Shareholder Litigation Rights and Acquisition Decisions: Evidence from a Natural Experiment

Transcription

1 Shareholder Litigation Rights and Acquisition Decisions: Evidence from a Natural Experiment Iftekhar Hasan Fordham University and Bank of Finland ihasan@fordham.edu Marco Navone University of Technology Sydney marco.navone@uts.edu.au Thomas To University of Sydney thomas.to@sydney.edu.au Eliza Wu University of Sydney eliza.wu@sydney.edu.au 1

2 Shareholder Litigation Rights and Acquisition Decisions: Evidence from a Natural Experiment ABSTRACT We examine the effect of shareholder litigation rights on acquisition decisions using an experimental design that exploits the reduction in litigation threat generated by a U.S. Ninth Circuit Court of Appeals ruling. We find that firms in the Ninth Circuit states acquire larger targets and that acquirer returns are lower after the ruling, especially for acquirers with weaker corporate governance. Further analysis shows that the value destruction is due to managers freedom to conduct empire-building acquisitions with overvalued equity. Overall, our findings imply that the threat of shareholder litigation significantly reduces managers incentives to make self-serving acquisitions. JEL classification: G34, K22, M41 Keywords: Shareholder litigation, corporate governance, corporate acquisitions, earnings management 2

3 1. Introduction The finance literature recognizes that agency problems result in managers not always make acquisitions that maximize shareholder value. Jensen s (1986) free cash flow hypothesis argues that managers have an incentive to grow their firms beyond an optimal size in order to increase the resources under their control. Lang et al. (1991) provide empirical evidence supporting this hypothesis. Furthermore, Masulis et al. (2007) show that managers of firms with poor corporate governance are most likely to be those engaging in empire-building acquisitions that destroy shareholder value. Given that acquisitions are one of the largest forms of corporate investment, it is important to establish ways of reducing managers incentives to pursue self-serving acquisitions at the cost of shareholder value. In this study, we address this issue by examining the effect of the threat of shareholder litigation on acquisition decisions. In the United States, securities class action litigation provides a mechanism through which shareholders can sue management for managerial misconduct. Such lawsuits are designed to provide recourse in the event that other governance mechanisms are not effective: for example, in firms with poor corporate governance. 1 However, there is an ongoing debate on whether shareholder litigation rights actually serve as an effective corporate governance tool in practice. On the one hand, several prior studies show empirical evidence supporting the active role of shareholder litigation as an effective governance tool. For example, Ferris et al. (2007) show that lawsuits are more likely to be 1 Class action litigation is particularly important in the United States because ownership is widely dispersed and, hence, often does not provide a sufficient incentive for shareholders to constantly monitor the management of a firm. For example, based on a sample of U.S. firms, Crane and Koch (2018) find that 95% of ownership positions are less than 2%, and that the top 20% of owners have individual positions that are, on average, only 2% of a firm. 3

4 filed against firms with agency conflicts, and that significant improvements in corporate governance occur after a lawsuit is filed. Similarly, Houston et al. (2018) document that the threat of shareholder litigation encourages voluntary disclosure practices. Then, Appel (2018) finds that the threat of shareholder litigation incentivizes managers to implement shareholder-friendly governance provisions because managers prefer to preempt lawsuits in order to avoid subsequent reputational penalties. 2 Following this line of thought, our managerial discipline hypothesis predicts that the threat of shareholder litigation improves managers acquisition performance by reducing their incentives to conduct selfserving acquisitions at the expense of shareholder value. Here, managers are aware that shareholders have the right to file lawsuits once they find out that the managers have produced misleading information about an acquisition. This can cause the company s stock price to fall significantly after an announcement or when financial and operational problems surface during the post-acquisition integration phase and predicted synergies do not materialize. For example, a class action was filed against BT Office Products International, Inc. in 1996, stating that the company s 10Q, management discussion and analysis, and certain press releases were false and misleading with respect to the company s acquisition strategy, which consequently resulted in material losses for shareholders. Ultimately, BT Office Products agreed to pay USD 1.48 million to settle the case. However, other studies have found that shareholder litigation is frivolous. That is, a significant number of lawsuits are filed by plaintiffs who expect to lose at trial, but who file the suits anyway, expecting the defendant to settle the case to avoid the considerable costs of litigation. Such studies argue that shareholder litigation is simply a method by which 2 Niehaus and Roth (1999), Fich and Shivdasani (2007), and Brochet and Srinivasan (2014) provide empirical evidence that lawsuits impose heavy reputational penalties on managers and directors. However, Agrawal et al. (1999) and Helland (2006) find contrasting results. 4

