Corporate fraud and their financial consequences: An empirical study for listed U.S. firms

(1)

University of Groningen

Faculty of Economics and Business

Corporate fraud and their financial consequences:

An empirical study for listed U.S. firms

Name: Vincent van den Broek Student number: S2813165 Supervisor: Lammertjan Dam

Number of words: 8,903

E-mail: v.a.g.van.den.broek@student.rug.nl

(2)

Corporate fraud and their financial consequences:

An empirical study for listed U.S. firms

Abstract

In this study, I examine the financial consequences of corporate fraud. I use a sample of 71 U.S. listed firms, that are identified by the SEC for manipulating their financial statements. Consistent with prior research, stock prices decline significantly after a fraud disclosure. This thesis shows an average decline of 7.76 % at the day of disclosure and a decline of 8.06 % for the day after for a 21-day event window. For the long-term, I test the financial consequences for the firms’ market value, short-term debt and total debt in the year of fraud disclosure and the three years thereafter. Corporate fraud only affects the firms’ market value negatively in the year of disclosure. For the short-term debt and total debt, corporate fraud negatively affects the firm in the year of disclosure and the two years thereafter. These findings with regard to debt, indicates that funders see the perceived corporate fraud as a risk and restrict the availability of financing sources. The third year after fraud disclosure, corporate fraud does not affect short-term debt and total debt anymore.

JEL Classification: G14, G32, G38

(3)

1. Introduction

This thesis investigates the short- and long-term financial consequences of fraud. This is of particular importance since fraud is still a serious problem of modern business. Every two years, the Association of Certified Fraud Examiners (ACFE) publish a report on occupational fraud and abuse. In their most recent survey of 2016, they had analyzed 2,410 fraud cases in 114 countries that caused a total loss of more than $ 6.3 billion. 23 % of those cases resulted in a loss of at least $ 1 million and of the categories of occupational fraud, financial statement fraud still caused the greatest median loss per scheme1_.

It is well known that fraud can harm firms. But most prior research related to the financial consequences of corporate fraud analyze the reason, mechanism and characteristics (i.a. Dechow et al. (1996); Farber (2005); Burns and Kedia (2005); Johnson et al. (2009)) and the following negative stock market reaction ( i.a. Feroz et al (1991); Dechow et al. (1996)). However, the financial effects after the short-term market reaction have not been a subject in many studies. Only a few studies concerning consequences after a fraud revelation have been carried out. Chen et al. (2011) investigated the long-term financial consequences of corporate fraud on bank loans in China. Where others like Agrawal et al. (1999) examine the effects of fraud disclosure on subsequent changes in the board of directors and top management.

Despite the fact that fraud still occurs in firms, little is known about the financial

consequences for the long-term. More specific what happens with the market value, short-term debt and total debt in the long run. Therefore this thesis gives an answer to the following research question: “What are the financial consequences for firms’ market value and debt after a fraud disclosure and the years thereafter?”

In order to answer this question, I do two types of analysis. First, I perform an event study to measure the short-term financial consequences. Here I assume a semi-strong efficient market where all publicly available information is reflected in the stock price. Based on stock returns it is possible to see in an event study, if certain events have an influence on the value of a firm. In this case the events are the revelation of corporate fraud for 71 U.S. listed firms who

1_{ACFE: “Report to the nations on occupational fraud and abuse, 2016 global fraud study”,}

(4)

manipulate their financial statements, taken from the Auditing Enforcement Releases database, during the years 2000 – 2013. Second, I execute an unbalanced panel data regression on the same sample to measure, the long-term financial consequences for firms after a fraud disclosure. I test the relation between fraud and the market value, short-term debt, total debt for a period that starts in the year of disclosure until three years thereafter. The results show that the stock prices strongly decreases after a fraud disclosure in the short term. In the long run corporate fraud negatively affects the market value and debt. Fraud affects the firms’ market value only in the year of fraud revelation and fraud influence firms’ debt in the disclosure year and the two years thereafter.

As a contribution to the literature, this thesis provides new evidence on the financial consequences of corporate fraud. Where previous papers have focused on the reason, mechanism and characteristics and stock market reaction, this thesis investigates the long-term effects of corporate fraud on the firms’ market value and debt in the U.S. Further, this thesis shows how long corporate fraud affects firms’ market value and debt.

(5)

2. Literature and hypothesis development

This section, presents related literature and main hypotheses. The literature focuses mostly on the characteristics, corporate governance mechanisms of fraudulent firms and the short-term reaction of the market. Fraud occurs in many industries all over the world, where fraudulent misreporting happen nowadays more in high-growth industries (Qiu and Slezak, 2012). Qiu and Slezak (2012) show that fraud appears in new economy (high-tech) industries with considerable growth potential. Further, they find that there is a strong fraud incentive in good times, but fraud detection more likely occurs for high-growth industries when they slump.

Povel et al. (2007) results are similar regarding good times. They study the relationship between fraud and monitoring decisions. Povel et al. (2007) show that the link between fraud and good times becomes stronger as the monitoring costs decreases. Monitoring costs

decreases, when investors have reasonable expectations and their monitoring focus moves to companies that have to deal with negative public information. The companies that are no longer monitored can therefore commit fraud more easily. At the end of a boom the fraud peaks and since investors do not monitor, companies can misuse their information to commit fraud.

Another characteristic that determines fraud is corporate governance. Dechow et al. (1996) examine the capital market consequences of earning manipulation for firms subject to SEC enforcement actions issued between 1982 -1992. They find that an important motivation to commit fraud is the desire to attract additional financing at low cost. According to Dechow et al. (1996), this happens more likely in firms where the board of directors is dominated by insiders and where there is less likely an audit committee. A later study by Wang (2011), supports the outcome that additional financing is an incentive to commit fraud. Wang (2011) created an empirical model, which predicts that fraud will be committed in the case there is easy access and a strong need for external capital.

(6)

more fraudulent relative to their control sample, Beasley (1996) did not find a significant relationship between duality and fraud. Dechow et al. (1996) also mentioned that fraudulent firms don’t have an audit committee, Beasley (1996) tests the presence of an audit committee and he finds that they do not seem to affect the likelihood of financial statement fraud.

Beasley (1996) also examine the multiple directorships of outside members and the amount of time that they can spend on the company. When those directors participate in more boards and get more responsibilities, it is more likely that the probability of fraud increases. This in contrast with Ferris et al. (2003), who did not find a relation between fraud and multiple directorships performance of outside members.

Besides the interests of the company, managers also have their own personal interests to commit fraud. According to many authors, there is an important role for compensation schemes. Burns and Kedia (2005) provide evidence that there is a positive association between the use of stock-based compensation and fraudulent misreporting. By using

compensation schemes, it ensures that managers do their job, but it can also create incentives to inflate share prices and manipulate financial reports. Those results are in line with Efendi et al. (2005) who state that when the CEO has a large amount of “in-the-money” stock options, the tendency of misstatement significantly increases. Johnson et al. (2009) researched the managerial incentives that lead to fraudulent behavior. They find that the likelihood of fraud is unrelated to the incentives from restricted stock, unvested- and vested options, but is positively related to the incentives from unrestricted stock.

