. ~r: “a" .53. -. , 9553mm gaff .~ ~ '57:? ”“1221“ 8 . ,3: .. ; .. ”LB. 94- 1:?“ ..‘ u. . . ‘ ‘ i. ‘ 5Li§$3€3i42 J3. K" " «WW. I ‘1' 3'V ‘ of “Us ‘ p :6 ’J . .‘.| ...‘ L‘"."'I.1g’( f» . a .. \ Q " A $2 \n- w « 43:17: .5' 333': , 7? . x #5? 4 g. as?) or“ ,3, - ”L‘fifé‘ifi '- “tn ’9 P ’ 4' "3:1 4%: This is to certify that the dissertation entitled An Empirical Analysis of the Relation Between Corporate Governance and Management Fraud presented by Mark Swearingen Beasley has been accepted towards fulfillment of the requirements for PhD degree in Accounting v / - Major professor Date QW— MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 . illllflilillllllMWNIH“NW\IH‘HUHIHWIHHI 3 1293 010275 ———.—. LIBRARY “— M‘Chigan State University PLACE N RETURN BOXto romovothlo checkout from your record. TOAVOIDFINESrotunonorbdoroddoduo. DATE DUE DATE DUE DATE DUE MSU loAn Afflmotlvo ActionlEquol Opportunity Institution WM‘ AN EMPIRICAL ANALYSIS OF THE RELATION BETWEEN CORPORATE GOVERNANCE AND MANAGEMENT FRAUD By Mark Swearingen Beasley A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Accounting 1994 ABSTRACT AN EMPIRICAL ANALYSIS OF THE RELATION BETWEEN CORPORATE GOVERNANCE AND MANAGEMENT FRAUD By Mark Swearingen Beasley This study is an empirieal examination of the relation between board of director composition and the occurrence of management fraud. Economic theory of the firm suggests that unique board of director composition may help to reduce management fraud. This study exploits variations in board of director composition to examine this theory, which no prior study has empirically tested. The research methodology involves logit cross-sectional regression analysis that examines differences in board of director composition between seventy-five fraud firms and seventy-five no-fraud firms. Each of the seventy-five fraud firms was matched with a no-fraud firm on the basis of firm size, industry, national stock exchange where the firm’s stock trades, and time period because review of the management fraud and corporate governance research indieates that these variables may be associated with both management fraud and board of director composition. In addition, four control variables were included in the logit model to control for differences in other non-board characteristics. The logit regression results confirm the predicted relation between board of director composition and the occurrence of management fraud. No-fraud firms have signifieantly higher percentages of outside members on the board of directors than fraud firms, and the outside directors of no-fraud firms have significantly greater ownership levels in the firm, longer tenures on the board of directors, and fewer outside directorships in other firms. Managers who serve on the board of directors of fraud and no-fraud firms differ in the extent of their ownership in the firm only when they hold moderate levels of outstanding shares of common stock - between 5 % and 25 96 . Chairpersons of fraud firms are not more likely to hold managerial positions, such as CEO or president, than Chairpersons of no-fraud firms, and CEO tenure does not differ between fraud and no-fraud firms. Finally, this study demonstrates the importance of an active audit committee of the board given that no-fraud firms compared to fraud firms are significantly more likely to have an active audit committee. ACKNOWLEDGMENTS I would like to express my sincere appreciation to my dissertation committee - Dr. Alvin Arens, Dr. Kathy Petroni, Dr. Mary Bange, and Dr. Frank Boster for their invaluable encouragement and sincere commitment of time and effort on my behalf. I especially want to thank my chairperson, Al Arens, for the tremendous support he has provided throughout my doctoral program. I am grateful to the American Institute of Certified Public Accountants for their financial support. I am also grateful to the Institute of Internal Auditors for their generous support, which made possible the timely completion of this dissertation. I would like to especially thank Dr. C. Wayne Alderman, Dr. Daniel M. Guy, and Mr. Donald L. Neebes for being such positive role models for me. I express deep appreciation to my parents, Bill and Ann Beasley, for their love, patience and dedication throughout my life and for showing me how to accomplish life’s difficult goals. Thank you to my sisters, Melanie and Melinda, who are always cheering me on and to my wife’s parents, Jim and Janice Johnson, for allowing me to take their daughter to Michigan for four fun years. I would like to thank my son, Johnson, for showing me the real joy of life. Finally, I am deeply grateful to my wife, Beth, who has endured this process with me. Her love, encouragement, patience, and confidence in me has made this possible. Without her, I would have never made it. iv TABLE OF CONTENTS LIST OF TABLES ...................................... vii LIST OF FIGURES ...................................... viii CHAPTER 1 - INTRODUCTION .............................. 1 1.1 Overview of Hypotheses Examined ........................ 3 1.2 Motivation ....................................... 4 1.3 Overview of Sample Selection and Research Design .............. 8 1.4 Organization of Dissertation ............................. 9 CHAPTER 2 - THEORY AND HYPOTHESES DEVELOPMENT ......... 10 2.1 Underlying Theory: The Monitoring Role of the Board of Directors . . . 10 2.2 The Management Fraud Literature ........................ 14 2.3 The Corporate Governance Literature ...................... 18 2.4 Development of Hypotheses ............................ 20 2.4.1 Representation of Outsiders on Board .................. 21 2.4.2 Quality of Outside Board Members .................... 24 2.4.3 Board Members’ Ownership Stakes ................... 26 2436) WWW ......... 26 2.4.3(ii) WWW ............. 27 2.4.4 Management Power ............................. 30 24-46) WW .......... 31 2.4.4(ii) W ........................ 32 2.4.4(iii) MW ..................... 33 2.4.5 Active Audit Committee .......................... 33 2.5 Controlling for Board Size ............................. 36 2.6 Summary ....................................... 37 CHAPTER 3 - SAMPLE SELECTION AND DESCRIPTION ............ 38 3.1 Fraud Firm Selection ................................ 38 3.2 Matching Fraud with No-Fraud Firms ...................... 43 3.3 Board Composition Differences Between Fraud and No-Fraud Firms . . . 50 3.4 Effects of Differences in Other Firm Characteristics ............. 53 3.4.1 Growth ..................................... 58 3.4.2 Financial Health ............................... 60 3.4.3 Length of Time Publicly Traded ..................... 62 3.4.4 Blockholders ................................. 64 3.5 Summary ....................................... 65 CHAPTER 4 - RESEARCH DESIGN ........................... 66 4.1 Appropriateness of Logit Regression ...................... 66 4.2 Logit Regression Model - Outside Directors .................. 70 4.1.1 Variables of Interest ............................. 71 4.1.2 Control Variables .............................. 75 4. 3 Logit Regression Model- Independent Directors ............... 75 4. 4 Piecewise Regression- Non-Linear Effects of Firm Ownership ...... 81 4. 5 Summary ....................................... 86 CHAPTER 5 - EMPIRICAL RESULTS ......................... 88 5 .1 Logit Regression Results - Outside Directors ................. 88 5.2 Logit Regression Results - Independent Directors .............. 103 5 .3 Piecewise Logit Regression Analysis of Firm Ownership Levels ..... 108 5.4 Another Robustness Test ............................. 112 5 .5 Summary ...................................... 113 CHAPTER 6 - SUMMARY, CONCLUSIONS, LIMITATIONS, AND SUGGESTIONS FOR FUTURE RESEARCH ........... 118 6.1 Summary of Research Findings ......................... 118 6.1.1 Differences in Representation of Outside Directors ......... 120 6.1.2 Differences in Outside Director Characteristics ........... 120 6.1.3 Differences in Management Director Characteristics ........ 123 6.1.4 Differences in Other Characteristics .................. 124 6.2 Contributions to Management Fraud and Corporate Governance Research 125 6.3 Limitations ..................................... 127 6.4 Suggestions for Future Research ........................ 129 LIST OF REFERENCES ................................. 131 LIST OF TABLES Ialzlefitle 10 ll 12 Matching of Fraud Firms and No-Fraud Firms ............. Board Structure Statistics on 75 Fraud and 75 No-Fraud Firms . . . Non-Board of Director Differences Across Fraud and ......... No-Fraud Firms Descriptive Characteristics of Two Subsets of Outside ........ Directors: Grey and Independent Directors Cell Sizes for Low, Moderate, & High Outside and Management . . . . Ownership Levels, 75 Fraud and 75 No-Fraud Firms Outside Director Logit Regression Results, ............... 75 Fraud Firms Matched With 75 No-Fraud Firms Outside Director Logit Regression Results, 67 Fraudulent ...... Financial Reporting Firms Matched With 67 No-Fraud Firms Logit Regression Results - Using Growth in Net Sales, ........ 75 Fraud Firms Matched With 75 No-Fraud Firms Pearson Correlation Matrix of Independent Variables ......... Independent Director Logit Regression Results, ............ 75 Fraud Firms Matched With 75 No-Fraud Firms Piecewise Logit Regression Results for Analysis of Ownership Levels, 75 Fraud Firms Matched With 75 No-Fraud Firms Logit Regression Controlling For Size, Industry, and Stock ..... Exchange, 75 Fraud Firms Matched With 75 No-Fraud Firms ..46 ..51 ..59 ..79 85 ..90 ..95 100 101 114 LIST OF FIGURES Elam Title 1 22 Red Flag Indicators That Differ Significantly Across ......... l6 Fraud and No-Fraud Firms 2 Summary of Hypotheses About Board of Director Characteristics . . . . 20 and Management Fraud 3 Identification of 75 Fraud Firms ........................ 41 4 Source of Fraud Firm by Type of Management Fraud ........... 43 5 SIC Code Description for 75 Fraud Firms .................. 48 6 Linkage of Model Variables with Eight Hypotheses ............ 72 About Board of Director Characteristics 7 Summary of Logit Model Variable Definitions ............... 76 8 Summary of Hypotheses Testing ....................... 103 9 Summary of Empirical Findings ....................... 119 CHAPTER 1 - INTRODUCTION This study is an empirical examination of the relation between corporate governance and the occurrence of management hand. The focus of this study is on an important corporate governance mechanism: the board of directors. There are wide variations among firms in board of director composition, such as the degree of outsiders who serve on the board of directors, their affiliations with other organizations, and the degree of share ownership by management versus outsiders [Baysinger and Butler (1985), Jensen and Warner ( 1988)]. Economic theory of the firm suggests that unique board of director composition may help to reduce management fraud. This study exploits variations in board of director composition to examine this theory. Economic theory of the firm suggests that the board of directors is an important corporate governance mechanism within modern day corporations [Fama and Jensen (1983a)]. The board of directors is the ultimate internal control mechanism within the firm that arises out of the separation of decision control and residual risk-bearing [Fama (1980), Fama and Jensen (1983a)]. As the apex of decision control, one of the board of director’s primary responsibilities is to monitor management decisions and actions. Management fraud is one example of the agency problem that arises out of the separation of decision control and residual risk-bearing. Because most of the day-to-day actions of boards of directors are unobservable, management fraud provides a unique setting where the monitoring role of the board of directors can be examined ex post. Specifically, this study tests economic theory that suggests board of director composition impacts the board’s effectiveness as a monitor of management for the prevention of management fraud. 2 Little is known about the relation of board of director composition and the occurrence of management fraud. Previous studies examining characteristics of firms experiencing management fraud primarily identify ”red flag” indicators that suggest the presence of management fraud. These studies note the existence of ”weak internal control environments” for fraud firms [Merchant (1987), Loebbecke, Eining, and Willingham (1989), Bell, Szykowny, and Willingham (1991)]. However, none of these studies explicitly examines board of director composition. Separately, the corporate governance literature includes empirical studies that examine the effectiveness of the board of directors as a monitor of management in acute agency settings; however, none of these studies examines the agency problem of management fraud. While the primary purpose of this study is to test economic theory about the relation of board of director composition and management fraud, this study will also contribute to the management fraud and corporate governance literatures. Even though the purpose of this study is not to develop a predictive model of management fraud, this study contributes to the development of future management fraud predictive models by providing evidence of a relation between board of director composition and management fraud. It also contributes to the corporate governance literature by studying an acute agency problem not previously examined. The remainder of this chapter is organized as follows. Section 1.1 summarizes the underlying theory used to motivate the study of several hypotheses about the relation of board of director composition and the occurrence of management fraud. Section 1.2 discusses the motivation for empirically examining board of director composition in settings of management fraud. Section 1.3 briefly overviews the sample selection and 3 research design that is used to examine the hypotheses. Section 1.4 summarizes the organization of the remainder of the dissertation. 1.1 Overview of Hypotheses Examined Fama and Jensen (1983a) conceptualize that the board of directors in the modem- day corporation is created by stockholders who delegate responsibilities to the board because it is too costly for each stockholder to individually monitor management. This delegated responsibility for monitoring management makes the board of directors the ultimate internal control mechanism within the firm. Economic theory suggests that the board of directors is an important part of the governance structure of large business corporations and that board of director composition, such as degree of outside director representation, quality, and ownership are relevant factors for board effectiveness. This theory is used in this study to motivate several hypotheses about the relation of board of director composition and the occurrence of management fraud. The first hypothesis predicts that the board of directors is composed of fewer "outside” members for fraud firms than for no-fraud firms. The second and third hypotheses predict that outside members of the board of directors of fraud firms are of lower ”quality” and hold smaller ownership stakes than outside directors of no-fraud firms, respectively. The fourth hypothesis predicts that managers (i.e. , insiders) who serve on the board of directors have lower ownership stakes in fraud firms than managers of no-fraud firms. Critics of board of director governance often argue that the board of directors is not an effective monitor of management. They believe that boards of directors are ineffective because management can generally override outside director monitoring by 4 dominating the board of directors through management’s influence on the selection of outside directors, control of the agenda of board of director meetings, and delivery of internal information to outside members [Mace (1986), Patton and Baker (1987)]. This study develops and examines three additional hypotheses about board of director composition that may influence the extent of power management can use to override monitoring by outside directors. One hypothesis predicts that the chairperson of the board of directors holds a managerial position, such as chief executive officer (CEO) or president, more often for fraud firms than for no-fraud firms. A second hypothesis predicts the CEO’s tenure on the board of directors for fraud firms is longer than the tenure of CEO’s for no-fraud firms. A third hypothesis predicts that the average outside director tenure on the board of directors is shorter for fraud firms than for no-fraud firms. Agency theory suggests that one of the mechanisms that a board of directors may establish to minimize the occurrence of management fraud is the audit committee. Audit committees are designed to reduce information asymmetries between management and the board of directors by serving as a conduit for information flow to the board [Pincus, Rusbarsky, and Wong ( 1989)]. This study examines a final hypothesis predicting that the board of directors has an active audit committee less often for fraud firms than for no-fraud firms. 1.2 Motivation The primary motivation of this study is to provide empirical evidence of whether board of director composition and management fraud are related in the manner predicted by economic theory. While this is the first study to empirically test this economic theory 5 about board of director composition in a setting of management fraud, this is not the first acknowledgement of a possible link between board of director composition and management fraud. The board of directors as a corporate governance mechanism for the prevention of management fraud is often discussed by the financial press, regulators, and standard setters. Because of substantial estimated economic loss by investors and creditors, the issue of management fraud receives significant attention, often front page headlines, in the financial press.‘ In these reports, there is a documented perception of a relation between board of director composition and the occurrence of management fraud. For example, We: (April 30, 1993) reported that in the wake of material fraudulent financial reporting, the Leslie Fay Company announced its board of directors elected two additional outside members "to give its board a more independent character. " And, WM (August 30, 1993) reported that in response to Clayton Homes Inc. ’s alleged failure to internally investigate a possible management fraud, two of the firrn’s outside board members resigned. The significance of management fraud in today’s business community has received significant attention by regulators and standards-setters who often discuss the importance of the board of directors as a corporate governance mechanism that assists in the prevention of management fraud. The National Commission on Fraudulent Financial Reporting (commonly referred to as The Treadway Commission) was created in the mid- ‘ Examples of financial reports of management fraud by 111W include allegations of management fraud at Clayton Homes Inc. (August 17, 1993), Leslie Fay Company (February 23, 1993), Comptronix Corporation (December 14, 1992), Phar-Mor Corporation (August 4, 1992) and Cascade International (November 21, 1991). 