The Role of Accounting in the Design of CEO Equity Compensation

Author(s): Mary Ellen Carter, Luann J. Lynch and İrem Tuna

Source: The Accounting Review , Mar., 2007, Vol. 82, No. 2 (Mar., 2007), pp. 327-357
Published by: American Accounting Association

Stable URL: https://www.jstor.org/stable/30243469

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Accounting Association is collaborating with JSTOR to digitize, preserve and extend
access to The Accounting Review

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
THE ACCOUNTING REVIEW
Vol. 82, No. 2
2007
pp. 327-357

The Role of Accounting in the Design of

CEO Equity Compensation
Mary Ellen Carter
University of Pennsylvania

Luann J. Lynch
University of Virginia

Irem Tuna
University of Pennsylvania

ABSTRACT: We examine the role of accounting in CEO equity compensation design.

For a sample of ExecuComp firms in 1995-2001, we find that financial reporting con-
cerns are positively related to stock option use and total compensation, and negatively
related to the use of restricted stock. We confirm our findings by examining changes
in CEO compensation in firms that begin expensing options in 2002 or 2003. We find
that these firms reduce their option use and increase their restricted stock use after
starting to expense options but exhibit no decrease in total compensation. Taken to-
gether, our analyses suggest that favorable accounting treatment for options led to a
higher use of options and lower use of restricted stock than would have been the case
absent accounting considerations.

I. INTRODUCTION
n this paper, we examine the role of accounting in CEO equity compensati
ically, we investigate whether favorable accounting treatment for stock options t
available until very recently affected their use and the use of restricted stock. W
substantiate the role of accounting in equity compensation by examining whether fir
start expensing options shift CEOs' equity compensation away from options a
stricted stock.
Prior work models the choice of stock options and restricted stock (e.g., Lam
Larcker 2004; Oyer and Schaefer 2005; Hall and Murphy 2002; Feltham and
These models lead to different predictions about the preferable form of equity comp

We thank John Core, Dan Dhaliwal (editor), Joseph Gerakos, Tim Gray, Wayne Guay, Raffi Indje
Ittner, Dave Larcker, Francisco de Asis Martinez-Jerez, Tjomme Rusticus, two anonymous reviewer
participants at the 2005 American Accounting Association Management Accounting Conference, 20
Accounting Association Financial Accounting and Reporting Conference, Babson College, Boston C
University of Pennsylvania for their helpful comments. We gratefully acknowledge the financial s
Wharton School and the University of Virginia Darden School Foundation. We thank I/B/E/S for prov
forecast data.

Editor's note: This paper was accepted by Dan Dhaliwal.

Submitted April 2005
Accented August 2006

327

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 328 Carter, Lynch, and Tuna

based on the assumptions embedded in th

despite the prediction of some models th
options, empirical evidence suggests that
20 percent of ExecuComp firms granted
2001, while approximately 80 percent gr
Hall and Murphy (2002) suggest that on
models of the choice between options an
vorable accounting treatment for option
SFAS No. 123, Accounting for Stock-Bas
Board [FASB] 1995), firms could account
intrinsic value method prescribed by APB N
generally did not record compensation e
of expense that would have been recorde
firms expensed options under SFAS No.
doing so would decrease earnings. Indeed
firm may have been willing to incur re
save $1 in compensation expense.
However, prior literature is inconclusiv
options motivated their use. Dechow et
sition in response to FASB's 1993 Expo
resulted from concerns about the effec
Aboody et al. (2004a) find that firms vo
expense under SFAS No. 123 have signif
they find no significant relation between
expense after controlling for other fac
detect some evidence of a relation betw
Yermack (1995) and Bryan et al. (2000
specific settings suggests that accounting
contract terms (e.g., Carter and Lynch
for research directly examining the eff
2003); specifically, they say, "It is impo
accounting in motivating firms to eithe
the role of financial accounting for emp
to firms, but is not well understood by
The infrequent use of restricted stock
trary, the lack of consensus in the liter
accounting influences the use of options
No. 123(R) all make it important to ascer
compensation plan design.
Using a sample of ExecuComp firms an
that financial reporting concerns affect
CEO pay packages. We find that our prox
comprehensive than those of prior studie

Under the intrinsic value method, firms could avoid

with a fixed exercise price set at or above the pric
2 On March 31, 2004, the Financial Accounting St
new accounting for stock options. For fiscal years
for options using the fair-value method as defin
2004).

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 329

and negatively related to the use of restricted stock during a period when very few firms
were expensing stock options. This result suggests that the previously available favorable
accounting treatment for stock options has influenced equity compensation. In addition, w
find that our proxy for financial reporting concerns is positively related to total compen-
sation, suggesting that the once favorable accounting treatment for stock options may hav
lead to higher overall CEO pay.
We complement our findings by examining changes in CEO compensation in firms
upon their decision to expense options. This setting allows us to investigate the role of
accounting without having to rely on a proxy for those financial reporting concerns. A
decrease in the use of options once expensing began would be consistent with the favorable
accounting treatment encouraging their use. Further, it may explain the puzzling empirica
observations regarding the infrequent use of restricted stock during times of favorable ac
counting treatment for options and would support the assertion that accounting affects th
design of executive compensation.
Using a sample of ExecuComp firms that begin to expense stock options in 2002 and
2003, we find that their use of options in CEO compensation decreases upon expensing
them. We also find that they award more compensation in restricted stock relative to wha
they had granted previously. While we take steps to rule out the possibility that firms
decreased the use of options and then decided to expense them, our results are subject to
the caveat that we may not have completely ruled out this alternative explanation. Collec-
tively, our results are consistent with the favorable accounting treatment for options in the
pre-expensing period leading to an overweighting of options and an underweighting of
restricted stock in executive pay packages. Finally, we detect no decrease in total compen-
sation upon expensing. In combination with the positive association between financial re-
porting concerns and total CEO compensation in the pre-expensing period, this result sug
gests that firms find it difficult to downsize the large executive pay packages that may have
resulted from the favorable accounting treatment for stock options.
In response to the call for research on the impact of accounting on option use, our
results suggest that the method of accounting for options has affected decisions regarding
their use. We find that firms more concerned about earnings used more options in thei
equity compensation due to the favorable accounting treatment for options; once firms start
expensing stock options, they shift into restricted stock. Our analysis provides insight into
changes that may occur in CEO equity compensation now that the FASB has made option
expensing mandatory: while we may not see an overall decrease in CEO compensation, w
expect a decline in stock option use and an increase in the use of restricted stock. Our
results also help to reconcile the theoretical predictions regarding the use of restricted stock
with the empirical observation that few firms use restricted stock. Consistent with Hall and
Murphy (2002), our results suggest that models of firms' choices of equity compensation
methods should include the accounting considerations.
Section II discusses related literature. Section III presents the hypotheses. Section IV
examines the relation between financial reporting costs and the use of stock options befor
many firms began to expense stock options. Section V examines changes in CEO compen
sation upon firms expensing options. Finally, Section VI concludes.

II. BACKGROUND AND PRIOR LITERATURE

Models of the choice of stock options and restricted stock (e.g., Lambert a
2004; Oyer and Schaefer 2005; Hall and Murphy 2002; Feltham and Wu
different predictions about the preferable form of equity compensation base
sumptions embedded in the models and type of analysis employed. Specificall

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 330 Carter, Lynch, and Tuna

Schaefer (2005) show that options w

who are optimistic about the outlo
options more highly than restrict
of the options vary more with the
argue that the employee's risk aver
averse the employee, the larger the
the "risk cost." Consistent with t
erally dominate restricted stock i
options (equity with a non-zero ex
of the firm to the agent, thereby de
risk-averse employees and highly v
restricted stock as the only equity
tract if volatility further increases.
stock when it is substituted for ca
when they are granted in addition
that restricted stock is optimal when
and options are optimal when the
outcome.

Despite the predictions above that firms should prefer restricted stock under cer
conditions, empirical evidence suggests they rarely do so. This is potentially due to
reasons. First, it is possible that the conditions under which restricted stock is preferred a
seldom met. Alternatively, these models may be ignoring another factor that support
dominance of stock options over restricted stock. Hall and Murphy (2002) suggest tha
such missing, but key, factor in existing models of the choice between stock options
restricted stock is the previously favorable accounting treatment for stock options,
might have influenced their use. The accounting for stock options differed substan
from that for restricted stock. Firms granting restricted stock must record an expen
sociated with the grant.3 However, before SFAS No. 123(R), firms that granted stock option
were not required to record an expense if they granted a fixed number of stock op
with a fixed exercise price equal to or greater than the market price on the grant
Botosan and Plumlee (2001) report that for their sample of Fortune's September 1999 l
of the 100 fastest-growing companies, recording stock option expense would have decr
earnings per share (EPS) by 14 percent and return on assets (ROA) by 13.6 percent;
other words, if firms had expensed options, the effect would be substantial.
Many studies examine the use of stock options in compensation contracts for CE
executives, and non-executive employees (e.g., Core and Guay 2001; Ryan and Wiggi
2001; Bryan et al. 2000; Core and Guay 1999; Kole 1997; Yermack 1995; Gaver and Ga
1993; Smith and Watts 1992). Unlike stock options, prior research related to the us
restricted stock is limited, perhaps because of the low proportion of firms that incorp
restricted stock into the compensation plan. Kole (1997) and Gaver and Gaver (
examine the presence of plans to authorize the issuance of restricted stock, but can
examine actual grants of restricted stock due to data availability constraints.4 Bryan
(2000) and Ryan and Wiggins (2001) study actual restricted stock grants to CEOs du
1992-1997 and 1997, respectively. Ryan and Wiggins (2001) investigate the use of ca
bonuses, stock options, and restricted stock, but do not consider the impact of fin