The motives for firms to operate in an unethical manner, can ultimately lead to a disclosure of misconduct. According to Dyck et al. (2010), the most prominent actors “who blows the whistle” are not the SEC or the auditors, but the media, industry regulators and employees. A huge strand of literature has studied the effect of a fraud disclosure on the accused firm and their short-term financial consequences. The results of those studies are one: The abnormal returns for firms are negative (and significant) after the revelation of corporate fraud (Feroz et al. (1991); Dechow et al. (1996); Anderson and Yohn (2002); Palmrose et al. (2004)).

(7)

violations. Those declines are similar to that reported by Dechow et al. (1996), whose result shows on average a nine percent decline on the disclosure date for a subset of firms who manipulated their earnings. Like Feroz et al. (1991), Dechow et al. (1996) also obtain their data returns from CRSP, but they estimated the abnormal returns using the value-weighted-with-dividends index. Palmrose et al. (2004) examine the reaction of the market to a sample of 403 restatements from 1995 – 1999. They show, using market-adjusted abnormal returns, a negative market reaction to restatement announcements of – 9.2 percent over a two-day event window. According to their paper, most of the negative returns are associated with

restatements involving fraud. The sample contains 83 fraud observations and has on average a negative return of 20 percent for a two-day event window. That result is consistent with Anderson and Yohn (2002), who find that the returns for fraud restatements are more negative than the returns from restatements related to other issues. Over a seven-day announcement period the fraud restatements in their study, show a negative return of 12.85 percent. Palmrose et al. (2004) use the Accounting and Auditing Enforcement Releases (AAERs) database for the initial announcement date to see how the market reacts to fraud. However, they replicate those results using press release information to identify fraudulent misstatements since Feroz et al. (1991) states that those AAERs usually takes several years to eventuate. Finally, Karpoff and Lott (1993) include in their study market reactions to various types of fraud and document that reports of financial reporting fraud do trigger significant returns ( average -5 percent), whereas violations of regulations do not trigger significant returns (average -1 percent).

Literature shows that after a fraud disclosure the stock returns for fraudulent firms

significantly declines. To test if the effect is the same for the returns of fraud firms in this thesis, I empirically tests the following hypothesis:

H1: After the revelation of corporate fraud, the stock returns for fraudulent firms shows negative financial effects.

(8)

that the reputation loss is greater than the legal penalty for committing fraud. Only 6.5 percent of the loss represents court imposed costs, and 1.4 percent of the loss are for the penalties and fines. Further, they find that for private parties the external effects of corporate fraud are small and that firms earnings declines after a fraud announcement.

The outcomes of Karpoff and Lott (1993) are consistent with Karpoff et al. (2008). They examine 585 firms targeted by the SEC for financial misrepresentation from 1978 – 2002 and they find that not regulators, but the market imposes the highest penalties. Karpoff et al. (2008) find that; “ for each dollar that a firm misleadingly inflates its market value, those firms will lose, on average, this dollar when the fraud has occurred. On top of that, the firm loses $ 0.36 due to expected legal penalties $ 2.71 due to loss reputation. The loss reputation is even bigger for firms that survive the enforcement process and that loss is estimated at $ 3.83”. However, Karpoff et al. (2005) find the opposite for firms when they violate

environmental regulations. In that case, firms receive more legal and administrative penalties than reputation penalties. Armour et al. (2017) study a more recent British sample for

punishments of misconduct. Other than the US firms the enforcement process only contains one announcement. That announcement takes place when the British Financial Services Authority (FSA) concluded the research towards the accused firm and the penalty is set. They find that the reputational losses, measured in stock prices, for punished firms are nearly nine times the size of the fines that are imposed by the FSA. Further, Armour et al. (2017) find that those reputational losses are confined to misconduct that directly affects customers or investors, but not third parties.

The literature shows that the firm’s market value is affected after a revelation of corporate fraud. But no single study shows how long corporate fraud affects the market value. Therefore, this study tests the following hypothesis:

H2: In the long run, a corporate fraud disclosure has negative financial consequences for the firms’ market value.

According to Dechow et al. (1996), firms that commit fraud experience a significant increase of their cost of capital. Fraud allegations can lead to a revision of existing contracts, including bank loans, which is for firms a major source of financing. The bank loans gives us an

(9)

requirements, etc.). The Literature focus mostly on the restatements of financial statements and not on corporate fraud in particular. In case firms have to restate their financial

statements, banks need to revalue firms because the previous information was incorrect.

Graham et al. ( 2008) study the effects of financial restatements on bank loan contracting between January 1997 and June 2002. They use a sample of 237 firms that restated their financials and 19 of them allegedly committed fraud. They mention that about 10 percent of all the listed firms restated their financial statements at least once and the average market value increased from $ 500 million in 1997 to $ 2 billion in 2002. Graham et al. use a

regression model to examine the effect of restatement on the cost of bank debt. They find that the restatements by firms lead to a 42.6 % increase in the loan spread. Graham et al. (2008) also tested the effect of fraud on the cost of bank debt by including an interaction variable, that includes a fraud dummy, in the regression model. The outcome shows that fraudulent restatements have a much greater effect, with a 68.9 % ( 42.6 % + 26.3 %) increase of the loan spread. Chen et al. (2011) investigated the long-term economic consequences of corporate fraud in China. By studying 677 Chinese A-share market firms between 2000 – 2007, they test the effects of punishment for corporate fraud on the size and interest rates of bank loans in the year of fraud disclosure and the two years thereafter. They use two

regression models, one with- and one without the interaction variable Fraud * Performance. Chen et al. (2011) find that the loan interest rates significantly increases after a fraud disclosure for Chinese corporate fraud firms. Furthermore, after the revelation of corporate fraud, the bank loans become significantly lower than for non-fraudulent firms. Those results are consistent with the outcomes of Graham et al. (2008), who also find that after financial restatements the number of loan lenders declines. Moreover, Chen et al. (2011) show that corporate fraud also affects bank loans by influencing the relationship between bank loans and corporate performance indirectly.

All in all, literature has studied the financial consequences for fraudulent firms but to a limited extent. Therefore, this study tests the following hypothesis:

(10)

3. Methodology & Data

This chapter covers the methodology and the data selection process. Firstly, I describe the methodology in section 3.1, which consists of an event study and a regression analysis. In subsection 3.1.1. I explain the event study to measure the short-term effects of fraud and in subsection 3.1.2. I clarify the regression analysis to measure the long-term effects of fraud. Section 3.2 contains the sample selection process.

