We estimate habit formation in voting--the effect of past on current turnout--by exploiting transitory voting cost shocks. Using county-level data on U.S. presidential elections from 1952-2012, we find that precipitation on current and past election days reduces voter turnout. Our estimates...

Exploiting regression discontinuity designs in Brazilian, Indian, and Canadian first-past-the-post elections, we document that second-place candidates are substantially more likely than close third-place candidates to run in, and win, subsequent elections. Since both candidates lost the election...