Is There a Link Between Executive--pak Gudono

Accounting Research Center, Booth School of Business, University of Chicago
Is There a Link between Executive Equity Incentives and Accounting Fraud?
Author(s): Merle Erickson, Michelle Hanlon and Edward L. Maydew
Source: Journal of Accounting Research, Vol. 44, No. 1 (Mar., 2006), pp. 113-143
Published by: Wiley on behalf of Accounting Research Center, Booth School of Business,
University of Chicago
Stable URL: http://www.jstor.org/stable/3542319
Accessed: 22-03-2016 04:24 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
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].

Accounting Research Center, Booth School of Business, University of Chicago and Wiley are collaborating with
JSTOR to digitize, preserve and extend access to Journal of Accounting Research.

http://www.jstor.org

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

DOI: 10.1111/j.1475-679X.2006.00194.x

Journal of Accounting Research

Vol. 44 No. 1 March 2006

Printed in U.S.A.

Is There a Link between Executive

Equity Incentives and Accounting

Fraud?

MERLE ERICKSON,* MICHELLE HANLON,t AND

EDWARD L. MAYDEWt

Received 8 April 2004; accepted 3 August 2005

ABSTRACT

We compare executive equity incentives of firms accused of accounting

fraud by the Securities and Exchange Commission (SEC) during the period

1996-2003 with two samples of firms not accused of fraud. We measure equity

incentives in a variety ofways and employ a battery of empirical tests. We find no

consistent evidence that executive equity incentives are associated with fraud.

These results stand in contrast to assertions by policy makers that incentives

from stock-based compensation and the resulting equity holdings increase the

likelihood of accounting fraud.

1. Introduction

Some of the largest accounting frauds in history occurred in the last sev-

eral years, leading to the well-known upheaval in the accounting industry

*University of Chicago; tUniversity of Michigan; tUniversity of North Carolina. We have

benefited from comments from Ray Ball (editor), Robert Bushman, Ilia Dichev, Arthur Kraft,

Mark Lang, Shiva Rajgopal,John Robinson, Steve Rock, Phil Shane, Terry Shevlin,Joel Slemrod;

workshop participants at London Business School, Michigan State University, UC-Berkeley,

UCLA, the University of Colorado, the University of Florida, the University of North Carolina,

the University of Oregon, the University of Texas at Austin, and Washington University; and

an anonymous referee. Judson Caskey, Scott Dyreng, and Brad Lindsey provided valuable

research assistance. We appreciate funding from the University of Chicago Graduate School of

Business, the Ross School of Business at the University of Michigan and an Ernst & Young Faculty

Fellowship, and the University of North Carolina Kenan-Flagler Business School, respectively.

113

Copyright ?, University of Chicago on behalf of the Institute of Professional Accounting, 2006

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

114 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

and sweeping legislative and regulatory changes. These events have left legis-

lators, regulators, practitioners, and academics searching for causes of these

frauds. Understanding the underlying forces that gave rise to these frauds

is a necessary precursor to effectively preventing future occurrences. Many

have suggested that the explanation lies in the incentives and opportunities

for personal gain faced by executives.

One of the incentives that managers might be responding to is the increase

in the proportion of their wealth that is tied to stock-based compensation

and the resulting equity holdings. It is well documented that the use of stock

options as a form of executive compensation rose dramatically during the

1990s, as did other forms of pay-for-performance plans such as grants of

restricted stock and bonus plans tied to performance (Murphy [1999]). A

substantial body of theoretical work, beginning with Jensen and Meckling

[1976], shows that these plans can be useful in aligning the incentives of

managers with those of shareholders. Following this work, many empiri-

cal researchers have predicated their analyses on the premise that granting

options is consistent with firm value maximization (for examples see Dem-

setz and Lehn [1985], Himmelberg, Hubbard, and Palia [1999], Core and

Guay [1999], and Rajgopal and Shevlin [2002]). Moreover, a number of

empirical papers find results consistent with the incentive alignment view.

For example, Brickley, Bhagat, and Lease [1985] report positive stock price

reactions to announcements of long-term managerial compensation plans.

Another example is Lewellen, Loderer, and Rosenfield [1985], who find ev-

idence consistent with managers being less likely to make merger bids that

lower their stock prices when they hold more stock in the firm, consistent

with their interests being aligned with shareholders. There is even some

empirical evidence that, on average, executive stock options are effective

in generating positive future payoffs for the firm in terms of accounting

earnings (Hanlon, Rajgopal, and Shevlin [2003]).'

However, there are others that view option compensation differently.

Some argue that options are an inefficient way to compensate managers

(Jenter [2001], Meulbroek [2001], and Hall and Murphy [2002]), that

managers use option grants for their own benefit (Aboody and Kasznik

[2000], Yermack [1997]), or that stock options do not exhibit empiri-

cal relations consistent with the economic motivations behind granting

them (Yermack [1995]). Further, recent work by Bebchuk, Fried, and

Walker [2002] argues that executives have the power to influence their own

pay, that executives use that power to extract rents, and that their desire

to camouflage their rent extraction might lead to the use of inefficient

pay arrangements that provide suboptimal incentives, thereby reducing

1 In addition, in a study of new economy firms, Ittner, Larcker, and Lambert [2003] find that

lower than expected stock option grants and/or existing option holdings are associated with

lower accounting and stock price performance in subsequent years. However, they find little

consistent association between future performance and greater than expected option grants

and holdings.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 115

shareholder value. In related work, Bar-Gill and Bebchuk [2003] and

Goldman and Slezak [2006] model managerial incentives to report truth-

fully and show that performance-based compensation can induce managers

to misreport performance.Jensen [2003] asserts that nonlinearity in pay-for-

performance systems induces managers to lie and that such lying is so perva-

sive that firms would be better off adopting solely linear pay-for-performance

systems.

Given the increased use of performance-based compensation over time

and the theoretical and empirical arguments of its shortcomings, it is possi-

ble that the perceived increase in corporate fraud in the late 1990s and early

2000s was a predictable outcome of these changed incentives. Indeed, this

conclusion seems to have become accepted wisdom among policymakers

and regulators. For example, Alan Greenspan, Chairman of the Federal Re-

serve Board, in his semiannual monetary policy report to Congress on July

16, 2002 speculated about the causes of the accounting frauds as follows:

An infectious greed seemed to grip much of our business community...

Too many corporate executives sought ways to "harvest" some of those

stock market gains. As a result, the highly desirable spread of sharehold-

ing and options among business managers perversely created incentives

to artificially inflate reported earnings in order to keep stock prices high

and rising. This outcome suggests that the options were poorly structured,

and, consequently, they failed to properly align the long-term interests of

shareholders and managers, the paradigm so essential for effective corpo-

rate governance. The incentives they created overcame the good judgment

of too many corporate managers. It is not that humans have become any

more greedy than in generations in the past. It is that the avenues to express

greed had grown so enormously.

Further evidence is found in prosecutors' arguments in these financial

accounting fraud cases. For example, in the Bernie Ebbers case (the former

CEO of WorldCom), the prosecution argued that Ebbers was motivated to

commit fraud when the end of a wave of mergers and the beginning of

a meltdown in the telecom industry put pressure on the company's share

price. They argued that "Ebbers's personal fortune was largely based on

WorldCom shares, and he had borrowed nearly $400 million with those

shares as collateral" (Ackman [2005]). Ebbers maintained that he commit-

ted the fraud to keep up with Wall Street expectations and not for personal

gain-he held his shares until WorldCom filed for bankruptcy in July 2002,

three months after he was forced to resign and one month after the fraud

was uncovered (Ackman [2005]). Similar allegations about the link between

fraud and executive stock options were made in other cases (e.g., Health-

south). The purpose of this paper is to examine whether the empirical ev-

idence supports the claim that equity incentives resulting from stock-based

compensation are positively associated with the probability of accounting

fraud.

We begin by identifying all firms explicitly accused of accounting fraud

by the Securities and Exchange Commission (SEC) during the period

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

116 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

January 1996 to November 2003. Using actual SEC allegations of fraud

avoids biases that can occur when a sample is selected based on researcher

interpretations about which accounting problems are the result of fraud and

which are not. After some restrictions, including the availability of executive

compensation data, our sample consists of 50 firms accused of accounting

fraud by the SEC during this period (hereafter, the "fraud firms" or "fraud

sample"). We compare the fraud firms with two samples of firms not accused

of fraud. First, we compare the fraud firms with a matched sample of firms

(hereafter, the "matched sample") in which we select two matched firms

not accused of fraud, based on size, year, and industry, for each fraud firm.

Second, we compare the fraud firms with an unmatched sample of firms

consisting of all remaining firm-years (meaning all firm-years other than

those of the 50 fraud firms) on ExecuComp (hereafter, the "unmatched

sample").

We estimate a variety of equity incentive measures for the top five execu-

tives at each firm, focusing specifically on the expected change in value of

the executives' stock and option portfolio to a 1% stock price change (here-

after, "sensitivity"). In univariate and multivariate analyses for the matched

sample, we find no significant association between fraud and the sensitiv-

ity of equity holdings. In the unmatched sample, using logistic regression,

we find significant associations between fraud and sensitivity when we

include no control variables in the regression. However, these associa-

tions become insignificant when we include controls for corporate gover-

nance, the desire for external financing, financial performance, and firm

size.

In additional analyses, we examine whether there is evidence that execu-

tives at firms accused of fraud systematically sell stock and exercise options

during the period of the alleged fraud to a greater extent than executives

in the matched and unmatched samples of control firms. We do not find

significantly greater stock sales or option exercises by executives at fraud

firms compared to nonfraud firms, regardless of whether the comparison is

based on the matched sample or the unmatched sample.

Some caveats are in order. While our sample of fraud firms is based on SEC

allegations of fraud and is therefore free of researcher classification bias,

there is no perfect method of identifying firms that engage in accounting

fraud. What our tests rely on is that the probability that a firm engaged in

fraud is greater for firms accused of fraud by the SEC than it is for firms not

accused by the SEC.

