The Role of Activist Investors in the Market for Corporate Assets

Transcription

1 The Role of Activist Investors in the Market for Corporate Assets Adrian A. Corum Wharton Doron Levit Wharton April 15, 2015 Abstract This paper studies the role of blockholders and activist investors in the market for corporate control. We show that activist investors play an important role in dis-entrenching corporate boards who oppose value increasing takeovers. Importantly, we show that this role cannot be substituted by corporate buyers, even if the latter have the same governance expertise as activists. Moreover, we show that there is strategic complementarity between the search intensity of activist investors for firms that are likely to receive a takeover bid, and the search intensity of corporate buyers for targets with which they can create synergies. Among other implications, strategic complementarity implies that the aggregate M&A activity is positively related to the intensity of shareholder activism, and gives rise to episodes with high and low volume of transactions. Finally, our model provides interesting empirical predictions on the interaction between shareholder activism and M&A activity. In particular, we show that more severe agency problems in target firms provide activists with more investment opportunities, and hence, can result with a higher volume of M&A deals. Keywords: Takeover, Shareholder Activism, Corporate Governance, Proxy Fight, Search JEL Classification: D74, D83, G23, G34 The authors are from the University of Pennsylvania, Wharton School, Finance Department. (contact author) corum@wharton.upenn.edu and dlevit@wharton.upenn.edu, respectively. s: 1

2 Introduction This paper studies the role of blockholders and activist investors in the market for corporate control. Due to the separation of ownership and control, executives and board members of public companies enjoy benefits that are not shared with shareholders (e.g., Jensen (1986)). 1 In order to protect these private benefits of control, managers and directors may resist takeovers attempts that would otherwise create shareholder value. Under many jurisdictions, including the Delaware jurisdiction, the incumbent board has the power to block takeover attempts, for example, by putting in place poison pills or other defensive tactics. 2 The resistance of the incumbent board can be overcome only if its members are voted out in a contested director election ( proxy fight ). Indeed, the power of shareholders to replace the board is a central element of the modern public corporation with dispersed ownership. 3 Is the proxy fight an effective mechanism to overcome managerial resistance? To answer this question, we analyze a bargaining model with the following key element: if the negotiations with the target fail, the buyer has the option to challenge the incumbent board of the target during director elections. If the proxy fight succeeds, the buyer can use its control of the board to reach an acquisition agreement and close the deal. If the proxy fight fails, the incumbent board of the target can block the takeover. Winning a proxy fight and ousting the incumbent board, however, is not trivial. The buyer must convince the majority of target shareholders that voting for the alternative slate of directors is indeed in their best interest. Thus, a key assumption of the model is that a proxy fight (as well as the takeover itself) is successful only if target shareholders support the buyer. Our first result demonstrates that even if the buyer has the governance expertise that is required in order to run a proxy fight, any attempt to oust the target board is futile. That is, target shareholders never cast their vote in favor of the buyer s candidates, even if it implies keeping in office the entrenched incumbent board. The reason is simple. Target shareholders 1 Private benefits of control have different forms: excessive salaries and perks, investment in pet projects, prestige, publicity, or the pure pleasure of command. 2 We do not take a stand whether defensive tactics are beneficial or harmful to target shareholders. In fact, in the model, there are circumstances where target shareholders benefit from the ability of the board to unilaterally block the takeover. 3 Courts use the shareholder franchise as basis for allowing boards to block takeover bids: If the shareholders are displeased with the action of their elected representatives, the powers of corporate democracy are at their disposal to turn the board out. (Delaware Supreme Court, Unocal). See also Bebchuk (2007), who argues that the shareholder franchise is in fact a myth. 2

3 realize that if they elect the buyer s candidates, the buyer will abuse his new power. Specifically, since the buyer is the counterparty to the sale transaction, he will be tempted to use the control of the board to low-ball the takeover premium, extract value by tunnelling desired assets, or engaging in self-dealing. The inability to commit not to abuse the power of the board deters target shareholders from surrendering control in the first place. In other words, any proxy fight that is initiated by the buyer is doomed to fail. Without an alternative remedy, value increasing takeovers would fail even when an intervention mechanism such as a proxy fight exists. In turn, corporate buyers will have fewer incentives to search for targets with which they can create synergies, and as a result, more value is lost. Target shareholders benefit significantly from takeovers (e.g., see Betton et al. (2008)). Therefore, if among the shareholders of the target firm there are blockholders or activist investors, they may have the incentives to run a proxy fight themselves and thereby remove the opposition of the board to the takeover. In fact, hedge fund activists often demand firms to sell off all or part of their assets, and make governance-related changes (Brav et al. (2008)). Moreover, the idea that activist investors can put firms into play is also consistent with Greenwood and Schor (2009), who show that the abnormal returns to 13D filings stem from activists ability to force target firms into a takeover. 4,5 Can activist investors succeed where corporate buyers failed? To study this question, we introduce an activist investor into model. The activist chooses whether to search, identify which firm is a potential target, and then buy a stake in the target firm. 6 If the activist owns a stake in the target and the negotiations with the buyer fail, the activist can challenge the incumbent board during director elections. Our second result shows that, unlike corporate buyers, activist investors have the ability to win a proxy fight and facilitate the sell process. That is, target shareholders will vote out the resisting incumbent directors and elect the activist s candidates instead. This result holds even though the activist 4 Consistent with this view is a speech by a private equity fund manager Thomas H. Lee: I d like to thank my friends Carl Icahn, elson Peltz, Jana Partners, Third Point, he said. I d like to thank these funds for teeing up deals because they re coming in there and shaking up the management and many times these companies are being driven into some form of auction. See Will Credit Crisis End the Activists Run? ew York Times, August 27, Recent examples include the activist hedge fund Elliott Management has successfully pushed BMC Software and Compuware to their sale for $6.9 billion in 2013 and $2.5 billion in 2014, respectively (see The ew York Times, Elliott and Jana, Activist Investors, Are Behind 2 Big Buyouts, December 15, 2014). 6 In the model, by searching the activist privately learns which firm is a potential target. The activist then uses her informational advantage over the market maker to buy a stake in that firm for a relatively low price. 3

