The Dark Side of Shareholder Litigation: Evidence from Corporate Takeovers

Transcription

1 The Dark Side of Shareholder Litigation: Evidence from Corporate Takeovers Yongqiang Chu and Yijia Zhao * January, 2016 Abstract Exploiting staggered adoption of universal demand (UD) laws by 23 states between 1989 and 2005 as quasi-natural experiments, we show that reduced shareholder litigation threat improves corporate takeover efficiency. Using a difference-in-differences approach, we find that acquirers incorporated in states that adopt UD laws experience significantly higher abnormal announcement returns and better long-run post-merger operating performance. The results suggest that the threat of litigation destroys value ex ante as acquirers make suboptimal deals to reduce the risk of being sued. Keywords: Shareholder Litigation, Mergers and Acquisitions, Acquirer Returns, Value Destruction, UD Laws JEL Codes: G30, G34, G38 * Chu is from Moore School of Business at the University of South Carolina, 1014 Greene Street, Columbia, SC Chu can be reached at yongqiang.chu@moore.sc.edu. Zhao is from Wheaton College, 26 E. Main Street, Norton, MA Zhao can be reached at zhao_yijia@wheatoncollege.edu. We thank Brain Broughman (CELS discussant), Qianqian Huang (CICF discussant), Feng Jiang, Donghang Zhang, and seminar participants at the 2015 China International Conference in Finance (CICF), the 10 th Annual Conference on Empirical Legal Studies (CELS 2015) at Washington University in St. Louis, and University of Massachusetts at Boston for helpful comments. We are responsible for all errors. 1

2 1. Introduction Shareholders right to bring up litigations against corporate management is a key institutional feature of the legal system that governs business in the U.S.. Meanwhile, a long-lasting debate persists among both academia and law makers over the merit and effect of these shareholder litigations. On the one hand, some consider shareholder litigations as an essential element of external corporate governance mechanism (e.g. Bhagat and Romano 2002). They argue that the threat of lawsuits imposes ex ante disciplining forces on managers by providing shareholders with the opportunity to bring up issues against managers when shareholders feel that managers inappropriately exploit their positions of control. On the other hand, however, some cast doubts on the effectiveness of shareholder litigations and argue that in practice these litigations are subject to significant defects. The litigation agency problem posits that self-interested attorneys, acting as the agent of shareholder plaintiffs, have the incentive to serve their own interests instead of that of the shareholder plaintiffs (e.g. Coffee and Schwartz 1981; Macey and Miller 1991, Romano 1991). This conflict of interests leads to the concern that a large number of lawsuits may have been brought up based on frivolous charges but with attorneys fees as the true underlying motivation (noted as strike suits in Romano 1991 and Swanson 1992). 1 While serious questions have been raised about the possible abuse of lawsuits by lawyers and 1 It is important to note that a shareholder litigation is not necessarily initiated by shareholders. Indeed, in filing a derivative suit (which is the focal type of shareholder litigation in our paper), the plaintiff's attorney is often the principal mover. The attorney locates a possible derivative claim and then finds an eligible shareholder to serve as plaintiff. As a result, the attorney may have a much more direct and substantial financial interest in the case and its outcome than the plaintiff shareholder who is a purely nominal participant in the litigation. 2

3 whether the threat of unwarranted litigations impedes normal businesses, empirical evidence has been scarce. In this paper, we present the first effort to empirically identify how the threat of litigation distorts managers incentives ex ante and adversely affects shareholder wealth in corporate takeover transactions. 2 We use mergers and acquisitions (M&As) as a laboratory to examine the ex ante value destruction effect of shareholder litigation threat for two reasons. First, mergers and acquisitions are one of the most important decisions a company makes and can enhance or destroy shareholder value at a massive scale (e.g. Eckbo 2013). Second, mergers and acquisitions, especially the larger (and more lucrative) ones, are often targeted by shareholder litigations. 3 Therefore, merger and acquisition deals are a likely place to find the impact of litigation threat, if it has any effect. Specifically, we examine whether the threat of litigation leads to lower value creation for acquirer shareholders and, more importantly, whether this value effect is attributable to the distortion of acquirer mangers ex ante incentives by litigation threat in mergers and acquisitions. Because shareholder litigation often imposes direct and indirect costs on officers and directors (Brochet and Srinivasan 2014, Ferris et al. 2007, Fich and Shivdasani 2007, and Karpoff, Lee, and Martin 2008), officer and directors may be deterred by the threat of litigation from making optimal business decisions (Kinney 1994). In the case of mergers and acquisitions, litigations tend to target transactions that have greater 2 There can also be an ex post value destruction effect of shareholder litigation, i.e., shareholder litigation shifts wealth from shareholders to lawyers in the form of ex post legal fees. For example, Alexander (1991) argues that shareholder compensation in litigation settlements shifts wealth from current shareholders to former shareholders and the net effect is simply the destruction of shareholder value in the amount of the transaction costs of the litigation, a large part of which is wealth transferred to law firms. 3 For example, a recent Cornerstone Research report shows that about 93% of M&A deals valued over $100 million in 2014 were litigated. 3

