that a reasonable investor would consider the information im-
portant in making an investment decision. In SEC v. Tome,28 for example,
the United States District Court for the Southern District of New York,
using the reasonable investor standard to determine materiality in an in-
sider trading case, stated: "The significance of this information to investors
is highlighted by the temporary halting of trading in St. Joe securities and
by the virtually immediate jump in price of St. Joe stock from approxi-
mately $30 per share to approximately $45 per share when the Seagram
tender offer was publicly announced .... ,29
Historically, however, evidence that information is important enough
to impact a security price has not been a necessary condition for substan-
tiating materiality. Plaintiffs often have not used security returns to satisfy
21. Id.
22. JAcoBs, supra note 17, § 61.02[b][ii].
23. 401 F.2d 833 (2d Cir. 1968), cert. denied, 394 U.S. 976 (1969).
24. Id. at 848 (quoting Arthur Fleischer Jr., Securities Trading and Corporate Information
Practices: The Implications of the Texas Gulf Sulphur Proceeding, 51 VA. L. REV. 1271, 1289
(1965)).
25. Basic, Inc. v. Levinson, 485 U.S. 224, 238-39, 250 (1988).
26. Texas Gulf Sulphur, 401 F.2d at 849-50.
27. See, e.g., JACOBS, supra note 17, § 61.02[b] [ii].
28. 638 F. Supp. 596 (S.D.N.Y. 1986).
29. Id. at 623.
the materiality requirement without examining security returns associated

with the release of the information."
The SEC recently began to use stock price evidence to show materiality
in securities fraud cases, especially insider trading cases. The SEC applied
financial economics using evidence derived from economic theory and
prior empirical research in addition to information from the actual case.
First, economic theory and prior empirical research suggest the expected
stock price reaction to the release of information for the event in question.
For example, if the case involves suspected insider trading prior to an
earnings announcement, a determination is made as to the expected stock
price movement based on evidence accumulated from prior similar earn-
ings announcements. The second component in using financial economics
to evaluate the materiality of information is an examination of the stock
price movement associated with the release of information for the case in
question. A price change consistent with theory and prior evidence bolsters
the establishment of materiality; and the larger the price movement, the
more likely the information is material. Financial economics analysis is
especially applicable for cases where the stock-price movement attributable
to the release of information is relatively small, or where the stock exhibits
high volatility.
DISGORGEMENT
The other potential application of financial economics in SEC enforce-
ment actions involves disgorgement calculations. 3' Disgorgement requires
defendants to " 'give up the amount by which [they were] unjustly en-
riched.' " Thus, disgorgement depends on the profits the defendant made
from his or her fraudulent conduct, not the victims' losses. Additionally,
the Insider Trading Sanctions Act of 198433 allows for punitive penalties
up to three times the amount of disgorgement. 4 The application of fi-
nancial economics is especially relevant because the implication drawn
30. See, e.g., JACOBS, supra note 17, § 61.04.

31. For additional discussions of disgorgement in SEC cases, see MICHAEL J. KAUFMAN,
SECURITIES LITIGATION: DAMAGES (1992 & Supp. 1993) and Thomas C. Mira, The Measure of
Disgorgement in SEC Enforcement Actions Against Inside Traders Under Rule lOb-5, 34 CATH.
U.L. REv. 445 (1985).
32. SEC v. Tome, 833 F.2d 1086, 1096 (2d Cir. 1987) (quoting SEC v. Commonwealth
Chem. Sec. Inc., 574 F.2d 90, 102 (2d Cir. 1978)), cert. denied sub norn. Lombardfin S.p.A.
v. SEC, 486 U.S. 1014 (1988).
33. Pub. L. No. 98-376, 98 Stat. 1264 (codified in scattered sections of 15 U.S.C.).
34. 15 U.S.C. § 78u-l(a)(2) (1988). For cases that were influential in establishing the
method of calculating disgorgement damages, see infra text accompanying notes 35-84. This
Article does not address specifically the issue of calculating damages in private actions. The
methodology presented in this Article, however, can be applied easily to such suits. See
generally Bradford Cornell & R. Gregory Morgan, Using Finance Theory to Measure Damages
in Fraud on the Market, 37 UCLA L. REv. 883 (1990) (discussing Basic and the out-of-pocket
measure).
from the efficient markets hypothesis that security prices react quickly to
the release of new information reduces the subjectivity in estimating profits
from fraud.
Securities Fraud Disgorgement Cases

The courts first expanded the SEC's remedies for insider trading vio-
lations of rule 1Ob-5 beyond injunctions to include disgorgement in the
Texas Gulf Sulphur cases during the late 1960s and early 1970s. 35 In SEC
v. Texas Gulf Sulphur Co.,6 the defendants, officers and employees of Texas
Gulf Sulphur, purchased Texas Gulf Sulphur stock and call options a few
days before the public announcement of a major ore discovery on April
16, 1964. The district court accepted the SEC's argument that the de-
fendants be required to disgorge profits based on the difference between
the price they paid for the stock and the closing stock price the day after
the public announcement of the ore strike.3 7 The court noted that by the
day after the public announcement, "[t]he news was widely disseminated
by the news media and was available to the investing public."38 In sub-
sequent private suits arising from the insider trading, the court required
disgorgement based on the average of the highest prices for each of the
twenty days after the public announcement.3 9 In all cases, most of the
defendants sold shortly after the announcement, so there was little dif-
40
ference between paper and actual profits.
In the second major SEC disgorgement case, SEC v. Shapiro,41 the de-
fendant, Norman Berman worked for a merger boutique. In January and
February 1971, Berman purchased 1100 shares of Harvey's Stores stock
at prices ranging from $7.25 to $23.75 while negotiating a merger between
Harvey's Stores and Ridge Manor. On February 18, 1971, Harvey's an-
nounced an agreement in principle to merge with Ridge Manor-its stock
price closed at $24. A few days later, the proposed merger unravelled and
Harvey's stock price declined. Berman subsequently sold his stock at prices
ranging from $21 to $22 a share.
The district court held that Berman violated the insider trading laws by
exploiting material, nonpublic information regarding the proposed
35. Several cases arose as a result of insider trading in Texas Gulf Sulphur stock. The
most famous case is SEC v. Texas Gulf Sulphur, Co., 401 F.2d 833 (2d Cir. 1968), cert. denied,
394 U.S. 976 (1969).
36. 312 F. Supp. 77 (S.D.N.Y. 1970), aff'd in part, rev'd inpart, 446 F.2d 1301 (2d Cir.),
cert. denied, 404 U.S. 1005 (1971).
37. Id. at 93-94.
38. Id. at 93.
39. Reynolds v. Texas Gulf Sulphur, 309 F. Supp. 548, 563 (D. Utah 1970), aff'd inpart,
rev'd inpart sub nom. Mitchell v. Texas Gulf Sulphur, 446 F.2d 90 (10th Cir.), cert. denied,
404 U.S. 1004 (1971).
40. Id. at 558-62.
41. 494 F.2d 1301 (2d Cir. 1974).
merger. 42 The court required Berman to disgorge his paper profits based
on the average price, $23.80, of all transactions on February 18, the date
of the public announcement of the tentative merger agreement, on the
basis that he could have sold his stock at that price. 43 The Second Circuit
explained that "[o] nce public disclosure is made and all investors are trad-
ing on an equal footing, the violator should take the risks of the market
44
himself."
While the courts applied paper profits in the Texas Gulf Sulphur cases
and Shapiro, the SEC argued in favor of actual profits in a subsequent,
influential insider trading case-SEC v. MacDonald.45 The defendant, James
E. MacDonald, purchased stock in Realty Income Trust (RIT) based on
information regarding a pending acquisition and lease agreement.
MacDonald, chairman of the board of trustees of RIT, learned of the
agreement on December 15, 1975 at a board meeting and then purchased
100 shares at $4.25 the following day4 6 and 9500 shares at $4.625 on
December 23. On December 24, RIT issued a press release detailing the
acquisition and lease agreement, disseminated by Dow Jones News Service
and Reuters, and the stock price increased 19% to $5.50. RIT's stock price
steadily increased after the announcement for several days, and by the end
of the next month the price climbed to $7.125. 47 Figure 1 displays the
daily closing price for RIT from one week before MacDonald's trades,
December 8, 1975, through January 29, 1976.
MacDonald did not sell his shares immediately after the announcement;
instead he waited more than a year before selling at roughly $10 a share.
The SEC argued that MacDonald should disgorge his actual profits. 48 While
the district court accepted the SEC's argument, 49 the United States Court
of Appeals for the First Circuit reversed the disgorgement decision holding
that disgorgement should only be the amount of profit attributable to the
inside information. 50 The First Circuit stated that profit should be based
on the difference between the purchase price and the price "a reasonable
time after the inside information had been generally disseminated. ' 51
Moreover, "the court should consider the volume and price at which RIT
shares were traded following disclosure, insofar as they suggested the date
42. SEC v. Shapiro, 349 F. Supp. 46, 54-55 (S.D.N.Y. 1972), aff'd, 494 F.2d 1301 (2d Cir.
1974).
43. Id.at 56.
44. Shapiro, 494 F.2d at 1309.
45. 699 F.2d 47 (1st Cir. 1983).
46. MacDonald placed a limit order to buy up to 20,000 shares at $4.25 but only 100
shares were available at this price.
47. See infra Figure 1.
48. MacDonald, 699 F.2d at 52.
49. SEC v.MacDonald, Litigation Release No.0073, (1981 Transfer Binder] Fed. Sec. L.
Rep. (CCH) 98,009 (D.R.I. Apr. 23, 1981).
51. Id.
'5 2
by which the news had been fully digested and acted upon by investors.
The case then was remanded to the district court to determine a reasonable
53
time.
Upon remand, the district court used the price movement of RIT stock
as evidence of full assimilation of the information. 54 The district court
stated that the market did not fully digest the news immediately because
the price continued to rise for several days after the December 24 an-
nouncement.55 The court held that the price stabilized on January 13 and
used the average price of $6.50 on that day for disgorgement. 56 Thus, the
court of appeals and the district court on remand did not follow exactly
the approach of the courts in Texas Gulf Sulphur and Shapiro-the full
information price was the closing price on the day after the public release
of the information-claiming the information in MacDonald was "consid-
erably less spectacular" 5 than the Texas Gulf Sulphur ore strike. 58 Addi-
tionally, the district court recognized that part of the price increase sub-
sequent to the lease-agreement announcement corresponded to a
favorable Wall Street Journal story on December 31, 1975 regarding the
sale of properties by RIT.5 9 While the district court held that the Wall
Street Journal story brought creditability to the press release, 60 the Wall
Street Journal story did not pertain to the lease agreement announcement
that was the basis of the insider trading. 6' MacDonald therefore appealed
the finding of the district court and argued that the Wall Street Journal
article was "an intervening, superseding, cause of the RIT stock price surge
in early 1976."62 The court of appeals affirmed the judgment of the district
5
court, 6 saying it was "unable to conclude that the district court committed
'64
clear error in rejecting defendant's argument.
The MacDonalddecision is an important precedent for determining SEC
disgorgement calculation. First, the court of appeals in MacDonaldreversed
the lower court's method of profit calculation and recommended paper
profits as in Texas Gulf Sulphur and Shapiro, strengthening the use of paper
52. Id.
53. Id.
54. SEC v. MacDonald, 568 F. Supp. 111 (D.R.I. 1983), aff'd, 725 F.2d 9 (1st Cir. 1984).
55. Id. at 113.
56. Id. at 112.
58. Id.; MacDonald, 568 F. Supp. at 114 n.5.
59. MacDonald, 568 F. Supp. at 112 n.1. The Wall Street Journal reported that a Boston
firm planned to buy 15 of RIT's properties. During the three-day period surrounding the
announcement, RIT's stock price increased from $5.25 to $5.875. See Realty Income Trust
Says Boston Group Bids for 15 Properties, WALL ST. J., Dec. 31, 1975, at 8.
60. MacDonald, 568 F. Supp. at 113.
61. SEC v. MacDonald, 725 F.2d 9, 11 (1st Cir. 1984).
62. Id.
63. Id. at 10.
64. Id.at 11.
profits over actual profits. 65 Second, the MacDonald decision held that a
reasonable time must take place after the public release of information
and before complete dissemination occurs.66 Because the court of appeals
held that the pattern of the price and volume movements after the an-
nouncements should be considered in determining a reasonable time pe-
riod, 67 the SEC and the courts have had leeway in determining when a
"reasonable time" has taken place.
MacDonald also illustrates that the use of financial economics analysis
reduces the ambiguity in determining the reasonable time period. For
example, financial economic analysis could have been applied to show that
the price increase around the unrelated Wall Street Journal article was
specific to the information in that article and not at all to the information
that MacDonald used in his trading. Further, while the courts and the SEC
held that the price increases that continued until January 13 were related
to the inside information,68 it so happens that the overall stock market
increased substantially over this period. 69 To the extent that RIT's stock
price moved with the overall stock market, part of the increase in the price
of RIT over this period could have been due to general economic con-
ditions.
Disgorgement in Schedule 13D Violations

Historically, the SEC confined disgorgement to insider trading cases.
Recently, however, the SEC has obtained disgorgement for other violations
such as delinquent Schedule 13D filings. The SEC requires that purchasers
of more than five percent of the stock of a publicly traded company file
a Schedule 13D within ten business days after crossing the five percent
threshold. 0 If the holder is intent on acquisition, the Schedule 13D pro-
vides information about the acquisition group including a list of its mem-
bers and the group's intentions for the target company.7 The SEC also
requires the holder to file amendments to Schedule 13D after the occur-
rence of material changes in the information contained in the original
72
filing.
65. SEC v. MacDonald, 699 F.2d at 47, 55 (lst Cir. 1983).

