Wealth Effects in the Principal Agent Model

This paper addresses the question of how the principal's surplus and agency costs depend on the agent's wealth. Our main results are: If the agent has an additively separable utility function in income and effort and his degree of absolute prudence is smaller than three times the agent's degree of absolute risk aversion, then the principal's expected pay-off is smaller the richer the agent. For general utility functions, this result also holds if the first order approach is applicable and one further technical assumption is satisfied. If, however, the participation constraint does not bind, then richer agents are cheaper