3.1. Methodology

3.1.1. Measuring short-term effects of fraud

Event studies are one of the most frequently used methodologies to measure the impact of an event on a security return. By using an event study it is possible to examine if the firms’ abnormal returns during the ‘event window’ are positive or negative. An event window explains the amount of trading days that is examined before and after the event date (Seiler 2004, p. 424). In addition to the event window, I also use an estimation window, which is the period prior the event window, where there was no event and I estimate expected returns. For the fraud cases, I collect data from CRSP, where stock returns of all fraudulent firms are available. To evaluate the abnormal returns around a SEC fraud disclosure, I use the market model as a benchmark to calculate the abnormal returns. In equation (1) the market model is represented (Fama et al. 1969; Campbell et al. 1997):

(1)

(11)

(2)

Where N is the number of firms over which abnormal returns are averaged on day t. To draw overall inferences, the observations of the abnormal returns have to be aggregated. Therefor I use the cumulative abnormal return (CAR). The CAR is computed by summing the portfolio abnormal returns across time, where t1 and t2 are considered as the event window and thereby the next equation (3) will be used:

(3)

For the event study, I examine different lengths of event windows. I utilize the standard three-day event window (-1, +1) but also a two three-day (-0, +1) and twenty-one three-day (-10, +10). By using different lengths, it is possible to see if there is a potential relation between the length of the window and the abnormal returns. According to MacKinlay (1997), a longer event

window may acquire information prior the announcement and the day after the announcement which captures the effects on stock prices when the market closes. Finally, to test if the mean abnormal returns are significant, I perform a parametric test. I perform this parametric test for all the fraud cases without eliminating eventual outliers. This since outliers are not identified by the absolute size of the observation of an individual variable, but by the size of the residual from the regression model (Sorokina et al. 2013).

3.1.2. Measuring long-term effects of fraud

(12)

Variable Symbol Definition

Market value renewal CMV The difference between market value in year t + 1 and year t

Corporate fraud involvement Fraud yr. 0 Dummy variable that takes a value of 1 in the year of fraud disclosure, and 0 otherwise Fraud yr. 1 Dummy variable that takes a value of 1, one year after fraud disclosure, and 0 otherwise Fraud yr. 2 Dummy variable that takes a value of 1, two years after fraud disclosure, and 0 otherwise Fraud yr. 3 Dummy variable that takes a value of 1, three years after fraud disclosure, and 0 otherwise Performance ROA Return on total assets in year t

Growth SalesGrowth The financial ratio of the difference between sales in year t + 1 and year t Common shares outstanding CSO The difference between outstanding shares in year t + 1 and year t Common ordinairy shareholders COS The difference between number of shareholders in year t + 1 and year t Return on equity ROE Return on equity in year t

Size SIZE Natural logarithm of total assets in year t

financial consequences, that consists of four research models, whereby two of those models includes an interaction effect variable. The research models are as follows:

(4) (5) (6) (7)

Research models (4) & (6) examine the direct influence of corporate fraud on the dependent variables. In model (5) & (7) there is an interaction variable included between ROA and the dummy fraud to test if there is a combination effect on the proxies. Model (4) & (5) examine the influence of corporate fraud on the market value. Table I defines the variables I use for the market value model.

Table I

Definitions market value model variables

(13)

Variable Symbol Definition

Total debt renewal CTD The difference between total debt in year t + 1 and year t, divided by total assets year t CSD The difference between short term debt in year t + 1 and year t, divided by total assets year t Corporate fraud involvement Fraud yr. 0 Dummy variable that takes a value of 1 in the year of fraud disclosure, and 0 otherwise

Fraud yr. 1 Dummy variable that takes a value of 1, one year after fraud disclosure, and 0 otherwise Fraud yr. 2 Dummy variable that takes a value of 1, two years after fraud disclosure, and 0 otherwise Fraud yr. 3 Dummy variable that takes a value of 1, three years after fraud disclosure, and 0 otherwise Performance ROA Return on total assets in year t

Interest Interest Interest for debt (based on an American bond) in year t

Growth SalesGrowth The financial ratio of the difference between sales in year t + 1 and year t Operating cash flow CFO Operating cash flow divided by total assets in year t

Leverage LEV Total debt divided by total assets in year t Size SIZE Natural logarithm of total assets in year t

CMV is the change in market value in year t, and the relation of interest is with the fraud dummy variable. In case there is a fraud occurrence the fraud dummy should be negative and significant. I put four dummy variables into the model. I use these dummy variables in the “year of fraud disclosure” , “ one year after fraud disclosure”, “two years after fraud disclosure”, and “ three years after fraud disclosure”, respectively. The other variables are basic characteristics of the firms.

Model (6) & (7) explains the financial consequences of corporate fraud on debt, where Table II defines the variables. Here are CTD and CSD the change in total debt and the change in short-term debt in year t, and the relation of interest is with the fraud dummy variable. Chen et al. (2011) use a research model that is almost the same, but instead of total (short-term) debt they use total (short-term) bank loans as a dependent variable. Since Compustat do not (very well) represents bank loans & related interest rates for companies, I implement short-term debt and total debt as a dependent variable. In case there is a fraud occurrence the fraud dummy should be negative and significant. Further the same guidelines should be adopted as aforementioned for these regression models.

Table II

Definitions short-term debt and total debt model variables

(14)

3.2. Data selection

The sample of fraudulent companies is hand-collected from the United States Securities and Exchange Commission (SEC). The SEC takes enforcement actions against companies who have violated requirements of the securities exchange act of 1934. They publish these enforcement actions in the Accounting and Auditing Enforcement Releases (AAERs)

database2_{. I include firms in the corporate fraud sample, when those firms are accused by the} SEC for violating rule 10(b)-5. Rule 10(b)-5 deems that it is illegal for firms to manipulate their financial statements.

The procedure to select the sample of fraud cases is summarized in Figure I. At the start of this thesis, the AAERs database contains 2688 cases of SEC enforcement actions issued between halfway October 1999 and halfway September 2017. Firstly, I eliminate 110 AAERs since 2017 is a current year. I reduce the sample further by 2188 AAERs, due to auditor violations, no manipulated financial statement cases, incomplete SEC data and duplicates. Another 65 AAERs are dropped because some firms are not listed at the NYSE or NASDAQ. Of the remaining 325 AAER cases, firms with more than one fraud related announcement within five years are deleted. Finally, I exclude 139 firms for several reasons, like mergers or acquisitions shortly after the announcement or no available data in CRSP and Compustat. This results in a final sample of 71 firms and these are presented in Appendix A. I consult the webpage of the Securities Class Action Clearinghouse (SCAC)3, to determine the fraud detection date for the fraud firms in the sample. Following Feroz et al. (1991), if it is not possible to find an identification date at SCAC, I use the AAER date as a proxy for the detection date.

2_{The AAERs database provide varying degrees of detail on the nature of the misconduct, the}

individuals and entities involved and their effect on the financial statements. SEC website:

https://www.sec.gov/divisions/enforce/friactions.shtml

(15)

Figure I

Identification of 71 fraud firms

Selection of 71 fraud cases for stock listed companies subject to enforcement actions by the SEC for the period 1999 – 2017.

