I ,^ |UBmRiSSi g 'FEB 5 1988 ALFRED P. WORKING PAPER SLOAN SCHOOL OF MANAGEMENT TAKEOVER DF.FENSES AND SHAREHO LDFR VOTING David Austen- Smith University of Rochester and Patricia C. O'Brien Massachusetts Institute of Technology WP# 1823-86 Revised: November 1987 MASSACHUSETTS INSTITUTE OF TECHNOLOGY 50 MEMORIAL DRIVE CAMBRIDGE. MASSACHUSETTS 02139 TAKEOVER DEFENSES AND SHAREHOLDER VOTING * David Austen- Smith University of Rochester and Patricia C. O'Brien Massachusetts Institute of Technology WP# 1823-86 Revised: November 1987 * The authors gratefully acknowledge helpful comments from Jeff Banks, Harry DeAngelo, Linda deAngelo, John Parsons and Rick Ruback, and from seminar participants at MIT, Stanford and the University of Rochester. Remaining errors are soley the authors' responsibility. M 1.7. FEB Ll^'RAR'!^*^ 5 1988 RECEIVED Takeover Defenses and Shareholder Voting* David Austen-Smith Department of Political Science University of Rochester Rochester NY 14627 and Patricia C.O'Brien Sloan School of Management M.I.T. Cambridge MA 02139 October 1986 Revised: November 1987 *The authors gratefully acknowledge helpful comments from Jeff Banks, Harry DeAngelo, Linda DeAngelo, John Parsons, Scott Richard, Rick Ruback, David Scharfstein, and from seminar piu-ticipants at MIT, Stanford and the University of Rochester. Remaining errors are solely the authors' responsibility. Abstract Why do shareholders vote for anti-takeover devices which apparently lower the value of their firm? We address this question by constructing an agenda- setting model in which rational, informed, and value-maximizing shareholders vote on requests for such devices self-interested management with employment opportunities outside the furo. made by a We find sufficient conditions for the value of the firm to decline as a result of a request, although it is approved by shareholders. In our model, the apparently paradoxical voting behavior occurs because the expected takeover premium would be reduced more by rejection of the request than by approval, so shareholders rationally choose approval. Takeover Defenses and Shareholder Voting L Introduction Why do shareholders vote for anti-takeover devices which apparently lower the value of their firm? We address this question by constructing a model in which rational, informed, and value-maximizing shareholders vote on requests for such devices made by a self-interested management with employment opportunities outside the firm. We describe conditions under which the value of the firm declines as a result of the request, although it is approved by shareholders. In our model, the apparendy paradoxical voting behavior occurs because the expected takeover premium is reduced more by rejection of the request than by approval. A large and increasing number of amendments to corporate charters are specifically designed to increase the cost of transferring control. amendments in 1974 through 1979. Linn and McConnell (1983) find generally increasing incidence of such proposals among NYSE Responsibility Research Center amendments is DeAngelo and Rice (1983) repon over 250 proposed lists firms over the period ft-om 1960 to 1980. The Investor over 200 in 1985 alone. While the use of anti-takeover charter increasing, there is considerable disagreement over their effects on shareholders and on the market for corporate control. Two major hypotheses describing the motives for instituting anti-takeover chaner amendments are commonly described as "management entrenchment" and to the from management entrenchment view, incumbents "shareholder interests." ^ According are interested in job security and seek protection the takeover market, to the detriment of shareholders. However, a manager cannot change corporation's charter without winning voted approval ftom shareholders. the Two suggested explanations for shareholder approval of entrenching antitakeover devices are: (1) for a majority of shareholders, the costs of becoming informed about the effects of defensive charter exceed any potential benefits, and uninformed shareholders consistently give amendments their proxies to management; and (2) large shareholders wish to maintain friendly relations with management to ensure the benefits of future business, and large shareholders control sufficient shares to be pivotal in the vote. Shareholder irrationality is sometimes offered as a third alternative. explanation for the observed voting behavior must describe individuals shares to be pivotal, that is, sufficient shares to determine the The stockholder interests hypothesis recognizes a who control outcome of the free-rider Note problem sufficient vote. premia from takeover bidders, particularly dilution of minority interests are imperfect. Anti-takeover devices may benefit if may have remedies to limit shareholders by enforcing a level of collusion in takeover negotiations, to extract a higher premium. If this case, there is little no inconsistency in rational shareholders voting for these devices. evidence that shareholders benefit from such amendments. statistically insignificant by in collective action shareholders (Grossman and Hart (1980), Jarrell and Bradley (1980)). Shareholders difficulty colluding to extract larger any that However, is the there is DeAngelo and Rice (1983) find negative abnormal returns around the public announcement of proposed antitakeover amendments. Linn and McConnell (1983) find positive returns around the board meeting date at proxy mailing which amendments date. Jarrell, announcement of "shark are proposed, and insignificant negative returns around the Poulsen and Davidson (1985) find negative returns accompanying the repellents", viz. supermajority, classified board and authorized preferred amendments. In the context of our model, the management entrenchment and shareholder interests hypotheses do not necessarily lead to different predictions about shareholder value. Anticipatory takeover defenses raise the costs of acquiring control of the firm, thereby entrenching management, but lead to a higher potential premium for shareholders if a bid succeeds. managers contribute different value to a given firm, We presume that different by virtue of skill or experience. Only managers capable of producing higher value than the incumbent can mount successful takeover bids. Whether shareholders gain or quality of the lose as a result of the incumbent's defenses depends incumbent management relative to potential bidders, and the incumbent's upon the employment outside opportunities for rationally vote for anti-takeover it was before the shareholders are better off The paper is organized formal model is chaner amendments, although amendments were proposed. made developed We describe circumstances in which the firm. their wealth 2. lower afterwards than We also describe different circumstances in which by the amendments. as follows. In section 2 in section 3, we provide a brief overview of the model. and we summarize and discuss the Section 5 concludes. All proofs are confined to Appendix A. Appendix that illustrates is shareholders B The results in section 4. consists of an example our results. Overview We construct a two-period model of a firm with risk-neutral shareholders, with a manager who has employment opportunities outside the firm, and with the prospect of a management change by takeover. change bid is We consider the capital structure and assets of the firm to be fixed, and focus on the in value associated with a outstanding. During the charter change first in management. At the period, the incumbent amendments, on which shareholders made. Dep>ending upon the outcome of the employment with the firm versus vote. first expectations about what period no takeover period and the manager's preferences for the firm, the and the manager leaves the firm for may occur in first During the second period, a takeover bid may be employment outside shareholders and the incumbent manager of the manager may propose anti-takeover additional defenses in response to the bid. If the takeover bid their shares to the bidder start make decisions is manager may choose to erect successful, the shareholders sell alternative employment Both in the first period based upon rational the second period. A major implication of the results developed in sections 3 and 4 is that observations of a firm's value before and after the implementation of takeover defenses are not observations of the alternatives shareholders face when voting on these defenses. In our model, shareholders approve defenses face an unobserved, inferior alternative if who they choose rejection, and so rationally choose approval. Two central features of the model drive the results. First, the manager agenda on takeover defenses monopolistically, and shareholders may choose only They reject the managerial proposal. are unable to amend the proposal sets the to accept or to or counter-propose before the vote.