History of the ICP
Statisticians have long recognized that using market exchange rates to compare economies’ levels of economic activity can lead to misleading results. In particular, the differences between the size of high-income economies with high price levels and low-income economies with low price levels will appear larger than they actually are. This distortion can be avoided by using purchasing power parities (PPPs) instead of market exchange rates to undertake such comparisons.
In the 1500s scholars at the University of Salamanca, in Spain, first explored the idea of purchasing power parities and later in 1918 the Swedish economist Gustav Cassel developed a modern definition of PPPs. It was only in the 1940s that the British economist Colin Clark use PPPs to first estimate levels of real income in his study The Conditions of Economic Progress. The first edition of his study was published in 1940, followed by second and third editions in 1951 and 1957 (all published by MacMillan, London). The first edition covered the United States and 52 other economies. Other economies were linked through a series of bilateral comparisons with the United States. The results were then used to quantify the inter-economy spread in real income per capita and to provide an estimate of world income. Income was defined as consumer expenditure and did not include government expenditure or capital expenditure. For income per capita, total persons employed rather than total population was the denominator. The PPPs were calculated using Fisher’s ideal volume index formula, that is, the geometric mean of the Laspeyres and Paasche volume indexes Referred to as international units, they measured the purchasing power of national currencies over the period 1925–34 based on average prices for the period. In the second and third editions of his study, Clark increased the number of economies covered and refined the methodology applied.
Clark’s pioneering studies stimulated further research. In the 1950s, the Organisation for European Economic Cooperation (now the Organisation for Economic Co-operation and Development (OECD)) used purchasing power equivalents to compare the national products of France, Germany, Italy, the United Kingdom, and the United States (Gilbert and Kravis, 1954). The comparison was subsequently enlarged to include Belgium, Denmark, the Netherlands, and Norway (Gilbert and Associates, 1958). All final expenditures, including government and capital expenditures, were covered in the comparison. In the 1960s, the Economic Commission for Latin America carried out PPP-based comparisons of real product in 19 Latin American countries; the Council for Mutual Economic Assistance (COMECON) conducted PPP-based comparisons of national income between several central and eastern European centrally planned economies; and the Conference of European Statisticians approved a project to undertake PPP-based comparisons of consumption levels among a small group of market economies and centrally planned economies.
In 1965 the United Nations Statistical Commission (UNSC) discussed the problems inherent in market exchange rate–based comparisons and agreed that the United Nations Statistical Office (UNSO) should develop a more suitable methodology for making international comparisons of GDP. In 1968 the UNSC considered a report that outlined a research project to be run from 1968 to 1971 aimed at developing PPP-based comparisons. The report proposed using a small group of economies representative of different income levels, social systems, and geographical areas to test and assess methodology. The UNSC agreed that the project should proceed, and, because the UNSO had only limited resources, asked other international organizations and UN member economies to assist in the project. At this stage, the research endorsed by the UNSC was to cover gross domestic products (GDP) measured from both the expenditure and production sides of the national accounts. Even so, it was understood that the initial efforts would concentrate on the expenditure side because it was less difficult to implement in practice since a single set of expenditures was involved rather than both outputs and inputs, which gave rise to the added complexity of double deflation.
The “International Comparison Project” was launched in 1968 as a joint undertaking between the UNSO and the University of Pennsylvania, which established a special unit funded by a grant from the Ford Foundation. The World Bank became involved, providing financial assistance both directly and through a grant from the Scandinavian economies that was channeled through the World Bank. The U.S. Agency for International Development and the U.S. Social Science Research Council assisted with monetary contributions. The United Kingdom offered in-kind statistical support for the participating economies. The director of UNSO was responsible for supervising the project. The advisory board was set up to provide technical advice and considered detailed proposals for the project at a meeting held in October 1969.
One of the proposals discussed by the advisory board resulted in the ICP adopting a concept of consumption that summed the individual consumption expenditures of households and government to obtain an aggregate of total individual consumption called the consumption expenditure of the population (CEP). The objective in measuring the CEP was to minimize the effect on the volume comparisons of differences in institutional arrangements, particularly regarding the extent to which the government and private sectors provided health and education services in different economies. In this respect, the ICP was more than two decades ahead of the System of National Accounts 1993 (SNA93), which set out the concept of actual individual consumption (defined almost identically to the CEP) as an official national accounts measure.
