Heterogeneous reaction versus interaction in spatial econometric regional growth and convergence models

In this chapter, we present the main econometric specifications capturing spatial heterogeneity, or models of absolute locations, and examine how these specifications can be extended to further allow for spatial autocorrelation models of heterogeneous reaction by emphasizing the complex links between spatial heterogeneity and spatial autocorrelation. We detail some of these issues in the growth and convergence context in this chapter by first presenting the specifications allowing for discrete heterogeneity, that is, when different parameters are estimated following spatial regimes including a focus on recent papers dealing with the endogenous detection of convergence in the presence of spatial autocorrelation. Then we present continuous spatial heterogeneity models: geographically weighted regressions and models allowing for both continuous spatial heterogeneity and spatial autocorrelation. We conclude with some research directions.