Number of AAERs issued between halfway October 1999 and halfway

September 2017 2,688

Less:

- AAERs of current year (2017) (110)

- Auditors for their violations of auditing standards, no manipulated financial statement cases, incomplete SEC

data & duplicates (2,188)

- Companies not listed on the NYSE or NASDAQ (65)

- Cases of firms with more than one SEC announcement

within five years (115) - Firms who were merged or acquired, went bankrupt,

not available at CRSP and Compustat and several firms in

same industry in the same period (139)

Total number of fraud firms included in the study 71

(16)

Distribution of fraud firms by SIC Code

Two-digit SIC industry classification

SIC Industry obs SIC Industry obs

13 Oil and Gas Extraction 3 49 Electric, Gas and Sanitary Services 1 15 Building Construction - General Contractors & Operative Builders 1 51 Wholesale Trade - Nondurable Goods 1 17 Construction - Special Trade Contractors 1 53 General Merchandise Stores 2

20 Food and Kindred Products 2 59 Miscellaneous Retail 1

27 Printing, Publishing and Allied Industries 1 60 Depository Institutions 5 28 Chemicals and Allied Products 6 62 Security & Commodity Brokers 1

33 Primary Metal Industries 1 63 Insurance Carriers 4

35 Industrial and Commercial Machinery and Computer Equipment 7 64 Insurance Agents, Brokers and Service 2

36 Electrical 7 67 Holding and Other Investment Offices 2

37 Transportation Equipment 3 72 Personal Services 1

38 Measurement instruments 4 73 Business Services 11

45 Transportation by Air 1 79 Amusement and Recreation Services 1 48 Communications 1 87 Engineering, Accounting, Research, Management & Related Services 1

Total 71

Figure II

Distribution of fraud cases by year

The graph shows the number of enforcement actions per year. The vertical- and horizontal axis displays the number of fraud cases and the year of release, respectively. There are 71 cases studied during this period.

Table III represents the industry classification of the 71 firms. The industry that has the largest representation is Business services ( SIC 73) with 11 observations; followed by Industrial and Commercial Machinery (SIC 35) and Electrical (SIC 36) with each seven observations.

Table III

Distribution of fraud cases by industry

(17)

Variable Number of observations Mean Median Min. Max. Std. in thousands

Change market value 96,8 0.19 -0.01 -0.88 4.78 0.67

Change total debt 92,8 0.02 0.00 -0.35 0.43 0.08

Change short-term debt 92,8 0.00 0.00 -0.15 0.16 0.02

Performance 106,1 -0.03 0.02 -1.68 0.33 0.22

SalesGrowth 93,1 0.18 0.09 -0.68 4.59 0.46

Common shares outstanding 102,9 0.14 0.01 -0.61 6.03 0.47

Common ordinairy shareholders 69,2 0.12 -0.03 -0.93 16.00 0.97

Return on Equity 105,4 0.03 0.02 -3.70 3.26 0.48

Leverage 107,2 0.15 0.02 0.00 0.97 0.20

Natural Logarithm of Total Assets 105,6 6.35 6.32 1.48 12.07 2.09

Interest 123,6 4.11 4.30 0.00 6.60 1.47

Operating cash flow 106,1 0.03 0.0568 -1.15 0.37 0.17

For the years before and after the fraud disclosure, I use a clean sample from Compustat North America fundamentals annual database to measure the financial consequences for fraudulent firms. The clean sample period is from January 1996 till December 2016. This since firm fraud occurs years before the disclosure date. As was mentioned in the sample selection procedure, one of the requirements is that the fraudulent firms are listed on the NYSE or the NASDAQ. Therefore the clean sample only consists of (former) publicly traded firms on those indexes. In total the clean sample features 123,560 observations for 12,240 companies. On the basis of this clean sample, I do an unbalanced panel data regression. To do this regression I use 12 variables. To control for outliers these continuous variables, except Interest, are winsorized at lower and upper 1% levels. The change of the variables before and after winsorizing can be found in Appendix B. Table IV shows the firm-level financial characteristics for the 12,240 firms of the clean sample.

Table IV

Firm-level financial characteristics

The table displays the financial characteristics of the 71 fraudulent firms and the 12,169 companies of the clean sample for the period January 1996 – December 2016. The data is collected from Compustat North America fundamentals annual database. The variables in the table are the variables that are implemented in the research models.

(18)

4. The short-term and long-term effects of corporate fraud

In this chapter the financial consequences of corporate fraud are elaborated in the short-term and long-term. Firstly, I present the short-term effects after fraud disclosure in section 4.1. Afterwards, I show the long-term financial consequences on market value and debt in section 4.2.

4.1 Short-term effects of corporate fraud

As I mentioned in chapter 3, the event study that I execute consists of multiple event

windows. Of the 71 fraudulent firms, I exclude one firm from the event study since there are no returns available at the disclosure date and the two months thereafter in CRSP. Figure III presents the average abnormal returns of the firms who are accused of fraud for a 21-day event window. Ten days before the disclosure date the average abnormal returns slightly fluctuates around zero. On the event day itself the abnormal return drops on average with 7.76 %, and is statistically significant. For the day after the revelation the abnormal returns fall even at a higher rate (- 8.06 %) and are also statistically significant. Thereafter the market returns are fluctuating slightly around zero again.

Figure III

Mean abnormal returns of fraudulent firms

(19)

70 fraud accused firms

Event window Mean Min. Max. t-statistic

(0) -8.51% -70.35% 2.65% -5.58***

(0,+1) -14.33% -87.35 6.29% -6.10***

(-1,1) -15.61% -86.56% 10.76% -6.67***

(-10,10) -18.93% -148.92% 35.72% -5.22***

Table V shows the average cumulative abnormal returns (CAR) of the four different event windows and the associated parametric test. All the four CARs are negative and significant at a 1 % significance level. The one-, two-, three- and 21-day event windows are on average: - 8.51 %; -14.33 %; - 15.61 % and - 18.93 %, respectively. The negative CARs indicates that the firms value significantly decreases around or after a disclosure of corporate fraud.

Table V

CARs around the SEC fraud disclosure

Table VI reports the average cumulative abnormal returns (CAR) around the fraud disclosure by the SEC. Four different event windows are used. The sample consists of 70 fraud cases. The CARs are based on market model parameters, that are calculated over the period - 200 days to - 11 days before the announcement day. T-statistics at a significance level of 1 %, 5 % and 10 % are indicated by ***, ** and *, respectively.

For a robustness check I split the sample in two fraud cases, namely still active firms and inactive firms. The results are reported in Table VI. The 55 active firms who were involved with corporate fraud shows negative and significant CARs in the four event windows at 1 % significance level of respectively: - 8.67 %; - 15.05 %; - 16.27 %; and - 17.88 %. For the 15 inactive fraud firms, the CARs are also negative and significant but at a 5 % and a 10 % significance level. The CAR for the one-day event window is - 7.90 % at a 10 % significance level. The two-, three- and 21-day event window are significant at 5 % significance level and have a CAR of - 11.67 %; - 13.18 %; and - 22.77 %, respectively.

(20)

Event window Mean Min. Max. t-statistic 55 active firms (0) -8.67% -42.42% 2.65% -5.71*** (0,+1) -15.05% -70.99% 6.29% -5.93*** (-1,1) -16.27% -75.31% 10.76% -6.42*** (-10,10) -17.88% -83.48% 35.72% -5.13*** 15 inactive firms (0) -7.90% -70.35% -0.11% -1.73* (0,+1) -11.67% -87.35% 1.76% -1.96** (-1,1) -13.18% -86.56% 3.34% -2.25** (-10,10) -22.77% -148.92% 27.86% -2.00**

are negative and significant at a 1 % significance level and the outcomes can be found in Appendix C.

