University of Southern California Dissertations and Theses

Leveling the playing field: unbiased tests of the relative information content of book income and taxable income

(USC Thesis Other)

Leveling the playing field: unbiased tests of the relative information content of book income and taxable income

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

LEVELING THE PLAYING FIELD: UNBIASED TESTS OF THE RELATIVE
INFORMATION CONTENT OF BOOK INCOME AND TAXABLE INCOME

by

James Stekelberg

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)

December, 2013

Copyright 2013 James Stekelberg

2

ACKNOWLEDGMENTS
I would like to express my deepest gratitude to my dissertation chair, Robert Trezevant,
for his guidance and insightful comments throughout the development of this project. I also
thank my other dissertation committee members, Elizabeth Chuk, Geert Ridder, and Charles
Swenson, for their comments and support. Additionally, this paper has benefited from the helpful
feedback of Eric Allen, Bryce Schonberger, K.R. Subramanyam, Kara Wells, and workshop
participants at the University of Southern California, Arizona State University, Michigan State
University, Southern Methodist University, the University of Arizona, and the University of
Missouri. Most of all, I thank my wife, Ingrid Chen, and my parents, Ann and Bill Stekelberg,
for their unconditional support throughout my doctoral studies.

3

TABLE OF CONTENTS
ACKNOWLEDGMENTS ………………………………...…………………………………….. 2

LIST OF TABLES ………………………………………………………...…………………….. 5

LIST OF FIGURES …………………………………………………………………………....... 6

ABSTRACT ……………………………………………………………………………………... 7

SECTION 1: INTRODUCTION …………………………………………………...………….. 8

SECTION 2: BACKGROUND, PRIOR LITERATURE, AND HYPOTHESIS
DEVELOPMENT ……………………………………………………………… 14

2.1 Background Information on Book Income and Taxable Income ………………. 14

2.2 Prior Literature on the Information Content of Taxable Income ……….……… 16

2.3 Prior Literature on Market Mispricing of Book Income and
Taxable Income ………………………………………………………………… 18

2.4 Removing Market Mispricing from the Valuation Model ……………………... 20

2.5 Hypothesis Development ………………………………………………………. 22

SECTION 3: RESEARCH DESIGN …………………………………………………………... 25

3.1 Variable Definitions ……………………………………………………………. 25

3.1.1 Definitions of Measures of Income …………………………………….. 25

3.1.2 Definitions of Measures of Firm Value ………………………………... 26

3.1.3 Definitions of Measures of Book Earnings Quality and
Tax Planning …………………………………………………………… 28

3.2 Sample Selection ……………………………………………………………….. 30

3.3 Full Sample Empirical Tests …………………………………………………… 32

3.4 Cross-Sectional Empirical Tests ……………………………………………….. 35

SECTION 4: RESULTS ……………………………………………………………………….. 37

4.1 Descriptive Statistics and Correlations ………………………………………… 37
4

4.2 Regression Results: Full Sample Tests ………………………………………… 38

4.3 Regression Results: Book Earnings Quality Tests ……………………………... 39

4.4 Regression Results: Tax Planning Tests ……………………………………….. 41

SECTION 5: ADDITIONAL ANALYSES ……………………………………….…………… 44

5.1 Analysis of Observations in Extreme Quintiles of Book Earnings
Quality and Tax Planning ……………………………………………………… 44

5.2 Taxable Income vs. Cash Flows from Operations ………………………...…… 45

5.3 Untabulated Robustness Tests …………………………………………………. 48

SECTION 6: CONCLUSION AND FUTURE RESEARCH ……………………………...…... 51

REFERENCES ………………………………………………………………………................ 54

APPENDIX: VARIABLE DEFINITIONS ……………………………………………….......... 79

5

LIST OF TABLES

TABLE 1: Sample Selection …………………………………………………….…………. 58

TABLE 2: Descriptive Statistics ………………………………………………..………….. 60

TABLE 3: Correlations …………………………………………………………………….. 61

TABLE 4: Relative Explanatory Power of Book Income and Taxable
Income: Full Sample …………………………………………………………… 62

TABLE 5: Relative Explanatory Power of Book Income and Taxable
Income: by Discretionary Accruals Quintile …………………………………... 64

TABLE 6: Relative Explanatory Power of Book Income and Taxable
Income: by Cash Effective Tax Rate Quintile …………………………………. 66

TABLE 7: Relative Explanatory Power of Book Income and Taxable
Income: Low Discretionary Accruals / Low Cash Effective
Tax Rate Subsample …………………………………………………………… 68

TABLE 8: Relative Explanatory Power of Book Income and Taxable
Income: High Discretionary Accruals / High Cash Effective
Tax Rate Subsample …………………………………………………………… 70

TABLE 9: Relative Explanatory Power of Taxable Income and Cash
Flows from Operations: Full Sample and by Discretionary
Accrual and Cash Effective Tax Rate Quintiles ……………………………….. 72

TABLE 10: Relative Explanatory Power of Book Income and Cash
Flows from Operations: Full Sample and by Discretionary
Accrual and Cash Effective Tax Rate Quintiles ……………………………….. 74
6

LIST OF FIGURES

FIGURE 1: Graphical Illustration of Table 4 ……………………………………………….. 76

FIGURE 2: Graphical Illustration of Table 5 ……………………………………………….. 77

FIGURE 3: Graphical Illustration of Table 6 ……………………………………………….. 78

7

ABSTRACT
In this study, I document that at least a portion of the superior ability of book income
relative to taxable income to explain the market value of equity may be due to market mispricing
arising from investors’ fixation on book income and underemphasis on the information contained
in taxable income, rather than book income’s superior information content. I find that this result
generally intensifies as book earnings quality and tax planning decrease. Indeed, I show that once
market mispricing is removed from the valuation model, taxable income possesses statistically
equivalent or even superior ability relative to book income to explain firm value among firms
with particularly low book earnings quality and firms that engage in a relatively low degree of
tax planning. This study adds to the growing literature on the informativeness of firms’ tax-
related financial statement disclosures by demonstrating that prior research may have conducted
tests of value relevance that are inherently biased in favor of book income and, consequently,
understated the relative information content of taxable income.

8

SECTION 1: INTRODUCTION

A growing body of research examines the information content of firms’ tax-related
financial statement disclosures for equity valuation purposes (see Graham, Raedy, &
Shackelford, 2012, for a review). In particular, Hanlon, Laplante, and Shevlin (2005) and Ayers,
Jiang, and Laplante (2009) examine the information content of a firm’s estimated taxable income
relative to its book income, concluding that book income is superior to taxable income as a
summary measure of firm performance, even when book earnings are of relatively low quality
and when the firm engages in a relatively low degree of tax planning. Hanlon et al. and Ayers et
al. use contemporaneous stock returns as the performance criterion. However, other research
indicates that contemporaneous stock returns may be biased by investors’ overemphasis, or
fixation, on the information contained in book income (e.g., Sloan, 1996; Xie, 2001), and
underemphasis on the information contained in taxable income (e.g, Lev & Nissim, 2004;
Thomas & Zhang, 2011) when forecasting future earnings and pricing securities. These findings
suggest that Hanlon et al. and Ayers et al. may have understated the relative information content
of taxable income as compared to book income as a result of these sources of market
mispricing.
1

In this study, I conduct empirical tests that attempt to overcome biases related to market
mispricing. My findings indicate that the ability of taxable income relative to book income to
explain firm value increases after removing the effects of market mispricing from the valuation
model, suggesting that Hanlon et al. (2005) and Ayers et al. (2009) may have understated the
information content of taxable income relative to book income. Further, in contrast to Ayers et

1
In the remainder of the text, “market mispricing” refers to mispricing arising from investors’ overemphasis on
book income and underemphasis on taxable income when forecasting future earnings and pricing securities.
9

al., my results indicate that taxable income actually surpasses book income as a summary
measure of firm performance when book income is of particularly low quality. These findings
are important because the information content of taxable income is a fundamental concern in the
literature on the value relevance of the tax information reported in firms’ financial statements.
Further, this issue is also pertinent to the policy debate on book/tax conformity (e.g., Hanlon &
Shevlin, 2005). In particular, if the information content of taxable income has been understated
by prior studies, then so has the information that would be lost if taxable income were conformed
to book income.
In my empirical tests, I compare the abilities of book income and taxable income to
explain the market value of equity and the ex-post intrinsic value of equity (Subramanyam &
Venkatachalam, 2007).
2
The market value of equity is a function of investors’ potentially-biased
expectations of future earnings and terminal stock price, and is therefore likely tainted by the
same sources of market mispricing that confound the tests performed by Hanlon et al. (2005) and
Ayers et al. (2009). On the other hand, ex-post intrinsic value attempts to remove market
mispricing from the valuation model by replacing investors’ ex-ante expectations of future
earnings and terminal stock price in the residual income model with five years of their ex-post
realizations. Consequently, using ex-post intrinsic value allows me to conduct tests of the
relative information content of book income and taxable income that overcome documented
biases arising from investors’ fixation on book income and underemphasis on taxable income
when forecasting future earnings and pricing securities.

2
Hanlon et al. (2005) and Ayers et al. (2009) use returns (changes) rather than price (levels) specifications in their
studies. Subramanyam and Venkatachalam (2007) note that using ex-post intrinsic value precludes the use of a
changes specification. I argue that my results are comparable to those reported by Hanlon et al. and Ayers et al.
because Kothari and Zimmerman (1995) suggest that price and returns frameworks are theoretically equivalent.
10

Consistent with Hanlon et al. (2005), I find that book income possesses superior ability
relative to taxable income to explain the market value of equity. However, if at least part of this
superiority is attributable to market mispricing of the information contained in the two measures
of income, then I expect that book income’s superior ability to explain firm value will decrease
once market mispricing is removed from the valuation model. The results of my tests bear out
this prediction. Specifically, I document that when ex-post intrinsic value replaces the market
value of equity as the measure of firm value, the superior explanatory power of book income
relative to taxable income decreases by 26.54%.
3
Since ex-post intrinsic value attempts to
overcome biases related to investors’ fixation on book income and underemphasis on taxable
income, this result suggests that at least part of book income’s superior ability to explain the
market value of equity may be due to market mispricing, and not due to book income’s superior
information content.
Next, in cross-sectional tests, I follow Ayers et al. (2009) and partition my sample into
quintiles based on the absolute value of discretionary accruals (as a proxy for book earnings
quality) and five-year “long-run” cash effective tax rates, or ETRs (as a proxy for tax planning). I
then rerun my primary analyses within each quintile. Similar to Ayers et al., I find that although
the ability of taxable income relative to book income to explain the market value of equity
generally increases as book earnings quality and tax planning decrease, book income nonetheless
retains its superior explanatory power, regardless of the degree of book earnings quality or tax
planning.
As before, however, I contend that at least part of the superior ability of book income
relative to taxable income to explain the market value of equity (even in the presence of low

3
As discussed in greater detail in Section 4.2, this figure is calculated by comparing the ratio of the R
2
s when the
market value of equity is regressed separately on book income and taxable income to the ratio of the R
2
s when ex-
post intrinsic value is regressed separately on book income and taxable income.
11

book earnings quality and low tax planning) may be due to market mispricing and, as a result,
that book income’s superior explanatory power will decrease once market mispricing is removed
from the valuation model. Further, I expect that investors’ overreliance on book income and
underreliance on taxable income will lead to more severe mispricing as the information content
of taxable income increases. Ayers et al. (2009) show that the informativeness of taxable income
relative to book income increases as book earnings quality and tax planning decrease. Thus, I
predict that once market mispricing is removed from the valuation model, the decrease in the
superior explanatory power of book income will intensify as book earnings quality and tax
planning decrease.
The results of my tests are generally consistent with this prediction. Specifically, I find
that when ex-post intrinsic value replaces the market value of equity as the measure of firm
value, the superior explanatory power of book income relative to taxable income increases by
18.05% then decreases by 6.77%, 62.46%, 44.20%, and 233.31% in the first (highest book
earnings quality) through fifth (lowest book earnings quality) quintiles of discretionary accruals,
respectively. Similarly, when ex-post intrinsic value replaces the market value of equity as the
measure of firm value, the superior explanatory power of book income relative to taxable income
decreases by 20.38%, 17.19%, 26.33%, 37.50%, and 77.38% in the first (highest tax planning)
through fifth (lowest tax planning) quintiles of cash ETRs, respectively. These trends show that
the decrease in book income’s superior explanatory power generally intensifies as book earnings
quality and tax planning decrease. Indeed, I find that when ex-post intrinsic value replaces the
market value of equity as the measure of firm value, taxable income possesses statistically
equivalent or even superior ability relative to book income to explain firm value among firms
12

with particularly low book earnings quality and among firms that engage in a relatively low
degree of tax planning.
This study contributes to the literature on the valuation of the tax information reported in
firms’ financial statements. For example, Hanlon et al. (2005) document that book income
possesses superior ability relative to taxable income to explain contemporaneous stock returns.
However, I find that book income’s superior ability to explain firm value decreases once market
mispricing arising from investors’ fixation on book income and underemphasis on taxable
income is removed from the valuation model. In this way, I show that Hanlon et al. may have
understated the relative information content of taxable income by conducting empirical tests that
are inherently biased in favor of book income.
My findings also contribute to the emerging literature on the use of taxable income as an
alternative to book income as a summary measure of firm performance. Using contemporaneous
stock returns as the performance criterion, Ayers et al. (2009) show that the informativeness of
taxable income as a summary measure of firm performance is greatest when book earnings
quality is relatively low but, nonetheless, that book income retains its superior ability to explain
contemporaneous stock returns regardless of the degree of book earnings quality. However, my
results indicate that once market mispricing is removed from the valuation model, the ability of
taxable income to explain firm value actually surpasses that of book income among firms with
particularly low book earnings quality. Therefore, in contrast to Ayers et al., my findings suggest
that in certain subsamples, taxable income may actually be more informative than book income
as a summary measure of firm performance.
Finally, the results reported in this study contribute to the policy debate on book/tax
conformity. Although prior research generally concludes that there would be a loss of
13

information to the market if book income and taxable income were conformed to one measure,
my findings indicate that the information content of taxable income may be even greater than
prior studies suggest. Thus, I demonstrate that conforming taxable income to book income could
result in a more severe loss of information than previously proposed.
The remainder of this paper proceeds as follows. Section 2 provides background
information on book income and taxable income, discusses prior literature on the information
content and market mispricing of these two measures of income, and develops my hypotheses.
Section 3 details the research design of this study. Section 4 reports its empirical results, while
Section 5 discusses additional analyses conducted to assess the robustness of these results.
Section 6 concludes and discusses potential avenues for future research related to the findings
reported in this study.

