To control for product quality and eliminate the exchange rate volatility effect, we use the Japanese regional data to study the Penn effect - the positive relationship between price and income levels. Similarly to what is widely documented with international data, the price and income levels exhibit significant positive association across the Japanese prefectures. Furthermore, the intra-Japan Penn effect is driven essentially by prices of nontradables. The effect is also found stronger among rich prefectures than poor ones, as is the case with the international data. In explaining the Penn effect within Japan, we find that the measures of sectoral productivity do not behave in the way suggested by the Balassa-Samuelson hypothesis. On the other hand, the population density variable that captures the agglomeration effect offers a good explanatory power.
agglomeration effect Penn effect sectoral productivity differential tradables and non-tradables population density