4 potentials of generating large attorney fees but are otherwise value-enhancing from shareholders perspective. The threat of litigation thus may force corporate managers and directors to choose between maximizing shareholder wealth and minimizing (personal) legal liabilities in making acquisition decisions (Roger and Van Buskirk 2009), i.e., managers may pursue suboptimal acquisitions which reduce their risk of being sued at the expense of shareholders. This value destruction effect is due to the threat of shareholder litigation and is not a result of actual litigation, we therefore call it the ex ante value destruction effect of shareholder litigation. To examine any ex ante effect of shareholder litigation, however, is often challenging due to the difficulty to empirically capture the threat of litigation. Existing studies often rely on actual litigation events to measure litigation risk (e.g. Krishnan, Masulis, Thomas, and Thompson 2012 and Bourveau, Brochet, and Spira 2014), which renders two difficulties in identifying the causal effect of litigation. First, it can only uncover ex post effects of litigation but not ex ante effects of the threat of litigation. Second, it can be difficult to establish causality because the actual events of shareholder litigation are often correlated with unobservable factors and this can lead to omitted variable biases. In this paper, we overcome these difficulties by exploiting exogenous variation in the threat of litigation generated by the staggered adoption of universal demand (UD) laws, which makes shareholder derivative litigation more difficult, if not impossible. Derivative litigation is a primary form of litigation available to acquirer shareholders to challenge M&A transactions (Afsharipour 2012). 4 Between 1989 and 2005, 4 A derivative suit challenges a breach of fiduciary duty by directors and officers. In an acquisition transaction, for example, the plaintiff of the acquirer shareholders may sue the managers of the acquirer by alleging that the managers overpay for the target firm or assets after failing to conduct adequate due diligence. We look at 4

5 23 states sequentially adopted the universal demand (UD) laws (Appel 2015). After the adoption of UD laws, shareholders are required to seek board approval prior to initiating virtually every derivative litigation case. Since the board members themselves are usually the targets of these lawsuits, they seldom grant this approval. Consistent with the legislative mission, researchers have documented that UD laws created significant obstacles to derivative litigation and reduced the incidence of such lawsuits for firms incorporated in adopted states (e.g. Davis 2008 and Appel 2015). 5 Because UD laws were adopted by different states in different years, we employ a generalized difference-in-differences (DID) framework to examine the impact of UD laws on takeover outcomes. Specifically, using over 1,600 U.S. takeover transactions between 1980 and 2012, we examine the impact of UD laws on stock market reactions and postmerger performances. We find that the three-day acquirer cumulative abnormal return (CAR) is significantly higher if the acquirer s incorporation state has passed the UD law at the time of merger announcement. Specifically, the CAR is on average 4.04% higher if the UD law is passed in the acquirer s incorporated state. The increase in stock market reaction suggests that litigation threat is detrimental to acquirer shareholder wealth. While the baseline result is consistent with the ex ante value destruction effect of shareholder litigation, they may also be driven by the impact of shareholder litigation on deal completion uncertainty or ex post wealth transfer from the firm to lawyers. A merger acquirer shareholder litigation also because of the increasingly popular post-merger-close litigations. These post-close litigations are often against the acquirer s executives who are in charge after a merger is complete, as suggested by Feldman (2012). 5 More institutional details of derivative lawsuit involved in M&A deals and UD laws will be provided in the section 3. 5

6 deal may have to overcome greater hurdles to complete or may not be completed at all due to follow-up litigations. The UD laws, by reducing the threat of derivative litigation, decrease the uncertainty of deal completion, and therefore mechanically increase the acquirer announcement returns. In contrast, the ex ante value destruction hypothesis posits that the superior stock market reaction comes from more efficient merger decision-makings due to reduced litigation risk. To show that the deal completion uncertainty explanation does not drive our results, we proceed with the analysis of target announcement returns and long-term value creation of the merger deals. Any evidence of long-run value creation also indicates that the superior stock market reaction after UD laws does not only reflect the instant saving of legal costs. We first examine target cumulative abnormal returns (target CARs). If UD laws merely reduce the uncertainty of deal completion without having any long-run value impact, we should observe a positive effect of UD laws on target announcement returns. In contrast, if our baseline findings are driven by the ex ante value destruction effect, the impact of UD laws on target returns is ambiguous. Empirically, we find that UD laws have little effect on target CARs, which is not supportive of the reduced deal completion uncertainty explanation. Moreover, we examine the long-run operating performances (operating margin) of the newly merged entities after deal completion. We find that the UD laws lead to greater operation efficiency improvements, which again suggests that our baseline results are unlikely to be driven by the effect of UD laws on deal completion uncertainty. This result also suggests that our baseline findings are unlikely to be driven by ex post wealth transfer (in terms of tangible legal costs) from the firm to the lawyers because such an effect is 6

