The study of possible asymmetric effects of monetary policy at a spatially disaggregated scale has recently received renewed attention in the literature due to the introduction of EMU. To quantify the differences in monetary policy transmission different econometric approaches have been proposed. At the macro level both structural simultaneous equations models and structural vector autoregressions (SVAR) to address the issue. The current paper mainly builds on the SVAR approach and extends it by incorporating geographical information in model's specification, making use of the techniques commonly employed in spatial econometrics. While, to capture spatial interactions it would be necessary to adopt a VAR specification modelling jointly the given set of regions, this is generally not feasible using standard VAR models due to the shortage of degrees of freedom. In the proposed specification, information on spatial proximity is used to derive parameter constraints that make the joint estimation feasible for panels of moderate or large dimension, requiring time series of length comparable to that necessary for standard VAR estimatio Having introduced the model's specification, with specific reference to the issue of parameter identification, the paper deals with parameter estimation, that, in this, case is complicated by the complex simultaneos dependence structure. Finally, to test the model's empirical performance, the paper presents an application to the analysis of the differential monetary policy effects on the US states. Based on the estimation results, geographical heterogeneity in the impulse response function found out in previous studies appears to be confirmed.
展开▼
