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

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

1

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

2

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

1 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).

3

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 pre-

screened 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’

4

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.

5

Despite numerous high-profile frauds in the past and the high cost of fraud to market

participants, investors’ a priori decision processes vis-à-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

6

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 publically-

traded 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

2 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.

7

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

4 ―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.‖

8

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 1930’s, banks and others appeared to underweight the probability of this rare event.5

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

9

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 mortgage-

backed 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).

10

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 investor’s 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

11

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 manager’s 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.

12

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 investor’s 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,

9 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.

13

the survey method is the most effective means of addressing our hypotheses. In addition, given

the paucity of research related to investor’s 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 non-

fraud-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.

14

Participants were residents from all 50 states and Washington DC, approximately 50%

male, well educated (75% had a bachelor’s 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 = 0 + 1RELIANCE ON FINANCIAL + 2PERCEPTION OF FRAUD +

3RELIANCE ON FINANCIAL X PERCEPTION OF FRAUD + 4-37CONTROL

VARIABLES + ε (Model 1)

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).

15

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 + ε (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.‖

RELIANCE ON FINANCIAL = Mean reliance on direct financial statement information /

Mean reliance on non-financial statement information

(further described below and in Table 1).

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.‖

USE OF RED FLAGS = Mean use of red flags (further described below and in

Table 1).

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.

16

[Insert Table 1]

The means for reliance on financial statement information (item 7) and reliance on non-

financial 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).

17

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.

18

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 Cronbach’s alpha

levels exceeding the generally accepted threshold of .70 (Nunnally 1978).

19

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

20

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).

21

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 investor’s 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 straight-

forward, 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).

22

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 p-

values < 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.

23

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, Enron’s financial statements from 1995 include multiple red flags related to the accuracy of its

statements (Hubbard 2002).

24

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 investor’s 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-à-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

25

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

26

earnings fixation by naïve 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.

27

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.

28

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 ex-

ante 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-à-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.

29

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.

30

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 naïve 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 1983-

1989 survey of consumer finances. Review of Economics and Statistics 80 (2): 263-

275.

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

31

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): 1–36.

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.

32

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 Maddoff’s 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): 231-

244.

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 deficiencies—finding 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.

33

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.

34

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 auditor’s loss function and investors’ perceptions of audit quality:

Effects of regulatory change. Working Paper, University of Nevada—Las Vegas.

Tett, Gillian. 2009. Fool’s 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

35

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).

36

FIGURE 1

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

+

+ +

Reliance on

Financial Statement

Information

Importance of Fraud

Risk Assessment when

Investing

Use of Red Flags

Perception of the Rate of

Financial Statement Fraud

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

of red flags.

37

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. Cash flow statement c 5.30 (1.57)

3. Income statement c 5.23 (1.72)

4. Internal control effectiveness c 5.01 (1.66)

5. Statement of owner’s equity c 4.98 (1.59)

6. Notes to financial statements d 4.91 (1.59)

7. Mean reliance on financial statement information e 5.13 (1.47)

8. Stock price f 5.46 (1.75)

9. Advice from professionals f 5.06 (1.74)

10. Company risk f 4.98 (1.66)

11. Macroeconomic factors g 4.75 (1.60)

12. Non-financial information related to operations h 4.17 (1.77)

13. Advice from the media h 4.09 (1.77)

14. Advice from nonprofessionals h 4.03 (1.80)

15. Mean reliance on non-financial statement information i 4.69 (1.31)

16. RELIANCE ON FINANCIAL j 1.12 (.28)

Perception of Rate of Fraudulent Financial Reporting

17. PERCEPTION OF FRAUD j 5.06 (2.38)

Importance of Fraud Risk Assessment When Investing

18. FR j 5.24 (1.57)

38

Use of Red Flags k

19. SEC investigation l 5.28 (1.46)

20. Pending litigation l 5.21 (1.53)

21. Violation of debt covenant l 5.14 (1.57)

22. High management turnover l 5.14 (1.52)

23. Insider trades m

5.13 (1.59)

24. Abnormally high valuation ratios m

5.08 (1.53)

