Organizational theories

Q1B3-Shi-2017-External-corporate-governance-and-f.pdf

Home >Education homework help >Organizational theories

Strategic Management Journal Strat. Mgmt. J., 38: 1268 – 1286 (2017)

Published online EarlyView 27 October 2016 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2560 Received 18 September 2015; Final revision received 20 April 2016

EXTERNAL CORPORATE GOVERNANCE AND FINANCIAL FRAUD: COGNITIVE EVALUATION THEORY INSIGHTS ON AGENCY THEORY PRESCRIPTIONS

WEI SHI, 1* BRIAN L. CONNELLY,2 and ROBERT E. HOSKISSON3

1 Kelley School of Business, Indiana University, Indianapolis, Indiana, U.S.A. 2 Raymond J. Harbert College of Business, Auburn University, Auburn, Alabama, U.S.A. 3 Jesse H. Jones Graduate School of Business, Rice University, Houston, Texas, U.S.A.

Research summary: Agency theory suggests that external governance mechanisms (e.g., activist owners, the market for corporate control, securities analysts) can deter managers from acting opportunistically. Using cognitive evaluation theory, we argue that powerful expectations imposed by external governance can impinge on top managers’ feelings of autonomy and crowd out their intrinsic motivation, potentially leading to financial fraud. Our findings indicate that external pressure from activist owners, the market for corporate control, and securities analysts increases managers’ likelihood of financial fraud. Our study considers external governance from a top manager’s perspective and questions one of agency theory’s foundational tenets: that external pressure imposed on managers reduces the potential for moral hazard.

Managerial summary: Many of us are familiar with stories about top managers “cooking the books” in one way or another. As a result, companies and regulatory bodies often implement strict controls to try to prevent financial fraud. However, cognitive evaluation theory describes how those external controls could actually have the opposite of their intended effect because they rob managers of their intrinsic motivation for behaving appropriately. We find this to be the case. When top managers face more stringent external control mechanisms, in the form of activist shareholders, the threat of a takeover, or zealous securities analysts, they are actually more likely to engage in financial misbehavior. Copyright © 2016 John Wiley & Sons, Ltd.

INTRODUCTION

Strategy scholars and policymakers have devoted renewed attention in recent years to “external” mechanisms of corporate governance, such as the monitoring and control by stakeholders who are not inside the organization. For example, a recent review of this literature seeks to “bring external

Keywords: External governance mechanism; Financial fraud; Ownership; Takeover defenses; Securities analysts *Correspondence to: Wei Shi. 801 W. Michigan St, BS 4020, Kelley School of Business-Indianapolis, Indianapolis, IN 46202, phone: 317-274-0939. E-mail: [email protected]

corporate governance into the corporate governance puzzle” more fully (Aguilera et al., 2015). Gov- ernance research has yielded important insights about these external governance mechanisms (Cof- fee, 2006), but few have considered their potentially adverse ramifications. Toward this end, we incorpo- rate a behavioral perspective of managers into our understanding of external governance to highlight how the expectations imposed by external gover- nance could impose on managers’ motivation, and we thus uncover the potential harm such governance mechanisms might introduce.

Recent developments in agency theory research relax the theory’s assumption of purely economic

External Corporate Governance and Financial Fraud 1269

agents (Wiseman and Gomez-Mejia, 1998). For example, behavioral agency theory reevaluates predictions in view of more realistic assumptions about agent behavior, with particular emphasis on internal governance (Pepper and Gore, 2015). Researchers have incorporated prospect theory (Martin, Gomez-Mejia, and Wiseman, 2013) and equity theory (Pepper, Gosling, and Gore, 2015) into agency theory predictions about how compen- sation structures influence managerial behavior. We build on the notion of overlaying cognitive biases onto agency theory prescriptions and extend this approach to external governance mechanisms. In particular, we inquire into how agents feel about external monitoring and control and what this means for their intrinsic motivation to behave ethically.

Cognitive evaluation theory (Boal and Cum- mings, 1981; Deci, 1971, 1975) is particularly informative in this regard because it explains how external controls can actually be counterproductive. The fundamental tenet of cognitive evaluation the- ory is that intrinsically motivated behavior is a func- tion of a person’s need to feel self-determining in his or her decisions (Phillips and Lord, 1981). The theory asserts that external monitoring and controls “crowd out” an individual’s motivation to behave in ways the controls are designed to ensure (Frey and Jegen, 2001). In our context, this would sug- gest that pressure from external governance lessens managers’ feelings of autonomy, thereby decreas- ing their intrinsic motivation to behave in ways that the governance mechanisms are supposed to safe- guard against (Deci and Ryan, 2000). In this study, we ask whether external governance weakens man- agers’ intrinsic motivation to act in the interest of shareholders and behave appropriately in the con- text of financial reporting.

Managerial financial fraud (e.g., inappropriately booking revenue, improperly valuing assets, not disclosing material information) is a phenomenon that is drawing extensive industry and regula- tory attention (Eaglesham and Rapoport, 2015). In fact, in 2014 alone, the Securities and Exchange Commission (SEC) announced 93 investigations against publicly traded companies for alleged finan- cial misconduct. As a result, governance scholars are acutely interested in how to predict and prevent the occurrence of financial fraud. Agency theory suggests that internal governance reduces informa- tion asymmetry between those inside and outside the firm, and consequently, decreases the likelihood

of fraud (Dalton et al., 2007). We, however, sug- gest and find that pressure from external governance may impose hidden agency costs as managers shift their locus of causality outward and lose their intrin- sic motivation to ethically report their respective firm’s performance, thus resulting in a greater like- lihood of financial fraud.

Our study introduces a key behavioral consider- ation into agency theory’s predictions about exter- nal governance, uncovering some counterintuitive relationships. For instance, we found that the “high- est quality” principals (Higgins and Gulati, 2006) are positively associated with the likelihood of fraud. Conversely, organizational provisions that many thought would lead to managerial entrench- ment, such as poison pills and golden parachutes (Bebchuk, Cohen, and Ferrell, 2009), actually bear a negative association with the likelihood of finan- cial fraud.

THEORETICAL DEVELOPMENT

Agency theory

Recent agency theory formulations focus on how agents behave in boundedly rational ways (Wise- man and Gomez-Mejia, 1998). The behavioral agency model (BAM) was developed largely to overcome criticisms regarding static assumptions about executives’ risk preferences (Wiseman and Gomez-Mejia, 1998). Empirical research on behav- ioral agency theory to date has focused mainly on behavioral risk propensities and internal gov- ernance using prospect theory arguments (Chris- man and Patel, 2012). For instance, this line of study re-examines compensation risk, highlight- ing the importance of individual problem framing to explain how risk influences executive behavior (Larraza-Kintana et al., 2007; Martin et al., 2013).

Following this model, we extend agency the- ory by applying behavioral considerations to three forms of external governance (we define external as being those forms of governance that operate without full access to the firm’s inside informa- tion). Within agency theory, one form of external governance is a firm’s owners, which serve as a market-based governance mechanism (Baysinger, Kosnik, and Turk, 1991). From an agency per- spective, managers are also subject to the market for corporate control, which researchers sometimes describe as a governance mechanism of last resort

1270 W. Shi, B. L. Connelly, and R. E. Hoskisson

(Jensen and Ruback, 1983). More recently, agency theory scholars have begun to investigate rating agencies (i.e., securities analysts) as another form of external governance (Chen, Harford, and Lin, 2015; Wiersema and Zhang, 2011). There are other exter- nal forces that act on firms beyond the three men- tioned here (e.g., legal institutions, social activists), but we limit our investigation to these three because they are the most central and influential forms of external governance within the agency framework.

We also add to agency theory by considering a key cognitive bias that challenges the theory’s eco- nomic assumptions. Specifically, we theorize about how to incorporate trade-offs between intrinsic and extrinsic motivation into agency theory models (c.f. Boivie, Graffin, and Pollock, 2012; Pepper and Gore, 2015). We use cognitive evaluation theory (Deci and Ryan, 1985) to explain how external gov- ernance mechanisms introduce high expectations on managers who serve as an extrinsic motivational force, which could crowd out intrinsic motivation, so that the combined effect is actually the oppo- site of what was intended with respect to preventing fraudulent behavior.

Cognitive evaluation theory

The concepts underlying cognitive evaluation the- ory emerged from the study of how external pres- sure affects internal motivation to do what is right. Some described this in terms of a “crowding-out effect,” wherein excessive external rewards and punishments can subvert intrinsic motivation to behave ethically (Bertelli, 2006; Georgellis, Iossa, and Tabvuma, 2011). Originators of the theory pred- icated their ideas on the assumption that individuals have innate needs for autonomy and competence (Ryan and Deci, 2000). Autonomy concerns “the experience of acting with a sense of choice, volition and self-determination” and competence is about “the belief that one has the ability to influence important outcomes” (Stone, Deci, and Ryan, 2009: 77). The level of autonomy and competence that individuals perceive they have is a powerful deter- minant of their intrinsic motivation (Deci and Ryan, 2012; Gagne and Deci, 2005).

In the cognitive evaluation theory framework, when external mechanisms of control impinge on an individual’s sense of autonomy and control, it could thereby decrease his or her internal motiva- tion to behave in ways that the external controls were supposed to ensure (Osterloh, Frost, and Frey,

2002). Consistent with these ideas, a number of studies in management support the notion that exter- nal consequences could potentially reduce individu- als’ motivation to behave in ways that are consistent with their responsibilities to the firm (e.g., Barkema, 1995; Jacquart and Armstrong, 2013). For example, Osterloh and Frey (2000) argued that high lev- els of extrinsic motivators can curtail employees’ intrinsic motivation to engage in organizational cit- izenship behavior, thus hindering them from trans- ferring tacit knowledge. Similarly, Sundaramurthy and Lewis (2003) contended that top managers oftentimes perceive external controls as coercive, reducing their desire to put forth effort. Our study builds on these ideas to develop specific hypotheses about how some of the most commonly investigated mechanisms of external corporate governance affect the likelihood of managerial financial fraud.

HYPOTHESES

Financial fraud

Financial fraud occurs when managers take actions that deceive investors or other key stakeholders (Gande and Lewis, 2009; Shi, Connelly, and Sanders, 2016). It often involves corruption, lying about facts, failure to disclose material infor- mation, falsifying information about the firm’s performance, or covering up systematic problems (Baucus and Near, 1991). There may be benefits to financial fraud that motivate managers to engage in such actions, such as the appearance of improved performance or increases in contingent compensa- tion. However, financial fraud harms investors, and especially, those who hold the firm’s stock over long periods.

