The use of fixed-term contracts has proliferated during the past decade in many European countries due to the relaxation of their regulation. Policymakers aimed to reduce labor-market rigidities by offering to firms these flexible contracts with little or no dismissal costs but with a finite contract length. The analysis of these contracts has thus far focused on their effect on the overall employment rate. This study high-lights that in the evaluation of fixed-term contracts as policy instruments it is also important to look at their effect on productivity as a function of tenure and on the tenure distribution of employed workers. These two effects jointly determine the policy's overall productivity effect. I show that the liberalization of fixed-term contracts can have a significant effect on the productivity of employment relationships when match-specific learning is important. Moreover, the effect is different depending on the assumption about the nature of the learning process. I distinguish between two kinds of match-specific learning - learning-by-doing and learning about match quality - and show that under learning-by-doing the overall productivity effect is necessarily negative, while under learning about match quality the effect could be either negative or positive depending on how much experimentation improves in the presence of fixed-term contracts. I calibrate the model based on earlier empirical work and find that indeed the productivity effect is positive as output per worker increases by 0.6%.