Table VI

Robustness: Sample split in active- and inactive fraud firms

Table VII presents the average cumulative abnormal returns for 55 fraud cases from active firms and 15 fraud cases for firms who are inactive. There are four event windows used and the cumulative abnormal returns are based on market model parameters, that are calculated over the period -200 days to -11 days before the announcement day. T-statistics at a significance level of 1 %, 5 % and 10 % are indicated by ***, ** and *, respectively.

Overall the outcomes of the event study support the first hypotheses. This means that there is significant evidence of negative abnormal returns on the day that a party discloses information of a firm regarding corporate fraud.

4.2 Long-term effects of corporate fraud

This section presents the outcomes of the unbalanced panel data regression for the long-term financial consequences of corporate fraud. Firstly, I examine the outcomes of the regression model with the proxy change in market value, and after that I discuss the outcomes of the proxy change in short-term debt and the proxy change in total debt.

(21)

CMV

Fraud yr. 0 Fraud yr. 1 Fraud yr. 2 Fraud yr. 3

Model (4) Model (5) Model (4) Model (5) Model (4) Model (5) Model (4) Model (5)

Variable Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

Fraud -0.282*** -0.283*** 0.059 0.054 0.202** 0.210** 0.004 0.013 (0.091) (0.092) (0.092) (0.093) (0.088) (0.089) (0.083) (0.085) Fraud*ROA -0.122 0.325 -0.633 -0.454 (0.981) (1.020) (0.817) (0.830) ROA 0.496*** 0.496*** 0.496*** 0.496*** 0.496*** 0.496*** 0.496*** 0.496*** (0.200) (0.200) (0.200) (0.200) (0.200) (0.200) (0.200) (0.200) SalesGrowth 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) CSO 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) COS 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) ROE 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) SIZE -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Constant 0.302*** 0.302*** 0.302*** 0.302*** 0.302*** 0.302*** 0.302*** 0.302*** (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) 0.131 0.131 0.131 0.131 0.131 0.131 0.131 0.131 Observations 61,945 61,945 61,945 61,945 61,945 61,945 61,945 61,945 Groups 7,160 7,160 7,160 7,160 7,160 7,160 7,160 7,160 Table VII

The long-term effects of corporate fraud on the market value

(22)

The coefficient of the relation between Fraud yr. 0 and CMV is - 0.282 and significant at a 1% significance level. This indicates that in the year of corporate fraud disclosure, the firms market value strongly decreases with - 28.2 %. For the years after the fraud disclosure there is no negative relation between the firms market value and fraud. After two years there is a significant relation, but this is positive and will not have a negative impact on the firms market value. The outcomes partly supports the second hypotheses, that there are negative financial consequences for the market value, but only in the year of fraud disclosure. The interaction variable Fraud * ROA is negative in Fraud yr. 0, Fraud yr. 2 and Fraud yr. 3 but not significant. This indicates that corporate fraud does not indirectly influences the relation between market value and performance.

Table VIII and IX give the estimation results of the relation between the change in total (short-term) debt and fraud. Table VIII shows the outcomes from the relation between the change in short-term debt (CSD) and corporate fraud. The coefficient of the relation between Fraud yr. 1 and CSD is - 0.010 and significant at a 1% significance level. This means that a year after fraud disclosure there is a small decline in debt, which could indicate that it is more difficult to get short-term money from financers. For the years thereafter there is a negative but not significant relation, which means that fraud does not have any influence anymore on borrowing short-term money.

(23)

The financial consequences after corporate fraud disclosure for debt, are partly in line with the findings of Chen et al. (2011). They find that in the year of fraud disclosure and the two years thereafter, the short-term bank loans and all the bank loans for firms in China are significantly negatively related. In their study the bank loans decrease by two to five percent. This study shows that fraud also influences U.S. listed firms for the long term, but debt (loans) is not for every year negatively related to corporate fraud.

(24)

CSD

Fraud -0.003 -0.002 -0.010*** -0.010*** 0.002 0.002 -0.004 -0.004 (0.003) (0.003) (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) Fraud*ROA 0.024 -0.009 -0.030 0.013 (0.021) (0.013) (0.024) (0.031) ROA 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Interest 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) SalesGrowth 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) CFO -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) LEV -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 SIZE 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) 0.008 0.008 0.008 0.008 0.008 0.008 0.008 0.008 Observations 85,374 85,374 85,374 85,374 85,374 85,374 85,374 85,374 Groups 9,663 9,663 9,663 9,663 9,663 9,663 9,663 9,663 Table VIII

The long-term effects of corporate fraud on short-term debt

(25)

CTD

Fraud -0.022** -0.025** -0.001 -0.001 -0.023** -0.023** 0.004 0.005 (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) (0.009) (0.010) Fraud*ROA -0.093 0.005 -0.013 -0.052 (0.068) (0.036) (0.079) (0.100) ROA 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Interest 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) SalesGrowth 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) CFO -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) LEV 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) SIZE 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) 0.106 0.106 0.106 0.106 0.106 0.106 0.106 0.106 Observations 85,545 85,545 85,545 85,545 85,545 85,545 85,545 85,545 Groups 9,678 9,678 9,678 9,678 9,678 9,678 9,678 9,678 Table IX

The long-term effects of corporate fraud on total debt

(26)

5. Conclusions & recommendations

This thesis studies the financial consequences of a fraud disclosure on the stock returns in the short run and the change in market value and change in debt in the long run. It uses a sample of 71 (former) listed U.S. firms who manipulated their financial statements.

The results in the short term shows that the revelation of corporate fraud leads to significant large negative abnormal returns for the fraudulent firms. This result is in line with those found elsewhere in the literature (i.a. Feroz et al (1991); Dechow et al. (1996)).

In the long run, corporate fraud only negatively affects the firms’ market value in the year of fraud disclosure. This indicates that investors punish firms after the announcement that firms manipulated their financial statements.

Short-term debt is negatively affected in the first year after fraud disclosure and this could indicate that it is more difficult for firms to get short-term loans. For total debt the outcomes are different than for short-term debt. Total debt is negatively related to corporate fraud in the year of- and the second year of fraud disclosure. Since total debt also consists of long-term debt, this could mean that it is more difficult to close long-term debt contracts. So financers see the disclosure of fraud as a problem to provide loans. But from year three and onwards, fraud does not seem to have any more influence on raising capital for firms.

The outcomes of debt are slightly different than the results by Chen et al. (2011). In the study of Chen et al. (2011) the firms already have issues with getting bank loans in the year of fraud disclosure, and this thesis shows there are consequences a year after the disclosure for short-term debt. This indicates that the loan (debt) issuers in the U.S. will not reduce the short-short-term loans directly after a fraud disclosure in comparison to China.

(27)

other words, all the variables that were downloaded from Compustat were incomplete. Further, the clean sample contains probably more fraudulent cases but those are not recognized.

