The timing of elections is flexible in many countries. We study this optimization, by first creating a Bayesian learning model of a mean-reverting political support process. We then explore optimal electoral timing, modeling it as a renewable American option with interacting waiting and stopping...