Oliver Hart Prize Lecture: Incomplete Contracts and Control

374 

The Nobel Prizes

have a good alternative: it may be very expensive for the power plant to transport 

coal from a different coal mine given that it is located next to this one.

Economists refer to this situation as the “hold-up” problem. The coal mine 

can hold up the power plant because the power plant, by locating next to the coal 

mine, has become dependent on it. The next point to realize is that although it 

may be impossible to write a contract that is complete enough to avoid hold-up, 

this does not mean that the parties will be unable to anticipate hold-up. Indeed, 

the assumption made in the theory is that the power plant does anticipate that 

it will be at the mercy of the coal mine, and that a substantial share of its future 

profit may be expropriated by the coal mine. Fearing such expropriation, the 

power plant may choose not to become so dependent on the coal mine in the 

first place. For example, it may locate at an equal distance between several coal 

mines rather than right next to this one, even though this may increase the cost 

of transporting coal.

At the risk of belaboring the point, it is worth pinpointing the source of 

the coal mine’s hold-up power. It arises because the owner of the coal mine has 

residual rights of control over the mine. In this case the key residual right of 

control is the decision about what kind of coal to mine: high-ash-content or 

low-ash-content.

Is there anything the power plant can do to avoid this situation? Short of 

writing a better contract, one thing it can do is to buy the coal mine in advance. 

That way the power plant as owner of the coal mine will have the key residual 

control right. The coal mine can no longer extract a high price by threatening to 

produce high-ash-content coal: the power plant can order the coal mine man-

ager to mine low-ash-content coal. In an extreme case, if the coal mine manager 

threatens to disobey the order, the power plant can fire the manager and replace 

her with someone else.

One consequence is that the power plant may now be willing to become 

dependent on the mine. Given that it does not fear hold-up, it may locate next 

to it. Thus the theory identifies a benefit of integration, where integration in this 

case means the purchase of the coal mine by the power plant. The value of inte-

gration is that the power plant may undertake efficiency-enhancing relationship-

specific investments—in this case locating next to the coal mine—that it would 

not carry out if it was protected only by an incomplete contract.

So far we have discussed the benefits of integration. But just as the transfer 

of residual control rights from the coal mine to the power plant empowers the 

owner of the power plant, it disempowers the owner of the coal mine, and this is 

likely to have costs in terms of her incentive to make relationship-specific invest-

ments. Assume that the coal mine was previously an owner-managed firm. After 

---

Incomplete Contracts and Control 

375

the acquisition by the power plant the coal mine manager stays on but is now an 

employee of the power plant. Suppose that the coal mine manager has an idea 

about how to run the mine more efficiently. When the coal mine was separate, 

the manager had the authority (residual control rights) to implement, and benefit 

from, this idea. Now that the manager is an employee she has to get permission to 

implement the idea from her boss: the owner of the power plant has veto power. 

The owner of the power plant can use his veto right to extract some of the gains 

from the idea for himself. Knowing that she faces the risk of expropriation, the 

coal mine manager’s incentive to innovate is reduced.

So integration has costs as well as benefits. Whether it makes sense for the 

power plant to purchase the coal mine will depend on whether the distortion 

in the power plant’s investment is more important than the distortion in the 

coal mine manager’s investment. It is also worth noting that a further possibil-

ity is for the coal mine to purchase the power plant. This is not the same as the 

power plant purchasing the coal mine since now the residual control rights are 

concentrated in the hands of the coal mine manager rather than the power plant 

manager. Finally, the theory can be generalized beyond the case of two managers 

and two assets to many assets and many workers, and to more general ownership 

structures, such as joint and shared ownership; see Hart and Moore (1990). That 

paper also shows that synergistic assets should be owned together and that assets 

should be owned by indispensable human capital.

Let us pause to make some observations. First, the formal models in Gross-

man and Hart (1986) and Hart and Moore (1990) take the view that ex post 

renegotiation of an incomplete contract occurs under conditions of symmetric 

information—both parties can see what has been left out of the contract—and 

that, given that there are no wealth constraints, the bargaining proceeds effi-

ciently, as in Coase (1960). Inefficiencies arise solely because ex ante relationship-

specific investments are distorted.

Second, the distortion in ex ante investments can be overcome if these invest-

ments are contractible. In this case the parties can write a contract that speci-

fies, say, that the power plant must locate next to the coal mine in return for an 

upfront payment: in effect the mine compensates the power plant for its later 

hold-up power. For the theory to work, one has to suppose that some aspects of 

the investment are not contractible (or are costly to contract on): for example, 

even if the location decision is contractible, whether the power plant installs a 

boiler that burns this particular mine’s coal efficiently might not be. Similarly, 

one must assume that the coal mine manager’s investment in innovation is non-

contractible (a highly plausible assumption).

Third, the focus on distortions in noncontractible investments distinguishes 

---