Nevertheless, any classification of firms into fraud and nonfraud cate-

gories will inevitably contain both Type I and Type II errors. If we assume

that the Type I error is relatively small in firms accused of fraud by the SEC

relative to other sample selection methods (e.g., size of discretionary ac-

cruals), then our tests are best interpreted as joint tests of the likelihood

of engaging in fraud and being caught by the SEC. It is unlikely, however,

that the SEC catches all fraud, and thus the sample contains an unknown

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 117

amount of Type II error.2 While in an ideal world we would have a perfect

sample, we simply cannot reliably identify firms that commit fraud and get

away with it. One aspect of the potential undercounting of fraud is that it

could induce a seeming association between compensation and fraud even

if none existed. This could occur if the SEC is more likely to investigate,

or investigates with more scrutiny, accounting problems that occur in firms

that pay high levels of stock-based compensation to their executives. In that

case, the SEC's own selection criteria could induce an association between

compensation and SEC investigations and/or allegations of fraud.

The paper proceeds as follows. In section 2 we discuss prior related liter-

ature. In section 3 we discuss the sample selection. Section 4 contains the

results and section 5 concludes.

2. Prior Research on the Determinants ofFraud and Related Issues

While the set of papers that examine accounting fraud is small in number,

the broader context of this study relates to compensation and financial re-

porting incentives. For example, Kedia [2003], Cheng and Warfield [2005],

Ke [2002], and Gao and Shrieves [2002] examine the relation between

compensation and earnings management measures (e.g., discretionary ac-

cruals, earnings thatjust meet or beat analysts' forecasts, small positive earn-

ings announcements, etc.). Richardson, Tuna, and Wu [2003] and Efendi,

Srivastav, and Swanson [2004] examine the relation between compensation

and restatements, and Denis, Hanouna, and Sarin [2005] examine the re-

lation between compensation and class action lawsuits. Our study focuses

on firms accused of intentionally (fraudulently) misstating earnings. While

earnings management, restatements, and fraud share certain traits and are

all phenomena worthy of study, they are not the same. Fraud and earnings

management differ in that earnings management can be within or outside

of generally accepted accounting principles (GAAP), whereas alleged fraud-

ulent accounting is invariably outside of GAAP. In addition, accruals models

are useful, in part because they can be applied to the entire market, but are

subject to well known limitations (see Dechow, Sloan, and Sweeney [1995],

Guay, Kothari, and Watts [1996], and Healy and Wahlen [1999]). Fraud and

restatements differ in that restatements do not necessarily reflect a prior in-

tent to deceive, whereas fraud by definition involves intent to deceive. For

example, Palmrose and Scholz [2002] examine 492 restatements and find

that only 11% result in an Accounting and Auditing Enforcement Release

(AAER) issued against the firm and only 38% result in litigation. This is

consistent with restatements reflecting a wide variety of behavior, ranging

from genuine disagreements over the application of GAAP to specific sets

2 Although one could argue that very large magnitude frauds are more difficult for managers

to hide indefinitely, and thus there would be less misclassification of frauds involving large

magnitudes.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

118 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

of transactions constituting fraud. Examining the purported causes of fi-

nancial accounting fraud is important in its own right, as evidenced by the

widespread impact of the financial accounting frauds of recent years. The

remainder of our discussion of the literature is specifically focused on papers

related to accounting fraud.

We are aware of only two published papers that examine possible links

between equity incentives and fraud. Dechow, Sloan, and Sweeney [1996]

examine the causes and consequences of firms subject to SEC enforcement

actions during the years 1982-1992. They find that firms subject to en-

forcement actions are more likely to have boards of directors dominated

by management, a CEO who is also Chairman of the Board, a CEO who is

also the firm's founder, no audit committee, and no outside blockholder.

While not the focus of their study, Dechow, Sloan, and Sweeney [ 1996] find

no significant evidence that sample firms had earnings-based bonus plans

or that officers and directors made unusual amounts of stock sales during

the manipulation period. Since the end of the sample of Dechow, Sloan,

and Sweeney [1996] in 1992, stock-based compensation has come to dwarf

earnings-based bonuses as a form of incentive compensation. We extend

the work of Dechow, Sloan, and Sweeney [ 1996] by examining the effects of

stock-based incentives (e.g., the sensitivity of stock options and stock held)

on the propensity of executives to engage in alleged fraud.

Beneish [1999] finds that managers of firms subject to AAERs are more

likely to sell their own stock during periods in which the earnings man-

agement is taking place than are managers of control firms. In contrast to

Dechow, Sloan, and Sweeney [1996], Beneish [1999] does not find evidence

that managers are motivated to inflate earnings in advance of an equity is-

suance by the firm.3

Recently, a number of working papers analyzing accounting fraud and

executive compensation have appeared. Johnson, Ryan, and Tian [2005]

find that executives of firms accused of fraud have greater unrestricted stock

incentives to increase stock price than do executives of firms not accused

of fraud. However, at the same time, and consistent with our results, they

report that the fraud firms do not have greater incentives from vested stock

options, unvested stock options, or restricted stock.4 Peng and Roell [2003]

3Another study, Beasley [1996], finds that outside directors decrease the probability of

fraud, as do higher tenure and ownership by outside directors. Beasley [1996] finds that the

presence of an audit committee has no effect on the probability of fraud, but does not examine

executive compensation.

4 There are a number of sample and design differences between our paper and that of

Johnson, Ryan, and Tian [2005]. In addition, if we rank our equity incentive variables, neither

of our test variables is consistently significant in both the matched and unmatched sample.

However, the rank of unrestricted stock is significantly associated with the likelihood of fraud,

with a p-value of 0.06 in the matched sample and a p-value of 0.02 in the unmatched sample,

consistent with the Johnson, Ryan, and Tian [2005] median results for this variable. The un-

ranked values are insignificantly related to the likelihood of fraud, however, with p-values of

0.797 and 0.230 in the matched and unmatched samples, respectively.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 119

find that incentive pay in the form of options (but not restricted stock or

base pay) increases the probability of private securities litigation.5 While we

can also find a relation between equity incentives and alleged fraud under

certain specifications, we find that the results are not robust to alternative

specifications and inclusion of control variables. Thus, we conclude that

there is no consistent evidence of a link between executive equity incentives

and accounting fraud.

3. Sample Selection and Description

To identify firms accused of accounting fraud, we searched SEC AAERs

for the word "fraud" during the years 1996-2003. AAERs are summaries

of the SEC's accounting-based enforcement actions and describe the SEC's

investigations of alleged violations of accounting provisions of the securities

laws. We require AAERs to contain allegations of fraud to be included in

our sample because AAERs often involve nonfraudulent issues (Feroz, Park,

and Pastena [1991]).6

Table 1 summarizes our sample selection procedure. From the set of

AAERs during the period January 1, 1996 to November 19, 2003, we iden-

tify 287 instances (excluding repeat accusations) in which the SEC uses the

word "fraud." We exclude AAERs that are not relevant to the subject of this

inquiry, such as alleged fraud by brokers and dealers, government person-

nel, or charitable organizations. We also require sufficient disclosure about

executive compensation to compute our estimates of executive wealth sen-

sitivity to stock price changes. We first look to the S&P ExecuComp database

(ExecuComp). For the firms without ExecuComp data, whenever possible,

we hand collect the necessary data from proxy statements filed with the SEC.

This results in a sample of 50 firms that are both accused of fraud by the

SEC and for which we have compensation data.

Table 1 also presents information on the two comparison samples used

in this study. First, we compile a matched sample of firms not accused of

accounting fraud. For each of the 50 firms accused of accounting fraud,

we select two matched firms not accused of accounting fraud, matching

on industry (two-digit Standard Industrial Classification [SIC] code), year,

and firm size (total assets). Thus, we have 100 matched sample firms for

the matched sample tests. Second, we also compare the 50 firms accused

of fraud with all the remaining firm-years on ExecuComp. After employing

data requirement screens, the unmatched sample contains 13,033 firm-year

observations (not including the 50 fraud firm observations).

5 In recent theoretical work, Goldman and Slezak [2006] develop an agency model in which

stock-based compensation is a double-edged sword, inducing managers to exert productive

effort but also inducing managers to divert valuable resources to misrepresent performance.

Similarly, Bar-Gill and Bebchuk [2003] model managerial incentives to commit fraud.

6 For more detail on AAERs and the SEC's process in investigating firms, see Pincus, Holder,

and Mock [1988], Feroz, Park, and Pastena [1991], and Dechow, Sloan, and Sweeney [1996].

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

120 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

TABLE 1

Description of Sample Selection Procedure

Sample of firms accused of fraud

Total Accounting and Auditing Enforcement Releases (AAERs) issued from 287

January 1, 1996 through November 19, 2003 attributable to alleged accounting

fraud (nonduplicates)

Less:

Fraud allegations not against a firm or its executives (e.g., allegations (57)

involving broker dealers, investment advisors, auditors, city government

personnel, charitable organizations)

AAERs unrelated to fraudulent financial statements (1)

AAERs that do not specify the year of the alleged fraud (1)

AAER firms for which data are not available (not on Compustat or (178)

ExecuComp, no proxy statements are available, missing variables, or fraud

predates 1992 when ExecuComp coverage begins)

Sample of firms accused of fraud by the SEC 50

Samples of firm-years not accused of fraud

A: Matched sample offirms

Firms matched based on two-digit industry code, year, and total assets 100

B: Unmatched sample

Total firm-years available on ExecuComp through 2001 17,427

Less:

Firm-years of firms accused of fraud (191)

Firm-years with missing required ExecuComp compensation or stock (685)

holding variables

Firm-years with missing required Compustat financial statement data (3,518)

Sample of firm-years not accused of fraud by the SEC 13,033

We require nonmissing option data to be included in the sample. The category "Firm-years

with missing required Compustat financial statement data" includes firm-years in the financial

services industry for which our measure of ex ante financing cannot be obtained.

Table 2 presents the name of each of the 50 sample firms accused of fraud

by the SEC. As the list indicates, the sample captures the well-publicized

alleged accounting frauds of recent years as well as some that were less

publicized.

In table 3, we break down the number of fraud and nonfraud firms for the

unmatched sample by industry (panel A) and by year (panel B). To make

panel A of manageable size, we break out only those two-digit SIC codes for

which we have one or more firms accused of fraud. Most two-digit SIC codes

have no firms accused of fraud by the SEC during the time period in this

study and we lump those together into an "all other" category.

