Download This Paper - SSRN

Earnings management, lawsuits, and stock-for-stock acquirers’ market performance Guojin Gong, Henock Louis*, Amy X. Sun Smeal College of Business, Pennsylvania State University, University Park, PA, 16802

Abstract There is a positive association between stock-for-stock acquirers’ pre-merger abnormal accruals and post-merger announcement lawsuits. The market only partially anticipates the effects of post-merger announcement lawsuits at the merger announcement and the post-merger announcement long-term market underperformance is largely limited to litigated acquisitions. The evidence indicates that lawsuits are a contributing factor to the post-merger announcement underperformance, which is a very important finding, given the puzzling nature of the postmerger underperformance and the heightened interest in explaining this phenomenon. The evidence also suggests that it is important that investors not only undo the direct stock price effects of earnings management but also factor the contingent legal costs associated with earnings management.

JEL classification: G14; G34; M41; M43 Keywords: Stock-for-stock merger; earnings management; lawsuit; market efficiency This paper benefits from comments and suggestions by Christine Cheng, S.P. Kothari (the editor), Sugata Roychowdhury (discussant at the 2007 American Accounting Association annual conference), Hal White, Paul Zarowin (a referee), an anonymous referee, and workshop participants at Indiana University, Penn State University, Purdue University, Washington University, and the 2007 American Accounting Association annual conference. *E-mail address: [email protected]

Earnings management, lawsuits, and stock-for-stock acquirers’ market performance Abstract There is a positive association between stock-for-stock acquirers’ pre-merger abnormal accruals and post-merger announcement lawsuits. The market only partially anticipates the effects of post-merger announcement lawsuits at the merger announcement and the post-merger announcement long-term market underperformance is largely limited to litigated acquisitions. The evidence indicates that lawsuits are a contributing factor to the post-merger announcement underperformance, which is a very important finding, given the puzzling nature of the postmerger underperformance and the heightened interest in explaining this phenomenon. The evidence also suggests that it is important that investors not only undo the direct stock price effects of earnings management but also factor the contingent legal costs associated with earnings management.

1

1. Introduction Louis (2004a) suggests that the reversal of the effects of pre-merger abnormal accruals is a significant determinant of stock-for-stock acquirers’ long-term underperformance. While Louis (2004a) partly attributes long-term underperformance after stock-for-stock mergers to the reversal effects of pre-merger earnings management, he does not consider the potential litigation costs associated with pre-merger earnings management. In this study, we analyze whether postmerger announcement lawsuits are associated with pre-merger abnormal accruals and the potential effects of lawsuits on acquirers’ market performance. We posit that, by subjecting stock-for-stock acquirers to lawsuits, pre-merger earnings management can have an indirect effect on acquirers’ performance around and after the merger announcement, in addition to the direct effect associated with post-merger accrual reversals. Post-merger lawsuits are likely to be very costly for various reasons. First, plaintiffs generally receive large monetary settlements in acquisition-oriented class action suits (Thompson and Thomas 2003). Second, lawsuits tend to distract management at the very moment when it should be concentrating on integrating the merging partners. For instance, the members of the HewlettPackard (HP) and Compaq integration team maintained that lawsuits severely hampered their integration plan and decision making (Swartz 2002). Many analysts believe that the distraction and the ensuing loss in employee morale caused by these lawsuits seriously hurt HP to the benefit of its competitors (see, e.g., Sullivan 2002). Some analysts then conclude that mergers are likely to destroy value, not necessarily because the acquisitions are bad, but because firms that are embroiled in merger battles fall behind their competitors (Etzel 2002). After analyzing the association between pre-merger announcement abnormal accruals and post-merger announcement lawsuits, we examine whether the market anticipates the

2

potential lawsuits and their consequences at the merger announcement. It is well documented that the average stock-for-stock acquirer experiences significant market losses at the merger announcement. Louis (2004a) suggests that earnings management is a significant determinant of acquirers’ losses in the days leading to stock-for-stock merger announcements. Other potential explanations for the losses include the hubris hypothesis (Roll 1986), the signaling hypothesis (Leland and Pyle 1977; Jensen and Ruback 1983; DeAngelo, DeAngelo, and Rice 1984; and Travlos 1987), and the winner’s curse hypothesis (Varaiya and Ferris 1987). However, no study has examined whether merger announcement losses reflect the increased probability that an acquirer will face a lawsuit. In a fully efficient market, the probability of a lawsuit should be reflected in the market reaction to the merger announcement. However, extant studies suggest that the market does not efficiently process the valuation implications of stock-for-stock mergers. They find that acquirers underperform in the years after the merger announcement (Loughran and Vijh 1997; Louis 2004a; Moeller, Schlingemann, and Stulz 2005). As Jensen and Ruback (1983, p. 20) comment, the “post-outcome negative abnormal returns [after takeovers] are unsettling because they are inconsistent with market efficiency.” Thus, a priori, it is not clear whether, and to what extent, the market reaction to the merger announcement impounds the probability of a lawsuit. To examine the association between the merger announcement abnormal return and the probability of a lawsuit, we use an instrumental variable approach. In a first step, we estimate the probability that an acquirer would be sued, using ex-ante predictors of lawsuits. In a second step, we analyze the association between the merger announcement abnormal return and the probability of a lawsuit. We use the two-stage estimation process because of the potential endogeneity in the relation between lawsuits and performance.

3

A negative association between the probability of a post-merger announcement lawsuit and the market reaction to the merger announcement would be consistent with the notion that the merger announcement return impounds the probability of a lawsuit. However, it is plausible that the probability of a lawsuit is only partly impounded at the merger announcement. In this case, the post-merger stock performance will also be related to the probability of a post-merger announcement lawsuit. Thus, we also analyze the association between stock-for-stock acquirers’ long-term abnormal returns and the ex-ante probability of a post-merger announcement lawsuit. Consistent with our conjectures, we find that pre-merger abnormal accruals are a strong determinant of post-merger lawsuits. The effect of abnormal accruals is significant even after controlling for the post-merger abnormal return, which suggests that pre-merger earnings management has a first order effect on the likelihood of a lawsuit. We also find evidence that the market anticipates the lawsuits at merger announcements. There is a significantly negative association between the market reaction to a merger announcement and the probability that a stock-for-stock acquirer is subsequently sued. Further analyses suggest, however, that the market reaction to the merger announcement only partially reflects the probability of a lawsuit. First, we find that stock-for-stock acquirers’ long-term market underperformance is largely limited to litigated acquisitions. The average stock-for-stock acquirer that does not face a merger-related lawsuit experiences a statistically insignificant abnormal return of approximately -11.4% relative to a match firm over the four years after the merger announcement. In contrast, the average acquirer in litigated mergers experiences an abnormal return of -77.7% over the same horizon. Second, and more importantly, we find a very strong negative association between the likelihood of a lawsuit and the long-term market performance over the four years after the merger announcement. Therefore, post-merger

4

announcement long-term market performance can be predicted using lawsuit-related information that is available at the time of the merger announcement. We do not claim that lawsuits are the only cause of the post-merger announcement longterm underperformance. The evidence only indicates that lawsuits are a contributing factor to the underperformance, which is a very important finding, given the puzzling nature of the postmerger underperformance and the heightened interest in explaining this phenomenon. The remainder of the study is organized as follows. The next section discusses related studies. Section 3 describes the sample selection process. Section 4 presents univariate analyses. We conduct multivariate regression analyses in Section 5. The study concludes in Section 6.

2. Related studies and motivation Prior studies suggest that managers have strong incentives to inflate earnings prior to stock-for-stock mergers and provide supporting evidence that, on average, acquiring firms report large income-increasing abnormal accruals prior to merger announcements (Erickson and Wang 1999; and Louis 2004a). Jensen (2005) also contends that overvalued firms tend to inflate earnings and to undertake stock acquisitions to create the illusion of growth to satisfy market expectations. However, Ball and Shivakumar (2008) argue that, because large corporate events are “characterized by higher than usual litigation and regulatory risk from inflating earnings,” managers should be dissuaded from manipulating their reported earnings around these events. Notwithstanding Ball and Shivakumar’s (2008) argument, it is observed that acquirers often face lawsuits after stock-for-stock mergers. In addition, the most common complaint in these lawsuits is that managers have misguided investors by issuing false and misleading statements. As Ball and Shivakumar (2008) suggest, the lawsuits should caution managers

5

against inflating earnings prior to acquisitions. However, the large incidence of lawsuits after stock-for-stock mergers is an indication that managers and directors might not be as cautious as they should. Furthermore, the increased exposure to lawsuits does not necessarily mean that managers would not manage earnings. It implies that they may be dissuaded from managing earnings but, if they do manage earnings, they would be more likely to face lawsuits. Our study is closely related to DuCharme, Malatesta, and Sefcik (2004) who analyze the interaction between stock issuances, abnormal accruals, and lawsuits. However, the two studies are different in several aspects. We address three questions: (1) whether post-merger announcement lawsuits are associated with pre-merger abnormal accruals, (2) whether investors anticipate the potential lawsuits at merger announcements, and (3) whether post-merger lawsuits are a driver of the post-merger announcement market underperformance. DuCharme, Malatesta, and Sefcik (2004) address the first question in the context of stock issuances with mixed results and do not address the second and the third questions. DuCharme, Malatesta, and Sefcik (2004) find a significant association between abnormal accruals and lawsuits after seasoned equity offerings (SEOs). However, they fail to document a significant correlation between abnormal accruals and lawsuits after initial public offerings (IPOs), which is consistent with Ball and Shivakumar (2008) who conclude that the average IPO firm does not inflate earnings prior to the IPO. Therefore, because DuCharme, Malatesta, and Sefcik’s (2004) results are mixed, they cannot be generalized to other settings such as stock-forstock mergers. One reason why the IPO setting is likely different from the stock-for-stock merger setting is that, as Ball and Shivakumar (2008) recognize, an IPO firm is transitioning from private to public status and is subjected to very intensive scrutiny during the transition period by third

