Hanlon Dan Heitzman (2010)

Journal of Accounting and Economics 50 (2010) 127178
Contents lists available at ScienceDirect
Journal of Accounting and Economics

journal homepage: www.elsevier.com/locate/jae
A review of tax research$

Michelle Hanlon a,n, Shane Heitzman b
a
Massachusetts Institute of Technology, 100 Main Street, Cambridge, MA 02142, USA
b
University of Oregon, Eugene, OR 97403, USA
a r t i c l e in fo abstract
Available online 17 September 2010 In this paper, we present a review of tax research. We survey four main areas of the
JEL classication: literature: (1) the informational role of income tax expense reported for nancial
H20 accounting, (2) corporate tax avoidance, (3) corporate decision-making including
G10 investment, capital structure, and organizational form, and (4) taxes and asset pricing.
G30 We summarize the research areas and questions examined to date and what we have
M40 learned or not learned from the work completed thus far. In addition, we provide our
opinion as to the interesting and important issues for future research.
Keywords: & 2010 Elsevier B.V. All rights reserved.
Tax
Investment
Corporate decision-making
Earnings quality
Governance
Tax avoidance
1. Introduction
In this paper, we review tax research in accounting as well as tax research in economics and nance to the extent that it is
related to or is affected by research in accounting. Shackelford and Shevlin (2001) provide a careful and thorough review of
empirical tax research in accounting in the prior Journal of Accounting and Economics review volume. Shackelford and Shevlin
limit their review to research published in accounting outlets and describe the development of the relatively young archival,
microeconomic-based income tax literature that arose from the Scholes and Wolfson framework. In his discussion of the
review, Maydew (2001) emphasizes the need for tax researchers in accounting to think more broadly and to incorporate
more theory and evidence from economics and nance. We agree. Tax research has a long history in many disciplines; this
fact cannot be ignored. Our goal in this paper is to integrate the theoretical and empirical tax research from accounting,
economics, and nance, to summarize what is known and unknown, and to offer suggestions for future research.
The multidisciplinary nature of tax research is what makes tax research exciting (yes, exciting), yet difcult. Tax research
can be difcult not only because one has to follow tax studies in accounting, nance, economics, and law (through academic
institutions, governmental agencies, and policy think tanks), but also because different disciplines often use different
languages and have different perspectives.1 For example, economists generally focus on tax compliance, tax incidence
$
This paper was prepared for the 2009 Journal of Accounting and Economics Conference. We appreciate comments from Scott Dyreng, Mihir Desai
(discussant), S.P. Kothari (editor), John Long, Ed Maydew, Lillian Mills, Tom Omer, Sonja Rego, Doug Shackelford, Terry Shevlin, Joel Slemrod, Cliff Smith, David
Weber, Ryan Wilson, Jerry Zimmerman, George Zodrow, and conference participants at the JAE Conference as well as at the 2009 University of North Carolina
Tax Symposium and the discussant there, Bill Gentry. All errors and omissions are our own. Heitzman appreciates funding from the University of Rochester.
n
Corresponding author.
E-mail address: [email protected] (M. Hanlon).
1
Slemrod (1992a) characterizes these differences as economists studying what corporations actually do and accountants studying what companies say they
do. Gentry (2007) characterizes the difference between economists and accountants as economists believing that theory is reality and accountants believing that
perception is reality. The best, however, may have been when a law professor began his talk at a tax conference of economists, accountants, and lawyers by
stating, Maybe we should be like the U.N. Security Council and give each person earphones so the talks can be simultaneously translated into our own languages.
0165-4101/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jacceco.2010.09.002
128 M. Hanlon, S. Heitzman / Journal of Accounting and Economics 50 (2010) 127178
(such as who bears the corporate tax), investment and economic growth effects (such as how taxes affect investment),
and optimal tax policy (for example, whether a consumption or an income tax is better at minimizing distortions). In nance,
taxes are viewed as a market imperfection in a Miller and Modigliani type of world. This view leads to studies on whether
taxes affect rm value (e.g., whether dividend taxes affect expected returns), rm nancial policy decisions (e.g., whether
taxes affect a rms use of leverage), and investor portfolio decisions (e.g., the role of international tax considerations in
portfolio allocation).2
Tax research in accounting examines some of the same questions as tax research in economics and nance. In addition,
those working in accounting utilize specic knowledge of nancial accounting rules and an understanding of the
institutional details of tax and nancial reporting (i.e., our comparative advantages) to identify and examine research
questions. For example, divergent reporting incentives for tax and nancial accounting purposes lead to empirical studies
of the tradeoffs between tax costs and nancial accounting earnings. One outcome of this research is the ability to put
bounds on managers value of incremental accounting earnings because we can measure the cash tax cost incurred to alter
the measure of the earnings. Conversely, by understanding how managers balance tax incentives with external reporting
incentives, the same research can provide evidence on the effectiveness of tax policy.
Tax accounting researchers also leverage off another comparative advantage, an understanding of information
contained in the income tax disclosures in the nancial statements, to examine whether the tax accounts provide
incremental information about current and future earnings and rm value. In addition, we also use our understanding of
tax return data and the income tax data in the nancial reports to derive measures of rms tax avoidance activities and to
address important questions about the determinants and consequences of tax avoidance. Finally, accountants also have a
comparative advantage in studying information asymmetry problems (e.g., asymmetry between the rm and the state and
between shareholders and managers) where measurement and information are core issues (Slemrod, 2005).
Although we discuss research at the intersection of accounting, nance, and economics, we do not advocate that accounting
researchers broaden the research scope in ways that clearly do not reect a comparative advantage (e.g., corporate tax
incidence or the optimality of a consumption tax). Rather, we advocate that accounting researchers incorporate and/or extend
the theories and evidence from economics and nance that are relevant for tax research in accounting. Further, we encourage
accounting researchers to increasingly leverage our comparative advantages to examine real corporate decisions, issues
important in tax policy debates, and the incentive structures involved in corporate tax reporting.
To focus the paper and maintain a manageable length, we limit the review to certain areas of tax research. We focus on
tax issues facing businesses because this is where the majority of tax research in accounting is done and because research
on businesses overlaps with the scope of most other areas of accounting research. Even then, we are forced to retain only
certain areas. First, we discuss the corporate reporting of income taxes in the nancial statements and the information tax
disclosures provide about current and future earnings. In this section, we also discuss research on the policy proposal to
conform book and tax income measurement and review the empirical evidence on the effect of such conformity on the
informativeness of accounting earnings. Second, we discuss the recent theoretical models of corporate tax avoidance,
evaluate the empirical measures of tax avoidance, and summarize the recent evidence on the causes and consequences of
corporate tax avoidance. Third, we address the importance of taxes on business decisions by reviewing the literature on the
effect of taxes on business investment, corporate capital structure (including estimation of the marginal tax rate),
organizational form, and other decisions. Notably, we incorporate theory and evidence on the tradeoffs between tax and
nancial reporting incentives for these real decisions. Finally, we review the literature on the effects of investor-level
taxes on asset prices. A brief summary of our ndings and recommendations in our main areas follows.
First, the last decade has produced a great deal of evidence about the information contained in the income tax expense
and in book-tax differences reported in, or estimated from, the nancial statements. The information in tax expense or
book-tax differences, if any, derives from the idea that taxable income is an alternative measure of performance, or a
benchmark measure of performance for nancial accounting earnings. We emphasize that literature examining the
information in the tax accounts about current and future earnings constitutes a subset of nancial accounting research.
This stream of literature is not about tax avoidance. The research is often, although not always, conducted by tax
researchers because of a comparative advantage in knowledge of both tax and accounting rules. We provide a taxonomy of
the literature in an attempt to clarify some common misunderstandings. The evidence to date suggests that book-tax
differences provide information about current and future earnings (e.g., earnings persistence and future earnings growth)
and potentially indicate pre-tax earnings management. Recent research partitions the book-tax difference measures to
hone in on the underlying causes of these relations. An important issue in this area is that the literature employs several
measures of book-tax differences, but it is often not clear why a particular measure is used for the research question at
hand. We provide examples in our review.
Second, the corporate tax avoidance literature is young, but very active. Recent advances in the theory of corporate tax
avoidance in a principalagent (and principalagentgovernment) setting should help progress the literature along several
dimensions. Moreover, recent work broadens our thinking by incorporating the effects of corporate governance. Empirical
tests of these theories offer many potentially exciting avenues for future research. A primary issue in the empirical tax
2
We acknowledge that there is a large body of tax literature in the law eld as well. Much of that literature is focused on transaction planning at a
detailed level, but some is economics-based. We reference the economics-based literature throughout the paper where applicable.
M. Hanlon, S. Heitzman / Journal of Accounting and Economics 50 (2010) 127178 129
avoidance literature is the researchers denition and measurement of tax avoidance. We provide guidance in this area
by summarizing the various measures used, highlighting the difference between measures of conforming and
non-conforming tax avoidance, and discussing the benets and limitations of each with regard to the inferences that
can be made. Our key message is that researchers must carefully consider the underlying construct that is most
appropriate for their particular research question, select an empirical proxy (or proxies when appropriate) that best ts
that construct based on logical reasoning, and recognize the limitations on the inferences that can be drawn given the
measurement of the proxies and attributes of the sample.
Third, taxes potentially affect many real corporate decisions but their order of importance is still an open question.
We review the literature on the tax effects in such decisions and focus, in particular, on the interactions and tradeoffs
between tax and nancial reporting incentives for these real decisions. Shackelford and Shevlin (2001) summarize that the
eld has much evidence that under certain conditions, such as high debt levels, rms will tradeoff taxes for higher
accounting earnings when making reporting decisions and accounting method choices. Since Shackelford and Shevlin,
research has provided evidence that even rms that fraudulently report accounting earnings will at times pay taxes on
those earnings, thus sacricing cash ow to alter a reported accounting number (Erickson et al., 2004).3 In our review, we
focus more on the tax and nancial accounting tradeoffs involved in the operating and structural (real) decisions. We are
not saying the effect of book-tax tradeoffs on earnings management is an unimportant question. Rather, the unlocked
potential for future research is likely in the context of real corporate decisions. The key point we highlight in this area is
that taxes and the book-tax tradeoff can affect real decisions such as investment and capital structure which in turn affect
economic activity and have implications for the structure and efcacy of tax policy. In other words, nancial accounting
effects may mitigate or exacerbate tax incentives.4 Recent papers include the specic effects on the accounting for income
tax the effective tax rate and deferred tax accounts as a consideration in the tradeoff. The implications of the accounting
for income taxes have only recently been explored in this literature, but it appears from these studies that income tax
expense effects are important for tax policy efcacy as well. It is in this area of tax and accounting tradeoffs (or
interactions) for real decisions that tax accountants may have the greatest opportunities to contribute to tax policy
debatesdebates at which accountants are often absent. In light of the increasing national decit and tax-based stimulus
packages from the federal government, the effectiveness of tax policy has perhaps never been more important.
Finally, theoretical and empirical studies of the effect of dividend taxation on equity prices and rm behavior have been
conducted for well over thirty years. However, we still lack clarity on a number of fundamental research questions. Part of
the problem can be traced to inconsistent theoretical assumptions about the role of investor taxation in securities markets.
For example, an important issue is understanding the conditions for a capital market equilibrium under which an
identiable marginal investor matters in the sense of Miller and Scholes (1978) or all investors matter in the sense of
Brennans (1970) after-tax CAPM. A related debate is whether institutional ownership can be used to identify the tax
characteristics of investors. We provide an overview of the various motivations for, and the approaches taken in, studying
the effects of dividend taxation. The literature examining capital gains taxation has arguably had more success
documenting the role of taxes in pricing and trading decisions.
As we move beyond the set of established research areas, we believe there are a number of open areas to explore.
First, we do not have a very good understanding of loss rms, the utilization and value of tax-loss carryforwards, and how
the existence of losses affects the behavior (e.g., tax and accounting reporting and real decisions) of any of the involved
parties. Obvious settings where losses may be important include start-up rms, ownership changes (including merger and
acquisition transactions), and nancial crises.5
Second, we do not have much evidence on the taxation of nancial securities or nancial institutions. Often, nancial
institutions are dropped from the sample because of concerns over regulatory differences. After reviewing the literature, it
seems clear that we do not even know much about the variety of leases used and even less about the tax (and accounting
for tax) implications of more sophisticated nancial instruments. Third, more work on privately held rms may be
important beyond using them as a comparison group for publicly held rms. These rms have different ownership
structures, different nancial reporting incentives, and constitute a large portion of our economy.
3
Erickson et al. (2004) identify a setting and structure tests that avoid the need to compute as if tax and accounting numbers (as if the rm had
made a different decision) which Shackelford and Shevlin argue biased results in prior research. See also Badertscher et al. (2009) for an extension of
Erickson et al. to restatement rms. Note also that Seida et al. (2005) develop a model of tax motivated intertemporal income shifting that relates the
costs associated with shifting income to the amount of income shifted as called for by Shackelford and Shevlin (2001).
4
That tax and non-tax tradeoffs are important in these real (versus reporting) decisions is not a new idea and we do not claim rights to it. For
example, some book-tax tradeoff studies about a rms decision to use LIFO inventory methods involved tests of additional purchases of inventory or
inventory liquidations. Shackelford and Shevlin also discuss tradeoff papers dealing with compensation structure (e.g., Matsunaga et al.,1992), capital
structure (e.g., Engel et al., 1999), and asset sales and divestitures (e.g., Bartov, 1993; Klassen, 1997; Maydew et al., 1999). See Shackelford and Shevlin
(2001) for a review. However, the implications beyond a book-tax tradeoff were not generally emphasized or at times even discussed in these prior
papers. Neubig (2006a) discusses the policy effects of nancial accounting concerns and Shackelford et al. (2010) model and discuss the option value of
tax and accounting reporting exibility (we discuss these papers more below).
5
See Erickson and Heitzman (2010) for a discussion and descriptive evidence on rms that protect their net operating loss carryforwards using
shareholder rights plans, namely poison pills.
Finally, research involving international tax issues has increased over time (we incorporate this research throughout
our review) and will continue to be in demand. Studies on both foreign direct investment and foreign portfolio investment
have grown substantially in recent years and will likely continue to grow. How taxes affect these investment decisions are
important topics for many reasons, one of which is the design of tax policy for worldwide activities. It is important to note
that even though many companies now think globally, tax rules and regulations are still determined on a national basis.6
As a result, much of the research is country specic. Looking forward, the potential future adoption of International
Financial Reporting Standards (IFRS) in the U.S. may change taxable income to the extent book and tax reporting are
aligned and may affect book-tax differences, LIFO inventory accounting (unless otherwise repealed rst), and transfer
pricing comparisons.
The paper proceeds as follows: we begin by discussing the nancial reporting of corporate income taxes and the
informational role of the income tax expense disclosures in Section 2 (we review the accounting for income taxes under
GAAP in Appendix A). In Section 3 we review the theory and evidence of corporate tax avoidance, and provide a detailed
discussion of the empirical measures of tax avoidance. In Section 4, we discuss the literature on the effect of taxes and the
book-tax tradeoff on real corporate decisions, including investment, nancial policy, and organizational form, among
others. In Section 5 we critique the literature on investor-level taxation and asset prices. In Section 6, we conclude.
2. The informational role of accounting for income taxes
The income tax expense occupies just one line on the income statement; footnote disclosure provides additional detail. If
one thinks about the comparisons in the nancial accounting literature of accrual earnings to cash ows, one can imagine a
similar comparison to taxable income as a type of benchmark for accrual accounting (pre-tax) earnings. In addition, because
the income tax expense is an accrual-based expense, portions of it can potentially be manipulated to affect after-tax earnings.
In this section, we discuss the primary conceptual sources of differences between book and taxable incomes (a brief
discussion of the rules is in Appendix A). We then present a taxonomy of the studies of GAAP income tax disclosures to date
and a summary of what we have learned from these studies within the context of the larger literature on the information
content of accounting information.7 Finally, we conclude the current section with a discussion of policy calls for conformity
of income reported for both tax and nancial reporting purposes. This policy option has important potential consequences
for nancial reporting about which nancial accountants should be cognizant. We want to emphasize that this entire
section and the literature reviewed herein are about nancial accounting implications and inferences and not about
corporate income tax avoidance. We review the literature on tax avoidance in Section 3.
2.1. Sources of differences between book and taxable income
Conceptually, the differences between taxable income and accounting income, i.e., book-tax differences, are driven by a
wide variety of factors. The most basic of these is that the two systems have very different objectives and these different
objectives lead to different rules.8 In theory, nancial accounting standards follow from a conceptual framework in which
the objective of GAAP is to capture the economics of transactions in order to provide useful information for decision
makers, such as equity investors and contracting parties. Tax rules are written under a much more political process.
Lawmakers can enact tax rules to raise revenue, encourage or discourage certain activities, and attempt to stimulate the
economy. The calculation of taxable income is accrual based (for large, non-agricultural companies) but is really a hybrid of
cash basis and accrual basis that does not allow companies to estimate expenses in advance of the cash payments or defer
revenue received until earned. In addition, the tax rules focus on the location of the earnings so that the appropriate
jurisdiction(s) can tax the income. In contrast, for nancial accounting, a consolidated nancial statement includes all
income (and losses), no matter where earned, from controlled entities. The two income measures are linked, or aligned, to
some degree but do not need to conform for many transactions (we discuss the policy proposal of book-tax conformity
below).
Another potential source of differences between accounting earnings and taxable income, at least a suspected source, is
aggressive reporting for book or tax purposes. For example, when managers manipulate earnings upward, they may have a
choice between reporting taxable income at the higher amount and paying taxes on the inated earnings or reporting taxable
income at the unmanaged, lower amount and recording a book-tax difference in the nancial statements. On one hand, the
least costly method of managing earnings would be to record a book-tax difference and not report the managed earnings in
taxable income because this saves cash taxes. On the other hand, if book-tax differences provide information to the market
about the earnings management, then recording the book-tax difference may reduce the credibility of reported earnings.
6
However, tax authorities around the world are coordinating more and the IRS Commissioner has indicated that joint audits with other countries tax
authorities are being planned. See Commissioner Doug Shulmans speech to the New York State Bar Association Taxation Section Annual Meeting in
New York City, January 26, 2010
7
Graham et al. (2009) provide a detailed review of many of the papers in the accounting for income tax area.
8
While it is known that the rule differences contribute a great deal to book-tax differences, Seidman (2009) attempts to quantify the extent to which
the rule differences drive the amount by looking at changes to the accounting rules.
An example on the tax side is a rm that is aggressive for tax reporting purposes and takes action to lower reported
taxable income. The question would then be whether the rm reports a similar position for nancial accounting purposes
and reports lower income to shareholders. On one hand, the least costly path may be to report the book-tax difference and
get the best of both worlds (i.e., low taxable income and high accounting income). On the other hand, if the taxing
authority uses the book-tax differences as an indicator of some form of tax aggressiveness, this increases the cost of the tax
aggressiveness by increasing the likelihood of detection.9
For any specic rm at any time, it is likely that all of these factors are operating simultaneously to some extent to
generate the level of book-tax differences the rm reports. Most authors recognize the many possible determinants of
book-tax differences and attempt to develop models or identify settings where a particular determinant can be studied or
identied.
2.2. A taxonomy of the literature and review of the evidence
In our view there are at least three categories of, or motivations for, research on the accounting for income taxes.
The rst category consists of studies about whether and to what extent tax expense and tax disclosures provide
information about current or future earnings (including the markets interpretation of the information). The underlying
theory here is that taxable income may provide an alternative measure of performance, less subject to earnings
management. The second category of studies examines whether the income tax accruals (e.g., valuation allowance,
contingency reserve) are used to manipulate after-tax earnings.10 The third (and much smaller) category examines
whether changes in the valuation allowance reveal inside information about the future earnings of the rm.
2.2.1. Inferences from book-tax differences about current and future earnings.
The idea that book-tax differences are informative about earnings quality in pre-tax accruals has been around for a
while and was formalized to some extent in accounting textbooks such as Revsine et al. (1998), Palepu et al. (2000), and
Penman (2001). The argument is that accounting accruals reect more discretion than the tax laws allow, thus, temporary
differences between book and tax income reveal something about discretion in non-tax accounting accruals (e.g., bad debt
accrual, warranty expense, deferred revenue, etc.). This theory does not generally extend to permanent differences because
permanent differences are not driven by the accounting accruals process.
The typical suspicion is that rms that report higher pre-tax book income than taxable income have manipulated book
income upward. Indeed, after the Enron and WorldCom scandals, some policy makers and members of the press suggested
that there should be additional disclosure of book-tax differences and that companies with large book-tax differences
might be doing some manipulation somewhere, either for book or for tax or for both.11 However, most studies are agnostic
in making predictions based on the sign of book-tax differences because negative book-tax differences (where book income
is less than taxable income) could generate earnings quality concerns as well. For example, rms taking a big bath or
using cookie-jar reserves often cannot deduct for tax purposes the write-offs and expense accruals involved, thus
generating negative book-tax differences.
A variety of empirical approaches have been employed in an attempt to examine the information in book-tax
differences for earnings quality. Overall the results are quite consistent across the studiesbook-tax differences contain
information about nancial accounting earnings quality. For example, rms that report large (in absolute value) book-tax
differences appear to have weaker earningsreturn relations (Joos et al., 2000) and total book-tax differences are positively
associated with prior earnings patterns, nancial distress, and bonus thresholds (Mills and Newberry, 2001).12 In addition,
the evidence suggests that in rm-years with a small earnings increase (a la Burgstahler and Dichev, 1997), there is a
relatively larger deferred tax expense (Phillips et al., 2003), suggesting that deferred tax expense is (temporary book-tax
differences are) informative about pre-tax earnings management. The evidence in Hanlon (2005) indicates that earnings
and the accrual portion of earnings are less persistent when rms have large (in absolute value) temporary book-tax
differences. This implies that book-tax differences have information about the persistence of pre-tax earnings and pre-tax
accruals. Hanlon (2005) also examines whether the market understands that when rms have large book-tax differences
the accrual portion of earnings is less persistent than when book-tax differences are small.13 The results suggest that it
does and Hanlon (2005) concludes that the reported book-tax differences are potentially a red ag for investors.
9
See Cloyd et al. (1996) and Mills (1998) for related research.
10
See Healy and Whalen (1999), Schipper and Vincent (2003), and Dechow et al. (2010) for reviews of the earnings quality and earnings management
literatures.
11
See Grassley and Charles (2002), Weisman (2002), and Reason (2002).
12
The paper by Mills and Newberry (2001) shows that book-tax differences are less informative about tax positions for private rms because private
rms will more likely conform book and tax, reducing the size of the book-tax differences for a given tax position. Conversely, public rms have higher
non-tax costs and thus report higher book-tax differences. Mills and Newberry (2001) use a measure of temporary and permanent book-tax differences
between pre-tax earnings and taxable income.
13
There are limitations to the paper such as the short time period, the possibility of correlated variables with book-tax differences (in other words, it
could be something else that tells investors about the lower persistence, not the book-tax differences), and general problems in estimating book-tax
differences, even temporary differences.
It is important to note two points from the latter two papers discussed above. First, the book-tax differences in Phillips
et al. (2003) and Hanlon (2005) consist only of temporary differences because these are the differences most closely related
to pre-tax accruals and most often hypothesized to provide information about pre-tax earnings quality. Second, these two
papers do not examine whether deferred tax expense itself is managed. Rather, the papers examine whether the deferred
tax expense can inform us about pre-tax earnings quality or management.
Lev and Nissim (2004) document a positive relation between future earnings growth and the ranked ratio of taxable
income to book income. The authors compute the ratio using after-tax earnings and adjust taxable income to be after-tax
by reducing it by a tax computed using the statutory rate (i.e., a cross-sectional constant). The authors also test a ranked
variable of deferred tax expense (dened as the ratio of deferred tax expense to assets) but nd only weak evidence of a
relation with future earnings growth. Thus, the authors conclude that their measure of total book-tax differences holds
information about future earnings growth but temporary differences do not. Lev and Nissim also report that returns can be
predicted using the tax-to-book ratio.14 Examining a different market participant, credit ratings agencies, Ayers et al.
(2010) nd that revisions in credit ratings appear to be more negative when rms have large changes in book-tax
differences.15
As discussed above, book-tax differences can arise from a variety of sources. Recent studies attempt to further
investigate why book-tax differences are associated with earnings properties and with market participants valuations
(e.g., lower persistence, the relation with credit ratings, etc.) by identifying rms more likely to have managed earnings and
those more likely to have tax planned or avoided taxes. One way to disentangle the rms is to sort observations by the level
of book-tax differences and then by the extent of total or discretionary accounting accruals. Following this approach, Ayers
et al. (2010) nd that the association between credit ratings and book-tax differences is weaker when book-tax differences
are driven by tax planning. Employing the same partitioning methodology, Blaylock et al. (2010) report evidence that
Hanlons (2005) results are driven by rms more likely to have managed earnings upward.16
One issue in the literature is that different studies use different measures of book-tax differences. For example, some
focus on temporary differences (Phillips, 2003; Hanlon, 2005), while others focus on total differences between tax and
pre-tax book income (Mills and Newberry, 2001), and yet others focus on total differences between after-tax income
measures (Lev and Nissim, 2004).17 It is certainly appropriate to adopt different measures for different research questions,
but it seems that studies often adopt different measures to address similar questions without providing a reason why.
We discuss the specic measures and links to earnings quality next with the hope of clarifying the issues.
The main issue for research design is that most measures of a total difference between book income and taxable
income will usually contain many items that are not really book-tax differences (e.g., tax credits which do not affect either
income measure) and it will include permanent differences that are not related to accounting accruals (e.g., municipal
bond interest). The concern is over interpretation. Temporary differences reect information about pre-tax accruals and
are consistent with the theory that book-tax differences provide information about management in other non-tax
accrualsbad debt, warranty expense, etc. Measures of total book-tax differences include much more than temporary
differences, which may be useful in certain research settings (e.g., when examining market participants use of the
information). However, total book-tax differences are a very noisy proxy to use to test whether the information in the tax
expense is telling us anything about earnings management in other pre-tax accrual accounts. Thus, studies that use total
book-tax differences cannot directly draw inferences about pre-tax earnings quality (especially if consistent results are not
obtained when temporary differences are tested alone). In other words, the research question should dictate the measure
used (e.g., is the interest in pre-tax accruals or after-tax income?) and the measure used will dictate the inferences that can
14
Weber (2009) examines whether Lev and Nissims return results are attributable to mispricing or risk. Weber concludes that the result is due to
mispricing because the association between the ratio and returns only exists for rms with weaker information environments and disappears when
controls for analysts forecast errors are included.
15
Schmidt (2006) seems similar to the above but in fact is different in many regards. Schmidt examines whether ETR changes are transitory or have
some persistence and whether the market understands the implications for future earnings. He uses the interim reporting requirements of the nancial
accounting rules (APB 28 and SFAS 1977) to decompose the change in the ETR into an initial change (rst quarter) and revised changes (quarters 24).
Schmidt (2006) nds signicant positive association between the tax change components of earnings and that the initial (Q1) change to the ETR is more
persistent than a change in the other quarters. He also documents that the market underweights the forecasting ability of the tax change component of
earnings but that the mispricing seems to be in all the later quarters. The paper does not hinge on the ETR revealing information about other accounts,
which then provides information about future earnings properties. Schmidt (2006) is a nice contribution to the literature but is not in the spirit of the
literature above, however, and thus, is not included in the discussion.
16
The strategy of partitioning rms into those more likely to be tax planners and those more likely to be managing earnings is from Ayers et al.
(2009) where the research question is about the relative information in taxable income versus book income. The authors extend Hanlon et al. (2005) and
document that taxable income has more incremental information relative to book when tax planning is less likely and earnings management is more
likely. In general, we think this split of rms is useful and further innovations including more detailed partitioning or identication of the sources of book-
tax differences will likely further our understanding. However, we recognize that categorizing book-tax differences by source or management intent for
the transaction that generated the difference is difcult. In addition, identifying individual book-tax differences from the tax footnote is complicated
because only the balance sheet amounts are disclosed and changes in these accounts do not often tie to the income statement effect (likely due to merger
and acquisition activitysee Collins and Hribrar (2002) and Hanlon (2003)). Moreover, which items rms disclose in the tax note is to some extent a
choice variable.
17
In addition, while most papers use levels, some papers use changes in book-tax differences. See Phillips et al. (2003) for a discussion of changes
versus levels. Ayers et al. (2009) use changes because they are examining changes in credit ratings. Further yet, some papers compute discretionary
book-tax differencesan issue we address more fully when reviewing the tax avoidance literature.
be drawn. In our opinion, the difference in inferences is an important issue to recognize in papers using book-tax
differences.
2.2.2. Are earnings managed through the tax accounts?

