CFOs versus CEOs:Equity incentives and crashes$
Jeong-Bon Kim a,1,Yinghua Li b,Liandong Zhang a,n
a Department of Accountancy,City University of Hong Kong,Tat Chee Avenue,Kowloon,Hong Kong
b Baruch College,The City University of New York,One Bernard Baruch Way,New York,NY10010-5585,USA
a r t i c l e i n f o
Article history:
Received8October2010
Received in revid form
18November2010
Accepted12December2010
Available online8April2011
JEL classification:
D89
G12
G17
G34
M52
Keywords:
Equity incentives
Crash risk
Compensation
Corporate governance
CFO
早上跑步能减肥吗a b s t r a c t
Using a large sample of U.S.firms for the period1993–2009,we provide evidence that
the nsitivity of a chieffinancial officer’s(CFO)option portfolio value to stock price is
significantly and positively related to thefirm’s future stock price crash risk.In contrast,
wefind only weak evidence of the positive impact of chief executive officer option
nsitivity on crash risk.Finally,wefind that the link between CFO option nsitivity
and crash risk is more pronounced forfirms in non-competitive industries and tho怎样写学习计划
with a high level offinancial leverage.
&2011Elvier B.V.All rights rerved.
1.Introduction
大理自由行攻略
The paration of ownership and control in modern
corporations creates a conflict of interest between share-
holders and managers(Jenn and Meckling,1976).Execu-
tive compensation contracts are one potential mechanism
for aligning the interests of shareholders and managers.
However,Jenn and Murphy(1990b)show that during
the1970s and1980s,corporate managers were not paid for
performance but were compensated,like bureaucrats,bad
on the size of the organization.Furthermore,Jenn(1993)
argues that such inefficient executive pay schemes created
empire-building incentives,which exacerbated the excess
capacity problem that had emerged since19732and con-
tributed to the widespread value destruction by corporate
America during the era.To address this problem of empire
building,Jenn and Murphy(1990a)recommend increasing
the u of equity-bad compensation,which they believe
属鼠今年多大
to be an effective tool for aligning the interests of managers
and shareholders by exposing managers’wealth to their
firms’stock prices.Perhaps partially inspired by Jenn
and Murphy(1990a),the u of stock-and option-bad
Contents lists available at ScienceDirect
journal homepage:/locate/jfec
Journal of Financial Economics
0304-405X/$-e front matter&2011Elvier B.V.All rights rerved.
doi:10.1016/j.jfineco.2011.03.013
$We are especially grateful to Amy Hutton(our referee)and Bill
Schwert(the Editor)for their many insightful and constructive sugges-
tions.We appreciate helpful comments from Paul Brockman,Qiang
Cheng,Mo Khan,Franco Wong,Yong Yu,and the colleagues at City
University of Hong Kong.J.-B.Kim and L.Zhang acknowledge partial
financial support for this rearch from GRF grant(Project#:
CityU144310)and start-up grants of City University of Hong Kong.All
白痰咳嗽errors are our own.
n Corresponding author.Tel.:þ852********;fax:þ852********.江苏高考政策
E-mail address:jeong.bon.kim@cityu.edu.hk(J.-B.Kim),
yinghua.(Y.Li),liandong.zhang@cityu.edu.hk(L.Zhang).
1Tel:þ852********.
2Jenn(1993)describes the year1973as the beginning of the
‘‘Third Industrial Revolution.’’
Journal of Financial Economics101(2011)713–730
compensation incread dramatically during the 1990s.3Arguably,this shift in compensation structure toward greater pay for performance may have discouraged manage-rial empire-building behavior and contributed to the enor-mous value creation by porations in the 1990s.