14

SECTION 2: BACKGROUND, PRIOR LITERATURE, AND
HYPOTHESIS DEVELOPMENT

2.1. Background Information on Book Income and Taxable Income
In the United States, firms report one measure of income to shareholders on their income
statements and another measure of income to tax authorities on their tax returns. I refer to these
measures of income as book income and taxable income, respectively. Book income is calculated
pursuant to Generally Accepted Accounting Principles (GAAP), the objective of which is to
provide relevant and faithfully-represented accounting information to financial statement users
such as current or potential investors and creditors (FASB, 2010). On the other hand, taxable
income is calculated according to the tax law, the objective of which is to raise revenue for the
federal government and achieve policy objectives such as encouraging or discouraging certain
activities and supporting certain industries (Scholes, Wolfson, Erickson, Maydew, & Shevlin,
2008).
Book income and taxable income are aligned to a large extent, as evidenced by prior
research documenting that firms must frequently make tradeoffs between reporting relatively
high book income but paying higher taxes or reporting relatively low book income but paying
lower taxes (e.g., Erickson, Hanlon, & Maydew, 2004; Guenther, Maydew, & Nutter, 1997).
Despite many similarities, however, considerable differences exist between book income and
taxable income due to the different objectives and sources of GAAP rules and tax law. Many of
these differences arise from the fact that book income is calculated under the accrual basis
whereas taxable income, although generally accrual-based, is in many ways calculated pursuant
to a method of accounting that is closely related to the cash method. For example, in calculating
15

taxable income, firms are often not permitted to deduct expenses before cash is paid. As a result,
the determination of taxable income lacks much of the subjectivity and discretion inherent in the
determination of book income. The U.S. Supreme Court summarized this view in the landmark
tax case Thor Power Tool Co. v. Commissioner:
“Financial accounting… is hospitable to estimates, probabilities, and reasonable
certainties. The tax law, with its mandate to preserve the revenue, can give no quarter to
uncertainty.”
4

This relative lack of subjectivity has led some observers to suggest that taxable income can be
used as a “baseline” measure of firm performance that is not as subject to manipulation as is
book income (e.g., Palepu & Healy, 2007; Revsine, Collins, Johnson, & Mittelstaedt, 2011;
Seida, 2003).
Book income is reported to financial statement users on the firm’s income statement,
while taxable income is reported to tax authorities on the firm’s tax return. Despite many
experts’ calls for increased transparency of firms’ tax-related information (e.g., Lenter,
Shackelford, & Slemrod, 2003; Mills & Plesko, 2003), firms are not currently required to
publicly disclose taxable income. However, GAAP does require that firms report some limited
tax information in their financial statements. Income tax expense, calculated pursuant to GAAP
and reported in the firm’s financial statements, is intended to represent total tax expense related
to the current period’s book income. Total tax expense is separated into current and deferred
components, where current tax expense is intended to represent taxes payable (or receivable) in
the current period, while deferred tax expense records the effect of items that are recognized in
different periods for book and tax purposes.
5
In theory, then, current tax expense is analogous to

4
Thor Power Tool Co. v. Comm., 99 S. Ct. 773 (USSC, 1979).
5
Items that affect deferred tax expense are called temporary book-tax differences (as opposed to permanent book-
tax differences) and arise from timing differences between book and tax treatment. There is a large literature on the
information content of temporary book-tax differences (e.g., Blaylock, Shevlin, & Wilson, 2012; Hanlon, 2005).
16

the tax due to (or receivable from) tax authorities on the current period’s tax return, and thus
grossing-up (dividing) current tax expense by the top U.S. statutory corporate tax rate can
provide financial statement users with an estimate of a firm’s taxable income.
Hanlon (2003) identifies a number of reasons why this estimate of taxable income likely
does not equal actual taxable income as reported on the firm’s tax return. Briefly, these reasons
relate to accounting for employee stock options; reserves for uncertain tax positions; intraperiod
tax allocation among continuing operations, discontinued operations, and extraordinary items;
tax credits; differential tax rates faced by multinational firms; and differing consolidation rules
between book and tax. It is generally not possible to adjust the estimate of taxable income to
more precisely account for these issues in a large-sample study such as this one. Nonetheless, as
discussed in greater detail in Section 5.2, I conduct robustness tests to at least partially address
some of the concerns identified by Hanlon. The results of these analyses are qualitatively
unchanged from the results of my primary tests. Further, despite the issues identified by Hanlon,
Plesko (2000, 2003) matches publicly-available financial statement data with confidential tax
return data and concludes that current tax expense is a reasonable approximation of actual taxes
owed in the current year.
6

2.2. Prior Literature on the Information Content of Taxable Income
Motivated by a growing gap between firms’ reported book income and estimated taxable
income, some policymakers and academics (e.g., Desai, 2005), have called for increased or even
complete conformity between book income and taxable income in order to constrain book

The primary finding of this stream of literature is that temporary book-tax differences may provide insight into the
quality of book income, or more specifically, the persistence of book income.
6
Specifically, Plesko (2000, 2003) reports a regression coefficient of .986 when actual tax liability (before tax
credits) is regressed on current federal tax expense as reported in the firm’s financial statements.
17

earnings management and overly aggressive tax planning by forcing firms to make book/tax
tradeoffs. Hanlon et al. (2005) study the potential consequences of book/tax conformity by
examining the value relevance of taxable income as compared to book income. Since the
objective of GAAP is to provide relevant and faithfully-represented information about firm
performance to financial statement users, Hanlon et al. predict that book income should be
relatively more informative to investors than is taxable income, which is not designed to be a
measure of firm performance per se. However, the authors also expect that investors may rely on
taxable income as an alternative to book income as a summary measure of firm performance due
to the lack of subjectivity inherent in the determination of taxable income as compared to book
income.
As predicted, Hanlon et al. (2005) document that taxable income provides additional
information content incremental to book income, but that book income possesses superior
information content relative to taxable income, in explaining contemporaneous stock returns.
Based on this evidence, Hanlon et al. conclude that book income and taxable income each
contain value-relevant information incremental to the other, but that book income is superior to
taxable income as a summary measure of firm performance when contemporaneous stock returns
are used as the performance criterion.
Ayers et al. (2009) extend Hanlon et al. (2005) by examining settings in which book
income may be a relatively less informative, and taxable income may be a relatively more
informative, summary measure of firm performance. Ayers et al. draw their hypotheses from
comments such as those made by Seida (2003), who suggests that investors can benchmark book
income against taxable income to evaluate book earnings quality, and Desai (2006), who
speculates that aggressive tax planning can obfuscate the relation between taxable income and
18

firm value. Based on these observations, Ayers et al. predict that the usefulness of taxable
income as an alternative to book income as a summary measure of firm performance is greatest
when book income is of relatively low quality, but least when firms engage in a relatively high
degree of tax planning.
Consistent with these predictions, Ayers et al. (2009) find that the ability of taxable
income relative to book income to explain contemporaneous stock returns is greatest among
firms in the lowest quintile of book earnings quality and, conversely, lowest among firms in the
highest quintile of tax planning. Notably, however, the authors find that book income retains its
superior ability relative to taxable income to explain contemporaneous stock returns across all
quintiles of book earnings quality and tax planning. That is, regardless of the degree of book
earnings quality or tax planning, book income remains superior to taxable income as a summary
measure of firm performance when contemporaneous stock returns are used as the performance
criterion. These findings are again consistent with the notion that book income is intended to be
an informative measure of firm performance, whereas taxable income is not.

2.3. Prior Literature on Market Mispricing of Book Income and Taxable Income
In summary, Hanlon et al. (2005) and Ayers et al. (2009) provide compelling evidence
that book income possesses superior ability relative to taxable income to explain
contemporaneous stock returns, even when book earnings quality and the degree of tax planning
are relatively low (i.e., when book income should be relatively less informative and when taxable
income should be relatively more informative). However, an underlying assumption in value
relevance studies such as Hanlon et al. and Ayers et al. is that investors correctly impound the
information contained in the measures of income under examination (Aboody, Hughes, & Liu,
19

2002). Prior research suggests that this is likely not the case with regards to both book income
and taxable income.
First, numerous studies, such as Sloan (1996) and Xie (2001), document that investors
fixate on book income when forming expectations of future earnings and pricing securities. If
investors fixate on book income, it follows that the observed superior ability of book income
relative to taxable income to explain contemporaneous stock returns could be at least partially
attributable to the fact that the market overemphasizes its reliance on book income, rather than
due to book income’s superior information content.
Second, a separate line of research provides evidence that the superior ability of book
income relative to taxable income to explain contemporaneous stock returns could be at least
partially due to the fact that investors do not fully appreciate the implications of taxable income
for future earnings. For example, Lev and Nissim (2004) document that the ratio of taxable
income to book income is associated with future earnings growth. However, since the authors
also find that this ratio is associated with future abnormal stock returns, Lev and Nissim
conclude that contemporaneous stock prices do not fully reflect the information contained in
taxable income. In a related study, Weber (2009) attributes at least part of the market mispricing
of the information contained in taxable income to the fact that analysts underemphasize this
information when making earnings forecasts. More recently, Thomas and Zhang (2011) find that
the unexpected component of tax expense, which the authors argue is informative about a firm’s
core profitability, is associated with future abnormal stock returns. This result again suggests that
contemporaneous stock prices do not fully reflect the information contained in taxable income,
despite its apparent usefulness in forming earnings expectations.
20

The previous findings discussed in this section are consistent with the broader behavioral
literature on the limited attention and incomplete revelation hypotheses (e.g., Bloomfield, 2003;
Hirshleifer & Teoh, 2003), which posit that investors tend to overemphasize (underemphasize)
their reliance on information that is easy (difficult) to gather or process. Since book income is
clearly and separately stated on the income statement, whereas constructing an estimate of
taxable income requires that investors use and comprehend complicated tax footnote disclosures,
the limited attention and incomplete revelation hypotheses would predict that investors both
overemphasize their reliance on book income and underemphasize their reliance on taxable
income when forming expectations of future earnings and pricing securities.

2.4. Removing Market Mispricing from the Valuation Model
To summarize the discussion thus far, Hanlon et al. (2005) and Ayers et al. (2009)
document that book income possesses superior ability relative to taxable income to explain
contemporaneous stock returns. However, other research suggests that the empirical tests
conducted by Hanlon et al. and Ayers et al. may be inherently biased in favor of book income
because of market mispricing arising from investors’ fixation on book income and
underemphasis on the information contained in taxable income. Thus, it is unclear to what extent
the superior ability of book income relative to taxable income to explain contemporaneous stock
returns is due to these sources of market mispricing rather than due to book income’s superior
information content.
Hanlon et al. (2005) and Ayers et al. (2009) use contemporaneous stock returns to
measure firm performance. In contrast, in this study I compare the relative abilities of book
income and taxable income to explain both the market value of equity and the ex-post intrinsic
21

value of equity. While the market value of equity is likely contaminated by the same sources of
market mispricing that bias the tests conducted by Hanlon et al. and Ayers et al., ex-post intrinsic
value, a measure of fundamental firm value developed by Subramanyam and Venkatachalam
(2007), attempts to remove market mispricing from the valuation model.
Ex-post intrinsic value is calculated by expressing two popular valuation models, the
dividend discount model and the residual income model formalized by Ohlson (1995), over
three-year and five-year finite time horizons. In brief, ex-post intrinsic value determined using
the dividend discount model is calculated by replacing investors’ expectations of future
dividends and terminal stock price in the traditional dividend discount model with three or five
years of their ex-post realizations. Similarly, ex-post intrinsic value determined using the residual
income model is calculated by replacing investors’ expectations of future earnings, book value,
and terminal stock price in the traditional residual income model with three or five years of their
ex-post realizations. The calculation of ex-post intrinsic value is expressed mathematically and
discussed in greater detail in Section 3.1.
Subramanyam and Venkatachalam (2007) contend that by removing investors’
potentially-biased expectations of future earnings and terminal stock price from the valuation
model, ex-post intrinsic value allows researchers to conduct tests of the relative information
content of alternative measures of income (such as book income and taxable income, in the case
of this study) that are not confounded by market mispricing. In their empirical tests,
Subramanyam and Venkatachalam find that book income possesses superior ability relative to
operating cash flows to explain both the market value of equity and ex-post intrinsic value. Since
ex-post intrinsic value is arguably unbiased by market mispricing, the authors conclude that book
income’s superior ability relative to operating cash flows to explain firm value (Dechow, 1994)
22

is due to book income’s superior information content and not due to mispricing arising from the
market’s fixation on book income.