5 shareholders and lawyers extract wealth from the defendant corporation, because many lawsuits are triggered by unexpected declines in stock prices (Romano, 1991; Rhode, 2004; Helland, 2006). Jensen (1993) suggests that the legal incentive to avoid lawsuits encourages managers to minimize downside risk rather than to maximize shareholder value. Concurring with this view, Lin et al. (2018) find that managers of firms that face a higher threat of shareholder litigation invest less in R&D and generate fewer patents. In this case, managers are less innovative because they prefer to avoid the potential lawsuits resulting from project failure and, thus, a decline in the stock price (Holmstrom, 1989). Based on these findings, our wealth extraction hypothesis predicts that managers have legitimate incentives to conduct acquisitions that minimize downside risk rather than maximize shareholder value. As a result, the threat of shareholder litigation is expected to be detrimental to acquisition performance. Investigating the effect of ex ante shareholder litigation risk on acquisition performance is challenging because litigation is often correlated with unobservable factors. In this study, we overcome this potential omitted variable bias by exploiting the U.S. Ninth Circuit Court of Appeals ruling of July 2, 1999, In re: Silicon Graphics Inc. Securities Litigation, which provided exogenous variations in the threat of shareholder litigation for a subset of firms. The reason we choose the Ninth Circuit ruling is as follows. Despite the Private Securities Litigation Reform Act (PSLRA) of 1995 passed by the U.S. Congress to make it more difficult for shareholders to file frivolous lawsuits, the incidence of securities class actions has not decreased (Choi et al., 2009). The Ninth Circuit Court s 1999 ruling mandated a stricter interpretation of the PSLRA pleading standards, and became a turning point in curtailing the filing of frivolous lawsuits. Under common law, laws are created by courts using the doctrine of precedent, requiring that judges make decisions by considering 5

6 past cases that have set a precedent for future cases. The Ninth Circuit ruling effectively discourages shareholder plaintiffs from proceeding with a legal action unless there is clear evidence of intentional managerial misbehavior. Empirical evidence shows that the ruling resulted in a 43% drop in the number of class action lawsuits in the Ninth Circuit. In comparison, other circuits have seen a 14% increase in such cases (Crane and Koch, 2018). We compare the acquisition performance of U.S. firms headquartered in the Ninth Circuit states with that of firms headquartered in other states around the time of the ruling. This strategy effectively isolates the effect of litigation risk on acquisition performance from unobservable covariates. Using corporate acquisitions announced between 1996 and 2003 in the U.S. market, we examine the performance of such transactions for the four-year periods before and after the 1999 Ninth Circuit Court ruling. We find that, after the ruling, the five-day cumulative abnormal returns (CARs) around the acquisition announcement dates are, on average, 0.96 percentage points lower for acquiring firms headquartered in the Ninth Circuit states, vis-à-vis firms in other states. We interpret these results to mean that the reduction in the threat of litigation after the ruling increases the probability of managers undertaking value-destroying, self-serving acquisitions. We address the possibility of a selection bias by using propensity score matching, because a firm may choose the location of its headquarters based on firm-specific needs. With a matched sample of firms that exhibit similar characteristics, except for their corporate headquarters being located in or outside the Ninth Circuit states, we continue to find that managers of the treatment firms (i.e., firms headquartered in the Ninth Circuit states) make worse acquisitions than do managers of the matched control firms after the ruling. Furthermore, we show that this result is not driven by acquisitions made by high-tech firms during the tech bubble around the time of the ruling. Furthermore, we show that the 6

7 effect is more pronounced in deals made by acquirers with poor corporate governance. This makes intuitive sense, because the threat of shareholder litigation was likely playing an important governance role in these firms prior to the ruling. Next, we investigate the mechanisms through which managers destroy shareholder value. First, consistent with the empire-building motives of managers, we find that managers in the Ninth Circuit states began acquiring larger firms when it became more difficult for shareholders to litigate. Second, after the ruling, managers in the Ninth Circuit states heavily inflate earnings in the quarter before making an acquisition. Moreover, this relationship holds only in the case of acquisitions funded by stock, suggesting that managers have had an incentive to pursue empire-building acquisitions using overvalued equity since the ruling, fueling a spate of value-destroying, self-serving acquisitions. As a final test, we investigate whether the Ninth Circuit s ruling affects the likelihood of CEO replacement, especially when CEOs are engaged in value-decreasing acquisitions. Lehn and Zhao (2006) document that poor decisions related to mergers and acquisitions (M&As) tend to increase the likelihood of a CEO being fired, in addition to the substantial effect of such decisions on shareholder value. We find a reduction in the probability of CEO replacement (and forced CEO replacement) in the Ninth Circuit states after the ruling compared with that of firms in other states. Moreover, the likelihood of CEO replacement is not significantly affected by their engagement in value-decreasing acquisitions. This finding suggests that the increase in managerial entrenchment since the ruling is another reason why managers in Ninth Circuit firms are able to conduct value-destroying, empire-building acquisitions. Our study contributes to the extant literature on corporate governance in the following two ways. First, we find a relation between the exogenous change in shareholder 7

