Freitas-Groff's dissertation studies whether policies update as voters change their views, or whether policies can stay stuck for long periods of time. To measure this, he assembles a novel dataset containing the statutory history of close U.S. state-level referendums since 1900. In a regression discontinuity design, he estimates that narrowly- passed policies are 40 percentage points more likely to be operative 20, 40, or even 100 years later than those that narrowly fail across a wide range of topics and political contexts. Using additional novel data on international referendums and congressional legislation, he replicates this result more generally. In a game-theoretic model, he proposes a novel explanation for policy persistence based on time-varying political interest. The finding of strong persistence implies many policies are misaligned with social welfare and provides (previously absent) support for event-study designs built around long-term policy changes.
