Market Reaction to. The SEC Enforcement Action

Transcription

1 Market Reaction to The SEC Enforcement Action Master Thesis Department Accountancy Faculty of Economics and Business Studies Tilburg University Haicha Sha ANR: Supervisor: Jeroen Suijs

2 Abstract This study examines the stock price performance of fraudulent firms by the time of the SEC investigation announcement. We investigate: (1) whether the market reacts differently to the disclosures of informal investigation and formal investigation; (2) during the informal investigation stage, whether the magnitudes of stock price decline will be similar between wrongly judged firms and correctly judged firms; (3) whether the market reactions to the SEC investigation announcement differ depending on types of violation. We find that the market responses almost equally to both types of investigation disclosure. However, the market s uncertainty of the stock s value prior to the actual disclosure differs greatly between the informal and formal investigation events. The results also suggest that investors react less negatively to wrongly judged companies than to correctly judged companies, which reflects that investors do not completely trust the SEC investigation or they can anticipate the outcome of the investigation. Finally, we observe that companies involved in overstating income (including premature revenue recognition, understating expenses, and overstating assets) face the harshest punishment from the market. 1

3 Content I. Introduction... 3 II. The SEC Investigation Process... 4 III. Prior Research on Fraud Investigations... 6 IV. Research Questions... 8 V. Sample Selection... 9 VI. Methodology VII. Statistic Description VIII. Empirical Results IX. Conclusion and Limitation X. References

4 I. Introduction The Securities and Exchange Commission (the SEC) was formed as a result of the stock market crash of Some of the causes of the crash were found to be a precrash speculative frenzy, artificially inflated trading activity, false and misleading information published by companies listed on the exchange, and inside trading. Congress passed the Securities Act of 1933 and the Securities Exchange Act of 1934 to regulate companies that wanted to raise capital through the financial markets. The SEC is formed in 1934 to enforce these laws. Its enforcement program consists of investigations and subsequent injunctive actions or administrative proceedings against offending registrants and auditors. The SEC plays a vital role in shaping and enforcing accounting standards. Its goals are to anticipate emerging reporting problems, to maintain the credibility of the disclosure system as well as the integrity of the market, and to protect investors. The efficient markets hypothesis suggests that the market rapidly impounds relevant information into stock prices. Most prior related studies reveal that companies being accused by the SEC experience an economically and statistically significant negative stock price reaction to the announcement of the investigation. The purpose of this paper is to provide insight on the market reaction to the SEC enforcement action during the year 2008 to Following prior researches, we also identify our sample of fraudulent firms through AAERs. In our first research question, we separately analyze market reactions to the informal and formal investigation events by measuring the extent of stock return decline and changes in implied volatility of options for both types of investigation. We find that the market responses almost equally to both types of investigation disclosure. However, the market s uncertainty of the stock s value prior to the actual disclosure of the investigation differs greatly between the informal and formal investigation events. Then we test investors belief in the SEC investigation by comparing magnitude of stock return decline of wrongly judged companies (firms being investigated but later found no evidence of fraud) with that of correctly judged companies. The result suggests that investors react less negatively to wrongly judged companies than to 3

5 correctly judged companies, which reflects that investors do not completely trust the SEC investigation or they can anticipate the outcome of the investigation At last, we classify the sample of firms according to the type of violation to examine whether violation types will influence the severity of the market reaction. We find that companies involved in overstating income (including premature revenue recognition, understating expenses, and overstating assets) face the harshest punishment from the market. Our study makes several contributions to the existing literature. Firstly, previous researches hardly study the market reaction to the informal investigation disclosure, but our research focus on both the formal and informal investigations. Our result suggests that informal investigation announcement play a role in influencing investors expectation on the stock value. Secondly, we study the firms that are wrongly judged by the SEC. The result indicates that investors might carefully interpret and evaluate the information about the investigation before the final outcome. In short, our study provides the empirical evidence to the academic relevance in investor behavior and the SEC investigation. It also helps investors to anticipate the magnitude of the stock price decline associated with the SEC investigation. The remainder of this paper is organized as follows. Section 2 and 3 will introduce the SEC investigation process and prior researches on fraud. Section 4 will give the research questions of this paper. Section 5 to Section 7 will present the sample selection, statistic description, and methodology of the research. Then Section 8 will show the empirical results of the study. The final section will contain the conclusion and limitation. II. The SEC Investigation Process Due to limited funds and personnel, the SEC normally does not actively search for financial statement fraud. As a result, the SEC usually obtains the tips of potential fraud from an interested person, an auditor, or a news article. Pincus, Holders, and Mock (1988) report that the sources of the SEC enforcement action could be (1) 4

