An Application of Spatial Poisson Models to Manufacturing Investment Location Analysis

The influence product markets, agglomeration, labor, infrastructure, and government fiscal attributes had on manufacturing investment flows in Indiana between 2000 and 2004 were estimated using Poisson regression, geographically weighted regression, and a spatial general linear model. Counties with access to urbanization economies, product markets, available labor, a high-quality workforce, and transport infrastructure were more likely to attract manufacturing investment. These effects were magnified to some extent when inter-county spatial effects were modeled. The distributional assumptions of the spatial models are different, but both methods are useful for understanding the spatial context of the factors influencing manufacturing investment flows