However,the solution to yesterday’s problem can sow the eds of today’s problem.The widespread u of equity-bad compensation coincides with veral recent catastrophic events,including the dot-com bubble in the late 1990s,the 2001–2002corporate scandals,and the recent financial crisis.This unfortunate coincidence has led regulators,the media,and academics to question whether the large portfolios of stocks and options held by managers were the culprit in the financial disasters.Specifically,there is an argument that the high nsitivity of managers’wealth to stock price afforded by stock option holdings motivates managers to engage in short-termist behavior to inflate current share prices at the expen of long-term firm value (Bebchuk,2009).The primary purpo of this paper is to investigate whether equity incentives,particularly stock option incentives,are related t
o future firm-specific stock price crash risk.
Using a dynamic rational expectations model with asymmetric information,Benmelech,Kandel,and Veronesi (2010)show that stock-bad compensation not only induces managers to exert costly effort,but also incentivizes them to conceal bad news about future growth options.Such concealment of bad news can lead to vere over-valuation and a subquent crash in stock price.While bad news hoarding is a somewhat unintended conquence of the compensation contract in the Benmelech,Kandel,and Veronesi (2010)framework,Bolton,Scheinkman,and Xiong (2006)prent a multiperiod agency model showing that incumbent investors u stock-bad compensation to intentionally encourage managers to adopt short-termist behavior to boost the speculative component of the share price.The key tension of this model is the conflict of interest between the incumbent investors and future generations of investors.Although bad on somewhat different under-lying assumptions,the models of both Benmelech,Kandel,and Veronesi (2010)and Bolton,Scheinkman,and Xiong (2006)predict that equity incentives induce managers to engage in short-termist behavior,such as bad news hoard-ing,to inflate short-term share price.
Though not focusing on equity incentives,Jin and Myers (2006)and Bleck and Liu (2007)offer more analysis of how bad news hoarding can lead to stock price crashes.For example,Jin and Myers (200
6)argue that there is an upper limit to the amount of bad news that managers can absorb or successfully accumulate.When the accumu-lated bad news reaches this upper limit,it will come out all at once,leading to a large and sudden price decline.Moreover,Bleck and Liu (2007)argue that hiding bad news prevents investors and the board of directors from discerning negative net prent value (NPV)projects at an
early stage and forcing managers to take timely abandon-ment actions.As a result,the bad performance of negative NPV projects accumulates and eventually materializes,which results in ast price crashes.
This study conducts a simple test to show whether the predicted (positive)relation between equity incentives and crash risk is consistent with real world data.Our empirical strategy involves the identification of an empirical proxy for managers in the bad news hoarding and crash story.Previous empirical rearch on managerial compensation has largely focud on chief executive officers (CEO).However,recent rearch provides evidence suggesting that the incentives of chief financial officers (CFO)could be more influential in a decision tting where sophisticated finan-cial experti is required.For example,Jiang,Petroni,and Wang (2010)find that CFO equity incentives are more important than CEO incentives in determining earnings management.Chava and Purnanandam (201
0)show that while CEO incentives are associated with capital structure and cash holding decisions,CFO incentives dominate in debt maturity choices and earnings smoothing decisions.Becau our predicted link between equity incentives and crash risk is largely built on the manipulation of information flow by managers,we expect that CFO incentives play a more important role in this tting.However,Benmelech,Kandel,and Veronesi (2010),Kedia and Philippon (2009),and McNichols and Stubben (2008),among others,argue that managers also hide bad news by mimicking the investment behavior of firms with truly high growth poten-tial.Thus,CEO incentives may also be significant becau we expect CEOs to be more influential in investment decisions.Therefore,the differential impact of CFOs’versus CEOs’incentives on crash risk is an empirical question.
Following Bergstresr and Philippon (2006),we measure the strength of CEO/CFO equity-bad incentives as the dollar change in the value of the stock or option holdings of a CEO/CFO given a one-percentage-point increa in the com-pany stock price.Firm-level crash risk is proxied by the probability of extreme,negative firm-specific weekly returns,the negative skewness of firm-specific weekly returns,and the asymmetric volatility of negative and positive stock returns (Chen,Hong,and Stein,2001;Hutton,Marcus,and Tehranian,2009;Kim,Li,and Zhang,in press ).Using a sample of U.S.firms from the Compustat Executive Compensation databa (ExecuComp)during the period 19
93–2009,we find that the strength of CFO option incentives is significantly and positively related to future stock price crash risk.In contrast,we find only weak evidence that the strength of CEO option incentives is positively related to crash risk.More impor-tantly,we find that CFO option incentives dominate CEO option incentives in determining future crash risk when we include both CFO and CEO option incentives in the regression.This result suggests that CFOs are more influential in firms’bad news hoarding decisions.