This thesis utilizes the AAERs database to investigate corporate fraud. For future research it is recommendable to use a clean sample as a starting point to investigate the financial

consequences of corporate fraud on fraudulent companies. In this case you have a completely unbiased overview of stock listed firms and you can look if those firms have been involved with corporate fraud. Added to that, it is maybe then possible by consulting the financial reports to fill in the missing variable values.

Since bank loans and the related interest rates for firms are not (good) represented in Compustat, I have used short-term- and total debt as dependent variables for this thesis. An interesting follow-up study might be, to measure the long-term financial consequences on bank loan contracting for fraudulent firms in the U.S. In this way it is possible to see how corporate fraud can affect financing contracts between a fraudulent firm and its bank.

(28)

References

• Agrawal, A., Jaffe, J.F., Karpoff, Jonathan M., 1999. Management turnover and governance changes following the revelation of fraud. Journal of Law and Economics 42, 309 – 342.

• Anderson, K., Yohn, T., 2002. The effect of 10K restatements on firm value, information asymmetries, and investors' reliance on earnings. Working paper. Georgetown University.

• Armour, J., Mayer, C., Polo, A. 2017. Regulatory sanctions and reputational damage in financial markets. Journal of Financial and Quantitative Analysis, vol. 52, issue 04, 1429 - 1448.

• Beasly, M.S., 1996. An Empirical Analysis of the relation between the Board of the Director Composition and Financial statement Fraud. The Accounting Review 71 (October), 443 – 465.

• Burns, N., Kedia, S., 2005. The impact of performance-based compensation on misreporting. Journal of Financial Economics, 2006, vol. 79, issue 1, 35 - 67.

• Campbell, J., Lo, A.H., McKinlay, C., 1997. The Econometrics of Financial Markets Princeton University Press, Princeton, NJ.

• Chen, Y., Zhu, S., Wang, Y., 2011. Corporate fraud and bank loans: Evidence from China. China Journal of Accounting Research 4, 155 – 165.

• Dechow, P.M., Sloan, R.G., Sweeney, A.P., 1996. Causes and Consequences of Earnings Manipulation: An Analysis of Firms Subject to Enforcement Actions by the SEC. Contemporary Accounting Research, 13(1), 1 - 36.

• Dyck, A., Morse A., Zingales L., 2010. Who Blows the Whistle on Corporate Fraud? The Journal of Finance 65, 6 (2010): 2213 - 2253.

• Efendi, J., Srivastava, A., Swanson, E.P., 2005. Why do corporate managers misstate financial statements? The role of option compensation and other factors. Journal of financial economics, 667 – 708.

• Fama, E.F., Fisher, L., Jensen, M.C., Roll, R., 1969. The Adjustment of Stock Prices to New Information. Int. Econ.Rev., 10(1), 1 - 21.

(29)

• Feroz, E.H., Park K., Pastena S., 1991. The financial and market effects of the SEC’s accounting and auditing enforcement releases. Journal of Accounting Research 29, 107 – 142.

• Ferris, S.P., Jagannathan, M., Pritchard. A.C., 2003. Too busy to mind the business? Monitoring by directors with multiple board appointments. The Journal of Finance 58, 1087–1111.

• Graham, J., Li, S., Qiu, J., 2008. Corporate misreporting and bank loan contracting. Journal of Financial Economics 89(1), 44 – 61.

• Johnson, S.A., Ryan Jr, H.E., Tian, Y.S., 2009. Managerial Incentives and Corporate Fraud: The Sources of Incentives Matter. Review of Finance 13, 115 -145.

• Karpoff, J. M., Lee, D. S., Martin, G.S., 2008. The Cost to Firms of Cooking the Books. Journal of Financial and Quantitative Analysis 43, 581 - 611.

• Karpoff, J. M., Lott Jr, J.R., 1993. Reputational penalty firms bear from committing criminal fraud. Journal of Law & Economics, vol. 36., 757 – 802

• Karpoff, J. M., Lott Jr, J.R., Wehrly, E.W., 2005. The Reputational Penalties for Environmental Violations: Empirical Evidence. Journal of Law and Economics, 48(2), 653-675.

• MacKinlay, A.C., 1997. Event Studies in Economics and Finance, Journal of Economic Literature, 35, 13 - 39.

• Palmrose, Z.-V., Richardson, V., Scholz, S., 2004. Determinants of market reactions to restatement announcements. Journal of Accounting and Economics 37, 59 – 89. • Poval, P., Singh, R., Winton, A., 2007. Booms, Busts, and Fraud. Review of Financial

Studies, 20(4): 1219 - 1254.

• Qiu, B., Slezak, S.L., 2008. The Strategic Interaction between Committing and Detecting Fraudulent Reporting. Unpublished Working paper.

• Seiler, M.J., 2004 Performing Financial Studies: A Methodological Cookbook. Pearson Education, Inc., New Jersey, USA.,

• Sorokina, N., Booth, D.E., Thornton Jr. J.H., 2013. Robust Methods in Event Studies: Empirical Evidence and Theoretical Implications. Journal of Data Science, 11, 575 – 606.

(30)

# Year Release date SIC Ticker Company Market Active

1 2000 31-3-2000 6311 CNO Conseco, Inc. NYSE A

2 2000 27-10-2000 2086 KO The Coca-Cola Company NYSE A 3 2001 30-1-2001 2834 ELN Elan Corporation, plc NYSE I

4 2001 2-3-2001 3812 RTN Raytheon Company NYSE A

5 2001 29-3-2001 3662 CAMP California Amplifier, Inc. NASDAQ A 6 2001 17-4-2001 7373 BVSN BroadVision, Inc. NASDAQ A 7 2001 30-4-2001 5331 DG Dollar General Corporation NYSE A 8 2001 25-7-2001 3674 TUNE Microtune, Inc. NASDAQ I 9 2001 27-7-2001 4813 Q.2 Qwest Communications International Inc. NASDAQ I 10 2001 27-6-2001 5912 CVS CVS Caremark Corporation NYSE A

11 2001 30-7-2001 7372 MFE McAfee, Inc. NASDAQ I

12 2001 31-8-2001 7379 NTES NetEase.com, Inc. NASDAQ A 13 2001 14-12-2001 7370 TTWO Take-Two Interactive Software, Inc. NASDAQ A 14 2002 6-2-2002 3570 LTRX Lantronix, Inc. NASDAQ A 15 2002 14-2-2002 6798 SUI Sun Communities, Inc. NYSE A 16 2002 22-2-2002 7372 CA Computer Associates International, Inc. NASDAQ A 17 2002 22-3-2002 2834 BMY Bristol-Myers Squibb Company NYSE A 18 2002 30-4-2002 7373 GRB Gerber Scientific, Inc. NYSE I 19 2002 3-6-2002 1389 HAL Halliburton Company NYSE A 20 2002 16-7-2002 6021 COF Capital One Financial Corporation NYSE A 21 2002 5-8-2002 7311 IPG Interpublic Group of Companies, Inc. NYSE A 22 2002 12-9-2002 3678 TYC Tyco International Ltd. NYSE A 23 2002 3-10-2002 2834 SGP Schering-Plough Corporation NYSE I