At least three tax-related items are thought to be available for earnings management: the valuation allowance, the tax
contingency reserve, and the amount of foreign earnings designated as permanently reinvested. Note that this set of the
literature is not about inferring information about pre-tax accruals, but testing whether managers use the tax accounts to
manage earnings. The earnings management literature is large, with a mixture of analyses of summary measures of
accruals and individual accounts (see Dechow et al., 2010 for a review). The literature we discuss here includes both types
of analyses.
Why would the tax expense and the related accounts be different from other accounts and do they warrant a separate
investigation? First, the items are separately disclosed (except the tax contingency reserve prior to FIN 48, discussed
below) and thus provide specic accounts to test for earnings management (fullling the request by Healy and Whalen,
1999). Second, changes in these accounts affect earnings dollar for dollar (i.e., there is no tax effect). Third, these accounts
affect only after-tax earnings. Thus, to the extent debt and compensation contracts are based on pre-tax earnings these
accounts may be managed less often for contracting reasons, enhancing the ability to detect the effect of pure stock market
pressures. Finally, the within-rm dynamics may be different for these accounts as they may require sign-off by (or at
least input from) the tax department. Thus, the accounts are sufciently different to warrant independent study.
Dhaliwal et al. (2004b) study earnings management through the total tax expense line (i.e., through the GAAP ETR
rather than through any one component of the expense). The authors provide evidence that rms lower the projected
GAAP ETR from the third to the fourth quarter when the company would otherwise miss the analyst consensus forecast,
implying that managers lower accrued tax expense to meet analysts forecasts.18 However, the change in GAAP ETR could
also arise through tax planning transactions.19 For example, Desai (2003) provides specic examples of rms engaging in
tax shelters where the main objective is to increase accounting earnings. Evidence in Cook et al. (2008) suggests that the
results in Dhaliwal et al. are caused by both management of tax accruals and tax avoidance behavior, where tax avoidance
is proxied for by high tax fees paid to the auditor.20 We turn next to the evidence on earnings management through specic
tax accounts: the valuation allowance, the tax contingency reserve, and the designation of foreign earnings as permanently
reinvested.
The valuation allowance is a contra-asset established against deferred tax assets based on managements expectations
of future taxable income amounts and timing. The justications for changes in the account are difcult for the auditor to
verify (because managers use expectations of earnings for many years into the future), opening the door for manipulation
(Miller and Skinner, 1998). Changes in the valuation allowance affect earnings directly by changing the amount of deferred
tax benet recognized. Specically, an increase in a valuation allowance for a given deferred tax asset will decrease
earnings (decrease the amount of deferred tax asset recognized) and a decrease in a valuation allowance for a given
deferred tax asset will increase earnings.
Evidence of earnings management through the valuation allowance is somewhat mixed. Miller and Skinner (1998)
nd no evidence of earnings management in this account, while Bauman et al. (2001) nd limited evidence and conclude
that earnings management in the account is not widespread.21 However, Schrand and Wong (2003) report that banks
manage the valuation allowance to achieve analysts consensus forecasts and historical earnings per share targets in some
cases. Frank and Rego (2006) present more comprehensive evidence consistent with managers using the valuation
allowance to meet analyst forecasts, but do not nd evidence of earnings management through the account for any other
target.
These papers contribute to the literature by showing that a specic tax account that potentially involves a great deal of
manager discretion is used in certain contexts to manage earnings, although not to the extent some feared when the
standard was set. In our opinion, the incremental contribution of additional studies on the use of the valuation allowance
to manage earnings will be limited. There are likely more interesting issues to address. For example, Skinner (2008)
examines consequences of the inclusion of deferred tax assets (with no valuation allowance) as part of regulatory capital
for nancial institutions in Japan. In the U.S., rules in place since 1995 only allow a small portion of deferred tax assets to be
used as regulatory capital for banks because of the subjectivity and uncertainty in valuation. Recently, groups have
18
The accounting for the effective tax rate falls under APB 28 Interim Financial Reporting, which requires rms to adjust the GAAP ETR each quarter to
the estimated annual GAAP ETR.
19
An interesting follow-up study by Gleason and Mills (2008) provides evidence consistent with the market understanding and discounting the
reward to rms for meeting a target through managed tax expense; however, note that the company is still rewarded more than having missed the target
so managing the expense is not useless even though the market discounts the reward.
20
This idea of actual tax planning as opposed to accrual management being used to manage the tax expense is similar in spirit to Roychowdhurys
(2006) examination of real transactions for earnings management more generally.
21
Miller and Skinner (1998) state that their primary question is whether managers comply with the provisions of SFAS 109 and thus conclude that
their earnings management tests are not powerful. They study 200 rms with large deferred tax assets, not rms with incentives to manage earnings. The
Bauman et al. (2001) paper provides detailed discussion on how to measure the earnings effect of a change in the valuation allowance and concludes that
not all changes in the contra-asset account result in an income effect as had been assumed in prior studies, see their paper for further discussion. See also
Visvanathan (1998) for more evidence on valuation allowances and earnings management.
questioned whether this limitation should be lifted.22 Additional research would be useful in assessing nancial
institutions setting of valuation allowances and, if the rules change, an examination of rm behavior at that time could
prove useful to such regulatory policy debates.
The next account potentially used to manage earnings is the tax contingency reserve. Historically, the tax contingency
reserve, previously known as the tax cushion, was determined under SFAS No. 5Accounting for Contingencies. Additional
guidance was recently issued in 2006 in FASB Interpretation Number 48 (FIN 48), Accounting for Uncertainty in Income
Taxes.23 In essence, the standard and interpretation require managers to accrue a liability for tax assessments from
expected challenges to tax positions taken in the current and prior periods. The following is a simplied example. Suppose
a company takes a risky tax position and reduces its tax liability by $1,000. Management is required to estimate how much
additional tax would likely be assessed upon a future audit. Suppose they estimate the likely additional tax liability to be
$200.24 The company then records an additional $200 of income tax expense and a corresponding accounting liability even
though the company did not pay this amount to the tax authorities and the position has not yet been challenged. This
adjustment to tax expense is recorded in the current tax expense (although if the position relates to a temporary
difference, it is at times recorded in deferred tax expense25). The magnitude of the effect on current tax expense, and
resulting estimates of taxable income, are unknown before FIN48 and difcult to correct.
Unlike the valuation allowance, the tax contingency reserve is not limited by the amount of deferred tax assets, was
very rarely disclosed prior to FIN 48 (Gleason and Mills, 2002), and involves more managerial discretion.26 Thus, it is an
account potentially available for earnings management. However, whether it is managed is hard to test because of lack of
disclosure before FIN 48. Evidence in Gupta and Laux (2008) suggests that the pre-FIN 48 tax contingency reserve, when
disclosed by rms, was used to manage earnings to meet or exceed analyst forecasts.
FIN 48 requires disclosure of the tax contingency reserve, renamed in the interpretation as unrecognized tax benet
or UTB.27 Recent disclosures by rms following the implementation of FIN 48 reveal that this account is large for many
rms. For example, as of January 1, 2007, Merck disclosed $7.4 billion in unrecognized tax benets; General Electric
disclosed $6.8 billion; and AT&T disclosed $6.3 billion.28 Some early papers on FIN 48 nd that rms have incentives to
reduce their tax contingency reserves to avoid disclosure under FIN 48, and that doing so prior to the adoption of the new
standard allowed rms to increase earnings rather than record an adjustment to stockholders equity (Blouin et al., 2010).
The nal suspected method of managing earnings through the tax accounts is via the designation of permanently
reinvested earnings. Firms can potentially manage earnings by either increasing or decreasing the amount of certain
foreign source earnings designated as permanently reinvested (see Appendix A for details on the accounting). Krull (2004)
nds that managers use discretion in the designation of permanently reinvested earnings (PRE) to manage reported
earnings to meet analyst forecasts but not other targets.
Overall, there is some evidence that the tax expense accrual is used to manage earnings. As rules change through time
for computation, disclosure, or other factors (e.g., the inclusion of deferred tax assets as part of regulatory capital), the
extent to which rms can manage earnings through these accounts may change signicantly. Thus, while further
contributions in the earnings management through tax accounts area are hard to envision, constantly changing rules
surrounding the accounts may create opportunities for further study. Another consideration is that for any of these
accounts there will be real transactions that can masquerade as earnings management that researchers need to be
22
These concerns were expressed in a September 25, 2009 letter from the The Clearing House Association and the American Bankers Association,
which together represent nearly all U.S. depository institutions, to the four agencies responsible for determining regulatory capital (Board of Governors of
the Federal Reserve System, the Federal Deposit Insurance Corporation, the Ofce of the Comptroller of Currency, and the Ofce of Thrift Supervision). The
letter addresses the agencies earlier concerns over the proper valuation of deferred tax assets.
23
FIN 48 became effective in the rst quarter of 2007 for most calendar-year companies. This interpretation was issued because the FASB was
concerned that diverse accounting practices had developed resulting in inconsistency in the criteria used to recognize, derecognize, and measure
benets related to income taxes (FIN 48). The interpretation uses the terminology unrecognized tax benets (UTBs) for the portion of the tax benets
from uncertain tax positions that are not recognized for nancial reporting purposes (instead of the old terms tax cushion or tax contingency reserve).
The new disclosure requirements include a tabular reconciliation of the beginning and ending balances of the UTBs, expected future changes in UTBs, and
the amount of the UTBs that would affect income if released.
24
When the account was governed under SFAS No. 5 it was not clear exactly how the probability of audit should be taken into account when
assessing the amount of reserve to record and thus rms adopted a wide range of practices with regard to audit probability. Under FIN48, the company is
to assume the tax authority will audit the position.
25
Although a reading of SFAS 109 would seem to indicate that all of the tax contingency reserve should be in current tax expense, anecdotally we
know that some amounts are recorded in deferred tax expense. If recorded in current tax expense before FIN 48, the liability may have been recorded in
an other liability account. See for example Microsofts 2005 annual report in which the company discloses that it had $3 billion in tax contingency
liabilities and $961 million in legal contingencies included in other long-term liabilities. Most companies did not provide such detailed disclosures
pre-FIN 48.
26
See Gupta and Laux (2008) for a discussion of the tax contingency disclosure and recognition requirements prior to FIN 48. Gleason and Mills
(2002) provide one of the earliest studies on the tax contingency reserve and show that rms often do not disclose contingent tax reserves (estimated
using tax return data relative to nancial statement data) even when the estimated reserve amount met general materiality standards. However, Gleason
and Mills did not examine whether the account was used to manage earnings.
27
See FIN 48 for a detailed discussion. FIN 48 changes the threshold for disclosure and recognition of tax positions. Moreover, it changes the
measurement of the tax liability the rm must recognize. Under FIN 48, the rm records the liability for uncertain tax positions meeting the threshold as
unrecognized tax benets. We discuss the use of FIN 48 as a proxy for tax avoidance below.
28
Source: Credit Suisse Peeking Behind the Tax Curtain by David Zion and Amit Varshney. May 18, 2007. For the 361 calendar-year-end rms in the
S&P 500 for which Credit Suisse could obtain disclosures, the total unrecognized tax benets as of January 1, 2007 was $141 billion.
cognizant of when designing and interpreting empirical tests. For example, most valuation allowance studies control for
the economic factors outlined in SFAS 109 for establishing the account, in particular the role of future expected
performance. In a more specic example, as mentioned above, there is evidence that rms released a signicant portion of
tax contingency reserves prior to adoption of FIN 48, which increased accounting earnings and has been attributed to
earnings management (Blouin et al., 2010). However, rms also appear to be settling with the tax authorities during this
time period in attempts to clear the slate and resolve the uncertainty before the implementation of FIN 48. Blouin et al.
document this economic explanation as well.
A deeper understanding of institutional details will allow us to reconcile data and results across studies. For example,
Graham et al. (2010b) report that 75% of the rms in their sample designate 100% of unremitted foreign earnings as
permanently reinvested (PRE). If these data are correct and generalizable, it suggests that there is little practical variation
in the discretion applied to the designation. Thus, while Krull (2004) argues that rms use discretion within the
account to manage earnings (and we do not dispute her results), such a conclusion does not easily square with the
observation that the PRE designation seems to be xed in many rms. An alternative explanation is that to
manage earnings rms must repatriate (or reinvest) more cash than they otherwise would which would be a real
decision rather than a simple accounting designation. Another possible explanation is that accounting policies at
companies may have become xed over time if managers have come under greater scrutiny about their PRE
designations post-SOX than when her tests were conducted. Thus, the difference above could be a result of different
sample periods.
2.2.3. Do changes in the valuation allowance reveal managers private information about future performance?
Changes to the valuation allowance depend in part on the managers expectation of future taxable income and the
likelihood that deferred tax assets will be realized in the future. To the extent taxable income is an alternative measure of
performance, changes in the valuation allowance can provide forward-looking information about future economic
performance. Anecdotes and examples in the press and nancial accounting textbooks suggest that a rms valuation
allowance can be indicative of future earnings problems.29 There is some empirical evidence that valuation allowance
changes have incremental predictive ability for future earnings and cash ows (Jung and Pulliam, 2006), and that investors
appear to use valuation allowance disclosures to infer managements expectations of future performance (Kumar and
Visvanathan, 2003).30
2.3. Book-tax conformity
There is an ongoing debate in the tax literature and among policymakers about whether or not the U.S. should conform
its two sets of income measures book and tax into one common measure.31 The most common proposal is to tax
accounting earnings (or slightly adjusted accounting earnings). Proponents claim benets such as reduced compliance
costs, a broader base and lower rate, possible elimination or at least downsizing of the IRS, and a forced tradeoff such that
rm management will tell the truth when they report the income number.32 For example, management will not want to
overstate earnings or the rm will have to pay more tax, and they will not want to understate earnings to save taxes if
shareholders punish management for reporting a lower number. These benets, if attainable, would be hard to argue
against. Opponents claim that conformity would not be as simple as the proponents argue, the rate would not be able to be
lowered as much as claimed because of the behavioral response on the part of rms, and a substantial cost would be
incurreda decrease in the information contained in the accounting earnings number.33
While there are many differences between book and taxable incomes in the U.S., it is important to note that the two
systems are at some level currently aligned. Indeed, there is a stream of literature that tests whether earnings management
occurs around tax acts (i.e., companies shift taxable income to be in lower-taxed years and the evidence of shifting is
signicant in nancial accounting earnings). Examples of these studies include Scholes et al. (1992), Guenther (1994) and
Maydew (1997). Thus, while the U.S. has a dual reporting system and the two income amounts can, and often do, diverge a
great deal, there is basic alignment in the fundamental accounts (e.g., many types of sales, much of cost of goods sold, etc.).
The experience around the world varies. For example, Germany historically has had a much higher degree of book-tax
conformity than the U.S. (see Harris et al., 1994). However, many European practices involve maintaining one set of
company accounts the single accounts that are conformed, and one set of accounts the group accounts that are not
conformed. While many point to higher conformity around the world, the single accounts which have the greater
29
Common examples are Bethlehem Steels 2001 valuation allowance increase (see Weil and Liesman, 2001) and GMs valuation allowance increase
of $39 billion in 2008 (Green and Bensinger, 2007).
30
Prior studies document a relation between valuation allowance changes and contemporaneous stock prices (e.g., Amir et al., 1997; Ayers, 1998;
Amir and Sougiannis, 1999).
31
Other countries have either moved away from a conformed system or more toward a conformed system over time. In addition, upon the adoption
of International Financial Reporting Standards the European Union (EU) seriously considered making the income measured under IFRS the common
consolidated tax base. The EU opted not to adopt such a plan at this time (see Schon, 2005 for a discussion).
32
See Desai (2003, 2005, 2006), Graetz (2005), Murray (2002), Rossotti (2006), and others.
33
See Shackelford (2006), Hanlon et al. (2009), Hanlon and Shevlin (2005), and McClelland and Mills (2007).
conformity are not those presented to the capital markets in many cases. Thus, in some sense the companies still have two
sets of books.34
In examining what would happen here in the U.S. if book-tax conformity were adopted, the ideal research design cannot
be employed since the U.S. has not switched from a full book-tax conformity system to a non-conformed system (or vice
versa). In addition, to properly analyze all aspects one would have to be able to obtain an accurate measure of the
compliance costs for tax and accounting purposes both before and after the change. In the absence of such a natural
experiment, how has the issue been studied? Most of the research to date has primarily focused on one testable item:
whether there is a decrease in information in accounting earnings after book-tax conformity increases. Alternative
approaches to testing this question include cross-country studies, comparisons of information contained in book and
estimated taxable incomes, examinations of settings in which conformity changed, and simulations.35 While there are
weaknesses with each of these approaches, they all essentially document a substantial cost of conformity in terms of a loss
of information in nancial accounting earnings. In addition, the simulations show that the tax rate necessary to achieve
revenue neutrality is higher than the attainable rates suggested by proponents of book-tax conformity (see Hanlon and
Maydew (2009) who estimate a 26% revenue neutral rate and McClelland and Mills (2007), who estimate a 2835%
revenue neutral rate).
For tax accountants and nancial accountants, book-tax conformity is an important proposal. In some ways, it is more
important for nancial accounting than for tax. Depending on how much inuence governments have on the standard
setters of the conformed income, the number reported (and process to determine the rules to compute the number) could
be closer to the current taxable income than current book income.36 If that were to occur, much of the philosophy of the
FASB and the SEC (i.e., having an independent standard-setting board, full disclosure rules, etc.) would be eliminated (see
Hanlon and Shevlin (2005) for a discussion).37
President George W. Bushs Tax Reform Panel suggested the topic of conformity be studied further. What additional
research is needed? The evidence suggests there will a substantial cost in terms of the information loss in accounting
earnings should book-tax conformity be adopted. We have little evidence about anything else. Some think that the
potential adoption of International Financial Reporting Standards (IFRS) will open the door to a one accounting standard
world, which could then open the door to a worldwide-consolidated tax base (Shaviro, 2009). The costs estimated in terms
of information loss in accounting earnings will still apply and, should the IFRS-as-tax-base model be considered, this cost
would need to be evaluated. Further evidence on a broader set of costs and benets would be valuable to inform this
debate.38
2.4. Summary and suggestions for the future
Do differences between book and taxable incomes provide information about current and future earnings? Do rms
appear to manage the tax expense line? What are the costs and benets of having companies report the same income
measure to tax authorities and to shareholders? These are some of the broad questions examined to date in this area of
research.
In sum, the evidence is consistent with book-tax differences containing information about pre-tax earnings quality.
Book-tax differences are by denition differences between book and tax reporting of the same transaction. Thus, if
information is available elsewhere about the transaction, an open question is whether investors get the information from
the book-tax differences or from some other source (this is a question with accounting earnings more generally). This issue
is often recognized by the authors of these studies and examining the information in the book-tax differences is still
important if the accounts could provide investors with the information. Future research will likely benet from
investigating the conditions under which there is more or less information and perhaps from sorting out the sources of the
information, if possible. Researchers should be careful in choosing the measure of book-tax differences that are appropriate
for the research question. For example, if one is interested in examining pre-tax accrual quality than a measure of
temporary book-tax differences is appropriate.
34
See Atwood et al. (2010), Burgstahler et al. (2006), and Goncharov and Werner (2009) for a discussion of these issues. Note that Goncharov and
Werner (2009) attempt to reconcile results in the extant literature with the use of single and group accounts. The authors argue that the nancial
reporting incentives play an economically signicant role in the preparation of the single accounts. Further, see the nancial accounting literature on
comparison of accounting across countries more generally (e.g., Ball et al., 2000) and a discussion of the international studies on book-tax conformity
(Hanlon et al., 2009).
35
See Guenther et al. (1997), Ali and Hwang (2000), Ball et al. (2000), Ball et al. (2003), Guenther and Young (2000), Hanlon et al. (2005), Hanlon and
Shevlin (2005), McClelland and Mills (2007), Hanlon et al. (2009), Ayers et al. (2009), Atwood et al. (2010), and Hanlon and Maydew (2009).
36
See Hanlon and Shevlin (2005) and also a recent article by Kessler (2008) on the inuence of the EU on the IASB. In all likelihood, government
inclination to interfere with accounting standard setting would increase should those standard setters be determining taxable income as well.
37
See Kothari et al. (2010) for a discussion of the uses of accounting information and the use of that information by multiple stakeholders/
constituents.
38
One alternative is partial conformity, taking into account the potential harm to accounting earnings with each item considered (i.e., working from
the current system to a more conformed system instead of conforming and starting over creating differences). Clearly, the rst items to change are the
special provisions in the tax code that provide tax breaks or punishments to particular taxpayers or for certain activities. Removing these would make the
system simpler, increase conformity, and retain the information in nancial accounting earnings. But like all of the tax code, these provisions are very
political in nature and their removal would be a shift in the tax law setting process currently in the U.S.
The evidence on whether rms manage the tax expense line item, for the most part, suggests the answer is yes.
However, an interesting take-away from this literature is that given all the discretion and judgment necessary to establish
the tax accrual accounts, there is not the level of earnings management evidence that many likely suspected. Future
contributions in this area seem limited; however, regime shifts, rule changes, and changes in cross-sectional determinants
may open avenues for future research.
Book-tax conformity is an important issue to both nancial accountants and tax researchers. This policy alternative is
seen by some as a non-starter but at times gains support in Washington and in other governments around the world.
The most important issue in this area is to consider all the costs, including those generated by any reporting responses on
the part of the companies and those from political posturing on the part of the government.
3. Tax avoidance
In this section, we shift our attention to the corporate reporting of taxable income to tax authorities. There is
widespread interest and concern over the magnitude, determinants and consequences of corporate tax avoidance and
aggressiveness. Shackelford and Shevlin (2001) call for research on the determinants of tax aggressiveness and Weisbach
(2002) asks why there is not more tax sheltering (later coined the undersheltering puzzle). These are certainly important
topics of research. The challenge for the area is that there are no universally accepted denitions of, or constructs for, tax
avoidance or tax aggressiveness; the terms mean different things to different people. Similar to nancial accounting
research on earnings quality, the lack of an accepted denition should not stop research on the topic. Quite the contrary;
the more good research on the topic, the more likely an accepted denition will take shape.
Let us rst discuss our conceptual denition of tax avoidance in an attempt to minimize confusion over semantics for
the rest of our review (and perhaps in future research). We dene tax avoidance broadly as the reduction of explicit taxes.
This denition conceptually follows that in Dyreng et al. (2008) and reects all transactions that have any effect on the
rms explicit tax liability. This denition does not distinguish between real activities that are tax-favored, avoidance
activities specically undertaken to reduce taxes, and targeted tax benets from lobbying activities.39 However, we do not
limit our conceptual term tax avoidance to the Dyreng et al. (2008) empirical measure; we dene tax avoidance very
broadly. If tax avoidance represents a continuum of tax planning strategies where something like municipal bond
investments are at one end (lower explicit tax, perfectly legal), then terms such as noncompliance, evasion,
aggressiveness, and sheltering would be closer to the other end of the continuum. A tax planning activity or a tax
strategy could be anywhere along the continuum depending upon how aggressive the activity is in reducing taxes.40
However, much like art, the degree of aggressiveness (beauty) is in the eye of the beholder; different people will often have
different opinions about the aggressiveness of a transaction. The individual studies we discuss often use different terms to
describe the tax reporting behavior (aggressiveness, sheltering, evasion, noncompliance, etc.). Clearly, most
interest, both for research and for tax policy, is in intentional actions at the aggressive end of the continuum (e.g., evasion).
However, to minimize the focus on semantics, we will, for the most part, discuss the literature using the generic term tax
avoidance.
We rst discuss the theory of corporate tax avoidance. This includes the relationships between the shareholders,
management, and the government and the potential agency issues that can arise in corporate tax avoidance. We also
discuss the recent re-incorporation of corporate governance into the tax avoidance literature. This involves consideration
of rm-level governance structures and the effect on responses to taxation and also a consideration of tax authorities
acting as additional governance mechanisms for the rm.
We then focus our discussion on the measurement of tax avoidance. Because a wide range of proxies are currently used
in the literature and because of the signicant policy implications of research on tax avoidance, researchers must be careful
to consider whether the measure they choose is appropriate for their particular research question. In our review, we
attempt to clarify the information that can be inferred from the different measures and clearly state what information
cannot. Most importantly, we point out that most measures in the literature capture only non-conforming tax avoidance;
that is, tax avoidance transactions accounted for differently for book and tax purposes. Conforming tax avoidance, in which
nancial accounting income is reduced when the tax strategy is employed, is not captured by most measures. We discuss
the empirical measures before discussing the literature in order to review what inferences can and cannot be made from
studies given the limitations of the empirical measures. Finally, we discuss research on the determinants and
consequences of tax avoidance.
39
We do not distinguish between technically legal avoidance and illegal evasion for two reasons. First, most of the behavior in question surrounds
transactions that are often technically legal. Second, the legality of a tax avoidance transaction is often determined after the fact. Thus, avoidance captures
both certain tax positions (e.g., municipal bond investments) as well as uncertain tax positions that may or may not be challenged and determined illegal.
Weisbach (2003) discusses a similar problem with denitions. He points out that lawyers and economists are quick to classify avoidance as legal tax
planning and evasion as illegal tax planning as if one can determine the legality of a tax structure easily. A problem with tax shelters is that it is almost
always ambiguous whether the transaction is permissible or not.
40
One denition of tax aggressiveness is found in Slemrod (2004) and Slemrod and Yitzhaki (2002) in which aggressive tax reporting encompasses a
wide range of transactions whose primary intent is to lower the tax liability without involving a real response by the rm and is a subset of tax avoidance
activities more generally. Intuitively, aggressiveness is thought of as pushing the envelope of tax law.
3.1. Theory of corporate tax avoidance
The factors affecting individual tax avoidance and compliance are well studied in public economics (Slemrod and
Yitzhaki, 2002). In theory, individual tax compliance is determined by tax rates, the probability of detection and
punishment, penalties, and risk-aversion (Allingham and Sandmo, 1972) as well as intrinsic motivations such as civic duty.
Many of these factors apply to the corporate taxpayer as well. However, as Slemrod (2004) points out, additional
issues arise in widely held corporations because of the separation between ownership and control. Risk-neutral
shareholders expect managers acting on their behalf to focus on prot maximization, which includes going after
opportunities to reduce tax liabilities as long as the expected incremental benet exceeds the incremental cost. Thus, tax
avoidance is not, in and of itself, a reection of agency problems. However, separation of ownership and control can lead to
corporate tax decisions that reect the private interests of the manager. Thus, the challenge for shareholders and boards of
directors is to nd the combination of control mechanisms and incentives that minimize these agency costs (Jensen and
Meckling, 1976).
Slemrod (2004), Chen and Chu (2005), and Crocker and Slemrod (2005) lay the theoretical foundation for understanding
corporate tax avoidance within an agency framework. Chen and Chu (2005) examine corporate tax avoidance under a
standard principalagent model and focus on the efciency loss due to the separation of management and control. Crocker
and Slemrod (2005) examine the compensation contract of the executive (e.g., CFO or tax director) who determines the
rms deductions from taxable corporate income and focus specically on the relative efcacy of tax compliance penalties
levied on the principal versus the agent. Most of the literature prior to these studies assumes the rm makes the tax
reporting decision with no agency considerations.
The separation of ownership and control implies that if tax avoidance is a worthwhile activity, then the owners ought to
structure appropriate incentives to ensure that managers make tax-efcient decisions. By tax-efcient, we mean corporate
tax decisions that increase the after-tax wealth of the rms owners, that is, where the marginal benets of the transaction
exceed the marginal costs. This can be done explicitly based on tax outcomes or implicitly via contracts linking pay to
after-tax returns or stock price. This approach can be extended to generate a set of predictions about the relation between
tax reporting, incentive structures, and rm value.
Another perspective has recently been introduced into the literature by Desai et al. (2007b) who propose a situation in
which self-interested managers structure the rm in a complex manner in order to facilitate transactions that reduce
corporate taxes and divert corporate resources for private use (which presumably includes manipulating after-tax earnings
for private gain). Desai et al. (2007b) posit that in such a scenario, a strong tax authority can provide additional monitoring
of managers, and thus, the incentives of the outside shareholders are aligned with the tax authority to reduce diversion of
resources by insiders.41 The authors extend their theory to several aspects of the tax system. First, they predict that a
system of high tax rates but weak enforcement may increase managerial diversion from the tax authority as well as from
outside shareholders. Thus, the authors argue that outside shareholders benet when tax enforcement increases because it
increases the probability that diversion will be detected. Based on a sample of Russian rms undergoing an increase in tax
enforcement following the election of Vladimir Putin in 2000, tax payments increased, related party trades were curtailed,
and tax haven entities were abandoned.42 Important to the interpretation that tax authorities reduce agency problems,
shareholders of rms in the oil industry appeared to approve of the increased enforcement, as stock prices for tax avoiding
rms rose signicantly around government enforcement actions.43
The second point Desai et al. (2007b) make is that corporate governance affects the response of rms to changes in
corporate tax rates. When governance is weak, an increase in the tax rate results in more diversion lowering corporate tax
revenues. When corporate governance is strong, an increase in the corporate tax rates will yield higher corporate tax
revenues. The authors nd evidence consistent with their conjecture through tests across countries.44
41
A third possible scenario is where the inside managers collude with the tax authority such that the tax authority looks away from divertive
activities in exchange for extra payments.
42
It is important to note that tax authorities are likely not interested in resolving agency problems per se. The effect on shareholders is a byproduct of
the tax authorities interest in controlling managerial diversion, which hurts both tax authorities and shareholders. In addition, the authors follow one
particular company and show that reported company income soared, and for the rst time the company issued a dividend to shareholders. The authors
note that the Financial Times reported that to comply with the new laws the companies must show the true extent of their nancial operations to outside
shareholders, who are just as keen to have a share of the proceeds as the tax inspector (A. Jack, Financial Times, September 17, 2001).
43
While the Desai et al. (2007b) theory about tax authority monitoring reducing managerial diversion is certainly interesting and the examples and
empirical results in the paper are compelling, the generalizability of data from one industry in Russia is somewhat difcult. A separate test of the theory is
in Guedhami and Pittman (2008) who examine the cost of debt for private rms 144A bond issues in the U.S. using the probability of a face-to-face IRS
audit from TRAC (Transactional Records Access Clearinghouse) as the measure of IRS monitoring. They nd evidence consistent with the Desai
et al. predictionthe higher the IRS audit rates, the lower the yield spread (the lower the cost of debt). The authors conclude that the IRS is an active
third-party monitor and alleviates the information asymmetry between controlling shareholders and outside investors resulting in lower interest rates
for the rms. This line of research is intriguing. In public companies, most would agree a stronger nancial reporting or securities law enforcement agency
would be preferable to a strong IRS. It is somewhat hard to believe that nancial markets perceive the IRS as competent enough to monitor given the
plethora of stories about how the agency has few resources and little talent to match that of corporate tax departments. Further research is necessary to
make broader policy implications about resource allocation or the impact on tax revenues. There are limitations to the TRAC data to be sure (e.g., it is
highly determined by size and year); however, the data may prove a valuable proxy for IRS monitoring in future research.
44
The control premium varies across country and is from negotiated control block sales as computed by Dyck and Zingales (2004).
These papers provide theory and predictions about the relationships and incentives of the various parties insiders,
outside shareholders, and the state in corporate tax avoidance settings. In addition, the studies make predictions about
cross-sectional variation in tax avoidance activities (including determinants) and the market consequences of those
activities. These studies help advance our understanding of the relation between corporate governance and tax avoidance.
We include in our review a number of papers that integrate the implications of these theories. This literature is relatively
young and as a result, most studies have examined rm-level determinants without much examination of executives and
their incentives and the position of the state as a minority shareholder.45
3.2. Measuring tax avoidance
In this section, we discuss the various measures of tax avoidance in the literature. Our main point is that not all
measures are appropriate for all research questions. We rst discuss where one can obtain information on rm-level taxes
and then discuss the most common tax avoidance measures in detail.
Corporations report taxable income on their tax returns and also report income tax expense and income tax assets and
liabilities on their GAAP nancial statements. Thus, estimates of taxable income and tax payments, important factors in
measuring tax avoidance, could be obtained from either source. Most tax avoidance measures are obtained from nancial
statement data because tax returns are not publicly available and access is granted to only a few (we discuss potential problems
with tax return information below). Yet, it is well known that there are many problems with computing estimates of taxable
income from nancial statements. Hanlon (2003) and McGill and Outslay (2004) and others discuss and show the potential
issues in conducting such an exercise. In essence, there is a lack of disclosure in nancial statements about taxable income and/
or the actual cash taxes paid or to be paid on the current years earnings. FASBs task in designing nancial accounting standards
is in establishing rules under which companies represent their economic performance for external stakeholders to evaluate that
performance and estimate future performance. FASB is not inherently interested in disclosure of tax data, except to the extent
those items affect GAAP earnings and the GAAP balance sheet. There are potential reasons why outsiders want to know taxable
income, for example, to use as a benchmark for accounting earnings and to evaluate corporate tax avoidance and/or corporate
tax citizenship (see Hanlon, 2003; Lenter et al., 2003).
Knowing that nancial statement data leads to errors in estimating taxable income, it is important to consider what the
alternative source access to tax returns would and would not solve. Because of the different consolidation rules between
book and tax (see Appendix A and Hanlon, 2003), it is nearly impossible to match any one tax return to any one set of nancial
statements (see Mills and Plesko, 2003). Thus, even with tax return data it is very difcult (if not, impossible) to ascertain how
much tax is being paid on the reported accounting earnings or cash ow reported in a led 10K. Looking at the tax return alone,
one could compare to the accounting earnings for the entities included in the tax return (via the schedule M-1 or M-3), but
again if one is interested in benchmarking to accounting earnings for some certain group of entities (the consolidated set for
GAAP), then this approach may not provide the solution. In addition, tax regulations and enforcement are conducted at a
national level, so using U.S. tax returns only provides information about the U.S. portion of the activity of a U.S. multinational
corporation. Further, if one wants the markets interpretation of the information in a rms taxable income, then research must
use data available to the market (i.e., something other than tax return data). Finally, research conducted using tax return data is
not replicable. There are settings where tax return data are valuable, to be sure, but many times researchers say ideally we
would use tax return data or tax return data provides the truth when that is not necessarily the case.
The measures of tax avoidance that we review are listed in Table 1. We list 12 measures of tax avoidance commonly
used in the literature.
3.2.1. Effective tax rate measures

These are computed by dividing some estimate of tax liability by a measure of before-tax prots or cash ow. These
measures capture the average rate of tax per dollar of income or cash ow. Understanding what the numerator captures is
essential. In Table 1 we describe the inferences that are possible across the various numerators. Most importantly, the
numerator in the GAAP ETR (dened as total income tax expense divided by pre-tax accounting income) is total income tax
expense. A tax strategy that defers taxes (e.g., more accelerated depreciation for tax purposes) will not alter the GAAP ETR.
In addition, several items that are not tax planning strategies, such as changes in the valuation allowance or changes in the
tax contingency reserve could affect the GAAP ETR. The GAAP ETR is the rate that affects accounting earnings. The Cash
ETR, on the other hand, is computed using cash taxes paid in the numerator, and is affected by tax deferral strategies but is
not affected by changes in the tax accounting accruals. The annual Cash ETR could mismatch the numerator and
denominator if the cash taxes paid includes taxes paid on earnings in a different period (e.g., from an IRS audit completed
in the current year) while the denominator includes only current period earnings.
Depending on the research question, a more important consideration may be the denominator of the ETR. Most effective
tax rates use pre-tax GAAP earnings as the denominator and thus can only capture non-conforming tax avoidance
45
Note that the idea that the taxing government is a minority shareholder is an old one included in the original Scholes-Wolfson text and earlier
work (see Desai et al. (2007b) for a review). However, the attention given to other stakeholders beyond the rm itself has been limited in the extant
research until recently.
140
Table 1
Measures of tax avoidance.
Measure Computation Description Impact Reect deferral Reect non- Reect Computable
accounting strategies? conforming conforming by
earnings? avoidance? avoidance? jurisdiction?
GAAP ETR Worldwide total income tax expense Total tax expense per dollar of pre-tax Yes No Yes No Yes
Worldwide total pretax accounting income
book income
a
Current ETR Worldwide current income tax expense Current tax expense per dollar of pre-tax Maybe Yes Yes No Yes
book income
Cash ETRb Worldwide cash taxes paid Cash taxes paid per dollar of pre-tax book No Yes Yes No No
M. Hanlon, S. Heitzman / Journal of Accounting and Economics 50 (2010) 127178

income
P
Long-run cash ETRc P Worldwide cash taxes paid Sum of cash taxes paid over n years divided No Yes Yes No No
Worldwide total pretax accounting income by the sum of pre-tax earnings over n years
ETR Differentiald Statutory ETRGAAP ETR The difference of between the statutory ETR Yes No Yes No No
and the rms GAAP ETR
e
DTAX Error term from the following regression: The unexplained portion of the ETR Yes No Yes No No
ETR differential Pre-tax book differential
income =a +b Controls +e
Total BTDf Pre-tax book income ((U.S. CTE +Fgn CTE)/ The total differences between book and Yes for a portion, Yes Yes No Yes (U.S.)
U.S. STR) (NOLt NOLt-1)) taxable incomes no for a portion
Temporary BTD Deferred tax expense/U.S. STR No Yes Yes No Yes (U.S.)
g
Abnormal total BTD Residual from BTD/TAit = bTAit + bmi +eit A measure of unexplained total book-tax Yes for a portion, Yes Yes No No
differences no for a portion
Unrecognized tax Disclosed amount post-FIN48 Tax liability accrued for taxes not yet Yes If uncertain Yes, some Yes, some No
benetsh paid on uncertain positions
i
Tax shelter activity Indicator variable for rms accused of Firms identied via rm disclosures, Depends on the Shelter may be a Not overall- Not overall- Unlikely
engaging in a tax shelter the press, or IRS condential data type of shelter deferral strategy; measure is measure is
but not an transaction transaction
overall measure based based
Marginal tax ratej Simulated marginal tax rate Present value of taxes on an additional No Yes Yes Yes Not with
dollar of income existing data
a
May impact accounting earnings if the item that changes the Current ETR is not a temporary difference.
b
A more direct measure of taxes actually paid but numerator and denominator may be unaligned. The measure is more volatile year-to-year than the two measures above. Can also deate by pre-tax
income adjusted for special items.
c
See Dyreng et al. (2008). Measured generally over 310 years. Longer may be better but fewer available rms. Eliminates some of the volatility in Cash ETR. Can also deate by pre-tax income adjusted for
special items.
d
If using the same U.S. statutory tax rate for all rms, comparing GAAP ETRs yields similar inferences.
e
The terms on the right-hand side can vary depending on the research question. Model is only as good as the variables included as determinants. What variables to include depends on how one interprets
the actions of the manager with regard to that constructaction taken to reduce taxes or the reduction of tax is a byproduct.
f
Grossing up current tax expense by the statutory tax rate to estimate taxable income is subject to well-known measurement error (Hanlon, 2003). Subtracting the change in the NOL is intended to capture
changes in taxable income that are not captured by the current tax expense because the rm is a tax-loss rm and current tax expense is thus reported as zero (or a negative if they have NOL carrybacks).
Researcher should conduct sensitivity tests for the cases where measurement error is likely the highest as in Hanlon et al. (2005).
g
See Desai and Dharmapala (2006). A variety of other right-hand side variables could be included depending on what the research question calls for in terms of controls. TA (Total Accruals) intended to
control for earnings management.
h
The measure is a nancial accounting accrual subject to the conservative or aggressive nature of the rm for nancial accounting purposes.
i
The measure will not include rms that are not caught nor will it include rms that can otherwise avoid tax successfully and do not engage in shelters.
j
See Shevlin (1990), Graham, (1996a, b), Blouin et al. (2010), Graham and Kim (2009). This measure is not really a measure of avoidance but may provide information when comparing rms with varying
importance for nancial accounting earnings (e.g., private companies versus public companies).
(e.g., ETRs will not reect the tax benets of interest deductibility). Thus, if a rm that does not face a strong nancial
accounting constraint (e.g., a private rm) avoids most of its explicit taxes by reporting lower accounting earnings as well as
lower taxable income (i.e., conforming tax avoidance), this type of tax avoidance would not be captured by the typical effective
tax rate measures. As a result, the researcher must be very careful when making inferences about overall tax avoidance if the
sample under consideration contains rms with differing levels of importance placed on nancial accounting earnings.
3.2.2. Long-run effective tax rates

Dyreng et al. (2008) develop a long-run cash ETR measure estimated as the sum of cash paid for income taxes over ten years
scaled by the sum of pre-tax income (net of special items) over the same period.46 The main benet of the measure is the long-
run nature of the computation which avoids year-to-year volatility in annual effective tax rates. Using the long time periods
allows the use of cash taxes paid in the numerator because the long-run measure avoids much of the mismatch of cash taxes
and earnings.47 The use of cash taxes paid is benecial because it avoids tax accrual effects present in the current tax expense.
However, note that all ETR variants, including the long-run measure, (1) reect all transactions that have any effect on
the rms explicit tax liability, (2) do not distinguish between real activities that are tax-favored, avoidance activities
specically undertaken to reduce taxes, and targeted tax benets from lobbying activities, (3) do not directly capture
implicit taxes (although if the denominator is lower for a given amount of tax, the ratio will be higher),48 and (4) do not
capture conforming tax avoidance because the measure uses book income as the denominator. Further, some may
interpret the low ETRs (including the long-run measure) as the result of upward earnings management (if the taxes remain
constant) and not due to any specic tax savings behavior. But if taxes remain constant as the rm manages earnings
upward over the long run, then the rm is still avoiding taxes on the overstated accounting earnings. In addition, earnings
management through accruals should reverse over the long run, to some extent mitigating the concerns about the effect of
earnings management in the long-run measure.49
3.2.3. Book-tax differences

It seems intuitive that book-tax differences could provide information about tax avoidance behavior given our previous
discussion of the sources of book-tax differences.50 However, compared to the earnings quality studies in which
researchers can correlate book-tax differences with outcomes such as future earnings patterns, the information in book-tax
differences about tax avoidance is harder to document because valid tax outcomes are difcult to obtain. Mills (1998)
documents that rms with large book-tax differences (measured on the tax return and using deferred tax expense from the
nancial statements) are more likely to be audited by the IRS and have larger proposed audit adjustments. Wilson (2009)
nds that book-tax differences are larger for rms accused of engaging in tax shelters than for a matched sample of non-
accused rms. The evidence from these studies suggests that book-tax differences capture some element of tax avoidance.
Of course, book-tax differences by denition only capture non-conforming tax avoidance. Thus, this measure cannot be
used to compare tax avoidance activities across rms with varying levels of importance on nancial accounting earnings.
Further, as both authors recognize, it is important to keep in mind that in both studies the rms in the sample are rms
that were caught. If the tax authority uses large book-tax differences to identify tax avoidance, then what the studies
capture is the tax authoritys model of tax avoidance.51
3.2.4. Discretionary or abnormal measures of tax avoidance