The first observation one can make regarding panel A is that the inci-

dence of fraud does not seem to be highly concentrated in a few industries.

The 50 firms accused of fraud are spread over 21 different two-digit SIC

codes. The industry with the largest absolute number of alleged frauds is

business services (SIC 73), which has 10 firms accused of fraud. Business ser-

vices firms are over twice as likely as the average firm to be accused of fraud,

accounting for 8.52% of the nonfraud firms and 20% of the alleged frauds.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 121

TABLE 2

List ofFirms Accused of Accounting Fraud by the SEC

Time Period of

Number Company Name Alleged Fraud

1 Adelphia Communications Corporation 1999-2001

2 Advanced Technical Products Inc. 1998-1999

3 Anicom Inc. 1998-2000

4 Aremissoft Inc. 2000-2001

5 Ashford.com Inc. 2000-2001

6 Atchison Casting Corporation 1997-2000

7 Bausch & Lomb Inc. 1993-1994

8 Critical Path Inc. 2000-2001

9 Cyberguard Corporation 1997-1998

10 Cylink Corporation 1998

11 Diagnostek Inc. 1992-1993

12 Dynegy Inc. 2001-2002

13 Enron Corporation 1997-2001

14 Jo Ann Stores Inc. 1992

15 Fine Host Corporation 1996-1997

16 Gateway Inc. 2000

17 Guilford Mills Inc. 1997-1998

18 HBO & Co. 1997-1999

19 Healthsouth Corporation 1999-2002

20 IGI Inc. 1995-1997

21 Indus International Inc. 1999

22 International Thoroughbred Breeders 1997

23 K Mart Corporation 2001

24 Safety Kleen Corporation 1998-2000

25 Legato Systems Inc. 1999-2000

26 Material Sciences Corporation 1996-1998

27 Max Internet Communications Inc. 2000

28 Micro Warehouse Inc. 1994-1996

29 Microstrategy. Inc. 1998-2000

30 Oak Industries Inc. 1995-1996

31 Physician Computer Network Inc. 1996-1997

32 Premier Laser Systems Inc. 1998

33 Qwest Communications International Inc. 2000-2001

34 Rite Aid Corporation 1998-2000

35 Saf T Lok Inc. 1997-1998

36 Signal Technology Corporation 1996-1998

37 Structural Dynamics Research Corporation 1992-1994

38 Sunbeam Corporation 1997-1998

39 Sunrise Medical Inc. 1994-1995

40 Swisher International Inc. 1996

41 Symbol Technologies Inc. 1998-2002

42 System Software Assoc Inc. 1994-1996

43 Texlon Corporation 1999

44 Thomas & Betts Corporation 1998-1999

45 Thor Industries Inc. 1996-1998

46 Tyco International LTD 1997-2002

47 W R Grace & Co. 1992-1995

48 Waste Management Inc. 1992-1997

49 Worldcom Inc. 1999-2002

50 Xerox Corporation 1997-2000

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

122 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

TABLE 3

Sample of Firms Accused of AccountingFraud by the SEC and Firms Not Accused (Unmatched Sample)

by Industry and by Year

Panel A: By Industry

12

No Allegation Alleged Fraud

of Fraud Firms Firms

Two digit Percent of Percent of

SIC code Industry Name Total Total

22 TEXTILE MILL PRODUCTS 1.03% 2.00%

23 APPAREL & OTHER FINISHED PDS 1.03% 2.00%

28 CHEMICALS & ALLIED PRODS 8.35% 6.00%

33 PRIMARY METAL INDUSTRIES 3.00% 2.00%

34 FABR METAL, EX MACHI TRANS EQ 1.73% 6.00%

35 INDL, COMML MACHY, COMPUTER EQ 7.17% 12.00%

36 ELECTR, OTH ELEC EQ, EX CMP 8.36% 4.00%

37 TRANSPORTATION EQUIPMENT 2.94% 2.00%

38 MEAS INSTR; PHOTO GDS; WATCHES 4.83% 6.00%

48 COMMUNICATIONS 2.06% 6.00%

49 ELECTRIC, GAS, SANITARY SERV 8.24% 4.00%

50 DURABLE GOODS-WHOLESALE 2.53% 2.00%

51 NONDURABLE GOODS-WHOLESALE 1.41% 4.00%

53 GENERAL MERCHANDISE STORES 1.55% 2.00%

58 EATING AND DRINKING PLACES 2.02% 2.00%

59 MISCELLANEOUS RETAIL 1.89% 10.00%

67 HOLDING, OTHER INVEST OFFICES 0.41% 2.00%

73 BUSINESS SERVICES 8.52% 20.00%

79 AMUSEMENT & RECREATION SVCS 0.84% 2.00%

80 HEALTH SERVICES 1.90% 2.00%

99 NONCLASSIFIABLE ESTABLISHMNT 0.32% 2.00%

OTHER INDUSTRIES 29.87% 0.00%

100.00% 100.00%

Panel B: By year

No Allegation of Fraud Firms Alleged Fraud Firms

34

Year Percent of Total Percent of Total

1992 7.4% 10.0%

1993 9.4% 2.0%

1994 9.7% 6.0%

1995 9.9% 4.0%

1996 10.8% 12.0%

1997 11.4% 20.0%

1998 11.4% 18.0%

1999 10.8% 12.0%

2000 10.1% 12.0%

2001 9.1% 4.0%

100.0% 100.0%

Firms are categorized as those accused of accounting fraud and those not accused. In column (1)

we report the firms not accused of fraud by industry as a percentage of all firms not accused of fraud.

In column (2) we report the percentage of all firms accused of fraud in the sample that are drawn

from the respective industry. In column (3) we report the firm-years in which there is no accusation

of fraud as a percentage of all firm-years not accused of fraud (unmatched sample). In column (4) we

report the percentage of all firms accused of fraud in the sample by the year in which the alleged fraud

began.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 123

Miscellaneous retail (SIC 59) is also overrepresented in the alleged

fraud sample; firms in this industry are about five times more likely

than the average firm to be accused of fraud, accounting for 1.89%

of the nonfraud sample and 10% of the fraud sample. Other overrep-

resented industries in the fraud sample include communications (SIC

48) and fabricated metal (SIC 34). Overall, however, the data in this

table suggest that accounting fraud is not isolated in any particular

industry.

We excluded banks and savings and loans from our sample because an im-

portant control variable-the proxy for the desire for external financing-

was not available on Compustat for these firms. However, this restriction

only excluded three firms, suggesting that unlike the alleged accounting

frauds of the late 1980s (see Erickson, Mayhew, and Felix [2000]), the

alleged frauds of the 1990s and early 2000s do not appear to be concen-

trated in banks and savings and loans.

Panel B of table 3 presents the breakdown of SEC fraud allegations by

year, where year represents the first year the alleged fraud took place (not

the year in which the SEC made the allegation). The alleged frauds begin

as early as 1992 and as late as 2001. The number of frauds beginning in a

given year range from a low of one in 1993 to a high of 10 in 1997. Overall,

there is no discernable trend over time and the alleged frauds are not overly

clustered in any given year.

For each of the fraud firms, we collect the method of earnings overstate-

ment as documented by the SEC (while in theory the SEC could issue an

AAER for understating earnings, all of the AAERs that we observed were

for overstating earnings). In untabulated statistics, we find that 56% of

the firms were accused of overstating their revenue. Specific SEC accusa-

tions include: reporting "false revenues," "improperly recognizing" revenue,

"channel stuffing," "prematurely recognizing revenue," "inflating" revenue,

"recording revenue from the sale... prior to shipment," and recognizing

revenue on "invalid or nonexistent sales." The second most common accu-

sation consists of various types of cost or expense understatement, applying

to 34% of the firms. Examples of such expense understatement accusations

contained in SEC AAERs include: "created fictitious assets," "overstating

inventory," "improperly capitalizing expenses," "improper capitalization of

millions of dollars of company expenses," and "including fake items in in-

ventory." Other more specific sources of overstatement include purchase ac-

counting or merger-related accounting entries (e.g., "cookie jar reserves"),

barter transactions, lease accounting manipulation, and overstatement of

inventory.

We also examine the level of management accused in the AAER. In unt-

abulated results, we find that the highest levels of management are most

often accused of involvement in the accounting fraud. In more than 50% of

the cases, the CEO or the CFO was accused of perpetrating the accounting

fraud that led to the AAER.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

124 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

4. Tests of the Relation between Executive Equity Incentives and

Alleged AccountingFraud and Estimates of Managerial Benefits

from Accounting Fraud

4.1 MEASUREMENT OF EXECUTIVE EQUITY INCENTIVE VARIABLES

In this section, we introduce variables that measure the expected benefit

the executive could obtain from artificially inflating the stock price of the

firm through fraudulent actions. For the firms accused of fraud, all variables

are measured in the year immediately preceding the first alleged fraud year.

However, for some firms, the fraud years started prior to 1992, and thus

data on ExecuComp are not available for these firm-years. In those cases,

and where the fraud started no more than two years earlier, we substitute the

year 1992 for the year prior to fraud in order to keep the firm in the sample.

All equity incentive variables are aggregated over the entire management

team listed in the proxy statement for the fiscal year. Generally, this is the five

most highly compensated executives. The variables are defined as follows:

SENSITIVITY = the change in value of the top five managers'

stock, restricted stock, and stock option portfolio

in response to a 1% change in the stock price.

(We discuss this measure further below.)

VESTED STOCK AND

OPTION SENSITIVITY = the change in value of the top five managers'

portfolio of exercisable stock options and unre-

stricted stock in response to a 1% change in the

stock price. (We discuss this measure further

below.)

We define SENSITIVITY as the expected change in the top five executives'

firm-based equity wealth from a 1% change in stock price. This change in

wealth consists of the sum of the change in the value of the executives' em-

ployee stock option portfolios, restricted stock holdings, and stock holdings.