4 does not have a higher governance related expertise than the buyer in running proxy fights. 7 Intuitively, unlike the buyer, target shareholders trust the activist since both of them sit on the sale side of the negotiating table. Since the activist will try to maximize the value of her stake, shareholders are confident that the activist will use her control of the board to negotiate the highest takeover premium possible. 8 Taken together, our analysis suggests that activist investors play an important role in the market for corporate control, a role that cannot be substituted by corporate buyers. Decoupling the capacity to dis-entrench incumbent managers from corporate buyers is essential for overcoming inefficient managerial resistance. A case in point is Valeant s failed acquisition bid for Allergan in Prior to its unsolicited bid for Allergan, Valeant teamed up with the hedge fund activist Pershing Square with the intention that Pershing Square will build a significant toehold in Allergan, and then push for its sale to Valeant. Our analysis suggests that collaborations between activists and buyers are likely to fail. Indeed, the sophisticated maneuver of Valeant and Pershing Square has failed. 9 Essentially, by teaming up with a buyer, the activist loses its unique ability to relax managerial entrenchment since target shareholders can no longer trust the activist to advance their goals once she obtains control. Without the trust of target shareholders, the activist would be as ineffective as a buyer when challenging the leadership of the incumbent board during director elections. If activist investors can relax managerial entrenchment and thereby facilitate a sell process, they might also have the incentives to search for firms that are likely to receive a takeover bid, and profit from the increase in share value. Whether a given firm is likely to receive a takeover bid depends on its unique characteristics as well as on the intensity of search by potential buyers. Firms will try to identify potential targets with which they can create synergies by performing valuations, carrying out due diligence, and hiring the help of professionals such as lawyers, investment bankers, and financial advisors. Either way, the search process requires 7 Usually, activist investors own 5-10% of the shares of the target firm while buyers seek to acquire 100% of the firm. Therefore, all else equal and relative to the buyer, the activist incurs larger fixed costs per share when running a proxy fight. 8 Consistent with this result, activists have been successful in getting their nominees to be elected for board seats. According to the estimate of Institutional Shareholder Services (ISS), the percentage of cases where the activists won board seats has increased from 43% in 2012 to %68 in While Valeant s bid failed, Allergan eventually was bought by Actavis. evertheless, from the perspective of Valeant, the takeover attempt failed. See The Flaws in Valeant s Activist Deal Effort ew York Times, ovember 18,

5 corporate resources, and in particular, the attention of the firm s senior executives and board members. Firms will actively search for targets only if they believe that corporate assets that fit their strategy are also available for sale. For this reason, the existence of activist investors can relax the search friction in the market for corporate control. Our third result shows that there is strategic complementarity between the search intensity of activists for firms that are likely to receive a takeover bid, and the search intensity of corporate buyers for targets with which they can create synergies. Intuitively, since search is costly, and since the presence of an activist investor in the target firm relaxes the opposition of its board to the takeover, buyers have stronger incentives to search for a target if the likelihood that the identified target has an activist investor is higher. Similarly, activists have stronger incentives to search if the identified target is more likely to receive a takeover bid. Therefore, activists will search more intensively if buyers, due to their increased search effort, are more likely to identify these firms as targets. Strategic complementarity has three important implications. First, the aggregate volume of M&A activity is positively related to the intensity of shareholder activism (which can be measured by the volume of 13D filings). Second, small changes in the economic environment have a significant effect on the aggregate volume of M&A deals. For example, a regulatory change that marginally ease the access of shareholders to the ballot can have a large impact on aggregate volume of transactions. On the margin, activist investors will increase their search effort since an easier proxy access enhances their ability to dis-entrench the incumbent board of the target. Since the activist is more powerful and more likely to own a stake in the target, buyers will also exert more effort in searching for potential targets. Consequently, potential targets are more likely to be identified by a buyer. This effect amplifies the incentives of the activist to search for a target, and therefore, further increases the incentives of the buyer to search. For the same reason, other polices that ease the proxy fight process (for example, reimbursement of challengers cost of running the proxy fight or the declassification of the board) would have a large positive effect on the aggregate volume of M&A transactions. Moreover, factors that effect the activist s cost of buying a large stake in the target firm, such as disclosure requirement (e.g., the filing of schedule 13D) and market liquidity, will also have a large effect on the aggregate volume of M&A transactions. Finally, strategic complementarity implies that there can be multiple equilibria, characterized by the aggregate 5