Desai and Dharmapala (2006, 2009) compute a measure of abnormal book-tax differences by regressing total book-tax
differences on total accruals, where total accruals is intended to control for accounting earnings management. The residual
46
Dyreng et al. adjust net income for special items to retain more rms in the sample (reduce the number of loss rms). The authors estimate the
tests without making the adjustment and retain fewer rms, but the basic inference in terms of the proportion of rms that sustain low long-run rates is
the same (and in sensitivity they estimate the results using cash ow as the denominator, avoiding the issue altogether). However, this same adjustment
will not be necessary or even right for all research questions.
47
The mismatch of cash taxes and earnings results from (1) the timing of cash tax payment through estimates and with the tax returns, (2)
settlements with tax authorities (earnings recorded in prior periods but tax paid to tax authorities later in time), (3) any potential U.S. residual taxes on
foreign earnings which are not paid until the earnings are repatriated in cash (whereas the earnings are included in income when earned).
48
Implicit taxes is a term popularized by Scholes and Wolfson (1992) to describe the effect of tax preferences on asset prices and expected pre-tax
returns. An investment in a tax-preferred asset is said to bear implicit taxes when the pre-tax returns are lower than the returns on a fully taxed asset of
identical risk (e.g., municipal bond investments).
49
Dyreng et al. (2008) suggest an alternative specication using cash ows from operations as a scalar in order to partially address the earnings
management issue and as a way to eliminate the use of book earnings as a benchmark and capture some types of conforming tax avoidance behavior.
Note that scaling by cash ows only allows the capture of conforming tax avoidance if the tax avoidance strategy is accruals based (e.g., the expense is
accrued for both book and tax purposes). However, if the tax avoidance does not involve accruals but instead reduces both cash ows and taxable income
(taxes paid) then again both the numerator and denominator will be reduced and the measure will not capture this type of tax avoidance.
50
Recognize that book-tax difference measures are closely related to effective tax rate measures; in general, the book-tax difference measures
subtract one measure of income from the other and the effective tax rate measures are a ratio of some measure of tax (expense or paid) to a measure of
income.
51
Desai (2003) and others point to the increasing difference between book and tax income over the 1990s as evidence of aggressive tax reporting.
(Desai in later papers attempts to control out other causes of book-tax differences.) However, Shevlin (2002) argues that one must be cautious when
drawing inferences about the levels and trends in tax avoidance based on book-tax differences. See also Treasury (1999), Manzon and Plesko (2002),
Yin (2003), Hanlon and Shevlin (2005), and Mills et al. (2002).
is used to proxy for the construct of tax avoidance. Separately, Frank et al. (2009) estimate the discretionary portion (DTAX)
of their PERMDIFF measure (PERMDIFF is essentially the difference between the effective and statutory tax rates multiplied
by pre-tax accounting income). These discretionary measures are similar in notion to the Jones (1991) model of
discretionary accruals and could be useful conceptually in the sense that the discretionary portion attempts to remove
underlying determinants that are not driven by intentional tax avoidance leaving the portion driven by intentional tax
avoidance (some notion of intentional actions to avoid taxes) in the residual.
Like the Jones model, regression-based partitions are only as good as the model used and the validity of the proxies
employed for the known determinants. Because we lack good structural models of book-tax differences and effective tax
rates, the discretionary models could introduce additional error. In addition, because taxes affect many decisions but are
not often rst-order drivers it is difcult to know what variables to include in the known (not tax driven) determinants
and which effects to let fall into the residuals. For example, should the extent of a rms foreign operations or amount of
research and development spending be classied as items to be included in the rst stage regression as known
determinants of GAAP ETR (because the manager that took these actions could have done so for non-tax reason and tax
savings were just a byproduct)? Or, should the effects of these variables be included as part of the overall intentional tax
avoidance measure (the residual) since we know both of these can be intentionally used by the rm (and managers) to
reduce taxes? This is a difcult issue and one for which we do not have an answer. However, researchers should consider
their research question and the implications of the model when structuring the research design for their particular
question and when making inferences from the results.
The measure developed by Frank et al. (2009) has been used in subsequent papers and thus, similar to our earlier
discussion of the long-run cash ETR from Dyreng et al. (2008), we discuss the measure in some detail to be clear on the
types of questions that the measure can and cannot address. The intent of the measure is to capture items that alter, and in
particular reduce, the rms GAAP ETR which then raises bottom-line earnings, all else constant. The authors label of
PERMDIFF for the total amount is somewhat unfortunate, however, because the measure captures much more than
permanent differences.52 This has led to some misapplication of the measure and misinterpretation of results. The
PERMDIFF measure is essentially an ETR differential and can be calculated as the statutory tax rate minus the GAAP ETR.
To get the dollar gure comparable to Frank et al. (2009), multiply the ETR differential by pre-tax book income. Because it
is a function of GAAP ETR, this measure does not capture conforming tax avoidance behavior or tax deferral strategies (and
thus cannot be used to make inferences about overall tax avoidance, especially if the sample being analyzed has both
private and public rmsi.e., rms with varying degrees of importance of accounting earnings). The measure will capture
the effects of tax credits (in fact, it overweights the effects of credits), foreign operations in jurisdictions with different tax
rates (e.g., if earnings are designated as permanently reinvested), and any other item that affects GAAP ETRs. In addition,
the PERMDIFF measure only captures tax avoidance that helps the company boost accounting earnings by reducing the
GAAP ETR. In sum, the effect of these items should be considered when motivating studies using this measure (i.e., are
these the types of transactions the research question is about?).
Beyond the details of the measures described above, an important issue is the elds perception of the meaning of
permanent book-tax differences and temporary book-tax differences as it pertains to tax avoidance.53 With regard to
permanent book-tax differences, some have made the argument that permanent differences likely reect aggressive tax
reporting (meaning at the aggressive end of the tax avoidance continuum). Indeed, there are claims that the ideal tax
shelter is one that generates a permanent difference and claims that permanent differences reect the most egregious
type of tax avoidance (i.e., the most aggressive behavior). It is important to carefully consider whether these statements are
true and why these statements are made.
What evidence do we have? If one thinks that engaging in a tax shelter indicates some notion of tax aggressiveness
(recognizing that rms that do not shelter, or at least do not get caught, may indeed be avoiding more taxes through all
available means) then we can look to the available anecdotal and empirical evidence on tax shelters. Anecdotally, we
observe that transfer pricing disputes have resulted in some of the largest tax settlements with the IRS in history. Transfer
pricing, however, results in no book-tax difference. We observe other transactions such as the Lease-in-Lease-Out (LILO)
and Contested Liability Acceleration Strategy (CLAS) that generate temporary differences (including depreciation
differences). Yet others, such as the Corporate-Owned Life Insurance shelter generate a permanent difference (and part
of the tax benets are interest, which yields no book-tax difference). Thus, some shelters generate a temporary book-tax
difference, some a permanent difference, and some do not generate a difference between book and tax incomes at all.54
The data are limited but can shed some initial light on the issue. Wilson (2009) reports that both temporary and
permanent book-tax differences are signicant in predicting tax shelter involvement. Indeed, when Wilson (2009) employs
a broad control sample only the temporary book-tax differences are signicant. Further evidence is found in Lisowsky et al.
52
Frank et al. (2009) use DTAX as the name of the discretionary portion.
53
As we discussed above, temporary book-tax differences are directly related to pre-tax accounting accruals and thus the conjecture is that they
could provide information about earnings management. Again most of the earnings quality type studies test only for associations with earnings
properties but they do acknowledge in general that book-tax differences, including temporary book-tax differences, can to some extent be driven by tax
avoidance as well.
54
We thank Pete Lisowsky and Ryan Wilson for comments and consultation on this section. See Graham and Tucker (2006), Hanlon and Slemrod
(2009), and the U.S. Treasury (1999) for descriptions of the shelters.
(2010) who examine 101 tax shelters disclosed to the Internal Revenue Service as Reportable Transactions.55 The authors
report that out of the 101 disclosed Reportable Transactions only 48 report any book-tax difference at all on the tax return.
Of those that did, 12 report a related permanent book-tax difference ($1.4 billion total) and 38 report a related temporary
difference ($2.5 billion total; note the same transaction can have part of the effect as a temporary difference and part as a
permanent difference).56 Thus, although more evidence is needed, the premise that a tax position leading to a permanent
difference is more tax aggressive than a tax position that does not is currently unsupported.
So why do people say that ideal shelters are those that generate permanent differences? The most likely explanation
is the nancial accounting implications of such a difference. A tax strategy that generates a permanent book-tax difference
reduces the rms effective tax rate and increases accounting earnings (and can be timed to when such an increase is
desirable). It is ideal precisely because of the nancial accounting benets. Thus, tests that correlate some measure of
permanent differences and measures of nancial accounting aggressiveness may be documenting that rms that care
about nancial accounting are more aggressive for nancial accounting (i.e., it is somewhat circular) rather than the
implied conclusion that rms that are aggressive for tax are aggressive for nancial accounting. We do not intend to
minimize the importance of studying low ETR rms or transactions that lower GAAP ETRs. Our message here is that
researchers should be careful not to conclude the transactions to achieve this effect are the most aggressive for tax
purposes per se (e.g., investments in municipal bonds will lower the GAAP ETR but these investments are not aggressive
for tax purposes).
3.2.5. Unrecognized tax benets (UTB)

This proxy is measured as the levels and or changes in UTB, the accounting reserve for future tax contingencies.57 For
research using UTB as a measure of tax avoidance, it is important to keep in mind that UTBs are potentially driven by two
underlying determinants. The rst is obviously taxes. Higher UTBs represent more uncertainty in the rms tax positions
and thus, are likely indicative of the degree of tax avoidance.58 The second determinant, and the more problematic one for
research design, is nancial reporting incentives. The amount of unrecognized tax benet recorded for nancial accounting
(i.e., the amount of the accounting accrual for potential future tax assessments) is an accounting accrual subject to the
judgment of management. The accrual ultimately affects bottom-line earnings which is precisely why researchers study
the account for indications of earnings management (see Section 2 discussion above). Further, consider a manager that
takes an aggressive tax position to reduce taxable income but also wishes to take an aggressive position for nancial
accounting. The manager is unlikely to record a reserve for the tax associated with the tax position. If one considers the
conjecture that tax avoidance strategies are sometimes adopted with the primary purpose of increasing accounting
earnings, the point is especially salient the reserve will not be recorded (i.e., all benets will be recognized so as to
increase earnings) and the tax avoided will not be captured in the UTB. Thus, the tax contingency or UTB is not a clean
measure of tax avoidance by any stretch. We are not claiming that the correlation of other measures of tax avoidance and
the UTB will not be signicant. However, not all tax avoidance will be included in the UTB (or other measures) and thus it is
hard to interpret what a signicant correlation really means. Studies of UTBs should recognize the dual nature of the
account and discuss the implications for their research question, design, and results.
3.2.6. Tax shelter rms

This classication of tax avoidance is advantageous if a researcher is interested in identifying cases of intentional tax
planning behavior at a transaction level at the aggressive end of the tax avoidance continuum. However, there are several
important tradeoffs. First, there are potential selection biases. Similar to the statements above, the sample of tax shelter
rms identies only rms that were either caught and formally charged or that disclosed the shelter under recent
disclosure rules for certain transactions. Second, the use of a tax shelter is likely endogenous. Firms that can otherwise
avoid taxes may not need to engage in shelters, and rms that for whatever reason cannot otherwise avoid taxes may be
the ones that engage in (resort to) tax shelters. Thus, tax sheltering rms may not be the rms that avoid the most taxes if
one adopts a more general denition of tax avoidance.59 In other words, the shelters are single transactions and may not
capture the rms overall avoidance behavior. As a result, studies that correlate a measure of tax avoidance with tax shelter
use may or may not be establishing support for the validity their tax avoidance measure. Whether tax shelter proxies are
appropriate measures of tax avoidance obviously depends on the research question. However, the authors interpretations
of the evidence must clearly recognize the limitations imposed (as well as the benets garnered) by using a tax shelter
sample.
55
There are two types of reportable transactions, listed and non-listed. See Lisowsky (forthcoming) and Lisowsky et al. (2010) for a description of
listed and reportable transactions.
56
The Lisowsky et al. (2010) paper is an early working paper and using listed transactions has some limitations which the authors recognize.
However, at a rst pass their data reveal some evidence on the extent to which listed transactions involve conforming tax avoidance and the extent to
which listed transactions involve temporary versus permanent differences.
57
Amounts are disclosed by rms under Financial Accounting Standards Board Interpretation No. 48, Accounting for Uncertainty in Income Taxes
(FIN 48) issued in 2006. See discussion in Section 2.
58
Lisowsky (2010) reports an empirical link between tax sheltering and his measure of the pre-FIN 48 tax contingency reserve.
59
We recognize that there is not an accepted denition of tax shelter either. We do not have the space to go into this issue but refer readers to the
papers on tax shelters such as Bankman (1999, 2004), Graham and Tucker (2006), Wilson (2009), and Hanlon and Slemrod (2009).
To summarize, prior and current research employs a variety of metrics to measure corporate tax avoidance. We cannot
overemphasize that not all measures are equally appropriate for every research question. In addition, many proxies in
essence capture the same thing some measure of book-tax differences and researchers should, when employing
multiple measures, consider which measures should yield different results and why (or why not).60 We have attempted to
provide guidance on the appropriateness of certain measures for general types of research questions and samples. In the
following section, we discuss the studies on the determinants and consequences of tax avoidance and we specically
highlight areas where the measure used for tax avoidance limits the interpretation of the results.
3.3. Determinants of tax avoidance
Several studies examine the relation between rm-level characteristics and tax avoidance using a number of the proxies
discussed above (e.g., GAAP ETR, tax shelter use, etc.). For example, Gupta and Newberry (1997) examine a wide range of
determinants of GAAP ETRs. Rego (2003) reports evidence that suggests the scale of international operations leads to more
tax avoidance opportunities resulting in lower GAAP ETRs. Both studies examine only non-conforming tax avoidance.
In addition, rms accused of using tax shelters have larger book-tax differences, more foreign operations, subsidiaries in
tax havens, higher prior-year effective tax rates, greater litigation losses, and less leverage (Wilson, 2009; Lisowsky,
forthcoming).61
A young, but growing empirical literature incorporates agency predictions into an analysis of corporate tax avoidance.
One aspect is that if avoidance activities create value and compensation incentives align the managers interest with
shareholders, then rms that use more after-tax performance-based incentives should engage in more tax avoidance.
Consistent with this notion, Phillips (2003) nds evidence from survey data that compensating business unit managers on
after-tax income leads to lower GAAP ETRs. However, Desai and Dharmapala (2006) extend the theories in Slemrod (2004)
and Desai et al. (2007b) by modeling the effect of incentive compensation and governance structures on tax avoidance at
the rm level and nd a negative association between equity-based compensation and tax avoidance (measured by
abnormal book-tax differences). Employing cross-sectional variation in their tests, the authors report that the negative
association holds only among rms with weaker shareholder rights and lower institutional ownership. Desai and
Dharmapala attribute this difference to the theoretical predictions in Desai et al. (2007b) discussed above. If managers
engage in tax avoidance in order to increase managerial diversion, then increasing equity incentives to further align
managers and shareholders will decrease diversion, which then decreases tax avoidance engaged in to accomplish the
diversion.
Ownership structure may also be important. While the tax policy in place can have important implications for the
development of corporate ownership patterns (Desai et al., 2007b), ownership patterns can have an important effect on tax
avoidance (Desai and Dharmapala, 2008). Firms with concentrated ownership, such as the family rms examined in Chen
et al. (2010), may avoid more taxes because controlling owners benet more from the savings. On the other hand, these
rms may avoid fewer taxes because these long-term concentrated holders have a longer horizon and may be more
sensitive to the total costs of avoidance arising from reputation effects and suspicions of diversion from minority
shareholders. Chen et al. document that family rms avoid fewer taxes than non-family rms, a result that appears
consistent with Desai and Dharmapala (2006). That is, family-owned rms are willing to forgo tax benets to avoid
concerns by minority shareholders of family rent seeking masked by tax avoidance activities. However, the result is also
consistent with a more basic model from the evasion literature, which links aggressive tax reporting to an individuals risk
aversion and intrinsic motivation. On a relative basis, high-ownership family rms are more likely to behave like
individuals.
As mentioned above, an important issue highlighted by the growing research on tax avoidance is how to measure
avoidance and how to interpret the results when rms in the sample place different levels of importance on accounting
earnings. We use the Chen et al. (2010) paper discussed above as an example. Chen et al. employ four measures
of tax avoidance: the GAAP ETR, the Cash ETR, total book-tax differences, and abnormal total book-tax differences. It is
important to note that each of these measures of tax avoidance capture only non-conforming tax avoidance. In other
words, if family-owned rms place less importance on accounting earnings, which is plausible if they are under less capital
market pressure, they may be willing to voluntarily conform book and tax treatment, i.e., take a tax deduction and record
an accounting expense. The measures used in Chen et al. cannot pick up such conforming tax avoidance. The paper
provides a contribution to the literature nonetheless, but interpretation of the results must be done with care. The results
in the paper show that family-owned rms engage in less non-conforming tax avoidancethat is, they pay more in tax per
dollar of book income. What the paper does not reveal is whether the family-owned rms generally engage in more tax
avoidance, including conforming tax avoidance. While the literature at present does not have a good measure of
conforming tax avoidance, alternative measures that would test tax avoidance where tax is not measured relative to
accounting earnings include a ratio of cash taxes paid to cash ows from operations, or possibly the marginal tax rate in
60
This message is analogous to that about the earnings quality literature in Dechow et al. (2010).
61
Zimmerman (1983) examines the effect of size on tax avoidance, as measured by GAAP ETRs (Zimmermans intent was not testing tax avoidance
per se but rather the effect of political costs, proxied for by size, on rm behavior). See also Hanlon et al. (2007) and Mills et al. (1998).
order to obtain a broader perspective on the rms activities (although the use of the marginal tax rate alone would not be
sufcient).
Recent studies examine items beyond rm-level characteristics and ownership. For example, when the tax department
is considered a prot center, GAAP ETRs are lower but Cash ETRs are not (Robinson et al., 2010). Armstrong et al. (2010)
report similar results with respect to the GAAP ETR versus the Cash ETR when testing tax director compensation contracts.
In addition, separate from rm-level characteristics and incentives, the top executives appear to have a signicant effect on
tax avoidance using both GAAP and Cash ETRs (Dyreng et al., 2010).
Overall, the eld cannot explain the variation in tax avoidance very well, although strides have been made in linking
avoidance to rm characteristics, manager effects, ownership, governance, and incentive structures.62 Some plausible
explanations are that the theory on corporate tax avoidance in an agency framework is relatively young and is not well
developed or sufciently incorporated into the empirical literature at this time. In addition, empirical measures of tax
avoidance that rely on nancial statements have known limitations in part because they capture variation in tax avoidance
as well as the choice between conforming and non-conforming tax avoidance.63 In addition, reliable empirical measures of
some of the interesting cross-sectional determinants, such as governance, are difcult to obtain because corporate
governance is endogenous.64 Finally, tax avoidance may be highly idiosyncratic and determined by a number of factors and
interactions, not all of which can be measured.
3.4. The consequences of tax avoidance
Tax avoidance has a number of potential consequences. The consequences may be direct, i.e., taking a deduction for an
otherwise non-deductible expense increases cash ow and investor wealth, or indirect, i.e., the increased deduction lowers
the marginal benet of the interest tax shield and may change the rms capital structure decisions (e.g., Graham and
Tucker, 2006). One potential consequence is that the rm may be identied by tax authorities and forced to pay additional
taxes and perhaps interest and penalties, representing a decrease in cash ow and investor wealth.65 Moreover if the rm
avoids taxes through investing in tax-advantaged assets (see Berger, 1993), implicit taxes could act to mitigate the effect of
tax avoidance on shareholder wealth. For example, while the literature contains conicting evidence on the role of tax
incentives on the price of investment assets (see Goolsbee, 1998a; Hassett and Hubbard, 1998), it is not clear whether
rms that are successful at tax avoidance sacrice pre-tax cash ows or incur other non-tax costs, and if they do, to what
extent.
Tax avoidance also has potential consequences for managers, shareholders, creditors, and the government. Consider the
case of shareholders. If risk-neutral shareholders demand that managers take actions to maximize after-tax cash ows,
then tax avoidance is a natural byproduct of managerial decision making if managers are provided the right incentives.
If managers optimally avoid taxes (on average), and if investors form unbiased beliefs about the extent and payoff from tax
avoidance, then no association should emerge between tax avoidance and rm value or stock returns. However, this
assumes the right incentives are provided, the incentives work perfectly, and managers and shareholders understand all
the risks and rewards of avoiding taxes.
Desai and Dharmapala (2009) nd that abnormal book-tax differences (their proxy for tax avoidance) have no average
association with rm value measured by market-to-book. However, there is cross-sectional variation in the association; it
depends on the level of institutional ownership. Firms with high institutional ownership have a stronger positive
association between book-tax differences and market-to-book, which suggests that the value shareholders place on
corporate tax avoidance depends on their ability to control the manager, consistent with governance differences explaining
cross-sectional variation in the consequences of tax avoidance.
Hanlon and Slemrod (2009) analyze the market reaction to news about a rms involvement in tax shelters. They nd a
negative, albeit relatively small, stock price reaction of 1.04% on the rst press mention of the shelter.66 Firms in more
consumer-oriented elds have a more negative reaction and the reaction also depends on investor perceptions of the rms
avoidance level. Firms that appear less likely to avoid tax have less negative event returns which the authors interpret as
an indication that the market is positively surprised.
Frischmann et al. (2008) examine the market reaction to events surrounding the passage of FIN 48. They argue that if
the market expected tax costs to rise, there would be a negative reaction, and if the market expected disclosures to
improve, there would be a positive reaction. They document very little in terms of reaction to the events surrounding
passage of the rule and a small positive return for rms surrounding the rst disclosures under the rule. An interesting test
62
Gupta and Newberry (1997) explain 3848% of GAAP ETR based on basic economic characteristics of the rm. Rego (2003) reports R2s of 1014% in
similar regressions.
63
One way perhaps to improve our understanding may be to try to explain long-run rates from Dyreng et al. (2008) as these presumably remove
random variability from the measure. Note also proxies based on tax return data or compliance statistics have their own set of limitations.
64
See Armstrong et al. (2010) for a discussion of corporate governance.
65
See Crocker and Slemrod (2005) for research on the effects of increasing penalties.
66
The results of this study must be interpreted in the correct light. The sample rms are those that were publicly accused of using a tax shelter,
which means that these rms were caught. Thus, the reaction may be different for these rms than for rms with less extreme tax avoidance and/or tax
avoidance that is not caught.
would be whether the stock price reaction to FIN 48 depends on the strength of corporate governance (under the theory in
Desai and Dharmapala, 2006). One would expect the response to the news of the rule enactment to be negative (or less
positive) for well-governed rms but positive for poorly governed rms.
3.5. Summary and suggestions for the future
Why do some corporations avoid more tax than others? How do investors, creditors, and consumers perceive corporate
tax avoidance behavior? What non-tax costs are borne by the shareholders of tax avoiding rms? These are interesting
questions worthy of study.
Our overarching concerns are with the divergent proxies for tax avoidance and, more importantly, the strength of the
inferences that can be made given the chosen proxies and the research question at hand. Further, it is difcult to validate
measures of tax avoidance. For example, some have attempted to use the sample of known tax shelter rms to validate
nancial statement-based tax reporting proxies. However, engaging in a tax shelter is likely endogenous. Even the unrecognized
tax benets disclosed under FIN 48 is not the panacea because this account is also affected by nancial reporting incentives.
There are many interesting avenues for future research. For example, careful thought should be given to the implications
of Chen and Chu (2005), Crocker and Slemrod (2005), and Desai and Dharmapalas (2006) principalagenttax authority
models and predictions. The theory from Desai and Dharmapala (2006) that rent extraction and tax avoidance require
complementary technologies is an interesting angle and is underexplored. In addition, we look forward to future research
that focuses on new methodology for measuring tax avoidance and perhaps methods by which to identify conforming tax
avoidance and methods to identify rms that consistently pursue strategies toward the more aggressive end of the
continuum. Another area where work could be done is a more serious examination of the effects of ownership structure. In
addition, it may be interesting to consider the role of bondholders on tax avoidance (and vice versa).67 However, the theories
relating tax avoidance to ownership structure and bondholder-stockholder conicts is less developed.
Moreover (although this statement is likely biased), the eld could push the manager effect concept of Dyreng et al.
(2010) to other areas. For example, what is the role of directors and their connections to other aggressive tax-reporting
rms? An emerging literature in nance and other areas considers the role of social linkages of directors and executives
across rms, which has been used to explain co-movement in investment. A similar test could be done about co-movement
in tax avoidance activities. The following is an illustrative quote:
Hustlers are also striking at the top, going after chief nancial ofcers and board members. When Merrill Lynch was
promoting its now-famous installment sales shelter in the early 1990s, Merrill board member Robert Luciano, who
was then chairman of Schering-Plough, took the idea to executives of AlliedSignal, on whose board he also sat.
Schering used the scheme too. [(Novack and Saunders, December 1998, Forbes)]
Perhaps a more fundamental question is who makes the tax decisions for the rm? What is the role of the general
counsel in tax decisions? How much control do the top executives have over the tax director, and how is their performance
monitored? What are the cross-sectional determinants that guide the balance of power (e.g., decentralization versus
centralized management)? These are interesting issues we hope to see resolved in the coming years.
In addition, it may be interesting to correlate tax avoidance associated with environmental or social responsibility
constructs (or even divorce rates). Another possible area is whether tax avoidance affects mergers and acquisitions.
Specically, how do acquirers handle targets with prior uncertain tax positions that could lead to future tax liabilities?
Do these aggressive positions affect acquisition price, that is, does the acquirer price tax risk? To what extent do rms
purchase tax indemnity insurance to protect against tax risk and how is this insurance priced?68 We look forward to future
studies on these topics and others. An examination of these topics may allow the tax accounting eld to push the
boundaries of the traditional areas of tax research in accounting.
4. Taxes, book-tax tradeoffs, and real corporate decisions
In this section we review the literature on the effect of taxes on various corporate decisions. We focus on a few key areas
in the interest of space. The effect of taxes on decision-making is important for several reasons, one of which is that
governments use tax policy to provide incentives and disincentives for certain actions (e.g., investment) and thus the
degree to which taxes actually impact corporate decisions determines whether such policies are effective. In addition,
determining the effect of non-tax factors, such as nancial accounting costs, and their tradeoff or interaction with taxes
offers insights into why some tax policies may not affect behavior to the degree intended or may, at times, lead to
unintended consequences. We review the literature in three main areas investment, capital structure, and organizational
form by covering research from all disciplines, highlighting the accounting measurement issues, and incorporating an
examination of the importance of nancial accounting and capital market incentives. We briey discuss the literature in
67
Ayers et al. (2010) report that credit rating agencies are more cautious of rms with large book-tax differences and that the result is attributable to
rms with a higher likelihood of earnings management and not to rms that are likely tax avoiders.
68
See Wolfe (2003) and Logue (2005).
several other areas where the literature is less developed. The effect of taxes and the tradeoff or interactive effect of taxes
and nancial accounting implications (or managerial accounting costs or agency costs) on real corporate decisions is an
area where accounting researchers are contributing and should continue to focus.
4.1. Investment
Investment is the fundamental source of rm value and economic growth. Finance textbooks devote considerable
attention to capital budgeting theory and techniques. Corporate taxes can play a role in the managers investment decision
because the amounts, timing, and even uncertainty of tax payments and deductions affect the calculation of a projects net
present value and hence the decision to invest. In addition, tax incentives can potentially interact with nancial reporting
effects in ways that affect investment decisions. We discuss research on capital investment, research and development
spending, and investment location.69
4.1.1. Theory of investment and taxes: a brief background

The fundamental decision rule for investment is simple: invest as long as the marginal benets exceed the marginal
costs. The typical approach to studying the link between investment and taxes is based on the neoclassical theory of
investment. Early applications of this theory are based on a derivation of the user cost of capital in which the cost of
investment is a function of the required returns to debt and equity and an adjustment for corporate taxes (Hall and
Jorgenson, 1967). In this model, corporate taxes on prots increase the cost of investment, while allowances for
depreciation and investment tax credits reduce it.
More recent attempts to understand neoclassical theory of investment are based on applications of q theory (Tobin,
1969). In accounting, q theory underlies recent research on nancial reporting and investment decisions (e.g., Biddle and
Hilary, 2006; McNichols and Steubben, 2008; Biddle et al., 2009; Bushman et al., 2008). The q theory basically holds that a
rm will invest as long as the value of the marginal investment to shareholders exceeds its cost, or alternatively, when the
ratio of value to cost of the marginal investment marginal q exceeds 1. Investment theory is an important line of
research in nance and economics and is summarized in Hassett and Hubbard (2002) and Hassett and Newmark (2008).70
The q theory approach to understanding investment can be extended to include taxes and is often used as the framework
for linking investment to tax incentives (Hall and Jorgenson, 1967; Summers, 1981).71
4.1.2. Taxes and investment: some evidence

Traditionally, economists have investigated the effect of corporate taxes on investment using aggregate data, drawing
inferences from time-series changes in tax rates or tax regimes.72 Using these methods, the literature has had little
success documenting a link between tax incentives and investment. Hines (1998) states, The apparent inability of tax
incentives to stimulate aggregate investment spending is one of the major puzzles in the empirical investment literature.
Hassett and Hubbard (2002, p. 1316) conclude that, the tendency for a number of aggregate variables to move together
over the business cycle makes it difcult to isolate effects of individual fundamentals on investment using time-series
data. The endemic problems plaguing aggregate time-series analysis are rst, a change in tax rates is likely endogenous,
meaning in response to some macroeconomic factors that could also affect investment, and second, it is difcult to control
for the effect of contemporaneous non-tax shocks on investment.73
The lack of an observed correlation between aggregate investment and taxes over time implies that more powerful
methods are needed to indentify tax effects. As a result, the economics literature has incorporated more cross-sectional
analyses that exploit variation in tax incentives across rms, asset types, and locations to identify tax effects. While it is
somewhat difcult to draw macroeconomic inferences about aggregate investment from these micro-level studies, recent
69
One could consider merger and acquisition (M&A) activity as well, which we discuss in a separate section below.
70
Measurement of q varies in the literature. As noted by Desai and Goolsbee (2004) the q in corporate nance is usually estimated by scaling the
market value of assets by the book value of assets, while in public nance, q is estimated by scaling the market value of assets to the rm by an estimate
of their replacement cost. See Erickson and Whited (2000), Hayashi (1982), Hassett and Hubbard (2002), Summers (1981), Fazzari et al. (1988), Kaplan
and Zingales (2000), Hennessy and Whited (2007), Desai and Goolsbee (2004), and Cummins et al. (2006) for further discussions of the q literature.
71
Such models can also be extended to incorporate investor-level taxes, but require making assumptions about the price-setting investor(s) and the
role of dividend taxes and dividend taxation on investment incentives. We ignore investor tax effects on investment here, but return to the link between
investor taxes and nancial asset prices in a later section.
72
As a clarifying point, it is important to consider the tax rate used to investigate whether and to what extent investment varies with taxes. To study
the question, economists generally use the marginal effective tax rate, which is based on tax codes and measures the wedge between the pre- and post-
tax return on the marginal investment project after taking into account the return on the investment, ination and the amount and timing of depreciation
deductions and investment tax credit. This is very different from either the effective tax rate calculated in accounting or the marginal tax rate used in
corporate nance. The effective tax rate (or GAAP ETR) is the ratio of the tax expense to pre-tax earnings reported in the rms accounting statements. The
marginal tax rate (MTR) is the rm-specic present value of tax on an additional dollar of income. The computation of the MTR takes into account the
carryover provisions of net operating losses (NOLs) and is discussed more fully below. See Edgerton (2009b) for a good discussion of issues with the METR
and a recent attempt to adjust for effects of tax-loss carrybacks and carryforwards in his analysis of investment responses to tax incentives.
73
See Hassett and Hubbard (2002), Desai and Goolsbee (2004), and Hassett and Newmark (2008) for reviews of the literature on taxes and
investment.
empirical studies appear to have reached a consensus that the elasticity of investment with respect to the tax-adjusted
user cost of capital ranges between 0.25 and 1 (Hassett and Hubbard, 2002; Hassett and Newmark, 2008).74
The evidence on the response of investment to taxes using cross-sectional analysis can be illustrated in the context of
bonus depreciation provisions that were introduced to induce investment following September 11, 2001. The provisions
provided for an immediate deduction of up to half of the cost of certain assets put in place during a specied time period.
Although these provisions seem large, little evidence exists that the bonus depreciation provisions of the post-9/11 era
increased aggregate investment (Desai and Goolsbee, 2004). However, as House and Shapiro (2008) note, bonus
depreciation affects the after-tax cost of investments differently. The authors show using asset-level data that investment
in assets for which the bonus incentives would provide the greatest benet qualifying assets with the longest lives when
the acceleration of the deductions matters most did increase in response to the incentives. This result is both a timing
effect (investment was made earlier than it otherwise would have been) and a substitution effect (investment in less
tax-favored assets likely declined).75 Whether there is a permanent aggregate effect on investment is not answered by
their study, however, they do document that investment decisions are sensitive to tax policy, a result that is difcult
to show using aggregate investment data.76 Additional evidence that exploits other cross-sectional variation, such as
rm-level effects across countries or surrounding tax reforms is summarized in Hassett and Hubbard (2002) and Hassett
and Newmark (2008).
Evidence on the link between investment and taxation can be interpreted within the framework of Slemrod (1992b)
who summarizes rm responses to taxation into three categories: (1) the timing of economic transactions, which are likely
the most responsive to tax incentives, (2) nancial repackaging and accounting alterations, and (3) real decisions of rms
and individuals. The evidence at the micro-level suggests that there was an increase in investment in the targeted type of
assets. However, this increase could have come at the cost of lower investment in other types of assets, lower investment
in the targeted assets in later periods, and could partially reect mere changes in the accounting classication rather than
real differences in purchases.
Although few studies in accounting test the effect of taxes on investment, and certainly not aggregate real investment,
accounting research can contribute to this area. Empirical research in economics and nance increasingly relies on
cross-sectional studies using nancial statement data (see, e.g., Barnett and Sakellaris, 1998; Bond and Cummins, 2000;
Erickson and Whited, 2000; Desai and Goolsbee, 2004; Cummins et al., 2006). Measurement issues are important,
and institutional knowledge of the nancial accounting behind estimates of investment, measurement of q, and nancial
constraints becomes crucial.77 In addition, any investment decision will affect pre-tax accounting earnings through
deprecation or expensing. If nancial reporting effects are important, the effect on accounting earnings likely provides
incentives or disincentives for investment that could have implications for interpretation of the results. In addition,
different types of policy choices may or may not affect accounting earnings, perhaps through the income tax expense
(e.g., the investment tax credit or accelerated/bonus depreciation). Edgerton (2009d) compares the effectiveness of tax
policies that do and do not affect accounting prots and nds that policies that do not affect accounting prots
(e.g., accelerated depreciation) are less effective in stimulating investment than those that increase accounting prots
(e.g., investment tax credits).
4.1.3. Investment in intangibles

Investment in intangible assets, e.g., through research and development programs, is both signicant in magnitude
and directly affected by tax policy and targeted incentives. Hassett and Newmark (2008) suggest that inquiry into
the drivers of investment in intangible assets is ripe for exploration. A line of research in accounting beginning at least
with Berger (1993) examines the effectiveness of tax incentives for R&D spending, and evidence suggests that spending
on R&D is sensitive to the tax rate and credit incentives (see also Klassen et al., 2004; Gupta et al., 2006; Chen and
Gupta, 2009).
However, under certain conditions, nancial reporting incentives have the potential to mitigate the response of
investment to tax incentives. For example, in the accounting literature there is evidence that rms cut R&D spending to
boost earnings (e.g., Bushee, 1998). In addition, a recent study by Brown and Krull (2008) intersects nancial accounting
74
Increasing tax incentives to invest may have little impact on the quantity of new investment if supply is inelastic in which case increases in the
price of capital goods offset the increased tax benet (in other words spending may increase but part of the increased spending is for higher priced goods
(implicit taxes) rather than more goods). A very detailed examination of investment responses in the form of R&D spending to tax incentives controlling
for nontax incentives is Berger (1993). He nds that the credit induced $1.74 of additional spending per dollar of revenue lost by the Treasury. Berger also
documents a signicant implicit tax cost from the credit. Evidence of implicit taxes is not easy to produce for capital assets. Goolsbee (1998a) concludes
that capital goods prices do respond to changes in investment tax incentives. Hassett and Hubbard (1998) argue that U.S. rms are price-takers in a world
market for capital and, hence, shifts in U.S. demand arising from tax incentives will have no impact on the prices they face. House and Shapiro (2008) and
Desai and Goolsbee (2004) nd no evidence that supply prices reacted to bonus depreciation incentives.
75
See, however, Cohen and Cummins (2006) who nd no effect of the incentives on investment quantities and Edgerton (2009c) who nds no
evidence on changes in the relative prices of new and used construction machinery (used machinery did not qualify for the incentives). In addition, see
Edgerton (2009b) for an examination of the policys effectiveness in light of the presence of loss rms. In essence, he concludes that the incentives were
5% less effective than they would have been if all rms were fully taxable (and declines in cash ows reduced effectiveness much more).
76
See also Desai and Goolsbee (2004) for a summary of the literature and additional work on investment and tax policy.
77
For example, Bushman et al. (2008) argue that the research on investment and nancing constraints (e.g. Fazzari et al., 1988) is confounded
because the measure of nancing constraints, cash ows, also capture accruals that reect investment in working capital.
and tax reporting incentives in the R&D investment decision.78 Stock option exercises by R&D employees generate
R&D tax credits that decrease income tax expense and increase earnings, helping managers meet their earnings
target. The authors document that the tax credits from these option exercises reduce the amount by which rms decrease
R&D spending in periods with strong incentives to manage earnings. This evidence suggests that tax incentives
and the effect on income tax expense for nancial accounting mitigate real earnings management through R&D and,
conversely, that nancial reporting incentives may mitigate the effectiveness of tax incentives on real investment.
Additional theory and evidence along these lines would be useful to the academic literature on investment decisions.
The tax and book effects are at times conicting and could confound interpretations unless both effects are considered
simultaneously. Thus, an understanding of the relative importance of accounting choice and the value of accounting
earnings is important for policymakers if accounting outcomes can mitigate the intended response of tax policies.79 For
example, in Neubigs (2006b) testimony before Congress, he discusses the importance of the accounting effects for
tax policy, and argues that corporate executives prefer lower statutory rates over full expensing for capital investment in
part because lower rates increase accounting earnings.
4.1.4. Investment location decisions

