The International Comparison of Prices Program (ICP)

Origins of the United Nations International Comparison Programme

The national income accounting framework is the standard statistical device for describing countries' economic affairs. Entries in the usual System of National Accounts (SNA) are maintained by most members of the United Nations. A very influential figure in the original development of national accounts was Simon Kuznets, a Penn faculty member from 1936-54, who for this and other pioneering work won the third Nobel Prize awarded in Economics, in 1971. Richard Stone, who helped draft the initial SNA in 1952, and oversaw most of the 1968 SNA, was also awarded the Nobel Prize in 1984. 

The 1968 SNA is a very effective data system for describing the details of a country's economic condition at a point in time or over a period of time. Unfortunately, it did not permit effective comparisons between different countries. The intertemporal viability of country comparisons had no interspatial counterpart.  There had been individual efforts such as the work of Colin Clark in the Conditions of Economic Progress (1940) and of Milton Gilbert and Irving Kravis (1954), and associates at the Organization for European Economic Co-operation (OEEC) who provided a more systematic model of how interspatial comparisons might be done. There was also a major effort of COMECON countries who undertook bilateral comparisons between the Soviet Union and various countries of Eastern Europe beginning in the 1960s. And the UN Economic Commission for Latin America also undertook purchasing power comparisons in the 1960s. The first systematic multilateral set of purchasing power comparisons was that of the International Comparison Programme (ICP) of the United Nations that was begun in 1968 under the overall umbrella of the Statistical Office with Zoltan Kenessey in New York and and Irving Kravis at Penn as its first operational director. Irving Kravis, a student when Kuznets was at Penn, helped assemble an international adisory group and secured major funding from the Ford Foundation to initiate the work in 1968. Alan Heston joined the project in 1968 and Robert Summers joined in 1971. The initial work greatly benefited from the experience of economists and statisticians who were associated with the COMECON work, including Bodan Szulc, Giorgi Syallagi, Laslo Drechsler, E. Krzeczkowska, Margaret Mod, L. Zienkowski, and Zoltan Kenessey, who was recruited by the UN Statistical Office to coordinate with the unit at Penn. Early in the work the ICP received moral and material support from the World Bank and many individual countries. This early work involved many meetings and some of the participants are pictured at an early conference supported by the Rockefeller Villa Serbolini in Bellagio. Partly as a result of this initial work, the 1993 SNA in which Peter Hill was heavily involved, included international comparisons involving conversions country national accounts at PPPs.

Over the years a number of so-called benchmark studies have been conducted with the active co-operation and participation of the United Nations, and since 1980 Eurostat at the EU, the Organization for Economic Cooperation and Development (OECD), and the UN Regional Commissions have been most active in the work. Prices of different goods and services---standardized with respect to quality---were collected in many countries. Appropriately computed indexes based on these prices are applied to entries in the countries' SNAs to produce valid comparisons of the countries' relative standings, both with respect to the countries' national incomes and also the fine detail of the compositions of their incomes. Participating countries in the benchmark comparisons worked closely with the ICP groups in New York and Philadelphia. The group at Penn evolved a framework of multilateral comparisons building on the work of others and developing some new methods for organizing and aggregating the data discussed in Kravis, Kenessey, Heston and Summers (1975). The estimation of the benchmark studies for 1967, 1970 and 1975 were carried out at Penn, and the last of these studies is posted on this website and the World Bank. (see Publications and World Income and Product at the World Bank site). After the 1975 comparison the ICP was organized on a regional basis with the EU, OECD and the UN regional commissions coordinating individual countries. Alan Heston did the work on the 1980 benchmark study when he was at the United Nations Statistical Office on leave from Penn. Since the 1980 benchmark Penn participation in benchmark studies has taken a different form: regional organizations (UN regional commissions, often with the assistance of the World Bank, the European Union, OECD, etc.) made the relevant comparisons for the relative standings of their member countries. Our contribution has been to combine these results into world benchmark comparisons that became the basis for extending the work across countries and time in the form of Penn World Table (See About PWT) Because the benchmark comparisons are rich sources of data for economic analysis, providing much more detail than is available other sources for comparisons across countries, we have made them available at a world level in as comparable a format as we could manage. As described below, these may be downloaded by interested users. Also, if users of benchmark data are primarily concerned with the Eurostat and OECD countries, their benchmark comparisons are more detailed, and both organizations make available their data to researchers with a justification for its use. (See OECD Statistics, Eurostat).

