The Partisan Divergence in U.S. Local Policy, 1972-2017

Abstract

This paper examines the determinants and trends of fiscal policy differentiation across U.S. local governments over the period 1972 through 2017. We provide a machine-learning-based measure of policy differentiation based on the predictive power of high-dimensional public budgets for socioeconomic and political characteristics. We find increasing polarization along political lines, in the sense that fiscal policies have become more predictive of political preferences. In contrast, the relationship between fiscal policies and other socio-demographic attributes, such as rural-urban, has remained relatively stable. Finally, we measure fiscal political polarization by state and find that it is strongly correlated with spatial political sorting.

Keywords: Fiscal Policy, Policy Polarization, Machine Learning