6

parties such as underwriters (Booth and Smith 1986; Beatty and Ritter 1986; Carter and Manaster 1990; Carter, Dark, and Singh 1998) and venture capitalists (Megginson and Weiss 1991; Barry, Muscarella, Peavy, and Vetsuypens 1990; Brav and Gompers 1997; Chemmanur and Loutskina 2006). Stock-for-stock acquirers could also be subjected to intense scrutiny by the target managers. However, as Shleifer and Vishny (2003) contend, the managers of a target firm could agree to a merger even if they know that the transaction is not in the best interest of their shareholders. Shleifer and Vishny (2003) suggest that the acquirer can buy their agreement through the acceleration in the exercise of stock options or by granting them generous severance pay. Target managers can also agree to mergers for reasons of retirement or illiquid stock ownership. Cai and Vijh (2007) also suggest that target CEOs who face high illiquidity discounts tend to leave after the acquisitions are completed. Therefore, these CEOs are not very concerned about acquirers’ long-term performance. As Louis (2004a) argues, the process leading up to a merger is different from the process leading up to an SEO. In particular, mergers are results of negotiations between acquirers and targets, whereas IPOs and SEOs involve only the issuing firms. In addition, in an IPO or an SEO, shareholders make their own decisions as to whether to acquire the issuer’s shares whereas, in a stock-for-stock merger, typically, the decision is practically imposed on the target’s shareholders by the target’s managers and directors. Given that, generally, the incentives of shareholders and managers are not perfectly aligned, issuers’ incentives to manage earnings, the probability of post-issuance lawsuits, and the associations between these factors and the long-term stock performance could vary across the different settings. Furthermore, while DuCharme, Malatesta, and Sefcik (2004) analyze the determinants of lawsuits after IPOs and SEOs, and the correlation between post-issuance lawsuits and long-term

7

stock performance, they do not analyze the extent to which the market anticipates the lawsuits. In addition, they document a significant correlation between long-term abnormal returns after stock issuances and post-issuance lawsuits, but they do not analyze whether lawsuits contribute to post-issuance underperformance. Because market performance is a determinant of lawsuits, an association between market performance and lawsuit does not imply that lawsuits are a cause of the underperformance. To assess whether lawsuits contribute to the post-merger announcement underperformance, we first estimate the probability of lawsuits using ex-ante determinants of lawsuits and then analyze the association between the estimated probability of lawsuits and the post-merger long-term abnormal returns.

3. Sample selection and descriptive statistics 3.1 Sample selection The study covers stock-for-stock mergers of U.S. companies that were announced between January 1996 and December 2002, inclusively. We start the sample in 1996 because data from our main source of information on lawsuits, the Securities Class Action Clearinghouse at Stanford University, are available starting in 1996. Necessary proxy statements in the historical Electronic Data Gathering, Analysis, and Retrieval (EDGAR) archives are also available starting in 1994. We end the sample in 2002 because we need to compute the abnormal returns over the four years after the merger announcements. The sample of mergers is obtained from the Security Data Company (SDC)’s online database of domestic mergers and acquisitions. A transaction is included in the sample if it satisfies the following criteria: (1) The merger is successfully completed;

8

(2) The completion date of the merger is available on SDC; (3) The target is coded on SDC as either a publicly traded company or a privately held company; (4) The total assets of the target are reported in SDC; (5) The acquirer is a nonfinancial company;1 (6) The necessary quarterly earnings announcement date for the acquirer is available on Compustat; (7) The acquirer has the necessary quarterly data on Compustat to estimate the abnormal current accrual for the earnings reports that take place within 91 days starting 89 days prior to the merger announcement; (8) In addition to the required data in (6) and (7), the acquirer has necessary quarterly Compustat data on equity book value, price, number of shares outstanding, and total assets; (9) The acquirer has a match firm within the parameters defined in Section 4.4 and there are necessary data on the Center for Research in Security Prices (CRSP) to compute post-merger announcement long-term abnormal returns for both the acquirer and the match firm; (10) There are necessary data on CRSP to compute pre-merger announcement abnormal returns for both the acquirer and the match firm; (11) The acquirer has the necessary data on CRSP to compute the merger announcement abnormal return; and (12) The equity holding of the acquirer’s Chief Executive Officer (CEO) is available on either the Wharton Research Data Services’ executive compensation database or from filings with the Securities and Exchange Commission (SEC). SDC contains 1,280 stock-for-stock mergers that satisfy conditions (1) to (5) over the period 1996–2002. A total of 405 of the bidders do not have (cusip) matches on CRSP. An additional 329 observations do not satisfy conditions (6) to (8). Another 26 observations do not satisfy condition (9) and an additional 11 observations do not satisfy condition (10). Finally, 14 bidders do not satisfy condition (12). In sum, there are 495 acquisitions that satisfy the sample selection criteria. Note that, relative to other studies, the sample size in this study is quite large. 1

Current assets (quarterly data item 40) and current liabilities (quarterly data item 49) necessary to compute the accruals, and auditor identity (annual data item 149) are not available on Compustat for the banking industry.

9

For instance, Erickson and Wang (1999) use a sample of 55 stock-for-stock acquirers and Louis (2004a) uses a sample of 236 stock-for-stock acquirers. To identify post-merger lawsuits, we search various sources for lawsuit information over the three years after the merger announcement. We first search the Securities Class Action Clearinghouse at Stanford University. Then we search Factiva, LexisNexis, ProQuest, America Newspapers, and Google.com for news releases using various keywords. The lawsuits must be caused by the mergers, use the mergers as justifications, or involve merger-related insider trading allegations. The percentage of merger-related lawsuits is 20.81% (103 out of 495).

3.2 Descriptive statistics Table 1 presents the distribution of the sample by industry and year, conditional on whether the acquirer is sued or not. Panel A of Table 1 presents the industry distribution of the sample. The largest concentrations of lawsuits are in business services, electronic and electrical equipment, and machinery and computer equipment, with 43.69%, 13.59%, and 8.74% of the sample, respectively. Note, however, that these industries also have the largest concentrations of stock-for-stock mergers in general. Business services, electronic and electrical equipment, and machinery and computer equipment represent 28.06%, 13.01%, and 8.93% of the acquisitions that are not associated with lawsuits, respectively. We present the time-series distribution of the sample in Panel B of Table 1. The ratio of litigated to non-litigated acquisitions is highest in 2000 and 2001. However, in general, it does not appear that the lawsuits are clustered in any specific year. Table 2 reports the characteristics of the sample. Acquisitions associated with lawsuits are less likely to be accounted for by the pooling-of-interest method than acquisitions that are not associated with lawsuits. Post-merger lawsuits are positively associated with the number of

10

stock-for-stock mergers that an acquirer recently completed and the total value of these transactions. Consistent with the notion that many lawsuits are filed because the acquirers have “deep pockets,” there is a positive association between the lawsuits and the market value of the acquirers. The medians of the value of the transactions and the book-to-market ratios are significantly larger for acquirers that are subsequently sued but the means are not. Consistent with Field, Lowry, and Shu (2005), lawsuits are positively associated with prior stock price volatility, prior share turnover, and technology firms.

4. Univariate analysis 4.1 Abnormal accruals prior to the merger announcement Following Teoh, Welch, and Wong (1998a, 1998b) and Louis (2004a), among others, we proxy for earnings management by the residual from a current accruals model. Using all firms that have the necessary data on Compustat, for each calendar quarter and two-digit SIC-code industry, we estimate the following model: 4

CACCi = Σλj-1Qj,i + λ4(∆SALESi-∆ARi) + λ5 LCACCi + λ6ASSETi + εi,

(1)

j=1

where CACC is current accruals, Qj is a binary variable taking the value one in fiscal quarter j and zero otherwise, ∆SALES is the quarterly change in sales, ∆AR is the quarterly change in accounts receivable, LCACC is the lag of CACC, ASSET is total assets at the beginning of the quarter; and ε is the regression residual.2 Consistent with Gong, Louis, and Sun (2008), we

2

CACC = ∆CA - ∆CL - ∆CASH +∆STD, where ∆CA is change in current assets (Compustat quarterly data item #40); ∆CL is change in current liabilities (#49); ∆CASH is change in cash and cash equivalents (#36); and ∆STD is change in debt included in current liabilities (#45). We measure current accruals using balance sheet information because using cash flow statement information substantially reduces our sample size. Hribar and Collins (2002) suggest that studies that use balance sheet accrual estimates are potentially contaminated because of non-articulation events such as mergers and divestitures. It is not clear whether this problem also affects quarterly accrual estimates. Nonetheless,

