A reasonable model of the labour market over the business cycle should predict, among other things, that (a) in very low states of product demand there may be too little employment from an efficiency perspective, but as the state improves employment will increase until ultimately it is efficiently deployed, and (b) in low states of demand, a worker's welfare level will be "low" and as the state of the world improves so will the worker's welfare, except, possibly, at high levels of demand where the worker's utility may start to fall. Surprisingly, there does not exist a labour contract based model that is consistent with predictions (a) and (b). In fact, the standard results in the literature are if leisure is a normal good then there will be too much employment in essentially all states of the world and the welfare of the worker declines as the state of the world improves. In this paper a labour contracting model is constructed that is consistent with the above mentioned predictions. Two necessary ingredients in the model are the possibility of financial distress in low demand states and "partial provability" in contracting. Financial distress can be viewed as frustrating renegotiation and, thus, inefficient outcomes are possible in equilibrium. Partial provability—the ability of an informed player to make verifiable claims or statements to an uninformed player—eliminates certain kinds of inefficient outcomes. In particular, it eliminates the possibility that, in equilibrium, there is too much employment. This last result is interesting in itself because it is commonly believed that normality of leisure necessarily implies that labour contracting models will generate employment levels that are too high from an efficiency perspective.

ABSTRACT Chapter 11 structures complex negotiations between creditors and debtors that are overseen by a bankruptcy court. We identify conditions where the court should sometimes err in determining which firms should be liquidated. Such errors affect actions by both good and bad entrepreneurs. We first characterize the optimal error rate without renegotiation. When creditors and debtors can renegotiate to circumvent an error‐riven court, for one class of actions a blind court that ignores all information is optimal. For another class, the court should place the burden of proof on the entrepreneur. The robust feature is that the court should sometimes err.