6 19808 with the objective of identifying causal factors that can lead to fraudulent financial reporting. In 1987, the Treadway Commission issued numerous recommendations, some of which suggest changes in the structure of boards of directors.2 More recently, the AICPA’s Public Oversight Board stated in its 1993 W that “the responsibilities of corporate boards and their audit committees for the integrity of management and financial reports should be pinpointed and reinforced and the appropriate authorities should adopt measures to assure that it is” (p. 50). Congress is considering H.R. 574, ”Financial Fraud Detection and Disclosure Act, " that would place certain responsibilities on boards of directors to inform the Securities and Exchange Commission (SEC) when notified by auditors that adequate remedial actions have not been taken against management in cases involving illegal acts. Recently, the Federal Deposit Insurance Corporation (FDIC) implemented new requirements for insured depository institutions to establish audit committees made up of independent directors who, for certain large institutions, must include individuals with banking or financial expertise and cannot include "large customers” of the institutions. Because the public looks to the independent auditor to detect management fraud, auditing professional standards-setters also have a vested interest in obtaining knowledge about the empirical relation between board of director composition and the occurrence of management fraud. Palmrose (1987) notes that management fraud accounts for about half of the litigation cases against auditors. Auditing professional standards highlight the 2 For example, the Treadway Commission (1987 , p. 40) stated that audit committees composed of independent directors would help reduce the occurrence of fraudulent financial reporting. 7 importance of the board of directors in the financial reporting process by requiring the auditor to ”obtain sufficient knowledge of the control environment to understand management’s and the board of director’s attitude, awareness, and actions concerning the control environment” [par. 20 of AICPA SAS No. 55, “Consideration of the Internal Control Structure in a Financial Statement Audit']. Interestingly, however, while auditing professional standards include ”red flag” indicators of the possibility of management fraud in SAS No. 53, "The Auditor’s Responsibility to Detect and Report Errors and Irregularities, " those indicators do not address potential board of director characteristics that may be uniquely associated with the occurrence of management fraud. Future allegations of management fraud are likely to continue. In one study, eighty-seven percent of managers surveyed were willing to commit fraud in one or more cases presented to them. Over half were willing to overstate assets, forty-eight percent were willing to establish insufficient return reserves for defective products, and thirty- eight percent would pad a government contract WW, March 1, 1990, p. 1]. In another study, seventy-six percent of surveyed firms report that they experienced fraud within the last year, and over two-thirds of the firms believe that fraud will become more of a problem in the future [KPMG Peat Marwick (1993)]. The issue of management fraud and the likelihood of its continued existence is of significance to numerous affected parties. As a result, this study’s empirical examination of the relation between board of director composition and the occurrence of management fraud may provide important insights for board of director governance policies. 8 1.3 Overview of Sample Selection and Research Design The firms examined in this study consist of 75 fraud firms that are matched with 75 no-fraud firms (the control firms). The fraud firms are those that have allegedly experienced management fraud as reported in an Accounting and Auditing Enforcement Release by the Securities and Exchange Commission (SEC) from 1982 through 1991 or as reported under the caption of ”Crime-White Collar Crime” in mm Index from 1980 through 1991.3 The fraud firms are matched with control firms based on industry, firm size, national exchange where common stock is traded, and time period because board of director composition may vary systematically with these variables [Baysinger and Butler (1985), Rosenstein and Wyatt (1990)], and they may be associated with management fraud. Additionally, other firm-specific characteristics are included in the regression model (described later in this section) to control for other endogenous factors that may be associated with both board of director composition and the likelihood of management fraud. This study’s definition of management fraud is limited to two types. The first type of management fraud includes fraudulent financial reporting whereby management intentionally issues materially misleading financial statement information to outside users. The second type of management fraud includes misappropriation of assets by top management. For purposes of this study top management includes the chairperson, vice chairperson, chief executive officer, president, chief financial officer, and controller. This study does not include cases of fraud by employees not considered as top 3 The SEC began issuing Accounting and Auditing Enforcement Releases in 1982. 9 management because those employees are generally not subject to direct monitoring by the board of directors. The statistical methodology underlying the empirical test is a logit regression in which the dependent variable (FRAUD) is dichotomous; a fraud is known to exist or is not known to exist. Three cross-sectional logit models are examined. The first logit model examines the hypotheses using a definition of outside directors that is consistent with the national stock exchanges. The second logit model examines the hypotheses using a more restrictive definition of an outside director that is consistent with previous corporate governance research. The third model includes a piecewise logit regression model that explores whether levels of firm ownership held by outside and management board of directors are linearly related to the occurrence of management fraud. 1.4 Organization of the Dissertation The remainder of the dissertation is organized into five chapters. Chapter Two develops the underlying economic theory and summarizes previous empirical research to motivate eight hypotheses about board of director composition and the occurrence of management fraud. Chapter Three describes the sample selection process by explaining how fraud firms are identified and matched with no-fraud firms. Chapter Four details the research design of the study and Chapter Five contains the empirical results of the study. Chapter Six includes a summary of the study and describes inherent limitations associated with the study’s research design. CHAPTER 2 - THEORY AND HYPOTHFSES DEVELOPMENT This chapter develops the underlying theory for eight hypotheses about the relation of board of director composition and the occurrence of management fraud. Such theory suggests that an important function of the board of directors is to monitor management. This chapter builds on that theory and related empirical research to highlight characteristics of the board of directors that may impact the board’s effectiveness in monitoring management for the prevention of management fraud. Section 2.1 describes the underlying economic theory that motivates the board of directors as an important monitor of management. Section 2.2 highlights the management fraud literature and indicates how this study contributes to that body of research by examining characteristics of the board of directors not previously empirically examined. Section 2.3 highlights the corporate governance literature and emphasizes how this study contributes to that research by examining an acute agency problem - management fraud - not previously explored. Section 2.4 builds on the underlying economic theory and previous empirical research to motivate eight hypotheses about specific board of director characteristics examined in this study. Section 2.5 emphasizes the importance of controlling for board size when examining these characteristics. Section 2.6 summarizes this chapter. 2.1 Underlying Theory: The Monitoring Role of the Board of Directors The purpose of this section is to describe the underlying economic theory of the fum that suggests the board of directors has an important responsibility to monitor management for the prevention of management fraud. As described in this section, this theory suggests that stockholders engage a board of directors to minimize agency 10 11 problems that arise out of the separation of decision control and residual risk-bearing. This study uses the setting of management fraud as an agency problem example to test the theory about the monitoring role of the board of directors. This economic theory is based on the view that firms are legal fictions that serve as a nexus for a set of contractual relationships among self-interested individuals whereby ownership and control are separate [Alchian and Demsetz (1972), Jensen and Meckling (1976), and Fama (1980)]. Such separation arises when one or more individuals (the principal(s)) engage another individual (the agent) to perform some service on the principal’s behalf. Contracts between the principal and agent are designed to limit divergences from the principal’s interests; however, because such contracts are not costlessly written and enforced there will be some divergences between the agent’s decisions and the principal’s interests. Economic theory suggests that there are both external and internal corporate governance mechanisms designed to minimize divergences that arise from the separation of ownership and decision control of the firm. External corporate governance mechanisms include the alienability of shares, limited liability, product and capital market competition, the market for corporate control, the managerial labor market, and corporation law [Williamson (1984)]. There are also potentially important internal mechanisms, such as competition among firm managers [Fama (1980), Fama and Jensen (1983a)], monitoring of holders of large share blocks [Shleifer and Vishny (1986)], and the focus of this study: the board of directors. The board of directors as a corporate governance mechanism receives its authority from stockholders of corporations who often effectively delegate important 12 responsibilities to boards of directors. Portfolio theory suggests that this delegation occurs because stockholders generally diversify their risks by owning securities in numerous firms. Such diversification creates a free-rider problem where no individual stockholder has a large enough incentive to devote resources to ensure that management is acting in the stockholders’ interests. This lack of sufficient incentive occurs because it is costly for all stockholders to be involved in decision control [Grossman and Hart (1980), Fama and Jensen (1983a)]. While stockholders, as residual claimants, generally retain approval rights of board membership, control over most other decision functions are separated from residual risk-bearing in corporations. The stockholder’s delegation of internal corporate governance to the board of directors makes the board the common apex of decision control within both large and small corporate organizations [Fama and Jensen (1983a)]. The board of directors ratifies and monitors important decisions, and chooses, dismisses, and rewards important decision agents. The board of directors makes collusion between top-level management more difficult by decomposing decision management performed by managers and decision control performed by the board of directors. The board of directors often delegates most decision management functions and many decision control functions to internal agents, but it retains ultimate control over internal agents thereby making it the top-level court of appeals of the internal agent market [Fama and Jensen (1983a)]. The board of directors ensures the establishment of an appropriate internal control system within the firm and monitors top management’s compliance with such system. As the ultimate internal control mechanism within the firm, the board of directors seeks to minimize the 13 expropriation of stockholder wealth by management. This study focuses on one example of such expropriation: management fraud. Management fiaud is one example of the agency problem where managers act self-interestedly in an attempt to expropriate stockholder wealth. As noted in Chapter 1, this study includes two types of management fraud where top management expropriates stockholder wealth by 1) misappropriating assets of the firm or 2) fraudulently reporting financial information by issuing materially misleading financial statements to outside users (e.g. , current and future investors). Because most of the day-to—day actions of boards of directors are unobservable, management fraud provides a unique setting whereby the characteristics of the board of directors that may affect the board’s ability to effectively monitor management can be examined ex post. Management fraud often occurs when internal controls are weak [Loebbecke, Eining, and Willingham (1989)]. Firms in which a fraud by top management has occurred represent situations in which the board of directors may have failed to establish a system of internal controls or inadequately monitored management’s compliance with such system of controls. This study examines whether there is a relation between occurrences of management fraud and board of director composition. This study contributes to existing research because little is known about differences in boards of directors of fraud firms as compared to other firms. Neither previous research on management fraud nor corporate governance research examines whether there are unique board of director characteristics for firms experiencing management fraud. Sections 2.2 and 2.3 provide an overview of previous management fraud and corporate governance research. Specific 14 empirical findings from these earlier studies are discussed more extensively as part of the development of hypotheses in Section 2.4. 2.2 The Management Fraud Literature Prior to the early 1980’s, there had been little empirical management fraud research and most of the speculations noted lacked empirical support [Albrecht, Romney, Cherrington, Payne, and Roe (1982)]. The earliest comprehensive study about management fraud dates back to 1978 when Peat Marwick commissioned experts in many relevant disciplines to participate in a multidisciplinary symposium about management fraud [Elliott and Willingham ( 1980)]. Most of the management fraud research conducted during the 19803 is descriptive. Based on the analysis of fraud cases as well as the review of research in other disciplines such as organizational behavior, psychology, and criminology, researchers highlight characteristics of firms experiencing management fraud that include both financial ratios and non-financial characteristics [Elliott and Willingham (1980), Albrecht, Romney, Cherrington, Payne and Roe (1982), Merchant (1987), National Commission on Fraudulent Financial Reporting (1987), Loebbecke and Willingham (1988), Loebbecke, Eining, and Willingham (1989)]. These researchers suggest that the identified factors are possible predictors of management fraud and commonly refer to them as "red flag“ indicators of fraud [Sorenson and Sorenson (1978), Romney, Albrecht, and Cherrington (1980), Loebbecke, Eining, and Willingham (1989), AICPA’s SAS No. 53 (1992)]. Interestingly, this previous management fiaud research does not empirically examine whether the ”red flag" characteristics identified are unique to firms experiencing 15 management fraud. The earlier studies only include firms where management fraud was alleged to have occurred and excludes firms where fraud was not present. A recent study by Bell, Szykowny, and Willingham (1991) attempts to validate these ”red flag” characteristics identified in the previous management fraud research by empirically examining whether the ”red flag” characteristics of fraud firms differ from no-fraud firms. Their examination is based on surveys of audit partners who previously served on a fraud or no-fraud firm engagement. The particular ”red flag" characteristics they examine are based on factors summarized in the AICPA’s SAS No. 53, ”The Auditor’s Responsibility to Detect and Report Errors and Irregularities, " and in Loebbecke and Willingham (1988). Bell et al. (1991) identify those factors that are present on an univariate basis significantly more often for fraud firms compared to no- fraud firms. Figure 1 summarizes the twenty-two factors they identify as being significantly different between fraud and no-fraud firms. Thirteen factors are from SAS No. 53 and 9 factors are from Loebbecke and Willingham (1988). While these factors are significantly different between fraud and no-fraud firms on an individual basis, Bell et al. (1991) note that some factors that are significant on an stand-alone basis may be highly correlated and not incrementally significant when combined with other factors in a predictive model. They use their survey results to build a decision aid predictive model for assessing the likelihood of management fraud and find that not all of the 22 factors are significant in the combined model. It is important to note that the "red flag” indicators in Figure 1 do not address board of director composition. However, several of the significant factors imply that the l6 22 Management Fraud Predictive Factors‘ 13 Factors From SAS No. 53: “ Weak internal control environment Management decisions dominated by single person or group Management attitude unduly aggressive Management places undue emphasis on earning projections Management's reputation in business community is poor Inadequate profitability relative to industry Organization is decentralized without adequate monitoring Doubt about the entity’s ability to continue as a going concern Many contentious or difficult accounting issues Significant difficult-to—sudit transactions Management is overly evasive when responding to audit inquiries Management has engaged in frequent disputes with auditors Accounting personnel exhibit inexperience or laxity in performing duties (Misststements in prior periods) 9 Factors From Loebbecke and Willingham (1988): Company is in a period of rapid growth Company has inexperienced management A conflict of interest exists within the company Company is confronted with adverse legal circumstances Auditor’s experience with management indicates degree of dishonesty Client personnel exhibits strong personality anomalies Management places undue pressure on auditors Management has engaged in opinion shopping L..._.._.