We develop a Principal-Agent model to analyze the optimal composition of the compensation policy with both monetary and nonmonetary incentives. We characterize nonmonetary benefits as symbols to capture a large set of non-wage compensations such as fringe benefits, status, identity (or...