As a result, external stakeholders attempt to cur- tail executive misbehavior in the form of finan- cial fraud by increasing their levels of monitoring and control (Davidoff, 2013). Standard economic approaches, including agency theory, consider the relative costs and benefits of fraud to determine the extrinsic motivation necessary to ensure that individuals will not engage in such scandalous behavior (Becker, 1976). Working within these the- oretical frames, a large body of empirical work has found external governance mechanisms that focus on monitoring and disciplining managers for misbehavior can reduce the likelihood of financial fraud (Beasley et al., 2000; Chen et al., 2006). Few,

External Corporate Governance and Financial Fraud 1271

however, have considered the psychological impact of these external corporate controls.

Mechanisms of external governance

The first external governance mechanism we con- sider is the firm’s owners (Brav et al., 2008; Hoskisson, Castleton, and Withers, 2009). Researchers examining shareholder influence on firm outcomes focus largely on institutional investors, such as mutual funds, hedge funds, pension funds, banks, insurance companies, and endowments (Goranova, Dharwadkar, and Bran- des, 2010). One type of investor resides at the extreme with respect to his or her ability to monitor and control: the dedicated institutional investor (Bushee, 1998, 2004). Porter (1992) described dedicated owners as being those who maintain large, long-term holdings concentrated in a small number of firms. These owners have incentive to monitor executive behavior and are able to understand rich and complex information about firms in which they invest (Higgins and Gulati, 2006). As such, they introduce high expectations on managers because they are closely attuned to managerial performance.

Dedicated institutional investors have a unique variety of tools at their disposal to control man- agers and demand results, which makes them an unusually potent force of external governance (Connelly et al., 2010a; Goranova and Ryan, 2014). By definition, dedicated institutional investors own substantial portions of firms in their portfolios. Thus, they are endowed with immense power over top managers because their exit would almost cer- tainly be followed by a sizeable drop in the firm’s stock price (Bushee, 2004). Under the threat of exit, this class of investors can demand that managers offer consistently high performance maintained over time (Koh, 2007; Maffett, 2012). Dedicated institutional investors also affect managerial expectations by exercising voice-based governance (Filatotchev and Toms, 2006; Goranova and Ryan, 2014). Dedicated owners frequently undertake shareholder resolutions, launch proxy contests, and initiate media campaigns to coerce managers (Wahal and McConnell, 2000). This group of owners is particularly adept at leading activism activities among shareholders (Gaspar, Massa, and Matos, 2005). They often support activist shareholders who push managers to maximize

performance, which can give rise to excess pressure faced by managers (Martin, 2011).

Traditional agency theory predicts that a higher level of dedicated institutional ownership should be associated with a reduced likelihood of moral hazard (Sharma, 2004). One of the main reasons is that managers should be fearful of the neg- ative repercussions of being caught, which they might know is more likely to occur when the firm has high levels of dedicated institutional owner- ship. Information asymmetry between principals and agents should be lower for dedicated owners as compared to other types of owners (Weiss and Beckerman, 1995). This is because dedicated own- ers have extensive resources to devote to moni- toring managerial behavior, and given the nature of their holdings, are motivated to monitor man- agers carefully (Connelly et al., 2010a). Close mon- itoring and low levels of information asymmetry should constrain self-serving managerial manipu- lations of financial information by increasing the risk of detection (Hadani, Goranova, and Khan, 2011). In other words, agency theory highlights the notion that dedicated investors could heighten managerial concerns about being caught for wrong- doing, thus mitigating the likelihood of financial fraud.

Incorporating cognitive evaluation theory, on the other hand, uncovers a hidden problem with this agency theory prediction by accounting for how CEOs might feel about the external expec- tations that come with high levels of dedicated ownership. As one observer noted: “The perception that activism creates greater value for all share- holders has won the sympathy and support of major institutional investors that traditionally have remained passive when it comes to engaging with the companies in their portfolios” (Duffy, 2015). The heavy hand of dedicated institutional investors could prompt managers to shift from an internal to an external locus of causality, making them potentially less concerned with doing business honestly than they are with outward perceptions of compliance (Osterloh and Frey, 2004). This shift in managers’ locus of causality helps explain why traditional agency theory predictions about external governance may not apply, and in fact, we expect to see results that are more consistent with steward- ship arguments with a focus on intrinsic motivation (Arthurs and Busenitz, 2003; Sundaramurthy and Lewis, 2003).

1272 W. Shi, B. L. Connelly, and R. E. Hoskisson

In the cognitive evaluation theory framework, external intervention could crowd out managers’ intrinsic motivation to act ethically. This is espe- cially so, given that such monitoring is often focused on continually positive financial returns, so managers may become more likely to compro- mise their ideals by engaging in financial fraud that appears to meet dedicated owners’ persistent exter- nal expectations, even though managers know it is wrong. Argyris (1964) was one of the first to recognize the potential for this phenomenon when he noticed that strict governance has a paradoxical effect: It leads to continuously expanding control, but at the same time, reduces managerial loyalty. When dedicated institutional ownership is high, top managers, subject to unrelenting external expec- tations from dedicated institutional investors and activists, may feel compelled to make financial reporting decisions not from their own beliefs, but merely to satisfy the expectations of the firm’s own- ers. Therefore, we suggest the following hypothesis:

Hypothesis 1: A firm’s level of dedicated institu- tional ownership is positively associated with the likelihood of financial fraud.

Another commonly considered form of exter- nal governance is the market for corporate con- trol (Jensen and Ruback, 1983). The market for corporate control imposes external pressure on managers to deliver consistently positive financial earnings reports because, if they do not, other management teams may attempt to gain control of the company (Hitt et al., 1996). It is diffi- cult to directly measure the extent to which this governance mechanism is at work because it is an unobservable force until it is activated (i.e., until the poor performing firm is acquired). How- ever, we can view the extent to which executives are exposed to the market for corporate control by looking at the firm’s takeover defense provi- sions (Humphery-Jenner, 2014). Although takeover defenses are internal, researchers often use them as a means of examining the extent to which managers are subject to the external governance of the mar- ket for corporate control (Humphery-Jenner, 2014; Kabir, Cantrijn, and Jeunink, 1997).

Common takeover defenses include supermajori- ties, staggered board appointments, poison pills, and golden parachutes. Kini, Kracaw, and Mian (2004) argued that the disciplinary function of the

market for corporate control is largely ineffective when firms have takeover defenses such as these. As a result, external expectations to perform that arise from the market for corporate control are likely to be less dogged when managers enjoy more and better takeover protections. In contrast, managers experience greater pressure and higher expectations from the market for corporate control when their company has fewer, or weak, takeover defenses (Mahoney, Sundaramurthy, and Mahoney, 1997).

Traditional agency theory predicts that a higher level of takeover defenses (and thus, an ineffec- tive market for corporate control) should be asso- ciated with a greater likelihood of moral hazard (McGurn, 2002). Agency theorists would argue that these types of provisions give rise to managerial entrenchment and increase agency costs (Bebchuk et al., 2009; Gompers, Ishii, and Metrick, 2003). As a result, though managers generally want takeover defenses, most existing studies focus on how they can be bad for shareholders (Mahoney et al., 1997; Sundaramurthy, Mahoney, and Mahoney, 1997). From a purely economic view of the agent, takeover defenses can reduce or even eliminate the poten- tially negative outcomes associated with commit- ting financial fraud. If they are less concerned about the consequences of being caught, managers may be more likely to inflate numbers or adjust financial reports to garner private benefits.

Cognitive evaluation theory, on the other hand, offers a different perspective. Top managers of firms with strong takeover defense protection may not be overly concerned about employment safety, even if their companies fail to meet external perfor- mance expectations, and thereby, become takeover targets. Put differently, top managers are more likely to make decisions that reflect their own values and beliefs when the company has ample takeover defenses in place. In the absence of those provisions, though, failing to meet performance expectations would increase top managers’ employ- ment risk (Kacperczyk, 2009). Without takeover defenses, the threat of a takeover could alter the risk propensity of top managers, potentially lead- ing them to make short-term financial reporting decisions. Takeover defense provisions protect top managers from the external pressure of the mar- ket for corporate control, affording managers an extra measure of decision autonomy and allow- ing them to think about the long term when reporting their performance (Wang, Zhao, and He, forthcoming).

External Corporate Governance and Financial Fraud 1273

The imposing presence of external expectations from the market for corporate control could crowd out managers’ intrinsic motivation to behave ethically, and vice versa, removing those external expectations could activate managers’ intrinsic sense of duty to do what is right (Kanfer, 1990), as is consistent with stewardship theory (Sundara- murthy and Lewis, 2003). In fact, being vulnerable to the market for corporate control may be par- ticularly salient to the problem of financial fraud. Financial fraud encompasses a range of actions that are almost universally harmful to shareholders and generally have one thing in common: failure to disclose material financial information. Takeover defenses could make managers more willing to disclose critical information simply because they know it is the right thing to do. Stated formally,

Hypothesis 2: The number of a firm’s takeover defense provisions is negatively associated with the likelihood of financial fraud.

Researchers in management, law, finance, and accounting have, through the years, explored how the expectations of external financial markets via rating agencies influence firm behaviors (Ben- ner and Ranganathan, 2012; Chen et al., 2015). Securities analysts raise questions with top man- agers about firm performance and strategies dur- ing conference calls and distribute information to investors through reports and media outlets (Lang, Lins, and Miller, 2004). These reports generally include forecasts of the firm’s expected future stock price and the analysts’ recommendations about whether to “buy,” “hold,” or “sell” the firm’s stock (Bradshaw, 2004; Schipper, 1991).

This line of research has shown that analysts play an important role in shaping the expectations imposed on managers to undertake actions (Gen- try and Shen, 2013). One way analysts impose pressure on managers is via their effects on stock price (Zuckerman, 2000). This occurs even when analysts make recommendations based on stock repurchase plans that could have vague or inde- terminate stock price effects (Zhu and Westphal, 2011). Positive or negative recommendations have implications for stock purchase behavior and help determine the value of a firm’s stock. As a result, there is a growing body of evidence that managers are highly attentive to analysts’ recommendations and the resultant changes in stock price (Martin,

2011; Rao and Sivakumar, 1999). For example, one study shows that firms covered by a large number of financial analysts have less innovative activity because financial analysts impose pressure to deliver consistently positive short-term financial results (He and Tian, 2013).