8 litigation rights and agency problems in acquiring firms. Many studies have examined the role of internal/external corporate governance in the context of M&As. For instance, Byrd and Hickman (1992) examine tender offers, and find that independent boards are associated with higher bidder returns. Masulis et al. (2007) show that acquiring firms with higher anti-takeover provisions exhibit announcement returns that are more negative. Krishnan et al. (2012) report that shareholder litigation reduces the likelihood of completing a merger or acquisition. Whereas the extant literature uses conventional governance mechanisms (e.g., anti-takeover provisions, board characteristics, or shareholder litigation) to depict their role in M&As, few studies exploit regulatory reforms that could affect the quality of corporate governance related to M&As. Second, our study extends the growing body of literature on shareholder litigation that documents the role of retail shareholders in corporations. Crane and Koch (2018) find that corporate ownership structures have shifted from retail shareholders to institutional shareholders since the Ninth Circuit s ruling. Hopkins (2017) documents an increase in misreporting by firms since the threat of shareholder litigation decreased. Lin et al. (2018) show that shareholder litigation rights discourage managers from engaging in innovation, because it necessarily includes the possibility of project failure and stock price reductions, which gives shareholders an opportunity to file lawsuits related to a breach of fiduciary duty. We find evidence indicating that corporate managers do consider the threat of shareholder litigation when making decisions related to M&As. The rest of the paper is organized as follows. Section 2 introduces the legal background of M&A lawsuits. Section 3 describes the data and explains the construction of the variables used in this study. Section 4 examines the impact of the threat of shareholder litigation on acquisition decisions. Section 5 identifies the channel through which value 8

9 destruction takes place and describes several robustness tests. Lastly, Section 6 concludes the paper. 2. Institutional Background In this section, we first describe the different kinds of M&A lawsuits that are filed by shareholders. Then, we discuss the history of securities litigation reform and the specific ruling by the Ninth Circuit Court of Appeals that generates the exogenous variation in the threat of shareholder litigation. 2.1 SHAREHOLDER LITIGATION IN M&As Managers and directors have a fiduciary duty to a firm s shareholders, which means that they are legally obliged to act in the best interests of the shareholders. However, in reality, the separation of ownership and control provides managers with an incentive to maximize their own interests at the expense of shareholders interests (Jensen and Meckling, 1976). The M&A literature documents that managers often conduct empire-building acquisitions at the expense of shareholder value (Jensen, 1986; Masulis et al., 2007; Harford et al., 2012), and one way in which shareholders can respond to such wrongdoing is to file a lawsuit. In the United States, lawsuits related to M&As can be filed as class actions, derivative lawsuits, and other less common forms. Class actions and derivative lawsuits are both forms of shareholder-representative litigation. Here, the plaintiff's law firm will pursue the matter at the request of a particular named shareholder and on behalf of all shareholders that are adversely affected by the company's actions. Empirical evidence shows that most M&A lawsuits are filed in the form of class actions. For example, using a sample of M&A lawsuits, Krishnan et al. (2012) find that 87.6% are class action suits, while 9

10 only 3.4% are derivative suits. 3 Similarly, Crutchley et al. (2015) report that M&As and earnings management are the most commonly cited reasons for federal class action suits. 2.2 BACKGROUND OF THE SECURITIES LITIGATION REFORM AND THE NINTH CIRCUIT COURT RULING Prior to 1995, all types of class action litigation were governed by the same rule (Federal rules of Civil Procedure). Under this rule, shareholders were able to bring lawsuits against firms with minimal evidence of fraud. The result was a significant number of lawsuits with dubious merit being filed on a regular basis (Johnson et al., 2001). These abuses imposed excessive burdens on firms and led to the promulgation of the Private Securities Litigation Reform Act (PSLRA). The PSLRA, enacted in 1995, introduced hurdles to curb potentially frivolous lawsuits. While the PSLRA has contributed to a less litigious environment for all firms, the pleading standards of the law, as a practical matter, were interpreted differently by various U.S. circuit courts, of which the interpretation by the Ninth Circuit Court in the Silicon Graphics case on July 2, 1999 was the most stringent. 4 The Ninth Circuit s ruling requires that, prior to forming a class, plaintiffs must establish that the defendant acted with deliberate recklessness in making the misrepresentation giving rise to the claim. In contrast, in other circuits, proving mere recklessness is sufficient. This remarkable 3 Shareholders prefer to bring class actions to challenge M&A transactions because derivative lawsuits involve a number of procedural hurdles. More importantly, financial recoveries from a derivative lawsuit go to the corporate treasury rather than to the shareholders. The limited use of derivative lawsuits makes the exogenous variation in shareholder litigation rights generated by universal demand (UD) laws less useful in our research setting, because UD laws only impose significant obstacles to initiating derivative suits. 4 In re: Silicon Graphics Inc. Securities Litigation, 183 F.3d 970 (9th Cir. 1999). The case was filed in the Northern District Court of California on behalf of the purchasers of Silicon Graphics, Inc. stock, who claimed that the company made misleading statements about its performance and that corporate insiders profited from the inflated stock price. The Ninth Circuit Court of Appeals dismissed the case on July 2, 1999, concluding that general allegations that management received internal reports at odds with public statements are insufficient unless plaintiffs also plead the details of those reports, such as authors, recipients and particular contents contradicting the public statements. 10

