This paper is concerned with the in the sense used in a recent paper by Kravis and Lipsey entitled Toward an Explanation of National Price Levels [15]. The price level of a particular country relative to the base country is defined as the ratio of the purchasingpower parity (PPP) to the exchange rate. It has become generally recognized' that the PPP need not equal the exchange rate even in long-run free trade equilibrium and the explanation of the discrepancy is a natural subject of scientific interest. There is a voluminous literature on temporary deviations of PPP from the exchange rate that are explained in terms of monetary disturbances and random shocks. There is a much smaller literature on long-run or structural factors that explain why PPP does not coincide with the exchange rate. The present paper falls in the second category; we shall assume the economy is in monetary and balance-of-payments equilibrium. The term as used in Kravis and Lipsey [15] may be confusing because that term usually refers to the price of goods in terms of money. The term price seems more appropriate, in that the concept refers to a ratio of two prices. The price is what the tourist has in mind when he goes to a poor country and remarks, My, but prices are low here. A major feature of the existing literature on long-run or structural explanations of the real price level is the emphasis on the distinction between tradables and nontradables; it is typically assumed that the law of one price holds for tradables but not for nontradables, which are dominated by services. The productivity differential model, as stated by Balassa [2] for example, assumes that international differences in labor productivity are greater in commodities (tradables) than in services. It follows that services will be relatively cheap in poor (low-productivity) countries and that, consequently, the real price level (relative to a base country) will rise with per capita income. Recently, Bhagwati [5] has shown that the low relative price of services in poor countries can be explained by a capital-labor model without recourse to the assumption that production functions differ internationally. In Bhagwati's model, rich countries export capital-intensive commodities and poor countries export labor-intensive commodities, but even though commodity prices are equalized through trade, complete specialization prevents factor prices from being equalized; since services are assumed to be labor-intensive relative to any commodity, they are cheaper in poor countries than in rich countries.