7 unlikely to persist and to affect long-run performance. These findings, together with the baseline results on acquirer CAR, suggest more efficient M&A decision-makings after UD laws, thereby are consistent with the ex ante value destruction effect of litigations threats. Next, in an attempt to identify the possible channel through which the threat of litigation destroys value, we examine how the passage of UD laws affects the probability of having common directors on the boards of both the acquirer and the target (connected boards deals). Connected boards deals are often targets of shareholder litigation because an argument of breach of fiduciary duty of loyalty can easily be made. 6 Nonetheless, recent studies such as Cai and Sevilir (2012) show that connected boards in fact lead to greater value creation for acquirer shareholders by reducing information asymmetry between acquirers and targets. We find that the probability of having connected boards increases after UD laws, suggesting that acquirers tend to avoid deals involving connected boards when faced with a greater threat of litigation (due to a higher likelihood of being sued). This result clearly illustrates the incentives of acquirer managers (or board of directors) in trading off maximizing shareholder wealth and reducing litigation liability. We then conduct an array of robustness checks. To check the validity of the baseline DID estimation, we first follow Bertrand and Mullainathan (2003) to study the dynamics of the effect of the UD laws. The concern is that the treated states and control states may have different trends in terms of merger outcomes, and the DID estimation simply picks up that trend difference. If that is the case, we should find the effect occurs even before the passing of the UD laws. In contrast, we find that the effect only occurs after the passing 6 In most states, exculpation statutes exist which relieve directors (while usually not officers) from liability to the company for breach of the fiduciary duty of care if they are believed to act in good faith. This makes the allegation of breach of fiduciary duty of loyalty more frequent. 7

8 of the laws, which suggests that the baseline results are unlikely to be driven by pre-existing trend differences between the states that have passed the UD laws and those that have not. We also show our DID estimation does not suffer from the over-rejection problem (Bertrand, Duflo, and Mullainathan 2004). A second concern is that the passing of the UD laws may be correlated with other state-level unobservable factors that also affect merger outcomes. To address this concern, we explore the potential cross-sectional heterogeneity of the effect of the UD laws on merger outcomes. Merger litigations often only target large deals because of the potential of extracting large legal fees. It is therefore expected that the effect should concentrate in large deals, which is why we only focus on large deals in our baseline sample. If the effect is driven by other unobservable factors, however, it should also affect small deals. We therefore conduct a placebo test on small deals to ascertain whether the baseline effect is most likely due to the UD laws instead of due to other unobservable factors. We find that the UD laws have no significant effect on smaller deals, suggesting that the baseline results are unlikely to be driven by other unobservable factors that are correlated with the passing of the UD laws which would have alike impact on large and small deals. Finally, we address several other endogeneity concerns. We first show that the positive effect of UD laws on acquirer announcement returns is not affected by the exclusion of firms incorporated in Delaware. We conduct this test because many firms choose to incorporate in the state of Delaware due to Delaware s special legal environment. In this case, the choice of the state of incorporation becomes endogenous and can be correlated with unobservable factors that can potentially affect merger outcomes. Excluding firms incorporated in Delaware alleviates this concern. We also show that the 8

9 positive effect of UD laws on acquirer announcement return remains intact if we only include firms incorporated in their headquarters states. Additionally, we check and confirm that UD laws affect acquirer announcement return in a similar fashion when we include firm headquarters state fixed effects or acquirer firm fixed effects. We find that the main results documented in our paper are not the effect of lobbying by interest groups, too. To the best of our knowledge, this is the first paper to document and to identify the ex ante value-destruction effect of litigation threat in corporate takeovers. Commentaries have long argued that the America appears to be overly litigious. Our findings have important implications on the role played by shareholder litigation in the efficiency of resource allocation and the welfare of the economy as a whole. The rest of the paper is organized as follows. We develop three competing hypotheses on the effect of shareholder litigation in section 2; section 3 provides more institutional details of derivative litigation and the UD laws; we describe our sample construction and summary statistics in section 4; section 5 presents the empirical results; and section 6 concludes. 2. Hypotheses In this section, we outline three competing hypotheses on the effect of shareholder litigation threat, which correspond to three different channels through which the threat of litigation can affect merger outcomes and shareholder wealth. The first is the disciplinary channel. Some studies suggest that litigation can discipline managerial behaviors, that is, managers and board of directors are more likely to pursue shareholder value-maximizing policies when facing a greater likelihood of being sued for misconducts (e.g. Niehaus and Roth 1999; DuChrame et al. 2004; Fich and Shivadasani 2007; Brochet and Srinivasan 9

10 2014; and Appel 2015). Under the disciplinary view, merger outcomes are likely to get worse following the passage of the UD laws because less litigation threat may exacerbate managerial agency problems. We therefore have the following hypothesis. Hypothesis 1: Under the disciplinary channel, less litigation threat (caused by UD laws) leads to worse merger outcomes. The second channel is the ex ante value destruction channel. It has been long argued that the threat of frivolous litigation may discourage management from taking necessary risk and experimenting new ideas (e.g. Block, Radin, and Maimone 1993; Kinney 1994; and Shaner 2014). Empirically, for example, the accounting literature has shown that managers respond to the threat of litigation by reducing information disclosure, a behavior consistent with less risk-taking (Johnson, Kasznik, and Nelson 2001 and Roger and Van Buskirk 2009). Because mergers and acquisitions are often risky and could subject the officers and directors to litigation, officers and directors may avoid pursuing deals at all or, when they do, they may choose to make deals that are less likely to expose themselves to litigation. However, the legally safer deals are not necessarily the ones that maximize shareholder wealth. Although it is often thought that officers and directors face a very small chance of personal liability due to the wide use of indemnification contracts and directors and officers (D&O) insurance (Black, Cheffins, and Klausner 2006), the reality is that settlements and judgments in derivative suits are usually not indemnifiable by the company (Ferris et al. 2007). In fact, a recent ACE report points out that directors become increasingly more 10