66. Id. at 53-54.
67. Id. at 55.
68. MacDonald, 568 F. Supp. at 113.
69. See infra Figure 1.
70. 17 C.F.R. § 240.13d-l(a) (1993); see also id. § 240.13d-101. Schedule 13D implements
section 13(d) of the Exchange Act. 15 U.S.C. § 78m(d) (1988).
71. 17 C.F.R. § 240.13d-101.
72. Id. § 240.13d-2(a).
Stock prices rise in response to Schedule 13D filings.7 3 Therefore, share-

holders who file a Schedule 13D later than required may be able to pur-
chase shares subsequent to their Schedule 13D, triggering purchase at
prices lower than they would have paid otherwise. This means that filing
a delinquent Schedule 13D can lead to extraordinary profits that are sub-
ject to disgorgement. The SEC first sought and obtained disgorgement
for an improper Schedule 13D filing in 1988 in SEC v. First City Financial
Corp.74 During February and March of 1986, First City Financial Corp.
(First City), controlled by the Belzberg family, purchased over nine percent
of the stock of Ashland Oil Company (Ashland). On March 25, 1986, First
City announced its stake in Ashland. The next day, First City proposed a
merger in a letter to Ashland and filed a Schedule 13D with the SEC
detailing its ownership. Ashland responded to the proposal the following
week by announcing the repurchase of the stake from First City.
The SEC alleged that First City and its vice president, Marc Belzberg,
violated section 13(d) of the Exchange Act by filing the Schedule 13D later
than required.75 As of February 28, First City held 4.9% of Ashland stock.
According to the SEC, on March 4 Bear Stearns purchased a large block
of shares on behalf of First City through a put and call agreement.7 6 The
SEC claimed that this purchase gave First City beneficial ownership of the
shares held by Bear Stearns and hence First City should have filed the
Schedule 13D by March 17 rather than on March 26.7 Thus, the SEC
argued that First City purchased Ashland stock from March 17 to March
26 at prices that did not incorporate properly the impact of crossing the
78
five percent threshold.
The SEC sought disgorgement of approximately $2.7 million, reflecting
the difference between the price paid by First City for 890,100 shares
purchased between March 17 and March 26 (average purchase price of
$48 a share) and the price at which Ashland repurchased the shares from
First City on April 2 ($51 a share).7 9 The court approved the SEC's dis-
73. See Wayne H. Mikkelson & Richard S. Ruback, An Empirical Analysis of the Interfirm
Equity Investment Process, 14 J. FIN. EcoN. 523 (1985); Clifford G. Holderness & Dennis P.
Sheehan, Raiders or Saviors? The Evidence on Six Controversial Investors, 14 J. FIN. ECON. 555
(1985).
74. 890 F.2d 1215 (D.C. Cir. 1989).
75. Id. at 1223.
76. Id. at 1219. Under a put and call agreement, a broker buys stock for its own account
with the understanding that its client, the investor, can purchase the stock from the broker
at a set price, plus interest and commissions. To protect itself from market risks, the broker
has the right to put the stock to the investor at the same price. In First City, the defendants
argued that a misunderstanding occurred between Bear Steams and First City in that First
City merely meant to tell Bear Steams that buying Ashland stock would be a good investment,
not to buy and hold the stock for it. See id. at 1217-20.
77. Id. at 1220-23.
78. Id. at 1223-24.
79. Id. at 1230.
gorgement estimate, noting that the precise measure of ill-gotten gains

was not actual profits but that "disgorgement need only be a reasonable
approximation of profits causally connected to the violation." 80
In First City, the court explicitly chose not to use financial economics
to estimate the savings to the defendants of a late Schedule 13D filing,
stating that "[d]espite sophisticated econometric modelling, predicting
stock market responses to alternative variables is, as the district court
found, at best speculative."'" Moreover, the court rejected the testimony
of an expert witness employing financial economics who argued that factors
in addition to the Schedule 13D filing contributed to the price increase
82
during that period.
Disgorgement and Financial Economics

Financial economics analysis can be quite useful in estimating the
amount of profits a wrongdoer must disgorge. When disgorgement is used
as a penalty, the defendant must disgorge the profits realized from his or
her fraudulent conduct. Financial economics can be used to provide un-
biased estimates of these profits. This methodology is especially useful
when the actual profits realized from the securities transactions do not
equal the profits directly attributable to the fraudulent actions. To date,
this analysis has not been used universally for disgorgement calculation,
in part because it may appear complicated to the courts. As the court of
appeals in SEC v. First City Financial Corp.83 said, "[i]f exact information
were obtainable at negligible cost, we would not hesitate to impose upon
the government a strict burden to produce that data to measure the precise
amount of the ill-gotten gains."8a4 Despite these concerns, financial eco-
nomics can play an important role in calculating disgorgement.
EVENT STUDY METHODOLOGY

An event study is a statistical technique that estimates the stock price
impact of occurrences such as mergers, earnings announcements, and so
forth.8 5 The basic notion is to disentangle the effects of two types of in-
formation on stock prices-information that is specific to the firm under
80. Id. at 1231.

81. Id.
82. Id. at 1232.
83. 890 F.2d 1215 (D.C. Cir. 1989).
84. Id. at 1231.
85. See generally Cynthia Campbell & Charles Wasley, Measuring Security PricePerformance
Using Daily NASDAQ Returns, 33 J. FIN. ECON. 73 (1993); LAURENTIUS MARAIS & KATHERINE
SCHIPPER, APPLICATION AND EvENT STUDY METHODS IN LITIGATION SUPPORT (1992); Glenn
V. Henderson, Jr., Problems and Solutions in Conducting Event Studies, 57 J. RISK & INS. 282
(1990); Pamela P. Peterson, Event Studies: A Review of Issues and Methodology, 28 Q. J. Bus.
& ECON. 36 (1989); Stephen J. Brown & Jerold B. Warner, Using Daily Stock Returns: The
Case of Event Studies, 14 J. FIN. ECON.3 (1985).
question (e.g., dividend announcement) and information that is likely to

affect stock prices marketwide (e.g., change in interest rates). Eugene F.
Fama, Lawrence Fisher, Michael Jensen and Richard Roll from the Uni-
versity of Chicago were the first researchers to apply this methodology.
Their seminal work examined the stock price reaction to stock splits and
86
subsequently was published by the InternationalEconomic Review in 1969.
Event study methodology has its foundation in the efficient markets
hypothesis. This well-known hypothesis states that security prices reflect
all available information."7 While theoreticians have developed various def-
initions of this basic statement, for event studies the relevant definition is
that stock prices reflect all publicly available information. Numerous event
studies in the academic finance, accounting, economics, marketing, and
legal literatures incorporated the idea that if stock prices reflect all public
information, price changes around public announcements is due generally
to the arrival of new information stemming from that announcement.
Consistent with the efficient markets hypothesis, studies have shown that
stock prices react quickly to the arrival of new information, often within
88
a matter of seconds.
The execution of an event study is quite simple. It involves the identi-
fication of an event that causes investors to change their expectations about
the value of a firm. The investigator compares a stock price movement
contemporaneous with the event to the expected stock price movement if
the event had not taken place. There are three basic steps in conducting
an event study: (i) define the event window; (ii) calculate abnormal stock
86. Eugene F. Fama et al., The Adjustment of Stock Prices to New Information, 10 INT'L ECON.
REV. 1 (1969).
87. See Eugene F. Fama, Efficient Capital Markets: A Review of Theory and Empirical Work,
25J. FINANCE 385 (1970) and Fama, supra note 1, for reviews of the literature on the efficient
markets hypothesis. From time to time, the efficient markets hypothesis comes under intense
criticism. For example, in the aftermath of the stock market crash of 1987, many commen-
tators suggested that the crash invalidated the efficient markets hypothesis. For the most
part, however, this hypothesis withstood such criticism and continues to be the most viable
theory offered. In fact, two studies reconcile the stock market crash with the efficient markets
hypothesis. See Mark L. Mitchell & Jeffry M. Netter, Triggeringthe 1987 Stock Market Crash:
Antitakeover Provisions in the Proposed House Ways and Means Tax Bill?, 24 J. FIN. EcoN. 37
(1989) and CharlesJ. Jacklin et al., Underestimationof Portfolio Insuranceand the Crash of October
1987, 5 REV. FIN. STUD. 35 (1992). More importantly, with respect to event studies described
herein, even critics of the efficient markets hypothesis concur that the efficient markets
hypothesis is relevant. See Lawrence H. Summers, Does the Stock Market Rationally Reflect
Fundamental Values?, 41 J. FINANCE 591, 596 (1986). This is because, as Fama points out,
there is little debate that individual firm's stock prices respond quickly to the release of new
information about that firm. See Fama, supra note 1, at 1601. The debate about the efficient
markets hypothesis is more concerned with the determinants of overall market fluctuations.
88. See generally James M. Patell & Mark Wolfson, The Intraday Speed of Adjustment of Stock
Prices to Earnings and Dividend Announcements, 13 J. FIN. ECON. 223 (1984); Larry Y. Dann
et al., Trading Rules, Large Blocks, and the Speed of PriceAdjustment, 4 J. FIN. EcON. 3 (1977).
price performance around the event; and (iii) test for statistical significance
of the abnormal stock price performance.
DEFINING THE EVENT WINDOW

The first step in the event study is selecting an event window. The event
window is the period when information about the event becomes available
to the stock market and thus may affect the relevant company's stock price.
For most publicly-traded corporations, the event is disseminated publicly
by newswire sources such as Dow Jones & Company and Reuters. In the
case of Dow Jones & Company, the news is distributed via the Dow Jones
Broadtape immediately after receipt of the news release from the corpo-
ration or government agency. For the relatively important news releases,
Dow Jones also reports the event in the Wall Street Journal on the next
business day. 9
The efficient markets hypothesis is influential in determining the length
of the event window. Because the efficient markets hypothesis, supported
by considerable empirical evidence, suggests that stock prices react quickly
to the release of new information, in many cases the event window will be
relatively short, sometimes as short as one trading day. In determining the
length of an event window, an important tradeoff exists. The longer the
event window, the more likely the window includes the period during which
all the new information about the event is released. The tradeoff, however,
is that long event windows may include noise and information from other
events, making it difficult to isolate the impact of the relevant event.
The extent of the difficulty in defining the event window length varies
across events. In those instances where the release of new information is
a complete surprise to the market, it is relatively easy to establish the
beginning of the event period. Consider an airline crash, for example. 90
Because airline crashes are unanticipated, the first day of the event period
is either the day of the crash or the subsequent trading day if the crash
occurred after the market close. 91 Even when it is easy to identify the
beginning of the event window, it can be difficult to establish the end of
the event window. In the airline example, the end of the period would
depend on when all of the relevant information regarding the crash was
made available to market participants. 92 For some crashes, it may take
89. For information regarding news releases sent to Dow Jones & Company and their
subsequent dissemination and impact upon the stock market, see Mark L. Mitchell &J. Harold
Mulherin, The Impact of Public Information on the Stock Market (1993) (unpublished manuscript,
on file with The Business Lawyer, University of Maryland School of Law) and Robert Thompson
et al., Attributes of News About Firms: An Analysis of Firm Specific News Reported in the Wall Street
Journal Index, 25 J. FINANCE 245 (1987).
90. See Mark L. Mitchell & Michael Maloney, Crisisin the Cockpit? The Role of Market Forces
in Promoting Air Travel Safety, 32 J. LAw & EcoN. 329 (1989).
91. Id. at 340.
92. Id.
several days or perhaps even weeks before the market receives all the
relevant information; in these cases, a longer event window is more nec-
essary than for a crash in which all information is available within a few
hours following the crash. 93 In most cases, however, the bulk of the in-
formation is released at the announcement of the event. Because the mar-
ket processes information rapidly, it is conventional to expand the window
only a short period after the announcement. The current academic stan-
dard is to extend the event period to the close of trading on the day after
94
the release of the pertinent information.
For those events that are subject to leakage, defining the beginning of
the event window can be problematic. Consider the case of a merger in
which the target company is rumored to be "in play" prior to the an-
nouncement. 95 For such a case, the event window should begin prior to
the actual merger announcement, perhaps as long as a week or two. Ideally,
the first day of the event window corresponding to a merger would be the
date on which investors began trading on news about the upcoming
merger, regardless of whether the news was based on rumors, inside in-
formation, a Schedule 13D filing, or a public announcement that merger
talks were in process. In practice, this date is difficult to define and some
degree ofjudgment is required generally based on price and volume move-
ments prior to the merger announcement.
With respect to securities fraud cases, there is substantial variation in
the complexity of determining the length of an event window. In some
fraud cases, choosing the appropriate event window is straightforward. An
example is an insider trading case where the information used by the
investor is revealed subsequently in a single public announcement. On the
other hand, in many securities fraud cases the relevant information is
revealed slowly over time, while during the same period investors receive
other, sometimes unrelated, information about the firm(s) in question. In
the latter case, it is relatively difficult to choose an appropriate window.
The main advice is to carefully identify the exact dates during which the
information in question reached the market, and then restrict the window
to a short period if possible, generally two or three days around each
release of new information.
93. Id.
94. This is particularly true when the researcher examines a sample of several occurrences
of the same type event such as a merger announcement. For a single event that is generally
the norm in a securities fraud case, depending upon market factors, the window often can
extend beyond the close of trading the day after the public announcement.
95. See Gregg A. Jarrell & Annette B. Poulsen, The Returns to Acquiring Firms in Tender
Offers: Evidence From Three Decades, 18 FIN. MGMT. 12 (1989) and Lisa K. Meulbroek, An
Empirical Analysis of Illegal Insider Trading, 47 J. FINANCE 1661 (1992), for a discussion of
stock price movements prior to major events. Gregg Jarrell and Annette Poulsen document
evidence of substantial stock-price run-up in target firms prior to takeover announcements.
Lisa Meulbroek shows that insider trading often accounts for a large part of this stock price
run-up.
CALCULATE ABNORMAL STOCK PRICE