6 market surveillance programs of the American and New York Stock Exchanges and the National Association of Securities Dealers; (2) public complaints, tips, referrals from other low enforcement agencies, and financial press information; (3) reviews of 1993 and 1994 Securities Acts filings. Then the SEC staff from the Division of Corporate Finance will scrutinize reports for violations of routine screening criteria and for suspicious subjective factors. When it warrants a further scrutiny, the SEC will proceeds to an informal investigation and invite people with relevant information to provide documents and testimony. During this stage, the SEC need not formally notify its target. Later, depending on the outcome of the informal investigation, the SEC staff will drop the case or initiate a formal investigation. The formal investigation grants subpoena power of greater access to company documents and executives. If the SEC notifies formally to the target company, the investigation may become public. As a matter of policy, the SEC makes its enforcement activities public only when it issues a formal investigation that the target firm violates securities law and seeks a settlement with the target. A formal order may also be issued without a prior informal investigation (McDowell, 2005). As a matter of fact, the SEC has more targets for formal investigation than it can practically pursue, and formal investigations are both costly and highly visible. Therefore, the SEC ranks candidates for formal investigation according to the probability of success and potential message value. Investors usually receive news of the alleged fraud prior to the SEC formal announcement of the investigation. However, there is significant variability in the timing and the manner that investors learn of an SEC investigation. Not only the SEC does not usually announce public about its initiation of an investigation, but also many target firms choose not to disclose the investigation until the SEC gives the formal notice of its enforcement action. On the other hand, knowing that investors will inevitably learn of the investigation, some firms choose to voluntarily disclosure that it is being informally investigated. As a whole, the source of the first public disclosure of investigation can be a press release (informal and formal investigation), through a filing made by the company with the SEC (informal and formal 5

7 investigation), or the SEC Commission s announcement (formal investigation). (Nourayi, 1994). At last, when the outcome of the SEC s formal investigations indicates alleged securities law violations, the SEC will file a final settlement. Figure 1 summarizes the normal sequence of major event leading up to the SEC s investigation and action related to fraudulent firms. Figure 1 Event Process Fraud occurs (Informal investigation) Formal investigation Investigation settlement and AAER release III. Prior Research on Fraud Investigations Feroz, Park, and Pastena (1991) examine 224 AAERs issued between 1982 and 1989, and explore questions related to the SEC s enforcement programs. They find that, in general, firms engaged in fraud underperform the market despite the benefit of the fraud schemes. They obtain the cumulative abnormal return decline of 13 percent during the two-day event window surrounding the first disclosure of a violation. Nourayi (1994) also studies the stock prices responses to the SEC s enforcement action, but he applies a longer event window: 15 days before through 15 days after the announcement. He finds that release of information by a company before the SEC s official announcement causes significantly larger negative reaction. The violation type does not produce significant reactions, but the magnitude of negative abnormal returns varies directly with the severity of the Commission s action. Moreover, the market reaction is most unfavorable when companies are named as the defendant/respondent. Dechow, Sloan, and Sweeny (1996) examine 92 fraudulent firms identified by AAER during the period, and investigate (1) earnings manipulations, (2) 6

8 internal governance structure, and (3) capital market consequences for firms that violate GAAP. They find that these firms manage earnings mainly to attract external financing at a low cost, but experience a significant increase in the cost of capital when the manipulation is made public. Besides, they also find that manipulating earnings is associated with an increase in the bid-ask spread, a drop in analyst following, an increase in short interest, and an increase in the dispersion of analysts earnings forecasts. Beasley (1996) uses AAERs and a news search to identify 75 fraudulent firms during the period, and examines the relation between board of directors composition and the occurrence of financial statement fraud. He finds that non-fraud companies have a significantly higher percentage of outside directors than fraud firms. Palmrose, Richardson, and Scholz (2004) study the determinants of market reactions to restatement announcements. They examine the market reaction to a sample of 403 restatements from 1995 to They find more negative returns are associated with restatements involving fraud, affecting more accounts, decreasing reported income, and attributed to auditors or management. There appeared to be additional penalty for announcements that do not quantify the restatement. Finally, they provide evidence on the relation between restatement announcements and analyst earnings forecast dispersion, bid-ask spreads and subsequent revisions in analyst earnings forecasts. Karpoff, Lee, and Martin (2008) examine the penalties imposed on the 585 firms targeted by the SEC for financial misrepresentation from 1978 to The penalties imposed by the market are huge due to the estimated reputational loss. For each dollar that a firm misleadingly inflates its market value, on average, it loses an additional $3.08 due to expected legal penalties and lost reputation. Based on the research done by Feroz et al. (1991), Christensen, Paik, and Williams (2010) examine market reactions surrounding the market discovery date and the date that the SEC announces its investigation. Their results suggest that investors do not fully realize the severity of a fraud until the SEC begins its formal investigation. This result is inconsistent with market efficiency. Moreover, they find that firm characteristics would predict the likelihood that a firm is investigated by the SEC. The logit regression formulated by authors suggests that in addition to ROA and 7