__.__.__l.__L.__=_—__ Management displays significant disrespect for regulatory bodies ‘ Source: Bell, Szykowny,.and Willingham (1991) Figure 1 22 Red Flag Indicators That Differ Significantly Across Fraud and No-Fraud Firms 17 board of directors of fraud firms may be ineffective in monitoring management. Examples include: Company has a weak internal control environment. Management decisions are dominated by single individual or group. Management exhibits strong personal anomalies. Management is unduly aggressive. Management’s reputation in business community is poor. Company has an inexperienced management. Accounting personnel exhibit inexperience or laxity in performing duties. Other management fraud research examines implications of using ”red flag“ checklists as decision aids in the audit risk evaluation process. Results from these studies are mixed. Pincus (1989) finds that auditors who do not use ”red flag” checklists outperform those who do in an experimental setting. Hackenbrack (1993) finds that auditors have different opinions about the amount of fraud risk indicated by specific ”red flag” indicators and concludes that one reason for this disagreement is that auditors with different client experience (e.g. , large versus small clients) have systematically different perceptions of the importance of a selected ”red flag” factor. The collective review of this management fraud research suggests that an important component of the firm - the board of directors - has not been explicitly examined. These studies suggest the importance of the board of directors by consistently noting the significance of ”weak internal control environments” for many of the firms experiencing fraud [Albrecht and Romney ( 1986), Merchant ( 1987), Loebbecke, Eining, and Willingham (1989), Bell, Szykowny, and Willingham (1991)]. For example, Loebbecke et al. (1989) note that ”Our findings support the importance of the control environment. . . . Where controls are weak, a significant condition exists that would allow either management fraud, defalcations, or an error to occur.” (p. 25). Pincus’s (1989) l8 believes that auditors who do not use ”red flag" checklists outperform those who do in her experimental setting because non-users are more likely to consider additional items, such as the competence and strength of the board of directors and/or the audit committee, which were not included on the checklists provided to auditors in her study. The purpose of this study is to test economic theory about the monitoring role of the board of directors in a setting of management fraud. By doing so, this study also provides empirical evidence about characteristics of a potential red-flag indicator - the board of directors - that is excluded from existing predictive models. While the purpose of this study is not to develop a predictive model of management fraud, this study contributes to the development of future management fraud predictive models by providing evidence of a relation between board of director composition and management fraud. Such evidence suggests the importance of considering board of director composition when developing future predictive models. 2.3 The Corporate Governance Literature This section briefly highlights the focus of previous corporate governance empirical research. Details about specific empirical findings are not presented in this section. Instead, findings relevant to this study are included where appropriate in the development of hypotheses in Section 2.4. Because most of the day-to-day actions of boards of directors are unobservable, attempts by empirical researchers to isolate the monitoring effects of boards of directors either consider some aspect of firm performance or concentrate on the boards’ observable actions for acute agency problems. None of these studies examines the issue of management fraud but they suggest that there may be a link between board of director l9 composition and monitoring management for the purpose of minimizing agency problems such as management fraud. Studies that examine the relation of board of director composition and firm performance find a weak positive relation at best [MacAvoy, Cantor, Dana, and Peck (1983), Baysinger and Butler (1985)] while others find no relation [Hermalin and Weisbach (1991)]; however, these studies are criticized for the lack of control of the multitude of endogenous and exogenous factors that influence firm performance [Hermalin and Weisbach (1991)]. Other studies find that boards of directors, as well as other internal monitoring mechanisms, monitor management by forcing top management turnover when firms perform poorly, particularly for firms with high proportions of outside directors [Coughlan and Schmidt (1985), Warner, Watts, and Wruck (1988), Weisbach (1988)]. Poor stock performance leads to changes in board of director composition with inside directors being replaced with outside directors [Hermalin and Weisbach (1988)]. Additional studies support the monitoring role performed by outside directors by examining board of director composition for firms experiencing acute governance problems. In general, these studies suggest that the board of directors, particularly outside directors, serve as effective monitors of management in situations involving corporate takeovers [Brickley and James ( 1987), Byrd and Hickman (1992), Kini, Kracaw, and Mian (1993)], management buyouts [Lee, Rosenstein, Rangan, and Davidson (1992)], greenmail payments [Klein and Rosenfeld (1988), Kosnik (1987), (1990)], and firms with golden parachutes (Cochran, Wood, and Jones (1985), Singh and Harianto (1989)]. 20 This study contributes to the existing corporate governance literature by empirically examining whether there is a relation between board of director composition and management fraud. This study expands the corporate governance literature by examining the board of directors in an acute agency setting not previously explored. 2.4 Development of Hypotheses Motivation of eight hypotheses about board of director composition and management fraud is provided in Sections 2.4.1 through 2.4.5 that follow. Figure 2 summarizes these eight hypotheses. Predicted Relation With Occurrence of Hypotheses Management Fraud Representation of Outside Directors | H1: % of Outside Members on Board Inverse I Quality of Outside Directors I H2: Quality of Outside Members on Board Inverse I | Ownership Stakes In Firm: fl H3: Held By Outside Directors on Board Inverse " H4: Held By Management on Board Inverse fl Management Power: H5: Chairperson is also CEO or President Direct H6: CEO’s Tenure on Board Direct H7: Average Outside Director Tenure on Board Inverse Audit Committees: H8: 5 of Firms with Active Audit Committees Inverse — r I Figure 2 Summary of Hypotheses About Board of Director Characteristics and Management Fraud 21 2.4.1 Representation of Outsiders on the Board. Economic theory of the firm suggests that the composition of individuals who serve on the board of directors is an important factor in creating a board that is an effective monitor of management actions. For reasons discussed in the paragraphs that follow, boards of directors are generally composed of both firm management and outsiders (non-employees). Understanding the representation of managers and outsiders on the board of directors is important as noted by Baysinger and Butler (1985) who state that “discussion of the role of the board in a theory of corporate governance without discussing board composition is as inappropriate as discussing the theory of the firm and ignoring the internal structure of the organization" (p. 121). Because the board of directors must be able to use information from the internal monitoring system, the board of directors of a corporation often includes several of the organization’s top managers [Fama and Jensen (1983a)]. Management’s presence on the board of directors can improve the amount and quality of information from the internal monitoring system. Because inside board of director members (those who are the firm’s top managers) are generally more influential than outside members, the board of directors is not an effective device for decision control unless it limits the decision discretion of individual top managers [Fama and Jensen (1983a)]. Managers employ huge informational advantages due to their full-time status and insider knowledge. As a result, the board of directors can easily become an instrument of management, sacrificing the interests of stockholders [Williamson (1984)]. Domination by top management on the board of directors may lead to collusion and transfer of stockholder wealth [Fama (1980)]. 22 The viability of the board of directors as a market-induced mechanism for low- cost internal transfer of control should be enhanced by the inclusion of outside (non- management) directors who are disciplined for their services by the market, which prices outside directors according to their performance as referees [Fama (1980)]. The purpose of an outside board of director is to act as an arbiter in disagreements among internal managers and carry out tasks that involve serious agency problems between internal managers and residual claimants [Fama and Jensen (1983a)]. Baysinger and Butler (1985) note that ”corporate reform proposals predict (implicitly) that corporations with boards having a higher proportion of monitoring [independent] directors will better serve shareholders’ objectives than corporations with boards having a smaller proportion of such directors" (p. 114). Existing empirical research provides evidence about the importance of including outside directors on the board for purposes of monitoring management. For example, Weisbach (1988) finds that the positive relation between poor firm stock performance and subsequent CEO turnover is strongest for firms with boards of directors with high proportions of outside directors. Lee, Rosenstein, Rangan, and Davidson (1992) find that shareholder wealth increases in management buyouts when boards of directors are dominated by outside directors. Kosnik (1987, 1990) finds that firms resisting greenmail payments have more outside directors relative to boards of directors of firms not resisting greenmail. Brickley and James (1987) find managerial consumption of perquisites represented by excessive expenditures in salaries in the banking industry is negatively related to the percentage of outside members on the board of directors. Mayers, Shivdasani, and Smith (1994) find that mutual life insurance firms relative to stock life 23 insurance firms have boards that employ larger fractions of outside directors. They attribute this difference to the fact that ownership rights are inalienable in mutual life insurance companies because ownership rights are held by policyholders and such rights are not separable from policies. Without such separation, a hostile takeover is impossible. Thus, mutual boards of directors have more outside directors because they serve as a major substitute monitoring mechanism for external capital markets. While there is numerous recent empirical research on corporate governance, Weisbach (1988) notes that ”understanding the role of the outside directors remains an important and unresolved question” (p. 432). Hermalin and Weisbach (1991) note that "the extent to which boards oversee management and to which this monitoring depends on the composition of the board are important and unresolved empirical questions” (p. 101). Trends in practice suggest there is perceived value in the role played by outside directors. The percentage of outsiders present on boards of directors is increasing with outside directors comprising a board majority of 94% of manufacturing firms polled in 1992 compared with 86% in 1989 and 71% in 1972 W, August 19, 1993, (p. 1)]. Evidence suggests that stockholders value outside directors as exhibited by the positive abnormal stock return Rosenstein and Wyatt (1990) find when management-chosen outside directors are added to the board of directors. The requirements of the national stock exchanges also suggest that inclusion of outside directors on the board of directors is important. The national stock exchanges specify certain audit committee composition requirements, which in turn affect board of director composition. In June 1978, the New York Stock Exchange (NYSE) established 24 a requirement that firms must have audit committees composed entirely of independent directors. An independent director is one who is not a part of current management. The other exchanges are less strict. The American Stock Exchange (AMEX) recommends but does not require audit committees composed entirely of independent directors. In 1987, the National Association of Securities Dealers (NASDAQ) established a requirement that audit committees be composed of at least a majority of independent directors. The previously discussed underlying theory, prior empirical research, and anecdotal observations suggest that the composition of the board of directors may be related to the prevention of management fraud. Particularly, the above suggests that having a higher percentage of outside directors increases the board of director’s effectiveness as a monitor of management for the prevention of management fraud. The following hypothesis is examined: H1: The proportion of outside members on the board of directors is lower for firms experiencing management fraud compared to control firms. 2.4.2 Quality of Outside Board Members. The mere presence of outsiders on the board of directors does not ensure that the board is an effective monitor of management. Boards of directors with similar percentages of outside directors may vary in their effectiveness as a monitor of management depending on whether outside directors have incentives to maintain reputations as quality directors. Fama (1980) argues that the external market for outside directorships provides an incentive for outside directors to maintain reputations as decision experts. The presence of this external market encourages outside board members to use their performance as 25 an outside director to signal that (1) they are decision experts, (2) they understand the importance of diffuse and separate decision control, and (3) they can work within such decision control systems [Fama and Jensen (1983a)]. This market rewards and punishes outsiders based on their performance as a director. The above discussion suggests that the board of director’s ability to effectively monitor management may be a function of the quality of outside directors who serve on the board. Given that the occurrence of management fraud is an example of where the board of directors has ineffectively monitored management, there is an expectation that outside directors for fraud firms are of lower quality as compared to outside directors of no-fraud firms. The following hypothesis is examined: H2: The quality of outside members on the board of directors is lower for firms experiencing management fraud compared to control firms. The number of additional directorships held by outside directors is used as a proxy for outside director quality, consistent with Byrd and Hickman (1992) and Shivdasani (1993). This proxy for outside director quality is based on Fama’s (1980) view that the external market for outside directors rewards and punishes outside directors for their director performance. According to this view, the market rewards quality directors with additional directorships and punishes directors for poor performance by restricting their opportunities to serve on boards of directors of other firms.1 ‘ Kaplan and Reishus (1990) find that top managers in poor performing firms (e. g. , dividend reducing firms) have fewer opportunities to serve as outside directors for other firms. Gilson (1990) finds that directors who leave boards of distressed firms hold approximately one-third fewer directorships three years after their departures. 