The expense is equal to the value of the shares granted, amortized over the vesting period of the shares
4 Kole (1997) examines compensation plans in 1980; Gaver and Gaver (1993) examine compensation pl
1985. Both studies examined time periods prior to the required disclosure of restricted stock grants.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 331

reporting costs on compensation plan design. Bryan et al. (2000) examine various deter-
minants of equity grants, but find no evidence that financial reporting costs are related to
either options or restricted stock.
Despite the research that exists regarding the choice of stock options and restricted
stock, and despite the suggestion by Hall and Murphy (2002) that one important but often-
omitted factor in studies that model that choice is the favorable accounting for stock op-
tions, prior literature provides inconclusive evidence on whether the accounting for stock
options motivates their use. Dechow et al. (1996) find that opposition in response to FASB's
1993 Exposure Draft proposing the expensing of stock options resulted from concerns about
reporting higher levels of executive compensation, but find no systematic evidence that the
opposition resulted from concerns about the effect on earnings of recording the expense.
Aboody et al. (2004a) examine firms' decisions to voluntarily recognize stock-based com-
pensation expense under SFAS No. 123. They find that these decisions are related to the
extent of capital market participation, private incentives of executives and the board of
directors, the level of information asymmetry, and political costs. Despite finding that firms
voluntarily recognizing an expense have significantly lower SFAS No. 123 expense than
other firms, they find no significant relation between the decision to expense and the mag-
nitude of that expense after controlling for other factors. They do find significant positive
announcement returns for firms announcing their decision earlier, consistent with the an-
nouncement serving as a signal about reporting transparency and favorable future prospects.
Core and Guay (1999) find their proxy for financial reporting costs is positively related
to the use of options for CEOs. Matsunaga (1995) finds some evidence of a weak relation
between the use of options and financial reporting costs, although he points out that incon-
sistencies in his results across methods of estimation and time suggest the need for addi-
tional research. Kimbrough and Louis (2004) find that firms alter the proportion of com-
pensation from options to meet certain earnings benchmarks, particularly when they expect
to issue shares the following year. Other literature related to options suggests that firms
may be motivated by accounting considerations to alter the terms of option contracts (see,
e.g., Carter and Lynch 2003, 2005). However, Yermack (1995) and Bryan et al. (2000) do
not find a reliable relation between options and financial reporting costs. This mixed evi-
dence has led to a call for research on the effect that accounting standards have on the use
of stock options (Core et al. 2003).
Based on the results of prior work, the role of accounting in equity compensation is
unresolved. The lack of consensus in prior literature on whether the accounting for stock
options promoted their use might stem from the proxies for financial reporting concerns
used in these studies. Most proxies focus on (1) potential debt covenant violations, or (2)
the extent to which earnings meet prior year earnings.5
Our analysis sheds light on the role of accounting in equity compensation in two ways.
First, we develop a more comprehensive measure of firms' concerns about the effect of an
expense on earnings. It incorporates multiple reasons why a company could incur financial

5 Yermack (1995) uses interest coverage as a proxy for financial reporting costs, since firms with low interest
coverage may be closer to violating debt covenants. Matsunaga (1995) uses a measure of (1) the extent to which
"as if stock options were expensed" income is below a target level (a random walk with drift), because the firm
enters into agreements that are either implicitly or explicitly based on reported income, and (2) the extent to
which the firm uses income-increasing accounting methods. Bryan et al. (2000) use several measures similar to
the first measure in Matsunaga (1995) and an interest coverage measure. Core and Guay (1999) use, as a proxy
for financial reporting costs, whether retained earnings limit the firm's ability to pay dividends and repurchase
stock. Specifically, a firm is constrained if [(year-end retained earnings + cash dividends and stock repurchases
during the year)/prior year's cash dividends and stock repurchases] is less than 2.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 332 Carter, Lynch, and Tuna

reporting costs due to expensing stoc

research. In addition, our proxy reco
empirically, firms typically do not ex
year. Most of these prior studies men
packages substantially each year in respo
Our proxy captures a more general co
Second, in addition to using a more
conclusions by using the natural setti
pensing options. This setting allows u
compensation but eliminates the nee
concerns.

III. HYPOTHESIS DEVELOPMENT

The role that accounting plays in equity compensation is still an open d
would argue that the favorable accounting treatment did not affect firms'
stock options. In other words, firms provide details regarding their options
nancial statement footnotes (SFAS No. 123), and financial statement users ca
those details into their assessment of the firm (e.g., Aboody et al. 2004b).
would expect no association between perceived financial reporting costs from
compensation expense and the design of CEO equity compensation prior to fi
options. We also would expect no change in CEO equity compensation de
decision to expense options. On the other hand, financial statement users m
porate footnote details into their assessment of the firm, perhaps because th
on reported accounting earnings (Hand 1990) or because it is costly to adjust
details in the footnotes (see, e.g., Espahbodi et al. 2002). If managers believ
the case, then we would expect firms that perceive higher financial costs
compensation expense to rely more on options prior to expensing them. Furt
expect a decrease in the use of options upon the decision to expense them if
of equity compensation would have otherwise been preferable.
Theory provides mixed predictions on the role of restricted stock in eq
sation depending on the assumptions in the underlying models. Despite re
being optimal in certain settings, few firms use it in executive compensation
that restricted stock is optimal, but as suggested by Hall and Murphy (200
considerations not modeled actually overshadowed their use. If firms limite
restricted stock because of differences in the accounting for options and re
under SFAS No. 123, we would expect firms that perceived higher financia
porting compensation expense to have relied less on restricted stock as a f
compensation for CEOs. In addition, we would expect an increase in the us
stock upon the decision to expense options.
Finally, some have claimed that an increased use of stock options in res
vorable accounting treatment led to higher overall levels of executive com
e.g., Anderson et al. 2002). If so, then we would expect a positive associ
perceived financial reporting costs and total CEO compensation. In addition
total compensation upon firms' expensing options would suggest that the fav
ing treatment for options resulted in higher levels of compensation. Altern
crease in total compensation upon firms' expensing options may also be co
such allegations. If the favorable accounting led to an increase in executive
in a pre-expensing regime, but firms find it difficult to decrease hefty exec
afterwards, then they may substitute other forms of compensation for opti

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 333

decrease the overall level of compensation if other forms of compensation were preferable
but for the earnings effect.

IV. THE RELATION BETWEEN FINANCIAL REPORTING CONCERNS AND

EQUITY COMPENSATION PRIOR TO EXPENSING OPTIONS
In this section, we examine whether financial reporting concerns are associated with
equity compensation in 1995 to 2001, before a large number of firms began expensing
stock options. Specifically, we are interested in the extent to which the favorable accounting
treatment for stock options may have (1) motivated the use of options, (2) deterred the use
of restricted stock, and (3) led to higher overall executive compensation. We estimate sep-
arate regressions of stock options, restricted stock, and total compensation on proxies for
firms' financial reporting concerns and control variables that have been documented to
explain compensation design.

Proxies for Financial Reporting Concerns

Our proxies for financial reporting concerns expand on variables used in prior research
by encompassing several capital markets effects of earnings characteristics documented in
the literature. First, the literature has identified circumstances under which particular earn-
ings patterns or characteristics have positive effects on stock price (or negative effects when
those patterns or characteristics no longer exist). Two important circumstances are (1) the
need to maintain increasing earnings patterns (Barth et al. 1999; Burgstahler and Dichev
1997; Degeorge et al. 1999) and (2) pressure to meet analyst expectations (Burgstahler and
Eames 1999; Degeorge et al. 1999; Abarbanell and Lehavy 2003; Bartov et al. 2002;
Kasznik and McNichols 2001). Indeed, some of these studies suggest that firms enjoy a
stock market premium that increases in the length of the string of earnings increases or
meeting analyst expectations, making it more costly to break the string. In these circum-
stances, managers may act as if sustaining these patterns will prevent negative capital
markets effects.
The literature also has identified capital-raising and contracting circumstances that may
focus managers' attention on maintaining higher earnings. Indeed, Espahbodi et al. (2002)
find that the need for additional capital and the possibility of debt covenant violations are
related to concerns about expensing of stock options during the SFAS No. 123 debate.
Accordingly, we consider three additional circumstances that may lead to concerns about
reported earnings: (1) the need to meet debt covenants (Dichev and Skinner 2002; Watts
and Zimmerman 1986, 1990; Sweeney 1994; DeFond and Jiambalvo 1994), (2) the need
to access the equity markets (Teoh and Wong 1998; Teoh et al. 1998; Richardson et al.
2004), and (3) the need to issue debt (Anthony et al. 2004).
Since the use of stock options did not require the recording of compensation expense
in our sample period, firms concerned about these capital markets effects of lower earnings
may be more likely to use options in executive compensation. That is, firms with patterns
of earnings increases, pressure to meet analyst expectations or debt covenant requirements,
or plans to access capital markets may be revealing a higher general concern about the
level of earnings. As a result, we expect these firms to be more likely to use stock options.
Based on prior literature discussed above, our proxies for concerns about financial
reporting costs are: (1) the proportion of quarters (of all quarters that the firm appears on
I/B/E/S) that the firm's EPS was equal to or greater than the prior year same quarter
(EPS_INCR); (2) the proportion of quarters (of all quarters that the firm appears on
I/B/E/S) that the firm met or beat analysts' EPS forecasts (BEAT_FCST): (3) the firm's
ratio of debt to assets (LEVERAGE); (4) extent to which the firm accesses the equity markets

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
334 Carter, Lynch, and Tuna

in the upcoming year (ISSUE_EQ); a

markets in the upcoming year (ISSUE
Compustat and analyst forecasts from
Since each proxy measures a firm's o
error and because we are interested in
costs rather than in any one particu
characteristics using principal compo
error inherent in each individual com
in a comprehensive yet parsimonious
values greater than 1, and for each fa
with a value of greater than 0.45.7
INCR, BEAT_FCST, and LEVERAGE, w
associated with participating in the c
EQ and ISSUE_DEBT, which captures
to access the capital markets.8 If the
the use of options, deterred the use
compensation, then we would expect
and stock options and total compensat
We validate our proxies for a firm's
of-sample analysis of firms that have
cial reporting costs. Since under SFA
unvested options over the vesting per
tions and avoid an expense under the
1 and FINRPT_2 are greater for firms t
as these firms appear to have done so
No. 123(R) (Choudary et al. 2006). We
of options from Bear Steams Equit
Company, Inc. 2006). We calculate
data) for 2,192 firms on Compustat, o
mean (median) of FINRPTI is signific
than those firms that are not, at p <

6 We calculate LEVERAGE as [(Compustat #

ISSUEEQ as [(the increase from year t to yea
- Compustat #88)/Compustat #6]. If this cal
We calculate ISSUE_DEBT as [(the increase f
Compustat #6]. If this calculation yields a neg
information from Compustat to proxy for equ
provide a more direct measure of equity and
losing a significant portion of our sample
DEBT) using balance sheet information and
mation for firms for which we have both is
7 Interestingly, LEVERAGE has a negative load
violating debt covenants. Rather, it is consisten
being concerned about meeting the expectati
we convert that variable to the proportion of a
8 For the first factor, loadings for EPSINCR,
For the second factor, loadings for ISSUE_E
9 We also estimate the regressions using th
coefficients (5 coefficients in each regression
coefficients are significant at least at p < 0.10
ponents clouds our ability to draw inferences w
we use principal components analysis to addr
common underlying construct represented by

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 335

greater for firms accelerating vesting than those firms that are not, though the difference is
not significant (p < 0.14 and p < 0.75). Since firms began accelerating in 2004 and we
measure future access to capital markets (FINRPT_2) in 2003, the lack of significance may
result from our inability to measure this variable in 2005. Overall, this analysis provide
some support that FINRPT_I and FINRPT_2 are capturing firms' overall concerns about
financial reporting costs.