11 sensitive to the threat of personal settlement contributions. 7 For example, in the recent case of Activision Blizzard, several individual defendants (officers and directors) contributed to the total settlement of $275 million. In addition, derivative suits (indeed any lawsuits) are likely to impose non-pecuniary reputation damage to officers and directors. Studies such as Brochet and Srinivasan (2014), Ferris et al. (2007), Fich and Shivdasani (2007), and Karpoff, Lee, and Martin (2008) all find that litigation imposes indirect costs on officers and directors. Consequently, the potential direct and indirect costs may deter officers and directors from pursuing value maximizing deals that come with greater litigation risk, that is, managers trade off the potential loss against maximizing shareholder value when pursuing risky merger deals. Meanwhile, litigation threat can also distort merger advisors incentives, i.e., merger advisors may also trade-off potential liability arising from litigation and business judgement. In another recent massive derivative litigation case involving the acquisition of McMoRan by Freeport, Credit Suisse, the financial advisor of Freeport, paid $16.25 million in settlement. In this case, the financial advisor s potential liability may also distort its incentives in giving financial advice. One example of deals that may enhance shareholder wealth but at the same time may also expose directors to greater litigation risk is deals with connected boards, that is, some directors of the acquirers also serve on the boards of the targets. The common directors are often accused of having conflict of interests. However, as documented in Cai and Sevilir (2012), such deals are, on average, value-enhancing as common directors 7 ACE Report, 2005, Derivative suits: Recent development, 59. One mitigating factor is the increased popularity of D&O insurance policies affording coverage for only non-indemnifiable losses (i.e., Side-A policies), which also suggests that directors become concerned with potential personal liability arising from derivative litigations. Nonetheless, firms have to pay extra premium for these policies and these costs are ultimately born by investors. 11

12 mitigate asymmetric information between the acquirers and targets and lead to more efficient deal negotiations and better merger outcomes. The UD laws, by increasing the difficulty of derivative litigation, help insulate officers and directors from potential litigation risk, and hence mitigate the potential distortionary effect in their effort to pursue value-maximizing merger deals. We therefore develop the following hypothesis under the ex ante value destruction channel as our second hypothesis: Hypothesis 2: Under the ex ante value destruction channel, less litigation threat (caused by UD laws) leads to better merger outcomes. The third channel through which the UD laws can also positively affect acquirer CAR is the ex post value destruction channel. Under this value destruction channel, while shareholder litigation may not have any effect on officers or directors incentives, litigation helps lawyers to extract wealth from the firm in the form of settlement payments (Romano 1991; Speiss and Tkac 1997; Johnson et al 2000). In the context of mergers and acquisitions, litigation may also significantly increase the uncertainty of merger completion and the time it takes to complete (Krishnan et al 2012). In both cases, litigation may destroy firm value ex post and therefore leads to lower announcement returns ex ante. However, the ex post value destruction channel should have minimal, if any, effect on longrun post-merger operating performance, as the effect is likely to occur before the completion of the merger and hence likely to be short-term. We therefore develop the third hypothesis: 12

13 Hypothesis 3: Under the ex post value destruction channel, less litigation threat (caused by UD laws) leads to higher announcement returns but not necessarily better post-merger operating performance. Note that that both the second and third hypotheses predict better acquirer CAR after the passage of UD laws. The major difference between the two hypotheses is that the ex ante value destruction channel (i.e. the second hypothesis) suggests that savings from legal costs/deal completion due to less litigation threat is only part of the total value creation. Therefore, to pin down which channel is at work, we will need to analyze the sources of value creation such as long-term synergy gains. 3. Derivative Litigations and the Universal Demand (UD) Law 3.1 Derivative Litigation Under U.S. law, corporate directors and executive officers owe fiduciary duties to the corporation and its shareholders (Becker and Stromberg 2012). 8 These duties include the duty of care and the duty of loyalty (Macey and O Hara 2003). The duty of care demands that directors and officers exercise reasonable care, prudence, and diligence in the management of the corporation. The Delaware Supreme Court has characterized the duty of loyalty as the rule that requires an undivided and unselfish loyalty to the corporation demands that there shall be no conflict between duty and self-interest. If these duties are breached, directors and officers may be held personally liable for any damages caused by the breach. 8 From a legal perspective, a corporation is a creature of statute that exists independently of its shareholders (Appel (2015)). 13

