Investor Perceptions about Financial Statement Fraud and Their Use of Red Flags

Joseph F. Brazel Department of Accounting College of Management North Carolina State University Campus Box 8113 Nelson Hall Raleigh, NC 27695 919-513-1772 joe_brazel@ncsu.edu

Keith L. Jones* Department of Accounting George Mason University Enterprise Hall, MSN 5F4 Fairfax, VA 22030-4444 703-993-4819 kjonesm@gmu.edu

Rick C. Warne Department of Accounting George Mason University Enterprise Hall, MSN 5F4 Fairfax, VA 22030-4444 703-993-1763 rwarne@gmu.edu

September 2010

We are grateful for the helpful comments provided by Jagadison Aier, Tina Carpenter, Brooke Elliott, Frank Hodge, Molly Mercer, Jason Smith, Steve Smith, and the Office of the Chief Accountant of the Securities and Exchange Commission. This study was funded by a research grant from the Financial Industry Regulatory Authority Investor Education Foundation. * Corresponding Author

Electronic copy available at: http://ssrn.com/abstract=1460820

Investor Perceptions about Financial Statement Fraud and Their Use of Red Flags ABSTRACT We draw upon literature in psychology to develop a model that examines nonprofessional investors use of financial statement information, their perception of the frequency of fraud occurrence, the importance they place on fraud risk assessment, and ultimately, their use of fraud red flags. To test our model, we administered a survey to 194 nonprofessional investors. Our model predicts and we find that the positive relation between investor reliance on financial statement information and the importance of fraud risk assessment becomes stronger as investor perceptions of the rate of fraud increase (i.e., when investors are more likely to question the validity of financial statement information). Investors who perceive fraud risk assessment as an important activity appear to act on these perceptions. We find a positive association between the importance of making fraud risk assessments and investors use of fraud red flags (e.g., analyses of accruals, management turnover) when making investment decisions. With respect to the importance of red flags documented in the literature, additional analysis reveals that investors tend to focus on SEC investigations, pending litigation, violations of debt covenants, and high management turnover. In contrast, investors rely less on company size, age of the company, the need for external financing, and the use of a non-Big 4 auditor. We also illustrate that investors rely more on analysts, regulators, external auditors, and audit committees to detect and report fraud. Investors expect less from low/mid-level employees, upper management, the media, and short-sellers. Finally, we provide initial empirical evidence that nonprofessional investors may be achieving higher market returns by assessing fraud risk and using red flags when investing. Keywords: fraud risk; fraud detection; investor; red flags Data availability: Contact the authors

Electronic copy available at: http://ssrn.com/abstract=1460820

I. INTRODUCTION Investors experience significant financial losses when fraud occurs at publicly-traded companies such as Enron and WorldCom. Glass Lewis & Co. (2005) report that investors lost nearly $900 billion in market capitalization from 1997 to 2004 due to high-profile frauds. According to a 2006 report by the North American Securities Administrators Association, investors lose $40 billion annually due to securities fraud (NASAA 2006). Some experts suggest that the rate of fraudulent financial reporting (hereafter, fraud) will likely increase during the current economic recession (Mintz 2009). Nonprofessional investors play an important role in the capital markets, owning approximately 34% of all corporate equity (Bogle 2005). Despite their potential exposure to fraud, little research to date has examined if and how nonprofessional investors evaluate fraud risk before making investment decisions. We conduct a comprehensive survey that asks nonprofessional investors to describe their perceptions about financial statement fraud and how they protect themselves against fraud. Our objective is to address and study the links between the following questions: To what extent do nonprofessional investors rely on financial statement information? What are their perceptions regarding how often financial statement fraud is perpetrated? How important is assessing the risk of fraud prior to investing? Do nonprofessional investors analyze red flags (e.g., high management turnover) to avoid fraudulent investments? Thus, the primary purpose of this study is to gain an understanding of nonprofessional investor (hereafter, investors) perceptions and actions with respect to fraud. Overall, our evidence provides a reference point for future academic research and establishes a link between prior behavioral research related to investor decision-making processes and archival research that has identified fraud red flags. We are not

aware of any prior research that investigates investors perceptions, judgments, and actions related to financial statement fraud. Prior research in psychology reveals that, as decision makers rely more heavily on an information source, they become increasingly concerned about the credibility of that source (Coleman and Irving 1997). With specific reference to accounting information and the potential for fraud, prior accounting research indicates that the credibility of information provided by management influences investors judgments (Jennings 1987; Mercer 2004). Thus, the more investors rely on accounting information, the more likely they are to perform due diligence by searching for evidence of fraud. However, prior research also finds that individuals tend to underweight the probability of a rare event when they derive through experience the probability that the event will occur. On the other hand, if information regarding the rare event is received via description, individuals tend to overweight the probability of the event (Hertwig et al. 2004). For example, a doctor will tend to underweight the probability of a side-effect occurring from the use of a particular drug because the doctor has likely experienced few, if any, instances of the side-effect. A patient with no experience with the drug will likely overweight the probability of the side-effect if the patient relies on news reports for descriptive data about the side-effect (Hertwig et al. 2004). Since descriptive data on the probability of fraud is scarce, investors are likely to rely on their own direct and indirect experiences with fraud. With respect to fraud experiences, investors can be divided into two segments: (1) those that have experienced fraud in their investment portfolio (or have experienced fraud through close contacts) and (2) those that have not experienced fraud. Material financial statement fraud is a rare event.1 Thus, a large portion of

While the actual rate of fraudulent financial reporting is unknown, research related to financial statement auditors and fraud detection suggests that the rate is very low (e.g., Loebbecke et al. 1989; Nieschwietz et al. 2000).

investors have not directly or indirectly experienced financial statement fraud and are likely to underweight the probability that fraud will occur. Consequently, regardless of the extent to which they rely on financial statement information, these investors will not foresee a great need to assess fraud risk when investing. However, for investors who perceive a high rate of fraud in the market (likely caused by direct or indirect experiences with fraud), we do expect to observe a positive relation between financial statement reliance and the importance placed on fraud risk assessment. Therefore, our theoretical model predicts that as investors rely more on financial statements to make investment decisions, they perceive fraud risk assessment to be a more important investment activity. However, this positive association is driven by investors who perceive that a higher rate of fraud exists. We then examine if investors who perceive fraud risk assessment as more important act on these perceptions by using red flags to avoid potentially fraudulent investments. To test our model, we administered a survey to 194 nonprofessional investors. We prescreened participants to ensure that they had purchased or sold individual company stocks within the prior 12 months. Our sample consists of a geographically diverse group of active investors from all 50 states and the District of Columbia. Similar to Elliott et al. (2008), we employ the survey method because we examine independent variables that cannot be manipulated between investors (e.g., investors perceptions related to the rate of financial statement fraud) and dependent variables that are not publicly available (e.g., investors use of fraud red flags during investment). The model we examine is most effectively addressed using the survey method. We find that the positive relation between investors use of financial statement data and the importance of making fraud risk assessments is positively moderated by investors

perceptions of the current rate of fraudulent financial reporting. For investors, we also observe a positive relation between the importance placed on fraud risk assessment and the extent of red flag usage. While extensive research describes the usefulness of red flags to identify if fraudulent financial reporting exists (e.g., Beasley 1996; Lee et al. 1999, Brazel et al. 2009a), we are the first study to provide detailed empirical evidence on investors use of these red flags. In addition to the results related to our model, we report several key findings with respect to investor perceptions about fraud and their use of red flags. First, investors rely more on analysts, regulators, external auditors, and audit committees to detect and report fraud. In contrast, investors rely less on low/mid-level employees, upper management, the media, and short-sellers. We are not aware of any other study that compares the responsibility to detect fraud (at least perceived by investors) across such a wide spectrum of capital market participants. Second, investors tend to focus on the following red flags: SEC investigations, pending litigation, violations of debt covenants, and high management turnover. In contrast, investors rely less on company size, company age, the need for external financing, and the use of a non-Big 4 auditor. Third, a considerable amount of archival research has been devoted to understanding the negative relation between accruals and future market returns. Additional research has examined whether this link is driven by nonprofessional or professional investors (e.g., Sloan 1996; Ali et al. 2000). Our study offers a unique contribution to this research stream. We provide evidence that investors who specifically evaluate the accrual component of earnings (a red flag if abnormally large (Lee et al. 1999)) achieve higher market returns. We also illustrate that nonprofessional investors may be attaining higher returns by assessing fraud risk and using the most predictive red flags identified in prior research.

Despite numerous high-profile frauds in the past and the high cost of fraud to market participants, investors a priori decision processes vis--vis fraud are largely unknown. Levitt and Dubner (2005) posit that little is known about fraud-related issues because there is a paucity of good data. We offer descriptive evidence of, and key insights into, investors fraud-related perceptions, judgments, and actions. Our findings can guide and fuel future research into the decision processes of nonprofessional investors in relation to fraud. Our model and descriptive data can inform policymakers and help standard-setters implement financial reforms to better protect individual investors (e.g., Zweig 2009). Furthermore, the empirical evidence derived from our survey can help future researchers design appropriate experimental materials or collect relevant archival data to better understand investor behavior. The remainder of the paper is organized as follows. Section II describes the background and develops our hypotheses. Sections III and IV provide the methods and results of the study, respectively. Section V concludes the paper.