-^ Second, the manager's request changes the subsequent decision-making environment. The two critical features, generate the voting monopolistic agenda- setting and the real effect of the request conundrum discussed power of the manager ensures in the Introduction. itself, The monopwlistic agenda-setting properly-chosen request, the shareholders can do no better that, for a than to approve. In certain circumstances, the value of the firm following approval of a managerial request for takeover defenses will be lower than the pre-proposal value. Nevertheless, the value would be no greater, and may be lower, value and the post-rejection value if the request is rejected. The wedge between pre-proposal induced by the real effect of the request, described below. is Previous research on the effects of the implementation of takeover defenses has compared the pre-proposal value of the firm with its post-proposal or post-vote value. In our model, the comparison for inferences about shareholder rationahty in voting is, instead, the value under approval of the request versus the value under rejection. Our model describes the potential acquirers. manager as a The model rests self-interested intermediary between shareholders and on a reduced form specification of an outside managerial labor market, our characterization of the mechanism that produces the real effect of the request for takeover defenses. Other mechanisms change in the may exist with decision-making environment introduce this real effect, since it the same general property of inducing a We find the outside labor market a natural way to defines the manager's opportunities in the event of a successful takeover. In the model, the manager's value added to the firm and utility schedule are knowledge at the time of the vote. The common knowledge assumption obtain even when common is used here to emphasize shareholders are fully informed when recognize the impUcations of their vote: they do not that the voting conundrum can they vote. In our model, shareholders make mistakes or have regrets with respect to their decision. Note, for it to result in a however, change of the vote. This is the time of the vote them worse would not result off. by the proposal employment, but they leam from the proposal. we expect that retaining some known We precisely at the time information asymmetry at the incidence of shareholders' approving measures which, ex We return to this point in the fourth section. change in a to shareholders that manager's utility schedule is clearly extreme, but would increase must by conveyed the stan of the first period, shareholders are not perfectly utility for leam everything, so that they post, leave At in value. informed about the manager's assume that information in firm value if it were possible Similarly, the proposal to eliminate incentive problems entirely using contingent claims contracts. Agenda control issues would not arise: shareholders would hedge themselves ex ante against the possibility of voting empirical observation of the phenomenon, it on value-reducing amendments. Given does not appear unrealistic to assume the non-existence of complete contigent claims maricets. We assume that contracts to hedge against value-reducing behavior by management are impossible or too costly to write. considered further in section One final aspect is worth noting at the outset In the firm will rise as a result of a request for takeover defenses. some circumstances, The empirical measures arc mixed, admitting the possibility measures increase shareholder wealth. the value of results to date that on the some such We describe circumstances in which the takeover defenses arc value-increasing. However, the focus of this paper is value of the firm will fall, 3. is 5. of the model effects of voted anti-takeover This issue to provide a rational choice to describe circumstances in which the answer to our opening question. Model We assume that, by virtue of skill, experience or effort, different managers can contribute different value continuum added to the fixed M, which may be capital structure infinite, possible managers.^ Denote by ra € and investments of the finn. There is a of possible firm values based on the value added by different M the value of the firm under incumbent management at the start of the period of the two-period model, absent the possibility of a takeover.^ first the value of the firm to shareholders will depend on both m and the expected premium to be received in the event of a successful takeover, as discussed below. The firm values under different managers, with density function f(), assumed knowledge. For technical reasons, we assume Of course, is distribution F(-) that lim.j^oo[l-F(01/f(t) be to < <»: on M of common this is not a strong restriction.^ The firm is owned by a set of risk-neutral shareholders, interested only in maximizing the expected value of their shares. the first We presume that there is a corporate charter in place at the start of period, and that this charter does not prohibit dilution of shareholders' claims by potential acquirers." The corporate charter in place at the stan of our optimally to the current manager. This can be true because shareholders to write contracts completely exhausting all model it is is not presumed to be tailored prohibitively costly for the initial possible eventualities. Further, in a diffuse-ownership corporation each current shareholder has little incentive to incur the costs of revising the charter. Heuristically, the model of takeovers we have in mind In the second period, Nature draws a potential alternative whether or not to make a takeover bid for the firm. new manager buys bid, the the firm incumbent remains the firm in this last period Given is the in control. is in the it manager from If a bid is thereafter. If there is Since the Ufe of the firm in the second period, the is no bid or is can choose successful, then the if there is an unsucessful only two periods, the value of first period expected value of the firm all shareholders. let discount factors be equal to one for The expected premium depends on the probability of a successful bidder appearing in the second period, and the expected size of the bidder. In M, who value of the net assets under the incumbent management, m, and the expected second period. Without loss of generality, manager and made and the set determined by the value added by the second period manager. model of takeovers sum of the premium the this and controls for the second period is the following. what follows, we first premium paid by describe the role of takeover defenses, next explain more that fully the incumbent manager's utility for work backwards 3.1 employment and to consider the incumbent's the second period takeover process, and then and shareholders' decisions in the first period7 Takeover Defenses The manager is endowed with two potential bidder's perspective, x minimum price control variables, x and y, both takeover defenses. and y are indistinguishable, and are simply increments From a in the necessary to acquire the finn. Thus, x and y arc measured in the same units as m, reduce the probability of a successful bid appearing, and increase the expected premium, conditional on a successful bid. From the perspective of the incumbent manager and the shareholders, however, x and y are quite distinct. Defense x is anticipatory bidder appearing. X. it The manager Majority voting granted. : among Amendments can be put in place only in the is period, prior to the potential obliged to ask the shareholders to approve anticipatory defense the shareholders determines whether or not the manager's request is to the proposal are not permitted, so the shareholders can only accept or reject the manager's proposal. Examples