• Selected
• Entire List

Abstract:
This paper analyzes the allocation of decision-making authority when the principal has reputation concerns. The principal can either keep the authority, or delegate the authority to the agent, who has better information. An outside evaluator cannot observe who makes the decision. Thus, the principal can share the blame with the agent when a decision is verified to be wrong. Delegation policy can then help to manipulate the principal's ex post reputation. We show that the principal keeps too much authority compared with the optimal level from the evaluator's point of view. A situation of passing the buck may also occur: when the evaluator believes that the agent is also responsible for the decision making, the principal has less incentive to make the right decision.

Delegation, Reputation Concerns.

Abstract:
This paper invesigates the optimal compensation scheme for workers in a team who value not only absolute but also relative incomes. A worker is said to be more ambitious if his utility places more weight on relative income. In this case the firm can exploit the worker's preference for relative comparison to design a compensation scheme that induces the same effort level with lower cost. Under the optimal compensation scheme, the workers' wages are shown to depend on relative performance, and exhibit wage compression. More importantly, even if the production technology calls for absolute performance evaluation in the traditional principal-agent framework, the optimal wage structure still relies on relative performance. Finally, in contrast to past literature, worker heterogeneity is shown to reduce the firm's profit.

relative performance, absolute performance, promotion tournament, wage contract, relative income comparison

Abstract:
This paper considers a two-stage sequential investment problem where the principal wishes to cancel the project if it fails in the first stage. Suppose that only the first-stage agent can observe the outcome in the first stage. There are two organizational forms to choose from: integration, where a single agent is in charge of investments in two stages; and separation, where two different agents are in charge of the two stages. Integration gives rise to a smaller wage cost of inducing high effort in both stages; however, in order to obtain the correct information to cancel the project, separation may have some advantage in terms of saving the information rent.

Optimal organization, sequential investment, cancellation option

Abstract:
This paper considers a two-stage sequential investment problem where the principal wishes to cancel the project if it fails in the first stage. Suppose that only the first-stage agent can observe the outcome in the first stage. There are two organizational forms to choose from: integration, where a single agent is in charge of investments in two stages; and separation, where two different agents are in charge of the two stages. Integration gives rise to a smaller wage cost of inducing high effort in both stages; however, in order to obtain the correct information to cancel the project, separation may have some advantage in terms of saving the information rent. We show that when the effort cost in the first stage is sufficiently small, the principal prefers separation because the first-stage agent has less incentive to lie about the outcome.

Optimal organization, sequential investment, cancellation option

Abstract:
This paper studies a finite horizon version of Baron and Ferejohn's (1989) majoritarian bargaining with incomplete information. Our devised model essentially blends Spence's signaling and the coalition formation of majoritarian bargaining. The main findings include: (i) oversized coalitions may arise in equilibrium and allowing for delay can be optimal for a proposer; (ii) whether being regarded as a high type makes a player better or worse off is not predetermined a priori but depends on two conflicting concerns: the "offer" concern à la Spence's signaling (the higher the type, the higher the offer) and the "coalition" concern in majoritarian bargaining (the higher the type, the lower the probability of being a member of the majoritarian coalition); (iii) players take action to distinguish themselves via delay; however, separating equilibria often fail to exist due to the incapability or unprofitability of exercising separation; (iv) separating equilibria, if they exist, can have a "no-envy" property, and pooling equilibria may coexist with the "least-cost" separating equilibrium even after refining the equilibrium set by the intuitive criterion; and (v) equilibria may be non-stationary and exhibit the novel feature that pooling proposals arise if a game lasts an even number of sessions, but not if there is an odd number of sessions.

Majoritarian bargaining, incomplete information, minimum winning coalition

Abstract:
Judicial torture to extract information or elicit confession was a common practice in pre-modern societies, both in the East and the West. Moreover, often it was applied not only on the suspects, but also on the witnesses and plaintiffs as well. This paper proposes a positive theory for judicial torture. It is shown that torture reflects the magistrate's attempt to balance type I and type II errors in decision-making, by forcing the guilty to confess with higher probability than the innocent, and thereby decreases type I error at the cost of type II error. In that case, torturing the witnesses or the plaintiff might also serve the same function, as it helps to screen the cases so that only those with greater merits enter the court. When the information revealed during investigation improved as a result of technological advance, a judicial system based on torture became inferior to one based on evidence. This result is then used to explain the historical development of the judicial system.

Torture, Type I and Type II errors, Evidence

Abstract:
By modeling judicial torture as a war of attrition, the paper derives the optimal strategies of the magistrate and the accused defendant as functions of their characteristics and the nature of uncertainty. Torture can occur as an equilibrium outcome in which both parties take costly actions to overcome informational barriers. Whether the magistrate will torture, and its result if he does, is shown to depend on how he evaluates the loss of type II error against the torturee's pain, his belief on how likely it is that the defendant is guilty, and the defendant's disutility of being tortured relative to the legal penalty of crime.

torture, type I and II errors, evidence, war of attrition, imperfect information