We empirically analyse regional inflation using data from Japan where there is no regulation to impede the free movement of labour, capital, goods and services across regions. In particular, our analysis will focus on the geographical location of regions and the productivity (known as the Balassa-Samuelson (BS)) effect as explanations for the dynamics of regional inflation. Based on a spatial model which is consistent with the theoretical specification of the BS, we have confirmed that, while it is a relatively small country in terms of land area, both spatial location and productivity are important determinants of regional inflation.