II. BACKGROUND AND DEVELOPMENT OF HYPOTHESES Standard setters have clearly indicated that nonprofessional investors should be a primary consideration when assessing the utility of financial statement information. Statement of Financial Accounting Concept No.1 asserts that financial reporting should provide information that can be used by all nonprofessionals as well as professionals who are willing to learn how to use it properly. Efforts may be needed to increase the understandability of financial information (FASB 2008, p. 11). Indeed, Congress (Public Law [107-204] 2002) and the Securities and Exchange Commission (Cox 2005) have explicitly stated their intent to protect nonprofessional investors. The Financial Industry Regulatory Authority (FINRA) states that one

of its primary objectives is to help investors build their financial knowledge and provide them with essential tools to better understand the markets and basic principles of saving and investing.2 In his State of the Union address, President Barack Obama noted, We need to make sure consumers and middle-class families have the information they need to make financial decisions (White House 2010). In short, the decision-making processes of nonprofessional investors matter to standard setters, The White House, Congress, the SEC, and investor protection groups. Given the red flags that accompanied high profile frauds at Enron and Bernard L. Madoff Investment Securities LLC, policy-makers should be particularly interested in ways investors can use red flags to protect themselves from fraud (Hubbard 2002; Markopolos 2010). However, little is known about if and how nonprofessional investors consider the risk of fraud or use red flags when making investment decisions. This dearth of research is troubling because fraudulent financial reporting at publicallytraded companies has a significant impact on investors.3 Enron investors lost a reported $60 billion (Vinod 2002), and trial testimony revealed that investors in WorldCom lost up to $200 billion (Rakoff 2003). The recent $50 billion fraud committed by Bernard Madoff indicates that investors continue to suffer serious consequences from financial statement fraud (Feiden and Zambito 2008). Without first understanding investors decision-making processes with respect to fraud, it is difficult for future researchers, policy makers, and investor protection groups to develop methods to protect investors. A baseline must be established to describe how investors

The FINRA is the largest independent regulator for all securities firms doing business in the United States. All told, FINRA oversees nearly 4,800 brokerage firms, about 170,400 branch offices and approximately 643,000 registered securities representatives. Created in July 2007 through the consolidation of NASD and the member regulation, enforcement and arbitration functions of the New York Stock Exchange, FINRA is dedicated to investor protection and market integrity through effective and efficient regulation and complementary compliance and technology-based services (see www.finra.org). The FINRA sponsored this research in an effort to improve investor protection with respect to fraud. 3 We use the term fraud in the context of fraudulent financial reporting as described in Statement on Auditing Standards No. 99 (AICPA 2002) as opposed to misappropriation of assets.

currently behave with respect to fraud. Then, future research can more effectively evaluate the potential impact of new policies to protect investors.4 Investor Perceptions about Financial Statement Fraud and their Use of Red Flags Prior research has shown that accounting information, including its presentation format, affects investors judgments and investment decisions (Maines and McDaniel 2000; Elliott 2006; Warne 2010). Recent research has provided insights into investors judgment processes in relation to accounting information sources (Elliott et al. 2008), comprehensive income (Maines and McDaniel 2000), pro-forma earnings (Elliott 2006), fair-market valuations (Warne 2010), and auditor opinions over internal controls (Smith 2010). This research stream provides interesting insights into the relation between accounting information and the investment decisions of nonprofessional investors. However, research has yet to examine how investors address the possibility that the accounting information may intentionally contain material misstatements (i.e., be fraudulent). Though investors may benefit from considering whether fraud exists in a company before making an investment decision, research to date has not examined if or how investors deliberately perform such fraud-related activities. We draw upon psychology research and propose a model of investor decision processes with respect to fraud. Figure 1 provides an illustration of our proposed model. We discuss the relations in the model and develop hypotheses to test the model below. [Insert Figure 1] Investor Fraud Risk Assessments Research in psychology suggests that, unless investors have been victims of fraud (or have experienced fraud through a close contact); they are unlikely to perform fraud risk

In a complex world where people can be atypical in an infinite number of ways, Levitt and Dubner (2009, 14) note the great value in discovering the baseline.

assessments. Individuals have difficulty assessing small-probability outcomes (i.e., rare events) such as the occurrence of a major fraud. Prior research indicates that individuals either underweight or overweight the probability of a rare event depending on how they learn about the likelihood of the event (Hertwig et al. 2004; Erev and Haruvy 2010). Individuals learn about the likelihood of rare events from description or from experience. For example, when individuals have access to a weather forecast they assess the probability of a hurricane from description. However, when individuals decide whether to back-up their hard drive, they assess the risk that their hard drive will be lost from experience because it is unlikely that they possess descriptive data about the probability of a lost hard drive. In the case of decision-making from description, individuals tend to overweight the probability of a rare event, which is consistent with Prospect Theory (Hertwig et al. 2004). In the case of decision-making from experience, individuals tend to underweight the probability of a rare event due to the likelihood that they have not experienced the rare event (Friedman and Massaro 1998; Hertwig et al. 2004; Erev and Haruvy 2010). For example, underweighting the probability of a national collapse in the real estate market would account for mortgage underwriters originating risky loans and banks investing so heavily in mortgage-backed securities. Historical data about mortgage default rates was scarce which led banks and underwriters to rely on experience (Tett 2009). Even when speculation of a real estate bubble began to surface, banks still continued to invest heavily in mortgage-backed securities (Tett 2009). Since there had not been a national collapse in the real estate market since the 1930s, banks and others appeared to underweight the probability of this rare event.5

It is important to note that historical data on mortgage defaults was available for certain regions of the country, and certain regions had experienced drops in real estate values over extended periods of time. However, national data was limited. In addition, banks incorrectly assumed that there was very little correlation between home prices across the country. For example, it was assumed that real estate values in Las Vegas would have little, if any, correlation

Investors normally make decisions with respect to financial statement fraud by experience since the true likelihood of fraud is generally unknown and descriptive data is unavailable. Given that fraud is a rare event, it can be assumed that the majority of investors have not experienced fraud either directly in their own portfolio or indirectly through a close contact. If investors have not experienced investing in a fraudulent company, then they will likely underweight the probability that an individual investment is fraudulent. In addition, nonprofessional investors may simply rely on others who are more qualified (e.g. regulators, auditor) to assess and detect fraud.6 Therefore, it is possible that investors generally underweight the risk of fraud and place little importance on assessing fraud risk themselves. On the other hand, prior research in psychology also indicates that source credibility is an important consideration for decision makers (Coleman and Irving 1997). Individuals examine the source of information when determining whether to rely on that information. When information comes from an outside source, decision makers search for knowledge and reporting biases (Eagly et al. 1978). Specifically, in ambiguous situations when the decision is important, source credibility significantly influences individuals judgments (Chaiken and Maheswaran 1994).7 In a capital market context, investment decisions involve uncertainty and financial statement information from an outside party (the company and its management). Thus, these findings suggest that source credibility and fraud risk assessment would be important to investors who rely heavily on financial information. Consequently, despite investors underweighting the
with real estate values in Miami. Thus, banks felt they could diversify the risk of a drop in real estate values in one region by investing in portfolios of mortgages from across the country. With limited data and no recent history of a national collapse in real estate, banks apparently underweighted the possibility that the entire country could experience a simultaneous decline in real estate values. See Tett (2009) for a more thorough description of these events and how banks began to rely on experience, versus descriptive data, when making investments in mortgagebacked securities. 6 We measure and control for investor perceptions that, regulators, auditors, etc. will detect fraud (see RELIANCE ON OTHERS (item 15) in Table 2). 7 In an accounting context, research has shown that auditors attend to the reliability of a source when making judgments (e.g., Hirst 1994).

probability of fraud, we posit that a positive association could exist between the extent to which an investor relies on financial statement data and the importance of making fraud risk assessments (see Figure 1). However, as described above, we hypothesize that aforementioned relation between financial statement reliance and fraud risk assessment is influenced by an investors perception of the rate of fraudulent financial reporting. For investors who rely heavily on financial statements, but through experiences perceive the rate of fraud to be low, it is unlikely that they will perceive fraud risk assessment as an important activity. On the other hand, an investor in Enron or whose acquaintances held WorldCom stock would not consider fraud to be a rare event. Investors who through experience consider fraud to be a more common event will be (1) less likely to underweight the probability of its occurrence and (2) more likely to protect themselves by assessing the risk of fraud. We therefore predict that the positive relation between investor reliance on financial statement data and the importance they place on making fraud risk assessments will be positively moderated by investor perceptions of the rate of fraud. We formally state Hypothesis One as follows: H1: The positive relation between investor reliance on financial statement information and the importance of fraud risk assessment becomes stronger as investor perceptions of the rate of fraud increase.

Investor Use of Red Flags Prior accounting research has examined firms that engaged in financial statement fraud and documented the characteristics of fraud firms. Hogan et al. (2008) and Dechow et al. (2010) summarize the red flags that fraud firms typically exhibit. Notably absent from the literature is any evidence regarding investors use of fraud red flags prior to making investment

10

decisions. In fact, whether investors actually use red flags, regardless of the importance investors place on making fraud risk assessments, is an unanswered research question. We expect that as investors perceive fraud risk assessment as a more important investment activity, they should have greater motivation to engage in behaviors designed to avoid investments in fraudulent companies and the losses that follow. A primary and logical way to avoid fraudulent investments is to evaluate red flags previously shown to discriminate fraud firms from non-fraud firms (e.g., boards of directors with high proportions of insiders (Beasley 1996)). Thus, when investors place more value on fraud risk assessment, they are more likely to act on this perception by analyzing red flags prior to investment.8 We formally state Hypothesis Two as follows: H2: The importance of fraud risk assessment is positively associated with investor use of red flags.

While we expect the aforementioned positive relation, we acknowledge that variation in the use of fraud red flags (i.e., investor action) may not occur. Just as a managers strategy does not necessarily lead to actions or successful implementation (Auer and Reponen 1995), investors opinions regarding fraud risk assessments may not lead to their use of fraud red flags for a variety of reasons. First, an investor may not have the domain-specific knowledge to
8

The intuition behind our second hypothesis follows the Elaboration Likelihood Model (Petty and Cacioppo 1986). This theory suggests that individuals who are highly motivated and have the ability to scrutinize issue-relevant arguments follow the central route route of persuasion. Motivated individuals exhibit a willingness to expend cognitive effort in evaluating the merits and attributes of a persuasive message. Motivation is directly correlated with personal relevance. On the other hand, individuals follow the peripheral route when their motivation (and/or ability) is relatively low. Under the peripheral route, attitudes and opinions are formed by simple positive and negative cues due to an absence of significant cognitive effort. In the context of financial statement fraud, an investor who believes fraud assessment is very important (due to prior experiences with fraud) will likely spend significant cognitive effort collecting and evaluating a wide range of red flags. However, an investor who believes fraud is extremely rare will likely underweight its likelihood and will expend little cognitive effort. Therefore, his/her assessment of fraud will be based on peripheral cues. For example, if an investor with little motivation were asked to assess fraud risk at Company X, the investor might think, the CEO seems like a very trust-worthy individual and their products are very popular and innovative, thus it seems unlikely Company X is committing fraud. While most investors might not admit to low cognitive effort with respect to fraud risk assessment, we do expect that investors who place greater emphasis on assessing fraud risk (i.e., greater motivation) will also be more likely to use red flags.