Theoretical analys such as that of Benmelech,Kandel,and Veronesi (2010)do not discriminate between option and stock incentives.However,we find no significant empirical relation between the strength of stock incentives and crash risk.This result is consistent with Burns and Kedia (2006),who find that option incentives,but not stock incentives,have a significant
3
Other possible drivers for the increasing level of stock options include tax rules introduced in 1994(Section 162(m)of the Internal Revenue Code),the accounting treatment of stock options (the ‘‘intrinsic value’’method),and the lack of cash flows of new economy firms.See Hall and Murphy (2003)for more discussions on this issue.
J.-B.Kim et al./Journal of Financial Economics 101(2011)713–730
714
impact on afirm’s misreporting.Thus,the results suggest that option holdings provide more powerful incentives for managers to inflate short-term share prices, possibly becau the loss to manager wealth from option holdings is limited in the event of a stock price crash.
Giroud and Mueller(2010,2011)find that the incen-tive provided by afirm’s corporate governance system matters only if thefirm operates in non-competitive industries.The authors suggest that rearchers take into consideration the disciplining role of the product market when examining the conquence of corporate govern-ance arrangements.Following their suggestion,we con-duct subsample analysis forfirms with high and low product market competition.Consistent with the argu-ment that agency problems are more vere for non-competitive industries,wefind that the positive relation between CFO option incentives and crash risk is signifi-cant only for the subsample offirms in non-competitive industries.
Finally,some regulators and academics argue that excessive risk taking induced by stock options contributed to the recentfinancial ,Bebchuk,2009). Motivated by this argument,we examine
whether the nsitivity of option portfolio values to stock return ,vega)is positively related to crash risk. However,wefind no evidence of such a relation.Expand-ing on our bad news hoarding story,we conjecture that the hiding of excessive risk taking rather than risk taking itlf leads to crashes.4Usingfinancial leverage as a proxy for the ex ante incentive to mask risk taking,we show that the positive relation between CFO option incentives and crash risk exists only for the subsample offirms with high ,firms with high ex ante incentive to mask risks).5
This study contributes to the literature by identifying a perver effect of option-bad compensation:it can increa future stock price crash risk.The results in this paper are largely supportive of the predictions of the theoretical model of Benmelech,Kandel,and Veronesi (2010),with two important differences.First,Benmelech, Kandel,and Veronesi(2010)predict that both stock and option incentives can increa crash risk,while wefind that only option incentives matter empirically.Second, Benmelech,Kandel,and Veronesi(2010)do not discrimi-nate between CEO and CFO incentives,while wefind that CFOs’option incentives dominate tho of CEOs in deter-mining crash risk.This difference between CEO and CFO incentives has attracted increasing attention from the recent literature,partially motivated by the recent Secu-rities and Exchange Commission(SEC)rule of requiring the disclosure of CFO pay(Chava and Purnanandam,2010; Jiang,Petroni,and Wang,2010).
Our study also extends a long and large literature on the relation between equity incentives andfirm performance orfirm value(Core,Guay,and Larcker,2003; Frydman and Jenter,2010;Murphy,1999).Using account-ing return or Tobin’s q as a proxy forfirm performance,this literature,in general,finds some inconclusive evidence, partly due to difficulties in rearch design.6Instead of focusing on the average and concurrent valuation effect,we examine the impact of equity incentives on future extreme outcomes.This exerci can add significantly to the litera-ture,becau extreme outcomes reflect an extraordinary cumulative effect that can provide more valuable insight into the true nature of a phenomenon(Kim,Li,and Zhang, in press;Taleb,2007).
