World Comparison Real Gross Domestic Product and Purchasing Power

III. REGIONAL RESULTS

There is an abundance of information contained in the 1985 regional comparison results. These comparisons produced a wealth of statistics revealing price and expenditure conditions in a total of 64 countries, that is, almost one third of the countries of the globe. Since the inception of the ICP the number of participating countries has increased significantly (see table E), though the coverage is still far from being universal.

The reasons why certain countries did not join the programme, or others that had participated earlier dropped out, are manifold. For a large number of countries participation is simply not possible without outside support, mainly technical assistance that requires appropriate funding. Such funding has been available for certain groups of countries, while it could not be secured for others in phase V. While insufficient financial support might adversely influence the decision of some countries regarding participation in the ICP, this explanation may not apply in the case of all nations outside the programme. Some countries may have different priorities in their national statistical services that, from their point of view, would outweigh the importance of obtaining internationally comparable national accounts and price data, or the benefits from participation. While reasons can be very complex, the fact is that in phase V no countries from Latin America or Western Asia were covered, in spite of the intentions and efforts of the coordinating organizations.

In the present report only some features of the regional comparisons with reference year 1985 are highlighted. For more detailed results and information on how the regional comparisons were conducted, the reports pertinent to each comparison should be consulted. The relevant publications are listed in the References.

In addition to the published data, more detailed results may be available, upon request, from the organizations coordinating each regional comparison. Regions, however, may have different policies regarding whether basic national data on prices and expenditures can or cannot be released for research purposes, depending upon the approval or restrictions imposed by the participating countries concerned. Inquiries about the availability of such detailed information could be directed to the respective agencies acting as regional centres.

1. AFRICAN COMPARISON

At the outset of the ICP work in phase 1, more than two decades earlier, one African country-Kenya-was included in the International Comparison Programme. For 1985, the regional comparison involved 22 African countries. Participating countries in the 1985 African comparison were Benin, Botswana, Cameroon, Congo, Cote d'lvoire, Egypt, Ethiopia, Kenya, Madagascar, Malawi, Mali, Mauritius, Morocco, Nigeria, Rwanda, Senegal, Sierra Leone, Swaziland, United Republic of Tanzania, Tunisia, Zambia and Zimbabwe. Seven African countries, namely, Benin, Congo, Egypt, Mauritius, Rwanda, Sierra Leone and Swaziland, were newcomers in the programme for 1985.

The key role in the work was undertaken by EURO- which coordinated and managed the African regional comparison, including data collection and processing work. EUROSTAT also provided substantial funding for the African exercise. Conducting the African comparison took four years of work and involved efforts concentrated in an extensive regional programme, headed by EUROSTAT working in close cooperation with the national statistical offices of the participating countries.

The classification scheme applied in the African comparison for 1985 comprised 170 basic headings where consideration was given to the minimum requirement of data availability at the detailed level for the world comparison. 'The number of basic headings for Private Consumption (including government expenditure on health and education) was 126; for Public Consumption 6; for Gross Fixed Capital Formation 36; and for Changes in Stocks and Balance of Exports and Imports I each. The number of products selected for pricing in each basic heading varied considerably. The total number of representative items exceeded 1,000 goods and services; for Private Consumption alone 973 items were selected, 174 specified items were used to price Equipment Goods (Transport, Machinery and Other Equipment), and for the comparison of construction prices 10 structures were chosen. Price surveys were conducted between December 1985 and May 1987; in most countries they were completed in the first half of 1986. In order to estimate prices in 1985, temporal adjustments were made to some of the prices reported. While the actual conduct of the surveys was the responsibility of the national statistical offices, survey documents and technical advice were provided by EUROSTAT.

