Geographical accessibility to bank branches and its relationship to new firm formation in Sweden via multiscale geographically weighted regression.

The dynamics of banking accessibility and its relationship with new firm formation can vary across space and time due to spatial contexts. With increasing bank branch closures in Sweden, it is even more critical to understand where interventions are needed and at which scale, in order to implement effective policy. Thus, spatial context is incorporated into the analysis with the use of the multiscale geographically weighted regression (MGWR) model. To overcome a potential endogeneity problem, an instrumental variable approach is applied. A two-stage least squares (2SLS) model is employed in which two instrumental variables are utilised to instrument the physical accessibility to the nearest bank branch in 2013. The MGWR results show that the geographical distance to the nearest bank branch has a negative association with the new firm formation in all Swedish municipalities, with little-to-no spatial heterogeneity. Further robustness tests show that the relationship between the geographical distance to the nearest bank branch and new firm formation are weaker in 2007, which shows temporal changes in the strength of the relationship over time. A policy implication from the analysis shows it is ideal to focus on the geographical accessibility to bank branches in all the Swedish municipalities with a single policy. QC 20211130