2.5. Hypothesis Development
In this study, I conduct tests analogous to those performed by Subramanyam and
Venkatachalam (2007) to examine the extent to which the superior ability of book income
relative to taxable income to explain firm value decreases once market mispricing is removed
from the valuation model. To do this, I first test the relative abilities of book income and taxable
income to explain the market value of equity. These tests are equivalent to those conducted by
Hanlon et al. (2005) except that my tests use a price (levels) rather than a returns (changes)
specification. Consistent with Hanlon et al., I expect that book income will possess superior
ability relative to taxable income to explain the market value of equity.
However, the superior ability of book income relative to taxable income to explain the
market value of equity may be at least partially due to market mispricing arising from investors’
overemphasis on book income and underemphasis on taxable income, rather than due to book
income’s superior information content. Since ex-post intrinsic value attempts to overcome biases
arising from market mispricing, I predict that book income’s superior ability to explain firm
value will decrease when ex-post intrinsic value replaces the market value of equity as the
measure of firm value. This leads to my first formal hypothesis:
H1: The superior ability of book income relative to taxable income to explain
firm value will decrease when ex-post intrinsic value replaces the market
value of equity as the measure of firm value.

23

If mispricing is severe enough, this hypothesis suggests it is possible that the explanatory power
of taxable income may even surpass that of book income when ex-post intrinsic value replaces
the market value of equity as the measure of firm value.
My next hypotheses make cross-sectional predictions regarding settings in which taxable
income should be particularly informative about firm value relative to book income. As
discussed earlier, Ayers et al. (2009) find that although the ability of taxable income relative to
book income to explain contemporaneous stock returns increases as book earnings quality and
tax planning decrease, book income retains its superior explanatory power, regardless of the
degree of book earnings quality or tax planning. My tests are equivalent to those performed by
Ayers et al. except that I use a price (levels) rather than a returns (changes) specification. As
such, I expect that in my tests the ability of taxable income relative to book income to explain the
market value of equity will increase as book earnings quality and tax planning decrease, but that
book income will retain its superior explanatory power, regardless of the degree of book earnings
quality or tax planning.
However, I argue that the superior ability of book income relative to taxable income to
explain the market value of equity (even in the presence of low book earnings quality and low
tax planning) may be at least partially due to market mispricing rather than due to book income’s
superior information content. Thus, as before, I expect that book income’s superior explanatory
power will decrease once market mispricing is removed from the valuation model. Further, I
predict that investors’ overreliance on book income and underreliance on taxable income will
lead to more severe mispricing as the information content of taxable income (relative to book
income) increases. Ayers et al. (2009) suggest that the informativeness of taxable income relative
to book income increases as book earnings quality decreases. Therefore, I expect that once
24

market mispricing is removed from the valuation model, the decrease in book income’s superior
explanatory power will intensify as book earnings quality decreases. This logic leads to my next
formal hypothesis:
H2: When ex-post intrinsic value replaces the market value of equity as the
measure of firm value, the decrease in the superior ability of book income
relative to taxable income to explain firm value will intensify as book
earnings quality decreases.

Ayers et al. (2009) also suggest that the informativeness of taxable income relative to
book income increases as tax planning decreases. As before, I predict that investors’ overreliance
on book income and underreliance on taxable income will lead to more severe mispricing as the
information content of taxable income (relative to book income) increases. Thus, I expect that
once market mispricing is removed from the valuation model, the decrease in book income’s
superior explanatory power will intensify as tax planning decreases. This discussion leads to my
final formal hypothesis:
H3: When ex-post intrinsic value replaces the market value of equity as the
measure of firm value, the decrease in the superior ability of book income
relative to taxable income to explain firm value will intensify as tax planning
decreases.

Note that these hypotheses imply that once market mispricing is removed from the valuation
model, it is possible that the explanatory power of taxable income may even surpass that of book
income as book earnings quality and tax planning decrease.

25

SECTION 3: RESEARCH DESIGN

3.1. Variable Definitions
In this section I provide detailed definitions of the variables employed in this study.
Variable definitions are also summarized in the Appendix.

3.1.1. Definitions of Measures of Income
As discussed in Section 2, the purpose of this study is to conduct tests of the relative
information content of book income and taxable income that attempt to overcome biases related
to market mispricing. Following Hanlon et al. (2005) and Ayers et al. (2009), I calculate book
income as pretax book income less minority interest. I define this variable as PTBI. Using pretax
book income rather than (post-tax) net income maintains its comparability with taxable income,
which is itself a pretax figure.
To construct an estimate of a firm’s taxable income using its publicly available financial
statement data, I again follow Hanlon et al. (2005) and Ayers et al. (2009) by grossing-up
(dividing) worldwide current tax expense (the sum of current federal plus foreign tax expense)
by the applicable top annual U.S. statutory corporate tax rate
7
, then subtracting the change in tax
net operating loss carryforwards from the result.
8
I define this variable as TI. Mathematically, TI
is expressed as:

7
The top annual U.S. statutory corporate tax rate applicable during my sample period is 46% for years 1986 and
earlier, 40% for 1987, 34% from 1988 to 1992, and 35% from 1993 onwards.
8
Mills, Newberry, and Novack (2003) document a number of issues with using tax net operating loss data reported
in Compustat. In Section 5.3, I discuss robustness tests designed to address these issues.
26

where TAXFED is current federal tax expense, TAXFO is current foreign tax expense, STR is the
top U.S. statutory corporate tax rate applicable in period t, and ΔTAXNOL is the change in tax net
operating loss carryforwards from period t-1 to period t.
9
As discussed earlier, although this
estimate of taxable income likely contains a degree of measurement error, it represents a
reasonable approximation of actual taxable income as reported on the firm’s current year income
tax return.

3.1.2. Definitions of Measures of Firm Value
I compare the relative abilities of book income and taxable income to explain the market
value of equity and the ex-post intrinsic value of equity. To provide some assurance that the
market has had time to respond to the current period’s earnings, I calculate the market value of
equity as the stock price per common share three months after the end of the fiscal year,
multiplied by the number of common shares outstanding on the same date. I define this variable
as MVE.
Following Subramanyam and Venkatachalam (2007), I calculate ex-post intrinsic value
in four ways: using the dividend discount model and the residual income model, each expressed
over three-year and five-year finite time horizons.
10
For expositional clarity, I only report results
from using ex-post intrinsic value calculated based on the five-year residual income model. In
untabulated robustness tests discussed in greater detail in Section 5.3, I find that Pearson
correlations among the four measures of ex-post intrinsic value are all greater than 0.90 (all p <

9
Following Hanlon et al. (2005) and Ayers et al. (2009), if either TAXFED or TAXFO is missing, I calculate
worldwide current tax expense as total tax expense less deferred tax expense.
10
In theory, the dividend discount model and the residual income model are equivalent, since the residual income
model is merely an algebraic manipulation of the dividend discount model (assuming clean surplus accounting).
However, empirical estimates of the two ex-post intrinsic value measures are generally not equal due to
measurement errors over finite time horizons.
27

0.0001) and confirm that the results of this study are qualitatively unchanged when alternate ex-
post intrinsic value measurements are used.
To express the ex-post intrinsic value of equity pursuant to the five-year residual income
model, I begin with the infinite time horizon residual income model formalized by Ohlson
(1995):

∑

where MVE is the market value of equity, BV is the book value of equity, X is accounting
earnings (net income), and ρ is one plus the discount rate. I empirically estimate the discount rate
as the annual equal-weighted realized return from period t-1 to period t for the size and book-to-
market decile portfolio of which the firm is a member (Barber & Lyon, 1997).
11,

12
In sensitivity
tests discussed in greater detail in Section 5.3, I find that my results are robust to alternative
discount rate estimates.
Next, the residual income model can be expressed over a finite time horizon of N periods
by adding a terminal value term to the model as follows:

∑

where terminal value is equal to the market value of equity at time t+N less the book value of
equity at time t+N.
Finally, to calculate ex-post intrinsic value, which I define as IV, I replace expectations of
future earnings, book value, and terminal value in the residual income model with five years of
their ex-post realizations as follows:

11
Several studies (e.g., Claus & Thomas, 1999; Gebhardt, Lee, & Swaminathan, 2001) use analysts’ earnings
forecasts as inputs into the residual income model to estimate a firm-specific implied cost of equity capital, i.e., to
solve for ρ. This approach is not appropriate for the purposes of my study because research suggests that analyst
forecasts are biased by analysts’ fixation on book income (Bradshaw, Richardson, & Sloan, 2001) and failure to
properly comprehend the implications of taxable income for future earnings (Weber, 2009). Hence, using analyst
forecasts to estimate the discount rate would introduce into my measure of ex-post intrinsic value the very same
biases that the measure is designed to overcome.
12
Return data is obtained from Ken French’s website.
28

∑

As discussed earlier, prior research documents that investors fixate on book income and
underemphasize the information contained in taxable income when forming expectations of
future earnings and pricing securities. As shown above, the market value of equity is a function
of these potentially-biased expectations. In contrast, ex-post intrinsic value removes investors’
expectations of future earnings and terminal stock price from the valuation model. Therefore,
using ex-post intrinsic value instead of the market value of equity as a measure of firm value
allows me to conduct tests of the relative information content of book income and taxable
income that are arguably unbiased by these sources of market mispricing.
13
Note that since
market value at year t+3 or t+5 is used to express terminal value in the ex-post intrinsic value
calculation, this statement assumes that market mispricing of current book income and taxable
income does not persist three years or five years hence. This is likely a valid assumption, since
Aboody et al. (2002) note that “measurement errors [between intrinsic value and current market
value] tend to be resolved in no more than three years.”

3.1.3. Definitions of Measures of Book Earnings Quality and Tax Planning
In my cross-sectional tests, I employ the absolute value of discretionary accruals as a
proxy for book earnings quality. I define this variable as DACC. Consistent with Ayers et al.
(2009), I calculate DACC as the absolute value of the residual from the Jones (1991) model of

13
A number of studies, such as Francis, Olsson, and Oswald (2000), calculate intrinsic value using ex-ante analyst
forecasts of accounting attributes rather than their ex-post realizations. However, as mentioned earlier, prior research
shows that analysts fixate on book income when making earnings forecasts and that analysts fail to understand the
implications of taxable income for future earnings. Therefore, using estimates of intrinsic value calculated using
analyst forecasts of accounting attributes is not appropriate for my research design.
29

discretionary accruals as modified by Dechow, Sloan, and Sweeney (1995), run cross-sectionally
by each industry-year
14
with at least 10 observations, as follows:

[ (

)] [

] (

)
where TACC is total accruals (calculated as the change in current assets plus the change in short
term debt and less the change in current liabilities, the change in cash, and depreciation and
amortization expenses); TA is total assets; ΔSALES is the change in net sales from period t-1 to
period t; ΔREC is the change in accounts receivable from period t-1 to period t; and PPE is gross
property, plant, and equipment. Following Kothari, Leone, and Wasley (2005), I scale all
variables by lagged total assets and include an intercept term in the model to mitigate issues
related to heteroskedasticity and omitted size variable bias. Following Ayers et al., I partition
observations into quintiles based on the magnitude of DACC, where a higher value of DACC
suggests a relatively lower degree of book earnings quality.
15

I employ the five-year “long-run” cash effective tax rate (cash ETR), as a proxy for the
degree of tax planning undertaken by the firm. I define this variable as CETR. Following Dyreng,
Hanlon, and Maydew (2008) and Ayers et al. (2009), I calculate CETR as the sum of cash tax
expense from the current plus four prior years divided by the sum of pretax book income less
special items from the current plus four prior years. Mathematically, CETR is expressed as:
CETR it =
∑

∑

14
In this and all other tests, I use two-digit SIC codes to define industries. SIC codes are only available in
Compustat for years beginning in 1988. To remedy this issue, I backfill missing SIC codes for earlier years with the
first available SIC code.
15
An implicit assumption in this discussion is that discretionary accruals affect book earnings quality but not
“taxable income earnings quality,” i.e., that discretionary accruals result in book-tax differences. In a study of firms
that restated book earnings, Badertscher, Phillips, Pincus, and Rego (2008) provide evidence in support of this
assumption by documenting that the “overwhelming majority” of restatement firms use nonconforming earnings
management strategies (strategies that affect book income and taxable income differently) rather than conforming
earnings management strategies (strategies with the same book and tax treatments).
30

where CTE is cash tax expense, PTBI is pretax book income as previously defined, and SI is
special items. Using a “long-run” cash ETR aggregated over a multi-year time horizon
overcomes measurement issues, such as noise related to annual fluctuations in cash taxes paid or
refunded, inherent in using annual cash ETRs to measure tax planning. Further, using a cash
ETR rather than a GAAP ETR calculated using book tax expense omits items such as valuation
allowances and reserves for uncertain tax positions that are likely not associated with the level of
a firms’ tax planning activities. To control for variation in tax planning opportunities across
industries and intertemporal changes in statutory corporate tax rates, I follow Ayers et al. and
rank observations into industry-year quintiles based on the magnitude of CETR, where a higher
value of CETR indicates greater cash tax expense and therefore suggests that the firm engages in
a relatively lower degree of tax planning.
Guenther, Jones, and Njoroge (2012) note that inferences regarding the information
content of estimated taxable income are sensitive to and can be significantly affected by the
presence of extreme observations in the data. Therefore, to mitigate issues related to extreme
observations and/or data errors, I winsorize (reset) all regression variables at the 1% and 99%
levels.