11 decision was largely unanticipated (Johnson et al., 2000) and has since been applied to all securities class actions filed with the Ninth Circuit Court. Crane and Koch (2018) document that, after the ruling, the number of class actions in the Ninth Circuit dropped by 43%. In comparison, the number of class actions in other circuits increased by 14%. Securities class action litigation can be brought in any of the federal circuit courts, because shareholders are often geographically dispersed. However, Cox et al. (2009) show that plaintiffs rarely switch the filing venue in order to avoid the heightened pleading standard. They find that 85% of the class action cases are filed in the circuit of the defendant corporation s headquarters, and that those filed in out-of-home circuits are more likely to be low-value cases. There are two reasons for this. First, out-of-home circuit filings are usually consolidated into the home circuit of the defendant, either by the federal legal process or by the defendant s motion to relocate the lawsuit, because it is more efficient to litigate the case where the defendant s officers, main witnesses, and documents are located. Second, plaintiffs rarely seek to file in another circuit for fear of significant delays and costs involved in the litigation consolidation process. Hence, the Ninth Circuit ruling essentially reduces the shareholder litigation rights for firms headquartered in states under the Ninth Circuit s jurisdiction (Alaska, Washington, Oregon, Idaho, Montana, California, Nevada, Arizona, and Hawaii). Furthermore, Cox et al. (2009) show that, despite numerous efforts by the Supreme Court, differences in the pleading standards across circuits persist, which suggests that the effect of the ruling is likely to be long term. Overall, the ruling provides an ideal quasi-natural experiment to evaluate whether and how shareholder litigation rights shape acquisition decisions and outcomes. 11

12 3. Data Sources and Construction of Variables 3.1 DATA SOURCES We obtain our sample of corporate acquisitions for the period between 1996 and 2003 from the Securities Data Company (SDC) M&A database in order to examine the post Reform period, before and after the Ninth Circuit Court of Appeals ruling. We require a minimum deal value of USD 1 million. Furthermore, we include only those deals where the acquiring firm controls less than 50% of the target's stock before the announcement, but owns 100% of the target's stock after the transaction. We obtain accounting data from Compustat, financial market data from CRSP, and governance data from ExecuComp and Riskmetrics. In order to determine the location of a firm s headquarters in a relevant period, we use the business address disclosed in the annual 10-K filings, provided by Bill McDonald, and supplement this with Compustat records when these data are missing. 5 Our final sample consists of 3,326 acquisitions made by 1,020 acquirers. 3.2 FIRM AND DEAL CHARACTERISTICS Following the acquisition literature, we control for a vector of firm and deal characteristics that may affect a firm's acquisition decisions. The variable definitions are provided in the Appendix. Our firm-level controls include ln(assets), leverage, Tobin s Q, free cash flow, sales growth, and board governance quality. All firm characteristics are measured at the fiscal year-end prior to the acquisition announcement. Our deal-level controls include the stock price run-up, the relative deal size, a high-tech deal indicator, a cross-border indicator, a cross-industry indicator, the method of payment, and the target s public status. Following Jenter and Lewellen (2015), we combine multiple governance 5 The data are publicly available from Bill McDonald s personal webpage, 12

13 measures into a broader index of governance quality. The measures we use include the CEO Chair duality, board independence, and G-index. To construct the index, we split each of the governance measures into two groups, with higher values indicating better governance, and cumulate the ranks (0 1). We then divide the cumulated ranks by the number of measures available for the firm-year to obtain the governance index score. 6 Table 1 provides the summary statistics of the variables used in this study. We winsorize all continuous variables at the 1st and 99th percentiles. We find that 24% of the acquisitions are made by firms headquartered in the Ninth Circuit states, which allows for a significant number of treated firms in our difference-in-differences (DID) setting. In terms of acquirer characteristics, an average acquirer in our sample has a leverage ratio of 22%, Tobin s Q of 2.75, and free cash flow of 4%, which are consistent with the figures reported in other recent M&A studies (e.g., Yim, 2013; Huang et al. 2014). The figures for the deal characteristics show that 20% of the acquisitions are cross-border deals, 44% are crossindustry deals, and 32% are funded entirely by cash. 4. Shareholder Litigation Rights and Announcement Returns In this section, we analyze the effect of shareholder litigation rights on deal announcement returns using both our full sample and a matched sample. Then, we conduct subsample tests to identify the source of value destruction. 4.1 BASELINE RESULTS To examine how the threat of shareholder litigation affects acquisition performance, we estimate the following DID model: 6 Combining different measures into a single governance index helps us to control for multiple aspects of firm governance, without sacrificing the sample size. We divide the governance measure at the median rather than into terciles, as in Jenter and Lewellen (2015), because CEO Chair is a binary variable. 13

14 CAR i,t+1 = α + β (1) Treat i Post t + β (2) Treat i + γz i,t + Year t + Industry j + State k + ε i,t,(1) where t denotes a year, i denotes a firm, j denotes an industry, and k denotes an incorporation state. The dependent variable, CAR i,t+1, denotes the five-day CAR centered on the acquisition announcement date. Following Masulis et al. (2007), abnormal stock returns are calculated by estimating the market model for each acquirer over a 200-day period ending 11 days before the announcement date ( 210, 11), using the CRSP valueweighted return as the benchmark market index. The indicator variable Treat differentiates and controls for differences between the treatment and control groups, taking the value one when the firm is located in a Ninth Circuit state, and zero otherwise. Post is a time dummy that takes the value one for fiscal years after 1999, and zero for years from 1996 to We do not include observations from years prior to 1996 because the PSLRA in 1995 significantly affected the litigation environment governing securities class actions. Then, β (1) is the DID coefficient of interest used to identify the difference in the treatment effect resulting from the Ninth Circuit ruling. We also include a set of control measures, Z i,t, identified in prior studies as the set of firm and deal characteristics that are likely to affect a firm s acquisition decisions (all variables used are provided in the Appendix). In addition, Year t and Industry j capture the time and industry fixed effects, respectively. Following the recent literature on the effects of state laws, we include firms incorporation state fixed effects, State k, and cluster the standard errors at the headquarters state level in all regressions to account for differences in incorporation and state-level laws and regulations related to a firm s headquarters (Gormley and Matsa, 2017; Houston et al., 2018). We report the DID estimation results in Table 2. We first report the results without the control variables in Column 1. We find that the coefficient estimate of Treat Post is negative and significant at the 5% level. We then add firm-level controls and report the 14