26 2.4.3 Board Members’ Ownership Stakes. Agency dreary suggests that a high stake in a company’s outstanding equity should provide individual directors with a strong incentive to promote firm activities that increase a firm’s value because this increases the value in the director’s own investment [Jensen and Meckling ( 1976)]. Empirical evidence by Feroz, Park, and Pastena (1991) suggests that the firm’s owners have an incentive to prevent management fraud to protect their investment in the firm. They find that allegations by the SEC of financial reporting violations (due to either error or management fraud) are associated with average two-day abnormal returns of ~13 % . Thus, the agency theory argument and empirical evidence suggests that as the extent of ownership in the firm by both outside directors as well as managers who serve on the board of directors increases, the occurrence of management fraud should decrease. This relationship is further developed in Sections 2.4.3(i) and 2.4.3(ii). WWW While the agency definition of board of director monitoring presumes that outside directors serve to protect the interests of stockholders, such protection should increase as outside members’ degree of ownership in the firm increases. Monitoring the performance of top management requires time and effort. Without a personal financial interest in the firm or control over a large block of votes, an outside director will be more reluctant to second-guess poor management decisions [Morck, Shleifer, and Vishny (1988)]. Outside directors with high equity ownership interests in the firm are less likely to engage in decisions that have negative consequences for stockholder wealth [Walkling and Long (1984)]. Outside directors whose ownership stakes in the firm are low appear more like an employee rather than an investor because they receive only cash compensation as a director. A 27 director with a sizeable stake in a firm is more likely to question and challenge management’s proposals [Mace (1986), Patton and Baker (1987)]. And, the presence of an outside director with large ownership in the firm who asks discerning questions frequently encourages other outside directors to get involved [Mace (1986)]. Jensen (1993) argues that encouraging outside board members to hold substantial equity interests would provide better incentives for monitoring top management. Recent empirical studies support the importance of firm ownership by outside directors. Kosnik (1990) finds outside directors’ resistance of greenmail payments is most likely if outside directors own a large amount of equity relative to their cash compensation. Shivdasani (1993) finds equity ownership by outside directors of hostile targets is significantly lower than that by outside directors of non-targets. These empirical studies suggest that ownership in the firm provides incentives for outside directors to monitor management closely to prevent management fraud. The following hypothesis is examined: H3: The extent of firm ownership by outside members on the board of directors is lower for firms experiencing management fraud compared to control firms. WWW As stressed by Berle and Mans (1932), when managers hold little equity in the firm and shareholders are too dispersed to enforce value maximization, corporate assets may be deployed to benefit managers rather than shareholders. Jensen and Meckling (1976) theorize that stock ownership by management can reduce the underlying agency problem: the more stock management owns, the stronger their motivation to work to raise the value of the firm’s stock. As their stakes rise, managers pay a larger share of agency costs and are less likely to 28 squander corporate wealth. This suggests a negative relation between the extent of management’s ownership stake in the firm and the occurrence of management fraud. Empirical research suggests that management ownership in the firm can serve as a substitute for other controls over management actions. When managers’ ownership stakes in the firm represent the bulk of their personal wealth, it affects their incentives [Jensen and Murphy (1985), Jensen and Warner (1988)]. Jensen and Murphy (1990) show that the vast majority of direct incentives of top managers comes through stock ownership. Because of the significance of firm ownership to management’s personal wealth, top management’s ownership in the firm may lead to less demand for alternative anti-agency measures such as a strong board of outside directors in firms where management owns a large fraction of stock. Weisbach (1988) finds the fraction of outside directors is negatively correlated with shareholdings of top management, which is consistent with the view that monitoring by outside directors and the direct incentives created by management’s stock ownership are substitute methods of control. This study focuses on two types of management fraud: misappropriation of assets and fraudulent financial reporting. It is possible that the negative relation between the extent of management ownership and the occurrence of management fraud may depend on the type of management fraud committed. The negative relation between management ownership in the firm and management fraud is expected to hold for occurrences of misappropriation of assets. If management owns 01% of outstanding shares, their net gain from misappropriating $1 dollar of assets is ($1-a($1)). As management ownership of outstanding shares increases, the net gain from misappropriating assets decreases. 29 The negative relation between management ownership in the firm and management fraud is expected to hold for fraudulent financial reporting occurrences when management has a long-term investment horizon and believes that the probability of detection is high. In that case, while fraudulently reporting financial information may artificially increase firm value, management’s assessment that the benefits of this artificial increase are temporary and do not exceed the costs of a decline in value of their stock holdings (as well as other penalties) when the fraudulent reporting is revealed. Thus, as stock ownership by management increases, the perceived net cost to management increases, consistent with the expected negative relation. The following hypothesis is examined: H4: The extent of ownership in the Turn held by managers who serve on the board of directors is lower for firms experiencing management fraud compared to control f'mns. Alternatively, it is possible that the negative relation may not hold when management has a short-term investment horizon and believes that the probability of detection is low. In that case, the benefits of fraudulently reporting financial information and artificially inflating firm value exceed the costs of a decline in value of their stock holdings (as well as other penalties) when the financial reporting is revealed. Thus, as stock ownership by management increases, the perceived net benefit to management increases, which is not consistent with the hypothesis. As a result, additional analysis will be performed for the subset of fraudulent financial reporting cases of management fraud examined in this study. As discussed in Chapter 3, the vast majority of management fraud occurrences examined in this study represent occurrences of fraudulent financial reporting and not misappropriation of assets. 30 The additional analysis of fraudulent financial reporting occurrences may offer insights about the relation between the extent of management ownership and the occurrence of fraudulent financial reporting. Due to the small number of occurrences of misappropriation of assets, a separate analysis of asset misappropriation will not be performed. 2.4.4 Management Power. The ability of outside board of director members to effectively monitor management may be impacted by management’s ability to exert power to override such monitoring. Jensen (1993) argues that board culture is an important component of board of director failure. The great emphasis on politeness and courtesy at the expense of truth and frankness in boardrooms is both a system and cause of failure in the control system. He argues that by rewarding consent and discouraging conflicts, chief executive officers (CEOs) have the power to control the board of directors. The CEO’s power to control the board of directors is often attributed to the belief that the CEO has by far the strongest voice in determining who is on the board of directors, even though boards have nominating committees [Mace (1986), Vancil ( 1987), Patton and Baker (1987)]. As a result, directors’ ties to management and the CEO are often stronger than the ties to stockholders because directors are captives of top management. Non-management directors refrain from overt criticism of management’s behavior to not jeopardize board of director seats [Patton and Baker ( 1987)]. Additionally, the CEO often determines the board of director’s agenda and information given to the board of directors. This limitation on information severely 31 hinders the ability of even highly trained board members to contribute effectively to the monitoring and evaluation of the CEO and other top management [Jensen (1993)]. The management fraud literature suggests that management power may affect management’s ability to act fraudulently. Loebbecke, et.al., (1989) find that in seventy- five percent of the fraud cases examined, operating and financial decisions are dominated by a single person. This suggests that the extent of power held by managers who serve on the board of directors may allow those managers to override monitoring by the board of directors for purposes of committing management fraud. This study examines three measures of board of director characteristics that may influence management’s ability to override monitoring by outside directors. These three measures are in addition to any power management derives from significant stock ownership (ownership is addressed in connection with Hypothesis 4). The first measure considers whether the chairpersons of fraud firms are more likely to hold a managerial position in the firm relative to chairpersons of no-fraud firms. The second measure considers whether there are differences in the CEO’s tenure on the board of directors between fraud and no-fraud firms. The third measure considers whether there are differences in outside director tenures on boards of fraud firms relative to no-fraud firms. These three measures are further developed in Sections 2.4.3(i) through 2.4.3(iii) that follow. Wu, In many firms, it is common for the chairperson to also hold managerial positions in the firm. The function of the chairperson is to run board of director meetings and oversee the process of hiring, firing, evaluating, and compensating the CEO. Clearly, the CEO cannot perform this function 32 apart from his or her personal interests [Jensen (1993)]. Without the direction of an independent leader, it is much more difficult for the board of directors to perform its critical function. When the chairperson of the board of directors yields great power by holding a management position in the firm, the decision processes of the board of directors appear to be dominated by an individual. In a company where the chairperson is also the CEO or president, power is concentrated in one individual and possibilities for checking and balancing powers of the CEO or president are eliminated [Chaganti, Mahajan, and Sharrna (1985)]. Such situations signal the absence of separation of decision management and decision control [Fama and Jensen (1983a)]. For the board of directors to be effective, it is important to separate the CEO and chairperson/president positions [Jensen (1993)]. The following hypothesis is examined: H5: The chairperson of the board of directors holds managerial positions more often for firms experiencing management fraud compared to control firms. W The extent of the CEO’s tenure of service on the board of directors may indicate the extent of power held by that individual. An established CEO has relatively more power than a new CEO [Hermalin and Weisbach (1988)]. A CEO who has successfully maintained a position on the board of directors for long periods of time may use his/her seniority to override monitoring by outside members for purposes of committing management fraud. The following hypothesis is examined: H6: The CEO’s tenure on the board of directors is longer for firms experiencing management fraud compared to control firms. 33 In addition to examining whether CEO tenure differs between fraud and no-fraud firms as predicted by hypothesis 6, it is important to control for differences in CEO tenure across fraud and no-fraud firms because CEO tenure may effect who is selected to serve on the board as a director. Hermalin and Weisbach ( 1988) find that insiders are added to the board of directors toward the end of a CEO’s tenure to be groomed as potential successors and that insiders leave just before and after a CEO change. Wm The potential for CEO power may be further enhanced if the average tenure of outside directors is low. In such situations, the CEO may be capable of exerting power over more recently appointed, shorter tenured outside directors for purposes of overriding outside director monitoring. The outside director’s lack of seniority may affect his/her ability to scrutinize top management. Newer members on the board of directors may be more susceptible to group pressures to conform. Kosnik (1990) finds that outside directors are significantly more likely to resist greenmail payments as their average tenure on the board of directors increases. The following hypothesis is examined: H7: The average outside director’s tenure on the board of directors is shorter for f’u'ms experiencing management fraud compared to control firms. 2.4.5 Active Audit Committees. Often the board of directors delegates the responsibility for the oversight of financial reporting to an audit committee [The National Commission on Fraudulent Financial Reporting ( 1987), AICPA’s SAS No. 53, AICPA’s Public Oversight Board (1993)]. Audit committees can be viewed as monitoring mechanisms that are voluntarily employed in high agency cost situations to improve the quality of information flows between principal and agent. Audit committees provide a 34 direct line of communication between the board of directors and the auditor thereby reducing the information asymmetries between management and the board [Pincus, Rusbarsky, and Wong ( 1989)]. The audit committee enhances the board of director’s capacity to act as a management control by providing the board of directors with more detailed knowledge and complete understanding of financial statements and other financial information issued by the company. The existence of an audit committee may be perceived as indicating higher quality monitoring [Pincus et al. (1989)] Despite decades of encouragement, audit committees were rare until the late 1970s and are still not universal [Pincus et al. ( 1989)]. Audit committees were first suggested as vehicles of communication between external auditors and boards of directors in the aftermath of McKesson and Robbins fraud case in the 1930s. Despite the growth in the number of audit committees, Pincus et al. (1989) report that a 1988 followup study on the implementation of the National Commission on Fraudulent Financial Reporting recommendations notes that companies continue to not create audit committees. That study surveyed 8564 public companies, receiving 1014 replies, and found that while 82% of respondents (including NYSE companies) had an audit committee, only 53% of smaller companies had audit committees. These results must be interpreted with caution due to the possibility of non-response bias.2 The audit committee can play an important role in preventing and detecting management fraud [The National Commission on Fraudulent Financial Reporting ( 1987)]. 2 Recall that the audit committee requirements of the national stock exchanges are discussed in Section 2.4.1. 35 According to Sommer (1991), an audit committee may often be the first non-management personnel to identify a potential irregularity. Sommer (1991) notes that having an audit committee as part of the board governance structure and having an effective audit committee are different matters. Respondents to the KPMG Peat Marwick (1993) fraud survey believe that one of the factors contributing to the occurrence of management fraud is that firms have inadequate audit committees. The AICPA’s Public Oversight Board (1993) reports ”that in too many instances the audit committees do not perform their duties adequately and in many cases do not understand their responsibilities” (p. 50). Research on audit committee effectiveness is limited. Whether audit committees are actually discharging their responsibilities remains insufficiently understood [Kalbers and Fogarty (1993)]. While an audit committee is designed to improve the quality of information between the principal and agent, the effectiveness of the audit committee is eliminated if the audit committee is never allowed to meet. Proxy statements disclose the number of meetings per year held by the audit committee. For purposes of this study, an "active” audit committee represents an audit committee meeting at least once during the year. The following hypothesis is examined: H8: The extent to which the board of directors has an active audit committee 8 lower for firms experiencing management fraud compared to control firms. 36 2.5 Controlling for Board Size It is important to control for differences in the size of the board of directors when examining these eight hypotheses. In the case of the representation by outside directors, a given percentage of outside director representation on the board, say 25 % , translates to one outside director for a board size of four members and to three outsiders for a board size of twelve. While the two firms have the same percentage (25 %) of outsiders on the board of directors, the effectiveness of those outsiders may differ between the two firms. For the firm with twelve members, the three outsiders may be able to band together with stronger voice to influence board of director action more effectively than the sole outsider on the board of directors of the firm with only four board members. Board size may affect other board of director characteristics examined such as audit committee formation. Small-sized boards may not believe there is a need to create an audit committee separate of the board of directors. Board size is included in this study as a control variable rather than testing a hypothesis about differences in board size due to conflicting expectations about the effects of board size on management fraud. Some researchers believe that a smaller board of directors plays a controlling function whereas a larger board of directors may not be able to function effectively as a controlling body leaving management relatively free [Chaganti, Mahajan, and Sharma (1985)]. This view is consistent with Jensen (1993) who believes larger boards are easier for the CEO to control. Others believe larger boards may be valuable for the breadth of its “services". Chaganti et al. (1985) find that firms filing for Chapter 11 bankruptcy protection have smaller boards than no—failed firms suggesting that a larger board is more effective in preventing corporate failure. 37 2.6 Summary This chapter describes how the occurrence of management fraud provides a unique setting to test the theory of the board of directors as a monitor of management. Using management fraud as an example of the agency problem that arises from the separation of decision control and residual-risk hearing, this chapter builds upon agency theory and existing management fraud and corporate governance research to motivate the examination of whether there are differences in eight characteristics of boards of directors between fraud and no-fraud firms. Chapter 3 describes how fraud and no-fraud firms are selected for this study. CHAPTER 3 - SAMPLE SELECTION AND DESCRIPTION This chapter describes the sample used in this study to examine the eight ‘ hypotheses developed in Chapter 2. The sample consists of 150 publicly traded firms. Seventy-five of the 150 firms represent the "fraud firms" because each of these firms had an occurrence of management fraud publicly reported during the period 1980 - 1991. Each of the fraud firms was matched with a no-fraud firm thereby creating a choice- based sample of 75 fraud and 75 no-fraud firms. Section 3.1 explains how the fraud firms were identified. Section 3.2 describes how a no-fraud firm was matched with each fraud firm. Section 3.3 highlights univariate differences in board of director composition between fraud and no-fraud firms. Section 3.4 highlights differences in other firm characteristics between fraud and no-fraud firms that will be considered in this study because they may be associated with the likelihood of management fraud and board of director composition. Section 3.5 summarizes this chapter. 