PERFORMANCE
The next step in the event study is to examine the stock price perfor-
mance around the event. The goal is to isolate the effect of the event on
the contemporaneous stock price movement. Stated differently, the in-
vestigator attempts to determine whether the stock price behavior around
the event is abnormal. A large abnormal stock price movement occurring
at the same time the market receives news about an event suggests that
the event caused the abnormal price movement. Furthermore, the link
between the event and the price movement is even stronger if there is no
other new information reaching the market at the same time that could
affect the stock price.
The simplest way to evaluate abnormal stock price performance is to
visually examine the stock price movement around the event and assess
whether it appears small or large. Of course, the degree to which the stock
price movement is small or large depends not only on the absolute value
of the movement but also on the movement relative to historical patterns
and to contemporaneous overall market movements.
Calculationof Stock Returns

In finance terminology, the change in a stock price over a given period
is known as the stock price return. The return is expressed as:
r = [(P, - P0) + DIV,]/P 0
where
P, = price at end of period
P0 = price at beginning of period
DIV, = dividend paid during period.
Thus, the return is simply the change in the stock price during the period
plus any payout of dividends during the period, relative to the stock price
at the beginning of the period. 96 This discussion focuses on daily stock
96. Researchers often express returns in logarithmic form as

r = natural logarithm(1 + [(P, - P0) + DIV,]/P1
or as
r = natural logarithmt[P, + DIV]/P}.
The logarithmic return is a continuously compounded return whereas the return described
in the text is a simple return. For practical purposes the distinction between these two return
measures is relatively minor. One benefit of the logarithmic return method is that in statistical
terminology, the transformation makes the distribution of the returns closer to a normal
distribution, thus improving the validity of statistical testing. For ease of exposition, the simple
return measure is focussed upon. Further, it is also the case that the simple return measure
provides better estimates for disgorgement purposes.
price returns, which is the standard time interval used in most event studies,
although returns can be calculated over any increment of time such as
hours or months. In securities fraud litigation, daily stock price returns
are typically the appropriate measure. In some cases, an examination of
hourly, weekly, or monthly data may be warranted-in such cases, the
methodology as described can be applied similarly.
An example of a major event to examine abnormal stock market per-
formance is the Tylenol poisonings of 1982. 91 On September 30, 1982,
Johnson & Johnson, the maker of Tylenol, announced that three people
died as the result of ingesting cyanide-laced Tylenol capsules. 98 Four more
deaths were reported within the next two days. 99 The Tylenol poisonings
resulted in 125,000 stories in the print media alone-an event unprece-
dented in American business.
To the extent that investors expected the Tylenol poisonings to reduce
future cash flows to the stockholders ofJohnson &Johnson, its stock price
should have declined in response to the announcement of the poisonings.
According to the efficient markets hypothesis, the stock price decline will
occur quickly. Correspondingly, the return to Johnson & Johnson stock
on September 30, 1982, the day that Johnson & Johnson revealed the
Tylenol poisonings, is:
-6.50% = [$46.125 - $43.125]/$46.125

where $46.125 and $43.125 are the closing prices on September 29 and
30, respectively.
Calculation of Standard Deviation

A decline of 6.50% on a given day appears quite large, especially for a
blue-chip firm such as Johnson & Johnson. It is necessary to perform
statistical tests, however, to determine that the 6.50% decline did not occur
by chance. One approach is to compare the return to a series of returns
over some prior period. The comparison period typically ranges from 100
to 300 trading days. For the Johnson & Johnson example, the trading days
for the one-year period ending on September 29, 1982, the day before
the public announcement of the poisonings, are used. There are 253 trad-
ing days during this period. Interestingly, for only one day during the
prior year did Johnson & Johnson's stock move more in absolute value
than on September 30, 1982. That day is August 17, 1982 when Johnson
& Johnson's stock price increased 7.19%. It so happens that this large
positive return is likely due in part to the overall stock market increase of
97. See Mark L. Mitchell, The Impact of External Partieson Brand-Name Capital: The 1982
Tylenol Poisonings and Subsequent Cases, 27 EcoN. INQUIRY 601 (1989).
98. Id. at 601.
99. Id.
4.45% that day. 100 The fact that there is only one return during the prior
one-year period that is of the magnitude of the September 30, 1982 decline
suggests this decline is significant.
To assess the significance of the -6.50% return on September 30, 1982,
a well-known metric of variation in statistics is relied on, the standard
deviation. This metric measures the dispersion in a variable around its
mean value. The standard deviation for stock returns is formally expressed
as:
,7- -(r,TY
S =
N-1
where f is the mean return over the sample period and N is the number
of trading days in the sample period. As the formula indicates, the greater
the variation around the mean value in the sample, the larger the standard
deviation. Suppose for example that all the returns had the same value.
In such a case, there would be no dispersion around the mean value and
thus the formula would indicate a value of zero. Note that the term (r, -
f) is squared-the rationale is the magnitude of the deviation of returns
from the mean value is what matters, not whether a return is above or
below the mean. The division by N - 1 adjusts for the number of returns
in the sample. The intuition behind the N - 1 term is straightforward. If
this divisor were not included, the calculated standard deviation would
increase in magnitude as the number of returns increased. Thus, the nu-
merator keeps track of total deviations while the denominator keeps track
of the number of deviations. In this light, the ratio represents an average
deviation between an observation and its mean.
The standard deviation of the daily return forJohnson &Johnson stock
during the 253 trading days period prior to the Tylenol poisonings is
1.84%. What can be inferred about the impact of the Tylenol poisonings
on Johnson & Johnson's stock price when there was a - 6.50% return on
September 30, 1982 and the historical daily mean return and standard
deviation of Johnson & Johnson's stock is known? The most common way
to analyze this question is to consider the statistical significance of the
daily return.
Testing the Statistical Significance of a Stock Return

Once a researcher identified an event window and calculated the return
during that event window, he or she then can determine the statistical
significance of the return. The question is whether the absolute value of
the return is large enough so that the researcher can indicate with con-
fidence that the return is relatively unusual. The importance of the his-
100. For discussion of controlling for general market movements, see infra text accom-
panying notes 106-113.
torical average and standard deviation of the daily returns is highlighted

in making the assertion that a given daily return is different from the
typical daily return. There is an additional consideration in this analysis-
the role of the normal distribution.
Many statistical tests rely on the assumption that the data of interest is
normally distributed. The normal distribution is attractive because it is a
good description of a wide variety of random variables including stock
returns. In the normal distribution, the values of the variable are distrib-
uted symmetrically around the mean value and are not concentrated
0 1
around extreme values. A normal distribution has the familiar bell-shape.
Also important is that a variable that is distributed normally can be de-
scribed by its mean and standard deviation. For example, for a normally
distributed random variable, the probability is 68.3% that a randomly se-
lected value will lie within one standard deviation of the mean value. Sim-
ilarly, the probability is 95.5% that a randomly selected value will lie within
two standard deviations of the mean value. Expressed differently, there is
only a 4.5% chance that a randomly selected observation will not fall within
two standard deviations of the mean value. Finally, very few observations
fall outside the boundary of three standard deviations from the mean
value-the probability is 99.7% that a randomly selected observation will
lie within three standard deviations of the mean value.
While visual displays of stock returns suggest returns tend to follow a
bell-shaped distribution, prior statistical research indicates that they are
not distributed precisely normally. 0 2 Even so, researchers have shown that
the normal distribution is an appropriate approximation for event study
analyses.' 03 Throughout the remainder of this Article, it is assumed that
stock returns actually adhere to the normal distribution so that appropriate
hypothesis tests can be constructed to determine whether stock price move-
ments during event windows are statistically significant.
To calculate probability values, the normal distribution must be trans-
formed into the standard normal distribution. A standard normal distri-
bution has a mean value of zero and a standard deviation of one. One can
calculate z-statistics with this standard distribution-the z-statistic is ex-
pressed as:
z-statistic = (observed value - mean value)/standard deviation.
Most standard statistics texts include a table of the cumulative standard
normal distribution. 104 The table reports for various values of the z-statistic
101. See generally LYMAN Or, AN INTRODUCTION TO STATISTICAL METHODS AND DATA
ANALYSIS (3d ed. 1988).
102. See Eugene F. Fama, The Behavior of Stock Market Prices, 38 J. BUSINESS 34 (1965).
103. See Brown & Warner, supra note 85, for a detailed exploration of the distribution
of stock returns. Stephen Brown and Jerold Warner state that "[t]he non-normality of daily
returns has no obvious impact on event study methodologies." Id. at 25.
104. See generally OTT, supra note 101.
the probability that a z of that value or greater will occur. Thus, a re-
searcher usually will convert an observation drawn from a normal distri-
bution into a z-value in order to assess the significance of that value.
The methodology discussed in the previous paragraph is phrased more
formally in terms of hypothesis testing. In general, a test of significance
aims to answer the question of whether an observed difference is real or
simply occurred by chance. In statistical tests, the researcher usually sets
out a null hypothesis which states that an observed difference occurred
by chance. If the null hypothesis is rejected because a test statistic (such
as a z-statistic) is greater than a specified value, then it is unlikely the
difference occurred by chance. This result often is called a finding of
statistical significance.
For example, researchers apply decision rules to determine whether a
given value is significantly different from the mean value. An often used
convention is the five percent rule-values greater than or equal to 1.96
standard deviations from the mean value are considered significantly dif-
ferent from the typical value because there is only a five percent chance
that a randomly selected value will be 1.96 or more standard deviations
from the true mean. Thus, if the calculated z-statistic has an absolute value
of 1.96 or greater, the observed value could be considered significant at
the five percent level. The decision rule may be more stringent. For ex-
ample, there is only about a one percent likelihood that a randomly selected
value will lie outside 2.58 standard deviations or more from the average
value. Thus, if the z-statistic is greater than or equal to 2.58, the observed
value can be considered significant. A third commonly used decision rule
is ten percent-here, the probability is ten percent that a randomly selected
value will lie 1.65 standard deviations or more from the mean value. Gen-
erally, researchers use a decision rule based on one percent, five percent,
or ten percent significance levels.
Stock returns provide a good example of a test of whether an observation
is significantly different from the mean. In the case of daily stock returns,
the mean daily return is very close to zero: the mean annual return on
the stock market over the past thirty years was roughly twelve percent with
a corresponding mean daily return of 0.045%. Because the daily return
is so small, it is assumed that it is zero for statistical tests and thus a test
of whether a daily stock return is different from the mean is just a test of
whether a daily return is different from zero. Therefore, the z-statistic is
simply the daily return divided by the standard deviation. If the z-statistic
is 1.96 or greater (based on a decision rule of five percent), the results
indicate the daily return is significantly different from the mean return.
In an article published in the Virginia Law Review in 1991,105 the authors
(with Jonathan R. Macey and Geoffrey P. Miller) provided guidelines as
to the magnitudes of daily stock price returns that are statistically different
105. Macey, supra note 8.