Sample
Our sample consists of 6,242 CEO-year observations from ExecuComp for 1995
through 2001.10 Table 1 presents descriptive statistics for the sample. Table 1, Panel A
describes our sample selection. Out of 11,693 CEO-year observations with compensation
data over the 1995-2001 period, 3,460 lack data to measure our financial reporting cost
variable, and an additional 1,991 observations lack data for the control variables.
We obtain compensation data from ExecuComp. We obtain compensation from stock
options (ExecuComp variable BLK_VALUE), compensation from restricted stock
(ExecuComp variable RSTKGRNT), and total compensation (ExecuComp variable TDC1)
for all executives identified by ExecuComp as the CEO." We obtain financial statement
data from Compustat.
The use of stock options increased steadily throughout the sample period (Table 1,
Panel B). Specifically, the percent of sample firms granting options to CEOs increased from
76.5 percent in 1995 to 82.3 percent in 2001. Firms in the sample used very little restricted
stock compared with options. However, the use of restricted stock to compensate CEOs
increased steadily throughout the study period, from 18.0 percent of firms in 1995 to 21.6
percent of firms in 2001 (Table 1, Panel B). Panel C of Table 1 shows that the average
firm in our sample awards $3.1 million and $0.4 million in options and restricted stock,
respectively.

Multivariate Analysis
We estimate the following regression using pooled data for 1995 to 2001:12

DEP_VBLjt = oo + aOtFINRPT_Ijt + 0o2FINRPT_2jt + O3DEV_INCt

+ o4CASH_CONSTRj, + oaDIV_YLDjt + L6EARN_VOL,
+ O7EQ_CONSTRjt + oa8TENUREjt, + oLLNASSETt
+ atoBOOKMKTj, + atiRETj, + o121npreDEPVBLjt + Ejt (1)

where:

Dependent variables:
In_OPTj, - natural log of value of stock options granted to CEO of firm jin year t;
In_RSTKj, = natural log of value of restricted stock granted to CEO of firm j in year t; and

10 This time period allows us to examine the relation between financial reporting costs and stock options after
SFAS No. 123 was issued and before a large number of firms began expensing stock options.
" We eliminate part-year executives because compensation in those years for those executives may not be repre-
sentative of annual compensation. For example, new executives frequently get hire-on equity grants (Doubleday
and Fujii 2001).
12 To mitigate the influence of outlying observations, we winsorize values DEV_INC, CASH_CONSTR, DIV_YLD,
EARN_VOL, EQCONSTR, TENURE, BOOKMKT, and RET that are below (above) the 1st (99th) percentile.
Because of censoring in the equity grant data, we estimate Equation (1) using Tobit when ln_OPT and In_
RSTK are the dependent variables.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 336 Carter, Lynch, and Tuna

TABLE 1
Descriptive Statistics

Panel A: Sample Selection

Number of CEO-year observations in 1995-2001 with compensation data 11,693
Less: CEO-year observations missing data to measure financial reporting concerns 3,460
Less: CEO-year observations missing control variables 1,991
Final sample 6,242

Panel B: Proportion of F
Percentage of Firms Granting 1995 1996 1997 1998 1999 2000 2001
Stock options to CEOs 76.5% 76.9% 76.8% 80.6% 80.8% 82.6% 82.3%
Restricted stock to CEOs 18.0% 18.8% 18.6% 19.9% 18.9% 19.4% 21.6%
Both options and restricted stock to CEOs 15.7% 16.1% 16.3% 17.9%
n 783 853 858 890 925 919 1,014

Panel C: CEO Compensation and Fin

Variable Mean Std. Dev. Q1 Median Q3
OPT ($) $3,078,648 $13,006,230 $119,676 $743,737 $2,365,379
RSTK ($) $432,694 $8,410,645 $0 $0 $0
FINRPT_ 1:
EPSINCR 0.735 0.167 0.619 0.735 0.869
BEAT_FCST 0.625 0.147 0.524 0.613 0.724
LEVERAGE 0.768 0.175 0.652 0.776 0.918

FINRPT_2:
ISSUE_EQ 0.046 0.138 0.000 0.004 0.024
ISSUE_DEBT 0.059 0.126 0.000 0.001 0.061
DEV_INC 0.064 1.086 -0.615 0.036 0.713
CASH_CONSTR 0.023 0.088 -0.024 0.012 0.057
DIV YIELD 0.013 0.016 0 0.006 0.021
EARN_ VOL 0.008 0.027 0.000 0.001 0.004
EQ_ CONSTR 0.134 0.152 0.048 0.090 0.163
TENURE 7.794 7.458 2 5 11
LNASSET 7.396 1.617 6.206 7.228 8.427
BOOK_MKT 0.470 0.365 0.232 0.391 0.606
RET 0.180 0.539 -0.149 0.107 0.392

Variable Definitions:

OPTj, = value of stock options granted to CEO of firm j in year t;

RSTKj, = value of restricted stock granted to CEO of firm j in year t;
FINRPT_Ijt = factor created from principal component analysis equally weighting standardized values of
EPS_INCR, BEAT_FCST, and LEVERAGE for firm j in year t;
EPS_INCRj, = proportion of quarters on I/B/E/S, as of the end of year t, that firm j has reported increase
in EPS over the same quarter in the prior year;
BEAT_FCSTj, = proportion of quarters on I/B/E/S, as of the end of year t, that firm j has beat analysts'
EPS forecast;
LEVERAGEj, = (1 - debt/assets) for firm j in year t (see footnote 7);
FINRPT_2j, = factor created from principal component analysis equally weighting standardized values of
ISSUE_EQ and ISSUEDEBT for firm j in year t;
ISSUE_EQj, = future increase in equity capital for firm j (year t+ 1) scaled by assets, 0 otherwise;
ISSUE_DEBTJ, = future increase in debt capital for firm j (year t+ 1) scaled by assets, 0 otherwise;
DEV_INCj, = In (actual incentive level/predicted incentive level) for year t-1, where actual incentive
level is the delta of the equity portfolio and predicted incentive level is estimated from a
model based on Core and Guay (1999) for the CEO in firm j;

(continued on next page)

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 337

TABLE 1 (continued)

CASHCONSTR, = the three-year average over year t-3 to t-1 of [(Common and preferred dividends - cash
flow from investing - cash flow from operations)/total assets] for firm j;
DIV_ YLDj, = the three-year average over year t-3 to t-1 of [dividends per share/price per share at the
end of the year t] for firm j;
EARN_VOL, = the square of the standard deviation of ROA, where the standard deviation of ROA is
calculated over ten years prior to year t for firm j;
EQ_CONSTRj, = (executive options outstanding at the end of the year t-1/the three-year average over year
t-3 to t-1 of percent of total options granted to executives)/total shares outstanding at the
end of year t- 1 for firm j;
TENUREj, = the number of years the CEO has been in that position (if missing, the number of years at
the firm) in firm j as of the end of year t;
LNASSETj, = natural log of total assets at the end of year t for firm j;
BOOK_MKT,, = book value of equity/market value of equity at the end of year t for firm j; and
RETj, = cumulative 12-month returns for year t for firm j.

ln_TCj, = natural log of total compensation to CEO of firm j in year t.

Independent variables:
FINRPT_ l, = factor created from principal component analysis equally weighting
standardized values of EPS_INCR, BEATFCST, and LEVERAGE for
firm j in year t;
FINRPT_2I, = factor created from principal component analysis equally weighting
standardized values of ISSUE_EQ and ISSUE_DEBT for firm j in
year t;
DEV_INC, = In (actual incentive level/predicted incentive level) for year t-l,
where actual incentive level is the delta of the equity portfolio and
predicted incentive level is estimated from a model based on Core and
Guay (1999) for the CEO of firm j;
CASH_CONSTRJ, = the three-year average over year t-3 to t-1 of [(Common and pre-
ferred dividends - cash flow from investing - cash flow from
operations)/total assets] for firm j;
DIV_YLDj, = three-year average over year t-3 to t-1 of [dividends per share/price
per share at the end of the year t] for firm j;
EARN_VOLi, = square of the standard deviation of ROA, where the standard deviation
of ROA is calculated over ten years prior to year t for firm j;
EQ_CONSTRj = (executive options outstanding at the end of the year t-1/the three-
year average over year t-3 to t-1 of percent of total options granted
to executives)/total shares outstanding for firm j;
TENUREj, = the number of years the CEO has been in that position (if missing,
the number of years at the firm) for firm j as of the end of year t;
LNASSETj, = natural log of total assets for firm j at the end of year t;
BOOK_MKT, = book value of equity/market value of equity at the end of year t for
firm j;
RET, = cumulative 12-month returns for year t for firm j;
ln_pre_OPTj = natural log of value of stock options granted to CEO of firm j in year
t- l;
In_pre_RSTK), = natural log of value of restricted stock granted to CEO of firm j in
year t-1l; and
In_pre_TCj, = natural log of total compensation to CEO of firm jin year t-1.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 338 Carter, Lynch, and Tuna

To test our hypotheses, we examine

FINRPT_I and FINRPT_2.