11

identify and interpret relevant red flags from information sources (cf. Anderson 1982). Second, since decision makers often disregard information inconsistent with their prior expectations (Fischer et al. 2008), some investors may attend to confirming/positive information rather than utilizing red flags that may call into question the validity of financial statements. Third, because investors have access to a large quantity of publicly-available information, such quantities may exceed an investors processing abilities or overshadow red flag data (e.g., Kahneman 1973; Hasher and Zacks 1979). Fourth, investors who perceive fraud risk assessment to be important may not assess fraud risk via red flags, and instead choose to use some other form of information to assess fraud risk. Finally, the possibility exists that investors utilize red flags equally regardless of how important they perceive fraud risk assessment. Thus, whether investors who perceive fraud risk assessment to be important actually act on this perception by using red flags is an open empirical question.9

III. METHOD Sample Similar to Elliott et al. (2008), we employ the survey method to test our hypotheses because we examine independent variables that cannot be manipulated between investors (e.g., investors perceptions related to the rate of financial statement fraud) and dependent variables that are not publicly available (e.g., investors use of fraud red flags during investment). Thus,

By way of comparison, auditors are required to assess fraud risk and act on their fraud risk assessments by altering their audit testing, further analyzing red flags, etc. Unlike investors, auditors have a responsibility to detect material fraud; they receive training and guidance related to fraud; and they are likely to have had more than one direct experience related to fraud (AICPA 2002; Brazel et al. 2010). However, despite these advantages, prior research has found that auditors have difficulty acting on their fraud risk assessments (e.g., Zimbelman 1997; Asare and Wright 2004). In sum, auditors, who are at an advantage with respect to fraud, have difficulty with fraud-related analyses/testing. Therefore, nonprofessional investors, who are less knowledgeable than auditors, may perceive fraud risk assessment as important, but fail to act on this perception by analyzing red flags.

12

the survey method is the most effective means of addressing our hypotheses. In addition, given the paucity of research related to investors consideration of fraud, the survey method allows us to examine multiple relations simultaneously (see Figure 1). Thus, we are able to shed light on multiple forms of investor perceptions, judgments, and actions in a single study. The survey method also allows for the collection and analysis of rich descriptive data that can serve as a starting point for future empirical research in this important domain. Our survey results can help future researchers design appropriate experimental materials or collect relevant archival data to better understand investor behavior. One hundred ninety-four nonprofessional investors completed an online survey for this study. The survey was titled Survey on Investor Beliefs and we collected our fraud -related data along with obtaining responses on a number of non-fraud-related topics. This variety in our survey, which concealed the purpose of our study, is illustrated by the large number of nonfraud-related control variables that we collected and discuss below. Greenfield Online (http://www.greenfield-ciaosurveys.com) distributed the survey. For the purposes of our survey, Greenfield screened their database for participants that actively traded individual shares of stock (vs. simply investing in a mutual fund). We further screened participants by requiring that they answer yes to the following statement in order to complete the survey: I have bought or sold individual company stock in the last 12 months. Greenfield distributed the survey to 1,178 participants. Thus, our response rate is 16.5%, which is comparatively high given the response rates of previous investor surveys (e.g., the response rate for Elliott et al. [2008] was approximately 3%). Participants completed the survey from August 21 - 25 of 2008. Our data collection occurred prior to the current economic recession.10,11

10

Because not all individuals responded to our survey, we examined the potential for non-response bias. Oppenheim (1992) recommends comparing data from late respondents to early respondents as a way of assessing this bias.

13

Participants were residents from all 50 states and Washington DC, approximately 50% male, well educated (75% had a bachelors degree or higher), used a wide variety of investment strategies, were, on average, between 40-49 years old, and had an average of 6-10 years of investing experience. These demographic data, as well as other control variables, will be further discussed in our review of descriptive statistics below. Given that researchers commonly use MBA students to proxy for individual nonprofessional investors, we obtained a relatively diverse and experienced sample of active investors. Regression Models In Hypothesis One we posit that the positive relation between investor reliance on financial statement information and the importance of fraud risk assessment becomes stronger as investor perceptions of the rate of fraud increase. To test Hypothesis One, we estimate the following model via ordinal logistical regression: FR = + 1RELIANCE ON FINANCIAL + 2PERCEPTION OF FRAUD + 3RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD + 4-37CONTROL VARIABLES + (Model 1)
0

Our first hypothesis is supported if the interaction term RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD ( 3) is positive and significant. Hypothesis Two predicts that the
Accordingly, we compared the responses from the first quartile of respondents to those of the last quartile of respondents. There were no statistically significant differences between early and late responses on any of our hypothesized variables. In addition, we asked Greenfield to screen their database to ensure that all respondents were nonprofessional investors. To determine if all respondents were indeed nonprofessional investors, we asked respondents to supply their occupation. One respondent noted that they worked for a stock broker and potentially could be a professional investor. On the other hand, the respondent could have worked in a nonprofessional investor capacity for the broker (e.g., worked in the accounting or marketing departments). Our results are qualitatively the same if we remove this participant from our analyses.
11

The CBOE Volatility Index (VIX) is a key measure of market expectations of near-term volatility conveyed by S&P 500 stock index option prices. Since its introduction in 1993, the VIX has been considered by many to be the world's premier barometer of investor sentiment and market volatility (see http://www.cboe.com/micro/vix/vixwhite.pdf). In short, higher indices are indicative of greater market fear. During the period our data was collected the highest measure of the index was 21.22, whereas in mid-September 2008 the index rose above 30 and did not fall below 30 until June 1, 2009 (see Lauricella 2009 and http://www.cboe.com/micro/vix/historical.aspx). As of August 26, 2010, the VIX was 26.17 and the 52-week range for the index was 15.23 - 48.20 (http://finance.yahoo.com/q/bc?s=%5EVIX&t=2y).

14

importance of the fraud risk assessment is positively associated with investor use of red flags. To test Hypothesis Two, we estimate the following model via ordinary least squares regression: USE OF RED FLAGS =
0

1FR

2-35CONTROL

VARIABLES + (Model 2)

Hypothesis Two is supported if the variable FR ( 1) is positive and significant. We define our variables as follows: FR = Importance of fraud risk assessment, relative to other factors, when making buy/sell decisions for stock that you currently hold, measured on a scale where 1 = not at all important and 7 = extremely important. = Mean reliance on direct financial statement information / Mean reliance on non-financial statement information (further described below and in Table 1). = In your opinion, how often do managers of publiclytraded companies commit financial statement fraud, measured on a scale where 1 = 0% of the time and 11 = 100% of the time. = Mean use of red flags (further described below and in Table 1).

RELIANCE ON FINANCIAL

PERCEPTION OF FRAUD

USE OF RED FLAGS

The control variables are defined in Appendix A. Hypothesized Variables Financial Statement Reliance Descriptive statistics for our hypothesized variables are presented in Table 1. We present mean responses and standard deviations for our measures of direct financial statement reliance (items 1-6), and the same data for reliance on non-financial statement information (items 8-14). All of these items were measured via a scale where 1 = very unimportant and 7 = very important. While Hodge and Pronk (2006) and Elliott et al. (2008) provide evidence that nonprofessional investors do indeed rely on financial and nonfinancial information when investing, we contribute to this literature stream by providing detail with respects to the types of financial and nonfinancial information these investors are likely and unlikely to use.

15

[Insert Table 1] The means for reliance on financial statement information (item 7) and reliance on nonfinancial statement information (item 15) were 5.13 and 4.69, respectively. Consistent with Elliott at al. (2008), we calculate a relative measure of financial statement reliance (item 16) for each participant by dividing item 7 by item 15. This measurement is termed RELIANCE ON FINANCIAL and is an independent variable in Model 1 above. With respect to the importance of financial statement information, investors appear to rely more on balance sheet data (see statistical test in footnote b, Table 1) and rely less on the footnotes to the financial statements (footnote d, Table 1). In relation to non-financial statement information, investors seem to rely more on stock price information and advice from professionals (footnote f, Table 1) and rely less on non-financial information related to operations, advice from the media, and advice from nonprofessionals (footnote h, Table 1).12 Finally, from non-tabulated analyses of frequency distributions, the majority of investors deem the following information sources important (i.e., rated the information source as 5, 6, or 7 on our scale): stock price information (77.4%), the balance sheet (73.8% of respondents), cash flow statement (73.3%), income statement (72.4%), and advice from professionals (69.5%) Perception of the Rate of Fraud and the Importance of Fraud Risk Assessment Table 1 also provides descriptive statistics for our variables PERCEPTION OF FRAUD (item 17) and FR (item 18). We asked participants, In your opinion, how often do managers of publicly-traded companies commit financial statement fraud? Participants responded on a scale, 1 = 0% of the time and 11 = 100% of the time. The mean response was 5.06, indicating that investors perceive fraudulent financial reporting as an issue (approximately a 40% rate). This

12

Our results are qualitatively the same if we employ item 7 from Table 1 (mean reliance on financial statement information) in our analyses instead of RELIANCE ON FINANCIAL (item 16 from Table 1).

16

rather high rate may result from the fact that approximately 25% of our sample reported owning shares of a company when it was found to have committed financial statement fraud (see item 2 in Table 2).13 Thus, relating to our theory presented in Section II, a quarter of our sample has directly experienced fraud and are unlikely to underweight its likelihood. While fraud is the most extreme form of earnings management, this high perception of fraud also appears consistent with research by Graham et al. (2006) who find, in a survey of 401 senior financial executives, that the presence of earnings management is pervasive. One CFO in the study stated, You have to start with the premise that every company manages earnings (Graham et al. 2006, p. 30). Thus, investors may perceive earnings management activities by management to be fraudulent, yet understand/accept that such activities are commonplace in the market and not substantial enough to warrant considerable attention. Likewise, what constitutes a material fraud is likely lower for investors than those involved in the financial reporting process (Jennings et al. 1987), which would explain participants high perceptions of fraud in our study. Last, in an experimental study of nonprofessional investors, Brazel et al. (2009b) report that investors perceive the rate of fraudulent financial reporting to be a relatively high 35%. Given this perception of the rate of fraud, one would expect that fraud risk assessment, relative to other activities, would be fairly important to investors when making buy/sell investment decisions. Participants provided data on the importance of fraud risk assessment (FR) on a scale where 1 = not at all important and 7 = very important. The mean response was
13

Interestingly, in a non-tabulated analysis, we find PERCEPTION OF FRAUD to be positively, but not significantly, correlated with OWNED THE STOCK OF A FRAUD COMPANY (p = 0.11). Thus, as described in Section II, investors perceptions of the rate of fraud are likely affected by both their own direct experiences with losses due to fraud as well as indirect experiences with losses incurred by others around them. OWNED THE STOCK OF A FRAUD COMPANY is clearly a measure of investors own direct experiences with fraud. When we replace PERCEPTION OF FRAUD with OWNED THE STOCK OF A FRAUD COMPANY in our analysis, we do not observe support for Hypothesis One. Thus, it is important to consider both direct and indirect investor fraud experiences when predicting investor behavior with respect to fraud.