We aggregate across the three to derive a total dollar measure of the sensi-

tivity for each executive. We aggregate across the firm's top five executives

to get a total dollar measure for the top management team of the firm. We

also examine the relation of the vested portion of this portfolio to being

accused of fraud, because these holdings are those the manager can im-

mediately benefit from and thus may provide greater incentives to commit

fraud.

The expected wealth changes from stock and restricted stock are esti-

mated by multiplying the market value of the stock holdings at year-end

(the year prior to the alleged accounting fraud for the alleged fraud firms)

by 1%. For stock options, we use the method of calculating their sensitiv-

ity to stock price as described by Core and Guay [2002]. Although details,

such as the number of options, exercise price, and time to maturity, are

available from ExecuComp or the current year proxy statement for current

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 125

year grants, much of these data are unavailable in the current year proxy

statement for prior grants. The one-year approximation method described

in Core and Guay [1999, 2002] requires information only from the most

recent proxy statement to estimate the sensitivity of the option portfolio to

a change in stock price. The sensitivity to stock price for each option held

is estimated as:

[a(optionvalue)/la (price)] * [price * 0.01] = e-dTN(Z) * [price * 0.01]

where:

d = natural logarithm of expected dividend yield over the life of the

option;

T = time to maturity of the option in years;

N = cumulative normal probability function; and

Z = [ln(S/ X) + T(r - d + a2/2)]/a T1/2 (where S is the price of the un-

derlying stock, X is the exercise price of the option, r is the natural

logarithm of the risk-free interest rate, and a is the expected stock-

return volatility over the life of the option).7

4.2 CONTROL VARIABLES

We include a variety of control variables defined as follows:

CEO= CHAIR = indicator variable that is set to one if the

Chairman of the Board is also the CEO.

NUMMTGS = the number of board meetings held dur-

ing the fiscal year (ExecuComp variable

Nummtgs) .

FINANCING = an ex ante measure of a firm's desire

for external financing. It is an indica-

tor variable coded 1 if the firm's variable

FREECASH is less than -0.5 and 0 other-

wise. FREECASHt is defined as:

ICash from operations, (Compustat data #308)

Average Capital Expenditurest-3to t-1 (data #128)

Current Assetst-_ (data #4)

7 We also conducted our analysis using the mix of compensation (a ratio computed as the

Black-Scholes value of executive stock option grants and the value of restricted stock grants

divided by the executives' total compensation, i.e., salary, bonus, the value of stock option

grants, and the value of restricted stock grants) in the year prior to fraud as the measure of

the manager's equity compensation incentives. The coefficient on this mix of compensation

variable is positive and significant in both samples when no controls are included (p-values

of 0.001 in the matched sample and 0.0001 in the unmatched sample). In the presence of

our control variables, the coefficient on the mix variable becomes less significant although

still significant at the 0.10 level (p-value of 0.07 in both samples). We exclude the mix of

compensation from our main tests because this variable does not represent the theoretical ex

ante incentives of management. We thank the referee for pointing this out to us.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

126 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

LEVERAGE = total debt (Compustat data #34 + #9)

scaled by total assets (data #6).

MARKET VALUE OFEQUITY = the market value of equity of the firm

(Compustat data #199 times #25).

ALTMAN'S Z SCORE = proxy for risk of financial distress calcu-

lated based on Altman [1968] as updated

by Begley, Ming, and Watts [1996].8

BOOK TO MARKET = the book value of shareholder's equity

(Compustat data #216) divided by the

market value of equity.

EARNINGS TO PRICE = net income (Compustat data #172) per

share divided by end-of-year stock price.

RETURN ON ASSETS = net income (Compustat data #172) di-

vided by year-end assets (data #6).

SALES GROWTH = the percentage change in sales (Compu-

stat data #12) from the prior year to the

current year (for the fraud firms from

two years prior to the fraud to the year

prior to the fraud).

AGE OFFIRM = the length of time in years the firm has

been publicly traded (from the Center

for Research in Security Prices [CRSP]).

M&A IN YEAR OF FRAUD = an indicator variable set equal to one

if the firm had an acquisition that

contributed to sales in the prior year (ac-

quisition in the first year of fraud for

fraud firms). (Variable is set equal to one

if data #249 > 0, otherwise variable is set

equal to zero.)

TOTAL ASSETS = the firm's total assets (data #6). For firms

accused of fraud, total assets are mea-

sured as of the year preceding the al-

leged fraud.

STOCK VOLATILITY = the standard deviation return volatility

of returns calculated over 60 months.

For firms available on ExecuComp, we

use the variable BS_ VOLATILITY. For the

firms not available on ExecuComp, we

compute the variable using the CRSP

database.

CEO TENURE = the number of years the CEO has been

the CEO of the company. For firms

available on ExecuComp we compute

8We also estimate the regressions using the proxy for distress as developed by Shumway

[2001], and our inferences are unchanged.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 127

the tenure as the number of years be-

tween the variable BECAMECEO and the

current year; for fraud firms, the cur-

rent year is the year prior to fraud.

For firms not available on ExecuComp,

we compute the variable in the same

manner using data obtained from proxy

statements.

MISSING CEO TENURE = an indicator variable set equal to one if

the CEO TENURE variable is missing and

zero otherwise.

We include two governance variables. The first is CEO= CHAIR. Hermalin

and Weisbach [2001] argue that the greatest factor affecting the Board

of Directors' effectiveness is its independence from the CEO. Moreover,

Dechow, Sloan, and Sweeney [1996] find that having a Chairman who is

simultaneously the CEO increases the likelihood of manipulating earnings.

We also include the number of board meetings (NUMMTGS) as a proxy for

the monitoring and effort contributed by directors (Adams [2000], Vafeas

[1999]).9

We include FINANCING to control for the need to obtain external

financing. As FREECASH becomes more negative the firm becomes closer to

exhausting its internal funds and thus has an incentive to manipulate earn-

ings in anticipation of accessing the capital markets. Because the relation is

not likely linear, we follow Dechow, Sloan, and Sweeney [1996] and create

an indicator variable set equal to one if the firm will likely want external

financing within the next two years (i.e., FREECASH less than or equal to

-0.5) and zero otherwise. This cutoff assumes that if a firm requires exter-

nal financing within the next two years it will start taking action now to raise

the desired funds.

We include LEVERAGE and ALTMAN's Z score to control for the possibil-

ity that financially distressed firms have greater incentives to commit fraud

than firms that are not distressed (Altman [1968], Begley, Ming, and Watts

[1996]). We also include several performance metrics in order to control

for the effect of poor financial performance or weakened financial stability.

Poorly performing firms may resort to accounting fraud in order to cover up

their deficient performance. Specifically, we include EARNINGS TO PRICE,

BOOK TO MARKET, and RETURN ON ASSETS, all measured in the year

prior to fraud to control for financial performance. We include MARKET

9 In addition to the governance variables listed above, we also include the Governance In-

dex metric compiled by Gompers, Ishii, and Metrick [2003], where available, for our matched

sample firms. This variable is a measure of shareholder rights based on 24 corporate gover-

nance provisions collected by the Investors Responsibility Research Center. Because we lose 18

observations from our fraud sample, we do not retain this variable in our main analyses. Our

inferences are unchanged in the subsample for which we can include this variable, and the

Governance Index variable is not significant in the tests.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

128 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

VALUE OFEQUITY to control for size and SALES GROWTH to control for

the possibility that poorly performing (slow-growing) or high-growth firms

may commit fraud in order to increase or sustain their sales growth.10 We

also include AGE OF THE FIRM to control for the incentives to commit

fraud provided by an initial stock offering or by having newly issued stock.

Finally, we include an indicator variable, M&A IN YEAR OF FRAUD, which

takes on a value of one if a portion of those sales are from an acquisition.

Prior literature provides evidence consistent with firms managing earnings

prior to an acquisition in order to raise their stock price (Erickson and Wang

[1999], Louis [2004]).

Compensation structure is a choice variable, so using it as an explana-

tory variable runs the risk of endogeneity problems. For example, in cases

in which monitoring is more difficult, firms may use a higher level of in-

centive (stock-based) compensation (Demsetz and Lehn [1985], Smith and

Watts [1992], Gaver and Gaver [1993] and Core, Holthausen, and Larcker

[1999]). It is exactly in those cases in which monitoring is difficult that we

might also expect to see a higher likelihood of fraud. This is especially true

in the institutional setting considered in this study, because our sensitivity

variable represents risk the executive bears in his compensation structure

as it varies with the stock price of the firm. If firms provide more incentives

through pay risk when monitoring is more difficult and if fraud is more

likely when monitoring is more difficult then we may see the association be-

tween the structure of executive compensation and being accused of fraud

simply because monitoring is difficult in these firms and not because the

stock-based pay provided incentives for managers to engage in fraud.

We include proxies for the determinants of equity incentives as indepen-

dent variables in our regression in an attempt to control for the endogeneity

described above." We follow the model of equity incentives in Core and

Guay [1999]. We include CEO TENURE and STOCK RETURN VOLATILITY

10 In addition, to investigate whether our sample of alleged fraud firms had slower growth

during the fraud period, which may have led to their being caught, we compare the earnings'

(data item #172) growth rates during the fraud years with those of our matched sample of firms.

We find that the fraud firms and nonfraud firms did not have significantly different average

annual growth rates during the fraud period.

11 We recognize that the use of an instrumental variables (IV) two-stage least squares approach

is the standard textbook fix for endogeneity. However, finding appropriate instruments is ex-

tremely difficult and when one uses inappropriate instruments ". . .it can easily be the case that

the IV estimates are more biased and more likely to provide the wrong statistical inference

than simple OLS estimates that make no correction for endogeneity" (Larcker and Rusticus

[2005]) (also see Bound, Jaeger, and Baker [1995]). Thus, we do not implement a two-stage

IV estimation in our main tests. Further, because the endogeneity in our tests is the correlated

omitted variables type of endogeneity, we can include a set of determinants of the endogenous

regressor (SENSITIVITY) in our regression, mitigating the inconsistency caused by the corre-

lated omitted variables (assuming of course we still do not have a correlated omitted variable).

However, we note that, in the interest of completeness, we estimate a two-stage least squares IV

model using stock price as an instrument for SENSITIVITY. In this estimation, the inferences

regarding the relation between fraud and equity incentives are unchanged (i.e., no significant

relation).

