Differential regional effects of monetary policy: A geographical SVAR approach

The study of possible asymmetric effects of monetary policy at a spatially disaggregated scale has received increasing attention in the literature. Different econometric approaches have been proposed to quantity the differences in monetary policy transmission, such as large-scale simultaneous equations models or structural vector autoregressive (SVAR) models. This article builds on the SVAR approach and extends it by incorporating geographical information, using spatial econometric techniques. The author employs information on spatial proximity to derive parameter constraints, enabling joint estimation for medium and large-sized panels. Moreover the use of spatial a priori information makes it possible to identify and estimate contemporaneous spatial spillover effects. Specific attention is paid to parameter identification when introducing the model. Subsequently, the author discusses parameter estimation, which is complicated by, the simultaneous spatial dependence structure and the need to impose complex parameter constraints. An empirical application is provided with respect to U.S. states.