of voted, anticipatory defenses include staggered board elections, super-majority voting provisions, "fair price" form of amendments first to the corporate charter, amendments, "poison pills" that take the and establishment of classes of stock with differential voting rights. For convenience, direct costs to the However, the manager bears an whether the request is manager of implementing x are normalized to zero. indirect cost for requesting anticipatory defense, re gardless of granted. This cost is in tenns of his outside value ~ the utility payoff he can expect to receive in the second period, conditional on being ousted from the firm. Our presumption is that efforts on the part of incumbent managers explicit offer indicate that the manager is, to erect takeover defenses in the for instance, more interested in personal security than the welfare of the shareholders; and this lowers his outside market value. Let level of anticipatory defense x. If absence of an X denote a requested X is approved by shareholder vote, then the implemented level of 8 this defense however > 0] is x = X; and that the if it is not approved then x above discussion impUes = 0. Evidently, that the states of the world X= [x implies x = 0. Note = I X = 0] and [x = I X are distinct: this is crucial to our model. The second type of takeover defense available to the manager, y, is responsive . Responsive defenses are implemented only to fight an expUcit takeover bid in the second period, conditional on a bid materializing. Unlike anticipatory defenses, the manager implements responsive defenses without obtaining shareholder approval. Examples of non-voted, responsive defenses include anti-trust lawsuits, targeted repurchases, "poison pills" private placements of stock with parties friendly to implemented by the board of directors, incumbent management, and counter-offers to acquire the would-be acquirer. The incumbent defenses. These As bears direct costs, measured in units of his utiUty, for implementing responsive described below, the manager receives direct costs of firom diverting time engaging and effort in a takeover battle from his usual utility may be managing from the activity of managing the firm. thought of as the disutiUty he receives activities into the less-preferred activity of fighting a takeover. The distinction we make between anticipatory and responsive takeover defenses, that the former are voted by shareholders while the latter are not, is purposes, and does not do violence to empirical reality. an abstraction. From 1968 It is a useful one for our to 1979, the Williams Act required that a takeover bid be outstanding for at least ten business days. Currently, SEC's Rule 14e-l, adopted in 1979, sets this minimum However, voting decisions by shareholders time for a takeover bid arc governed by at twenty business days. state corporation law requirements regarding disclosure and maiUng of formal proxies. Delaware General Corporation that notice be mailed to shareholders of record ehgible to vote at least Law requires twenty days prior to the meeting. While these requirements do not preclude the use of voted defenses as responsive tactics (the Martin Marietta - Bendix case most voted defenses are put is in place a counterexample), they surely when no make it more explicit takeover threat exists." difficult. Indeed, 3.2 The Second Period: The Incumbent's Utility for Employment We suppose that there is no utility yalue to making a bid per se, so that, giyen the value m of the finn under the incumbent manager, only managers with values n make takeover bids. If there is any administrative cost of the firm's shares), then no manager n < to bidding m will make an offer. > m find (i.e., it worthwhile to a cost other than the price Hereafter, we assume such a cost, but, to avoid cluttering the notation, suppose the cost constant and take in m. Hence, n > m is The condition n > that the m is not sufficient for a successful takeover bid because of the possibility the potential bidder's value n y, to be implicit a necessary, but not sufficient, condition for a successful takeover bid. incumbent will erect takeover defenses, as described above. defenses x and it there is must exceed If takeover defenses are used, m by at least the amount of takeover defenses. Given a takeover bid must be at least equal to [m+x+y] in order to win. Will any more be paid? Because shareholders are interested only in maximizing wealth, and given that dilution available to the acquirer, the it is sufficient to incumbent No acquirer wishes limit epsilon will shareholders [x+y]. Hereafter, denote the incumbent is more than [m+x+y] pay any more than necessary be zero. Hence, the premium above — conditional on To derive to pay e > the expected to persuade to take to remove over the furo, so in the m that the incumbent secures for the losing a takeover battle despite defenses x and y premium by them is — is precisely n. premium, consider first surely in control of the firm in the first the incumbent's objective function. period, and the investment structure of the enterprise are taken as given. Therefore, it is The and financial necessary only to examine his second period payoff under two possible employment alternatives: within the firm, or, if a bid succeeds, outside the firm. If there is utility V(m) no takeover bid fi-om in the second period, the managing the responsive defense y, as firm. If there is incumbent stays a takeover bid, the in control and receives manager may choose to erect described above. Let c(y) be the direct costs of implementing responsive 10 defense is assumed increasing convex y, convenient to assume lim.y_^o ^'(y) = Because of these is in y, c' costs, the incumbent's ultimately unsuccessful is v= V - ^' > 0, c" > 0, with c(0) = "°"^ °^ °"^ principal second period 0. results For technical reasons, depend on it this.^ he must fight a takeover bid which utility if c. In the event of a successful takeover bid, the manager's second period utility is the utility he receives from his takeover is (o(x, y X) = W(7:, X) I is takeover premium he diminishing marginal is We assume the is able to extract rate. As from the acquirer, Wj > 0, is increasing concave in the Wj j < 0, he is rewarded at a described above, the manager suffers a decline in outside value, when no explicit takeover threat exists. W2 We assume ^2 = -«>, so that requests for guaranteed job security result in extreme loss of outside We further assume the cross-effects of the decline due to the request on the increase due to value. extracting the if employment ousted from the firm. His outside value for requesting anticipatory defenses lim.^-^oo premium are non-increasing, Wj2 ^ 0. That is, they are not then they dampen, not amplify, each other. supj)Ose that at the stan of the employment in the firm to his outside alternative, enough approved requests W is period, before any request first The above assumptions imply that c(y). rewarded by the outside labor market for extracting a higher premium for shareholders, given he but - the manager's gross outside value, or utihty for alternative incumbent < 0, alternative outside the firm, net of the costs borne in engaging in the battle: (1) W employment (x manager of requesting two effects To make is V(m) > W(0, the in strictly decreasing in x. may be problem separable, nontrivial, we made, the manager prefers 0). that the manager's utility for alternate = X), everywhere decreasing the employment is, for large This does not exclude the possibility approved requests. This decrease is the indirect cost to the anticipatory defenses, mentioned above. Conditional on anticipatory defenses x, the manager can raise his outside value in the event of a successful bid by using responsive defenses y to increase the premium received by shareholders. 