11

include the lag of the accrual variable and total assets in the discretionary current accrual model. All the variables, including the indicator variables, are scaled by total assets at the beginning of the quarter. After deflating the model, ASSET is transformed into a column of ones, which allows us to estimate the model with the standard intercept. We also require at least 20 observations for each estimation. To mitigate the effect of outliers and errors in the data, for each calendar quarter, we delete the top and bottom one-percentiles of the deflated CACC, ∆SALE-∆AR, and LCACC. We also delete the top and bottom one-percentiles of the estimated abnormal accruals after adjusting them for performance. We adjust the estimated abnormal accruals for performance based on Kothari, Leone and Wasley’s (2005) recommendation. Consistent with Louis (2004a), for each quarter and each industry (two-digit SIC code), we create five portfolios with at least four firms each by sorting the data into quintiles based on return-on-assets from the same quarter of the previous year. The performance-matched abnormal accruals for a sample firm are the firm-specific abnormal accruals minus the average abnormal accruals for its respective industry-performance matched portfolio. In addition to controlling for performance, the portfolio-benchmarking approach controls for random effects arising from other events that may affect accruals or other managerial incentives to manage earnings. As Kothari, Leone, and Wasley (2005) note, the success of the benchmarking approach is predicated on the assumption that the differences between the

to assess the extent to which our results might be biased, we replicate them for the sample firms that have working capital accrual information from the cash flow statement. We define current accruals as follows: CACC = - CHAR CHINV - CHAP - CHTAX - CHOTHER, where CHAR is change in accounts receivable (data item #103), CHINV is change in inventories (#104), CHAP is change in accounts payable (#105), CHTAX is change in accrued income tax (#106), and CHOTHER is net change in other current assets and liabilities (#107). Using this definition reduces our sample size to 214 observations, with 53 acquisitions that generated lawsuits and 161 acquisitions that did not; however, it does not change any of our inferences.

12

abnormal accruals of the sample firms and those of the control portfolios proxy for earnings management that relates solely to the mergers.3 Table 3 presents the abnormal accrual estimates. Consistent with Erickson and Wang (1999) and Louis (2004a), the average abnormal current accrual for the quarterly earnings announcement that immediately precedes the merger announcement is significantly positive. However, we also find that acquirers that face lawsuits after the merger announcements report significantly more abnormal accruals than those that do not. Acquirers’ average (median) abnormal accruals are 0.5% (0.4%) of total assets for non-litigated acquisitions and 1.7% (1.4%) for litigated acquisitions. The differences are significant below the 1% level. These results are consistent with our conjecture that post-merger lawsuits are associated with pre-merger abnormal accruals.

4.2 Abnormal returns prior to and around the merger announcement Table 4 reports the acquirers’ abnormal returns over the year ending two days before the merger announcement and the market adjusted returns over the three days centered on the earnings announcement. The pre-merger abnormal returns are the differences between the acquirers’ raw returns and the raw returns of their matched firms. The matching process is described in the next section. We find that acquirers that face lawsuits after merger announcements experience significantly lower abnormal returns in the year prior to the merger announcements than those that do not. Actually, the average pre-merger abnormal return is insignificantly negative for the acquirers that face lawsuits and significantly positive for those that do not. Relative to the returns

3

See Campbell and Stanley (1963) and Cook and Campbell (1979) for a discussion of the implications of using a benchmarking approach in general.

13

of the matched firms, acquirers’ average abnormal returns are 33.9% for non-litigated acquisitions and -6.6% for litigated acquisitions, which is consistent with the conjecture that past performance is negatively associated with lawsuits. As we discuss in the next section, acquirers that face lawsuits after merger announcements significantly underperform those that do not over the years after the merger announcement. Therefore, the superior pre-merger market performance of acquirers in non-litigated acquisitions suggests that pre-merger stock performance is not a good indication of undervaluation or overvaluation. Table 4 also reports acquirers’ abnormal returns around the merger announcement. Consistent with prior studies, the average acquirer’s abnormal return around stock-for-stock merger announcements is negative. However, we find that the average market reaction to merger announcements that are followed by lawsuits is about twice as negative as the average market reaction to merger announcements that are not followed by lawsuits (-3.8% versus -1.9%). This result is open to two plausible explanations. The merger announcement return could affect the likelihood of post-merger announcement lawsuits. Alternatively, the market could anticipate the lawsuits and adjust the acquirers’ stock prices accordingly at the merger announcements. Further analyses will help to disentangle these two potential explanations.

4.3 Post-merger announcement long-term performance of acquirers Following Barber and Lyon (1997) and Kothari and Warner (1997), we use a match-firm procedure to estimate the acquirers’ post-merger market performance. We estimate the long-run performance of an acquirer as the difference between its raw buy-and-hold return and the raw buy-and-hold return of a match firm. Following Barber and Lyon (1997), we choose the match firm from all firms with a market value of equity between 70% and 130% of the market value of

14

equity of the acquirer at the end of the fiscal quarter prior to the merger announcement. From this set of firms, we match the acquirer with the firm that has the closest book-to-market ratio.4 Panel A of Table 5 reports the acquirers’ long-term abnormal returns over the four years after the merger announcement for the full sample. Consistent with extant studies, we find that acquirers experience significantly negative post-merger announcement long-term abnormal returns. However, we also find that the long-term abnormal returns are largely limited to litigated acquisitions. In general, there is little evidence that the average acquirer that is not involved in post-merger announcement lawsuits experiences significant post-merger announcement abnormal returns. For these acquirers, the average abnormal return over the four years starting two days after the merger announcement is -11.4% (t-value = -1.27). In contrast, acquirers in litigated acquisitions experience an average abnormal return of -77.7% (t-value = -6.02). To assess the extent to which our results are affected by the market bubble of the late 1990s, we partition the sample period into the bubble period and the post-bubble period. The results are qualitatively similar across the two sub-periods. Some firms have multiple stock-for-stock acquisitions over relatively short periods and, consequently, the abnormal returns after these acquisitions overlap. To mitigate the effects of multiple acquisitions on our inferences, if a sample firm has multiple stock-for-stock mergers within a twelve-month period, we delete the earlier acquisitions and keep only the most recent one in the sample. The restriction does not qualitatively change the results. As reported in Panel B of Table 5, the average abnormal return for the restricted sample is -11.6% (t-value = -1.19) for acquirers in non-litigated mergers and -74.3% (t-value = -4.99) for acquirers in litigated acquisitions.

4

We use the same set of matched firms throughout the paper.

15

Barber and Lyon (1997) and Kothari and Warner (1997) suggest that the match-firm approach controls for the new listing, survivor, rebalancing, and skewness biases that arise in studies of long-run abnormal returns. However, Fama (1998) and Brav, Geczy, and Gompers (2000) suggest that events may be correlated in calendar time and that Barber and Lyon’s (1997) method may not fully account for the correlation. The problem with using a calendar-time portfolio approach in our setting, however, is that the number of litigated acquisitions is small, resulting in few or even zero observations in many months. The portfolios with few observations tend to have undue influences on the results. Nonetheless, to ensure that our inferences are not driven by cross-sectional correlations, we estimate average monthly long-term abnormal returns using Fama and French’s (1993) three-factor model. Untabulated results show that, for the full sample, the average monthly abnormal returns over the four years starting the month after the merger announcement are -0.3% (t-value = -0.72) and -1.8% (t-value = -2.62) for the nonlitigated acquisitions and the litigated acquisitions, respectively. The results are similar for the restricted sample. The average monthly abnormal returns are -0.3% (t-value = -0.69) and -1.5% (t-value = -2.23) for the non-litigated acquisitions and the litigated acquisitions, respectively.

5. Multivariate analysis 5.1 Determinants of post-merger lawsuits In this section, we analyze the determinants of post-merger announcement lawsuits using a multivariate probit regression. More specifically, we use the following model: SUITi = α1ABCAi + α2ABRETYM1i + α3VOLATILITYi + α4TURNOVERi + α5POOLi + α6INTRAi + α7COLLARi + α8LNUMBERMi + α9LTVALUEi + α10LVALUEi + α11LMKTVi + α12BIG4i + α13TECHi + α14RETAILi + α15REGi + Year fixed effectsi + εi (2)

16

where SUIT is an indicator variable taking the value one if the acquisition generates a lawsuit within the three years starting the day after the merger announcement date and zero otherwise; ABCA is the acquirer’s abnormal current accruals for the quarterly earnings announcement that precedes the merger announcement; ABRETYM1 is the acquirer’s abnormal return for the year ending two days prior to the merger announcement; VOLATILITY is the standard deviation of the acquirer’s daily returns over the year ending two days prior to the merger announcement; TURNOVER is [1 – Пt(1 - volume traded on day t / total shares outstanding on day t)], measured over the year ending two days prior to the merger announcement; POOL is a binary variable taking the value one if the merger is accounted for by the pooling-ofinterest method and zero if it is accounted for by the purchase method; INTRA is a binary variable taking the value one if the target and the acquirer are located in the same state and zero otherwise; COLLAR is a binary variable taking the value one if the merger has a collar and zero otherwise; LNUMBERM is the log of one plus the number of stock-for-stock mergers that the acquirer executes over the two years prior to the merger announcement as reported by SDC; LTVALUE is the log of one plus the total value of the stock-for-stock mergers that the acquirer executes over the two years prior to the merger announcement as reported by SDC; LVALUE is the log of the value of the current transaction as reported by SDC; LMKTV is the log of the acquirer’s market value of equity at the end of the quarter prior to the merger announcement; BIG4 is a binary variable that is equal to one if the acquirer is audited by a Big 4 audit firm and zero otherwise;5 TECH is a binary variable taking the value one if the acquirer is a technology company [SIC codes 2833–2836, 3570–3577, 3600–3674, 7371–7379 or 8731–8734] and zero otherwise; RETAIL is a binary variable taking the value one if the acquirer is a retail company [SIC codes 5200–5961] and zero otherwise; and 5

We use Big 4 generically to designate Big 4, Big 5, and Big 6 audit firms, depending on the period.