External expectations from financial analysts may come in two forms. First, sell recommen- dations reduce a firm’s stock price, and there- fore, impose external pressure on managers to take action so that the firm’s performance bounces back (Stickel, 1995). Second, buy recommendations also introduce external pressure because managers are likely to feel the burden of high earnings expecta- tions when analysts are recommending their stock to capital markets (Barsky, 2008; Mishina et al., 2010). In contrast to buy and sell recommendations, a hold recommendation should result in sharehold- ers devoting less attention to firms, so this represents the lowest level of external pressure from analysts, and in fact, most recommendations are hold.

Agency theory envisions securities analysts col- lectively as an external governance mechanism, keeping managers in check by reducing information asymmetry between principals and agents (Jensen and Meckling, 1976). Seeking to gain investor fol- lowing, analysts are concerned with obtaining the most comprehensive information they can get about publicly traded firms and issuing insightful rec- ommendations to shareholders (Womack, 1996). Investors pay close heed to analysts’ recommenda- tions, lending particularly close attention to those firms to whom the analysts recommend attention (Beunza and Garud, 2007; Brown, Wei, and Wer- mers, 2013). Thus, in the agency framework, we should expect external expectations that arise from security analysts’ recommendations to be nega- tively associated with financial fraud.

Again, cognitive evaluation theory introduces a different perspective on securities analysts’ ratings. Some scholars have found that high performance expectations can actually lead to fraudulent behavior owing to the increased pres- sure leaders feel because of those expectations (Schweitzer, Ordóñez, and Douma, 2004). This may be particularly true of analysts’ ratings because missing analyst forecasts can precipitate drops in stock price, reduced managerial com- pensation (Chen et al., 2015), or even dismissal (Wiersema and Zhang, 2011). In this sense, “sell” recommendations can impose great pressure on managers. Similarly, when analysts recommend

1274 W. Shi, B. L. Connelly, and R. E. Hoskisson

“buy” it also introduces pressure on managers, wherein top managers are compelled to meet high expectations. Research suggests that this pressure may be accentuated by stock market overreaction to these analysts’ recommendations (Brown et al., 2013). When external pressure from analysts is high, it could lead top managers to make decisions that violate their codes of conduct but help achieve tangible objectives (e.g., meeting analyst expec- tations or turning around company performance) (Hirsch and Pozner, 2005). Stated differently, external pressure from analysts could introduce high performance expectations on managers, making them less concerned with doing the right thing than they are with outward perceptions of compliance. As a result, we expect this external governance mechanism could have an adverse effect on managerial behavior, as follows:

Hypothesis 3: Pressure from security analysts’ recommendations is positively associated with the likelihood of financial fraud.

METHODS

Sample

We tested our hypotheses on a longitudinal data set covering the years 1999 – 2012. The sample used in this study starts with all firms in the in the S&P 1500 index during our sampling window as well as a few other large public firms included in the Investor Responsibility Research Center (IRRC). The dependent variable data are from the SEC Accounting and Auditing Enforcement Releases (AAERs). Data on institutional ownership, takeover defense provisions, and securities analysts are from Thomson Reuters Institutional (13F) Holdings, the IRRC, and Thomson Reuters IBES, respectively. Financial data are from Compustat and the Center for Research in Security Prices (CRSP). We col- lected top manager compensation and governance data from ExecuComp and Risk Metrics.

Dependent variable

The dependent variable of our study is commit- ment of financial fraud. Since 1982, the SEC has issued AAERs during or at the conclusion of an investigation against a company, an auditor, or an

individual for alleged accounting or auditing mis- conduct (Dechow et al., 2011). The SEC takes enforcement actions against firms that it identifies as having violated the financial reporting require- ments of the Securities Exchange Act of 1934. Given budget constraints, the SEC chooses firms for enforcement action when there is strong evi- dence of accounting manipulation. In general, firms that the SEC selects have already admitted restating earnings or having unusually large write-offs (e.g., Enron and Xerox) (Dechow et al., 2011).

To identify fraud commitment years accurately, we read each AAER entry to identify the years when financial fraud actually occurred and matched them to our independent and control variables based on fraud commitment years. We identified a total num- ber of 265 cases of fraud commitment firm years for our sample firms. The primary advantage of our chosen operationalization is that firms selected for SEC enforcement are almost certainly guilty of fraudulent financial reporting (i.e., Type I error is low) (Dechow et al., 2011). Fraud commitment receives a value of 1 if a firm commits financial fraud in a year and is later detected by the SEC, and 0 otherwise.

Independent variables

Our first independent variable is Dedicated insti- tutional ownership. We followed Bushee (2001) to identify dedicated institutional investors among all the institutional investors reported in Thomson Reuter Institutional (13F) Holdings. This approach relies on a factor and cluster analysis to classify institutional investors into different types. The clas- sification is based on portfolio turnover, momen- tum trading strategies, and portfolio diversification strategies (Bushee, 2001). We identified dedicated institutional investors, which are low on all three factors, for each sample firm and calculated their average holdings across four quarters for each year. Dedicated institutional ownership is the ratio of total shares held by dedicated institutional investors to total shares outstanding (Connelly et al., 2010b).

Our second independent variable is takeover defenses. We measured the number of Takeover defense provisions as the sum of the following six indicator variables: (1) staggered board, (2) limita- tion on amending bylaws, (3) limitation on amend- ing the charter, (4) supermajority to approve a merger, (5) golden parachute, and (6) poison pill. Findings by Bebchuk et al. (2009) suggest these six

External Corporate Governance and Financial Fraud 1275

takeover defense provisions play the most important role in shielding managers from the market for cor- porate control and are of greatest relevance to man- agerial entrenchment. Because the IRRC published takeover defense provision data every other year until 2004, we followed existing work by Bebchuk et al. (2009) and replaced the four years in our sam- pling window that did not have IRRC coverage with data from immediately subsequent years.

The third independent variable is Analysts’ recommendation pressure. We measured this as the sum of the average percent of sell recommendations and the average percent of buy recommendations issued by securities analysts across different quarters for each year. We focus on sell and buy recommendations because these two types of recommendations have important implications on the trading behavior of individual investors and fund managers (Stickel, 1995). Sell recom- mendations suggest that rated stocks are likely to underperform relative to the market or its previous performance whereas buy recommendations sug- gest that rated stocks are likely to outperform the market within the next 6 – 12 months. As we argue, sell and buy recommendations introduce power- ful external earnings expectations on managers, while hold recommendations present the least pressure.

Control variables for fraud commitment

Our chosen method of analysis, bivariate probit models, mandates that we develop separate models with fraud commitment as one dependent variable and fraud detection as a different dependent variable (Wang, 2013; Wang, Winton, and Yu, 2010). We follow Wang’s (2013) guidelines for control vari- ables that could increase the likelihood of fraud commitment.

First, we control for a range of firm-level char- acteristics. We control for Firm performance using return on assets (ROA). We also control for Firm size using the natural logarithm of total assets. The fraud literature suggests that managers of firms with high levels of External financing need are more likely to commit fraud than managers of firms with lower need (Teoh, Welch, and Wong, 1998). We follow Demirguc-Kunt and Maksimovic (1998) to measure this as a firm’s asset growth rate in excess of the maximum internally financeable growth rate: asset growth rate — ROA/(1-ROA). We control for

Firm leverage using the ratio of total short- and long-term debt to total assets.

We also control for three variables related to firm risk-taking activities that are associated with fraud commitment (Wang, 2013). These are Capital expenditure intensity, measured as capital expen- diture divided by total sales revenues; R&D inten- sity, measured as R&D expenditure divided by total sales revenues; and Acquisition intensity, measured as total annual acquisition expenditure divided by total sales revenues.

We control for a number of executive charac- teristics as well. We control for Top management team (TMT) equity ownership, measured as the total percent of equity ownership held by all the top executives reported in ExecuComp. We control for TMT option pay, measured as the ratio of total TMT option pay value to total TMT pay. We con- trol for Board independence, measured as the total number of independent outside directions divided by board size, and for the percent of Directors appointed by CEOs, measured as the ratio of direc- tors appointed by CEOs to board size. We control for Outside directors’ ownership, measured as the ratio of shares held by outside directors to total shares outstanding, and for CEO duality, which receives a value of 1 if a CEO is also Board Chair. We control for Analysts’ coverage, measured as the number of analysts covering a firm. Last, we control for the Post-SOX period, which is 1 for years after (and including) 2002, and 0 otherwise because gov- ernance reforms triggered by the Sarbanes-Oxley Act (SOX) may deter managers from committing financial fraud.

Control variables for fraud detection

Fraud detection describes when firms commit fraud and the SEC catches them for doing so. To model the likelihood of fraud detection we include some control variables that overlap with the control vari- ables used to model the likelihood of fraud com- mitment. Control variables used in both the fraud commitment (P(F)) and fraud detection (P(D|F)) are possible indicators not only that a firm will com- mit fraud, but also that it will be caught. When we include a control in both fraud commitment and fraud detection models, we operationalize the vari- able the same in both models.

However, our models for the likelihood of fraud detection contain some unique control vari- ables that we do not use in our models of fraud

1276 W. Shi, B. L. Connelly, and R. E. Hoskisson

commitment. For instance, the SEC could be more likely to select for investigation firms that operate in industries where securities lawsuits are common, so we control for Abnormal industry litigation. To measure this, we calculate industry litigation intensity, measured as the natural logarithm of the total market value of all litigated firms in an indus- try year (using two-digit SIC codes). Abnormal industry litigation is the annual deviation from the average litigation intensity in an industry. We also control for Abnormal ROA, which can flag a firm as a potential problem, using the residual from the regression: ROA1 = 𝛼0 + 𝛼1ROA0 + 𝛼2ROA- 1 + 𝜀. Similarly, we control for Annual stock returns because the SEC may target for enforcement firms with sharp changes in stock returns. For the same reason, we control for Abnormal return volatility, measured as the demeaned standard deviation of monthly stock returns in a year, and Abnormal stock turnover, measured as the natural logarithm of the demeaned monthly turnover in a year. Last, we include a measure of total Institutional ownership in our fraud detection models because research shows that institutional investors could play a role in discovering fraud (Dyck, Morse, and Zingales, 2010).

METHODS AND RESULTS

Table 1 summarizes the descriptive statistics, including means, standard deviations, and cor- relations of variables used in this study. Table 2 presents the results of bivariate probit models. P(F) models the likelihood of fraud commitment, and P(D|F) models the likelihood of fraud detection given fraud commitment.