from zero. For the stocks of the largest equity-value New York Stock Ex-
change (NYSE) listed firms, a stock price movement of 2.86% could be
considered significantly different from zero at the five percent level. In
contrast, for the smallest equity-value NASDAQ firms, the necessary price
movement to be considered significant at the five percent level is 10.02%.
Therefore, because stock price volatility varies widely across firms, infer-
ences about the significance of a firm's stock returns are made with respect
to a comparison with that stock's own return history.
Statistical tests of significance are useful both in establishing materiality
and in calculating disgorgement. A finding that a stock return associated
with the release of information is large enough that it is unlikely that the
return occurred by chance is strong evidence that the information was
important. Therefore, if that information was used allegedly in securities
fraud, the finding that the associated stock return is large enough to be
statistically significant implies the information is material. Furthermore, a
finding of statistical significance for stock returns data used in calculations
of disgorgement is an indication that the estimates are accurate.
For example, suppose a firm's stock price increases seven percent on
the day that management releases a favorable earnings announcement.
Suppose, also, that the prior day an insider of the firm purchased stock
based on his or her knowledge of the forthcoming announcement. The
insider subsequently is charged with illegal insider trading. A finding that
the seven percent return on the earnings announcement day is statistically
significant is strong empirical evidence that the news was important. Stated
differently, it is unlikely that the seven percent increase in the stock price
occurred by chance. Furthermore, in calculating profits for disgorgement
based on the stock price increase on the announcement day, if the return
is statistically significant, then a more credible argument can be made that
the seven percent return represents the value of the defendant's inside
information.
Returning to the Johnson & Johnson example, recall that the standard
deviation during the year prior to the poisonings was 1.84%. Dividing the
return to Johnson & Johnson stock of -6.50% by this standard deviation
yields a z-statistic of 3.53. It is highly improbable that a randomly selected
return from Johnson & Johnson's return history would yield a value that
is more than 3.53 standard deviations away from zero. Thus, one can claim
with a high degree of confidence that Johnson & Johnson's stock price
decline on September 30, 1982 did not occur by chance and thus the
decline is likely due to the public announcement of the Tylenol poisonings.
It is preferable, however, to correct for overall market movements before
calculating the significance of abnormal returns.
Net of Market Stock Price Performance

When the market receives new information about the future cash flows
of a company, the stock price quickly moves to a new value reflecting the
new information. Often the information is firm specific in nature-e.g.,

earnings announcements. In addition, stock prices of individual firms move
in conjunction with overall stock market movements that are caused by
changes in underlying economy-wide factors. Thus, it is important to ac-
count for these marketwide movements, especially during periods when
the market is volatile. The best example is the fall of 1987 when market
volatility was extremely high around the stock market crash. 0 6 The basic
method for accounting for marketwide factors subtracts the marketwide
return from the individual stock's return. This estimate is called the net-
of-market return.
Several choices are available as proxies for a marketwide return. Two
well known measures are the Dow Jones Industrial Average (DJIA) and
the Standard & Poor's 500 Index (S&P 500). The DJIA is limited somewhat
as a market index as it contains only thirty stocks and thus large movements
in this index often can be driven by changes in just a few stocks. With the
S&P 500, this problem is less severe. While the S&P 500 is an acceptable
index to proxy for the overall market, this Article uses an even broader
measure-an index based on all stocks on the NYSE, American Stock Ex-
change (ASE), and NASDAQ. This index comes from the Center for Re-
search in Security Prices (CRSP) at the University of Chicago and is one
of the more comprehensive indexes available. It should be restated, how-
ever, that the S&P 500 is an appropriate proxy as well. Because the cor-
relation between these two indexes is close to one, similar results usually
are obtained regardless of the choice of index.
On September 30, 1982 whenJohnson & Johnson's stock price dropped
6.50%, the overall market, as proxied by the CRSP value-weighted NYSE,
ASE, and NASDAQ index, dropped as well, declining 0.89%. Thus, the
net-of-market to Johnson & Johnson stock was -5.61%. As a result, it can
be argued that the overall market decline can account for some of the
decline in Johnson & Johnson's stock price that day. Even so, the net-of-
market return is still quite large. To put this net-of-market return in per-
spective, it was calculated for the prior 253 trading days. Over this one-
year period, a net-of-market return of this magnitude never occurred. The
closest in absolute value took place on March 9, 1982, when Johnson &
Johnson's net-of-market return was 5.17%. This evidence provides addi-
tional support of the notion that the public announcement of the Tylenol
poisonings caused a major revaluation of Johnson & Johnson stock. To
test for statistical significance, the standard deviation of Johnson & John-
son's net-of-market return over the prior 253 trading day period is com-
106. For examples of event studies that examined specific announcements during the
October 1987 stock market crash, see Mitchell & Netter, supra note 87 andJeffry M. Netter
& Mark L. Mitchell, Stock-Repurchase Announcements and Insider TransactionsAfter the October
1987 Stock Market Crash, 18 FIN. MGMT. 84 (1989).
puted-this sample standard deviation is 1.42%. °7 Thus, the net-of-market

return of -5.61% is highly statistically significant as it is roughly four
standard deviations away from a mean return of zero. Stated differently,
the z-statistic is 3.95, which is substantially greater than the z-statistic of
2.58 necessary for significance at the one percent level.
Beta-Adjusted Stock Market Performance

Although net-of-market returns, in many cases, provide an appropriate
estimate of the stock price effects of new information, there are instances
where computation of market-adjusted returns requires a more refined
analysis to account for the fact that not all stocks are affected identically
by economy-wide factors. That is, the stock returns of some firms move
proportionately more than the market in reaction to economy-wide news;
the stock returns for some firms track the overall market very closely; and
the stock returns of other firms are relatively insensitive to marketwide
patterns. The methodology for this adjustment is the market model and
requires an estimation of the relation between the stock returns of the
individual firm and the returns of the overall market index during a com-
parison period (also known as the estimation period) which typically pre-
cedes the event window. The performance of the firm's stock during the
event window is then compared to the predicted performance during the
event window. The predicted performance is based on the firm's stock
price relationship with the market over the control period.
The first step is to estimate the market model:
Rit = a + liRmt + cit
which assumes that the return to a stock i at time t is a function of the

market return, R, plus a random error term, fi, that is uncorrelated with
the market return. The market model decomposes the return on a stock
into two parts, one part due to factors influencing the market and one
part due to variables specifically related to the firm itself. The term 13,
often referred to as beta, measures the sensitivity of a firm's stock returns
to overall market returns. Although on average the returns on stocks vary
proportionately with the returns on the market index, the returns on
individual stocks typically vary more or less than the returns on a market
index. For example, a beta of 1.5 indicates that a stock's return typically
increases (or decreases) fifteen percent when the market increases (or
decreases) ten percent. Similarly, a beta of 0.6 means that a stock's return
typically increases (or decreases) only six percent in conjunction with a
ten percent market increase (or decrease). In sum, firms that are relatively
107. Note that the net-of-market daily return standard deviation of 1.42% is less than the
standard deviation (1.84%) of Johnson & Johnson's actual returns over this period. This
difference is attributed to the fact that the actual return incorporates marketwide as well as
firm-specific factors, and thus is more volatile than the net-of-market return.
sensitive to market movements (e.g., airlines) typically have betas that are
greater than one, highly diversified firms have betas that are close to one,
and firms that are relatively insensitive to market movements (e.g., regu-
08
lated utilities) have betas that are less than one.
The market model is estimated with regression analysis. The estimation
period for this market model equation typically ranges from 100 to 300
trading days preceding the event under study. That is, the researcher uses
the estimates of a and 3, and the movement of the market to predict how
the stock price of the firm would have changed during the event period
if there were no firm-specific information released during the event period.
The difference between the predicted return and the actual return on a
given date during the event window is known as the abnormal return. The
abnormal return expressed as:
AR, = r, - (&, + ,R)
measures the impact of the event on stock i at time t.

As an example, apply the market model to the Tylenol case. To calculate
the abnormal return to Johnson & Johnson stock on September 30, 1982,
first the market model for the 253 prior trading days is estimated. As in
the computation of the net-of-market return, the overall stock market is
proxied by the CRSP value-weighted index of NYSE, ASE, and NASDAQ
stocks. The estimated beta for Johnson &Johnson for the one-year period
prior to the crash was 1.29. A beta of 1.29 suggests that Johnson & John-
son's stock typically increased (or decreased) 12.9% when the overall mar-
ket increased (or decreased) 10%. The intercept or alpha term is virtually
zero as it is only 0.0975%.
The abnormal return on September 20, 1982 to Johnson & Johnson
stock is
-5.46% = -6.50% - 0.0975% - (1.29 X -0.89%),
where -6.50% is the actual return to Johnson & Johnson stock, 0.0975%
is the estimate of alpha from the market model, 1.29 is Johnson & John-
son's beta estimate and -0.89% is the market return on September 30,
1982.109 Notice the abnormal return of -5.46% is not as negative as the
net-of-market return of -5.62%. The net-of-market return approach as-
sumes beta is 1.0, whereas the estimated beta for Johnson & Johnson is
1.29. Thus, more of the decline in Johnson & Johnson's stock price on
108. The term a (i.e., the intercept of the market model), also known as alpha, represents
the mean return on the stock when the market return equals zero. Although over time and
on average, alpha approximates zero for most companies, it can be significantly different
from zero for different intervals.
109. There is a very slight rounding error in these calculations as the calculated abnormal
returns are based on parameter estimates and stock returns using five decimal places instead
of the two decimal places as indicated for exposition purposes throughout the text.
September 30 is accounted for by market factors when it is adjusted with

the estimated beta than when assuming beta is 1.0.110
As before, statistical tests are necessary to estimate the confidence that
the abnormal return is different from zero. In computing the significance
of the abnormal stock price performance using the actual return and the
net-of-market return, those returns simply are compared to the standard
deviations of the actual and the net-of-market returns over the prior 253
trading days, respectively. In computing the statistical significance of the
abnormal returns, however, the significance tests are more complex than
in the case of the net-of-market return, yet the intuition is still the same."'
In this case, the researcher estimates the standard error of forecast"12 for
the abnormal return as
2
2
Sar = ts (1 + 1/Ne + (R - P-) 2/CSSI,)1/
where S2 is the estimated residual variance from the regression model for
the estimation period, N. is the number of trading days in the estimation
period, k is the estimation period sample mean of the market return,
and CSSR is the corrected sum of squares of the market return during
the event period. This measure is essentially the standard deviation of
Johnson &Johnson's returns during the prior 253 trading days accounting
for the relation of the returns with the stock market, plus terms to account
for the number of observations in the estimation period and overall market
deviations on the event date.' ' The estimated standard error of forecast
for September 30, 1992 is 1.42%.
110. It should be noted that recent works, see Eugene F. Fama & Kenneth R. French, The
Cross-Section of Expected Stock Returns, 47 J. FINANCE 427 (1992) and Eugene F. Fama &
Kenneth R. French, Common Risk Factors in the Return on Stocks and Bonds, 33 J. FIN. EcoN.
3 (1993), suggest additional risk factors to the overall market, such as firm size and market/
book equity, should be accounted for when calculating abnormal returns. Under certain
conditions, in a securities fraud case in which the information is released over a long period
of time, the additional factors may alter the calculation of the abnormal returns. In the
Johnson & Johnson example and the cases that follow, however, the Fama and French mul-
tifactor model does not alter the results.
111. For recent articles that describe statistical tests in event studies, see Ekkehart Boehmer
et al., Event-Study Methodology Under Conditions of Event-Induced Variance, 30 J. FIN. EcoN.
253 (1991) and Imre Karafiath & David E. Spencer, Statistical Inference in Multiperiod Event
Studies, 1 REv. QUANTITATIVE FIN. & Accr. 353 (1991). See also supra note 85.
112. See generally JOHN JOHNSTON, ECONOMETRIC METHODS (1984).
113. The term s' (0.0002) is the estimated residual variance from the regression model.
The square root of this term is 0.0141 or 1.41% and is simply the standard error of the
regression model. Notice that this term is virtually identical to the standard deviation of the
net-of-market returns over the estimation period. The major distinction is that the standard
error from the regression model is a measure of the variation in Johnson & Johnson's return
accounting for a more precise relation with the overall market. The first term in brackets is
simply 1. The second term, 1/N, accounts for the number of days in the estimation period.
The longer the estimation period, the more precisely estimated the market model parameters.
This term is generally very small as estimation periods typically range from 100 to 300 days.
The third term accounts for large stock market movements on the event date. The larger
Similar to the prior statistical tests, the abnormal return is compared

with the standard error of forecast to determine significance. The abnor-
mal return of -5.46 is nearly four times greater than the standard error-
the z-statistic is -3.85. This indicates that even after accounting for beta,
Johnson & Johnson's stock price declined significantly on September 30,
1982, when Johnson & Johnson revealed that Tylenol was laced with cy-
anide.
Cumulative Abnormal Returns