17

REG is a binary variable taking the value one if the acquirer is a regulated company [SIC codes 4812–4813, 4833, 4841, 4811–4899, 4922–4924, 4931, or 4941] and zero otherwise.6 We control for ABRETYM1 because past performance is likely to have an inverse relation with post-merger announcement lawsuits. We control for VOLATILITY and TURNOVER because Field, Lowry, and Shu (2005) find that prior stock volatility and prior share turnover are significant determinants of lawsuits. Field, Lowry, and Shu (2005) also conjecture that lawsuits could be related to regulated, retail, and technology firms. We therefore control for TECH, RETAIL, and REG. We control for POOL because, to ensure that a transaction met the pooling accounting method criteria, an acquirer needed the close cooperation of the target’s managers, which reduces the risk that the target’s managers would claim that they have been misled. The pooling-of-interest accounting method is not permitted for acquisitions initiated after June 30, 2001. We include COLLAR in the model because a collar protects the directors and managers against claims that they did not properly safeguard the interest of their shareholders. We control for INTRA because firms located in the same geographical area tend to share a common investor base (Coval and Moskowitz 2001; Brown et al. 2008; Grinblatt and Keloharju 2001; Pirinsky and Wang 2006), which could affect investors’ decision to sue an acquirer. We include the acquirers’ market value of equity (LMKTV) since many lawsuits are allegedly filed because the acquirers have “deep pockets.” We control for LNUMBERM, LTVALUE, and LVALUE because these factors are likely to increase the likelihood of a lawsuit. DuCharme, Malatesta, and Sefcik (2004) also find a positive correlation between a Big 4 audit firm and the likelihood that a client gets sued after stock issuances. However, because DuCharme, Malatesta, and Sefcik (2004) do not directly control for the size of the client, the Big 4 variable could capture the effect of size. While Big 4 firms presumably have “deep pockets” and are therefore more likely to attract 6

The sample does not include financial companies.

18

lawsuits, a more prestigious auditor could reduce the risk of lawsuits by increasing the perceived quality of the financial reports. The results of the probit model are reported in Table 6. In addition to our basic specification (Column 1), we use two other models (Column 2 and Column 3) in which we assess the potential effects of controlling for the market reaction to the merger announcement and the long-term abnormal returns after the merger announcement. Our basic model controls only for pre-determined variables. In general, the data fit the models quite well. The models classify about 73.7% to 76.1% of the observations correctly, the Nagelkerke’s (1991) pseudo-R2s range from 52.7% to 55.0%, and chi-squares testing whether the coefficient estimates are jointly zero are very significant. Consistent with our expectations, we find a significantly positive association between pre-merger abnormal accruals and the likelihood of a lawsuit, even after controlling for the other potential determinants of lawsuits, with a p-value of 0.005. We also find that the pre-merger announcement abnormal return, the pre-merger announcement stock volatility, and the number of acquisitions recently completed are strong determinants of the lawsuits. The univariate results indicate that lawsuits are positively correlated with the total value of stock-for-stock acquisitions recently completed, the value of the current acquisition, size, share turnover, and technology firms. However, we have no evidence that the coefficients of these variables are significantly positive after controlling for the other determinants of the lawsuits. In Column 2 of Table 6, we control for the potential effects of the market reaction to the merger announcement and the long-term abnormal returns after the merger announcement. We find that the effect of abnormal accruals is significant even after controlling for the post-merger announcement abnormal return, with a p-value of 0.008. This result suggests that pre-merger

19

earnings management likely has a first order effect on the likelihood of a lawsuit, which is incremental to the effect of the post-merger long-term performance. We also observe that, although the pre-merger abnormal return is a strong determinant of lawsuits, the market reaction to the merger announcement is not, after controlling for the other potential determinants of lawsuits. Although the market reaction does not seem to be a significant driver of the lawsuits, it is still plausible that the market impound the probability of the lawsuits.7 We test this hypothesis in the next section.

5.2 Probability of lawsuits and the market reaction to merger announcements Given the potential adverse effects of lawsuits, we expect a negative association between the probability of a lawsuit and the market reaction to a merger announcement. As explained earlier, the relation between a lawsuit and the market reaction to a merger announcement can be endogenous. The results reported in Table 6 suggest that the market reaction to the merger announcement is not a significant driver of post-merger announcement lawsuits, which alleviates the endogeneity concern. Nonetheless, to ensure that we mitigate potential endogeneity effects, we use an instrumental variable approach. In a first step, we use equation (2) (from the previous section) to estimate the probability that an acquirer would be sued, using ex-ante predictors of the lawsuit. In a second step, we analyze the association between the merger announcement abnormal return and the probability of a lawsuit. More specifically, we model the market reaction to the merger announcement as follows:

7

The coefficient on the market reaction is not statistically significant whether we consider the period prior to 2000 or the period afterwards. The coefficients on abnormal accruals and the pre-merger abnormal return are somewhat similar across the two subperiods. For instance, in Model 1, the coefficient on abnormal accruals is 4.649 (p-value = 0.059) over the period 1996-1999 and 4.552 (p-value = 0.059) over the period 2000-2002, and the coefficient on the abnormal return is -0.079 (p-value = 0.186) over the former period and -0.112 (p-value = 0.026) over the later period. Although the p-values are lower for the subsamples than for the full sample (due particularly to the reduction in the sample size), the magnitudes of the coefficients are similar. The coefficient on abnormal accruals is 4.599 (pvalue = 0.005) and the coefficient on the abnormal return is -0.090 (p-value = 0.013) for the full sample.

20

CARi = α0 + α1PSUITi + α2ABCAi + α3BMi + α4RSIZEi + α5PRIVATEi + α6POOLi + α7SAME_INDi + α8CEOSHi + εi, (3) where CAR is the acquirer’s market-adjusted return for the three days centered on the merger announcement date; PSUIT is the probability of a lawsuit estimated as the predicted value of SUIT from model (2); ABCA is the acquirer’s abnormal current accruals for the quarterly earnings announcement that precedes the merger announcement; BM is the acquirer’s book-to-market ratio at the end of the quarter prior to the merger announcement; RSIZE is the ratio of the total assets of the target to the total assets of the acquirer; PRIVATE is a binary variable taking the value one is the target is a privately held company and zero if it is a publicly traded company; POOL is a binary variable taking the value one if the merger is accounted for by the pooling-ofinterest method and zero if it is accounted for by the purchase method; SAME_IND is a binary variable taking the value one if the two merging partners are in the same two-digit SIC code and zero otherwise; and CEOSH is the total number of the acquirer’s common shares controlled by the chief executive officer (CEO) of the acquirer divided by the number of common shares outstanding.

We control for book-to-market ratios because Lang, Stulz, and Walking (1989) and Servaes (1991) report a significant correlation between Tobin’s q and bidders’ abnormal returns. Rau and Vermaelen (1998) also argue that high book-to-market acquirers tend to be more cautious before engaging in major transactions and their acquisitions are less likely to be motivated by hubris. Louis (2005) also finds a positive association between book-to-market ratios and acquirers’ abnormal returns around merger announcements. We control for the relative size of the target because prior studies find a negative correlation between relative size and bidders’ abnormal returns around merger announcements

21

(Scanlon, Trifts, and Pettway 1989; Louis 2004a, 2004b). We also control for private acquisitions because many studies find a positive average market reaction to announcements of stock-for-stock acquisitions of privately owned companies whereas the average reaction to acquisitions of publicly traded companies is negative (Chang 1998; Fuller, Netter, and Stegemoller 2002; Moeller, Schlingemann, and Stulz 2005; Louis 2005). We include the pooling variable because Lys and Vincent (1995) and Pandit (2003) suggest that, on average, pooling transactions were bad investments that were driven by accounting earnings instead of cash flows. We include SAME_IND, a measure of industry relatedness, because extant studies document a positive correlation between a firm’s stock return and the level of concentration in its lines of business (see Lang and Stulz 1994, among others). One suggested explanation relates to the coinsurance hypothesis; as two unrelated firms merge, the total value of their options to declare bankruptcy declines (Lewellen 1971; Scott 1977). Another explanation is that managers are more efficient in operating firms that focus on a single line of business as opposed to diversified firms. While diversification reduces the overall risk of the firm, it does not create value for investors because they can achieve their desired level of diversification on their own. The agency problems that lead managers to engage in non-value-maximizing investments could be mitigated as the CEO’s equity ownership in the company increases. However, as the stake of the CEO increases in the firm, he may have an incentive to undertake less profitable projects if they reduce the risk of his undiversified personal holding (Amihud and Lev, 1981). In addition, a CEO’s incentive to use overvalued equity as currency is likely to increase in his ownership in the firm. Therefore, a priori, it is not clear how the market reaction to the merger announcement will be correlated with CEO ownership.