17

5.24. This mean response indicates that, despite fraud being a rare event, investors perceive fraud risk assessment to be a relatively important investment activity. As stated previously, our study is the first academic study to examine such investor perceptions related to fraud and to determine if they are associated with investor behavior (e.g., use of red flags). Use of Red Flags Prior research in the areas of fraud, restatements, and earnings management has created an extensive list of red flags that correlate with financial statement quality. For example, Dechow et al. (1996) find that companies in need of external financing may have incentives to manipulate revenues in anticipation of accessing the capital markets. Also, companies have incentives to manage earnings prior to an acquisition or merger in order to raise their stock price (Erickson and Wang 1999; Louis 2004). Beasley (1996) concludes that fraud firms tend to have a higher proportion of insiders (i.e., employees) on their boards of directors, and Efendi et al. (2007) find a link between equity-based compensation and fraudulent financial reporting. In short, there are a large number of potential red flags that prior empirical research has validated and investors might use to assess fraud risk and avoid investing in fraudulent companies. In Table 1, we provide data on the use of 21 red flags (items 19 39). For the use of each red flag, participants responded to a scale where 1 = never and 7 = often. For each participant, we calculate their mean use of red flags. We label this variable USE OF RED FLAGS, which becomes the dependent variable in Model 2 above.14 The mean use of red flags for our sample is 4.91 (item 40 in Table 1). With respect to the importance of red flags, we find that investors tend to focus on SEC investigations, pending litigation, violation of debt covenant, and high management

14

In all instances where we use the mean of multiple variables to create one variable (e.g., USE OF RED FLAGS), factor analyses were performed. Without exception, all factor analyses indicated that the items satisfactorily loaded in excess of .50 on one factor (Nunnally 1978). Also, all tests of measurement reliability provided Cronbachs alpha levels exceeding the generally accepted threshold of .70 (Nunnally 1978).

18

turnover (see statistical test in footnote l, Table 1). Investors rely less on the following red flags: age of the company, need for external financing, company size, and use of a non-Big 4 auditor (footnote n, Table 1). While extensive research has examined the usefulness of red flags to identify fraudulent financial reporting, we are the first to provide empirical evidence on investor use of these red flags. Control Variables As previously stated, during the course of their investment activities, investors are likely influenced by a host of factors/information that are (and are not) fraud-related (e.g., reliance on others to detect and report fraud, trading strategy). To ensure the reliability of our results, we control for numerous variables that could potentially affect investor perceptions related to fraud, their use of red flags, and their overall investment activities. We present descriptive statistics for these control variables in Table 2. These data also highlight the diversity and appropriateness of our sample. First, we document that 100% of our participants reported buying or selling individual company stock in the 12 months prior to our survey (item 1). Intuitively, one would expect that prior fraud experiences (item 2), the perception that losses due to fraud can be recovered (item 3), and reliance on other parties to detect and report fraud (items 4 - 14) could impact our hypothesized variables. As noted previously, 25% of our sample reported that they held shares of a company that committed financial statement fraud (item 2). While we use the mean RELIANCE ON OTHERS to detect and report fraud (item 15) as a control variable in our analyses, Table 2 illustrates that investors see various capital market participants as more and less responsible in this area. For each party, reliance was measured on a scale ranging from 1 = not at all to 7 = completely. We are aware of no other study that compares the responsibility to detect fraud (at least perceived by investors) across such a wide spectrum of capital market

19

players. Investors appear to rely more on regulators, external auditors, analysts, and audit committees to detect and report fraud (see statistical test in footnote d, Table 2).15 Investors expect less from upper management, low/mid-level employees, the media, and short-sellers (footnote f, Table 2). Overall, mean RELIANCE ON OTHERS to detect and report fraud was moderate (mean = 4.56, item 15). [Insert Table 2] Consistent with Elliott et al. (2008) and Barber and Odean (2001), we control for a number of variables related to participants investing experiences, activities, and returns (items 16 21). These data suggest that our sample consists of a diverse and experienced set of active investors. For example, we measured investing experience (item 16) by asking, How many years have you been actively buying/selling the stocks of individual companies (as opposed to mutual funds, etc.)? Participants responded on a scale where 1 = less than one year and 6 = more than 20 years. The mean response for our sample was 3.34 (i.e., between 6 10 years). Although not tabulated, 32 of our participants (16% of our sample) have actively invested for over 20 years. As suggested by prior research (e.g., Markowitz 1952; Elliott et al. 2008), we control for participant trading strategies/risk preferences (items 22 30) and, given that fraud may be more prevalent in certain industries (e.g., Dechow et al. 2010), we control for the industries in which participants invested most heavily (items 31 38). The results indicate that our sample of investors employed a diverse set of investment strategies and invested heavily in a wide array of industries. Finally, consistent with Bertaut (1998), Masters (1989), Barber and Odean (2001), and Elliott et al. (2008), we control for a host of demographic data (items 39 46). Our sample
15

Since our results suggest that investors rely heavily on auditors to detect fraud, we confirm the relevancy of research that investigates jurors verdicts when auditors fail to detect financial statement fraud (e.g., Lowe and Reckers 1994; Kadous 2001; Cornell et al. 2009).

20

appears to be well educated (items 39 42), approximately half male/female (item 44) and, on average, between 40 49 years old (item 45). Correlation Matrix A correlation matrix is presented in Table 3. To present a parsimonious correlation matrix, control variables were excluded from presentation in Table 3 if they were not significantly correlated (p < 0.05) with at least two of the four hypothesized variables (e.g., reliance on financial statement information, importance of fraud risk assessment). Reducing this constraint from two to one led to a substantially larger correlation matrix. Of particular note is that neither reliance on financial statement information (Reliance) nor the investors perception of the rate of fraudulent financial reporting (Rate) is significantly correlated with the importance of fraud risk assessment (FR). However, as Hypothesis One posits, these two variables may positively interact to affect FR. [Insert Table 3] Consistent with Hypothesis Two, FR is significantly, positively correlated with the use of red flags (Flags). We will test Hypotheses One and Two in multivariate settings in the next section. It is also interesting to note that reliance on financial statement information (Reliance) is not significantly correlated with the use of red flags (Flags). While one might expect a straightforward, positive relation between the two variables, it appears that this relation is more complex and may involve moderating and mediating variables as described by our model (see Figure 1). In the next section we analyze the moderating and mediating links depicted by our model. With respect to multicollinearity, all reported regression analyses of our models provide variance inflation factors for all of our variables that are less than 3.00 and substantially below the standard threshold of 10 (e.g., Neter et al. 1996; Kennedy 1998).

21

IV. RESULTS Hypothesis One Testing Table 4 provides the results of Hypothesis One testing. Hypothesis One predicts that the positive relation between investor reliance on financial statement information and the importance of fraud risk assessment becomes stronger as investor perceptions of the rate of fraud increase. Thus, Hypothesis One would be supported by regression results that provide a positive and significant interaction between RELIANCE ON FINANCIAL and the PERCPETION OF FRAUD on FR. For presentation purposes, only effects related to our control variables with pvalues < 0.10 are presented in our tables related to hypotheses testing. As predicted by Hypothesis One, we find a significant and positive interaction between RELIANCE ON FINANCIAL and PERCEPTION OF FRAUD (A X B in Table 4) on FR (p < 0.05). Thus, our results provide support for Hypothesis One. [Insert Table 4] To illustrate the form of this interactive effect, we partition the sample into two groups: high and low PERCEPTION OF FRAUD. We partition the sample at the median PERCEPTION OF FRAUD (5.00) and delete observations at the median. We then re-perform the regression analysis described above for each of the two groups (removing the interaction term). For the high PERCEPTION OF FRAUD group, the relation between RELIANCE ON FINANCIAL and FR is significant and positive (p = 0.06). For the low PERCEPTION OF FRAUD group, the relation between RELIANCE ON FINANCIAL and FR is not significant (p = 0.18). Thus, the form of the interaction presented in Table 4 is consistent with the form of the interaction posited by Hypothesis One.

22

Consistent with our theory of investors typically underweighting the likelihood of fraud, we observe a positive but not significant (p = .66) association between RELIANCE ON FINANCIAL and FR. Because many investors likely underweight the probability of fraud, as investors rely more heavily on financial statement data, they do not simply perceive fraud risk assessment to be more important. It is when they rely on financial statements and perceive fraud to be a more likely event (i.e., PERCEPTION OF FRAUD is higher) that we observe the positive relation between RELIANCE ON FINANCIAL and FR. Hypothesis Two Testing Hypothesis Two predicts a positive relation between the importance of fraud risk assessment and investor use of red flags. Table 5 presents the results of our Hypothesis Two testing. We find support for Hypothesis Two as the relation between FR and USE OF RED FLAGS is positive and significant (p < 0.01). [Insert Table 5] Discussion of Control Variables With respect to the direct effects of control variables in Tables 4 and 5, several observed relations contribute to our understanding of fraud-related investor judgments and decisions and should spur future research. First, we note that investors who most often invest in the financial services and manufacturing/energy industries are most likely to consider fraud risk assessment important and use red flags, respectively. Given the recent crisis in the financial services market, investors in this industry are likely even more concerned with fraud risks today. Whether red flags are more transparent/easier to analyze in the manufacturing and energy industries, and potentially less so in other industries, is a fruitful area of research.16 Second, it is interesting to

16

For example, Enrons financial statements from 1995 include multiple red flags related to the accuracy of its statements (Hubbard 2002).

