A smooth coefficient quantile regression approach to the social capital–economic growth nexus

Abstract

This analysis assesses the role of social capital in generating heterogeneity in growth processes across U.S. counties by estimating growth regressions, using the novel semiparametric smooth coefficient quantile regression method in which parameters are unspecified functions of a measure of social capital. The results indicate substantial differences across the quantiles of economic growth in the profile shapes of the coefficient estimates over the level of social capital. Moreover, the coefficient function estimates are highly nonlinear over the level of social capital, providing evidence that the growth process that links initial income, education attainment, ethnic diversity, inequality, population density, and government activity to growth varies with social capital in a nonlinear way.