The computations made by EUROSTAT included appropriate edit checks and additional computing for the countries which did not provide national average prices, but only submitted price observations. Transitive basic parities were derived by using the Elteto-Koves-Szulc (EKS) method; for the aggregation of basic PPPs and calculating real values for aggregated categories of the GDP, the Geary-Khamis (G-K) method was applied. In the presentation of results of the comparison for the 22 African countries in 1985, the real values for each country are expressed in Purchasing Power Standard (PPS). Instead of giving preference to the currency of a single country, a numeraire was chosen which combined the currencies of all countries in the region.

The results show a nearly 16-fold difference between the first- and the twenty-second-ranked countries (Mauritius and Ethiopia, respectively). Per capita real GDP in Mauritius is over three times the African average, while Ethiopia accounts for one fifth of the African average level. Countries with per capita GDP around the African average were Zimbabwe, Nigeria and Senegal in 1985.

The report on the African ICP results for 1985 emphasized four major potential applications of the information developed in the comparisons.

First, potential application of ICP results is considered in costing investment projects. As is stated in the report, specific surveys undertaken within the ICP on prices, quantities, quality, the cost of installation etc. should make it much easier for the authorities to assess the investment costs of projects they are planning.

Second, the use of the information is placed in the context of competitiveness; detailed price data obtained from the comparison can be used to compare international export prices with local prices for specific capital goods, so that their competitiveness can be assessed.

Third, national authorities are expected to analyse ICP data to promote national price and income policies. The report on the African comparison states that it is obviously important for such policy measures that national price structures and differences in price levels between different regions should be known, even if only approximately, and improved knowledge of regional price variations obtainable from the ICP could help the authorities to improve the supply situation (by improving transport, storage facilities, packaging and marketing etc).

Fourth, it is thought that ICP information could help in understanding the structural imbalances of a country by considering the corresponding structures in another country. Such a comparison may be particularly rewarding when it concentrates on countries at different stages of economic development.

In addition to the specific advantages of the ICP information for analytical and research purposes in Africa, the organizational and methodological improvements engendered in the national statistical services of the participating countries were among the significant gains of the work. In a number of African countries, before they got involved in the regional comparison, price surveys had been irregularly taken and were restricted to large towns (or to the capital city only), and the sample of household consumer products was often limited and not strictly defined. Similar improvement in reporting national accounts data was experienced. For example, it was found that nearly all the 15 countries that participated in both the 1980 and 1985 phases revised their national accounts estimates.

The results of the African regional comparison are fully incorporated in the 1985 global comparison.

TABLE E.COUNTRY PARTICIPATION IN THE PHASES OF THE ICP

Country/Area		Phases
		I	II	III	IV	V
		Benchmark Year
		1970 a	1973 b	1975	1980	1985
Africa
	Benin					x
	Botswana				x	x
	Cameroon				x	x
	Congo					x
	Cote d'lvoire				x	x
	Egypt					x
	Ethiopia				x	x
	Kenya	x	x	x	x	x
	Madagascar				x	x
	Malawi			x	x	x
	Mali				x	x
	Mauritius					x
	Morocco				x	x
	Nigeria				x	x
	Rwanda					x
	Senegal				x	x
	Sierra Leone					x
	Swaziland					x
	Tunisia				x	x
	United Republic of Tanzania				x	x
	Zambia			x	x	x
	Zimbabwe				x	x
Asia and Oceania c
	Bangladesh					x
	Hong Kong				x	x
	India	x	x	x	x	x
	Indonesia				x
	Iran, Islamic Republic of		x	x		x
	Israel				x
	Japan	x	x	x	x	x
	Malaysia		x	x
	Nepal					x
	Pakistan			x	x	x
	Philippines		x	x	x	x
	Republic of Korea		x	x	x	x
	Sri Lanka			x	x	x
	Syrian Arab Republic			x
	Thailand			x		x
	Australia					x
	New Zealand					x
North America c
	Canada				x	x
	United States of America	x	x	x	x	x
	Mexico			x
Central and South America c
	Argentina				x
	Bahamas					x
	Barbados					x
	Bolivia				x
	Brazil.			x	x
	Chile				x
	Colombia	x	x	x	x
	Costa Rica				x
	Dominican Republic				x
	Ecuador				x
	El Salvador				x
	Grenada					x
	Guatemala Honduras				x
	Jamaica			x		x
	Panama				x
	Paraguay				x
	Peru				x
	Saint Lucia					x
	Suriname					x
	Trinidad and Tobago					x
	Uruguay			x	x
	Venezuela				x
Europe c
	Austria			x	x	x
	Belgium		x	x	x	x
	Denmark			x	x	x
	Finland				x	x
	France	x	x	x	x	x
	Germany, Federal Republic of	x	x	x	x	x
	Greece				x	x
	Hungary	x	x	x	x	x
	Ireland			x	x	x
	Italy	x	x	x	x	x
	Luxembourg			x	x	x
	Netherlands		x	x	x	x
	Norway				x	x
	Poland			x	x	x
	Portugal				x	x
	Romania			x
	Spain			x	x	x
	Sweden					x
	Turkey					x
	United Kingdom of Great Britain and Northern Ireland	x	x	x	x	x
	Yugoslavia			x	x	x