Control Variables
We include control variables that are expected to influence CEO compensation. We
include control variables for the choice of equity versus non-equity compensation (deviation
from optimal incentives, cash constraints, equity constraints, risk aversion), variables that
control for the form of the equity compensation used (dividend yield, volatility), and var-
iables capturing standard determinants of compensation (size, investment opportunities, and
performance).13
Deviation from predicted equity incentive levels (DEV_INC). Because firms grant
equity to align the interests of executives with those of shareholders, changes in the CEO's
equity portfolio from selling shares would change the incentive alignment and may require
the firm to grant equity to adjust the overall level of equity incentives. We use the proxy
developed in Core and Guay (1999) to control for equity grants that may result from the
firm's need to adjust equity incentives. If the incentives inherent in an executive's equity
portfolio are above (below) the predicted incentive level, then a lower (greater) use of equity
grants is expected. Accordingly, we expect a negative relation between the deviation from
this predicted equity incentive level and compensation from restricted stock and stock
options.
We measure the deviation of the CEO's equity incentive levels from its predicted level
(DEV_INC) as In(actual incentive level/predicted incentive level) following the procedure
in Core and Guay (1999).14 That is, the actual incentive level is measured as the delta of
the CEO's equity portfolio. The predicted level is determined from estimating a model of
the level of equity incentives as a function of firm size, firm risk, growth opportunities,
length of CEO employment, and free cash flow, including industry and yearly indicator
variables.
Cash constraints (CASH_CONSTR). The use of restricted stock or stock options as
compensation requires no cash outlay. Firms experiencing a shortage of cash may use equity
as a substitute for cash compensation (Yermack 1995; Dechow et al. 1996). Accordingly,
we expect a positive relation between cash shortfall and both restricted stock and stock
options. Consistent with Core and Guay (1999), we measure cash constraints (CASH_
CONSTR) as the three-year average of [(Common and preferred dividends - cash flow
from investing - cash flow from operations)/total assets] so that a larger number represents
a greater cash shortfall.

13 It is possible that some of the variables determining the choice of equity versus non-equity compensation have
varying degrees of importance in explaining the choice between options and restricted stock. For example, we
might see more use of restricted stock than options when firms are more cash constrained because restricted
stock is seen as a better substitute for cash compensation (Hall and Murphy 2002). If firms are more equity-
constrained, we might see more restricted stock because the firm can commit less shares of restricted stock than
options to convey the equivalent value to the executive. If executives are more risk-averse, then we might see
more restricted stock because it always provides some positive intrinsic value to the executive (i.e., offers
protection on the downside). Since the focus of our paper is on the role of accounting in equity compensation,
we do not develop formal hypotheses around these control variables but acknowledge that they may represent
trade-offs firms make.
14 We do make a few modifications worth mentioning. First, our measure of the risk-free rate is from ExecuComp,
and we make no attempt to match that measure against the maturity of the options. Second, in eliminating new
option grants from executive equity portfolios, we first take those out of unexercisable options. If that results
in a negative number, then we take all new option grants out of exercisable options. Despite these simplifying
modifications, we are able to replicate the results in Core and Guay (1999) using their sample and time period.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 339

Proximity to constraints on issuing equity (EQ_CONSTR). Firms that are close to

their constraint on issuing equity may be more likely to offer cash compensation than equity
compensation. Accordingly, we expect a negative relation between proximity to constraints
on issuing equity and the use of stock options and restricted stock.
Ideally, we would measure the proximity to constraints on issuing equity as the ratio
of issued to authorized shares. However, that information is not available in electronic form.
Thus, we measure proximity to constraints on issuing equity as an estimate of the total
options outstanding/total shares outstanding at the end of the year prior to the year of
interest. We proxy for total options outstanding by dividing executive options outstanding
[ExecuComp variable UXNUMEX + ExecuComp variable UEXNUMUN] at the end of th
year prior to the year of interest by the average in the prior three years of percent of total
options granted to executives [ExecuComp variable PCTTOTOPT].
Risk aversion of executives (TENURE). Executives who are more risk-averse may
prefer the certainty of fixed compensation over the uncertainty of performance-based com-
pensation. Accordingly, we expect a negative relation between the risk aversion of execu-
tives and the use of options and restricted stock.
We proxy for risk aversion using the length of time the CEO has held that position.
Individuals in the same position longer likely feel more stable and secure and therefore are
likely to be less risk-averse since the likelihood of CEO turnover decreases with tenure
(Sebora 1996; Farber 1999; Allgood and Farrell 2000, 2003). Accordingly, we expect
positive relation between TENURE and both stock options and restricted stock.
We measure TENURE as the number of years the CEO has held that position (calculated
from ExecuComp variable BECAMECEO), or if missing, the number of years an executiv
has been with the firm (calculated from ExecuComp variable JOINED_CO).
Dividend Yield (DIV_YLD). Stock options typically are not dividend protected (div-
idends do not accrue to the option holder), making options less valuable to the executive
Firms that pay high dividends may be less likely to use options as a form of compensation
Accordingly, we expect a negative relation between dividend yield and the use of stock
options. Restricted stock typically is dividend protected (dividends do accrue to the stock-
holder), making restricted stock more valuable to the executive in dividend-paying firms.
Firms that pay high dividends may be more likely to use restricted stock as a form of
compensation. Accordingly, we expect a positive relation between dividend yield and the
use of restricted stock.
Our proxy is the average dividend yield over the three-year period ending the year
prior to the year of interest, where dividend yield (ExecuComp variable DIVYIELD) is
measured as dividends per share/price per share at the end of the year. We use a three-
year average because it likely better captures future expectations than using than just one
(the prior) year data.
Volatility (EARN_VOL). Higher volatility firms may be more likely to grant restricted
stock because restricted stock offers the holder some downside protection. Alternatively,
higher volatility firms may be less likely to grant restricted stock because the likelihood of
stock options being in the money is greater with higher volatility firms. Accordingly, we
make no prediction regarding the relation between volatility and either options or restricted
stock.
We measure volatility as the variance of ROA, where the variance of ROA is calculated
over ten years prior to the year of interest using data from Compustat.
Standard economic determinants of compensation: size, investment opportunities,
performance. First, as firm size increases, it may become more difficult to monitor man-
agers' actions (see Smith and Watts 1992). If so, then larger firms may be more likely to

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 340 Carter; Lynch, and Tuna

use incentive compensation plans.

and both stock options and restri
(Compustat #6) at the end of the
Second, since it is more difficult
investment opportunities, firms wit
equity compensation to link manag
We include book-to-market ratio a
pustat #216 - Compustat #130)/(C
opportunities and expect a negativ
restricted stock.
Third, Murphy (1985) finds that
firm performance. Accordingly, w
and options, restricted stock, and
performance as cumulative monthl
We include industry indicator var
al. (1998).'5 This allows us to captu
tries and thus control for attracti
(Oyer and Schaefer 2005). Furtherm
differences in pay practices across
We include the previous year's o
In_pre_RSTK) or the previous ye
determinants of compensation desi
Finally, we include yearly indicato
White standard errors to adjust fo
the data.

Results

Table 2, Panel A presents the results of the regression of the natural logarithm of
compensation from stock options on our independent variables. The amount of compen-
sation from stock options is positively related to both FINRPT_I and FINRPT_2 (p < 0.01
and p < 0.05, respectively), as expected. Compensation from stock options is greater for
larger firms, higher growth firms, and firms with better performance, and smaller for firms
with CEO equity incentives above predicted levels and firms with higher dividend yield, as
expected. The sign of the coefficient on TENURE is contrary to predictions. It is possible
that being in a position longer signals that an executive is risk-averse, so firms rely less
heavily on variable compensation.'7 The coefficient on EQ_CONSTR also is contrary to
predictions but likely reflects that our proxy for proximity to equity constraints is higher,
by construction, for those firms that grant more options. In general, though, the results in
Panel A are consistent with our hypothesis and suggest that firms more concerned about

15 Barth et al. (1998) classify firms into 15 industry groupings according to four-digit SIC codes.
16 Our conclusions are unchanged if this variable is excluded.
17 It is also possible that longer tenure may indicate a CEO with greater wealth and thus greater risk aversion,
suggesting a need for more pay, which firms offer in the form of equity compensation. Alternatively, it is
possible that more talented CEOs with longer tenure demand higher equity-based pay. We thank an anonymous
reviewer for suggesting these two alternative explanations.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 341

TABLE 2
Regressions of the Level of CEO Compensation from Stock Options, Restricted Stock, and
Total Compensation in 1995 through 2001 on Concern about Financial Reporting Costs
Other Factors

Variable Predicted Sign Coefficient z-statistic

Panel A: Natural Logarithm of CEO Compensation from St

lnOPTj, = a + aotFINRPT_1j, + a2FINRPT_2j, + ot3DEV

+ OaLDIVYLDj, + o6EARN_VOL., + ao7EQ_CONST

+ oaLNASSETt, + aoLBOOK_MKTjt + ao lRETjt + o,21n_pre_OPTjt + Fjt"
Intercept ? 0.45 1.09
FINRPT_I + 0.23 2.94***
FINRPT_2 + 0.11 2.05**
DEVINC -0.45 -7.63***
CASH_CONSTR + 0.06 0.10
DIV_YLD - -13.70 - 3.44***
EARN_ VOL ? 1.36 0.73
EQ_CONSTR 1.99 4.53***
TENURE + -0.06 -7.11***
LNASSET + 0.48 12.01***
BOOK_MKT -0.50 -3.18***
RET + 0.38 4.26***
In_preOPT + 0.38 15.41***

n 6,242
Wald X2 statistic (p-value) 1,243.61 (0.00)
Panel B: Natural Logarithm of CEO Compensation from Restricted

In_RSTKjt, = a + atFINRPT_lj, + o2FINRPT_2t, + a3DEV_INCjt

+ asDIV_YLDj, + oQ6EARN_VOLj, + a7EQ_CONSTRj, +
+ aoLLNASSETt + atoBOOKMKTj, + aiRETj, + a12ln-pr
Intercept ? -11.56 -9.21**
FINRPT_I - -0.62 -2.54***
FINRPT_2 0.14 0.80
DEVINC -0.43 -2.85***
CASH_CONSTR + -2.46 -1.29
DIV_YLD + 32.83 2.66***
EARN_VOL ? -11.21 - 1.43
EQ_CONSTR - - 1.09 -0.87
TENURE + -0.11 -4.77***
LNASSET + 0.50 4.60***
BOOK_MKT 0.09 0.21
RET + 0.84 2.68***
In_pre_RSTK + 1.58 33.72***

n 6,242
Wald x2 statistic (p-value) 2,347.84 (0.00)

(continued on next page)

The Accounting Review, March 2

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 342 Carter, Lynch, and Tuna

TABLE 2 (continued)