3.2. Sample Selection
My initial sample consists of 47,366 U.S.-incorporated, non-financial/non-utility firm-
year observations (representing 5,579 distinct firms) with five subsequent years of data available
on Compustat, drawn from the period 1983 to 2005. Following Hanlon et al. (2005) and Ayers et
al. (2009), I do not include foreign, financial (SIC codes 4900-4999), or utility (SIC codes 6000-
6999) firms in my sample because these firms likely face different accounting rules and tax laws
31

than other firms in my sample. I begin my sample in 1983 to maintain consistency with Ayers et
al. I end my sample in 2005 because the calculation of ex-post intrinsic value requires five years
of subsequent data availability.
From this initial sample, I first delete 20,669 observations missing data necessary to
calculate book income, taxable income, market value of equity, or ex-post intrinsic value. Then,
consistent with Subramanyam and Venkatachalam (2007), I delete 1,375 observations with a
negative market value of equity or a negative ex-post intrinsic value because negative firm value
has no practical meaning or indicates a data error. I next delete 6,470 observations with negative
book income or negative taxable income because prior research documents a diminished or non-
specified relation between earnings and firm value among loss firms (e.g., Burgstahler & Dichev,
1997; Hayn, 1995) and because of inference issues related to combining samples of profit and
loss firms when studying the information content of estimated taxable income (Guenther et al.,
2012). This selection procedure results in a final sample of 18,852 firm-year observations
representing 3,426 distinct firms available for my full sample tests. Table 1, Panel A summarizes
the selection procedure for my full sample.
16

Since my measures of discretionary accruals and cash ETRs impose considerable (and
considerably different) data requirements on my sample, I follow Ayers et al. (2009) and
construct two additional subsamples for my cross-sectional tests related to book earnings quality
and tax planning. The subsample for my tests related to earnings quality begins with my full
sample of 18,852 observations. From this full sample, I delete 2,676 observations missing data
necessary to calculate discretionary accruals, resulting in a subsample of 16,176 firm-year

16
The calculation of ex-post intrinsic value requires five years of subsequent data. Further, I delete loss firms from
my sample of observations. I acknowledge that these data restrictions may introduce survivorship bias into my
sample. Consistent with this notion, I find that firms in my sample are on average larger and more profitable than
firms in the Compustat universe. As a result, I caution that the results reported in this study may not generalize to
small or unprofitable firms.
32

observations representing 2,759 distinct firms available for my tests related to earnings quality.
Table 1, Panel B summarizes the selection procedure for my subsample for tests related to
earnings quality.
The subsample for my tests related to tax planning also begins with my full sample of
18,852 observations. From this full sample, I first delete 9,087 observations missing data
necessary to calculate the five-year cash ETR. Note that data on cash tax expense is taken from
the statement of cash flows, which is only available for years beginning in 1988. Since the
calculation of my five-year cash ETR measure requires four years of lagged cash tax expense,
the sample period for my tests related to tax planning is effectively limited to the years 1992 and
later. I acknowledge that this restriction results in a significant loss of observations from my
primary tests, whose sample selection period begins in 1983.
Next, again following Ayers et al. (2009), I delete 613 observations with a negative
numerator (i.e., the five-year sum of cash taxes paid) or denominator (i.e., the five-year sum of
pretax book income less special items) in the five-year cash ETR calculation, plus another 206
observations with five-year cash ETRs greater than 100%. These data restrictions result in a
subsample of 8,946 firm-year observations representing 2,023 distinct firms available for my
tests related to tax planning. Table 1, Panel C summarizes the selection procedure for my
subsample for tests related to tax planning.

3.3. Full Sample Empirical Tests
I first test the relative abilities of book income and taxable income to explain the market
value of equity for my full sample of observations. To do this, I run a “horse race” between the
following two regression models:
33

(1)
and

(2)
where all variables are as defined earlier and in the Appendix.
Models (1) and (2) are equivalent to those employed by Hanlon et al. (2005) in their tests
of the relative information content of book income and taxable income, except that I use a levels
(price) specification rather than a changes (returns) specification in my tests. Hanlon et al.
document that book income possesses superior ability relative to taxable income to explain
contemporaneous stock returns. Consistent with this result, I expect that book income will also
possess superior ability relative to taxable income to explain the market value of equity.
Therefore, I predict that the R
2
of Model (1) is greater than the R
2
of Model (2), i.e., that the ratio
of the R
2
of Model (1) to the R
2
of Model (2), hereafter defined as RATIO
MVE
, is greater than
100%. Note that since each model has the same dependent variable and uses the same sample of
observations, it is possible to test the statistical significance of the difference in the explanatory
power of the independent variable(s) in each model by using the Vuong (1989) test of the
difference in R
2
between two non-nested models. The Vuong test has been extensively employed
in prior accounting research that studies the relative information content of alternative measures
of firm performance (e.g., Ayers et al., 2009; Dechow, 1994; Dhaliwal, Subramanyam, &
Trezevant, 1999; Hanlon et al., 2005; Subramanyam & Venkatachalam, 2007).
I contend that empirical tests of the relative abilities of book income and taxable income
to explain the market value of equity may be inherently biased in favor of book income due to
market mispricing arising from investors’ fixation on book income and underemphasis on the
information contained in taxable income. Thus, it is unclear to what extent book income
34

possesses superior ability relative to taxable income to explain the market value of equity
because of market mispricing rather than due to the superior information content of book
income. To examine this issue, I replace the market value of equity with ex-post intrinsic value
as the dependent variable in Models (1) and (2), then run a “horse race” between the resulting
two regression models:

(3)
and

(4)
where all variables are as defined earlier and in the Appendix.
As discussed earlier, using ex-post intrinsic value as the dependent variable in place of
the market value of equity allows me to conduct empirical tests that are arguably unbiased by
investors’ fixation on book income and underemphasis on taxable income. Since I expect that the
superior ability of book income relative to taxable income to explain the market value of equity
is at least partially attributable to these sources of market mispricing, H1 predicts that book
income’s superior ability to explain firm value will decrease when ex-post intrinsic value
replaces the market value of equity as the measure of firm value. As such, I expect that the ratio
of the R
2
of Model (3) to the R
2
of Model (4), hereafter defined as RATIO
IV
, will be less than
RATIO
MVE
. In an extreme case, this hypothesis implies that the R
2
of Model (4) may even be
greater than that of Model (3), i.e., that RATIO
IV
will be less than 100%, if the superior ability of
book income relative to taxable income to explain the market value of equity is sufficiently
attributable to market mispricing.

35

3.4. Cross-Sectional Empirical Tests
In cross-sectional tests, I partition my sample into quintiles based on the absolute value of
discretionary accruals and five-year cash ETRs, where greater values of these variables suggest
relatively lower degrees of book earnings quality and tax planning, respectively. I then run a
“horse race” between Models (1) and (2) above within each quintile. Ayers et al. (2009) show
that the ability of taxable income relative to book income to explain contemporaneous stock
returns is greatest among firms in the lowest quintiles of book earnings quality and tax planning.
Nonetheless, the authors find that book income retains its superior ability relative to taxable
income to explain contemporaneous stock returns across all quintiles of book earnings quality
and tax planning. Models (1) and (2) are equivalent to those employed by Ayers et al. in their
tests, except that I use a levels (price) specification rather than a changes (returns) specification.
Consistent with the results reported by Ayers et al., I expect that RATIO
MVE
will decrease as
book earnings quality and tax planning decrease, but that RATIO
MVE
will nonetheless remain
greater than 100% in each quintile of book earnings quality and tax planning.
As discussed earlier, the superior ability of book income relative to taxable income to
explain the market value of equity (even in the presence of low book earnings quality and low
tax planning) may be at least partially due to market mispricing rather than due to book income’s
superior information content. As before, to conduct tests that attempt to overcome biases due to
market mispricing, I replace the market value of equity with ex-post intrinsic value as the
measure of firm value, resulting in Models (3) and (4) above. I then run a “horse race” between
these models within each quintile of book earnings quality and tax planning.
I expect that the superior ability of book income relative to taxable income to explain
firm value will decrease when ex-post intrinsic value replaces the market value of equity as the
36

measure of firm value. Further, H2 and H3 predict that the decrease in the superior explanatory
power of book income relative to taxable income will intensify as book earnings quality and tax
planning, respectively, decrease. Therefore, H2 predicts that RATIO
IV
will be less than
RATIO
MVE
in each quintile of book earnings quality, and that the decrease will intensify as book
earnings quality decreases. Similarly, H3 predicts that RATIO
IV
will be less than RATIO
MVE
in
each quintile of tax planning, and that the decrease will intensify as tax planning decreases. Note
that these hypotheses imply that, as book earnings quality and tax planning decrease, it is
possible that the explanatory power of taxable income may even surpass that of book income
once market mispricing is removed from the valuation model. Thus, H2 and H3 suggest that it is
possible that RATIO
IV
may fall below 100% as book earnings quality and tax planning decrease.

37

SECTION 4: RESULTS

4.1. Descriptive Statistics and Correlations
Table 2 reports descriptive statistics for the primary variables examined in this study.
Mean (median) book income is $157.89 million ($19.74 million), while mean (median) taxable
income is $136.36 million ($16.29 million). Mean (median) market value of equity is $1,734.44
million ($195.45 million), while mean (median) ex-post intrinsic value is $3,119.93 million
($262.01 million). The mean (median) absolute value of discretionary accruals is 0.08 (0.05).
Finally, the mean (median) five-year cash ETR is 0.32 (0.32).
Table 3 presents Pearson (below the diagonal) and Spearman (above the diagonal)
correlations among these variables. I note that book income and taxable income are highly
correlated, with a Pearson correlation of 0.95. This correlation is consistent with the notion that
book income and taxable income are aligned to a large extent. I find a Pearson correlation of
0.73 between the market value of equity and ex-post intrinsic value, suggesting that these two
measures are distinct expressions of firm value. Further, book income is more highly correlated
with the market value of equity than is taxable income, with Pearson correlations of 0.86 and
0.82, respectively. On the other hand, the Pearson correlations between book income and ex-post
intrinsic value and between taxable income and ex-post intrinsic value are 0.68 and 0.67,
respectively. These correlations provide initial evidence in support of my hypothesis that the
superior ability of book income relative to taxable income to explain firm value will decrease
when ex-post intrinsic value replaces the market value of equity as the measure of firm value.

38

4.2 Regression Results: Full Sample Tests
Table 4 reports (and Figure 1 illustrates) results from estimating Models (1) and (2) to
test the relative abilities of book income and taxable income, respectively, to explain the market
value of equity.
17
I find that the R
2
of Model (1) is 73.15%, while the R
2
of Model (2) is 68.26%.
Thus, RATIO
MVE
is 73.15% / 68.26% = 107.16% (z-stat. = 5.63), indicating that, as expected
and consistent with Hanlon et al. (2005), book income possesses superior ability relative to
taxable income to explain the market value of equity. However, the market value of equity is
potentially biased by investors’ fixation on book income and underemphasis on the information
contained in taxable income when forecasting future earnings and pricing securities. Therefore, it
is unclear to what extent the superior ability of book income relative to taxable income to explain
the market value of equity is due to these sources of market mispricing rather than book
income’s superior information content.
To conduct tests of the relative information content of book income and taxable income
that attempt to overcome biases related to market mispricing, I next estimate Models (3) and (4),
which replace the market value of equity with ex-post intrinsic value as the dependent variable.
The results of these tests are reported in Table 4 (and illustrated in Figure 1). The results indicate
that book income possesses superior ability relative to taxable income to explain ex-post intrinsic
value, with R
2
s of 46.62% and 44.29% in Models (3) and (4), respectively. Thus, RATIO
IV
is
only 46.62% / 44.29% = 105.26% (z-stat. = 3.53), which is less than RATIO
MVE
of 107.16% in
both magnitude and statistical significance. This decrease between RATIO
MVE
and RATIO
IV

represents a 26.54% decrease (calculated as (105.26% - 107.16%) / (107.16% - 100%)) in the
superior explanatory power of book income relative to taxable income when ex-post intrinsic

17
For expositional clarity, I do not report coefficient estimates on the PTBI and TI variables, since my tests are
exclusively concerned with the explanatory power, or R
2
, of each model. I note that the coefficients on PTBI and TI
are positive and highly statistically significant (all p < 0.0001) in all models across all tests.
39

value replaces the market value of equity as the measure of firm value. This finding provides
support for H1 and suggests that book income’s superior ability relative to taxable income to
explain the market value of equity may be at least partially due to the effects of market
mispricing and not due to book income’s superior information content, as has been previously
assumed in the literature.

