Investor reaction to disclosures of employee fraud.
Article from: Journal of Managerial Issues Article date: September 22, 1995 Author: Lukawitz, James M. ; Steinbart, Paul John More results for: employee fraud

Fraud is a serious problem facing American businesses, with estimates of the annual losses due to employee theft and embezzlement exceeding $100 billion. Companies have been criticized, however, for failing to report cases of discovered employee fraud. A common rationale for not disclosing such news is the fear of negative publicity. The objective of this study is to determine whether or not such disclosures do indeed create negative publicity for the firm. Investors are one of the important "publics" for any company. Therefore, this study examines investors' reactions to disclosures of employee fraud. Specifically, the stock market's reaction to stories of employee fraud published in the Wall Street Journal was examined to determine whether such disclosures were indeed treated as "bad news." The results indicate that such disclosures are not associated with any negative stock price reactions. Thus, there was no evidence that one important segment of the public, investors, reacted negatively to disclosures of employee fraud.

Employee fraud is a serious problem for American businesses. Estimates of annual losses due to employee theft and embezzlement across all industries exceed $100 billion (Lary, 1989; Wells, 1994). It has been argued that disclosure and subsequent prosecution of all cases of discovered fraud might serve as a deterrent (Albrecht et al., 1994; Dycus, 1991; Richards and Knotts, 1989; but for an opposing view see Wells, 1990). Management has been criticized, however, for not adequately responding to discovered cases of employee fraud (Dycus, 1991). For example, studies (O'Donoghue, 1987; Richards and Knotts, 1989; Straub and Nance, 1990) have found that less than one-half of the discovered cases of computer-related frauds are ever reported to the authorities. Indeed, it is estimated that less than 20% of all discovered frauds are ever reported (Cushing and Romney, 1993).

A common rationale for management's failure to report discovered cases of fraud is fear of adverse publicity (Richards and Knotts, 1989; Gerlin, 1995). The objective of this paper is to test the validity of that belief. To do so, the reactions of one important segment of the public, investors, are examined. Focusing on investor reactions provides a powerful test of whether any embarrassment or reputational damage to the company arising from disclosure of employee fraud has economic consequences. If disclosure of employee frauds is indeed "bad news," such disclosures should be associated with a negative stock price reaction. The lack of negative market reaction to disclosures of employee frauds would indicate that investors did not perceive any significant long-term financial repercussions arising from the incident.

BACKGROUND AND HYPOTHESES

The word fraud encompasses many types of acts; hence, it is important to carefully define how it is used in this study. The authoritative auditing literature distinguishes two types of irregularities: management fraud and employee frauds (AICPA, 1988). Management fraud refers to such things as fraudulent financial reporting that results in misleading financial statements. The term employee frauds refers to the misappropriation of assets, thus including such acts as theft and embezzlement.

This study focuses on the effect of disclosing discovered cases of frauds committed by employees. A recent study by Karpoff and Lott (1993) examined investors' reactions to disclosures of management fraud. They found that the stock market did react negatively to disclosures of management fraud, including such acts as defrauding the government, consumers, and other stakeholders. Karpoff and Lott did not, however, include incidences of fraud by employees in their sample. Therefore, this study extends their research by examining the stock market's reaction to disclosures of employee fraud.

Although Karpoff and Lott found that investors reacted negatively to disclosures of management fraud, there are several reasons to question whether investors react similarly to disclosures of employee fraud. First, the authoritative auditing literature (AICPA, 1988) states that employee frauds are often immaterial in amount. This means that the amounts are too small to seriously affect the company's financial position. Consequently, investors may not react to such disclosures. Second, the disclosure itself indicates that the fraud has been detected and stopped. Therefore, there may be no effect on the company's future performance. Stock prices reflect investors' perceptions of future events; therefore, investors may not react negatively to news of employee fraud that has been detected and stopped. Third, it has been argued that companies publicly disclose cases of employee fraud only when they feel they can recover most or all of the loss (Pae, 1989). If that argument is correct, then any stigma attached to the fraud may be offset by the public's expectation that the loss will be recovered.

On the other hand, the concept of materiality can also be used to support the possibility that there will be a negative stock price reaction to disclosures of employee frauds. The argument is that materiality is not solely a function of the magnitude of the event, but is also affected by the nature of the event. Indeed, Statement of Financial Accounting Concepts No. 2 notes that "items too small to be thought material if they result from routine transactions may be considered material if they arise in abnormal circumstances" (FASB, 1980: paragraph 123). Fraud is (hopefully) an "abnormal" circumstance. Consequently, although employee frauds may be immaterial in size, investors may consider them to be material in nature. Specifically, disclosure of employee frauds may be taken as a signal indicating that weaknesses exist in a company's internal control structure. Investors may believe that the control weaknesses that permitted the disclosed fraud to occur make it likely that other employee frauds exist that have not yet been detected. Therefore, investors may react negatively to such disclosures.

