Inter-country comparison of ‘real’ (PPP) incomes: Revised estimates and unresolved questions

Adjusting for purchasing power parities of national currencies makes a significant difference to inter-country income comparisons. However, PPP estimates require detailed price data which are not readily available for most countries. Kravis, Heston and Summers have used regression analysis to extend their PPP estimates from the 16 countries for which they have detailed prices to any country for which data are available on the share of trade in GDP and on inflation. The first part of this paper argues that a variant of this ‘short-cut’ approach, one which relies on measures of human-resource development rather than on trade and inflation, is preferable on theoretical and statistical grounds. The second part of the paper raises a more fundamental issue. It argues that the treatment of services in their detailed analysis, and hence in any short-cut approach derived from it, is biased. This bias tends to exaggerate the PPP incomes of countries with relatively good human-resource development, and so produces misleading estimates of relative incomes among poor countries and of the gap between poor and rich countries.