6 volume of M&A transactions. The existence of multiple equilibria suggests that due to effective shareholder activism, the market for corporate assets can experience episodes of high volume of transactions ( hot markets ) and episodes of low volume of transactions ( cold markets ), without any apparent changes in the underlying fundamentals of the economy. In this respect, the extent of M&A activity is self-fulfilling. In the model, activist investors buy stakes in firms for two reasons: either because they believe that the company is likely to be a target (selection effect) or because they believe that their presence will help putting the company into play (treatment effect). The selection effect always exists since activists always invest in firms they believe are likely to be a target. However, the treatment effect exists only if the activist has the ability and incentives to relax the opposition of the incumbent board to the takeover. Therefore, the aforementioned strategic complementarity and its implications exist only if the treatment effect is present. Generally, the model provides a framework to study the implications of the treatment and the selection effects, and thereby helps creating identification strategies for empirical research. For example, market liquidity affects the average price the activist pays for the shares of the target. If market liquidity decreases then prices increase, and the activist s expected profit decreases. Since investment is less profitable, the activist will reduce her search intensity. If the activist cannot or does not have the incentives to relax managerial resistance in companies in which she owns a stake, then a decrease in market liquidity will reduce activist s search intensity, but will have no effect on the buyer s search intensity. By contrast, if the activist can and has the incentives to relax managerial resistance, due to strategic complementarity, lower market liquidity will also reduce the buyer s search intensity. Therefore, market liquidity should affect the aggregate volume of M&A activity if and only if the treatment effect is present. Finally, our model provides interesting empirical predictions on the interaction between shareholder activism and M&A activity (for example, the magnitude of target abnormal returns around merger announcement with and without a pre 13D filing by an activist investor). We show that when activists cannot or do not have the incentives to relax managerial resistance (no treatment effect), the volume of M&A deals decreases with severity of the agency problems in target firms. However, when activists can and have the incentives to relax managerial resistance, this relationship can reverse. That is, more severe agency problems result with a higher volume of M&A deals. Intuitively, when the incumbent board has high private benefits 6

7 of control, it is likely to resist takeover bids, and the volume of M&A deals should be low. However, higher managerial resistance provides activist investors with more opportunities to make profit on their abilities to put firms into play. Due to the strategic complementarity, buyers will increase their search intensity, and overall, the aggregate volume of M&A deals will increase. An interesting implication of this result is that the sign of the relationship between the severity of agency problems in the cross section of target firms and the volume of M&A deals can also point the existence of the treatment effect. Overall, our paper connects two important strands of the corporate finance literature: the literature on blockholders and shareholder activism (for a survey, see Edmans (2014)) and the literature on the market for corporate control (for a survey, see Becht et al. (2003)). Shleifer and Vishny (1986) studied the role of large shareholders in takeovers. Their focus was on the ability of the large shareholder to overcome the free-riding problem by making a bid for the target. By contrast, in our paper, the activist does not have the capacity to buyout the target. Moreover, we abstract from the free-riding problem among target shareholders, and instead focus on contested director elections as a mechanism through which managerial resistance can be relaxed. The ability of the target to resist takeover bids was studied by Bagnoli, Gordon, and Lipman (1989), Baron (1983), Berkovitch and Khanna (1990), Hirshleifer and Titman (1990), Harris and Raviv (1988), and Ofer and Thakor (1987). Unlike these papers, here, activist investors can dis-entrench the board, a feature which give rise to the strategic complementarity between activist s and buyer s search efforts. 1 Setup Consider an economy with a buyer, 2 potential target firms, and an activist investor. The market for corporate control has the two main frictions. First, the acquirer (we use the terms buyer and acquirer interchangeably) has to search for a target firm whose acquisition will create value. Second, the incumbent manager and directors of each potential target firm enjoy private benefits from control, and hence, may resist takeover bids, including those which add value. The expected standalone value of target firm i is q > 0, where q are common knowledge. The acquirer can add value to exactly one of the target firms. All firms are ex-ante 7

