"Most contracts, whether between voters and politicians or between house owners and contractors, are incomplete. ""More law,"" it typically is assumed, increases the likelihood of contract performance by increasing the probability of enforcement and/or the cost of breach. This paper studies a contractual relationship where the first mover has to decide whether she wants to enter a contract without knowing whether the second mover will perform. We analyze how contract enforceability affects individual performance for exogenous preferences. Then we apply a dynamic model of preference adaptation and find that economic incentives have a non-monotonic impact on behavior. Individuals perform a contract when enforcement is strong or weak but not with medium enforcement probabilities: Trustworthiness is ""crowded in"" with weak and ""crowded out"" with medium enforcement. In a laboratory experiment we test our model's implications and find support for the crowding prediction."