9 Altman Z-score of the year prior to AAER, total and abnormal accruals are significant explanatory factors. IV. Research Questions Since investors will value the immensity of the investigation through the information disclosed by company, there appear to be a clear distinction between informal and formal investigations. As a consequence, when the company decides to disclose the informal or formal investigation, it will carefully select the wording informal or formal in the news release, as informal and formal provide signal messages to the market. Therefore, we will first examine the effect of informal and formal investigation announcement to the stock market performance. We expect that market will react differently regarding the type of the investigation, as different message will trigger different market reactions. And we expect the formal investigation will face hasher punishment, since the formal investigation is likely to confirm that the target firm behaves dishonestly in reporting financial statement. RQ1: The market will react differently to the disclosure of informal investigation and formal investigation. Due to limited funds and personnel, the SEC normally does not actively search for financial statement fraud. Therefore, when the public learns about an informal inquiry, investors will become suspicious to the target firms. If investors believe that the SEC correctly identifies the real fraudulent firms during the informal investigation stage, they will react similarly to all target firms. In prior study, Christensen et al. (2010) hold the view that investors do not fully realize the severity of a fraud until the SEC begins its formal investigation. In order to test whether investors totally trust the SEC, we will divide our sample into two groups: Fraud Group of firms being investigated but later found involved in the fraud (that is they are correctly judged by the SEC), and Non-Fraud Group of firms being investigated but later found no evidence of fraud (that is they are wrongly judged by the SEC). We will examine whether investors respond equally to the target firms in two groups during the informal and formal 8

10 investigation stage. For this research question, we expect that investors will not completely believe in the SEC s investigation. Although many companies choose not to leak the information of the informal investigation, due to different incentives some companies choose to voluntarily disclose the informal investigation. Receiving this message, the investors may take a skeptical attitude towards the issue. They may question whether the SEC correctly identifies the real fraudulent firms. Thus, we predict that firms in Non-Fraud Group will face a lower magnitude of stock prices decline than firms in Fraud Group. RQ2: During the informal investigation stage, the magnitudes of stock price decline will be different between firms in Fraud Group and those in Non-Fraud Group. Prior research suggests that not all types of fraud are equal to investors. Feroz et al. (1991) discover that, in their sample period, the SEC most often pursued overstatements of accounts receivable and inventories resulting from premature revenue recognition and delayed write-off. Then Christensen et al. (2010) examine different types of violations that lead to fraud investigation and find that the market responds more negatively to investigation related to revenue recognition, asset overstatement, and insider trading allegation than other types of violations. Our study will focus on a more recent time period, from 2008 to We will examine whether market reactions to the SEC investigation announcement still differ depending on types of violations in the current environment when investors are more responsive to the fraud. And we anticipate that investors regard overstating income as most serious. RQ3: Market reactions to the SEC investigation announcement will differ depending on types of violations. V. Sample Selection In order to obtain the sample of firms for Fraud Group, we rely on the Accounting and Audit Enforcement Release (AAER). Most prior studies have utilized AAERs to identify fraudulent firms. These reports provide an objective criterion for identifying firms with fraudulent financial statements, and the SEC lists the fraud type committed 9

11 in the AAERs. To obtain the sample of firms for Non-Fraud Group, we are reliant on the key-word searches on LexisNexis News Library. Since the SEC does not publicly disclose the initiation of an investigation, we are reliant on the companies to disclose the event to investors through its annual filings with the SEC as well as LexisNexis News Library database to gather the disclosure date of the SEC s formal and informal investigation. The sample time period for events to occur is defined to be 2008 to From AAER list and announcement in LexisNexis News Library, we are able to backtrack from the original 10-k filing, looking in previous year s annual 10-k, quarterly 10-Q, and current 8-k filings, as well as press releases to arrive at the exact date that the investigation was made public. For each investigation reference, we also look for the data relating to the informal or formal investigation announcement, depending on which was originally found. If there were no press releases or 8-k filings with the SEC that referenced the investigation, we have to rely on the first occurrence of the investigation within a quarterly or annual filing as the date that the investigation was made public. Or based on key-word searches (e.g. the SEC investigation, the SEC inquiry) on LexisNexis News Library database to identify the exact date available in public news media. Since some companies may have chosen not to disclose an investigation within its financial reports, the data sample is not considered to be the entire universe of the SEC investigations during the sample period. We just consider the companies with disclosure thus it could produce some bias to the investigation. It will be discussed in the part of limitations. Furthermore, stock prices and daily returns are available on the Center for Research in Security Prices (CRSP). We collect totally of 105 samples for Fraud Group and 27 samples for Non-Fraud Group. More detail would be discussed in Section 8 Statistic Description. 10