Analysis

Corporate fraud is a rare event. Given the low rate of occurrence within our population of firms, exam- ining the data using hazard models or conditional logistic regressions with matched pairs could be appropriate (Carberry and King, 2012). Yet, there are two latent processes associated with corporate fraud: firms that engage in fraud (i.e., fraud com- mitment) and those that the SEC actually catches in the act of fraud (i.e., fraud detection). We are interested in the former, but can only observe the latter. Traditional methods are limited to examining the observable firms that have been caught in the act

of fraud, ignoring firms that have committed fraud but have not (or not yet) been caught. The under- lying assumption is that firms that are cheating and getting away with it are comparatively much fewer than those that cheat and are caught, but this may not be an accurate assumption.

We attempt to address this problem methodologi- cally by using bivariate probit regressions with par- tial observability, following the works of Wang et al. (2010) and Wang (2013). Bivariate probit regres- sions model fraud detection and fraud commitment simultaneously, thus mitigating biases caused by the presence within our sample of firms that have engaged in fraud but have not yet been detected. This is an important distinction because our theory describes why firms might actually engage in finan- cial fraud, not whether the SEC catches them in the act. Of course, neither the traditional methods nor our bivariate probit models account for the possi- bility that firms may have had an SEC enforcement action against them when, in fact, they did nothing wrong. However, given the extensive nature SEC investigations and the burden of proof they must overcome, we expect it is a safe assumption that there are few, if any, firms that are innocent victims of the SEC enforcement.

To explain how bivariate probit models work, let F∗

i represent firm i’s propensity to commit fraud,

and D∗ i

represent the firm’s likelihood of being detected conditional on fraud being committed. The reduced form model is then:

F∗ i = xF,i𝛽F + ui, (1)

D∗ i = xD,i𝛽D + vi, (2)

where xF,i is a row vector with variables that explain the propensity for firm i to commit fraud, and xD,i is a second-row vector with variables that explain the firm’s likelihood of getting caught conditional on fraud commitment. The variables ui and vi are zero-mean disturbances with a bivariate normal distribution and variances normalized to unity because we cannot estimate the variances. The correlation between ui and vi is 𝜌 (Wang, 2013).

To model fraud commitment, we transform F∗ i

into a binary variable Fi, where Fi = 1 if F ∗ i >

0, and Fi = 0 otherwise. To model fraud detection conditional on fraud commitment, we transform D∗

i into a binary variable in the same way. We cannot observe all the realizations of F∗

i and D∗

i , but note

External Corporate Governance and Financial Fraud 1277

T ab

le 1.

D es

cr ip

ti ve

st at

is ti

V ar

ia bl

e M

ea n

S .D

. M

in M

ax 1

2 3

4 5

6 7

8 9

1 F

in an

ci al

fr au

d 0.

01 7

0. 12

8 0.

00 0

1. 00

0 1.

00 0

2 D

ed ic

at ed

ow ne

rs hi

p 0.

04 5

0. 06

7 0.

00 0

0. 35

0 0.

07 6

1. 00

0 3

T ak

eo ve

r de

fe ns

e 2.

56 9

1. 34

4 0.

00 0

6. 00

0 −

0. 05

4 −

0. 09

3 1.

00 0

4 A

na ly

st s’

pr es

su re

0. 56

0 0.

22 5

0. 00

0 1.

00 0

0. 06

1 0.

07 8

− 0.

08 4

1. 00

0 5

R O

A 0.

09 0

0. 08

5 −

0. 23

0 0.

35 2

− 0.

01 8

0. 01

8 −

0. 05

4 0.

15 6

1. 00

0 6

F ir

m si

ze (L

og )

7. 80

7 1.

66 3

4. 58

4 12

.4 41

0. 05

6 −

0. 01

7 0.

00 1

− 0.

05 6

− 0.

09 2

1. 00

0 7

E xt

er na

l fi

na nc

in g

ne ed

0. 15

9 1.

26 5

− 5.

76 8

6. 97

1 0.

01 8

− 0.

00 1

0. 00

5 0.

03 4

0. 02

0 −

0. 03

2 1.

00 0

8 F

ir m

le ve

ra ge

0. 21

8 0.

17 4

0. 00

0 0.

76 8

0. 02

7 0.

06 7

0. 04

9 −

0. 02

3 −

0. 12

7 0.

25 5

− 0.

02 8

1. 00

0 9

C ap

it al

ex pe

nd it

ur e

ra ti

o 0.

06 7

0. 11

0 0.

00 0

0. 80

1 −

0. 01

1 0.

03 5

0. 00

8 0.

09 1

− 0.

06 3

0. 03

0 0.

03 9

0. 13

3 1.

00 0

10 R

& D

in te

ns it

y 0.

03 7

0. 07

6 0.

00 0

0. 44

1 0.

02 3

0. 02

9 −

0. 05

9 0 .

11 1

− 0.

22 1

− 0.

22 3

0. 04

3 −

0. 20

3 0.

03 7

1. 00

0 11

A cq

ui si

ti on

in te

ns it

y 0.

03 2

0. 08

6 0.

00 0

0. 55

3 0.

01 4

− 0.

01 2

0. 00

8 0.

09 2

− 0.

01 6

− 0.

03 4

− 0.

00 9

0. 08

1 0.

01 7

0. 13

5 12

A na

ly st

s’ co

ve ra

ge 10

.6 52

7. 84

5 0.

00 0

55 .0

00 0.

00 8

0. 04

6 −

0. 06

6 0.

06 1

0. 15

9 0.

51 6

0. 00

1 −

0. 02

8 0.

11 3

0. 08

5 13

In st

it ut

io na

l ow

ne rs

hi p

0. 68

9 0.

24 7

0. 00

0 1.

00 0

− 0.

05 1

0. 14

4 0.

12 9

0 . 08

6 0.

16 6

− 0.

09 0

− 0.

00 8

− 0.

05 4

− 0.

01 8

0. 02

9 14

T M

T ow

ne rs

hi p

0. 01

2 0.

03 4

0. 00

0 0.

22 8

− 0.

04 0

− 0.

13 3

0. 04

1 −

0. 04

4 0.

00 3

− 0.

12 1

0. 00

7 −

0. 07

2 −

0. 03

0 −

0. 01

7 15

T M

T op

ti on

pa y

0. 27

5 0.

24 5

0. 00

0 0.

89 8

0. 10

6 0.

23 3

− 0.

09 7

0. 19

0 0.

02 9

− 0.

02 3

0. 03

6 −

0. 09

3 0.

04 0

0. 31

8 16

B oa

rd in

de pe

nd en

ce 0.

71 6

0. 15

6 0.

00 0

1 . 00

0 −

0. 07

9 −

0. 14

0 0.

24 6

− 0.

09 8

− 0.

03 3

0. 19

3 −

0. 01

2 0.

05 0

− 0.

00 3

0. 00

3 17

C E

O ap

po in

te d

di re

ct or

0. 67

8 0.

29 9

0. 00

0 1.

00 0

0. 03

4 0.

02 4

− 0.

07 9

0. 04

6 0.

00 7

− 0.

05 3

0. 01

8 −

0. 00

9 0.

03 0

− 0.

00 5

18 O

ut si

de di

re ct

or ow

ne rs

hi p

0. 04

5 0.

08 7

0. 00

0 0.

54 4

− 0.

01 0

0. 05

0 −

0. 15

7 −

0. 00

7 0.

02 9

− 0.

14 8

− 0.

00 1

− 0.

02 5

− 0.

03 2

− 0.

06 1

19 C

E O

du al

it y

0. 59

6 0.

49 1

0. 00

0 1.

00 0

0. 02

8 0.

07 3

0. 01

1 0.

01 5

0. 00

4 0.

13 6

− 0.

00 4

0. 07

8 0.

02 8

− 0.

06 9

20 A

bn or

m al

in du

st ry

li ti

ga ti

on 0.

65 5

0. 82

4 0.

00 0

4. 72

9 0.

03 9

0. 01

3 −

0. 05

5 0.

04 6

− 0.

06 8

0. 02

2 0.

04 1

− 0.

00 9

0. 15

7 0.

22 1

21 A

bn or

m al

R O

A 0.

01 3

0. 07

3 −

0. 25

2 0.

23 3

− 0.

02 0

0. 01

1 −

0. 05

5 0.

16 2

0. 99

2 −

0. 08

5 0.

02 2

− 0.

12 3

− 0.

06 0

− 0.

21 7

22 A

nn ua

l st

oc k

re tu

rn s

0. 07

4 0.

44 3

− 0.

79 6

1. 78

8 −

0. 03

0 0.

01 2

0. 02

4 −

0. 00

6 0.

03 4

− 0.

02 6

− 0.

04 9

− 0.

02 0

− 0.

01 6

− 0.

01 7

23 A

bn or

m al

re tu

rn vo

la ti

li ty

0. 11

4 0.

06 9

0. 01

7 1.

07 4

0. 03

6 0.

06 0

0. 00

6 0.

08 6

− 0.

25 4

− 0.

28 7

0. 03

6 −

0. 02

7 0.

05 7

0. 26

0 24

A bn

or m

al st

oc k

tu rn

ov er

(L og

) 11 .7

44 1.

50 2

4. 66

3 18

.5 84

0. 05

6 −

0. 06

3 −

0. 00

9 0.

07 3

0. 05

2 0.

64 7

0. 02

2 0.

04 7

0. 09

0 0.

16 7

V ar

ia bl

e 11

12 13

14 15

16 17

18 19

20 21

22 23

11 A

cq ui

si ti

on in

te ns

it y

1. 00

0 12

A na

ly st

s’ co

ve ra

ge 0.

01 2

1. 00

0 13

In st

it ut

io na

l ow

ne rs

hi p

0. 05

6 0.

30 8

1. 00

0 14

T M

T ow

ne rs

hi p

0. 00

4 −

0. 05

5 0.

07 1

1. 00

0 15

T M

T op

ti on

pa y

0. 05

1 0.

15 3

− 0.

06 0

− 0.

12 2

1. 00

0 16

B oa

rd in

de pe

nd en

ce −

0. 00

3 0.

12 1

0. 18

7 −

0. 02

0 −

0. 12

4 1.

00 0

17 C

E O

ap po

in te

d di

re ct

or 0.

00 6

− 0.

01 4

− 0.

02 7

0. 13

8 0.

