Inefficiencies in Local Infrastructure: Evidence from Drinking Water in California
Paper joint with Giulia Brancaccio and Karam Kang.
Abstract: We develop an empirical framework to quantify inefficiencies that arise when infrastructure investments are chosen by locally elected officials, and apply it to California's drinking-water systems. Using new data on system performance and investment, we show that weaker electoral competition increases water pollution and dampens the effectiveness of federal investment aid. We then build a dynamic model in which residents differentially value water quality and vote accordingly, while office-seeking politicians weigh their own private preferences against improved re-election prospects. To separately identify residents' preferences, we exploit housing transactions at water-system boundaries; to distinguish policymakers' preferences from electoral incentives, we leverage within-jurisdiction changes in electoral rules and electorate composition. We estimate a substantial willingness to pay for investment, with unmet needs concentrated where electoral competition is weakest. Using the estimated model, we simulate the counterfactual effectiveness of federal aid under imperfect representation of constituents.