11 Given the anticipatory defenses in place after the first pericxl, X > and x e {0, X } , and observing a bidder with value n, the incumbent chooses responsive defense y in the second period to his utility. Define y* as the response maximize bid is successful and the y*(x X) = argmax.y ^ r^ (2) From I we obtain (1), doi/dy (4) 82(o/9y2 da/dy = first- Wj = (3) Setting manager leaves which maximizes the incumbent's net utility if the the firm: 0)(x, y X). I and second-order conditions: - c'; = Wii-c"<0,Vy. implicidy defines y*. By (4), y* is unique for all X, x > 0. We now define the incumbent's net utility from fighting and retaining his job, that bid which Observing a bidder with value n ^ fails. is incumbent's net utility v(n, x) The manager will fight to co. so increases his outside The Second - Once utility; CO his net utility is he chooses y > y = n-m-x. Thus v, the as a function of n and x: that doing co. Period: Takeover Bids if any bidding x and y are fixed and known, only a potential bidder with value n > offer. cost, Moreover, only bids which can succeed will be made, as and no utility value to making unsuccessful bids. At the is indeed known. However, since responsive, the incumbent's selection of y depends on the particular n drawn by Nature in the second period. schedules, if dominates v, however, he will fight only to the extent beginning of the second period, the anticipatory defense x defense y win from employment with the firm v exceeds then he will leave to collect [m+x+y] can make a successful is will fighting a c(n-x-m). win as long as We argued above that, long as there is from fighting an unsuccessful bid can be written = V(m) his net outside utility 3.3 incumbent cosUy, the incumbent's best winning response n-m-x. Since y (5) m-i-x, the is, W, V, and c, are By the second pericxl, common we suppose the incumbent's utility and cost knowledge, along wiUi his contribution to firm value m. 12 made her draw, Likewise, once Nature has the value n of the potential acquirer is knowledge. Therefore, the potential acquirer bidder's best offer (that acquirer makes no offer at response (m-f-x+y < Though the fighting, c, n), m+x-Hy > n), then then she makes by surely, employment is sufficient to because of the bidding cost ~ beat the the potential capable of topping the incumbent's best against the incumbent inside the firm, V, or it is utility, not. In in W* = win control of the firm. I defined by examining the scenarios prior to the takeover W(x-t-y*, X), is either less than his utility what follows, we discuss the process which each of these two cases in turn. Formally, denoting the bid which can succeed by n, and defining CO* [inf.n is The manager faces one of two (O. determines bids and responsive defenses (6) the manager's definition of credibility he prefers to bear the costs of His highest attainable gross outside minimum is bid, and the winner pays the premium n = x+y. manager's choice between v and for when n the smallest offer necessary to The minimum bid which can succeed battle. - If the potential acquirer is all. incumbent loses if, y.^^ If this utility-maximizing response is is, if capable of calculating the incumbent manager's A response y to a bid n is credible credible response y to any bid. utility-maximizing response is common = W* - c(y*): ifW*>V; V = CD], V = CO*], otherwise. n(xlX) ={ [inf n In the first case, that allows the where manager to I W* > V, the minimum bid which can succeed will be below the one attain his maximum outside utility. To see this, note that at the beginning of the second period, absent a takeover bid, the incumbent prefers his job inside the firm to outside for employment, or V > W(x-K), X). ^ some y between zero and prefers outside employment to Thus, a bidder whose value n illustrated in Figure y*, Since by hypothesis ^ V = W or equivalently v = co. employment with is slightly Above it must be this point, the that, incumbent the firm, and so prefers to surrender control. above the point where v = 1(a). [HGURE ABOUT HERE] 1 V < W*, then (O can win. This case is 13 premium paid by In this case, the the successful bidder may depend on the bidder's value n. Define n* to be the bid correspx)nding to the incumbent's highest attainable outside utiUty: n*(x X) (7) I where y* first let is n € =m+X+ y*, defined as above. Note that n* (n, n*]. The incumbent's >n co: this is and only best response willingness-to-pay for the firm by choosing y outside value, if is to y = y* and then leaving the firm Consider Figure and 1(a) fight to extract all the bidder's = n-m-x, and then to leave the firm to collect his net Observing is, the bidder does not the firm. Although the firm may attain manager's value added, the new manager is assumed that the anticipatory defenses in the this, the incumbent optimizes by choosing to obtain his outside value, (O*. pays only n* to acquire the firm. That chosen by the manager W* > V. clearly credible. Now let the bidder's value n exceed n*. presume if Any pay bidder with value n > n* all that the value n in the second period x put in place second period, set the to consume by is willing to pay for virtue of the the surplus, n in the first period, minimum she - new n*. That we is, along with the responses acquisition price but do not prevent acquirers from capturing benefits above this price. Now consider the other case, where W* < V. Here the manager is willing to undertake costly responsive defenses y until the cost of doing so drives his net utiUty from continued employment with the firm below his the incumbent prefers alternative fight maximum employment which nets him co. attainable outside utility. In Figure 1(b), observing any n in the firm, for Thus, for any n < n, which he receives is a, will n, outside the manager's utility-maximizing action and defeat the bidder. Any successful bidder, with n > value added. This utility v, to his < is to pay n* regardless of individual true because the incumbent, observing a bidder who can contribute a firm value of n > n, recognizes defeat, but chooses y = y* to maximize his personal outside value. From value the above discussion, the premium n paid by the successful bidder of value m contributed by the incumbent, can be defined: (8) 7t(n, X I X) = min.[n - m, n* - m]. n, above the 14 Inspection of (6) and Figure and win. On 1 reveal that, for bids n the other hand, any potential acquirer < n, credible that the incumbent will fight it is who can bid n >n will be successful. Therefore, the probability of a successful bid appearing in the second period premium she will pay, described by depends on n and on the incumbent's (8), [1-F(n)]. is The utility function. Together, these two pieces of information determine the expected second-period premium, evaluated by shareholders and the manager in making their 3.4 The first period choices. Shareholder Voting on the Incumbent's Proposal First Period: We now specify the actions of the incumbent and the shareholders in the first period, which are based on their expectations about the second period. Viewed from the expected second period payoff is a function of anticipatory defenses utility fhjm retaining firom alternative his job times the probability employment times x. period, the incumbent's This expected payoff no bidder exceeding n appears, plus is his his utility the associated probability of a successful bidder appearing: U(x X) = F(n)V + 8[F(n*) (9) first I - F(n)]-! (o(x, n-m-x X)-f(n)dn I Q + {5[1-F(n*)] + [l-5][l-F(n)]}cD(x,n*-m-x where 5 = 1 iff n < n* (equivalently, iff incumbent's best response The first V < W*), and 5 = may depend on otherwise. the bidder's value n if period expected value of the firm under IX), As described above, the V < W*. management m is the sum of m and the expected premium: (10) S(xlX) = m + 6[F(n*)-F(n)]|[n-m]-f(n)dn + {5[1-F(n*)] + [l-5][l-F(n)]}-[n* - m] with 6 defined as above. Since anticipatory defenses x require shareholder approval, the manager's optimization problem is to choose a request, constraint of shareholder approval: first period X ^ 0, to maximize his expected utility, subject to a 1 15 max. U(xlX) (11) subject to: The S(X X) ^ S(0 I I X). constraint says simply that shareholders will vote rationally if the manager asks However, if X> for no anticipatory defense, is no it is X = 0, then the constraint trivially binds. less than its value if the request is rejected. manager's outside value in the second period if request. Clearly, then the constraint says that approval of the request must result in an expected value of the firm that even on the manager's not granted - we have in is Because the incumbent reduced by his request, X, in the general that S(0 I 0) t^ S(0 X) for any I first X > 0. period In other words, the alternative that rational shareholders compare against the manager's request 0), their situation prior to notion is any request, but central to the