22

The results are reported in Table 7. The evidence indicates that the market at least partly anticipates the lawsuits at the merger announcements. After controlling for the other determinants of the market reaction to merger announcements, we find a significantly negative association between the market reaction to a merger announcement and the probability that an acquirer is subsequently sued. Consistent with prior studies, we also find that the market reaction to a merger announcement is positively associated with book-to-market ratios and acquisitions of privately held targets.

5.3 Probability of lawsuits and post-merger announcement long-term stock performance Next, we analyze the extent to which the market fully anticipates the post-merger announcement lawsuits. In a fully efficient market, the probability of a post-merger lawsuit should be fully reflected in the market reaction to the merger announcement and, consequently, there should not be any association between the post-merger announcement stock performance and the ex-ante probability of a lawsuit. However, extant studies suggest that the market does not efficiently process the valuation implications of stock-for-stock mergers (Jensen and Ruback 1983; Loughran and Vijh 1997; Louis 2004a; Moeller, Schlingemann, and Stulz 2005). Therefore, it is plausible that market reactions to merger announcements do not fully reflect the probability of post-merger announcement lawsuits. To assess whether the post-merger announcement stock performance is associated with the ex-ante probability of a lawsuit, we again use an instrumental variable approach. In the first step, we use equation (2) (from Section 5.1) to estimate the probability that an acquirer would be sued, using ex-ante predictors of the lawsuit. In a second step, we analyze the association between the abnormal return over the four years starting two days after the merger announcement and the probability of a lawsuit.

23

We condition the analysis on the acquirer’s pre-merger abnormal accruals, the acquirer’s book-to-market ratio, the relative size of the merging partners, acquisitions of privately held companies, the industry relatedness of the merging firms, the method used to account for the transaction, and the acquirer’s CEO stock ownership. More specifically, we estimate the following regression model: ABRETYP4i = α0 + α1PSUITi + α2ABCAi + α3BMi + α4RSIZEi + α5PRIVATEi + α6POOLi + α7SAME_INDi + α8CEOSHi + εi,

(4)

where ABRETYP4 is the long-run abnormal returns measured over the four years after the merger announcement using the match-firm approach suggested by Barber and Lyon (1997).8 The other variables are defined as before. It is important to note that, with the exception of abnormal accruals, there is no evidence that the variables in the lawsuit prediction model [equation (2) in Section 5.1] are associated with long-term abnormal returns after stock-for-stock mergers. Prior studies identify only two factors that are associated with long-term market performance after stock-for-stock mergers: pre-merger abnormal accruals (Louis 2004a) and book-to-market ratio (Rau and Vermaelen 1998). Rau and Vermaelen (1998) argue that high book-to-market acquirers tend to be more cautious before engaging in major transactions and that their acquisitions are less likely to be motivated by hubris. Accordingly, they suggest that acquirers’ long-term market performance is positively associated with book-to-market. However, Louis (2004a) finds no evidence that the long-term market performance is associated with book-to-market ratios. Therefore, so far, abnormal accrual is the only factor that is found to be reliably associated with long-term market performance. 8

We do not use Fama and French’s (1993) three-factor model to estimate the long-term abnormal returns because the abnormal returns obtained under this procedure are averaged across firms every month and, hence, are not conducive to cross-sectional regression analyses. We could compute coefficient estimates by month or by year, and then use the means and standard deviations of the time-series of the coefficient estimates to make inferences (Fama and MacBeth 1973). However, there are too few observations to run monthly or annual regressions.

24

The results on the association between the probability of a lawsuit and post-merger announcement long-term market performance are reported in Table 8. We find a significantly negative association between the probability of a post-merger lawsuit and post-merger long-term abnormal returns, suggesting that the market reaction to a merger announcement only partially reflects the probability of a lawsuit. The association is significant even after controlling for premerger abnormal accruals. Therefore, part of the association between the post-merger long-term abnormal return and the probability of a post-merger lawsuit is incremental to the effect of the pre-merger abnormal accruals on the long-term performance. Overall, the results suggest that, not only are the lawsuits associated with the post-merger underperformance, but they are also likely a driver of the underperformance. Consistent with Louis (2004a), there is a negative association between the post-merger long-term abnormal returns and the pre-merger abnormal accruals. However, the association weakens after controlling for the probability of post-merger lawsuits, suggesting that some of the effects of abnormal accruals on post-merger stock performance are related to the effect of the abnormal accruals on the lawsuits. In general, there is no evidence that the long-term abnormal return is associated with the other control variables. We include them in the model to ensure that our inferences are not due to the omission of potentially relevant correlated variables. However, the results are qualitatively similar if we exclude them from the model.

6. Conclusion We analyze the interactions among stock-for-stock acquirers’ pre-merger accrual inflation, post-merger announcement lawsuits, and acquirers’ market performance. First, we document a significantly positive association between pre-merger announcement abnormal accruals and post-merger announcement lawsuits. The abnormal accrual effect is significant even

25

after controlling for post-merger announcement abnormal returns. Second, we find a significantly negative association between the market reaction to a merger announcement and the probability of a post-merger announcement lawsuit. Finally, we find a significantly negative association between post-merger announcement long-term abnormal returns and the probability of a post-merger lawsuit. The evidence that the market reaction to a merger announcement is negatively associated with the probability of a lawsuit indicates that the market anticipates post-merger announcement lawsuits at merger announcements. However, the subsequent evidence that the probability of a lawsuit is also related to post-merger announcement long-term abnormal returns suggests that the market reaction to the merger announcement only partially reflects the probability of a lawsuit and that the market likely underestimates the contingent losses associated with stock-for-stock mergers. In particular, the evidence suggests that, not only are the lawsuits associated with postmerger announcement underperformance, but they are also likely a driver of the underperformance. This result is very important, given the puzzling nature of the post-merger underperformance and the heightened interest in explaining this phenomenon. The study also provides evidence on the potential legal costs of earnings management. The results suggest that the legal costs associated with earnings management are very destructive. It is therefore important for investors to not only undo the direct stock price effects of earnings management but also factor the contingent losses associated with earnings management.

26

References Amihud, Y. and B. Lev, 1981, Risk reduction as a managerial motive for conglomerate mergers. Bell Journal of Economics 12, 605-617. Ball, R. and L. Shivakumar, 2008, Earnings quality at initial public offerings: Managerial opportunism or public-firm conservatism. Journal of Accounting and Economics (forthcoming). Barber, R., and J. Lyon, 1997, Detecting long-run abnormal stock returns: The empirical power and specifications of test statistics. Journal of Financial Economics 43, 341-372. Barry, C., C. Muscarella, J. Peavy, and M. Vetsuypens, 1990, The role of venture capital in the creation of public companies: Evidence from the going public process. Journal of Financial Economics 27, 447-471. Beatty, R., and J. Ritter, 1986, Investment banking, reputation, and the underpricing of initial public offerings. Journal of Financial Economics 15, 213-232. Booth, J., and R. Smith, 1986, Capital raising, underwriting and the certification hypothesis. Journal of Financial Economics 15, 261-281. Brav, A., C. Geczy, and P. Gompers, 2000, Is the abnormal return following equity issuances anomalous? Journal of Financial Economics 56, 209-240. Brav, A., and P. Gompers, 1997, Myth or reality? The long-run underperformance of initial public offerings: Evidence from venture and nonventure capital-backed companies. Journal of Finance 52, 1791-1821. Brown, J., Z. Ivković, P. Smith, and S. Weisbenner, 2008, Neighbors matter: Causal community effects and stock market participation. Journal of Finance (forthcoming).

27

Cai, J., and A. Vijh, 2007, Incentive effects of stock and options holdings of target and acquirer CEOs. Journal of Finance 62, 1891-1933. Campbell, D. and J. Stanley, 1963, Experimental and quasi-experimental designs for research. Chicago: Rand McNally. Carter, R., F. Dark, and A. Singh, 1998, Underwriter reputation, initial returns, and the long-run performance of IPO stocks. Journal of Finance 53, 285-311. Carter, R., and S. Manaster, 1990, Initial public offerings and underwriter reputation. Journal of Finance 45, 1045-1068. Chang, S., 1998, Takeovers of privately held targets, methods of payment, and bidder returns. Journal of Finance 53: 773-784. Chemmanur, T. and E. Loutskina, 2006, The role of venture capital backing in initial public offerings: Certification, screening, or market power? Working paper. Cook, R., 1977, Detection of influential observations in linear regression. Technometrics 19, 1518. Cook, T. and D. Campbell, 1979, Quasi-experimentation: Design and analysis issues for field settings. Boston: Houghton-Mifflin. Coval, J., and T. Moskowitz, 2001, The geography of investment: Informed trading and asset prices. Journal of Political Economy 109, 811-841. DeAngelo, H., L. DeAngelo, and E. Rice, 1984, Going private: Minority freezeouts and stockholder wealth. Journal of Law and Economics 27, 367-401. DuCharme, L., P. Malatesta, and S. Sefcik, 2004, Earnings management, stock issues, and shareholder lawsuits, Journal of Financial Economics 71, 27-49.

