Oliver Hart Prize Lecture: Incomplete Contracts and Control

382 

The Nobel Prizes

If the government owns the prison the same problem does not arise. Just 

as the power plant can require the provision of low-ash-content coal if it owns 

the coal mine, the government can forbid the prison warden from hiring cheap, 

unskilled guards.

Hart et al. (1997) cite some evidence that the level of violence is indeed 

higher in private prisons.

18

Of course, private provision can yield some benefits too. In the case of the 

coal mine, we argued that the manager will have a greater incentive to innovate 

when the mine is independent. The same is true of prisons. The warden of a pri-

vate prison will have a greater incentive to find socially efficient ways of saving 

money or to develop socially valuable rehabilitation programs.

19

 This might tip 

the balance in favor of private provision in situations where innovation matters 

and violence is a relatively small problem, e.g., half-way houses or youth correc-

tional facilities. However, in maximum security prisons, where the prevention 

of violence by prisoners against guards and other prisoners is paramount, Hart 

et al. (1997) conclude that the case for private provision is weak.

Hart et al. (1997) use the same logic to argue that private provision makes 

sense for garbage collection, does not make sense for the army, police or foreign 

policy, and may or may not make sense for schools and health care. Competi-

tion strengthens the case for privatization since actions that reduce quality will 

elicit a negative market response. Competition may work fairly well in the case of 

schools and hospitals, but it is hard to imagine it operating in the case of prisons.

Perhaps the most valuable lesson of the analysis is that it suggests that the 

public-private choice should be seen as a matter of efficiency, not ideology.

5. FOUNDATIONS

The property rights theory described in Section 2 is based on the idea that own-

ership of non-human assets is a source of bargaining power when contracts are 

incomplete. The formal models of Grossman and Hart (1986) and Hart and 

Moore (1990) justified contractual incompleteness by appealing to the idea that 

it may be difficult to describe in advance what kind of good a buyer wants from 

a seller; this may depend on a future state of the world and there may be many 

such states. In contrast, once the state is realized, it is easy to describe the good 

and hence a perfect spot contract can be written. Unfortunately, when bargain-

ing takes place ex post, ex ante investments will already have been sunk, hold-

up is possible, and, anticipating this, the parties will choose these investments 

inefficiently.

The challenge is to turn this informal story into a formal one.

---

Incomplete Contracts and Control 

383

It turns out that this is not easy. Consider first the issue of bargaining power. 

Return to the power plant-coal mine example. Suppose that it is desirable for 

the power plant to be in a strong bargaining position ex post in order to give it 

an incentive to locate next to the coal mine ex ante. Why not allocate bargaining 

power directly rather than through asset ownership? For example, the initial 

contract could specify that whenever renegotiation occurs it will proceed by the 

power plant making an offer to the coal mine that the coal mine can only accept 

or reject. Any attempt by the coal mine to make a counter-offer will be heavily 

penalized.

Indeed imagine that this is all the contract says; it does not specify the quan-

tity, quality or price of coal at all!

Consider the situation where the power plant wants the coal mine to produce 

low-ash-content coal. Then the power plant will make an offer to the coal mine 

for low-ash-content coal in return for a payment that slightly exceeds the cost 

c of mining low-ash-content coal; say the price p = c + ε. Assume that there is a 

deadline by which the deal must be done. The coal mine can always threaten to 

reject such a low offer, but the power plant can ignore such threats, confident that 

at the last moment the coal mine will accept the offer since ε is better than noth-

ing. This is the power of subgame perfection. As a result the power plant obtains 

all the ex post surplus from the transaction, and will locate next to the coal mine.

Indeed, as Hart and Moore (1988) show, it is not even necessary to stipulate 

in the contract that the coal mine will be penalized for making counter-offers. 

The contract could instead state that both sides can make offers and that neither 

party needs to accept an offer before trade occurs; instead they can trade and 

then produce a signed acceptance later. Suppose that absent any signed docu-

ment, trade is assumed to be a gift from the coal mine to the power plant. Imag-

ine now that the power plant makes the proposal described above while the 

coal mine proposes a much higher price, p′, say. Then trade will occur at price 

p. Why? On the one hand the coal mine can ensure itself p by trading and then 

signing and disclosing the power plant’s offer in the event of a dispute. On the 

other hand the power plant will never sign and disclose the coal mine’s proposal 

in the event of a dispute since it would prefer to reveal nothing and claim that 

the coal is a gift.

Of course, allocating all the bargaining power to the power plant gives it 

an incentive to locate next to the coal mine, but it does not give the manager of 

the coal mine an incentive to innovate. Suppose that the manager finds a way 

to reduce the cost of coal from c to c′. Then the power plant will change its offer 

from c + ε to c′ + ε. The coal mine’s innovation will be completely expropriated 

by the power plant. This suggests that asset ownership may be useful after all. 

---