This paper is cloly related to tho of Bergstresr and Philippon(2006),Burns and Kedia(2006),and Jiang, Petroni,and Wang(2010),which link CEO or CFO equity incentives to earnings management.However,our study provides additional insights into this literature,becau earnings management is only one of many ways for managers to withhold bad news(Hutton,Marcus,and Tehranian,2009).7For example,Kim,Li,and Zhang (forthcoming)argue that managers can also hide bad news through complex tax shelters.In our empirical tests, we explicitly control for earnings management andfind that the relation between option incentives and crash risk remains significant.
Finally,our rearch is related to the emerging litera-ture examining the caus for the recentfinancial crisis, although our investigation is more general and does not focus on the crisis.Fahlenbrach and Stulz(2010)investi-gate whether bank performance during the crisis is related to CEO incentives before the crisis andfind evidence that CEOs with stronger equity incentives per-formed wor during thefinancial crisis for a sample of large banks.8Our results of the positive relation between managers’incentives and crash risk are somewhat con-sistent with the authors’findings.
The remainder of the paper is structured as follows. Section2conducts a review of the related literature. Section3describes the data and the measurement of key variables.Section4prents the empirical analysis and Section5prents our conclusions.
2.Related literature and empirical predictions
Stock and option holdings tie managers’wealth to a firm’s stock price,and have long been viewed as an effective tool to align managers’incentives to shareholder interests.One of the underlying assumptions for this belief is that stock price is an unbiad indicator of thefirm’s
4Here,excessive risk taking is considered to be bad news,becau the disclosure of it can depress s
tock prices.
5Financial leverage could be a proxy for many other things,and thus we suggest that readers exerci caution in accepting our explana-tions.Plea refer to Section4.3.3for more details on the motivation,the design,and the limitation of this test.
6Section2provides a short review of this literature.
7In addition,some recent studies fail tofind a significant positive relation between equity incentives and ,Armstrong, Jagolinzer,and Larcker,2010;Erickson,Hanlon,and Maydew,2006). Thus,mixed evidence of the relation between equity incentives and earnings management also makes our study more necessary.Moreover, thefinancial statement’s bottom line is only one of many ways of conveying information(Lambert,2010).
8In particular,the authorsfind that banks with a larger‘‘dollar gain fromþ1%’’(a measure similar to the equity incentive measures in our paper)had significantly lower stock and accounting performance during the crisis.
J.-B.Kim et al./Journal of Financial Economics101(2011)713–730715
fundamental value.However,there is both anecdotal and empirical evidence showing that managers can manipu-late market expectations and that a firm’s stock price can deviate from its fundamental value for an extended period of time (Peng and Roell,2008).For example,Enron’s managers were able to conceal Enron’s bad performance through means such as earnings management,tax shelter-ing,and related party transactions,which led the market to overvalue Enron’s stock for a prolonged period in the late 1990s.Empirically,Sloan (1996)shows that managers can manipulate the share price through accounting accruals.The above evidence suggests that equity incen-tives can have the perver effect of inducing managers to inflate a firm’s short-term share price without improving its true underlying performance.
Recent literature on the conquences of equity incen-tives has focud on earnings management.The main-tained assumption of this line of rearch is that managers can successfully inflate share prices by manip-ulating financial statement bottom lines.9Cheng and Warfield (2005),Bergstresr and Philippon (2006),Burns and Kedia (2006),and Efendi,Srivastava,and Swanson,2007),among others,show a positive relation between CEO equity incentives and earnings manage-ment.More recently,Chava and Purnanandam (2010)and Jiang,Petroni,and Wang (2010)provide an important extension to this line of rearch by showing that CFO incentives domina
te CEO incentives in determining earn-ings management.Admittedly,the evidence in the litera-ture is not unanimous.For instance,Erickson,Hanlon,and Maydew (2006)find no significant positive relation between executive equity incentives and accounting irre-gularities.Moreover,Armstrong,Jagolinzer,and Larcker (2010)find a modest negative relation between executive equity incentives and accounting fraud.The mixed evidence above may reflect the difficulties in capturing managerial earnings management behavior empirically,a problem that has long been recognized by the accounting literature (Dechow,Ge,and Schrand,2010).