15 results in Column 2. Again, the coefficient estimate of Treat Post is negative and significant at the 5% level. Lastly, we add deal characteristics as additional controls to the model. Specifically, we add an indicator variable to control for whether an acquisition is in the hightech industry, because the Ninth Circuit ruling was implemented around the time of the tech bubble in 2000, to which firms in the Ninth Circuit states (especially California) were significantly exposed. 7 We report the results in Column 3. Here, we find that the coefficient estimate of Treat Post is negative and significant at the 5% level after adding deal-level controls. In terms of economic magnitude, the coefficient estimate of Treat Post indicates that, after the ruling, acquirers in the Ninth Circuit states experienced a reduction of 0.96 percentage points in announcement returns compared with those of acquirers located in the other circuits. This negative market reaction represents a loss of USD 23.9 million in shareholder value for the median Ninth Circuit firm in our sample. Overall, our results support our managerial discipline hypothesis. 4.2 PROPENSITY-SCORE MATCHED SAMPLE Next, we use a matched sample to address the possibility that the location of a firm s headquarters might be chosen based on specific firm characteristics. We employ the propensity score matching approach to find the closest non-treated firm for each of our treatment firms located in the Ninth Circuit states. Specifically, we randomly select a firm located outside of the Ninth Circuit states for each treatment firm (with no replacement), matched on the firm s average logarithm of total assets (size), average leverage, and average Tobin s Q ratio using data for the period 1996 to Moreover, we require that the treatment firms and control firms operate within the same 48 Fama French industries. 7 We further mitigate this problem by using a propensity-score matched sample that requires an exact match on the 48 Fama French industries in Section 4.2. In addition, we conduct DID estimations after excluding all high-tech firms in the matched sample. 15

16 The final sample contains 102 treatment firms and an equal number of matched control firms. Panel A of Table 3 shows that the treatment and control firms are well matched. The test statistics show there are no significant differences between the treatment and control firms in the three matching variables prior to the ruling. Panel A of Fig. 1 shows a similar trend in the level of CAR between the treatment and control firms prior to the ruling, suggesting that the parallel-trend assumption of the DID estimation is valid. The figure shows what appears to be a delayed effect of the Ninth Circuit ruling due to the tech bubble that occurred at around the same time. We control for this systematic event in all regression models using an indicator variable showing whether the acquisition took place within the high-tech industry. Using the matched sample, we first conduct a baseline DID test to reaffirm that a reduction in the threat of shareholder litigation affects acquisition performance. Panel B of Table 3 shows the results. Consistent with our earlier finding, the results in Columns 1 and 2 show that, after the ruling, acquisitions conducted by firms located in the Ninth Circuit states generate significantly lower abnormal announcement returns compared with those of acquisitions conducted by firms located in other states. We find that the economic significance of the difference is stronger than that reported in the baseline regressions (2.19 percentage points vs percentage points) after removing systemic differences between the treatment and control groups. In addition, we conduct placebo tests using an extended sample period from 1996 to With the exception of 1999 and 2000, we find no evidence of a similar effect if we use any other (placebo) year between 1996 and 2015 as our treatment year (unreported, but available upon request). Using a matched sample removes systemic differences between the treatment and control groups. However, a potential concern that remains is that the acquisition 16

17 performance may not be changing in response to the ruling, but rather to some other event that happened around the same time as the ruling that affects our treatment firms more than it does our control firms. As mentioned previously, one such event could be the tech bubble. Specifically, firms on the West Coast (especially California) were particularly exposed to this event. Hence, it is possible that the deterioration in acquisition performance around this time for our treatment firms (Panel A of Fig. 1) was caused by the tech bubble rather than by the Ninth Circuit ruling. To rule out this concern, we remove all high-tech acquirers from the sample. Panel B of Fig. 1 plots the average CARs for the treatment and matched control firms not in the high-tech industry. The result clearly shows a significant deterioration for the treatment firms around 1999 to 2000, compared with the control firms, as a result of the Ninth Circuit ruling. In addition, we conduct DID estimations after excluding acquirers in the high-tech industry, and find that our previous finding holds, despite the significant drop in the sample size, as shown in Columns 3 and 4 of Table SUBSAMPLE TESTS Next, we conduct cross-sectional tests to identify the types of firms that completed value-destroying deals after the easing of the threat of litigation. Shareholder litigation is designed to provide recourse in the event that all other governance mechanisms fail. This governance mechanism is unlikely to be needed in firms with good corporate governance, because constant monitoring enables the board of directors and institutional blockholders to identify and discipline managers seeking self-serving acquisitions. In contrast, we expect the threat of shareholder litigation to play an important governance role in firms with poor corporate governance. Without an adequate level of monitoring, managers in the latter firms have an incentive to conduct self-serving acquisitions, and one of the only ways to prevent them from doing so is the threat of shareholder litigation. Therefore, we expect 17

