Lecture on the first SFB/TR 15 meeting, Gummersbach, July, 18 - 20, 2004: The existing literature on the comparison of tournaments and piece rates as alternative incentive schemes has focused on the case of unlimited liability. However, in practice real workers' wealth is typically restricted. Therefore, this paper compares both schemes under the assumption of limited liability. The results show that if the cost function is sufficiently convex, first-best effort will be more likely implemented under piece rates than under tournaments. Moreover, if first-best implementation is not achieved and workers earn positive rents, efforts and profits will be larger for piece rates than for tournaments given sufficiently convex costs. While tournaments offer a partial insurance due to their fixed prizes, piece rates may not work any longer if potential losses become prohibitively high. Finally, if risk is sufficiently high, piece rates will dominate tournaments despite the partial insurance effect of tournament compensation. Since effort costs and risk may depend on an individual worker's characteristics, on the characteristics of his job and on his hierarchical position, these findings have important implications for the choice of incentive schemes and the allocation of workers in firms.