14 State law permits shareholders to initiate judicial proceedings to vindicate shareholder interests protected by the concept of fiduciary duty (Ferris et al. 2007). Derivative lawsuit is a special subset of shareholder litigation. A derivative action allows shareholders to bringing suits against executive officers or directors on the corporation s behalf to address a breach of fiduciary duty. The name derivative implies that the shareholder is asserting his rights derivatively. For example, if a manager mismanages the company and causes a general decline in the company s value, the primary recipient of harm from this managerial action is the corporation and shareholders injury is indirect in nature since it results from the damages to the corporation. In this case shareholders can sue the manager derivatively on behalf of the corporation. In acquisition transactions, the acquirer s shareholders are not losing their status as shareholders. Therefore they are generally limited to bring derivative lawsuits when alleging that directors and officers have breached their fiduciary duties to the corporation (Afsharipour 2012). 9 Ferris et al. (2007) show that 7% of allegations in their sample of derivative lawsuits between 1982 and 1999 are directly related to mergers and acquisitions. In practice, a number of well-known merger deals involve acquirer shareholders derivative litigations. For example, on June 29, 2011, Oracle Corp. announced that it agreed to acquire 100% ownership of the privately-held Pillar Data System, Inc. Pillar Data System is majorityowned by Lawrence J. Ellison, Oracle s founder and CEO at that time. Each share of Pillar common and preferred stock was exchanged for rights to receive an earn-out in After the acquisition was announced, plaintiffs led by the City of Roseville Employees 9 Though allegations related to acquisitions are usually brought as class action lawsuit (by target shareholders), they often initiate parallel derivative claims as well. 14

15 Retirement System and the Southeastern Pennsylvania Transportation Authority filed derivative suits alleging that the Oracle Board acted disloyally and caused Oracle to waste corporate assets by allowing Ellison to force Oracle to acquire Pillar in a transaction with an unfair price structure, exposing Oracle to substantially increased risk. Another example involves the Freeport-McMoRan Copper & Gold Inc. ( Freeport ) s acquisition of McMoRan Co. ( MMR ). Freeport was a Phoenix-based copper and gold mining company and MMR was a New Orleans-based developer of oil and natural gas. On December 5, 2012, Freeport announced it had signed definitive agreements to acquire MMR for $3.4 billion in cash. Plaintiffs led by several pension funds and the Amalagmated Bank filed derivative suits alleging that the company overpaid for MMR. The plaintiffs also claimed that the transaction was tainted because the companies had overlapping directors and ownership stakes. There are several reasons why derivative suits may not deliver their purported effects. First, there are serious concerns about the impact of the divergence of incentives between the plaintiff shareholders and their attorneys (e.g. Jefferies 2014). 10 The divergence of incentives rises from different monetary values provided by lawsuits to the two parties. According to Jefferies (2014), most merger litigation cases settle and, of those that settle, an overwhelming majority of cases do not result in any monetary benefit for the plaintiff shareholders. Yet, virtually all lawsuits result in the award of attorneys fees. Both cases mentioned above ended up with settlements with hefty fees paid to plaintiff s lawyers. 10 As pointed out by Appel (2015), monetary settlements of derivative suit are circular in nature. Any financial recovery from a derivative suit is paid to the corporate treasury because suits were brought on behalf of the corporation. So shareholders do not directly receive a payment. Moreover, in most cases, a large if not all portion of the payouts are covered by directors and officers (D&O) insurance. 15

16 In the Oracle case, the company agreed to pay $15 million in attorneys fees to help settle the lawsuit. In the Freeport case, the Wall Street Journal estimated the attorney fee for this case around $35 million. Therefore, people worry that derivative suits may have been abused by attorneys who may initiate litigation for the only sake of forcing a settlement and generating large legal fees. 11 Legal studies have also expressed concerns about the unintended deterrence purpose of a derivative suit. For example, Kinney (1994) explicitly suggests that A problem with this deterrence purpose (of a derivative suit) may be that management is also deterred from risk-taking and trying innovative ideas. He points out that a corporation s competitive position can be at risk if the corporation over-emphasizes on playing it safe and avoiding experimentation. 3.2 Universal Demand (UD) Law A unique procedural aspect of a derivative suit is that it is designed to give management an opportunity to decide whether it should bring a suit against the wrongdoers. Prior to instituting a derivative action, it is required by state laws that a shareholder must first make a demand on the board of directors to take remedial action (via litigation or other means) to deal with the alleged wrongdoing. However, because derivative suits usually name the board members as defendants, it is anticipated that directors almost inevitably decide against proceeding with the litigation (Swanson 1992). Once a demand of a 11 In 2012, the U.S. Chamber of Commerce s Institute for Legal Reform released a study on merger-related litigation. The report concludes: If the allegations were real, there would be intense law enforcement focus on such a hotbed of fraud by the Securities and Exchange Commission, the Department of Justice, and State Attorneys General. That has not happened. The only plausible explanation for this lawsuit epidemic is abuse of the litigation system by trial lawyers. 16