28

Erickson, M., and S. Wang, 1999, Earnings management by acquiring firms in stock for stock mergers. Journal of Accounting and Economics 27, 149-176. Etzel, B., 2002, A chill wind on tech mergers: HP-Compaq controversy could stall further M&A activity in sector. Investment Dealers Digest, April 15. Fama, E., 1998, Market efficiency, long-term returns, and behavioral finance. Journal of Financial Economics 49, 283-306. Fama, E., and K. French, 1993, Common risk factors in the returns on stocks and bonds. Journal of Financial Economics 33, 3-56. Fama, E., and J. MacBeth, 1973, Risk, return, and equilibrium: Empirical tests. Journal of Political Economy 81, 607-636. Field, L., M. Lowry, and S. Shu, 2005, Does disclosure deter or trigger litigation? Journal of Accounting and Economics 39, 487-507. Fuller K., J. Netter J, and M. Stegemoller, 2002, What do returns to acquiring firms tell us? Evidence from firms that make many acquisitions. Journal of Finance 57, 1763-1793. Gong, G., H. Louis, and A. Sun, 2008, Earnings management and firm performance following open-market repurchases. Journal of Finance 63, 947-986. Grinblatt, M., and M. Keloharju, 2001, How distance, language, and culture influence stockholdings and trades. Journal of Finance 56, 1053-1073. Hribar, P., and D. Collins, 2002, Errors in estimating accruals: Implications for empirical research. Journal of Accounting Research 40, 105-134. Jensen, M., 2005, Agency costs of overvalued equity. Financial Management 34, 5-19. Jensen, M., and R. Ruback, 1983, The market for corporate control: The scientific evidence. Journal of Financial Economics 11, 5-50.

29

Kothari, S., A. Leone, and C. Wasley, 2005, Performance matched discretionary accruals. Journal of Accounting and Economics 39, 163-197. Kothari, S., and J. Warner, 1997, Measuring long-horizon security price performance. Journal of Financial Economics 43, 301-339. Lang, L., and R. Stulz, 1994, Tobin’s q, corporate diversification and firm performance. Journal of Political Economy 102, 1248-1280. Lang, L., R. Stulz, and R. Walking, 1989, Managerial performance, Tobin’s q and the gain from successful tender offers. Journal of Financial Economics 24, 137-154. Leland, H. and D. Pyle, 1977, Informational asymmetries, financial structure, and financial intermediation. Journal of Finance 32, 371-387. Lewellen, W., 1971, A pure financial rationale for the conglomerate merger. Journal of Finance 26, 521-537. Loughran, T. and A. Vijh, 1997, Do long-term shareholders benefit from corporate acquisitions? Journal of Finance 52, 1765-1790. Louis, H., 2004a, Earnings management and the market performance of acquiring firms. Journal of Financial Economics 74, 121-148. Louis, H., 2004b, The cost of using bank mergers as defensive mechanisms against takeover threats. Journal of Business 77, 295-310. Louis, H., 2005, Acquirers’ abnormal returns and the non-Big 4 auditor clientele effect. Journal of Accounting and Economics 40, 75-99. Lys, T., and L. Vincent, 1995, An analysis of value destruction in AT&T’s acquisition of NCR. Journal of Financial Economics 39, 353-378.

30

Megginson, M., and K. Weiss, 1991, Venture capital certification in initial public offerings. Journal of Finance 46, 879-904. Moeller, S., F. Schlingemann, and R. Stulz, 2005, Wealth destruction on a massive scale? A study of acquiring-firm returns in the recent merger wave. Journal of Finance 60, 757-782. Nagelkerke, N., 1991, A note on general definition of the coefficient of determination. Biometrika 78, 691–692. Pandit, S., 2007, Accounting choice, announcement returns, and operating performance in stockfor-stock acquisitions. Working paper – Ohio State University. Pirinsky, C. and Q. Wang, 2006, Does corporate headquarters location matter for stock returns? Journal of Finance 61, 1991-2015. Rau, P., and T. Vermaelen, 1998, Glamour, value, and the post-acquisition performance of acquiring firms. Journal of Financial Economics 49, 223-253. Roll, R., 1986, The hubris hypothesis of corporate takeovers. Journal of Business 59, 197-216. Scanlon, K., J. Trifts, and R. Pettway, 1989, Impacts of relative size and industrial relatedness on returns to shareholders of acquiring firms. Journal of Financial Research 12, 103-112. Scott, J., 1977, On the theory of conglomerate mergers. Journal of Finance 32, 1235-1250. Servaes, H., 1991, Tobin’s Q and the gains from takeovers. The Journal of Finance 46, 409-419. Shleifer, A., and R. Vishny, 2003, Stock market driven acquisitions. Journal of Financial Economics 70, 295-311. Sullivan, B., 2002, HP/Compaq – Analysts to Hewlett – It’s time to move on. IDG Data, April 15. Swartz, J., 2002, Members of H-P integration team distracted by lawsuit. USA Today, April 15, B.03.

31

Teoh, S., I. Welch., T. Wong, 1998a, Earnings management and the long-run market performance of initial public offerings. Journal of Finance 53, 1935-1974. Teoh, S., I. Welch, and T. Wong, 1998b, Earnings management and the underperformance of seasoned equity offerings. Journal of Financial Economics 50, 63-100. Thompson, R., and R. Thomas, 2003, The new look of shareholder litigation: Acquisitionoriented class actions. Vanderbilt Law Review. Available at: http://ssrn.com/abstract=407580 or DOI: 10.2139/ssrn.407580. Travlos, N., 1987, Corporate takeover bids, methods of payment, and bidding firms’ stock returns. Journal of Finance 42, 943-963. Varaiya, N., and K. Ferris, 1987, Overpaying in corporate takeovers: The winner’s curse. Financial Analysts Journal 43, 64-70.

32

Table 1 Industry and event-year distribution of the sample Panel A. Sample distribution by industry SIC code Industry

10 13 16 20 23

Metal mining

24

Lumber and wood products

26 27 28 29

Paper and allied products

30 32 33 34

Rubber and plastics products

35 36 37 38 39

Machinery and computer equipment

42 45 48

Motor freight transportation

49 50 51 53 57

Utility services

58 59

Eating and drinking places

Oil and gas exploration Heavy construction Food and kindred products Apparels

Printing and publishing Chemicals Petroleum Refining and Related Industries

Stone, clay, glass, and concrete products Primary metals Fabricated metal products

Electronic and electrical equipment Transportation equipment Measuring equipment Miscellaneous Manufacturing Industries

Transportation By Air Communications

Wholesale – durable Wholesale – nondurable General Merchandise Stores Home Furniture, Furnishings, and Equip.

Miscellaneous retail

33

Non-litigated acquisitions

Litigated acquisitions

# of obs. Percentage

# of obs. Percentage

3 14 1 2 1

0.77 3.57 0.26 0.51 0.26

0 1 0 0 0

0.00 0.97 0.00 0.00 0.00

2

0.51

0

0.00

1 1 25 2

0.26 0.26 6.38 0.51

1 0 6 0

0.97 0.00 5.83 0.00

2 1 3 3

0.51 0.26 0.77 0.77

0 0 0 0

0.00 0.00 0.00 0.00

35 51 1 40 1

8.93 13.01 0.26 10.2 0.26

9 14 0 2 0

8.74 13.59 0.00 1.94 0.00

3 1 21

0.77 0.26 5.36

0 0 8

0.00 0.00 7.77

17 1 9 3 0

4.34 0.26 2.3 0.77 0.000

1 1 2 0 1

0.97 0.97 1.94 0.00 0.97

3 3

0.77 0.77

2 0

1.94 0.00

70 73 79

Hotels and other lodging places

80 82 83 87

Health services

Business services Recreation services

Educational Services Social Services Engineering and management services

Total

3 110 3

0.77 28.06 0.77

0 45 0

0.00 43.69 0.00

9 1 1 15

2.3 0.26 0.26 3.83

3 0 0 7

2.91 0.00 0.00 6.80

392

100.00

103

100.00

Panel B. Sample distribution by year Year

Non-litigated acquisitions

Litigated acquisitions

1996 1997 1998 1999 2000

# of obs. 39 62 79 81 67

Percentage 9.95 15.82 20.15 20.66 17.09

# of obs. 10 10 13 19 31

Percentage 9.71 9.71 12.62 18.45 30.10

2001 2002 Total

42 22 392

10.71 5.61 100.00

16 4 103

15.53 3.88 100.00

34

Ratio of litigated to non-litigated acquisitions 0.256 0.161 0.165 0.235 0.463 0.381 0.182 0.263

Table 2: Descriptive statistics Non-litigated acquisitions (N = 392)

Litigated acquisitions (N = 103)