3.1 Fraud F‘u'm Selection Identifying a sample for this study began with the search of publicly traded firms experiencing management fraud. The sample was limited to public rather than privately- held firms because the data examined in this study includes information only available in proxy statements and financial statements filed with the SEC. Two sources were used to identify the fraud firms. The first source of fraud firms was Accounting and Auditing Enforcement Releases (AAERs) issued by the SEC. A firm reported in an AAER was included as a sample fiaud firm if the SEC accused top management of violating Rule 10(b)-5 of the 38 39 1934 Securities Exchange Act (the 1934 Act).2 Rule 10(b)—5 requires the intent to deceive, manipulate, or defraud [Commerce Clearing House SEC Accountant’s Handbook (1993)] . The second source of fraud firms was W W caption of “Crime-White Collar Crime. " Other possible captions, such as "Fraud,” ”Management Fraud, " 'Embezzlement, " are not provided in the mm. The only other related captions are "Crime” and “Crime-Organized Crime; " however, those two captions include articles about occurrences not related to management fraud, such as drug-trading, murder, and tax evasion. In most cases, the firms identified in the 2151 Index were also reported in AAERs. However, because the time lag between a management fraud occurrence and the subsequent reporting in an AAER can be lengthy, some of the management fraud occurrences noted in the W are yet to be reported in an AAER. Therefore, all fraud firms reported by the m that were not addressed in an AAER were included as a sample fraud firm. A fraud firm identified from these two sources is included in the sample if the appropriate proxy and financial statement data filed with the SEC in the fiscal year preceding the first occurrence of the management fraud is available.2 Such proxy and financial statement data were hand collected from the Q-Data SEC Files (the Q Files) that ‘ Feroz, Park, and Pastena (1991) note that the SEC only pursues cases where the probability of SEC success is high and where the allegations involve material violations. 2 For some firms, the proxy examined may have been filed with the SEC in the fiscal year the management fraud took place but before the fraud was discovered. For example, if the management fraud occurred in 1986 and 1987 , the 1986 proxy may be examined if the 1985 proxy is not available in the Q Files. 40 are on microfiche. Information about the specific financial reporting periods affected by the management fraud was obtained from the AAER or W51 Index. For fraud firms identified from AAERs, the related AAER notes the time period of the alleged fraud. For fiaud firms identified from the W. the related 3151 articles were reviewed to determine when the first report of the alleged fraud appeared. The AAERs and the mundex appear to be reasonable sources for identifying management fraud occurrences for two reasons. First, almost all of the applicable AAERs contain a disclosure that management personnel involved in the management fraud consented to the final judgment action imposed by the SEC.3 Second, for fraud firms identified by review of the W. management personnel involved has(ve) either resigned, been terminated, or indicted by a grand jury. Management’s consent, resignation, termination, or indictment disclosed by these two sources suggests a high level of seriousness of the management fraud allegation. Thus, the fraud occurrences included in this study appear to represent serious instances of management fraud. These two sources provided a sample of seventy-five fraud firms for examination in this study. As noted in Figure 3, sixty-seven of the seventy-five fraud firms came from the review of 1982-1991 AAERs, which include AAERs till-#348. The remaining eight fraud firms came from the review of the 1980-1991 M. AAERs and the We; issues after 1991 were not reviewed to allow at least two subsequent years 3 In a small number of AAERs, the SEC did not disclose whether or not management consented to the SEC’s final judgment. This lack of disclosure of management’s consent does not imply that management is challenging the SEC’s allegation of management fraud. Instead, it appears that the SEC inadvertently omitted the consent disclosure. Number of Accounting & Auditing Enforcement Releases (AAERs) 1982-1991 Less: AAERs not involving management fraud (e.g., unintentional misapplication of GAAP) or AAERs expanding other AAERs (e.g., duplicate AAERs for same firm) AAERs affecting firms with no available proxy or financial statement data AAERs affecting banks or insurance firms experiencing management fraud Add: Allegations of management fraud reported by the mm but not 8 reported in an AAER _.__M #h______,_ , ~__ ._ _ _____ _j [ Total number of fraud firms included in study I L Figure 3 Identification of 75 Fraud Firms to verify that the related matched no—fraud firms have not experienced management fraud (the process of matching no-fraud firms with each fraud firm is described in Section 3.2). Also, as summarized in Figure 3, 198 of the 348 AAERs were excluded because they do not involve management fraud as defined for this study (e. g. , they involve unintentional misapplications of GAAP) or they represent AAERs that expand other AAERs (e.g. , duplicate AAERs involving the same firm), 64 AAERs were excluded due to the lack of proxy and financial statement data availability, and three were excluded because no-fraud firms could not be identified using the matching criteria specified in Section 3.2.‘ Finally, sixteen AAERs were excluded because they involve banks and insurance institutions. Banks and insurance institutions were excluded to be consistent with Hermalin and Weisbach (1988) who note that bank boards have characteristics that differ ‘ The Q Files do not contain financial statements for all public companies. 42 from industrial firms and Mayers, Shivdasani, and Smith (1994) who suggest that insurance company boards may have unique board characteristics.’ Figure 4 shows that the majority of fraud firms represent occurrences of fraudulent financial reporting rather than misappropriation of assets.6 Sixty-seven of the seventy-five fraud firms experienced fraudulent financial reporting and eight firms experienced misappropriation of assets. Thus, 89.3 % of the management fraud cases in this study represent instances of fraudulent financial reporting. This is consistent with the findings of the National Commission on Fraudulent Financial Reporting (1987) that note 87% of the SEC enforcement actions in 1982-1986 dealt with fiaudulent financial reporting. As noted in Figure 4, almost all of the fraudulent financial reporting sample firms in this study came from the review of AAERs. Many of these fraudulent financial reporting occurrences were also reported in the W; however, only two additional fraudulent financial reporting occurrences were found in the review of the wee; that were not also covered in an AAER. Two of the misappropriation of asset occurrences were identified during the review of the AAERs. The large number of 5 Fama and Jensen (1983b) discuss unique characteristics of financial organizations. They note that there is a special form of diffuse control inherent in the redeemable claims of financial organizations. Specifically, claimholders (i.e. , bank depositors and insurance policyholders) can independently withdraw resources that deprive management control over assets. ‘ One might argue that the fraudulent financial reporting instances also include an implicit misappropriation of assets. For example, management may receive additional compensation if fraudulently reported financial information increases bonuses that are paid as a function of accounting-based earnings. 43 fraudulent financial reporting occurrences in the AAERs reflects either the more common occurrence of fraudulent financial reporting rather than misappropriation of assets in the population or the SEC’s enforcement bias towards fraudulent financial reporting occurrences. All other misappropriation of assets occurrences came from the review of the Minder. Source of Fraud Firms Type of Accounting & Auditing Enforcement MW * Management fiaud Releases (AAERs) “-348 “Crime - White Collar” Total 1982-1991 1980-1991 fiandulent 65 2 67 Financial Reporting Misappropriation of 2 6 8 8 75 ‘ Figure 4 Source of Fraud Firm by Type of Management Fraud 3.2 Matching Fraud Firms with No-Fraud Firms A comparison sample of seventy-five no-fraud firms was created by matching a no-fraud firm with each fraud firm based on the following requirements: 1. Steekjgehme, The common stocks of a fraud firm and its matched no-fraud firm trade on the same national stock exchange (NASDAQ, AMFX, NYSE). 2. Einujjze, All firms within the particular national stock exchange category per the annual COMPUSTAT tape that are in the same industry (see step 3) as the fraud firm were selected if those firms are similar in firm size. Firms are considered similar in firm size if the current market value of common equity is within 1; 30% of the current market value of common equity for the fraud firm in the year preceding the year of the management fraud.7 2 2 Kaplan and Reishus (1990) created a comparison sample using a cutoff of 150%. While this study’s use of i30% may appear as a large range, most of the fraud firms and related control firms are within 120% . Given that the mean market value of 44 3. Industry, All firms identified in steps 1 and 2 were reviewed to identify a no- fraud firm within the same four-digit SIC code as the fraud firm. The no-fraud firm selected was the one that had a current market value of common equity closest to the current market value of common equity of the fraud firm (or total assets if market data was not available). If no four-digit SIC code firm match was identified, the same procedure was performed to identify a firm with the same three-digit SIC code. If no three-digit match was identified, the same procedures were performed to identify a two-digit SIC code match. 4. Med, A no—fraud firm identified in steps 1 through 3 was included in the final sample if proxy and financial statement data was available for the time period used to collect data from the proxy and financial statements of the related fraud firm. The matching of no-fiaud firms will result in some misclassification if a firm classified as a no-fraud firm had an occurrence of management fraud that has yet to be detected. To minimize this likelihood, the W was reviewed from 1980 through 1994 to verify that there was no report of a management fraud for that no-fraud firm. Also, all AAERs were reviewed to verify that there was no SEC enforcement action against the no-fraud firm. The intent of this procedure was to reduce the likelihood of such misclassification error. Misclassification errors should by minimal given that the likelihood of a management fraud in a random sample is assumed to be small. Note, however, that the misclassification of a no-fraud firm biases against observing the hypothesized relations. common equity of the fraud firms is $127.6 million, the related control firm size could range from $89.3 million to $165.9 million. There is no reason to believe that such a range has a significant effect on board characteristics. ’ If market value information is not available on the COMPUSTAT tape or in the Daily Stock Price Record, fraud and no—fraud firms were matched based on total assets at the end of the fiscal year preceding the occurrence of the management fraud. 45 As a result of the matching procedures 1 through 4, fraud and no-fraud firms should not differ significantly by the type of national stock exchange where a firm’s common stocks trade, firm size, industry, and time period. It is important that the hand and no-fraud firms are similar in these variables because prior research suggests that board characteristics may vary systematically with these variables [Baysinger and Butler (1985), Rosenstein and Wyatt (1990), Shivdasani (1993), Mayers, Shivdasani, and Smith (1994)], and they may also be associated with management fraud as discussed further in Section 3.4. Finally, such matching is consistent with the matching process used in previous corporate governance empirical studies. Based on the descriptive information discussed in the next paragraph, the matching of these four variables appears reasonable. Table 1 shows that the fraud and no-fraud firms are of similar size based on total assets, net sales, and current market value of common stock.’ Fraud firms have mean (median) total assets of $103.2 million ($11.1 million), and no—fraud firms have mean (median) total assets of $79.6 million ($12.5 million). The mean (median) net sales for fraud firms are $102.3 million ($13.0 million) as compared to $93.1 million ($12.9 million) for the no-fiaud firms, respectively. For the subset of 50 fraud and 50 no-fraud firms with available common stock market value information, the mean (median) current market value of common equity is $127.6 million ($26.6 million) for fraud firms as compared to a mean (median) of $124.6 million ($23.7 million) for no-fraud firms. None of the above size measures are statistically different between fraud and no-fraud firms based on paired data t-tests and Wilcoxon matched-pair signed-rank tests. 9 Fifiy firms were matched on current market value of common stock. However, because market value information was not available for twenty-five fraud firms, no-fraud firms for those twenty-five fraud firms were matched based on total assets. 46 Table 1 Matching of Fraud Firms and No-Fraud Firms (3 in thousands) hand Firms No-fiaud Firms Mean Mean [Median] [Median] (Standard Deviation) (Standard Deviation) Total Assets $103,192 $79,626 [11,130] [12,487] (316,734) (221,187) n-75 n-=75 Net Sales $102,285 $93,078 [13,043] [12,936] (262,875) (257,451) n=75 n=75 Current Market $127,630 $124,590 Value of Equity‘ [26,563] [23,660] (263,370) (257,690) n=50 n=50 Stock Traded on: NASDAQ 62 62 4 4 9 9 Match Bmed On: 4 Digit SIC Codes 19 3 Digit SIC Codes 32 2 Digit SIC Codes 21 75 first Year of Fraud: 1979 - 3 1982 - 9 1985 -11 1988 - 3 1980 - 6 1983 -13 1986 - 5 1989 - 6 1981 - 3 1984 - 4 1987 -11 1990 -_1 75 _ ‘ Market price information was available for fifty of the seventy-five fraud firms. Thus, no-fraud firms were matched based on current market value of equity for those fifiy firms. For the remaining twenty- five fraud firms, no-fraud firms were matched based on total assets. 47 Table 1 also shows that the sample fraud firms and related no—fraud firms are also closely matched based on national stock exchange, industry, and time period. The sample includes sixty-two fraud firms whose common stocks trade on the NASDAQ Exchange, four firms whose shares trade on the AMEX, and nine firms whose shares trade on the NYSE. All no-fraud firms trade on the same national exchange as the fraud firm. The sample includes fraud firms representing fifty-seven different four-digit SIC codes, which does not suggest clustering by industry type. Figure 5 includes a list of those industries. Nineteen of the seventy-five no-fraud firms have the same primary four-digit SIC code as the fraud firm, and thirty-two have the same primary three-digit SIC code. For twenty-four of the fraud firms, a suitable no-fraud firm match could only be obtained by matching two-digit SIC codes. Finally, the years when the fraud firms experienced a management fraud range from 1979-1990. SICCodes Numberof SlCCodeDescs-iptlonforfiaudfirms ma Firms 1311 2 Crude Petroleum & Natural Gas 2035 1 Pickled Fruits & Vegetables 2295 1 Coated Fabrics 2328 1 Men’s, Youth’s, Boy’s Work Clothing 2341 1 Women's Underwear & Nightwear 2451 2 Mobile Homes 2621 1 Paper Mills 2819 1 Industrial Inorganic Chemicals 2833 1 Medicinal Chemicals and Botanical Products 2834 1 Pharmaceutical Preparations 3241 1 Cement, Hydraulic 3411 1 Metal Cans 3499 l Fabricated Metal Products 3555 1 Printing Trades Machinery & Equipment 3561 1 Pumps and Pumping Equipment 3571 5 Electronic Computers 3572 1 Computer Storage Devices 3573 3 Flexible Magnetic & Memory Disks 3577 1 Computer Peripheral Equipment 3612 1 Power, Distribution, Special Transformers 3635 1 Household Vacuum Cleaners 3643 1 Current Carrying Wiring Devices 3651 1 Radio & Television Receiving Sets 3662 1 Radio & Television Transmitting Equipment 3663 1 Radio, TV, Communication Equipment 3674 2 Semiconductor Related Devices 3693 1 Radiographic X—Ray Apparatus 3811 1 Engineering, Laboratory, Scientific Instruments 3822 1 Automatic Regulating Controls Figure 5 SIC Code Description for 75 Fraud Firms (Figure 5 continued on next page) 49 SICCode Numberof FraudFirms SICCodeDescriptionforfiaudfirms ape—Museums—NHUNfiflH—nuHNNHH—HHwN—sh} Electronic Measurement Instruments Manufacturer of Measurement and Control Devices Surgical Medical Instruments Games, Toys - Children Trucking Contract Aviation Services Telephone Communication Distributor of Electric Power Metals Service Centers-Wholesales Commercial Machines &. Equipment Wholesale Construct. Equipment Professional Equipment and Supplies Scrap & Waste Materials Mobile Home Dealers Radio, TV, Electric Stores Eating Places Cemetery Subdividers and Developers Management Investment Co. Oil Royalty Traders Advertising Agencies Prepackaged Sofiware CMP Integrated Systems Design CMP Processing Data Preparation Business Services Research & Development Laboratories Equipment Rental & Leasing Services General Medical & Surgical Hospitals Commercial Physical, Biological Research figure 5 (cont’d) 50 3.3 Board Composition Differences Between Fraud and No-Fraud Firms The purpose of creating this sample of fraud and no—fraud firms is to examine whether there are differences in board of director composition between fraud and no- fraud firms in a manner predicted by economic theory. This section highlights that there are univariate differences in board of director composition between fraud and no-fraud firms examined in this study. Table 2 contains univariate descriptive information about board of director characteristics for fraud and no-fraud firms. As reported in Table 2, fraud firms have on average (median) 6.20 (6.0) individuals serving on the board of directors while no- fraud firms have on average (median) 6.72 (6.0) board of director members. The board sizes of firms included in this study are smaller than boards of directors of firms examined in other corporate governance studies. For example, the board sizes of hostile takeover targets and control firms in Shivdasani (1993) are 11.32 and 10.96, respectively. The difference in board size is attributed to the heavy concentration of smaller NASDAQ firms in this study as compared to the primary focus on larger AMEX and NYSE firms in corporate governance studies such as Shivdasani (1993). While similar in size, the composition of boards of directors differs across firms with boards of fraud firms having significantly fewer (p < .01) outside members and more management directors than no-fraud firms. Outside directors represent all directors who are not current employees of the firm. Fraud firms have boards with 50.4% (50%) of its members on average (median) composed of outside directors whereas no-fraud firms have boards with 64.7% (64.3%) of its members on average (median) composed of 51 Table 2 Board Structure Statistics on 75 Fraud and 75 No-Fraud Firms [median] (standard deviation) No-Fraud *. ["1. (m) Variable hand Variable Name Firms (n=75) Firms (n=75) Average size of board of directors BOARDSZ 6.200 6.720 [6.000] [6.00] (2.557) (2.633) % Non-managers on board %OUTBOARD 50.4 64.7‘“ [50.0] [64.31‘“ (22.1) (15.9) Average I of other directorships QUALBOARD .999 .901 held by non-managers on board [.670] [.750] (1.077) (.794) Cumulative % shares held by non- OWNBOARD 5.40 12.0‘“ I] managers on board [1.40] [4.701‘“ (8.40) (15.1) Cumulative % shares held by MGTBOARD 30.3 21.3‘“ managers on board [26.5] [16.91‘“ (21.6) (18-6) 5 of firms where Chairperson is BOSS .853 .733‘ also CEO or President CEO’s board tenure (in years) CEOTENURE 8.847 10.560 [7.000] [8.000] (7.006) (3537) Average board tenure for outside OU'I'I'ENURE 3.786 6587“" directors (in years) [2.000] [5 3001““ (3.835) (4.531) % with active audit committees ACTIVEAC 26.7 56.0‘" Significantly different across firm type at less than the .10, [.05], (.01) level (one-sided when difference in direction predicted, two-sided otherwise) based on paired Meets (or chi-square test) for means or Wilcoxon matched-pair sign-rank test for medians. 