17 derivative suit has been made and rejected by the board, courts will generally follow the deferential business judgment rule and almost never allow shareholders to proceed further. To prevent directors from wrongly blocking a derivative suit, courts have developed the futility exception to the demand requirement. Under this exception, the demand requirement is excused when shareholders plead that demand would have been futile since a majority of the directors are the wrongdoers. Shareholders prefer claiming that demand would have been futile because if demand is refused, most courts apply the business judgment rule. Kinney (1994) suggests that many plaintiffs avoid demand by filing suit and pleading to the court that demand would have been futile. 12 Yet, several deficiencies of the demand futility exception have been brought up. Most of the criticisms lie in the difficulty for proving demand futility. For example, in response to the filing of a shareholder lawsuit, the board of directors can form a special litigation committee (SLC) consisting of independent and disinterested directors to review the situation. 13 In this case, it is hard to persuade a judge that the board is not in a position to receive and evaluate demand. Also, the futility exception propagates inefficiency from the perspective of judicial effort. In fact, Swanson (1992) argues that most derivative litigations turn the focus from the alleged breach of fiduciary duty by directors and officers of the corporation to the issue of demand futility. Concerned with the burdensome procedural difficulties of the demand futility, the American Bar Association (ABA) proposes the idea of eliminating the futility exception to the demand requirement in the Model Business Corporation Act (MBCA). According to 12 A number of courts have held that when shareholders choose to demand, they have demonstrated that facts showing futility are absent. 13 Again, the SLC is subject to the critic that it suffers from structural bias in favor of the corporation. 17

18 the MBCA, demand should be required in all derivative actions instead of being excused for futility. Beginning with Georgia and Michigan in 1989, 23 state legislatures have adopted this concept (the Universal Demand (UD) law). UD laws are considered as significant obstacles to derivative litigation. Kinney (1994) notes The effect of the MBCA approach is that if a majority of the board or committee is independent and the corporation is likely to have ensured that it is - then the court will dismiss the derivative suit. Davis (2008) and Appel (2014) conclude that UD laws significantly reduce the incidence of derivative litigation and weaken the deterrence function of derivative litigation. Following Appel (2014), we use the staggered adoption of UD laws as exogenously generated variations in the difficulty of raising derivative lawsuits and examine its impact on takeovers. 4. Data, Variables, and Summary Statistics 4.1 Data and Sample The initial merger sample is drawn from the SDC database and includes mergers and acquisitions announced between January 1, 1980 and December 31, We choose this sample period to ensure having enough observations both before and after UD laws, since the UD laws were adopted between 1989 and We apply the following filters to refine our data: 1. The acquirer is a publicly traded U.S. firm. 2. The target is a publicly traded U.S. firm. 3. The acquirer owns less than 50% of the target before the acquisition, and more than 50% after the acquisition. 18

19 4. The acquirer and the target are not financial or regulated firms. 5. The merger status is Completed". We then match the SDC mergers and acquisitions data with Compustat by matching on the CUSIP numbers or ticker symbols. The matching process ends up with 2,869 deals that satisfy the above criteria. Following Bouwman, Fuller, and Nain (2009), we further exclude deals with transaction value less than $50 million. We choose to focus on large M&A deals because shareholder litigation often only targets these large deals that have potentials to pay large amounts of settlement fees. The information on acquirer incorporation state is obtained from Compustat. We supply the missing incorporation state with information from the SDC, if available. Finally, we have 1,626 unique deals with nonmissing acquirer incorporation state information and with sufficient daily returns from CRSP to calculate acquirer CAR. The number of observations used for empirical tests will depend on the availability of additional variables included Variable Definitions The key dependent variable of our analysis is the acquirer abnormal announcement return. We follow Moeller, Schlingeman, and Stulz (2004) to calculate the acquirer cumulative abnormal returns (CARs) during a three-day window that spans from one day before to one day after the event day (-1, +1), where event day 0 is the acquisition announcement date. Using the CRSP-equally weighted return index as the market return, we estimate the market model parameters over the 200-day period from event day -205 to event day -5 (Brown and Warner 1985 and Moeller, Schlingeman, and Stulz 2004). The 14 In unreported results, we exclude a few deals that were announced around earnings release dates (equal, or plus or minus one day) and find that our main result remains robust. 19

20 key independent variable is UD Law, which equals one if the acquirer s incorporation state has passed the UD law in a given year, and equals zero otherwise. We follow the existing literature (e.g., Moeller, Schlingeman, and Stulz 2004) to construct other control variables as follows. Log (Deal Value) is the natural logarithm of the merger transaction value in $ million. This variable captures the absolute size of the merger. We also include another measure of the relative size of a merger, Relative Deal Size. Following Cai and Sevilir (2012), we define Relative Deal Size as the ratio of the merger deal value over the acquirer s market capitalization. Studies show that acquisitions with smaller deal size relative to the size of the acquirer are associated with better acquirer announcement returns (e.g. Eckbo and Thorburn 2000 and Moeller, Schlingeman, and Stulz 2004). We include a dummy variable, Tender Offer, to indicate whether an acquisition is solicited as a tender offer. Jensen and Ruback (1983) document that tender offer acquisitions are associated with better bidder announcement returns. To account for the potential impact of deal attitude, we use a dummy variable, Hostile, which equals one if the bid is hostile according the SDC and zero otherwise. In addition, we take into account the competitiveness of the bid by adding a dummy variable, Competing Bidder, which is set to be one if the target receives multiple bids and zero otherwise. We include two dummy variables, All Cash Deal and All Stock Deal, to capture the effect of the payment method. All Cash Deal equals one if the deal is paid all with cash and All Stock Deal equals one if the deal is paid all with stocks. The literature shows that the method of payment is related to the stock market s response to acquisition announcements. 20

