Can Contract Theory Explain Social Preferences?

For several decades, a growing body of research has shown that humans do not always

choose to maximize material payoffs. Economists following the lead of psychologists Daniel

Kahneman and Amos Tversky (1979) and Matthew Rabin (1993) have built on such research

to suppose that individuals are attentive to fair distribution rewards between themselves as

well as personal payoffs. (Ernst Fehr and Klaus Schmidt (1999)) An alternative approach,

suggested by Elizabeth Hoffman, Kevin McCabe and Vernon Smith (1996) argue that laboratory

subjects that perceive a potential for future interaction as approximated by social

distance act on a preference for reciprocity.

Both approaches capitalize on the power of psychology to enhance understanding of

economic exchanges between particular people in specific laboratory conditions. Yet, psychology

is not only concerned with the problem of explaining behavior, but it is also in the

business of modifying behavior. On of the shelves of the psychology section of any bookstore

one will find numerous works devoted to helping individuals modify their behavior so that

they drink and smoke less, and learn to have more rewarding social interactions at work and

in private life.

Economists suppose that individuals maximize rewards not because they believe that

people do so in every case, but because the maximizing supposition provides a useful unified

model of behavior in designing better economic institutions. The model not only rationalizes

the profit motive, but explains how even successful institutions may be destroyed by the

common human desire to maximize possessionsвЂ"if that leads to malfeasance, theft, or general

corruption.

In this note I suggest that economics may also learn more frompsychology about positive

behaviors вЂ" such as the complex dynamics of honesty, fairness, and trust. Rather than insist

1

that people are endowed with a stable set of unchanging preferences, we may ask instead

how small modifications in preferences can lead to significant improvements in economic

performance.

Specifically, I suggest that contract theory can be a starting point for a larger enquiry

into exactly how preferences might be modified to enhance market performance. I consider

two substantive themes. First I show how contract theory may help to anchor a theory of

fairness. Second, I show that a small taste for trustworthiness can lead to a large increase

in cooperation in a relational contract.

Fairness is a crucial theme because notions about it are decisive in determining whether

a party has breached an agreement. For example, an authoritative and widely used treatise

on contract law by Edward A. Farnsworth has twenty index entries under the heading “Good

Faith and Fair Dealing”. The Uniform Commercial Code of the United States requires all

parties to a contract to act in “good faith,” defined to mean “the observance of reasonable

commercial standards of fair dealing. ”

Yet “fair” is never precisely defined. The next section outlines a model of fairness developed

with Lorne Carmichael (Carmichael and MacLeod (2003)). Following the insights

of Williamson (1975) and Hart (1995), we suggest that one can develop a concept of fairness

based upon the idea that it is optimal to reward sunk investment, and hence “fair” bargains

should take this into account. This may help explain the observations of Kahneman,

Knetsch and Thaler (1986) and many others who identify the endowment effect - individuals

value more highly assets they own than those with similar attributes that they must purchase.

Contract theory predicts that this effect arises from the investment, either psychic or

pecuniary, that owner has made into the asset that increases the value from ownership.

My second point begins with the observation that we invest in teaching our children

to abide by their obligations as a matter of principle, a behavior that makes little sense if

we expect them to simply maximize their material payoff. Such training will never produce

completely trustworthy offspring, but it may be expected to produce children who experience

2

some remorse or disutility from breaking promises. ("You’re grounded.") In section II, I

modify a behavioral model introduced by Hart and HolmstrÐ"¶m (1987) that supposes a few

individuals are inherently honest but most not. We relax this assumption and suppose that

all individuals experience some disutility from breaching a contract.

Section II shows that a small amount of remorse is enough to support a high level of

cooperation in a long term relationship. Moreover, the theory makes some predictions on

the form of the optimal contract that