23

note the positive relation between investor returns and fraud risk assessment. In analyses that follow, we provide evidence of links between investor returns and the use of specific red flags (e.g., accrual levels). Third, we observe a counter-intuitive negative relation between the value of the investors portfolio and the use of red flags. Why investors, with potentially more to lose due to fraud, are less likely to analyze fraud red flags is an interesting question for future research. Fourth, we observe that investors with higher levels of education are more likely to use red flags. Perhaps the training received or the complex analytical skills required at higher levels of education are necessary to collect and analyze red flags. How investors can use both their general and domain-specific knowledgebases to effectively use fraud red flags could be addressed by future research. Last, the significant relation between gender and fraud risk assessment suggests that male investors perceive fraud risk assessment to be a more important activity than female investors. Future research can shed light on why this relation exists (e.g., male investors are currently more risk adverse vis--vis fraud than female investors, recent publicized frauds have been largely perpetrated by men and male investors may be more sensitive to fraud incentives faced by managers). Mediation Analysis of the Overall Model As Figure 1 illustrates, the importance of fraud risk assessment (FR) should mediate the interactive effect of RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD on the USE OF RED FLAGS. Following Baron and Kenney (1986), we conduct a mediation analysis to determine the validity of the model presented in Figure 1. Statistical evidence of FR mediating the relation between RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD on the USE OF RED FLAGS first requires that the interaction significantly affect the USE OF RED FLAGS. In non-tabulated regression analysis, controlling

24

for all variables described above with the exception of FR, we find the interaction positive and statistically significant (p = 0.03). Second, the interaction must affect FR. Our test of Hypothesis One finds this relation to be positive and significant. Third, FR must also be significantly correlated with the USE OF RED FLAGS. Our test of Hypothesis Two finds this relation to be positive and significant. Lastly, when both FR and the interaction are included in Model 2: (1) FR must be significant; and (2) the interaction term must either be insignificant (full mediation) or its significance must decline (partial mediation). Non-tabulated regression results find (1) FR is positive and significant (p < 0.01); and (2) the significance level for the interaction has dropped from the previously described level (p = 0.03) to a lower level of significance (p = 0.08). These results point to FR partially mediating the interactive effect of RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD on the USE OF RED FLAGS. Specifically, investors that rely more on financial statements perceive that fraud risk assessment is a more important investing activity. However, this relation is driven by the rate at which investors believe that fraud occurs in the capital markets. In turn, investors who perceive fraud risk assessment to be an important part of investing appear to act on these perceptions. They are more likely to use red flags to avoid investing in companies that might be committing financial statement fraud. Use of Accruals and Investor Portfolio Returns Sloan (1996) finds investors fixate on earnings and have difficulty distinguishing between earnings derived from cash flows and earnings derived from accruals. Consequently, Sloan finds a negative association between accruals and future abnormal stock returns. Ali et al. (2000) posit that the negative association between accruals and future abnormal returns is due to

25

earnings fixation by nave or nonprofessional investors. Contrary to their expectation, Ali et al. (2000) find that the negative association between accruals and stock returns is stronger for larger firms which are more likely to be followed by analysts and held by institutions, and weaker for smaller firms which are less likely to be of interest to these sophisticated market participants. This counter-intuitive result suggests that any failure to appreciate the valuation implications of accruals is more pronounced for sophisticated investors than for nonprofessional investors. Given our dataset related to nonprofessional investors, we are in a unique position to add to this research stream. Specifically, we have a measure of investor usage of accrual data (item 36, Table 1) and their twelve month return on their personal investment portfolios (item 21, Table 2). Similarly, Elliott et al. (2008) use survey data to study variables associated with investor returns (e.g., types of information used, experience levels). In a non-tabulated regression controlling for the variables used in Elliott et al. (2008), we find the relation between the consideration of accruals by nonprofessional investors and their market returns to be positive and significant (p = 0.02).17 As illustrated in Table 4, we also observe a positive and significant association between FR and investor returns. Thus, we are able to provide initial empirical evidence that nonprofessional investors may be benefitting from using accrual data and assessing fraud risk when investing. Use of Ex-Ante Red Flags and Investor Portfolio Returns Table 1 illustrates that the red flags investors report using most often are usually manifest later in the fraud discovery process and are usually revealed to investors ex-post (i.e., after the

17

We do not include the variable training from Elliott et al. (2008) for two reasons. First, it was specific to trainings provided by the investment club from which their sample was derived. Second, it was not statistically significant in their analysis.

26

fraud has been detected and reported publicly).18 Consequently, it is questionable as to whether investors attention and reaction to such ex-post red flags would (a) reduce the likelihood that they experience losses due to fraud, and (b) increase their portfolio returns. Table 1 also documents that, in practice, there are several red flags that investors are less likely to use (e.g., need for external financing, use of a non-Big 4 auditor, see footnote n). Given this understanding, we explore whether ex-ante fraud indicators (e.g., large difference between revenue growth and non-financial measure growth; auditor change), that investors report to typically use, can help investors more effectively screen investments, avoid financial losses related to fraud, and achieve higher portfolio returns. We develop a measure of the use of ex-ante red flags for each participant (averaging their responses to items 22 35 from Table 1). Similar to above, we include this measure in the model used by Elliott et al. (2008). In non-tabulated analyses, we find the relation between the use of ex-ante red flags by nonprofessional investors and their portfolio returns to be positive and significant (p < 0.01). As such, we provide evidence of an association between the use of ex-ante red flags and the achievement of higher investment returns. V. CONCLUSION This paper models nonprofessional investors use of financial statement information, their perception of the frequency of fraud, the importance of assessing fraud risk, and their use of fraud red flags. Investors are often victims of fraudulent financial reporting; however, very little prior research investigates investors perception of fraud and how investors protect themselves from fraudulent financial reporting. To shed light on these issues, we administered a survey to

18

In Table 1, investors report to use four fraud red flags relatively more often than other red flags: SEC investigations, pending litigation, violations of debt covenants, and high management turnover (see footnote l). Though high management turnover may occur before or after the market discovers the fraud, the other three red flags can be considered ex-post fraud indicators.

27

194 nonprofessional investors. We find that the positive relation between investor reliance on financial statement information and the importance of fraud risk assessment becomes stronger as investor perceptions of the rate of fraud increase. We also consider whether investors act on their fraud risk assessments by utilizing various fraud red flags in an effort to avoid potentially fraudulent investments. We find a positive association between the importance of making fraud risk assessments and investors use of fraud red flags when making investment decisions. With respect to the importance of red flags, our additional analysis reveals that investors tend to focus on SEC investigations, pending litigation, violations of debt covenants, and high management turnover. In contrast, investors seem to rely less on company size, age of the company, the need for external financing, and the use of a non-Big 4 auditor. Additionally, we illustrate that investors appear to rely more on analysts, regulators, external auditors, and audit committees to detect and report fraud. Investors rely less on low/mid-level employees, upper management, the media, and short-sellers. Finally, we provide initial empirical evidence that nonprofessional investors who are more apt to consider reed flags (accruals in particular; and exante red flags in general) achieve higher market returns. Our model and our comprehensive set of control variables provide much detail into the factors that investors consider important when making investment decisions vis--vis fraud. As current policymakers have become increasingly concerned with the behavioral aspects of the market (e.g., Zweig 2009), our descriptive results and model should inform future policies aimed to protect investors from fraud. Our results also provide an important first step in examining how investors both fall prey to, and avoid investments in, fraud firms. The results may help policymakers as they determine what types of company information (i.e., red flag-related data) should be made readily available or transparent to investors.

28

Future research can use our descriptive results to develop more refined models related to the sources of information that investors utilize when making decisions. Because fraud risk assessments and red flag usage appear to be concerns for some investors, experimentally manipulating fraud red flags in investment settings and measuring investor reactions (e.g., investment decisions) would be a fruitful area of research. Future research may also investigate which investor characteristics lead to more appropriate fraud risk assessments. Continuation of such research will help standard setters make informed public-policy decisions designed to protect individual investors from financial statement fraud.

29

REFERENCES AICPA. 2002. Statement on Auditing Standards No. 99: Consideration of Fraud in a Financial Statement Audit. New York, NY: AICPA. Ali, A., D. Daqing, H. Lee-Seok, and M. Trombley. 2000. Accruals and future stock returns: Tests of the nave investor hypothesis. Journal of Accounting, Auditing and Finance. 15 (Spring): 161-181. Anderson, J.R. 1982. Acquisition of cognitive skill. Psychological Review. 89(4): 369-406. Asare, S.K., and A.M. Wright. 2004. The effectiveness of alternative risk assessment and program planning tools in a fraud setting. Contemporary Accounting Research 21 (2): 325-352. Auer, T., and T. Reponen. 1995. Information systems strategy formation embedded into a continuous organizational learning process. In Managing Information and Communications in a Changingin a Changing Global Environment, ed.M. Khosrowpour, 182-191. Idea Group Inc., Hershey, PA. Barber, B., and T. Odean. 2001. Boys will be boys: Gender, overconfidence, and common stock investment. Quarterly Journal of Economics 116 (1): 261-292. Baron, R. M., and D. A. Kenney. 1986. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology 51 (6): 1173-1182. Beasley, M. 1996. An empirical analysis of the relation between Board of Director composition and financial statement fraud. The Accounting Review 71 (October): 443-465. Bertaut, C. 1998. Stockholding behavior of U.S. households: Evidence from the 19831989 survey of consumer finances. Review of Economics and Statistics 80 (2): 263275. Bogle, J. 2005. The ownership of corporate America rights and responsibilities. Remarks by John C. Bogle, Founder and Former Chairman, The Vanguard Group, 20th Anniversary Meeting of the Council of Institutional Investors, April 11, 2005. Website, http://johncbogle.com/speeches/JCB_CII0405.pdf Brazel, J. F., K. L. Jones, and M. Zimbelman. 2009a. Using nonfinancial measures to assess fraud risk. Journal of Accounting Research. 47 (December): 1135-1166. Brazel, J. F, K. L. Jones, R. C. Warne. 2009b. Do nonprofessional investors react to fraud red flags? Working Paper, North Carolina State University. Brazel, J. F., T. D. Carpenter; and J. G. Jenkins. 2010. Auditors use of brainstorming in the 30

consideration of fraud: Reports from the field. The Accounting Review (forthcoming).