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 129

proxies for the two variables in Core and Guay's [1999] model not included

in our model for other reasons.'2 CEO TENURE is predicted to be posi-

tively related to the level of equity incentives. As described by Core and

Guay [1999], over time, uncertainty about a CEO's ability is resolved. As risk

borne by the CEO due to uncertainty about his ability is reduced, it is possible

to impose more incentive risk on him, ceteris peribus. In addition, equity

incentives can be used to mitigate horizon problems (Dechow and Sloan

[1991]) so equity incentives may be increased as retirement approaches.

As discussed above, Demsetz and Lehn [1985] hypothesize that firms op-

erating in less predictable environments have higher monitoring costs and

because of these higher monitoring costs firms in noisier environments will

have higher managerial ownership. Executives in firms in which monitor-

ing is difficult will likely also have easier opportunities to commit financial

accounting fraud. Thus, we include STOCK VOLATILITY as a proxy for this

uncertainty in our regression.

4.3 UNIVARIATE TESTS

In table 4, we present descriptive statistics for the alleged fraud firms

and both samples of nonfraud firms (matched sample of 100 nonfraud

firms and the unmatched sample of nonfraud firms). For each variable we

present the mean, standard deviation, median, and lower and upper quar-

tiles. Variables that represent dollars are shown in millions and all variables

are winsorized (reset) at the 1% and 99% levels.'" For the firms accused of

fraud, all variables are measured as of the year prior to the beginning of the

fraud.

Comparing the fraud firms with the matched sample of firms reveals sev-

eral significant differences in control variables but no statistically significant

differences in the equity incentive variables. Notable differences are that the

fraud firms have a higher stock return volatility, are, on average, younger

firms that have CEOs with a marginally shorter tenure, have higher SALES

GROWTH prior to the fraud commencing, and have more instances of an

12 We include the BOOK TO MARKET ratio and MARKET VALUE OF EQUITYas described
above. We control for industry membership by matching on industry in the matched sample and

by including one-digit industry indicator variables in the unmatched sample. Core and Guay

[1999] include a variable for free cash flow problems set equal to the ratio of operating cash flow

less dividends to total assets if the firm has low growth opportunities and zero otherwise. The

variable follows Jenson [1986], who argues that the combination of low growth opportunities

and high free cash flow creates agency problems that can be mitigated with higher levels of

equity incentives. This is the only variable in Core and Guay's [1999] model that is insignificant

in their tests. Because of its insignificance and because we already include a free cash flow

variable following Dechow, Sloan, and Sweeney [1996] that serves as a proxy for the firm's

external financing needs, and several proxies for growth (MARKET TO BOOK and SALES

GROWTH), we do not also include Core and Guay's [1999] specific free cash flow variable.

13We winsorize the matched sample variables over the distribution of the matched sample

plus alleged fraud firms (150 firms) and we winsorize the unmatched sample over the distri-

bution of the unmatched sample plus alleged fraud firms (13,083 firms).

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

TABLE 4

Descriptive Statistics on Financial Statement Variables and Market-based Measures for Firms Accused of Fraud by the SEC and for the Matched and Unmatched Samples

Standard Lower Upper in Mean in Median

Variable Mean Deviation Quartile Median Quartile (p-value) (pvalue)

SENSITIVITY

Accused of fraud 1.906 3.640 0.146 0.419 1.644

Matched sample 1.513 7.304 0.109 0.342 1.084 0.346 0.127

Unmatched sample 1.019 2.204 0.123 0.326 0.878 0.092 0.080

VESTED STOCK AND OPTION SENSITIVITY

Accused of fraud 1.537 3.239 0.092 0.301 1.136

Matched sample 1.238 6.615 0.059 0.276 0.782 0.374 0.187

Unmatched sample 0.810 1.938 0.075 0.220 0.627 0.119 0.087

CEO=CHAIR

Accused of fraud 0.76 0.43 1.00 1.00 1.00

Matched sample 0.76 0.43 1.00 1.00 1.00 0.791 0.792

Unmatched sample 0.71 0.45 0.00 1.00 1.00 0.415 0.415

NUMMTGS

Accused of fraud 7.20 3.65 4.00 6.00 9.00

Matched sample 6.99 3.24 5.00 6.00 8.00 0.747 0.449

Unmatched sample 7.07 2.79 5.00 6.00 9.00 0.806 0.502

FINANCING

Accused of fraud 0.140 0.351 0.000 0.000 0.000

Matched sample 0.020 0.141 0.000 0.000 0.000 0.281 0.202

Unmatched sample 0.032 0.177 0.000 0.000 0.000 0.035 0.001

LEVERAGE

Accused of fraud 0.206 0.184 0.037 0.185 0.325

Matched sample 0.184 0.161 0.013 0.176 0.314 0.491 0.587

Unmatched sample 0.229 0.176 0.070 0.223 0.349 0.347 0.269

MARKET VALUE OFEQUITY ($ MILLIONS)

Accused of fraud 5,012.90 11,563.87 168.20 412.10 4,256.96

Matched sample 5,759.75 17,836.23 212.51 503.66 2,749.90 0.868 0.858

Unmatched sample 3,718.95 9,247.22 346.66 881.11 2,774.60 0.433 0.141

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

ALTMAN'S Z SCORE

Accused of fraud 1.071 1.125 0.389 0.844 1.422

Matched sample 1.197 1.404 0.451 0.984 1.675 0.414 0.383

Unmatched sample 1.044 1.050 0.367 0.789 1.350 0.861 0.803

BOOK TO MARKET

Accused of fraud 0.040 0.353 0.150 0.325 0.481

Matched sample 0.454 0.341 0.252 0.358 0.551 0.380 0.172

Unmatched sample 0.487 0.364 0.243 0.409 0.633 0.085 0.018

EARNINGS TO PRICE

Accused of fraud -0.011 0.107 -0.030 0.020 0.050

Matched sample -0.065 0.588 0.015 0.042 0.060 0.380 0.014

Unmatched sample 0.020 0.123 0.018 0.045 0.067 0.069 0.001

RETURN ONASSETS

Accused of fraud -0.023 0.178 -0.035 0.037 0.087

Matched sample -0.017 0.339 0.021 0.057 0.087 0.304 0.070

Unmatched sample 0.042 0.102 0.020 0.051 0.089 0.013 0.022

SALES GROWTH

Accused of fraud 0.650 1.386 0.064 0.219 0.607

Matched sample 0.179 0.345 0.027 0.116 0.255 0.022 0.012

Unmatched sample 0.188 0.369 0.015 0.102 0.249 0.008 0.001

AGE OF FIRM

Accused of fraud 12.39 13.20 2.50 7.96 18.51

Matched sample 18.46 19.88 5.25 9.63 25.01 0.030 0.043

Unmatched sample 20.66 18.99 5.84 14.42 29.01 0.001 0.001

M&A IN YEAR OF FRAUD)

Accused of fraud 0.30 0.46 0.00 0.00 1.00

Matched sample 0.15 0.36 0.00 0.00 0.00 0.049 0.031

Unmatched sample 0.15 0.36 0.00 0.00 0.00 0.013 0.001

TOTAL ASSETS ($ MILLIONS)

Accused of fraud 3,904.24 7,870.00 100.26 281.02 3,294.46

Matched sample 4,141.72 9,726.75 136.30 278.41 2,236.01 0.885 1.000

Unmatched sample 2,994.32 5,909.17 299.39 794.58 2,518.75 0.418 0.008

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

TABLE 4 - Continued

Standard Lower Upper in Mean in Median

Variable Mean Deviation Quartile Median Quartile (p-value) (p-value)
STOCK VOLATILITY

Accused of fraud 0.534 0.264 0.334 0.482 0.669

Matched sample 0.400 0.172 0.267 0.363 0.513 0.002 0.004

Unmatched sample 0.376 0.205 0.243 0.348 0.485 0.000 0.000

CEO TENURE

Accused of fraud 4.820 5.927 0.000 2.500 9.000

Matched sample 6.540 5.589 2.000 5.500 10.000 0.084 0.014

Unmatched sample 6.037 7.072 1.000 4.000 9.000 0.154 0.142

MISSING CEO TENURE

Accused of fraud 0.140 0.351 0.000 0.000 0.000

Matched sample 0.060 0.238 0.000 0.000 0.000 0.151 0.103

Unmatched sample 0.100 0.300 0.000 0.000 0.000 0.423 0.346

All monetary amounts are in $ millions. Variables are as defined as follows: SENSITIVITY is a measure of how much the value of the top five executives' portfolio of stock options,

restricted stock and stock changes in response to a one percent change in stock price. This variable is measured using the methodology described in Core and Guay [2002]. For the

firms accused of fraud this variable is measured in the year prior to the alleged fraud. VESTED STOCKAND OPTION SENSITIVITY is a measure of how much the value of the top five

executives' portfolio of exercisable stock options and unrestricted stock changes in response to a one percent change in stock price. This variable is measured using the methodology

described in Core and Guay [2002]. For the firms accused of fraud this variable is measured in the year prior to the alleged fraud. CEO = CHAIR is an indicator variable taking

the value one if the CEO is also the Chairman of the Board, and zero otherwise. NUMMTGS is the number of times the firm's board meets, annually. FINANCING is an indicator

variable set equal to one when the firm's free cash flow divided by lagged current assets is less than -0.5 and zero otherwise. LEVERAGE is measured as the firm's debt divided by

total assets. For firms accused of fraud, IEVERAGE is measured as prior to the fraud. MARKET VALUE OFEQUITY is the market value of shareholders' equity. For firms accused of

fraud, it is measured the year prior to the alleged fraud. ALTMANS Z SCORE is a proxy for risk of financial distress calculated based on Altman [1968] as updated by Begley et al.