8 symmetric. If the buyer takes over this firm, the synergy is drawn from distribution F with a full support over [0, ). If the acquirer takes over any other firm, the acquisition destroys value. We assume that the value destruction is sufficiently large to deter the buyer from approaching a company without a proper search and due diligence process. 10 Moreover, we assume that ex-ante, each firm is equally likely to be the target. In order to find the firm with which it can create a synergy and learn about the level of this synergy, the buyer has to exert effort and start a search process. The cost of searching for a target is c > 0. If the buyer incurs this cost, then he identifies the target. We assume that c is the private information of the buyer and it is drawn from distribution G, with full support on [0, ). If the buyer searches and identifies firm i as a target, he starts negotiating a deal with the board of that firm. We assume that the synergy level becomes publicly known by the incumbent board and shareholders of firm i once it is approached by the acquirer and the negotiations start. The parties negotiate a cash offer for 100% of the shares of the target. There are two rounds of negotiations, t {1, 2}. In each round, the proposer is decided randomly and independently of the other round. With probability s (0, 1), the target board is the proposer and with probability 1 s the proposer is the buyer. The proposer in each round makes a take it or leave it offer. Thus, parameter s can be interpreted as the relative bargaining power of the target. We denote the offer made by the proposer in round t by π t. In each round, if the offer is accepted by the target board, the deal is brought to target shareholders for a vote. The deal closes only if the majority of target shareholders approve the acquisition. At the voting stage, shareholders are playing undominated strategy, and hence, approve the deal if and only if the takeover offer is higher than the perceived long-term standalone value of the target, q. If the offer is accepted by shareholders, shareholders get π t for each share they own, and the buyer gets q + π t. If the first offer is rejected, a proxy fight stage takes place. As we explain in details below, the proxy fight may change the composition of the target board. Once the proxy fight ends, a second round of negotiations starts between the target board (which may be populated with new directors) and the acquirer. The second round of negotiations follows the same protocol as the first round. If the second offer is rejected, the game ends and the 10 Specifically, suppose that if a firm is not the target then a merger will create dis-synergy < 0. The expected synergy in every firm is E[ ]+( 1) < 0. So, at the best case scenario, the buyer will pay q for the firm, and the expected profit will is bound from above by E[ ]+( 1) < 0. Therefore, without knowing that firm i is a target the buyer will not approach its board. 8

9 standalone value of the target firm is realized. The second friction in the market for corporate control stems from the agency problems in the target firm. We assume that the incumbent board of each firm i owns α I shares of the firm, but it also has private benefits from keeping the firm independent. The private benefits are lost if the target is acquired or if the incumbent board losses a proxy fight. We denote these private benefits by b > 0, which are common knowledge. 11 Going forward, it will be connivent to define b α I. The target board is entrenched and it has the ability to veto any takeover attempt as long as it has control. For example, the board can implement anti-takeover measures such as issuing shareholder rights plans (poison pills). The only way through which a takeover can be consumed in spite of the incumbent board s resistance, is by replacing the board. We assume that if the buyer and the incumbent board do not reach an agreement in the first round of negotiations, or if shareholders vote down a proposed transaction, a proxy fight to replace the incumbent board can be initiated. A proxy fight can be initiated either by the buyer, or by the activist investor, if the activist owns a stake in the target firm. At the proxy fight stage, the buyer and the activist simultaneously decide whether to challenge the incumbent board by proposing an alternative slate of directors at the director elections. If the buyer or the activist decide to launch a proxy fight, then they incur a fixed cost κ > 0. This cost is not reimbursed by the firm, and captures the effort that is exerted by the challenger when campaigning against the incumbent. We assume that the buyer and the activist has the same expertise of running a proxy fight. 12 Once the proxy fight is initiated, shareholders of the target decide whether to vote for the incumbent board or the other competing teams. When shareholders elect directors they play undominated strategies. If the incumbent board wins the proxy fight, it retains control. However, if the incumbent board loses, it is replaced by directors who represent the objective of the winning party. That is, if the activist wins the proxy fight, the buyer will negotiate in the second round with the activist, whose objective, as described in details below, is to maximize the value of the firm. If the acquirer wins the proxy fight, then he sits on both sides of the negotiation table at the second round. This allows the buyer to transfer some assets of the target to his company 11 In Section 3, we consider the case of a management buyout, changing the assumption on the private benefit to b < 0, i.e., there we regard b as a private cost. 12 This assumption goes against our result that only the activist is capable of removing the incumbent board from office. 9

