UC Berkeley

This dissertation considers how income inequality affects attitudes towards tax policies and income redistribution. In the first paper, I adopt traditional models of the relationship between income inequality and policy preferences to incorporate income volatility. I find that greater income volatility can lead voters to prefer less progressive taxes depending on the distribution of the income shocks. Because income volatility and inequality are positively correlated, with each potentially having an opposite effect on public opinion, my model predicts that support for increasing taxes on upper incomes may not rise in the face of growing income inequality. In the second paper, I estimate the casual effect of short term, temporary changes in household income on individuals' tax progressivity preferences through a series survey experiments. I find evidence that volatility does in fact diminish preferences for tax progressivity. Respondents preferred significantly less progressive taxes when described households that had more volatile incomes regardless of the pattern of volatility over time. For the third and final paper, I develop a new, more powerful permutation procedure for analyzing two-way factorial experiments. The method is a non-parametric alternative to traditional analysis of variance (ANOVA). In Monte Carlo simulations, the procedure was better able to disentangle main factor and interaction effects than regression-based ANOVA tests, particularly when the design was imbalanced. I developed the procedure for use in the analysis of the income volatility experiments.