As noted earlier, event windows can extend beyond one trading day.
For these cases, the abnormal returns can be cumulated to create the
cumulative abnormal return formally expressed as
T
CART = HI (1 +
t,1
AR,) - 1
where T is the length of the event window. 14 The CAR measures the total
impact of the event on the firm. Generally, event studies report both the
AR and CAR over the event window. For event windows where the in-
formation was released over several days, the CAR often is emphasized.
Analogously, simple returns and net-of-market returns can be cumulated
as well.
Again, the Tylenol example illustrates measuring abnormal performance
over multiday periods. A ten-day event window is constructed, covering
the period from September 30 through October 13, 1982. Table 1 reports
cumulative measures ofJohnson &Johnson's stock price performance over
this ten-day period in Panels A-C respectively." 5 Panel A displays the actual
returns performance, Panel B displays the net-of-market returns perfor-
mance, and Panel C displays the abnormal returns performance.
the absolute value of the overall stock market movement on the event date, the larger this
term, and hence the larger the standard error of forecast. Consider for example the market
crash of 1987 when the overall stock market fell approximately 20%. An abnormal return
of 4% on this day would be less significant than an abnormal return of 4% on a day when
the market was flat. In general, the second two terms are very small. For example on Sep-
tember 30, the sum of the three terms in brackets was 1.008, resulting in a standard error
of forecast of 1.42%.
114. Intuitively, cumulative abnormal returns would appear to be simply the sum of the
abnormal returns over the event window (this actually would be the case for logarithmic
returns). The problem with a simple summation of the abnormal returns is that the base in
the calculation changes from day to day. Therefore, the sum of abnormal returns does not
yield a holding period return. Note that the product of (1 +AR) over the event period is
used rather than simply summing the abnormal returns. To illustrate, consider a price move-
ment from $4 to $5 (AR = 25%) on one day and then back to $4 (AR = 20%) the next
day. Summing the abnormal returns would yield a value of 5%, yet the holding period return
is zero. Taking the product of (1 + .25) and (I + -0.20) yields the correct cumulative
return of zero.
115. See infra Table 1.
To illustrate the cumulation technique, consider the actual returns on

September 30 and October 1 of -6.51% and 1.74%, respectively. The
cumulative return is given by the product of (1 - 0.0651) X (1 + .017)
- 1. Thus, the two-day cumulative return is -4.88%. This process con-
tinues throughout the length of the event window. As of the tenth day of
the event window, October 13, the cumulative return to Johnson & John-
son stock is -8.94%. Also, as shown in Table 1, there is a z-statistic that
corresponds to the cumulative return." 6 With respect to the cumulative
return of -8.94% over the ten-day window, the z-statistic is 1.54. As this
z-statistic is slightly less than 1.65, the -8.94% cumulative return misses
statistical significance at the ten percent level.
The cumulative net-of-market returns in Panel B display strikingly di-
vergent estimates from that of the cumulative return. Here, the cumulative
net-of-market return on October 13, 1982 is -18.91%. The reason that
the cumulative net-of-market return is considerably more negative than
the cumulative return is that there were large overall stock market gains
during this two-week period. That is, Johnson & Johnson's stock would
have realized even a larger absolute decline in value had the market not
increased during this event window. Similar results are revealed in Panel
C which shows the cumulative abnormal returns, based on the market
model estimates. In this case, the cumulative abnormal return is -22.39%
by the tenth day following the public announcement of the Tylenol poi-
sonings. Given that Johnson & Johnson's beta exceeds one, its stock price
should have outperformed the market over this event window in the ab-
sence of the Tylenol poisonings-this fact accentuates the negative cu-
mulative abnormal performance realized by Johnson & Johnson.
The sharply contrasting differences between the returns to Johnson &
Johnson and the market-adjusted returns illustrate an important point that
is relevant for securities fraud cases. Occasionally, announcements by cor-
porations occur contemporaneous with large overall stock market move-
ments. These market movements must be accounted for to isolate the stock
price impact of the firm's announcement. It is then possible to assess the
materiality of the information and the value of the information contained
in the announcement for disgorgement purposes.
A hypothetical example using the Johnson & Johnson facts shows how
it may be important to account for market movements in using financial
116. The z-statistics for the cumulative returns simply are computed as the square root
of the sum of the variances associated with each of the daily returns over the event window.
For example, to compute the z-statistic for the cumulative return of -4.88% over the first
two days of the event window, the standard error of the forecast for each of the two days
is squared to obtain the variance for each of these two days. The square root of the sum of
the two variances is the standard error for a two-day cumulative return. The reason to convert
to variances before summing is that mathematically, variances can be summed whereas stan-
dard deviations cannot. Note that the variances are cumulated based on logarithmic returns
due to better specified distributional properties.
economics in securities fraud cases. Suppose an employee of Johnson &

Johnson found out at his job on September 30 that there were poisonings.
He then either sold stock or bought puts in Johnson & Johnson before
the public received any news of the poisonings. In an insider trading case
brought against this employee, financial economics could be used to show
the information he had was material and to calculate disgorgement. As
stated previously, the cumulative abnormal return of -22.39% is much
more negative than the cumulative return of -8.94% over the period the
news about the poisonings became public. Because the insider's infor-
mation was about factors that affected Johnson & Johnson's stock price
and not the overall market, the cumulative abnormal return is theoretically
a better measure of the materiality of the information than the cumulative
return. In this case the very large cumulative abnormal return of -22.39%
significantly buttresses the claim of materiality.
For the same reason, cumulative abnormal returns are also theoretically
a better measure than cumulative returns in calculating profits for dis-
gorgement. Cumulative abnormal returns only measure the impact of firm-
specific information, in this case the news about poisonings. In a real sense,
the value of the information to the trader is best represented by how the
information would have affected the stock price in the absence of any
other factors-the cumulative abnormal return. Finally, note that this
analysis is even more intuitively appealing if the employee in his trading
hedged against overall market movements.
In addition, note the role of statistical tests in this example. The cu-
mulative abnormal return of -22.39% has an associated z-statistic of
-4.96. That z-statistic indicates that the cumulative abnormal return is
highly significant-at the one percent level. Therefore, it is very unlikely
that the negative cumulative abnormal return occurred by chance, which
is strong evidence that the information about the Tylenol poisonings led
to the negative cumulative abnormal return. The finding of statistical sig-
nificance is thus strong evidence the information was material and it boosts
the credibility of disgorgement estimates based on the cumulative abnor-
mal returns.
USE OF FINANCIAL ECONOMICS IN SEC

ENFORCEMENT CASES
Two roles for financial economics in securities fraud cases-determining
materiality and calculating disgorgement-have been suggested. The event
study methodology provides the basis for these two roles. The following
five recent SEC cases provide evidence of staff economists participating
in the determination of materiality and disgorgement. 1 7
117. Only cases for which relevant information may be obtained from the SEC under the
Freedom of Information Act are discussed. Cases that would require consideration of any
nonpublic information are not discussed. For example, financial economics analyses at the
SEC may suggest that potential suits not proceed. Such investigations are not mentioned
here.
118
INSIDER TRADING BY AN EXECUTIVE RECRUITER
In September 1986, Artel Communications, a small fiber-optic telecom-
munications firm traded on NASDAQ, fired its chief executive officer and
hired Ingoldsby Associates to recruit a replacement. Three months later
Ingoldsby Associates recommended Robert Bowman for the position, and
on February 4, 1987 Artel's board voted to offer the job to Bowman. On
the morning of February 9, Artel informed Ingoldsby Associates that Bow-
man accepted. Later that day, Michael 0. Ingoldsby, president of In-
goldsby Associates, purchased 23,500 shares of Artel stock for approxi-
mately $72,000 (average price of $3.06 per share). Artel announced the
appointment the following morning; Reuters Ltd. and Dow Jones News
Service reported the appointment at 11:17 a.m. and 4:13 p.m., respec-
tively. In April 1989, the SEC charged Ingoldsby with insider trading based
on information he misappropriated from Artel.
Table 2 displays Artel's stock market performance during the period
surrounding Bowman's appointment. 119 Over the two-day period, February
10 (announcement date) and the prior day, Artel's stock price increased
from $2.25 to $3.75. The abnormal return on the announcement day is
20.53% and is 45.38% on the prior day. The large abnormal return on
February 9, the day prior to the announcement, is likely due to trading
by Ingoldsby and leakage of the information. As noted supra, stock prices
often move significantly prior to a material announcement as the infor-
mation leaks out. For example, Ingoldsby's purchase of 23,500 shares on
the 9th was very large relative to prior days. The average daily number of
shares traded for Artel over the prior year was 13,948. Thus, the trades
by Ingoldsby alone on February 9 exceeded the daily average by 69%. The
total trading volume of 72,000 shares on February 9 was more than five
times higher than the daily average. His own trades and the fact that he
told his broker that an important announcement was about to take place
118. The facts for this case originate from SEC v. Ingoldsby, Litigation Release No. 12,461,
[1990 Transfer Binder] Fed. Sec. L. Rep. (CCH) 95,351 (May 15, 1990) and Thomas
Newkirk & Catherine Shea, Civil Penaltiesand the Securities and Exchange Commission's Recent
jury Trial Experience Under the Insider Trading Sanctions Act, in SECURITIrs ENFORCEMENT
INsTITUTr 289 (PLI Corp. Law & Practice Course Handbook Series No. 741, 1991).
119. Abnormal and cumulative abnormal returns are calculated in the same manner as in
the Johnson &Johnson example. To maintain continuity in this Article, in all cases the value-
weighted CRSP index of all NYSE, AMEX, and NASDAQ stocks are used as the market proxy.
There are instances, however, in the actual cases at the SEC, depending on the facts of the
case, that staff economists also tried different market indexes as well as industry indexes in
the estimates. Likewise, in the financial economic analysis of these cases at the SEC, economists
also tried several different estimation periods ranging from 100 to 300 days to calculate
betas. The purpose was simply to test all alternatives so as to verify the robustness of the
estimates. Here, however, for reasons of simplicity only the results based on the broad-based
CRSP index and on estimation periods that cover the 253 days (one year) prior to the event
are presented.
at Artel may have provided signals to other investors that firm-specific