21 Travlos (1987) and Amihud, Lev, and Travlos (1990) report that stock-financed deals have lower CARs. We then follow the literature to include acquirer and target firm characteristics in various specifications. Acquirer and target firm characteristics are all measured at the fiscal year-end immediately prior to the acquisition announcement. We first control for the acquirer s size using the natural logarithm of its stock market capitalization, Log (Acquirer Market Cap.). The literature suggests that the acquirer size is an important determinant of acquirer performance (Moeller, Schlingemann, and Stulz 2004). Prior studies show that the merger announcement return is related to an acquirer s Tobin s Q (e.g. Lang, Stulz, and Walking 1991; Servaes 1991; and Moeller, Schlingemann, and Stulz 2004). We define Tobin s Q as the ratio of an acquirer s market value of assets over its book value of assets. Following Masulis, Wang, and Xie (2007), we also control for the acquirer s financial leverage (Acquirer Leverage) and cash holdings (Acquirer Cash Holdings). According to Masulis, Wang, and Xie (2007), an acquirer s leverage may serve as a governance mechanism because higher debt reduces future free cash flows (Jensen 1986). Harford (1999) shows that acquirers with more cash holdings are more likely to engage in empire building activities in mergers and acquisitions. We further include acquirer s profitability as a control variable. Acquirer Profitability may capture management quality which may be correlated with acquisition quality. Under some model specifications, we include the same set of target firm characteristics. The detailed definition of these variables can be found in Appendix B. We winsorize all variables at their 1/99 percentiles to mitigate the impact of extreme values. 21

22 4.3 Summary Statistics Summary statistics of the main variables are presented in Table 1. Consistent with prior work, acquirers on average experience negative cumulative abnormal returns around the announcement. Targets generally experience large positive abnormal returns on announcement. Overall, the bidders are much larger in firm size than the targets. Bidders also tend to have greater growth opportunities than targets. The summary statistics of other deal, acquirer, and target characteristics are similar to those reported in the existing literature. 5. Empirical Analysis 5.1 The Impact of UD Laws on Acquirer Announcement Returns To evaluate the impact of universal demand (UD) law on acquirer CAR, we use the generalized difference-in-differences (DID) regression framework, in which the treatment is the adoption of UD laws. Specifically, we estimate the following: Acquirer CAR = UD Law+ X Y Z (1) ijkst i i j j k k s t l ijkst The dependent variable is the cumulative abnormal return (CAR) of merger deal i that involves acquirer firm j and target firm k. The acquirer s incorporation state is s and the deal occurs in year t. As described before, UD Law is an indicator variable for whether the acquirer is incorporated in a state that has adopted the UD law at time t. X i is a vector of deal characteristics. Y j is a vector of acquirer firm characteristics. Z k is a vector of target firm characteristics. s is the acquirer state fixed effects and t is the year fixed 22

23 effects. l is the acquirer industry fixed effects. In this experimental design, the treatment group is firms in states with UD laws. The state fixed effects therefore capture the variation that comes from the permanent differences between the treatment and control groups. The year fixed effects control for time varying trends common to both treatment and control groups. The coefficient on UD Law is the difference-in-differences estimator, which captures the average change in the acquirer CAR by firms in states with UD laws relative to the average change in acquirer CAR by firms in states without UD laws. The staggered adoption of UD laws provides an experiment with multiple treatment and control groups. This design reduces biases and noise associated with one single treatment (Roberts and Whited 2012). Because the main independent variable UD Law is measured at the incorporate state level, we cluster the standard error at the acquirer incorporation state level (Bertrand, Duflo, and Mullainathan 2004). We report our baseline regression results in Table 2. Column (1) includes only the UD Law dummy variable with other fixed effects. The estimated coefficient of UD Law is positive and statistically significant at the 1% level. We present this result without any other covariates because other control variables (deal characteristics and firm characteristics) may also be affected by UD laws. On the other hand, adding otherwise exogenous covariates would improve estimation efficiency. In Column (2), we add a set of merger deal characteristics. The coefficient on the UD Law dummy remains positive and significant at the 1% level. In Column (3), we further include acquiring firm s characteristics as covariates, in addition to those deal characteristics. We find that the coefficient on UD Law is still positive at the same statistical significance level. In Column (4), we replace acquirer firm characteristics with target firm characteristics while still keep 23

24 deal characteristics in the regression. The coefficient estimation of UD Law stays positive and significant. Finally, we include all the control variables in the regression in Column (5); and find that UD Law still has a positive and significant impact on the acquirer CAR. In general, Table 2 shows that acquirer CAR has a robust and positive association with the adoption of UD laws. This finding indicates that acquirers on average experience a significant increase in the cumulative abnormal returns upon deal announcement after the acquirers states adoption of UD laws. Overall, the evidence is consistent with both the ex ante and the ex post value destruction channel of shareholder litigation (Hypothesis 2 and 3), but inconsistent with the discipline channel (Hypothesis 1). The coefficients on other control variables in Table 2 are mostly consistent with prior studies. Similar to Cai and Sevilir (2012), acquisitions with greater deal size relative to the size of acquirer are associated with lower acquirer CARs. Consistent with Travlos (1987) and Amihud, Lev, and Travlos (1990), we find that cash-financed deals yield higher acquirer CAR. We also find that tender offers are related to better acquirer announcement returns (Jensen and Rubak 1983). The coefficient estimate of Hostile is not significant under all model specifications. This is similar to Schwert (2000) who finds no significant economic difference between deals described as friendly or hostile. Acquirer s cash holdings are negatively related to merger announcement returns, consistent with the agency cost argument (Jensen 1986 and Harford 1999). 5.2 Ex Ante versus Ex Post Value Destruction We then move on to distinguish between the ex ante and the ex post value destruction effect. While the positive effect of UD laws on acquirer announcement returns 24