A number of earlier studies examine the relation between equity incentives and firm value,again produ-cing mixed results.Morck,Shleifer,and Vishny (1988)and McConnell and Servaes (1990)show nonmonotonic rela-tions between managerial ownership and firm value.Mehran (1995)finds that firm performance is positively related to the percentage of equity held by management and to the percentage of their compensation that is equity-bad.Habib and Ljungqvist (2005)find a positive relation between firm value and CEO stock holdings,but a negative relation between firm value and option holdings.In contrast,Himmelberg,Hubbard,and Palia (1999)find no relation between managerial ownership and firm performance after controlling for firm fixed effects.However,Zhou (2001)argues that fixed effects estimators may not detect an effect of managerial ownership on performance,even if the effect exists,becau fixed
effects estimation relies on within-firm variations and managerial ownership typically changes slowly over time within a firm.
This paper extends the prior rearch by examining the relation between equity incentives and future stock price crash risk.Our empirical exerci is mainly moti-vated by the theoretical predictions in the model of Benmelech,Kandel,and Veronesi (2010).Using a hidden action model,the authors show that equity incentives induce managers to conceal bad news about future growth options,and this bad news hoarding by managers leads to an overvaluation of a firm’s stock,which eventually results in a crash of the stock price.In a similar vein,Jin and Myers (2006)show analytically that the hoarding and accumulation of bad news for an extended period lead to an abrupt decline in stock price when a tipping point is crosd.
Moreover,managers’short-termist behaviors are not limited to the manipulation of financial information.To support the preten of strong investment opportunities,managers can also choo suboptimal investment policies (Benmelech,Kandel,and Veronesi,2010;McNichols and Stubben,2008;Kedia and Philippon,2009).For instance,McNichols and Stubben (2008)find that managers over-invest in property,plant,and equipment in the period of overstated earnings.Similarly,Kedia and Philippon (2009)show that during periods of inflated performance,firms
hire and invest excessively.In Benmelech,Kandel,and Veronesi’s (2010)theoretical model,the suboptimal invest-ment policy after the slowdown in growth rate eventually leads to undercapitalization and a stock price crash.
Bad on Benmelech,Kandel,and Veronesi (2010),we predict that managerial equity incentives are positively related to future crash risk.In addition,our empirical exerci also incorporates recent developments in the executive compensation literature.Specifically,following Jiang,Petroni,and Wang (2010),we parately and jointly examine the association between CFO and CEO equity incentives and crash risk.Fuller and Jenn (2010)argue that increasing the proportion of stock options in execu-tive compensation makes the prervation and enhance-ment of short-term stock price a personal priority for both CEOs and CFOs.Jiang,Petroni,and Wang (2010)find that CFO equity incentives are more strongly related to earn-ings management than CEO equity incentives.Moreover,Chava and Purnanandam (2010)argue that CFOs are more influential in decisions requiring financial experti,such as earnings smoothing.Thus,there is a reason to expect CFO equity incentives to have a different effect on crash risk from that of CEO equity incentives,becau bad news hoarding requires financial experti and CFOs are generally in direct charge of processing financial informa-tion about the firm and disminating it to the stock market.H
owever,Feng,Ge,Luo,and Shevlin (2011)argue that CFOs are simply CEOs’agents and that they engage in accounting manipulations becau of CEO pressure.Thus,it is an empirical question whether CEO or CFO incentives matter more in determining bad news hoarding and crash risk.
Drawing on the findings of prior rearch,we also parately examine the incentives induced by stock
9
Jiang,Petroni,Wang (2010)provide evidence that CFO incentives to manipulate earnings are stronger for firms with higher return-earn-ings relations.