03 1

− 0.

11 3

1. 00

0 18

O ut

si de

di re

ct or

ow ne

rs hi

p −

0. 01

9 −

0. 12

7 −

0. 16

6 0.

12 0

− 0.

01 5

− 0.

36 5

− 0.

05 9

1. 00

0 19

C E

O du

al it

y −

0. 00

5 0.

04 1

− 0.

05 3

− 0.

00 5

0. 05

6 0.

06 7

0. 37

6 −

0. 17

4 1.

00 0

20 A

bn or

m al

in du

st ry

li ti

ga ti

on 0.

02 9

0. 12

2 −

0. 01

9 0.

01 3

0. 12

4 −

0. 04

2 0.

04 2

− 0.

03 4

− 0.

02 4

1. 00

0 21

A bn

or m

al R

O A

− 0.

03 0

0. 15

8 0.

16 3

0. 00

1 0.

02 0

− 0.

02 8

0. 00

4 0.

02 5

0. 00

2 −

0. 06

3 1.

00 0

22 A

nn ua

l st

oc k

re tu

rn s

− 0.

04 0

0. 01

0 0.

04 9

− 0.

02 8

− 0.

07 8

0. 01

7 −

0. 01

2 0.

02 7

− 0.

01 6

− 0.

00 2

0. 04

7 1.

00 0

23 A

bn or

m al

re tu

rn vo

la ti

li ty

− 0.

01 0

− 0.

13 0

− 0.

09 0

0. 00

8 0.

21 1

− 0.

12 0

0. 04

0 0.

04 7

− 0.

03 5

0. 10

0 −

0. 25

4 −

0. 00

8 1.

00 0

24 A

bn or

m al

st oc

k tu

rn ov

er (L

og )

0. 00

4 0.

67 1

0. 11

3 −

0. 04

7 0.

17 8

0. 20

3 −

0. 04

8 −

0. 18

5 0.

03 9

0. 12

3 0.

05 4

− 0.

04 5

0. 05

1 1.

00 0

N =

14 ,7

29 .T

he ab

so lu

te va

lu e

of co

rr el

at io

n gr

ea te

r th

an 0.

02 si

gn ifi

ca nt

at p <

0. 05

fo r

tw o-

ta il

ed te

st s.

1278 W. Shi, B. L. Connelly, and R. E. Hoskisson

Table 2. Bivariate probit models with fraud commitment and detection as dependent variables

Variables Model 1 Model 2 Model 3

P(F) P(D|F) P(F) P(D|F) P(F) P(D|F)

Constant −2.949 −5.231 −2.664 −4.979 −3.119 −4.994 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

ROA 0.664 −0.967 −0.719 (0.415) (0.383) (0.587)

External financing need 0.032 0.027 0.044 (0.041) (0.263) (0.009)

Firm leverage 0.201 0.302 0.157 (0.478) (0.373) (0.532)

TMT ownership −1.415 −1.525 −1.420 (0.003) (0.002) (0.000)

TMT option pay 0.510 0.525 0.450 (0.165) (0.048) (0.127)

Board independence −0.883 −1.415 −0.566 (0.106) (0.068) (0.302)

CEO appointed directors 0.991 1.023 1.007 (0.000) (0.000) (0.000)

Outside director ownership −6.024 −6.778 −5.767 (0.020) (0.087) (0.027)

CEO duality 0.010 0.061 0.096 (0.916) (0.688) (0.307)

Post-SOX −0.479 −0.612 −0.541 (0.003) (0.000) (0.018)

Firm size 0.000 0.464 −0.014 0.380 −0.036 0.455 (0.999) (0.020) (0.814) (0.001) (0.697) (0.095)

Capital expenditure ratio −1.910 2.033 −1.999 −0.117 −2.748 3.900 (0.004) (0.533) (0.013) (0.941) (0.003) (0.371)

R&D intensity 0.236 0.412 −1.327 3.780 −0.694 1.272 (0.801) (0.831) (0.161) (0.151) (0.603) (0.805)

Acquisition intensity −1.098 13.311 −0.469 6.337 −1.238 13.464 (0.181) (0.202) (0.575) (0.217) (0.394) (0.571)

Analysts’ coverage −0.189 −0.443 −0.122 −0.700 −0.046 −0.407 (0.015) (0.110) (0.097) (0.007) (0.876) (0.616)

Institutional ownership 0.806 2.267 2.121 (0.376) (0.013) (0.255)

Abnormal industry litigation 0.293 0.425 0.334 (0.052) (0.010) (0.059)

Abnormal ROA −1.008 2.926 0.938 (0.608) (0.379) (0.816)

Annual stock returns 0.440 0.296 0.500 (0.210) (0.060) (0.354)

Abnormal return volatility 0.233 0.817 0.911 (0.912) (0.668) (0.801)

Abnormal stock turnover 0.315 0.496 0.302 (0.030) (0.001) (0.279)

Dedicated ownership 1.913 (0.026)

Takeover defense −0.202 0.397 (0.008) (0.107)

Analysts’ pressure 1.056 −1.587 (0.005) (0.247)

Observations 15, 845 15, 032 14, 729 Chi-squared 2, 180 285.6 2, 433 Log-likelihood −986.3 −854 −793.1

P-values in parentheses. Standard errors clustered by two-digit SIC codes. We do not control for dedicated ownership P(D|F) in Model 1 because dedicated ownership is included in total institutional ownership.

External Corporate Governance and Financial Fraud 1279

that Zi = Fi × Di, (3)

where Zi = 1if firm i has committed fraud and been detected, and Zi = 0 otherwise. With Φ as the bivariate standard normal cumulative distribu- tion function, the empirical model for estimating Zi is

P ( Zi = 1

) = P

( FiDi = 1

) = P

( Fi = 1, Di = 1

)

= F ( xF,ibF, xD,ibD, 𝜌

) (4)

P ( Zi = 0

) = P

( FiDi = 0

) = P

( Fi = 0, Di = 0

)

+ P ( Fi = 1, Di = 0

) = 1 − F

( xF,ibF, xD,ibD, 𝜌

) .

(5)

Poirier (1980) and Feinstein (1990) suggest that the conditions for full identification of the model parameters have two requirements. First, xF,i and xD,i must not include the same variables. As noted in the variable section, we mainly follow Wang (2013) to identify key variables that influence the propensity of fraud commitment and the likeli- hood of fraud detection. The second requirement is that predictor variables need to exhibit sub- stantial variation in the sample. As a result, the model can be identified more easily if xF,i and xD,i include continuous instead of indicator vari- ables (Wang, 2013). This explains why we did not include industry dummy variables and year dummy variables in our bivariate probit models because inclusion of too many dummies without sufficient variation can lead to estimation failure. Given that we cannot include industry fixed-effects in regres- sions, we cluster standard errors by two-digit SIC codes to address potential correlations among resid- uals of firms in the same industry (Khanna, Kim, and Lu, 2015). We then estimate bivariate pro- bit models using the maximum-likelihood method, as follows:

L ( 𝛽F, 𝛽D, 𝜌

) = ∑

zi=1 log(P

( Zi = 1)

)

+ ∑

zi=0 log(P

( Zi = 0)

) =

N∑

i=1

{ zi log

[ Φ ( xF,i𝛽F,

xD,i𝛽D, 𝜌 )] + ( 1 − zi

) log

[ 1−Φ

( xF,i𝛽F, xD,i𝛽D, 𝜌

)]} .

(6)

Results

P(F) in Model 1 of Table 2 introduces the first inde- pendent variable, dedicated institutional ownership. The coefficient estimate of dedicated investors is positive (𝛽 = 1.913, p = 0.026), lending support to Hypothesis 1, which suggests a positive relationship between the level of dedicated institutional own- ership and the likelihood of committing financial fraud. In terms of economic significance, when ded- icated institutional ownership increases from mean (0.045) to mean plus one standard deviation (0.112), holding all other variables at their means, the per- centage increase in the likelihood of firms’ commit- ting financial fraud is 36%. In P(D|F) in Model 1, the coefficient estimate of institutional ownership is positive but statistically not significant, suggesting that institutional ownership may not influence fraud detection.

Hypothesis 2 states that takeover defenses are negatively associated with the likelihood of com- mitting financial fraud. The estimated coefficient for takeover defense provisions in P(F) of Model 2 is negative and is associated with a p-value of 0.008 (𝛽 = -0.202, p = 0.008), consistent with Hypothe- sis 2. In terms of economic impact, when the number of takeover defense provisions increases from zero to one, holding all other variables at their means, the percentage decrease in the likeli- hood of firms’ committing financial fraud is 37%. In P(D|F) in Model 2, the coefficient estimate of takeover defense is statistically not significant, indi- cating that takeover defense provisions may not bear a relationship with the likelihood of fraud detection.

Hypothesis 3 states that earnings pressure from analysts’ recommendations is positively associ- ated with the likelihood of committing financial fraud. The estimated coefficient for this inde- pendent variable in P(F) of Model 3 is posi- tive and is associated with a p-value of 0.005 (𝛽 = 1.056, p = 0.005), supporting Hypothesis 3. In terms of economic magnitude, when analysts’ pres- sure increases from mean (0.560) to mean plus one standard deviation (0.785), holding all other vari- ables at their means, the percentage increase in the likelihood of firms’ committing financial fraud is 82%. In P(D|F) in Model 2, the coefficient esti- mate of analysts’ pressure is statistically not sig- nificant, indicating that analysts’ pressure may not bear a relationship with the likelihood of fraud detection.

1280 W. Shi, B. L. Connelly, and R. E. Hoskisson

Additional analyses

In our main study and analysis, we theorize that and test whether external monitoring and pressure from dedicated investors, the market for corporate control, and financial analysts result in higher lev- els of financial fraud. One could make additional related arguments, though, that a moderate amount of external pressure may be necessary and dis- courage top managers from committing financial fraud, even if too much external pressure exacer- bates fraud. Therefore, in supplementary analyses, we test whether there were curvilinear relationships between our independent variables and dependent variable. However, we failed to find statistical sup- port for such relationships. Results for this analysis and other unreported results are available from the authors on request.

Also, as described in our theory and methods, we examine external pressure from analysts’ rec- ommendations in terms of both buy and sell recom- mendations. To investigate this further in a post-hoc manner, we subsequently considered the individual relationships between buy/sell recommendations and fraud commitment. We found that the coef- ficient estimate of the percent of buy recommen- dations is positive and is statistically significant, whereas the coefficient estimate of the percent of sell recommendations is positive but statistically not significant. This suggests that high earnings expec- tations from positive analysts’ recommendations appear to exert a stronger influence on top man- agers’ motivation to commit financial fraud com- pared to the pressure managers feel from having to turn around performance in the face of negative recommendations.