results If the constraint binds at rejecting the request. Since on is the voting some S(0 X), the result I Appendix is not S(0 they reject the request. This conundrum. X > 0, then shareholders are indifferent between accepting and managers are not indifferent, they circumstances by slighdy perturbing the request, X, to induce shareholders: as if - A shows, such a perturbation is can insure acceptance strict in such preference on the part of always available. From a game-theoretic perspective, this induces a unique subgame perfect voting rule for the shareholders (Banks and Gasmi (1987)): vote for the management's proposal with probability one S(X X) > S(0 X), and vote I I against the proposal with probability one otherwise S(0 X)). Therefore, without loss of generality, I we can if (i.e. and only S(X X) < I adopt the convention that in cases of shareholder indifference, shareholders always "vote with the management". We turn now to some results. ^^ 4. Results We now analyze the voting conundrum discussed in the Introduction in the context of the model described in section 3. The manager requests shareholders vote on the request anticipatory defenses in the first period, and The outcome of the vote can if affect both the probability of a 16 takeover bid appearing in the second period, and the premium shareholders receive if a takeover bid succeeds. In pan, these effects are driven by the fact that the manager has responsive defenses available to counter a particular takeover bid, and his utility maximizing use of responses will be affected by the outcome of the second period all cases, the probability of a bid appearing in the a result of a request for anticipatory defenses, and in falls as firm to shareholders period. In first is lower if they reject the proposal than been requested. The value of the firm to shareholders less than than if the it would have been value under acceptance is if if it all cases the value of the would have been if they accept the proposal no defenses had may be greater or no defenses had been requested. The voting conundrum occurs larger than the value under rejection, so rational voting implies acceptance; but the value under acceptance lower than is it would have been if no defenses had been We provide a sufficient condition for this to occur. requested. There are three possible conditions beginning of the second period: (1) the manager at the requested no anticipatory defenses, (2) the manager made a request which was voted down by shareholders, and (3) the manager's request was approved. Proposition the second period response to a takeover bid which maximizes the manager's outside value, puts an ordering on y*, 1 conditional on the first-period outcome. Proposition Corollary 1: 1: y*(0 0) > y*(0 X) > y*(X X) > I I X + y*(X I than if 0. X) > y*(0 X). I If a first-period request is rejected strictiy larger I the request is by shareholders, the second-period maximizing response approved. However, from Corollary 1 , total is defenses (anticipatory plus responsive) are smaller if the request is rejected. Since total defenses determine the premium received premium if in a successful bid, this implies that shareholders can expect a smaller they reject the manager's request and a successful bidder emerges. 17 An interesting response is feature evident from Proposition never larger after a request, even if the 1 is that the second-period request is denied, than Thus, the request carries no implicit threat of "scorched earth" responses Rather, as premium we show if below, the request alters the if if maximizing no request is made. shareholders deny it manager's incentives to fight for a higher a successful bidder appears. Anticipatory and responsive defenses are substitutes in their effects on n, the required to acquire the firm. shareholders in the Comparative first minimum Generally, the higher the level of anticipatory defense approved by period, the lower will be the manager's choice of responsive defenses. on y*, the maximizing second-period response (Appendix A, statics bid (a3) - show (a7)) that this response is non-increasing in the size of the first-period request. how In evaluating to vote, value-maximizing shareholders consider both the size of the premium they can expect from a successful bid, and the probability that such a bid will emerge. Proposition 2 describes the probability of a successful takeover bid, as a function of requested anticipatory defenses and shareholder approval of the defenses. Proposition 2: X) increasing in X, x e {0, X}; (a) n(x (b) VX>0,n(XIX)>n(OIX)>n(OIO). I is strictly Proposition 2 states that the minimum bid which can succeed against incumbent management increases with the level of anticipatory defense requested. Since the probability that a successful bid will emerge, [1 - F(ii)], falls as the minimum successful bid rises, pan (a) of Proposition 2 implies that the probability of takeover declines with increases in the level of requested anticipatory defenses, whether or not these are approved. Part (b) implies that the probability of takeover largest if if the no anticipatory defenses request is approved Notice are requested, declines if there that part (b) is not an is a request, and may immediate consequence of part is decline (a): the more 18 change leave from in state it" [x = X X > 0] I = X > 0] is not incremental because of the "take 1 or it nature of the shareholders' decision. The value successful bid to shareholders is from anticipatory takeover defense depends on both the premium made, which generally increases with the probability of a successful bid, are to [x level of anticipatory defense, which generally decreases. Proposition 3 shows unambiguously worse off if they reject if a and on the that shareholders management's request for anticipatory defenses than they were before the request ForX>0, Proposition 3: S(0 0) > S(0 X). 1 I Proposition 3 follows directly from part (b) of Proposition Proposition 1. The will be is result manager's proposal, because the now higher. Moreover, no higher than effects are driven along with the first inequality of probability that a successful bid will appear in the second period after rejection of the incumbent 2, it for any successful bid the would have been by the manager's minimum if is lowered bid which can succeed against the premium received by shareholders no request had been made, and may be lower. These utility for employment outside the firm, which declines as a of the request for anticipatory defenses. The voting (incentive compatibility) constraint in the model requires that approval of a request leave shareholders at least as well off as rejection. Proposition 3 states that shareholders are worse off if they reject the request than they latter circumstance, no request, To complete our description leave them worse off, shareholder value if is would have been if no request had been made. (Note not available to shareholders deciding how to vote that this on a request.) of the apparent paradox that shareholders vote for defenses which we require a comparison of shareholder value if the request is approved no request is made: S(X X) versus S(0 condition for the apparent paradox to occur. I 1 0). with Proposition 4 gives a sufficient 19 Proposition 4: V(m) < W(y*(0 3X => According to Proposition his maximum 1 0), 0) e (0, oo): 4, if the S(0 0) > S(X I manager's I X) > S(0 X). I utility fix)m employment with the firm is less than gross outside utility before any request for anticipatory defenses, then requests exist which shareholders will approve, but important to observe that this which leave them worse off than they were before. maximum gross outside utility is available only if, in the period, Nature draws a bidder with value n was > n*(0 I 0). Recall that we assumed from employment utility before defenses x or y are considered, ftoposition 4 concerns the incumbent's gross outside utiUty: that in Figure 1(a), by setting x With Proposition which is, 4, we have shown who are fully worse off by them. An X = 0. is, maximum The simation is we hoped informed artifact at the to describe. That is, rational, value-maximizing time of the vote, will approve the defenses but are of our assumption of identical, risk-neutral shareholders and holders of large blocks may well-diversified portfolios. pictured the existence of requested