The preceding discussion leads to the first hypothesis to be tested, stated in its alternative form as follows:

H1: There will be a negative stock price reaction to disclosures of employee frauds.

Effect of Required Disclosure: The Banking Industry

Banks are currently required to report all discovered employee frauds in excess of $1,000 to the authorities (12 Code of Federal Regulations, Section 21.11). This prevents bank managers from concealing embarrassing employee frauds or those which they do not want revealed. Other types of companies do not have this mandatory disclosure requirement. Therefore, the probability of disclosure may differ for banks versus other types of companies. Indeed, almost 40% of the fraud disclosures in our sample involve banks. Consequently, we also examine banks and nonbanks separately. Therefore, two additional hypotheses are tested, stated in their alternative form as follows:

H2: There will be a negative stock price reaction to disclosures by banks of employee frauds.

H3: There will be a negative stock price reaction to disclosures by nonbanks of employee frauds.

RESEARCH DESIGN

Sample Selection

The sample was selected by searching the Wall Street Journal Index for the previous decade under the headings of crime, computer crime, white collar crime, fraud, embezzlement and theft. Three screening criteria were used in the search. First, only stories about employee frauds were included in the sample. Stories about fraudulent financial reporting, defense contract or other government fraud, violations of regulations and defrauding of stakeholders were excluded because the information content of such disclosures was already examined by Karpoff and Lott (1993). Second, stories indicating that the company had gone bankrupt were excluded because it would not be possible to separate the information content associated with the fraud story from that associated with bankruptcy. Third, only the first story about a specific fraud was included in the sample. Karpoff and Lott (1993) found that the information content associated with disclosures of management fraud was limited to the first story about that event.

Application of those three criteria yielded a sample of 144 stories about employee frauds. The average number of stories is only 12 each year, indicating that disclosures of employee fraud are relatively rare events. That number was then reduced to the final sample size of 46 stories as follows:

1. The corporate index was searched for other stories about the company within a five-day window centered on the date of the fraud story. If there was a major economic announcement such as earnings, the signing of a major contract, or dividends paid out by the firm during this period, the firm was removed from the sample since it would be impossible to separate the market reaction of the fraud from the other economic news. This led to the elimination of 43 frauds for which there were conflicting events.

2. Only companies that were listed either on the Center for Research in Security Price (CRSP) daily return tape or the Standard and Poors Daily Stock Price Record were retained. This led to the elimination of an additional 53 cases for which stock price data was not readily available.

3. Two cases (Enron, with a loss of $85,000,000 and Chemical Bank, with a loss of $33,000,000) were eliminated because they were disproportionately large compared to the remaining 46 stories, which had an average loss of less than $2,000,000.(1)

Method

Two tests are used to examine investor reactions to disclosures of employee fraud. The first test examines the average standardized cumulative abnormal returns across companies. The abnormal return measures unusual fluctuations in a company's stock prices that differ from its relationship with the average market fluctuation. The abnormal return for company i on day t is calculated as follows:

A[R.sub.i,t] = [R.sub.i,t] - ([a.sub.i] + [b.sub.i][R.sub.m,t]) (1)

where

[R.sub.i,t] = the return on company i's stock on day t,

[a.sub.i], [b.sub.i] = ordinary least squares (OLS) coefficient estimates from regressing company i's individual daily stock returns on the market returns, and

[R.sub.m,t] = the market return on the CRSP equally-weighted index on day t.

The market model parameters are estimated over a 60-day period from day t-61 through day t-2, where t = 0 is the date the story about the employee fraud appears in the Wall Street Journal. The cumulative abnormal returns for each company i ([CAR.sub.i]) are calculated by summing the abnormal returns across one- and two-day periods to allow sufficient time to serve any reaction to the fraud disclosure. The cross-firm and interval returns are then normalized according to the procedures outlined by Patell (1976).(2) If the abnormal returns for either day t-1, day t = 0, or the two-day window are significantly negative, it would provide evidence that investors do indeed treat disclosures of employee frauds as bad news.

The second test examines whether the proportion of companies which experience a negative stock price reaction to disclosures of employee fraud is significantly different from chance. A binomial test is applied to the sign of the abnormal returns experienced by companies that disclose news of employee frauds. A significant test result indicates that companies which make such disclosures are likely to experience a negative stock price reaction to such news. Note, however, that this test is weaker than the first test because it does not examine whether the sizes of the observed negative price reactions are themselves significant.