J.-B.Kim et al./Journal of Financial Economics 101(2011)713–730
716
holdings and option holdings.Peng and Roell(2008) analytically show that options have a more powerful impact than stock awards on managers’incentives to engage in share price manipulation,given their higher pay-performance elasticity.Burns and Kedia(2006)argue that option holdings create a more powerful incentive than stock holdings for managers to inflate short-term sha
夜光玫瑰re prices at the expen of long-term value.This is becau the loss to option holdings is limited when future price declines occur.In contrast,the payoff from stock holdings has a symmetric relation to share price,which expos managers’wealth to price declines as well as price appreciation.Thus,we expect that,compared to stock holdings,option holdings induce more aggressive bad news hoarding behavior by managers,which,in turn, leads to higher future crash risk.
3.Data and variable measurement
3.1.Data
The initial sample consists offirm-year obrvations in the ExecuComp databa during the period1993–2009.We then delete obrvations with missing Compustat account-ing data and missing Center for Rearch in Security Prices (CRSP)price,return,and trading volume data.We also excludefirms with a year-end share price that is lower than$1.Thefinal sample includes29,638firm-year obr-vations.The exact number of obrvations ud in our regression analys varies,depending on the data require-ment for the variables included in the regression.Table1 prents a comparison of annual obrvations as well as the percentage offirms experiencing crashes for the Compustat univer and the ExecuComp sample.It shows that about 30%of Compus
tatfirms are covered by the ExecuComp databa.In addition,Table1shows a slightly higher crash frequency forfirms in the ExecuComp databa than tho in the Compustat univer.
3.2.Measuringfirm-specific crash risk
This study employs three measures of crash risk, which are constructed following previous studies in the crash risk literature.Since we are interested infirm-specific factors that contribute tofirm-specific crash risk, wefirst estimatefirm-specific weekly returns for each firm and year.Specifically,we define thefirm-specific weekly return,denoted by W,as the natural log of one plus the residual return from the expanded market model regression10
r j,t¼a jþb1j r m,tÀ2þb2j r m,tÀ1þb3j r m,tþb4j r m,tþ1
þb5j r m,tþ2þe j t,ð1Þwhere r j,t is the return on stock j in week t,and r m,t is the return on the CRSP value-weighted market index in week t.We include the lead and lag terms for the market index return to allow for nonsynchronous trading (Dimson,1979).Thefirm-specific weekly return forfirm j in week t,W j,t,is measured by the natural log of one plus the residual return in Eq.(1),that is,W j,t¼lnð1þe j:tÞ.
Following Hutton,Marcus,and Tehranian(2009)and Kim,Li,and Zhang(in press),this paper defines crash weeks in a givenfiscal year for a givenfirm as tho
Table1
Yearly frequencies of stock price crash events.
门庭若市什么意思
This table prents descriptive statistics on the frequencies of stock price crash events for both the Compustat univer and the ExecuCompfirm sample from1993to2009.Stock price crash is defined in Appendix A.
Compustat univer ExecuCompfirms
Fiscal year No.of
firms
No.offirms
with price crash
Percentage offirms
with price crash
No.of
firms
No.offirms
with price crash
Percentage offirms
with price crash
19935,3608250.1541,5533170.204 19945,8328050.1381,6482640.160 19956,6988640.1291,7192390.139 19966,9428470.1221,8192490.137 19977,4059480.1281,8702710.145 19987,4361,3310.1791,9103460.181 19996,8888950.1301,8112610.144 20006,7461,0120.1501,7362860.165 20016,4981,4430.2221,75
44880.278 20026,0919930.1631,7773130.176 20035,7097880.1381,8132680.148 20045,5539440.1701,7653020.171 20055,5039800.1781,6783020.180 20065,4148120.1501,7842660.147 20075,2289410.1801,7003030.170 20085,0511,3740.2721,5204740.276 20094,6463620.0781,482930.063
Total103,00016,1640.15729,6385,0400.170
10All the empirical results are qualitatively unchanged if we also
include industry index return(and its lead and lag terms)in the
expanded market ,Hutton,Marcus,and Tehranian,2009).
Results are available upon request.
J.-B.Kim et al./Journal of Financial Economics101(2011)713–730717