18 that the value-destroying acquisitions conducted by firms in the Ninth Circuit after the ruling are more likely to be conducted by firms with weaker corporate governance standards. To test this hypothesis, we conduct DID estimations conditional on standard proxies for firms corporate governance quality. Following the literature, we measure the quality of corporate governance using the G-index and the degree of institutional monitoring. The first proxy we use to measure the quality of corporate governance is the G-index, first introduced by Gompers et al. (2003). The G-index is based on 24 anti-takeover provisions, where a higher value corresponds to worse corporate governance. Masulis et al. (2007) find that acquirers with more anti-takeover provisions are more likely to conduct empire-building acquisitions that destroy shareholder value. Hence, we follow Masulis et al. (2007) and use the G-index as a measure of corporate governance. We classify firms as having a higher (lower) quality of corporate governance if the firm has a G-index score of 9 or less (above 9). Columns 1 and 2 of Table 4 present the results. We find that the value destruction stems from firms with a higher G-index score, because the managers of firms located within the Ninth Circuit states are likely to have greater freedom in conducting selfserving acquisitions after the ruling. The second proxy we use to measure the quality of corporate governance is institutional monitoring. Recent evidence shows that institutional investors play a significant role in corporate governance. Aggarwal et al. (2015) show that institutional investors value their votes and use the proxy process to affect corporate governance. McCahery et al. (2016) find that 45% of the surveyed institutional investors have had private discussions with the corporate board outside of the management s presence. Furthermore, Liu et al. (2018) find that institutional investors monitoring strengthens board oversight. We measure institutional monitoring as the number of institutional owners (blockholders) that hold more 18

19 than 5% of a firm s shares, because we expect that institutional investors who own a larger portion of a firm s shares will have a greater incentive and ability to discipline managers. Columns 3 and 4 of Table 4 show the results. Here, we find that the value destruction is concentrated in acquirers with a lower number of blockholders. 5. The Underlying Mechanism 5.1 EMPIRE-BUILDING WITH OVERVALUED EQUITY M&As and earnings management are the two most commonly cited reasons for federal class action suits (Crutchley et al., 2015), and studies have found that the two reasons often co-exist. For example, Louis (2004) shows that acquiring firms often overstate their earnings in the quarter before a stock swap announcement in an attempt to boost their stock price. Furthermore, Gong et al. (2008) find that acquirers who manipulate their earnings before stock offers are more likely to attract subsequent lawsuits. We expect that the reduction in the threat of shareholder litigation for managers of firms incorporated in the Ninth Circuit states will enable them to manipulate earnings before an acquisition to boost their stock prices, and then to use the overvalued stock to fund empire-building acquisitions. These managers are less afraid of losing control of their empire by paying with stock, because most of them are employed by firms with poor corporate governance (as shown in Table 4) and shareholders have very little power in poorly governed firms. Our view is shared by Jensen (2005), who suggests that the enormous amount of value destruction for acquirers can be explained by the agency costs of overvalued equity. This view is supported by the empirical findings of Moeller et al. (2005), who show that large-loss bidders have significantly higher market-to-book ratios and are more likely to finance their deals with equity. 19

20 To test our predictions, we first test whether managers in the Ninth Circuit states are more likely to conduct larger acquisitions and use equity financing after the Ninth Circuit ruling. We measure acquisition size using the relative deal size, and define an acquisition as being equity financed if 80% or more of an acquisition is funded by stock. Table 5 reports the results. Columns 1 and 2 show that managers in the Ninth Circuit states did begin acquiring larger firms after it became more difficult for shareholders to litigate. In Columns 3 and 4, the results from the logit models show that these managers are also more likely to use equity financing after the ruling. Second, we examine acquirers earnings management using the modified Jones model and quarterly financial data from Compustat to estimate their abnormal accruals in the quarter prior to a deal announcement (Jones, 1991; Dechow et al., 1995). Note that the sample size shrinks dramatically owing to missing quarterly values for the variables required to calculate the abnormal accruals. Table 6 presents the results. We find that, compared with managers of firms in other states, managers in Ninth Circuit states significantly manage earnings upwards in the quarter before acquisitions after the Ninth Circuit ruling, and that this relationship holds only in the case of stock offers. Overall, these findings suggest that managers had the freedom to conduct self-serving acquisitions using overvalued equity after it became more difficult for shareholders to litigate. In order to confirm that the stock-financed acquisitions are value-destroying, selfserving acquisitions, we examine the effect of the threat of shareholder litigation on acquisition performance for all-cash deals and for stock-financed deals separately. To rule out the possibility that stock-financed acquisitions may yield a lower CAR owing to the market s anticipation of overvalued equity being used in stock offers for Ninth Circuit firms, rather than because of their self-serving nature, we examine the change in the operating 20

