A Spatial Explanation for the Balassa–Samuelson Effect∗

We propose a model of urbanization and development to explain why the price level is higher in rich countries. There are two sectors: manufacturing, which is freely tradable, and non-tradable services, which have to locate near customers in big cities. As countries develop, total factor productivity increases simultaneously in both sec-tors. However, because services compete with the population for scarce land, labor productivity will grow slower in services than in manufacturing. Services become more expensive, and the aggregate price level becomes higher. The model hence provides a theoretical foundation for the Balassa–Samuelson assumption that productivity growth is slower in the non-tradable sector than in the tradable sector. The key implications of the model are borne out in the data: while the positive correlation between income and the price level is very strong among countries where land is scarce (densely populated, highly urban countries), there is virtually no such correlation among rural countries.