I consider a common agency model under adverse selection with a risk averse agent. Contracting takes place ex ante when all players have symmetric, although incomplete, information. The coordination problem between principals leads to more distortion in the optimal policy from the first best compared to the case of risk neutrality. In contrast with the risk neutral case the principals are unable to screen completely the agent's preferences if she/he is sufficiently risk averse. However, if the agent is almost risk neutral the output is separating, but the transfer schedules keep track of asymmetric contractual externality. When risk aversion goes to zero the transfers become truthful as in the complete information case.