12 VI. Methodology Similar to most prior studies, we will apply the event study methodology. Event studies measure security price changes in response to events. A single event study typically analyzes the average security price reaction to instance of the same type of event experienced by many firms. The basic idea is to find the abnormal return attributable to the event being studied by adjusting for the return that stems from the price fluctuation of the market as a whole. A main event window of three days (starting on one day before the day of investigation disclosure date and ending on one day after the disclosure [-1,1]) is used. Short-horizon event studies are more reliable than long-horizon event studies. While long-horizon methods have improved, serious limitations still remain. Kothari and Warner (1997) find that inferences from long-horizon tests require extreme caution. The tests are highly susceptible to the joint-test problem and have lower power. Therefore, the interpretation of long-horizon results could be problematic. Shorthorizon methods are relatively straightforward and trouble-free. (Christensen et al. 2010, McDowell 2005, Nourayi 1994). Besides, we will apply a market-adjusted model based on a value-weighted index (with dividend) to estimate abnormal returns. This model subtracts the CRSP market index return from a company s daily return to obtain the market-adjusted abnormal return (AR) for each day and company. The daily abnormal returns are summed to calculate the cumulative abnormal return (CAR) for the given time period. VII. Statistic Description During the period from 2008 to 2011, a total of 105 SEC litigation releases are reported and there are 27 firms being investigated but later found no evidence of fraud. The tests of market reactions are based on the market sample of firms that meet two criteria: (1) press release disclosing the error, its investigation, or its settlement are reported in the LexisNexis or available 8-k filings and (2) stock price data are 11

13 available for the 3-day window surrounding the disclosure of informal, formal investigation and final settlement through the CRSP Daily Returns File. Table 1 illustrates these criteria eliminate some firms from the original samples. Table 1: Sample Firms Named as Targets of AAER Fraud Group Non-Fraud Group Firms being informally/formally investigated Firms with missing disclosure dates 1 1 Firms with insufficient price data 32 2 Firms in our sample Firms being both informally and formally investigated in our sample 43 3 Firms only being informally investigated in our sample 0 11 Firms only being formally investigated in our sample As a result, the data sample of Fraud Group consists of 72 firms, where 43 of them are being both informally and formally investigated and 29 only being formally investigated. The data sample of Non-Fraud Group consists of 24 firms, where 11 of them are only being informally investigated and 10 of them only being formally investigated. A significant period of time generally elapses between the market s discovery of a firm s fraudulent conduct and the completion of the SEC investigation of the case as well as the issuance of an AAER report. In the sample of Fraud Group, the average lag time is 36 months with a standard deviation of 20 months. However, this period varies widely for each individual case from a minimum of 3 months to a maximum of 87 months. Figure 2 illustrates the number of sample firms of Fraud Group during the four-year period from 2008 to In 2008, there are 16 fraud firms. In 2009, the number increased to 23. Finally, the SEC released 18 fraudulent firms in 2010 and 15 fraudulent firms in

14 Figure 2 Frequency of Sample Firms by year Figure 3 illustrates the industry distribution of sample fraud firms. We find that firms in service in manufacturing industries (SIC codes which includes rubber, leather, stone, metal machinery, electronic equipment, transportation equipment etc.) are highly represented in our sample. 35,0% 30,0% 25,0% 20,0% 15,0% 10,0% 5,0% 0,0% Figure 3 Industry Distribution of Sample Firms Table 2: Industry Distribution of Sample Firms of Fraud Group SIC Codes Nr Industry Description Mineral and Construction Industries Manufacturing: Food, Tobacco, Textile, Lumber, Furniture, Paper, Printing, Chemicals, and Petroleum. 13