levels of anticipatory defense The differ made is that all votes will be unanimous. Empirically, shareholders will differ: for example, managers shares, his X = 0. result in the voting behavior shareholders larger than his initial gross outside utility: that considering his optimal choice of y at = second the incumbent's utiUty in the firm It is who own from holders of small positions who have results will apply to non-identical shareholders as long as the pivotal voter is risk-neutral and value-maximizing. The sufficient condition for the result in Proposition current and alternative employment In 4 depends on the manager's utihty Appendix B, we construct an example of a manager with utility satisfying this condition, whose optimizing choice of anticipatory defense by shareholders, and them worse off than in will leave no sense pathological, and demonstrates requested by managers with in if be approved no request had been made. The example that these levels utilities satisfying the will of anticipatory defense could be antecedent of Proposition 4. is 20 It is worth noting that under some circumstances, shareholders can be made better off by implementing some anticipatory defense. By Proposition 3, a necessary condition for this to occur S(X X) > S(0 is that the manager's request, X, be strictly interior to the constraint set; is not, however, sufficient In our example in Appendix B, shareholders would like some anticipatory defense, but the manager prefers more and, despite i.e. I 1 X). It the constraint not binding, the request results in a decline in shareholder wealth. We remarked in section 2 that, in order for the stock price to fall as a result of a request, some information must be conveyed by the request. The information conveyed concerns the manager's from employment outside the utility firm. In the circumstances described clear that any probability weighting of S(0 revision place when some 0) with prior probability on the X> outcome it is S(X X) or S(0 X) implies a downward I I that the manager will prefer no which are associated with a drop The primary implication of our model for inteipreting the empirical shareholder wealth before and after the vote. as an unambiguous drop anticipatory defense, in the value of the firm. inferences about whether shareholders vote rationally cannot be management 4, the first possibility is eliminated. In these circumstances, as long as shareholders there exist requests is that 1 by Proposition work which has preceded made from it a comparison of We describe the alternative to voting with in shareholder value, n^t as a return to the pre-proposal status quo. Therefore, pre-proposal shareholder wealth is not the correct benchmark for determining shareholder rationality. Our model suggests that empirical investigation of shareholder voting and takeover defenses should consider the manager's outside employment opportunities, and the manager's 5. skill relative to potential bidders. Conclusion We have provided a rational choice model of shareholder voting on anticipatory takeover defenses. In our model, it is feasible for informed, value-maximizing shareholders to approve measures which leave them worse off than they were before the measures were requested. What 21 drives this result is that a manager, in requesting such measures, lowers his outside market value. Consequently, the manager's optimal response to an actual takeover bid rejected than An open he had made no request Shareholders recognize if question devices, and allow is him why is different if the request this in evaluating shareholders allow the manager to act first how is to vote. to request anti-takeover discretion over the level of responsive defenses. Shareholders will not generally want to prohibit the use of responsive defenses as we have characterized them. This is true because, although their existence reduces the probability of a successful bid appearing, responsive defenses allow shareholders to receive higher premia from some bidders. In fact, if a bidder appears with sufficiently high value added (n > n*), shareholders would like the incumbent to use more responsive defense than he shareholders cannot will know ex ante which choose to maximize his own utility. bidder will appear in the future, and Since may have difficulty discerning the utility schedules of current and future managers, prohibiting takeover defenses The preclude value-increasing actions as well as value-reducing ones. empirical fact that managers are given control over voting agendas, and are not prohibited by shareholders from engaging in responsive tactics suggests that complete contingent contracts are costly to write. However, offer as an observation that recent charter targeted repurchases, which, if may implemented, Finally, we offer is amendments which may we prohibit the use of "greenmail," or be viewed as shareholders' prohibiting a particular responsive defense unlikely to result in a higher two remarks. First, it is premium for them. frequently asserted (cf. Easterbrook and Fischel (1983)) that the alienability of ownership claims protects shareholders firom detrimental management entrenchment However, unless shareholders tactics. takeover defenses by management, they cannot SEC proxy mailing requirements demand proposal date. any drop Once the proposal in share value is sell anticipate the proposal of a voting claim in response to the proposal. that the record date for shareholder voting announced, the constituency is fixed. precede the Our model suggests should occur with the announcement of the proposal, not with the vote. Second, our model implies that changes in shareholder wealth associated with voting on that 22 takeover defenses may be positive or negative, depending on the manager's value added and schedule. In cases where shareholders' wealth the manager's request If there positive. is is reduced, the reduction should occur at the date of any detectable change at the date of the vote, This follows from Proposition 3 and the voting constraint: the shareholders' wealth is unambiguously reduced shareholders' wealth under rejection interpretation, empirical results is no utility if it should be first states that they reject a request; the second ensures that larger than their wealth under approval. On this which find shareholders' wealth declines following implementation of anticipatory defenses are not evidence of shareholder irrationality or ignorance, but reflect an informed choice of the lesser of two declines in value. 23 Appendix A: Proofs Proposition X > 0. Let 1: Then: y*(0 0) > y*(0 X) > y*(X X) > I 1 I 0. Proof: (i) y*(X X) > follows from assuming I (u) y*(0 ) 1 Wj c'(y) = 0. I I Wi(y*(0 I I 0), 0) - Wj Wi(y*(0 X), X) > 1 j < 0, W ^2 ^ and X > 0: 0. (3) to obtain: Wi(y*(0 (a.l) - 0), 0) I Wi(y*(0 X), X) = I c'(y*(0 I 0)) - c'(y*(0 I X)). implies the RHS(a.l) < 0: contradiction, But c" > (iii) everywhere, and lim.y_^ > y*(0 X) Suppose y*(0 0) < y*(0 X). Then, by Use > y*(0 X) > y*(X X) I I Suppose y*(0 X) < y*(X X), and again use the I I WjPC + y*(X (a.2) Then But X Remark > c" > 1: I X) implies RHS(a.2) 0, so that W j | < I 1 I 1: Let X > 0. I Wi(y*(0 X), X) = c'(y*(X X)) - I I > iff - c'(y*(0 I X)). 0. implies LHS(a.2) y*(0 0) = y*(0 X) [X + y*(X X)] > y*(0 Corollary X), order condition (3) to get: first W12 = 0. < 0: contradiction. X> Hence, x = and II W12 = imply: 0). Then: [X + y*(X X)] > y*(0 X). Proof: Use (3) and Proposition I I 1 with Wj < and c" > j 0. II Comparative Statics: (a.3) dy*(X X)/dX = -[W^ j+Wi2]/[Wi (a.4) dy*(0 X)/dX = -Wj2/[Wi I I j-c"] < j-c"] 0, < 0. with the inequality strict iff W12 < 0. 