Space, Time, and Local Employment Growth: An Application of Spatial Regression Analysis

Abstract

Local and regional employment growth is generally studied either by searching for local qualitative explanatory factors such as governance, synergy between firms, and milieu effects, or by searching for general growth factors using statistical techniques. The body of work that relies on this approach has tended, in keeping with economics' nomothetic tradition, to assume that local and regional growth factors are constant over space. The focus of this paper is on exploring the spatial stationarity of employment growth factors in Canada, but it also seeks to clarify some of the broad principles behind spatial regression techniques in order to provide a point of entry and a conceptual framework for empirical researchers. To do so, we apply a recently developed technique, Geographically Weighted Regression (GWR), and we explore the method's advantages and limits for answering our research question. We find evidence that growth factors differ across Canada, but we also conclude that the GWR technique, given the number and shape of regions available for our analysis and given certain limitations that are currently inherent to the method, can only provide tentative and exploratory results.