information was imminent.
The stock price remained relatively flat until February 18 when it rose
to $4.50 (abnormal return is 20.25%). On this date, a federal holiday, a
February 16 article published in FiberOptics News became publicly available.
It reported that Bowman would remain a director at Telco Systems where
he previously was chairman. 20 Fiber Optics News indicated considerable
potential for synergies between the two firms as they were direct com-
petitors.' 21 The only other large stock price movement in February oc-
curred towards the end of the month when Artel's stock dropped roughly
15% on February 25, from $4.25 to $3.625. This decline occurred fol-
lowing a Dow Jones Broadtape story at 6:20 p.m. the prior day reporting
fourth-quarter 1986 losses for Artel.
In order to prove insider trading, the SEC must establish that the inves-
tor possessed material, nonpublic information. 22 For this case, two factors
warranted the use of financial economics in establishing materiality-the
relative importance of a management appointment and the fact that Artel
was a thinly-traded stock. First, would a reasonable investor consider a
managerial appointment important? The majority of insider trading cases
involve corporate control transactions where a large price movement gen-
erally occurs. This is not the case with managerial appointments, however,
as empirical studies indicate a small, positive return at the announcement
of top management appointments. 2 Poorly managed firms 2 4 or firms in
financial distress 125 exhibit larger returns, but at most the average abnormal
return is only about three percent. These studies, however, indicate sub-
stantial variation across the abnormal returns and thus the stock price
movement associated with the Bowman announcement is not incompatible
26
with the academic evidence.
The second step in establishing materiality is the significance of the price
movement. As indicated in the Artel example, the positive abnormal re-
120. Bowman Crowned Chairman, CEO of Artel Communications, FIBER OPTics NEWS, Feb.
16, 1987, at 1.
121. Id.
122. See generally Dirks v. SEC, 463 U.S. 646 (1983); Chiarella v. United States, 445 U.S.
222 (1980).
123. See Eugene P. H. Furtado & Vijay Karan, Causes, Consequences, and Shareholder Wealth
Effects of Management Turnover: A Review of the Empirical Evidence, 19 FIN. MGmrr. 60 (1990).
124. See Michael S. Weisbach, Outside Directors and CEO Turnover, 20 J. FIN. EcON. 431
(1988).
125. See Karl-Adam Bonnier & Robert F. Bruner, An Analysis of Stock Price Reaction to
Management Change in Distressed Firms, 11 J. Accr. & EcON. 95 (1989).
126. According to the review article by Eugene P. H. Furtado and Vijay Karan, see supra
note 123, the research has centered on NYSE and AMEX firms. It is likely the stock price
reaction is larger for NASDAQ firms, especially small ones such as Artel, which had an equity
value of only about $6 million prior to the announcement, because managerial capital ar-
guably represents a greater proportion of total value at smaller firms.
turns are very large on the announcement day, February 10, and the prior
day, as well as February 18 in conjunction with the Fiber Optics News story.
Even so, it is especially important to test for statistical significance because
Artel is a thinly-traded stock. For example, during 1986 the daily average
trading volume for Artel's stock was only 25% of the average daily trading
volume for the average NASDAQ stock and only about 5% of the average
daily trading volume for all publicly traded stocks. As noted supra, small
stocks are considerably more volatile than large stocks-thus, the abnormal
returns must be larger before a given abnormal return can be considered
significantly different from zero. This general fact is true for Artel as well.
For example, the standard deviation of Artel's stock returns over the prior
one-year period is 5.57%. Thus, for a given day, Artel's stock return must
be about 11% before significance can be asserted. In contrast, recall from
the Tylenol example that the standard deviation of Johnson & Johnson's
return was 1.84%. Here, for this large NYSE firm that exhibits less volatile
price movements, significance on a given day can be asserted with a much
smaller stock price movement.
In spite of the fact that Artel's stock price is very volatile, the abnormal
returns on the announcement day, February 10, the prior day, and Feb-
ruary 18 (associated with the Fiber Optics News article), the positive ab-
normal returns are highly statistically significant largely because of the
magnitude of these abnormal returns.
Even though the abnormal returns are statistically significant around
the information release dates, trading volume also is very high on these
dates. It is possible that the high trading volume might have created tem-
porary price pressure on Artel stock. The stock price, however, remains
stable until the negative earnings announcement at the end of the month,
buttressing the SEC's establishment of materiality. Furthermore, the cu-
mulative abnormal return remains significant at the five percent level
throughout the entire month.
This analysis suggests that Ingoldsby could have profited from the inside
information, thus warranting disgorgement. Ingoldsby purchased 23,500
shares for $72,000 (average price of $3.06). On February 24, Ingoldsby
sold 3000 shares at $4.125 and one day later he sold 1000 at $3.875. He
then held on to the remaining 19,500 shares until the spring of 1989 when
he sold at an average price of $2.00. Thus, Ingoldsby realized actual losses
of about $17,000 on his investment based on inside information. As noted
supra,12 7 however, paper profits, not actual profits, theoretically provide a
better benchmark with which to calculate disgorgement.
In calculating paper profits, the full information price is the first date
when Artel's stock price fully reflected the information regarding the Bow-
man hiring. While most of the price reaction occurred at the announce-
ment, Artel's price also increased significantly one week later subsequent
127. See supra text accompanying note 65.

to the Fiber Optics News story which noted that Bowman would remain a
director at Telco Systems, increasing the likelihood of potential business
combinations between the two competitors. If Ingoldsby knew that Bow-
man would remain a director at Telco, then the full information price
would reflect the February 18 price increase. The SEC argued that the
full information price extend out to February 18 to reflect the price impact
of the Fiber Optics News story.'28 In contrast, Ingoldsby claimed the court
should follow the standard set by Texas Gulf Sulphur and base disgorgement
on the closing price the day after the announcement. l2 9 Thus, he argued
for a full information disgorgement price of $3.75, the closing price on
February 11.s0
The court accepted the SEC's argument and held that the full infor-
mation price was $4.50, the closing price on February 18.11 The court
calculated disgorgement at $24,663, accounting for commission costs
32
($0.10 per share), bid ask spread ($0.25 per share), and other factors.
Thus, while Ingoldsby actually lost money on his transactions, he was re-
quired to disgorge the paper profits realized from his trading. In outlining
its decision, the court stated:
Artel was a relatively small company with limited media attention and
exposure. Although a story regarding the new Artel president ran in
the Wall Street Journal on February 11, 1987, I find that the news
was not fully disseminated, absorbed and digested by the investing
public until after the Bowman articles appeared in the fiber optics
33
trade publications.1
In sum, this case illustrates the application of financial economics in an
insider trading action. The event study technique is applied to show that
information about a pending managerial appointment is material. Addi-
tionally, the event study analysis is used to calculate the value of inside
information for disgorgement calculations.
128. The SEC and the court cited the MacDonald case as precedent for extending the
event window a few days beyond the date of the original announcement. SEC v. Ingoldsby,
Litigation Release No. 12,461, [1990 Transfer Binder] Fed. Sec. L. Rep. (CCH) 95,351,
at 96,694-95 (May 15, 1990). Despite the apparent similarities between the two cases, how-
ever, the argument for a longer window is stronger in this case than it was in MacDonald. As
previously discussed, there are two problems with the district court's reasoning in the
MacDonald case. First, in MacDonald the follow-up Wall StreetJournal story was unrelated to
the information on which MacDonald traded. Second, the overall stock market increased a
great deal during the long event window used in MacDonald.Both of these factors are different
in the Ingoldsby case. In Ingoldsfy, the subsequent Fiber Optics News story is directly related
to Ingoldsby's inside information. Also, there were no market swings during the long window
in Ingoldsby that could have caused the significant increase in Artel's stock price.
129. Id. at 96,694-95.
130. Id. at 96,695.
131. Id.
132. Id.
133. Id.
OFFICER OF ACQUIRER SELLS TARGET

134
STOCK AND
BUYS CALL OPTIONS IN ACQUIRER
On March 9, 1987, Robert Slattery, divisional vice president of Reebok
International, purchased Reebok call options and sold stock of Avia Group
International. On the following day, Reebok announced an agreement to
acquire Avia in a move to expand market share in the athletic shoe industry.
In September 1988, the SEC charged Slattery with trading on the basis
of inside information.
On February 11, 1987, one month prior to his alleged insider trading,
Slattery purchased 1000 shares in Avia at $19 per share.1 35 Coincidentally,
twelve days later, at a Reebok executive meeting, Slattery learned of Ree-
bok's interest in acquiring Avia at about $16 a share. For the next two
weeks, Reebok and Avia officials engaged in merger negotiations, culmi-
nating in Reebok's announcement of the acquisition on March 10. During
this period, Slattery participated in Reebok's investigation of Avia's facil-
ities. On the day before the merger announcement, Slattery sold his 1000
shares of Avia stock at $26 a share. He also purchased twenty March call
option contracts on Reebok stock with exercise price $40 at $75 per con-
tract and twenty April call option contracts with exercise price $40 at $156
per contract.
Table 3 displays the stock and option price performance for Reebok and
the stock price performance for Avia during the period surrounding the
merger proposal. 36 On the merger announcement date, March 10, 1987,
Reebok's stock price increased from $37.50 to $41.75. The stock market
impact of the merger proposal occurred largely on the announcement day
as Reebok's stock price did not move very much on any of the days sur-
rounding the announcement. Because Slattery purchased Reebok calls,
Table 3 also displays Reebok option prices for March 40 calls and for April
40 calls. In both cases, the price of the calls increased greatly on March
10: the March 40 calls increased from $0.8125 to $2.50 and the April 40
calls increased from $1.875 to $4.125.
The primary issue in this case concerned the establishment of materi-
ality.' 37 This issue is especially important here because Slattery's trading
pattern before the announcement-sell shares in the target and buy calls
in the acquirer-is unprofitable in a typical merger. Thus Slattery may
argue that his trading behavior could not have been based on information
regarding a subsequent merger announcement as it is well known that the
stock price of the target almost always increases upon acquisition an-
134. The facts of the case originate from SEC v. Slattery, Litigation Release No. 11,856,
1988 SEC LEXIS 1758 (Sept. 2, 1988).
135. The SEC did not charge Slattery with buying the Avia stock based on inside infor-
mation, only the subsequent sale of Avia shares was considered illegal. Id.
137. Slattery, 1988 SEC LEXIS 1758.
nouncement. Furthermore, some financial economists documented that

the average stock price reaction to acquiring firms for acquisition an-
nouncements during the 1980s was negative.13 8
While the average announcement-period return may have been negative
for some samples and close to zero for most studies, large price increases
for the acquiring firm are not that uncommon. In a study of 401 acqui-
sitions during 1982-86 (Average Abnormal Return = -0.08%, z-statistic
= -0.45), however, Mark L. Mitchell and Kenneth Lehn found that in
51 of the acquisitions, the stock price of the acquiring firm increased more
than five percent during a three-day window surrounding the announce-
ment. 3 9 Furthermore, researchers recently documented positive stock
price reactions to acquirers when the merging firms operate in the same
40
line of business.1
Moreover, as displayed in Table 3, the Reebok abnormal return of 9.90%
on March 10, the announcement day, is statistically significant at the one
percent level (z-statistic is 2.80). For the ten trading days surrounding the
announcement, none of the abnormal returns are statistically different
from zero, indicating that there was little leakage prior to the merger
announcement and that the initial price reaction on March 10 captured
the full impact of the announcement (notice that the cumulative abnormal
remains fairly steady after March 10). The combination of the large stock
price and option price movement for Reebok on the day of the merger
announcement and the prior empirical evidence suggests Slattery pos-
sessed material, non-public information.
Further evidence that Slattery had material information comes from the
fact that he sold Avia stock the day before the announcement, and Avia's
stock price fell from $25 to $17.25 on the announcement day. This stock
price decline contrasts sharply with the large, positive price reaction that
almost always occurs with target firms, and illustrates the facts of this
merger are unique. Insiders and venture capitalists owned most of the
stock in Avia. While Avia never went public, roughly 800,000 converted
debentures (about eight percent of the common stock) from a 1981 fi-
nancing traded in the "pink sheet" over-the-counter market. Prior to the
Reebok merger offer, it was known publicly that Avia was in the process
of planning a public offering at $11 to $14 a share. As Table 3 indicates,
investors in the small pink sheet market for Avia stock anticipated that the
price after the public offering would rise well above the $11 to $14 range.
138. See Jarrell & Poulsen, supra note 95; Michael Bradley et al., Synergistic Gains from
CorporateAcquisitions and Their Division Between the Stockholders of Target and Acquiring Firms,
21 J. FIN. ECON. 3 (1988).
139. Mark L. Mitchell & Kenneth Lehn, Do Bad Bidders Become Good Targets?, 98 J. POL.
EcoN. 372 (1990).
140. See Neil W. Sicherman & Richard H. Pettway, Acquisition of Divested Assets and Share-
holders' Wealth, 42 J. FINANCE 1261 (1987); Randall Morck et al., Do Managerial Objectives
Drive Bad Acquisitions?, 45 J. FINANCE 31 (1990).
Consequently, when Reebok announced the merger pact with Avia at a

price of $16.50, the share price of Avia fell sharply from $25 to $17.25.
Therefore, Slattery's trades that on the surface appear surprising were
actually predictably profitable. He bought his 1000 shares of Avia a month
earlier at $19 a share. By the day of his trades (March 9), Slattery knew
that Reebok's merger was likely to go forward. Consequently, he had ma-
terial information given that Avia's stock price was trading at about $25
rather than $16.35. The 31% decline in Avia's price in the pink sheet
market suggested a material price change even for a stock with thin trading.
Disgorgement calculation is simple in this case because Slattery's security
transactions clustered around the announcement date. For the Reebok
call options, the SEC computed the difference between the purchase price
on March 9 and the sale price on March 10, adjusting for commission
costs. 41 For the sale of the Avia stock on March 9, the SEC computed
disgorgement as the difference between the price of $26 at which he sold
the 1000 shares on March 9 and the closing price of $16.50 on March
10,142 the merger announcement date, applying the arguments of the ef-
ficient markets hypothesis that stock price rapidly adjusts to the release
of new information. 43 Neither admitting or denying the allegations, Slat-
tery agreed to settle with the SEC by 1disgorging profits of $11,129 and
44
paying a penalty of the same amount.
145
DELINQUENT SCHEDULE 13D FILING
On December 18, 1987, Francis Spillman, president of Pizza Inn, bought
50,000 shares of a chain of chicken restaurants called Winners Corpo-
ration, increasing his stake from 4.3 1%to 5.56%. Because Spillman crossed
the five percent threshold with this purchase, SEC rules required him to
file a Schedule 13D within ten calendar days, reporting his ownership stake
and intention for Winners. 14 6 Spillman, however, did not file a Schedule
13D until January 6, eight days later than required. During the period
between the required filing date and the actual date of filing (December
29 through January 5), Spillman bought an additional 45,000 shares in-
creasing his stake to 6.9%. When Spillman filed the Schedule 13D on
January 6 reporting the 6.9% stake, he also revealed a tender offer con-
sideration at $4.25 per share for the remainder of the stock. Winners
141. Slattery, 1988 SEC LEXIS 1758.