RESULTS

Table 1 shows that the size of reported frauds in the final sample ranged from $14,500 to $12,000,000, with an average size of slightly less than $2 million. Fifteen (32.6%) of the stories dealt with frauds of less than $1 million and four involved losses of less than $100,000. Although the average size of the frauds at the banks ($2,770,830) was larger than the average size of the fraud at the nonbanks ($1,419,543), the difference in size was not significant (two-tailed t = 1.63 p[greater than].10).

Table 1. Descriptive Statistics of the Sample

Non banks
Company Loss

AMC EntertainmentN/A
AMR125,000
ARA Services 250,000
American International Group 1,700,000
Anheuser Busch N/A
Atek Metals700,000
CNA Financial3,000,000
Coca Cola5,000,000
Deere & Company 1,000,000
Dun & Bradstreet 1,500,000
GTE500,000
Genesco 5,500,000
Global Natural Resources N/A
Jet Industries N/A
L.L. Luria & Sons 31,000
LTV (1)(*) N/A
LTV (2)(*) 14,500
MCA294,000
Marietta Corp. 385,000
Marriott 150,000
NCR300,000
Pittston Co. (1)(*) 1,000,000
Pittston Co. (2)(*) 1,000,000
Revlon Group 1,100,000
Tonka2,000,000
URS 2,500,000
Valley Industries2,600,000
Woolworth2,000,000
Avg. Nonbank loss1,419,543
Avg. Total Loss 1,973,918

Banks

CompanyLoss

American Fletcher4,000,000
American Republican400,000
Bank of New York (1)(*) 1,200,000
Bank of New York (2)(*) 5,500,000
Bankers Trust NY 1,200,000
Chase Manhattan900,000
Citicorp14,800
Comfed BancorporationN/A
Crocker National 6,000,000
Financial Corporation of America N/A
First Chicago 400,000
First City Bancorp TX 18,484
Gibraltar Financial 3,400,000
Harris Bankcorp 1,300,000
Manufacturers Hanover1,000,000
J.P. Morgan 6,000,000
NCNB12,000,000
Shawmut National 1,000,000

Avg. Bank loss 2,770,830

* There were two separate incidents for these three companies.
Hypothesis 1: Investor Reactions (All Firms)

Table 2 presents the data about the stock price reactions to the 46 disclosures of employee frauds in our sample. Hypothesis 1 predicted that disclosures of employee frauds would result in a negative stock price reaction for all firms. The all-companies column in Table 2 provides the results of testing that hypothesis.

Panels A and C show the results of the test involving abnormal returns across companies. Panel A shows the average abnormal returns for each day for companies disclosing news of employee frauds. The data show that the abnormal returns are not significantly less than zero on either day t-1 or day t = 0. Moreover, Panel C shows that the cumulative abnormal [TABULAR DATA FOR TABLE 2 OMITTED] returns over the two-day window [-1,0] are also not significantly less than zero. Thus, the first method of testing finds no evidence of a negative market reaction to disclosures of employee fraud.

Panels B and D show the results of the test of the proportion of companies experiencing a negative stock price reaction to disclosures of employee fraud. Panel B shows that the proportion of firms experiencing an abnormal negative stock price reaction is not significant on either day t-1 or day t = 0. Panel D shows that the proportion of firms experiencing negative abnormal stock price reactions across the two-day window [-1,0] is also not significantly different from chance.

In summary, the data presented in Table 2 do not support hypothesis 1. Across all types of firms, there is no evidence of any negative stock market reaction to disclosures of employee frauds.

Hypotheses 2 and 3: Effects of Type of Company

Hypotheses 2 and 3 deal separately with investors' reactions to disclosures of employee frauds by banks and nonbanks, respectively. The column labeled Banks in Table 2 provides data concerning hypothesis 2. Panel A of Table 2 shows that the average abnormal stock returns for banks only were not significantly less than zero on either day t-1 or day t = 0. Panel C corroborates that finding, showing that the average cumulative stock price returns across both days were not significantly less than zero. Thus, there is no evidence of significant negative market reactions to disclosures of employee fraud by banks.

Panel B of Table 2 shows that the proportion of banks experiencing negative stock price returns on day t-1 was not significantly different from chance. However, on day t = 0, the day that the story about employee fraud appears in the Wall Street Journal, two-thirds of the banks suffered negative stock price reactions. That proportion is marginally different from chance (p = .08), providing some weak evidence in support of hypothesis 2. Panel D shows that the proportion of banks experiencing cumulative negative stock price returns over the two-day [-1,0] window was not statistically significant.

