Exploring the meso-determinants of apartment prices in Polish counties using spatial autoregressive multiscale geographically weighted regression

The aim of this study is to identify the meso-determinants of apartment prices in Polish counties using recently developed spatial autoregressive multiscale geographically weighted regression (MGWR-SAR), which can simultaneously deal with spatial dependence, spatial heterogeneity, as well as different spatial scales of modelled processes. Estimation results indicate that MGWR-SAR outperforms other studied models, especially in terms of goodness-of-fit indicators. Moreover, estimates of the above model reveal that relationships between the dependent variable and covariates operate at different spatial scales. Furthermore, among the studied determinants only two, describing unemployment rate and standard of apartments, are significant in all counties.