Parties bound by an incomplete contract have an incentive to renegotiate after acquiring new information. The issue of the parties' investment in the relationship before renegotiation is analyzed in a simple two-period procurement model. The firm invests in the first period. It then learns its production cost, and the sponsor learns its value for the project. Williamson's underinvestment presumption is shown to hold under very general assumptions about bargaining and about the ex post asymmetry of information as long as the firm's investment is not observable by the sponsor. The introduction of a cancellation fee may well lead to even less investment contrary to what is sometimes argued. The role of ex ante price fixing as an alternative way to reintroduce some form of commitment and the problem of cost overruns are discussed. Last, it is shown that if investment is observable by the sponsor andthus may become a joint decision variable, the two parties may choose to under- or overinvest.

Parties bound by an incomplete contract have an incentive to renegotiate after acquiring new information. The issue of the parties' investment in the relationship before renegotiation is analyzed in a simple two-period procurement model. The firm invests in the first period. It then learns its production cost, and the sponsor learns its value for the project. Williamson's underinvestment presumption is shown to hold under very general assumptions about bargaining and about the ex post asymmetry of information as long as the firm's investment is not observable by the sponsor. The introduction of a cancellation fee may well lead to even less investment contrary to what is sometimes argued. The role of ex ante price fixing as an alternative way to reintroduce some form of commitment and the problem of cost overruns are discussed. Last, it is shown that if investment is observable by the sponsor andthus may become a joint decision variable, the two parties may choose to under- or overinvest.

This paper shows that direct (product) competition acts as a tournament between retailers when informational problems (or transaction costs) prevent the manufacturer from using contracts based on their relative performances. Anticompetitive restraints such as exclusive territories and resale price maintenance (which, we show, are not necessarily good substitutes) may or may not be privately desirable. It also shows that privately desirable anticompetitive restraints may not be socially desirable.

The paper emphasizes the use of accounting data in regulatory or procurement contracts when the supplier (1) has superior information about the cost of the project and (2) invests in cost reduction. The main result states that, under risk neutrality, the supplier announces an expected cost and is given an incentive contract linear in cost overruns. This (optimal) contract moves toward a fixed-price contract as the announced cost decreases. An investment choice is then introduced and the use of a rate-of-return regulation is studied.

The paper emphasizes the use of accounting data in regulatory or procurement contracts when the supplier (1) has superior information about the cost of the project and (2) invests in cost reduction. The main result states that, under risk neutrality, the supplier announces an expected cost and is given an incentive contract linear in cost overruns. This (optimal) contract moves toward a fixed-price contract as the announced cost decreases. An investment choice is then introduced and the use of a rate-of-return regulation is studied.

We develop a duopoly model in which exit occurs because of the existence of fixed costs or opportunity costs. Each firm enters the market knowing its own cost, but not that of its opponent. As times goes on, each firm becomes increasingly pessimistic about the cost of its remaining rival. The time of exit is the only strategic variable, so that our model is a of attrition. In contrast to the classic war of attrition, however, we assume that with positive probability each firm's costs may be low enough that staying in forever is a dominant strategy. Thus our model, unlike the classic one, has a unique equilibrium.