25 is consistent with the ex ante value destruction effect of shareholder litigation, it can also be driven by the effect of shareholder litigation on the uncertainty of deal completion. Krishnan et al (2012) document that M&A offers subject to litigation are completed at a significantly lower rate than offers not subject to litigation. Together with the assumption that the deals, if completed, create value for the acquirers, it implies that UD laws can mechanically increase the acquirer announcement returns by simply decreasing the deal completion uncertainty or increasing the rate of deal completion. To show that the baseline results in Table 2 are not driven by this explanation, we carry out three different sets of tests. First, we examine how the UD laws affect target announcement returns (target CAR). Target firms often experience substantial positive announcement stock returns upon merger announcements, which reflect the expected gain of the merger from target shareholders perspective. The expected gain is the product of realized gain if deal is completed and the probability that the deal will be completed. Therefore, if shareholder litigation threat decreases deal completion rate, it should negatively affect target announcement returns. In other words, target CAR should be positively related to UD Laws if reduced threats of litigation (due to UD laws) improve the odds of having deals done. To test this, we replace the dependent variable in Equation (1) with target CAR over the event period (-1, +1). The results are presented in Panel A of Table 3. In contrast to the results presented in Table 2, the coefficients on UD Law are neither statistically nor economically significant, which suggests that acquirers states UD laws have limited effect on target returns. This finding indicates that the baseline results are unlikely to be driven by increased rates of deal completion due to UD laws. 25

26 To further show that the threat of litigation destroys value ex ante, we examine the effect of the UD laws on long-run operating performance. Long-run operating performance of completed deals is unlikely to be affected by the uncertainty associated with deal completion. Furthermore, long-run performance is also unlikely to be affected by ex post wealth extraction because such wealth extraction often has only a temporary effect. If UD laws also positively affect long run operating performance, we can further rule out the possibility that the baseline results are driven by either increased deal completion rates or ex post wealth extraction (legal fees). Specifically, we study changes in the operating profit margin (operating income before depreciation, interest, and tax (EBIDTA) over net sales, also known as return on sales (ROS)) of the merged entity four years after the acquisition, compared with the salesweighted average of similar measures in the acquirer and the target four years before the acquisition. The operating profit margin is often used to compare the profitability across companies of different sizes (Farris, Bendle, Pfeifer, and Reibstein 2010), which is suitable for studying profitability dynamics in the context of a merger. For each fiscal year in the four-year period prior to the deal announcement, we calculate the industry adjusted ROS of the acquirer and the target by subtracting the median ROS in their industry based on the three-digit SIC codes. In a similar fashion as in Healy, Palepu, and Ruback (1992), we calculate the weighted average industry-adjusted ROS of a portfolio of the acquirer and the target for each fiscal year, where the portfolio weights are the sales of the acquirer and the target at the beginning of that fiscal year. Then we use the four-year mean of the yearly portfolio ROS as a measure of the pre-merger ROS of the acquirer and the target. For the four years after the deal completion year, we simply use the four-year mean of the merged 26

27 firm s industry-adjusted ROS as our measure of post-merger ROS. We then calculate the difference between post-merger ROS and pre-merger ROS, denoted as measure of the change in operating performance of the combined company. ROS, as one We examine how UD laws affect ROS and present the results in the Panel B of Table 3. We include the same set of control variables as in our baseline regressions. In Columns (1) to (4) of Table 3, the coefficient estimates on UD Law are positive and statistically significant at the 10% level. In Column (5), in which we include all deal, acquirer, and target characteristics, this coefficient becomes more significant at the 5% level. This result shows that the post-deal operating profit margin improvement is significantly better for mergers after UD laws, suggesting these mergers create greater value for the combined entities in the long run. In addition, we also analyze the changes of ROS around mergers for the acquirer alone. We do this test because shareholders of the acquirer may only pay attention to the operating performance dynamic of the acquirer itself. To obtain the ROS for the acquirer itself as an individual unit, we calculate the pre-merger ROS as the average of the fouryear industry-adjusted ROS of the acquirer alone. We then calculate ROS as the difference between the post-merger ROS of the merged entity and the pre-merger ROS of the acquirer. We present this analysis in the Panel C of Table 3. We find that UD laws also have a positive effect on the change of ROS of acquirers alone. In summary, the results in Table 3 are all consistent with the ex ante value destruction effect (Hypothesis 2) but inconsistent with the ex post value destruction effect of shareholder litigation (Hypothesis 3). 27