Until 1993, the ICP was conducted in phases; after 1993 it was organized by rounds. Phase I had two stages. The first stage was a pilot study based on data collected for 1967 for six economies (Hungary, India, Japan, Kenya, the United Kingdom, and the United States). The second stage was run for 1970 and included four additional economies (Colombia, France, Germany, and Italy) that had not been able to report the necessary data for 1967. The outcome consisted of different sets of estimates, including multilateral comparisons between all 10 economies for GDP and a range of expenditure components for 1970. The results of Phase I were published in 1975 in A System of International Comparisons of Gross Product and Purchasing Power (Kravis et al. 1975). The details presented in this publication include the overall results of the multilateral comparison for 1970, a variety of bilateral comparisons for both 1967 and 1970, and the outcomes of various experiments on important issues such as rents, motor vehicle prices, and the consistency of different quantity comparisons.
Phase II included six more economies (Belgium, the Islamic Republic of Iran, the Republic of Korea, Malaysia, the Netherlands, and the Philippines), initially to enable a broader comparison for 1970, but mainly to update the PPPs and associated price and volume measures to 1973. Results for the 16 economies were published in 1978 in International Comparisons of Real Product (Kravis, Heston, and Summers 1978).
Thirty-four economies participated in Phase III for reference year 1975. In the earlier phases, the detailed characteristics of items in the U.S. consumer price index were used as the starting point for developing the ICP item lists. Later, they were modified in consultation with some of the participating economies, including India and the COMECON group, to make the ICP item specifications more generally applicable for example, by removing characteristics such as brand name that were specific to the United States. The greater diversity of economies in Phase III meant that the range of items to be priced had to be further expanded so that all participating economies could price a sufficient number of items representative of their expenditures. At this point, the ICP considered the pros and cons of continuing with a single global comparison or moving to regional comparisons that would be linked to produce worldwide results. The trade-off involved in regionalizing the project was improved comparisons between economies within a region but at the expense of the comparisons between economies in different regions because of the difficulties inherent in linking results between regions. In the end, however, Phase III went ahead as a single global comparison, although some regional results were presented as having been calculated for the relevant economies from the globally based results. The results of this phase were published in 1982 in World Product and Income: International Comparisons of Real Gross Product and Purchasing Power (Kravis, Heston, and Summers 1982).
Phase IV saw some major developments in the program. The first was that the number of participating economies almost doubled, from 34 to 60. The second was that the ICP shifted from a research project to a regular operational part of the UNSO work program. With this development, the University of Pennsylvania’s participation in the day-to-day running of the project ended, although it continued to advise on methodological issues. The third significant change was the regionalization of the ICP. The principal reason for regionalization was the large number of economies now involved worldwide, making it no longer feasible to organize comparisons centrally. Another factor was the decision by the Organisation for Economic Co-operation and Development (OECD) to set up a PPP program for its member economies in conjunction with the PPP program being run by Eurostat (the statistical office of the European Union) for economies in what is now the European Union. In addition to the Eurostat-OECD comparison, Africa, Asia, and Latin America participated in Phase IV as regions. The regions were linked using the bridge economy approach in which selected economies priced a range of item specifications from another region to provide a bridge or link between their region and the other region. The reference year for Phase IV was 1980.
The reference year for Phase V was 1985. It saw only a small increase in the number of participating economies, from 60 to 64, with some new economies replacing some that had been in Phase IV but had then dropped out of Phase V. Once again, a regional approach was adopted. The regions were Africa, Asia, the Caribbean, alongside the Eurostat-OECD comparison. In addition, three central and eastern European economies were added to the Eurostat-OECD comparison using Austria as a bridge. The bridge economy approach was again used to link the regions, but some of the links were problematic because of the difficulties several bridge economies encountered in collecting prices for a sufficiently broad range of items from the other region.
Phase VI, conducted with 1993 as the reference year, was the most ambitious phase to date, seeking to produce PPP-based comparisons for 118 economies. In the end, however, only 83 were covered. From the outset, this phase was beset by difficulties. Lack of funding was the major problem, although the lack of overall coordination also led to some major deficiencies in the final outcome. Regional comparisons were undertaken for Africa, Asia, Eurostat-OECD, and Western Asia, but not for Latin America. Moreover, there was no global comparison because it proved virtually impossible to link the regions. In response to these problems, in 1997 the UNSC commissioned a major review of the ICP before further phases were attempted.
The report on the review was presented to the UNSC in 1999. It concluded that PPPs and PPP-related statistics were needed, but that the ICP was not producing these data on a timely and regular basis for a sufficient number of economies as required by potential users. Poor management and insufficient resources at all levels - central, regional, and national - were identified as the principal reasons for the difficulties. Other important contributory factors were inadequate documentation, heavy data requirements that did not take into account the circumstances of individual economies, lack of uniformity in the execution of activities across regions, lack of confidence among economies that others were following guidelines and standards consistently, and failure to involve economies in the editing and calculation stages of the exercise. The report recommended that the UNSC not sanction a new round until at least the management and resource issues had been resolved.