Chaiken, S and D. Maheswaran. 1994. Heuristic processing can bias systematic processing: Effects of source credibility, argument ambiguity, and task importance on attitude judgment. Journal of Personality and Social Psychology 66: 460-473. Coleman, D. F. and P.G. Irving. 1997. The influence of source credibility attributions on expectancy theory predictions of organizational choice. Canadian Journal of Behavioural Science 29: 122-131. Cornell, R. M., R. C. Warne, and M. E. Eining. 2009. The use of remedial tactics in negligence litigation. Contemporary Accounting Research 26 (Fall): 767-787. Cox, C. 2005. Speech by SEC chairman: Speech to SEC staff. August 4. Securities and Exchange Commission. Available at http://www.sec.gov/news/speech/spch080405cc.htm Dechow, P.M., R.G. Sloan, and A.P. Sweeney. 1996. Causes and consequences of earnings manipulation: An analysis of firms subject to enforcement actions by the SEC. Contemporary Accounting Research 13 (Spring): 136. Dechow, P. M., W. Ge, C. R. Larson and R. G. Sloan. 2010. Predicting material accounting misstatements. Contemporary Accounting Research (forthcoming). Available at SSRN: http://ssrn.com/abstract=997483. Eagly, A., W. Wood, and S. Chaiken. 1978. Causal inferences about communicators and their effect on opinion change. Journal of Personality and Social Psychology 36: 424-435. Efendi, J., A. Srivastava, E.P. Swanson. 2007. Why do corporate managers misstate financial statements? The role of option compensation and other factors. Journal of Financial Economics 85: 667-708. Elliott, W. B. 2006. Are investors influenced by pro forma emphasis and reconciliations in earnings announcements? The Accounting Review 81 (1): 113-133. Elliott, W.B., F.D. Hodge, K.E. Jackson. 2008. The association between nonprofessional investors information choices and their portfolio returns: The importance of investing experience. Contemporary Accounting Research. 25 (Summer): 473-498. Erev, I. and E. Haruvy. 2010. Learning and the Economics of Small Decisions. Working Paper, University of Texas at Dallas. Available at: http://www.utdallas.edu/~eeh017200/papers/LearningChapter.pdf Erickson, M. and S.W Wang. 1999. Earnings management by acquiring firms in stock for stock mergers. Journal of Accounting and Economics 27: 149-176.

31

Financial Accounting Standards Board. 2008. Statement of Financial Accounting Concepts No. 1, Objectives of financial reporting by business enterprises. Stamford, CT: FASB. Feiden, D. and T. Zambito. 2008. Losses rise in collapse of Bernie Maddoffs ponzi scheme. New York Daily News. (December 16). Fischer, P.; S. Schulz-Hardt; and D. Frey. 2008. Selection exposure and information quality: How different information quantites moderate decision makers preference for consistent and inconsistent information. Journal of Personality and Social Psychology 94(2): 231244. Friedman, D. and D. W. Massaro. 1998. Understanding variability in binary and continuous choice. Psychonomic Bulletin and Review 5(3): 370-389. Glass Lewis & Co. (GLC). 2005. Control deficienciesfinding financial impurities analysis of the 2004 and early 2005 of deficiency disclosures. Control deficiencies trend alert (June 24). Available at: http://www.glc.com. Graham, J. R.; C. R. Harvey; and S. Rajgopal. 2006. Value destruction and financial reporting decisions. Financial Analysts Journal. 62: 27-39. Hasher, L. and R.T. Zacks. 1979. Automatic and effortful processes in memory. Journal of Experimental Psychology: General. 108: 356-388. Hertwig, R.; G. Barron; E. U. Weber; and I. Erev. 2004. Decisions from experience and the effect of rare events in risky choice. Psychological Science 15(8):534-9. Hirst, D. E. 1994. Auditors sensitivity to source reliability. Journal of Accounting Research 32 (1): 113-126. Hodge, F. D., and M. Pronk. 2006. The impact of expertise and investment familiarity on investors use of online financial reporting. Journal of Accounting, Auditing and Finance 21 (Summer): 297-292. Hogan, C.E., Z. Rezaee, R.A. Riley, and U. Velury. 2008. Financial statement fraud: Insights from the academic literature. Auditing: A Journal of Practice and Theory 27 (November): 231-252. Hubbard, G. D. 2002. Warning signs found in Enron reports from 1995. Birmingham Business Journal. May 3. Jennings, M., D.C. Kneer, and P.M.J. Reckers. 1987. A reexamination of the concept of materiality: Views of the auditors, users, and officers of the court. Auditing: A Journal of Practice & Theory (Spring): 104-115.

32

Jennings, R. 1987. Unsystematic security price movements, management earnings forecasts, and revisions in consensus analyst earnings forecasts. Journal of Accounting Research 25: 90-110. Kadous, K. 2001. Improving jurors' evaluations of auditors in negligence cases. Contemporary Accounting Research. 18 (Fall): 425-444. Kahneman, D. 1973. Attention and Effort. Englewood Cliffs, NJ: Prentice-Hall. Kahneman, D. and A. Tversky. 1984. Choices, Values and Frames. American Psychologist 39 (April): 341-350. Kennedy, P. 1998. A Guide to Econometrics. Cambridge, MA: MIT Press. Lauricella, T. 2009. Investors see silver lining, as bad news clouds outlook. The Wall Street Journal. (May 4). Lee, T.A., R.W. Ingram, and T.P. Howard. 1999. The difference between earnings and operating cash flow as an indicator of financial reporting fraud. Contemporary Accounting Research 16 (Winter): 749-786. Levitt, S. D. and S. J. Dubner. 2005. Freakonomics: A rogue economist explores the hidden side of everything. New York, NY: HarperCollins. Levitt, S. D., and S. J. Dubner. 2009. Superfreakonomics: Global cooling, patriotic prostitutes, and why suicide bombers should buy life insurance. New York: HarperCollins. Loebbecke, J.K., M.M. Eining, and J.J. Willingham. 1989. Auditors' experience with material irregularities: Frequency, nature, and detectability. Auditing: A Journal of Practice & Theory 9 (1):1-28. Louis, H., 2004. Earnings management and the market performance of acquiring firms. Journal of Financial Economics 74: 121-148. Lowe, D. J., and P. M. J. Reckers. 1994. The effects of hindsight bias on jurors' evaluations of auditor decisions. Decision Sciences 25 (3): 401-26. Maines, L. A. and L. S. McDaniel. 2000. Effects of comprehensive income characteristics on nonprofessional investors judgments: The role of financial statement presentation format. The Accounting Review 75 (2): 179-207. Markopolos, H. 2010. No One Would Listen: A True Financial Thriller. Hoboken, NJ: John Wiley. Markowitz, H. 1952. Portfolio selection. The Journal of Finance 7 (1): 77-91.

33

Masters, R. 1989. Study examines investors risk-taking propensities. Journal of Financial Planning 2 (3): 151-155. Mercer, M. 2004. How do investors assess the credibility of management disclosures? Accounting Horizons 18: 185-96. Mintz, S.L. 2009. The gauge of innocence. CFO Magazine. (April 1). North American Securities Administrators Association (NASAA). 2006. NASAA Year in Review 2006: Advancing a Legacy of Investor Protection. Available at: http://www.nasaa.org/content/Files/2006YIR.pdf Nieschwietz, R., J. Schultz, and M. Zimbelman. 2000. Empirical research on external auditors detection of financial statement fraud. Journal of Accounting Literature 19: 190-246. Neter, J., M.H. Kutner, C.J. Nachtsheim, and W. Wasserman. 1996. Applied Linear Statistical Models. Fourth Edition. New York: McGraw-Hill. Nunnally, J.D. 1978. Psychometric Theory. New York: McGraw-Hill. Oppenheim, A.N. 1992. Questionnaire Design, Interviewing, and Attitude Measurement. New York, NY: Pinter Publishers. Public Law 107-204. 2002. 15 USC 7201. 107th United States Congress. H.R. 3763. July 30. Rakoff J.S. 2003. Opinion and Order, Securities and Exchange Commission v. WorldCom, Inc. United States District Court, Southern District of New York. 02 Civ. 4963 (JSR). July 7. Sloan, R.G. 1996. Do stock prices fully reflect information in accruals and cash flows about future earnings? The Accounting Review 71 (July): 289-315. Smith, J. L. 2010. The auditors loss function and investors perceptions of audit quality: Effects of regulatory change. Working Paper, University of NevadaLas Vegas. Tett, Gillian. 2009. Fools Gold: How the bold dream of a small tribe at J.P. Morgan was corrupted by Wall Street greed and unleashed catastrophe. New York: Free Press. Vinod, H. D. 2002. Winners and losers in multiple failures at Enron and some policy changes. Available at SSRN: http://ssrn.com/abstract=300542 or DOI: 10.2139/ssrn.300542 Warne, R.C. 2010. The effects of fair-market valuations of non-current assets on the judgments of nonprofessional investors. Working Paper, George Mason University. White House 2010. Remarks by the President in State of the Union Address. Available at http://www.whitehouse.gov/the-press-office/remarks-president-state-union-address

34

Zimbelman, M.F. 1997. The effects of SAS No. 82 on auditors attention to fraud risk factors and audit planning decisions. Journal of Accounting Research 35 (Supplement): 75-97. Zweig, J. 2009. About time: Regulation based on human nature. The Wall Street Journal. (June 20-21).

35

FIGURE 1 Investor Perceptions about Financial Statement Fraud and their Use of Red Flags

Perception of the Rate of Financial Statement Fraud

+ Reliance on Financial Statement Information

Importance of Fraud Risk Assessment when Investing

Use of Red Flags +

This figure provides a model of investor perceptions about financial statement fraud and their use of red flags.

36

TABLE 1 Descriptive Statistics Hypothesized Variables Response [n = 194] Mean (Std. Dev.)

Variables

Reliance on Financial Statement and Non-Financial Statement Information When Investing a 1. Balance sheet b 5.32 (1.64) 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Cash flow statement c Income statement c Internal control effectiveness c Statement of owners equity c Notes to financial statements d Mean reliance on financial statement information e Stock price f Advice from professionals f Company risk f Macroeconomic factors g Non-financial information related to operations h Advice from the media h Advice from nonprofessionals h Mean reliance on non-financial statement information i RELIANCE ON FINANCIAL j 5.30 (1.57) 5.23 (1.72) 5.01 (1.66) 4.98 (1.59) 4.91 (1.59) 5.13 (1.47) 5.46 (1.75) 5.06 (1.74) 4.98 (1.66) 4.75 (1.60) 4.17 (1.77) 4.09 (1.77) 4.03 (1.80) 4.69 (1.31) 1.12 (.28)

Perception of Rate of Fraudulent Financial Reporting 17. PERCEPTION OF FRAUD j Importance of Fraud Risk Assessment When Investing 18. FR j

5.06 (2.38)

5.24 (1.57)

37

Use of Red Flags k 19. 20. 21. 22. 23. 24. 25. SEC investigation l Pending litigation l Violation of debt covenant l High management turnover l Insider trades m Abnormally high valuation ratios m Large difference between cash flow from operations and net income m Anticipated merger or acquisition m Abnormally high revenue growth m Large change in a reserve account m Material weakness in internal control m Equity-based compensation m Recent stock or debt issuance m Large difference between revenue growth and non-financial measure growth m Abnormal decline in non-financial measures m Auditor change m Number of insiders on board of directors m Age n Need for external financing n Size n 38 5.28 (1.46) 5.21 (1.53) 5.14 (1.57) 5.14 (1.52) 5.13 (1.59) 5.08 (1.53)

5.01 (1.54) 5.01 (1.49) 5.00 (1.50) 4.98 (1.51) 4.93 (1.59) 4.89 (1.48) 4.83 (1.44)

26. 27. 28. 29. 30. 31. 32.