15 Manufacturing: Rubber, Leather, Stone, Metal, Machinery, Electronic Equipment, Transportation Equipment, etc Transportation, Communications, and Utilities Wholesales Trade (durable and non-durable) and Retail Trade (building materials, general merchandise, food, automotive, apparel, home furnishing, dining, etc.) Financial Services, Insurance, and Real Estate Industries Service Industries: Hotel, Personal Services, Business Services, Automotive Repair, Motion Pictures, Amusement and Recreation Services. Services Industries: Health, Legal, Educational, Social, Museums, Engineering, Accounting, Management, etc Services Industries: Disposal, membership, etc. Figure 4 categorizes sample firms by the type of fraud committed and Table 3 presents the description of fraud type. The most common fraud type identified by the SEC during this sample period is overstated income. The second most common type is unlawful use of assets. Figure 4 Type of Fraud Committed by Sample Firms Overstated Income False documentation Unlawful use of assets Security fraud 14

16 Table 3: Type of Fraud Committed by Sample Firms Category Example Nr Revenue Recognition: fictitious sales, accelerated sales, not in accordance with GAAP (signing return agreement, contractual agreement not yet fulfilled, bill and hold, etc.), Overstated cookie jar reserve Income Understated Expenses: COGS, debt, taxes, bad debt, loss, 25 depreciation Overstated Assets: receivables, inventory, capitalized expenses, cash, long-term assets False Fraudulent dissemination of false and misleading press documentation releases 11 Unlawful use of assets Bribery, theft, assist other companies in committing fraud 20 Security fraud Inside trading, stock option granting practice 16 VIII. Empirical Results A window of -1 to +1 days from the event date is used to calculate the cumulative abnormal returns for the informal and formal investigations. Table 4 shows a detailed result of market reaction to the SEC investigation announcement. Table 4: Market-adjusted Cumulative Abnormal Stock Return Comparisons around Event Dates Panel A: All Sample Firms (Including Fraud Group and Non-Fraud Group) Cumulative Abnormal Return (CAR) Informal Investigation Formal Investigation Settlement Nr Mean Standard Deviation Comparison to zero (t-statistic) 57 a *** 85 b *** 96 c * a: Total firms of informal investigation in two groups: 43 in Fraud Group plus (3+11) in Non-Fraud Group b: Total firms of formal investigation in two groups: (43+29) in Fraud Group plus (3+10) in Non-Fraud Group c: Total firms of settlement in two groups: (43+29) in Fraud Group plus ( ) in Non-Fraud Group 15

17 Panel B: Sample of firms in Fraud Group Cumulative Abnormal Return (CAR) Informal Investigation Formal Investigation Settlement Nr Mean Standard Deviation Comparison to zero (t-statistic) *** *** Panel C: 43 Sample of firms being formally and informally investigated in Fraud Group Cumulative Abnormal Return (CAR) Informal Investigation Formal Investigation Settlement N Mean Standard Deviation Comparison to zero (t-statistic) *** *** Panel D: 29 Sample of firms only being formally investigated in Fraud Group Cumulative Abnormal Return (CAR) Informal Investigation Formal Investigation Settlement Nr Mean Standard Deviation Comparison to zero (t-statistic) NA NA NA NA **

18 Panel E: Sample of firms in Non-Fraud Group Cumulative Abnormal Return (CAR) Informal Investigation Formal Investigation Outcome of Being Cleared Nr Mean Standard Deviation Comparison to zero (t-statistic) * 24 d d: There are 3 firms being both informally and formally investigated, therefore the total is 24 ( ). ***, **, * indicate significance at 1%, 5%, 10% level, respectively, using two-sided tests. First of all, we take a look at the market response to all samples of firms. In Panel A, during the window around the disclosure of informal investigation, the mean cumulative abnormal return is %. During the window of formal investigation, mean cumulative abnormal return is %. Both declines are significant at 1% level. However, when the investigation reaches a settlement, market reacts positively with a mean cumulative abnormal return of 1.492%, significant at 10%. This indicates that market penalizes fraudulent firms immediately when the news of the potential fraud is revealed. Contrary to our expectation, there is no significant difference between stock return declines surrounding the informal and informal investigation announcements. And the stock return turns to favorable when the case is settled. Since an increasing cost could occur during the investigation, cease of investigation and settlement are good news to the investors. Then we break down the total samples into two groups. Panel B, Panel C and Panel D contain the result of samples in Fraud Group, and Panel E sums up the result of samples in Non-Fraud Group. In Panel B of all samples of firms in Fraud Group, the mean cumulative abnormal return around the disclosure of informal investigation is , while the mean cumulative abnormal return of formal investigation is %, only slightly higher than that of informal investigation. Both numbers are significant at 1% level. Besides, in Panel E of all sample firms in Non-Fraud Group, the mean cumulative abnormal return around the disclosure of informal investigation is , while the mean 17

