Chapter 10 Corporate Governance 10.1 The market for corporate control The agency problem that arises from the separation of ownership and control (Berle and Means, 1932) has been a major focus of the literature on corporate finance and the theory of the firm over the last twenty years. Various insti- tutional arrangements exist to deal with this agency problem and one that has attracted a lot of attention is the market for corporate control. Manne (1965) suggested that, if a publicly traded company is badly managed and the usual methods of corporate governance (board of directors, proxy bat- tles, etc.) are not e?ective in disciplining the management, a hostile takeover allows an outsider to acquire a controlling interest in the firm, change the management, and realize an increase in shareholder value. Grossman and Hart (1980) provided a formal analysis of how the mar- ket for corporate control functions and pointed out the existence of a free- rider problemthat may prevent takeoversfrommaximizingshareholdervalue. Here is a brief sketch of the model. The manager of a firm chooses an action a ∈ A and the resulting value of the firm is denoted by V(a). Suppose the firm is under the control of an incumbent manager who for some reason (e.g., incompetence or private benefits) is not maximizing shareholder value. The optimal action is a ? but the manager chooses ˉa. If a raider acquires control of the firm and changes the action from ˉa to a ? , social surplus increases by V(a ? )?V(ˉa) and this gain in surplus can be shared between the raider and the shareholders. A ‘hold-out’ problem arises because the existing shareholders anticipate 1 2 CHAPTER 10. CORPORATE GOVERNANCE an increase in value if the raider successfully takes control of the firm. The shareholders will be unwilling to tender their shares unless they are paid the full anticipated value. If takeovers are costly, the raider will undertake a takeover only if he anticipates a positive profit. But if the raider has to pay the full price he gets no profitfromthetakeover. To make this argument precise, consider the following game form: ? The raider o?ers a price p for the shares of the firm and pays a fixed cost C>0 (the cost of organizing the tender o?er). ? Each shareholder has a single share, which he can tender or retain. ? If the raider acquires a fraction 0 <γ<1 of the shares, he acquires control and can choose the action that maximizes the value of the firm. If the fraction of shares tendered is less than γ,theo?er fails and the incumbent management is left in control. The equilibria of this game can be characterized by looking at these stages in reverse order. ? At the last stage, if the raider has acquired a fraction g ≥ γ of the shares, he gets control, chooses the optimal action a ? , and the value of the firm is V(a ? ). If he acquires a fraction g<γ, the incumbent manager remains in control, the firm’s policy is unchanged, and the value of the firm is V(ˉa). ? At the second stage, the shareholder receives the price p if he tenders his share. If he holds onto his share and the o?er fails, his share is worth V(ˉa). If he holds onto his share and the o?er succeeds, his shareisworthV(a ? ). Thus, he will tender his share if p>V(ˉa) (resp. p>V(a ? )) and hold onto it if p<V(ˉa) (resp. p<V(a ? )). ? At the first stage, the raider must o?er a price that equals the share- holders’ reservation price to succeed. Thus, the o?er can succeed only if p ≥ V(a ? ). If the raider acquires a fraction g ≥ γ of the shares, his profitis (p?V(a ? ))g?C<0. 10.1. THE MARKET FOR CORPORATE CONTROL 3 So it appears that a takeover cannot succeed. Grossman and Hart suggest that dilution of the existing shareholders’ property rights may provide the raider with su?cient profit to undertake the raid. Suppose that the dilution ratio is φ,thatis,theraidercancapturea fraction φ of the minority shareholders’ property rights by self-dealing, etc. Then the price o?ered must satisfy p ≥ (1?φ)V(a ? ) and a successful tender o?er is possible if φV (a ? ) ≥ C. Bagnoli and Lipman (1987) point out that the Grossman Hart model with a continuum of shareholders is special. Abstract: We noted at the outset that most of the literature on takeovers assumes atomistic stockholders. As we pointed out, however, there are many large firmsforwhichthisassumptionis obviously inappropriate. This led us to consider the finite stock- holder game. We showed that there are substantial di?erences between the finite game and the atomistic stockholder models. In particular, because some stockholders must be pivotal and hence cannot free ride, successful takeovers are possible without exclusion. Since the equilibrium outcome in the finite stockholder