4.82 (1.56) 4.79 (1.47) 4.74 (1.60) 4.71 (1.64) 4.65 (1.62) 4.64 (1.46) 4.59 (1.47)

33. 34. 35. 36. 37. 38.

39. 40.
a

Use of non-Big 4 auditor n USE OF RED FLAGS j

4.42 (1.63) 4.91 (1.23)

Importance of financial statement information and non-financial statement information when deciding to buy or sell a companys stock, measured on a scale where 1 = very unimportant and 7 = very important. b In non-tabulated t-tests, mean response for financial statement item was significantly greater than (p-value < .05) the mean reliance on financial statement information (item 7). c In non-tabulated t-tests, mean response for financial statement item was not significantly different than (p-value > .05) the mean reliance on financial statement information (item 7). d In non-tabulated t-tests, mean response for financial statement item was significantly less than (p-value < .05) the mean reliance on financial statement information (item 7). e Calculated as the mean response to items 1-6. f In non-tabulated t-tests, mean response for non-financial statement item was significantly greater than (p-value < .05) the mean reliance on non-financial statement information (item 15). g In non-tabulated t-tests, mean response for non-financial statement item was not significantly different than (p-value > .05) the mean reliance on non-financial statement information (item 15). h In non-tabulated t-tests, mean response for non-financial statement item was significantly less than (p-value < .05) the mean reliance on non-financial statement information (item 15). i Calculated as the mean response to items 8-14. j RELIANCE ON FINANCIAL = Mean reliance on financial statement information / Mean reliance on non-financial statement information (item 7 / 15). PERCEPTION OF FRAUD = In your opinion, how often do managers of publicly-traded companies commit financial statement fraud, measured on a scale where 1 = 0% of the time and 11 = 100% of the time. FR = Importance of fraud risk assessment, relative to other factors, when making buy/sell decisions for stock that you currently hold, measured on a scale where 1 = not at all important and 7 = extremely important. USE OF RED FLAGS = Mean use of red flags (mean of items 19 - 39). k How often do you consider the following factors in assessing the risk of financial statement fraud in companies that you are screening for investment or in firms that you currently hold in your personal investment portfolio, measured on a scale where 1 = never and 7 = often. l In non-tabulated t-tests, mean response for red flag was significantly greater than (p-value < .05) the mean for use of red flags (item 40). m In non-tabulated t-tests, mean response for red flag was not significantly different than (p-value > .05) the mean for use of red flags (item 40). n In non-tabulated t-tests, mean response for red flag was significantly less than (p-value < .05) the mean for use of red flags (item 40).

39

TABLE 2 Descriptive Statistics Control Variables Response [n = 194] Mean (Std. Dev.)

Variables Screening Question 1. % bought or sold individual company stock in the last the last twelve months a Fraud-related Measures 2. % that OWNED THE STOCK OF A FRAUD COMPANY b 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. LOSS RECOVERY b Rely on regulators to detect and report fraud c, d Rely on external auditors to detect and report fraud c, d Rely on analysts to detect and report fraud c, d Rely on audit committees to detect and report fraud c, d Rely on other investors to detect and report fraud c, e Rely on internal auditors to detect and report fraud c, e Rely on internal controls to detect and report fraud c, e Rely on upper management to detect and report fraud c, f Rely on low/mid-level employee to detect and report fraud c, f Rely on media to detect and report fraud c, f Rely on short sellers to detect and report fraud c, f RELIANCE ON OTHERS b

100.00

24.74 3.37 (1.78) 5.02 (1.46) 4.86 (1.45) 4.79 (1.31) 4.56 (1.48) 4.54 (1.34) 4.52 (1.50) 4.52 (1.39) 4.34 (1.48) 4.32 (1.54) 4.31 (1.50) 4.29 (1.59) 4.56 (1.06)

Investing Experience, Activity, and Return b 16. INVESTING EXPERIENCE 17. 18. TIME ALLOCATED TRADING ACTIVITY 40

3.34 (1.53) 2.64 (1.63) 2.10 (1.33)

19. 20. 21.

DIVERSIFICATION OF INVESTMENTS VALUE OF PORTFOLIO RETURN ON INVESTMENTS

1.82 (1.06) 4.25 (1.99) 6.12 (2.64)

Investment Strategies and Industries b 22. RELY ON OTHERS VS. OWN ANALYSIS 23. 24. GROWTH STOCK STRATEGY A STRATEGY BASED ON YOUR FAMILIARITY WITH THE COMPANY HIGH YIELD STOCK STRATEGY A STRATEGY BASED ON TECHNICAL ANALYSIS LOW-RISK STOCK STRATEGY VALUE STOCK STRATEGY MOMENTUM STRATEGY LAST YEARS WINNER STRATEGY % invested heavily in ENERGY % invested heavily in HIGH TECH/COMMUNICATIONS % invested heavily in MANUFACTURING % invested heavily in HEALTHCARE/PHARMACEUTICALS % investing heavily in FINANCIAL SERVICES % invested heavily in RETAIL % investing heavily in MISCELLANEOUS INDUSTRIES % invested heavily in GOVERNMENT/NOT-FOR-PROFIT

4.04 (1.49) 4.75 (1.41)

4.70 (1.67) 4.52 (1.47) 4.44 (1.58) 4.41 (1.63) 4.38 (1.43) 3.90 (1.58) 3.69 (1.64) 41.23 40.72 32.98 29.38 25.77 22.16 6.70 5.15

25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

41

Demographic Data b 39. EDUCATION 40. 41. 42. 43. 44. 45. 46.
a

3.22 (1.14) 74.22 15.97 9.79 17.01 51.51 4.31 (1.41) 3.18 (1.32)

% with at least an UNDERGRADUATE DEGREE % with UNDERGRADUATE BUSINESS-RELATED DEGREE % with GRADUATE BUSINESS-RELATED DEGREE % CERTIFIED GENDER AGE HOUSEHOLD INCOME

Participants were asked to respond to the following statement (screening question): I have bought or sold individual company stock in the last twelve month. Participants could respond yes or no. b OWNED THE STOCK OF A FRAUD COMPANY = Have you ever owned the stock of an individual company when it was found to have been committing financial statement fraud, measured 1 = yes and 0 = no. LOSS RECOVERY = If you held the stock of a firm that committed financial statement fraud, how likely do you believe it is that you would recover your losses through shareholder lawsuits, measured on a scale where 1 = extremely unlikely and 7 = extremely likely. RELIANCE ON OTHERS = Mean reliance on others to detect and report fraud (mean of items 4 - 14). INVESTING EXPERIENCE = How many years have you been actively buying/selling the stocks of individual companies (as opposed to mutual funds, etc.), measured on a scale where 1 = less than one year and 6 = more than 20 years. TIME ALLOCATED = In an average week, how much time do you spend thinking about and evaluating stocks that you are screening for possible investment or that you currently hold in your personal investment portfolio, measured on a scale where 1 = less than one hour and 7 = more than 10 hours. TRADING ACTIVITY = Approximately, how many times, on average, do you buy or sell stocks of individual companies in a one-year period, measured on a scale where 1 = 1-5 times and 5 = more than 20 times. DIVERSIFICATION OF INVESTMENTS = In how many individual companies do you own stock (i.e., directly owning shares, not via a mutual fund, or pension), measured on a scale where 1 = 1-5 companies and 5 = more than 20 companies. VALUE OF PORTFOLIO = What is the approximate value of your personal investment portfolio, measured on a scale where 1 = less than $10,000 and 8 = more than $1,000,000.

42

RETURN ON INVESTMENTS = Over the last twelve months, what was the approximate return on your personal investment portfolio, measured on a scale where 1 = less than -20% and 11 = more than 20%. RELY ON OTHERS VS. OWN ANALYSIS = To what extent are your decisions to buy or sell stocks based on your own analysis relative to the advice of others, measured on a scale where 1 = based completely on my own analysis and 7 = based completely on the advice of others. GROWTH STOCK STRATEGY, A STRATEGY BASED ON FAMILIARITY WITH THE COMPANY, HIGH YIELD STOCK STRATEGY, A STRATEGY BASED ON TECHNICAL ANALYSIS, LOW-RISK STOCK STRATEGY, VALUE STOCK STRATEGY, MOMENTUM STRATEGY, LAST YEARS WINNER STRATEGY = Each strategy measured with the following question and scale: How often do you use the following investment strategies in your decisions to buy or sell stocks, where 1 = never and 7 = often. ENERGY, HIGH TECH/COMMUNICATIONS, MANUFACTURING, HEALTHCARE/PHARMACEUTICALS, FINANCIAL SERVICES, RETAIL, MISCELLANEOUS INDUSTRIES, GOVERNMENT/NOT-FOR-PROFIT = Measured with one question: In what industries do you most often buy and sell stocks of individual companies, response coded 1 if participant selected the industry, 0 otherwise. Participants could select more than one industry. EDUCATION = Please indicate the highest level of education you have completed, measured on a scale where 1 = high school and 5 = post-graduate degree UNDERGRADUATE DEGREE = coded 1 if participant obtained an undergraduate degree or higher, 0 otherwise. UNDERGRADUATE BUSINESS-RELATED DEGREE = coded 1 if participant obtained an undergraduate business-related degree, 0 otherwise. GRADUATE BUSINESS-RELATED DEGREE = coded 1 if participant obtained an graduate business-related degree, 0 otherwise. CERTIFIED = coded 1 if person obtained CPA, CFA, or CFP, 0 otherwise. GENDER = Coded 1 if male, 0 otherwise. AGE = Measured on a scale where 1 = under 20 and 8 = 80 or above. HOUSEHOLD INCOME = What is your total annual household income, measured on a scale where 1 = $0 - $30,000 and 6 = more than $150,000. c To what extent do you rely on the following parties to detect and report financial statement fraud in companies that you are screening for investment or in firms that you currently hold in your personal investment portfolio, measured on a scale where 1 = not at all and 7 = completely. d In non-tabulated t-tests, mean response for reliance was significantly greater than (p-value < .05) the mean for others to detect and report fraud (item 15). e In non-tabulated t-tests, mean response for reliance was not significantly different than (p-value > .05) the mean for others to detect and report fraud (item 15). f In non-tabulated t-tests, mean response for reliance was significantly less than (p-value < .05) the mean for others to detect and report fraud (item 15).