19 cumulative abnormal return of formal investigation is %, slightly lower than that of informal investigation. Therefore, we obtain the same result that there is no significant difference between stock return declines of two types of investigation, which contradicts our expectation on RQ1. Further insights into the 43 sample firms that are being both informally and formally investigated are shown in Panel C. The mean cumulative abnormal return of informal investigation is %, which is lower than that of formal investigation of %. Both numbers are significant at 1% level. This result is consistent with what Christensen et al. have found: an additional stock price decline at the time of the SEC s investigation announcement after market s first discovery of the fraud. On the other hand, Panel D indicates that the cumulative abnormal return of those 29 sample firms only being formally investigated goes down by 5.326%. Although the punishment is harsher to these 29 sample firms (-5.326%) compared with those 43 firms (-3.999%) for formal investigation events, the former faces only one time stock price drop while the latter experiences twice. As a result, magnitude of overall stock price decline is greater for those 34 firms being both informally and formally investigated. Due to this difference, we believe that the informal investigation plays a crucial role during the event and market may not equally regard the informal and formal investigations. To further examine this question, we will look into the events that lead up to the disclosure. Put and call option implied volatility will be used to analyze the market s uncertainty of stock value prior to the investigation disclosure, as implied option volatility has been found to be a very good predictor of future volatility for stock indexes and individual stocks (McDowell, 2005). The implied volatility of at the money puts and calls with 30 days to maturity are compared for each day leading up to the event for informal and formal investigations. 30-day maturity options are chosen because of their high liquidity. For each trading day prior to the event, the average implied volatility across the sample for each option is computed. The option data are available in the database of Option Metrics volatility surface. However, option data are not available for all stocks within the investigation samples. The table below provides the number of stocks with available option data that are included in the daily average implied volatilities. 18

20 Annual Implied Volatility Informal Investigation Formal Investigation Option Data Option Data 30 Day ATM Call and Puts We plot the movements of changes in implied volatilities in Figure 5, 6, and 7. Figure 5 illustrates the changes in implied volatilities of informal investigation. Figure 6 illustrate the changes in implied volatilities of formal investigation and figure 7 plots the fluctuation of implied volatilities of firms only being formal investigated (19 observations). 64% Figure 5 Informal Investigation Option Volatility 59% 54% 49% 44% Trading days prior to event 30 Day ATM Call 30 Day ATM Put 19

21 Annual Implied Volatility Annual Imlied Volatility 64% Figure 6 Formal Investigation Option Volatility 59% 54% 49% 44% Trading days prior to event 30 Day ATM Call 30 Day ATM Put Figure 7 Formal Investigation Volatility of 19 Stocks 0,64 0,59 0,54 0,49 0, Trading days prior to event 30 Day ATM Call 30 Day ATM Put Figure 5 shows a relative steady increase across the 30-day period. The implied volatilities for the put and call options in Figure 6 and 7 show more fluctuations. It begins with a decrease in volatility then shows a sudden rise at about 13 to 15 trading days prior to the event. After that, there is a steady, slight decline, then followed by a second slight rise. At the time just before the formal investigation event, there is a small swing of the implied volatility. 20

22 Informal investigations could be motivated by earnings restatement (McDowell, 2005). Earnings restatements increase the uncertainty about the firms value, as the investors attempt to assess the implications for future earnings and cash flows. Besides, firms internal investigation may also motivate the SEC to begin an inquiry (Feroz, et al, 1991). So it gives rise to additional uncertainty due to the possibility of an SEC investigation and the most probable outcome if it were to occur. If the event is partially anticipated, some of the market behavior related to the event will show up in the pre-event period. Therefore all uncertainty above is reflected in the steady rise in implied volatilities prior to informal investigation announcement. Comparing to the disclosure of informal investigation, the announcement of formal investigation seems to be more unexpected. During the 10 to 15 days prior to the formal investigation event, the market s uncertainty of future price is declining. The sudden rise in implied volatility from 13 to 15 days before the announcement suggests the existence of information events prior to the investigation announcement that alert the investors. Figure 7 shows that 19 firms only being formally investigated experience an even more fluctuated movement and overall volatilities are higher. It is consistent with the fact that without previous informal investigation, the sudden formal investigation is more unexpected to the market. In sum, to answer the RQ1 of different market reaction to the informal and formal investigation, results indicate that the market responses almost equally to both types of investigation disclosure. However, the market s uncertainty of the stock value prior to the actual disclosure of the investment differs greatly between the informal and formal investigation events. The formal investigation is more unexpected to the investors. As a whole, the market s uncertainty is relatively high, which is reflected in the significant negative cumulative abnormal return and the jump in the implied volatility as a result of the disclosure. Moreover, when the informal investigation turns to the formal, there is additional stock price decline, but investors react approximately the same when they previously learn an informal inquiry. This is consistent with the notion that the market reaction is not complete until a further market penalty is imposed at the time of SEC s investigation announcement. We also find that the market reacts favorably when the investigation is ended. 21