10 regardless of the outcome of the negotiations. This results in a value transfer of ηq i from the target to the acquiring firm, where η (0, 1). Parameter η is determined by the strength of the legal environment and investor protection, and all the results hold even when η is arbitrarily close to zero. 13 When the proxy fight stage ends, the game proceeds to the second offer stage, which is also the last stage. We assume that if the incumbent board loses a proxy fight, not only it loses its private benefits from control but it also suffers an additional dis-utility which captures either embarrassment or the lose of reputation. 14 Similar to the buyer, the activist does not know initially which one of the firms is a potential target for the buyer, but she can incur a cost c and find out. We assume that c is the private information of the activist and it is drawn from distribution H, with full support on [0, ). The search costs of the buyer and the activist are independent of each other. Moreover, whether or not the activist searches for a target is the activist s private information. We also assume the activist searches at the same time that the buyer searches for the target. 15 Thus, the activist does not learn about the exact value of before she realizes whether the buyer arrives. Whether or not the activist engages in a search, she can buy a stake in each firm. We assume that the activist trades with a risk-neutral and competitive market maker, who sets the prices to be the expected value of the firm, given the total order-flow and any other available public information. Each firm has a separate market maker. The order flow for firm i is composed of liquidity traders, who submit an order to buy a stake of a size L (0, 1 α I ) in the firm with probability 1, and with probability 1 they submit no buy order. We assume 2 2 that the liquidity trades are independent across firms. We let market price of firm i be p (z i ), where z i 0 is the total order-flow in firm i. The market maker of firm i observes z i, but it cannot observes its source (that is, informed or uninformed) and it cannot observe order flows in other firms. If the activist bought a stake in firm i, he announces his position in the firm to the public (e.g., by filing schedule 13D). We assume that if the buyer did not incur the cost of searching, he would not notice the presence of the activist. Later on we relax this assumption. Finally, we assume that the activist s objective is fully aligned with the objective 13 As we discuss after Proposition 1, the results continue to hold even if η = This cost can be arbitrarily small, and its role is to ensure that if the incumbent board is sure he is going to lose his private benefit from control, he is better off surrendering control voluntarily, as opposed to being forced out of office. 15 Later, we consider a variant of the model where the buyer observes in which of the firms the activist has bought a stake. 10

11 of other shareholders given its long position in the firm. 16 focus our attention on pure strategies. All agents are risk-neutral, and we 2 Analysis Consider the set of Perfect Bayesian Equilibria of the game. We solve the game backward. First note that if the buyer does not approach firm i, then the firm will remain independent under the control of the incumbent board with probability one. Indeed, even if the activist owns a stake in firm i, she has not incentives to incur the cost of a proxy fight and replace the incumbent board. That is, removing the incumbent board from office does not create any value to shareholders or the activist. This result is a direct implication of our assumption that the private benefit that are consumed by the incumbent board are not on the expense of the standalone long-term value of the target. In the extension section we relax this assumption. Suppose the buyer identifies the firm i as a target and he approaches the firm. We start with the second round of negotiations. The negotiations in the second round depend on who control the board of the target. Generally, there are three scenario to consider. In the first scenario, the buyer controls the board. With control, the buyer uses the board s authority to sign on a deal that offers target shareholders the lowest amount they would accept. Moreover, by controlling the target board, the buyer can tunnel ηq from the target s standalone value. Therefore, in this case, the buyer offers shareholders (1 η)q and shareholders, who at this point cannot prevent the buyer from tunneling ηq of the target s value, accept this offer. In the second scenario, the incumbent board retains control of the target. The target board will relinquish control if and only if the buyer offers at least q +. In turn, the buyer will never agree to pay more than q +. Therefore, the parties reach an agreement if and only if. If then with probability (1 s) the buyer offers π 2 = q +, which is always accepted by the incumbent board and the shareholders, and with probability s the board offers π 2 = q +, which is also always accepted by the buyer and the shareholders. In the third scenario, the activist controls the target board. The activist is unbiased, and hence, she is willing to sell the firm for any offer higher than q. Therefore, the buyer and the activist always reach an agreement. With probability 1 s the buyer offers π 2 = q which is always accepted by the 16 In Section 3, we consider biased activists, where such activists have a preference for early liquidation of their stake. 11

12 activist and the shareholders, and with probability s the activist offers π 2 = q +, an offer which is always accepted by the buyer and the shareholders. Lemma 1 In the second round of negotiations, the target is acquired by the buyer unless the incumbent board retains control and >. Shareholders expected value conditional on is given by (1 η) q if the buyer controls the board, q + s + (1 s) if the incumbent board retains control and, Π SH = q if the incumbent board retains control and >, q + s if the activist controls the board. The next result uses Lemma 1 to characterize the preferences of target shareholders in the proxy fight stage. Proposition 1 Consider the proxy fight stage: (i) Whether or not the activist runs a proxy fight, shareholders never elect the buyer s team, if the buyer decides to run a proxy fight. (ii) Shareholders reelect the incumbent board unless > and the activist runs a proxy fight, in which case, shareholders elect the activist s team. Part (i) of Proposition 1 makes a simple yet powerful observation: if the buyer chooses to run a proxy fight, he will always lose the fight, whether or not the activist is also challenging the incumbent board and regardless of the incumbent board s private benefits. According to Lemma 1, target shareholders are always worse off when the buyer controls the board relative to the alternative teams, and apart from the knife-edge case where η = 0, they are actually strictly worse off. 17 The reason behind this result is the buyer s inability to commit not to abuse power once he controls the board of the target. Indeed, once the buyer wins the proxy 17 If η = 0 and then shareholders are also strictly better off voting for the incumbent board in the proxy fight. This can be seen from expression (1). In the knife-edge case η = 0 and >, shareholders are indifferent between letting the buyer controlling their firm and letting the incumbent board retaining control. However, as Part (ii) of Proposition 1 indicates, they would strictly prefer giving control to the activist. (1) 12