Variable Predicted Sign Coefficient z-statistic

Panel C: Natural Logarithm of Total CEO Compensation

In_TCj, = ao + aiFINRPT_lj, + c 2FINRPT_2j, + t3D

+ a5DIV_YLDj, + 0L6EARN_VOLj, + oL7EQ_CONS

+ otgLNASSET,, + o~oBOOK_MKT,, + allRET,, + o

Intercept ? 2.35 16.44***
FINRPT_I + 0.06 3.86***
FINRPT_2 + 0.04 2.50**
DEV_INC ? -0.05 -4.37***
CASH-CONSTR ? -0.26 - 1.84*
DIV_YLD ? -3.11 -3.89***
EARN_VOL ? 1.68 3.90***
EQCONSTR ? 0.80 7.02***
TENURE ? -0.00 -2.60**
LNASSET + 0.26 17.49***
BOOK_MKT ? -0.23 -7.43***
RET + 0.21 8.38***

ln_preTC + 0.48 18.42***

n 6,242
Adjusted R2 0.58

***, **, * Significant at 1 percent, 5

Coefficients on yearly and industry
standard errors.
Variable Definitions:
InOPTj, = natural log of value of stock options granted to CEO of firm j in year t;
ln_RSTKj, = natural log of value of restricted stock granted to CEO of firm j in year t;
In_TCj, = natural log of total compensation to CEO of firm j in year t;
FINRPTljt = factor created from principal component analysis equally weighting standardized values of
EPS_INCR, BEAT_FCST, and LEVERAGE for firm j in year t;
EPS_INCRj, = proportion of quarters on I/B/E/S, as of the end of year t, that firm j has reported increase
in EPS over the same quarter in the prior year;
BEAT_FCSTj, = proportion of quarters on I/B/E/S, as of the end of year t, that firm j has beat analysts'
EPS forecast;
LEVERAGEj, = (1 - debt/assets) for firm j in year t (see footnote 7);
FINRPT_2I, = factor created from principal component analysis equally weighting standardized values of
ISSUE_EQ and ISSUE_DEBT for firm j in year t;
ISSUE_EQJ, = future increase in equity capital for firm j (year t+ 1) scaled by assets, 0 otherwise;
ISSUE_DEBTjt = future increase in debt capital for firm j (year t+ 1) scaled by assets, 0 otherwise;
DEV_INCj, = In (actual incentive level/predicted incentive level) for year t-1, where actual incentive
level is the delta of the equity portfolio and predicted incentive level is estimated from a
model based on Core and Guay (1999) for the CEO in firm j;
CASH_CONSTRJ, = the three-year average over year t-3 to t-1 of [(Common and preferred dividends - cash
flow from investing - cash flow from operations) / total assets] for firm j;
DIVYLDj, = the three-year average over year t-3 to t-1 of [dividends per share/price per share at the
end of the year t] for firm j;
EARNVOL, = the square of the standard deviation of ROA, where the standard deviation of ROA is
calculated over ten years prior to year t for firm j;
EQ_CONSTRJ, = (executive options outstanding at the end of the year t-l1/the three-year average over year
t-3 to t-1 of percent of total options granted to executives)/total shares outstanding at the
end of year t- 1 for firm j;

(continued on next page)

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 343

TABLE 2 (continued)

TENUREj, = the number of years the CEO has been in that position (if missing, the number of years at
the firm) in firm j as of the end of year t;
LNASSETjt = natural log of total assets at the end of year t for firm j;
BOOK_MKTt = book value of equity/market value of equity at the end of year t for firm j;
RETj, = cumulative 12-month returns for year t for firm j;
In_pre_OPTj, = natural log of value of stock options granted to CEO of firm j in year t-1;
In_pre_RSTK, = natural log of value of restricted stock granted to CEO of firm j in year t-1; and
In_pre_TCj, = natural log of total compensation to CEO of firm j in year t-1.

financial reporting costs paid more option-based compensation, avoiding recognition of

compensation expense.'8
Table 2, Panel B presents the results of the regression of the natural logarithm of
compensation from restricted stock on our independent variables. The relation between th
use of restricted stock and FINRPTI is negative and significant (p < 0.01), as expected,
suggesting that firms with greater financial reporting concerns use less restricted stock
Unexpectedly, FINRPT_2 is not related to the use of restricted stock. That the coefficien
on FINRPTI is negative and significant as expected but the coefficient on FINRPT_2 is
not related to the use of restricted stock suggests that it is firms' overall concerns about
financial reporting costs (captured by FINRPTI) rather than the intent to access the capita
markets (captured by FINRPT2) that are important in their decisions to grant restricte
stock. It is also consistent with firms' compensation decisions about restricted stock bein
made over longer-term horizons (a stickiness in the restricted stock decision) rather than
in response to the decision to access capital markets.
Larger firms and firms with better performance use more restricted stock; firms wit
CEO equity incentives above predicted levels use less restricted stock. Further, firm
with higher dividend yields use more restricted stock, as expected. Also, the positive relation
between dividend yield and restricted stock is consistent with expectations, as options ar
not dividend protected, while restricted stock is. There is no significant relation between
cash constraints and restricted stock, similar to what we see with options. This is inconsis
tent with Hall and Murphy's (2002) suggestion that firms would prefer restricted stock
when it is a substitute for cash compensation. Again, similar to Table 2, Panel A, the sign
of the coefficient on TENURE is contrary to predictions, possibly because being in the jo
longer signals that an executive is risk-averse, so firms rely less heavily on variab
compensation.
Table 2, Panel C presents the results of the regression of the natural logarithm of tota
compensation on our independent variables. The relation between total compensation an
both FINRPTI and FINRPT_2 is positive and significant (p < 0.01 and p < 0.05, re
spectively). While we cannot draw definite conclusions regarding this relation from th
regressions in Panel C, it is consistent with the favorable accounting treatment for stock
options resulting in higher executive pay packages. We will explore this further in Sec
tion V.

18 It is possible that exercisable and unexercisable options provide different performance incentives, particular
with regard to CEO turnover, and thus may affect the granting of options. While this is likely captured by the
industry indicator variables, we include the proportion of options held by the CEO that are unexercisable, an
our results are unchanged. We thank an anonymous reviewer for this suggestion.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 344 Carter, Lynch, and Tuna

Robustness Tests
We conduct several additional tests to assess the robustness of our results to alternative
specifications (results untabulated). First, to mitigate possible concerns that equity grants
may be influenced by managers' expectations about future prospects of the firm, we also
estimate Equation (1) including future stock returns proxied by year t+ 1 cumulative 12-
month stock returns, and our conclusions are unchanged.
Second, it is possible that our proxy for financial reporting concerns is correlated with
prior performance and that it is prior performance, not financial reporting concerns, that
leads to increases in option grants. To address this possibility, we estimate Equation (1)
including first, average return on assets in years t-3 to t-1 and, alternatively, cumulative
stock returns from year t-3 to year t-1 to capture prior performance. Our conclusions are
unchanged.
Third, we estimate Equation (1) including as the dependent variable and prior year's
compensation the proportion of total compensation from stock options and restricted stock,
and our conclusions are unchanged. We also estimate Equation (1) scaling the dependent
variable and prior year's compensation by sales. Our conclusions from the options and
restricted stock regressions are unchanged. The results from the total compensation regres-
sion differ slightly from those in Table 2. While our conclusions on FINRPT_2 are con-
sistent, the coefficient on FINRPT_I is of the predicted sign, but not significant at conven-
tional levels, inconsistent with the results in Table 2. However, we also estimate Equation
(1) with total compensation scaled by assets as an alternative size scalar, and results are
consistent with those in Table 2. Overall, our conclusions are unchanged.
Collectively, our results suggest that the favorable accounting for stock options affected
equity compensation. Even after controlling for standard determinants of compensation and
other factors that may influence the choice between stock options and restricted stock,
concerns about reported earnings have significant explanatory power when examining the
amount of compensation from options and restricted stock.

V. EXAMINATION OF CHANGES IN COMPENSATION BY FIRMS

EXPENSING STOCK OPTIONS
Results from Section IV suggest that accounting played a role in firms' choi
compensation method. To substantiate this finding, we examine, for firms t
expense stock options, changes in CEO compensation from the average of th
before expensing to the year of and year after first expensing.19 By eliminating t
reporting benefits of stock options, these firms no longer have the ability to avoid
an expense with any form of equity compensation. Using this sample of firms
examine the role of financial reporting concerns without having to rely on a proxy
concerns. We interpret the results from our tests here as evidence that expen
a reduced use of stock options. It should be noted, though, that an alternative
that firms reduce their use of options and then decide to expense them. We in
possibility in "Robustness Tests and Additional Analysis" subsection.

Sample Selection
From Bear Stearns Equity Research dated December 16, 2004 (Bear Stearns
pany, Inc. 2004), we obtain a sample of 824 firms that began to expense stoc
1995 to 2004. According to Bear Steamrns, almost 50 percent of these firms ha

19 An alternative sample could have been firms expensing options after the effective date of S
However, proxy data for the fiscal years subject to the new accounting rule are not yet available.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 345

capitalization greater than $1 billion, and 35 percent are international firms.20 We focus ou
analysis on the 206 firms that began to expense stock options in either 2002 or 2003 and
that have both ExecuComp and Compustat data in the year of their first option expensing.21
Table 3 provides a description of those firms, as well as firms not expensing options. Thirty-
two percent of firms expensing options are financial institutions, 10 percent are utilities,
and 23 percent are in manufacturing industries (food, textiles, chemicals, durable manufac
turers, and computers).22

Univariate Analysis
Changes in Compensation in Firms Expensing Options
Table 4, Panel A presents univariate statistics regarding compensation levels both before
and after expensing for our sample of 206 firms that expense stock options.23 The mean
(median) decrease from the period before expensing in compensation from stock options
is $1,430.2 ($356.8) thousand in the year of first expensing and $1,709.2 ($589.4) thousand
in the year after first expensing (all significant at p < 0.01). The mean (median) proportion
of CEO compensation from options decreases from 46.5 percent (44.2 percent) to 38.
percent (37.1 percent) in the year of first expensing and further to 32.4 percent (30.3
percent) in the year after first expensing (all significant at p < 0.01). In addition, the percent
of firms granting options declines significantly (p < 0.01) from 88.7 percent to 68.9 percen
(64.3 percent) in the year of (after) first expensing. These changes are consistent with firm
shifting away from options upon deciding to expense them.
The proportion of our sample firms granting restricted stock increases significantly from
42.8 percent to 48.1 percent in the year of first expensing (p < 0.10) and to 55.0 percent
in the year after first expensing (p < 0.05). In addition, the level of compensation from
restricted stock in those firms rises significantly. The mean (median) increase from the
period before expensing in compensation from restricted stock is $508.8 ($313.1) thousand
in the year of first expensing and $1,093.4 ($969.3) thousand in the year after first expensing
(all significant at either p < 0.01 or p < 0.05). In addition, the mean (median) proportion
of compensation from restricted stock increases from 19.4 percent (18.6 percent) to 27.7
percent (25.2 percent) in the year of expensing and 28.9 percent (26.4 percent) in the year
after expensing (all significant at p < 0.01). This increase may reflect a shift from options
toward restricted stock in providing longer-term performance incentives, suggesting tha
restricted stock was previously underweighted in equity compensation. Finally, firms ex-
pensing options show no significant change in total compensation.