As described in the analysis section above, our chosen methodology is a unique approach to inves- tigating the likelihood of fraud. Therefore, we con- firm our results using an alternative method based on a matched-pair sample (Arthaud-Day et al., 2006; Cumming, Leung, and Rui, 2015; Gomulya and Boeker, 2014; O’Connor et al., 2006). For each fraud firm, we found a control firm that is most sim- ilar to fraud firm in terms of firm size (log assets) and Tobin’s q and belongs to the same Fama and French 12 industry classification (Fama and French, 1997). We require that control firms have never been charged with fraud. To analyze the matched-pair sample, we used conditional logistic regressions, which recognize the conditional nature of the prob- abilities that matched-pair samples create (Manski

and Lerman, 1977). Conditional logistic regressions control for time-invariant paired fixed-effects, but cannot address potential biases caused by fraud that has been committed but not detected. In unreported results, we find support for our three hypotheses.

Supplementary study

The analyses above suffer from some limitations common to studies of archival data. For example, our control variables cannot account for all possible alternative explanations and the control variables we include could overlap with the explained vari- ance of our predictors, thus changing what it is our predictors are actually measuring (Audia, Locke, and Smith, 2000). Another problem with archival studies of market data is that we cannot actually measure people’s thoughts, but rather are limited to viewing outcomes. This may be important because our theory deals with intrinsic motivation, and with archival analyses, we make inferences about how managers are motivated based on how we see them behaving.

Therefore, we develop a supplementary study, the details of which we provide in Appendix S1, to observe managerial decision making directly in a more controlled environment (Priem, Walters, and Li, 2011). We used a policy-capturing approach in a survey-based study with strengths, and limitations, that are complementary to our prior analyses. In short, we ask executives to imagine they are the CEO of a small but publicly traded company. The company received a sale a few days into Q1 of the current fiscal year, but it would help the CEO if he or she could book it in Q4 of the prior year, which is when most of the work for the sale actually occurred anyway. Results are presented in Table 3.

We analyze the data using hierarchical linear modeling to account for both within-person and between-person variance (Spence and Keeping, 2010). The dependent variable is Managers’ reac- tion, which is how they would actually report the sale. For each scenario, participants select their response on a seven-point Likert scale to the state- ment “Based solely on the information provided here, I might consider reporting the sale in Q4 of the prior year.” Answers range from “strongly dis- agree” to “strongly agree.” We inform respondents that they should report the sale in Q1, but report- ing it in Q4 would engender both personal benefits and risk to the firm. In other words, indicating they

External Corporate Governance and Financial Fraud 1281

Table 3. Hierarchical regression of intent to commit fraud

Model 1 Model 2 Variables Controls Main effects

Experience −0.071 −0.071 [0.031] [0.031]

Gender −0.658 −0.658 [0.326] [0.326]

Locus of causality 0.131 0.131 [0.839] [0.839]

CEO power −0.050 −0.050 [0.281] [0.275]

Shareholder pressure 0.086 [0.064]

Takeover defense −0.077 [0.097]

Analysts’ pressure 0.121 [0.009]

Race (dummy variables) Included Included Education (dummy variables) Included Included Raised place (dummy variables) Included Included Constant 4.802 4.802

[3.231] [3.231] Chi-squared 17.63 30.65 Log-likelihood −722.8 −722.9

N = 456. P-values based on Huber-White robust errors reported in brackets.

would report the sale in Q4 is equivalent to intent to commit fraud.

Model 1 in Table 3 shows the effects of the control variables. Model 2 in Table 3 shows our supplementary study’s measurement of the main effects advanced in Hypotheses 1 through 3, with the dependent variable being intent to commit fraud. Consistent with the results from our main study, the coefficient estimate of pressure from shareholders is positive (𝛽 = 0.086, p = 0.064), consistent with Hypothesis 1. Also in Model 2, the coefficient esti- mate of takeover defenses is negative (𝛽 = -0.077, p = 0.097), consistent with Hypothesis 2. The drop in significance level may be due to the small number of observations used in this study. The coefficient estimate of analysts’ pressure is positive (𝛽 = 0.121, p = 0.009), supporting Hypothesis 3.

DISCUSSION

In this study, we extend agency theory by examin- ing external mechanisms of corporate governance in view of managerial cognitions. We find empirical support for the notion that external governance can

dampen managers’ intrinsic motivation to act in the interest of shareholders, increasing their likelihood of financial fraud. Each of the three external gov- ernance mechanisms under investigation — activist shareholders, the market for corporate control, and rating agencies — provides unique explanatory value in the context of financial fraud, and each runs counter to traditional agency predictions.

Key contributions

Our results hold the potential for contributing to the literature in several ways. Foremost, behav- ioral agency theory incorporates cognitive biases into agency theory assumptions about internal governance (Martin et al., 2013; Wiseman and Gomez-Mejia, 1998), but this stream of research devotes less attention to external governance. Our study adds to the academic community’s understanding of agency theory by introducing a key behavioral consideration into agency theory’s predictions about external governance. In so doing, our study questions the utility of external gover- nance mechanisms for reigning in the potential for moral hazard, in our case, in the form of managerial financial fraud. Whereas the governance literature has prescribed (over the long years of research in corporate governance) several alignment mech- anisms that policymakers expect to work most of the time, our study shows that some of these mechanisms may not work as expected.

Our study also potentially provides new insights into the consequences of investor activism. Agency theory suggests that dedicated institutional investors, who are well known for their activist approach to ownership (Goranova and Ryan, 2014), can mitigate agency problems because they have incentive to monitor managers and the ability to bring about change due to the size of their holdings (Shleifer and Vishny, 1997). Existing studies lend empirical support to the notion that dedicated investors can be conducive to mitigating managerial risk aversion and encouraging managers to make decisions that create long-term value for the firm (Bushee, 1998; Connelly et al., 2010b; Hoskisson et al., 2002). However, researchers have devoted little attention to possible trade-offs of the powerful expectations imposed on managers via the external pressure of dedicated investors. Our results show that an ownership structure laden with dedicated investors can have unanticipated consequences in the form of diminished intrinsic motivation,

1282 W. Shi, B. L. Connelly, and R. E. Hoskisson

resulting in higher instances of managerial finan- cial fraud. There is an ongoing policy debate about the effects of activist shareholders (Benoit and Hoffman, 2015; Gandel, 2015). Our study informs this debate by showing that powerful shareholders could perhaps be hurting more than helping as they magnify the Wall Street “expectations game,” and consequently, managers could abandon their own beliefs in order to satisfy the expectations of shareholders (Martin, 2011).

Our findings may also contribute to research on takeover defenses. Existing studies generally suggest that the market perceives takeover defense provisions negatively because they result in reduced firm value (Gompers et al., 2003; Mahoney and Mahoney, 1993; Sundaramurthy et al., 1997). Research shows that managerial entrenchment brought on by takeover defenses can cushion man- agers’ exposure to the market for corporate control. Our study, however, adds that such protection can provide a positive effect wherein managers maintain higher levels of intrinsic motivation, akin to stewardship theory, as opposed to being motivated mainly by extrinsic factors, which are central to agency theory (Sundaramurthy and Lewis, 2003). Future research might tease out even further this juxtaposition of the principles and assumptions underlying cognitive evaluation theory, and relatedly, stewardship theory versus those that underlie agency theory.

We also contribute to a nascent stream of literature that explores the benefits of takeover defense provisions. For instance, Danielson and Karpoff (2006) find that firms that have adopted poison pills witness modest operating performance improvement over time. Similarly, findings by Kacperczyk (2009) and Wang et al. (forthcoming) suggest that an exogenous increase in takeover protection leads managers to focus on strategic decisions that can increase shareholders’ long-term interests. Our results suggest that the benefits of takeover defense provisions extend beyond issues of performance as they discourage managers from engaging in financial fraud, which is detrimental to the interests of shareholders and other stakeholders. Relatedly, researchers might also consider how different kinds of corporate governance provisions operate in different ways, such as those that prevent actions that could lead to a takeover versus those that take effect only if a takeover occurs.

Last, our study illustrates the power that secu- rities analysts wield. Although analysts play an

important role as information intermediaries that can help reduce information asymmetry between investors and companies, their recommendations introduce powerful expectations on managers to perform, which could influence their financial reporting decisions. While much of the research on rating agencies focuses on how investors react to analysts’ recommendations, our study builds on the more limited amount of research (Gentry and Shen, 2013; Zhu and Westphal, 2011) that explores how managers react, and behave differently in response, to analysts’ recommendations. Our study has focused on one type of external rating agen- cies (i.e., financial analysts), but future research might explore how other external rating agencies (e.g., journalists) shape managers’ performance expectations, and thereby, influence managerial motivations to engage in financial fraud (Shani and Westphal, 2016; Westphal and Deephouse, 2011).

The focus of our study is on how performance expectations from external governance mechanisms crowd out top managers’ intrinsic motivation, lead- ing to higher instances of financial fraud. Future research might consider whether and how this gen- eralizes to managerial misconduct. Further, schol- ars could extend this research by considering how cognitive evaluation theory applies to internal gov- ernance devices, such as boards of directors and top manager compensation designs. Similarly, scholars might also examine whether internal and external governance devices are complements or substitutes. It could be, for example, that dedicated institutional investors substitute for vigilant boards because both introduce strong expectations on managers. If this is the case, then the monitoring role of vigilant boards could become less important when dedi- cated investors are present. Conversely, one could make an argument that dedicated investors comple- ment vigilant boards. In this case, the two work together to inflict such weighty expectations that managers reach a tipping point where they feel they have to resort to fraud so that they do not let every- body down.

CONCLUSION

In sum, our findings suggest that policymakers may face a paradox in regulating corporate governance. Imposing strict external monitoring and control can decrease top managers’ intrinsic motivation and reduce their focus on internal values, potentially

External Corporate Governance and Financial Fraud 1283

leading them to commit financial fraud. However, granting top managers too much freedom from external performance pressure could result in some managers extracting personal gains at the expense of shareholders. Perhaps managers can “earn the right” to autonomy over time as they demonstrate that they consistently act in the best interest of shareholders, despite who may or may not be looking over their shoulders.