Mean

Median

Mean

Median

PRIVATE POOL SAME_IND

0.173 0.520 0.566

0.000 1.000 1.000

0.204 0.388** 0.563

0.000 0.000** 1.000

INTRA COLLAR RSIZE

0.329 0.138 0.369

0.000 0.000 0.144

0.272 0.107 0.277*

0.000 0.000 0.144

0.589 343.075

0.000 0.000

1.777*** 1585.000**

1.000*** 3.234***

1354.000 11949.000 0.294

174.998 1166.000 0.200

1709.000 26444.000** 0.294

294.746*** 2642.000*** 0.158**

BIG4 VOLATILITY TURNOVER CEO_OWN

0.959 0.044 0.768 0.062

1.000 0.040 0.838 0.016

0.961 0.053*** 0.829*** 0.058

1.000 0.050*** 0.908*** 0.013

TECH RETAIL REG

0.485 0.023 0.082

0.000 0.000 0.000

0.602** 0.029 0.087

1.000** 0.000 0.000

NUMBERM TVALUE LVALUE MKTV BM

Notes: PRIVATE is a binary variable taking the value one is the target is a privately held company and zero if it is a publicly traded company; POOL is a binary variable taking the value one if the merger is accounted for by the pooling-ofinterest method and zero if it is accounted for by the purchase method; SAME_IND is a binary variable taking the value one if the two merging partners are in the same two-digit SIC code and zero otherwise; INTRA is a binary variable taking the value one if the target and the acquirer are located in the same state and zero otherwise; COLLAR is a binary variable taking the value one if the merger has a collar and zero otherwise;

35

RSIZE, relative size, is the ratio of the target’s total assets to the acquirer’s total assets; NUMBERM is the number of stock-for-stock mergers that the acquirer executes over the two years prior to the merger announcement as reported by SDC; TVALUE is the total value of the stock-for-stock mergers that the acquirer executes over the two years prior to the merger announcement as reported by SDC; VALUE is the value of the current transaction as reported by SDC; MKTV is the acquirer’s market value of equity at the end of the quarter prior to the merger announcement; BM is the acquirer’s book-to-market ratio at the end of the quarter prior to the merger announcement; BIG4 is a binary variable that is equal to one if the acquirer is audited by a Big 4 audit firm and zero otherwise; VOLATILITY is the standard deviation of the acquirer’s daily returns over the year ending two days prior to the merger announcement; TURNOVER is [1 – Пt(1 - volume traded on day t / total shares outstanding on day t)], measured over the year ending two days prior to the merger announcement; TECH is a binary variable taking the value one if the acquirer is a technology company [SIC codes 2833–2836, 3570–3577, 3600–3674, 7371–7379 or 8731–8734] and zero otherwise; RETAIL is a binary variable taking the value one if the acquirer is a retail company [SIC codes 5200–5961] and zero otherwise; and REG is a binary variable taking the value one if the acquirer is a retail company [SIC codes 4812–4813, 4833, 4841, 4811–4899, 4922–4924, 4931, or 4941] and zero otherwise. *** **

, , and * indicate that the difference between the two groups of firms is significant at the 1, 5, and 10 percent levels in a two-tail test, respectively. We use the t-test for the mean and the Wilcoxon two-sample test for the median. The t-test assumes unequal variances.

36

Table 3: Acquirers’ abnormal current accruals prior to the merger announcement Non-litigated acquisitions (N = 392)

Litigated acquisitions (N = 103)

Differences

Mean

0.005++ (2.26)

0.017+++ (3.91)

0.012+++ (2.52)

Median

0.004++ (4411)

0.014+++ (1342)

0.010+++ (3.03)

Notes: Abnormal accrual is measured in the quarter prior to the earnings announcement preceding the merger announcement. It is proxied by the difference between the unexplained current accrual (UECA) of an acquirer and the average UECA of a portfolio matched on industry and performance four quarters prior to the estimation quarter. UECA is the residual of the following regression: CACCi = Σλj-1Qj,i + λ4(∆SALESi-∆ARi) + λ5 LCACCi + λ6ASSETi + εi, where CACC is current accruals, Qj is a binary variable taking the value one in fiscal quarter j and zero otherwise, ∆SALES is the quarterly change in sales, ∆AR is the quarterly change in accounts receivable, LCACC is the lag of CACC, ASSET is total assets at the beginning of the quarter; and ε is the regression residual. T-values for the means are in parentheses. We also report the Wilcoxon Signed Rank statistic and the Z-statistics from the Wilcoxon two-sample test for the median in parentheses. +++ and ++ indicate significance at the 1 and 5 percent levels in a one-tail test, respectively. The test of mean difference assumes unequal variances.

37

Table 4: Average acquirers’ abnormal returns prior to and around the merger announcement Non-litigated acquisitions (N = 392)

Litigated acquisitions (N = 103)

Mean difference

Abnormal returns over the year prior to the merger announcement

0.339+++ (3.64)

-0.066 (-0.29)

-0.405+ (-1.64)

Abnormal returns around the merger announcement

-0.019+++ (-3.48)

-0.038+++ (-2.85)

-0.019+ (-1.29)

Notes: The pre-merger announcement abnormal returns are the buy-and-hold abnormal returns over the period the year ending two days prior to the merger announcement. They are computed as the difference between the raw returns of the sample firms and the raw returns of the matched firms, using the matching approach suggested by Barber and Lyon (1997). The abnormal returns around to the merger announcement are the market-adjusted return for the three days centered on the merger announcement date. T-values are presented in parentheses. +++, ++, and + indicate significance at the 1, 5, and 10 percent levels in a one-tail test, respectively. The test of mean difference assumes unequal variances.

38

Table 5: Acquirers’ average long-term abnormal returns after the merger announcement Panel A: Full sample All acquisitions

Non-litigated acquisitions

Litigated acquisitions

Full period: 1996 – 2002

-0.253+++ (-3.27) {N = 495}

-0.114 (-1.27) {N = 392}

-0.777+++ (-6.02) {N = 103}

Bubble period: 1996 – 1999

-0.234+++ (-2.35) {N = 313}

-0.101 (-0.89) {N = 261}

-0.897+++ (-5.80) {N = 52}

Post-bubble period: 2000 – 2002

-0.286++ (-2.32) {N = 182}

-0.142 (-0.95) {N = 131}

-0.655+++ (-3.15) {N = 51}

All acquisitions

Non-litigated acquisitions

Litigated acquisitions

Full period: 1996 – 2002

-0.235+++ (-2.79) {N = 442}

-0.116 (-1.19) {N = 358}

-0.743+++ (-4.99) {N = 84}

Bubble period: 1996 – 1999

-0.225++ (-2.05) {N = 274}

-0.108 (-0.87) {N = 233}

-0.890+++ (-5.10) {N = 41}

Post-bubble period: 2000 – 2002

-0.251++ (-1.91) {N = 168}

-0.131 (-0.84) {N = 125}

-0.603+++ (-2.52) {N = 43}

Panel B: Restricted sample

Notes: The abnormal returns are buy-and-hold abnormal returns over the four years starting two days after the merger announcement, using the match-firm approach of computing abnormal returns suggested by Barber and Lyon (1997). Some firms have multiple stock-for-stock acquisitions over relatively short periods and, consequently, the abnormal returns after these acquisitions overlap. To mitigate the effects of multiple acquisitions on our inferences, we delete the earlier acquisitions and keep only the most recent one in the sample if a sample firm has multiple stock-for-stock mergers within a twelvemonth period. These results are reported under the label “Restricted sample.” T-values for the means are in parentheses. +++ and ++ indicate significance at the 1 and 5 percent levels in a one-tail test, respectively. The tests of mean differences assume unequal variances.

39

Table 6: Probit regression of post-merger lawsuits on abnormal accruals and other potential determinants of the lawsuits (N = 495) Parameter

(1) Coefficient ChiEstimate Square

(2) Coefficient ChiEstimate Square

pvalue

(3) pvalue

Coefficient ChiEstimate Square

pvalue

ABCA

4.599

8.046

0.005

4.280

6.955

0.008

4.656

8.310

0.004

ABRETYM1

-0.090

6.134

0.013

-0.088

5.692

0.017

_

_

_

CAR

_

_

_

-0.219

0.133

0.716

-0.353

0.358

0.550

ABRETYP4

_

_

_

-0.131

7.818

0.005

_

_

_

VOLATILITY

15.510

11.622

0.001

13.901

8.902

0.003

15.121

11.118

0.001

TURNOVER

0.067

0.031

0.861

0.080

0.042

0.839

-0.042

0.012

0.912

POOL

-0.344

3.810

0.051

-0.339

3.608

0.058

-0.300

2.947

0.086

INTRA

-0.221

1.940

0.164

-0.213

1.751

0.186

-0.242

2.337

0.126

COLLAR

-0.077

0.137

0.711

-0.106

0.252

0.616

-0.040

0.037

0.847

LNUMBERM

0.554

4.363

0.037

0.588

4.877

0.027

0.620

5.566

0.018

LTVALUE

-0.040

0.502

0.479

-0.045

0.613

0.434

-0.052

0.857

0.355

LVALUE

0.085

2.583

0.108

0.077

2.000

0.157

0.070

1.739

0.187

LMKTV

0.117

4.152

0.042

0.112

3.678

0.055

0.110

3.701

0.054

BIG4

0.036

0.009

0.923

0.041

0.012

0.911

0.093

0.063

0.801

TECH

0.122

0.552

0.457

0.097

0.334

0.563

0.114

0.479

0.489

RETAIL

0.567

1.681

0.195

0.500

1.236

0.266

0.593

1.911

0.167

REG

-0.011

0.002

0.969

-0.013

0.002

0.964

0.051

0.032

0.859

40

Yes

Yes

Yes

Likelihood ratio statistics (p-value)

255.41 (0.000)

263.51 (0.000)

249.16 (0.000)

Score statistics (p-value)

220.27 (0.000)

223.62 (0.000)

215.72 (0.000)

Wald statistics (p-value)

190.72 (0.000)

188.98 (0.000)

187.93 (0.000)

74.1 (25.6)

