Document Type

Working Paper

Publication Date


Document Number

Working Paper Number 2018-05


Although the role of spatial dependence has been considered in studying the relationship between economic diversity and regional economic performance, the existing literature seldom mentions model uncertainty, which mainly arises from at least two sources. One source of model uncertainty is the choice of an appropriate spatial weight matrix that describes the spatial interactions between two regions, which can be specified in a variety of ways. The second source of model uncertainty is choosing a set of control variables to model the diversityperformance relationship. To overcome these limitations, a Bayesian Model Averaging (BMA) method is used to address model uncertainty when studying the effects of economic diversity on short-term employment growth and long-term economic stability among 359 Metropolitan Statistical Areas (MSA) in the contiguous U.S. The potential spatial spillovers are also considered through spatial regression models. This empirical analysis suggests that ignoring model uncertainty can impact the estimates and our understanding of economic diversity, and it also confirms that economic diversity of neighbors plays an important role in regional economic development.