Districts, spillovers, and government overspending

The author considers the overspending bias in legislatures when the benefits of public policies are concentrated in particular districts but the costs of financing them are spread over the entire political jurisdiction. He formalizes this idea in a simple theoretic framework, in the context of externalities between districts. His main prediction is that greater districting leads to bigger government, but the effects are mitigated if there are positive spillovers of government spending between districts. Institutional forms of government that concentrate decision-making power can curtailthe overspending bias. He presents evidence on these predictions from a cross section of U.S. city governments. His main findings are that: 1) If we take an average city, keep its population and other characteristics constant, but divide it into a greater number of political districts, we get substantially greater government spending per capita. 2) Greater jurisdictional heterogeneity and income equality are associated with bigger government. 3) At-large electoral systems are not less sensitive to overspending than district electoral systems are. 4) Strong-mayor forms of city government, especially those in which mayors have veto power, can curtail the overspending bias. These findings are robust to controlling for socioeconomic characteristics of cities and to alternative measures of government size.