[1996]. BOOK TO MARKET is the book value of shareholders' equity divided by the market value of shareholder's equity. For the firms accused of fraud, the ratio is measured as

of the year preceding the alleged fraud. EARNINGS TO PRICE is the earnings to price ratio. For firms accused of fraud, the ratio is measured as of the year preceding the alleged

fraud. RETURN ON ASSE'TS is net income divided by assets. For the firms accused of fraud, the ratio is measured as of the year prior to the alleged fraud. SALES GROWTH is the

percentage change in sales from the prior year to the current year (for the fraud firms from two years prior to the fraud to the year prior to the fraud). AGE OFFIRM is the length

of time in years the firm has been publicly traded (from CRSP). M&A IN YEAR OF FRAUD is an indicator variable set equal to one if the firm had an acquisition that contributed to

sales in the prior year (acquisition in the first year of fraud for fraud firms). (Variable is set equal to one if data item #249>0, otherwise variable is set equal to zero.) TOTAL ASSETS

are the firm's total assets. For firms accused of fraud, total assets are measured as of the year preceding the alleged fraud. STOCK VOIATILITY is the standard deviation volatility of

returns calculated over 60 months. For firms in the Execucomp database this variable is the BSVolatility variable in Execucomp. We compute the variable in the same manner for

the hand-collected firms. (CO TENURE is the number of years that the CEO has been CEO of the firm. For the firms in the ExecuComp database, the variable is computed as the

number of years between the variable BE(CAMECEO and the current year. For fraud years it is measured as the year prior to fraud. For hand-collected firms, we compute the variable

in the same manner using dates obtained from the proxy statement. MISSING CEO 7TENNURE is an indicator variable set to one if the CEO TENURE variable is missing for the firm. All

variables are winsorized at the 1% and 99% levels. All p-values are from two-tailed tests.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 133

acquisition occurring in the first year of fraud. In terms of median differ-

ences, the results are similar with the exception of the fraud firms having a

lower EARNINGS TO PRICE.

There are a number of differences between the fraud firms and the un-

matched sample of firms. With regard to the compensation variables, the

data in table 4 indicate that SENSITIVITY is not statistically different be-

tween the fraud firms and the unmatched sample using a two-tailed test

(p = 0.092). The VESTED STOCK AND OPTION SENSITIVITY of the fraud

firms also is not statistically different from that of the firms not accused of

fraud. In terms of control variables, the fraud firms are again, on average,

younger firms, they have higher SALES GROWTH prior to the fraud com-

mencing, have greater STOCK VOLATILITY, and have more instances of an

acquisition occurring in the first year of fraud. In addition, the fraud firms,

on average, have a lower RETURN ONASSETS and have a greater likelihood

of needing external financing. Median differences are similar but indicate

that the fraud firms have fewer TOTAL ASSETS, have lower EARNINGS TO

PRICE, and have lower BOOK TO MARKET ratios. The lower accounting

returns and greater need for external financing for the alleged fraud firms

is consistent with the idea that these firms are somewhat troubled, possi-

bly providing at least a part of the motivation for the alleged accounting

fraud.

4.4 MULTIVARIATE TESTS

In this section, we examine whether executive compensation is associated

with alleged accounting fraud after controlling for corporate governance

factors, the desire for external financing, performance metrics, firm size,

return volatility, and other factors. The conceptual model for the logit anal-

ysis is as follows:

Pr(Fraud) = f (equity incentives, governance, performance metrics,

external financing, other controls). (1)

We begin with the analysis of the 50 fraud firms compared with the matched

sample of firms not accused of accounting fraud.

4.4.1. Matched Sample Results. Table 5 presents the results using the

matched sample design. The dependent variable equals one if the firm

was accused of accounting fraud by the SEC and zero otherwise. Neither

SENSITIVITY nor VESTED STOCK AND OPTION SENSITIVITY is signifi-

cantly associated with accounting fraud in any of the regressions, which we

estimate both with and without control variables. Columns A and B present

the results for the SENSITIVITY variable without and with control variables,

respectively. Columns C and D present the results for the variable VESTED

STOCK AND OPTION SENSITIVITY in order to test the association with the

equity holdings that the executives can sell immediately. Among the con-

trol variables, STOCK VOLATILITY is significant while the other control

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

TABLE 5

Results of a Logit Regression Comparing 50 Firms Accused of Fraud by the SEC with the Matched Sample ofFirms

ABCD

Predicted Chi-Square Chi-Square Chi-Square Chi-Square

Variable Sign Coefficient p-value Coefficient p-value Coefficient p-value Coefficient p-value

INTERCEPT -0.736 0.000 -2.047 0.107 -0.731 0.000 -2.040 0.108

SENSITIVITY ? 0.020 0.322 0.001 0.721

VESTED STOCK AND OPTION SENSITIVITY ? 0.021 0.346 0.008 0.755

CEO=CHAIR + 0.146 0.391 0.146 0.391

NUMMTGS -0.033 0.306 -0.033 0.307

FINANCING + 0.478 0.376 0.475 0.377

LEVERAGE + 2.250 0.084 2.243 0.084

MARKET VALUE OFEQUITY ? 0.000 0.942 0.000 0.961

ALTMAN'S Z + -0.310 0.130 -0.312 0.129

BOOK TO MARKET ? -0.351 0.621 -0.355 0.617

EARNINGS TO PRICE ? 2.227 0.443 2.220 0.444

RETURN ON ASSETS ? -0.120 0.935 -0.112 0.939

SALES GROWTH ? 0.807 0.096 0.806 0.096

AGE OFFIRM - -0.013 0.209 -0.013 0.207

M&A INFIRST YEAR OF FRAUD + 0.474 0.169 0.475 0.168

STOCK VOLATILITY + 2.729 0.017 2.7291 0.017

CEO TENURE ? -0.012 0.770 -0.0117 0.769

MISSING CEO TENURE ? 1.393 0.057 1.3983 0.056

Model Likelihood Ratio 1.066 0.302 34.450 0.005 0.950 0.330 34.419 0.005

N 150 150 150 150

This table presents the results of logistic regressions where the dependent variable is an indicator variable set equal to one for firms accused of financial accounting fraud and

zero otherwise. Our sample includes 50 firms accused of financial accounting firms and 100 firms matched to the alleged fraud firms based on industry (2 digit SIC code), year, and

firm size (total assets) in the year prior to the alleged fraud. All variables are defined in table 4. All p-values are two tailed except where a sign is predicted.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 135

variables are insignificant. These results suggest that once industry and size

have been controlled for via matching, uncertainty in the operating envi-

ronment as proxied by STOCK VOLATILITY is the only significant predictor

of fraud.

4.4.2. Unmatched Sample Results. In this section, we present the results of

the same tests described in the prior section, but we now compare the 50

fraud firms with the broad, unmatched sample (13,000+ firm-year obser-

vations) of nonfraud firms with data on ExecuComp. Table 6 presents the

basic regression results. The results in table 6 columns A and C reveal a sig-

nificantly positive association between fraud and SENSITIVITY (p = 0.007),

and VESTED STOCKAND OPTION SENSITIVITY (p = 0.012) when no con-

trol variables are included in the regression. Thus, in a large sample with no

additional controls, there is a statistically positive association between exec-

utives' equity portfolio incentives and the likelihood of committing fraud.

We then add the control variables to the regression. Because the sample is

unmatched, we also include industry fixed effects at the one-digit SIC level

(coefficients are not tabulated in the interest of brevity). After adding these

controls, we find that, consistent with table 5, SENSITIVITY and VESTED

STOCK AND OPTION SENSITIVITY are not significantly associated with ac-

counting fraud (p = 0.266 and p = 0.242, respectively). The coefficients on

several control variables are significant. For example, the fraud firms gener-

ally have higher STOCK VOLATILITY (p = 0.005, one-tailed) and are more

likely to have sales attributable to an acquisition in the first year of fraud (p =

0.024, one-tailed).

4.5 MANAGERIAL BENEFITS DERIVED FROM THE ACCOUNTING FRAUD

We also examine ex post stock option exercises and stock sales to deter-

mine whether there is evidence that executives at fraud firms sold more

equity than executives at nonfraud firms. There are reasons to believe that

even if equity incentives are associated with fraud, ex post stock sales and

option exercises would not be associated with fraud. For example, execu-

tives may not want to draw attention to the fraud by engaging in unusual

levels of selling. In addition, if the fraud is perpetrated in order to buy some

time to become profitable or increase (nonfraudulent) performance and

the executive does this with the belief that they will not be caught, then the

executive will want to delay sales until the firm's price is even higher than

what it is during the fraud period. Finally, executives can use the shares

as collateral for loans, thus extracting wealth without the actual sale of the

stock. For these reasons, we believe the ex ante measures of incentives we

use in our main analysis provide a better test of our research question. How-

ever, we perform the tests on sales of shares and exercise of stock options

for completeness.

To provide empirical evidence on these questions, we compare option

exercises, stock sales, and cash compensation over the fraud period for the

fraud firms relative to the nonfraud firms. We report these results in table 7.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

TABLE 6

Results of a Logit Regression Comparing 50 Firms Accused ofFraud by the SEC with the Unmatched Sample

ABCD

Predicted Chi-Square Chi-Square Chi-Square Chi-Square

Variable Sign Coefficient p-value Coefficient p-value Coefficient p-value Coefficient p-value

INTERCEPT -5.703 0.000 -7.084 0.000 -5.684 0.000 -7.095 0.000

SENSITIVITY ? 0.101 0.007 0.060 0.266

VESTED STOCK AND OPTION SENSITIVITY ? 0.109 0.012 0.067 0.242

CEO=CHAIR + 0.596 0.041 0.600 0.040

NUMMTGS -0.027 0.299 -0.026 0.306

FINANCING + 0.848 0.052 0.848 0.052

LEVERAGE + -0.760 0.199 -0.755 0.200

MARKET VALUE OFEQUITY ? 0.000 0.299 0.000 0.228

ALTMAN'S Z + -0.257 0.062 -0.256 0.062

BOOK TO MARKET ? -0.452 0.339 -0.455 0.337

EARNINGS TO PRICE ? 2.222 0.174 2.232 0.171

RETURN ON ASSETS ? -2.919 0.052 -2.916 0.052

SALES GROWTH ? 0.371 0.172 0.372 0.171

AGE OF FIRM -0.018 0.068 -0.018 0.067

M&A IN FIRST YEAR OF FRA UD + 0.668 0.024 0.666 0.024

STOCK VOLATILITY + 1.849 0.005 1.861 0.005

CEO TENURE ? -0.022 0.407 -0.022 0.401

MISSING CEO TENURE ? 0.286 0.525 0.277 0.540

Model Likelihood Ratio 5.246 0.022 67.824 0.000 4.609 0.032 67.934 0.000

N 13,083 13,083 13,083 13,083

This table presents the results of logistic regressions where the dependent variable is an indicator variable set equal to one for firm-years where the firm is accused of financial

accounting fraud and zero otherwise. Our sample includes 50 firms accused of financial accounting fraud and 13,033 firm-years (all other firm-years on ExecuComp) for which no

accusation of fraud is made. All variables are defined in table 4. One digit industry indicator variables are included in the regression estimation but coefficients are suppressed above

for simplicity. All p-values are two tailed except where a sign is predicted.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

