Spatial patterns in the location decisions of US credit unions

Building on central place theory, we offer a series of simple location models for US credit unions. Credit unions are historically viewed as a community-based strategy designed to fill voids in spatial financial markets. Unlike banks and more traditional retail financial institutions, our results show that credit unions operate in areas with a low concentration of retail banks. This suggests that credit unions do not follow the "herd" mentality that dominates traditional retail banking institutions. Instead they cluster around common bonds of association. That is credit unions are fundamentally capturing a different market than traditional banks and do not behave in a pro-competitive manner with banks. In the spirit of Bitter et al. (J Geogr Syst 9:7-27, 2007), we compare the overall performance of geographically weighted regression with a set of both spatial and non-spatial estimators. We find that the spatial estimators generally performed better than the aspatial estimators, and there is significant spatial heterogeneity in some of the determinants of credit union concentrations. Results on individual variables of interest tend to be stable across the different estimators lending confidence to our hypothesis that credit unions serve niche markets.