ACKNOWLEDGEMENTS

Guidance and comments provided by SMJ Editor Will Mitchell and two anonymous SMJ reviewers have significantly improved this article.

REFERENCES

Aguilera R, Desender K, Bednar MK, Lee JH. 2015. Con- necting the dots: bringing external corporate gover- nance into the corporate governance puzzle. Academy of Management Annals 9(1): 483 – 573.

Argyris C. 1964. Integrating the Individual and the Orga- nization. Wiley: New York.

Arthaud-Day ML, Certo ST, Dalton CM, Dalton DR. 2006. A changing of the guard: executive and director turnover following corporate financial restatements. Academy of Management Journal 49(6): 1119 – 1136.

Arthurs JD, Busenitz LW. 2003. The boundaries and limitations of agency theory and stewardship theory in the venture capitalist/entrepreneur relationship. Entrepreneurship: Theory and Practice 28(2): 145 – 162.

Audia PG, Locke EA, Smith KG. 2000. The paradox of success: an archival and a laboratory study of strategic persistence following radical environmental change. Academy of Management Journal 43(5): 837 – 853.

Barkema HG. 1995. Do top managers work harder when they are monitored? Kyklos 48(1): 19 – 42.

Barsky A. 2008. Understanding the ethical cost of organi- zational goal-setting: a review and theory development. Journal of Business Ethics 81(1): 63 – 81.

Baucus MS, Near JP. 1991. Can illegal corporate behavior be predicted: an event history analysis. Academy of Management Journal 34(1): 9 – 36.

Baysinger BD, Kosnik RD, Turk TA. 1991. Effects of board and ownership structure on corporate R&D strat- egy. Academy of Management Journal 34(1): 205 – 214.

Beasley MS, Carcello JV, Hermanson DR, Lapides PD. 2000. Fraudulent financial reporting: consideration of industry traits and corporate governance mechanisms. Accounting Horizons 14(4): 441 – 454.

Bebchuk L, Cohen A, Ferrell A. 2009. What matters in corporate governance? Review of Financial Studies 22(2): 783 – 827.

Becker GS. 1976. The Economic Approach to Human Behavior. University of Chicago Press: Chicago, IL.

Benner MJ, Ranganathan R. 2012. Offsetting illegiti- macy? How pressures from securities analysts influence incumbents in the face of new technologies. Academy of Management Journal 55(1): 213 – 233.

Benoit D, Hoffman DL. 2015. Taking sides on activist investors. Wall Street Journal 20 March. Available at: http://www.wsj.com/articles/strine-urges-closing-of- 10-day-investment-disclosure-window-1426791548.

Bertelli AM. 2006. Motivation crowding and the federal civil servant: evidence from the U.S. Internal Rev- enue Service. International Public Management Jour- nal 9(1): 3 – 23.

Beunza D, Garud R. 2007. Calculators, lemmings or frame-makers? The intermediary role of securities ana- lysts. Sociological Review 55: 13 – 39.

Boal KB, Cummings LL. 1981. Cognitive evaluation theory: an experimental test of processes and out- comes. Organizational Behavior and Human Perfor- mance 28(3): 289 – 310.

Boivie S, Graffin SD, Pollock TG. 2012. Time for me to fly: predicting director exit at large firms. Academy of Management Journal 55(6): 1334 – 1359.

Bradshaw MT. 2004. How do analysts use their earn- ings forecasts in generating stock recommendations? Accounting Review 79(1): 25 – 50.

Brav A, Jiang W, Partnoy F, Thomas R. 2008. Hedge fund activism, corporate governance, and firm performance. Journal of Finance 63(4): 1729 – 1775.

Brown NC, Wei KD, Wermers R. 2013. Analyst recom- mendations, mutual fund herding, and overreaction in stock prices. Management Science 60(1): 1 – 20.

Bushee BJ. 1998. The influence of institutional investors on myopic R&D investment behavior. Accounting Review 73(3): 305 – 333.

Bushee BJ. 2001. Do institutional investors prefer near-term earnings over long-run value? Contemporary Accounting Research 18(2): 207 – 246.

Bushee BJ. 2004. Identifying and attracting the “right” investors: evidence on the behavior of institutional investors. Journal of Applied Corporate Finance 16(4): 28 – 35.

Carberry EJ, King BG. 2012. Defensive practice adoption in the face of organizational stigma: impression man- agement and the diffusion of stock option expensing. Journal of Management Studies 49(7): 1137 – 1167.

Chen GM, Firth M, Gao DN, Rui OM. 2006. Owner- ship structure, corporate governance, and fraud: evi- dence from China. Journal of Corporate Finance 12(3): 424 – 448.

Chen T, Harford J, Lin C. 2015. Do analysts matter for governance? Evidence from natural experiments. Journal of Financial Economics 115(2): 383 – 410.

Chrisman JJ, Patel PC. 2012. Variations in R&D invest- ments of family and nonfamily firms: behavioral agency and myopic loss aversion perspectives. Academy of Management Journal 55(4): 976 – 997.

Coffee JC. 2006. Gatekeepers: The Professions and Cor- porate Governance. Oxford University Press: Oxford, UK.

Connelly BL, Hoskisson RE, Tihanyi L, Certo ST. 2010a. Ownership as a form of corporate governance. Journal of Management Studies 47(8): 1561 – 1589.

http://www.wsj.com/articles/strine-urges-closing-of-10-day-investment-disclosure-window-1426791548

1284 W. Shi, B. L. Connelly, and R. E. Hoskisson

Connelly BL, Tihanyi L, Certo ST, Hitt MA. 2010b. Marching to the beat of different drummers: the influence of institutional owners on competitive actions. Academy of Management Journal 53(4): 723 – 742.

Cumming D, Leung TY, Rui O. 2015. Gender diversity and securities fraud. Academy of Management Journal 58(5): 1572 – 1593.

Dalton DR, Hitt MA, Certo ST, Dalton CM. 2007. The fundamental agency problem and its mitigation: inde- pendence, equity, and the market for corporate control. Academy of Management Annals 1: 1 – 64.

Danielson MG, Karpoff JM. 2006. Do pills poison operat- ing performance? Journal of Corporate Finance 12(3): 536 – 559.

Davidoff SM. 2013. A push to end securities fraud lawsuits gains momentum. New York Times 15 October.

Dechow PM, Ge WL, Larson CR, Sloan RG. 2011. Predict- ing material accounting misstatements. Contemporary Accounting Research 28(1): 17 – 82.

Deci EL. 1971. Effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology 18(1): 105 – 115.

Deci EL. 1975. Intrinsic Motivation. Plenum Press: New York.

Deci EL, Ryan RM. 1985. Intrinsic Motivation and Self-Determination in Human Behavior. Plenum: New York.

Deci EL, Ryan RM. 2000. The “what” and “why” of goal pursuits: human needs and the self-determination of behavior. Psychological Inquiry 11(4): 227 – 268.

Deci EL, Ryan AM. 2012. Self-determination theory. In Handbook of Theories of Social Psychology, Lange PAMV, Kruglanski AW, Higgins ET (eds). SAGE: Los Angeles, CA.

Demirguc-Kunt A, Maksimovic V. 1998. Law, finance, and firm growth. Journal of Finance 53(6): 2107 – 2137.

Duffy D. 2015. CEOs need a “healthy paranoia” of activist investors. Available at: http://www.cnbc.com/ 2015/08/12/ceos-need-a-healthy-paranoia-of-activist- investors-commentary.html (accessed 12 Dec 2015).

Dyck A, Morse A, Zingales L. 2010. Who blows the whistle on corporate fraud? Journal of Finance 65(6): 2213 – 2253.

Eaglesham J, Rapoport M. 2015. SEC gets busy with accounting investigations. Wall Street Journal 20 January.

Fama EF, French KR. 1997. Industry costs of equity. Journal of Financial Economics 43(2): 153 – 193.

Feinstein JS. 1990. Detection controlled estimation. Jour- nal of Law and Economics 33(1): 233 – 276.

Filatotchev I, Toms S. 2006. Corporate governance and financial constraints on strategic turnarounds. Journal of Management Studies 43(3): 407 – 433.

Frey BS, Jegen R. 2001. Motivation crowding theory. Journal of Economic Surveys 15(5): 589 – 611.

Gagne M, Deci EL. 2005. Self-determination theory and work motivation. Journal of Organizational Behavior 26(4): 331 – 362.

Gande A, Lewis CM. 2009. Shareholder-initiated class action lawsuits: shareholder wealth effects and industry

spillovers. Journal of Financial and Quantitative Anal- ysis 44(4): 823 – 850.

Gandel S. 2015. SEC’s Mary Jo White goes after share- holder activism and Bill Ackman deal. Fortune 23 March. Available at: http://fortune.com/2015/03/19/ mary-jo-white-activist-investors-ackman.

Gaspar JM, Massa M, Matos P. 2005. Shareholder invest- ment horizons and the market for corporate control. Journal of Financial Economics 76(1): 135 – 165.

Gentry RJ, Shen W. 2013. The impacts of performance relative to analyst forecasts and analyst coverage on firm R&D intensity. Strategic Management Journal 34(1): 121 – 130.

Georgellis Y, Iossa E, Tabvuma V. 2011. Crowding out intrinsic motivation in the public sector. Journal of Public Administration Research and Theory 21(3): 473 – 493.

Gompers P, Ishii J, Metrick A. 2003. Corporate governance and equity prices. Quarterly Journal of Economics 118(1): 107 – 155.

Gomulya D, Boeker W. 2014. How firms respond to financial restatement: CEO successors and external reactions. Academy of Management Journal 57(6): 1759 – 1785.

Goranova M, Dharwadkar R, Brandes P. 2010. Owners on both sides of the deal: mergers and acquisitions and overlapping institutional ownership. Strategic Manage- ment Journal 31(10): 1114 – 1135.

Goranova M, Ryan LV. 2014. Shareholder activism: a mul- tidisciplinary review. Journal of Management 40(5): 1230 – 1268.

Hadani M, Goranova M, Khan R. 2011. Institutional investors, shareholder activism, and earnings man- agement. Journal of Business Research 64(12): 1352 – 1360.

He J, Tian X. 2013. The dark side of analyst coverage: the case of innovation. Journal of Financial Economics 109(3): 856 – 878.

Higgins MC, Gulati R. 2006. Stacking the deck: the effects of top management backgrounds on investor decisions. Strategic Management Journal 27(1): 1 – 25.