4.3. Regression Results: Book Earnings Quality Tests
Turning to my cross-sectional analyses, I next test H2, which predicts that when ex-post
intrinsic value replaces the market value of equity as the measure of firm value, the decrease in
the superior ability of book income relative to taxable income to explain firm value will intensify
as book earnings quality decreases. To test this prediction, I first estimate Models (1) and (2)
within quintiles partitioned on the absolute value of discretionary accruals, where greater
discretionary accruals suggest relatively lower quality book earnings. Table 5 reports (and Figure
2 illustrates) the results of these tests. Consistent with Ayers et al. (2009), I find that although the
ability of taxable income relative to book income to explain the market value of equity is greatest
among firms in the highest quintile of discretionary accruals (i.e., among firms with the
relatively lowest quality book earnings), book income nonetheless retains its superior
explanatory power in each quintile of discretionary accruals. Specifically, RATIO
MVE
is equal to
107.98% (z-stat. = 3.53), 106.20% (z-stat. = 2.01), 114.12% (z-stat. = 3.09), 110.00% (z-stat. =
3.12), and 104.89% (z-stat. = 1.70), in the first (lowest) through fifth (highest) discretionary
accrual quintiles, respectively. However, as argued earlier, book income’s superior ability
relative to taxable income to explain the market value of equity, even as book earnings quality
40

decreases, may be at least partially due to the effects of market mispricing and not due to the
superior information content of book income.
To conduct tests that are unbiased by market mispricing, I estimate Models (3) and (4),
which replace the market value of equity with ex-post intrinsic value as the dependent variable,
within each discretionary accruals quintile. Results of these tests are reported in Table 5 (and
illustrated in Figure 2). The results indicate that book income remains statistically superior to
taxable income only among those observations in the lowest discretionary accruals quintile (i.e.,
among firms with the relatively highest quality book earnings), with RATIO
IV
equal to 109.42%
(z-stat. = 3.11) in this quintile. On the other hand, among observations in the middle three
discretionary accruals quintiles, book income and taxable income possess statistically equivalent
abilities to explain ex-post intrinsic value, with RATIO
IV
equal to 105.78% (z-stat. = 1.12),
105.30% (z-stat. = 0.96), and 105.58% (z-stat. = 0.79) in the second through fourth quintiles,
respectively. Lastly, among observations in the highest discretionary accruals quintile, I find that
the R
2
of Model (4) is actually significantly greater than the R
2
of Model (3), with RATIO
IV

equal to only 93.97% (z-stat. = -2.46) in this quintile. This result indicates that taxable income
actually possesses superior ability relative to book income to explain ex-post intrinsic value
among firms with the lowest quality book earnings.
Finally, comparing RATIO
MVE
and RATIO
IV
indicates that, when ex-post intrinsic value
replaces the market value of equity as the measure of firm value, the superior ability of book
income relative to taxable income to explain firm value decreases (in both magnitude and
statistical significance) in each quintile of discretionary accruals except the first. Specifically, I
find that the superior explanatory power of book income increases by 18.05% then decreases by
6.77%, 62.46%, 44.20%, and 223.31% in the first through fifth quintiles of discretionary
41

accruals, respectively (note that as before, these figures are obtained by calculating (RATIO
IV
–
RATIO
MVE
) / (RATIO
MVE
– 100%) in each quintile). Although somewhat mixed in the middle
quintiles, this trend generally provides support for H2 that the decrease in the superior
explanatory power of book income relative to taxable income will intensify as book earnings
quality decreases. Indeed, these results indicate that once market mispricing of the information
contained in book income and taxable income is removed from the valuation model, the
explanatory power of taxable income actually surpasses that of book income among firms with
particularly poor book earnings quality.

4.4. Regression Results: Tax Planning Tests
I now turn to the results of my tests of H3. This hypothesis predicts that, when ex-post
intrinsic value replaces the market value of equity as the measure of firm value, the decrease in
the superior ability of book income relative to taxable income to explain firm value will intensify
as the degree of tax planning in which the firm engages decreases. To test this prediction, I first
estimate Models (1) and (2) within quintiles partitioned on the magnitude of cash ETRs, where a
greater cash ETR suggests that the firm engages in a relatively lower degree of tax planning.
18

Table 6 reports (and Figure 3 illustrates) the results of these tests. Consistent with Ayers et al.
(2009), I find that although the ability of taxable income relative to book income to explain the
market value of equity increases monotonically as cash ETRs increase (i.e., as tax planning
decreases), book income nonetheless retains its superior explanatory power in each quintile of
cash ETRs. Specifically, RATIO
MVE
is equal to 134.74% (z-stat. = 3.89), 114.31% (z-stat. =

18
Note that since observations are ranked by industry-year, quintiles are generally not of equal sizes for this set of
tests. For example, when an industry-year contains only two observations, SAS ranks these observations into the
second and fourth quintiles. Consequently, this industry-year will not have a first, third, or fifth quintile.
42

3.11), 111.62% (z-stat. = 1.97), 108.16% (z-stat. = 2.30), and 105.04% (z-stat. = 1.77), in the
first (lowest) through fifth (highest) cash ETR quintiles, respectively.
As argued earlier, the superior ability of book income relative to taxable income to
explain the market value of equity, even as the degree of tax planning in which the firm engages
decreases, may be at least partially due to the effects of market mispricing and not due to the
superior information content of book income. As before, to conduct tests that are unbiased by
market mispricing, I estimate Models (3) and (4), which replace the market value of equity with
ex-post intrinsic value as the dependent variable, within each cash ETR quintile. Results of these
tests are reported in Table 6 (and illustrated in Figure 3). I find that book income remains
statistically superior to taxable income only among those observations in the lowest two cash
ETR quintiles (i.e., among those firms that engage in the relatively highest degree of tax
planning), with RATIO
IV
equal to 127.66% (z-stat. = 2.23) and 111.85% (z-stat. = 2.22) in the
first and second quintiles, respectively. On the other hand, among observations in the highest
three cash ETR quintiles, RATIO
IV
is equal to only 108.56% (z-stat. = 1.50), 105.10% (z-stat. =
1.02), and 101.14% (z-stat. = 0.11) in the third through fifth quintiles, respectively. These results
indicate that book income and taxable income possess statistically equivalent abilities to explain
ex-post intrinsic value among firms that engage in a relatively low degree of tax planning.
Lastly, comparing RATIO
MVE
and RATIO
IV
indicates that, when ex-post intrinsic value
replaces the market value of equity as the measure of firm value, the superior ability of book
income relative to taxable income to explain firm value decreases (in both magnitude and
statistical significance) in each cash ETR quintile. Specifically, I find that the superior
explanatory power of book income decreases by 20.38%, 17.19%, 26.33%, 37.50%, and 77.38%
in the first through fifth quintiles of cash ETRs, respectively (note that as before, these figures
43

are obtained by calculating (RATIO
IV
– RATIO
MVE
) / (RATIO
MVE
– 100%) in each quintile).
Although this trend is not as pronounced as that of the book earnings quality test results, it
generally supports H3 that the decrease in the superior explanatory power of book income
relative to taxable income will intensify as tax planning decreases. In fact, these results indicate
that once market mispricing of the information contained in book income and taxable income is
removed from the valuation model, taxable income and book income possess statistically
equivalent abilities to explain firm value among firms that engage in a relatively low degree of
tax planning.

44

SECTION 5: ADDITIONAL ANALYSES

5.1 Analysis of Observations in Extreme Quintiles of Book Earnings Quality and Tax
Planning
My first set of additional analyses examines only those observations in both the lowest
quintile of discretionary accruals and lowest quintile of cash ETRs (i.e., those firms with the
relatively highest quality book earnings and highest degree of tax planning) and those
observations in both the highest quintile of discretionary accruals and highest quintile of cash
ETRs (i.e., those firms with the relatively lowest quality book earnings and lowest degree of tax
planning). My earlier predictions imply that when ex-post intrinsic value replaces the market
value of equity as the measure of firm value, the decrease in the superior explanatory power of
book income relative to taxable income will be especially muted in the first set of firms, since I
expect that the results of my tests will be tempered as book earnings quality and tax planning
increase. Conversely, my earlier predictions imply that when ex-post intrinsic value replaces the
market value of equity as the measure of firm value, the decrease in the superior explanatory
power of book income relative to taxable income will be especially pronounced in the second set
of firms, since I expect that the results of my tests will be intensified as book earnings quality
and tax planning decrease.
Table 7 reports results of estimating Models (1) – (4) within the low discretionary
accruals / low cash ETR subsample. I find that RATIO
MVE
is 121.37% (z-stat = 2.79), while
RATIO
IV
is 118.34% (z-stat = 2.15). Comparing these ratios indicates that when ex-post intrinsic
value replaces the market value of equity as the measure of firm value, the decrease in the
superior explanatory power of book income relative to taxable income is only 11.84% in this
45

subsample, compared to a 26.54% decrease in my full sample (reported in Table 4). This finding
is consistent with my prediction that the results of this study will be muted among firms with
both relatively high quality book earnings and a relatively high degree of tax planning.
Next, Table 8 reports results of estimating Models (1) – (4) within the high discretionary
accruals / high cash ETR subsample. I find that RATIO
MVE
is 122.52% (z-stat = 2.86) while
RATIO
IV
is 80.89% (z-stat = -2.30). The latter ratio indicates that the ability of taxable income to
explain ex-post intrinsic value actually surpasses that of book income among firms in this
subsample, consistent with the results reported in Table 5, which show that taxable income
possesses superior ability relative to book income to explain ex-post intrinsic value among firms
with particularly low quality book earnings. Finally, comparing these ratios indicates that when
ex-post intrinsic value replaces the market value of equity as the measure of firm value, the
decrease in the superior explanatory power of book income relative to taxable income is
184.86% in this subsample, compared to a decrease of only 26.54% in my full sample (reported
in Table 4). This finding is consistent with my prediction that the results of this study will be
especially pronounced among firms with both relatively low quality book earnings and a
relatively low degree of tax planning.

5.2. Taxable Income vs. Cash Flows from Operations
As noted earlier, taxable income is calculated pursuant to a method of accounting that
resembles the cash method of accounting in many respects. Therefore, the question remains
whether taxable income is merely a proxy for cash flows from operations and, consequently,
whether the results of this study are similar whether taxable income or cash flows from
46

operations is used as an alternative to book income as a summary measure of firm performance.
To address this concern, I run a number of supplementary analyses discussed below.
First, in untabulated robustness tests, I find that taxable income and pretax cash flows
from operations (calculated as cash flows from operations plus cash taxes paid, defined as CFO)
are correlated at 0.68 (p < .0001), indicating that taxable income and cash flows from operations
are not merely interchangeable. Further, in additional untabulated robustness tests, in regressions
of firm value on book income, pretax cash flows from operations, and taxable income, I find that
taxable income provides incremental explanatory power to book income and cash flows from
operations in explaining both the market value of equity and ex-post intrinsic value. This result
demonstrates that taxable income contains information separate from the information contained
in both book income and cash flows from operations in explaining firm value.
Next, I rerun my primary tests using pretax cash flows from operations instead of book
income in order to examine the relative abilities of cash flows from operations and taxable
income to explain firm value. The results of these tests are reported in Table 9. The results
indicate that taxable income is statistically superior to cash flows from operations in explaining
both the market value of equity and ex-post intrinsic value for the full sample. Further, the results
indicate that taxable income and cash flows from operations possess statistically different
abilities to explain the market value of equity and ex-post intrinsic value in all but three (out of
10) of the quintiles of discretionary accruals and cash ETRs. Since taxable income and cash
flows from operations possess statistically different explanatory powers across a variety of
specifications and samples of observations, these results again suggest that taxable income and
cash flows from operations are not interchangeable.
47

Lastly, I rerun my primary tests using pretax cash flows from operations instead of
taxable income as an alternative to book income as a summary measure of firm performance in
order to examine whether and how the results differ from those of my primary analyses (note that
these tests are similar in spirit to those performed by Subramanyam and Venkatachalam, 2007).
The results of these tests are reported in Table 10. The results indicate that, consistent with
Subramanyam and Venkatachalam and with my primary tests using taxable income instead of
cash flows from operations, book income is superior to cash flows from operations in explaining
both the market value of equity and ex-post intrinsic value, although book income’s superiority
declines when ex-post intrinsic value is used as the measure of firm value. This result is
consistent with the fact that one of the arguments made in this study with respect to taxable
income also applies to cash flows from operations – namely, that investors fixate on book
income and underemphasize cash flows when forecasting future earnings and pricing securities
(Sloan, 1996). Next, in cross-sectional tests using the market value of equity as the measure of
firm value, again the results are similar to those of my primary tests. Specifically, I find that
book income is superior to cash flows from operations in explaining the market value of equity,
regardless of the degree of book earnings quality or the degree of tax planning in which the firm
engages. However, this difference is not statistically significant among firms in the highest
quintile of cash ETRs, unlike in my primary tests, in which all differences were statistically
different.
The results of cross-sectional tests using ex-post intrinsic value as the dependent variable
differ considerably from those of my primary tests. Specifically, I find that book income remains
superior to cash flows from operations in explaining ex-post intrinsic value, regardless of the
degree of book earnings quality or the degree of tax planning in which the firm engages
48

(although this difference is not statistically significant among firms in the fourth quintile of
discretionary accruals or cash ETRs). This is in contrast to my primary test results, which show
that taxable income possesses superior ability relative to book income to explain ex-post intrinsic
value among firms with particularly low book earnings quality. Further, in my primary tests I
find that the decrease in book income’s superiority to taxable income generally intensifies as
book earnings quality and tax planning decrease. In contrast, I find no such pattern when cash
flows from operations is used in place of taxable income. To summarize, the results discussed in
this section are inconsistent with the notion that taxable income and cash flows from operations
are merely proxies for one another.