In summary, there is little evidence that investors reacted negatively to disclosures of employee frauds at banks. Although there was a marginally significant percentage of banks that did experience negative returns on the date such disclosures

peared in the Wall Street Journal, the results in panels A and C indicate that the size of those negative returns was not significantly different from zero.

The column labeled nonbanks in Table 2 provides data concerning hypothesis 3. Panel A shows no evidence of abnormal negative stock price returns on either day t-1 or day t = 0. Panel C shows no evidence of any negative cumulative stock price returns for the two-day window [-1, 0]. Thus, there is no evidence of any significant negative stock price reaction to disclosures of employee frauds by nonbanks.

Panel B shows that the percentage of nonbanks disclosing news of employee frauds which experienced a negative stock price reaction on either day t-1 or t = 0 was not significant either. Panel D shows similar results for the two-day window [-1, 0].

In summary, hypothesis 3 is not supported; there is no evidence of a negative stock price reaction to disclosures of employee frauds at non-banks.

DISCUSSION AND CONCLUSION

Management has been criticized for not disclosing news of discovered employee frauds. It has been argued that one reason management is reluctant to make such disclosures is fear of adverse publicity. This study tested the validity of that belief by examining investor reaction to stories of employee fraud published in the Wall Street Journal. Across all companies, no evidence that investors reacted to such stories as "bad news" was found. There was some weak evidence that a majority of banks that were the victims of employee fraud experienced negative stock price reactions on the day that news of those frauds was published, but the size of the negative price reaction was not statistically significant.

Limitations

Before discussing the implications of these results, several limitations of the study should be noted. The first limitation is that the sample was obtained from only one source, the Wall Street Journal. Investors may have other sources of news about employee frauds. If these other sources are more timely, then disclosure in the Wall Street Journal may not be "news" to these investors. Thus, the failure to find a reaction to such disclosures in this study may be due to investors having become aware of such news from other sources. On the other hand, the Wall Street Journal is the most widely read source of business news and is routinely used by researchers as the source for testing investors' reactions to a variety of disclosures. Moreover, Karpoff and Lott (1993) used the Wall Street Journal as the source in their study and found significant stock market reactions to disclosures of fraud by management (not employees).

The second limitation involves the size of the frauds examined in this study. Most of these frauds were small, especially when considered in terms of total assets or revenues. The failure to find a market reaction may be due to the fact that these employee frauds are immaterial. Market participants may only be concerned with employee frauds when they are large enough to materially affect the firm's financial operations.

Several additional analyses were conducted to test this possibility. First, the size of the fraud in absolute terms, as a percentage of assets, and as a percentage of income, was compared with the company's abnormal returns. None of these measures of size were found to be significantly related to abnormal returns. Second, the analyses were rerun for only those firms for which the size of the fraud was greater than one percent of total assets or one-half of one percent of net income. Again, no significant relationship with abnormal returns was found. Therefore, the results presented in Table 2 appear to indicate that investors do not react negatively to disclosures of employee frauds of the magnitude reported in this study.

Nevertheless, it is important to note that fifty-three firms were lost from the sample due to the inability to gather their stock returns. This exceeds the number retained in the sample, and represents over one-third of all the employee frauds reported during the time period of our study. Most likely, these firms were too small to be a part of the stock return databases. Employee frauds may invoke a market reaction for these small firms where the size of the fraud may be significant in relationship to assets or net income. Further research on whether these small firms are more apt to report employee theft and any market reaction to these reported thefts needs to be undertaken.

The third limitation involves the different requirements for disclosure of employee fraud across companies. Nonbanks are not required to disclose employee frauds, and therefore they may choose which frauds, if any, to disclose. Banks, on the other hand, must disclose all employee frauds that are larger than $1,000. Table 2 did find some weak evidence of a negative stock market reaction to disclosures by banks, but no evidence at all of any reaction to disclosures by other types of companies. It is not possible to tell whether these different results are due to the nature of the disclosure (voluntary versus mandated) or to the type of company (bank versus nonbank). Additional research in this area is needed.

Finally, this preliminary study hypothesized that the report of employee frauds would be "bad news" or "no news" to the firm. We did not examine the potential that such disclosure is "good news" to the firm. The discovery of the fraud may lead to changes in a firm's accounting control system to prevent future frauds which would favorably impact future cash flows. In addition, recovery from insurance or directly from the embezzler may reduce or negate any cost of the initial fraud. Thus, it is possible that some frauds are "bad news" to the company whereas others are "no news" or "good news." Consequently, our failure to find any market reaction may be due to our sample containing a mixture of both types of frauds. This is a topic that also needs to be addressed in future research.

