| The first chapter analyzes, in the context of one-sided delegated bargaining, how a principal (a seller) should design the delegation contract in order to provide proper incentives for her delegate (an intermediary) AND gain strategic advantage against a third party (a buyer). In a context with risk-neutral players, moral hazard and adverse selection problems in the delegation relationship, and absence of commitment effect, a linear contract is optimal. When delegation contracts have commitment value, the seller can gain substantially by imposing a minimum price, above which she pays the delegate a commission. The interaction between commitment (through minimum price) and incentives depends on the nature of the agency problem. Incentives and commitment are substitute when the delegate's unobservable effort improves his bargaining position, but are neither substitute nor complement when his effort increases chances of finding a buyer. In most cases, the seller's strategic manipulation of the delegation contract may cause bargaining failures between the delegate and the buyer.; The second and third chapters examine, in a principal-monitor-agent framework, the use of the timing to monitor the agent as a contracting instrument. First, we show that, under some conditions, the optimal contract with an ex post monitor depends on obtaining a signal after the agent exerted effort and not on what is being monitored (agent's effort or productivity). This result extends to a self-interested monitor for any possible level of collusion between the agent and the monitor (i.e., information concealment or manipulation). In the third chapter we analyze the optimal monitoring timing in different collusion environments. Ex ante monitoring (supervising) is optimal for weak punishment and an informative signal, and ex post monitoring (auditing) is optimal otherwise. This result extends to collusion with hard and non-forgeable information, but may be reverted when information is hard and forgeable or soft. In these cases, collusion imposes costs to both auditing and supervising and, under some circumstances, supervising is more likely to be optimal when the signal is noisy even for unbounded punishments, or when punishment schemes are weak (independently of the signal's noise). |