52 outside directors. Both mean and median differences are statistically significant at the .01 level using paired t-tests and Wilcoxon matched-pair signed-rank tests. Outside members of both fraud and no-fraud firms hold, on average, one directorship in another firm. Recall from Chapter 2 that the number of additional directorships in other firms is the proxy for outside director quality. Thus, based on this proxy, the Table 2 descriptives suggests that outside director quality does not differ between fraud and no-fraud firms on a univariate basis. Cumulative shares of common stock held by both outside and management directors differ significantly between fraud and no—fraud firms. Outside director mean (median) cumulative common stock ownership in fraud firms of 5.4% (1.4%) is significantly lower at the .01 (.01) level as compared to mean (median) outside director cumulative common stock ownership in no-fraud firms of 12.0% (4.7%). Management director mean (median) cumulative common stock ownership in fraud firms of 30.3 % (26.5%) is significantly higher at the .01 level (.01 level) as compared to mean (median) management director cumulative common stock ownership in no-fraud firms of 21.3 % (16.9 %). The chairperson holds managerial positions of CEO or president in 85 % of fraud firms and 73.3 % of no—fraud firms and such difference is statistieally significant at the .10 level. While the CEO’s tenure on the board is, on average, 8.9 years for fraud firms as compared to 10.6 years for no-fraud firms (not significantly different), the average tenure of outside directors is statistically longer for no-fraud firms compared to fraud firms. Average (median) tenure on the board is 3.8 years (2.0 years) for outside 53 directors of fraud firms relative to an average (median) tenure of 6.6 (5.8) years for no- fraud firms. No-fraud firms are significantly (p< .01) more likely to have an active audit committee compared to fraud firms. Sixty-seven percent of the no-fraud firms have an audit committee of the board whereas only 41% of the fraud firms have an audit committee (not reported in Table 2). While a firm may have an audit committee, the firm may be “window dressing" by creating an audit committee that never meets. While 56% of no-fraud firms have an audit committee that met at least once during the year, only 27% of fraud firms have audit committees that met at least once during the year prior to the fraud. 3.4 Effects of Differences in Other Firm Characteristics While the prior section documents that there are univariate differences in board of director characteristics between fraud and no-fraud firms, it is important that other non-board of director characteristics that are likely to be associated with the occurrence of management fraud and board of director composition be controlled for when testing the hypotheses. Failure to consider variables that may be correlated with the occurrence of management fraud and board of director composition may bias tests of the hypotheses. This section highlights procedures that serve to minimize the potential for correlated omitted variables in this study. For an ideal test of the hypotheses, the fraud and no-fraud firms would only differ in board of director composition and whether or not a management fraud has occurred. In other words, the ideal test would include fraud and no-fraud firms whose probabilities for fraud are identical based on all non-board of director characteristics. Unfortunately, 54 due to the extensive list of possible red flag indicators of management fraud identified in management fraud research, such matching of fraud and no-fraud firms is not practical. Section 3.2 describes how fraud firms were matched with no-fraud firms on the basis of national stock exchange, industry, firm size, and time period. As noted in that section, these variables were used to match fraud and no-fraud firms because prior corporate governance research shows that board of director composition may vary systematically with these variables [Baysinger and Butler (1985), Rosenstein and Wyatt (1990), Shivdasani (1993), Mayers, Shivdasani, and Smith (1994)], and they are likely to be associated with occurrences of management fraud. The association of these four matching variables with management fraud and board of director composition is discussed next. Firm size is likely to be associated with both management fraud and board of director composition. Management fraud research notes that decentralized firms are more likely than other firms to experience management fraud. Given that firm size may be associated with organizational structure - centralization versus decentralization - firm size is likely to be associated with occurrences of management fraud. Firm size is also likely to be associated with board of director composition. Larger firms are likely to have more outside directors because of their expertise in monitoring and project evaluation [Shivdasani (1993)]. Serving on the board of directors of a large corporation can also enhance a director’s visibility and reputation, which will increase the likelihood that he/she will serve on additional boards of directors [Shivdasani (1993)]. Beeause 55 firm size is likely to be associated with both management fraud and board of director composition, it is important to match firms on the basis of firm size. The industry in which the firm operates is likely to be associated with both management fraud and board of director composition. Management fraud research notes that increased management pressure due to inadequate profitability relative to industry peers often leads to management fraud. Industry trends such as increased competition may place undue emphasis on the firm’s profits relative to the industry. As a result, industry trends and the probability for management fraud are likely to be associated. Additionally, management fraud research notes that firms experiencing many contentious and difficult accounting issues are more likely to experience management fraud. Because difficult accounting issues can be industry-specific, industry and the occurrence of management fraud are likely to be associated. The firm’s industry is also likely to be associated with board of director composition. Firms in certain industries could require outside directors with a greater amount of industry-specific experience. Matching on the basis of industry serves as a control for the opportunities of directors to serve on other boards [Shivdasani (1993)]. Because of the association of industry with both management fraud and board of director composition, it is important to match firms on the basis of industry. It is also important to match firms in the same time period because the likelihood of management fraud and changes in board of director composition are likely to vary across time. Economic conditions change over time, and those changes may affect the likelihood of management fraud. Recessionary periods may place undue pressure on management leading to concern about whether the entity will be able to survive. In 56 response to that pressure, management may act fraudulently. The corporate governance literature notes that board of director composition varies across time as well. Baysinger and Butler (1985) find that board of director composition has been changing over time. They note that between 1970 and 1980, the proportion of outside directors increased. As noted in Section 2.4.1 of Chapter 2, WW [August 19, 1993, p. 1] reports similar increases through the early 19905. Because the likelihood of management fraud and board of director composition are likely to vary over time, it is important that fraud and no-fraud firms be matched in the same time period. Finally, it is important to match firms on the basis of the national stock exchange where the firms’ common stocks trade because the type of exchange is likely to be associated with both management fraud and board of director composition. Firms that trade on the NASDAQ exchange are likely to have different characteristics from AMEX and NYSE firms. Some of those characteristics may be associated with occurrences of management fraud. For example, developing companies are likely to initially trade on the NASDAQ exchange. Developing companies may have newer, less experienced management personnel, and management fraud research notes that management fraud is more likely when management and accounting personnel are inexperienced. Thus, the type of national stock exchange may be associated with occurrences of management fraud. Additionally, the type of national stock exchange is associated with board of director composition, given that each of the exchanges have differing composition requirements. For example, as discussed in Section 2.4 of Chapter 2, the national stock exchanges have different requirements for how many non-management directors must serve on the audit committee. Because the type of national stock exchange is likely to 57 be associated with both the occurrence of management fraud and board of director composition, it is important to match fraud and no-fraud firms on the basis of national stock exchange. In addition to the four variables used to match fraud and no-fraud firms, there may be other firm-specific characteristics associated with the occurrence of management fraud and board of director composition that should be considered when evaluating the hypotheses. There are numerous red flag indicators of management fraud. However, only those that are likely to be associated with board of director composition are important to this study because their omission may otherwise create a potential correlated omitted variable bias. Based on the review of the management fraud and corporate governance literatures, four additional firm-specific characteristics were identified for inclusion in this study. They were identified because they are likely to be associated with both the . occurrence of management fraud and board of director composition. The four additional variables included in this study are: l. Extent of firm growth. 2. Financial health of firm. 3. Length of time the firm’s stock has publicly traded. 4. Extent of monitoring by blockholders. Motivation for the inclusion of these four variables in this study, as well as how they differ between fraud and no-fraud firms, is described in subsections 3.4.1 through 3.4.4 that follow. These variables are included as control variables in the logit regression model (described extensively in Chapter 4) because it is more practical to include them 58 as control variables rather than consider them as part of the matching process. Table 3 contains the univariate descriptive information about these variables. 3.4.1 Growth. The extent of recent firm growth may be associated with the likelihood of management fraud and board of director composition. Thus, it is important to include a measure of firm growth in this study when evaluating the hypotheses. One of the most significant management fraud red flag indicators noted in the management fraud research is the presence of rapid company growth. Bell et al. , (1991) note that if the company has been experiencing rapid growth, management may be motivated to misstate the financial statements during a downturn to give the appearance of stable growth. In high growth situations, responsibility for overall decision making and segments of key decisions are spread across a number of individuals, and no one individual has sufficient authority or information to prevent or stop illegal activities. Extensive growth through a rapid expansion of sales, acquisition of a new division or firm, or entry into an unfamiliar line of business can weaken internal controls. Weak or non-existent internal controls can make fraud easier to commit and detection less likely [National Commission on Fraudulent Financial Reporting (1987)]. The extent of company growth may also be associated with board of director composition. In rapid growth situations, needed modifications to rules, procedures, and other control mechanisms, like the board of directors, often lag behind the growth of the firm. Necessary changes to board of director composition, such as the addition of outside members, may follow high growth periods. As a result, high growth firms may have few outside directors. Thus, firm growth and board of director composition may be correlated. 59 Table 3 Non-Board of Director Differences Across Fraud and No-Fraud Firms Fraud firms No-fiaud firms Mean Mean Variable [Median] [Median] Name (Standard Deviation) (Standard Deviation) Average 2-year growth in GROWTH 103.8 51.2" total assets (in %) [41,2] [20.7]... (199.3) (125-9) % of firms in financial TROUBLE .467 .280‘“ trouble before year of fraud length of the stock has AGEPUB 5.273 8.744‘" traded on national exchange [1.500] [10.5001'“ (in years) (6.402) (6.313) % of common stock held by BLOCKl-IID 6.12 7.73 unaffiliated blockholders [0.00] [0.00] (10.2) (13.2) t. [“l. c") Significantly different across firm type at less than the .10, [.05], and (.01) level (one-sided) based on paired t-tests (or chi-square test) for means and Wilcoxon matched-pair sign-rank test for medians. 60 As reported in Table 3, fraud firms and no-fraud firms differ significantly in the extent of growth in assets for the two years preceding the year of the management fraud. The mean (median) growth in assets for fraud firms is 103.8% (41.2%) which is statistically greater than the mean (median) growth in assets for no-fraud firms of 51.2% (20.7%) at the .05 (.01) level based on paired t-tests (Wilcoxon matched-pair signed-rank tests). Because the extent of firm growth may be associated with management fraud and board of director composition, the variable, GROWTH, is included as a control variable in this study. GROWTH represents the average change in total assets for the two years ending before the first year of the management fraud occurrence. '° 3.4.2 Financial Health. The extent of the firm’s financial health may be associated with the likelihood of management fraud and board of director composition. Thus, it is important to include a measure of financial health in this study when evaluating the hypotheses. The management fraud literature notes that the degree of financial health may be associated with the likelihood of management fraud [Bell et al. (1991)]. Poor financial performance may cause management to place an undue emphasis on earnings and ‘° Fraud and no-fraud firms also differ significantly (in the same direction) in mean (median) average two year growth in net sales at the .05 (.01) level. When average two- year growth in net sales is used, the results discussed in Chapter 5 are not substantively different. . 61 profitability thereby increasing the likelihood of management fraud. Bell et al. , ( 1991) identify three red flag indicators that suggest management fraud is likely when: 1. There is inadequate profitability relative to the industry. 2. Management places an undue emphasis on earnings projections. 3. There is doubt about an entity’s ability to continue as a going concern. All three of these indicators suggest an association of financial health and the likelihood of management fraud. The corporate governance literature suggests that the degree of financial health may also be associated with board of director composition. Gilson (1990) finds that a firm’s financial distress causes significant changes in board of director composition with boards shifting to a higher number of directors who are creditors and blockholders subsequent to the onset of financial distress. Hermalin and Weisbach (1988) find that poor performance leads to changes in board composition with inside directors being replaced with outside directors. Other studies find that top management (i.e. , president and CEO) turnover occurs subsequent to poor firm performance [Coughlan and Schmidt (1985), Weisbach (1988), Warner, Watts, and Wruck (1988)]. Because the degree of financial health may be associated with the likelihood of management fraud and board of director composition, the variable TROUBLE is included in this study as a control variable. Using a measure of financial trouble consistent with DeAngelo and DeAngelo (1990) and DeAngelo, DeAngelo, and Skinner (1994), TROUBLE is dichotomous with the value of one when a firm has reported at least three annual net losses in the six-year period preceding the first year of the management fraud. Otherwise, TROUBLE has a value of zero. 62 It is important to control for financial health in this study given that fraud and no- fraud firms differ significantly in financial health, as reported in Table 3. Forty-seven percent of the fraud firms are in financial trouble whereas only twenty-eight percent of the no-fraud firms are in financial trouble. The difference between financial health is statistieally significant at the .01 level." 