TABLE 7

Stock Option Exercises, Stock Sales, and Cash Compensation during Fraud Period

Alleged Fraud Firms versus Matched Sample Alleged Fraud Firms versus Unmatched Sample

Alleged Fraud Matched Alleged Fraud Unmatched

Firm Sample Sample Difference p-value Firm Sample Sample Difference p-value

Stock Options Exercises

# options exercisedt/# exercisable optionstl

Mean 24.4% 27.3% -2.9% 0.645 24.4% 22.9% 1.5% 0.774

Median 0.7% 9.4% -8.7% 0.009 0.7% 5.2% -4.5% 0.085

# options exercisedt/ (# total options heldt-1 + # option grantst)

Mean 7.4% 8.3% -0.9% 0.640 7.4% 8.8% -1.4% 0.411

Median 0.8% 4.0% -3.2% 0.009 0.8% 2.2% -1.4% 0.049

# options exercisedt/ (# options exercisedt + # ending exercisable optionst)

Mean 12.3% 14.3% -2.0% 0.396 12.3% 13.2% -0.9% 0.643

Median 1.7% 6.8% -5.1% 0.017 1.7% 4.1% -2.4% 0.079

$ realized from exerciset/ ($ realized from exerciset + intrinsic value of exercisable optionst)

Mean 22.5% 21.9% 0.6% 0.873 22.5% 20.0% 2.5% 0.414

Median 3.0% 10.2% -7.2% 0.115 3.0% 7.5% -4.5% 0.369

Stock Sales

# net shares soldt/# shares heldt-1

Mean 0.3% 2.9% -2.6% 0.258 0.3% -0.6% 0.9% 0.664

Median 0.0% 0.0% 0.0% 0.363 0.0% 0.0% 0.0% 0.783

$ value of net shares soldt/$ value of shares heldt-1

Mean 1.2% 2.9% -1.7% 0.389 1.2% -0.5% 1.7% 0.365

Median 0.0% 0.0% 0% 0.816 0.0% 0.0% 0.0% 0.538

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

T ABLE 7 - Continued

Alleged Fraud Firms versus Matched Sample Alleged Fraud Firms versus Unmatched Sample

Alleged Fraud Matched Alleged Fraud Unmatched

Firm Sample Sample Difference p-value Firm Sample Sample Difference p-value

Cash Compensation during Fraud Period (in $ millions)

Salary + bonus over fraud period

Mean $4.478 $3.742 $0.736 0.104 $4.478 $3.142 $1.336 0.001

Median $2.434 $2.804 -$0.370 0.716 $2.434 $2.506 -$0.072 0.331
Salary + bonus scaled by average total assets over the fraud period

Mean 0.7% 0.7% 0.0% 0.957 0.7% 0.5% 0.2% 0.113

Median 0.4% 0.4% 0.0% 0.377 0.4% 0.3% 0.1% 0.425
Salary + bonus + other cash compensation over fraud period

Mean $5.217 $3.960 $1.257 0.022 $5.217 $3.295 $1.922 0.001
Median $2.854 $2.917 -$0.063 0.489 $2.854 $2.562 $0.292 0.108
Salary + bonus + other cash compensation scaled by average total assets

Mean 0.7% 0.7% 0.0% 0.834 0.7% 0.5% 0.2% 0.046

Median 0.4% 0.4% 0.0% 0.534 0.4% 0.3% 0.1% 0.692

Difference in means p-values is based upon a t-test, while difference in medians is based upon Wilcoxon signed rank test. Variables are defined as follows: #
options exercisedt is the variable SOPTEXSH from ExecuComp; exercisable optionst_- is the variable UEXNUMEX from ExecuComp; total options heldtl
is the sum of the variables UEXNUMEX and UEXNUMUN and current grants is the variable SOPTGRNT from ExecuComp; $ realized from exercise is the
variable SOPTEXER from ExecuComp and the intrinsic value of exercisable optionstl is the variable INMONEX from ExecuComp. Salary and Bonus are
both from ExecuComp variables of the same name. Assets are taken from Compustat (data item #6). For stock sales we gathered the data from Thompson

Financial Insider Filing Data Feed. The number of shares sold is the net of acquisitions (negative values) and sales, excluding sales of shares just gained from
option exercise (positive values) (variable NSHR). The shares held are as of the beginning of the year (SHRHELD). The value of the transactions is computed
as the stock price at the time of the sale or purchase (TPRICE) multiplied by the number of shares in the transaction. All variables are summed over the top 5
executives for ExecuComp variables and over all officers for Thompson variables. All variables are winsorized (reset) at 99% and 1%.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 139

We find no evidence that fraud firm executives exercise stock options (mea-

sured in number of options and value) to a greater degree than executives

at the nonfraud firms in either the matched sample or the unmatched sam-

ple. For example, fraud firm executives exercise approximately 24% of their

beginning of the year exercisable options while the matched (unmatched)

sample executives exercise 27% (23%). The median nonfraud firm execu-

tive actually exercised more options than the median fraud firm executive

during the fraud period; however, these results are only marginally signifi-

cant for the unmatched sample. Table 7 also indicates that fraud firm exec-

utives do not sell significantly more (measured at the mean and median) of

their stock holdings than do managers at nonfraud firms.

In terms of cash compensation, we find that fraud firm executives earned,

on average, $4.48 million per year over the fraud period in terms of salary

and bonus and $5.22 million per year over the fraud period in terms of

total cash compensation.14 With the average number of years of fraud for

our sample being 2.3 years, this translates into the top five executives in our

average firm earning about $12 million dollars in total cash compensation

while perpetrating the fraud. In contrast, the executives at the matched

sample of firms earned only $3.96 million per year on average in total cash

compensation during this same period. This difference is significant at the

0.02 level, but the medians are not significantly different (p = 0.49). In

order to control for size differences we scale the compensation measures

by the total assets of the firm. After doing this, we find no evidence of a

statistical difference in the annual cash compensation of the fraud firms

compared with the matched sample of nonfraud firms during the fraud

period and a 0.2% difference, on average, in the scaled measure relative to

the unmatched sample (p = 0.046).

We interpret the data in table 7 as failing to support the conclusion that

fraud firm managers exercised more stock options during the fraud period

than did managers at nonfraud firms. Similarly, consistent with the findings

of Dechow, Sloan, and Sweeney [1996], we interpret the data in table 7 as

failing to support the conclusion that fraud firm managers sold more of

their stock holdings than did managers at nonfraud firms.

5. Conclusion

We examine whether the incidence of alleged accounting fraud is associ-

ated with executive equity incentives. Some of the largest accounting frauds

in history occurred in the last several years, leading to the well-known up-

heaval in the accounting industry and sweeping legislative and regulatory

changes. These events have left legislators, regulators, practitioners, and

academics searching for causes of these frauds. Understanding the underly-

ing forces that gave rise to these frauds is a necessary precursor to effectively

14 Total cash compensation is computed as the sum of the variables salary + bonus + all other

pay from ExecuComp.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

140 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

preventing future occurrences. Many have suggested that the explanation

lies in the incentives and opportunities for personal gain faced by executives.

Our sample consists of firms explicitly accused of accounting fraud by the

SEC during the periodJanuary 1996 to November 2003. After requiring the

availability of certain financial data, we have a final sample of 50 firms that

were accused of fraud by the SEC, which we call the fraud firms. We compare

the fraud firms with two samples of firms that were not accused of fraud.

First, we compare the fraud firms with a matched sample of firms in which

the matches are based on size, year, and industry. Second, we compare the

fraud firms with an unmatched sample of firms consisting of all remaining

firm-years on ExecuComp.

We estimate a variety of equity incentive measures for the top five execu-

tives at each firm, focusing specifically on the expected change in value of

the executives' stock and option-based portfolio to a 1% stock price change

(which we denote "sensitivity"). In additional analyses, we examine whether

there is evidence that executives at firms accused of fraud systematically

sell stock and exercise options during the period of the alleged fraud to a

greater extent than executives in the matched and unmatched samples of

control firms.

The empirical analysis reveals no consistent evidence to support the con-

clusion that the probability of accounting fraud is increasing in the sensitivity

of executives' total equity or vested stock and stock option-based wealth to

changes in stock prices. We also find that managerial exercises of stock op-

tions and managerial stock sales are not significantly higher for fraud firms

than for nonfraud firms. Thus, these results stand in contrast to assertions

by policymakers that incentives from stock-based compensation and the re-

sulting equity holdings increase the likelihood of accounting fraud.

REFERENCES

ABOODY, D., AND R. KASZNIK. CEO "Stock Option Awards and the Timing of Corporate Volun-

tary Disclosures."Journal of Accounting and Economics 29 (2000): 73-100.

ACKMAN, D. "Bernie Ebbers Guilty." Forbes.com. March 15, 2005.

ADAMS, R. B. "What Do Boards Do? Evidence from Committee Meeting and Director Compen-

sation." Working paper, Federal Reserve Bank of New York, 2000.

ALTMAN, E. "Financial Ratios, Discriminant Analysis and the Prediction of Corporate

Bankruptcy." Journal of Finance 23 (1968): 589-609.

BAR-GILL, O., AND L. BEBCHUK. "Misreporting Corporate Performance." Working paper, Har-

vard Law School, 2003.

BEASLEY, M. "An Empirical Analysis of the Relation Between Board of Director Compensation

and Financial Statement Fraud." The Accounting Review 71 (1996): 443-65.

