Oliver Hart Prize Lecture: Incomplete Contracts and Control

---

371

Incomplete Contracts and Control

Prize Lecture, December 8, 2016

by Oliver Hart

Department of Economics, Harvard University, USA

1. INTRODUCTION

The work on incomplete contracts cited by the prize committee began in the 

summer of 1983, but it may be useful to say a bit about how I reached that point. 

As a graduate student, first at the University of Warwick and then at Princeton 

University, with a degree in mathematics behind me, I was drawn to general 

equilibrium theory, and my Ph.D. thesis was on general equilibrium theory with 

incomplete markets.

1

 Although I ended up focusing on optimality and existence 

problems that could arise even in exchange economies, one of my primary inter-

ests was the theory of production. In a complete market Arrow-Debreu economy 

with perfect competition, it makes sense for a firm to maximize profit or net 

market value. But with incomplete markets, what is the generalization of this 

goal? More fundamentally, what happens if shareholders disagree about what 

the firm should do?

2

I started to work on this topic after my thesis and as a result of a serendipitous 

assignment continued the work with Sanford Grossman in the summer of 1976.

3

At some point we decided that, interesting though the disagreement between 

shareholders was, an empirically more important conflict was that between 

managers and shareholders.

4

 This led first to work on corporate takeovers as 

a mechanism for disciplining management and then to a paper on debt as a 

bonding device.

5

 At some stage we realized that since we were studying ways to 

incentivize management, maybe we should analyze directly the optimal incentive 

scheme between an owner and a manager. This led to Grossman and Hart (1983), 

a paper squarely in the principal-agent tradition.

---

372 

The Nobel Prizes

This rather circuitous path helps to explain how my thinking evolved from 

markets to contracts as the unit of analysis, and provides the backdrop to the 

summer of 1983. Sitting in Grossman’s University of Chicago office, the two of us 

were considering what to work on next. After some discussion we decided that 

a question that was ripe for analysis was: Why would one firm ever buy another 

firm rather than conduct business with that firm through a contract? In other 

words, what are the limits of contracts and why do we have firms?

Of course, this was hardly a new question: there is a literature on the bound-

aries of the firm that goes back to Coase (1937) and includes Oliver Williamson’s 

many works (see, e.g., Williamson (1975)), and Klein et al. (1978). I think that 

it is fair to say that we were aware of this literature without being intimate with 

it. (Intimacy came later.) One thing that we knew for sure is that the literature 

was informal (or in the prize committee’s felicitous language, “not formalized”). 

As economic theorists with a formal training we thought that we might be able 

to add something.

We worked on the difference between firms and contracts for ten very intense 

days. With apologies to John Reed these were ten days that shook my world.

6

 I 

recall that initially we thought that the difference had to do with authority. An 

employer can choose the task of an employee.

7

 But what is the conceptual dif-

ference between this and a requirements contract between two firms where one, 

the buyer say, can choose how many widgets q to buy from the other, the seller, 

with payment determined by a pre-agreed schedule: p = p(q)? One could say 

that the buyer has authority over q in this case. Perhaps more seriously, once we 

are in a world where a buyer obtains private information about a demand shock 

and a seller obtains private information about a supply shock, then, according to 

mechanism design theory, the quantity q should depend on both parties’ reports 

about their shocks: neither party should have authority over q.

At some point it dawned on us that we were thinking about the issue the 

wrong way. We were viewing the problem in complete contracting terms. But 

what if the contract between the buyer and seller is incomplete?

2. INCOMPLETE CONTRACTS

The formal literature to that point was all about complete contracts. These are 

contracts where everything that can ever happen is written into the contract. 

There may be some incentive constraints arising from moral hazard or asym-

metric information but there are no unanticipated contingencies.

Actual contracts are not like this, as lawyers have realized for a long time. 

They are poorly worded, ambiguous, and leave out important things. They are 

---

Incomplete Contracts and Control 

373

incomplete. At some stage Grossman and I realized that a critical question that 

arises with an incomplete contract is, who has the right to decide about the 

missing things? We called this right the residual control or decision right. The 

question is, who has it?

Further thought led us to the idea that this is what ownership is. The owner 

of an asset has the right to decide on how the asset is used to the extent that its 

use is not contractually specified. This naturally leads to a theory of the difference 

between contracts and firms. Think of a firm as consisting of assets. If firm A 

and firm B sign an arms-length (incomplete) contract, then the owner of firm A 

has residual control rights over the A assets and the owner of firm B has residual 

control rights over the B assets. In contrast if, say, firm A buys firm B, then the 

owner of firm A has residual control rights over the A and B assets.

Why should it matter who has residual control rights? Residual control 

rights are like any other good: there is an optimal allocation of them. Sometimes 

it is more efficient for one owner to hold all the residual control rights, and 

sometimes it is more efficient for these control rights to be split between several 

owners. Which is the case will determine whether firms A and B should merge 

or stay as separate entities.

Grossman and I constructed a formal model along these lines (see Grossman 

and Hart (1986)), and I developed the ideas and model further in work with John 

Moore (see Hart and Moore (1990)). Collectively these papers are often referred 

to as “property rights theory” (PRT).

It is useful to illustrate the model with a real-world example. Consider a 

power plant that locates next to a coal mine with the purpose of burning coal 

to make electricity.

8

 One way to regulate the transaction is for the power plant 

to sign an arms-length long-term contract with the coal mine. Such a contract 

would specify the quantity, quality, and price of coal for many years to come. But 

any such contract will be incomplete. Events will occur that the parties could not 

foresee when they started out.

For example, suppose that the power plant needs the coal to be pure but that 

it is hard to specify in advance what purity means, given that there are many 

potential impurities. Imagine that ten years into the relationship, ash content is 

the relevant impurity and that high-ash-content coal is more expensive for the 

power plant to burn than low-ash-content coal but cheaper for the coal mine to 

produce. Given that the contract is incomplete, the coal mine may be within its 

rights under the contract to supply high-ash-content coal.

The power plant and coal mine can, of course, renegotiate the contract. How-

ever, the coal mine is in a strong bargaining position. It can demand a high price 

for switching to low-ash-content coal. The reason is that the power plant does not 

---