Information and Incentives in Organizations

This thesis asks about the value of information for providing incentives in principal agent models with hidden action and limited liability. The classical literature deals with information in the form of signals that the principal (and the agent) receive after the agent's effort choice. It shows that such signals are beneficial if and only if they are informative about the agent's effort. In contrast to these papers, the first three chapters deal with situations where the principal and the agent observe the signal realization before the agent's effort choice. The fourth chapter considers ex post information. Chapter 1 endogenizes the timing of the signal and asks whether the principal prefers the agent to receive an additional signal before or after the agent chooses his effort. For decision problems it is well known that ex ante information is (weakly) beneficial. I show that there is no difference between incentive and decision problems if the signal is uninformative about the agent's effort. In contrast, if the signal is informative ex ante information does strictly worse. Chapter 2 assumes that the signal the agent observes is the output of a colleague. It identifies a positive effect that acts on the incentives of this colleague and interacts with the effects from Chapter 1. This can make ex ante information optimal even when the colleague's output is informative. I relate these findings to the organizational structure of a firm and its internal transparency. Chapter 3 keeps the timing fixed and asks whether the principal benefits from an additional signal that the agent observes before his effort choice. My finding that such additional information is not always beneficial contrasts with the classical result that more (ex post) information is (weakly) better. Chapter 4 considers ex post information and asks how the principal's desire to receive more information about the agent's effort leads to inefficient job assignments. A simple trade-off between incentive provision and job assignments explains the Peter Principle and delivers predictions consistent with empirical evidence from personnel economics