142. Id. The March 10 closing price of $16.50 used in disgorgement differs from the price
given in Table 3, as the latter price is the daily average of the bid and ask prices.
143. Id.
144. Id.
145. The facts of the case originate from SEC v. Spillman, Litigation Release No. 12,321,
1989 SEC LEXIS 2384 (Dec. 13, 1989).
146. 17 C.F.R. §§ 240.13d-l(a), -2(a) (1993). During the ten days following December 18,
Spillman purchased an additional 19,900 shares.
rejected the potential offer, and three months later Spillman began re-
ducing his stake; by October Spillman had sold all of his Winners' stock.
In December 1989, the SEC charged Spillman with violating section
13(d) of the Exchange Act. 147 Specifically, the SEC argued that during the
period from December 29, 1987 to January 5, 1988, Spillman purchased
Winners' stock at prices that did not reflect the information that should
have been reported in the Schedule 13D. 148 Academic research in 1985
supports this argument; 149 these studies document significant, positive
price reactions to announcements of Schedule 13D filings. The SEC sought
disgorgement of the savings Spillman realized in purchasing the 45,000
shares from December 29 to January 5.150
As noted supra, the SEC first sought disgorgement for Schedule 13D
violations in First City where the court required disgorgement of the actual
profits.1 5' The court used actual profits, which it recognized were only a
reasonable approximation of the "ill-gotten gains," because they were easy
to calculate-the difference between the price the defendants received
when they sold their stock back to the corporation and the price they paid
for the stock. 5 2 In First City the defendants sold their stock shortly after
the 13D filing so that the paper profits were similar to the actual profits.
In this case, however, actual profits might not be a reasonable approxi-
mation of the paper profits from the late filing because Spillman waited
three months before beginning to sell some of his shares.
Table 4 displays the stock price performance for Winners for December
11, 1987 through January 22, 1988. l15 Table 4 also reports the ratio of
daily volume to the mean volume over the prior year in Winners' stock
and the ratio of the volume accounted for by Spillman to the mean volume
over the prior year. Table 4 shows Winners' stock price rose during the
period when Spillman was purchasing large numbers of Winners' shares,
before filing the required Schedule 13D. Specifically, from December 11
through January 5 the cumulative Abnormal Return for Winners' stock
was roughly 70% a period during which Spillman accounted for 45.5% of
the trading volume. For example on December 16, when Spillman pur-
chased 107,500 shares (accounting for all but 3000 shares traded that
day), Winners' price increased from $1.375 to $1.50, and on December
18, when Spillman purchased 50,000 shares (accounting for 42% of shares
traded), Winners' price increased from $1.50 to $2.25. Besides accounting
for a substantial amount of the trading volume on these days, the volume
of Spillman's trades often exceeded historical volume. For example, on
147. 15 U.S.C. § 78m(d) (1988).

148. Spillman, 1989 SEC LEXIS 2384.
149. See supra note 73.
150. Spillman, 1989 SEC LEXIS 2384.
151. See supra text accompanying notes 74-82.
152. SEC v. First City Fin. Corp., 890 F.2d 1215, 1231 (D.C. Cir. 1989).
December 16, Spillman's volume was 12.2 times greater than the average
daily volume over the prior year.
An argument that Spillman should not be required to disgorge any
profits from the late Schedule 13D filing rests on the observation he simply
could have purchased the 45,000 shares during the ten-day period after
crossing the five percent threshold but before the required filing date (that
is, by December 28). The stock price evidence, however, suggests Spillman
could not have purchased the 45,000 shares in this period without in-
creasing the market price. Therefore, it is likely that, to the extent that
the late Schedule 13D filing enabled Spillman to spread his purchases over
a longer period, he was able to buy shares more cheaply; otherwise he
would have purchased them during the authorized time period.
The disgorgement issue focuses on the determination of the price that
Spillman would have had to pay for the 45,000 shares purchased from
December 29 throughJanuary 5 had he filed on December 28, as required,
rather than January 6. On January 6 when Spillman announced his Sched-
ule 13D filing, Winners' stock price increased from $2.50 to $3.00 (ab-
normal return = 19.98%, z-statistic = 3.72). Thus, one could argue that
if Spillman filed on December 28, Winners' stock price would have closed
at $3.00 rather than $1.875 (actual closing price on December 28th and
29th). 154 Accordingly, Spillman should disgorge the difference between the
price he paid for the 45,000 shares and $3.00.
Winners' stock price performance after the Schedule 13D filing suggests,
however, that $3.00 may be too high a price to use for disgorgement.
While the price increased from $2.50 to $3.00 on January 6, the date of
the 13D filing announcement, it began to fall two days later. The decline
in the stock price suggests the $3.00 price partly reflected the market
overestimating the probability of a successful tender offer. Furthermore,
the data in Table 4 indicate that the market reacted to Spillman's trades
during the period around the filing. Spillman often accounted for a large
proportion of the total volume and his large purchases likely led to price
pressure, as evidenced by the fact that the price increased on those days
in which he made relatively large purchases. In addition, in almost all cases
Spillman made his purchases at the high price of the day, likely due to his
large share amounts.
Consequently, the SEC based the price that Spillman should have paid
for the 45,000 shares on the estimated price at which he could have sold
these shares following the Schedule 13D filing. As argued supra, it is un-
likely that Spillman could have sold the shares for $3.00 because Winners'
154. It could be argued that had Spillman filed on December 28, Winners' stock price
would not have risen fully to $3.00, because the $3.00 closing price on January 6 reflected
the purchases of the 45,000 shares during the period December 29 to January 5. Holderness
and Sheehan, however, find little evidence of a relation between the size of the purchase by
an investor filing a Schedule 13D and the stock price reaction. See Holderness & Sheehan,
supra note 73, at 565.
stock was thinly traded and Spillman's purchases often accounted for a
large amount of the trading volume. Thus, the SEC used the average price
($2.675) over the two-week period following the Schedule 13D filing as a
basis for disgorgement calculation, under the assumption that he could
have sold his shares over this longer period without having a large market
impact. Spillman agreed to a permanent injunction against further vio-
lations and paid disgorgement of roughly $24,000, without admitting or
denying the allegations. 55
CHAIRMAN LEAKS PRIVATE INFORMATION TO

ANALYSTS1 5 6
On May 15, 1987, Ultrasystems Corporation announced lower earnings
than expected for the first quarter of 1987. The day before, Phillip J.
Stevens, founder and chairman of Ultrasystems, called several analysts who
followed Ultrasystems and informed them of the bad news. The analysts
in turn advised their clients who then sold the stock that day and the
following morning prior to the announcement at 3:14 p.m. In March 1991,
the SEC charged Stevens with violation of insider trading by providing the
negative information to select analysts in advance of the press release. 57
On May 15, the day of the press release, the price of Ultrasystems' stock
dropped about eight percent (negative abnormal return of roughly seven
percent), a decline that was statistically significant. In the typical insider
trading case, the establishment of materiality relies partly on the stock
price reaction at the announcement of the information. This case is unique
because Ultrasystems' stock price did not move very much subsequent to
the 3:14 p.m. announcement on May 15, thus putting into question the
materiality of the information. Instead almost all of the statistically sig-
nificant abnormal price decline on May 15 occurred by 10:15 a.m. The
price fall during the first few minutes of trading on May 15, however,
occurred just after large sell orders from the analysts' clients, thus indi-
cating materiality-in this case, the stock market just responded to the
negative information hours prior to its official actual release rather than
afterwards. 158 Neither admitting nor denying the insider trading allega-
155. SEC v. Spillman, Litigation Release No. 12,321, 1989 SEC LEXIS 2384 (Dec. 13,
1989).
156. The facts of the case originate from SEC v. Stevens, Litigation Release No. 12,813,
1991 SEC LEXIS 451 (Mar. 19, 1991).
157. Id. This case is the first in which the SEC charged that a corporate insider violated
insider trading laws by providing material information to analysts. While the analysts did not
pay Stevens for the inside information, the SEC argued that Stevens benefitted by enhancing
his reputation among analysts. Id. Stevens' reputation recently was tarnished as a result of
issuing positive forecasts only to be followed by unexpected bad earnings announcements,
such that some analysts actually ceased covering Ultrasystems.
158. See Meulbroek, supra note 95, for an empirical analysis of insider trading cases. She
shows that stock prices often move prior to material announcements as a result of insider
trading. Id. at 1675.
tions, Stevens settled with the SEC by disgorging more than $125,000
based on the losses the59analysts' clients avoided by selling the stock prior
to the announcement.
CORPORATE RAIDER SELLS STOCK IN TARGET FIRM

IMMEDIATELY FOLLOWING TAKEOVER BID160
On February 29, 1988, Mesa Limited Partnership, of which T. Boone
Pickens Jr. is the general partner, publicly announced a 3.8% stake in
Homestake Mining and offered to acquire the remaining shares at $20 a
share. Upon this announcement, Homestake's stock price jumped from
$14 to $18. That same day, Mesa began selling its stock in Homestake as
well as some call options. Mesa continued to sell shares in Homestake for
the next several days without issuing a press release informing the public
of its sale transactions. In September 61990, the SEC charged Mesa with
negligence without fraudulent intent.' '
The SEC claimed Mesa's February 29 press release was misleading be-
cause it did not reveal Mesa's intentions to sell shares in Homestake, en-
abling Mesa to sell Homestake stock at inflated prices. The primary issue
was whether Mesa's initial reporting of the ownership stake can be con-
sidered material, aside from the offer to acquire the remaining shares of
Homestake. If the 3.8% ownership disclosure is material, then the decision
to sell the stake is also likely a material event and hence should be disclosed
62
promptly.
As stated earlier, the announcement of a Schedule 13D filing generally
results in a significant, positive stock price change. Hence, the disclosure
of ownership stakes is typically material. Furthermore, an examination of
Mesa's prior announcements of ownership stakes in various corporations
indicates a large, positive stock price reaction. This suggests that investors
considered the announcement of the 3.8% stake in Homestake to be ma-
terial. Thus, the decision not to disclose promptly the sale of the stake
can be considered material as well. Mesa settled with the SEC without
159. Stevens, 1991 SEC LEXIS 451.

160. The facts of the case originate from SEC v. Mesa Ltd. Partnership and T. Boone
Pickens, Litigation Release No. 12,637, [1990 Transfer Binder] Fed. Sec. L. Rep. (CCH)
95,492 (Sept. 27, 1990).
161. Id. at 95,572.
162. An argument could be made that Mesa sold the stake to hedge its investment in
Homestake by covering potentially large losses in the event of unsuccessful merger. There
are a few isolated cases involving this strategy. For example, in 1989 when Alfred Checchi
acquired NWA, he sold more than half of a 4.9% stake in NWA at a time when it still was
not certain that the acquisition would go forward. The explanation was his intention to use
the proceeds of the sale to pay the merger proposal expenses. See Checchi Cut NWA Stake to
1.9%, Filing Indicates, WALL ST. J., June 26, 1989, at C9.
admitting or denying the allegation by disgorging $2.3 million in profits