^a In phase 1, there were 6 countries which also provided benchmark data for 1967.
^b In phase 11, 6 additional countries provided benchmark data for 1970, which were combined with revised data for the original 10 benchmark countries; in addition, all estimates were also made for 1973 based on some original and some updated price data.
^c Countries in this table are listed by geographical areas, since regionali- of the ICP became effective in phase IV and onward. However, one should bear in mind that the composition of ICP regions does not necessarily correspond to the geographically contiguous areas as shown above.

2. CARIBBEAN COMPARISON

The Caribbean comparison for 1985 covered the following seven countries: Bahamas, Barbados, Grenada, Jamaica, Saint Lucia, Suriname, and Trinidad and Tobago. Except for Jamaica, this was the very first time that these Caribbean countries ever participated in the ICP.

The Caribbean work was coordinated by the Statistical Office of the European Communities (EUROSTAT), assisted by its consulting agency, the European Center for World-wide Cost of Living Comparisons (EuroCost), in data collection and preparation of the study on the regional results. The two organizations worked in close cooperation with the national statistical authorities of the countries involved, in consultation with the Caribbean Community secretariat. Substantial financial support for this regional programme was provided by EUROSTAT.

Work on the Caribbean comparison was started later than phase V activities in other regions. The overall methodology, data needs and organization of the project were reviewed at a workshop held at Kingston, Jamaica, in 1987. In the Caribbean comparison with reference year 1985, the overall classification scheme for which expenditure data were required comprised 170 basic headings. The number of basic headings under each of the major GDP aggregates corresponded to those in the classification scheme of the 1985 African comparison. The list of products for household consumption and machinery and equipment used in the Caribbean price comparison was based on the one drawn up for the phase V comparison in Africa. After some modifications aiming at reflecting items typical in the Caribbean and omitting African commodities uncharacteristic in the Caribbean countries, but with most items originating from the African list, a total of 876 products for consumption and 170 for equipment goods were priced. Basic data were collected by the national statistical offices assisted by EUROSTAT and EuroCost experts. The price surveys were carried out between May and October 1989 in the countries concerned, typically during a one-month period. In order to estimate the price levels for 1985, the ICP reference year of phase V, national price index information was used to take price changes between 1985 and 1989 into account. The study on the regional results, "Comparison of price levels and economic aggregates me results of seven Caribbean countries" (EUROSTAT, Luxembourg, 1991), contains a complete listing of the deflators used for adjusting surveyed prices.

Data processing was undertaken by EUROSTAT. The calculation of basic parities was based on the EKS method; for the aggregation the G-K method was chosen. To present comparison results, a numeraire was chosen that combined the currencies of all seven Caribbean countries compared. The GDP of each country in absolute and per capita terms is expressed in national currency, in United States dollars (converted by official exchange rates) and in Purchasing Power Standard (PPS).