Hirsch P, Pozner J-E. 2005. To avoid surprises, acknowl- edge the dark side: illustrations from securities analysts. Strategic Organization 3(2): 229 – 238.

Hitt MA, Hoskisson RE, Johnson RA, Moesel DD. 1996. The market for corporate control and firm innovation. Academy of Management Journal 39(5): 1084 – 1119.

Hoskisson RE, Castleton MW, Withers MC. 2009. Com- plementarity in monitoring and bonding: more intense monitoring leads to higher executive compensation. Academy of Management Perspectives 23(2): 57 – 74.

Hoskisson RE, Hitt MA, Johnson RA, Grossman W. 2002. Conflicting voices: the effects of institutional owner- ship heterogeneity and internal governance on corpo- rate innovation strategies. Academy of Management Journal 45(4): 697 – 716.

Humphery-Jenner M. 2014. Takeover defenses, inno- vation, and value creation: evidence from acquisi- tion decisions. Strategic Management Journal 35(5): 668 – 690.

http://www.cnbc.com/2015/08/12/ceos-need-a-healthy-paranoia-of-activist-investors-commentary.html

http://fortune.com/2015/03/19/mary-jo-white-activist-investors-ackman

External Corporate Governance and Financial Fraud 1285

Jacquart P, Armstrong JS. 2013. Are top executives paid enough? An evidence-based review. Interfaces 43(6): 580 – 589.

Jensen MC, Meckling WH. 1976. Theory of the firm: man- agerial behavior, agency costs and ownership structure. Journal of Financial Economics 3(4): 305 – 360.

Jensen MC, Ruback RS. 1983. The market for corporate control: the scientific evidence. Journal of Financial Economics 11(1 – 4): 5 – 50.

Kabir R, Cantrijn D, Jeunink A. 1997. Takeover defenses, ownership structure and stock returns in the Nether- lands: an empirical analysis. Strategic Management Journal 18(2): 97 – 109.

Kacperczyk A. 2009. With greater power comes greater responsibility? Takeover protection and corporate attention to stakeholders. Strategic Management Journal 30(3): 261 – 285.

Kanfer R. 1990. Motivation theory and industrial and orga- nizational psychology. In Handbook of Industrial and Organizational Psychology, Dunnette MD, Hough LM, Triandis HC (eds). Consulting Psychologists Press: Mountain View, CA.

Khanna V, Kim E, Lu Y. 2015. CEO connectedness and corporate fraud. Journal of Finance 70(3): 1203 – 1252.

Kini O, Kracaw W, Mian S. 2004. The nature of disci- pline by corporate takeovers. Journal of Finance 59(4): 1511 – 1552.

Koh PS. 2007. Institutional investor type, earnings man- agement and benchmark beaters. Journal of Accounting and Public Policy 26(3): 267 – 299.

Lang MH, Lins KV, Miller DP. 2004. Concentrated con- trol, analyst following, and valuation: do analysts mat- ter most when investors are protected least? Journal of Accounting Research 42(3): 589 – 623.

Larraza-Kintana M, Wiseman RM, Gomez-Mejia LR, Welbourne TM. 2007. Disentangling compensation and employment risks using the behavioral agency model. Strategic Management Journal 28(10): 1001 – 1019.

Maffett M. 2012. Financial reporting opacity and informed trading by international institutional investors. Journal of Accounting and Economics 54(2 – 3): 201 – 220.

Mahoney JM, Mahoney JT. 1993. An empirical investi- gation of the effect of corporate charter antitakeover amendments on stockholder wealth. Strategic Manage- ment Journal 14(1): 17 – 31.

Mahoney JM, Sundaramurthy C, Mahoney JT. 1997. The effects of corporate antitakeover provisions on long-term investment: empirical evidence. Managerial and Decision Economics 18(5): 349 – 365.

Manski CF, Lerman SR. 1977. The estimation of choice probabilities from choice based samples. Econometrica 45(8): 1977 – 1988.

Martin R. 2011. Fixing the Game: How Runaway Expec- tations Broke the Economy, and How to Get Back to Reality. Harvard Business Review Press: Boston, MA.

Martin GP, Gomez-Mejia LR, Wiseman RM. 2013. Exec- utive stock options as mixed gambles: revisiting the behavioral agency model. Academy of Management Journal 56(2): 451 – 472.

McGurn PS. 2002. Classification cancels corporate accountability. Stanford Law Review 55(3): 839 – 844.

Mishina Y, Dykes BJ, Block ES, Pollock TG. 2010. Why “good” firms do bad things: the effects of high aspirations, high expectations, and prominence on the incidence of corporate illegality. Academy of Manage- ment Journal 53(4): 701 – 722.

O’Connor JP, Priem RL, Coombs JE, Gilley KM. 2006. Do CEO stock options prevent or promote fraudulent financial reporting? Academy of Management Journal 49(3): 483 – 500.

Osterloh M, Frey BS. 2000. Motivation, knowledge trans- fer, and organizational forms. Organization Science 11(5): 538 – 550.

Osterloh M, Frey BS. 2004. Corporate governance for crooks? The case for corporate virture. In Corporate Governance and Firm Organization, Grandori G (ed). Oxford University Press: Oxford, UK; 191 – 211.

Osterloh M, Frost J, Frey BS. 2002. The dynamics of motivation in new organizational forms. International Journal of the Economics of Business 9(1): 61 – 77.

Pepper A, Gore J. 2015. Behavioral agency theory: new foundations for theorizing about executive compensa- tion. Journal of Management 41(4): 1045 – 1068.

Pepper A, Gosling T, Gore J. 2015. Fairness, envy, guilt and greed: Building equity considerations into agency theory. Human Relations 68(8): 1291 – 1314.

Phillips JS, Lord RG. 1981. Causal attributions and per- ceptions of leadership. Organizational Behavior and Human Performance 28(2): 143 – 163.

Poirier DJ. 1980. Partial observability in bivariate probit models. Journal of Econometrics 12(2): 209 – 217.

Porter ME. 1992. Capital Choices: Changing the Way America Invests in Industry. Harvard Business School Press: Boston, MA.

Priem RL, Walters BA, Li SL. 2011. Decisions, decisions! How judgment policy studies can integrate macro and micro domains in management research. Journal of Management 37(2): 553 – 580.

Rao H, Sivakumar K. 1999. Institutional sources of boundary-spanning structures: the establishment of investor relations departments in the Fortune 500 indus- trials. Organization Science 10(1): 27 – 42.

Ryan RM, Deci EL. 2000. Self-determination theory and the facilitation of intrinsic motivation, social develop- ment, and well-being. American Psychologist 55(1): 68 – 78.

Schipper K. 1991. Analysts’ forecasts. Accounting Hori- zons 5: 105 – 121.

Schweitzer ME, Ordóñez L, Douma B. 2004. Goal setting as a motivator of unethical behavior. Academy of Management Journal 47(3): 422 – 432.

Shani G, Westphal J. 2016. Persona non grata? Deter- minants and consequences of social distancing from journalists who engage in negative coverage of firm leadership. Academy of Management Journal 59(1): 302 – 329.

Sharma VD. 2004. Board of director characteristics, insti- tutional ownership, and fraud: evidence from Australia. Auditing: A Journal of Practice and Theory 23(2): 105 – 117.

Shi W, Connelly BL, Sanders WG. 2016. Buying bad behavior: tournament incentives and securities class

1286 W. Shi, B. L. Connelly, and R. E. Hoskisson

action lawsuits. Strategic Management Journal 37(7): 1354 – 1378.

Shleifer A, Vishny RW. 1997. A survey of corporate governance. Journal of Finance 52(2): 737 – 783.

Spence JR, Keeping LM. 2010. The impact of non-performance information on ratings of job performance: a policy-capturing approach. Journal of Organizational Behavior 31(4): 587 – 608.

Stickel SE. 1995. The anatomy of the performance of buy and sell recommendations. Financial Analysts Journal 51(5): 25 – 39.

Stone DN, Deci EL, Ryan RM. 2009. Beyond talk: creat- ing autonomous motivation through self-determination theory. Journal of General Management 34(3): 75 – 91.

Sundaramurthy C, Lewis M. 2003. Control and collabo- ration: paradoxes of governance. Academy of Manage- ment Review 28(3): 397 – 415.

Sundaramurthy C, Mahoney JM, Mahoney JT. 1997. Board structure, antitakeover provisions, and stock- holder wealth. Strategic Management Journal 18(3): 231 – 245.

Teoh SH, Welch I, Wong TJ. 1998. Earnings management and the underperformance of seasoned equity offerings. Journal of Financial Economics 50(1): 63 – 99.

Wahal S, McConnell JJ. 2000. Do institutional investors exacerbate managerial myopia? Journal of Corporate Finance 6(3): 307 – 329.

Wang TY. 2013. Corporate securities fraud: Insights from a new empirical framework. Journal of Law, Economics, and Organization 29(3): 535 – 568.

Wang TY, Winton A, Yu XY. 2010. Corporate fraud and business conditions: evidence from IPOs. Journal of Finance 65(6): 2255 – 2292.

Wang H, Zhao S, He J. Forthcoming. Firm-specific knowl- edge assets and CEO turnover. Strategic Management Journal.

Weiss EJ, Beckerman JS. 1995. Let the money do the mon- itoring: how institutional investors can reduce agency costs in securities class actions. Yale Law Journal 104(8): 2053 – 2127.

Westphal JD, Deephouse DL. 2011. Avoiding bad press: interpersonal influence in relations between CEOs and journalists and the consequences for press reporting about firms and their leadership. Organization Science 22(4): 1061 – 1086.

Wiersema MF, Zhang Y. 2011. CEO dismissal: the role of investment analysts. Strategic Management Journal 32(11): 1161 – 1182.

Wiseman RM, Gomez-Mejia LR. 1998. A behavioral agency model of managerial risk taking. Academy of Management Review 23(1): 133 – 153.

Womack KL. 1996. Do brokerage analysts’ recommenda- tions have investment value? Journal of Finance 51(1): 137 – 167.

Zhu DH, Westphal JD. 2011. Misperceiving the beliefs of others: how pluralistic ignorance contributes to the persistence of positive security analyst reactions to the adoption of stock repurchase plans. Organization Science 22(4): 869 – 886.

Zuckerman EW. 2000. Focusing the corporate product: securities analysts and de-diversification. Administra- tive Science Quarterly 45(3): 591 – 619.

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article:

Appendix S1. Policy capturing study.