3.4.3 Length of Time Publicly Traded. The length of time that a firm’s common stock has traded publicly on a national exchange may be associated with the likelihood of management fraud and board of director composition. Thus, it is important to include a measure of the length of time the firm’s common stock has publicly traded when evaluating the hypotheses. The management fraud literature suggests that the length of time that a firm’s common stock has traded in public markets may be associated with the likelihood of management fraud. The National Commission on Fraudulent Financial Reporting (1987, p. 29) notes that new public companies may have a proportionately greater risk of management fraud because management may be especially pressured to meet earnings expectations, given that they are new registrants in the market. Research on management fraud shows that management’s undue emphasis on meeting earnings projections is a significant red flag indicator of management fraud. The length of time that a firm’s common stock has traded in public markets may also be associated with board of director composition. Before trading on a national " The extent of financial trouble for both fraud and no-fraud firms is high. This is most likely due to the nature of the firms included in this sample. As previously noted, the bulk of this sample consists of NASDAQ firms. Given that newer, developing companies typically trade on the NASDAQ before trading on the AMEX or NYSE, the high percentage of financial trouble for both fraud and no—fiaud firms appears reasonable. 63 exchange begins, a firm must make the necessary changes in board structure to satisfy the requirements of the exchange. Changes in board of director composition such as adding outside directors directly affects outside director tenure, which is a variable of interest in this study. When new registrants add outside directors to the board of directors, outside director tenure will likely be shorter than the outside director tenure of firms whose common stocks have traded publicly for long periods of time. As reported in Table 3, fraud firms and no-fraud firms differ significantly in the length of time their respective common stocks have traded publicly on a national exchange. The mean (median) length in years of public trading of common stock on a national exchange for fraud firms is 5.27 years (1.5 years) which is statistically shorter than the mean (median) length of public trading of common stock for no-fraud firms of 8.74 years (10.5 years) at the .01 (.01) level based on paired t-tests (Wilcoxon matched- pair signed-rank tests). Given that the length of time that the firrn’s common stock has traded publicly may be associated with the likelihood of management fraud and board of director composition, a measure of the length of time that the firm’s common stock has traded publicly is included as a control variable in this study. The variable AGEPUB represents the number of years the firm’s stock has traded on a national stock exchange. Information about the year when the firm began trading on a national stock exchange is obtained from the annual report or proxy statements examined, if disclosed. If there is no disclosure about the time period when the firm initially went public, the date of the initial filing of securities with the SEC is obtained from the SEC Workload File that is on microfiche. — 64 3.4.4 Blockholders. Large blockholders of common stock may be associated with the likelihood of management fraud and board of director composition. Thus, it is important to include a measure of the extent of blockholder ownership in the firm when evaluating the hypotheses of this study. Large blockholders may reduce the likelihood of management fraud. Shleifer and Vishny (1986) note large block shareholders have incentives to monitor management and serve as an additional control mechanism. Large institutional investors are likely to closely scrutinize firm operations and hold boards responsible for corporate performance. Shivdasani (1993) finds that block ownership unaffiliated with management increases the likelihood of a hostile takeover attempt consistent with the view that blockholders serve as a corporate governance mechanism by facilitating takeover attempts to replace ineffective management. Increased monitoring by large blockholders may reduce the likelihood of management fraud. Large blockholders may also be associated with board of director composition. large blockholders may be able to influence who is selected to be a member of the board of directors. Gilson (1990) finds that increases in outside director representation on the board of directors is associated with increases in blockholder ownership in periods subsequent to a firm’s poor performance. Brickley and James (1987) find a negative relation between concentration of stock ownership and proportion of outside directors for banks in states that restrict acquisitions. As reported in Table 3, blockholders hold on average (median) 6.1% (0%) of the outstanding common shares of fraud firms whereas blockholders hold on average (median) 7.7% (0%) of the outstanding common shares of no-fraud firms. The mean 65 (median) is not statistically different based on paired t-tests (Wilcoxon matched-pair ciated with re variable, i represents Mmm... mmmmmmrm %mwm w mammm myu.w.m.mmwam.m -.mmm mass“..- 762.55.. an NQHKI ignore: a. “Nu. HEL DU magengmenm Knew—:8 man—v. C2... mmsmmmomos Oemmnmogmwum are $689.30: om wmmmmanr 5.9318 $9er mm magism 5.3::an 39.. AFC 2.5 @0958... Ame—.521 AUUV $2.8m 5 .3835 .5..— mnmnmaan =w..m...em. One 3:60.63. cm :5 mag. .m .0 339.8 v.9... mesmmmnzos SE. $.5an F300 «3:88 9.5.3.8.. 9.15m "rm me: See 3...... 2e mmmr v55 82.2.30: 5 mega—mcsm 95 via 9.23.. 288 8.522. :5 9.8295 817... 358...»— On ..mmwonmm no SEQ. "Em .m mzmnrea. <9: ammvosmm mm 8,535.05. 6... 9m 8&2m 0: :5 2.33. SE neg—3m mmmammne 9.8 0... <9... Swanson. 28 3.52.055 33m. man 3: 38 3.. .5 Fr 25:88. 38mm 2.929. :6 mo=o.$..m 9.8395 men. 3»: v.9... 83123 ..mmwonmm 8 a... 3988 :08... $102. H. UK :8 :2... 3.. names: Raven—m3 2.15. 2.2.... a... 283%.? 3.5m $539 On 3. :8 9.8 3. 27.9.. <0: 59.3.3 v5: :3an :5 25.5%.. <8 . . Zo : so. 23.» N. Sam near? om :6 3.0883“ ”homage—m 0.. $6.... 9: 80< "a: .25: 0:8 .35 .58 0:80: .8 box: 20.: 2: 88.....Eeo £5: 0>wo< a 0.809;.— .e Eden 05 .0 330300.30 00:5 acre 0:»... ..0 2:5. Duo :05 .83... a. Etc 0:80: ..0 2802:. ..0 0.00.. 05 :0 95:8 9.0 6: .35. 0:80: :5 0:8 0003.3 8.5.... .0: 88.. 2:8 ONO 0 .258 0:0... ..0 23.2520 :05 30302.. .0 ONO .0 2038.. 3.03:2. .58 38-0: ..0 «0020.320 :05 .0028... .0 OH. 20.. 8 be...— 80— 2: .58 0:80: .0 2020.520 ”m: ..0 20:10.. 1.8.3.3.: 20.. 8 box: 20... .0: 2: menu .58 ..0 2: 2020.355 a $8.8 38-0: 5...» 00300.30 0.. 8 20M: 20:. 2e ARWN 8 *0 8:20:30 ..EE 0:2,. .0 p.802... 2.050900... :8 0:..8 0:8 ..0 8w::. 853E 5.23 8:20:30 5 .0220... .05 $02.0 2...... .380: .0: E 8.3. 8:20:30 .2»... 20.. 0.802.: 305035.: 3.: .0:- 38 0003.00 .050 .0: .00: BE 0... 5 8:20:30 30:6»..qu .0 308.0 0:... 0 2809.5 Eastman—z .0 823.302an .208 0:8 ..0 2802:. 023:0 :05 . 2:2... 0:8 ..0 2802:. 028:0 .35 .030— nm 2...... .58 8.0:: .2%... ..0 2: .58 38.0: ..0 2802:. 028:0 ”NS .0: .0 2802:. 028:0 .3 20.. .58 .050 5 820.802.: 1:05:00- ..0 .0852 0 ..EE 0:8 ..0 0.802.: 023:0 :05 2:5. .25... 0:0... ..0 2:08 0w....0>: .0w:0_ 0:... ..EG 28-0: ..0 2802:. 028:0 ”h: .802... 028:0 :05 .080— 8 2...... 0:80.. ..0 2809:: 035:0 ..0 2:08 0w20>< 0 $5.... 3...... ..0 2802:. 023:0 :05 EC... .58 0:2. ..0 8:20:30 .802... 033:0 .581... s. 8.3. 82256 .23.. 22. 208.... 85:0 a: S... 8...... a 2E: .586: .e 238.... 8.2.5 .3 32. E... 2.. c. .2550 e 2800.5 028:O .0 Sufi-300.30 .20.... .58 :8 2802:. 023:0 ..0 833020.. .23.. 0:.— ..Eu 0:802 3: SE .50... :05 2:8 .580: .8 .052 a. 2802:. 023:0 ..0 5.885.290“ e 0.88.5 033:6 ..e guano-2&0: 2.858... .38.. ”uses... .83.... 121 6.1.2 Differences in Outside Director Characteristics. The characteristics of outside directors who serve on boards of directors of fraud firms differ significantly from characteristics of outside directors who serve on boards of directors of no-fraud firms. Outside directors of no-fraud firms have significantly higher ownership levels in the firm, longer tenures on the board of directors, and fewer outside directorships in other firms. These findings are discussed further in the following paragraphs. Agency theory suggests that a high stake in a company’s outstanding equity should provide individual directors with a strong incentive to promote firm activities that increase a firm’s value because this increases the value of the director’s own investment. Thus, agency theory suggests that as the extent of ownership in the firm held by outside directors increases, the occurrence of management fraud should decrease. The results of this study are consistent with this prediction given that outside directors of no-fraud firms have significantly higher ownership stakes in the firm than outside directors of fraud firms. Additional analysis based on a piecewise logit regression model indicates that as outside director ownership in the firm increases, particularly above 5 96 , the likelihood of management fraud decreases. This finding suggests that holding a personal financial stake in the firm encourages outside directors to spend the time and effort necessary to effectively monitor management for the prevention of management fraud. Interestingly, the relation between firm ownership and the occurrence of management fraud holds for grey directors but does not hold for independent directors. Perhaps, the incentives from owning shares in the firm helps grey directors overcome potential conflicts of interests with management to effectively prevent the occurrence of 122 management fraud. However, for independent directors, the separation from management through the lack of any non-board affiliation enables them to effectively monitor management for the prevention of management fraud, and no significant incentives are derived from increasing independent director ownership in the firm. Outside directors of no-fraud firms have significantly longer average tenures on the board of directors of those firms. This empirical finding suggests that managers of fraud firms may be able to override outside director monitoring in order to commit management fraud when outside directors have recently joined the board of directors. When outside director tenure is short, managers may be able to take advantage of outside directors’ lack of seniority to avoid monitoring by those directors. Results also suggest that the average tenure on the board of directors for independent directors of no-fraud firms is longer than the average tenure of independent directors of fraud firms. However, there appears to be no significant difference in average tenure of grey directors of fraud and no-fraud firms. Surprisingly, the empirical results are not consistent with the prediction that no- fraud firms have higher quality outside directors than fraud firms. Boards with similar percentages of outside directors may vary in their effectiveness as a monitor of management depending on whether outside directors have an incentive to maintain reputations as quality directors. Fama (1980) argues that the external market for directors provides an incentive by rewarding high quality directors with additional directorships and punishing low quality directors with fewer directorship opportunities. Using the number of additional directorships held by outside directors as a proxy for outside director quality, the finding that outside directors of fraud firms have significantly 123 higher numbers of additional directorships than outside directors of no-fraud firms suggests that fraud firms have higher quality outside directors compared to outside directors of no—fraud firms. This finding is opposite of the expected relation. An explanation of this unexpected result may be that the proxy used to measure outside director quality is inappropriate, particularly given the heavy concentration of small firms in this study. Outside directors of small firms are likely to serve as outside directors of other similar-sized firms. Such firms may require outside directors to spend more time with the day-to-day monitoring than the time required to serve on larger, more widely-known firm boards. Also, small firms may have difficulty attracting high quality directors. If this is the case, as the number of additional directorships increases, the quality of monitoring by outside directors at the individual firm level deteriorates. 6.1.3 Differences in Management Director Characteristics. The results of this study suggest that managers who serve on the board of directors of fraud and no-fraud firms differ in the level of ownership in the firm only when they hold moderate levels of outstanding common shares of the firm - between 5 96 and 25 % of the outstanding shares. This findings suggests that increases in firm ownership for managers serving on the board of directors within the range of 5% to 25% decreases the likelihood of management fraud. That finding is consistent with agency theory, which predicts that stock ownership by management in the firm decreases agency costs. However, the relation between increases in ownership held by management directors and the occurrence of management fraud is limited to moderate levels of ownership given that ownership levels held by management directors do not appear to differ between fraud and no-fraud firms when managers own small percentages (less than 5 96) or large percentages (greater 124 than 25 %). Apparently, at both low and high levels of firm ownership, there are other incentives for management that offset incentives to prevent management fraud derived from firm ownership. The overall results suggests that the relation between ownership held by management directors and the occurrence of management fraud may not be linear. The results also suggest that chairpersons of the board of directors of fraud firms are not more likely to hold managerial positions of CEO or president than chairpersons of no-fraud firms. Given that eighty-five percent of fraud firms and seventy-three percent of no-fraud firms have chairpersons who hold one of these two managerial positions, it appears that this lack of separation of board of director and management positions is not that uncommon in most firms. Additionally, the results indicate that the average tenure of CEOs on the board of directors does not differ significantly across fraud and no-fraud firms. 6.1.4 Differences in Other Characteristies. This study documents the importance of an active audit committee for the prevention of management fraud. The findings highlight that boards of directors of no-fraud firms are significantly more likely than boards of fraud firms to have an audit committee that meets at least once during the year. This result is consistent with the view of many corporate governance reform proponents who believe that the audit committee ean be an effective deterrent of management fraud. These proponents believe audit committees with oversight responsibility for the financial reporting process are critical for effective governance by the board of directors for the prevention of management fraud. 125 Additionally, while this study does not include any hypotheses predicting a relation between board of director size and the occurrence of management fraud, the results of this study indicate that fraud and no-fraud firms differ in board size. Fraud firms are significantly more likely to have a larger board of directors than no-fraud firms. This finding is consistent with the view that larger boards of directors are not able to function effectively as a controlling body leaving management relatively free [Chaganti, Mahajan, and Sharma (1985)]. In summary, this study documents that there is a relation between board of director composition and the occurrence of management fraud. It highlights that the composition, ownership, and tenure of members of the board of directors, particularly outside directors, as well as the presence of an active audit committee are important factors that affect the board of directors’ ability to prevent management fraud. 6.2 Contributions to Management Fraud and Corporate Governance Research This study contributes to existing research because little is known about the relation between board of director composition and management fraud. Prior research has not empirically tested economic theory suggesting that there is a relation between board of director composition and the occurrence of management fraud. By providing initial empirical results of this economic theory prediction, this study expands knowledge about the effectiveness of the board of directors as a corporate governance mechanism designed to prevent agency problems such as management fraud. This study contributes to research on management fraud by highlighting how characteristics of boards of directors differ between fraud and no-fraud firms. While prior management fraud research identifies numerous red flag indieators of firms likely 126 to experience management fraud, none of those indieators address unique characteristics of boards of directors of firms experiencing management fraud. Many of the red flag indieators imply that the board of directors of fraud firms may be ineffective monitors of management by consistently noting the significance of ”weak internal control environments" for many of the firms experiencing management fraud. By providing a more focused analysis of the internal control environment, particularly the board of directors, this study provides empirieal support about differences in one aspect of the internal control environment - the board of directors - betrveen fraud and no-fraud firms. Recall that the purpose of this study is to test economic theory about the monitoring role of the board of directors in a setting of management fraud. While the purpose of this study is not to develop a predictive model of management fraud, this study contributes to the development of future management fraud predictive models by providing evidence of a relation between board of director composition and the occurrence of management fraud. Such evidence suggests the importance of including board of director composition in future predictive models. This study also contributes to the corporate governance literature. Previous research on corporate governance considers the relation of board of director composition with either some aspect of firm performance or some type of acute agency problem. No prior study has empirieally examined the relation of board of director composition with the agency problem of management fraud. Many of the empirical findings from this study provide additional support for a couple of the corporate governance reform proposals suggested by groups such as the National Commission on Fraudulent Financial Reporting and the AlCPA’s Public 127 Oversight Board. In addition, these findings are relevant to standards-setters such as the Auditing Standards Board, which is responsible for developing auditing professional standards. Current standards set certain management fraud detection responsibilities for auditors and provide guidance to increase audit effectiveness in the form of red flag predictive indicators for consideration by auditors. Given that current standards are silent as to board of director characteristics that may be associated with occurrences of management fraud, empirieal findings in this study may provide further insights for consideration by the Auditing Standards Board when making future modifieations to existing authoritative standards. 