The UNSC’s response was to ask the World Bank to consult with other interested parties and propose a strategy to address the deficiencies identified by the review and to draw up an implementation plan for a new round of the ICP. The plan involved mobilizing funds from a variety of sources and establishing a governance infrastructure to provide effective management and coordination between the global center and the regions and between the regions and the participating economies. It also involved providing complete and clearly written documentation on the ICP’s technical and procedural guidelines and standards. Such guidelines would allow economies to participate in a full comparison covering GDP or in a partial comparison covering actual final consumption, using, as far as possible, regular national statistical programs to obtain price and national accounts data for the ICP and linking participation in the ICP to national statistical capacity building.
The UNSC considered the implementation plan in 2000 and again in 2001. It was reluctant to start another round of the ICP before adequate funding had been secured. However, after the World Bank embarked on a successful major fund-raising exercise, the UNSC agreed, at its 33rd session, to a new round in 2002. At the same time the UNSC reviewed and endorsed a new strategic framework for the ICP, including an international governance arrangement and a broad implementation plan.
The new round was launched in 2003 and ended in 2008. The reference year was 2005. Regional comparisons were organized by the ICP regional coordinating agencies - the African Development Bank (AfDB), the Asian Development Bank (ADB), assisted by the Australian Bureau of Statistics, the Statistical Office of the Commonwealth of Independent States (CIS-STAT) with the State Statistical Service of the Russian Federation, the United Nations Economic Commission for Latin America and the Caribbean with Statistics Canada (UN-ECLAC), and the United Nations Economic and Social Commission for Western Asia (UN-ESCWA), and by Eurostat and the OECD. The ICP Global Office was established at the World Bank to provide overall coordination and to ensure technical and procedural uniformity across the regions. The Global Office was also responsible for organizing the Ring comparison that, by comparing a small number of economies from each region across regions, provided the means to link the regional comparisons in one global or worldwide comparison. The results of the regional and global comparisons were published at the end of 2007 and the beginning of 2008.
ICP 2005 was generally considered to be a success. It produced estimates of the relative price levels of GDP and its principal aggregates for 146 economies, including the major emerging ones such as Brazil, China, India, Indonesia, the Russian Federation, and South Africa, and its results were published on a timely basis in 2008 in Global Purchasing Power Parities and Real Expenditures: 2005 International Comparison Program (World Bank 2008). An important contributory factor was the governance structure that the World Bank had put in place prior to the start of the exercise to ensure that the ICP regional coordinating agencies would deliver, within a common time frame, regional results that would be consistent across regions and that could be combined in a global comparison. The governance structure was retained after ICP 2005 to commence preparations for the next round of the ICP proposed for 2011. The proposal was approved by the UNSC at its 39th session in 2009, and the UNSC requested the World Bank host the Global Office and take on the global program coordination of the 2011 round.
The ICP 2011, with its considerably expanded coverage of 199 countries, brought a broader acceptance compared to earlier exercises, and the wide availability of reliable PPPs referenced to 2011, and published in 2014, increased their use both topically and across the globe. Notably the international poverty line was updated in 2015 to $1.90 a day reflecting PPPs for 2011. The major improvements in the program were documented by the Friends of the Chair group, who were asked by the UNSC at its 45th Session, in March 2014, to evaluate the 2011 cycle. The group observed that the 2011 round had put the program on a firm methodological basis by introducing approaches such as the global core list and applying major technical innovations. Specifically, the provision of technical assistance to countries, the broad documentation of metadata, and the further development of ICP operational guides and handbooks, contributed significantly to the knowledge of staff conducting the work around the world.
The 47th Session of the UNSC, held in March 2016, discussed the future of the ICP, in light of the recommendations of the Friends of the Chair group in its evaluation of the ICP 2011. As a result, the UNSC instituted the ICP as a permanent element of the global statistical work program to be conducted at more frequent intervals from the 2017 cycle onwards.
At the same time the UNSC also endorsed the strengthening of the governance structure consisting of the ICP Governing Board, the Inter-Agency Coordination Group (IACG) and the Technical Advisory Group (TAG), with its intermittent task forces, and the implementing agencies at the national, regional and global level. The structure ensures efficient functioning and balanced representation of countries and coordinating agencies in the governing bodies. The Global Office was established as a permanent unit at the World Bank, responsible for the global coordination, data validation and calculation of global results, and related day-to-day organizational activities. With regards to the methodology of the ICP, the UNSC agreed that no major changes should be introduced and that a research agenda, to be developed by the Technical Advisory Group, should focus on methodological improvements to be considered for future comparison cycles.