5.3. Untabulated Robustness Tests
Finally, I conduct a number of additional tests to ensure that the results reported in this
study are robust to a battery of alternative specifications. In particular, these robustness tests at
least partially address concerns regarding the use of tax net operating loss data reported in
Compustat; concerns regarding potential measurement error in estimated taxable income;
concerns regarding estimates of the discount rate used in the calculation of ex-post intrinsic
value; and alternative expressions of ex-post intrinsic value.
First, Mills et al. (2003) identify a number of issues with using tax net operating loss data
reported in Compustat. More specifically, these issues generally relate to Compustat reporting
that a firm has a U.S. tax net operating loss carryforward when, in fact, it does not, and
alternatively, Compustat reporting that a firm does not have a U.S. tax net operating loss
carryforward when, in fact, it does. Mills et al. note that these issues generally concern the use of
Compustat data to estimate U.S. taxable income, in contrast to this study, in which I estimate
49

worldwide taxable income. Nonetheless, in untabulated robustness tests, I rerun all of the
primary analyses conducted in this study after dropping observations with tax net operating loss
carryforwards reported in Compustat and find that my results are qualitatively unchanged. I also
rerun all of my primary tests using an estimate of taxable income omitting the change in tax net
operating loss carryforwards from the calculation, and again find that my results are qualitatively
unchanged.
Next, as noted earlier, Hanlon (2003) identifies a number of potential sources of
measurement error in the estimate of taxable income used in this study. Two of these issues are
the existence of tax credits and the fact that multinational firms may have a large amount of
income taxed at rates other than the top U.S. statutory corporate tax rate. Although it is generally
not possible to correct for these issues in a large-sample study such as this one, I at least partially
address these concerns by re-running the primary tests performed in this study after dropping
observations with high research and development expense (which I define as firms in the top
quartile of research and development expense scaled by pretax book income), because these
firms are likely to have high research and development tax credits. I also re-run my primary tests
after dropping firms with high foreign income (which I define as firms whose ratio of foreign
income to total income is greater than 50%), because these firms are likely to have high foreign
tax credits and/or a significant amount of income taxed at rates different from the top U.S.
statutory corporate tax rate. The results of these analyses are qualitatively unchanged from the
results of my primary tests.
Third, I conduct additional tests to ensure that my results are robust to alternative
discount rate measurements in the calculation of ex-post intrinsic value. Specifically, I use the
monthly risk-free (t-bill) rate plus a flat 6% risk premium and a flat discount rate of 10% as
50

alternative discount rate measurements, and find results that are qualitatively similar to those of
my primary tests. The results of these robustness tests are consistent with the results reported by
Penman and Sougiannis (1998), who find that the relative accuracy of alternative valuation
models is not sensitive to the choice of discount rate, and Subramanyam and Venkatachalam
(2007), who likewise find that their results are robust to a variety of alternative discount rate
measurements.
Finally, as noted earlier, in my primary tests I calculate ex-post intrinsic value pursuant to
the residual income model expressed over a five year finite time horizon. In my last set of
untabulated robustness tests, I follow Subramanyam and Venkatachalam (2007) and re-run my
primary tests using three alternative expressions of ex-post intrinsic value. Specifically, I confirm
that my results are qualitatively unchanged when I calculate ex-post intrinsic value pursuant to
the residual income model expressed over a three year finite time horizon. I also confirm that my
results are qualitatively unchanged when I calculate ex-post intrinsic value pursuant to the classic
dividend discount model expressed over both three year and five year finite time horizons.
Lastly, I find that Pearson correlations among the four measures of ex-post intrinsic value
employed in this study are all greater than 0.90 (p < .0001). These results are consistent with
those reported by Subramanyam and Venkatachalam, who find that their conclusions are
qualitatively unchanged when alternative expressions of ex-post intrinsic value are employed.

51

SECTION 6: CONCLUSIONS AND AVENUES FOR FUTURE RESEARCH

In this study, I conduct tests of the relative information content of book income and
taxable income that are arguably unbiased by the effects of market mispricing arising from
investors’ fixation on book income and underemphasis on the information contained in taxable
income. I report three primary findings. First, I find that the ability of book income relative to
taxable income to explain firm value decreases when investors’ potentially-biased expectations
of future earnings and stock price are removed from the valuation model. Second, I document
that this result generally intensifies as book earnings quality and tax planning decrease. Third, I
find that once market mispricing is removed from the valuation model, taxable income actually
possesses statistically equivalent or even superior ability relative to book income to explain firm
value among firms with particularly low book earnings quality and firms that engage in a
relatively low degree of tax planning.
This study contributes to the literature on the information content of firms’ tax-related
financial statement disclosures. Using contemporaneous stock returns as the performance
criterion, prior studies generally conclude that book income is superior to taxable income as a
summary measure of firm performance, even in the presence of low book earnings quality and
low tax planning. By conducting tests that overcome biases related to market mispricing, I am
able to show that the information content of taxable income relative to book income may be
greater than prior studies suggest and, in some cases, that the ability of taxable income to explain
firm value actually surpasses that of book income once market mispricing is removed from the
valuation model. These findings also inform the policy debate on whether book income and
taxable income should be conformed to one measure. By demonstrating that the information
52

content of taxable income may have been understated by prior research, my results indicate that
conforming taxable income to book income could result in a more significant loss of information
to the market than originally believed.
There are a number of potential avenues for future research related to the findings
reported in this study. One possible extension of the current study would be to examine whether
the empirical evidence is consistent with investors underweighting the information contained in
taxable income because the information is difficult to gather and understand — an assertion that
has been previously assumed, but not empirically tested, in the literature. To do this, future
research could examine whether abnormal returns associated with the information contained in
taxable income are concentrated in small firms, firms with relatively low analyst following, and
firms with relatively low levels of institutional ownership. One would expect that abnormal
returns would be concentrated in these firms because they have relatively less sophisticated
investors than do other firms, and less sophisticated investors would be more likely to ignore or
misunderstand the complex information reported in firms’ tax footnotes. Firm characteristics,
such as tax aggressiveness or the opacity of the firm’s information environment, could also
impact shareholders’ abilities to comprehend firms’ tax disclosures.
Another possible avenue for future study is to examine why, if the information contained
in taxable income provides information that is useful to investors, firms do not choose to disclose
their taxable income or, at least, more information about their tax activities in the footnotes to
their financial statements. One way that future work may be able to study this issue is to
empirically examine the characteristics of firms whose shareholders have called for increased
disclosure of the firm’s tax-related information via proxy proposals, in the spirit of work by Ferri
and Sandino (2009) who conduct a similar analysis in the context of firms voluntarily expensing
53

employee stock options. It may be interesting to study whether and how firm characteristics,
such as tax aggressiveness and corporate governance, differ between these “shareholder activist”
firms and other firms, and whether such firms’ tax-related behavior (such as becoming more or
less tax aggressive) changes following shareholder calls for increased tax disclosures.
Finally, future research related to this study could more directly examine the association
of book income and taxable income with future accounting attributes. For example, similar to
Dechow (1994) who studies the relative abilities of book income and cash flows from operations
to predict future earnings and future cash flows, future research could examine whether book
income or taxable income better predicts future earnings and future cash flows. Further research
could also examine in what settings taxable income may be superior to book income in
predicting future accounting attributes. Although this research design would be quite similar to
the one employed in this study, the benefit of this alternative design is that it overcomes potential
criticisms of the ex-post intrinsic value model and, instead, more directly tests the predictive
abilities of book income and taxable income for future earnings and cash flows.
.

54

REFERENCES

Aboody, D., Hughes, J., & Liu, J. (2002). Measuring value-relevance in a (possibly) inefficient
market. Journal of Accounting Research, 40, 965-986.

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, 15-
54.

Badertscher, B., Phillips, J., Pincus, M., & Rego, S. (2009). Earnings management strategies and
the trade-off between tax benefits and detection risk: to conform or not to conform? The
Accounting Review, 84, 63-97.

Barber, B., & Lyon, J. (1997). Detecting long-run abnormal stock returns: the empirical power
and specification of test statistics. Journal of Financial Economics, 43, 341-372.

Blaylock, B., Shevlin, T., & Wilson, R..( 2012). Tax avoidance, large positive temporary book-
tax differences, and earnings persistence. The Accounting Review, 87, 91-120.

Bradshaw, M., Richardson, S., & Sloan, R. (2001). Do analysts and auditors use information in
accruals? Journal of Accounting Research, 39, 45-74.

Bloomfield, R. (2002). The incomplete revelation hypothesis and financial reporting. Accounting
Horizons, 15, 311-328.

Burgstahler, D. & Dichev, I. (1997). Earnings, adaptation, and equity value. The Accounting
Review, 72, 187-216.

Claus, J., & Thomas, J. (1999). Equity premia as low as 3%? Evidence from analysts’ earnings
forecasts for domestic and international stock markets. Journal of Finance, 56, 1629-
1666.

Dechow, P. (1994). Accounting earnings and cash flows as measures of firm performance: the
role of accounting accruals. Journal of Accounting and Economics, 18, 3-42.

Dechow, P., Sloan, R., & Sweeney, A. (1995). Detecting earnings management. The
Accounting Review, 70, 193-225.

Desai, M. (2005). The degradation of reported corporate profits. Journal of Economic
Perspectives, 19, 171-192.

Desai, M. (2006). Testimony of Mihir Desai Associate Professor Harvard University before the
Subcommittee on Select Revenue Measures (Committee on Ways and Means, House of
Representatives May 9).

55

Dhaliwal, D., Subramanyam, K.R., & Trezevant, R. (1999). Is comprehensive superior to net
income as a measure of firm performance? Journal of Accounting and Economics, 26,
43-67.

Dyreng, S., Hanlon, M., & Maydew, E. (2008). Long-run corporate tax avoidance. The
Accounting Review, 83, 61-82.

Erickson, M., M. Hanlon, & E. Maydew. (2004). How much will firms pay for earnings that do
not exist? Evidence of taxes paid on allegedly fraudulent earnings. The Accounting
Review, 79, 387-408.

Ferri, F. & Sandino, T. (2009). The impact of shareholder activism on financial reporting and
compensation: the case of employee stock options expensing. The Accounting Review,
84, 433-466.

Financial Accounting Standards Board. (2010). Statement of financial accounting concepts no.
8., conceptual framework for financial reporting. Stamford, CT: FASB.

Francis, J., Olsson, P., & Oswald, D. (2000). Comparing the accuracy and explainability of
dividend, free cash flow, and abnormal earnings equity value estimates. Journal of
Accounting Research, 38, 45-70.

Gebhardt, W., Lee, C., & Swaminathan, B. (2001). Toward an implied cost of equity capital.
Journal of Accounting Research, 39, 135-176.

Graham, J., Raedy, J., & Shackelford, D. (2012). Research in accounting for income taxes.
Journal of Accounting Economics, 53, 412-434.

Guenther, D., Jones, S., & Njoroge, K. (2012). How research design choices affect inferences
about the value relevance of tax expense. Working paper, University of Oregon and the
College of William and Mary.

Guenther, D., Maydew, E., & Nutter, S. (1997). Financial reporting, tax costs, and book-tax
conformity. Journal of Accounting and Economics, 23, 225-258.

Hanlon, M. (2003). What can we infer about a firm’s taxable income from its financial
statements? National Tax Journal, 56, 831-863.

Hanlon, M. (2005). The persistence and pricing of earnings, accruals, and cash flows when firms
have large book-tax differences. The Accounting Review, 80, 137-156.

Hanlon, M., Laplante, S., & Shevlin, T. (2005). Evidence for the possible information loss of
conforming book income and taxable income. Journal of Law and Economics, 48, 407-
442.

56

Hanlon, M. & Shevlin, T. (2005). Book tax conformity for corporate income: an introduction to
the issues. Tax Policy and the Economy, Poterba, J., editor. Cambridge, MA: National
Bureau of Economic Research.

Hayn, C. (1995). The information content of losses. Journal of Accounting and Economics, 20,
125-153.

Hirshleifer, D. & Teoh, S. (2003). Limited attention, information disclosure, and financial
reporting. Journal of Accounting and Economics, 36, 337-386.