24 (a.5) dco(X, y (a.6) dco(0, y Note Since n*(x X)/dX = W2 < 0. y, I (a.5) I < ^ I X. x= If (a.6) 0, then I from Let X) is X > 0, x e {0, (a.3) sign. is differentiable in given by is (a.4). and collecting terms, as c"(y*(x I ambiguous I. an*(0 X)/9X 9n*(x X)/9X ^ (<) 1: a priori has I Substituting I (a.7) + W2, which I + dy*(X X)/9X. Lemma W X) = x + y*(x X) + m, and y*(x X) I differentiable in 1 I any that at X)/dX = I I I is I is x = X, then 9n*(X I X)/3X we obtain: X)) ^ (<) -Wi2(x+y*(x X} Then: n(x X) . If X, n*(x X) differentiable in I X), X). X; and an(x X)/dX > I 0. Proof: (i) n(x I differentiable. from section First, recall A* = [v(n*(X I X), X) A* with Differentiating - 3 that n(x I y*(X X) co(X, I respect to I A* can change Suppose there V(m) < So by case a (6): : By exists an (>) I X Consider (6), In this case, X)]. Then X) iff X n*(X X) iff X < (>) X. I X), I < X is unique and: (>) X. < X- n(X X) is I n(X X) I X > 0, and define: 0. X such that A* = 0. (>) Let x = sign at most once as x increases, and only from negative to positive. W(X+y*(X n(X X) < I X ^ 0. X at x = X, and using (3) gives: aA*/ax = -W2(X+y*(X X), X) > Therefore, X) > x+m, Vx, impUciUy defined by: V(m) is differentiable because W - W(n(X X)-m, X) = 0. I differentiable. In particular, is (a.8) dn(X X)/aX = -W2(n(X X)-m, X)AVi(ii(X X)-m, X); (a.9) lim.x^ x- I I VX I and, case p : Consider X [^Il(-'-)/9X] > X. = -W2(n*(2L 20-m, 2DAVi(n*(X X)-m. 1 I X). < X, 25 By (6), n(X X) t Differentiability of is impliciUy defined by: n(X X) follows from I an(X X)/dX = (a.lO) [c'(n£X I V(m) - c(n(X X)-X-m) I differentiability X)-X-m) I where Wi*(X) = Wi(n*(X X)-ni, X), I i =1, of c and W. y*(X X) X) = 0>(X, - I I 0. In particular, using (3): Wi*(X) - W2*(X)]/c"(n(X X)-X-m), - I Hence, 2. lim.x_^x+[^(')/9^1 = (a.ll) [c'(n*(X 1 2D-X-m) - Wi*QD - W2*(2D]/c'(n*(X I X)-X-m) = -W2(n*(X X)-m, 2DAVi(n*(X X)-m, 2D; I the second equaUty follows differentiable The proof of this case is immediate from I Together, (a.9) and (3). ) complete an(x X)/9X > I case case a [3 x = : : x= X<X (a.lO). By > c" I Moreover, Wi*(X) = (3), X) < n(X X). Therefore, (a.l2) (a.8) and (a.9). X>X Consider c'(y*(X I X)) = c'(n*(X X)-X-m). Since A* > I Wj+W2 < Wj and (a. 10) is strictly I X > 0. Similar reasoning as before gives n(0 follows on implicit differentiation of (6) for Remark Lemma I 2: (a.7) and 1 imply n(X X) > n(0 X); I I Proof: a V(m) < W(X+y*(X I positive. Since c' > 0, Lemma for x = X. and 9n(0 X)/9X > : n*(X V y implies, Therefore, the numerator of . completes the proof of the Lemma 2: 0, c'(n(XIX)-X-m)>Wi*(X)>0. Now suppose x = case (a. 1 1 everywhere when x = X. is (ii) this from another application of X) the argument that n(x I X), X) that U(x X) and S(x X) I I I X) differentiable in X, X < X and for X > X. are differentiable in X. II ) 26 By (6) and the premise, V(in) = W(n(X X) - m, I V(m) < W(n*(0 X) - m, X) => V(m) = W(n(0 X) - m, X), by (a.l) I I => X). There are then two possibilities: (6) W(n(X X) - m, X) = W(n(0 X) - m, X) I I => n(X X) = n(0 X). I I (a.2) V(m) > W(n*(0 X) => V(m) = m/X) y*(0 X) X) + c(n(0 X) [0)(0, I - I definition of y*(0 I I W(n(X X) - m, X) => By - I cCn(0 X) I X), co(0, y*(0 I Therefore,W(n(X X) - I m, X) I c(n(0 - m) = - - m)], by (6) (o(0, y*(0 X) I X) X) > W(n(0 X) I I X) I - m) > W(n(0 X) - I X). I m, X) - c(n(0 - m, X) - I c(n(0 X) I - X) m). - m) => n(X X) > n(0 X). I I This proves the proposition for case (a). case B : V(m) > W(X+y*(X X), X) I => V(m) = co(X, y*(X X) X) + c(n{X X)-X-m), by I By Corollary 1 and X > 0, W(X+y*(X I I 13) V(m) = (0(0, y*(0 X) I I W(y*(0 X), X), X) > Hence, V(m) > W(n*(0 X)-m, X). So by (a. (6). I I I X). (6), X) + c(n(0 X) I - m). Therefore, (a. By 14) co(X, Proposition 1 y*(X X) X) I and Corollary co(0, - I 1, c' > is strictly X and x = differentiating (a. 13), we obtain an(0 X)/ax = -W2(y*(0 X), X)/c'(n(0 I I (a. 10), (a.l6) I > - c(n(X X)-X-m). I positive; hence, the RHS(a.l4) must 0, this implies, n(OIX)>nCXIX)-X. Now impUcitly Using I I LHS(a.l4) likewise be strictly positive. Since (a.l5) y*(0 X) X) = c(n(0 X)- m) we have X)-m) > 0. VX > 0: [8n(X X)/dX I I VX > 0: - dniO X)/dX] = + {[W2(y*(0 I I { 1 - X), X)/c'(n(0 [Wi(X+y*(X I X)-m)] - I X), X)/c'(n(X [W2(X+y*(X I I X)-X-m)] X), X)/c'(n(X I X)-X-m)]). 27 By the first term of (a. 12), By CoroUary (a. 1, 16) in {) VX > 0. nonnegative is Consider the second term in {•}. W2 < 0, and Wi2 - 0' > W2(y*(0 17) (a. X), I X) > W2(X+y*(X X), X). I By(a.l5)andc">0. (a. < c'(n(X X)-X-m) < 18) I and c'(n(0 I X)-m). 18) imply that the second term of Together, (a. 17) When X = 0, both terms in (a. {• vanish. Therefore, } (a. 16) in {•} is also positive VX > 0. VX > 0: X [n(X X) I - n(0 I X)] = j [an(r as required (with equality iff Remark 3: from [x = Remark Notice that I X = 0). 8n(0 r)/ar]dr > 0, - I II I X > 0] is I 1 : this is because the change in state not incremental. Remark 2 and Lemma 2 justify S(X X) = S(0 X), then I r)/8r Lemma 2 is not implied by Lemma X X > 0] to [x = 4: Together, I the claim there exists a penurbation in made in section 3 that, if x X -- say, X' -- such that S(X' I X') = X and > S(0 I X'). Lemma 3: For any Proof: Given x = X > 0, n(0 0, v(n, x) I X) > n(0 = V(m) - W2 < 0, W(y, X) < W(y, 0), Vy > 0. Proposition 2: (a) n(x (b) I X) is strictly Vx = X > 0, I 0). c(n-m). Since So by (6), n(0 c' I > 0, v(n, x) is strictly X) > n(0 0), VX increasing in X, x € {0, X); and, n(X X) > n(0 X) > n(0 Proof: The proposition follows from I I I Lemmas 1, 2 and 3. II I 0). > decreasing in 0. H n. By 28 X > 0, Proposition 3: For By Lemma Proof: By Proposition case a : By n*(0 X) 1, By n*(0 1 n: 0))]. 1 0), on a continuous variable a c.d.f. defined with the inequality strict iff W12 < at any y < y*(0 I 0). I (8): 010)^ 7i(n. 7i(n, I : <. I n*(0 X) < n*(0 case p F is n ^ n(0 X). (a-20) If fact that < [1-F(n(0 X))] I I 1 and the 3, [1-F(n(0 (a.19) S(0 0) > S(0 X). n e (n(0 I I X). Vn > 0), the inequality in (a.20) is strict I n*(0 X). I n(0 X)). 0), I an argument in section 3, only managers with value n > n(-l-) will make takeover bids. Therefore: (a.21) 7:(n, casey n < n(0 : By the I 0) I > 7i(n, 7i(n, 010) = Together, (10) and Lemma 4: X) s 0. 0). same argtmient (a.22) I (a. as case 7t(n, 19) I {3: X) = 0. (a.22) yield the desired result. - II lim.x_>oo S(X X) = m. I Proof: By assumption, Wj ^ < > Xi, V(m) > W(X+y*(X S(X X) = I By Remark (a.23) I and lim.x_^oo X), X). Hence, by (10), m + [l-F(n(X 2, S(- •) 1 is ^2 = -««. I X))]-[n*(X I Therefore, for sufficiently large X, say VX > Xj, X)-m]. differentiable in X. In particular, VX > Xj, 8S(X X)/ax = I [l-F(n(X I X))]-dn*(X I X)/dX We prove the Lemma by first showing 3S(X I - {f(n(X X)/9X < I X)ydn(X X)/aX-[n*(X X)-m]). for all finite I I X > X2, X2 sufficiently X 29 By Lemma large. 1 and the assumption strictly positive for all finite case a dn*(l-)/9X < : for for I dn*(X X)/9X > : (7) and By I X))]/f(n(X < therefore, (a. 10), VX > X2. (7) [l-F(t)] = 0, the By CoroUary W(y*(0 X"), X") obtain. I [l-F(n(X" (a.25) By Corollary 1, of X", n*(0 X") > I Vn < n, Vn e Vn > Given (10), (a.25) 7i(n, (n, 3X 1 hypothesis, € I (10). for all (0, 00) : an*(X X)/dX > I 0, Therefore, there exists a Since lim.j_^^ II S(0 0) > S(X I such that both Then by case I X) > S(0 X). I V(m) < W(X"+y*(X" (a.l) of Lemma 2, = [1-F(n(0 X"))] = I an argument of section I X"), X"), and n(X" X") = n(0 X") = I I the premise of the Proposition 3, I I 7i(n, 7t(n, 7i(n, X") = I X" X") = I X" X") > and (a.26) imply I that S(X" 7c(n, I n. Hence: and the choice only managers with value n > n will (8): X" X") = V(m) < [1-F(n)]. I n*(0 X"), From 0), 0) n*(0 X")], I By W2 = -«>. X > X2. I By and lim.x_^j^ 3S(X X)/3X < n*(X" X") ^ n*(0 X"). From n. < that takeover bid. Using this and (a.26) and x = X, X, X2, such X"))] I 1 00. 3X" > 1, Lemma lim.x_>oorRHS(a.24)] = Proposition 4: V(m) < W(y*(0 Proof: m]. Wj j if: I - Therefore, by °°. X, aS(X X)/dX < that Lemma follows from => is of ambiguous sign. imply and y > sufficiently large value of is X)/aX-[n*(X X) for sufficiently large X. 1 I sufficiently large I I Also by assumption, «>. dn(X X)/dX > Hence, For 1. X)) < dn(X I assumption, lim.^^oo [[l-F(0]/f(t)] < lini.x_^oorLHS(a.24)] on the RHS(a.23) {•} X > X2. I [l-F(n(X (a.24) term in X ^ X2. sup.