Public Opinion and Policy Responsiveness in Small Electorates: Institutions and Spending in American School Districts

Assessing policy responsiveness, the degree to which public policies correspond to citizen preferences, is a central challenge to political scientists. However, one aspect of policy responsiveness -- whether policy responsiveness is enhanced or diminished by various institutional arrangements -- remains largely unexplored. The researchers address this by focusing on spending outcomes in America's school districts. More specifically, they examine how much actual spending levels correspond to local preferences (after controlling for economic resources). Because school districts display enormous variation in several key institutions and their governance arrangements were specifically intended to enhance or diminish policy responsiveness, school districts represent an ideal unit of analysis to pursue these theoretical questions.

The researchers examine how the correspondence between local educational spending levels and citizen preferences varies by three aspects of institutional design: fiscal independence, referendum requirements, and rules of electoral representation. Using data on over 9000 US school districts they show that reforms resulting from Civil Rights challenges to electoral systems were successful in improving the policy responsiveness of districts with even small minority populations. In addition, they show that one innovation championed by the Progressives -- fiscal independence-- lessens policy responsiveness while another such reform -- the referendum -- strengthens the opinion-policy linkage.

To explore these substantive themes, they use Bayesian hierarchical models to estimate public opinion for various voter types within each state. These estimates are combined using demographically based post-stratification weights to estimate public opinion in each school district. The research support is for (1) refining and validating the estimation method, (2) applying the new measure to empirical questions on policy responsiveness, (3) adding a dynamic aspect to the analysis, and (4) generalizing the method to partisanship, ideology and other policies.

In this stage, the broader value of the research is its contribution to several literatures by addressing important and unresolved questions. The analysis of fiscal independence contributes to a better understanding of American political development and the intention of Progressives to insulate educational policy making from the public. The analysis of the referendum will contribute to a growing empirical literature on direct democracy. In particular, it is possible to test certain aspects of the median voter model used in these studies because the investigators have actual estimates of the median voter's preferences. The analysis of electoral systems will add substantially to literature on racial politics and vote dilution. Prior studies show that the shift from at-large to ward-based electoral systems increased the descriptive representation of African Americans (and Hispanics, in Texas). The investigators go further by seeing whether these gains also led to increases substantive representation and n policy responsiveness.