game is quite di?erent from the atomistic stockholder outcome, the natural question to ask is under what conditions the atom- istic stockholder outcome obtains for firms which are su?ciently widely held. We showed that the atomistic stockholder outcome does not obtain in the infinite stockholder game. We also showed that the di?erence between the finite and atomistic stockholder outcomes may not vanish in the limit. We argued that atom- istic stockholder models may provide a reasonable approximation to the outcome for takeovers with any-and-all bids if the firm is not su?ciently valuable relative to the dispersion of stock owner- ship. Otherwise, the finite stockholder model is likely to provide a more accurate prediction, so that exclusion is not necessary for successful takeovers. Since, all else equal, stockholders generally benefit more from takeovers without exclusion, our analysis sug- gests that stockholders would prefer to invest in firms which are valuable relative to the dispersion of stock ownership. This, in turn, suggests that a given firm’s stock will not be “too” widely 4 CHAPTER 10. CORPORATE GOVERNANCE held relative to its value. This seems like an interesting topic for future research. The essential idea is captured by the following game. Suppose there is a finite number of shareholders i =1,...,n and shareholder i holds a fraction θ i of the firm’sshares. Thegameisthesameasaboveexceptthateach shareholder i can tender a fraction t i ≤ θ i ofhissharesandtheraidsucceeds if and only if X i t i ≥ γ. If the tender price is p the payo? to shareholder i is u i (p,t)= ? (θ i ?t i )V(a ? )+t i p if P i t i ≥ γ (θ i ?t i )V(ˉa)+t i p if P i t i <γ. Suppose that V(ˉa) <p<V(a ? ). Since we assume the o?er succeeds if and only if P i t i ≥ γ, each shareholder will minimize his o?er subject to this constraint. Any further reduction would cause the o?er to fail and his payo? would fall. For any p>V(ˉa) it is optimal for agents to submit the maximum t i if the tender o?er is expected to fail and the minimum consistent with P i t i ≥ γ if it is expected to succeed. In a SPE the raider will o?er p ≤ V(ˉa) and the shareholders will choose to o?er amounts t i such that P i t i ≥ γ. If p<V(ˉa) there exists a trivial continuation equilibrium in which t i =0 for all i if θ i <γfor each i. The equilibrium constructed here depends crucially on the assumption that the fraction of the shares needed for control is known with certainty so that every shareholder is pivotal. Introducing a small amount of uncertainty could upset this equilibrium. Holmstrom and Nalebu? (1992) study mixed strategy equilibria of the finite game. Abstract: This paper reexamines Grossman and Hart’s (1980) in- sight into how the free-rider problem excludes an external raider from capturing the increase in value it brings to a firm. The inability of the raider to capture any of the surplus depends criti- cally onthe assumptionof equal andindivisible shareholdings—the one-share-per-shareholder model. In contrast, we show that once shareholdings are large and potentially unequal, a raider may cap- ture a significant part of the increase in value. Specifically, the 10.2. BENEFITS OF MANAGERIAL INDEPENDENCE 5 free-rider problem does not prevent the takeover process when shareholdings are divisible. Grossman and Hart (1988) study the design of the firm’s corporate charter to optimize the role of takeovers in maximizing the value of the firm. There is a tradeo? between making the firm too di?cult to take over and thus protecting incumbent management and making it too easy and allowing the existing shareholders to be exploited in a corporate control contest. Abstract: This paper analyzes the optimality of the one share-one vote rule. The authors focus on takeover bids as a mechanism for allocating control. They assume two types of control benefits– benefits to security holders and private benefits to the controlling party. One share-one vote maximizes the importance of benefits to security holders, relative to benefits to the controlling party, and, hence, encourages the selection of an e?cient management team. However, one share-one vote does not always maximize the reward to security holders in a corporate control contest. Su?- cient conditions are given for one share-one vote to be optimal overall. The paper also includes a discussion of the empirical evidence. 10.2 Benefits of managerial independence The agency approach assumes that the manager is in control of the firm, that his interests are opposed to the interests of the shareholders, and that the shareholders maximize their interess by exerting control over his actions. This is a useful complement to the traditional idea that managers maximize shareholders’ preferences. How realistic is this view of the modern publicly traded company? In this section, we present a model of managerial indepen- dence and show that maximum control may not be optimal. We assume that the interests of managers and shareholders are imper- fectly aligned. Specifically, the manager has an incentive to overinvest.How- ever, the manager also has superior information about the e?cient level of investment. The essential idea is that the shareholders may want to give the manager discretion in order to take advantage of his superior information, even if discretion is costly because it allows overinvestment. 6 CHAPTER 10. CORPORATE GOVERNANCE The value of the firm is assumed to be a function v(x,θ)=(θ?x/2)x. oftheamountinvestedx ≥ 0 and a random variable θ, uniformly distributed on an interval[0,M], which can be interpreted as the profitability of invest- ment. The manager’s preferences are represented by a utility function u(x,θ) ≡ v(x,θ + a)=(θ + a?x/2)x, where a>0. A Pigovian tax t(x)=?ax achieves the first best. We assume that no such schemes are available. 10.2.1 Delegation without Commitment Delegation without commitment is a special case of the “cheap talk” game introduced by Crawford and Sobel (1982). A strategy for the manager is afunctionf :[0,M] → [0,M] and the shareholders’ strategy is a function g :[0,M] → R + . The shareholders beliefs are represented by a function μ :[0,M] → ?[0,M],where?[0,M] denotes the set of probability distribu- tions on [0,M].Thenμ(m) is the shareholders’ probability distribution over possible values of θ when the manager announces m. The equilibrium con- ditions require that each player is choosing a best response and that beliefs are consistent with Bayes’ rule wherever possible. (i) g(m) ∈ argmax R M 0 (θ?x/2)xdμ(m); (ii) f(θ) ∈ argmax(θ + a?g(m)/2)g(m); (iii) μ(m)=unif f ?1 (m), for almost all m. If Gis the range of the function g, then the manager is e?ectively choosing the level of investment from the set G and condition (ii) merely requires the manager to choose optimally from G for each value of θ.Theconcavityof the manager’s objective function implies that the set f ?1 (x) is convex for every x ∈ G. Furthermore, the number of these sets must be finite, as the next lemma shows. Lemma 1 Suppose that x and x 0 belong to G and are chosen in equilibrium and x<x 0 .Thenx + a<x 0 . 10.2. BENEFITS OF MANAGERIAL INDEPENDENCE 7 Without loss of generality, we can identify the manager’s strategy with a finite list of intervals {(θ k ,θ k+1 )} K k=1 ,whereθ 1 =0and θ K+1 = M, such that all manager types θ ∈ (θ k ,θ k+1 ) send the same signal, which causes shareholders to choose an investment level x k . Theorem 2 Let {(θ k ,x k )} K k=1 be a sequence satisfying θ 1 =0and θ k <θ k+1 and the following conditions: (i) x k =(θ k + θ k+1 )/2, for k =1,...,K, where θ K+1 = M; (ii) (θ k+1 + a)=(x k + x k+1 )/2, for k =1,...,K?1. Then there exists a perfect Bayesian equilibrium (f,g,μ) such that (θ k ,θ k+1 ) ? f ?1 (m k ) and g(m k )=x k ,fork =1,...,K. Conversely, for any perfect Bayesian equilibrium (f,g,μ),thereexistsasequence{(θ k ,x k )} K k=1 satisfying conditions (i) and (ii) and such that (θ k ,θ k+1 ) ? f ?1 (m k ) and g(m k )=x k , for k =1,...,K 10.2.2 Delegation with Commitment By the revelation principle, we can restrict attention to direct revelation mechanisms. A direct revelation mechanism is a function g :[0,M] → R + , where g(θ) is the investment specified by the shareholders when the manager reports his type to be θ. The manager will report his type truthfully if the mechanism is incentive-compatible and the optimal (incentive-compatible) mechanism maximizes the shareholders’ payo?E[(θ?g(θ)/2)g(θ)] subject to the incentive-compatibility constraint. As in the case without commitment, the manager is e?ectively choosing an element from the range of g, G = g([0,M]). It will be convenient to use this representation of the mechanism in our analysis. To avoid pathological cases, we assume that G is a closed set. Lemma 3 If g is an optimal, incentive-compatible mechanism, the graph G is an interval. Theorem 4 If g ? :[0,M] → R + is an optimal incentive-compatible mech- anism for the shareholders, then for some value of x 1 ≤ M, the mechanism has the form g ? (θ)=min{θ + a,x 1 },?θ ∈ [0,M]. 8 CHAPTER 10. CORPORATE GOVERNANCE The optimal mechanism involves putting an upper bound x 1 on the amount of investment but apart from that the manager can choose the level of investment that he wants. Although there is some constraint on managerial discretion, it will be very small when M is large or a is small. In other words, when uncertainty is large and/or the divergence between the interests of the manager and shareholders not too great, the manager will be given almost complete discretion. One is led to speculate that if the support of θ were un- bounded, the optimal mechanism could give managers complete discretion. In any case, this exercise indicates that there may be circumstances in which shareholders are best served by the separation of ownership and control, in spite of the existence of private benefits that distort the manager’s decision. 10.2.3 Burkart, Gromb and Panunzi (1997) This is a variant of solving the agency problem by “selling the firm to the agent”. If the conflict of interest between shareholders and managers leads to ine?ciency, reducing the claim of the shareholders may lead to greater ef- ficiency. Too much control by shareholders is not a good thing. The manager must have some discretion to pursue his own interests and reap private ben- efits; otherwise he will not have an incentive to make an e?ort on behalf of the firm. The problem is that shareholders, once they start to micro-manage, cannot commit themselves to reward the manager in a way that is consistent with optimal incentives. So the firm has to be constituted in a way that re- stricts shareholder power, in other words, commits them to leave some rents for the manager. Burkart, Gromb and Panunzi argue that ownership struc- ture can act as such a commitment device. By having dispersed ownership of outside equity, shareholders are e?ectively precommitting not to interfere with the managers. Each shareholder’s ownership will be su?ciently small that there will be little incentive to monitor. 10.3 Competition Over the last twenty years, the literature on corporate governance–or cor- porate finance for that matter–has focused on the agency problems facing shareholders under separation of ownership and control. The assumption underlying this literature seems to be either that 10.3. COMPETITION 9 ? government intervention is required to solve problems of corporate gov- ernance, or that ? the US economy is underperforming because of the corporate gover- nance problems inherent in the separation of ownership and control of publicly traded companies compared with other systems. While it is possible to build theoretical models to substantiate these claims, the empirical evidence is less clear. There is, of course, anecdotal ev- idence of agency problems. But, the equity premium puzzle and the success of publicly traded companies in the US and UK, suggest that shareholders have done quite well by comparison with bondholders. An alternative approach is to focus not on the governance of individual firms but instead to focus on the e?ect of competition among firms. Corpo- rate governance can be regarded as a technology. Competition forces firms to adopt the most e?cient technology. If a new technology arrives and is not adopted, it will be implemented by a competitor. Whatever ine?ciencies exist at any date are to be regarded as the state of the art technology. It has been argued (see, e.g., Alchian (1950) and Stigler (1958)) that com- petition in product markets is a very powerful force for ensuring good corpo- rate governance. If the managers of a firm waste or consume large amounts of resources, the firm will be unable to compete and will go bankrupt. There seems little doubt that competition, particularly international competition, is a powerful force in disciplining management. 10.3.1 Competition and managerial slack One idea studied in the corporate governance literature is that competition between di?erent organizational forms may be helpful in limiting e?ciency losses. If a family-owned business has the sole objective of maximizing share value, it may force all the corporations in that industry to do the same thing. An early attempt to model product-market competition as a mechanism to disciplinemanagersis foundinHart(1983). Onthe supplyside, Hartassumes that there is a large number of small firms. A fraction ν are traditional profit maximizers; these are called entrepreneurial firms. The remaining fraction 1?ν are operated by managers who maximize their own interests; 10 CHAPTER 10. CORPORATE GOVERNANCE these are called managerial firms. The firms have identical cost functions C(w,q,L),wherew is the input price, q istheoutputlevel,andL is the level of managerial e?ort. Managerial e?ort and input prices are assumed to be substitutes, in the sense that greater e?ort compensates for higher input costs: C(w,q,L)= ? C(Φ(w,L),q). The cost index Φ(w,L) is increasing in the input price w and decreasing in managerial e?ort L. Ex ante, the input prices are independently and identi- cally distributed across firms. Ex post, there is no aggregate uncertainty: the cross-sectional distribution of input prices is non-stochastic and proportional to the ex ante probability distribution. The manager takes output and input prices as given and decides how much output to produce and how much managerial e?ort to exert in order to maximize his own preferences. An incentive problem arises because the manager can observe his input price w and his e?ort L, but the shareholders cannot. Thus, a manager who faces a low input price may choose to shirk: instead of achieving high profits for the shareholders he exerts a low level of e?ort and claims that profits are low because the input price is high. The manager’s preferences are assumed to be additively separable in in- come and e?ort: the von Neumann-Morgenstern utility function is H(U(I)? V (L)),whereI is the manager’s income and L is his e?ort. The manager is infinitely risk averse: his utility-of-income function is very flat above ˉ I and very steep below ˉ I. The manager’s reservation utility is ˉ U.Inorder to be acceptable to the manager, a managerial contract must guarantee the manager an income that is at least ˉ I and never call on the manager to make an e?ort greater than ˉ L,where U( ˉ I)?V( ˉ L)=H ?1 ( ˉ U). These restrictive assumptions are chosen to make the problem analytically tractable, but it turns out that they are crucial for the substantive results as well, as we shall see below. Since the manager must be paid a fixed income and exert a fixed amount of e?ort to achieve his reservation utility, this is the only outcome that is consistent with e?ciency. If the shareholders could observe the manager’s e?ort L, they could achieve the first best by o?ering the manager a contract that pays him ˉ I as long as he exerts an e?ort L = ˉ L. Since they cannot do this, they must settle for the second best. It is assumed that the shareholders 10.3. COMPETITION 11 know the distribution function F(w) and the equilibrium product price p and can observe the firm’s ex post profit level. Let π(p,Φ) denote the maximum profits when the product price is p and the cost index is Φ.Ifthemanager follows the first best rule of setting L = ˉ L, the profits vary between a low of π(p,Φ(w max , ˉ L)) and a high of π(p,Φ(w min , ˉ L)). The manager must receive a salary of ˉ I (otherwise he gets less than his reservation utility ˉ U)andthereis no point giving him more (since his utility function is flat above ˉ I). Thus, the best that the owners can do is to o?er him a salary of ˉ I as long as profits are equal to π(p,Φ(w max , ˉ L)) or greater. The managers will accept this contract, which requires them to work flat out (i.e., choose L = ˉ L)whenw = w max ;but for any lower price w<w max they can slack by taking less e?ort. Precisely, for any w<w max the manager chooses L(w) so that π(p,Φ(w,L(w))) = π(p,Φ(w max , ˉ L)). Under this contract, the equilibrium level of aggregate output will be strictly lower and the output price will be strictly higher than in the first best. This analysis assumes that the input prices for di?erent firms are in- dependent, so the law of large numbers implies that there is no aggregate uncertainty. Suppose, on the other hand, that the input prices are perfectly correlated across firms. Then the equilibrium product price is stochastic and positively correlated with input prices. It is assumed that the shareholders cannot observe the equilibrium price or the profits of other firms in the in- dustry. As before, it is optimal for the shareholders to pick some input price ?w and insist that the manager work flat out at that price. For any other price w 6=?w the manager will continue to produce the same profits, but will engage in some slack. The amount of managerial slack in an individual firm is measured by the proportional increase in input prices that could be absorbed without any change in profits if slack could be eliminated. The average amount of slack is denoted by X ind in the market with independent input prices and X corr in the market with perfectly correlated input prices. Given an equilibrium of the kind described above and some additional assumptions, it can be shown that X corr ≤ X ind . In the independent case, the product price is constant across states of nature, because the cross-sectional distribution of input prices is non-stochastic. When an individual manager faces an input price w<w max , he can reduce his e?ort because he only needs to produce profits greater than or equal to π(p,Φ(w max , ˉ L)), the amount that he would produce by 12 CHAPTER 10. CORPORATE GOVERNANCE working flat out when the input price is w max . However, when input prices are perfectly correlated, there is a second e?ect to consider. Suppose that ?p is the product price in the state with the input price ?w. In a state of nature where w<?w, the output price p will be lower than ?p.Thisisbecauseallthe entrepreneurial firms also face a lower input price w<?w and are producing more output, thus pushing down the output price. The manager has to exert more e?ort to compensate for the fall in the output price, and this e?ect o?sets his ability to slack as a result of the fall in input price. It canbe shownthatX ind is the same as the amount of slack that would be realized in a monopolistic firm whose owners were faced by the same incentive problem in controlling the manager. In this sense, competition does not have an e?ect on managerial slack in the independent case. However, it can be shown that an increase in the fraction of entrepreneurial firms ν does reduce X corr . The more entrepreneurial firms there are in the market, the stronger the incentive e?ect of competition on managers will be. We remarked above that the simplifying assumptions about the man- agers’ risk aversion turn out to be crucial for the substantive conclusions of Hart’s analysis. Scharfstein (1988) shows that the result that increased competition reduces managerial slack can be reversed when the manager’s marginal utility of income is strictly positive. Perhaps this should not be too surprising. Consider a di?erent model, say a Cournot oligopoly model in which managerial e?ort is unobserved and the manager receives a fixed share of profits. Greater competition (in the form of a larger number of firms in the market) will reduce profits and hence the managers’ incentive to work. There is a trade-o?, of course, between the interest of the shareholders in greater profits and the interest of consumers in getting more output at a lower price, so this is not to say that greater competition is a bad thing. Still, it suggests that the e?ects of competition will generally be ambiguous. The role of the market in Hart (1984) and Scharfstein (1988) is to provide information. If the owners could observe the profits of the entrepreneurial firms, they could use that information as a benchmark to condition the re- wards to the managers. This is the approach taken by Holmstrom (1982) and Nalebu? and Stiglitz (1983), for example. In Hart’s model this information is unavailable, so an indirect channel must be used. Competition translates a fall in input prices into a fall in output prices, which in turn are translated into lower profits unless the managers work harder to keep profits up. In a recent study, Schmidt (1997) addresses a related question in a model without hidden information. Schmidt (1997) observes that increased compe- 10.3. COMPETITION 13 tition may threaten the survival of a firm by forcing it into bankruptcy and asks what e?ect this will have on managerial slack. Schmidt’s model assumes that the manager is risk neutral and wealth-constrained and that he su?ers a penalty (lost rents, foregone opportunities) when the firm goes bankrupt. The manager is required to engage in e?orts that reduce future production costs. Costs are assumed to take on two values, high or low, and greater e?ort reduces the probability of high costs. The actual costs are observed after the manager has made the e?ort, and at that point the owner decides whether to liquidate the firm or not. Greater competition lowers the price that the firm receives for its output and, other things being equal, increases the risk that the owner will find it optimal to liquidate the firm. Schmidt discovers that greater competition has two e?ects on the man- ager’s optimal e?ort. The first is the threat-of-liquidation e?ectthatweex- pect to find: the manager has a directly increased incentive to work harder to avoid liquidation. There is also an indirect e?ect, since the cost to the owner of providing incentives to take high e?ort is reduced. So the threat- of-liquidation e?ect unambiguously raises managerial e?ort. The second e?ectisambiguousandresultsfromthefactthatincreased competition reduces profits and may reduce the benefits of a cost reduction. As a result, the owner may be disinclined to pay the manager the high rents necessary to achieve a cost reduction. If the value of a cost reduction is de- creasing in the degree of competition, the net e?ect of increased competition may be to lower managerial e?ort. The second e?ect occurs only if the manager is paid more than his reser- vation level. Schmidt (1997) cites a study by Aghion, Dewatripont, and Rey (1995) that treats a special case of his model. In their model, the manager is always paid his reservation level and so the e?ect of increased competition is unambiguous. These studies describe a mechanism by which competition in the product market