Jones, J. (1991). Earnings management during import relief investigations. Journal of
Accounting Research, 29, 193-228.

Kothari, S.P., Leone, A., & Wasley, C. (2005). Performance matched discretionary accruals.
Journal of Accounting and Economics, 39, 163-197.

Kothari, S.P. & Zimmerman, J. (1995). Price and return models. Journal of Accounting and
Economics, 20, 155-192.

Lenter, D., Shackelford, D., & Slemrod, J. (2003). Public disclosure of corporate tax return
information: accounting, economics, and legal perspectives. National Tax Journal, 56,
803-830.

Lev, B. & Nissim, D. (2004). Taxable income, future earnings, and equity values. The
Accounting Review, 79, 1039-74.

Mills, L., Newberry, K., & Novack, G. (2003). How well does Compustat NOL data identify
firms with U.S. tax return loss carryovers? Journal of the American Taxation Association,
25, 1-17.

Mills, L. & Plesko, G. (2003). Bridging the reporting gap: a proposal for more informative
reconciling of book and tax income. National Tax Journal, 56, 865-893.

Ohlson, J. (1995). Earnings, book values, and dividends in equity valuation. Contemporary
Accounting Research, 11, 661-687.

Palepu, K. & Healy, P. (2007). Business analysis and valuation: using financial statements.
Mason, OH: South Western College Publications.

Penman, S. & Sougiannis, T. (1998). A comparison of dividend, cash flow, and earnings
approaches to equity valuation. Contemporary Accounting Research, 14, 343-383.

Plesko, G. (2000). Book-tax differences and the measurement of corporate income. NTA
Proceedings 1999, 171-176.

57

Plesko, G. (2003). An evaluation of alternative measures of corporate tax rates. Journal of
Accounting and Economics, 35, 201-236.

Revsine, L., Collins, D., Johnson, B., & Mittelstaedt, F. (2011). Financial reporting and
analysis. New York, NY: McGraw Hill – Irwin.

Scholes, M., Wolfson, M., Erickson, M., Maydew, E., & Shevlin, T. (2008). Taxes and business
strategy: a planning approach. Upper Saddle River, NJ: Pearson-Prentice Hall.

Seida, J. (2003). Enron: the Joint Committee on Taxation’s investigative report. Testimony
before the Senate Committee on Finance (108
th
Congress, 1
st
Session). S. Hrg. 108-117
(February 13).

Sloan, R. (1996). Do stock prices fully reflect information in accruals and cash flows about
future earnings? The Accounting Review, 71, 305-333.

Subramanyam, K.R. & Venkatachalam, M. (2007). Earnings, cash flows, and the ex-post
intrinsic value of equity. The Accounting Review, 82, 457-481.

Thomas, J. & Zhang, F. (2011). Tax expense momentum. Journal of Accounting Research, 49,
791-821.

Vuong, Q.H. (1989). Likelihood ratio tests for model selection and non-nested hypotheses.
Econometrica, 57, 307-333.

Weber, D. (2009). Do analysts and investors fully appreciate the implications of book-tax
differences for future earnings? Contemporary Accounting Research, 26, 1175-1206.

Xie, H. (2001). The mispricing of discretionary accruals. The Accounting Review, 76, 357-73.

58

TABLE 1

Sample Selection

Panel A: Full Sample

Firm-Years

Distinct Firms

All U.S.-incorporated, nonfinancial/nonutility firm-years with 47,366

5,579

five-year-ahead data available on Compustat, 1983-2005

Less: missing data necessary to calculate book income,
taxable income, market value of equity, or ex-post intrinsic
value
(20,669)

(1,350)

Less: negative market value of equity or ex-post intrinsic
(1,375)

(137)

value

Less: negative book income or taxable income
(6,470)

(666)

Final Sample for Full Sample Tests 18,852

3,426

Panel B: Earnings Quality Tests Subsample

Firm-Years

Distinct Firms

Full Sample

18,852

3,426

Less: missing data necessary to calculate discretionary
accruals
(2,676)

(667)

Final Sample for Earnings Quality Tests 16,176 2,759

Notes: this table reports the sample selection procedure followed in this study.

59

TABLE 1, CONT.

Sample Selection
Panel C: Tax Planning Tests Subsample

Firm-Years

Distinct Firms

Full Sample

18,852

3,426

Less: missing data necessary to calculate the five-year cash
effective tax rate
(9,087)

(1,153)

Less: negative numerator or denominator in the five-year cash
effective tax rate calculation
(613)

(167)

Less: five-year cash effective tax rate greater than 100%
(206)

(83)

Final Sample for Tax Planning Tests 8,946

2,023

Notes: this table reports the sample selection procedure followed in this study.

60

TABLE 2

Descriptive Statistics

Variable n Mean Median Std Dev Min P25 P75 Max

PTBI 18,852 157.89 19.74 488.74 0.12 4.68 80.75 3,654.00

TI 18,852 136.36 16.29 420.26 0.00 3.88 67.64 3,065.71

MVE 18,852 1,734.44 195.45 5,539.77 3.28 46.85 878.61 42,275.28

IV 18,852 3,119.93 262.01 10,797.01 1.01 47.82 1,301.92 82,239.01

DACC 16,176 0.08 0.05 0.08 0.00 0.02 0.10 0.47

CETR 8,946 0.32 0.32 0.13 0.01 0.25 0.39 0.75

Notes: this table reports descriptive statistics for the primary variables examined in this study.
All variables are as defined in Appendix A. Variables are winsorized at the 1% and 99% levels.

61

TABLE 3

Correlations

Variable n PTBI TI MVE IV DACC CETR

PTBI 18,852 - 0.93*** 0.93*** 0.80*** -0.15*** -0.13***

TI 18,852 0.95*** - 0.89*** 0.76*** -0.14*** -0.04***

MVE 18,852 0.86*** 0.82*** - 0.81*** -0.13*** -0.15***

IV 18,852 0.68*** 0.67*** 0.73*** - -0.11*** -0.13***

DACC 16,176 -0.07*** -0.06*** -0.05*** -0.05*** - -0.02*

CETR 8,946 -0.09*** -0.06*** -0.10*** -0.09*** -0.02** -

Notes: this table reports Pearson (below the diagonal) and Spearman (above the diagonal)
correlations for the primary variables examined in this study. All variables are as defined in
Appendix A. For clarity, p-values are omitted. The symbols *, **, and *** indicate correlation
coefficients significant at the 10%, 5%, and 1% levels, respectively. Variables are winsorized at
the 1% and 99% levels.
62

TABLE 4

Relative Explanatory Power of Book Income and Taxable Income: Full Sample

MVE

IV

PTBI

R
2
(1) = 73.15%

R
2
(3) = 46.62%

TI

R
2
(2) = 68.26%

R
2
(4) = 44.29%

R
2
RATIO

RATIO
MVE
= R
2
(1) / R
2
(2) =
107.16%

RATIO
IV
= R
2
(3) / R
2
(4) =
105.26%

z-stat.

5.63***

3.53***

PERCENTAGE CHANGE IN THE

SUPERIORITY OF BOOK INCOME

(RATIO
IV
- RATIO
MVE
) / (RATIO
MVE
- 100%) = -26.54%

n = 18,852

Notes: this table reports the explanatory power, or R
2
, of separate regressions of book income (PTBI) and taxable income (TI) on the
market value of equity (MVE) and ex-post intrinsic value of equity (IV). Detailed variable definitions are provided in Appendix A.
Regression coefficients and t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference in explanatory power
63

between two non-nested models are reported in italics. The symbols *, **, and *** indicate z-statistics significant at the 10%, 5%, and
1% levels, respectively. Variables are winsorized at the 1% and 99% levels.
64

TABLE 5

Relative Explanatory Power of Book Income and Taxable Income by Discretionary Accruals Quintile

MVE

IV

PERCENTAGE CHANGE IN THE

SUPERIORITY OF BOOK INCOME

Quintile n PTBI TI RATIO
MVE
z-stat.

PTBI TI RATIO
IV
z-stat.

(RATIO
IV
- RATIO
MVE
) /

R
2
(1) R
2
(2) R
2
(1) / R
2
(2)

R
2
(3) R
2
(4) R
2
(3) / R
2
(4)

(RATIO
MVE
- 100%)

Q1 (Low DACC) 3,235 79.18% 73.33% 107.98% 3.53***

52.50% 47.98% 109.42% 3.11***

18.05%

Q2 3,235 76.03% 71.59% 106.20% 2.01**

45.37% 42.89% 105.78% 1.12

-6.77%

Q3 3,236 74.82% 65.56% 114.12% 3.09***

44.31% 42.08% 105.30% 0.96

-62.46%

Q4 3,235 68.72% 62.47% 110.00% 3.12***

44.69% 42.33% 105.58% 0.79

-44.20%

Q5 (High DACC) 3,235 68.90% 65.69% 104.89% 1.70*

37.22% 39.61% 93.97% -2.46**

-223.31%

This table reports the explanatory power, or R
2
, of separate regressions of book income (PTBI) and taxable income (TI) on the market
value of equity (MVE) and ex-post intrinsic value of equity (IV) within discretionary accrual (DACC) quintiles, where a higher quintile
indicates a lower degree of book earnings quality. Detailed variable definitions are provided in Appendix A. Regression coefficients
65

and t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference in explanatory power between two non-nested
models are reported in italics. The symbols *, **, and *** indicate z-statistics significant at the 10%, 5%, and 1% levels, respectively.
Variables are winsorized at the 1% and 99% levels.

66

TABLE 6

Relative Explanatory Power of Book Income and Taxable Income by Cash Effective Tax Rate Quintile

MVE

IV

PERCENTAGE CHANGE IN THE

SUPERIORITY OF BOOK INCOME

Quintile n PTBI TI RATIO
MVE
z-stat.

PTBI TI RATIO
IV
z-stat.

(RATIO
IV
- RATIO
MVE
) /

R
2
(1) R
2
(2) R
2
(1) / R
2
(2)

R
2
(3) R
2
(4) R
2
(3) / R
2
(4)

(RATIO
MVE
- 100%)

Q1 (Low CETR) 1,540 69.62% 51.67% 134.74% 3.89***

28.11% 22.02% 127.66% 2.23**

-20.38%

Q2 1,917 67.41% 58.97% 114.31% 3.11***

26.33% 23.54% 111.85% 2.22**

-17.19%

Q3 1,918 74.25% 66.52% 111.62% 1.97**

32.58% 30.01% 108.56% 1.50

-26.33%

Q4 1,917 76.10% 70.36% 108.16% 2.30**

30.10% 28.64% 105.10% 1.02

-37.50%

Q5 (High CETR) 1,654 70.92% 67.52% 105.04% 1.77*

38.07% 37.64% 101.14% 0.11

-77.38%

Notes: this table reports the explanatory power, or R
2
, of separate regressions of book income (PTBI) and taxable income (TI) on the
market value of equity (MVE) and ex-post intrinsic value of equity (IV) within five-year cash effective tax rate (CETR) quintiles,
where a higher quintile indicates that the firm engages in a relatively lower degree of tax planning. Detailed variable definitions are
provided in Appendix A. Regression coefficients and t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference
67

in explanatory power between two non-nested models are reported in italics. The symbols *, **, and *** indicate z-statistics
significant at the 10%, 5%, and 1% levels, respectively. Variables are winsorized at the 1% and 99% levels.

68

TABLE 7

Relative Explanatory Power of Book Income and Taxable Income: Low Discretionary Accruals / Low Cash Effective Tax Rate
Subsample

MVE

IV

PTBI

R
2
(1) = 78.66%

R
2
(3) = 42.13%

TI

R
2
(2) = 64.81%

R
2
(4) = 35.45%

R
2
RATIO

RATIO
MVE
= R
2
(1) / R
2
(2) =
121.37%

RATIO
IV
= R
2
(3) / R
2
(4) =
118.84%

z-stat.

2.79***

2.15**

PERCENTAGE CHANGE IN THE

SUPERIORITY OF BOOK INCOME

(RATIO
IV
- RATIO
MVE
) / (RATIO
MVE
- 100%) = -11.84%

n = 296

Notes: this table reports the explanatory power, or R
2
, of separate regressions of book income (PTBI) and taxable income (TI) on the
market value of equity (MVE) and ex-post intrinsic value of equity (IV) for those observations in both the lowest quintile of
discretionary accruals (DACC) and lowest quintile of cash effective tax rates (CETR). Detailed variable definitions are provided in
69

Appendix A. Regression coefficients and t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference in
explanatory power between two non-nested models are reported in italics. The symbols *, **, and *** indicate z-statistics significant
at the 10%, 5%, and 1% levels, respectively. Variables are winsorized at the 1% and 99% levels.

70

TABLE 8

Relative Explanatory Power of Book Income and Taxable Income: High Discretionary Accruals / High Cash Effective Tax
Rate Subsample

MVE

IV

PTBI

R
2
(1) = 77.30%

R
2
(3) = 41.82%

TI

R
2
(2) = 63.09%

R
2
(4) = 51.70%

R
2
RATIO

RATIO
MVE
= R
2
(1) / R
2
(2) =
122.52%

RATIO
IV
= R
2
(3) / R
2
(4) =
80.89%

z-stat.