Using the Benchmark Comparisons

The published record of benchmark comparisons includes regional and world comparisons. These are described through 1985 in publications of the World Bank, including the published comparisons of the Penn group. What is made available here is the benchmark world that have been put together for the development of the Penn World Table. These estimates will differ from published estimates in three basic respects. First, we have used the latest national accounts totals. Second we have tried to make the methodology comparable for the treatment of services in all comparisons through 1985. And third, after 1975, we have combined regional comparisons into world comparisons without fixity, the practice of maintaining the relationship between countries found in a region even when countries outside the region are introduced into the comparison. The files include 16 countries for 1970, 34 for 1975, 60 for 1980, 65 for 1985, and 115 for 1996. Between 1985 and 1996 there were a number of regional comparisons, including the beginning of annual estimates within Eurostat and 3 year benchmarks for the OECD beginning in 1990. Many countries in Africa, Asia, and the former Soviet Union also took part in 1993 benchmark comparisons. In an attempt to produce a world comparison, the World Bank then updated the 1993 benchmark ICP comparisons for 14 ESCAP countries, 22 African countries, 12 Caribbean countries and 8 ECWA countries to 1996 and combined these with the results for 9 South American countries for 1996, plus the 52 countries that the OECD covered, that included non-member countries within the former Soviet Union. (The total of 117 double counts Japan in the OECD and ESCAP and Egypt in Africa and ECWA). The linking of the various regions was done in different ways, usually with a link country like Japan for ESCAP, and the United States for Africa, and the Western Hemisphere. The linkage was carried out at the world level for only 32 headings and the estimates are subject to much more error than earlier benchmarks, and the World Bank has not thought them of sufficient accuracy to justify publication.

The interested user may download the basic input data for each of the 5 benchmarks. These include 10 tables for each benchmark, 5 for detailed headings, and 5 for aggregations, where for 1996 there are about as many aggregations in earlier benchmarks as there are detailed headings for 1996.  All expenditures are expressed on a per capita basis and a separate table provides the exchange rates, populations and supercountry weights used in each benchmark. Table 1 provides expenditures per capita in national currencies and Table 3 the price parities for heading and aggregations. Individual price comparisons underlie the heading parities but these are not in generally available. The basic output of the Geary aggregations is a set of expenditures in international dollars of each year. (See box at right.) These are given in the detailed and aggregate Table 5, and the entries are comparable across and down country columns. An $I has the purchasing power over all of GDP (but not the components) of a US$ in current prices of each benchmark year.  Detailed Table 3 has been expressed as the parity for each heading in national currencies per US $, where it is understood that the US =3D 1.0.  In order to provide more information, we have put what we term the international price in the US column.  In the aggregate Table 3, the same procedure has been followed.  (If a user wishes to express these parities relative to Earthea, the total of the countries in our comparisons, they can also be obtained by dividing each entry in Table 1 by the corresponding entry in Table 5 where the US will not be 1.0; the international price is simply the reciprocal of the purchasing power parity of the US.) 

Future Benchmark Comparisons

There have been several reviews of benchmark comparisons.   One is the Ian Castles report:  Review of the Eurostat OECD Programme, 1997. This is on the OECD website under Statistics Directorate/Purchasing Power Parities, as well as papers from the Joint World Bank-OECD Seminar on Purchasing Power Parities, Washington, 30 January-2 February, 2001.  This site also has links to the report of Jacob Ryten on the United Nations programme.  As a follow up to the 2001 meeting, the World Bank held a conference in 2002 that was to spell out how a coordinated world comparison in 2003 or 2004 might be organized and it also included the Ryten report. (See WorldBank: Data and Statistics / ICP). It is clear that the comparisons carried on by Eurostat and OECD are continuing and include many non-members. A major question remains of how to include other world areas. 


Gilbert, Milton and Irving B. Kravis (1954), An International Comparison of National Products and the Purchasing Power of Currencies: A Study of the United States, The United Kingdom, France, Germany and Italy, OEEC, Paris.

Kravis, I. B., Z. Kenessey, A. Heston and R. Summers (1975), A System of International Comparisons of Gross Product and Purchasing Power Baltimore:  Johns Hopkins University Press.

United Nations Economic Commission for Latin America (1963), Measurement of Price Levels and the Purchasing Power of Currencies in Latin America, 1960-1962, Mar del Plata Argentina, May. E/CN.12/653.