13 fight, he is always tempted to tunnel ηq of the target value and offer shareholders the lowest amount possible, (1 η)q. Shareholders anticipate the abuse of power, and therefore, they never give the buyer control of the firm s board, unless it is part of an acquisition agreement. The buyer anticipates his lose in the proxy fight, and in order to avoid spending the cost κ, he avoids running the proxy fight. The next corollary follows directly from this observation. Corollary 1 The buyer never runs a proxy fight once the first round of negotiations fails. Corollary 1 implies that the buyer will never launch a proxy fight to replace the incumbent and entrenched board, even though he has the technological ability to do so. This observation is true regardless of the cost of running a proxy fight κ, whether or not the activist runs a proxy fight, and regardless of the magnitude of the private benefits of the incumbent board,. It is worth noting that, in many circumstances, the buyer will choose not to run a proxy fight even if he expects shareholders to vote for him. Indeed, by controlling the target, the buyer is able to reduce the expected premium paid to shareholders from s + (1 s) to zero. If this amount is smaller than κ then the running a proxy fight remains unprofitable from the buyer s perspective. Part (ii) of Proposition 1 shows that shareholders will support the activist s proxy fight if and only if the incumbent board is sufficiently entrenched. Indeed, if > then shareholders understand that the buyer and the incumbent board will not be able to reach an agreement. Shareholders elect the activist, since the activist can relax the resistance of the board, and thereby facilitate a value-increasing takeover offer. If then shareholder expect the incumbent board and the buyer to reach an agreement. In this case, shareholders strictly prefer the incumbent board over the activist, since the board s resistance helps increasing the takeover premium, without endangering the existence of a deal. Effectively, the incumbent board s private benefit increases the target bargaining power in this case. The next result characterizes the incentives of the activist to run a proxy fight. Proposition 2 If the first round of negotiations fails, and the activist owns a stake of size α A in the target firm, then the activist runs a proxy fight if and only if Moreover, whenever the activist runs a proxy fight, she wins. κ sα A <. (2) 13

14 If the activist has no incentives to run a proxy fight even if she expects to win. Indeed, based on Proposition 1 and Lemma 1, the buyer never gets control of the target board, and if the incumbent board retains control, he will reach an agreement with the buyer, and expected premium would be s + (1 s). This premium is higher than s, the premium that the activist expects to negotiate with the buyer if she gets control of the board. As was discussed above when, the private benefits of the incumbent board is beneficial since it increases the relative bargaining power of the target. The target can commit to reject offers below q +, a commitment which is not credible if the activist controls the board. Instead, if < then the activist understands that if she chooses not to run proxy fight, then shareholders will retain the incumbent board, but the incumbent board and the buyer will not reach an agreement. The expected value of the activist s stake would be α A q. However, if the activist launches a proxy fight, then based on Part (ii) of Proposition 1 her challenge will succeed and she will control the board of the target. Based on Lemma 1, the expected value of the activist stake net of the cost of running the proxy fight would be α A (q + s ) κ. Comparing this term with α A q, and accounting for the condition < creates expression (2). Intuitively, expression (2) implies that the likelihood that the activist runs a proxy fight when the first round of negotiations fails increases with the target s bargaining power s, the stake of the activist α A, and the private benefits of the incumbent board. This likelihood clearly decreases with the cost of running a proxy fight, κ. The contrast between Proposition 2 and Corollary 1 shows that even though the buyer and the activist share the same technology of running a proxy fight (both face the same fixed cost κ), only the activist can utilize this capability. 18 Target shareholders are willing to delegate the negotiations on the takeover premium to the activist, since the activist has similar incentives to get the highest premium possible. By contrast, shareholders have no incentives to delegate this decision to buyer, who is obviously interested in offering shareholders the lowest premium possible. This lack of trust makes it impossible for the buyer to win the control of the target. Precisely for this reason, the activist s role in dis-entrenching the incumbent board is critical for the market of corporate control and cannot be substituted by the buyer. The next result accounts for the first round of negotiation, and characterizes the outcome of the interaction between the buyer and the target firm once the buyer identifies firm i as 18 In fact, since α A < 1, the effective cost of the activist from running a proxy fight is κ then the cost incurred by the buyer. α A which is higher 14

15 a target. ote that if the activist owns a stake in the target firm, then the presence of the activist and the size of her stake becomes public. If the activist did not buy a stake, then is it a common knowledge that the activist is absent. Proposition 3 Suppose the buyer identifies firm i as a target with a synergy level of. (i) If then regardless of the activist investor s presence and characteristics, the buyer pays q + s + (1 s) and takes over the target after the first round of negotiations. (ii) Suppose <. (a) If the activist investor owns a stake α A and κ sα A and takes over the target after the first round of negotiations. then the buyer pays q + s (b) If the activist investor does not own a stake in firm i or she owns a stake but < then the target remains independent under the control of the incumbent board. Proof. Suppose. Based on Proposition 1, whether or not the activist is present, the incumbent board retains control. Based on Lemma 1, all players expect the takeover to consume in the second round of negotiations, where the price is q + s + (1 s). Therefore, in the first round of negotiations, the incumbent board will reject any offer lower than q + s + (1 s), and the buyer will reject any offer higher than q + s + (1 s). If there are arbitrarily small waiting costs to either the buyer or the incumbent board, the deal will close in the first round. Suppose < and the activist is not present. Based on Proposition 1, the incumbent board retains control. Based on Lemma 1, the incumbent board and the buyer will not reach an agreement in the second round. Therefore, in the first round of negotiations, the incumbent board will reject any offer lower than q +, and the buyer will reject any offer higher than q +. Thus, the parties will not reach an agreement in the first round as well, and the target remains independent. Suppose < and the activist is present. According to Proposition 1 and Proposition 2, the buyer never controls the target board, and the activist runs and wins a proxy fight if and only if (2) holds. Thus, if < κ sα A κ sα A then the activist never runs a proxy fight. Therefore, the outcome is the same as in the case where the activist is not present. If κ sα A and the first round of negotiations fails, the activist will run a proxy fight and take control of the 15

