Using a formal principal-agent model, I investigate the relation between monetary gift-exchange and incentive pay, while allowing for worker heterogeneity. I assume that some agents care more for their principal when they are convinced that the principal cares for them. Principals can signal their altruism by offering a generous contract, consisting of a base salary and an output-contingent bonus. I find that principals signal their altruism by offering relatively weak incentives and a relatively high expected total compensation, but the latter does not necessarily hold. Furthermore, since some agents do not reciprocate the principal's altruism, the principal may find it optimal to write a contract that simultaneously signals his altruism and screens reciprocal worker types. I show that such a contract is characterised by excessively strong incentives and relatively high expected total compensation.