2.86***

-2.30**

PERCENTAGE CHANGE IN THE

SUPERIORITY OF BOOK INCOME

(RATIO
IV
- RATIO
MVE
) / (RATIO
MVE
- 100%) = -184.86%

n = 318

Notes: this table reports the explanatory power, or R
2
, of separate regressions of book income (PTBI) and taxable income (TI) on the
market value of equity (MVE) and ex-post intrinsic value of equity (IV) for those observations in both the highest quintile of
discretionary accruals (DACC) and highest quintile of cash effective tax rates (CETR). Detailed variable definitions are provided in
71

Appendix A. Regression coefficients and t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference in
explanatory power between two non-nested models are reported in italics. The symbols *, **, and *** indicate z-statistics significant
at the 10%, 5%, and 1% levels, respectively. Variables are winsorized at the 1% and 99% levels.

72

TABLE 9

Relative Explanatory Power of Taxable Income and Cash Flows from Operations: Full Sample and by Discretionary Accruals
and Cash Effective Tax Rate Quintiles

MVE

IV

Quintile n TI CFO RATIO
MVE
z-stat.

TI CFO RATIO
IV
z-stat.

R
2
(1) R
2
(2) R
2
(1) / R
2
(2)

R
2
(3) R
2
(4)
R
2
(3) / R
2

(4)

Full Sample 14,147 67.81% 55.38% 122.44% 12.39***

39.71% 34.66% 114.57% 7.23***

Q1 (Low DACC) 2,759 74.02% 56.25% 131.59% 5.49***

42.58% 41.26% 103.20% 1.09

Q2 2,760 69.37% 53.90% 128.70% 5.11***

35.30% 31.89% 110.69% 3.38***

Q3 2,760 70.56% 65.68% 107.43% 3.56***

41.50% 38.41% 108.04% 2.50**

Q4 2,760 59.96% 67.34% 89.04% -2.57**

31.94% 43.57% 73.31% -4.09***

Q5 (High DACC) 2,759 54.21% 62.00% 87.44% -2.82***

31.74% 36.71% 86.46% -3.99***

Q1 (Low CETR) 1,491 50.23% 45.03% 111.55% 1.90*

22.56% 20.43% 110.43% 1.38

Q2 1,855 60.11% 66.62% 90.23% -1.77*

24.02% 42.37% 56.70% -6.43***

Q3 1,834 65.75% 58.25% 112.88% 2.12**

30.23% 46.14% 65.52% -4.39***

Q4 1,855 69.06% 63.29% 109.16% 1.71*

27.31% 39.51% 69.12% -3.12***

Q5 (High CETR) 1,597 68.23% 73.91% 92.31% -0.97

37.47% 33.51% 111.82% 1.88*

Notes: this table reports the explanatory power, or R
2
, of separate regressions of taxable income (TI) and pretax cash flows from
operations (CFO) on the market value of equity (MVE) and ex-post intrinsic value of equity (IV) for the full sample, within
73

discretionary accrual (DACC) quintiles, and within five-year cash effective tax rate (CETR) quintiles. CFO is calculated as cash flows
from operations plus cash taxes paid. Detailed definitions of other variables are provided in Appendix A. Regression coefficients and
t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference in explanatory power between two non-nested
models are reported in italics. The symbols *, **, and *** indicate z-statistics significant at the 10%, 5%, and 1% levels, respectively.
Variables are winsorized at the 1% and 99% levels.

74

TABLE 10

Relative Explanatory Power of Book Income and Cash Flows from Operations: Full Sample and by Discretionary Accruals
and Cash Effective Tax Rate Quintiles

MVE

IV

PERCENT CHANGE IN THE

SUPERIORITY OF BOOK INCOME

Quintile n PTBI CFO RATIO
MVE
z-stat.

PTBI CFO RATIO
IV
z-stat.

(RATIO
IV
- RATIO
MVE
) /

R
2
(1) R
2
(2) R
2
(1) / R
2
(2)

R
2
(3) R
2
(4) R
2
(3) / R
2
(4)

(RATIO
MVE
- 100%)

Full Sample 14,147 79.96% 55.38% 144.38% 15.27***

46.57% 34.66% 134.36% 11.12***

-22.58%

Q1 (Low DACC) 2,759 84.28% 56.25% 149.83% 7.41***

53.12% 41.26% 128.74% 4.99***

-42.32%
Q2 2,760 80.26% 53.90% 148.91% 7.21***

42.62% 31.89% 133.65% 5.67***

-31.20%
Q3 2,760 83.44% 65.68% 127.04% 5.72***

46.41% 38.41% 120.83% 4.42***

-22.97%
Q4 2,760 75.51% 67.34% 112.13% 2.49**

45.38% 43.57% 104.15% 1.36

-65.76%
Q5 (High DACC) 2,759 70.85% 62.00% 114.27% 2.77***

41.38% 36.71% 112.72% 3.54***

-10.88%

Q1 (Low CETR) 1,491 75.88% 45.03% 168.51% 6.25***

39.20% 20.43% 191.87% 9.78***

34.10%
Q2 1,855 88.65% 66.62% 133.07% 3.92***

55.66% 42.37% 131.37% 3.32***

-5.15%
Q3 1,834 87.92% 58.25% 150.94% 5.43***

55.52% 46.14% 120.33% 2.02**

-60.09%
Q4 1,855 70.63% 63.29% 111.60% 1.80*

41.63% 39.51% 105.37% 1.10

-53.73%
Q5 (High CETR) 1,597 75.73% 73.91% 102.46% 0.29

36.13% 33.51% 107.82% 1.66*

217.51%

Notes: this table reports the explanatory power, or R
2
, of separate regressions of book income (PTBI) and pretax cash flows from
operations (CFO) on the market value of equity (MVE) and ex-post intrinsic value of equity (IV) for the full sample, within
discretionary accrual (DACC) quintiles, and within five-year cash effective tax rate (CETR) quintiles. CFO is calculated as cash flows
75

from operations plus cash taxes paid. Detailed definitions of other variables are provided in Appendix A. Regression coefficients and
t-statistics are omitted for clarity. Results of Vuong (1989) tests of the difference in explanatory power between two non-nested
models are reported in italics. The symbols *, **, and *** indicate z-statistics significant at the 10%, 5%, and 1% levels, respectively.
Variables are winsorized at the 1% and 99% levels.

76

FIGURE 1

Graphical Illustration of Table 4

Panel A: Relative Abilities of Book Income (PTBI) and
Taxable Income (TI) to Explain the Market Value of Equity
(MVE)

Panel B: Relative Abilities of Book Income (PTBI) and
Taxable Income (TI) to Explain Ex-Post Intrinsic Value
(IV)

65.00%
66.00%
67.00%
68.00%
69.00%
70.00%
71.00%
72.00%
73.00%
74.00%
75.00%
PTBI TI
40.00%
41.00%
42.00%
43.00%
44.00%
45.00%
46.00%
47.00%
48.00%
49.00%
50.00%
PTBI TI
77

FIGURE 2

Graphical Illustration of Table 5

Panel A: Relative Abilities of Book Income (PTBI) and Taxable Income (TI) to Explain the Market Value of Equity (MVE), by
Discretionary Accruals (DACC) Quintile

Panel B: Relative Abilities of Book Income (PTBI) and Taxable Income (TI) to Explain Ex-Post Intrinsic Value (IV), by
Discretionary Accruals (DACC) Quintile

60.00%
65.00%
70.00%
75.00%
80.00%
Q1 (Low DACC) Q2 Q3 Q4 Q5 (High DACC)
PTBI TI
35.00%
40.00%
45.00%
50.00%
55.00%
Q1 (Low DACC) Q2 Q3 Q4 Q5 (High DACC)
PTBI TI
78

FIGURE 3

Graphical Illustration of Table 6

Panel A: Relative Abilities of Book Income (PTBI) and Taxable Income (TI) to Explain the Market Value of Equity (MVE), by
Cash Effective Tax Rate (CETR) Quintile

Panel B: Relative Abilities of Book Income (PTBI) and Taxable Income (TI) to Explain Ex-Post Intrinsic Value (IV), by Cash
Effective Tax Rate (CETR) Quintile

50.00%
55.00%
60.00%
65.00%
70.00%
75.00%
80.00%
Q1 (Low CETR) Q2 Q3 Q4 Q5 (High CETR)
PTBI TI
15.00%
20.00%
25.00%
30.00%
35.00%
40.00%
45.00%
Q1 (Low CETR) Q2 Q3 Q4 Q5 (High CETR)
PTBI TI
79

APPENDIX

Variable Definitions

______________________________________________________________________________

Variable Definition
______________________________________________________________________________

PTBI Book income, calculated as pretax book income less
minority interest.

TI Taxable income as estimated from firms’ financial
statement data, calculated as worldwide current tax expense
(the sum of current federal tax expense and current foreign
tax expense, or total tax expense less deferred tax expense
if either current federal or current foreign tax expense is
missing) grossed- up (divided) by the applicable top annual
U.S. statutory corporate tax rate, all less the change in tax
net operating loss carryforwards.

MVE Market value of equity, calculated as share price three
months after the end of the fiscal year multiplied by the
number of common shares outstanding on the same date.

IV Ex-post intrinsic value determined pursuant to the residual
income model expressed over a five-year finite time
horizon. Ex-post intrinsic value is calculated by replacing
expectations of future earnings, book value, and terminal
stock price in the residual income model with five years or
their ex-post realizations.

DACC The absolute value of discretionary accruals, calculated
using the Jones (1991) model of accruals as modified by
Dechow et al. (1995), run cross-sectionally by each
industry- (two digit SIC code) year with at least 10
observations, including an intercept and scaling all
variables by lagged total assets as in Kothari et al. (2005).
A higher value of DACC indicates relatively lower book
earnings quality.
______________________________________________________________________________
80

APPENDIX, CONT.

Variable Definitions

______________________________________________________________________________

Variable Definition
______________________________________________________________________________

CETR The five-year “long-run” cash effective tax rate developed
by Dyreng et al. (2008), calculated as the sum of cash taxes
paid in the current plus four prior years divided by the sum
of pretax book income (PTBI as defined above) less special
items in the current plus four prior years. A higher value of
CETR indicates that the firm engages in a relatively lower
degree of tax planning.
______________________________________________________________________________

Abstract (if available)

Abstract In this study, I document that at least a portion of the superior ability of book income relative to taxable income to explain the market value of equity may be due to market mispricing arising from investors’ fixation on book income and underemphasis on the information contained in taxable income, rather than book income’s superior information content. I find that this result generally intensifies as book earnings quality and tax planning decrease. Indeed, I show that once market mispricing is removed from the valuation model, taxable income possesses statistically equivalent or even superior ability relative to book income to explain firm value among firms with particularly low book earnings quality and firms that engage in a relatively low degree of tax planning. This study adds to the growing literature on the informativeness of firms’ tax-related financial statement disclosures by demonstrating that prior research may have conducted tests of value relevance that are inherently biased in favor of book income and, consequently, understated the relative information content of taxable income.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Voluntary disclosure responses to mandated disclosure: evidence from Australian corporate tax transparency

PDF

Increased access to capital: evaluation of the New Market Tax Credit Program in New York

PDF

Real asset liquidity and asset impairments

PDF

CEO reputation: who benefits -- the firm and the CEO?

PDF

Does mandatory adoption of international financial reporting standards decrease the voting premium for dual-class shares: theory and evidence

PDF

Three essays on macro and labor finance

PDF

Marketing strategies with superior information on consumer preferences

PDF

Share repurchases: how important is market timing?

PDF

The EITC, labor supply, and child development

PDF

Essays on the competition between new and used durable products

PDF

Does mandatory adoption of International Accounting Standards reduce the cost of equity capital?

PDF

Accrual quality and expected returns: the importance of controlling for cash flow shocks

PDF

The effects of a customer's comparative processing with positive and negative information on product choice

PDF

Internal capital markets and competitive threats

PDF

The role of accounting information in the sentiment-price relation

PDF

Essays on information and financial economics

PDF

Expectation dynamics and stock returns

PDF

“What you see is all there is”: The effects of media co‐coverage on investors’ peer selection

PDF

Essays on digital platforms

PDF

Gone with the big data: institutional lender demand for private information

Asset Metadata

Creator Stekelberg, James Michael (author)

Core Title Leveling the playing field: unbiased tests of the relative information content of book income and taxable income

School Leventhal School of Accounting

Degree Doctor of Philosophy

Degree Program Business Administration

Degree Conferral Date 2013-12

Publication Date 09/11/2013

Defense Date 08/20/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,tax,tax accounting,taxable income

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Trezevant, Robert (committee chair), Chuk, Elizabeth (committee member), Ridder, Geert (committee member), Swenson, Charles (committee member)

Creator Email jstekelberg@email.arizona.edu,stekelbe@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-326353

Unique identifier UC11294069

Identifier etd-Stekelberg-2031.pdf (filename),usctheses-c3-326353 (legacy record id)

Legacy Identifier etd-Stekelberg-2031.pdf

Dmrecord 326353

Document Type Dissertation

Format application/pdf (imt)

Rights Stekelberg, James M.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA