Hart 2001 JEL Financial Contracting

7/31/2019 Hart 2001 JEL Financial Contracting

1/23

American Economic Association

Financial ContractingAuthor(s): Oliver HartSource: Journal of Economic Literature, Vol. 39, No. 4 (Dec., 2001), pp. 1079-1100Published by: American Economic AssociationStable URL: http://www.jstor.org/stable/2698520

Accessed: 04/11/2008 09:17

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=aea.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access toJournal

of Economic Literature.

http://www.jstor.org
http://www.jstor.org/stable/2698520?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=aeahttp://www.jstor.org/action/showPublisher?publisherCode=aeahttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2698520?origin=JSTOR-pdf


2/23

of Economic LiteratureXXXIX (December 2001) pp. 1079-1100

Financial ContractingOLIVER HART'

1. IntroductionFINANCIAL CONTRACTINGmightbedescribed as the theory of whatkinds of deals are made between finan-ciers and those who need financing. Letme motivate the subject matter of thisarticle with the following questions:(A) Suppose an entrepreneur has an ideabut no money and an investor has

money but no idea. There are gains-from trade, but will they be realized?If the idea (project) gets off theground, how will it be financed?(B) We see companies around the worldwith a wide variety of financial struc-tures. Almost all companies haveowners (i.e., shareholders or equityholders). Some have other claimants,e.g., creditors, preferred sharehold-ers, etc. Why? Does this matter, forexample, for corporate efficiency orinvestment behavior? What deter-mines a company's debt-equity ratio,

that is, the ratio of the market valueof its debt to the market value of itsequity?2Questions like these have been thefocus of much of the very large corpo-rate finance literature that has devel-oped over the last forty years, and theyhave also been studied in the morerecent financial contracting literature.My plan is to summarize some of the

older literature (section 2) and thenmove on to some more recent thinking(sections 3 and 4). Section 2 will be de-liberately brief and will not do justiceto the older literature. Fortunately,there are excellent surveys by MiltonHarris and Artur Raviv (1991), AndreiShleifer and Robert Vishny (1997), andLuigi Zingales (2000) that the readercan consult to supplement what I haveto say. (The latter two papers alsohave insightful things to say about thefinancial contracting literature.)

2. Established Viewsof Financial StructureThe modern corporate finance litera-

ture starts with the famous Modiglianiand Miller (MM) theorem (Franco

1 HarvardUniversity and London School of Eco-nomics. This article is a revised version of theNancy L. Schwartz Lecture delivered at North-western University in June 2000. I would like tothank Philippe Aghion, Patrick Bolton, BengtHolmstrom, John Moore, Andrei Shleifer, andJean Tirole for helpful comments; Fritz Foley forexcellent research assistance;and Ehud Kalai andMort Kamien for inviting me to give the lecture. Iwould also like to acknowledge research supportfrom the National Science Foundation through theNational Bureauof Economic Research.

2 Post-war, the value of long-term debt of largeU.S. corporations has been about half the value ofequity. See Franklin Allen and Douglas Gale(2000).1079


3/23

1080 Journal of Economic Literature, Vol. XXXIX(December 2001)Modigliani and Merton Miller 1958).This striking irrelevance result can beparaphrased as follows:

Modigliani-Miller (MM): In an ideal world,where there are no taxes, or incentive or in-formation problems, the way a project or firmis financed doesn't matter.A simple (too simple) way to under-stand this result is the following. A proj-ect can be represented by a stream ofuncertain, future cash flows or (net)revenues. Each future revenue isequivalent to some amount of cash to-day; the exact amount is obtained by

applying a suitable discount factor (ifthe future revenue is uncertain, wemight apply a higher discount factor).Now add all the cash equivalents to-gether to obtain the total value of theproject-its present value, V, say.Suppose the project costs an initialamount C. Then the project is worthundertaking if and only if V > C, that is,if and only if it contributes positive netvalue. Now we get to MM.3 The finan-ciers of the project-who put up theC-have to get their C back. They canget it back in a variety of ways: theycould be given a share s of future reve-nues, where sV = C. Or they could getsome debt (riskless or risky) that has apresent value equal to C. But, howeverthey get it back, they must get C, andsimple arithmetic tells us that the en-trepreneur who sets up the project willget the remainder V - C. That is, fromthe entrepreneur's point of view (andfrom the financiers') the method offinancing doesn't matter.4Merton Miller (who, sadly, died re-

cently) used to illustrate MM with oneof Yogi Berra's famous (mis-)sayings:"You better cut the pizza in four piecesbecause I'm not hungry enough to eatsix."5 Apart from the crumbs, thisseems to sum up the proposition prettywell.MM, although an enormously impor-tant benchmark, does not seem to de-scribe the world very well. To give oneexample of a problem, if MM wereempirically accurate, we might expectfirms to use no debt or large amountsof debt, or firms' debt-equity ratiosto be pretty much random. However,Raghuram Rajan and Zingales (1995)find that similar, systematic factors de-termine the debt-equity ratio of firmsin different countries. In fact, I thinkthat it would be fair to say that, since itsconception, MM has not been seen as avery good description of reality; thus,much of the research agenda in corpo-rate finance over the last forty years hasbeen concerned with trying to find"what's missing in MM."Researchers have focused on twoprincipal missing ingredients: taxes andincentive problems (or asymmetric in-formation). In both cases the idea isthat, because of some "imperfections,"V is not fixed and financial structurecan affect its magnitude.2.1 Taxes

The simplest tax story is the follow-ing. In many countries, the tax authori-ties favor debt relative to equity: in par-ticular, interest payments to creditorsare shielded from the corporate incometax while dividends to shareholders arenot. As a result, it is efficient for a firmto pay out most of its profits in theform of interest-this reduces its taxbill and thus increases the total amountavailable for shareholders and creditors

3 Actually the result that the project should beundertaken if and only if V > C can also bethought of as being part of MM.4 This informaljustification of MM can easily bemade rigorous for the case where the entrepre-neur and investors are risk neutral. If the partiesare risk averse, however, a more subtle, "home-made leverage" argument is required. See JosephStiglitz (1974). 5 See Berra (nd).


4/23

Hart: Financial Contracting 1081taken together. (Of course, this increasein firm value is at the expense of societysince the treasury receives less taxrevenue,)This simple tax story is too simple: itsuggests that we should see muchhigher debt-equity ratios than we actu-ally do. For this reason, it has beenelaborated on in various ways.6 But ex-tensions of the theory, however in-genious, do not seem to be adequate toexplain the data: for example, Rajan andZingales (1995) find that, while taxes in-fluence debt-equity ratios, other factorsare important too.

In- fact, in the last few years theliterature has focused on a differentdeparture from MM:incentives.2.2 Incentive (Agency) Problems

The most famousincentivepaperin thecorporate finance li'terature is MichaelJensen and William Meckling (1976).Jensen and Mecklingargue that the valueof the firm or project V is not fixed, asMM assume: rather it depends on theactions of management, specificallytheir consumption of "non-pecuniarybenefits" (perks). Perks refer to thingslike fancy offices, private jets, the easylife, etc. These benefits are attractiveto management but are of no interestto shareholders-in fact they reducefirm value. Moreover, it is reason-able to assume that they are ineffi-cient in the sense that one dollar ofperks reduces firm value by more thana dollar.7Jensen and Meckling use these ideasto develop a trade-off between debt andequity finance. Consider a manager (or

entrepreneur) who initially owns 100percent of a firm. This manager willchoose not to consume perks since eachdollar of perks costs more than a dollarin market value (and as owner he bearsthe full cost). Now suppose the man-ager needs to raise capital to expand thefirm. One way to do this is to issue eq-uity to outside investors. However, thiswill dilute the manager's stake-he willnow own less than 100 percent of thefirm. As a result, he will consume perks,since the cost of these is borne at leastpartially by others. As noted, this is in-efficient since total value (firm valueplus the value of perks) will fall.Alternatively, suppose the managerborrows to raise capital. At least forsmall levels of debt, this does not dilutethe manager's stake. The reason is thatthe debt must be paid back for sure (itis riskless), which means that, in a mar-ginal sense, the manager still owns 100percent of the firm (his payoff is V - D,where D is the value of the debt). As aresult, he bears the full cost of perksand will not take them.So far it looks as if borrowing is anefficient way to raise capital. However,Jensen and Meckling argue that borrow-ing becomes costly when debt levels arelarge. The debt then becomes risky,sincethere is a chance that it won't be repaid.At this point, the manager will have anincentive to gamble with the firm's as-sets, e.g., to engage in excessively riskyinvestments. The reason is that if anexcessively risky project succeeds, thefirm's profits are high and the benefici-ary is the firm's owner-that is, themanager himself (recall that he has 100percent of the equity); whereas if theproject fails, the firm's profits are lowand the losers are the firm's creditorssince the firm is bankrupt.According to Jensen and Meckling,the optimal debt-equity ratio or capitalstructure for the firm is determined at

6 See, e.g., Miller (1977).7 This assumption makes sense since managerscan typically consume perks only in quite narrowways; that is, if unconstrained, they might preferto spend an extra dollar on their children's educa-tion rather than a fancy office, but the formerwould look suspicious whereas the latter can bedefended (to shareholdersand society).


5/23

1082 Journal of Economic Literature, Vol. XXXIX (December 2001)the point where the marginal benefit ofkeeping the manager from taking perksis offset by the marginal cost of causingriskybehavior.The effects that Jensen and Mecklingemphasize are clearly important. How-ever, their analysishas a theoretical short-coming. The incentive problem thatJensen and Meckling focus on is whateconomists call an agency problem, i.e.,a (potential) conflict of interest betweenan agent who takes an action (in thiscase, the manager choosing the level ofperks) and a principal who bears the con-sequences of that action (other share-holders or creditors). There is a largeeconomics literature on agency prob-lems, but the main finding from thatliterature is that the best way to dealwith them is to put the agent on anoptimal incentive scheme.An illustration may be helpful. Sup-pose you (the principal) hire me (theagent) to sell silverware; my job is todrive around the suburbs, knock onpeople's doors, and try to interest themin knives and forks. You may be worriedthat I will sit in my car listening to rapmusic and not selling your product. Onesolution is to pay me a fixed amount perset of silverware I sell (a piece-rate)rather than a fixed wage per hour. (Oryou might use a combination of thetwo.)Applying this logic to the presentcontext leads to the conclusion that themanager's salary should be geared tofirm performance, that is, the managershould be put on an incentive scheme,I =f (V), where V is firm market value.But this can be done independently ofthe firm's financial structure, that is, in-dependently of whether the manager isa shareholder. (In the silverware exam-ple, I did not have to become a share-holder of the silverware firm to workhard.) Moreover, given an optimal in-centive scheme, the manager's prefer-

ence for'borrowing rather than issuingshares disappears.8In other words, a question unan-swered by Jensen and Meckling's analy-sis is: why use financial structure ratherthan an incentive scheme to solve whatis reallyjust a standardagencyproblem?Before we move on, it is worth men-tioning another strand of the agencyliterature that focuses on private in-formation possessed by managers ratherthan managerial actions. (This part ofthe literature corresponds to the ad-verse selection version of the moral haz-ard problem studied by Jensen andMeckling.) A leading example of this lit-erature is Stewart Myers and NicholasMajluf (1984). Like Jensen and Meck-ling, Myers and Majluf consider a man-ager who needs capital to expand thefirm. Myers and Majluf ignore perks,but suppose that the manager has bet-ter information about the profitabilityof the existing firm, i.e., assets in place,than investors. In particular, imaginethat the manager knows that these areworth a lot, whereas investors do not.Then, if the manager acts on behalf ofcurrent shareholders (e.g., because heholds equity in the firm himself), hewill not want to raise capital by issuingnew shares. The reason is that the newshares will be sold at a discount relativeto their true value, which dilutes thevalue of the current shareholders' stake.Instead the manager will raise capitalby issuing (riskless) debt. Riskless debtwill not sell at a discount-the firm willsimply pay the market interest rate onit. Hence no dilution will take place.8 This point is elaborated on in Philip Dybvigand Jaime Zender (1991). Paying the manager ac-cording to total market value V has the drawbackthat the manager may have an incentive to investin unprofitable projects in order to raise V. Thisproblem can be overcome by deducting the capitalraised from V before assessing the manager'ssalary,i.e., paying the manager according to valuenet of investment cost.


6/23

Hart: Financial Contracting 1083Thus Myers and Majluf provide anotherreason why MM fails: if managers havesuperior information, they will want tosell new securities whose return is in-sensitive to this information (risklessdebt being the most insensitive securityof all).Myers-Majluf are surely right thatprivate information is an important de-terminant of financial structure, and theeffect that they identify appears to beempirically significant. However, theiranalysis suffers from the same theoreti-cal weakness as Jensen-Meckling. Fi-nancial structure matters only becausemanagers are (implicitly) on a particularkind of incentive scheme. Specifically,Myers and Majluf assume that man-agers act on behalf of current share-holders, e.g., because they hold equitythemselves. But things don't have to bethis way. Suppose managers were paid afraction of the firm's total market valueV. Then managers wouldn't worry aboutselling new shares at a discount, sinceany loss in current shareholder value isoffset by a gain in new shareholdervalue and managers are paid on thebasis of the sum of the two. With thisincentive scheme, managers are happyto expand by issuing new equity, andfinancial structure no longer matters.9

3. Financial ContractingLiterature:Decision and ControlRightsWe have seen that incentive (agency)problems alone do not yield a very sat-isfactory theory of financial structure.

The recent financial contracting litera-ture (developed in the last fifteen yearsor so) adds a new ingredient to thestew: decision (control) rights.10This literature takes as its startingpoint the idea that the relationshipbetween an entrepreneur (or manager)and investors is dynamic rather thanstatic. As the relationship develops overtime, eventualities arise that could noteasily have been foreseen or plannedfor in an initial deal or contract be-tween the parties. For example, howmany people in 1980 could have antici-pated the fall of the Soviet Union or therise of the internet in the 1990s? In anideal world, a contract between a com-puter manufacturer (IBM, say) and asoftware producer (Microsoft), writtenin 1980, would have included a contin-gency about what would happen if theinternet took off-or for that matterwould have had a clause guardingagainst Microsoft becoming the domi-nant supplier of operating systems. Inpractice, writing such a contingent con-tract would have been impossible: thefuture was simply too unclear.Economists (and lawyers) use the term"incomplete" to refer to a contract thatdoes not lay out all the future contin-gencies. A key question that arises withrespect to an incomplete contract is: how

9There is in fact a strict advantage to puttingmanagers on an incentive scheme that rewardsthem according to total shareholder value, ratherthan current shareholder value. As Myers andMajluf show, the latter scheme may cause manag-ers to turn down some profitable new projects, be-cause the dilution effect on current shareholdervalue will be so great that they prefer not to in-vest. This inefficiency is avoided if managers arerewarded according to total shareholder value.John Persons (1994) argues that an incentivescheme where managers are paid a fraction of thefirm's total market value is not "renegotiation-proof":the board of directors (acting on behalf ofcurrent shareholders) will always revise it. How-ever, Persons does not explainwhy the board actson behalf of current shareholders or why theboard is given the power to revise the managerialincentive scheme.

10This recent literature should be contrastedwith an earlier literature based on costly state veri-fication; see Robert Townsend (1978) and Galeand Martin Hellwig (1985). In this earlier litera-ture, an optimal contract between an entrepreneurand investor was analyzed under the assumptionthat a firm's profitability is private information,but that this information can be made public at acost. This earlier literature did not stress contrac-tual incompleteness (as defined below) or focus onthe role of decision (control) rights.


7/23

1084 Journal of Economic Literature, Vol. XXXIX (December 2001)are future decisions taken? That is,given that an incomplete contract is si-lent about future eventualities, and giventhat important decisions must be takenin response to these eventualities, howwill this be done? What decision-mak-ing process will be used?It might be helpful to give some ex-amples. Consider a firm that has a long-term supplier. Advances in technologymight make it sensible for the firm in-stead to buy its inputs on the internet.Who makes the decision to switch?Or take a biotech firm that is engagedin trying to find a cure for diabetes.The firm has been pursuing a particulardirection, but new research suggeststhat a different approach might bebetter. Who decides whether the firmshould change strategy?Other examples concern whether afirm should undertake a new invest-ment, whether the firm's CEO shouldbe replaced, or whether the firm shouldbe closed down.The financial contracting literaturetakes the view that, although the con-tracting parties cannot specify whatdecisions should be made as a functionof (impossible) hard-to-anticipate-and-describe future contingencies, they canchoose a decision-making process in ad-vance. And one way they do this isthrough their choice of financial struc-ture. Take equity. One feature of mostequity is that it comes with votes. That is,equity-holders collectively have the rightto choose the board of directors, whichin turn has the (legal or formal) rightto make key decisions in the firm-specifically, the kinds of decisionsdescribed above.In contrast, take debt. Creditors donot have the right to choose the boardof directors or to take decisions in thefirm directly. However, they have otherrights. If a creditor is not repaid, shecan seize or foreclose on the firm's as-

sets or push the firm into bankruptcy.Moreover, if the firm enters bank-ruptcy, then creditors often acquiresome of the powers of owners.A rough summary is that shareholdershave decision rights as long as the firm issolvent, while creditors acquire decisionrights in default states.It is worth emphasizing the differ-ence between this perspective and thatdescribed in section 2. According toMM, the firm's cash flows are fixed andequity and debt are characterized bythe nature of their claims on these cashflows: debt has a fixed claim whileequity gets the residual. In Jensen andMeckling, the same is true except thatnow the allocation of cash flow claimscan affect firm value through manage-rial incentives. In neither case do votesor decision rights matter. In contrast,in the financial contracting literature,decision rights or votes are key, eventhough, of course, as we shall see, cashflow rights matter a lot too.It is also worth noting that there is animportant distinction between the kindsof decisionswe are talking about here andthe managerial actions we discu-ssed inthe context of Jensen-Meckling. Mana-gerial actions, e.g., the level of perks oreffort, are usually assumed to be non-transferable (or hard to transfer): onlythe manager can choose them. In con-trast, decision rights are (more easily)transferable: e.g., the decision aboutwhether to replace the CEO, say, canbe taken by one party (shareholders) orby another party (creditors). Hence, akey design question is: how should deci-sion rights be allocated in the initialcontract/deal between the parties? Tothis we now turn.3.1 TheAllocation of Decision Rights

The financial contractingliterature hastended to focus on small entrepre-neurial firms-rather than a publicly


8/23

Hart: Financial Contracting 1085traded company or corporation-andwe will do this too for the moment. Tomake things very simple, consider a sin-gle entrepreneur, a single investor, anda single project. The question is, howshould the right to make future deci-sions be allocated between the entre-preneur and the investor? Who shouldhave the right to replace the CEO orterminate the project?In order to answer this question, weobviously need a theory of why theallocation of decision-making authoritymatters. Various possibilities have beenadvanced. One approach is based on theidea that decision rights are importantfor influencing asset- or relationship-specific investments. Suppose individ-ual i is considering whether to investresources in learning about how tomake the project more profitable. If hecontrols the project, and has a good idea,he can implement this idea without in-terference from anyone else. This giveshim a strong incentive to have an idea.On the other hand, if someone else con-trols the project, i will have to get per-mission from this other person and mayhave to share the fruits of his idea withthem; this will dilute his incentives.The above approach has been used inthe theory of the firmll but has beenemployed less in the financial contract-ing literature. Instead, in this latter lit-erature, researchershave focused on howthe allocation of control rights affectsthe trade-off between cash flows andprivate benefits once the relationship isunderway.The best-known paper adopting thisapproach is Philippe Aghion and PatrickBolton (1992).12 Aghion and Bolton as-

sume that the project yields cash flowsin the amount of $V and private bene-fits in the amount of $B. Private bene-fits are similar to the non-pecuniarybenefits discussed in Jensen-Meckling;although private benefits may representthings like psychic value, we supposethat they have a cash equivalent, i.e., theycan be measured in dollars. The inves-tor is interested only in cash flows, whilethe entrepreneur s interested in both cashflows and private benefits. These differ-ent interests create a potential conflictbetween the entrepreneur and investor.Since private benefits (like decisionrights) are very important n what follows,it may help to illustrate them in the cur-rent context. Consider an entrepreneurwho has developed an idea for a project.The entrepreneur is likely to get somepersonal satisfaction from workingon theproject, or from the project succeeding,that is over and above any cash flowsreceived. Also, if the project succeeds, theentrepreneur's reputation is enhancedand he will do better in future deals.Personal satisfaction and reputationalenhancement are both examples of pri-vate benefits since they are enjoyed bythe entrepreneur but not the investor.Some private benefits are less in-nocuous. Someone who controls a proj-ect can decide who will work on theproject; the controller may choose toappoint relatives or friends to key posi-tions even though they are incompetent("patronage"). The controller may alsobe able to divert money from the proj-ect, e.g., he can set up other firms thathe has an ownership interest in, andchoose the terms of trade between theproject and these firms to suck cash outof the project. Patronage and diversionare also examples of private benefits.As noted, the existence of privatebenefits introduces a potential conflictof interest between the entrepreneurand the investor. How is this conflict

11 See Sanford Grossmanand Hart (1986), HartandJohn Moore (1990), and Hart (1995).12 For related work, see Bolton and DavidScharfstein (1990), Douglas W. Diamond (1991),Hart and Moore (1994) and Hart and Moore(1998).


9/23

1086 Journal of Economic Literature, Vol. XXXIX(December 2001)resolved? The answer is that this de-pends to a large extent on who hasthe right to make decisions once therelationship is underway.To understand this, consider a simplecase where the entrepreneur is allocateda fraction 0 of the project cash flows andthe investor receives the remaining (1 -0). Suppose that the project is set up atdate 0 and all decisions are taken andbenefits earned at date 1. The date 1objective functions of the entrepreneurand investor are then as follows:

Entrepreneur: Max B + OV,Investor: Max (1 - O)V_ Max V

It is also useful to write down the ob-jective of a planner who is concernedwith social (or Pareto) efficiency. In afirst-best world where lump sum distri-butions are possible the planner wouldmaximize the sum of the entrepreneurand investor's payoffs (since both aremeasured in money), i.e., social surplus,B +V.

Social Planner:Max B + V.It is clear that these three objectivefunctions are generally distinct. Thissuggests that it will indeed matterwhether the entrepreneur or the inves-tor makes ex post decisions. (The planner

is a mere construct and so will not havedecision-making authority!)For example, suppose the only deci-sion to be made concerns whether theproject should be terminated or contin-ued (at date 1). Assume that E's privatebenefit from continuation is $100, butthat $200 in resources can be saved ifthe project is terminated now ratherthan later. Also assume 0 = .1.From a social surplus or efficiencyperspective, the project should be ter-minated (the $200 loss exceeds the$100 private benefit and social surplusis represented by the sum of these).

This outcome will be achieved if the in-vestor makes the decision since sheputs no weight on private benefits, butnot if the entrepreneur does (given hisstake of 10 percent, he gains only $20from avoiding losses, but loses his fullprivate benefit of $100).On the other hand, suppose that thelosses from continuation are $80 ratherthan $200. Now it is efficient to con-tinue the project, and this time effi-ciency will be achieved if the entre-preneur has decision-making authority,but not if the investor does (since theinvestor is concerned only with lossavoidance) 13

13We have not considered renegotiation. Sup-pose the losses from continuation are $200. Wesaw that if the entrepreneur has control at date 1,he will keep the firm going even though this isinefficient. However, one thing the investor coulddo is to offer the entrepreneur a payment in re-turn for closing the firm down. The entrepreneurrequires at least $80 to make this worthwhile andthe investor is prepared to offer up to $180-sopresumablysomething in this range will be agreedupon. Similarly, f the losses from continuation are$80, and the investor has control, the entrepre-neur-if he has the money-could pay the inves-tor an amount between $72 and $92 to persuadeher not to close the firm down.In fact, in a world of perfect renegotiation, thefamous Coase theorem tells us that the allocationof decision rights doesn't affect the date 1 out-come at all: the parties will always arrive at theefficient outcome through bargaining.However, inthe present context, there is an importantimpedi-ment to renegotiation: the fact that the entrepre-neur is wealth-constrained. (Presumably this iswhy the entrepreneur approached the investor inthe first place. If he was not wealth-constrained,he could have financed the project himself.) Thus,while it may be relatively easy for the investor tobribe the entrepreneur to make a concession whenthe entrepreneur has control, it is harder for theentrepreneur to bribe the investor to make a con-cession when the investor has control. In fact wehave implicitly assumed that the entrepreneur hasno wealth, so that the only item of value he canoffer to give up is his fraction 0 of V; this may notbe enough to achieve efficiency. Note that he can'tgive up B directly because B is a nontransferableprivatebenefit.Since renegotiation complicates the basic story,without changing the fundamental message thatthe allocation of control matters, I will ignore it inwhat follows.


10/23

Hart: Financial Contracting 1087Consider the issue of contract designat date 0. The parties have two instru-ments at their disposal: the allocation ofcash flow rights, represented by 0, andthe allocation of control rights. (Forsimplicity, we have assumed that the

parties share cash flows in a linear man-ner, but nothing significant depends onthis-the investor could hold convert-ible, preferred stock, for example.) Forsimplicity assume that the entrepreneurmakes a take-it-or-leave-it offer to theinvestor at the contract-signing stage.Suppose also that both parties are riskneutral. Then the entrepreneur willchoose the contract to maximize hisexpected payoff subject to the inves-tor breaking even, i.e., recovering herinvestment cost C (on average).A simplification can be made. Sincethe investor's gross expected return isfixed at C, an optimal contract will alsomaximize the sum of the entrepreneurand investor's payoffs, i.e., (expected)social surplus, B + V, subject to theinvestor breaking even. It follows that,given two contracts, both of which havethe investor breaking even, the onethat generates greater expected socialsurplus is superior.It is useful to consider two polar con-tracts. At one extreme, suppose the en-trepreneur has all the cash flow rights(0 = 1) and all the decision rights. Thenthe entrepreneur's objective functionand the social planner's coincide, whichmeans that an efficient outcome is guar-anteed. Unfortunately, the investor getsnone of her money back! Thus, this con-tract is not feasible.At the other extreme, suppose thatthe investor has all the cash flow rights(0 = 0) and all the decision-makingrights. This contract maximizes the in-vestor's payoff and so the investor willmore than break even-or at least, if shedoesn't, the project can never go aheadat all. Unfortunately, this contract may

lead to the destruction of significantprivate benefits since the investor putsall the weight on cash flows. Note thatthis contract has a simple interpretation:the entrepreneur is a paid employee-he has no formal authority and gets aflat wage (actually zero!).The question is, where between thesetwo extremes does the optimal contractlie?There is one case where there is asimple answer. Suppose that, whateverdecision is taken at date 1, the projectyields a cash flow that is at least C (dis-counted back to date 0). Then the in-vestor can be given riskless debt withvalue C and the entrepreneur can be al-located all the equity, i.e., he is the re-sidual income claimant and has all thedecision rights. This contract is feasiblebecause the investor breaks even andoptimal because there is no inefficiency:the entrepreneur maximizesB + V.14Unfortunately, in a world of uncer-tainty, it is unlikely that the projectcash flows will be large enough to sup-port riskless debt of value C given anydecision. In order to understand what isoptimal then, imagine that the partiescan anticipate-and contract on-certainevents at date 1 (they are verifiable).'5An example of an event might be a situ-ation where the firm has low earningsand its product is not selling; in anotherevent the opposite may be true-thefirm has high earnings and its productis selling.

14 Note the importance of the condition that theproject yields at least C whatever decision istaken. For riskless debt to be optimal, it is notenough to suppose that the project can alwaysgenerate C ex post if some decision is taken; sucha decision might involve project termination, say,and the destruction of significant private benefits.In this case, it may be better to allocate decisionrights on an event-contingent basis, as describedbelow.15 An event is a subset of the set of all possiblestates of the world (i.e., all possible contingen-cies).


11/23

1088 Journal of Economic Literature, Vol. XXXIX(December 2001)The advantage of allocating cash flowand control rights to one party or theother will typically differ across theseevents. For example, in one event itmay be the case that a ruthless strategyof value (cash flow) maximization leads

to an approximately efficient outcomebecause private benefits aren't veryimportant. Recall the example whereclosing the firm down saved $L andwasted a private benefit of $100. If L =200, then indeed value maximizationgenerates an efficient outcome.On the other hand, in other events,private benefits may be relatively moreimportant, and value maximization maycause a significant loss of social surplus.This would be the case in the sameexample if L = 80.Aghion and Bolton show that theinvestor should have control-andcash flow rights-in the first kind ofevent, and the entrepreneur shouldhave control-and also possibly cashflow rights-in the second kind ofevent. The reason is that giving theinvestor control and cash flow rightsin the first kind of event generates anapproximately efficient (social surplusmaximizing) outcome and makes it eas-ier to satisfy the investor's break-evenconstraint; while giving the entrepre-neur control in the second kind of eventprevents inefficiency and hence is de-sirable as long as it is consistent withsatisfying the investor's break-even con-straint. (Giving the entrepreneur cashflow rights in the second kind of eventmay be useful to bring the entrepre-neur's objective function in line withthe social objective function.)A very rough summary of the Aghion-Bolton model is thus the following. Ifwe rank events from those where ruth-less value maximization is least ineffi-cient to those where it is most inefficient,then the investor should have control inthe first set and the entrepreneur in the

second set, where the cut-off is chosenso that the investor breaks even.How good a job does the Aghion-Bolton model do in explaining the fea-tures of real-world financial contracts?An interesting recent paper by StevenN. Kaplan and Per Stromberg (2001)argues that a good place to look is theventure capital sector (see also PaulGompers 1997).16 Venture capitalistsare private providers of equity capitalfor young growth-oriented firms (high-tech start-ups). Although venture capi-talists often represent several large richindividuals or institutions, they corre-spond quite well to the single investorof the Aghion-Bolton model. Similarly,the founder or founders of a start-upcompany can be represented withouttoo much of a stretch by a single entre-preneur. The distinguishing feature ofventure capital deals is that the majorparticipants have a close relationshipand are few in number.

Kaplan-Stromberg study 213 venturecapital (VC) investments in 119 port-folio companies (firms) by fourteen VCpartnerships. Most of these firms are inthe information technology and soft-ware sectors, with a smaller numberbeing in telecommunications. Kaplan-Stromberg's main findings (from ourpoint of view) are the following:(1) VC financings allow the parties to al-locate separatelycash flow rights, vot-ing rights, board rights, liquidationrights, and other control rights.(2) Cash flow rights, voting rights, con-trol rights, and future financings arefrequently contingent on observablemeasures of financial and nonfinan-

cial performance. For instance, theVCs may obtain voting control orboard control from the entrepreneurif the firm's EBIT-earnings before16 For related work on biotechnology alliances,see JoshuaLerner and Robert Merges (1998).


12/23

Hart: Financial Contracting 1089interest and taxes-falls below a pre-specified level or if the firm's networth falls below a threshold. Also,the entrepreneur may obtain morecash flow rights if the firm receivesapproval of a product by the Foodand Drug Administration (FDA) or isgranted a patent.(3) If the firm performs poorly, the VCsobtain full control. As firm perfor-mance improves, the entrepreneurretains/obtains more control rights.If the firm performs very well, theVCs retain their cash flow rights, butrelinquish most of their control andliquidation rights. The entrepre-neur's cash flow rights also increasewith firm performance.(4) VCs have less control in late roundsof financing (i.e., when the project isclose to completion).

At a broad level, these findings fitvery well with the Aghion-Boltonmodel. First, as that model emphasizes,cash flow rights and control (decision)rights are independent instruments, andindeed they are used independently:someone may be allocated significantcash flow rights without significant con-trol rights and vice versa. (To put itanother way, there can be a substantialdeviation from one share-one vote.)Second, as the Aghion-Bolton modelpredicts, to the extent that differentevents can be identified, the allocationof cash flow rights and control rightswill depend on these; here the eventscorrespond to performance as measuredby such things as earnings, net worth,or product functionality (FDA or patentapproval). Third, the fact that VCs havefewer control rights in late financingscan be understood as follows. In latefinancings, a firm requires less cashrelative to future profitability, i.e., theinvestment cost C is, in effect, lower.As we have seen, this makes it more

likely that the project cash flows cansupport something like riskless debt, inwhich case an efficient outcome can beachieved by giving all the control rightsand residual income rights to the entre-preneur. Under these conditions thereis no gain-and there can be a consider-able loss-from allocating control rightsto the investor.Interestingly, there is one strikingfinding of the Kaplan-Stromberg studythat, althoughconsistentwith the Aghion-Bolton model, does not necessarily fol-low from it. This is that control rightsand cash flow rights shift to the VC ifthe firm does poorly (number 3 in theabove list). This makes perfect sense ifwe can identify poor performance withan event where ruthless value maximi-zation leads to an approximatelyefficientoutcome, e.g., because private benefitsaren't very important relative to cashflows.17And indeed, this is quite plausi-ble, in the sense that in bad events itmay be efficient that the project be ter-minated or the entrepreneur removed asCEO, and this is exactly what a ruthlessvalue maximizer would do.However, things do not have to bethis way. It could be that ruthless valuemaximization leads to an efficient out-come in good events. For example,imagine that, if a start-up is very suc-cessful, the founding entrepreneur is nolonger the best person to run it, e.g.,because his creativity gets in the way ofthe professional approach to manage-ment that is now desirable. If the lossesfrom keeping the entrepreneur on arehigh enough, then it is efficient to re-place him. However, the entrepreneurmay resist replacement given his privatebenefit. Under these conditions, theonly way to obtain an efficient outcomeis to put control in the hands of the VC.

17 For a set of conditions guaranteeing this, seeHart and Moore (1998).


13/23

1090 journal of Economic Literature, Vol. XXXIX(December 2001)In other words, this is a case wherethe VC should have control if the firmperforms well, since it is in good eventsthat cash flows are important relative toprivate benefits.As noted, Kaplan-Stromberg do not

find this effect in the data, but thequestion is why? Possibly the answer isthat the Aghion-Bolton model ignoresan important variable: effort. That is, inreality, private benefits B and cashflows V are a function of ex ante effortas well as ex post decisions. An entre-preneur may have little incentive towork hard to ensure that a good eventoccurs, i.e., V is high, if his reward is tobe replaced by a ruthless investor.18 Inother words, ex ante effort considera-tions may explain why, empirically, con-trol shifts to the investor in bad ratherthan good events.4. Costly Intervention and the Diversityof Outside Claims

In section 3, we discussed how con-trol should be divided between an in-sider (the entrepreneur) and an out-sider (the investor). However, in manylarge companies in countries like theUnited States, the United Kingdom, orJapan, insiders, as represented by theboard of directors or management, donot have (voting) control in any stateof the world. Rather control rests withdispersed outside investors. Moreover,those outsiders hold diverse claims:some are shareholders and others arecreditors.In this section we discuss what maybe responsible for the diversity of out-side claims. We will argue that diversitycan be understood as part of an optimalmechanism when intervention by an

outside investor is costly (that is, the in-vestor has to expend time or resourcesto exercise control). Before we get intothe details of the analysis, it is worthemphasizing that neither the agencyapproach of section 2 nor the controlrights model of section 3 bears directlyon this question. The agency approach,as we have already argued, is really atheory of optimal incentive schemesrather than capital structure; while thecontrol rights model helps to explainthe optimal allocation of control be-tween insiders and outsiders, but notwhy, given a particular level of controlby insiders (in this case, zero), outsidershold heterogeneous claims, i.e., some areshareholders while others are creditors.In fact, one's first thought would bethat diversity is bad since it createsconflicts of interest between differentinvestors. Moreover, it is not clearwhy management should be affected bydiversity: why does it matter to themthat in good states of the world share-holders have control, while in badstates creditors have control (given thatmanagement never has control)?One approach to the diversity issue isbased on the existence of collective ac-tion problems. Imagine a large companythat has many (relatively small) share-holders. Then each shareholderfaces thefollowing well-known free-rider problem:if the shareholder does something toimprove the quality of management,then the benefits will be enjoyed by allshareholders. Unless the shareholder isaltruistic, she will ignore this beneficialimpact on other shareholders and sowill under-invest in the activity of moni-toring or improving management. Forexample, an individual shareholder willnot devote time and resources to per-suading other shareholders to vote toreplace an incompetent board of direc-tors. As a result, the management of acompany with many shareholders will

18 The entrepreneur may also have an incentiveto manipulate the accounts ex post, to make agood event look like a bad one, if he is likely to bereplaced in a good event (e.g., he could throwaway money).


14/23

Hart: Financial Contracting 1091be under little pressure to perform well.(The threat of a hostile takeover bid canovercome the shareholder passivity prob-lem to some extent, but, for all sorts ofreasons, is unlikely to eliminate it.)In contrast, individual creditors canin principle obtain the full benefits oftheir actions for themselves and thus donot face the same kind of free-riderproblem (at least outside bankruptcy).Suppose a creditor's debt is not repaid.Then she can seize some of the firm'sassets if her debt is secured; while ifher debt is unsecured she can obtain ajudgment against the firm and have asheriff sell off some of the firm's assets.She does not require other creditors toact.In fact,it is better forherif they do not,since there are then more assets to seize!So creditors impose discipline onmanagement in a way that shareholdersdo not. Specifically, whereas a managerwho faces a large number of small share-holders is unlikely to be penalized sig-nificantly if he fails to deliver highprofit or pay out large dividends, a man-ager who faces a large number of smallcreditors knows that he must repay hisdebts or he will be in trouble: his assetswill be seized or he will be forced intobankruptcy (which is assumed to beunpleasant for him). However, there isa trade-off: too much discipline can bebad. While some debt is good in orderto force management to reduce slack,too much debt is bad because it canlead to the bankruptcy and liquidationof good companies, and can preventmanagement from financing profitablenew projects. Various papers have ex-plored this trade-off and have used it toderive the optimal debt-equity ratio fora company.19The view that financial diversity oc-curs because of collective action prob-

lems is not entirely satisfactory for tworeasons. First, the existence of thesecollective action problems is assumed,not derived: in particular, it is supposedthat shareholders face these problemswhile creditors don't. However, thingsdon't have to be this way. One couldimagine a company that sets itself upso that each shareholder has the rightto liquidate a fraction of the company'sassets unilaterally-in fact we see justsuch an arrangementwith open-end mu-tual funds. Equally, one could imaginethat creditors are required to act by amajority vote to seize assets or push afirm into bankruptcy. If most compa-nies choose not to operate this way, weneed to explain this; we shouldn't justtake it as given.Second, most collective action mod-els of the trade-off between debt andequity assume that shareholders arecompletely passive. However, this viewis hard to square with the fact that com-panies routinely pay out cash to share-holders in the form of dividends andshare repurchases. If managers face nopressure from shareholders, one wouldexpect them to retain all their earnings:wouldn't they always be able to findsomething better to do with a dollarthan to pay it out to shareholders?There is a third problem with mostcollective action models of debt that isalso worth mentioning. In these modelsit is typically the case that debt mattersonly if a firm is close to bankruptcy.The reason is that, if not, then the firmcan pay off its current debts by borrow-ing against future income, i.e., currentdebt levels do not constrain manage-ment. However, the idea that debt mat-ters only if a firm is in extreme financialdistress does not seem very plausible.In recent years, Erik Berglof andErnst-Ludwig von Thadden (1994) andMathias Dewatripont and Jean Tirole(1994) have explored an alternative

19Representative contributions are Grossmanand Hart (1982), Jensen (1986), Myers (1977),Rene Stulz (1990), and Hart and Moore (1995).


15/23

1092 Journal of Economic Literature, Vol. XXXIX (December 2001)approachto diversitythat proceeds morefrom first principles. The basic idea be-hind these papers is that diversity isgood not because of the existence ofcollective action problems, but ratherbecause diversity changes incentives. Inparticular, suppose that a company hasa single investor (or group of homo-geneous investors)-say, a shareholderwith 100 percent control rights. Thisshareholder has the right and the abilityto intervene at any time; but assume, incontrast to what has gone before, thatintervention is costly. Then this investormay choose not to act because the costsof intervention exceed the benefits. Incontrast, if the company has several in-vestors with heterogeneous claims, it islikely that for at least one investor thebenefits of intervention exceed thecosts. If this investor also has the abilityto intervene, management will be underpressure. The conclusion is that hetero-geneous claimants can put more pres-sure on management than homogeneousclaimants when intervention is costly.It will be useful to present a verysimple model-in the form of a nu-merical example-that illustrates thisapproach. The model is based on (pre-liminary) joint work with Moore anddraws on the ideas of Jeffrey Zwiebel(1996), as well as those of Berglof andvon Thadden (1994) and Dewatripontand Tirole (1994).Let me begin with a verbal descrip-tion of the model. Consider a firm thathas some current earnings, and is alsoexpected to be profitable in the future.The manager or board of directors ofthe firm will have to decide how muchof the current earnings to pay out to in-vestors. It is plausible that the managerhas his own (selfish) reasons for notpaying out as much as the sharehold-ers would like, e.g., he might want toengage in empire-building activities orprotect against a possible future calam-

ity by buttressing the firm's financialposition. (The model below focuses onprotection against future calamitiesrather than empire-building.)20Suppose initially that the firm has nodebt, i.e., all investors are shareholders.Assume also that these shareholders havecontrol rights, do not face collectiveaction problems, and so could interveneto force the manager to disgorge someof the "free cash flow." However, inter-vention is costly, e.g., it requires theexpenditure of time or resources. Thenthe manager will pay out just enoughcash-d, say-to stop the shareholdersfrom intervening, i.e., such that the in-tervention cost equals the inefficiencygenerated by reinvesting earnings ratherthan paying them out.Now assume instead that the firmowes some money to short-term credi-tors that exceeds d-call the amount p.Suppose that creditors do not have anypowers that shareholders do not, i.e.,their cost of intervention is just thesame. The manager could announce thathe will pay the creditors less than whathe owes them-say d. However, credi-tors are unlikely to accept this-theywill choose to intervene. Why? The rea-son is that if they agree to accept drather than p, then the residual amountp - d will at best be postponed andpossibly even canceled (this is thenature of a debt claim). In either case,if we allow for discounting and uncer-tainty, creditors won't get much of theresidual. In contrast, if they intervene

20 Retaining some cash to protect against futurecalamities may be at least partly in the interest ofshareholders, of course. However, to the extentthat the manager obtains a private benefit fromcontinuation, he may retain excessive cash. Themodel focuses on this excessive element of cashretention. Excessive cash retention was a promi-nent feature of Kirk Kerkorian'sbattle against theChryslerboard in the 1990s; for another example,see the discussion of Japanese companies in theFinancial Times of October 17, 2000 ("TakeoverSpecialistTries a Different Tactic," p. 10).


16/23

Hart: Financial Contracting 1093t=O 1 2 3

Company set up Earningsy, = 50 Liquidityshockk EarningsY2 = 90at cost 56 kuniform (if firmsurvives)on [0,200]

Figure 1.now, they may be able to get all of theirp (assuming that the firm's cash flowplus asset value exceeds p). To this end,they will even be prepared to destroyvalue, e.g., liquidate productive assetsor cut off funds for good investmentprojects. The point is that creditors donot care about the firm's future profit-ability given that the beneficiaries of fu-ture profitabilityare shareholders ratherthan them.To put it very simply, shareholdersare soft because they are the residualincome claimants while creditors aretough because they are not.Now to the details of the model (or ex-ample). The model is a slightlymore com-plicated version of the Aghion-Boltonmodel of section 3, with the one impor-tant difference being that interventionis costly. There are four dates. At date 0the firm is set up at cost C. This amountmust be raised from outside investorssince the manager has no funds of hisown. At date 1 earnings of yi are real-ized. The' firm's manager (or board ofdirectors) then has a choice about howmuch of yl to pay out and how much toretain (retentions are placed in thefirm's bank account(s), which pay thegoing rate of interest). At this stage acontrolling outside investor can inter-vene to undo the manager's pay-outdecision-but this costs F. (F can be in-terpreted as the cost of learning wherethe firm'-sbank accounts are.) We sup-pose that the manager's motive forretaining funds is that the firm will behit by a liquidity shock at date 2. Whatthis means is that, because of s'ome

calamity (environmental, legal, etc.), thefirm will have to come up with $k tosurvive. Here k is a random variablewith a known distribution as of date 0;the realization of k becomes known (tothe manager and investors) at date 2.Finally, if it survives the liquidity shock,the firm earns y2 at date 3, which ispaid out to investors.For simplicity,we will assume C = 56,yi = 50, k is uniformly distributed on[0,200], y2 = 90, and the intervention costF = 18. Also, investors are risk neutral,and the marketinterest rate is zero.The time-line is illustrated in figure1. To simplify matters, we suppose thatthe manager's only interest is to maxi-mize his chances of surviving to date 3,i.e., of overcoming the liquidity shock.(In others words, he gets a fixed privatebenefit from running the firm betweendates 2 and 3 and he maximizes theexpected value of this benefit; he'suninterested in cash.) Given that themanager is uninterested in cash, it willnever be efficient for him to hold anyincome claims, i.e., all equity (in thesense of residual income rights) will beheld by outside investors.Denote by el the amount the man-ager pays out to investors at date 1 andby i = y1 - el the amount he retains. It'suseful to start with two polar cases.4.1 Investor OptimumFrom the point of view of investors(who hold all the income claims), thefirm should survive at date 2 if and onlyif k < 90. The reason is that this is therule that maximizes present value at


17/23

1094 Journal of Economic Literature, Vol. XXXIX(December 2001)date 2: the firm is worth saving only if itcosts less to save than it is worth (90).Moving backwards in time, we seethat this outcome can be achieved aslong as the firm pays out all its earningsat date 1. The point is that, at date 2,the manager can borrow up to 90against date 3 earnings; and this will en-sure that he can survive liquidity shockk if and only if k < 90.In other words, from the point ofview of investors, it is best to leave noslack in the firm at date 2. Under theseconditions, the firm's (present) value atdate 0 is given by

90VIO= 50 + 1/200 (90- k)dk=70.25.0

(IO stands for investor optimum.) Thefirst term represents the date 1 earningsthat are paid out, while the second termrepresents the expected going concernvalue from date 1 onwards: note that thefirm borrows k at date 2 whenever k 56, so that, if themanager can commit to the investor op-timum, the firm will be set up at date 0(there is enough value to compensatethe investors).4.2 Manager Optimum

Once the firm has been set up, themanager has a quite different goal fromthat of the investors: he wants the firmto survive to date 3. This means that hewants to retain as much earnings at date1 as possible. Suppose he retains i.Then he can add this to the 90 he canborrow against date 3 earnings and sur-vive any liquidity shock k such that k 140, the manager cannot save the firm anda question arises as to what happens to the date 1earnings of 50. We take the view that the managerengages in a partial rescue, specifically, he keepsthe firm going for a fraction X of the period be-tween dates 2 and 3, where 90 X + 50 = kk. Thatis, there are constant returns to scale with respectto time for X < 1, so that the cash injection re-quired to overcome the liquidity shock is propor-tional to the length of the period (k > 1 is as-sumed to be infeasible). Given this assumption,the date 1 earnings are totally dissipated, which iswhy they don't appearin the formu a for VMO.


18/23

Hart: Financial Contracting 1095shareholder decides whether to inter-vene. Intervention means that she canchoose to pay out more funds if she likes(obviously she'll choose to pay outeverything), but she has to incur theinterventioncost 18.

The manager will pay out just enoughto make the shareholder indifferent be-tween intervention and not. It is easy tosee that the equilibrium value of ei is10, i.e., i = 40. Given this the firm willbe saved at date 2 if and only if k < 130and its (present) value at date 0 will be130

Vs= 10 + 1/200 (130- k)dk= 52.25.0Note that Vio - Vs = 18, which is theintervention cost. In other words, it isindeed the case that, with ei = 10, theshareholder is just indifferent betweenintervening and not.Unfortunately,Vs < 56. In other words,total shareholder control is not enoughto get the firm financed at date 0!

4.4 Shareholder and Creditor ControlThere is a way to get the projectfinanced: it is to include a short-term creditor as well as a controllingshareholder.Suppose there is a creditor who isowed 20 at date 1. If the creditor is not

fully paid, she can choose to interveneat a cost of 18 (just like the share-holder). Assume that, if the creditorintervenes, she can seize retained earn-ings and pay herself the remainingamount she is owed and be reimbursedfor her intervention costs, i.e., if she isowed x, she can seize x + 18.Finally, suppose that, if the creditordecides not to intervene, her remainingdebts are canceled, i.e., she is entitledto nothing further at date 3.Some of these assumptions arestrong, but for our purposes this doesnot really matter. The debt claim can

always be structured with the above fea-tures and we are simply trying to showthat there is a way to get the firmfinanced at date 0.Note also that the assumption thatthe creditor can be reimbursed does notintroduce an asymmetry between thecreditor and the shareholder: the share-holder is also in effect reimbursed forintervening since, as the residual in-come claimant, she owns everything. Toput it another way, if the shareholderreimbursed herself formally, then shewould simply be transferring fundsfrom one pocket to another.

The timing is now as follows. First,the manager makes a payment to thecreditor-call the amountp. Second, thecreditor decides whether or not to in-tervene. Third, the manager makes apayment to the shareholder-call theamount d. Fourth, the shareholderdecides whether to intervene.I claim that the equilibrium is for themanager to pay 20 to the creditor andnothing to the shareholder and for nei-ther party to intervene. The reason isthe following. If the manager pays p


19/23

1096 Journal of Economic Literature, Vol. XXXIX (December 2001)90

30 + 1/200 (90- k)dk= 50.250

by seizing the remaining 30 in retainedearnings. But the gain from interventionequals 14.25, which is less than theintervention cost 18; i.e., shareholderintervention will not occur.So, under short-term debt and equity,the firm will be worth120

Vsc =20 +1/200 J(120 - k)dk= 560

at date 0, where the first-term representsdebt repayment and the second termrepresents the value of date 3 dividends.Since Vsc equals the date 0 investmentcost C, it follows that the firm can nowbe financed!Some comments on the model are inorder. First, it is worth rehearsing theintuition for the benefits of diversity. Asingle shareholder with full controlrights is not tough enough on manage-ment. The reason is that intervention iscostly and, although intervening per-mits the seizure of retained earningsthat would otherwise b-eused for unpro-ductive purposes (from the point ofview of the shareholder), the gross rateof return on these earnings is positive(in low k states they will be paid outas a dividend at date 3). So the gainsfrom seizing the retained earnings arenot that high. In contrast, a short-termcreditor has a very different objectivefunction: any funds left in the firm ac-crue to the shareholder, not to her (i.e.,their gross rate of return is zero), andso her incentive to intervene is muchgreater.

Note that it is the combination ofcash flow, rights and control rightsthat is vital here: it is important boththat the creditor has a claim that iscapped above (which makes her ineffect impatient), and that she has

the right to intervene in default states.A second comment concerns renego-tiation. One argument that can be lev-eled against the beneficial role we havefound for short-term debt and equity isthat we have ignored the possibility ofrenegotiation between the shareholderand the creditor. Suppose the managerdoes not repay the creditor the full 20she is owed, e.g., instead the managerpays only 10. It is easy to see that it isnot in the collective interest of theshareholder and the creditor for thecreditor to intervene since the slack inthe system (given by Vio - Vs) is only18, the cost of intervention. In fact, ifthe creditor intervenes, she is entitledto seize 28 and will gain 10 in netterms, while the shareholder's returnwill be reduced from

130 1021/200 (130 k)dk o 1/200 (102 k)dk,

0 0i.e., by 16.24. (The shareholder andcreditor's incremental payoffs sum toless than zero because the creditor seizesonly 28 rather than the 50 she wouldseize if the shareholder and creditorcould coordinate.)Since the shareholder's loss from in-tervention exceeds the creditor's gain,one might expect the shareholder tobribe the creditor not to intervene; i.e.,pay off her remaining debt. (One way todo this is through a debt-equity swap.)Of course, anticipating this, the man-ager has no incentive to pay the credi-tor the full 20 in the first place and thedisciplinary role of debt evaporates.Although this argument has someforce, it is far from decisive. What is re-ally involved is a matter of timing. It istrue that, if the shareholder has achance to bribe the creditor after themanager has decided how much to paythe creditor, then this reduces the man-ager's incentive to pay the creditor.


20/23

Hart: Financial Contracting 1097However, it is just,as plausible that theshareholder must decide whether tobribe the creditor before the managermakes his decision (i.e., the share-holder's move in the game comes first).In this case the shareholder won't bribethe creditor, knowing that, if she doesnot, the manager will prefer to pay thecreditor the full twenty rather than faceintervention.In other words, under the timingwhere the shareholder moves first, themodel is intact. Note that there is noinefficiency on the equilibrium path:any funds the manager pays out to thecreditor reduce slack and so the share-holder is happy to see these funds beingpaid out.Third, one might ask whether themanager could avoid the reduction inslack at date 1 by borrowing against fu-ture earnings. It is indeed possible forthe manager to borrow against futureearnings-since the firm is not close tobankruptcy-but this only makes mat-ters worse for him. To raise 20 at date 1the manager must promise d > 20 atdate 3 since the firm is not certain tosurvive the date 2 liquidity shock. Butthis means that at date 2 the conditionfor the firm to survive becomes 140 - d> k, since the manager arrives at date 2having already mortgaged d. As a resultthe manager is less likely to survivethan if he doesn't borrow, where thecondition for survivalis 120 > k.What the manager really wants to dois to issue new equity rather than bor-row. If the manager issues new equityof value 20, he can pay his date 1 debtwithout incurring any future obliga-tions: the condition for date 2 survivalbecomes k < 140 (given that he has 50in retained earnings). In fact, the man-ager can go further. Even if there is noshort-term debt, he could issue hugeamounts of new equity at date 1-to thepoint where the value of initial equity is

zero and use the proceeds to providea large financial cushion at date 2.Of course, arbitrarily large issues ofnew equity are not in the interest ofthe initial shareholder and so it makessense for the initial contract betweenthe manager and investors to limitthese. We have implicitly supposed thatno new equity issues are allowed at date1 (without shareholder permission), buta similar logic will work if a limitednumber of new shares can be issued,particularly if the time horizon extendsto greater than four dates: the managerthen faces a trade-off between issuingshares today to create more slack andissuing them in the future when an-other liquidity shock may hit the firm.Note that a limit on new equity is quiterealistic-in practice, companies are usu-ally authorized to issue a certain num-ber of new shares, but once this limit hasbeen reached, the company must getpermission from existing shareholdersto issue more (in our model, permissionwill not be given).A fourth comment concerns the costsand benefits of debt. In the model thebenefit of debt is that it is a way for themanager to commit to pay out free cashflow, which enables the manager to getthe firm financed; while the cost ofdebt is that the manager has less slackto guard against liquidity shocks, whichreduces his private benefit. However, inextensions of the model there would beother costs of debt. For example, if thedebt level at date 1 exceeded 50, thenthe manager would have to pay this byborrowing against date 3 earnings.However, this introduces debt overhangat date 2 (in the sense of Myers 1977):if d is owed at date 3, the condition forthe firm to survive the liquidity shockbecomes 90 - d > k, which means thatthe firm may fail to survive even whensurvival generates value for investors(survival generates value for investors


21/23

1098 Journal of Economic Literature, Vol. XXXIX (December 2001)whenever 90 > k). If the debt level be-comes even higher, the manager maybe forced to liquidate part of the firm atdate 1, i.e., sell off some key assets,which may again be inefficient.Finally, if the debt level becomeshuge, the firm may declare bankruptcy,and-depending on the bankruptcy pro-cedure-the firm may be turned over tothe creditor. However, this creates aproblem: the creditor will become theresidual income claimant and will act"soft" instead of "tough." So anothercost of debt is that, if it becomes toolarge, the disciplinary role of debt-depending as it does on the existenceof multiple claimants with conflictinginterests-is lost.This last observation has two interest-ing implications. First, it shows that inthis model moderate debt levels matter.If the debt level is very small, it doesnot affect what the manager pays out atdate 1 (this is true if p < 10); on theother hand, if the debt level is verylarge, the disciplinary role of debt islost. As noted earlier, this conclusioncontrasts with that of most collective ac-tion models of debt in the literature,where debt matters only when a firm isclose to bankruptcy. Second, for debt tohave a role it is essential that the firmcannot declare bankruptcy too easily.Specifically, a "no-fault"procedure thatallows any firm to declare bankruptcyand carry out an automatic debt-equityswap will be counterproductive sincethe manager can avoid the disciplinaryrole of debt. This may provide somejustification for the idea that a firm thatwants to declare bankruptcy shouldhave to convince a disinterested party, ajudge, say, that it cannot pay its debts.As a last comment, we should notethat the above model is consistent withthe payment of dividends (or repur-chase of shares-we have not distin-guished between the two)-something

which, as we have pointed out, manymodels of the debt-equity trade-off arenot.22 This is clearly true if the amountowed at date 1, p, is less than 10, sincewe saw that the manager will have topay the shareholder 10 - p to stop herfrom intervening. At the optimum p =10 and the dividend is zero. However, ifwe allow for uncertainty in yi, debt re-payments and dividends can both occurat an optimum. Suppose yi can be highor low. There is a class of cases suchthat when yi is low p is paid to thecreditor and there is no dividend andwhen yi is high p is paid to the creditorand on top of this a dividend is paid tothe shareholder.Whether this theory of dividends isadequate to explain the data is anothermatter. The theory suggests that thepayment of dividends should be quiteirregular and the amounts far from con-stant. There is, of course, a large em-pirical literature that finds the opposite:dividends are regular and smooth. How-ever, the recent evidence suggests thatthings are changing. Also, the idea thatdividends are the result of pressurefrom shareholders receives support froma recent cross-country comparison byRafael La Porta et al. (2000).