23 RQ 2 studies the investors trust in the SEC s investigation. During the window of informal investigation disclosure, whether the magnitudes of stock price decline will be similar between firms in Fraud Group and those in Non-Fraud Group. Table 5 summarizes the result for this study. During the window of disclosure of informal and formal investigations, Non-Fraud Group has the mean cumulative abnormal returns of % and % respectively. For both types of investigations, the magnitudes of stock return decrease of samples in Non-Fraud Group are all under those of samples in Fraud Group. Table 5: Market-adjusted Cumulative Abnormal Stock Return Comparisons around Event Dates (Fraud Group VS Non-Fraud Group) Cumulative Abnormal Return (CAR) Group & Nr Mean Standard Deviation Comparison to zero (t-statistic) Fraud (43) *** Informal Investigation Non-Fraud (14) Fraud (72) *** Formal Investigation Non-Fraud * (13) Settlement/Outcome of Fraud (72) being cleared Non-Fraud (25) ***, **, * indicate significance at the 1%, 5%, 10% level, respectively, using twosided tests. Investors generate information for trading from various sources as rumors, newspapers and financial statements. They under- or over- react to new information that is either good or bad. There are many studies in the psychology about how people interpret and evaluate the information that is brought to them. If it is in line with their own expectations, investors are more confident about the information (Daniel et al. 1988). And information that is more salient will be received as more confident and trustworthy for the public (Robinson et al. 1982). 22

24 In this study, investors react less severely to the informal investigation announcement of the sample of firms in Non-Fraud Group. Even during the formal investigation events they response less severely to the sample of firms in Non-Fraud Group than to Fraud Group. It seems that investors are more or less confident about the companies performance, and carefully interpret and evaluate the information about the investigation. Some investors might rationally expect the performance of the company stock, some might choose to wait and see. Therefore, we can say that the investors do not completely trust the SEC or they can anticipate the outcome of the investigation. When the investigation is settled, we can see that market reacts more positively to those wrongly judged firms (2.79%) than to those fraudulent firms (1.03%). Obviously, it is good news to investors that the wrongly judged firm is cleared. In RQ3, we explore whether certain types of potential violations are more salient to investors. Table 5 and Table 6 present the result of cumulative abnormal returns for each type of allegation of the informal and formal investigation events. Table 6: Market-adjusted Cumulative Abnormal Stock Return Comparisons of Panel A: Informal Investigation Cumulative Abnormal Return (CAR) Nr Different Fraud Type Mean Standard Deviation Comparison to zero (t-statistic) Overstated income *** False documentation Unlawful use of assets Security fraud * 23

25 Panel B: Formal Investigation Cumulative Abnormal Return (CAR) Nr Mean Standard Deviation Comparison to zero (t-statistic) Overstated income *** False documentation Unlawful use of assets ** Security fraud ***, **, * indicate significance at the 1%, 5%, 10% level, respectively, using twosided tests. The results suggest that there are significant negative stock returns for alleged violation related to overstated income (including revenue recognition, understated expenses, and overstated assets) for both types of investigation. The cumulative abnormal returns are % and % respectively, both significant at 1% level. It is probably because overstated income is the most frequent fraud type that is investigated by the SEC, and it has a big impact on overstating earnings numbers, which affect the company s value to a great extent. Previous studies mention that informal investigations could be motivated by earnings restatement (McDowell, 2005). And Wu (2002) finds an increase in proportion of earnings misstatement including revenue restatement. Palmrose and Scholz (2004) suggest that investors regard revenue restatements as more serious. The greater stock decline of informal investigation (-7.968%) seems to confirm these findings. Regarding other types of violation, the magnitude of stock decline of informal investigation disclosure is lower than that of formal investigation disclosure. Especially, alleged companies that unlawfully use assets have little influence on the market reaction (-0.585%) when the informal investigation is announced. Unlawful use of assets includes bribery, theft, and assisting other companies in committing fraud, respondents of which might be individual (e.g. the senior management team of the company) and have a relatively less effect on company s value. The little market reaction during the informal investigation may also reflect that investors suspect the likelihood of violation and wait for an official allegation. 24