20 Firms on the Bear Steams list are larger and more profitable than other firms on Compustat. This is consistent
with either (1) expensing firms being bigger, more profitable firms, or (2) Bear Steams' limiting its coverage to
bigger, more profitable firms. Since ExecuComp limits itself to larger, more profitable firms as well, we believe
that the Bear Steams list identifies virtually every firm on ExecuComp that expenses stock options. To the extent
that it does not, it would bias against our finding results. In addition, to control for any differences in industry
between the Bear Steams firms and other firms, we include in our regression indicator variables to control for
industry.
21 We focus on 2002 and 2003 because it is in those years that a substantial number of firms began to expense
options (see Table 3). We do not include firms beginning to expense options in 2004 because sufficient executive
compensation data are not yet available for those firms.
22 The relatively large proportion of expensing firms being financial institutions could reflect the August 2002
announcement by the Financial Services Forum of a list of its members that had agreed to expense options (see
Schrand 2004). Accordingly, we control for industry in our multivariate analysis.
23 Specifically, we compare the value for each variable in the year t and year t+ 1, separately, to the average value
for each variable over years t-2 and t-1, where year t is the year the firm first expenses options. A firm had
to award stock options and restricted stock in at least one of t-1 or t-2 to be included in the calculation for
that compensation component.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
346 Carter, Lynch, and Tuna

TABLE 3
Comparison of 206 Firms Beginning to Expense Stock Options in 2002 or 2003 and 1,48
Firms Not Expensing Stock Options in 2002 or 2003

Panel A: Firms Beginning to Expense Stock Options By Year

Number of Firms
Number of Firms Beginning to Expense
Beginning to Expense Stock Options with
Stock Options (on Data on ExecuComp
Year Bear Stearns Report) and Compustat
1995 2 2
1996 2 1
1997 0 0
1998 3 1
1999 2 0
2000 2 1
2001 7 2
2002, 145 43
2003a 537 163
2004 124 20

Total 824 233

Panel B: By-Ind
or 2003 and 1
Number of Firms Number of Firms not
Expensing on Expensing on
ExecuComp and ExecuComp and
Industry Descriptionb Compustat % Compustat %
Mining and construction 7 3 29 2
Food 4 2 37 3

Textiles, printing 16 8 82 6
Chemicals 7 4 44 3
Pharmaceuticals 1 0 55 4
Extractive industries 13 6 48 3
Durable manufacturers 15 7 333 22

Computers 3 2 114 8
Transportation 12 6 72 5
Utilities 20 10 75 5
Retail 17 8 178 12
Financial institutions 66 32 138 9
Insurance and real estate 16 8 20 1
Services 7 3 251 17
Other 2 1 7 0

Total 206 100% 1,483 100%

a In our analysis, we use the 206 fir

both ExecuComp and Compustat data
b Industry classifications based on B

The Accounting Review, March

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 345

(contiuedxpag)

TABLE4

(4.2%)371-58*0
morfegnahCtYB (861.3),40*79
(18.6%)254*73

DoNtExpensSckOi erofBgnispxE egnispxE (2,376.0)15-8*94 (4,80.6)17952

tahTsmriF384,1dn02opOkcSexEgB6fC)M(
"302ronistpOkcSexEgBmF6faCh)dM(:AlP Levlofcmpnsatir($0)4,7.132-*59 Proptinfcmesa46.5%38-*21 %ofirmsgantp8.769-1*432 Levlofcmpnsatirdk($0)1,79.24358* Proptinfcmesadk19.4%270*8 %ofirmsgantecdk42.8167*50 Totalcmpensi($0)7,93.14-68*52

TheAcountigRvw,Mar207

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 346 Carter, Lynch, and Tuna

citsa-)z(

TABLE4(contiued)

)26.3(*%1-54870
ecnrfiDomgahCtY B
AlenaPmorfBgispxE )68.1(*9-473,25 )04.2-(637,5*819

)29.-(*5461083$kcots )37.1-(*%542860kcots
dnalevhT.4023si"gpxErtfY;odna1evh02f .doireptahnkcs/gmfylu)verda(otcisnpmf
bsnoitpOkcSexENDahTmrF384,1fCg)dM(:BlP 63.1*"470,-8529)$(snoitpmrfaeclvL 26.3*%7-54890snoitpmrfaecP 93.2*%1-5784snoitpgarmf 43.2-*97516,0detcirsmofnaplvL 58.1-*%74290detcirsmofnapP 70.2-*%13584kcotsdeirgnamf 17.0-*954,2863)$(noitasepmclT .)snaidemcrfotukyhW-M( g,lvpsenc01datr5ifS* aTomitgehnflucrs,wiozethnuvabls1prctd9en.Th"BfoExpsigclumatedhvr
/snoitpmrfaecdlvhT.xyw,1-tna2sreobicfulv .doireptahnkcs/gmfylu)evitprdna(kcos
bTomitgaehnflucrs,wzv1pd9."BEx

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 349

Comparison to Changes in Compensation in Firms Not Expensing Options

The changes in CEO compensation in our sample of firms expensing options can stem
from these firms' decisions to expense options or from changes in other economic factor
that affect compensation. To control for macroeconomic factors, we compare the change
in CEO compensation in our sample to changes in CEO compensation over the same time
period in a sample of firms not expensing options. That sample comprises 1,483 firms
included on both ExecuComp and Compustat but not included on the Bear Steams list.
Table 4, Panel B presents univariate statistics regarding changes in compensation fo
our control sample of 1,483 firms not expensing options. Control firms also decrease com
pensation from options and the proportion of compensation from options, and there is
decline in the proportion of control firms granting options over the study period.24 However,
these changes are significantly smaller than changes exhibited by firms expensing options
Control firms increase pay from restricted stock, but this increase is significantly lower
than that by expensing firms (the mean [median] increase in compensation from restricte
stock is significantly lower than that for expensing firms in the year after, at p < 0.05
< 0.01]).25 In addition, the proportion of control firms granting restricted stock decreas
from 27.2 percent to 24.7 percent in the year of expensing, compared to an increase amon
the sample firms (p < 0.10), although that proportion does increase in the year after. In
total, these differences offer significant evidence of a shift by firms toward using restricted
stock upon deciding to expense options.
The change in total compensation for expensing firms is not significant at conventiona
levels (with the exception of the mean decrease in the year of expensing, which is significant
at only p < 0.10). Other than the median increase in the year after first expensing, there
is no difference between expensing and control firms in changes in total compensation
Combined with results related to stock options and restricted stock, these results are con
sistent with firms shifting from stock options to restricted stock and maintaining pr
expensing compensation levels.

Multivariate Analysis
In this section, we examine whether our inferences drawn from Table 4 hold after
controlling for both general economic trends and other factors that affect compensation
We estimate the following regression for the year of and year after first expensing:26

DEP_VBLj, = + ol1EXPENSER, + OL2DEV_INCt + oL3CASH_CONSTRt

+ a4DIV_YLDj + a5EARN_VOLi, + ct6EQCONSTRjt
+ a7TENUREij + aLNASSET7, + 9oLBOOKMKTt +
+ aaI"lnpre-DEP-VBLj + Ejt (2)
24 The decrease in reliance on options by non-expensing firms may reflect an expectation on beh
that option expensing will become mandatory and, as a result, a move to decrease reliance o
25 It is important to note that we understate the increase in pay from restricted stock for our s
our calculation requires firms to be granting restricted stock in years t-2 or t-1 and samp
significant increase in the number of firms granting restricted stock in year t. Use of restric
that begin granting in year t is excluded from our calculations. This understatement does not
firms since there is a decline in the number of firms granting restricted stock.
26 To mitigate the influence of outlying observations, we winsorize the current and lagged value
variables and values of DEVINC, CASH_CONSTR, DIV_YLD, EARN_VOL, EQ_CON
BOOK_MKT, and RET that are below (above) the 1st (99th) percentile. Because of censoring in
data, we estimate Equation (2) using Tobit when In_OPT and In_RSTK are the dependent var

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 350 Carter, Lynch, and Tuna

where:

Dependent variables:
ln_OPTj, = natural log of value of
t+ 1;
In_RSTKj, = natural log of value of restricted stock granted to CEO for firm j in year t or
t+1; and
ln_TCj, = natural log of total compensation for CEO for firm j in year t or t+ 1.

Independent variables:
EXPENSERj = 1 if firm j expenses options in 2002 or 2003, 0 otherwise;
In_pre_OPTj, = natural log of average value of stock options granted to CEO for firm j
in years t-2 and t-1;
ln_pre_RSTKj, = natural log of average value of restricted stock granted to CEO for firm
j in years t-2 and t-1; and
In_pre_TCj, = natural log of average total compensation for CEO for firm j in years
t-2 and t-1.

All other variables are as defined in Section IV.

We include industry indicator variables in each regression and estimate significance
using Huber/White standard errors to adjust for any potential heteroscedasticity and serial
correlation in the data.
The coefficient on EXPENSER captures the differences in the dependent variable be-
tween firms expensing options and firms not expensing options, after controlling for other
factors. If firms shift away from stock options after expensing options, then we would
expect a negative relation between compensation from stock options and EXPENSER. A
positive relation between compensation from restricted stock and EXPENSER would sug-
gest that firms shift into restricted stock once they begin expensing options and would be
consistent with firms having substituted options for restricted stock prior to expensing op-
tions. A negative relation between EXPENSER and total compensation would suggest that
favorable accounting treatment for stock options in a pre-expensing regime led to higher
levels of overall compensation and that expensing firms are adjusting compensation down-
ward. While no relation between EXPENSER and total compensation would be consistent
with the favorable accounting treatment from options having no effect on overall levels of
executive compensation, it is also consistent with it having led to higher levels but firms
finding it difficult to downsize executive pay packages.

