One of the most important parameters of a labor contract is its length, yet contract length has received relatively little attention by economists. The lack of attention to contract length might be justified if contract lengths were stable over time, but union labor contracts have fluctuated in length from one to five years. In this paper, we study the optimal length of labor contracts and determine what factors alter their lengths. In addition, some dynamic techniques are presented which may assist in the analysis of contract lengths in other contexts. Since virtually all multiperiod economic models make some implicit assumptions about contract length when the lengths of contracts are actually endogenous, this research may have applications in a wide variety of settings.