25. Large difference between cash flow from operations and net

income m

5.01 (1.54)

26. Anticipated merger or acquisition m

5.01 (1.49)

27. Abnormally high revenue growth m

5.00 (1.50)

28. Large change in a reserve account m

4.98 (1.51)

29. Material weakness in internal control m

4.93 (1.59)

30. Equity-based compensation m

4.89 (1.48)

31. Recent stock or debt issuance m

4.83 (1.44)

32. Large difference between revenue growth and non-financial

measure growth m

4.82 (1.56)

33. Abnormal decline in non-financial measures m

4.79 (1.47)

34. Auditor change m

4.74 (1.60)

35. Number of insiders on board of directors m

4.71 (1.64)

36. Age n 4.65 (1.62)

37. Need for external financing n 4.64 (1.46)

38. Size n 4.59 (1.47)

39

39. Use of non-Big 4 auditor n 4.42 (1.63)

40. USE OF RED FLAGS j 4.91 (1.23)

a Importance of financial statement information and non-financial statement information when

deciding to buy or sell a company’s 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).

40

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 100.00

Fraud-related Measures

2. % that OWNED THE STOCK OF A FRAUD COMPANY b 24.74

3. LOSS RECOVERY b 3.37 (1.78)

4. Rely on regulators to detect and report fraud c, d

5.02 (1.46)

5. Rely on external auditors to detect and report fraud c, d

4.86 (1.45)

6. Rely on analysts to detect and report fraud c, d

4.79 (1.31)

7. Rely on audit committees to detect and report fraud c, d

4.56 (1.48)

8. Rely on other investors to detect and report fraud c, e

4.54 (1.34)

9. Rely on internal auditors to detect and report fraud c, e

4.52 (1.50)

10. Rely on internal controls to detect and report fraud c, e

4.52 (1.39)

11. Rely on upper management to detect and report fraud c, f

4.34 (1.48)

12. Rely on low/mid-level employee to detect and report fraud c, f

4.32 (1.54)

13. Rely on media to detect and report fraud c, f

4.31 (1.50)

14. Rely on short sellers to detect and report fraud c, f

4.29 (1.59)

15. RELIANCE ON OTHERS b 4.56 (1.06)

Investing Experience, Activity, and Return b

16. INVESTING EXPERIENCE 3.34 (1.53)

17. TIME ALLOCATED 2.64 (1.63)

18. TRADING ACTIVITY 2.10 (1.33)

41

19. DIVERSIFICATION OF INVESTMENTS 1.82 (1.06)

20. VALUE OF PORTFOLIO 4.25 (1.99)

21. RETURN ON INVESTMENTS 6.12 (2.64)

Investment Strategies and Industries b

22. RELY ON OTHERS VS. OWN ANALYSIS 4.04 (1.49)

23. GROWTH STOCK STRATEGY 4.75 (1.41)

24. A STRATEGY BASED ON YOUR FAMILIARITY WITH THE

COMPANY 4.70 (1.67)

25. HIGH YIELD STOCK STRATEGY 4.52 (1.47)

26. A STRATEGY BASED ON TECHNICAL ANALYSIS 4.44 (1.58)

27. LOW-RISK STOCK STRATEGY 4.41 (1.63)

28. VALUE STOCK STRATEGY 4.38 (1.43)

29. MOMENTUM STRATEGY 3.90 (1.58)

30. LAST YEAR’S WINNER STRATEGY 3.69 (1.64)

31. % invested heavily in ENERGY 41.23

32. % invested heavily in HIGH TECH/COMMUNICATIONS 40.72

33. % invested heavily in MANUFACTURING 32.98

34. % invested heavily in HEALTHCARE/PHARMACEUTICALS 29.38

35. % investing heavily in FINANCIAL SERVICES 25.77

36. % invested heavily in RETAIL 22.16

37. % investing heavily in MISCELLANEOUS INDUSTRIES 6.70

38. % invested heavily in GOVERNMENT/NOT-FOR-PROFIT 5.15

42

Demographic Data b

39. EDUCATION 3.22 (1.14)

40. % with at least an UNDERGRADUATE DEGREE 74.22

41. % with UNDERGRADUATE BUSINESS-RELATED DEGREE 15.97

42. % with GRADUATE BUSINESS-RELATED DEGREE 9.79

43. % CERTIFIED 17.01

44. GENDER 51.51

45. AGE 4.31 (1.41)

46. HOUSEHOLD INCOME 3.18 (1.32)

a 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.‖

43

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 YEAR’S 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).

