A Multivariate Spatial Analysis for Anticipating New Firm Counts

摘要: This paper analyzes county-level firm births across the United States using a spatial count model that permits dependence, cross-correlation among different industry types, and over-dispersion commonly found in empirical data. Results confirm presence of autocorrelation (which can arise from agglomeration effects missing variables), industry-specific over-dispersion, positive, significant cross-correlations. After controlling for existing-firm counts 2008 (as an exposure term), parameter estimates inference suggest younger work force and/or clientele quantified each county’s median-age values) is associated with more (in 2009). Higher population densities new basic-sector firms, while reducing retail-firm starts. The modeling framework demonstrated here be adopted variety settings, harnessing very local, detailed data to evaluate effectiveness investments policies, terms generating business establishments promoting economic gains.