BEBCHUK, L.; J. FRIED; AND D. WALKER. "Managerial Power and Rent Extraction in the De-

sign of Executive Compensation." The University of Chicago Law Review 69 (2002): 751-

846.

BEGLEY, J.; J. MING; AND S. WATTS. "Bankruptcy Classification Errors in the 1980s: An Empir-

ical Analysis of Altman's and Ohlson's Models." Review of Accounting Studies 1 (1996): 267-

84.

BENEISH, M. "Incentives and Penalties Related to Earnings Overstatements that Violate GAAP."

The Accounting Review 74 (1999): 425-57.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 141

BOUND, J.; D. JAEGER; AND R. BAKER. "Problems with Instrumental Variables Estimation When

the Correlation Between the Instruments and the Endogenous Explanatory Variable Is

Weak." Journal of the American Statistical Association 90 (1995): 443-50.

BRICKLEY, J.; S. BHAGAT; AND R. LEASE. '"The Impact of Long-Range Managerial Compen-

sation Plans on Shareholder Wealth." Journal of Accounting and Economics 7 (1985): 115-

29.

CHENG, Q., AND T. WARFIELD. "Stock-Based Compensation, Earnings Management and the

Informativeness of Earnings." The Accounting Review 80 (2005): 441-76.

CORE,J., AND W. GUAY. "The Use of Equity Grants to Manage Optimal Equity Incentive Levels."

Journal of Accounting and Economics 28 (1999): 151-84.

CORE,J., AND W. GUAY. "Estimating the Value of Employee Stock Option Portfolios and Their

Sensitivities to Price and Volatility."Journal of Accounting Research 40 (2002): 613-30.

CORE, J.; R. HOLTHAUSEN; AND D. LARCKER. "Corporate Governance, Chief Executive Offi-

cer Compensation, and Firm Performance." Journal of Financial Economics 51 (1999): 371-

406.

DECHOW, P., AND R. SLOAN. "Executive Incentives and the Horizon Problem: An Empirical

Investigation."Journal of Accounting and Economics 14 (1991): 51-89.

DECHOW, P.; R. SLOAN; AND A. SWEENEY. "Detecting Earnings Management." The Accounting

Review 70 (1995): 193-225.

DECHOW, P.; R. SLOAN; AND A. SWEENEY. "Causes and Consequences of Earnings Manipulations:

An Analysis of Firms Subject to Enforcement Actions by the SEC." Contemporary Accounting

Research 13 (1996): 1-36.

DEMSETZ, H., AND K. LEHN. "The Structure of Corporate Ownership: Causes and Conse-

quences." Journal of Political Economy 93:6 (1985): 1155-77.

DENIS, D.; P. HANOUNA; AND A. SARIN. "Is There a Dark Side to Incentive Compensation?"

Working paper, Purdue University and Santa Clara University, 2005.

EFENDI, J.; A. SRIVASTAV; AND E. SWANSON. "Why Do Corporate Managers Misstate Financial

Statements? The Role of Option Compensation, Corporate Governance, and Other Factors."

Working paper, Texas A&M University, 2004.

ERICKSON, M.; B. MAYHEW; AND W. FELIX. "Why Do Audits Fail: Evidence from Lincoln Savings

and Loan."Journal of Accounting Research 38 (2000): 165-94.

ERICKSON, M., AND S. WANG. "Earnings Management by Acquiring Firms in Stock for Stock

Mergers. "Journal of Accounting and Economics 27 (1999): 149-76.

Federal Reserve Board (FRB). Testimony of Alan Greenspan, Federal Reserve Board's Semiannual

Monetary Policy Report to the Congress before the Committee on Banking, Housing, and Urban Affairs.

U.S. Senate July 16, 2002. Chairman Greenspan presented identical testimony before the

Committee on Financial Services, U.S. House of Representatives, on July 17, 2002.

FEROZ, E. H.; K. PARK; AND V. S. PASTENA. "The Financial and Market Effects of the SEC's

Accounting and Auditing Enforcement Releases." Journal of Accounting Research 29 (Supple-

ment 1991): 107-42.

GAO, P., AND R. SHRIEVES. "Earnings Management and Executive Compensation: A Case of

Overdose of Option and Underdose of Salary?" Working paper, Northwestern University

and University of Tennessee at Knoxville, 2002.

GAVER, J., AND K. GAVER. "Additional Evidence on the Association Between the Investment

Opportunity Set and Corporate Financing, Dividend, and Compensation Policies." Journal

of Accounting and Economics 16 (1993) 125-60.

GOLDMAN, E., AND S. SLEZAK. "An Equilibrium Model of Incentive Contracts in the Presence

of Information Manipulation. "Journal of Financial Economics (2006): Forthcoming.

GOMPERS, P.; J. ISHII; AND A. METRICK. "Corporate Governance and Stock Prices." Quarterly

Journal of Economics 118 (2003): 107-55.

GUAY, W.; S. KOTHARI; AND R. WATTS. "A Market-Based Evaluation of Discretionary Accrual

Models. "Journal of Accounting Research 34 (Supplement 1996): 83-105.

HALL, B.J., AND K. MURPHY. "Stock Options for Undiversified Executives. "Journal ofAccounting

and Economics 33 (2002): 3-42.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

142 M. ERICKSON, M. HANLON, AND E. L. MAYDEW

HANLON, M.; S. RAJGOPAL; AND T. SHEVLIN. "Are Executive Stock OptionsAssociated with Future

Earnings?" Journal of Accounting and Economics 36 (2003): 3-43.

HEALY, P., ANDJ. WAHLEN. "A Review of the Earnings Management Literature and Its Implica-

tions for Standard Setting." Accounting Horizons 13 (1999): 365-83.

HERMALIN, B. E., AND M. S. WEISBACH. "Endogenously Chosen Boards of Directors and Their

Monitoring of the CEO." American Economic Review 88 (1998): 96-118.

HIMMELBERG, C.; G. HUBBARD; AND D. PALIA. "Understanding the Determinants of Managerial

Ownership and the Link Between Ownership and Performance."Journal ofFinancialEconomics

53 (1999): 353-84.

ITTNER, C.; R. LAMBERT; AND D. LARCKER. "The Structure and Performance Consequences of

Equity Grants to Employees of New Economy Firms." Journal of Accounting and Economics 34

(2003): 89-127.

JENSEN, M. "Paying People to Lie: The Trust About the Budgeting Process." European Financial

Management 9 (2003): 379-406.

JENSEN, M., AND W. MECKLING. "Theory of the Firm: Managerial Behavior, Agency Costs and

Ownership Structure."Journal of Financial Economics 3 (1976): 305-60.

JENTER, D. "Understanding High-Powered Incentives." Working paper, Harvard Business

School, 2001.

JOHNSON, S.; H. RYAN; AND Y TIAN. "Executive Compensation and Corporate Fraud." Working

paper, Louisiana State University, 2005.

KE, B. "Why Do CEOs of Publicly Traded Firms Prefer Reporting Small Increases in Earnings

and Long Duration of Consecutive Earnings Increases?" Working paper, Pennsylvania State

University, 2002.

KEDIA, S. "Do Executive Stock Options Generate Incentives for Earnings Management?

Evidence from Accounting Restatements." Working paper, Harvard Business School,

2003.

LARCKER, D., AND T. RusTICus. "On the Use of Instrumental Variables in Accounting Research."

Working paper, University of Pennsylvania, 2005.

LEWELLEN, W.; C. LODERER; AND A. ROSENFELD. "Merger Decisions and Executive Stock Own-

ership in Acquiring Firms." Journal of Accounting and Economics 7 (1985): 209-31.

Louis, H. "Earnings Management and the Market Performance of Acquiring Firms."Journal of

Financial Economics 74 (2004): 121-48.

MEULBROEK, L. "The Efficiency of Equity-Linked Compensation: Understanding the Full Cost

of Awarding Executive Stock Options." Financial Management 30 (2001): 5-30.

MURPHY, K. J. "Executive Compensation," in Handbook of Labor Economics, edited by O.

ASHENFLETER AND D. CARD. Amsterdam: North Holland, 1999: 2485-2563.

PALMROSE, Z., AND S. SCHOLZ. "The Accounting Causes and Legal Consequences of Non-GAAP

Reporting: Evidence from Restatements." Contemporary Accounting Research 21 (2002): 139-

80.

PENG, L., AND A. ROELL. "Executive Pay, Earnings Manipulation and Shareholder Litigation."

Working paper, Baruch College/CUNY and Princeton University, 2003.

PINCUS, K.; W. HOLDER; AND T. MOCK. Reducing the Incidence of Fraudulent Financial Reporting:

The Role of the Securities and Exchange Commission. Los Angeles, CA: SEC Financial Reporting

Institute of the University of California, 1988.

RAJGOPAL, S., AND T. SHEVLIN. "Empirical Evidence on the Relation Between Stock Op-

tion Compensation and Risk Taking." Journal of Accounting and Economics 33 (2002): 145-

71.

RICHARDSON, S.; I. TUNA; AND M. WU. "Predicting Earnings Management: The Case of Earnings

Restatements." Working paper, University of Pennsylvania and Hong Kong University of

Science and Technology, 2003.

SHUMWAY, T. "Forecasting Bankruptcy More Accurately: A Simple Hazard Model." The Journal

of Business 74 (2001): 101-24.

SMITH, C., AND R. WATTS. "The Investment Opportunity Set and Corporate Financing,

Dividend, and Compensation Policies." Journal of Financial Economics 7 (1992): 117-

61.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions

EXECUTIVE EQUITY INCENTIVES AND ACCOUNTING FRAUD 143

VAFEAS, N. "Board Meeting Frequency and Firm Performance." Journal of Financial Economics

52 (1999): 113-42.

YERMACK, D. "Do Corporations Award CEO Stock Options Effectively?" Journal of Financial

Economics 39 (1995): 237-69.

WRMACK, D. "Good Timing: CEO Stock Option Awards and Company News Announcements."

Journal ofFinance 52 (1997): 449-77.

This content downloaded from 202.43.95.117 on Tue, 22 Mar 2016 04:24:22 UTC
All use subject to JSTOR Terms and Conditions