6.3 Limitations There are limitations that may reduce the generalizability of the results of this study. These limitations relate to the uniqueness of the type of management fraud examined in this study, biases of the sources used to identify fraud firms, potential misclassifications of fraud firms as no-fraud firms, potential correlated omitted variables bias, and alternative explanations for the documented relation of board of director composition and management fraud. These limitations are discussed next. First, conclusions of this study are limited to two types of management fraud - misappropriation of firm assets and fraudulent financial reporting. These types of management fraud were examined in this study because they are directly related to the , financial statement reporting activities of publicly traded firms. There are other types of fraudulent activities such as customer fraud, employee fraud, anti-trust violations, illegal mergers, tax evasion, and defense contract violations. The findings of this study may not be generalizable to other fraudulent activities. 128 Second, the results are limited to cases of management fraud for publicly traded firms investigated as part of an enforcement action by the SEC staff or reported in The W. To the extent that both the SEC and W’s selection of management fraud cases is not representative of the population of management fraud occurrences, the ability to generalize the results of this study is limited. Third, while procedures have been performed to minimize the potential misclassification of fraud firms as no-fiaud firms, such misclassifieation may have occurred. Recall, however, that such misclassification biases against findings consistent with the hypotheses. More importantly, it is assumed that the likelihood of management fiaud in a random sample is small. Fourth, while this study attempted to control for endogenous differences in characteristics of fraud and no-fraud firms, there may be certain unidentified variables that may be associated with both management fraud and board of director composition. The exclusion of such characteristics, if any, may create a correlated omitted variables bias that potentially affects conclusions about the tests of the hypotheses. Finally, there is an alternative explanation to the economic theory prediction about the relation between board of director composition and the occurrence of management fraud. As summarized in this study, economic theory argues that shareholders delegate responsibility to the board of directors for the oversight of management for purposes of minimizing agency problems like management fraud. This theory argues that the board of directors is an important monitor of management with ultimate control over activities within the firm. Critics of this theory argue that the board of directors is not an effective 129 monitor of management because the board of directors is generally controlled by top management. They argue that the CEO often selects individuals for service on the board. Consistent with this alternative view of the board, some may argue that the observed empirieal relation between board of director composition and the occurrence of management fraud is not evidence that certain characteristics of the board of directors serve to decrease the likelihood of management fraud. Instead, it is possible that managers who control the selection of individuals to serve on the board of directors use board of director composition to signal information about the quality of top management to investors. Perhaps, high quality managers signal information about their quality by creating boards of directors with unique characteristics, such as higher percentages of outside directors and active audit committees, to distinguish them from other firms controlled by lower quality managers. Unfortunately, the research design of this study cannot distinguish the economic theory prediction from this alternative view. 6.4 Suggestions for firture Research This study finds that there is a relation between board of director composition and the occurrence of management fraud. While prior management fiaud research notes that fraud firms often have weak internal control environments, none of these studies examine differences in board of director composition. This study is particularly relevant to management fraud researchers who are attempting to development management fraud predictive models. As future predictive models are developed, this study suggests that researchers consider the inclusion of board of director composition in those predictive models. 130 As noted in the prior section, this study does not examine the relation between board of director composition and the occurrence of other fraudulent activities. Future research on board of director composition may be able to provide additional insights about board effectiveness for the prevention of other illegal activities. This study focuses on board of director composition that firms have in place in the fiseal year prior to the occurrence of management fraud. One extension of this research would be to examine how board of director composition evolved over a period of time leading up to the occurrence of management fraud. Additionally, research on changes in board of director composition subsequent to the announcement of management fraud may provide further evidence of how shareholders modify board of director composition in response to evidence of failed board governance. Examining board of director characteristics in periods subsequent to management fraud may also provide additional evidence about the existence of an external market for corporate directors. Future research may be able to determine whether the external market punishes directors of firms experiencing management fraud by offering fewer directorships in other firms in periods subsequent to the discovery of management fraud. Finally, according to Kalbers and Fogarty (1993), research on audit committee effectiveness is limited. Perhaps, future research about differences in audit committee characteristics between fraud and no-fraud firms may provide additional insights about the effectiveness of audit committees in fulfilling their responsibilities for financial reporting activities. LIST OF REFERENCE LIST OF REFERENCES Albrecht, W. Steve, Marshall B. Romney, David J. Cherrington, 1. Reed Payne, and Allan J Roe WWW Prentice-Hall Inc” Englewood Cliffs, NJ, 1982. Albrecht, W. Steve and Marshall B. Romney, 'Red- -Flagging Management Fraud. A Validation," W, Vol. 3, JAI Press Inc. Greenwich, CF, 1986. Alchian, Armen A. and Harold Demsetz, "Production, Information Costs, and Economic Organization.“ Wm 1972,1’P- 777-795 - American Institute of Certified Public Accountants, "SAS No. 53, The Auditor’ s Responsibility to Detect and Report Errors and Irregularities," mm W. AICPA New York 1992 ,‘SAS No. 55, Consideration of the Internal Control Structure in a Financial Statement Audit " WM AICPA New York, 1992. The Public Oversight de8’ WW —Amunfing£mf_essjgn, AICPA, Stamford, CI‘, 1993. , “Meeting the Financial Reporting Nwds of the Future," White Paper, June 1993. Baysinger, Barry D. and Henry N. Butler, ”Corporate Governance and the Board of Directors. Performance Effects of Changes rn Board Composition, W W Vol 1 No 1 Fall 1985 PP 101124 Bell, Timothy B., Scott Szykowny, and John J. Willingham, “Assessing the Likelihood of Fraudulent Financial Reporting: A Caseaded Logit Approach, " Working Paper, December 1991 . Berle, A.A. Jr. and G.C. Means, - Macmillan, New York, NY, 1932. 131 132 Brickley, James A. and Christopher M. James, "The Takeover Market, Corporate Board Composition, and Ownership Structure: The Case of Banking, " W Wigs, Vol. XXX, April 1987, pp. 161-180. Byrd, John W. and Kent A. Hickman, "Do Outside Directors Monitor Managers?, " WWW Vol 32 1992 PP 195-221 Chaganti, Rajeswararao S., Vijay Mahajan, Subhash Sharma, "Corporate Board Size, Composition, and Corporate Failures in Retailing Industry," W W, Vol. 22, No. 4, July 1985, pp. 400-417. Cochran, P.L., R.A. Wood, and T.B. Jones, "The Composition of Boards of Directors and Incidence of Golden Parachutes, " WWW. September 1985, pp. 664-671. Cosslett, Steven R. "Efficient Estimation of Discrete Choice Models," C. F. Manski and D. McFadden, eds. Applications, MIT Press, Cambridge, MA, 1981. Couglan, Anne T. and Ronald M. Schmidt, "Executive Compensation, Management Turnover, and Firm Performance: An Empirical Investigation," W Amntinwdjmnomics. Vol 7 1985 P 43-66 DeAngelo, Harry and Linda DeAngelo, "Dividend Policy and Financial Distress: An Empirical Investigation of Troubled NYSE Firms, " W, Vol. XLV, No. 5, December 1990, pp. 1415-1431. ,Linda DeAngelo, and Douglas J. Skinner, "Accounting Choice m Troubled Companies," WWW. Volume 17 1994 PP 113- 143. Demsetz, Harold, "The Structure of Ownership and the Theory of the Firm," lonmalgf MM, June 1983, pp. 375-393. _and Kenneth Lehn, "The Structure of Corporate Ownership: Causes and ConsequenceS"1Qumal_o££nlitical_E&onmnx V01 93 1935 PP 1155 1177 Dopuch, N., R. W. Holthausen, and R. W. Leftwich, "Predicting Audit Qualifications With Financial and Market Variables," Wm, vol 62, July 1987, pp. 431-454. Elliott. Robert K. and John 1- Willingham, W W, Petrocclli Books Inc., New York, 1980. 133 Fama, Eugene F., "Agency Problem and the Theory of the Firm," loumaloflfiolifical Economy, 1980, pp. 288-308. _ and Michael C. Jensen, "Separation of Ownership and Control," W Wigs, Vol. XXVI, June 1983a, pp. 301-325. and Michael C. Jensen, "Agency Problems and Residual Claims," W W, Vol. XXVI June 1983b, pp. 327-349. Feroz, Ehsan H., Kyungioo Park, and Victor S. Pastena, "The Financial and Market Effects of the SEC’s Accounting and Auditing Enforcement Releases, Journal WW, Vol. 29, Supplement 1991, pp. 107-142. Gilson, Stuart C. "Bankruptcy, Boards, Banks, and Blockholders: Evidence on Changes in Corporate Ownership and Control When Firms Default," W W, Vol. 26, pp. 1-33. Grossman, Sarnford J. and Oliver D. Hart, "Takeover Bids, The Free-Rider Problem, and The Theory of the Corporation." WW. Spring 1980, pp. 42-64. Hackenbrack, Karl, "The Effects of Experience With Different Sized Clients on Auditor Evaluations of Fraudulent Financial Reporting Indicators," W W. Vol 12 No 1 Spring 1993 Hart, Oliver D., "The Market Mechanism As An Incentive Scheme," W W. 1984. PP- 366-332- Hcrmalin, Benjamin and Michael S. Weisbach, "The Determinants of Board Composition." WW. Vol. 19. No. 4. Winter 1988. pp. 589-606. _,"The Effects of Board Composition and Direct Incentives on Firm Performance," W. Vol 20.1991 PP 101-112- Herman. E.S.. WW. Cambridge University Press. Cambridge, MA, 1981. Hsieh, David A, Charles F. Manski, and Daniel McFadden, "Estimation of Retrospective Probabilities From Augmented Retrospective Observations," Journal WWW. Vol 80. PP 651-652 J ensen, Michael C. , "The Modern Industrial Revolution, Exit, and the Failure of Internal Control Systems, WW, Vol. XLVIlI, No. 3, July 1993, pp. 831-880. 134 _and William H. Meckling, "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure" lonmalnffiinancialfmnomics. 1976. PP 305- 360. __and Kevin Murphy, Performance Pay and Top Management Incentives," 19131131 MW.V01 98 1990. PP 225-264 __and J crold B. Warner, "The Distribution of Power Among Corporate Managers, Shareholders, and Directors" MW, Vol. 20, 1988, pp. 3-24. Kalbers, Lawrence P. and Timothy J. Fogarty, "Audit Committee Effectiveness. An Empirical Investigation of the Contribution of Power" W W, Vol. 12, No. 1, Spring 1993, pp. 24-49. Kaplan, Steven N. and David Reishus, "Outside Directorships and Corporate Performance," W, Vol 27, 1990, pp. 389-410. Karpoff, Jonathan M. and John R. Lott, "The Reputational Penalty Firms Bear From Committing Criminal Fraud." Wigs. Vol. XXXVI. forthcoming. Kennedy, Peter, W, The MIT Press, Cambridge, MA, 1989. Kesner, Idalenc F. , Bart Victor, and Bruce Lamont, "Board Composition and the Commission of Illegal Acts: An Investigation of Fortune 500 Companies, " AcademLQLManagemenLloumal. Vol. 29. No. 4. PP- 789-799. Kini, Omesh, William Kraeaw, and Shchmd Mian, "Corporate Takeovers, Firm Performance, and Board Composition, " Working Paper, February 1993. Klein, April and James Rosenfeld, "Targeted Share Repurchases and Top Management Change-t." WWW. Vol 20.1988 PP 493-506 Kosnik, Rita D., "Grecnmail: A Study of Board Performance m Corporate Governance," AW, Vol 32, 1987, pp. 163-185. _, "Effects of Board Demography and Directors’ Incentives on Corporate Greenmail Decisions" Wei. 1990. Vol. 33. PP- 129-150- KPMG Peat Marwick, MW, KPMG Peat Marwick, New York, 1993. Lee, Chun I. Stuart Rosenstein, Nanda Rangan, and Wallace N. Davidson III, "Board Composition and Shareholder Wealth: The Case of Management Buyouts," EinancialManazemcm. V0121. 1992 PP 53-72 135 Loebbecke, James K. and John J. Willingham, "Review of SEC Accounting and Auditing Enforcement Releases," Working Paper, January 1987. Loebbecke, James K. ,Martha M. Eining, and John J. Willingham, "Auditors’ Experience with Material Irregularities: Frequency, Nature, and Dctectability," W. Vol 9 No 1 Fall 1989 PP 1-28 MacAvoy, P., S. Cantor, J. Dana, and S. Peck, "ALI Proposals for Increases Control of the Corporation by the Board of Directors: An Economic Analysis," in Statement of the Business Roundtablc on the Ameriean Law Institute’s Proposed "Principles of Corporate Governance and Structure: Restatement and Recommendations," New York: Business Roundtable, 1983. Mace, Myles L., W» Harvard Business School Press, Boston, MA 1986. Maddala, G. S, "A Perspective on the Use of Limited-Dependent and Qualitative Variables Models In Accounting Research, W, Vol. 66, No. 4, October 1991, pp. 788-807. Manski, Charles F. and Daniel McFadden, "Alternative Estimators and Sample Design For Discrete Choice Analysis," C. F. Manski and D. McFadden, eds, 8mm Cambridge, MA, 1981. Mayers, David, Anil Shivdasani, and Clifford W. Smith, Jr., "Board Composition in the Life Insurance Industry," Working Paper, 1994 Merchant, Kenneth A. . ' ' ' - . - Wye, Financial Executives Research Foundation, Morristown, NJ, 1987. Morck, Randall, Andrei Shleifer, and Robert W. Vishny, "Management Ownership and Market Valuation: An ErnPirioal AnalysiS."Ioumal_of_F1nansial_Economiss. Vol. 20, 1988, pp. 293-315. National Commission on Fraudulent Financial Reporting, Wham CommissionmiraudulenLEirmciaLBemmM October 1987 Palepu, Krishna, "Predicting Takeover Targets. A Methodological and Empirical Analysis} WW. V018. March 1986 PP 3-35 Palmrose, Zoe-Vanna, "Litigation and Independent Auditors: The Role of Business Failures and Management Fraud." W. Spring 1987, pp. 90-103. 136 Patton, A. and J. Baker, "Why Do Not Directors Rock the Boat?," Waitress m, Vol. 65, pp. 10-12. Pincus, Karen V, "The Efficacy of a Red Flags Questionnaire for Assessing the Possibility of Fraud." WWW. Vol 14 Nos 1/2 1989, pp. 153-163. .Mark Rusbarsky, and J rlnaught Wong, "Voluntary Formation of Corporate Audit Committees Among NASDAQ Firms," MAW Peliey, Vol 8, 1989, pp. 239-265. Romney, Marshall B., Steve W. Albrecht, and David J. Cherrington, "Auditors and the Detection of Fraud," W, May 1980, pp. 63-69. Rosenstein, Stuart and J cffrcy G. Wyatt, "Outside Directors, Board Independence, and Shareholder Wealth," W, Vol. 26, 1990, pp. 175- 191. Shivdasani, Anil, "Board Composition, Ownership Structure, and Hostile Takeovers," Wigs. Vol 16. 1993. PP 167-198 Shleifer, Andrei and Robert W. Vishny, "Large Shareholders and Corporate Control, " Wm. Vol 95. 1986. PP 461-488 Singh, H. and F. Harianto, "Management-Board Relationships, Takeover Risk, and the Adoption of Golden Parachutes," AendemyeflMmagememleumal, March 1989, pp. 7-24. Sommer, A. A. Jr. "Auditing Audit Committees: An Educational Opportunity For Auditors," Aeeeunfingfiefim, June 1991, pp. 91-93. Sorenson, James E. and Thomas L. Sorenson, "Detecting Management Fraud: Some Organizational Strategies for the Independent Auditor," W W, Petrocclli Books, Inc., New York, 1980. Stone, Mary and John Rasp, "Tradeoffs m the Choice Between Logit and LOS for Accounting Choice Studies," W, Vol.66 N,o. 1, January 1991, pp. 170-187. Vancil, R, -. ’ : '_ - '. I, Boston: Harvard Business School Press, 1987. Vicknair, David, Kent Hickman, and Kay C. Carnes, "A Note on Audit Committee Independence: Evidence from the NYSE on "Grey" Area Directors," Aecmmfing Begum, Vol. 7, No. 1, March 1993, pp. 53-57. 137 Walkling, Ralph A. and Michael S. Long, "Agency Theory, Managerial Welfare, and Takeover Bid Resistance," WW, Vol. 15, pp. 54-68. Warner, Jerold B., Ross L. Watts, and Karen H. Wruck, "Stock Prices and Top Management ChangeS." IonmalnfEinanciaLEconomics. Vol 20. 1988. PP 461- 492. Weisbach, Michael 8., "Outside Directors and CEO Turnover," W Bunnies, Vol. 20, 1988, pp. 431-460. Weston, J .F., "The Tender Takeover," We, 1979, pp.74-82. Williams. H M andl S ShaPim. W W, New York, Camegie-Mellon University Press, 1979. Williamson, Oliver E. , "Organization Form, Residual Claimants, and Corporate Control." WWW. Vol. 26. 1983. PP- 351. _, "Corporate Governance," mm, Vol. 93, 1984, pp. 1197-1230. Zmijewski, M.E. , "Methodological Issues Related to the Estimation of Financial Distress Prediction Models," MW, Vol. 22, 1984 Supplement, pp. 59-82.