44

Table 3

Correlation Matrix

Variables a Reliance Rate FR Flags Recovery Others Time Return Momentum Growth Low Last Value High Technical Familiarity Misc

Rate -0.09

FR -0.01 0.08

Flags -0.08 0.12 0.51

Recovery -0.21 0.15 0.05 0.12

Others -0.24 -0.02 0.33 0.53 0.38

Time 0.04 0.06 0.23 0.32 0.18 0.21

Return -0.06 0.05 0.26 0.24 0.04 0.19 0.15

Momentum -0.17 0.09 0.22 0.30 0.41 0.43 0.28 0.12

Growth -0.01 -0.04 0.29 0.45 0.19 0.44 0.22 0.11 0.60

Low -0.12 0.02 0.30 0.42 0.28 0.37 0.12 0.13 0.42 0.50

Last -0.25 0.10 0.27 0.40 0.48 0.47 0.24 0.13 0.61 0.45 0.51

Value -0.15 0.12 0.30 0.52 0.15 0.43 0.40 0.15 0.34 0.42 0.36 0.44

High -0.16 0.13 0.41 0.49 0.20 0.49 0.33 0.27 0.40 0.45 0.46 0.47 0.55

Technical -0.21 0.15 0.29 0.42 0.42 0.49 0.29 0.16 0.59 0.43 0.39 0.56 0.51 0.62

Familiarity -0.18 -0.01 0.22 0.40 0.40 0.30 0.25 0.14 0.22 0.23 0.27 0.29 0.43 0.43 0.47

Misc 0.01 -0.15 -0.21 -0.20 -0.20 -0.15 -0.08 0.11 -0.07 -0.11 -0.13 -0.18 -0.08 -0.16 -0.19 -0.15

Gender -0.04 0.15 0.19 0.09 0.09 0.12 -0.02 0.14 0.10 0.01 0.06 0.11 -0.01 0.10 0.03 0.14 -0.13

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 YEAR’S 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.

45

TABLE 4

Hypothesis One Testing: Ordinal Logistic Regression Results for FR a

Independent Variables a

Estimated

Coefficient

Wald-

statistic p-value

RELIANCE ON FINANCIAL (A) 0.282 0.20 0.657

PERCEPTION OF FRAUD (B) -0.061 0.48 0.486

A X B 0.139 2.88 0.045

LOSS RECOVERY -0.248 5.72 0.017

RETURN ON INVESTMENTS 0.122 4.28 0.038

LAST YEAR’S WINNER 0.238 3.03 0.082

HIGH YIELD 0.245 2.73 0.099

GENDER 0.671 4.71 0.030

FINANCIAL SERVICES 0.795 4.46 0.035

HOUSEHOLD INCOME -0.266 3.84 0.050

Model Chi-Square statistic = 83.11 (p-value < .001)

R2

= .348

a 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 YEAR’S 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.‖

46

TABLE 5

Hypothesis Two Testing: Linear Regression Results for USE OF RED FLAGS a

Independent Variables a

Estimated

Coefficient t-statistic p-value

FR 0.199 4.31 <.001

RELIANCE ON OTHERS 0.312 4.01 <.001

VALUE OF PORTFOLIO -0.089 2.01 .056

MANUFACTURING 0.295 2.03 .044

ENERGY 0.427 2.88 .005

VALUE STOCK STRATEGY 0.147 2.30 .023

EDUCATION 0.154 2.52 .013

Model F-statistic = 7.15 (p-value < .001)

R2

= .613

a 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.‖

47

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

YEAR’S WINNER, VALUE, HIGH YIELD,

TECHNICAL ANALYSIS, FAMILIARITY WITH THE

48

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/NOT-

FOR-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.‖