Results

Table 5, Panels A, B, and C present the results of estimating Equation (2). These results
confirm inferences drawn from the univariate statistics in Table 4. Specifically, upon ex-
pensing options, firms pay less option compensation to CEOs (the coefficient on
EXPENSER in Panel A is negative and significant at p < 0.01). These firms increase their
use of restricted stock significantly more than do firms not expensing options (the coefficient
on EXPENSER in Panel B is positive and significant at p < 0.01). The shift away from
stock options upon expensing them confirms that accounting mattered in the decision to
grant options. That the shift into restricted stock accompanies the shift away from options
is consistent with an underweighting of restricted stock in CEOs' pay packages under the

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 351

TABLE 5
Regressions of the Level of CEO Stock Options, Restricted Stock, and Total Compensat
in the Year of and Year after First Expensing on an Indicator of the Expensing Decision
and Other Factors

Variable Predicted Sign Coefficient t-statistic

Panel A: Natural Logarithm of CEO Compensation from Sto

lnOPTj, = ao + oaEXPENSERj + o2DEV_INCj, + ot3CASH_

+ a5EARN_VOLt + oa6EQ_CONSTRjt + a7TENUREj,
+ a9BOOKMKT, + aotoRETjt + aOl1ln_preOPTJ, + EF

Intercept ? 0.34 0.47

EXPENSER -1.17 -3.15***
DEV_INC -0.15 - 1.69*
CASH_ CONSTR + 0.12 0.11
DIV_ YLD - -19.00 -2.64***
EARNVOL ? -2.71 -1.07
EQ_CONSTR 2.58 4.29***
TENURE + -0.06 -4.40***
LNASSET + 0.54 8.48***
BOOK_MKT -1.13 -4.47***
RET + -0.23 -1.59

ln_pre_OPT + 0.29 10.81"***

n 4,017
Wald X2 statistic (p-value) 482.68 (0.00)

Panel B: Natural Logarithm of CEO Compensation from Restrict

In_RSTKj, = a + a1EXPENSERj + t2DEV_INCj, + a3CASH_CO

+ ts5EARN _VOLjt + a6EQCONSTRjt + o7TENUREj, + ao8LNASSET,j

+ a9BOOKMKTj, + a oRETj, + atl1lnpre-RSTKjt + ej,.

Intercept ? -6.03 -3.58**

EXPENSER + 1.98 2.81***
DEVINC -0.15 -0.78

CASH_ CONSTR + -0.36 -0.12

DIVYLD + 38.80 2.51***
EARN VOL ? -2.10 -0.30
EQ_CONSTR -2.04 -1.31
TENURE + -0.14 -3.99***
LNASSET + 0.70 4.71***
BOOK_MKT -0.55 -0.99
RET + 0.17 0.45
ln_pre_RSTK + 0.56 20.16***

n 4,017
Wald x2 statistic (p-value) 947.15 (0.00)

(continued on next pag

The Accounting Review, Mar

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 352 Carter, Lynch, and Tuna

TABLE 5 (continued)

Variable Predicted Sign Coefficient t-statistic

Panel C: Natural Logarithm of Total CEO Compensation

In_TCj, = ao + otaEXPENSERj + o2DEV_INC, + ot3CA

+ a5EARN_VOLj, + at6EQ_CONSTR,, + 7TENUR
+ aotBOOK_MKTj, + oaoRETj, + a,,1ln_pre_TCjt +

Intercept ? 2.17 6.14***

EXPENSER 0 or - 0.04 0.87
DEVINC ? -0.07 -2.05**
CASH_ CONSTR ? -0.79 - 1.84*
DIV_YLD ? 0.16 0.08
EARN_VOL ? -0.85 -1.15
EQCONSTR ? 0.45 3.08***
TENURE ? -0.00 -0.66
LNASSET + 0.20 4.87***
BOOK_MKT ? -0.34 -2.60**
RET + 0.04 1.10

In_preTC + 0.55 7.63***

n 4,016
Adj. R2 0.50

***, **, * Significant at 1 p

Coefficients on industry in
errors.

Variable Definitions:

ln_OPTj, = natural log of value of stock options granted to CEO for firm j in year t or t+1;
ln_RSTKj, = natural log of value of restricted stock granted to CEO for firm j in year t or t+1;
In_TCj, = natural log of total compensation to CEO for firm j in year t or t+ 1;
EXPENSERj = 1 if firm j expenses options in 2002 or 2003, 0 otherwise;
DEV_INCj, In (actual incentive level/predicted incentive level) at the beginning of year t for firm j,
where actual incentive level is the delta of the equity portfolio and predicted incentive level
is estimated from a model based on Core and Guay (1999) for the CEO in firm j;
CASH_CONSTR, = the three-year average over year t-3 to t-1 of [(Common and preferred dividends - cash
flow from investing - cash flow from operations) /total assets] for firm j;
DIV_YLDj, = the three-year average over year t-3 to t-1 of [dividends per share/price per share at the
end of the year t];
EARN_VOLj, = the square of the standard deviation of ROA, where the standard deviation of ROA is
calculated over ten years prior to year t for firm j;
EQ_CONSTRJ, = (executive options outstanding at the end of the year t-1/the three-year average over year
t-3 to t-1 of percent of total options granted to executives)/total shares outstanding at the
end of year t- 1 for firm j;
TENUREj, = the number of years the CEO has been in that position (if missing, the number of years at
the firm) in firm j as of the end of year t;
LNASSETj, = natural log of total assets at the end of year t for firm j;
BOOK_MKTj, = book value of equity/market value of equity at the end of year t for firm j;
RETj, = cumulative 12-month returns for year t for firm j;
ln_pre_OPTj, = natural log of average value of stock options granted to CEO of firm j in years t-2 and
t- 1;
In_pre_RSTKj, = natural log of average value of restricted stock granted to CEO of firm j in years t-2 and
t- 1; and
In_pre_TCj, = natural log of average total compensation to CEO of firm j in years t-2 and t-1.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 353

previous regime.27 Our results suggest that firms likely have been willing to forgo some
benefits of restricted stock, such as those discussed in Section II, to get the favorable
accounting treatment of stock options.
We find no evidence of a change in total compensation (the coefficient on EXPENSER
in Table 5, Panel C is not significant at conventional levels). In combination with results
from Section IV, this result is consistent with the favorable accounting treatment for options
leading to higher levels of overall compensation and firms now finding it difficult to down-
size executive pay packages.28

Robustness Tests and Additional Analysis

We conduct several tests to assess the robustness of our results to alternative specifi-
cations. First, we estimate Equation (2) including as the dependent variable the change in
compensation from stock options, restricted stock, and total compensation from the two
years before to the year of and year after first expensing.29 The coefficient on EXPENSER
in our options regression is negative and significant at p < 0.05. The coefficient on
EXPENSER in our restricted stock regression is positive and significant at p < 0.05. The
coefficient on EXPENSER in our total compensation regression is positive and significan
at p < 0.05. Accordingly, our conclusions that firms shift away from options and into
restricted stock and that expensing does not result in a decrease in total compensation are
unchanged.
Second, we estimate Equation (2) including as the dependent variable and prior year's
compensation the proportion of total compensation from stock options and restricted stock
and our conclusions are unchanged. We also estimate Equation (2) scaling the dependent
variable (options, restricted stock, and total compensation) and prior year's compensation
by sales, and our conclusions are unchanged.
Finally, the voluntary nature of the expensing decision presents some challenges in
interpreting our results in this section. Specifically, it is possible that firms' choices about
expensing result in a selection bias that impacts our conclusions regarding a reduction in
options and an increase in restricted stock upon expensing options. We consider the use of
an instrumental variables approach to address this potential issue. However, concerns about
the instrumental variables approach raised by Larcker and Rusticus (2005) seem to apply
to our study. Specifically, we model the decision to expense options in the first stage using
Equation (1) in Aboody et al. (2004a). Our logit estimation has a Pseudo R2 of 0.20. W
then consider whether this model provides good instruments for our second stage analysis
as suggested in Larcker and Rusticus (2005). First, we obtain the residuals from the second
stage regressions of each of our dependent variables and regress them on all exogenous
variables. An indication of good instruments would be coefficients on the instruments tha

27 We also estimate Equation (2) with the natural log of salary and the natural log of bonus as dependent variables
(results not tabulated). We find no evidence of an increase in the use of salary or bonus upon expensing options.
These statistics provide no evidence that the shift away from options after expensing is accompanied by an
increased reliance on cash compensation.
28 We thank an anonymous reviewer for suggesting that possible concurrent changes in the corporate governance
environment may be contributing to our findings. To explore this possibility, we examine one aspect of corporate
governance, whether there are differences in the independence of the compensation committee (given that we
looking at executive compensation) between firms that expense and firms that do not expense both before and
after 2002. We find no difference in the mean or median percent of independent directors on the compensation
committee between expensers and firms not expensing options either before or after 2002. This suggests tha
changes in the independence of the compensation committee are not affecting the decision to alter compensation
in response to expensing options.
29 In this specification, we exclude from our independent variables the prior year's compensation. We also exclude
TENURE because in a changes specification, all observations have a value of 1 for this variable.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 354 Carter, Lynch, and Tuna

are close to zero. However, our tests indicate

We also examine the "unconstrained" secon
replace the first stage variables with the produ
the variable's original value. An indication
the instruments that do not differ much from
differ from zero and are of different signs. T
choose, we could get either a positive or a
option compensation. We conclude from thi
quality, suggesting that 2SLS in our contex
does OLS. Therefore, we rely on our result
Our inability to control for the voluntary
possibility of an alternative explanation for
pense options and as a result to decrease the
decrease the use of options and as a result
potential alternative explanation, we do tw
by expensing firms before the decision to e
proportion of firms granting options in the
suggest the opposite for our sample of fir
firms that granted options increased steadil
percent in 1999; 83.5 percent in 2000; and 8
statements for our sample of firms that expen
equity compensation plans. We obtain proxy
sample for the year in which they begin t
percent) indicate they are making changes t
firms that are decreasing their use of opti
restricted stock, and 32 that are doing both
that these changes resulted from their dec
these data as further support that the previ
has played a role in equity compensation an
deciding to expense them. However, this int
not have completely ruled out the alternativ

VI. CONCLUSION
We provide evidence that accounting does affect equity compensation usin
of ExecuComp firms and data from 1995 to 2001. Our proxy for firms' conce
financial reporting costs is positively related to the use of stock options and
related to the use of restricted stock, consistent with the favorable treatmen
options having led to an overweighting of options and an underweighting of restrict
in CEO compensation packages. In addition, we find that our proxy for financi
concerns is positively related to total compensation, consistent with the former
accounting treatment for stock options possibly leading to higher overall CEO
As further tests of the role of accounting in equity compensation, we examin
firms that expense stock options alter CEO equity compensation packages in r
the decision to expense options. Using a sample of firms that began to expense st
in 2002 and 2003, we examine changes in the structure of CEO pay packages c
with and after the decision to expense options. By eliminating the financial reportin
of stock options, firms expensing stock options no longer have an ability to avo

30 Other reasons firms provide include the desire to retain executives and avoid dilution from stock o

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 355

expenses with any form of equity compensation. Using this sample, we are able to test our
hypotheses without having to rely on a proxy for firms' financial reporting concerns.
Our findings confirm the role of accounting in equity compensation design. We find
that firms expensing options decrease compensation from options and increase compensa-
tion from restricted stock, even after controlling for standard economic determinants of
compensation and general economic trends. However, our results are subject to the caveat
that we may not have fully ruled out the alternative explanation that firms decreased the
use of options and then decided to expense them. We find no evidence of a decrease in
total compensation, possibly suggesting that the favorable accounting treatment for stock
options did not lead to higher levels of executive compensation. However, in combination
with the positive association between financial reporting concerns and total CEO compen-
sation in the pre-expensing period, this result suggests that firms find it difficult to downsize
the large executive pay packages that resulted from the favorable accounting treatment for
stock options.
That firms expensing stock options are granting fewer options and more restricted stock
suggests that firms shift toward restricted stock to provide longer-term performance incen-
tives and that there will likely be changes in CEO compensation now that SFAS No. 123(R)
is in effect. Though firms may have appeared to favor options, under a regime of mandatory
expensing, the role of options in executive compensation may be restricted. Overall, our
results support the assertion that accounting plays a role in executive compensation plan
design.