16 target. Based on Lemma 1, all players expect the takeover to consume in the second round of negotiations, where the price is q + s. Therefore, in the first round of negotiations, the incumbent board will reject any offer lower than q + s, and the buyer will reject any offer higher than q + s. Since the board suffers a dis-utility from losing a proxy fight, the deal will close in the first round at a price of q + s. Search Stage Suppose in equilibrium the buyer searches for a target with probability λ B [0, 1] and the activist searches for a target with probability λ A [0, 1]. Recall that the decisions of the buyer and the activist to search are private. We first consider the activist s decision to search and then buy a stake in one of the firms, and then analyze the search decision of the buyer. The next result considers the activist s decision to buy a stake in firm i in equilibrium. Lemma 2 Suppose in equilibrium the buyer searches for a target with probability λ B [0, 1] and the activist searches for a target with probability λ A [0, 1]. (i) If the activist exerts effort and searches, she buys L shares of firm i if it is identified as a target, and buys no shares of firm i otherwise. (ii) If the activist does not exert effort and 2, then she does not buy a positive stake in any firm. We assume that 2, hence part (ii) of Lemma 2 confirms that the activist never finds it optimal to buy a stake in any company without learning first whether the company is a potential target. More generally, Lemma 2 indicates that the activist buys a stake in firm i if and only if the activist exerts effort, searches, and identifies firm i as a potential target. If the activist finds that firm i is a potential target, the activist will buy exactly L shares to disguise her trade as a liquidity and uninformed order-flow. In all other cases, the activist finds it optimal not to trade and buy shares of the firms. Using Bayesian rules and some algebra which appears in the Appendix in the proof of 16

17 Lemma 2, the share price of firm i is given by, 19 p i (z i ; λ A, λ B ) = q + λ B 1 (s + (1 s))df ( ) + λ A if s df ( ) min{, } κ (s + (1 s))df ( ) + if s df ( ) min{, } κ z i = L z i = 2L (3) otice that the share price increases in λ B and λ A. A higher probability that the buyer searches for a target increases the likelihood of a transaction with firm i, and hence, the value of holding a share of firm i. Similarly, higher probability that the activist searches for a target increases the value of the share of firm i, since if a buyer arrives but the incumbent board of firm i refuses to relinquish control, the presence of the activist can relax the tension and facilitate a transaction. According to Lemma 2, if the activist does not search, she does not trade and her expected profit is zero. If the activist engages in search, she will find the target firm and buy L shares of that firm. The expected profit of the activist net of the search cost is Π A (c A, λ A, λ B ) = c A + L q + λ B = c A Lλ B [ 1 ( (s + (1 s))df ( ) + min{, κ } s df ( ) ) 1 2 p (L; λ A, λ B ) 1 2 p (2L; λ A, λ B ) (s + (1 s))df ( ) + ( ) ] 1 λ A min{, } s df ( ). κ The activist will exert effort and search for a target if and only if Π A (c A, λ A, λ B ) 0. ote that Π A (c A, λ A, λ B ) decreases in its first argument. Therefore, in any equilibrium there is a unique threshold c A 0 such that the activist searches for a target if and only if c A [0, c A ], where c A satisfies Π A (c A, λ A, λ B ) = 0. In equilibrium, λ A = H (c A ). Interestingly, Π A (c A, λ A, λ B ) increases in λ B and decreases in λ A. Higher market expectations that the activist will own a stake of the target firm increases the share price, and therefore, 19 The price when z i = 0 is immaterial for the analysis, since Pr [z i > 0] whenever the activist submits a buy order for firm i. (4) 17