helps discipline managers, but they are restrictive in several respects: ? First, they all take the agency approach, in which the main obstacle to achieving e?ciency is the principal-agent relationship between the manager and the shareholders. As we have indicated, some organiza- tions appear to function successfully even in the absence of external governance mechanisms. ? Secondly, the focus of the models on cost minimization seems to be too narrow. While cost minimization may be a useful proxy for other 14 CHAPTER 10. CORPORATE GOVERNANCE important managerial activities, it is not clear that this model captures all the important features of managerial behavior. ? Thirdly, casual empiricism suggests that “e?ort” is something that managers supply quite readily. The failure of management to maxi- mize the shareholders’ interests may come from other sources than lack of e?ort. With an alternative source of management failure, such as dif- ferences in inherent ability (adverse selection), risk shifting, or private benefits, the e?ects of competition may be di?erent. ? Fourthly, the entrepreneurial firms that force the managers to provide greater e?ort are a deus ex machina in these models. What happens in industries where all the firms are managerial? Perhaps competition among managerial firms will only ensure that corporate governance is equally ine?cient across firms. For example, if all the firms in a market are corporations which face an agency problem, they will all be able to survive if the ine?ciencies associated with corporate governance problems a?ect all firms equally. Competition between them may not lead to full e?ciency (Jensen and Meckling (1976)). ? Fifthly, the arguments of Hart and his followers assume that markets are perfectly competitive. As we have seen, things could be quite dif- ferent in markets that are imperfectly competitive. Many markets in which Fortune 500 companies operate are oligopolistic. These compa- nies compete on the product market, but it is not clear what e?ect im- perfect competition has on managerial slack. Possibly collusion among asmallnumberoffirms will take the form of high levels of managerial slack rather than high monopoly profits. Nonetheless, the idea that competition enhances the performance of manage- rial firms may have a broader application than these models suggest. We next review the empirical evidence on the e?ect of competition on performance. Although limited in quantity, this suggests that competition is important. 10.3.2 The E?ectiveness of Competition as a Control Mechanism The empirical evidence on the role of competition in ensuring corporate per- formance is sparse. Nickell (1996) suggeststhatthemostpersuasiveevidence 10.3. COMPETITION 15 consists of broad-brush observations. The firstoftheseisthelowlevelofpro- ductivity in Eastern Europe compared to Western Europe after competition was suppressed in the East under communist regimes. The second is the im- portance of domestic competition in ensuring that firms are internationally competitive, demonstrated by Porter (1990). The third is that deregulation is generally followed by productivity gains. Graham, Kaplan and Sibley (1983) document this for the case of the U.S. airline industry. A number of studies have provided detailed evidence of the e?ect of com- petition on performance. One question is the e?ect of competition on inno- vation. Geroski (1990) and Blundell, Gri?th and Van Reenen (1995) find that the more concentrated an industry and the higher are other measures of monopoly power the lower is the rate of innovation. Another is the re- lationship between competition and technical e?ciency. Caves and Barton (1990), Green and Mayes (1991) and Caves (1992) find that above a certain level of market concentration technical e?ciency is reduced. Caves (1980) points to the evidence in the management literature that competition leads to more e?cient decision-making structures in firms. Finally, Nickell (1996) and Nickell, Nicolitsas and Dryden (1997) find evidence that the higher is the level of competition the higher is the level of growth in productivity. Moreover, the latter paper documents that competition is a substitute for other corporate governance mechanisms. 10.3.3 Competition and corporate governance The Hart model studies competition between entrepreneurial and managerial firms ex post, that is, after the companies have been set up. In this section we consider an alternative model of the process by which companies are formed and argue that ex ante competition for capital will force companies to adopt incentive e?cient corporate governance. ? Competition and free entry will reduce profits to zero whatever agency costs the firm is subject to. Survival requires profit maximization sub- ject to incentive constraints. ? In equilibrium, each stakeholder will be receiving his outside option plus information rents. Competition for labor and capital raises factor pricesandputs pressureonine?cientfirms. Evenamonopolyis subject to this competition, unless it is also a monopsony. 