[an*(X X)/3X] = (a.3), 0), the term on the RHS(a.23) (a.7), the first all finite for all I By By V(m) > W(0, X > X2. all Then, 9S(X X)/9X < case B X. that 0; 7c(n, I I X") = n - m; X"). X") ^ S(0 X"). Therefore, I make a 30 C= Let X {X > I = X. There S(X X) ^ S(0 X)} I I are now two cases: pX e C X < oo & S(X (i) 0. ?t : I X) = S(0 X)] I => [S(0 0) > S(X X)], by Proposition I I (ii) [VX e C X : => \3X' € check > < C oo : , S(X X) > S(0 X)] 1), X' < m < n(0 I I oo I & S(0 1 0) > S(X' V(m) < W(y*(0 this last inequality, recall (Proposition 3; 0) < ~ and n*(0 I I I X')], 0), 0) 0) > m. by Lemma 4 and S(0 0) > m. To I by hypothesis. Hence, by II (6) and y*(0 0) 1 31 Appendix B: Example We claim in section 4 of the text that there exist managerial utilities satisfying the sufficient condition of Proposition 4, under which the manager would ask for, and shareholders approve, a level of anticipatory defense, X, such that S(0 1 0) > S(X I X). In this appendix, we justify our claim with an example. In the interests of computational simplicity, violated in the example: The example therefore some technical assumptions of the the support of F is not the viz. shows that the main whole real line, result (Proposition 4) assumptions. Also, there are choices of the distribution function do assumptions of the satisfy all the text, and which are and Wj is model are zero at x = y = does not depend on these F and the outside utility arbitrarily closely W, that approximated by the functional forms exploited in the example. Let F be uniform on manager the closed interval [-2,2], and assume the value added by the incumbent m = 0. is Let V(m)=1.57, W(7r, c(y) For X) = 2-[x+y]l/2 = . [x2]/2 + m, (2/3)y(3/2). this specification, U(-l) is strictly dU(X X)/9X = I incumbent's utility specification, X* = 0, that is, the we function X* the |X)int where incumbent's unconstrained optimizing choice of X. The U(x X) I is as defined in equation (9) of the text. obtain the following results, illustrated in Figure B.l.^^ .35 U(OiO) = 1.3623 U(X* IX*) =1.3766 au(x* x*)/ax = I concave. Denote by o.o Given this For the manager: 0. 32 For the shareholders: S(0 0) = 0.2592 1 S(X* X*) = 0.2582 I S(0 X*) = 0.2570 I so the condition S(0 shareholders. That is, request than they are worse than > S(X* X*) > S(0 X*) 0) 1 1 I by the second if they approve their starting value In this example, at x = S(0 I holds, and S(X* X*) I inequality, shareholders are it, but by the is the outcome worse off if they the final first inequality, for reject the outcome S(X* X*) I is 0). X = 0: aU(0 0)/ax = 0.095 > 0, 9S(0 0)/ax = 0.024 > 0, 1 I and so both the manager and the shareholders prefer a strictly positive level of manager's most-preferred x is strictly x. However, the greater than the shareholders'. [RGURE B.l ABOUT HERE] Notice that, for this parameterization, the manager's unconstrained optimal choice of anticipatory defense, X*, is the manager's unconstrained constrained low enough that the constraint, optimum were approximately optimum X* would be approximately this alternate result .39, by raising V(m), the incumbent's S(X X) > S(0 X), does I I 0.4 or above in this example, then the where S(X X) = S(0 X). (We can obtain I utility for current I employment.) In circumstance, shareholders are indifferent between approving or rejecting. accepted we could management introduce the convention that, Alternately, as described at the Remark 4 of Appendix when this To ensure that X indifferent, shareholders X by e > to is always vote with end of section 3 (and justified more formally A), the manager can reduce his request preference by the shareholders. not bind. If induce strict in 33 Footnotes. 1. DeAngelo and Rice (1983) provide 2. SEC disclosure requirements regarding management mail proxy determining a thorough discussion of these hypotheses. matters put to shareholder vote specify that the materials to shareholders eligible to vote. Thus, the record date for who may vote is prior to the proxy mailing date. The proxy materials include, in the case of charter amendments, the wording of the proposed change and a proxy form with which shareholders Therefore, may vote in absentia, at the proxy mailing indicating approval, rejection, and sometimes abstention. date, both the constituency for the vote This type of voting arrangement, termed "take studied in a political context by 3. We consider infinite M or leave Romer and Rosenthal in it" agenda control, has been extensively (1978, 1979). for this exposition. This Appendix B, we construct an example on a it and the proposal are fixed. is not essential to our arguments. In which the voting conundrum obtains when M is uniform finite interval. 4. We consider the value added by the incumbent to be fixed. Scharfstein (1987) considers circumstances in which the threat of takeovers can induce the manager to work harder, and consequendy, to increase the value of the firm. 5. For example, the uniform, normal, gamma and exponential distributions all satisfy this restriction. 6. Grossman and Hart (1980) have shown be made. Thus, some dilution Generally, dilution is is that if dilution is prohibited, no takeover bids necessary for the circumstances modelled here to be of will interest. possible unless specifically precluded by provisions of the corporate chaner or state corporation law. In the Grossman and Hart framework, our model can be interpreted as describing mechanisms by which shareholders and incumbent managers set the level of allowable dilution and acquirer- sj^ecific acquisition costs. 7. game Although our exposition in is not explicidy game-theoretic, the model which the manager, the shareholders. Nature, and is a sequential potential bidders are players. move It is 34 straightforward to describe the formal model as an extensive-form game. we employ subgame is perfect agents' optimal behavior Nash (in pure strategies), and by working backwards ftx)m the it is this final stage The equilibrium concept which motivates solving of the process to the for first (Selten, 1975). 8. Sec Gilson (1986) for discussion of these and other aspects of takeover law. The role of the assumption 9. level of y will lim.y_^Q that c'(y) = 0, is to insure that some stricdy positive always be chosen by the manager, whatever the value of x. Consequently, use the calculus in our analysis. Without the assumption, we have to consider the we can comer case explicidy: this adds very litUe and does not substantively alter our main results. Similarly, insisting on the cost function being make strictly convex in y is not essential for our results. It does, however, the arguments cleaner. 10. 1 1. The concept of credibility is simply that of subgame perfection: cf fn.7. This follows from the assumption that the incumbent prefers employment with the firm to the outside alternative at the stan of the 12. Proofs of all results are in The period, and that he rationally chooses X. Appendix A. These proofs unique subgame perfect equilibrium 13. first in also establish the existence of a pure strategies to the model. derivations of these results are available from the authors upon request. 35 References Banks,!, and Gasmi.F., Endogenous agenda formation in three-person committees, Social Choice and Welfare 4 (1987) 133-152. DeAngelo3. and Economics Rice,E., Antitakeover charter amendments and stockholder wealth, J.Financial 11 (1983) 329-360. EasterbrookJ^. and Fischel,D., Voting in corporate law, J.Law and Economics 26 (1983) 395-427. Gilson,R. The Law and Finance The Foundation of Corporate Acquisitions (1986), Mineola, New York: Press, Inc. Grossman, S. and Hart,0., Takeover bids, the free-rider problem, and the theory of the corporation, Bell J.Economics, Spring (1980) 42-64. Grossman, S. and Hart,0., The allocational role of takeover bids in situations of asymmetric information, J. Finance 36 (1981) 253-270. Jarrell,G. and Bradley,M., The economic offers, Jarrell,G., effects of federal and state regulations of cash tender J.Law and Economics 23 (1980) 371-407. Poulsen,A. and Davidson,L., Shark repellents and stock prices: the effects of antitakeover amendments since 1980, mimeo. Office of the Chief Economist, SEC (1985). Linn.S. and McConnell,!., on common An empirical investigation of the impact of 'antitakeover' amendments stock prices, J.Financial Romer,T. and Rosenthal Ji., Economics 1 1 (1983) 361-399. Political resource allocation, controlled agendas and the status quo, Public Choice 33 (1978) 27-43. Romer,T. and Rosenthal,H., Bureaucrats vs by direct voters: on the political economy of resource allocation democracy. Quarterly J.Economics 93 (1979) 563-587. Scharfstein,D., The paper (1987). disciplinary role of takeovers, M.I.T. Sloan School of Management working 36 for equilibrium points in extensive Selten,R., Reexamination of the perfectness concept games, International J.Game Theory 4 (1975) 25-55. fomi Figure 1(a) V, 0) V W(X, X) n(X I n*(X X)-m-X X)-m-X I Figure 1(b) V, CO W(X, X) n*(XIX)-m-X n(X I X)-m-X Figure B.l U, S U(XIX) X, h5 :? h 089 X Date Due Lib-26-67 MIT LIBRARIES 3 IDflD DDS 135 fi4D mmmmm f\ •^.. »t¥a&fiiai9MMMuitiatU»f f A 1 'T-Jt-mlSliiinuiJniWfcft! •». imm^mmm