Nationalization of U.S. Local Elections

Scholars have documented an upward trajectory of the nationalization of the American electorate, but frequently overlook local elections. Existing tools for measuring electoral nationalization possess notable drawbacks and more importantly, cannot be appropriately applied to local elections. I develop a measure of nationalization based on a k-means clustering algorithm and show that nationalization trends differ between subnational and local elections: congressional and gubernatorial elections have markedly nationalized but local elections have not followed the same rising trend and exhibit considerably lower degrees of nationalization. These findings imply that conclusions made exclusively based on subnational elections, elections that are federal or elected at the state-level, are not complete and political behavior has not converged across local units to the extent previously postulated. Furthermore, I find evidence that suggests election timing plays a larger role than partisanship in the dampening of nationalization in local elections.