21 performance of acquiring firms around the deal announcement. Specifically, we follow Harford and Schonlau (2013) and use the difference between the (Fama French) industryadjusted ROA in year t + 2 and year t 1 to capture the change in operating performance, because it may take longer than one year to fully reflect the transformation. Table 7 confirms that the value destruction stems primarily from stock-financed deals. Overall, our results indicate that the reduced threat of litigation has afforded managers the opportunity to conduct self-serving acquisitions using equity overvalued via inflated earnings and is responsible for the destruction of shareholder value. 5.2 PROBABILITY OF CEO TURNOVER Lehn and Zhao (2006) document that bad M&A decisions tend to increase the likelihood of a CEO being fired. However, our results show that the threat of being fired does not stop CEOs from conducting value-destroying, empire-building acquisitions. We conjecture that this is because a reduction in the threat of litigation increases managerial entrenchment, which decreases the probability of being fired. This is consistent with Appel s (2018) finding that a reduction in shareholder litigation rights allows managers to adopt classified boards, supermajority voting requirements, and poison pills. To investigate whether the Ninth Circuit s ruling affects the likelihood of CEO replacement, especially when CEOs are engaged in value-decreasing M&As, we merge our sample of acquirers with the Execucomp database to examine the probability of CEO turnover in these firms around the time of the ruling for the period 1996 to We obtain CEO turnover and forced CEO turnover data from Professor Andrea Eisfeldt s website. Forced CEO turnovers are identified based on news stories published in Factiva. 8 Table See Eisfeldt and Kuhnen (2013) for a detailed explanation of the information-collection process. 21

22 shows the logistic regression results. In Column 1, we find a reduction in the probability of CEO replacement for firms in the Ninth Circuit states compared with that of firms in other states after the ruling. In Column 2, we show that the likelihood of CEO replacement is not affected significantly by engaging in value-decreasing M&As (NegM&A). Furthermore, we examine forced CEO turnovers in Columns 3 and 4, and find similar results. Overall, these findings suggest that the lower threat of being fired after the ruling is another reason why managers in Ninth Circuit firms are able to conduct value-destroying, empire-building acquisitions ROBUSTNESS TESTS We conduct a set of additional robustness tests. The results are reported in Table 9. First, a number of papers argue that shareholder litigation affects disclosure and the information environment of firms. For example, Bourveau et al. (2018) find that firms increase their disclosure significantly after UD laws to make it more difficult for shareholders to claim derivative lawsuits. Houston et al. (2018) report that the threat of shareholder litigation encourages voluntary disclosure practices. Similarly, Hopkins (2017) concludes that the threat of shareholder litigation can discipline managers and deter financial misreporting. However, Boone et al. (2018) show that, although managers increase their level of voluntary disclosure owing to the reduced threat of shareholder litigation, overall, firms provide lower quality financial reporting after UD laws are passed. Another strand of literature argues that disclosure is related to M&A behavior. For example, Hope and Thomas (2008) conclude that disclosure requirements limit the ability of managers to engage in empire building. In this case, the aforementioned literature suggests that the reduction in the threat of shareholder litigation affects disclosure, which may, in turn, affect M&A behavior. To address this concern, we control for two variables that capture a firm s 22

23 information quality: Restate and Ln(Analysts). Restate is equal to one if a firm restates its financial statement in a given fiscal year, and zero otherwise. Ln(Analysts) is the logtransformed number of unique analysts (plus one) that issue a firm s earnings forecasts. Column 1 of Table 9 shows the results. We find that the coefficient of Treat Post remains negative and statistically significant. Furthermore, Restate is positively associated with acquirer announcement returns and is significant at the 5% level, 9 whereas Ln(Analyst) is negatively associated with acquirer announcement returns, but is non-significant. Second, we control for the staggered implementation of UD laws because these have been shown to affect acquisition performance (Chu and Zhao, 2018). Column 2 of Table 9 shows the results. We find that the estimated coefficient of Treat Post is of a similar magnitude to those in the main analysis and is statistically significant at the 5% level. Third, we control for other legal changes related to shareholder litigation and antitakeover provisions to ensure that these changes are not driving our results. Following Karpoff and Wittry (2018), we control for various state anti-takeover laws, including control share acquisition laws, business combination laws, fair-price laws, directors duties laws, and poison-pill laws. Column 3 of Table 9 shows the results. We again find that the estimated coefficient of Treat Post is of a similar magnitude and significance level to those in the main analysis after the addition of these control variables. Fourth, we control for whether target firms are affected by the Ninth Circuit ruling, because we expect that the change in managerial behavior within these firms may also be driving our results. Our sample size shrinks significantly because we remove cross-border deals for a better comparison, as well as acquisitions involving private targets because we 9 While this result indicates that restating firms tend to exhibit higher positive announcement returns, it cannot exclude the possibility that acquiring firms exhibiting better announcement returns are more likely to restate their financial statements than are those with worse announcement returns, for a given fiscal year. 23