24 2002 4-10-2002 6021 CMA Comerica, Inc. NYSE A

25 2002 16-10-2002 4522 AAWW Atlas Air Worldwide Holdings, Inc. NASDAQ A

26 2002 29-10-2002 3341 OMG OM Group, Inc. NYSE I

27 2002 13-12-2002 4931 NWE NorthWestern Corporation NYSE A 28 2003 17-10-2003 2834 CBM Cambrex Corporation NYSE A 29 2004 14-4-2004 6020 PFBI Premier Financial Bancorp, Inc. NASDAQ A 30 2004 27-5-2004 7373 BCSI Blue Coat Systems, Inc. NASDAQ I 31 2004 7-6-2004 1389 KEG Key Energy Services NYSE A 32 2004 30-6-2004 5122 CAH Cardinal Health, Inc. NYSE A 33 2004 22-7-2004 2851 FOE Ferro Corporation NYSE A 34 2004 2-8-2004 1731 IESC Integrated Electrical Services, Inc. NASDAQ A 35 2004 30-8-2004 7353 URI United Rentals, Inc. NYSE A

36 2004 14-10-2004 6411 AON Aon Corporation NYSE A

Appendix

A. Sample of 71 fraudulent firms

(31)

37 2004 27-10-2004 7372 AZPN Aspen Technology, Inc. NASDAQ A 38 2005 16-5-2005 1389 WG Willbros Group, Inc., NYSE A 39 2005 11-7-2005 6331 RNR RenaissanceRe Holdings Ltd. NYSE A

40 2005 21-10-2005 6035 FBP First BanCorp NYSE A

41 2005 3-11-2005 3820 FARO Faro Technologies, Inc. NASDAQ A 42 2005 14-12-2005 3711 NAV Navistar International Corporation NYSE A

43 2006 7-2-2006 3519 CMI Cummins Inc. NYSE A

44 2006 20-3-2006 3550 BRKS Brooks Automation, Inc. NASDAQ A 45 2006 12-6-2006 7310 MWW Monster Worldwide, Inc. NASDAQ A

46 2006 14-7-2006 3670 BRCM Broadcom Corp. NASDAQ I

47 2007 27-3-2007 1531 BZH Beazer Homes USA, Inc. NYSE A

48 2007 8-5-2007 3663 MSI Motorola, Inc. NYSE A

49 2007 5-9-2007 5311 SKS Saks Incorporated NYSE I

50 2007 24-10-2007 6324 WCG WellCare Health Plans, Inc. NYSE A 51 2007 19-12-2007 6020 HBAN Huntington Bancshares, Inc. NASDAQ A 52 2008 23-1-2008 3845 ARTC Arthrocare Corporation NASDAQ I 53 2008 24-4-2008 3842 SMA Symmetry Medical, Inc. NYSE I

54 2008 4-9-2008 3531 TEX Terex Corporation NYSE A

55 2008 24-9-2008 7993 BYI Bally Technologies, Inc. NASDAQ I 56 2008 29-12-2008 7200 STEI Stewart Enterprises Inc. NASDAQ I 57 2009 22-1-2009 3711 GM General Motors Corporation NYSE A 58 2009 31-7-2009 8742 HURN Huron Consulting Group Inc. NASDAQ A

59 2010 15-1-2010 3651 KOSS Koss Corporation NASDAQ A

60 2010 1-6-2010 3674 CSIQ Canadian Solar NASDAQ A

61 2010 29-9-2010 3572 FALC Falconstor NASDAQ A

62 2011 20-1-2011 6411 LPHI Life Partners NASDAQ A

63 2011 17-3-2011 2834 MED MEDIFAST INC NYSE A

64 2011 31-3-2011 3577 LOGI Logitect International NASDAQ A

65 2011 1-11-2011 2090 DMND Diamond Foods NASDAQ A

66 2011 2-11-2011 6311 PRU Prudential Financial, Inc. NYSE A

67 2012 9-2-2012 6211 INTL Intl FCStone NASDAQ A

68 2012 11-6-2012 6719 MLNK Modulink NASDAQ A

69 2012 18-12-2012 2711 TST TheStreet, Inc. NASDAQ I

70 2013 20-2-2013 3578 PAY Verifone NYSE A

(32)

Variable Obs. Mean Std. Dev. Min. Max

Before CMV 98,7 0.89 48.86 -1.00 10,63

After CMV 96,8 0.19 0.67 -0.88 4.77

Before CTD 94,7 -302.30 9,110,645.00 -28,000,000.00 56.76

After CTD 92,8 0.02 0.08 -0.35 0.43

Before CSD 94,7 17.29 267,159.00 -294,367.90 510,690.00

After CSD 92,82 0.00 0.24 -0.15 0.16

Before ROA 108,2 -0.77 2.69 -373.29 226.00

After ROA 106,1 -0.03 0.22 -1.68 0.33

Before SalesG 95,0 0.91 51.58 -58.77 11,651.00

After SalesG 93,1 0.18 0.46 -0.68 4.59

Before CSO 105,0 91.68 7,323.12 -1.00 1,312,238.00

After CSO 102,9 0.14 0.47 -0.61 6.03

B. Winsorized variables for regression models

These tables shows the winsorizing process for the variables that are used for the unbalanced panel data regression. The variables are winsorized at lower and upper 1 % levels. Observations are in thousands.

Change in Market value (CMV)

Change in total debt (CTD)

Change in short-term debt (CSD)

ROA (Performance)

SalesGrowth

(33)

Before COS 70,6 86.84 6,813.35 -1.00 1,175,489.00

After COS 69,2 0.12 0.97 0.93 16.00

Before ROE 107,5 0.02 73.48 -23,155.88 2,979.00

After ROE 105,4 0.25 0.48 -3.70 3.26

Before LEV 108,2 0.17 2.26 0.00 702.00

After LEV 107,2 0.15 0.20 0.00 0.97

Before SIZE 107,8 6.36 2.24 0.00 15.00

After SIZE 105,6 6.35 2.09 1.48 12.07

Before CFO 108,2 0.00 1.07 -161.00 10.00

After CFO 106,1 0.03 0.17 -1.15 0.37

Number of shareholders growth (COS)

ROE

Leverage (LEV)

Log Total Assets (Size)

(34)

Event window Mean Min. Max. t-statistic 30 manufacturing firms (0) -7.64% -70.35% 2.65% -2.75*** (0,+1) -10.11% -87.35% 4.08% -2.69*** (-1,1) -12.49% -86.56% 10.76% -3.17*** (-10,10) -16,32% -148.92 18.99% -2.64*** 14 finance firms (0) -7.55% -32.25% 2.48% -2.80*** (0,+1) -18.54% -68.25%% 1.50% -3.24*** (-1,1) -16.76% -67.27% 2.07% -3.17*** (-10,10) -21.89% -76.30% 4.97% -3.48*** 14 services firms (0) -9.50% -32.06% -0.11% -1.73*** (0,+1) -16.43% -70.99% 6.29% -1.96*** (-1,1) -18.62% -69.40% -0.93% -2.25*** (-10,10) -16.77% -85.32% 35.72% -2.00*