With the launch of the United Nations’ 2030 Agenda for Sustainable Development, the UNSC also emphasized the need to link the capacity-building activities of the ICP with efforts to enhance the statistical capacity of countries for monitoring progress towards the Sustainable Development Goals. The Commission also suggested exploring the closer alignment of ICP price surveys with consumer price index compilation by countries.
Data collection for the ICP’s 2017 cycle began in 2016 and continued through to the end of 2018. The number of participating countries decreased slightly to 176, with Fiji the only representative from the Pacific Islands. Argentina and Guyana joined the Latin America and the Caribbean comparison while Costa Rica and Colombia moved to the OECD exercise. Some other countries that participated in the 2011 cycle were affected by conflict or natural disasters and were not included in 2011. These included Guatemala and Venezuela, RB in Latin America and the Caribbean and Yemen in Western Asia. The AfDB oversaw the work of the 50 economies in that region, with Economic and Statistical Observatory for Sub-Saharan Africa (AFRISTAT) coordinating 30 of those, and the Common Market for Eastern and Southern Africa (COMESA) the remaining 20. Three economies within this region also participated in the Western Asia exercise (Egypt, Arab Republic; Morocco; and Sudan). The ADB coordinated the work of its 22 participating economies, while CIS-STAT oversaw its eight economies. UN-ECLAC coordinated the work of its 36 economies, with support from CARICOM for the 23 Caribbean islands. UN-ESCWA coordinated the 12 economies in that region. Eurostat and OECD provided results for their joint total of 49 economies and assisted with the special participation of Georgia and Ukraine. Iran participated through a special exercise linked with Western Asia. To encourage and prepare for future participation, several countries also benefitted through ICP capacity building efforts. These included Eritrea, Libya, Somalia, South Sudan, Turkmenistan, Uzbekistan and the Republic of Yemen.
Preparations for the planned ICP 2020 cycle began in 2019. However, the onset of the COVID-19 pandemic in early 2020 caused disruptions to statistical exercises across the world. National implementing agencies experienced lockdown and mobility-related challenges in conducting price surveys and compiling national accounts. The ICP Technical Advisory Group considered these challenges as possibly impacting the quality of results and advised the ICP Governing Board to move the benchmark year to 2021, which was subsequently enacted in April 2020. The cycle was renamed to the ICP 2021 cycle. The regional implementing agencies for the 2017 cycle continued in those roles for the 2021 cycle. There were 176 participating economies. Fifty-two economies participated in the Africa comparison with the addition of Somalia and South Sudan for 2021. Twenty-one economies participated in the Asia and Pacific comparison, with Myanmar dropping out for the 2021 cycle. Nine economies participated in the Commonwealth of Independent States with the addition of Uzbekistan. Thirty-two economies participated in the Latin America and the Caribbean region, with Guatemala joining the cycle, while Haiti, The Bahamas, Barbados, Sint Maarten, and Turks and Caicos did not participate. In Western Asia, sixteen economies participated, with Lebanon, Mauritania, the Syrian Arab Republic, and Tunisia joining the cycle. Five economies participated in both the Africa and Western Asia exercises (Egypt, Arab Republic; Mauritania, Morocco; Sudan; and Tunisia). A number of nonparticipating economies benefitted from ICP 2021 capacity-building activities, and some of these economies are aiming to subsequently participate in future ICP cycles.
References:
Clark, Colin. 1940; 1951, 1957. “The Conditions of Economic Progress”. MacMillan and Co Limited. ASIN B007T0H3J0. London.
Gilbert, M. and Kravis, I., OEEC. 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”. Paris.
Gilbert, M. and Associates, OEEC. 1958. “Comparative National Products and Price Levels: A Study of Western Europe and the United States”. Paris.
Commission of the European Communities et al. 1993. “System of National Accounts 1993”. Brussels/Luxembourg, New York, Paris, Washington, D.C. https://unstats.un.org/unsd/nationalaccount/docs/1993sna.pdf
Kravis, Irving B.; Kenessey, Zoltan; Heston, Alan Wiley; SUMMERS, ROBERT. 1975. “A system of international comparisons of gross product and purchasing power” (English). Washington DC ; World Bank. http://documents.worldbank.org/curated/en/199981467988893189/A-system-of-international-comparisons-of-gross-product-and-purchasing-power
World Bank. 2008. Global purchasing power parities and real expenditures: 2005 international comparison program (English). Washington, DC; World Bank Group. http://documents.worldbank.org/curated/en/842221468166145357/Global-purchasing-power-parities-and-real-expenditures-2005-international-comparison-program