REFERENCES
Abarbanell, J., and R. Lehavy. 2003. Can stock recommendations predict earnings mana
analyst forecast errors? Journal of Accounting Research 41 (March): 1-31.
Aboody, D., M. Barth, and R. Kasznik. 2004a. Firms' voluntary recognition of stock-bas
sation expense. Journal of Accounting Research 42 (May): 123-150.
- , - , and - . 2004b. SFAS No. 123 stock-based compensation expense and equity
market values. The Accounting Review 79 (April): 251-275.
Allgood, S., and K. Farrell. 2000. The impact of tenure on the firm performance-CEO-turnover re-
lation. Journal of Financial Research 23 (Fall): 373-390.
, and . 2003. The match between CEO and firm. Journal of Business 76 (April): 317-
341.

Anderson, S., J. Cavanagh, C. Hartman, S. Klinger, and H. Sklar. 2002. Excessive Excess 2002.
Washington, D.C.: Institute for Policy Studies.
Anthony, J. H., B. Bettinghaus, and D. B. Farber. 2004. Earnings management and the market response
to convertible debt issuance. Working paper, Michigan State University, East Lansing, MI.
Barth, M., W. Beaver, and W. Landsman. 1998. Relative valuation roles of equity book value and net
income as a function of financial health. Journal of Accounting and Economics 25 (February):
1-34.

, J. Elliot, and M. Finn. 1999. Market rewards associated with patterns of increasing earnings.
Journal of Accounting Research 37 (Autumn): 387-413.
Bartov, E., D. Givoly, and C. Hayn. 2002. The rewards to meeting or beating earnings expectations.
Journal of Accounting and Economics 33 (June): 173-204.
Bear Steams & Company, Inc. 2004. FASB Does It: FAS 123(R) Requires Stock Option Expensing.
December 16. New York, NY: Bear Steams.
. 2006. Employee Stock Options: 749 Companies Accelerate Stock Options' Vesting. January
6. New York, NY: Bear Steams.
Botosan, C., and M. Plumlee. 2001. Stock options expense: The sword of Damocles revealed. Ac-
counting Horizons 15 (December): 311-327.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
 356 Carter, Lynch, and Tuna

Bryan, S., L. Hwang, and S. Lilien. 2000. CEO sto

incentive-intensity, relative mix, and economic
661-693.
Burgstahler, D., and I. Dichev. 1997. Earnings management to avoid earnings decreases and losses.
Journal of Accounting and Economics 24 (December): 99-126.

a-, and M. Eames. 1999. Management of earnings and analyst forecasts. Working paper, Uni-
versity of Washington, Seattle, WA.
Carter, M. E., and L. J. Lynch. 2003. The consequences of the FASB's 1998 proposal on accounting
for stock option repricing. Journal of Accounting and Economics 35 (April): 51-72.
- , and - . 2005. Agency issues in compensation contract design: Evidence from the change
in accounting for stock option repricings. Working paper, University of Pennsylvania, Phila-
delphia, PA.
Choudary, P., S. Rajgopal, and M. Venkatachalam. 2006. Accelerated vesting of employee stock
options in anticipation of FAS 123-R. Working paper, Duke University, Durham, NC.
Core, J., and W. Guay. 1999. The use of equity grants to manage optimal incentive levels. Journal
of Accounting and Economics 28 (December): 151-184.
, and 1. 2001. Stock option plans for non-executive employees. Journal of Financial
Economics 61 (August): 253-287.
, -, and D. Larcker. 2003. Executive equity compensation and incentives: A survey.
Economic Policy Review (April).
Dechow, P., A. Hutton, and R. Sloan. 1996. The economic consequences of accounting for stock-
based compensation. Journal of Accounting Research 34 (Supplement): 1-20.
DeFond, M. L., and J. Jiambalvo. 1994. Debt covenant violation and manipulation of accruals. Journal
of Accounting and Economics 17 (January): 145-176.
Degeorge, F., J. Patel, and R. Zeckhauser. 1999. Earnings management to exceed thresholds. Journal
of Business 72 (January): 1-33.
Dichev, I. D., and D. J. Skinner. 2002. Large-sample evidence on the debt covenant hypothesis.
Journal of Accounting Research 40 (September): 1091-1123.
Doubleday, D., and S. Fujii. 2001. Weighing your options, equity strategies in a volatile market.
WorldatWork Journal (Second Quarter).
Espahbodi, H., P. Espahbodi, A. Rezaee, and H. Tehranian. 2002. Stock price reaction and value
relevance of recognition versus disclosure: The case of stock-based compensation. Journal of
Accounting and Economics 33 (August): 343-373.
Farber, H. S. 1999. Mobility and stability: The dynamics of job change in labor markets. In Handbook
of Labor Economics, Vol. 3, edited by O. Ashenfelter, and D. Card. New York, NY: Elsevier
Science.
Feltham, G., and M. Wu. 2001. Incentive efficiency of stock versus options. Review of Accounting
Studies 6 (March): 7-28.
Financial Accounting Standards Board (FASB). 1995. Accounting for Stock-Based Compensation.
Statement of Financial Accounting Standards No. 123. Norwalk, CT: FASB.
1. 2004. Share-Based Payment. Statement of Financial Accounting Standards No. 123(R). Nor-
walk, CT: FASB.
Gaver, J., and K. Gaver. 1993. Additional evidence on the association between the investment op-
portunity set and corporate financing, dividend, and compensation policies. Journal of Account-
ing and Economics 16 (January-July): 125-160.
Hall, B., and K. Murphy. 2002. Stock options for undiversified executives. Journal of Accounting and
Economics 33 (February): 3-42.
Hand, J. 1990. A test of the extended functional fixation hypothesis. The Accounting Review 65
(October): 740-763.
Kasznik, R., and M. McNichols. 2002. Does meeting earnings expectations matter? Evidence from
analyst forecast revisions and share prices. Journal of Accounting Research 40 (June): 727-
759.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms
The Role of Accounting in the Design of CEO Equity Compensation 357

Kimbrough, M., and H. Louis. 2004. Financial reporting incentives, corporate governance, and the
trade-off among alternative forms of executive compensation. Working paper, Harvard Univer-
sity, Boston, MA.
Kole, S. 1997. The complexity of compensation contracts. Journal of Financial Economics 43 (Jan-
uary): 79-104.
Lambert, R., and D. Larcker. 2004. Stock options, restricted stock, and incentives. Working paper,
University of Pennsylvania, Philadelphia, PA.
Larcker, D., and T. Rusticus. 2005. On the use of instrumental variables in accounting research.
Working paper, Stanford University, Stanford, CA.
Matsunaga, S. 1995. The effect of financial reporting costs on the use of employee stock options. The
Accounting Review 70 (January): 1-26.
Murphy, K. 1985. Corporate performance and managerial remuneration. Journal of Accounting and
Economics 7 (April): 11-42.
Oyer, P., and S. Schaefer. 2005. Why do some firms give stock options to all employees? An empirical
examination of alternative theories. Journal of Financial Economics 76 (April): 99-133.
, and . 2006. Costs of broad-based stock option plans. Journal of Financial Interme-
diation 15 (4): 511-534.
Richardson, S., S. H. Teoh, and P. Wysocki. 2004. The walk-down to beatable analysts' forecasts:
The role of equity issuance and insider trading incentives. Contemporary Accounting Research
21 (December): 885-924.
Ryan, H., and R. Wiggins. 2001. The influence of firm- and manager-specific characteristics on the
structure of executive compensation. Journal of Corporate Finance 7 (June): 101-123.
Sebora, T. C. 1996. CEO-board relationship evaluation: An exploratory investigation of the influence
of base rate factors. Journal of Managerial Issues 8 (Spring): 54-77.
Schrand, C. 2004. Discussion of firms' voluntary recognition of stock-based compensation expense.
Journal of Accounting Research 42 (May): 151-158.
Smith, C., and R. Watts. 1992. The investment opportunity set and corporate financing, dividend, and
compensation policies. Journal of Financial Economics 32 (December): 263-292.
Sweeney, A. 1994. Debt covenant violations and managers' accounting responses. Journal of Ac-
counting and Economics 17 (May): 281-308.
Teoh, S. H., and T. J. Wong. 1998. Earnings management and the long-run market performance of
initial public offerings. Journal of Finance 53 (December): 1935-1974.
-- , I. Welch, and T. J. Wong. 1998. Earnings management and the underperformance of seasoned
equity offerings. Journal of Financial Economics 50 (October): 63-99.
Watts, R., and J. Zimmerman. 1986. Positive Accounting Theory. Englewood Cliffs, NJ: Prentice Hall.
., and - . 1990. Positive accounting theory: A ten-year perspective. The Accounting Review
65 (January): 131-56.
Yermack, D. 1995. Do corporations award CEO stock options effectively? Journal of Financial Ec-
onomics 39 (October-November): 237-269.

The Accounting Review, March 2007

This content downloaded from

5.186.125.237 on Thu, 04 Feb 2021 09:39:35 UTC
All use subject to https://about.jstor.org/terms