While the effect of tax policies on aggregate investment has been difcult to document, research considering the tax
effects on investment location has been more successful. In addition, research in accounting has been more active in this
area relative to research on investment levels discussed above. Contrary to the Why did the chicken cross the road? joke
mentioned in Maydew (2001) (because taxes were lowerimplying this was a question with an obvious answer) and even
beyond his response about the interesting chickens (rms) are the ones that do not cross the road (or have to incur other
costs to cross the road), broader issues about investment location and taxes include: (1) how sensitive the chickens (rms)
are to taxes, (2) what happens to the economies on both sides of the road and to worldwide capital allocation because of
the road-crossing chickens (e.g., is there tax evasion on the high tax side via income stripping to the low tax side?), and (3)
how do or should policymakers respond? If countries desire to attract or retain investment then determining the effects of
the tax system on investment location is an important issue.
The largest area of research is on foreign direct investment (FDI), that is, signicant investment in foreign entities
(generally dened as a 10% voting interest) by U.S. taxpayers and investment (similarly dened) in U.S. entities by foreign
taxpayers.80 This research generally focuses on the allocation of capital in response to tax rate differences among
alternative locations and on the income shifting activities to low tax jurisdictions. A smaller, but related and important
area of research examines the issue of cross-border investment at the state level.81
Evidence from this literature consistently shows that host country taxes matter and that on average the tax elasticity is
negative, meaning that a decrease in host country tax rates (which increases the after-tax return on investment) leads to an
increase in FDI in that jurisdiction.82 Variation in this elasticity can be explained by industry-, time-, and country-level
factors. For example, capital ows to tax havens from non-manufacturing rms appear to be more responsive to taxes than
capital ows from manufacturers (see, e.g., Grubert and Mutti, 1991; Hines and Rice, 1994).83 Later studies show that
capital has become more sensitive to tax rates over time (Altshuler et al., 2000), that the openness of the trade regime
matters (Grubert and Mutti, 2000), and that cross-border merger and acquisitions are not as sensitive to host country tax
rates as other types of investment (Swenson, 2001).84
Hines (1999) estimates a signicant negative tax elasticity and concludes that high tax rates may generate a signicant
loss of FDI and may contribute to governments race to the bottom in competition for investment and that high home
taxation may lead rms to relocate their tax home. Hines also suggests that too much research focuses on the
responsiveness of FDI to taxes and that more is needed on the role of tax policy in affecting the form that FDI takes, the
possible importance of tax policy credibility and enforcement, and the relationship between tax and non-tax determinants
of FDI.
Traditional investigations of the link between taxes and FDI ignore the accounting treatment of taxes on foreign prots.
However, Shackelford et al. (2010) argue that U.S. rms are more likely to invest abroad if they value exibility in their tax
78
In addition, many rms now have greater investments in other intangibles which are typically subject to different expensing rules and faster
depreciating property (e.g., computer-based technology) quite possibly making rms investment responsiveness to targeted tax depreciation incentives
weaker (Desai and Goolsbee, 2004).
79
See also Robinson and Sansing (2008).
80
The ownership threshold to be counted as FDI is dened by country, but the OECD recommends a threshold of 10%. The Bureau of Economic
Analysis denes FDI as Ownership or control, directly or indirectly, by one foreign person, or entity, of 10% or more of the voting securities of an
incorporated U.S. business enterprise or an equivalent interest in an unincorporated U.S. business enterprise. This contrasts with foreign portfolio
investment (FPI), which covers cross-border investment in stocks and bonds for portfolio reasons. There is less research on FPI, but studying the tax
effects on FPI is an interesting and fast-growing area.
81
See Klassen and Shackelford (1998) and Gupta and Mills (2002) for examples.
82
See Hines (1997) and DeMooij and Ederveen (2003) for summaries. DeMooij and Ederveen provide a wide range of elasticities. See Hartman (1984,
1985) for early work and Slemrod (1990) for modications.
83
Grubert and Mutti (2000) use micro tax return data of more than 500 U.S. returns to construct an aggregated data set of average effective tax rates.
They estimate an elasticity of roughly 3.0 when the country has an open trade regime.
84
See also Wilson (1993), Kemsley (1998), and Single (1999) for accounting research on investment location decisions. Wilson (1993) and Single
(1999) examine other costs such as governance, customer location, and institutions. These are reviewed in Shackelford and Shevlin (2001).
and nancial reporting. The additional exibility in accounting is provided through APB 23 which allows rms to not
recognize the incremental U.S. income tax expense on foreign source earnings (see Appendix A for more details). By
designating earnings as permanently reinvested under APB 23, rms avoid an accrual for U.S. tax expense and increase
reported after-tax accounting earnings. Using this logic, Shackelford et al. essentially introduce non-tax nancial
reporting costs as a possible determinant of FDI. Consistent with this prediction, evidence from a recent survey of tax
executives by Graham et al. (2010a) indicates that 48% of publicly traded respondents with foreign earnings state that the
accounting effect from the deferral of income tax expense under APB 23 is an important factor in their decision to locate
operations overseas.
4.1.5. The reinvestment or repatriation decision

A manager also faces a decision about whether to reinvest the earnings on the rms foreign direct investments or repatriate
the funds to the domestic parent. Throughout the literature on the repatriation decision, there is some debate about the
relevance of a home country repatriation tax. For example, Hartman (1985) demonstrates that if the repatriation
and U.S. taxation of foreign earnings is inevitable and tax rates are constant (and these are crucial assumptions), then U.S.
repatriation taxes do not affect the decision of mature rms to either reinvest funds abroad or repatriate the earnings.85
Others argue that rms can avoid the repatriation tax and thus the tax is not a constraint on repatriating funds
(Altshuler and Grubert, 2003). Further, some estimates suggest that little U.S. tax is collected on foreign earnings (Grubert and
Mutti, 1995; Altshuler and Newlon, 1993; U.S. Treasury, 2007; U.S. Government Accountability Ofce, 2008; Dyreng and
Lindsey, 2009).
However, the assumptions for the irrelevance of repatriation taxes in Hartman (1985) are very specic and do not apply
in many contexts (e.g., rates of return and tax rates are not generally constant over time). In addition, Desai and Hines
(2004) estimate that the U.S. tax burden on foreign income is in the range of $50 billion per year, which they argue is
signicant. Tests of rm actions in a cross-section where the repatriation tax varies (Desai et al., 2001) and tests using
specic matched tax return data for the parent and subsidiaries (Altshuler and Newlon, 1993) provide evidence that
repatriations are sensitive to changes in the repatriation tax price. Further, Foley et al. (2007) provide evidence that
repatriation taxes can explain large cash balances.
Accounting concerns appear to play a role in the repatriation decision as well. Repatriation of earnings on investments
in low-tax jurisdictions triggers a U.S. tax liability and an increase in the reported income tax expense if the earnings were
previously designated as permanently reinvested. In their survey of tax executives, Graham et al. (2010a) report that 55% of
the respondents state that the income tax expense effect on the accounting income statement is an important factor in
their decision to repatriate or reinvest earnings.86
4.1.6. Corporate inversions to tax havens

The investment location discussion is not complete without a mention of tax havens. The OECDOrganization for
Economic Co-operation and Development (1998) denes a tax haven as a jurisdiction that imposes no or only nominal
taxes and offers itself as a place to be used by non-residents to escape taxes in their country of residence. The most notable
examples are Bermuda, the Cayman Islands, and Ireland. There are 35 such havens, 27 are island nations. These countries
hold less than 1% of the worlds population but host 5.7% of the foreign employment and 8.4 of the foreign property, plant,
and equipment of American rms (Hines, 2005). There are many interesting questions about tax havens that are beyond
the scope of our paper.87 What is relevant here is that tax havens exist (and, more broadly, that tax rates vary across the
world), and this has implications for investment location, debt location, accounting earnings, tax revenues and, as a result,
public policy.
In the late 1990s through 2001, a cluster of U.S. rms moved their tax homes outside of the U.S. in a transaction known
as a corporate inversion. Most rms claimed the inversions were designed to save U.S. taxes on foreign-sourced earnings.
Ingersoll-Rand expected to save $40 million a year, and Cooper Industries expected to save $55 million a year in U.S.
income taxes (Cloyd et al., 2003).88 However, policy makers were not only concerned about the opportunities to avoid U.S.
tax on foreign earnings, but about rms ability to shift domestic income out of the U.S. to the haven location. After the
proposed inversion of Stanley Works around the September 11, 2001 terrorist attacks and the anti-patriotism backlash
85
See Hartman (1985) and Scholes et al. (2009) for a model and discussion of a rms decision to reinvest or repatriate earnings.
86
The American Jobs Creation Act of 2004 granted a one-time dividend received deduction of 85% on repatriated foreign earnings for a specied time
period. Several studies examine the Act and companies responses to the Act. In general the results suggest that nancially constrained rms used the
cash to increase investment, and the remaining rms returning the repatriated funds to shareholders through share repurchases. See Blouin and Krull
(2009), Dharmapala et al. (forthcoming), and Faulkender and Petersen (2009). See Brennan (2008) for conicting results. Because of the way the tax act
and following interpretations and notices from the IRS were written, however, the Act cannot really be used to evaluate policy effects of specic tax
incentives to increase investment, although it can be used to test the response to a cash ow shock as the studies above do.
87
For example, some questions include (1) How do the economies of tax havens fare?, (2) How much investment do they attract?, and (3) What
happens to neighboring countries? See Hines (2005), Desai et al. (2005a, 2005b), Dharmapala and Hines (2006), Desai and Hines (2004), Hines and Rice
(1994), Desai et al. (2008), and Slemrod and Wilson (2009). These papers generally deal with macroeconomic tax policy effects, an area almost exclusively
dominated by economists, and in the interest of space we do not conduct a detailed review here.
88
There were several studies of this activity in terms of market reaction upon inversion announcement. The results of these studies were generally
mixed (see Cloyd et al., 2003; Desai and Hines, 2002; Seida and Wempe, 2004).
toward the company following its announcement, the U.S. government put stringent rules in place to prevent further
inversions.89 An even more recent set of inversions has occurred from the U.K to Ireland. For example, Shire and about a
dozen other companies in the FTSE 100 moved to Ireland in 2008, and most suspect the moves were in response to taxes
(Voget, 2009).
These tax-motivated restructurings have led to claims of reduced shareholder rights and potential degradation of
corporate governance.90 However there is very little evidence to support such claims.91 With hindsight, there are questions
about how these companies have fared after inversion, that is, did the restructuring lead to signicant non-tax costs and
implicit taxes, and have these rms lobbying costs increased to regain political favor? Seida and Wempe (2004) report that
effective tax rates do decrease for these rms.92 However, to our knowledge, we have no evidence about any other long-term
effects.
The inversion experience is still very relevant for understanding what lies ahead. Desai (2009) introduces the concept of
the decentering of the global rm, where he asserts that the relationship between rms and nation states has been
changing and continues to change in such a way that eventually it is plausible that rms will no longer be associated with a
particular country. In other words, there will not be U.S. multinationals or foreign multinationals. As nancial
accounting researchers study the effects of increased regulation from the Sarbanes-Oxley Act of 2002 (SOX) and the effects
on U.S. listings, onerous and expensive tax rules may similarly ship tax homes abroad and cause rms to separate the
location of the legal incorporation from the location of the executives, the listing location, and/or the main centers of
operations. This type of activity may have several implications. Desai (2009) argues that tax policies and some investor
regulations may have to be restructured because in many cases these are predicated on rms having a nationality. Most
certainly, the effects on corporate tax receipts will be an issue if many rms restructure or start-up with their tax homes
outside of the U.S.
4.1.7. Summary and thoughts for future research

Whether and to what extent taxation affects investment remains an important area of study. Although it is inherently
difcult to document aggregate shifts in investment in response to tax law changes, more evidence has been found
documenting cross-sectional responses in the timing, location, and type of investment.
One area of research ready for more work is the examination of how nancial reporting and tax incentives
jointly inuence the investment decision and a consideration of how this tradeoff or interaction affects tax policy.
There is already some evidence that suggests accounting quality affects investment decisions (e.g., McNichols and
Stubben, 2008; Biddle et al., 2009), but these studies generally ignore tax considerations. Moreover, there is some
indication from cross-country analysis that conformity in book and tax reporting distorts investment decisions (Cummins
et al., 1995). Furthermore, tax incentives to invest may be muted depending on how the investment and the tax incentives
affect the nancial statements, suggesting that accounting treatment of the tax incentive may be important (e.g., Edgerton,
2009d).
Foreign investment also continues to be an important area of research across accounting, nance, and economics.
Survey evidence in Graham et al. (2010a) suggests that the deferral of tax expense under APB23 is important
when rms make foreign investment decisions. Archival evidence to support this nding would be useful; however, the
research design would need to be careful and convincing that the tax and non-tax effects can be separated using archival
data.
The denition and measurement of investment and investment opportunities using accounting information are also
important issues. Research in nance and economics argues that correcting for measurement error in q is important.
Such measures, especially in nance, often wind up using some accounting measure of asset cost in the denominator
and thus are potentially confounded by known biases such as accounting conservatism as well as the accounting
methods for mergers and acquisitions. Moreover, investment in tangible assets, the measure of investment in many
studies, misses a lot of investment in the current economy, including advertising, research and development, and other
types of intangibles.
Finally, depending on the current and future administrations tax policy changes, corporate tax homes may be
established outside of the U.S. consistent with the decentering concepts in Desai (2009). A number of rms relocated their
tax headquarters to tax havens, but evidence on the long-run non-tax costs of undergoing inversions is limited.
89
For example, IRC sec. 7874 treats an expatriated entity that is 80% or more owned by the former shareholders of the domestic company as a US
company for tax purposes. In addition, the Stop Tax Haven Abuse Act included a provision that treats foreign corporations managed and controlled in
the U.S. as U.S. corporations for tax purposes (with an exception for CFCs).
90
Hanlon and Slemrod (2009) report statements from the press in this regard. For example, CalPERS (California Public Employees Retirement
System)) threatened to divest and block state purchases of stocks and bonds of U.S. companies that make use of foreign tax havens. The CalSTRS
(California State Teachers Retirement System) website (http://www.calstrs.com/Newsroom.Whatspercent20New/callforvote.aspx) contains an article
about institutional investors taking out a full page ad in USA Today calling for Ingersoll-Rand (which had moved its headquarters to Bermuda for tax
reasons) to move back to the U.S. citing a substantial loss of shareholder rights that goes with off-shore incorporation. Further, a Tax Briefs article from
Levin and Weiser, LLC (March 3, 2003, p. 6) states that CalPERS believes Bermuda incorporation makes it more difcult for Tyco shareholders to hold the
company, its ofcers and directors legally accountable in the event of wrongdoing.
91
One exception is Durney et al. (2009).
92
See also Dyreng and Lindsey (2009).
4.2. Capital structure
A long-standing question in corporate nance is the extent to which the corporate tax deduction for interest creates a
rst-order incentive to nance investment with debt. Modern attempts to address the question trace their roots to
Modigliani and Miller (1958) who show that in a world with no taxes, agency costs, bankruptcy costs, or information
asymmetries, the choice between debt and equity does not affect rm value. This means is that if capital structure does
affect rm value, it must be because debt nancing affects tax burdens, agency problems, the likelihood of bankruptcy and
the information environment.
Modigliani and Miller (1963) recognize that corporate interest deductions can create incentives for rms to use debt,
but Miller (1977) argues that personal taxes on interest income increase pre-tax yields on debt to the point where the
corporate tax advantage is completely offset. While Miller assumes all interest costs are fully deductible, there are limits on
the deductibility of interest payments. DeAngelo and Masulis (1980) show that when rms can exhaust their interest tax
shields, a rm-specic optimal capital structure exists. This basic framework for thinking about debt and taxes leads to a
number of testable predictions. In the time series, debt usage ought to be increasing in corporate tax rates, but decreasing
in the relative tax cost to investors. In the cross-section, leverage ought to be higher for rms that have higher expected
corporate tax rates, fewer non-debt tax shields, and expect to be protable in the future.
Auerbach (2002) and Graham (2008) provide comprehensive surveys of the empirical evidence on taxes and capital
structure choice. They conclude that capital structure decisions appear to respond to corporate tax incentives, but
agreement is not universal, especially with regard to the extent to which the tax incentives matter. Graham notes that
some academics (e.g., Myers et al., 1998) contend that, tax incentives are of third-order importance in the hierarchy of
corporate decisions. The apparent weakness of the evidence is generally explained by a combination of considerations. For
example, existing theories can lead to tests with limited power because they yield only qualitative predictions.93 In
addition, the corporate tax benet of leverage is potentially offset by the investor-level tax disadvantages of debt, and
these offsetting tax effects are difcult to disentangle. Finally, debt issues have long stated maturities and thus expected
future marginal tax rates, while ignored in the literature, are important.
An understanding of the relation between debt and taxes must also confront tax rate measurement issues and nancial
reporting issues which are the focus of our remaining discussion on capital structure. Empirical attempts to understand
capital structure decisions rely on accounting-based proxies for marginal tax benets that contain known shortcomings.
In addition, accounting rules, managerial discretion, and nancial innovation affect the measurement of nancial
instruments and contracts for book and tax reporting purposes. The non-reporting of many items may cause important
measurement error in estimates of rm leverage.
4.2.1. Estimating the tax benets of debt

The tax benets from incremental interest deductions depend on the rms marginal tax rate, dened here as the
present value of the corporate tax on an additional dollar of income or expense (Scholes et al., 2009). Firms with higher tax
benets from debt are expected to choose higher leverage ratios, all else constant. Early studies estimate the tax benets
using the existence of non-debt tax shields such as depreciation, investment tax credits, and net operating loss
carryforwards (e.g., Bradley et al., 1984; MacKie-Mason, 1990; Dhaliwal et al., 1992). The idea, known as the substitution
hypothesis, is that rms with substantial non-debt tax shields are less likely to nance with leverage. However, Bradley
et al. (1984) nd that leverage is positively associated with non-debt tax shields. This seemingly anomalous nding is due
to the fact that rms with high depreciation costs and investment tax credits likely have more assets in place and fewer
growth options, and hence are more likely to nance with debt. MacKie-Mason (1990) and Dhaliwal et al. (1992) address
some of the endogeneity issues by looking at incremental leverage choices and capturing the extent to which the rm has
exhausted incremental tax benets.94
Another approach to derive estimates of tax status is to categorize rms based on their current year net operating
losses (NOL) and income (for example, the dichotomous or trichotomous tax rates discussed in Shevlin, 1990 and
elsewhere). This can be a reasonable rst approximation (Graham, 1996b), but it ignores important dynamics of the tax
code that treat income and loss asymmetrically. For example, this static measure would classify a rm that has both
current losses and net operating losses as having a marginal tax rate of zero. But loss carryforward provisions allow the
rm to retain some of the tax benets into the future; the categorical variables cannot account for the effects of loss
carryovers.
The most recent and most sophisticated estimates of marginal tax rates are derived from a simulation procedure that
uses historical information about the rms level and volatility of estimated taxable income to project a path of expected
taxable income into the future (Shevlin, 1990; Graham, 1996a). Building in features of the tax code such as NOL provisions,
investment tax credits, and the Alternative Minimum Tax (AMT), this approach calculates the present value of tax
93
Furthermore, there is considerable debate over the correct theoretical models of leverage decisions. The debate is largely focused on comparing the
traditional tradeoff model that grew from the work of Modigliani and Miller with the pecking order model that stemmed from the work of Myers (1984).
See Frank and Goyal, 2009 for a review and summary of this debate.
94
Also, Ayers et al. (2001) report evidence consistent with the substitution hypothesis with regard to owner-debt for taxable corporations and for
non-owner debt across all types of organizational forms.
on an additional dollar of taxable income to estimate the marginal tax rate.95 Most studies using the simulated tax rate
nd consistent evidence of a positive association between a rms (pre-nancing) marginal tax rate and the use of debt.
The evidence also suggests that simulated marginal tax rates outperform static estimates based on categorical variables or
effective tax rates (Shevlin, 1990; Graham, 1996b; Plesko, 2003; Graham and Mills, 2008), however several caveats are
in order.
First, estimating taxable income from the nancial statements has known and probably unknown problems.96 Shevlins
and Grahams simulations estimate taxable income as pre-tax book income less deferred tax expense grossed-up by the
statutory tax rate (i.e., pre-tax book income [deferred tax expense/statutory tax rate]). Thus, any item that reduces
taxable income but is not in book income and not already accounted for in the deferred tax expense will not be accounted
for in estimated taxable income in the MTR estimation procedure. For example, prior to SFAS 123R, no expense related to
stock options was recorded for nancial accounting purposes and the difference between book and tax was not accounted
for through deferred taxes, leading to overstated estimates of the marginal tax rate (Hanlon and Shevlin, 2002).
Overstatement of the marginal tax rate potentially leads to the conclusion that rms do not use enough debt (e.g., Graham,
2000) and studies that recognize and adjust the estimate of marginal tax rates for known problems are better able to
identify the effect, if any, of corporate taxes on leverage (e.g., Graham et al., 2004).97
Additional sources of measurement problems include: (1) transactions that generate permanent book-tax differences,
(2) research and development tax credits, and (3) tax rate differentials between domestic and foreign operations (see
Appendix A). The effects of these items could be large. For example, a signicant portion of the income of multinational
corporations is sourced to foreign jurisdictions; thus, using the U.S. statutory tax rate in the MTR estimates may severely
overstate the MTR the rm faces. In many cases, adjusting the simulated tax rate for these differences is currently
impossible because of limited disclosure. Indeed, the MTR is estimated in the literature as a worldwide MTR, however,
rms likely face jurisdiction specic MTRs based on local statutory tax rates and net operating losses in each jurisdiction.
Graham and Tucker (2006) address a potential source of measurement error: corporate tax shelters. They conjecture
that sophisticated tax planning strategies (i.e., tax shelters) may reduce the true marginal tax rate but may not be captured
by the simulated marginal tax rates if the tax shelter creates a permanent book-tax difference (a difference not accounted
for in book income or deferred taxes, the two components used to estimate taxable income in the simulation procedure).
Consistent with their prediction, Graham and Tucker nd that rms accused of sheltering income from tax use less debt.
In other words, tax shelter use is a non-debt tax shield that simulated rates may miss and could be a partial explanation for
why rms appear underlevered in Graham (2000).98
If the nancial statements generate noisy measures of taxable income and true tax status, estimating the marginal tax
rate using actual tax return data would seem to represent a substantial improvement. Using condential tax return data,
Graham and Mills (2008) nd that the simulated tax rates using book estimates of taxable income appear superior to those
based on taxable income reported on the U.S. tax return. One explanation is that for multinational rms, consolidated
nancial accounting reports give a better view of worldwide tax status relative to U.S. tax return data that excludes the
income of foreign subsidiaries.99
Finally, the accuracy of simulated tax rates is subject to particular assumptions about the time-series properties
of taxable income. Shevlin (1990) and Graham (1996b) assume that changes in taxable income follow a random walk. This
approach has recently been criticized in Blouin et al. (forthcoming) as leading to biased estimates of the marginal tax rate.
95
Shevlin (1990) uses NOL dynamics in his simulation. Graham (1996a) extends Shevlin (1990) to incorporate the investment tax credit (ITC) and the
AMT. Importantly, these simulated tax rates can be estimated for income before nancing costs (that is, the marginal tax rate on the rst dollar of debt, as
in Graham et al., 1998) or after nancing costs (the marginal tax rate on the next dollar of debt, as in Graham, 1996a).
96
See Hanlon (2003).
97
Unfortunately, the stock option deduction gures must be hand-collected for periods prior to SFAS 123R, which makes this adjustment costly on a
large scale. However, the stock option deduction is less problematic under SFAS 123R. The amount of expense recognized for book purposes over the
vesting period of the option will now create a book-tax difference that will be reected in deferred tax benet (for the future deduction related to the
recognized expense; now it is a temporary difference). Thus, this portion of the deduction will be picked up by the typical simulation methodologies and
will no longer need separate adjustment. However, it is also the case that there will likely be a difference in the book expense and the actual tax deduction
in the future year because the book expense is an ex ante estimation of the future value of the option while the tax deduction is the actual difference
between the stock price and the strike price on the date of exercise. This latter difference will essentially be accounted for as under the old rules. The extra
tax deduction, if there is one, is not shown in book income or in the deferred tax expense and, thus, would not be incorporated into the conventional
estimation of taxable income in the marginal tax rate computation.
98
It is important to recognize the difference in this study and the option study noted above. Graham and Tucker use 44 rms that have received a
notice of deciency from the Internal Revenue Service. They examine whether rms accused of using tax shelters use less debt than a matched sample,
but do not adjust the marginal tax rate for any effects of the tax shelter. They cannot really adjust the marginal tax rate for the effects of the shelter
because for most rms in their sample they would not have the detailed data for which to make the adjustment and to the extent the shelters generate
temporary differences no adjustment would be necessary. In contrast, note that Graham et al. (2004) in their study of the tax benets of stock options
directly adjust the MTR because during the time of their study the entire effect of the stock option benets was not in book income nor in deferred taxes
(the two components used in the MTR estimation) and the amount of the needed adjustment could be estimated. It is important not to adjust the MTR for
items or transactions that affect book and tax differently but where the difference is captured by deferred tax expense because these items are already
included in the MTR estimation procedure.
99
Thus, even though the rate applied in the MTR estimation is likely erroneous when there is income in foreign jurisdictions (because the rate used is
the U.S. statutory rate), using nancial statement data includes the income from foreign jurisdictions whereas it is completely missing when looking at a
U.S. tax return for the rm.
Blouin et al. argue that the bias in the rates explains Grahams (2000) conclusion that rms are underlevered, that a
non-parametric approach is more appropriate for forecasting taxable income, and that once the estimates are done in this
manner, rms are not nearly as underlevered as Graham (2000) suggests. In response, Graham and Kim (2010) argue that the
non-parametric approach in Blouin et al. (forthcoming) is problematic and contend that an AR(1) model provides the best
estimates of simulated dynamic marginal tax rates. We look forward to a resolution of this matter and recognize these
studies in their attempt to rethink and improve upon the marginal tax rate estimation process.100
To summarize, many accounting and corporate nance studies rely on marginal tax rate estimates derived directly from
accounting information in order to test and empirically document the effects of taxes on debt use. The development of the
marginal tax rate estimation procedure (Shevlin, 1990 and Graham, 1996a) and its application to the corporate capital
structure literature are clearly signicant contributions. However, if the marginal tax rate estimates are incorrect and the
errors alter inferences from prior research, then improvements to the measure or developments of alternative proxies of
tax status will be valuable in furthering our understanding of tax effects on corporate capital structure.101,102
4.2.2. Measuring leverage

Nearly all rm-level analyses of capital structure decisions rely on the nancial statement representation of the rms
nancing sources. Leverage is often measured as long-term debt divided by book or market value of assets. However,
accounting rules, and the managers application of those rules, can lead to signicant problems in measuring true nancing
structure that are largely ignored in empirical research. For example, the balance sheet ignores many economic liabilities,
including operating lease commitments, debt of unconsolidated equity method investments, and until recently, pension
obligations.103
The classic framework that considers the tradeoff of tax and non-tax costs on the choice of nancing breaks down when
there is exibility in how nancial claims are classied for tax and non-tax purposes (Auerbach, 2002). Financial
instruments can be designed to give corporations the best of all worlds for tax, accounting, and regulatory purposes.
For example, structured nancing activities can provide interest deductions for tax purposes, but not show up at all on the
nancial statements. Mills and Newberry (2005) exploit condential tax return data to show that rms often report higher
interest expense on their tax return than in their nancial statements. This difference is at least partially explained by the
use of non-conforming nancing methods such as R&D limited partnerships, synthetic leases, and securitizations.104
Instruments like trust preferred securities were also used for a time because the rm could claim the tax deduction for
interest, but report the liability between the debt and equity sections on the balance sheet. Engel et al. (1999) document a
number of cases where rms incurred non-trivial issuance costs to retire debt by issuing trust preferred stock. In effect,
many rms were willing to incur real costs merely to lower their reported leverage while keeping the tax treatment, and
the underlying nancial risk, intact.
Another measurement issue arises when rms can allocate debt within the rm to exploit differences in tax treatment
across jurisdictions and activities; however, this allocation has no effect on the consolidated worldwide balance sheet. The
evidence is consistent with U.S. multinationals allocating debt based on jurisdiction-specic tax-loss carryforwards and
foreign tax credit considerations (Newberry and Dhaliwal, 2001). Similarly, Huizinga et al. (2008) provide evidence on tax-
motivated debt shifting for European rms. There is also aggregate evidence that rms allocate debt across taxable and
non-taxable activities within Real Estate Investment Trusts (REITs), a structure whose tax provisions allow the non-taxable
parent to carry out certain activities within a taxable subsidiary (Matheson, 2008).
It is well known to those who teach nancial reporting that accounting rules lead to misstatements of the rms liabilities.
One potential problem for research is that the link between debt and the marginal tax rate could in theory be driven by the
choice between on and off-balance sheet nancing rather than variation in real leverage. Because rms vary in their use
100
An additional point about the MTR estimation is that investment and nancing decisions involve cash ows over multiple periods, but marginal
tax rates estimate the tax benet of an interest deduction taken today. Understanding how these rates evolve over time and the effect of future expected
marginal tax rates on leverage are important issues.
101
An alternative way to estimate taxable income might be to gross-up the current tax expense by the statutory tax rate. Which method would have
more error is unknown, but using current tax expense would incorporate the effect of permanent differences and reect that tax credits reduce the
marginal tax rate. However, it would create other measurement problems because that method treats credits as book-tax differences and the use of
current tax expense may include more of the tax contingency reserve than the use of deferred tax expense.
102
One approach is to identify a set of rms whose marginal tax rate is known to be zero. A straightforward extension of Modigliani and Miller
suggests that taxable corporations ought to hold more debt than their tax-exempt counterparts (see also Scholes and Wolfson, 1989), all else constant.
Guenther (1992) and Gentry (1994) nd that non-taxable organizational forms such as partnerships have lower leverage and changes in leverage.
However, Omer and Terando (1999) present evidence consistent with some of that difference being attributable to risk and liability differences and not
taxes. Barclay et al. (2010) nd no difference in leverage between taxable real estate rms and REITs, suggesting that variation in corporate tax status
does not affect the debt decision in that setting. See Barclay et al. (2010) and Ayers et al. (2001) for a discussion of the predictions regarding debt and
taxes across organizational forms. A key concern in using organizational form as an instrument for tax benets of debt is the endogeneity of the
organizational form decision. Firms also choose organization form based on non-tax reasons. If those factors correlate with the variable of interest it
becomes difcult to draw strong conclusions from the data. However, efforts to modeling this decision are likely to encounter substantial difculties
because the characteristics of the rm have likely changed since the choice was rst made, and switching costs may prevent many rms from adopting an
alternative structure. One way around the problem is to consider leverage trends for rms that switch forms.
103
See Graham et al. (1998) who consider the effect of operating leases and Shivdasani and Stefanescu (2010) who consider the effect of
off-balance sheet pensions obligations.
104
See Altamuro (2006) and Zechman (2009) for further detail and evidence on incentives to use synthetic leases.
of off-balance sheet nancing for reasons tied to nancial reporting incentives, understanding to what extent the off-balance
sheet nancing affects the analysis of taxes on the choice between debt and equity is an open question.
4.2.3. Summary and thoughts for the future

The extent to which corporate taxes are important in the leverage decision and the extent to which rms leave tax
benets on the table are still open questions in which measurement and reporting issues are likely to play an important
role.105 Marginal tax rate estimates are affected by a variety of measurement issues. Moreover, accounting rules ignore
many liabilities in the measurement of reported leverage. Innovation in nancial instruments and the ability to structure
nancing within the rm creates valuable opportunities for tax arbitrage with limited non-tax consequences. Finally, the
recent credit crisis has reopened the classic debate about the tax code providing too many incentives for debt. Research that
addresses the issues above is clearly warranted.
4.2.4. Payout policy

Taxes, specically investor-level taxes, may also affect rm payout policy decisions. This important question has been
widely researched, and recent reviews provide in-depth synthesis and analysis (e.g., Allen and Michaely, 2003; Poterba,
2004; Graham, 2008; Kalay and Lemmon, 2008; DeAngelo et al., 2008). The basic conclusion from the literature is that
investor taxes do not appear to be a rst-order determinant of rm payout policies. Even recent evidence from the
substantial dividend tax cut of 2003 is mixed. To our knowledge, there is no tradeoff in terms of nancial accounting since
evidence suggests that repurchases offer not only lower individual-level taxes but also more nancial accounting benets
when managers xate on increasing earnings per share (see Bens et al., 2003; Graham et al., 2005).
4.3. Organizational form
In this section, we focus on the role of taxes on the organizational form choice of start-up and existing businesses,
as well as the interaction between organizational form, taxes, and the sale of the rm. Organizational form is a
fundamental issue for businesses and is discussed at length in Scholes et al. (2009).106
4.3.1. The choice of organizational form

The entrepreneurs choice of legal structure determines how it is taxed and thus, taxation is a factor in the choice of
organizational form. The available alternatives and the tax and non-tax considerations are well summarized in accounting,
legal, and entrepreneurial nance texts.107 Some forms are not separately taxed; business income ows through to the
owners and is taxed in the same period at the investor level. These ow-through options include sole proprietorships,
partnerships, limited liability companies, and S corporations. Most publicly traded entities are C-corporations organized
under subchapter C of the Internal Revenue Code.108 In a C-corporation, a tax is rst assessed on the taxable income
of the corporation and distributions to shareholders are then taxed at the shareholder level. This creates the oft-cited
double-tax penalty of the corporate form. However, not all distributions are taxed as dividends at the individual
level.109
Much of the research concludes that taxes affect organizational form choice, but the evidence also suggests that non-tax
factors often dominate the decision (Guenther, 1992; Ayers et al., 1996; MacKie-Mason and Gordon, 1997).110 Empirical tests
often exploit time-series changes in relative tax rates between corporations and individuals. For example, after the
enactment of the Tax Reform Act of 1986, the corporate form became much more tax-disadvantaged relative to forms that
105
Investor-level taxes are also expected to play a critical role. See Dhaliwal et al. (2006), Graham (1999), Green and Hollield (2003), Graham
(2008), and our discussion of investor-level taxes and asset pricing in a later section for more detail.
106
Throughout, we are referring to organizational form for tax purposes. Organizational form can represent a broader set of decisions such as
whether to operate in a single industry or many industries, to have centralized or decentralized management, to operate in a single jurisdiction or many
jurisdictions, and so on. The non-prot is an important organizational form in many industries such as education and health care. Although we focus our
attention on alternative for-prot structures, we acknowledge a number of recent accounting studies looking at the role of taxes and agency problems in
the modern non-prot, such as Yetman (2002), Yetman (2003), and Sansing and Yetman (2006).
107
Scholes et al. (2009) provides a thorough discussion and a basic mathematical framework for thinking about analyzing the effect of taxes on
returns to operating in corporate versus ow-through.
108
Exceptions include publicly traded partnerships and Real Estate Investment Trusts (REITs). Currently, a partnership can be publicly traded only if
its income is passive in nature. See IRC y7704. A REIT must distribute 90% of its income to shareholders. Technically, the 90% threshold is required to
maintain REIT status. REIT status allows the REIT to deduct dividends paid from taxable income. Dividends are treated as deductions, so the rm must
actually distribute 100% of taxable income to avoid corporate tax. Under recent IRS rulings, stock dividends count toward the 90% threshold as long as the
investor is taxed. Finally, we note that non-C-corporations can elect to be taxed as corporations under check-the-box regulations. See IRC y 301.7701-3
for more details.
109
Distributions in excess of the earnings and prots of the corporation are taxed as capital gains or not at all (i.e., return of basis), repurchases and
liquidating distributions are treated as capital gains, and many shareholders are tax exempt. See the Internal Revenue Code and the regulations there
under (and applicable case law) for more detail. For the basics, see IRC sections 301302 and section 331.
110
Ownership structure is also important when considering the importance of taxes on organizational form choice. For example, private equity rms
are organized as partnerships and not corporations, at least in part, because of the tax rules (Bank and Chefns, 2008). The partnerships tax status
currently affords the carried interest portion of the partners compensation to be taxed as a capital gain rather than ordinary income. For a brief summary
of tax and governance effects on ownership structure and patterns more generally, see Desai and Dharmapala (2008).
imposed a tax only at the individual level, providing incentives to undertake investment through non-corporate form.
Scholes and Wolfson (1991) document an explosion of S corporation elections surrounding the end of 1986. Seventy-ve
thousand elections were made throughout all of 1985, while 225,000 were made in the ve weeks surrounding year-end
1986.111
Because the organizational form choice made at start-up can be revisited throughout the rms life, rms that switch form
present an alternative method to identify the role of taxes. Aggregate level data suggests that important non-tax reasons exist
for using corporations and that the cost of switching forms is non-trivial (Gordon and MacKie-Mason, 1994; MacKie-Mason
and Gordon, 1997; Goolsbee, 1998b).112 Indeed, most public rms simply cannot switch out of C-corporation form and still
remain widely held. However, real estate rms can and do change form. The conversion of a taxable real estate corporation to
a REIT requires the rm to distribute accumulated earnings and prots (the tax reporting analog of retained earnings) as a
taxable dividend. Damodaran et al. (1997) document that tax savings are important in switching to REIT form; however, for
nancially distressed rms, the ability to retain exibility in investment and payout policies dominates the decision.113
Another dimension is how organizational form interacts with decisions to expand or nance certain activities. For
example, Petroni and Shackelford (1995) provide evidence consistent with property-casualty insurers structuring their
cross-state expansion, through licensing or subsidiary, in a manner that mitigates both state taxes and regulatory costs.
Shevlin (1987) and Beatty et al. (1995) examine how rms nance research and development activities through a
separate partnership jointly owned with another party or in house nanced by debt or equity and consider taxes,
accounting effects, and agency costs as part of the decision-making process. Examining the organizational form decision for
banks, Hodder et al. (2003) extend of the tradeoff decision to include nancial accounting considerations, specically the
effect of accounting for income taxes, on the organizational form decision. Banks with deferred tax assets appear less likely
to convert away from C-corporation status (to ow-through S-corporation status) because the write-off of deferred tax
assets potentially exposes the banks to regulatory capital issues. The authors also examine how banks that change
organizational form undertake real actions, such as selling assets and strategically setting dividends, to offset non-tax
costs. Detailed studies such as this not only contribute to the academic literature, but could be useful in policy debates.
Knowing how companies respond to tax law changes and which factors make them respond differently in a cross-section,
e.g., nancial accounting concerns, is important for estimating the economic effects of a tax law change (and in
documenting how important nancial accounting earnings are to rms).114
4.3.2. Summary and thoughts for future research

There has been a sizable body of research on the effects of taxation on choice of organizational form and how that
organizational form decision is related to other decisions of the rm such nancing and payout policies. Taxation appears
to be an important factor, but limited liability, transactions costs, and operational exibility are also important and reect
fundamental economic tradeoffs discussed by Scholes and Wolfson. Extending the literature to consider agency costs, the
accounting for income tax effects, and the real adjustments rms make to mitigate tax and non-tax costs ex post have been
among the recent contributions.
A number of interesting organizational form issues remain unanswered. For example, whether the tax system helps or
impedes rms that must adapt their organizational form to compete with new rms is an open question. An interesting
empirical observation is that many large, public corporations have ownership interests in ow-through entities. What are
the costs and benets of this organizational form choice?
In addition, there is a growing literature in nance, economics, and accounting that seeks to understand the causes and
consequences of entrepreneurial activity, the structure of nancial contracting between entrepreneurs and capital providers,
the design of the monitoring and control system in new ventures, and the exit decision.115 Taxes on operating income and
111
See Plesko (1995) for cross-sectional tests and Omer et al. (2000) for evidence on the choice in the natural resource industry (specically other tax
provisions built in gains are especially prevalent in that industry that offset the rate changes) during this time period. See also Ayers et al. (1996) for
evidence on small business organizational form choices.
112
MacKie-Mason and Gordon (1997) nd that reducing the tax burden on the non-corporate income leads taxpayers to shift assets out of corporate
form, but the sensitivity of organizational form changes to the tax disadvantage of the corporate form is economically small. Goolsbee (1998b) estimates
the deadweight loss of the corporate income tax to be around 5% based on data from 1900 through 1939. Goolsbees data begin nine years before the rst
corporate income tax and 13 years before the rst individual tax, which means that non-tax factors were the only driver of organizational form decisions
for a large fraction of the sample. Gordon and MacKie-Mason (1994) focus on estimating the non-tax benets of the corporate form by estimating the
incremental tax cost of the corporate form to shareholders. Based on estimates using aggregate data, non-tax benets average 3.8% of equity value.
113
A 2001 IRS ruling opened the door for taxable industrial rms to effectively reorganize their real estate assets as REITs via a REIT spin-off.
Goolsbee and Maydew (2002) estimate the tax savings to rms with large real estate holdings to be non-trivial. The upside of the transaction is that
shareholders can own the real estate through a tax-advantaged entity without triggering accumulated gains. However, for any taxable real estate rm,
stand-alone or subsidiary, the switch to REIT status is not free. There may be a one-time dividend depending on accumulated earnings and prots (which
determines the amount of payout that constitutes a dividend), the interest on debt nancing is not tax deductible, and limitations are placed on real
estate sales. Matheson (2008) reports that the use of taxable REIT subsidiaries grew four-fold to $68 billion in 2004 following the REIT Modernization Act
of 1999 which permits many REITs to earn so-called tainted income within a Taxable REIT Subsidiary (TRS) without jeopardizing their REIT status.
114
For excellent discussions of the importance of the accounting for income tax effects with respect to tax policy choices see Neubig (2006a, b).
115
See, for example, Cumming (2008), Hellmann (2006); Hellmann and Puri (2002); Kaplan et al. (2009), Kaplan and Stromberg (2004), and
Maksimovic and Phillips (2008). In addition, a recent stream of literature in accounting focuses on new venture activity and the reporting systems of the
rm (Davila and Foster, 2005; Allee and Yohn, 2009; Cassar, 2009). In the law literature, see Bankman (1994) and Bankman and Gilson (1999).
upon exit of the business are one factor to consider in the decision to undertake entrepreneurial activity.116 More research on
the effect of taxes in the start-up and venture capital-backed rm would be useful, especially given the magnitude of this
activity in the economy. For example, why are so many start-ups organized as C-corporations? How do taxes affect the
structure of compensation contracts and nancing policies in venture capital-funded rms? Does the monitoring provided by
tax authorities interact with other control mechanisms in the venture capital market (e.g., Desai et al., 2007a)?
Finally, as the U.S. considers various tax reform proposals, one possible scenario includes an increase in individual-level
tax rates and a decrease in corporate tax rates. Under such a scenario the tax incentives to be a C-corporation will increase
and the incentives to be a ow-through will decrease. To what extent will this affect organizational form choice? Will the
changes have a measurable effect on the activities of start-ups? We look forward to future research on these topics should
the U.S. adopt such a policy.
4.4. Taxes and other decisions: transfer pricing, mergers and acquisition, compensation, and trading
There are many other manager or corporate decisions that interact with taxes, nancial reporting, or other non-tax
costs. In the interest of space, we briey highlight selected issues in the areas of transfer pricing, mergers and acquisitions,
compensation, and trading by insiders. The research in each of these areas is more limited relative to the decisions
discussed above which presents opportunities, but perhaps more challenges as well.
4.4.1. Transfer pricing

Transfer pricing is the pricing of goods and services transferred between two or more units of the same company. There
are a myriad of issues involved including source of income and deductions for tax purposes, division performance
measurement within the rm, and customs. In 2001, 40% of U.S. international trade and 20% of world trade was intrarm
trade making the economic signicance of transfer prices substantial (Tang, 2002).
Tax issues arise when transfers are between units located in different jurisdictions sales between a unit in a low tax
jurisdiction and one in a high tax jurisdiction (whether state or country). Beyond the pricing of traditional inventory items
between divisions, transactions include the recent tax saving strategies where one company establishes a foreign subsidiary
or entity (partnership or corporation) in a low tax jurisdiction to own an intangible asset such as a patent or trademark. The
other divisions of the entity, usually in more highly taxed jurisdictions, pay a fee for the use of the patent or trademark.117
There is a relatively large literature on transfer pricing including studies of the effect of tax rates and regulation on
production and pricing decisions, investment incentives, allocation distortions, and distributional effects. Most of these
studies document that tax rates are a signicant factor in the transfer price established (Capithorne, 1971; Horst, 1971;
Samuelson, 1982; Halperin and Srinidhi, 1987; Harris, 1993; Jacob, 1996; Harris and Sansing, 1998; Sansing, 1999; Smith,
2002; Bartelsman and Beetsma, 2003; Clausing, 2003; and others). Baldenius et al. (2004) contend that a central issue for
rms is a tradeoff of the tax considerations of the entire company with the internal divisional performance measures.118
Possible alternatives for dealing with the issue are to compensate division managers on rm-wide performance, use
separate prices for tax purposes and internal purposes, or implement some adjustment procedure for internal evaluation
and control purposes. Baldenius et al. derive solutions for the optimal transfer price under different scenarios for a
multinational rm.119
Chan et al. (2006) extend the examination to relate rm organization, agency costs, and tax compliance. Chan et al.
study the effect of management centralization on tax compliance for foreign owned subsidiaries in China. The results
suggest that managers with autonomy over transfer prices (and who are evaluated by division performance) have fewer
tax authority audit adjustments to transfer prices than managers where the transfer price is set by the parent company.
The authors reason that the managers act in their self-interest, consistent with agency theory when given the autonomy to
do so resulting in greater tax compliance (i.e., less tax-motivated transfer pricing).
Transfer pricing issues have led to some of the largest tax settlements in U.S. history. Data obtained via survey or some
other source in the future could provide insights on many outstanding questions. For example, what are the cross-sectional
116
Gentry and Hubbard (2004) argue that asymmetry in tax treatment of gains and losses and progressivity of tax rates causes even risk-neutral
investors to demand higher returns to invest in assets with riskier returns. Thus, taxes affect aggregate risk-taking/entrepreneurial activity. (Graham and
Smith (1999), discuss the benets of hedging in this scenario.). Gilson and Schizer (2003) argue that taxes have a substantial effect on the use of
convertible preferred equity securities by VC investors. Cullen and Gordon (2007) show that taxes can affect entrepreneurial activity by changing the
tradeoff between business and wage income, by changing the effects of asymmetry in the tax code (through progressive rates and organizational form),
and by risk sharing with the government. See also Poterba (1989) and Guenther and Willenborg (1999).
117
Transfer pricing can apply to rent of property between divisions as well.
118
Arguably transfer pricing is not a real decision but merely a relabeling to internally price a real intracompany transaction. However, the pricing
decision potentially affects performance measurement within the company and is an important decision given the tax consequences.
119
Tang (2002) uses survey and case study evidence and concludes that in his sample, there are generally few conicts between tax and control
objectives because many companies use some adjustment mechanism to minimize the conicts between the two objectives. Baldenius et al. (2004)
acknowledge this and show in their models that when the product is sold to external parties and market prices are used to establish the arms length price
for tax purposes, internal transfers should be subjected to intracompany discounts that increase with the tax rate differential between the divisions. Most
managerial accounting textbooks say balancing the two objectives is an issue but due to the paucity of evidence stop there. As the IRC increases collection
efforts and focuses on transfer pricing the balancing act may become more delicate.
determinants of the various transfer pricing arrangements and how the transfer pricing issues have changed, if at all,
following the introduction of the advanced pricing agreements with the IRS where rms and the IRS can agree to transfer
prices and avoid dispute. How do rms balance divisional performance measurement and company-wide tax objectives?
In addition, how does the ability to set (tax motivated) transfer prices affect rms real investment, investment location,
and repatriation decisions, if at all?
4.4.2. Mergers and acquisitions

Taxes are potentially important in merger and acquisition activity because (1) a tax can be generated on the transaction,
(2) tax attributes, such as net operating losses, can be acquired and used to offset the acquiring rms tax liabilities (subject
to limits e.g., IRC Section 382), (3) there can be new tax planning opportunities post-acquisition, and (4) the targets tax risk
(e.g. the potential tax liability from prior aggressive tax avoidance) may affect deal negotiations. There may also be
nancial reporting interactions, for example, when allocating purchase price to the acquired assets.
The empirical literature on mergers and acquisitions is well established and covers an array of topics including the
pricing of acquisitions, the role of deal competition, information asymmetry and agency problems in the negotiation
process, and the effect of industry shocks and regulation on merger activity. We refer the reader to a number of
comprehensive surveys of the literature including Jensen and Ruback (1983), Jarrell et al. (1988), and Betton et al. (2008)
for a balanced background on the issues. The most successful advances on the link between taxes and mergers and
acquisition will nd their motivation and contribution to the broader literature on mergers and acquisitions. Our coverage
is largely focused on understanding the role of taxes at the transaction level.120
Evidence on the valuation of NOLs in M&A activity is largely mixed. Auerbach and Reishus (1988), Hayn (1989), and
Erickson (1998) document little evidence that NOL benets matter. Perhaps the growing number of loss rms could
provide sufcient power to detect the existence or absence of such effects. Because the valuation of a targets tax assets is
idiosyncratic, a useful approach is to model the acquirer-target choice in a world where there are alternative potential
bidders that have different valuations of a targets tax assets (e.g., by estimating pseudo-mergers as in Auerbach and
Reishus, 1988). Indeed, anecdotally, it seems some rms are better at making acquisitions to obtain tax benets. For
example, when the IRS unexpectedly lifted some of the loss limitation rules for nancial institutions during the recent
nancial crisis there was an increase in competition for and valuation of targets with NOLs.121
There is also some evidence that acquisitions create synergies by increasing tax planning opportunities. Huizinga and
Voget (2009) nd that when two rms from different countries merge, the new parents tax home is more likely to be in
the country with the lower total tax burden. In addition, Erickson and Wang (2007) predict that organizational form
matters and estimate that S-corporation shareholders can sell for higher prices because the sale can be structured to
increase future tax deductions for the buyer.122 Huizinga and Voget (2009) notwithstanding, much of the literature on deal
structuring is based on domestic acquisitions. An open question is how taxes affect acquisitions across borders. Do some
countries provide more favorable tax treatment of a corporate sale than others? Does this explain patterns in cross-border
merger activity and foreign investment?
The evidence on shareholder taxation and merger pricing is mixed.123 Erickson (1998) nds no evidence that target
shareholder tax liabilities affect acquisition price, which he attributes to marginal shareholders having high bases in target
stock (or low potential tax liability). Ayers et al. (2003) nd that cash acquisition premiums are increasing in individual
capital gains liabilities (using institutional ownership to identify this effect) and show that the effect is absent for tax-free
deals. See Landsman and Shackelford (1995) for evidence from the takeover of RJR Nabisco.124
There are a number of unexplored areas where tax and non-tax incentives are likely to intersect in the takeover market.
First, agency problems can be important in the context of merger negotiations (e.g., Wulf, 2004), and target rm CEOs and
directors have substantial power to block acquisition attempts. It is not uncommon for large shareholders such as family
founders to constrain the buyers payment method to stock if they want to defer taxation. Research that addresses the role of
120
Research on aggregate merger activity is centered on the role of market timing (e.g., Rhodes-Kropf et al., 2005) and industry shocks (e.g., Harford,
2005). With the exception of work by Ayers et al. (2000) who nd that the effect of goodwill tax deductions did not affect the level of M&A activity,
although it did affect the prices in applicable transactions.
121
See Crowell Moring Financial Services Alert dated October 6, 2008 Tax Notice Drives Wachovia Takeover Turmoil about the IRS changing the
loss limitation rules for banks and the immediate entrance of Wells Fargo as the acquirer of Wachovia at a much higher price than the agreed price for the
same acquisition by Citigroup.
122
An important implication of this analysis is that the owners consider the method and timing of exiting the investment when choosing an
organizational form.
123
The pricing impact is almost always measured by looking at premiums. Yet that requires an implicit assumption about the role of taxes on stock
prices in the absence of a merger.
124
Ayers et al. (2004) nd that the likelihood of a stock deal is increasing in the capital gains rate over the sample period, but is mitigated when
targets have greater ownership by institutions. Dhaliwal et al. (2004a) nd evidence that a taxable owner of a hospital demands a higher price than a tax-
exempt owner of a similar asset, consistent with taxable sellers demanding compensation for incurring an immediate tax liability. Later, we discuss the
validity of the various approaches to market equilibrium in the context of long-run pricing and dividend taxes. The issue is somewhat different in this
context because the price-setting shareholder is the investor whose reservation price is met to obtain control of the target. When institutions own more
of the target, individual capital gains taxes matter less, and ownership merely reects the relative shape of the supply curve faced by the acquiring rm
seeking to gain control of the target. Thus, knowledge of the underlying tax attributes of target shareholders could be important in determining the price
paid.
tax preferences of the CEO or other controlling shareholder on acquisition structure and price would be interesting. Second,
evidence in Ofcer (2007) suggests that purchase prices are lower for unlisted rms and argues that the discount arises because
the seller demands liquidity. Interestingly, the unlisted targets in Ofcers sample are largely subsidiaries, overlapping with
Erickson and Wang (2000) who nd a correlation between the tax structure of a subsidiary acquisition and the purchase price.
To what extent tax structuring affects the estimates of the discount is an open question. Third, many studies including Ayers
et al. (2004) consider publicly traded buyers. Because private company buyers almost always use cash, it is important to
understand how capital gains taxation affects the total mix of deals, not just those with public buyers. Evidence in Bargeron
et al. (2008) documents that private buyers offer lower premiums than public ones. Does this have a tax explanation?
Moreover, are there any book-tax interactions in takeovers that need to be considered? Finally, as mentioned in more detail
above, how the targets existing tax risk factors into the acquisition price or merger negotiations is completely unknown.
4.4.3. Executive compensation

Taxes also play a role in the design of compensation plans, including the choice between cash and equity compensation,
between different forms of equity incentives, and between current and deferred compensation. Because Shackelford and
Shevlin (2001) and Graham (2008) review this literature in detail, we keep our discussion brief and draw the following
observations. First, it is not uncommon for studies on executive compensation to control for the tax status of the rm
(e.g. Core and Guay, 1999), but there is little evidence that corporate tax rates affect the choice between cash and stock
compensation. Second, there is some evidence that companies alter the mix of incentive stock options and non-qualied
stock options around corporate and individual tax rate changes but rm-specic evidence is less convincing (see
Shackelford and Shevlin, 2001). There is a nancial accounting tradeoff considered in limited settings. For example,
Matsunaga et al. (1992) examine disqualifying dispositions of incentive stock options by rms and in a very detailed
manner incorporate the nancial accounting costs and benets as a factor in rms decisions to disqualify the options to
obtain tax benets.125 Finally, the literature still lacks evidence on deferred compensation plans despite the more
restrictive provisions of IRC 409A which became law in 2005 following the American Jobs Creation Act of 2004.126
4.4.4. Executive trading

Although the executives decision to trade stock is typically not a rm decision, we briey discuss some of the recent
literature on individual-level tax incentives and executive trading for the following reasons. First, if executives exercise
options or sell or hold shares as a result of taxes, portfolio incentives to maximize returns and take risky projects change,
possibly leading directors to adjust the compensation package.127 For example, like other taxable investors, the manager
faces a lock-in problem if the stock has appreciated. That is, the insider will face signicant tax costs from selling, but not
selling will only worsen the under-diversication problem. Evidence in Jin and Kothari (2008) suggest that managers that
face the greatest tax burden from selling equity divest fewer shares.128 This line of inquiry can be expanded a number ways
to consider alternative measures of the tax cost of selling, the effect of nancial contracts that allow the executive to
diversify while deferring the tax on the gain, whether the effect is different for shares divested by gift, and the effect of
changes in the capital gains tax rates in 1997 and 2003.
Second, there is also evidence that individual tax incentives caused some executives to backdate stock option exercises,
and this has consequences for the rm. Briey, the exercise of a non-qualied stock option generates an ordinary income
tax liability for the employee to the extent of the options intrinsic value on the day of exercise. If the employee holds the
shares to sell at a later date, any subsequent appreciation will be taxed as capital gains (and any loss available to offset
capital gains or to deduct to the extent of $3,000). For shares held long-term, the manager has a tax incentive to time these
exercise-and-hold transactions on days when the stock price is as low as possible. This converts ordinary income to lower-
taxed capital income.129 These tax incentives appeared strong enough to lead some executives to backdate the exercise
date. Before the enactment of SOX, Dhaliwal et al. (2009) nd that almost 14% of exercise-and-hold transactions by CEOs
are on the day with the lowest closing price of the month. The percentage increases to 22% among CEOs of rms implicated
in option grant backdating.
125
See McDonald (2003) and Blouin and Carter (2010) for research on rm and employee decisions to make an IRC Section 83(b) election to
accelerate tax on share-based compensation.
126
This is, in part, due to the lack of data on deferred compensation plans. Sundaram and Yermack (2007) point out that, while almost all companies
have deferred compensation plans, disclosure of amounts deferred by executives is generally not required under SEC disclosure rules. Recently enacted
deferred compensation rules under IRC 409A reduce the exibility of structuring deferred compensation arrangements by requiring that deferral
elections be made further in advance, that the process be well documented, etc. Failure to meet the requirements leads to immediate taxation as well as a
20% excise tax. Given the extensive use of deferred compensation plans, it would be helpful for researchers to identify new data sources. Such an analysis
would also improve understanding of how deferred compensation arrangements are structured.
127
We recognize there are many non-tax issues involved. See, for example, Seyhun (1986), Heath et al. (1999), Ofek and Yermack (2000), Core and
Guay (2001), Meulbroek (2001), and Malmendier and Tate (2005).
128
It is possible that opportunities exist to get the benets of the sale without the associated tax cost. Jagolinzer et al. (2007) document insiders use
of prepaid variable forward contracts (PVFs) to cash in on their holdings while at the same time deferring tax liability. Basically, PVFs provide discounted
cash proceeds to the insider in return for some future settlement that depends on rm performance. While the insiders claim to future appreciation is
capped, they were generally able to defer capital gains liability until sale.
129
Evidence in Goolsbee (2000) suggests that executives were more likely to exercise and hold their shares following the 1997 reduction in capital
gains tax rates.
The accelerated insider ling requirements mandated by SOX appeared to have largely curbed backdating of grants and
exercises (Heron and Lie, 2007; Dhaliwal et al., 2009), but a number of questions remain about the CEOs personal tax
decisions. First, are CEOs who are aggressive in personal tax matters more or less aggressive with the rms tax planning?
Second, given the rather small estimated tax savings from backdating documented by Dhaliwal et al. (2009), what does this
suggest about the expected costs of tax compliance? Third, why do executives exercise options and hold the shares instead
of immediately selling them? Exercising to start the clock rolling for capital gains treatment is generally thought to be
suboptimal (McDonald, 2003; Scholes et al., 2009). We have little understanding of the incentives and constraints that
drive the decisions of executives to exercise and hold shares.
Taxes may also provide incentives to gift shares, changing the executives equity holdings. CEOs and other executives
often have substantial wealth invested in their own company stock and can give their shares away to family and non-prot
organizations. There are well-recognized tax benets to donating appreciated stock. The gift is deductible at fair market
value (but escapes the capital gains tax that would be due if the stock were sold for cash and the after-tax cash donated).
An insider that commits to a xed-dollar donation has an incentive to time the transaction when stock price is high to
minimize the number of shares he has to give up.130
Consistent with a tax-based incentive to time stock gifts when stock price is high, Yermack (2009) nds that the
rms stock price subsequently declines by over 3% during the 20 days following large CEO stock gifts to their own family
foundation. He documents that $17.4 billion worth of stock was donated in a total of over 1,000 gifts over a two and
one-half year period ending in 2005. Of the $17.4 billion, $14.2 billion represented gifts to trust funds and other
entities controlled by the CEO. Gifts to donees not disclosed in SEC lings totaled $2.43 billion in CEO stock.131 Tax
considerations are likely to have a rst-order effect on the the timing and structure of gifts by executives and individuals
more broadly.132
5. Investor-level taxes and asset prices
The link between investor-level taxes and asset prices is perhaps the most widely studied area of common interest;
however, the underlying motivation for the research is often different across economics, nance, and accounting. In public
economics, for example, there has been a long-standing debate over the effect of dividend taxation on economic growth
and on rms marginal investments nanced with retained equity. In nance, the goal is more often to understand the
implications of dividend and capital gains taxation for capital structure, payout policy, and portfolio allocation. In
accounting, the primary focus is studying whether investor-level taxes affect asset prices. We rst review research on the
effect of investor-level dividend taxation and then move to the effects of the capital gains tax.
5.1. Dividend taxation
We begin with a brief discussion of the underlying theory surrounding dividend taxation and asset pricing from the
public economics literature. We then discuss the various methods of testing the effects of investor-level dividend taxes on
asset prices in the context of long-run returns, event studies, ex-dividend day pricing, and valuation models. We follow
with a summary of some of the key issues, including the ongoing debate over market equilibrium assumptions.
5.1.1. The economic effects of dividend taxation: a brief introduction

For the past several decades, research on the economics of dividend taxation has centered on three basic theories: the
irrelevance view, the traditional view, and the new view. At the core, the views are based on valuation models that, as a
result of different assumptions, offer competing predictions about how dividend taxation affects investment and dividend
policies through the rms cost of capital. Public economists interest in these models is to understand whether the
dividend tax distorts rm investment and payout decisions, and to understand the effects of moving to a dividend
exemption or imputation system. The different views offer different answers to each of those questions. We offer a brief
introduction to the competing views.133
130
But if an executive is interested in transferring control to descendants, he may want to transfer shares when stock price is low so as to maximize
the number of shares that can be transferred without triggering gift tax. Such a strategy would need to take place within broader estate tax planning
objectives.
131
Stock gifts represent an important means of disposing of shareholdings and must be disclosed to the SEC, but there is signicant opacity about the
recipient which limits the potential usefulness of the disclosures.
132
The incentive contracting literature measures a CEOs incentives as the sensitivity of CEO wealth to a change in stock price (Jensen and Murphy,
1990; Core and Guay, 1999). These incentives are usually expressed on a pre-tax basis. However, incentives from equity portfolios come from stock
options, restricted stock, and unrestricted shares. For unrestricted shareholdings, gains and losses are taxed at the long-term capital gains rate, and can be
deferred until death. Appreciation in restricted shares is effectively taxed at ordinary income rates. Options gains are taxed at the ordinary income tax
rate and cannot be deferred past the expiration date. This variation across incentives has been largely ignored in the literature. It would be useful to
understand the extent to which measuring portfolio incentives on an after-tax basis, rather than a before-tax basis, makes a difference in interpreting
prior results.
133
This discussion does not touch on the more complicated issues. Extant theory typically ignores what happens when rms can nance with debt.
Because projects are evaluated based on their net present value using a discount rate that factors in capital structure, a project nanced with new debt is
To be clear, we are not advocating that accounting research take up the task of trying to distinguish which of the views holds
empirically. However, we describe the alternative views to provide a background of the underlying theory of dividend taxation.
Under the tax irrelevance view, taxable individuals are infra-marginal, meaning a non-taxable entity (e.g., a pension
fund) or symmetrically taxed investor (e.g., a securities dealer) is the relevant price-setter (Black and Scholes, 1974; Miller
and Scholes, 1978, 1982). Thus, changing the tax rate on dividends does not affect expected returns, and so investment,
payout, and nancing decisions are unaffected.134 In contrast, the following two views implicitly or explicitly assume that
the individual-level tax on dividends affects the price of corporate equity.
Under the traditional view, there are non-tax benets from paying dividends, such as investor preferences for dividends
(Gordon, 1963), controlling agency problems (Jensen, 1986), or signaling performance (e.g., Bhattacharya, 1979). The manager
sets dividend policy at the point where the marginal benet of an extra dollar of dividends equals the marginal tax cost.
Reductions in the dividend tax rate can lower the required rate of return, lead to greater investment, and higher dividend payouts.
Economic models of the traditional view make two key assumptions: dividends are paid for non-tax reasons and the
marginal source of funds for investment is new equity issues (assuming no debt). Under the traditional view, reductions in
dividend tax rates increase share values and increase investment incentives for dividend-paying rms because they lower
the pre-tax required rate of return. In addition, a reduction in the tax lowers the marginal cost of paying dividends, leading
to higher dividend payouts. These so-called efciency consequences of dividend taxation lend support to proposals to
reduce or eliminate dividend taxation because, if eliminated, there would be fewer distortions.
In contrast, under the new view of dividend taxation, dividend taxation has no effect on investment nanced with retained
equity or on payout decisions and hence has no distortionary effects (Auerbach, 1979; Bradford, 1981; King, 1977). The
assumptions of this model are that retained earnings are the marginal source of investment funds and all retentions will
eventually be distributed as taxable dividends. The market value of equity capitalizes all expected taxes on current and future
dividends, even for non-dividend-paying rms. In addition, dividends are viewed as a residual after investment, implying
that dividend payouts should uctuate year to year. The model predicts that an increase in the dividend tax rate leads to
lower equity prices, but dividend yields per se do not explain rm value (because all taxes are already impounded into price,
even for non-dividend-paying rms) and investments and payout decisions are not affected. These implications apply when a
dividend tax rate change is expected to be permanent. Temporary changes could distort investment and payout decisions by
changing payout policy preferences. The assumptions of a constant tax rate, constant rate of return, and full distribution of
earnings as taxable dividends means that shareholders can get a dividend now or let the rm reinvest at a constant rate of
return and pay a bigger dividend later. The present value of the dividend tax is the same.135
Thus, under the irrelevance view and new view, dividend taxation does not cause distortions because it does not affect the
investment or payout decisions of the rm. Under the new view existing shareholders would, however, receive a windfall
gain upon a reduction in the dividend tax (Table 2).136 Because each theory relies on strict assumptions that are unlikely to
hold in reality (see Zodrow, 1991 for a good discussion), they should be viewed as benchmarks and tools that assist in
designing and interpreting empirical analysis.137 We next discuss the various methods used to examine the
effect of dividend taxes on asset prices: long-run return studies, event studies, ex-day pricing studies, and valuation models.138
(footnote continued)
implicitly nanced with a combination of debt and internal equity. A rm with existing debt that raises new equity may be closer to the traditional view.
But because dividend taxation does not affect the cost of debt, at least directly, the investment decisions of rms with more debt are likely to be less
sensitive to dividend taxation. This general proposition is not surprising if rms with extreme debt levels are compared. Further, the reality of tax law
changes makes a determination of what is temporary and what is permanent difcult, although this distinction is crucial for the new view models
implications for rm behavior.
134
Note the irrelevance view is consistent with a Modigliani and Miller world of no taxes. Miller and Modigliani (1961) show that in perfect and
complete capital markets (including no taxes) dividend policy does not affect the value of the rm. This implies that if dividend policy does affect rm
value, it does so because dividends signal managers private information, dividends reduce agency costs of free cash ow, or dividend policy affects
investors tax liabilities, which in turn affects price. See DeAngelo and DeAngelo (2006) for a discussion of the relevancy of Miller and Modigliani.
135
The new view is also referred to as the trapped equity view or tax capitalization view. A basic premise of this view is that earnings cannot be
distributed out of the corporation without incurring the dividend tax; repurchases are not allowed. This view is generally thought to apply to more
mature rms. In addition, recent studies relax the assumption that all retentions are distributed as dividends and still yield new view results of no
distortions.
136
See Table 2 for a summary of implications of the various views.
137
Both the new and traditional views are subject to some commonly mentioned criticisms. For example, rms can repurchase shares or be taken
over and hence never pay dividends, a result that seems inconsistent with new view assumptions. Dividend policy is sticky, which is inconsistent with the
new view implication that dividends are merely the residual. The traditional view has its own issues. It assumes that new equity is the marginal source of
nancing, but not many rms issue equity after the initial offering. The traditional view also depends on the validity of non-tax explanations for
dividends.
138
Shackelford and Shevlin (2001) discuss a line of research based on price level regressions and the incorporation of investor-level taxes into an
Ohlson (1995) type valuation model. These studies were written in an attempt to identify whether dividend taxes are capitalized into price and adopt the
assumptions of the new view. Harris and Kemsley (1999) and Collins and Kemsley (2000) interpret their model and results as evidence of dividend tax
capitalization at the highest individual statutory tax rates on rms level of retained earnings. At the time, Shackelford and Shevlin (2001) state that future
analysis would be needed to determine if the studies were revolutionary or implausible. Dhaliwal et al. (2003) and Hanlon et al. (2003) show that the
ndings are implausible and the models used are non-diagnostic with respect to dividend tax capitalization. While the research did serve to stimulate a
lot of thinking about the issues, it is important to note that this debate did not settle whether the new or traditional views apply empirically to dividend
taxation. The only end result of the debate is the conclusion that the models and tests used in the original studies cannot inform us about the pricing of
dividend taxes.
Table 2
The predicted economic effects of dividend taxes.
New view Traditional view Irrelevance view
Effects of a tax rate increase on investment, payout, and share value

@Investment 0 o0 0
@td
@dividend 0 o0 0
@td
@Value o0 o0 0
@td
@Value 0 o0 0
@td @d
Effect of dividend payout and taxes on expected returns

@R 40 ? 0
@d
@R 40 ? 0
@d@td
R is the expected pre-tax return, d is the dividend yield, and td is the individual level tax rate on dividends.
5.1.2. Long-horizon returns: the effect of dividend yield on expected returns

Following the theoretical structure of Brennan (1970), a common test of the effect of dividend policy and dividend taxes
on long-horizon asset prices follows the relation between expected return on dividend yield, controlling for systematic
risk, e.g.:
ERj a0 Rf a1 bj gdj 1
where Rj represents rm returns, Rf is the risk-free rate, bj is the covariance between rm and market returns, d is the
expected dividend yield of the rm, and g represents the price of dividends. That is, g denotes the compensation
demanded by investors to buy shares that pay tax-disadvantaged dividends.
One way to see the effect of dividends on expected returns is to rst assume that the non-tax benets of paying
dividends are zero and that rms must eventually distribute all excess cash ows as dividends. These are key assumptions
of the new view that allow the interpretation of g to be isolated from non-tax explanations. Let the after-tax required
return, r, for a given riskless security equal (1 td)d + (1 tcg)g, where td is the tax rate on dividend income, tcg is the
effective tax rate on capital gains that accounts for benets of deferral, d is the dividend yield, and g is the net appreciation
in share price. The pre-tax return, R, is the sum of d and g. After substitution, we get
r t tcg
R d d 2
1tcg 1tcg
That is, the pre-tax return is simply the after-tax required return grossed-up for expected capital gains taxes plus a
premium that compensates shareholders for receiving a portion of these returns as dividends. In more structured
asset pricing models such as Brennan (1970), Gordon and Bradford (1980), and Poterba and Summers (1985), the term
(td tcg)/(1 tcg) is the theoretical dividend tax premium reected in expected returns. In this simple setting, the dividend
yield coefcient (g) reects the tax rates of investors.
Interpreting the empirical evidence is more complicated if we assume there are non-tax reasons rms pay dividends, a
key assumption under the traditional view and the subject of a substantial body of research in nance (see DeAngelo et al.,
2008 for a recent survey). Under this view, the prediction for g depends on how these non-tax benets of dividends show
up in returns. If the non-tax benets of paying dividends are linearly related to expected dividend yield, are not adequately
controlled for by the other variables in the model, or if the dividend yield proxies for some unobservable risk factor, then it
may be impossible to denitively interpret the g as a tax effect. Extant research offers little intuition as to how the non-tax
aspects of dividend yield should show up in either prices or returns (see Fama and French, 1998 for a discussion). Creative
identication strategies are needed to isolate the effect of investor taxes.139
In an early empirical study, Black and Scholes (1974) nd no association between dividend yields and stock returns,
leading them to conclude that dividend policy has no permanent effect on rm value. Litzenberger and Ramaswamy (1979)
claim to nd evidence of dividend tax pricing using short-horizon returns, but Miller and Scholes (1982) argue that the
ndings are driven by information effects, not taxes.
Later studies document a positive association between dividend yields and long-horizon returns. But such yield effects
are difcult to attribute to dividend taxes (Chen et al., 1990). Naranjo et al. (1998) nd that the positive association
139
Importantly, it is nearly impossible to use evidence from long-run horizon return evidence to distinguish between the new and traditional views.
As Poterba and Summers (1985) note, the predictions are similar as long as one can isolate the effect of taxes from non-tax explanations. On the other
hand, the evidence does provide meaningful inferences on the validity of the irrelevance view.
between returns and dividend yields is too large to be explained by taxes alone and does not vary with changes in the
statutory tax rate on dividends. The breadth of the ndings leads Naranjo et al. to conclude, First, a yield effect clearly
exists. Second, we nd no evidence that it is attributable to taxes.140
In contrast, Dhaliwal et al. (2003) document evidence that suggests the signicant association between dividend yield
and expected returns is driven by taxes. They assert an equilibrium in which prices depend on the tax status of some
identiable marginal investor. Further, they argue that institutional ownership percentage captures meaningful variation
in the identity of this marginal investor. In their tests, the association between dividend yields and expected returns is
weaker when institutions own more, leading Dhaliwal et al. to conclude that the positive coefcient on dividend yield is
attributable to investor-level taxes.141 Dhaliwal et al. (2004c) present similar evidence using ex ante estimates of expected
returns that are based on derivations of the Ohlson (1995) model (e.g., Gebhardt et al., 2001; Claus and Thomas, 2001; and
others). Sialm (2009) examines the association between expected returns and the tax yield, measured as the component of
expected return attributable to investor-level dividend and capital gains taxes. His ndings suggest that shareholder taxes
increase expected returns on nearly a one-to-one basis.
These recent studies provide some of the rst evidence from long-run returns that dividend taxes are priced, but their
ndings raise a number of questions. For example, how important is the issue economically given the historical difculty of
documenting this effect?142 Is the identication strategy using institutional ownership both consistent with theory and
sufciently powerful to detect differences across rms? How should the results from an asset pricing model be interpreted
in the presence of non-tax benets? Finally, how does the evidence inform the debate on the links between dividend
taxation and investment or payout policy? We return to these issues below.
5.1.3. Event study predictions: the valuation of a dividend tax rate change
Event studies, usually surrounding a change in tax law, offer a potentially more powerful test of the valuation of
investor taxes. It is arguably easier to control for risk, agency costs, and information effects that are not expected to change
signicantly over a short window of time. However, not all events are created equal. An ideal event should involve an
economically signicant change in tax rates, be unexpected by market participants, be permanent in nature, and not run
concurrent with other tax or non-tax policy changes or other events that might affect price.
There is evidence to suggest that asset prices are affected by changes in investor tax rates. For example, stock prices
declined surrounding the 1993 announcement of an increase in top ordinary income tax rates from 31% to 39.6% and prices
declined more for rms with high dividend yields and less for rms with higher institutional ownership (Ayers et al.,
2002). In addition, Erickson and Maydew (1998) document signicant price declines in preferred stock, but not high yield
common stock, following a 1995 Treasury proposal to cut the dividends received deduction from 70% to 50%.
The Jobs and Growth Tax Relief Reconciliation Act of 2003 signicantly reduced the top individual statutory tax rate on
dividends from 35% to 15%, and included an expiration, or sunset, date that has varied over time. Auerbach and Hassett
(2007) examine stock price reactions to events leading up to the bills uncertain passage and nd that rms with high
dividend yields and rms that paid no dividends had the most positive price reactions.143 Dhaliwal et al. (2007) document
a decrease in accounting-based cost of capital estimates surrounding passage of the 2003 tax cut, which they attribute to
the reduction in dividend tax rates. Similar to Auerbach and Hassett (2003), they also nd that non-dividend-paying rms
beneted substantially from the tax cut.144
The evidence from price changes around tax law announcements is consistent with dividend taxes being capitalized
into price, albeit at low rates, and with the capitalization varying with dividend yield and ownership structure.145
140
Chen et al. (1990) add a bond risk premium measure to the model and form portfolios based on both yield and rm size. They measure dividend
yield using the sum of dividends over the prior year divided by beginning of year share price, and use a variety of structural models. Naranjo et al. (1998)
use an updated measure of dividend yield (based on most recent dividend and share price data) and the three-factor capital asset pricing model.
141
Interestingly, Dhaliwal et al. (2003) cast their ndings as evidence of dividend tax capitalization, a term that Poterba and Summers (1985) and
others use to describe the new view. But they also interpret the evidence as consistent with the traditional view. In fact, their results could be consistent
with both views and they cannot distinguish between them.
142
For example, Naranjo et al. (1998) nd no evidence that statutory tax rates explain the yield effect, while Dhaliwal et al. (2003) nd mixed
evidence. Sialm (2009) presents some of the strongest evidence that tax rates affect expected returns, but his measure of tax burden commingles tax rates
and dividend payouts. Ironically, the strongest evidence on dividend tax pricing comes from periods when tax-exempt investors made up a larger fraction
of shareholders and the personal tax-disadvantage of dividends was lowest.
143
However, because both the traditional and new views predict an increase in share prices in reaction to such a tax change, a test of the price
reaction will not distinguish between the views (see a similar discussion in Poterba and Summers, 1985). Auerbach and Hassett use cross-sectional
variation in yield in an attempt to design a distinguishing test. The authors contend that zero dividend paying rms having such a strong price reaction is
consistent with the new view because these rms have the most future dividends to pay.
144
Note that under the new view any positive impact of the dividend yield on short window returns around a tax rate reduction is attributable to the
timing of the path of future dividends, with a higher yield corresponding to a greater share of the rms future dividends being paid before the tax rate
reduction (Auerbach and Hassett, 2007). Thus, the zero dividend rms having a higher response is consistent with the new view, even though it seems
somewhat counterintuitive. However, Amromin et al. (2007) argue that the positive return on non-dividend-paying rms is not related to taxes since the
prices of non-dividend-paying rms around the world increased over the same period. See also Edgerton (2009a) for evidence from real estate investment
trusts.
145
In addition to price response, see Desai and Dharmapala (forthcoming) for a study of individual portfolio allocation in response to the Jobs and
Growth Tax Relief Reconciliation Act of 2003 (JGTRRA). This study reports a statistically and economically large response in terms of international
portfolio allocation of U.S. investors to equity holdings in foreign countries affected by the tax act.
The weight of the evidence from the papers discussed above seems to refute the irrelevance view, however, Amromin et al.
(2007) question whether the 2003 tax cut in particular had any aggregate effect on prices. In addition, tax regime changes
are endogenous and highly correlated with other developments in the economy, past, present and future (Romer and
Romer, 2008). The potential concerns that arise from treating tax rate changes as exogenous need to be considered in light
of the research question, tests, and results.
5.1.4. Ex-dividend day studies

A third approach to identifying the pricing of dividend taxes is to look at stock returns surrounding the ex-dividend day.
After the announcement of a dividend, an investor can purchase the shares with rights to the dividend. Beginning on the
ex-dividend day, the shares will sell without rights to the dividend. In the absence of taxes, transaction costs, and other
imperfections, the stock price is expected to decline on the ex-day by the amount of the dividend. However, when dividend
and capital gain income are taxed differently, a taxable buyer facing a penalty on dividend income prefers to wait until
after the stock goes ex-dividend, while a taxable seller prefers to accelerate the sale to just before the ex-day to avoid
paying ordinary income tax on the dividend. A well-known empirical fact is that the decline in stock price on the ex-day is
less than the full amount of dividend. Elton and Gruber (1970) argue that taking the ratio of the ex-day price decline to the
dividend, often known as the price-drop ratio, reects the tax status of the marginal investor.146 A large and growing
stream of literature provides both support and challenges to Elton and Grubers proposition. We discuss only a small
sample of studies here and refer readers to Allen and Michaely (2003) and Kalay and Lemmon (2008) for more
comprehensive surveys.
That the ex-day price drop is consistently less than the amount of the dividend is an empirical regularity in the U.S.,
although the price drop was closer to the full dividend amount in the U.S. before the income tax (Barclay, 1987). The
smaller-than-dividend price decline has been attributed to transactions costs (Kalay, 1982) and market microstructure
effects (Bali and Hite, 1998). In addition, Eades et al. (1994) nd that variation in ex-day behavior is unrelated to changes in
the tax regime and Frank and Jagannathan (1998) nd evidence of ex-day price drop ratios less than one in Hong Kong, a
country with no investor taxes.
If the ex-day price drop ndings are driven by individual taxes, the effect should weaken following the 2003 tax cut,
making this an ideal setting to test the ex-day tax explanation. Chetty et al. (2007) nd a large shift in ex-day pricing
around the time of the tax cut. However, they also document large year-to-year volatility in ex-day returns that is
fundamentally unrelated to taxes. Because of this extreme volatility, they conclude that they are unable to make any
inferences on the relation between dividend taxation and stock returns surrounding the 2003 tax cut. The time-series
volatility documented in Chetty et al. could have signicant implications for any study using time-series changes in tax
rates to identify tax effects in prices or returns.147
Can ex-day return patterns provide any information on how investor taxes affect the long-run pricing of stocks? In the static
clientele view of Elton and Gruber (1970), the ex-day return merely reects the tax status of a marginal investor in the security.
However, in addition to the non-tax explanations cited above, Allen and Michaely (2003) challenge this ex-day interpretation
by arguing that volume increases over time and the presence of heterogeneously taxed investors suggest dynamic clientele
effects in which ex-day returns are unlikely to provide any information about the effect of taxes on equilibrium pricing. Thus,
they argue that ex-day pricing effects are more likely to represent the tax incentives of short-term traders.
5.1.5. Evidence from valuation models

An alternative way to test whether stock prices are discounted for dividend taxes is to test the association between
dividends and value. Fama and French (1998) report a positive relation between yield and rm value and conclude that
substantial non-tax factors drive the yield effect and that dividends appear to be an informative signal about protability.
Their evidence does not suggest that taxes do not matter, but that any tax effect is swamped by other factors because of the
inability to isolate information content, agency costs, and risk explanations of dividends.148
5.1.6. Summary
Across the various methods of testing the implications of dividend taxation, the evidence is still mixed. There is growing
evidence that the irrelevance view does not hold and that taxes matter in the pricing of stocks. However, taxes generally
appear to be capitalized at low rates in event study tests, long-run return studies depend on cross-sectional variation using
a proxy of institutional ownership (discussed further below), and recent evidence in Amromin et al. (2007) suggests that
146
The price drop ratio can be derived by rst stating the condition under which the investor would be indifferent between buying before or after
stock goes ex-dividend (ignoring adjustment for risk). Let P0 and P1 be the stock prices before and after the stock goes ex-dividend, B is the basis in the
shares, D is the dividend amount, and td and tcg are the tax rates on dividends and capital gains. The indifference condition is P1 (P1 B)tcg +D(1 td) =
P0 (P0 B)tcg. After re-arranging terms, we get an expression for the price drop ratio in terms of taxes, specically, ((P0 P1)/D) =((1 td)/(1 tcg)).
147
Chetty et al. (2007) argue that more work employing trading data is needed to identify the effect of taxes and tax arbitrage on stock prices. A study
that might serve as an example here is Dhaliwal and Li (2006) who provide evidence that trade-level proxies for heterogeneity in investor tax status are
associated with ex-day trading volume.
148
See Gentry et al. (2003) and Li and Weber (2009) for evidence on dividend pricing in REITS. See Sialm (2009) for evidence on the association
between his measure of tax yield and market-to-book.
aggregate effects from a substantial dividend rate cut may have been trivial. While the documentation of any association
between dividend taxes and returns or prices has been difcult, empirical tests to distinguish between the traditional view
and the new view have been nearly impossible because of the endogenous nature of tax law changes and the fact that
temporary tax law changes makes the predicted results under the new view and traditional view very similar.
5.1.7. Remaining issues and questions for future research

5.1.7.1. How relevant is the marginal investor? It is often assumed that the equilibrium price of the rm is characterized by a
single marginal investor whose identity determines the pricing of taxes (Miller and Scholes, 1982; Harris and Kemsley, 1999;
Ayers et al., 2002; Dhaliwal et al., 2003). In that case, dividend tax effects depend on the tax rates of that marginal investor.149 This
approach is plausible if we consider two securities whose returns are perfectly correlated and that differ only with respect to their
tax treatment. For example, the Miller (1977) equilibrium is possible because he assumes that the two securities (debt and equity
in his case) have identical risk (see Schaefer, 1982 for a good discussion). But in general, an investor is unlikely to nd a pair of
securities with identical risk that vary only on the tax treatment of the returns (yields and tax rates). Because diversication is
important to investors, an alternative approach that accounts for risk-sharing incentives is likely more appropriate.
The alternative is the after-tax capital asset pricing model (CAPM) approach developed by Brennan (1970) and Gordon and
Bradford (1980). In the after-tax CAPM, heterogeneously taxed investors form portfolios to maximize after-tax return. In a
nutshell, taxable individuals hold dividend-paying stocks because they cannot replicate their risk exposure with non-dividend-
paying stocks, they just hold slightly less because of the tax penalty (see also Long, 1977). Thus, equilibrium asset prices reect the
aggregate preferences of investors and investors with the greatest wealth and least risk aversion have the greatest inuence on
price. All investors are marginal in the sense that a change in the taxation of any group potentially affects asset demand and prices.
Unlike the after-tax CAPM approach, the marginal investor approach suggests a potentially unstable equilibrium. Under the
marginal investor view, a change in the tax rates on an infra-marginal investor has no impact on securities prices as long as they
remain infra-marginal. Obvious clienteles should form according to tax status. However, if prices reect the tax preferences of
one group, individuals for example, then tax-based arbitrage opportunities arise for other groups such as tax-exempt entities,
and vice versa.150 Arbitrage opportunities may prevent stability in prices among close substitutes. Long (1978) provides an
interesting example of this based on the relative prices of differentially taxed shares issued by Citizens Utilities.
Recent research reiterates the importance of understanding the underlying equilibrium. Poterba (2007) states that it is
not clear which approach better describes market equilibrium. Guenther and Sansing (2006, 2010) point out that framing
the discussion in terms of some hypothetical marginal investor is unlikely to lead to useful insights because it is
inconsistent with the portfolio theory predictions of Brennan (1970) and Gordon and Bradford (1980). Greater
understanding of equilibrium will ultimately lead to more powerful empirical tests and more useful inferences. Future
work must be clear about which approach is assumed and why and whether the empirical evidence has anything to say
about the validity of the marginal versus after-tax CAPM approaches.151
5.1.7.2. Are dividend taxes priced differently across rms? The dividend premium (g) in Eq. (1) is in essence an exchange rate
between dividends and capital gains. Poterba (2007) notes that both the marginal investor and after-tax CAPM approaches
predict that the relative pricing of dividends and capital gains is the same across all stocks assuming no short-sale
restrictions. In other words, (td tcg)/(1 tcg) is an economy-wide constant that renders cross-sectional tests based on a
measure of rm-specic ownership irrelevant.
An open question is whether this exchange rate can actually be dened at the rm level. The issue applies to both the
marginal investor and after-tax CAPM approaches and is critical because recent studies that attempt to identify dividend
tax effects using institutional ownership data assume the dividend tax premium varies across rms (Ayers et al., 2002;
Dhaliwal et al., 2003). This latter way of thinking is based on the following reasoning: If the relative tax preference of
dividends to the price-setting investors varies across rms, then variation should be observed in the price of dividends
across rms partitioned by the investors tax status.
The use of institutional holdings as an identication strategy has produced the most consistent evidence that dividend
taxes affect security returns (Ayers et al., 2002; Dhaliwal et al., 2003, 2004c). But there are important theoretical and
empirical issues. First, it is unclear what conditions are required for the true dividend premium (g) to be larger in some rms
but not in others. Brennan (1970), which many empirical asset pricing studies rely on for theoretical motivation, derives the
result that the dividend premium is unaffected by the fraction of shares held by the tax-neutral investors in a given rm.
Second, one must also confront the empirical validity of using institutional ownership and its variants as the appropriate
proxy for the tax status of the price-setting investor. For example, institutional investors invest in different ways than
149
Under the irrelevance view, investors that are indifferent between dividends and capital gains are the marginal investors. Taxation of individuals
does not affect price.
150
Transactions costs have historically been used as a counter to this argument, i.e., that tax arbitrage opportunities are limited by transactions costs
born differently by the different investor groups. Based on trade-level evidence in Sias and Starks (1997), the fraction of institutional ownership is a proxy
for the likelihood that the marginal investor is indifferent between dividends and capital gains, at least over short horizons.
151
Theoretical discussions of the traditional and new views usually assume that the relevant investor is a taxable individual, but they are silent on
how tax heterogeneity among investors affects price. Both views are compatible with either a marginal investor or after-tax CAPM world while the
irrelevance view appears based on a marginal investor approach.
individuals for reasons that have nothing to do with taxes (see, e.g., Del Guercio, 1996; Gompers and Metrick, 2001; Bennett
et al., 2003; Blouin, 2009).152 This suggests the need to consider the determinants of institutional holdings before making
inferences about tax effects. Moreover, raw institutional ownership is at best an extremely noisy proxy for the tax status of a
rms investor base. Chetty and Saez (2005) argue that raw institutional ownership is clearly not a valid proxy for non-
taxable status because institutions include pension funds, insurance companies, corporations, endowments, and mutual
funds. Some are clearly tax-exempt, while others such as mutual funds ow through to fully taxable individuals. Because
there is substantial heterogeneity in the tax characteristics across institutions (and across non-institutions), drawing
meaningful inferences about taxes per se is very difcult. See Desai and Jin (forthcoming) for a good discussion of the
meaning of institutional ownership and an attempt to rene the measure in their analysis of tax clienteles.
5.1.7.3. Asset pricing in the open economy. As the securities markets become more global, identifying the effect of dividend
policy and dividend taxation based on U.S. tax rates alone becomes increasingly difcult. It may simplify our lives to
assume that the only investor is a taxable individual subject to U.S. tax rules. However, theory and evidence could become
nebulous when we consider the fact that not only is there substantial heterogeneity in the tax attributes of U.S. investors,
but that investors located in foreign jurisdictions face tax rules that may be completely different. This means that an
attempt to infer tax effects based on cross-sectional and time-series variation in U.S. taxes may mischaracterize the effect
of taxation if foreign investors and foreign investment matter (Carroll et al., 2003). On the other hand, data that crosses
jurisdictions provide more variation in tax rates. The challenge is how to leverage off that variation.
5.1.8. Thoughts for future research

In summary, several issues about the evolving literature on dividend taxes and asset prices appear relevant. First, we do
not have a good explanation for the inconsistent evidence on the association between dividends and expected returns;
more work is needed to understand why. Second, it is critical to recognize the role of portfolio theory in generating
predictions for equilibrium asset prices. The marginal investor approach appears inconsistent with portfolio theory, so
recognizing the implications for past and future research is vital. Third, it is highly likely that changes in tax rates are
correlated with changes in other tax and non-tax policies such as foreign trade policy, antitrust regulation, and so on. Thus,
the extent to which other concurrent or expected government policy changes could be driving the observed results is an
important issue, especially in event studies. Fourth, recent research documents extreme volatility in ex-dividend day
pricing that has nothing to do with taxes, raising questions about the ability of ex-day tests to identify tax effects. Fifth,
many asset pricing studies rely on cross-sectional tests using an ownership proxy to identify tax effects. More work, or at
least more clarity, on why this is or is not a valid approach will enhance our understanding on this topic, as will a
consideration of how endogeneity of ownership structure affects inferences. Finally, tests on the role of international
equity markets in asset pricing are important and can extend the literature in new ways.
5.2. Capital gains taxation: capitalization and lock-in
In the U.S., an individual that holds shares for a sufcient length of time before selling (currently 12 months) traditionally
faces a signicantly lower tax rate on realized gains relative to ordinary income.153 Even when statutory dividend and capital
gains tax rates are similar, the capital gains tax can be deferred to reduce the effective tax rate faced by the investor. An
individual can escape capital gains tax altogether by holding the shares until death and passing along stock to heirs or by
donating the shares to a charitable organization.154 However, capital losses are only deductible to the extent of capital gains
(although individuals can deduct an additional $3,000 of short-term capital losses per year against ordinary income).
There are two big questions driving the existing research on asset pricing consequences of capital gains taxation: First,
does a potential future capital gains tax reduce the price an investor is willing to pay for the asset? Second, does the capital
gains tax deter an investor from selling their shares and thus lead to reductions in supply and possible effects on price? There
are many questions within these broad queries, but all investigate the effect of the capital gains tax on economic behavior
and rm value. The literature appears to provide increasingly convincing evidence that capital gains taxes matter.155
5.2.1. Capital gains taxes and expected returns

One stream of literature suggests that the capital gains tax depresses stock prices because taxable investors demand
compensation for the anticipated tax due upon a future sale of the shares. This tax capitalization argument assumes that
152
Guenther and Sansing (2010) argue that if institutional investment is constrained more than non-institutional investment, then g can
theoretically be dened at the rm level.
153
Corporations, however, do not have a preferential rate for capital gains. Corporations instead have a tax preference for dividends. See Erickson and
Maydew (1998) for a paper exploiting this preference.
154
Under the estate tax rules, the asset basis is increased to the value of the asset upon death for appreciated assets (i.e., basis step-up) and no capital
gains tax is due on such appreciation should the heirs subsequently sell the asset. For estates large enough to be subject to the estate tax, the value upon
death is the value included in the estate, and this is then taxed at estate tax rates. A discussion of the estate tax is beyond the scope of this paper.
155
There are also studies that demonstrate cases in terms of investor trading behavior in which tax considerations are swamped by non-tax considerations
or behavioral aspects. For examples, see Seyhun and Skinner (1994), Shefrin and Statman (1985), Odean (1998), and Barber and Odean (2000, 2004).
price-setting investors are taxable and expect to realize capital gains in the future. The null is that capital gains taxes are
irrelevant; taxable investors have no effect on prices, can realize their capital gains through tax-advantaged vehicles, or
intend to hold the shares indenitely.
Most asset pricing studies using long-run returns tests focus on dividends, but implicit in the asset pricing regression is an
allowance for a potential capital gains tax. For example, Poterba and Summers (1985) model a simple expected return in a risk-
free security over a given period as R=(r/(1 tcg))+((td tcg)/(1 tcg))d. The rst term serves to gross-up the after-tax required
return for the expected capital gains tax (a function of the tax status of investors, tax rates, and holding period). The second
term is the compensation for receiving a dividend as discussed earlier. Most asset pricing tests focus on the implications of
dividend policy, and capital gains taxes are assumed to be irrelevant or to act only as a scalar. Because of the difculties
identifying capital gains tax effects using long-horizon returns, event studies offer a potentially more powerful setting.
Using event study methodology, Lang and Shackelford (2000) document signicant price increases following the
announcement of a cut in capital gains taxes in 1997. Moreover, they nd that the mean return for non-dividend-paying
stocks was 12.9% versus 6.1% for dividend-paying stocks, which is expected since rms that pay dividends leave less to be
taxed at the more favorable capital gains rates.156 Guenther and Willenborg (1999) nd that the initial public offering
prices for small rms were signicantly higher following a 1993 tax law change that reduced the capital gains tax on
shares of qualied small businesses.
5.2.2. Stock market realizations and the lock-in effect

The theory and empirical evidence on how capital gains taxes affect the decision to sell and whether that in turn affects
price is more developed.157 Given current and expected prices, the decision to sell is a function of what the investor paid
for the shares (their basis), the returns to alternative investments, the benets of portfolio rebalancing, the expectation of
discrete jumps in tax rates (e.g., crossing the long-term holding window, donating the security, death, etc.), and the
asymmetry in the treatment of gains and losses. Basic lock-in theory predicts that the probability of selling at current
prices is decreasing in the capital gains tax rate and increasing in basis.
Evidence on the role of capital gains taxes on selling behavior includes Feldstein et al. (1980) who nd that investors are
less likely to sell and realize gains when capital gains burdens, a function of tax rates and stock bases, are high. Ivkovich
et al. (2005) document strong evidence of lock-in effects in individual taxable trading accounts, but not in tax-deferred
accounts. Landsman and Shackelford (1995) document that the RJR Nabisco shareholders holding out for higher prices had
lower bases in their stock, consistent with lock-in predictions. Klein (2001) develops a long-horizon asset-pricing model
that allows for a lock-in effect in returns. He shows that rms with greater past appreciation have lower future expected
returns, which he interprets as investors having a disincentive to sell and incur capital gains liabilities, even when they
overweight that stock. Klein argues this is an explanation for the long-horizon return reversal phenomenon. However, the
ability of such an analysis to isolate capital gains effects is very much an open question.
Compounding the lock-in story is the fact that individuals must hold a capital asset long-term to qualify for
preferential capital gains tax treatment. A sharp drop in statutory rates could change investor behavior by accelerating
sales of stock with losses and delaying sales of stock with gains around the point holding period becomes long-term.
Shackelford and Verrecchia (2002) label this an intertemporal tax discontinuity and argue that it leads individuals who
are subject to short-term rates, but are overweighted in the rms stock, to sell less following a good news disclosure,
pushing prices above where they would have been in the absence of such an effect. Blouin et al. (2003) test these
predictions around earnings announcements. They nd that the taxes saved by holding the shares until the long-term rates
apply leads to higher announcement returns, which they argue supports the interpretation that tax discontinuities lead to
higher prices.158
The effect of capital gains taxes on expected returns works through the demand side and depresses prices whereas the
effect on selling behavior through lock-in enters on the supply side and increases prices. Few studies have considered their
joint effect on asset prices. The issue is important because, depending on the theory, capital gains taxes can both increase
and decrease asset prices. Dai et al. (2008) argue that a decrease in the capital gains tax rate shifts the demand curve up,
reecting the increase in prices buyers are willing to pay, and shifts the supply curve down, reecting the drop in the
reservation prices necessary to entice current owners to sell. The authors nd that following the announcement of the
1997 tax cut, but before its effective date, the tax capitalization effect dominates as in Lang and Shackelford (2000) and
does not appear to arise from a sellers strike.159 Importantly, they document that following the effective date of the
lower capital gains tax rates, stock price movements among rms with large past price appreciation and high individual
156
Blouin et al. (2009) extend the analysis of the 1997 tax cut to foreign rms with American Depository Receipts (ADRs) traded in U.S. markets. In
their setting, foreign rms have two sets of otherwise identical securities traded on two different markets, a nice control for risk. They compare the
reaction in the U.S. ADR market to the price reaction in the foreign home market, and they report evidence on tax effects through a difference in ADR
versus home country returns after the tax cut announcement, especially when arbitrage costs are high.
157
Important theoretical work by Constantinides (1983, 1984), Dammon and Spatt (1996), and Dammon et al. (2001) has advanced our understanding of
the effect of capital gains tax rules on optimal trading decisions. Poterba (1987) and Odean (1998) show that investors do realize substantial capital gains.
158
In a related study, Jin (2006) provides evidence that the documented price effects could result from more general lock-in effects.
159
Of issue to some extent is the magnitude of the tax effects and whether these are relatively consistent across studies or whether differences can be
explained. For example, Lang and Shackelford (2000) estimate effects that may appear implausibly high to be due from capital gains taxes alone. Future
research may focus more on the extent to which capital gains tax matters.
ownership were consistent with relief in the lock-in effect leading to increased supply and lower prices. The unique
features of the 1997 tax code allow Dai et al. to separate out the effects of tax capitalization and lock-in surrounding such
an act and contribute to the literature by explicitly explaining how the price effects can operate in both directions.160
6. Conclusions
Has tax research progressed since Shackelford and Shevlin (2001)? Undoubtedly, it has. However, there is much to be
done. Although the effects of taxes on real corporate decisions are at times difcult to document, they are important to
examine, especially given the focus on tax policy in federal stimulus efforts that compound growing budget decits.
Research in this area, especially using accounting researchers knowledge of nancial statements and institutional details,
will provide important contributions over the next era. In addition, further study of the tradeoff or interaction between tax
reporting and nancial reporting will be important to understand why some tax policies are more or less effective than
economists and policy makers might expect.
The literature on the informational role of the accounting for income taxes needs renement and clarication at this
stage. It is a relatively new area that has grown signicantly over the last 10 years and has been a contribution to the tax
and nancial accounting literatures (e.g., beyond documenting whether taxes matter as called for in Shackelford and
Shevlin, 2001). Given that many of the basic relations have been established, the quickly spreading literature will benet
from a more detailed examination of the sources of book-tax differences and their implications for nancial statement
analysis. Progress along the lines of distinguishing types and sources of book-tax differences (Ayers et al., 2010; Blaylock
et al., 2010) may be useful to further identify why many of the associations in the literature exist. In addition, a detailed
examination of when total book-tax differences have implications for nancial statement analysis relative to when
temporary book-tax differences hold information would be useful in understanding some of the documented results and as
a guide to future research efforts.
The relevance of tax avoidance research will increase as governments try to close the tax gap, increase compliance, and
collect more revenue. Meaningful extensions of the theoretical foundation of tax avoidance in a principalagent setting
should contribute to this growing literature. In addition, there is a need for creativity in empirical design to test these
theories and those regarding the interactions between the tax authority and corporate governance system as proposed by
Desai et al. (2007b). We predict this area will generate many interesting studies.
We provide a detailed discussion of the various empirical measures that have been used to measure tax avoidance.
We encourage researchers to think carefully about the measures given their research question and, just as importantly,
consider the inferences a reader should draw from the results given the measurement used. There are too many points on
this issue to reiterate here. But we cannot overemphasize how important it is to validly motivate and measure proxies for
tax avoidance and interpret results within the various limits of these measures (e.g., conforming versus non-conforming
tax avoidance and the use of temporary or permanent book-tax differences in the research design). We do not mean to stop
research on tax avoidance; rather, we intend to provide guidance for future work. Finally, more theory and evidence on the
causes and consequences of tax avoidance certainly has the potential to signicantly contribute to our understanding of
the role of taxes in the organization.
A signicant focus of our paper is on the role of taxes in business decisions. Investment and nancing policy,
organizational form, transfer pricing, mergers, and compensation decisions are all affected by taxes. The important point here
is that taxes are one factor that enters into managers costbenet tradeoff decisions, but it is often not the most important
factor. Understanding how important taxes are provides evidence useful for policy decisions. The bulk of the development of
these literatures has progressed in economics and nance. However, there is increasing reliance on nancial statements to
measure these effects as well as growing interest in the interaction between tax and accounting costs. Accounting researchers
have a lot to offer through their understanding of nancial reporting and managerial incentives in the reporting process.
In the long-standing literature on the effects of dividend and capital gains taxes on asset prices, there is a lot we do not
know. Recent studies suggest dividend taxes affect expected returns; however, the use of institutional ownership to parse
out these tax effects is subject to debate. Research that provides future guidance on the appropriate use and interpretation
of institutional ownership and other proxies for investor-level taxes in asset pricing studies will contribute to the
literature.
One issue that has been raised is that tax research in accounting needs more theory.161 This is likely true; almost every
applied eld could benet from a stronger theoretical foundation. Recent work on tax avoidance in a principalagent
160
In related studies, Reese (1998) examines samples of IPO holdings over different time periods, he nds that investors delay the sale of appreciated
assets to realize long-term capital gains and accelerate the sale of depreciated assets to lock-in short-term losses when short-term and long-term gains
are taxed differently. Poterba and Weisbenner (2001) show that capital gain tax rule changes induce investors to realize tax losses at year-end and that
this tax-loss trading contributes to turn-of-the year return patterns. Dammon et al. (2004) provide evidence consistent with the step-up at death and
resulting avoidance of the capital gains tax having real asset allocation effects. Despite the common theory that equity holdings should decrease with age,
these authors nd in their model that the optimal equity holding increases well into a investors lifetime due to the forgiveness of the capital gain at
death.
161
These calls are not asking for papers with math models that provide a counter example case in a specic setting but rather an overall theory or
framework to direct our research efforts.
setting (Crocker and Slemrod, 2005) and the role of the tax authority as an external monitor of management (Desai et al.,
2007b; Desai and Dharmapala, 2006) are excellent beginnings. Recent theoretical advances also help us understand the
nature of book-tax tradeoffs and the importance of nancial accounting incentives for tax policy (see Edgerton, 2009d;
Shackelford et al., 2009). Theories about how investor-level taxation affects asset prices (i.e., the new view, traditional
view, and irrelevance view) have led to a lot of confusion and disagreement, and little in the way of solutions or empirically
testable predictions. However, work by Brennan (1970), Gordon and Bradford (1980) and Guenther and Sansing (2010)
provide a more rigorous foundation for thinking about how taxes affect asset prices, portfolio allocations, and other
investment decisions when the simplistic marginal investor model falls short.
The tax area is exciting and big contributions can still be made that will advance our understanding of corporate and
individual decisions (both reporting and real), and perhaps more importantly, inform managerial decisions and tax
policy. As we write, the U.S. is (and other nations are) potentially approaching a major crossroads in tax policy. As
administrations propose changes to the tax rules, tax researchers in accounting should not sit on the sidelines. They should
position their research and inquiry in a manner that can inform debates and inject much needed data with regard to rm
behavior, real effects, and, to the extent possible, macroeconomic implications. We look forward to working on and reading
studies on these topics and more in coming years.
Appendix A. Accounting for Income Taxes (SFAS 109)
U.S. companies are required to compute income separately for nancial accounting and tax purposes. The two different
sets of rules result in two different income measures. As part of the computation of accounting income under GAAP,
companies are required to record an income tax expense.162 However, it is important to note that this is not the actual
expenditure for income tax liabilities. Under the current rules for the accounting of income taxes, the Financial Accounting
Standards Board (FASB) applies an accrual accounting-based balance sheet approach consistent with the rest of the
nancial statements. Thus, the tax expense on the income statement is not the actual cash tax paid but an accrual
accounting estimate of the taxes incurred on the nancial accounting earnings reported in a specic period (either due
now, to be paid in the future, or paid in the past). In the notes to the nancial statements, there is separate disclosure for
the income tax expense. This includes an estimate of what is actually due during the period the current tax expense as
well as an accrued expense (or benet) for the estimated taxes due in the future or paid in the past related to current year
transactions the deferred tax expense (or benet). However, the amount reported as current tax expense includes some
tax accruals and is affected by items accounted for through equity, and thus does not strictly equal the actual taxes paid on
the current years earnings.163
There are many differences between the computation of pre-tax accounting earnings and the computation of taxable
income. The differences are classied as temporary or permanent. Temporary differences are the most common. They
reect differences in the timing of recognition that will reverse in the future. For transactions that generate these
differences, the same amount is ultimately deducted for tax reporting as is expensed for nancial reporting (or the same
amount of income is includible under each system) but the deduction/expensing (revenue recognition) happens at
different points in time. Some common examples include depreciation, warranty and bad debt expense, and deferred
revenue. Because the accounting standards take a balance sheet approach, the objective in the accounting for these items is
to ensure that the correct liability (asset) is recorded on the accounting balance sheet for any liability (future benet) the
company has incurred to date. Thus, as an example, where tax deductions occur earlier than accounting expenses (e.g.,
depreciation), a deferred tax liability is recorded along with a deferred tax expense. So the expense incurred this period is
matched with the accounting earnings recorded this period and, more importantly from FASBs perspective, the correct
amount of deferred tax liability is recorded on the balance sheet.164
Other differences between the book and tax treatment are permanent and by denition do not reverse (e.g., municipal
bond interest is included in book income but not in taxable income).165 However, permanent differences are rare relative
162
Most of the principles of SFAS 109 are similar under International Financial Reporting Standards.
163
See Hanlon (2003) and McGill and Outslay (2004) for further details on problems with using current tax expense to approximate cash taxes paid.
However, some of the items they discuss have changed to some degree (e.g., the accounting for stock options and the accounting for the tax contingency
reserve).
164
Technically, the balance sheet approach is implemented as follows. Suppose a piece of equipment that cost $1,000 is depreciated in its rst year
$100 for accounting and $250 for tax purposes. Technically, FASB requires the computation of the deferred tax assets and liabilities, and the annual
changes in these generate the deferred tax expense. Thus, in the example, the book basis in the asset is $900 and the tax basis in the asset is $750. This
difference in asset bases is computed cumulatively over all years (here is just one year) as a $150 difference in bases. From this subtract the cumulative
bases difference at the beginning of the year, in this case zero, to yield the current year change of $150. Multiply the change in current year bases
difference by the tax rate, 35%, to yield a $52.50 deferred tax liability increase and a $52.50 deferred tax expense.
165
Technically, the term permanent difference is not used in SFAS 109. The concept of permanent differences under SFAS 109 is limited to events
recognized in the nancial statements that do not have tax consequences, such as tax-exempt interest. This type of permanent difference continues to
impact the calculation of current tax expense under SFAS 109.
to the frequency of temporary differences.166 These temporary book-tax differences and permanent book-tax differences
constitute the total difference between taxable income and pre-tax book income.
There are also accounting accruals related to taxes that do not exist for tax purposes and thus create additional
differences between taxable income and after-tax book income (but do not affect pre-tax income). For example, a valuation
allowance is a contra-asset account that must be established if a deferred tax asset (e.g., the future tax savings related to a
net operating loss carryforward) is more likely than not to not be realized. To accomplish this reduction in the deferred tax
asset, there is an increase in the reported tax expense. In another example, a tax contingency reserve (now renamed
unrecognized tax benets), must be established to accrue tax expense for potential future tax authority adjustments.
Furthermore, the consolidation rules for book purposes and tax purposes are different, often resulting in different
entities being consolidated for the annual report to the SEC than the tax return led with the tax authority.167 For example,
for tax purposes companies can elect to consolidate domestic subsidiaries owned 80% or more whereas for nancial
accounting, all subsidiaries greater than 50% owned are required to be consolidated.
There is an additional complexity with regard to foreign subsidiaries that is important in recent literature and thus
we provide a somewhat detailed discussion of the tax and accounting rules for foreign subsidiary earnings of a U.S.
multinational. Although the U.S. has a worldwide tax system, operating income from foreign subsidiaries of U.S.
multinationals is not included in the computation of taxable income in the U.S. This rule of not taxing these earnings
currently is called deferralthe taxation of the earnings is deferred until cash is actually (or effectively) repatriated to
the U.S. parent. To illustrate, assume a U.S. multinational earns $1,000 (for book and tax purposes) in a wholly owned
subsidiary in a foreign jurisdiction where the foreign tax rate is 10%. The company will include the $1,000 in the nancial
statements led with the SEC because the subsidiary is greater than 50% controlled and is thus consolidated for nancial
reporting purposes. The company will also include the $1,000 of income on its foreign tax return and pay $100 of tax to the
foreign jurisdiction. The U.S. consolidated company will, for nancial accounting purposes, record a foreign current tax
expense of $100 and a U.S. current tax expense of $0 on these earnings (since the earnings are not repatriated and no U.S.
taxation is assessed currently). Because the foreign earnings will be taxed in the U.S. when repatriated back to the U.S.,
a temporary difference is created and a deferred tax liability is recognized. However, APB 23 provides an exception to this
general rule of accrual accounting for income taxes. First, company management must make a decision: are the foreign
earnings permanently reinvested in the foreign jurisdiction, or does the company intend to repatriate the earnings? If the
company does not intend to bring the cash back to the U.S., they are to designate the earnings as permanently reinvested
for accounting purposes (under APB 23). As a result of this designation, a deferred tax liability (and deferred tax expense) is
not recorded for the amount of residual U.S. income taxes that will be owed upon repatriation (in this case, $250, the
incremental U.S. tax liability assuming a U.S. tax rate of 35%). Thus, the income tax expense recognized on those foreign
earnings is only the foreign taxes at the lower foreign tax rate. In such a case, the total tax expense for the company is
lower and after-tax income is higher, all else constant, than if the earnings were earned and taxed in the U.S.
There are also tax credits for items such as research and development expenditures and credits for foreign taxes paid
that reduce the tax owed but do not affect either taxable income or pre-tax accounting income. Because the credits reduce
the taxes owed, they reduce the current tax expense. A common way of estimating taxable income from nancial
statements is to gross-up the current tax expense by the statutory tax rate. The existence of tax credits is problematic in
this method because the credit is essentially converted in the estimation to a book-tax difference and leads to an
understatement of taxable income.168
The effective tax rate for accounting purposes (hereafter called the GAAP ETR) is dened as the worldwide total income
tax expense divided by worldwide pre-tax accounting earnings. Under U.S. GAAP, companies must reconcile the expected
tax expense (computed as accounting pre-tax earnings multiplied by the statutory tax rate) to the actual tax expense
(or the statutory tax rate to the effective tax rate).169
Some items in the nancial statements are reported net of tax, such as extraordinary items, discontinued operations,
and items in other comprehensive income. While these are taxes incurred, they are not part of the effective tax rate
166
At least we think this is the case. A descriptive-type examination of what the common permanent differences are and how common or uncommon
in both frequency and amount they are would be interesting. Specically, such a study would be useful in interpreting the literature that uses total book-
tax differences or ratios of taxable income to book income in nancial statement analysis studies and in thinking about research that claim permanent
differences are so important in tax avoidance strategies (e.g., just how frequent are these types of differences?). We discuss these studies and this gap in
the research further below.
167
Details are beyond the scope of this paper see Hanlon (2003). Also see Graham and Mills (2008) for a discussion of problems linking tax return
data to company nancial statement data.
168
Thus, a common sensitivity test when conducting studies using that approximation is to examine results after excluding rms with foreign
income or research and development expense. This helps mitigate any concern that measurement error in the computation of taxable income from the
rm having tax credits is driving the results.
169
The GAAP ETR is different from the marginal effective and marginal tax rates. The GAAP ETR is an average of two accounting numbers. It is not on
the margin and does not take into account the time value of money (i.e., deferred tax expense gets as much weight as the current tax expense). A current
effective tax rate can also be computed that is the current tax expense divided by pre-tax earnings. Also, each of these can be converted to U.S.-only or
foreign-only. However, if a residual U.S. tax is paid or accrued on foreign earnings, then the numerator (federal tax expense) will include those U.S. taxes
paid on the foreign earnings, but the denominator will include only the U.S. earnings if one is doing a U.S. only rate (U.S. tax divided by domestic pre-tax
book income), which will make the U.S. rate look higher than it is.
computation and do not factor into any of the measures of tax burden that are generally estimated from nancial
statements (because many rms do not disclose enough details to obtain the tax amount separately).
One important point of clarication is what does and what does not affect the GAAP ETR. A company whose investment
qualies for accelerated depreciation (or bonus depreciation provisions) will not have a lower GAAP ETR relative to
companies that do not take advantage of accelerated depreciation, assuming the level of investment is the same across the
rms. Accelerated depreciation creates a temporary difference that reduces the current tax expense (and taxes paid) but
increases the deferred tax expense by the same amount (assuming constant tax rates) and, thus, the GAAP ETR is
unchanged. Most temporary differences will operate in this same manner. While many teach this in intermediate
accounting classes, it often seems misunderstood in research (and referee reports). In addition, items that create no book-
tax difference at all (and are not subject to a rate differential and do not create a tax credit) do not decrease or increase the
GAAP ETR. A commonly misunderstood example is interest expense. A company that has debt and records interest expense
in the same amount for nancial accounting purposes and tax purposes will not have a lower GAAP ETR because of this
interest expense and deduction.170 The GAAP ETR is not affected by temporary differences, but is affected by permanent
differences, tax credits that directly reduce the taxes owed on a given income but do not change taxable income, items that
have a rate differential such as foreign earnings that are permanently reinvested, the incremental effects of state taxes, and
changes in the valuation allowance and many changes in the tax contingency reserve.
References
Ali, A., Hwang, L., 2000. Country-specic factors related to nancial reporting and the value relevance of accounting data. Journal of Accounting Research
38, 121.
Allee, C., Yohn, T., 2009. The demand for nancial statements in an unregulated environment: an examination of the production and use of nancial
statements by privately held small businesses. The Accounting Review 84, 125.
Allen, F., Michaely, R., 2003. Payout policy. In: Constantinides, G., Harris, M., Stulz, R. (Eds.), Handbook of the Economics of Finance, vol. 1. Elsevier
Science, Amsterdam, pp. 337429.
Allingham, M., Sandmo, A., 1972. Income tax evasion: a theoretical analysis. Journal of Public Economics 1, 323338.
Altshuler, R., Grubert, H., 2003. Taxes, repatriation strategies and multinational nancial policy. Journal of Public Economics 87, 73107.
Altshuler, R., Grubert, H., Newlon, T.S., 2000. Has U.S. investment abroad become more sensitive to tax rates? In: Hines Jr., J. (Ed.), International Taxation
and Multinational Activity. University of Chicago Press, Chicago, pp. 938.
Altshuler, R., Newlon, T.S., 1993. The effects of tax policy on the income repatriation patterns of U.S. multinational corporations. In: Giovannini, A.,
Hubbard, G., Slemrod, J. (Eds.), Studies in International Taxation. University of Chicago Press, Chicago, pp. 77115.
Altamuro, J., 2006. The determinants of synthetic lease nancing and the impact on the cost of future debt. Working Paper, Ohio State University.
Amir, E., Kirschenheiter, M., Willard, K., 1997. The valuation of deferred taxes. Contemporary Accounting Research 14, 597622.
Amir, E., Sougiannis, T., 1999. Analysts interpretation and investors valuation of tax carryforwards. Contemporary Accounting Research 16, 133.
Amromin, G., Harrison, P., Sharpe S., 2007. How did the 2003 dividend tax cut affect stock prices? Working Paper, Federal Reserve Bank of Chicago.
Armstrong, C., Blouin, J., Larcker, D., 2010. The incentives for tax planning. Working Paper, University of Pennsylvania.
Armstrong, C., Jagolinzer, A., Larcker, D., 2010. Chief executive ofcer equity incentives and accounting irregularities. Journal of Accounting Research 48,
225271.
Atwood, T., Drake, M., Myers, L., 2010. Book-tax conformity, earnings persistence and the association between earnings and cash ows. Journal of
Accounting and Economics 50, 111125.
Auerbach, A., 1979. Share valuation and corporate equity policy. Journal of Public Economics, 291305.
Auerbach, A., 2002. Taxation and corporate nancial policy. In: Auerbach, A., Feldstein, M. (Eds.), Handbook of Public Economics, vol. 1. Elsevier Science,
Amsterdam, pp. 12511292.
Auerbach, A., Hassett, K., 2003. On the marginal source of investment funds. Journal of Public Economics 87, 205232.
Auerbach, A., Hassett, K., 2007. The 2003 dividend tax cut and the value of the rm: an event study. In: Auerbach, A., Hines, J., Slemrod, J. (Eds.), Taxing
Corporate Income in the 21st Century. Cambridge University Press, New York.
Auerbach, A., Reishus, D., 1988. The effects of taxation on the merger decision. In: Corporate Takeovers: Causes and Consequences, NBER, pp. 157190.
Ayers, B., 1998. Deferred tax accounting under SFAS No. 109: an empirical investigation of its incremental value-relevance relative to APB No.11. The
Accounting Review 73, 195212.
Ayers, B., Cloyd, C., Robinson, J., 1996. Organizational form and taxes: an empirical analysis of small businesses. Journal of the American Taxation Society
18 (Suppl.), 4967.
Ayers, B., Cloyd, C., Robinson, J., 2001. The inuence of income taxes on the use of inside and outside debt by small businesses. National Tax Journal 54,
2756.
Ayers, B., Cloyd, C., Robinson, J., 2002. The effect of shareholder-level dividend taxes on stock prices: evidence from the Revenue Reconciliation Act of
1993. The Accounting Review 77, 933947.
Ayers, B., Jiang, X., Laplante, S., 2009. Taxable income as a performance measure: the effects of tax planning and earnings quality. Contemporary
Accounting Research 26, 1554.
Ayers, B., Laplante, S., McGuire, S., 2010. Credit rankings and taxes: the effect of book/tax differences on ratings changes. Contemporary Accounting
Research 27, 359402.
Ayers, B., Lefanowicz, C., Robinson, J., 2003. Shareholder taxes in acquisition premiums: the effect of capital gains taxation. Journal of Finance 58, 27832801.
Ayers, B., Lefanowicz, C., Robinson, J., 2004. The effect of shareholder-level capital gains taxes on acquisition structure. The Accounting Review 79, 859884.
Ayers, B., Lefanowicz, C., Robinson, J., 2000. The effects of goodwill tax deductions on the market for corporate acquisitions. The Journal of the American
Taxation Association 22, 3450.
Badertscher, B., Phillips, J., Pincus, M., Rego, S., 2009. Earnings management strategies: to conform or not to conform. The Accounting Review 84, 6398.
Baldenius, T., Melumad, N., Reichelstein, S., 2004. Integrating managerial and tax objectives in transfer pricing. The Accounting Review 79, 591615.
Bali, R., Hite, G., 1998. Ex dividend day stock price behavior: discreteness or tax-induced clienteles? Journal of Financial Economics 47, 127159
170
The effective tax rate might change because of mathematical proportions. For example, if the companys GAAP ETR is less than 35% for other
reasons and the rm then increases the level of debt, the GAAP effective rate will be reduced slightly because now the proportion of income taxed at a
lower rate is a greater percentage of the total (because there is less taxable income after the interest deduction). However, the interest expense itself does
not create a deduction that directly reduces the GAAP ETR; if reduced, it is a mathematical outcomeif the companys pre-interest GAAP ETR is 35%, it
will continue to be 35% after the interest expense (deduction).
Ball, R., Kothari, S.P., Robin, A., 2000. The effect of international institutional factors on properties of accounting earnings. Journal of Accounting and
Economics 29, 151.
Ball, R., Robin, A., Wu, J.S., 2003. Incentives versus standards: properties of accounting income in four East Asian countries and implications for acceptance
of IAS. Journal of Accounting and Economics 36, 235270.
Bank, S., Chefns, B., 2008. Tax and the separation of ownership and control. In: Schon, W. (Ed.), Tax and Corporate Governance, vol. 3. Springer, Berlin,
pp. 111160.
Bankman, J., 1994. The structure of Silicon Valley start-ups. UCLA Law Review 41, 17371768.
Bankman, J., Gilson, R., 1999. Why start-ups? Stanford Law Review 51, 289308
Bankman, J., 1999. The new market in U.S. corporate tax shelters. Tax Notes International 18, 26812706.
Bankman, J., 2004. The tax shelter problem. National Tax Journal 57, 925.
Barber, B., Odean, T., 2000. Trading is hazardous to your wealth: the common stock investment performance of individual investors. Journal of Finance 55,
773806.
Barber, B., Odean, T., 2004. Are individual investors tax savvy? Evidence from retail and discount brokerage accounts. Journal of Public Economics 88,
419442.
Barclay, M., 1987. Dividends, taxes, and common stock prices: the ex-dividend day behavior of common stock prices before the income tax. Journal of
Financial Economics 14, 3144.
Barclay, M., Heitzman, S., Smith, C., 2010. Debt and taxes: evidence from the real estate industry. Working Paper, University of Rochester.
Bargeron, L, Schlingemann, F., Stulz, R., Zutter, C., 2008. Why do private acquirers pay so little compared to public acquirers? Journal of Financial
Economics 89, 375390
Barnett, S., Sakellaris, P., 1998. Nonlinear response of rm investment to Q: testing a model of convex and non-convex adjustment costs. Journal of
Monetary Economics 42, 261288.
Bartelsman, E., Beetsma, R., 2003. Why pay more? Corporate tax avoidance through transfer pricing in OECD countries. Journal of Public Economics 87,
22252252.
Bartov, E., 1993. The timing of asset sales and earnings manipulations. The Accounting Review 68, 840855.
Bauman, C., Bauman, M., Halsey, R., 2001. Do rms use the deferred tax asset valuation allowance to manage earnings? Journal of the American Taxation
Association 23 (Suppl), 2748.
Beatty, A., Berger, P., Magliolo, J., 1995. Motives for forming research & development nancing organizations. Journal of Accounting and Economics 19,
411442.
Bennett, J., Sias, R., Starks, L., 2003. Greener pastures and the impact of dynamic institutional preferences. Review of Financial Studies 16, 12031238.
Bens, D., Nagar, V., Skinner, D., Wong, F., 2003. Employee stock options, EPS dilution, and stock repurchases. Journal of Accounting and Economics 36,
5190.
Berger, P., 1993. Explicit and implicit tax effects of the R&D tax credit. Journal of Accounting Research 31, 131171.
Betton, S., Eckbo, E., Thorburn, K., 2008. Corporate takeovers. In: Eckbo, E. (Ed.), Handbook of Corporate Finance, vol. 2. Elsevier, pp. 291430.
Bhattacharya, S., 1979. Imperfect information, dividend policy, and the bird in the hand fallacy. Bell Journal of Economics 10, 259270.
Biddle, G., Hilary, G., 2006. Accounting quality and rm-level capital investment. The Accounting Review 81, 963982.
Biddle, G., Hilary, G., Verdi, R., 2009. How does nancial reporting quality relate to investment efciency? Journal of Accounting and Economics 48,
112131
Black, F., Scholes, M., 1974. The effects of dividend yield and dividend policy on common stock prices and returns. Journal of Financial Economics 1, 122.
Blaylock, B., Shevlin, T., Wilson, R., 2010. Tax avoidance, large positive book-tax differences, and earnings persistence. Working Paper, University of
Washington.
Blouin, J., 2009. Discussion of dividend tax clienteles: evidence from tax law changes. The Journal of the American Taxation Association 31, 2328.
Blouin, J., Carter, M., 2010. Restricted stock and Section 83(b) elections. Working Paper, University of Pennsylvania.
Blouin, J., Core, J., Guay, W., Have the benets of debt been overstated? Journal of Financial Economics, forthcoming.
Blouin, J., Gleason, C., Mills, L., Sikes, S., 2010. Pre-empting disclosure? Firms decisions prior to FIN 48. The Accounting Review 85, 791815.
Blouin, J., Hail, L., Yetman, M., 2009. Capital gains taxes, pricing spreads and arbitrage: evidence from cross-listed rms in the U.S. The Accounting Review
84, 13211361.
Blouin, J., Krull, L., 2009. Bringing it home: a study of the incentives surrounding the repatriation of foreign earnings under the American Jobs Creation Act
of 2004. Journal of Accounting Research 47, 10271059.
Blouin, J., Raedy, J., Shackelford, D., 2003. Capital gains taxes and equity trading: empirical evidence. Journal of Accounting Research 41, 611651.
Bond, S., Cummins, J., 2000. The stock market and investment in the new economy. Some tangible facts and intangible ctions 31, 61124.
Bradford, D.E., 1981. The incidence and allocation effects of a tax on corporate distributions. Journal of Public Economics 15, 122.
Bradley, M., Jarrell, G., Kim, E., 1984. On the existence of an optimal capital structure: theory and evidence. Journal of Finance 39, 857878.
Brennan, M., 1970. Taxes, market valuation and corporate nancial policy. National Tax Journal 23, 417427.
Brennan, T., 2008. Cash-ow and market response to repatriation. Working Paper, Northwestern University School of Law.
Brown, J., Krull, L., 2008. Stock options, R&D, and the R&D tax credit. The Accounting Review 83, 705734.
Burgstahler, D., Dichev, I., 1997. Earnings management to avoid earnings decreases and losses. Journal of Accounting and Economics 24, 99126.
Burgstahler, D., Hail, L., Leuz, C., 2006. The importance of reporting incentives: Earnings management in European private and public rms. The
Bushee, B., 1998. The inuence of institutional investors on myopic R&D investment behavior. The Accounting Review 73, 305333.
Bushman, R., Smith, A., Zhang, F., 2008. Investment-cash ow sensitivities are really investment-investment sensitivities. Working Paper, University of
North Carolina.
Capithorne, L., 1971. International corporate transfer prices and government policy. Canadian Journal of Economics 4, 324341.
Carroll, R., Hassett, K., MacKie, J., 2003. The effect of dividend tax relief on investment incentives. National Tax Journal 56, 629651.
Cassar, G., 2009. Financial statement and projection preparation in start-up ventures. The Accounting Review 84, 2751.
Chan, H., Mo, P., Lo, A., 2006. Managerial autonomy and tax compliance: an empirical study on international transfer pricing. The Journal of the American
Chen, K.-P., Chu, C., 2005. Internal control vs. external manipulation: A model of corporate income tax evasion. RAND Journal of Economics 36, 151164.
Chen, M., Gupta, S., 2009. The incentive effects of R&D tax credits: an empirical examination in an emerging economy. Working Paper, National Chengchi
University and Michigan State University.
Chen, N., Grundy, B., Stambaugh, R., 1990. Changing risk, changing risk premiums, and the dividend yield effects. Journal of Business 63, S51S70.
Chen, S., Chen, X., Cheng, Q., Shevlin, T., 2010. Are family rms more tax aggressive than non-family rms? Journal of Financial Economics 95, 4161
Chetty, R., Rosenberg, J., Saez, E., 2007. The effects of taxes on market responses to dividend announcements and payments: What can we learn from the
2003 dividend tax cut? In: Auerbach, A., Hines, J., Slemrod, J. (Eds.), Taxing Corporate Income in the 21st Century. Cambridge University Press,
New York.
Chetty, R., Saez, E., 2005. Dividend taxes and corporate behavior: evidence from the 2003 Dividend Tax Cut. The Quarterly Journal of Economics 120,
791833.
Claus, J., Thomas, J., 2001. Equity premia as low as three percent? Evidence from analysts earnings forecasts for domestic and international stock markets.
Journal of Finance 56, 16291666.
Clausing, K., 2003. Tax motivated transfer pricing and US intrarm trade prices. Journal of Public Economics 87, 22072223.
Cloyd, C.B., Mills, L., Weaver, C., 2003. Firm valuation effects of the expatriation of U.S. corporations to tax-haven countries. Journal of the American
Taxation Association 25 (Suppl.), 87109.
Cloyd, B., Pratt, J., Stock, T., 1996. The use of nancial accounting choice to support aggressive tax positions: public and private rms. Journal of
Cohen, D.S., Cummins, J.G., 2006. A retrospective evaluation of the effects of temporary partial expensing. FEDS Working Paper Series.
Collins, J., Kemsley, D., 2000. Capital gains and dividend taxes in rm valuation: evidence of triple taxation. The Accounting Review 75, 405427.
Constantinides, G.M., 1983. Capital market equilibrium with personal tax. Econometrica 51, 611636.
Constantinides, G.M., 1984. Optimal stock trading with personal taxes. Journal of Financial Economics 13, 6589.
Cook, K., Huston, R., Omer, T., 2008. Earnings management through effective tax rates: the effects of tax planning investment and the Sarbanes-Oxley Act
of 2002. Contemporary Accounting Research 25, 447471.
Core, J., Guay, W., 1999. The use of equity grants to manage optimal equity incentive levels. Journal of Accounting and Economics 28, 151184.
Core, J., Guay, W., 2001. Stock option plans for non-executive employees. Journal of Financial Economics 61, 253287.
Crocker, K., Slemrod, J., 2005. Corporate tax evasion with agency costs. Journal of Public Economics 89, 15931610.
Cullen, J.B., Gordon, R., 2007. Taxes and entrepreneurial risk-taking: theory and evidence in the U.S. Journal of Public Economics 91, 14791505.
Cumming, D., 2008. Contracts and exits in venture capital nance. Review of Financial Studies 21, 19471982.
Cummins, J., Harris, T., Hassett, K., 1995. Accounting standards, information ow, and rm investment behavior. In: Feldstein, M., Hines, J., Hubbard, R.
(Eds.), The Effects of Taxation on Multinational Corporations. NBER, Cambridge, MA.
Cummins, J., Hassett, K., Oliner, S., 2006. Investment behavior, observable expectations, and internal funds. American Economic Review 96, 796810.
Dai, Z., Maydew, E., Shackelford, D., Zhang, H., 2008. Capital gains taxes and asset prices: capitalization or lock-in? Journal of Finance 63, 709742
Dammon, R., Spatt, C., 1996. The optimal trading and pricing of securities with asymmetric capital gains taxes and transaction costs. Review of Financial
Studies 9, 921952.
Dammon, R., Spatt, C., Zhang, H., 2001. Optimal consumption and investment with capital gains taxes. Review of Financial Studies 14, 583616.
Dammon, R., Spatt, C., Zhang, H., 2004. Optimal asset location and allocation with taxable and tax-deferred investing. Journal of Finance 59,
9991038.
Damodaran, A., John, K., Liu, C., 1997. The determinants of organization form changes: evidence and implications from real estate. Journal of Financial
Economics 45, 169192.
Davila, A., Foster, G., 2005. Management accounting systems adoption decisions: evidence and performance implications from startup companies. The
DeAngelo, H., Masulis, R., 1980. Optimal capital structure under corporate and personal taxation. Journal of Financial Economics 8, 329.
DeAngelo, H., DeAngelo, L., 2006. The irrelevance of the MM dividend irrelevance theorem. Journal of Financial Economics, 293315.
DeAngelo, H., DeAngelo, L., Skinner, D., 2008. Corporate payout policy. In: Constantinides, G., Allen, F., Lo, A., Stultz, R. (Eds.), Foundation and Trends in
Finance, vol. 3, pp. 95287.
Dechow, P., Ge, W., Schrand, C., 2010. Understanding earnings quality: a review of the proxies, their determinants and their consequences. Journal of
Accounting and Economics, this issue, doi:10.1016/j.jacceco.2010.09.001.
Del Guercio, D., 1996. The distorting effect of the prudent man law on institutional equity investments. Journal of Financial Economics 40, 3162.
DeMooij, R., Ederveen, S., 2003. Taxation and foreign direct investment: a synthesis of empirical research. International Tax and Public Finance 10,
673693.
Desai, M., 2003. The divergence between book income and tax income. In: Poterba, J. (Ed.), Tax Policy and the Economy, 17. MIT Press, Cambridge, pp. 169206.
Desai, M., 2005. The degradation of reported corporate prots. Journal of Economic Perspectives 19, 171192.
Desai, M., 2006. Reform alternatives for the corporate tax. Testimony before the U.S. House of Representatives Committee on Ways and Means,
Subcommittee on Select Revenue Measures, Corporate Tax Reform public hearing, Washington, DC (May 9).
Desai, M., 2009. The decentering of the global rm. World Economy 32, 12711290.
Desai, M., Dharmapala, D., 2006. Corporate tax avoidance and high-powered incentives. Journal of Financial Economics 79, 145179.
Desai, M., Dharmapala, D., 2008. Tax and corporate governance: an economic approach. In: Schon, W. (Ed.), Tax and Corporate Governance. Springer-
Verlag, Berlin, pp. 1330.
Desai, M., Dharmapala, D., Taxes, dividends, international portfolio choice. Review of Economics and Statistics, forthcoming.
Desai, M., Dharmapala, D., 2009. Corporate tax avoidance and rm value. Review of Economics and Statistics 91, 537546.
Desai, M., Dharmapala, D., Fung, W., 2007a. Taxation and the evolution of aggregate corporate ownership concentration. In: Auerbach, A., Hines, J.,
Slemrod, J. (Eds.), Taxing Corporate Income in the 21st Century. Cambridge University Press, New York, pp. 345383.
Desai, M., Dyck, I., Zingales, L., 2007b. Theft and taxes. Journal of Financial Economics 84, 591623.
Desai, M., Foley, C.F., Hines, J., 2001. Repatriation taxes and dividend distortions. National Tax Journal 54, 829851.
Desai, M., Foley, C.F., Hines, J., 2005a. Foreign direct investment and the domestic capital stock. American Economic Review 95, 3338.
Desai, M., Foley, C.F., Hines, J., 2005b. Foreign direct and domestic economic activity. NBER Working Paper No. 11717, October.
Desai, M., Foley, C.F., Hines, J., 2008. Capital structure with risky foreign investment. Journal of Financial Economics 88, 534553.
Desai, M., Goolsbee, A., 2004. Investment overhang and tax policy. Brookings Papers on Economic Activity 2, 285338.
Desai, M., Hines, J., 2002. Expectations and expatriations: tracing the causes and consequences of corporate inversions. National Tax Journal 55, 409440.
Desai, M., Hines, J., 2004. Old rules and new realities: corporate tax policy in a global setting. National Tax Journal 57, 937960.
Desai, M., Jin, L.. Institutional tax clienteles and payout policy. Journal of Financial Economics, forthcoming.
Dhaliwal, D., Erickson, M., Frank, M., Banyi, M., 2003. Are shareholder dividend taxes on corporate retained earnings impounded in equity prices?
Additional evidence and analysis. Journal of Accounting and Economics 35, 179200.
Dhaliwal, D., Erickson, M., Heitzman, S., 2004a. The effect of seller taxes on acquisition prices: evidence from purchases of taxable and tax-exempt
hospitals. Journal of the American Taxation Association 26, 123.
Dhaliwal, D., Erickson, M., Heitzman, S., 2009. Taxes and the backdating of stock option exercise dates. Journal of Accounting and Economics 47, 2749.
Dhaliwal, D., Gleason, C., Mills, L., 2004b. Last-chance earnings management: using the tax expense to meet analysts forecasts. Contemporary Accounting
Dhaliwal, D., Heitzman, S., Li, O., 2006. Taxes, leverage, and the cost of equity capital. Journal of Accounting Research 44, 691723.
Dhaliwal, D., Krull, L., Li, O., 2007. Did the 2003 Tax Act reduce the cost of equity capital? Journal of Accounting and Economics 43, 121150
Dhaliwal, D., Krull, L., Li, O., Moser, W., 2004c. Dividend taxes and implied cost of equity capital. Journal of Accounting Research 43, 675708.
Dhaliwal, D., Li, O., 2006. Investor tax heterogeneity and ex-dividend day trading volume. Journal of Finance 61, 463490.
Dhaliwal, D., Li, O., Trezevant, R., 2003. Is a dividend tax penalty incorporated into the return on a rms common stock? Journal of Accounting and
Economics 35, 155178
Dhaliwal, D., Trezevant, R., Wang, S., 1992. Taxes, investment related tax shields and capital structure. Journal of the American Taxation Association 14,
121.
Dharmapala, D., Foley, C.F., Forbes, K., 2010 Watch what I do, not what I say: the unintended consequences of the Homeland Investment Act. Journal of
Finance, forthcoming.
Dharmapala, D., Hines, J., 2006. Which countries become tax havens? NBER Working Paper #12802.
Durney, A., Li, T., Magnan, M., 2009. The governance of offshore rms: implications for nancial reporting. Working Paper Concordia University.
Dyck, I., Zingales, L., 2004. Private benets of control: an international comparison. Journal of Finance 59, 537600.
Dyreng, S., Hanlon, M., Maydew, E., 2008. Long-run corporate tax avoidance. The Accounting Review 83, 6182.
Dyreng, S., Hanlon, M., Maydew, E., 2010. The effects of executives on corporate tax avoidance. The Accounting Review 85, 11631189.
Dyreng, S., Lindsey, B., 2009. Using nancial accounting data to examine the effect of foreign operations located in tax havens and other countries on US
multinational rms tax rates. Journal of Accounting Research 47, 12831316.
Eades, K., Hess, P., Kim, E., 1994. Time-series variation in dividend pricing. Journal of Finance 49, 16171638.
Edgerton, J., 2009a. Effects of the 2003 dividend tax cut: evidence from real estate investment trusts. Working Paper, Federal Reserve Board of Governors.
Edgerton, J., 2009b. Investment incentives and corporate tax asymmetries. Working Paper, Federal Reserve Board of Governors.
Edgerton, J., 2009c. Taxes and used equipment: evidence from construction machinery. Working Paper, Federal Reserve Board of Governors.
Edgerton, J., 2009d. Investment, accounting, and the salience of the corporate income tax. Working Paper, Federal Reserve Board of Governors.
Elton, E., Gruber, M., 1970. Marginal stockholders tax rates and the clientele effect. Review of Economics and Statistics 52, 6874.
Engel, E., Erickson, M., Maydew, E., 1999. Debt-equity hybrid securities. Journal of Accounting Research 37, 249274.
Erickson, M., 1998. The effect of taxes on the structure of corporate acquisitions. Journal of Accounting Research 36, 279298.
Erickson, M., Hanlon, M., Maydew, E., 2004. How much will rms pay for earnings that do not exist? Evidence of taxes paid on allegedly fraudulent
earnings. The Accounting Review 79, 387408.
Erickson, M., Heitzman, S., 2010. NOL Poison Pills: Selectica v. Versata et al. Tax Notes (June).
Erickson, M., Maydew, E., 1998. Implicit taxes in high dividend yield stocks. The Accounting Review 73, 435458.
Erickson, M., Wang, S., 2000. The effect of transaction structure on price: evidence from subsidiary sales. Journal of Accounting and Economics 30, 5997.
Erickson, M., Wang, S., 2007. Tax benets as a source of merger premiums in acquisitions of private corporations. The Accounting Review 82, 359387.
Erickson, T., Whited, T., 2000. Measurement error and the relationship between investment and q. Journal of Political Economy 108, 10271057.
Fama, E., French, K., 1998. Taxes, nancing decisions, and rm value. Journal of Finance 53, 819843.
Fazzari, S., Hubbard, R.G., Petersen, B., 1988. Financing constraints and corporate investment. Brookings Papers on Economic Activity 19, 141206.
Faulkender, M., Petersen, M., 2009. Investment and capital constraints: repatriations under the American Jobs Creation Act. Working Paper, Northwestern
University.
Feldstein, M., Slemrod, J., Yitzhaki, S., 1980. The effects of taxation on the selling of corporate stock and the realization of capital gains. Quarterly Journal of
Foley, C., Hartzell, J., Titman, S., Twite, G., 2007. Why do rms hold so much cash? A tax-based explanation. Journal of Financial Economics 86, 579607.
Frank, M., Lynch, L., Rego, S., 2009. Tax reporting aggressiveness and its relation to aggressive nancial reporting. The Accounting Review 84, 467496.
Frank, M., Rego, S., 2006. Do managers use the valuation allowance account to manage earnings around certain earnings targets? Journal of the American
Taxation Association 28, 4365
Frank, M., Goyal, V., 2009. Tradeoff and pecking order theories of debt. In: Espen, Eckbo (Ed.), Handbook of Empirical Corporate Finance. Elsevier Science.
Frank, M., Jagannathan, R., 1998. Why do stock prices drop by less than the value of the dividend? Evidence from a country without taxes. Journal of
Frischmann, P., Shevlin, T., Wilson, R., 2008. Economic consequences of increasing the conformity in accounting for uncertain tax benets. Journal of
Gebhardt, W., Lee, C., Swaminathan, B., 2001. Toward an implied cost of capital. Journal of Accounting Research 39, 135176.
Gentry, W., 1994. Taxes, nancial decisions and organizational form: evidence from publicly traded partnerships. Journal of Public Economics 53,
223244.
Gentry, W., 2007. The future of tax research: a mostly economics perspective. Journal of the American Taxation Association 29, 95106.
Gentry, W., Hubbard, G., 2004. The effects of progressive income taxation on job turnover. Journal of Public Economics 88, 23012322.
Gentry, W., Kemsley, D., Mayer, C., 2003. Dividend taxes and share prices: evidence from real estate investment trusts. Journal of Finance 58, 261282.
Gilson, R., Schizer, D., 2003. Venture capital structure: a tax explanation for convertible preferred stock. Harvard Law Review 116, 875916.
Gleason, C., Mills, L., 2002. Materiality and contingent tax liability reporting. The Accounting Review 77, 317342.
Gleason, C., Mills, L., 2008. Evidence of differing market responses to beating analysts targets through tax expense decreases. Review of Accounting
Studies 13, 295318.
Gompers, P., Metrick, A., 2001. Institutional investors and equity prices. Quarterly Journal of Economics 116, 229259.
Goolsbee, A., 1998a. Investment tax incentives, prices and the supply of capital goods. Quarterly Journal of Economics 113 (1), 121148.
Goolsbee, A., 1998b. Taxes, organizational form, and the deadweight loss of the corporate income tax. Journal of Public Economics 69, 143152.
Goolsbee, A., 2000. Taxes, high-income executives, and the perils of revenue estimation in the new economy. American Economic Review 90, 271275.
Goolsbee, A., Maydew, E., 2002. Taxes and organizational form: the case of REIT spin-offs. National Tax Journal 55, 441456.
Goncharov, I., Werner, J.R., 2009. Reassessing the role of book-tax conformity. Working Paper, University of Amsterdam.
Gordon, M., 1963. Optimum investing and nancing policy. Journal of Finance 18, 264272.
Gordon, R., Bradford, D., 1980. Taxation and the stock market valuation of capital gains and dividends. Journal of Public Economics 14, 109136.
Gordon, R.H., MacKie-Mason, J., 1994. Tax distortions to the choice of organizational form. Journal of Public Economics 55, 279306.
Graetz, M., 2005. A fair and balanced tax system for the Twenty-rst century.
Graham, J., 1996a. Debt and the marginal tax rate. Journal of Financial Economics 41, 4174.
Graham, J., 1996b. Proxies for the corporate marginal tax rate. Journal of Financial Economics 42, 187221.
Graham, J., 1999. Do personal taxes affect corporate nancing decisions? Journal of Public Economics 73, 147185
Graham, J., 2000. How big are the tax benets of debt? Journal of Finance 55, 19011941
Graham, J., 2008. Taxes and corporate nance. In: Eckbo, E. (Ed.), Handbook of Corporate Finance; Empirical Corporate Finance Amsterdam. Elsevier
Science.
Graham, J., Hanlon, M., Shevlin, T., 2010a. Real Effects of Accounting Rules: Evidence from Multinational Firms Investment Location and Prot
Repatriation Decisions. Forthcoming, Duke University.
Graham, J., Hanlon, M., Shevlin, T., 2010b. Barriers to mobility: the lockout effect of U.S. taxation of worldwide corporate prots. National Tax Journal,
forthcoming.
Graham, J., Harvey, C., Rajgopal, S., 2005. The economic implications of corporate nancial reporting. Journal of Accounting and Economics 40, 373.
Graham, J., Kim, H., 2010. The effects of the length of the tax-loss carryback period on tax receipts and corporate marginal tax rates. Working Paper, Duke
University.
Graham, J., Lang, M., Shackelford, D., 2004. Employee stock options, corporate taxes and debt policy. Journal of Finance 59, 15851618.
Graham, J., Lemmon, M., Schallheim, J., 1998. Debt, leases, taxes, and the endogeneity of corporate tax status. Journal of Finance 53, 131162.
Graham, J., Smith, C., 1999. Tax incentives to hedge. Journal of Finance 54, 22412262.
Graham, J., Mills, L., 2008. Using tax return data to simulate corporate marginal tax rates. Journal of Accounting and Economics 46, 366388.
Graham, J., Raedy, J., Shackelford, D., 2009. Research in accounting for income taxes. Working Paper, Duke University.
Graham, J., Tucker, A., 2006. Tax shelters and corporate debt policy. Journal of Financial Economics 81, 563594.
Grassley, Sen, Charles E., 2002. US Senator Grassley Calls for Review of Corporate Return Disclosure Requirements. Tax Notes Today, October 11.
Green, J., Bensinger, G., 2007. GM reports $39 Billion loss on deferred tax charge. Bloomberg.com, November 7.
Green, R., Hollield, B., 2003. The personal tax advantages of equity. Journal of Financial Economics 67, 175216.
Grubert, H., Mutti, J., 1991. Taxes, tariffs and transfer pricing in multinational corporate decision making. Review of Economics and Statistics 73, 285293.
Grubert, H., Mutti, J., 1995. Taxing multinationals in a world with portfolio ows and R&D: is capital export neutrality obsolete? International Tax and
Public Finance 2, 439457
Grubert, H., Mutti, J., 2000. Do taxes inuence where U.S. corporations invest? National Tax Journal 53, 825883
Guedhami, O., Pittman, J., 2008. The importance of IRS monitoring to debt pricing in private rms. Journal of Financial Economics 90, 3858.
Guenther, D., 1992. Taxes and organizational form: a comparison of corporations and master limited partnerships. Accounting Review 67, 1745.
Guenther, D., 1994. Earnings management in response to corporate tax rate changes: evidence from the 1986 tax reform act. The Accounting Review 69,
230243.
Guenther, D., Maydew, E., Nutter, S., 1997. Financial reporting, tax costs, and booktax conformity. Journal of Accounting and Economics 23, 225248.
Guenther, D., Sansing, R., 2006. Fundamentals of shareholder tax capitalization. Journal of Accounting and Economics 42, 371384.
Guenther, D., Sansing, R., 2010. The effect of tax-exempt investors and risk on stock ownership and expected returns. The Accounting Review 85, 849875.
Guenther, D., Willenborg, M., 1999. Capital gains tax rates and the cost of capital for small business: evidence from the IPO market. Journal of Financial
Guenther, D., Young, D., 2000. The association between nancial accounting measures and real economic activity: a multinational study. Journal of
Gupta, S., Hwang, Y., Schmidt. A., 2006. An analysis of the availability and incentive effects of the R&D tax credit after the Omnibus Budget Reconciliation
Act of 1989. Working Paper, Michigan State University.
Gupta, S., Laux, R., 2008. Do rms use tax cushion reversals to meet earnings targets? Working Paper, Michigan State University.
Gupta, S., Mills, L., 2002. Multistate tax planning: benets of multiple jurisdictions. Journal of Accounting and Economics 33, 117139.
Gupta, S., Newberry, K., 1997. Determinants of the variability in corporate effective tax rates: evidence from longitudinal study. Journal of Accounting and
Public Policy, 134.
Hall, R., Jorgenson., D., 1967. Tax policy and investment behavior. American Economic Review 57, 391414.
Halperin, R., Srinidhi, B., 1987. The effects of U.S. income tax regulations transfer pricing rules on allocative efciency. The Accounting Review 62,
686706.
Hanlon, M., 2003. What can we infer about a rms taxable income from its nancial statements? National Tax Journal 56, 831863
Hanlon, M., 2005. The persistence and pricing of earnings, accruals, and cash ows when rms have large book-tax differences. The Accounting Review 80,
137166.
Hanlon, M., Krishnan, G., Mills, L., 2009. Do auditors use the information reected in book-tax differences? Working Paper, MIT.
Hanlon, M., Laplante, S., Shevlin, T., 2005. Evidence for the possible information loss of conforming book income and taxable income. Journal of Law and
Hanlon, M., Maydew, E., 2009. Book-tax conformity: implications for multinational rms. National Tax Journal 62, 127153.
Hanlon, M., Mills, L., Slemrod, J., 2007. An empirical examination of corporate tax noncompliance. In: Auerbach, A., Hines, J., Slemrod, J. (Eds.), Taxing
Corporate Income in the 21st Century. Cambridge University Press, New York, pp. 171210.
Hanlon, M., Myers, J., Shevlin, T., 2003. Dividend taxes and rm valuation: a re-examination. Journal of Accounting and Economics 35, 119153.
Hanlon, M., Shevlin, T., 2002. Accounting for tax benets of employee stock options and implications for research. Accounting Horizons 16, 116.
Hanlon, M., Shevlin, T., 2005. Book-tax conformity for corporate income: an introduction to the issues. Tax Policy and the Economy 19, 101134 (National
Bureau of Economic Research).
Hanlon, M., Slemrod, J., 2009. What does tax aggressiveness signal? Evidence from stock price reactions to news about tax shelter involvement. Journal of
Public Economics 93, 126141.
Harford, J., 2005. What drives merger waves? Journal of Financial Economics 77, 529560
Harris, D., 1993. The impact of U.S. tax law revision on multinational corporations capital location and income-shifting decisions. Journal of Accounting
Harris, D., Sansing, R., 1998. Distortions caused by the use of arms-length transfer prices. The Journal of the American Taxation Association, Spring, 4050.
Harris, T., Kemsley, D., 1999. Dividend taxation in rm valuation: new evidence. Journal of Accounting Research 37, 275291.
Harris, T., Lang, M, Moller, H.P., 1994. The value relevance of German accounting measures: an empirical analysis. Journal of Accounting Research 32,
187209.
Hartman, D., 1984. Tax policy and foreign direct investment in the United States. National Tax Journal 37, 475488.
Hartman, D., 1985. Tax policy and foreign direct investment. Journal of Public Economics 26, 107121.
Hassett, K., Hubbard, R.G., 1998. Are investment incentives blunted by changes in prices of capital goods? International Finance 1, 103125
Hassett, K., Hubbard, G., 2002. Tax policy and business investment. In: Auerbach, A., Feldstein, M. (Eds.), Handbook of Public Economics, vol. 3. Elsevier
Science, Amsterdam, pp. 12941343.
Hassett, K., Newmark, K., 2008. Taxation and business behavior: a review of the recent literature. In: Diamond, J., Zodrow, G. (Eds.), Fundamental Tax
Reform: Issues, Choices and Implications. MIT Press, Cambridge, pp. 191214.
Hayashi, F., 1982. Tobins marginal q and average q: a neoclassical interpretation. Econometrica 50, 213224.
Hayn, C., 1989. Tax attributes as determinants of shareholder gains in corporate acquisitions. Journal of Financial Economics 23, 121153.
Healy, P., Whalen, J., 1999. A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13, 365383.
Heath, C., Huddart, S., Lang, M., 1999. Psychological factors and stock option exercise. Quarterly Journal of Economics 114, 601627.
Hellmann, T., 2006. IPOs, acquisitions, and the use of convertible securities in venture capital. Journal of Financial Economics 81, 649679.
Hellmann, T., Puri, M., 2002. Venture capital and the professionalization of startup rms: empirical evidence. Journal of Finance 57, 169197.
Hennessy, C., Whited, T., 2007. How costly is external nancing? Evidence from a structural estimation. Journal of Finance 62, 17051745.
Heron, R., Lie, E., 2007. Does backdating explain the stock price pattern around executive stock option grants? Journal of Financial Economics 83, 271295
Hines, J., 1997. Tax policy and the activities of multinational corporations. In: Auerbach, A. (Ed.), Fiscal Policy: Lessons from Economic Research. MIT
Press, Cambridge, pp. 401445.
Hines, J., 1998. Investment ramications of distortionary tax subsidies. Working Paper, University of Michigan and NBER.
Hines, J., 1999. Lessons from behavioral responses to international taxation. National Tax Journal 52, 305322.
Hines, J., 2005. Do tax havens ourish? In: James, M.Poterba (Ed.), Tax Policy and the Economy, vol. 19. MIT Press, Cambridge, pp. 6599.
Hines, J., Rice, E., 1994. Fiscal paradise: foreign tax havens and American business. Quarterly Journal of Economics 109, 149182.
Hodder, L., McAnally, M., Weaver, C., 2003. The inuence of tax and non-tax factors on banks choice of organizational form. The Accounting Review 78,
297325.
Horst, T., 1971. The theory of the multinational rm: optimal behavior under different tariff and tax rates. Journal of Political Economy 79, 10591072.
House, C., Shapiro., J., 2008. Temporary investment tax incentives: theory with evidence from bonus depreciation. American Economic Review 98,
737768.
Huizinga, H., Laeven, L., Nicodeme, G., 2008. Capital structure and international debt shifting. Journal of Financial Economics 88, 80118.
Huizinga, H., Voget, J., 2009. International taxation and the direction and volume of cross-border M&As. Journal of Finance 64, 12171249.
Ivkovich, Z., Poterba, J., Weisbenner, S., 2005. Tax loss trading by individual investors. American Economic Review 95, 16051630.
Jacob, J., 1996. Taxes and transfer pricing: Income shifting and the volume of intrarm transfers. Journal of Accounting Research 34, 301312.
Jagolinzer, A., Matsunaga, S., Yeung, B., 2007. An analysis of insiders use of prepaid variable forward transactions. Journal of Accounting Research 45,
10551079.
Jarrell, G., Brickley, J., Netter, J., 1988. The market for corporate control: the empirical evidence since 1980. Journal of Economic Perspectives 2, 4968.
Jensen, M., 1986. Agency costs of free cash ow, corporate nance, and takeovers. American Economic Review 76, 323329.
Jensen, M., Meckling, W., 1976. Theory of the rm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3,
305360.
Jensen, M., Murphy, K.J., 1990. Performance pay and top-management incentives. Journal of Political Economy 98, 225264.
Jensen, M., Ruback, R., 1983. The market for corporate control: the scientic evidence. Journal of Financial Economics 11, 550.
Jin, L., 2006. Capital gain tax overhang and price pressure. Journal of Finance 61, 13991430.
Jin, L., Kothari, S., 2008. Effect of personal taxes on managers decision to sell unrestricted equity. Journal of Accounting and Economics 46, 2346.
Jones, J., 1991. Earnings management during import relief investigations. Journal of Accounting Research 29, 193228.
Joos, P., Pratt, J., Young, S., 2000. Book-tax differences and the pricing of earnings. Working Paper, Massachusetts Institute of Technology.
Jung, D., Pulliam, D., 2006. Predictive ability of the valuation allowance for deferred tax assets. Academy of Accounting and Financial Studies Journal 10,
4970.
Kalay, A., 1982. The ex-dividend day behavior of stock prices: a re-examination of the clientele effect. Journal of Finance 37, 10591070.
Kalay, A., Lemmon, M., 2008. Payout policy. In: Eckbo, E. (Ed.), Handbook of Corporate Finance: Empirical Corporate Finance, vol. 2. Elsevier Science,
Amsterdam.
Kaplan, S., Sensoy, B., Stromberg, P., 2009. Should investors bet on the jockey or the horse? Evidence from the evolution of rms from early business plans
to public companies. Journal of Finance 64, 75115.
Kaplan, S., Stromberg, P., 2004. Characteristics, contracts, and actions: evidence from venture capitalist analyses. Journal of Finance 59, 21772210.
Kaplan, S., Zingales, L., 2000. Investmentcash ow sensitivities are not valid measures of nancing constraints. Quarterly Journal of Economics 115,
707712.
Kemsley, D., 1998. The effect of taxes on production location. Journal of Accounting Research 36, 921941.
Kessler, G., 2008. Accounting standards wilt under pressure. The Washington Post, December 28, p. A01.
King, M., 1977. Public Policy and the Corporation. Chapman & Hall, London.
Klassen, K., 1997. The impact of inside ownership concentration on the trade-off between nancial and tax reporting. The Accounting Review 72,
455474.
Klassen, K., Pittman, J., Reed, M., 2004. A cross-national comparison of R&D expenditure decisions: tax incentives and nancial constraints. Contemporary
Klassen, K., Shackelford, D., 1998. State and provincial corporate tax planning: income shifting and sales apportionment factor management. Journal of
Klein, P., 2001. The capital gain lock-in effect and long-horizon return reversal. Journal of Financial Economics 59, 3362.
Kothari, S.P., Ramanna, K., Skinner, D., 2010. Implications for GAAP from an analysis of positive research in accounting. Journal of Accounting and
Economics, this issue, doi:10.1016.j.jacceco.2010.09.003.
Krull, L., 2004. Permanently reinvested foreign earnings, taxes, and earnings management. The Accounting Review 79, 745767.
Kumar, K., Visvanathan, G., 2003. The information content of the deferred tax valuation allowance. The Accounting Review 78, 471490.
Landsman, W., Shackelford, D., 1995. The lock-in effect of capital gains taxes: evidence from the RJR Nabisco leveraged buyout. National Tax Journal 48,
245259.
Lang, M., Shackelford, D., 2000. Capitalization of capital gains taxes: evidence from stock price reactions to the 1997 rate reduction. Journal of Public
Economics 76, 6985.
Lenter, D., Shackelford, D., Slemrod, J., 2003. Public disclosure of corporate tax return information: accounting, economics, and legal perspectives. National
Tax Journal 56, 803830.
Lev, B., Nissim, D., 2004. Taxable income, future earnings, and equity values. The Accounting Review 79, 10391074.
Li, O., Weber, D., 2009. Taxes and ex-day returns: evidence from REITs. National Tax Journal 62, 657676.
Lisowsky, P., 2010. Seeking shelter: empirically modeling tax shelters using nancial statement information. The Accounting Review 85, 16931720.
Lisowsky, P., Robinson, L., Schmidt, A., 2010. An examination of FIN 48: tax shelters, auditor independence, and corporate governance. Working Paper,
University of Illinois.
Litzenberger, R., Ramaswamy, K., 1979. The effect of personal taxes and dividends on capital asset prices: theory and empirical evidence. Journal of
Logue, K., 2005. Tax law uncertainty and the role of tax insurance. Virginia Tax Review 25, 339414.
Long, J., 1977. Efcient portfolio choice with differential taxation of dividends and capital gains. Journal of Financial Economics 5, 2553.
Long, J., 1978. The market valuation of cash dividends: a case to consider. Journal of Financial Economics 6, 235264.
MacKie-Mason, J., 1990. Do taxes affect corporate nancing decisions? Journal of Finance 45, 14711493
MacKie-Mason, J., Gordon, R., 1997. How much do taxes discourage incorporation? Journal of Finance 52, 477505
Maksimovic, V., Phillips, G., 2008. The industry life cycle and acquisitions and investment: does rm organization matter? Journal of Finance 63, 673709
Malmendier, U., Tate, G., 2005. CEO overcondence and corporate investment. Journal of Finance 60, 26612700.
Manzon, G., Plesko, G., 2002. The relation between nancial and tax reporting measures of income. Tax Law Review 55, 175.
Matheson, T., 2008. The development of taxable REIT subsidiaries, 20012004. IRS Statistics of Income Bulletin Spring., 196203.
Matsunaga, S., Shevlin, T., Shores, D., 1992. Disqualifying dispositions of incentive stock options: tax benets versus nancial reporting costs. Journal of
Accounting Research 30 (Suppl.), 3768.
Maydew, E., 1997. Tax induced earnings management by rms with net operating losses. Journal of Accounting Research 35, 8396.
Maydew, E., 2001. Empirical tax research in accounting: a discussion. Journal of Accounting and Economics 31, 389403.
Maydew, E., Schipper, K., Vincent, L., 1999. The impact of taxes on the choice of divestiture method. Journal of Accounting and Economics 28, 117150.
McClelland, J., Mills, L., 2007. Weighing benets and risks of taxing book income. Tax Notes 114, 779787.
McDonald, R., 2003. Is it optimal to accelerate the payment of income tax on share-based compensation? Working Paper, Northwestern University.
McGill, G., Outslay, E., 2004. Lost in translation: detecting tax shelter activity in nancial statements. National Tax Journal LVII, 739756.
McNichols, M., Stubben, S., 2008. Does earnings management affect rms investment decisions? The Accounting Review 83, 15711603
Meulbroek, L., 2001. The efciency of equity-linked compensation: understanding the full cost of awarding executive stock options. Financial
Management 30, 530.
Miller, G., Skinner, D., 1998. Determinants of the valuation allowance for deferred tax assets under SFAS No. 109. The Accounting Review 73, 213233.
Miller, M., 1977. Debt and taxes. Journal of Finance 32, 213233.
Miller, M., Modigliani, F., 1961. Dividend policy, growth, and the valuation of shares. Journal of Business 34, 411433.
Miller, M., Scholes, M., 1978. Dividends and taxes. Journal of Financial Economics 6, 333364.
Miller, M., Scholes, M., 1982. Dividends and taxes: some empirical evidence. Journal of Political Economy 90, 11181141.
Mills, L., 1998. Book-tax differences and Internal Revenue Service adjustments. Journal of Accounting Research 36 (2), 343356.
Mills, L., Erickson, M., Maydew, E., 1998. Investments in tax planning. Journal of American Taxation Association 20, 120.
Mills, L., Newberry, K., 2001. The inuence of tax and non-tax costs on book-tax reporting differences: public and private rms. Journal of the American
Taxation Association 23 (1), 119.
Mills, L., Newberry, K., 2005. Evidence of off-balance-sheet nancing from book-tax differences. Journal of Accounting Research 43, 251282.
Mills, L., Newberry, K., Trautman, W., 2002. Trends in book-tax income and balance sheet differences. Tax Notes 96, Special Report, pp. 11091124.
Mills, L., Plesko, G., 2003. Bridging the reporting gap: a proposal for more informative reconciling of book and tax income. National Tax Journal 56,
865893.
Modigliani, F., Miller, M., 1958. The cost of capital, corporation nance and the theory of investment. American Economic Review 48, 261297.
Modigliani, F., Miller, M., 1963. Corporate income taxes and the cost of capital: a correction. American Economic Review 53, 433443.
Murray, A., 2002. Narrowing tax gap should be priority of next congress. Wall Street Journal (October 8), A5.
Myers, S., 1984. The capital structure puzzle. Journal of Finance 39, 575592.
Myers, S., McConnell, J., Peterson, A., Soter, D., Stein, J., 1998. Vanderbilt University round table on the capital structure puzzle. Journal of Applied
Corporate Finance 11, 824.
Naranjo, R., Nimalendran, M., Ryngaert, M., 1998. Stock return, dividend yield, and taxes. Journal of Finance 53, 20292057.
Neubig, T., 2006a. Wheres the applause? Why most corporations prefer a lower tax rate. Tax Notes (April 24), 483486.
Neubig, T., 2006b. Testimony before the Senate Finance Committee Our business tax system: objectives, deciencies and options for reform. September
20, 2006.
Newberry, K., Dhaliwal, D., 2001. Cross-jurisdictional income shifting by U.S. multinationals: evidence from international bond offerings. Journal of
Odean, T., 1998. Are investors reluctant to realize their losses? Journal of Finance 53, 17751779
OECDOrganization for Economic Co-operation and Development, 1998. Fostering EntrepreneurshipThe OECD Job Strategy. OECD, Paris, p. 12.
Ofek, E., Yermack, D., 2000. Taking stock: equity-based compensation and the evolution of managerial ownership. Journal of Finance 55, 13671384.
Ofcer, M., 2007. The price of corporate liquidity: acquisition discounts for unlisted targets. Journal of Financial Economics 83, 571598.
Ohlson, J., 1995. Earnings, book values and dividends in security valuation. Contemporary Accounting Research 11, 661687.
Omer, T., Plesko, G., Shelley, M., 2000. The inuence of tax costs on organizational choice in the natural resource industry. Journal of the American
Omer, T., Terando, W., 1999. The effect of risk and tax differences on corporate and limited partnership capital structure. National Tax Journal 52,
699715.
Palepu, K., Healy, P., Bernard, V., 2000. Business Analysis and Valuation: Using Financial Statements. South Western College Publications.
Penman, S., 2001. Financial Statement Analysis and Equity Valuation. McGraw-Hill, Boston, MA.
Petroni, K., Shackelford, D., 1995. Taxation, regulation, and the organizational structure of property-casualty insurers. Journal of Accounting and
Phillips, J., 2003. Corporate tax-planning effectiveness: the role of compensation-based incentives. The Accounting Review 78, 847874.
Phillips, J., Pincus, M, Rego, S., 2003. Earnings management: new evidence based on deferred tax expense. The Accounting Review 78, 491521.
Plesko, G., 1995. The role of taxes in organizational choice: S conversions after the Tax Reform Act of 1986. Working Paper, MIT, Cambridge, MA.
Plesko, G., 2003. An evaluation of alternative measures of corporate tax rates. Journal of Accounting and Economics 35, 201226.
Poterba, J., 1987. How burdensome are capital gains taxes? Evidence from the United States. Journal of Public Economics 33, 157172.
Poterba, J., 1989. Venture capital and capital gains taxation. In: Summers, L. (Ed.), Tax Policy and the Economy, vol. 3. MIT Press, Cambridge, pp. 4768.
Poterba, J., 2004. Taxation and corporate payout policy. American Economic Review 94, 171175.
Poterba, J., 2007. Comments on: dissecting dividend decisions: Some clues about the effects of dividend taxation from recent UK reforms by Bond, S.,
Devereux, M., Klemm, A.. In: Auerbach, A., Hines, J., Slemrod, J. (Eds.), Taxing Corporate Income in the 21st Century. Cambridge University Press
New York, pp. 8592.
Poterba, J., Summers, L., 1985. The economic effect of dividend taxation. Working Paper, NBER.
Poterba, J., Weisbenner, S., 2001. Capital gains tax rules, tax-loss trading, and turn-of-the-year returns. Journal of Finance 56, 353368.
Reason, T., 2002. Align the Books? CFO Magazine, November 1, 2002.
Reese, W., 1998. Capital gains taxation and stock market activity: evidence from IPOs. Journal of Finance 53, 17991820.
Rego, S., 2003. Tax avoidance activities of U.S. multinational corporations. Contemporary Accounting Research 20, 805833.
Revsine, L., Collins, D., Johnson, W., 1998. In: Financial Reporting and Analysis. Prentice-Hall, Upper Saddle River, NJ.
Rhodes-Kropf, M., Robinson, D., Viswanathan, S., 2005. Valuation waves and merger activity: the empirical evidence. Journal of Financial Economics 77,
561603.
Robinson, J., Sikes, S., Weaver, C., 2010. Performance measurement of corporate tax departments. The Accounting Review 85, 10351064.
Robinson, L., Sansing, R., 2008. The effect of invisible tax preferences on investment and tax preference measures. Journal of Accounting and Economics
46, 389404.
Romer, C., Romer, D., 2008. A narrative analysis of postwar tax changes. Working Paper, University of California-Berkeley.
Rossotti, C., 2006. Testimony Before the Senate Committee on Finance, September 20, 2006.
Roychowdhury, S., 2006. Earnings management through real activities manipulation. Journal of Accounting and Economics 42, 335370.
Samuelson, L., 1982. The multinational rm with arms length transfer price limits. Journal of International Economics 13, 365374.
Sansing, R., 1999. Relationship-specic investments and the transfer pricing paradox. Review of Accounting Studies 4, 119134.
Sansing, R., Yetman, R., 2006. Governing private foundations using the tax law. Journal of Accounting and Economics, 363384.
Schaefer, S., 1982. Taxes and security market equilibrium. In: Sharpe, W., Cootner, C. (Eds.), Financial Economics: Essays in Honour of Paul H. Gootner.
Prentice-Hall, Englewood Cliffs, NJ, pp. 159178.
Schipper, K., Vincent, L., 2003. Earnings quality. Accounting Horizons 17 (Suppl.), 97110.
Schmidt, A., 2006. The persistence, forecasting, and valuation implications of the tax change components of earnings. The Accounting Review 81, 589616.
Scholes, M., Wilson, G., Wolfson, M., 1992. Firms responses to anticipated reductions in tax rates: the Tax Reform Act of 1986. Journal of Accounting
Scholes, M., Wolfson, M., 1989. Converting corporations to partnerships through leverage: theoretical and practical impediments. NBER Working Paper.
Scholes, M., Wolfson, M., 1991. The role of tax rules in the recent restructuring of US corporations. In: Bradford, D. (Ed.), Tax Policy and the Economy
vol. 5. MIT Press, Cambridge, pp. 125.
Scholes, M., Wolfson, M., 1992. Taxes and Business Strategy: A Planning Approach, rst ed Pearson Prentice-Hall, Upper Saddle River, NJ.
Scholes, M., Wolfson, M., Erickson, M., Maydew, E., Shevlin, T., 2009. Taxes and Business Strategy: A Planning Approach, fourth ed Pearson Prentice-Hall,
Upper Saddle River, NJ.
Schon, W., 2005. The David R. Tillinghast Lecture: the odd couple: a common future for nancial and tax accounting? New York University School of Law
Tax Law Review 58, 111
Schrand, C., Wong, F., 2003. Earnings management using the valuation allowance for deferred tax assets under SFAS No. 109. Contemporary Accounting
Seida, J., Randolph, D., Salamon, G., 2005. Quantifying the costs of intertemporal taxable income shifting: theory and evidence from the property-casualty
insurance industry. Accounting Review 80, 315348.
Seida, J., Wempe, W., 2004. Effective tax rate changes and earnings stripping following corporate inversions. National Tax Journal 57, 805827.
Seidman, J., 2009. Interpreting uctuations in the book-tax income gap as tax sheltering: alternative explanations. Working Paper, University of Texas.
Seyhun, N., 1986. Insiders prots, cost of trading, and market efciency. Journal of Financial Economics 16, 189212.
Seyhun, N., Skinner, D., 1994. How do taxes affect investors stock market realizations? Evidence from tax-return panel data. The Journal of Business 67 (2
(April)), 231262.
Shackelford, D., Shevlin, T., 2001. Empirical tax research in accounting. Journal of Accounting and Economics 31, 321387.
Shackelford, D., Slemrod, J., Sallee, J., 2010. Financial reporting, taxes, and real decisions; A unifying framework. Working Paper, University of North
Carolina.
Shackelford, D., Verrecchia, R., 2002. Intertemporal tax discontinuities. Journal of Accounting Research 40, 205222.
Shackelford, D., 2006. Testimony before the U.S. house of representatives committee on ways and Means. Subcommittee on Select Revenue Measures,
Corporate Tax Reform, May 9.
Shaviro, D., 2009. Internationalization of income measures and the U.S. book-tax relationship. National Tax Journal 62, 155168.
Shefrin, H., Statman, M., 1985. The disposition to sell winners too early and ride losers too long: theory and evidence. Journal of Finance 40, 777782.
Shevlin, T., 1987. Taxes and off-balance sheet nancing: research and development limited partnerships. The Accounting Review 62, 480509.
Shevlin, T., 1990. Estimating corporate marginal tax rates with asymmetric tax treatment of gains and losses. Journal of the American Tax Association 11,
5167.
Shevlin, T., 2002. Commentary: corporate tax shelters and book-tax differences. Tax Law Review 55, 427443.
Shivdasani, A., Stefanescu, I., 2010. How do pensions affect corporate capital structure decisions? Review of Financial Studies 23, 12871323.
Sialm, C., 2009. Tax changes and asset prices. American Economic Review 99, 13561383.
Sias, R., Starks, L., 1997. Return autocorrelation and institutional investors. Journal of Financial Economics 46, 103131.
Single, L., 1999. Tax holidays and rms subsidiary location decisions. Journal of the American Taxation Association 21, 2134.
Skinner, D., 2008. The rise of deferred tax assets in Japan: the role of deferred tax accounting in the Japanese banking crisis. Journal of Accounting and
Slemrod, J., 1990. The impact of the Tax Reform Act of 1986 on foreign direct investment to and from the United States. In: Slemrod, J. (Ed.), Do Taxes
Matter: The Impact of the Tax Reform Act of 1986. MIT Press, Cambridge.
Slemrod, J., 1992a. Why People Pay Taxes: Tax Compliance and Enforcement. University of Michigan Press, Ann Arbor, MI.
Slemrod, J., 1992b. Do taxes matter? Evidence from the 1980s. American Economic Review 82, 250256.
Slemrod, J., 2004. The economics of corporate tax selshness. National Tax Journal 57, 877899.
Slemrod, J., 2005. What corporations say they do, and what they really do: implications for tax policy and tax research. Journal of the American Taxation
Association 27, 9199.
Slemrod, J., Wilson, J., 2009. Tax competition with parasitic tax havens. Journal of Public Economics 93, 12611270.
Slemrod, J., Yitzhaki, S., 2002. Tax avoidance, evasion, and administration. In: Auerbach, A., Feldstein, M. (Eds.), Handbook of Public Economics, vol. 3.
Elsevier Science, Amsterdam, pp. 14231470.
Smith, M., 2002. Ex ante and ex post discretion over arms length transfer prices. The Accounting Review 77, 161184.
Summers, L., 1981. Taxation and corporate investment: a q-theory approach. Brookings Papers on Economic Activity 12, 67140.
Sundaram, R., Yermack, D., 2007. Pay me later: inside debt and its role in managerial compensation. Journal of Finance 62, 15511588.
Swenson, D., 2001. Tax reforms and evidence of transfer pricing. National Tax Journal 54, 725.
Tang, R., 2002. Current Trends and Corporate Cases in Transfer PricingQuorum Books, Westport, CT.
Tobin, J., 1969. A general equilibrium approach to monetary theory. Journal of Money, Credit, and Banking 1, 1529.
Treasury, H., 1999. Tackling Poverty and Extending Opportunity. HM Treasury, London.
U.S. Government Accountability Ofce, 2008. Report to the Congress on earnings stripping, transfer pricing and U.S. income tax treaties. GAO-08-950.
U.S. Treasury, 2007. Report to the congress on earnings stripping, transfer pricing and U.S. income tax treaties.
Visvanathan, G., 1998. Deferred tax valuation allowances and earnings management. Journal of Financial Statement Analysis 3, 615.
Voget, 2009. Impact of taxation on the location of corporate headquarters. Working Paper, Oxford University.
Weber, D., 2009. Do analysts and investors fully appreciate the implications of book-tax differences for future earnings? Contemporary Accounting
Research, 26
Weil, J., Liesman, W., 2001. Missed signals: stock gurus write off most big write-offs, but they shouldntspecial or unusual charges often hold vital
clues to companies prospects. The Wall Street Journal, December: p. A1.
Weisbach, D., 2002. Ten truths about tax shelters, 55 Tax L. Rev. 215.
Weisbach, D., 2003. Corporate tax avoidance. In: Proceedings of the National Tax Association Ninety Sixth Annual Conference, vol. 9.
Weisman, J., 2002. At rms, dual prot pictures: the gap grows between whats earned and whats taxed. The Washington Post, October 10, p. A01.
Wilson, P., 1993. The role of taxes in location and sourcing decisions. In: Giovannini, A., Hubbard, G., Slemrod, J (Eds.), Studies in International Taxation.
University of Chicago Press, Chicago, pp. 195231.
Wilson, R., 2009. An examination of corporate tax shelter participants. The Accounting Review 84, 969999.
Wolfe, R., 2003. Tax indemnity insurance: a valuable and evolving tool for managing tax risks. PLI/TAX 595, 604616.
Wulf, J., 2004. Do CEOs in mergers trade power for premium? Evidence from mergers of equals. Journal of Law, Economics and Organization 20, 60101.
Yermack, D., 2009. Deductio ad absurdum: CEOs donating their own stock to their own family foundations. Journal of Financial Economics 94, 107123.
Yetman, R., 2002. Tax motivated expense allocations by not-for-prot organizations. The Accounting Review 76, 297311.
Yetman, R., 2003. Nonprot taxable activities, production complementarities, and joint cost allocations. National Tax Journal 56, 789799.
Yin, G., 2003. How much tax do large public corporations pay? Estimating the effective tax rates of the S&P 500. Virginia Law Review December, 89.
Zechman, S., 2009. The relation between voluntary disclosure and nancial reporting: evidence from synthetic leases. Working Paper, University of
Chicago.
Zimmerman, J., 1983. Taxes and rm size. Journal of Accounting and Economics 5, 119149.
Zodrow, G., 1991. On the traditional and new views of dividend taxation. National Tax Journal 44, 497509.

Hanlon Dan Heitzman (2010)

Uploaded by

Copyright:

Available Formats

Hanlon Dan Heitzman (2010)

Uploaded by

Document Information

Original Description:

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

Hanlon Dan Heitzman (2010)

Uploaded by

Copyright:

Available Formats

Journal of Accounting and Economics 50 (2010) 127178

Contents lists available at ScienceDirect

Journal of Accounting and Economics

A review of tax research$

2. The informational role of accounting for income taxes

2.1. Sources of differences between book and taxable income

2.2. A taxonomy of the literature and review of the evidence

2.2.2. Are earnings managed through the tax accounts?

2.3. Book-tax conformity

2.4. Summary and suggestions for the future

3.1. Theory of corporate tax avoidance

3.2. Measuring tax avoidance

3.2.1. Effective tax rate measures

M. Hanlon, S. Heitzman / Journal of Accounting and Economics 50 (2010) 127178

3.2.2. Long-run effective tax rates

3.2.3. Book-tax differences

3.2.4. Discretionary or abnormal measures of tax avoidance

3.2.5. Unrecognized tax benets (UTB)

3.2.6. Tax shelter rms

3.3. Determinants of tax avoidance

3.4. The consequences of tax avoidance

3.5. Summary and suggestions for the future

4. Taxes, book-tax tradeoffs, and real corporate decisions

4.1.1. Theory of investment and taxes: a brief background

4.1.2. Taxes and investment: some evidence

4.1.3. Investment in intangibles

4.1.4. Investment location decisions

4.1.5. The reinvestment or repatriation decision

4.1.6. Corporate inversions to tax havens

4.1.7. Summary and thoughts for future research

4.2. Capital structure

4.2.1. Estimating the tax benets of debt

4.2.2. Measuring leverage

4.2.3. Summary and thoughts for the future

4.2.4. Payout policy

4.3. Organizational form

4.3.1. The choice of organizational form

4.3.2. Summary and thoughts for future research

4.4.1. Transfer pricing

4.4.2. Mergers and acquisitions

4.4.3. Executive compensation

4.4.4. Executive trading

5. Investor-level taxes and asset prices

5.1. Dividend taxation

5.1.1. The economic effects of dividend taxation: a brief introduction

New view Traditional view Irrelevance view

Effects of a tax rate increase on investment, payout, and share value

Effect of dividend payout and taxes on expected returns

5.1.2. Long-horizon returns: the effect of dividend yield on expected returns

5.1.4. Ex-dividend day studies

5.1.5. Evidence from valuation models

5.1.7. Remaining issues and questions for future research

5.1.8. Thoughts for future research

5.2. Capital gains taxation: capitalization and lock-in

5.2.1. Capital gains taxes and expected returns

5.2.2. Stock market realizations and the lock-in effect

Appendix A. Accounting for Income Taxes (SFAS 109)

You might also like