High-stakes electoral conflicts induce politicians to put politics before policy. One widely proposed solution is to improve policy-making processes: specifically, to implement rules that accommodate diverse interests and values by fostering impartiality, generality, transparency, deliberation, and consensus. The higher order problem, however, is that incentives to make better constitutional choices are often absent. This paper explores the potential of a “turn-taking institution” as a way to improve constitutional incentives. Turn-taking institutions lower the stakes of electoral conflict and foster the formation, enforcement and adaptation of broadly beneficial constitutional rules. We use a game-theoretic model to assess the robustness of these results when participants are impatient and have limited commitment horizons, when incumbents have the advantage, and when electoral accountability is biased against good policy.