C. Event study: Robustness check industries

(35)

CMV

Fraud -0.282*** -0.283*** 0.0586 0.054 0.202** 0.210** 0.004 0.013 (0.062) (0.061) (0.105) (0.110) (0.096) (0.095) (0.063) (0.060) Fraud*ROA -0.122 0.325 -0.633 -0.454 (0.801) (0.733) (0.735) (0.344) ROA 0.496*** 0.496*** 0.497*** 0.497*** 0.497*** 0.497*** 0.497*** 0.497*** (0.026) (0.026) (0.026) (0.026) (0.026) (0.026) (0.026) (0.026) SalesGrowth 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** 0.186*** (0.0104) (0.0104) (0.0104) (0.0104) (0.0104) (0.0104) (0.0104) (0.0104) CSO 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** 0.534*** (0.017) (0.017) (0.017) (0.017) (0.017) (0.017) (0.017) (0.017) COS 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) ROE 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** 0.055*** (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) (0.010) SIZE -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** -0.028*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Constant 0.302*** 0.302*** 0.303*** 0.303*** 0.303*** 0.303*** 0.303*** 0.303*** (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) 0.131 0.131 0.131 0.131 0.131 0.131 0.131 0.131 Observations 61,945 61,945 61,945 61,945 61,945 61,945 61,945 61,945 Groups 7,160 7,160 7,160 7,160 7,160 7,160 7,160 7,160

D. Regression analysis: Robustness checks

Robustness check: Standard errors adjusted for 7,160 clusters for the long-term effects of corporate fraud on the market value

(36)

CMV

Fraud -0.278*** -0.278*** 0.060 0.057 0.201** 0.210** 0.005 0.014 (0.091) (0.091) (0.092) (0.093) (0.088) (0.088) (0.083) (0.085) Fraud*ROA -0.056 0.249 -0.620 -0.483 (0.980) (1.018) (0.816) (0.828) ROA 0.471*** 0.471*** 0.471*** 0.471*** 0.471*** 0.471*** 0.471*** 0.471*** (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) (0.019) SalesGrowth 0.190*** 0.190*** 0.190*** 0.190*** 0.190*** 0.190*** 0.190*** 0.190*** (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) CSO 0.535*** 0.535*** 0.535*** 0.535*** 0.535*** 0.535*** 0.535*** 0.535*** (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) COS 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** 0.019*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) ROE 0.054*** 0.054*** 0.054*** 0.054*** 0.054*** 0.054*** 0.054*** 0.054*** (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) SIZE -0.027*** -0.027*** -0.027*** -0.027*** -0.027*** -0.027*** -0.027*** -0.027*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Constant 0.293*** 0.293*** 0.293*** 0.293*** 0.293*** 0.293*** 0.293*** 0.293*** (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) 0.105 0.105 0.105 0.105 0.105 0.105 0.105 0.105 Observations 61,945 61,945 61,945 61,945 61,945 61,945 61,945 61,945 Groups 7,160 7,160 7,160 7,160 7,160 7,160 7,160 7,160

Robustness check: OLS regression for the long-term effects of corporate fraud on the market value

(37)

CSD

Fraud -0.003 -0.002 -0.010*** -0.010*** 0.002 0.002 -0.004 -0.004 (0.004) (0.004) (0.004) (0.004) (0.002) (0.002) (0.003) (0.003) Fraud*ROA 0.024 -0.009 -0.030 0.013 (0.029) (0.015) (0.019) (0.022) ROA 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Interest 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) SalesGrowth 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) CFO -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** -0.020*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) LEV -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** -0.005*** 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 SIZE 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** -0.002*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) 0.008 0.008 0.008 0.008 0.008 0.008 0.008 0.008 Observations 85,374 85,374 85,374 85,374 85,374 85,374 85,374 85,374 Groups 9,663 9,663 9,663 9,663 9,663 9,663 9,663 9,663

(38)

CSD

Fraud -0.002 -0.001 -0.009*** -0.010*** 0.003 0.002 -0.004 -0.004 (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.013) Fraud*ROA 0.024 -0.008 -0.030 0.031 (0.021) (0.013) (0.024) (0.022) ROA 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** 0.009*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Interest 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) SalesGrowth 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) CFO -0.019*** -0.019*** -0.019*** -0.019*** -0.019*** -0.019*** -0.019*** -0.019*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) LEV -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 SIZE 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -0.001*** -0.001*** -0.001*** -0.001*** -0.001*** -0.001*** -0.001*** -0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) 0.008 0.008 0.008 0.008 0.008 0.008 0.008 0.008 Observations 85,374 85,374 85,374 85,374 85,374 85,374 85,374 85,374 Groups 9,663 9,663 9,663 9,663 9,663 9,663 9,663 9,663

Robustness check: OLS regression The long-term effects of corporate fraud on short-term debt

(39)

CTD

Fraud -0.022** -0.022** -0.001 -0.001 -0.023** -0.023** 0.004 0.005 (0.012) (0.011) (0.007) (0.007) (0.010) (0.010) (0.011) (0.011) Fraud*ROA -0.093 0.005 -0.013 -0.052 (0.098) (0.029) (0.082) (0.079) ROA 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** 0.044*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Interest 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) SalesGrowth 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** 0.028*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) CFO -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** -0.090*** (0.004) (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) LEV 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** 0.087*** (0.003) (0.003) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) SIZE 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** 0.004*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** -0.036*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) 0.106 0.106 0.106 0.106 0.106 0.106 0.106 0.106 Observations 85,545 85,545 85,545 85,545 85,545 85,545 85,545 85,545 Groups 9,678 9,678 9,678 9,678 9,678 9,678 9,678 9,678

Robustness check: Standard errors adjusted for 9,678 clusters for the long-term effects of corporate fraud on total debt

(40)

CTD

Fraud -0.022** -0.024** -0.002 -0.002 -0.023** -0.023** 0.002 0.003 (0.010) (0.010) (0.010) (0.011) (0.010) (0.010) (0.001) (0.001) Fraud*ROA -0.095 0.009 -0.024 -0.045 (0.069) (0.036) (0.080) (0.101) ROA 0.045*** 0.045*** 0.045*** 0.045*** 0.045*** 0.045*** 0.045*** 0.045*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) Interest 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) SalesGrowth 0.030*** 0.030*** 0.030*** 0.030*** 0.030*** 0.030*** 0.030*** 0.030*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) CFO -0.078*** -0.078*** -0.078*** -0.078*** -0.078*** -0.078*** -0.078*** -0.078*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) LEV 0.081*** 0.081*** 0.081*** 0.081*** 0.081*** 0.081*** 0.081*** 0.081*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) SIZE 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -0.024*** -0.024*** -0.024*** -0.024*** -0.024*** -0.024*** -0.024*** -0.024*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) 0.073 0.073 0.073 0.073 0.073 0.073 0.073 0.073 Observations 85,545 85,545 85,545 85,545 85,545 85,545 85,545 85,545 Groups 9,678 9,678 9,678 9,678 9,678 9,678 9,678 9,678

Robustness check: OLS regression the long-term effects of corporate fraud on total debt

(41)