43

Table 3 Correlation Matrix

Variables a Rate FR Flags Recovery Others Time Return Momentum Growth Low Last Value High Technical Familiarity Misc Gender Reliance -0.09 -0.01 -0.08 -0.21 -0.24 0.04 -0.06 -0.17 -0.01 -0.12 -0.25 -0.15 -0.16 -0.21 -0.18 0.01 -0.04 0.08 0.12 0.15 -0.02 0.06 0.05 0.09 -0.04 0.02 0.10 0.12 0.13 0.15 -0.01 -0.15 0.15 0.51 0.05 0.33 0.23 0.26 0.22 0.29 0.30 0.27 0.30 0.41 0.29 0.22 -0.21 0.19 0.12 0.53 0.32 0.24 0.30 0.45 0.42 0.40 0.52 0.49 0.42 0.40 -0.20 0.09 0.38 0.18 0.04 0.41 0.19 0.28 0.48 0.15 0.20 0.42 0.40 -0.20 0.09 0.21 0.19 0.43 0.44 0.37 0.47 0.43 0.49 0.49 0.30 -0.15 0.12 0.15 0.28 0.22 0.12 0.24 0.40 0.33 0.29 0.25 -0.08 -0.02 0.12 0.11 0.13 0.13 0.15 0.27 0.16 0.14 0.11 0.14 0.60 0.42 0.61 0.34 0.40 0.59 0.22 -0.07 0.10 0.50 0.45 0.42 0.45 0.43 0.23 -0.11 0.01 0.51 0.36 0.46 0.39 0.27 -0.13 0.06 0.44 0.47 0.56 0.29 -0.18 0.11 0.55 0.51 0.43 -0.08 -0.01 0.62 0.43 -0.16 0.10 0.47 -0.19 0.03 -0.15 0.14 -0.13 Rate FR Flags Recovery Others Time Return Momentum Growth Low Last Value High Technical Familiarity Misc

Pearson correlation statistic. Correlations with two-tailed p-values < .05 are italicized and in boldface type. a RELIANCE ON FINANCIAL (Reliance), PERCEPTION OF FRAUD (Rate), FR, and USE OF RED FLAGS (Flags) are defined in Table 1. LOSS RECOVERY (Recovery), RELIANCE ON OTHERS (Others), TIME ALLOCATED (Time), RETURN ON INVESTMENTS (Return), MOMENTUM STRATEGY (Momentum), GROWTH STOCK STRATEGY (Growth), LOW-RISK STOCK STRATEGY (Low), LAST YEARS WINNER STRATEGY (Last), VALUE STOCK STRATEGY (Value), HIGH YIELD STOCK STRATEGY (High), A STRATEGY BASED ON TECHNICAL ANALYSIS (Technical), A STRATEGY BASED ON FAMILIARITY WITH THE COMPANY (Familiarity), MISCELLANEOUS INDUSTRIES (Misc), and GENDER (Gender) are defined in Table 2.

44

TABLE 4 Hypothesis One Testing: Ordinal Logistic Regression Results for FR a Estimated Independent Variables a Coefficient RELIANCE ON FINANCIAL (A) 0.282 PERCEPTION OF FRAUD (B) -0.061 AXB 0.139 LOSS RECOVERY -0.248 RETURN ON INVESTMENTS 0.122 LAST YEARS WINNER 0.238 HIGH YIELD 0.245 GENDER 0.671 FINANCIAL SERVICES 0.795 HOUSEHOLD INCOME -0.266 Model Chi-Square statistic = 83.11 (p-value < .001) R2 = .348
a

Waldstatistic 0.20 0.48 2.88 5.72 4.28 3.03 2.73 4.71 4.46 3.84

p-value 0.657 0.486 0.045 0.017 0.038 0.082 0.099 0.030 0.035 0.050

FR = Importance of fraud risk assessment, relative to other factors, when making buy/sell decisions for stock that you currently hold, measured on a scale where 1 = not at all important and 7 = extremely important. RELIANCE ON FINANCIAL (A) = Mean reliance on financial statement information / Mean reliance on non-financial statement information (items 7 / 14 in Table 1). PERCEPTION OF FRAUD (B) = In your opinion, how often do managers of publicly-traded companies commit financial statement fraud, measured on a scale where 1 = 0% of the time and 11 = 100% of the time. A X B = Interaction term: RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD LOSS RECOVERY = If you held the stock of a firm that committed financial statement fraud, how likely do you believe it is that you would recover your losses through shareholder lawsuits, measured on a scale where 1 = extremely unlikely and 7 = extremely likely. RETURN ON INVESTMENTS = Over the last twelve months, what was the approximate return on your personal investment portfolio, measured on a scale where 1 = less than -20% and 11 = more than 20%. LAST YEARS WINNER, HIGH YIELD = Each strategy measured with the following question and scale: How often do you use the following investment strategies in your decisions to buy or sell stocks, where 1 = never and 7 = often. FINANCIAL SERVICES = Measured with one question: In what industries do you most often buy and sell stocks of individual companies, response coded 1 if participant selected the industry, 0 otherwise. Participants could select more than one industry. GENDER = Coded 1 if male, 0 otherwise. HOUSEHOLD INCOME = What is your total annual household income, measured on a scale where 1 = $0 - $30,000 and 6 = more than $150,000.

45

TABLE 5 Hypothesis Two Testing: Linear Regression Results for USE OF RED FLAGS a Estimated Independent Variables Coefficient FR 0.199 RELIANCE ON OTHERS 0.312 VALUE OF PORTFOLIO -0.089 MANUFACTURING 0.295 ENERGY 0.427 VALUE STOCK STRATEGY 0.147 EDUCATION 0.154 Model F-statistic = 7.15 (p-value < .001) R2 = .613
a
a

t-statistic 4.31 4.01 2.01 2.03 2.88 2.30 2.52

p-value <.001 <.001 .056 .044 .005 .023 .013

USE OF RED FLAGS = Mean use of red flags (mean of items 19 39 in Table 1). FR = Importance of fraud risk assessment, relative to other factors, when making buy/sell decisions for stock that you currently hold, measured on a scale where 1 = not at all important and 7 = extremely important. RELIANCE ON OTHERS = Mean reliance on others to detect and report fraud (mean of items 4 14 in Table 2). VALUE OF PORTFOLIO = What is the approximate value of your personal investment portfolio, measured on a scale where 1 = less than $10,000 and 8 = more than $1,000,000. MANUFACTURING and ENERGY = Measured with one question: In what industries do you most often buy and sell stocks of individual companies, response coded 1 if participant selected the industry, 0 otherwise. Participants could select more than one industry. VALUE STOCK STRATEGY = Each strategy measured with the following question and scale: How often do you use the following investment strategies in your decisions to buy or sell stocks, where 1 = never and 7 = often. EDUCATION = Please indicate the highest level of education you have completed, measured on a scale where 1 = high school and 5 = post-graduate degree.

46

APPENDIX A CONTROL VARIABLE DEFINITIONS: OWNED THE STOCK OF A FRAUD COMPANY = Have you ever owned the stock of an individual company when it was found to have been committing financial statement fraud, measured 1 = yes and 0 = no. LOSS RECOVERY = If you held the stock of a firm that committed financial statement fraud, how likely do you believe it is that you would recover your losses through shareholder lawsuits, measured on a scale where 1 = extremely unlikely and 7 = extremely likely. RELIANCE ON OTHERS = Mean reliance on others to detect and report fraud (further described below and in Table 2). INVESTING EXPERIENCE = How many years have you been actively buying/selling the stocks of individual companies (as opposed to mutual funds, etc.), measured on a scale where 1 = less than one year and 6 = more than 20 years. TIME ALLOCATED = In an average week, how much time do you spend thinking about and evaluating stocks that you are screening for possible investment or that you currently hold in your personal investment portfolio, measured on a scale where 1 = less than one hour and 7 = more than 10 hours. TRADING ACTIVITY = Approximately, how many times, on average, do you buy or sell stocks of individual companies in a one-year period, measured on a scale where 1 = 1-5 times and 5 = more than 20 times. DIVERSIFICATION OF INVESTMENTS = In how many individual companies do you own stock (i.e., directly owning shares, not via a mutual fund, or pension), measured on a scale where 1 = 1-5 companies and 5 = more than 20 companies. VALUE OF PORTFOLIO = What is the approximate value of your personal investment portfolio, measured on a scale where 1 = less than $10,000 and 8 = more than $1,000,000. RETURN ON INVESTMENTS = Over the last twelve months, what was the approximate return on your personal investment portfolio, measured on a scale where 1 = less than -20% and 11 = more than 20%. RELY ON OTHERS VS. OWN ANALYSIS = To what extent are your decisions to buy or sell stocks based on your own analysis relative to the advice of others, measured on a scale where 1 = based completely on my own analysis and 7 = based completely on the advice of others. TRADING STRATEGY = MOMENTUM, GROWTH, LOW-RISK, LAST YEARS WINNER, VALUE, HIGH YIELD, TECHNICAL ANALYSIS, FAMILIARITY WITH THE 47

COMPANY = Each strategy measured with the following question and scale: How often do you use the following investment strategies in your decisions to buy or sell stocks, where 1 = never and 7 = often. INDUSTRY = MANUFACTURING, RETAIL, GOVERNMENT/NOTFOR-PROFIT, ENERGY, HIGH TECH/COMMUNICATIONS, HEALTHCARE/PHARMACEUTICALS, FINANCIAL SERVICES, MISCELLANEOUS INDUSTRIES = Measured with one question: In what industries do you most often buy and sell stocks of individual companies, response coded 1 if participant selected the industry, 0 otherwise. Participants could select more than one industry. EDUCATION = Please indicate the highest level of education you have completed, measured on a scale where 1 = high school and 5 = post-graduate degree. UNDERGRADUATE DEGREE = Coded 1 if participant obtained an undergraduate degree or higher, 0 otherwise. UNDERGRADUATE BUSINESS-RELATED DEGREE = Coded 1 if participant obtained an undergraduate business-related degree, 0 otherwise. GRADUATE BUSINESS-RELATED DEGREE = Coded 1 if participant obtained a graduate business-related degree, 0 otherwise. CERTIFIED = Coded 1 if person obtained CPA, CFA, or CFP, 0 otherwise. GENDER = Coded 1 if male, 0 otherwise. AGE = Measured on a scale where 1 = under 20 and 8 = 80 or above. HOUSEHOLD INCOME = What is your total annual household income, measured on a scale where 1 = $0 - $30,000 and 6 = more than $150,000.

48