5. ConclusionsLet me conclude briefly. I havediscussed how economists' views offirms' financial structure decisions haveevolved from treating firms' profit-ability as given, to acknowledging thatmanagerial actions affect profitability,to recognizing that firm value dependson the allocation of decision or control

rights. I have tried to show that the de-cision or control rights approach is use-ful, even though it is at an early stage ofdevelopment, and that this approachhas some empirical content: it can22 Zwiebel (1996) is an exception.


22/23

Hart: Financial Contracting 1099throw light on the structure of venturecapital contracts and the reasons for thediversity of claims.I have been quite selective in thetopics I have covered. For instance, Ihave not discussed research on whycompanies in most countries other thanthe United States, the United Kingdom,and Japan often have controlling share-holders and exhibit deviations from oneshare-one vote. There has been somevery interesting recent empirical andtheoretical work on this topic,23 but adiscussion of this will have to wait foranother paper.

REFERENCESAghion, Philippe and Patrick Bolton. 1992. "An'Incomplete Contracts' Approach to FinancialContracting,"Rev. Econ. Stud. 59, pp. 473-94.Allen, Franklin and Douglas Gale. 2000. Compar-ing Financial Systems. Cambridge, MA: MITPress.Bebchuk, Lucian. 1999. "A Rent-Protection The-ory of Corporate Ownership and Control,"workingpaper 7203, NBER.Berglof, Erik and Ernst-Ludwig von Thadden.1994. "Short-Termversus Long-TermInterests:Capital Structure with Multiple Investors,"Quart. J. Econ. 109, pp. 1055-84.Berra, Yogi. "FamousYogi Berra Quotes," online:http://www.yogiberraclassic.org/.Bolton, Patrick and David Scharfstein. 1990. "ATheory of Predation Based on Agency Problemsin Financial Contracting,"Amer. Econ. Rev. 80,pp. 94-106.Dewatripont, Mathias and Jean Tirole. 1994. "ATheory of Debt and Equity: Diversity of Securi-

ties and Manager-Shareholder Congruence,"Quart. J. Econ. 109, pp. 1027-54.Diamond, Douglas W. 1991. "Debt MaturityStructure and Liquidity Risk,"Quart. J. Econ.106, pp. 709-37.Dybvig, Philip H. and Jaime F. Zender. 1991."Capital Structure and Dividend Irrelevancewith Asymmetric Information," Rev. Finan.Stud. 4, pp. 201-19.Gale, Doug as and MartinHellwig. 1985. "Incentive-Compatible Debt Contracts: The One-PeriodProblem," Rev. Econ. Stud. 52, pp. 647-69.Gompers, Paul A. 1997. "Ownership and Controlin Entrepreneurial Firms: An Examination ofConvertible Securities in Venture Capital In-vestments,"HarvardU. Grad. School Bus.

Grossman, Sanfordand Oliver Hart. 1982. "Corpo-rate Financial Structure and Managerial Incen-tives" in The Economics of Information andUncertainty. John J. McCall, ed. Chicago: U.Chicago Press, pp. 107-40.. 1986. "The Costs and Benefits of Owner-ship: A Theory of Vertical and Lateral Integra-tion,"J. Polit. Econ. 94, pp. 691-719.Harris, Milton and Artur Raviv.1991. "TheTheoryof Capital Structure,"J. Finance, 46, pp. 299-355.Hart, Oliver. 1995. Firms, Contracts, and Finan-cial Structure.Oxford:Oxford U. Press.Hart, Oliver and John Moore. 1990. "PropertyRights and the Nature of the Firm," J. Polit.Econ. 98, pp. 1119-58.. 1994. "A Theory of Debt Based on theInalienability of Human Capital," Quart. J.Econ. 109, pp. 841-79.. 1995. "Debt and Seniority: An Analysisof the Role of Hard Claims in ConstrainingManagement," Amer. Econ. Rev. 85, pp. 567-85. . 1998. "Default and Renegotiation: A Dy-namic Model of Debt," Quart. J. Econ. 113, pp.1-41.Jensen, Michael. 1986. "Agency Costs of FreeCash Flow, Corporate Finance and Takeovers,"Amer. Econ. Rev. 76, pp. 323-29.Jensen, Michael and William Meckling. 1976."Theory of the Firm: Managerial Behavior,Agency Costs, and Ownership Structure," J.Finan. Econ. 3, pp. 305-60.Kaplan, Steven N. and Per Stromberg. 2001. "Fi-nancial Contracting Theory Meets the RealWorld: An Empirical Analysisof Venture Capi-tal Contracts,"U. Chicago Grad. School Bus.La Porta, Rafael; Florencio Lopez-de-Silanes, An-drei Shleifer and Robert Vishny. 1998. "Lawand Finance,"J.Polit. Econ. 6, pp. 1113-55.. 2000. "Agency Problems and DividendPolicies around the World,"J. Finance 55, pp.1-33.Lerner, Joshua and Robert P. Merges. 1998. "TheControl of Technology Alliances: An EmpiricalAnalysisof the Biotechnology Industry,"J.Ind.Econ. 46, pp. 125-56.Miller, Merton. 1977. "Debt and Taxes,"J. Fi-nance 32, pp. 261-75.Modigliani, Franco and Merton H. Miller. 1958."The Cost of Capital, Corporation Finance andthe Theory of Investment," Amer. Econ. Rev.48, pp. 261-97.Myers, Stewart. 1977. "Determinants of CorporateBorrowing,"J.Finan. Econ. 5, pp. 147-75.Myers, Stewart and Nicholas Majluf. 1984. "Cor-porate Financing and Investment Decisionswhen Firms Have Information that InvestorsDo Not Have," J. Finan. Econ. 13, pp. 187-221.Persons, John C. 1994. "Renegotiation and the Im-possibility of Optimal Investment,"Rev. Finan.Stud. 7, pp. 419-49.Rajan, Raghuram G. and Luigi Zingales. 1995.23 See the many Papers by La Porta et al. (e.g.,La Porta et al. 1998); and LucianBebchuk (1999).


23/23

1100 Journal of Economic Literature, Vol. XXXIX (December 2001)"What Do We Know about Capital Structure?Some Evidence from International Data," J.Finance 50, pp. 1421-60.Shleifer, Andrei and Robert W. Vishny. 1997. "ASurvey of Corporate Governance," J. Finance52, pp. 737-83.Stiglitz, Joseph E. 1974. "On the Irrelevance ofCorporate Financial Policy,"Amer. Econ. Rev.64, pp. 851-66.Stulz, Rene. 1990. "Managerial Discretion and

OptimalFinancing Policies,"J. Finan. Econ. 26,pp. 3-27.Townsend, Robert. 1978. "OptimalContracts andCompetitive Markets with Costly State Verifica-tion,"J. Econ. Theory 21, pp. 265-93.Zingales, Luigi. 2000. "In Search of New Founda-tions,"J. Finance 55, pp. 1623-53.Zwiebel, Jeffrey. 1996. "Dynamic Capital Struc-ture under Managerial Entrenchment," Amer.Econ. Rev. 86, pp. 1197-215.

Date post:	05-Apr-2018
Category:	Documents
Upload:	obeidhamza
View:	219 times
Download:	0 times

Hart 2001 JEL Financial Contracting

Documents