Political Parties and the Tax Level in the American States: A Regression Discontinuity Design

With a regression discontinuity design I show that the partisan identity of the majority in the state House of Representatives has no causal effect on the tax level. This result goes against recent findings in the political economy literature. In the state Senate I find a significant discontinuity in the tax level, but I also find a discontinuity in the density of the forcing variable - which implies that we can not interpret the discontinuity in the Senate as a causal relation. Another contribution of the paper is to investigate under which conditions slim majorities in the American states (as opposed to close election) are appropriate for a regression discontinuity design.