Implications and Conclusion

This study's finding of no evidence that investors react negatively to disclosures of discovered cases of employee fraud has two important policy implications. First, it does not support the common argument that management should not make such disclosures because they result in negative publicity. There may be other reasons why management may not want to disclose all cases of discovered fraud (e.g., fear of being blamed for allowing it to occur and, therefore, losing one's job), but shareholder welfare does not appear to be a legitimate concern.

The second implication of our findings concerns the value of requiring companies to disclose cases of employee fraud. Public concern about fraud is increasing and there are calls for legislation to increase auditors' responsibility to find and report fraud (Goldstein and Dixon, 1989; Journal of Accountancy, 1993). Indeed, the public expects auditors to detect and report fraud (Humphrey et al., 1991). This study, however, finds no evidence that investors react to such disclosures. This finding suggests that there is no benefit to mandating disclosure of employee frauds. The limitations of this study noted above, however, suggest that additional research on investor reactions to fraud disclosures is warranted.

Future research should also examine whether other user groups (e.g., creditors) react to disclosures of employee fraud. In addition, there may be other benefits in requiring companies to disclose cases of employee fraud. For example, it has been argued that disclosure has a deterrent effect (Albrecht et al., 1994; Dycus, 1991; Richards and Knotts, 1989). Research investigating that possibility is also needed.

1 The analyses were also conducted on the sample of 48 including Enron and Chemical with no difference in results.

2 This approach adjusts the abnormal return by the standard error of the regression equation. Its purpose is to reduce the influence of large abnormal returns for companies where the regression model provides a poor fit.

References

12 Code of Federal Regulations: Banks and Banking. 1992. Washington, D.C.: Office of the Federal Register National Archives and Records Administration.

Albrecht, W.S., E.A. McDermott, and T.L. Williams. 1994. "Reducing the Loss of Fraud." Internal Auditor (February): 28-34.

American Institute of Certified Public Accountants. 1988. Statement on Auditing Standards No. 53: The Auditor's Responsibility to Detect and Report Errors and Irregularities. New York: AICPA.

Cushing, B.E., and M.B. Romney. 1993. Accounting Information Systems: A Comprehensive Approach. Reading, MA: Addison-Wesley.

Dycus, D.F. 1991. "Hear No Fraud, See No Fraud, There Is No Fraud." Corporate Controller (January/February): 13-17.

Financial Accounting Standards Board. 1980. Statement of Financial Accounting Concepts No. 2: Qualitative Characteristics of Accounting Information. Stamford, CT: FASB.

Gerlin, A. 1995. "How a Penney Buyer Made Up to $1.5 Million on Vendors' Kickbacks." The Wall Street Journal (February 7): A1, A12.

Goldstein, J.I., and C. Dixon. 1989. "New Teeth for the Public Watchdog: The Expanded Role of the Independent Accountant in Detecting, Preventing, and Reporting Financial Fraud." Business Lawyer 44 (February): 439-502.

Humphrey, C., S. Turley, and P. Mozier. 1991. "An Empirical Dimension to Expectations." Accountancy 108: (October): 30.

Journal of Accountancy. 1993. "Washington Update: AICPA Supports Fraud Detection Bill." (May): 15.

Karpoff, J.M., and J.R. Lott, Jr. 1993. "The Reputational Penalty Firms Bear from Committing Criminal Fraud." Journal Of Law and Economics 36: (October): 757-802.

Lary, B.K. 1989. "Why Corporations Can't Lock the Rascals Out." Management Review (October): 51-54.

O'Donoghue, J. 1987. "Strategies Found to be Effective in the Control of Computer Crime in the Forbes 500 Corporations." ACM SIG on Security, Audit & Control (Winter): 1-12.

Pae, P. 1989. "To Catch an Embezzler, Best Be Nimble, Be Quick." The Wall Street Journal (August 3): B1-B2.

Richards, T.C., and R. Knotts. 1989. "Top Management's View of Computer Related Fraud." ACM SIG on Security, Audit & Control (Winter): 34-43.

Straub, D.W., Jr., and W.D. Nance. 1990. "Discovering and Disciplining Computer Abuse in Organizations: a Field Study." MIS Quarterly (March): 45-60.

Wells, J.T. 1990. "Six Common Myths About Fraud." Journal of Accountancy (February): 82-88.

-----. 1994. "The Billion Dollar Paper Clip." Internal Auditor (October): 32-38.

James M. Lukawitz Associate Professor of Accountancy University of Memphis

Paul John Steinbart Ernst and Young Distinguished Professor of Accounting Saint Louis University

Cite this article
Pick a style below, and copy the text for your bibliography.