26 IX. Conclusion and Limitation In this study, we addressed three research questions about the SEC enforcement action: whether market reacts differently to the disclosure of informal investigation and formal investigation; during the informal investigation stage, whether the magnitudes of stock return decline will be similar between wrongly judged firms and correctly judged firms; whether market reactions to the SEC investigation announcement differ depending on types of violation. We find that the market responses almost equally to both types of investigation disclosure. However, the market s uncertainty of the stock s value prior to the actual disclosure of the investigation differs greatly between the informal and formal investigation events. And the formal investigation is more unexpected to the market. As a whole, the market s uncertainty is relatively high, which is reflected in the significant negative cumulative abnormal return and the jump in the implied volatility as a result of the disclosure. Moreover, when the informal investigation turns to the formal, there is additional stock price decline, but investors react almost the same with when they previously learn about an informal inquiry. What s more, we find that investors react less negatively to those wrongly judged companies than to correctly judged companies during both investigation events, which reflects that investors do not completely trust the SEC investigation or they could anticipate the outcome of the investigation. Finally, we find that companies involved in overstating income (including premature revenue recognition, understating expenses, and overstating assets) face the harshest punishment from the market. This paper has some limitations. First, given the fact that some companies do not disclose the informal investigation and companies that voluntarily disclose may have different incentives, there would be some bias in the sample selection of informal investigation, which could cause some errors in the result. Second, due to the fact that the SEC seldom mistakenly judges fraudulent firms, the samples in Non-Fraud Group are limited. As a result, the comparison between Non-Fraud Group and Fraud Group is not that persuasive. Third, since the information available in the LexisNexis News 25

27 Library is limited, there might be errors in the collected disclosure date (some dates are not available), which would lead different results of cumulative abnormal returns. Last, I did not take into accounts of some other factors, which might also favorably or unfavorably influence the firm s value. Thus we cannot rule out this possible biased result. Disclosure and regulated reporting is highly crucial, especially after the large-scale corporate scandals such as Enron, WorldCom, Xerox, Merrill Lynch and Arthur Anderson. This series of corporate fraud rattled investors and made them weary of company management. Therefore, the SEC would step up its enforcement actions to catch the fraud before investors benefit is harmed and renew the public s trust in the system. Research in market reaction on the SEC enforcement action increasingly interests researches since the behavior of investors changes over time. Our study takes into account more factors that related to investors, the SEC investigation, and fraud companies and provides the empirical evidence to the academic relevance in the area of investor behavior, the SEC fraud investigation and company management. Future researches could extend the investigation about investors behavior and deepen the understanding of investors trust in the SEC. Moreover, association between the market reactions and the amount of fine the fraudulent companies required to pay is also an interest research area. 26

28 X. References Christensen, Theodore E.; Paik, Daniel Gyung H.; Williams, Christopher D. (2010) Market Efficiency and Investor Reactions to SEC Fraud Investigations, Journal of Forensic & Investigative Accounting, Vol. 2, Issue 3, Special Issue Dana Mcleod (2005) Accounting Scandal Announcements: A Test of Market Efficiency, Journal of Business & Economic Research, Vol. 3, Nr. 3 Dechow, P.M.; Sloan, R.G.; Sweeney, A.P. (1996) Causes and Consequences of Earning Manipulation: Analysis of Firms Subjects to Enforcement Actions by the SEC Contemporary Accounting Research Vol. 13 No. 1, pp Feroz, Ensan H.; Park, Kyungjoo; and Pastena, Victor S. (1991) The Financial and Market Effects of the SEC s Accounting and Auditing Enforcement Releases, Journal of Accounting Research, Vol. 29, Studies on Accounting Institutions in Markets and Organizations, pp Karpoff, J.M., D.S. Lee, and G.S. Martin (2008) The Cost to Firms of Cooking the Books, Journal of Financial & Quantitative Analysis 43(3): Kothari, Warner (2006) Econometrics of Event Studies, Handbook of Corporate Finance: Empirical Corporate Finance, Vol. A Ch.1 McDowell, John (April 1, 2005) A Look at the Market s Reaction to the Announcements of SEC Investigations The Leonard N. Stern School of Business, Glucksman Institute for Research in Securities Markets Nourayi, Mahmoud M. (1994) Stock Price Responses to the SEC s Enforcement Actions Journal of Accounting and Public Policy, Vol. 13, pp Palmrose, Zoe-Vonna; Richardson, Vernon J.; Scholz, Susan (2004) Determinants of Market Reaction to Restatement Announcements Journal of Accounting and Economics vol. 37 pp Pastena, V. and Ronen, J. (1979) Some Hypotheses on the Pattern of Management s Informal Disclosure Journal of Accounting Research 17(2): Ronen, J (1977) The Effect of Insider Trading Rules on the Information Generation by Corporations The Accounting Review 52(2):