63
from sales of Homestake stock.
SUMMARY
Modern financial economics is becoming increasingly influential in se-
curities fraud law. The efficient markets hypothesis has provided a frame-
work for the analysis of certain questions and a basis for generating em-
pirical evidence on the value of information in individual cases. Clearly,
there are certain areas of securities law where the efficient markets hy-
pothesis continues to have an impact. Of particular importance is an em-
pirical technique derived from the efficient markets hypothesis-the event
study. Event studies are useful to establish, among other things, materiality
and calculate damages in securities fraud litigation. Event study analysis
already was applied in five SEC enforcement actions.
There are many areas in securities fraud litigation where empirical tech-
niques from financial economics may be useful. Indeed, event study tech-
niques potentially are much more valuable than described in this Article.
Event study analysis is useful at all stages of litigation to both defendants
and plaintiffs. The analysis is applicable, not just in SEC insider trading
cases, but in all types of securities fraud actions, including private suits.
Furthermore, by providing objective, relatively precise measures of the
importance of information and of illegal profits or damages, the impor-
tance of financial economics in securities fraud litigation will continue to
increase.
163. Mesa Ltd. Partnership, [1990 Transfer Binder] Fed. Sec. L. Rep. (CCH) at 95,573.
TABLE 1
Stock Price Performance for Johnson & Johnson
Following the 1982 llenol Poisonings
Panel A: Actual Returns Performance
Cumulative
Date Stock Price Return Z-statistic Return Z-statistic
Sept. 29 46.125
Sept. 30 43.125 -6.51 -3.55 -6.51 -3.55
Oct. 1 43.875 1.74 0.95 -4.88 -1.88
Oct. 4 41.250 -5.98 -3.29 -10.57 -3.33
Oct. 5 39.000 -5.46 -2.97 -15.45 -4.21
Oct. 6 41.750 7.05 3.84 -9.49 -2.31
Oct. 7 40.375 -3.29 -1.80 -12.47 -2.77
Oct. 8 42.625 5.57 3.04 -7.59 -1.56
Oct. 11 43.500 2.05 1.12 -5.69 -1.10
Oct. 12 41.500 -4.60 -2.51 -10.01 -1.82
Oct. 13 42.000 1.21 0.66 -8.94 -1.54
Panel B: Net-of-Market Returns Performance
Cumulative
Net-of-Market Net-of-Market
Sept. 29 46.125
Sept. 30 43.125 -5.62 -3.95 -5.62 -3.95
Oct. 1 43.875 0.54 0.38 -5.11 -2.54
Oct. 4 41.250 -5.74 -4.03 -10.56 -4.28
Oct. 5 39.000 -5.88 -4.13 -15.82 -5.56
Oct. 6 41.750 3.96 2.78 -12.48 -3.92
Oct. 7 40.375 -5.45 -3.83 -17.25 -4.95
Oct. 8 42.625 3.77 2.65 -14.13 -3.75
Oct. 11 43.500 -0.36 -0.25 -14.43 -3.58
Oct. 12 41.500 -4.61 -3.24 -18.38 -4.30
Oct. 13 42.000 -0.65 -0.45 -18.91 -4.20
TABLE 1 (continued)
Stock Price Performance for Johnson & Johnson
Following the 1982 Tylenol Poisonings
Panel C: Abnormal Returns Performance
Cumulative
Abnormal Abnormal
Sept. 29 46.125
Sept. 30 43.125 -5.46 -3.85 -5.46 -3.85
Oct. 1 43.875 0.10 0.07 -5.37 -2.67
Oct. 4 41.250 -5.76 -4.07 -10.83 -4.40
Oct. 5 39.000 -6.10 -4.31 -16.27 -5.73
Oct. 6 41.750 2.98 2.06 -13.78 -4.32
Oct. 7 40.375 -6.17 -4.31 -19.10 -5.47
Oct. 8 42.625 3.16 2.21 -16.54 -4.38
Oct. 11 43.500 -1.15 -0.80 -17.50 -4.33
Oct. 12 41.500 -4.72 -3.33 -21.39 -5.00
Oct. 13 42.000 -1.28 -0.89 -22.39 -4.96
Note: Returns are expressed in percents. Stock price data is from Center for Research in
Security Prices (CRSP) at the University of Chicago. Market model estimation period is
September 30, 1981 through September 29, 1982. Market proxy is CRSP value-weighted
index of NYSE, AMEX, and NASDAQ stocks. Beta estimate for Johnson & Johnson is 1.29.
TABLE 2
Stock Price Performance for Artel Communications
Surrounding the Announcement of Robert Bowman as Chief
Executive Office on February 10, 1987
Cumulative
Artel Artel Abnormal Abnormal
Date Price Volume Return Z-statistic Return Z-statistic
Feb. 4 $2.250 10,100 -6.22 -1.12 -6.22 -1.12

Feb. 5 2.250 4,400 -0.41 -0.07 -6.61 -0.84
Feb. 6 2.250 1,600 0.56 0.10 -6.08 -0.63
Feb. 9 3.250 72,000 45.38 8.20 36.54 3.30
Feb. 10 3.875 68,100 20.53 3.70 64.56 5.22
Feb. 11 3.750 65,200 -3.73 -0.67 58.43 4.31
Feb. 12 3.625 16,300 -2.50 -0.45 54.47 3.72
Feb. 13 3.750 6,800 2.47 0.44 58.28 3.72
Feb. 17 3.750 25,600 -1.39 -0.25 56.09 3.38
Feb. 18 4.500 57,000 20.25 3.67 87.70 5.01
Feb. 19 4.500 23,500 0.17 0.03 88.02 4.79
Feb. 20 4.625 23,600 3.09 0.56 93.83 4.89
Feb. 23 4.250 7,300 -6.92 -1.25 80.42 4.03
Feb. 24 4.250 7,900 0.17 0.03 80.72 3.90
Feb. 25 3.625 24,700 -14.74 -2.67 54.08 2.52
Feb. 26 3.625 6,200 0.61 0.11 55.02 2.49
Feb. 27 3.500 6,200 -3.50 -0.63 49.59 2.17
Notes: Returns are expressed in percents. Stock price data is from Center for Research in
Security Prices (CRSP) at the University of Chicago. Market model estimation period is
February 4, 1986 through February 3, 1987. Market proxy is CRSP value-weighted index of
NYSE, AMEX, and NASDAQ stocks. Beta estimate for Artel Communications is 0.96.
0D0
CD CD~
mt 0
0 0-4 -
II I I I I
0
E
o1 o~V) toCCD0) an0- 00
o ID.(Z0 -0 000
0
0
00 0 In~ 0
C.C3
0
000 000 x
000qCI
- -1 in4
SC-4
-'4
0 c 0 cir I-71- C
0 0
0Lo t- r- CI4 r- in LO LO = X) U) Ln
ci 0
0,0 en 00 en r-
(,q
CI4 r- 10 V r-
0000
r-
C 'o 00
v 4 ati
U
'0
U
E~ t-
000
0 UrI in~ 0
C11 0
0 0 0
LO~ LO~LO
r
-0 0-I L- 0-4 MC
N0
(.0qIIQ - 11
0 C4 C11 McCq0-
N bO
0 0 .1 0 0 ~ 0
co I I I 0 -
(U
(U
* .
OR~
00 0- 000 0
CU
0S
Cu..
- c-c) 0 0 0C ) 0- C -f
z0 o5c
0 iU) 0 if) 00
0
.0.U 0-- 0 if) 0 00 Cf) In-
If)if 0 U)
0 C-f) 0-
(.1 "'I -. C 0--! OR
CC) CC) ~
0D0- C11) (.0 r-

in cm C.
TABLE 4
Stock price performance for Winners Corporation Surrounding
Spillman's Delinquent Schedule 13D Filing on January 6, 1988
(required filing on December 28, 1987)
Spillman
Cumulative Volume/ Volume/
Winners Abnormal Z- Abnormal Z- Mean Mean
Date Price Return statistic Return statistic Volume Volume
Dec. 11 1.500 8.71 1.62 8.71 1.62 0.25
Dec. 14 1.375 -10.92 -1.99 -3.16 -0.41 5.46 1.14
Dec. 15 1.375 -0.06 -0.01 -3.22 -0.34 1.02 0.57
Dec. 16 1.500 7.32 1.35 3.87 0.36 12.54 12.20
Dec. 17 1.500 1.88 0.35 5.82 0.48 1.12 0.57
Dec. 18 2.250 47.96 8.80 56.57 4.26 13.53 5.67
Dec. 21 2.000 -11.22 -2.09 39.01 2.72 4.27 1.14
Dec. 22 2.000 0.21 0.04 39.30 2.57 5.46 0.11
Dec. 23 1.875 -7.29 -1.35 29.15 1.80 2.68 0.23
Dec. 24 2.000 7.15 1.33 38.39 2.25 0.42 0.22
Dec. 28 1.875 -3.58 -0.66 33.43 1.86 1.31 0.57
Dec. 29 1.875 0.65 0.12 34.31 1.83 6.01 2.22
Dec. 30 2.250 19.00 3.52 59.82 3.07 3.42 0.68
Dec. 31 2.250 0.44 0.08 60.53 2.99 0.90 0.90
Jan. 4 2.375 2.16 0.39 63.99 3.05 1.15 0.79
Jan. 5 2.500 4.26 0.79 70.97 3.28 1.12 0.51
Jan. 6 3.000 19.98 3.72 105.14 4.71 4.66
Jan. 7 3.000 -0.53 -0.10 104.06 4.53 3.95 1.25
Jan. 8 2.875 2.23 0.38 108.61 4.58 0.78 0.34
Jan. 11 2.875 -0.99 -0.18 106.56 4.38 0.70 0.57
Jan. 12 2.750 -3.33 -0.62 99.67 4.00 5.23 3.97
Jan. 13 2.500 -9.05 -1.68 81.60 3.20 1.21 0.57
Jan. 14 2.625 5.20 0.97 91.05 3.50 1.98 0.57
Jan. 15 2.625 -2.23 -0.41 86.78 3.26 0.84
Jan. 18 2.500 -4.58 -0.85 78.24 2.88 0.06
Jan. 19 2.500 1.01 0.19 80.04 2.89 0.77
Jan. 20 2.500 2.73 0.50 84.95 3.01 0.01
Jan. 21 2.625 5.06 0.94 94.30 3.28 6.13 4.62
Jan. 22 2.750 3.72 0.69 101.53 3.48 0.11
Notes: Returns are expressed in percents. Stock price and volume data is from Center for
Research in Security Prices (CRSP) at the University of Chicago. Market model estimation
period is December 11, 1986 to December 10, 1987. Market proxy is CRSP value-weighted
index of NYSE, AMEX, and NASDAQ stocks. Beta estimate for Winners is 1.03.
9L-Uef-Oc
"0
4)
IE
9L-ue-C L
S0)
U
0 0
-u
u4 9L-GOC- LE
U
--w
E
CL
ow 9L-DOe-9 L
0.
52
U, m) U) W) U) N U) (0 U) U) U) v U) m'
0) 00 r (D Lo It V)

Article 11 PDF

Uploaded by

Copyright:

Available Formats

Article 11 PDF

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

Article 11 PDF

Uploaded by

Copyright:

Available Formats

DATE DOWNLOADED: Tue Nov 15 21:07:56 2022

SOURCE: Content Downloaded from HeinOnline

Bluebook 21st ed.

ALWD 7th ed.

APA 7th ed.

Chicago 17th ed.

McGill Guide 9th ed.

AGLC 4th ed.

MLA 9th ed.

OSCOLA 4th ed.

By Mark L. Mitchell and Jeffry M. Netter*

beneficial because they allow the investigator to discern whether infor-

THE BASIS FOR FINANCIAL ECONOMICS ANALYSIS IN

6. Daniel R. Fischel, Use of Modern FinanceTheory in Securities FraudCases Involving Actively

The fraud-on-the-market theory is useful to those attempting to satisfy

Standards for Materiality

preliminary merger negotiations. 21 The third standard, market impact, de-

Materiality and Financial Economics

the materiality requirement without examining security returns associated

30. See, e.g., JACOBS, supra note 17, § 61.04.

Securities Fraud Disgorgement Cases

Disgorgement in Schedule 13D Violations

65. SEC v. MacDonald, 699 F.2d at 47, 55 (lst Cir. 1983).

Stock prices rise in response to Schedule 13D filings.7 3 Therefore, share-

gorgement estimate, noting that the precise measure of ill-gotten gains

Disgorgement and Financial Economics

EVENT STUDY METHODOLOGY

80. Id. at 1231.

question (e.g., dividend announcement) and information that is likely to

DEFINING THE EVENT WINDOW

CALCULATE ABNORMAL STOCK PRICE

Calculationof Stock Returns

96. Researchers often express returns in logarithmic form as

-6.50% = [$46.125 - $43.125]/$46.125

Calculation of Standard Deviation

Testing the Statistical Significance of a Stock Return

torical average and standard deviation of the daily returns is highlighted

105. Macey, supra note 8.

Net of Market Stock Price Performance

new information. Often the information is firm specific in nature-e.g.,

puted-this sample standard deviation is 1.42%. °7 Thus, the net-of-market

Beta-Adjusted Stock Market Performance

which assumes that the return to a stock i at time t is a function of the

AR, = r, - (&, + ,R)

measures the impact of the event on stock i at time t.

September 30 is accounted for by market factors when it is adjusted with

Similar to the prior statistical tests, the abnormal return is compared

Cumulative Abnormal Returns

To illustrate the cumulation technique, consider the actual returns on

economics in securities fraud cases. Suppose an employee of Johnson &

USE OF FINANCIAL ECONOMICS IN SEC

at Artel may have provided signals to other investors that firm-specific

127. See supra text accompanying note 65.

OFFICER OF ACQUIRER SELLS TARGET

nouncement. Furthermore, some financial economists documented that

Consequently, when Reebok announced the merger pact with Avia at a

141. Slattery, 1988 SEC LEXIS 1758.

147. 15 U.S.C. § 78m(d) (1988).

CHAIRMAN LEAKS PRIVATE INFORMATION TO