24 cannot identify their state of operation. Column 4 of Table 9 reports the results. Once again, we find that the estimated coefficient of Treat Post is negative and significant at the 5% level. Fifth, we extend our sample to 2015 to further address the concern that our results may be driven by the burst of the tech bubble, which occurred within our relatively short post-event window. We expect our findings to be similar when using the longer post-event window, because Cox et al. (2009) suggest that, despite numerous efforts by the Supreme Court, differences in pleading standards across circuits persist. Column 5 of Table 9 reports the results. We find that the estimated coefficient of Treat Post is negative and significant at the 5% level. Sixth, we limit our analysis to firms located in states on the border of the Ninth Circuit (i.e., the treated firms are those with headquarters in Montana, Idaho, Nevada, and Arizona, and the control firms are those with headquarters in Utah, New Mexico, Wyoming, and North Dakota) to address the possible selection bias resulting from firms not selecting their locations randomly. Column 6 of Table 9 reports the results. Because firms headquartered in these states completed only 68 M&A transactions during the period , we include only industry fixed effects in our regression in order to avoid multicollinearity. However, the coefficient of Treat Post remains negative and significant. Overall, our findings are robust to a different dimension of alternative tests. 6. Conclusion It is well recognized in the M&A literature that, in the absence of strong corporate governance, managers have an incentive to conduct empire-building acquisitions that destroy shareholder value. Shareholder litigation rights, a governance mechanism designed 24

25 to provide recourse in the event that all other governance mechanisms fail, should theoretically play an important external governance role in disciplining managers in poorly governed firms, thus increasing shareholder value. However, in practice, we often find that shareholder litigation is simply a method by which shareholders and lawyers extract wealth from the defendant corporation, because many lawsuits are triggered by unexpected decreases in stock prices. This encourages managers to conduct acquisitions that minimize downside risk rather than maximize shareholder value. As a result, the threat of shareholder litigation may also be detrimental to shareholder value. In this study, we examine the effect of the threat of shareholder litigation on acquisition decisions to determine whether shareholder litigation rights serve as a governance mechanism in preventing managers from conducting self-serving acquisitions at the expense of shareholder value. In order to establish a causal relation, we utilize the Ninth Circuit Court of Appeals ruling of July 2, 1999, In re: Silicon Graphics Inc. Securities Litigation, which generated an exogenous reduction in shareholder litigation rights for firms located in the Ninth Circuit states. Based on a DID methodology, we find that, after the ruling, firms located in the Ninth Circuit states experienced significantly lower deal announcement returns, especially for acquirers with weaker corporate governance. Furthermore, we find that the freedom of managers to conduct empire-building acquisitions using equity overvalued via inflated earnings is responsible for the value destruction. Overall, we show that the threat of shareholder litigation serves as an important governance tool in M&As because it significantly reduces managers incentives to engage in empire-building at the expense of shareholder value. 25

26 References Agrawal, A., Jaffe, J., Karpoff, J., Management turnover and governance changes following the revelation of fraud, J. Law Econ. 42, Aggarwal, R., Saffi, P. A., Sturgess, J., The role of institutional investors in voting: Evidence from the securities lending market, J. Financ. 70, Appel, I., Governance by Litigation (Working paper). Bourveau, T., Lou, Y, Wang, R., Shareholder litigation and corporate disclosure: Evidence from derivative lawsuits, J. Accounting Res. 56, Brochet, F., Srinivasan, S., Accountability of independent directors: Evidence from firms subject to securities litigation J. Financ. Econ. 111, Byrd, J., Hickman, K., Do outside directors monitor managers? Evidence from tender offer bids, J. Financ. Econ. 32, Choi, S. J., Nelson, K. K., Pritchard, A. C., The screening effect of the private securities litigation reform act, J. Empir. Legal Stud. 6, Chu, Y., Zhao, Y., The Dark Side of Shareholder Litigation: Evidence from Corporate Takeovers (Working paper). Cox, J. D., Thomas, R. S., Bai, L., Do differences in pleading standards cause forum shopping in securities class actions: Doctrinal and empirical analyses, Wisc. Law Rev Crane, A., Koch, A., Shareholder litigation and ownership structure: Evidence from a natural experiment, Manage. Sci. 64, Crutchley, C. E., Minnick, K., Schorno, P. J., When governance fails: Naming directors in class action lawsuits, J. Corp. Financ. 35, Dechow, P., Sloan, R., Sweeney, A., Detecting earnings management, Account. Rev. 70, Eisfeldt, A. L., Kuhnen, C., CEO turnover in a competitive assignment framework, J. Financ. Econ. 109, Ferris, S. P., Jandik, T., Lawless, R. M., Makhija, A., Derivative lawsuits as a corporate governance mechanism: Empirical evidence on board changes surrounding filings, J. Financ. Quant. Anal. 42, Fich, E., Shivdasani, A., Financial fraud, director reputation, and shareholder wealth, J. Financ. Econ. 86, Gompers, P., Ishii, J., Metrick, A., Corporate governance and equity prices, Q. J. Econ. 118, Gong, G., Louis, H., Sun, A., Earnings management, lawsuits, and stock-for-stock acquirers market performance, J. Account. Econ. 46, Gormley, T., Matsa, D., Playing it safe? Managerial preferences, risk, and agency conflicts, J. Financ. Econ. 122, Harford, J., Humphery-Jenner, M., Powell, R., The sources of value destruction in acquisitions by entrenched managers, J. Financ. Econ. 106, Harford, J., Schonlau, R., Does the director labor market offer ex post settling-up for CEOs? The case of acquisitions, J. Financ. Econ. 110, Helland, E., Reputational penalties and the merits of class action securities litigation, J. Law Econ. 49, Holmstrom, B., Agency costs and innovation, J. Econ. Behav. Organ. 12,