16 CHAPTER 10. CORPORATE GOVERNANCE ? With increasing returns to scale, ine?ciency threatens survival. With decreasing returns, the firm can always compensate for rent seeking by remaining small. Our argument is most powerful in the case of winner-takes-all industries, where competition threatens survival. Example AsourleadingexamplewetaketheJensenmodeloffreecashflow. A company requires a single manager and an investment of K 0 units of capital at date 0 to produce 2K 0 units of output at date 1.Therevenuecanbepaid out to investors at date 1 or be re-invested in the firm. If it is re-invested, it earnsalowrateofreturnwhichforsimplicityisassumedtobezero. The manager earns private benefits from investment, so will always re- invest any retained earnings. If some of the investment is in the form of debt, the manager is forced to pay out some of the earnings. Let D be the face value of the debt. Then, ignoring discounting, investors earn min{D,2K 0 )} Clearly, it is optimal to set D =2K 0 . Think of types θ ∈ [0,2K 0 ] as representing di?erent capital structures (facevaluesofdebt). Supposethereisanunlimitednumberofmanagers and a finite amount of capital. Let R be the opportunity cost of capital. Clearly, R will be set by the optimal type: R =2.Noothertypeoffirm will be able to raise capital. That is, θ<2K 0 implies the return on capital is min{θ,2K 0 } = θ<RK 0 . References Bagnoli, Mark and Bart Lipman.“Successful Takeovers without Exclusion,” University of Michigan Center for Research on Economic and Social Theory Working Paper: 87-13 (1987) 25. Berle, A. and G. Means (1932). The Modern Corporation and Private Prop- erty. Chicago: Commerce Clearing House, Inc. Blundell, Richard, Rachel Gri?th and John Van Reenen. “Dynamic Count Data Models of Technological Innovation,” Economic Journal 105 333-344. Burkart, M. , D. Gromb and F. Panunzi (1997). “Large Shareholders, Mon- itoring, and the Value of the Firm,” Quarterly Journal of Economics 112, 693-728. 10.3. COMPETITION 17 Caves, Richard. “Industrial Organization, Corporate Strategy and Struc- ture,” Journal of Economic Literature 18 (1980) 64-92. – Industrial E?ciency in Six Nations, Cambridge, MA: MIT Press, (1992). –andDavidBarton.E?ciency in US Manufacturing Industries Cambridge, MA: MIT Press, (1990). Crawford, V. and J. Sobel (1982). “Strategic Information Transmission,” Econometrica 50, 1431-51. Geroski, Paul. “Innovation, Technological Opportunity, and Market Struc- ture,” Oxford Economic Papers 42 586-602. Graham, David, Daniel Kaplan and David Sibley. “E?ciency and Competi- tion in the Airline Industry,” Bell Journal of Economics 14 (1983) 118-138. Green, Alison and David Mayes. “Technical Ine?ciency in Manufacturing Industries,” Economic Journal 101 (1991) 523-538. Grossman, Sanford and Oliver Hart. “Takeover Bids, the Free-Rider Prob- lem, and the Theory of Corporation,” Bell Journal of Economics 11 (1980) 42 — 64. – and –. “The Allocational Role of Takeover Bids in Situations of Asym- metric Information,” Journal of Finance 36 (1981) 253-70. –and–. “OneShare-OneVoteandtheMarketforCorporateControl,” Journal of Financial Economics 20 (1988) 175-202. Hart, Oliver. “The Market Mechanism as an Incentive Scheme,” The Bell Journal of Economics 14 (1983) 366-82. Hart, O. and J. Moore (1994). “A Theory of Debt Based on the Inalienability of Human Capital,” Quarterly Journal of Economics 109, 841-879. Holmstrom, Bengt and Barry Nalebu?. “To the Raider Goes the Surplus? A Reexamination of the Free-Rider Problem,” Journal of Economics and Management Strategy 1 (1992) 37-62. Jensen, M. (1986). “Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers,” American Economic Review 76, 323-29. Manne, Henry. “Mergers and the Market for Corporate Control,” Journal of Political Economy (1965), 110-20. Nickell, Stephen. “Competition and Corporate Performance,” Journal of Political Economy 104 (1996) 724-746. –, Daphne Nicolitsas and Neil Dryden. “What Makes Firms Perform Well?” European Economic Review 41 (1997) 783-796. Porter, Michael. The Competitive Advantage of Nations, London: Macmil- lan, (1990). 18 CHAPTER 10. CORPORATE GOVERNANCE Scharfstein, David. “Product Market Competition and Managerial Slack,” The RAND Journal of Economics 19 (1988) 147-155. Schmidt, Klaus. “Managerial Incentives and Product Market Competition,” Review of Economic Studies 64 (1997) 191-214.