New firm formation: Dynamics and determinants

Empirical studies of determinants of new firm formation to date have tended to yield diverse and even contradictory results. Three primary reasons for this have been advanced: (1) the paucity of suitable micro-level data (2) the failure to control for amorphous time- and place-specific influences that defy specification, and (3) the use of estimation techniques that do not handle adequately the effects of heteroscedasticity. This paper addresses all of these shortcomings by employing a unique data set composed of annual data on localized firm entry, exit and a variety of predictor variables that has been analyzed to yield heteroscedasticity-corrected estimates while controlling for unspecified place- and period-specific influences. We test a variety of models seeking to explain patterns of new firm formation in terms of macroeconomic, demographic, and labor market processes, patterns of industrial restructuring, availability of local financial capital, and local public sector spending. Our results suggest that regional patterns of new firm formation can be explained by variation in unemployment change rates, mean establishment size, prior firm entry and exit dynamics, and the availability of local financial capital. We find no evidence of influence attributable to population or income dynamics, unemployment level, or local government spending.