Robert Summers

American Economics Association Distinguished Fellow Citation, 1998.

Robert Summers has been eclectic in his interest in economics. He always is wondering what can be inferred from available behavioral data, how economic analysis can help explain phenomena, or what incentives would induce desired behaviors. Some of these characteristics are reflected in his early publications on topics such as an investigation of the small sample properties of alternative estimators, or what can be learned about risk aversion from contingent remainder evaluation. Whether writing professional articles or discussing sports or the behaviors of colleagues, he is always thinking as an economist and turning alternative possibilities over in his mind.

Summers' path-breaking work for which he is known internationally, however, has been a collaborative effort with Alan Heston (also an AEA Distinguished Fellow in 1998) and, initially, with the late Irving B. Kravis (an AEA Distinguished Fellow in 1991). Their work produced a series of books and articles on the measurement of consistent economic aggregates and prices across nations and over time. The team's major contribution has been in developing real product comparisons and estimating purchasing-power parities for an expanding number of countries.

The work was initiated on Western Europe in the 1950s, then expanded to a larger scale in the United Nations International Comparison Project. By the early 1990s multilateral techniques were worked out, inevitably involving compromises between practical operational and theoretically preferred procedures, for over 80 benchmark countries based on 400-700 prices for specific items per country and 150 well-specified expenditure categories. These benchmark estimates then were extended to over 130 countries. These estimates have made it possible to compare real quantities and price parities among countries of various levels of aggregation. Such comparisons indicate that exchange-rate-based comparisons may be quite misleading. They tend to overstate differences across countries in total product per capita and they tend to misrepresent the composition of output, with over estimates of investment and producer goods and underestimates for non-tradeables for countries with lower per capita incomes relative to those with higher per capita incomes.

The project has represented a major step towards the development of a full global national accounts system that permits interspatial as well as intertemporal quantitiy comparisons. The techniques and data developed by this project are widely used today, and the procedures have been adapted by a number of international organizations. Summers and Heston's 1991 Quarterly Journal of Economics article on "The Penn World Table (Mark 5): An Expanded Set of International Comparisons" has been one of the most cited papers of the 1990s. This article describes the status of the project as of the early 1990s, shows the sensitivity of the estimates to a range of assumptions, and provides guidance for potential users of these estimates. The Penn World Tables have been a major resource, used by many, to analyze various aggregate empirical dimensions of "new economic growth" models and other aspects of aggregate economic experience and shape much of what we think we currently know about international aggregate economic comparisons and empirical aspects of aggregate economic growth.