Online Program

Interdisciplinary Research in the Social Sciences: Multilevel Modeling of Municipal Expenditure Data (302807)

Keywords: multilevel models, random coefficient regression, municipal expenditure data

The Social Science Statistics Center at the University of Missouri was created in 2003 to increase interdisciplinary statistics research in the social sciences. Here, we present a collaborative effort with the MU Truman School of Public Affairs. The study queries whether the growing number of ethnic minority mayors and minority-governed municipalities affect the nature, quantity, and quality of services and programs funded by various municipal governments. Multilevel modeling, specifically random coefficient regression models, is used to investigate whether cities with minority mayors exhibit differences in expenditure behavior. Another important issue is whether institutional factors, such as form of government, affect expenditure trends. Using Peterson’s typology for municipal spending, we assemble a database for 3,110 municipal elections covering the years 1967–2007 for all American municipalities with populations exceeding 100K. Our multilevel approach includes time as a level-one predictor, allowing for level-two effects such as race and form of government. Our results find significance of minority status and form of government for some types of expenditures.