18 reduces the profit of the activist. On the contrary, higher market expectations that the buyer searches for a target increases the profits of the activist. Here, there are two effects. Similar to the effect of λ A on Π A, higher λ B increases the share price, and therefore, negatively effect Π A. On the other hand, higher λ B also increases the value of the activist s private information, namely, the knowledge of which of the firm is the target, and the fact that the activist decided to buy shares in the target and is planning to put pressure on the board to accept the future takeover bid. Since when the activist searches she knows which firm is the target but the market maker can only make imperfect inference, the positive effect of λ B on Π A always dominates. Overall, the activist is more likely to search, the higher is the likelihood that the buyer will do his due diligence. Formally, since Π A (c A, λ A, λ B ) decreases with c A and increases with λ B, the threshold c A increases in λ B. Consider the buyer s decision to search. If the buyer does not incur the search cost, he cannot identity which of the firms is a viable target. Since the expected synergy from an acquisition of an unidentified company is negative, the buyer does not approach any of the firms. In this case, the buyer s expected profit is zero. If instead the buyer engages in search, he identifies the target firm. Based on Lemma 2, the buyer knows if an activist is present, which happens with probability λ A, the activist will own a stake of size L in the target firm. Based on Proposition 3, the buyer s expected payoff net of the search cost is [ Π B (c B, λ A ) = c B + (1 s) ( )df ( ) + λ A min{, } κ df ( ) ]. (5) The buyer will exert effort and search for a target if and only if Π B (c B, λ A ) 0. ote that Π B (c B, λ A ) decreases in its first argument. Therefore, in any equilibrium there is a unique threshold c B 0 such that the buyer searches for a target if and only if c B [0, c B ], where c B satisfies Π B (c B, λ A) = 0. In equilibrium, λ B = G (c B ). Interestingly, Π B (c B, λ A ) increases in λ A, with a strict monotonicity if and only if > κ. Intuitively, the higher is the likelihood that the activist is present in the target firm, the more pressure the activist will put on the incumbent board to relinquish control and sell the firm, and hence, the higher is the likelihood that the buyer will consume the synergies from the acquisition. However, this argument works only if the threat of a proxy fight is credible when the resistance of the incumbent board is significant, that is > κ. Formally, since Π B (c B, λ A ) 18

19 decreases with c B and increases with λ A when > κ, the threshold c B increases in λ A. Proposition 4 An equilibrium always exists. target if and only if c B [0, c B ] where c B = ψ (c A ) > 0 and [ ψ (c A ) = (1 s) ( )df ( ) + H (c A ) In any equilibrium, the buyer searches for a min{, } κ df ( ) ]. (6) In addition, the activist searches for a target if and only if c A [0, c A ] where c A > 0 and it is a solution of The equilibrium is unique if κ. Π A (c A, H (c A), G (ψ (c A))) = 0. (7) As can be seen from (6), if κ then the presence of the activist does not change the incentives of the buyer to search for a target. The reason is that the activist s threat to run a proxy fight is not credible enough to relax the incumbent board s resistance. In this case, the equilibrium is unique. The activist selects and buys firms she believes are likely to be a target for a takeover, but the activist does not affect the likelihood that the firm becomes a takeover target. This is no longer true when > κ. In this case, the activist not only selects the target, but by owning a stake in the firm, the activist facilitates the takeover since she relaxes the resistance of the incumbent manager. Moreover, in this region the expected presence of the activist encourages the buyer to search more intensively for the target. So the activist affects the likelihood that the firm becomes a target directly and indirectly. Since c A increases in λ B and λ B = G (c B ), and since c B increases in λ A and λ A = H (c A ), the best response functions of the buyer and the activist are increasing. That is, the threshold below which the buyer decides to search is increasing in the threshold below which the activist decides to search, and vis-a-versa. In other words, the game exhibits strategic complementarity when > κ. As Proposition 4 suggests, strategic complementarity can lead to multiple equilibria. The existence of multiple equilibria suggests that due to effective shareholder activism, the market for corporate assets can experience episodes of high volume of transactions ( hot markets ) and episodes of low volume of transactions ( cold markets ), without any apparent change in to the underlying fundamentals. 19

20 The multiplicity of equilibria generated by strategic complementarity can be best seen when and. In this case, the target will sell if and only the activist is present. Therefore, the buyer will find it optimal to search if and only if the activist is present, and the activist will search if and only if she believes that the buyer will be present. In particular, as a special case of Proposition 4, (c A, c B ) must satisfy c B = (1 s) H (c A) c A = s L 2 G (c B) κ κ df ( ) df ( ) otice that (c A, c B ) = (0, 0) is always an equilibrium. That is, if the other player does not search for sure, the transaction will never take place, and hence, there are no incentives to search. However, if g (0) h (0) > s (1 s) L 2 [ ] 2 df ( ) then there is a sufficiently large mass of buyers and activists with low search costs, and an κ 1 (8) equilibrium where c A > 0 and c B > 0 also exists.20 3 Comparative Statics Let us focus our attention on two variables of interest. Probability of a takeover is given by G (c B) [ ] df ( ) + H (c A) df ( ), (9) min{, } κ 20 That can be seen by noting that c A must solve c A = s L 2 ((1 G s) H (c A ) κ df ( )) κ df ( ). The RHS converges to a finite number and the LHS to infinity. The derivative of the RHS is [ 2. h (c A ) κ df ( )] Therefore, if (8) holds, this derivative eval- ( s (1 s) L 2 g (1 s) H (c A ) ) κ df ( ) uated at c A = 0 is greater than one, and hence, an interior solution must exist. 20