76.1 (23.6)

73.7 (25.9)

0.537

0.550

0.527

Year fixed effects

Percent concordant (discordant) Max-rescaled R-square Notes:

SUIT is an indicator variable taking the value one if the acquisition generates a lawsuit within the three years starting the day after the merger announcement date and zero otherwise; ABCA is the acquirer’s abnormal current accruals for the quarterly earnings announcement that immediately precedes the merger announcement; ABRETYM1 is the acquirer’s abnormal return for the year ending two days prior to the merger announcement; CAR is the acquirer’s market-adjusted return for the three days centered on the merger announcement date; ABRETYP4 is the long-run abnormal returns measured over the four years starting two days after the merger announcement; VOLATILITY is the standard deviation of the acquirer’s daily returns over the year ending two days prior to the merger announcement; TURNOVER is [1 – Пt(1 - volume traded on day t / total shares outstanding on day t)], measured over the year ending two days prior to the merger announcement; POOL is a binary variable taking the value one if the merger is accounted for by the pooling-of-interest method and zero if it is accounted for by the purchase method; INTRA is a binary variable taking the value one if the target and the acquirer are located in the same state and zero otherwise;

41

COLLAR is a binary variable taking the value one if the merger has a collar and zero otherwise; LNUMBERM is the log of one plus the number of stock-for-stock mergers that the acquirer executes over the two years prior to the merger announcement as reported by SDC; LTVALUE is the log of one plus the total value of the stock-for-stock mergers that the acquirer executes over the two years prior to the merger announcement as reported by SDC; LVALUE is the log of the value of the current transaction as reported by SDC; LMKTV is the log of the acquirer’s market value of equity at the end of the quarter prior to the merger announcement; BIG4 is a binary variable that is equal to one if the acquirer is audited by a Big 4 audit firm and zero otherwise; TECH is a binary variable taking the value one if the acquirer is a technology company [SIC codes 2833–2836, 3570–3577, 3600– 3674, 7371–7379 or 8731–8734] and zero otherwise; RETAIL is a binary variable taking the value one if the acquirer is a retail company [SIC codes 5200–5961] and zero otherwise; and REG is a binary variable taking the value one if the acquirer is a retail company [SIC codes 4812–4813, 4833, 4841, 4811–4899, 4922–4924, 4931, or 4941] and zero otherwise.

42

Table 7: Conditional association between the market reaction to the merger announcement and the probability of post-merger lawsuits (N = 495) SUITi = α1ABCAi + α2ABRETYM1i + α3VOLATILITYi + α4TURNOVERi + α5POOLi + α6INTRAi + α7COLLARi + α8LNUMBERMi + α9LTVALUEi + α10LVALUEi + α11LMKTVi + α12BIG4i + α13TECHi + α14RETAILi + α15REGi + Year fixed effectsi (Step 1) + εi CARi = α0 + α1PSUITi + α2ABCAi + α3BMi + α4RSIZEi + α5PRIVATEi + α6POOLi + α7SAME_INDi + α8CEOSHi + εi (1)

(2)

(3)

Intercept

-0.031** (-2.12)

-0.051*** (-4.32)

-0.031** (-2.11)

PSUIT

-0.068++ (-2.03)

_

-0.071++ (-2.22)

ABCA

-0.030 (-0.25)

-0.110 (-0.94)

_

0.053+++ (3.59)

0.058+++ (3.97)

0.052+++ (3.59)

-0.009 (-1.14)

-0.009 (-1.04)

-0.010 (-1.17)

0.079+++ (5.86)

0.078+++ (5.75)

0.079+++ (5.85)

POOL

0.020 (1.86)

0.028 (2.68)

0.020 (1.85)

SAME_IND

-0.014 (-1.39)

-0.015 (-1.44)

-0.014 (-1.37)

CEOSH

-0.092** (-2.00)

-0.085* (-1.80)

-0.094** (-2.00)

Adj. R2

0.105

0.099

0.107

BM RSIZE PRIVATE

(Step 2)

Notes: CAR is the acquirer’s market-adjusted return for the three days centered on the merger announcement date; PSUIT is the probability of a lawsuit estimated as the predicted value of SUIT from Step 1 -- the variables and results for the first-step regression are presented in Table 5; ABCA is the acquirer’s abnormal current accruals for the quarterly earnings announcement that immediately precedes the merger announcement;

43

BM is the acquirer’s book-to-market ratio at the end of the quarter prior to the merger announcement; RSIZE is the ratio of the total assets of the target to the total assets of the acquirer; PRIVATE is a binary variable taking the value one is the target is a privately held company and zero if it is a publicly traded company; POOL is a binary variable taking the value one if the merger is accounted for by the pooling-ofinterest method and zero if it is accounted for by the purchase method; SAME_IND is a binary variable taking the value one if the two merging partners are in the same two-digit SIC code and zero otherwise; and CEOSH is the total number of the acquirer’s common shares controlled by the chief executive officer (CEO) of the acquirer divided by the number of common shares outstanding. The variables and results for the first-step regression are presented in Table 6. We find no evidence of influential observations using Cook’s (1977) distance statistic. +++, ++, and + (***, **, and *) indicate significance at the 1, 5, and 10 percent levels in a one- (two-) tail test, respectively.

44

Table 8: Conditional association between post-merger abnormal returns and the probability of post-merger lawsuits SUITi = α1ABCAi + α2ABRETYM1i + α3VOLATILITYi + α4TURNOVERi + α5POOLi + α6INTRAi + α7COLLARi + α8LNUMBERMi + α9LTVALUEi + α10LVALUEi + α11LMKTVi + α12BIG4i + α13TECHi + α14RETAILi + α15REGi + Year fixed effectsi (Step 1) + εi ABRETYP4i = α0 + α1PSUITi + α2ABCAi + α3BMi + α4RSIZEi + α5PRIVATEi + α6POOLi + α7SAME_INDi + α8CEOSHi + εi (Step 2) Variable

Full sample (N = 495)

Restricted sample (N = 442)

(1)

(2)

(3)

(1)

(2)

(3)

0.237 (1.03)

-0.140 (-0.77)

0.314 (1.39)

0.205 (0.80)

-0.170 (-0.86)

0.284 (1.12)

PSUIT

-1.354+++ (-2.61)

_

-1.612+++ (-3.27)

-1.393++ (-2.25)

_

-1.649+++ (-2.74)

ABCA

-2.957+ (-1.56)

-4.537+++ (-2.51)

_

-3.451++ (-1.68)

-4.590++ (-2.30)

_

BM

-0.115 (-0.51)

-0.017 (-0.07)

-0.153 (-0.68)

-0.002 (-0.01)

0.118 (0.48)

-0.038 (-0.15)

RSIZE

0.105 (0.83)

0.121 (0.95)

0.087 (0.69)

0.102 (0.77)

0.111 (0.83)

0.080 (0.61)

PRIVATE

0.030 (0.14)

0.003 (0.02)

0.015 (0.07)

0.038 (0.16)

0.039 (0.17)

0.008 (0.04)

POOL

-0.007 (-0.04)

0.141 (0.88)

-0.058 (-0.35)

-0.061 (-0.33)

0.092 (0.53)

-0.116 (-0.63)

SAME_IND

-0.281* (-1.79)

-0.292* (-1.85)

-0.266* (-1.70)

-0.241 (-1.40)

-0.248 (-1.31)

-0.227 (-1.32)

CEOSH

-0.498 (-0.68)

-0.343 (-0.47)

-0.614 (-0.85)

-0.437 (-0.55)

-0.317 (-0.40)

-0.593 (-0.75)

Adj. R2

0.019

0.007

0.016

0.011

0.002

0.007

Intercept

Notes: ABRETYP4 is the long-run abnormal returns measured over the four years after the merger announcement using the match-firm approach suggested by Barber and Lyon (1997); PSUIT is the probability of a lawsuit estimated as the predicted value of SUIT from Step 1 -- the variables and results for the first-step regression are presented in Table 5;

45

ABCA is the acquirer’s abnormal current accruals for the quarterly earnings announcement that immediately precedes the merger announcement; BM is the acquirer’s book-to-market ratio at the end of the quarter prior to the merger announcement; RSIZE is the ratio of the total assets of the target to the total assets of its acquirer; PRIVATE is a binary variable taking the value one is the target is a privately held company and zero if it is a publicly traded company; POOL is a binary variable taking the value one if the merger is accounted for by the pooling-ofinterest method and zero if it is accounted for by the purchase method; SAME_IND is a binary variable taking the value one if the two merging partners are in the same two-digit SIC code and zero otherwise; and CEOSH is the total number of the acquirer’s common shares controlled by the chief executive officer (CEO) of the acquirer divided by the number of common shares outstanding. The variables and results for the first-step regression are presented in Table 6. Some firms have multiple stock-for-stock acquisitions over relatively short periods and, consequently, the abnormal returns after these acquisitions overlap. To mitigate the effects of multiple acquisitions on our inferences, if a sample firm has multiple stock-for-stock mergers within a twelve-month period, we delete the earlier acquisitions and keep only the most recent one in the sample. The results are reported under the label “Restricted sample.” We find no evidence of influential observations using Cook’s (1977) distance statistic. +++, ++, and + (***, **, and *) indicate significance at the 1, 5, and 10 percent levels in a one- (two-) tail test, respectively.

46