The regional study produced by EuroCost officials contains abundant information on the methods applied and presents the Caribbean comparison results at the level of all 170 basic headings as well as the summary aggregates up to GDP. The results, however, have yet to be reviewed and approved by the participating countries, at a regional meeting scheduled for early 1994. Therefore, under these circumstances and for reasons explained in section IVB, this region could'not be integrated into the 1985 global comparison.

3. ESCAP COMPARISON

In the ESCAP regional comparison with reference year 1985, the following countries participated at full scale: Bangladesh, Hong Kong, India, Iran (Islamic Republic of), Japan, Pakistan, the Philippines, the Republic of Korea, Sri Lanka and Thailand. The above 10 countries took part in the entire programme aiming at estimating real GDP levels and corresponding PPPs. In addition, the participation of Nepal was limited to the comparison of consumption only. Japan joined the ESCAP comparison simultaneously with its participation in the 1985 OECD comparison; that made possible a convenient linking of the results obtained in the two studies. Both Japan and India have been long-time participants of the International Comparison Programme since its phase I onwards; the other ESCAP countries, except Bangladesh and Nepal, also had previous experience with ICP work.

Organizational and data processing activities of the comparison were shared by the Statistical Division of the United Nations Secretariat (UNSTAT) and the United Nations Economic and Social Commission for Asia and the Pacific (ESCAP). Technical assistance was provided also by experts of the World Bank, of the Statistical Office of the European Communities (EUROSTAT) and of the Statistics Bureau of Japan. Regional ICP work in Asia and the Pacific for 1985 was sponsored by the Government of Japan, the Asian Development Bank, the United Nations Development Programme and the World Bank.

Data inputs for the comparison were the result of the combined efforts of the participating countries; in their efforts countries received technical assistance from numerous sources, as mentioned above. Actual comparison work started with a regional seminar held in Bangkok in September 1986; another three regional workshops and meetings held between 1984 and 1990 contributed to the clarification of tasks and to the checking and evaluation of the results. Most of the data surveys took place during the first half of 1987. Data processing was undertaken by the Statistical Office of the United Nations Secretariat (UNSTAT).

In the ESCAP comparison 163 basic headings were distinguished. In all, for household consumption, 1,533 specifications were priced. It should be noted that GDP data submitted for the comparison, as in the case of many countries in other regions, are not exactly the same as the data reported for the same year in the United Nations National Accounts Statistics: Main Aggregates and Detailed Tables, 1985. The discrepancies are due to the countries' efforts to follow ICP principles to enhance comparability; however, the coordinating agencies are not aware of all of such corrections made.

On the basis of the prices submitted for the selected specifications, the PPPs were determined by using the CPD method. In order to give emphasis to the requirement of characteristicity, countries were asked to indicate the degree of characteristicity of each item priced so that items termed uncharacteristic would not be taken into account in the final calculation of PPPs. In this manner, care was taken to collect prices of items which were both representative and comparable; however, the requirement could not be equally satisfied in respect of each pair of countries. As in the practice of other regions, some countries usually tend to price items included in their national consumer price indices, while others favour conducting special surveys for the sake of comparability that, in some cases, may result in pricing higher-quality, more expensive commodities. Experience available from the 1985 ESCAP comparison indicates that, for example, Bangladesh tilted mainly to the former solution while India often followed the latter, and this might have resulted in their unusual relative volume indices.

The aggregation formula applied in the ESCAP comparison was the G-K. In this regional comparison the original national currency data converted by PPPs to values at international average prices of the countries involved are expressed in "Asian dollars", a nominal currency which plays the role of the numeraire only. Its purchasing power for the gross domestic product total (taking all participating countries together) equals the purchasing power of the United States dollar in 198 thus, the total gross domestic product of the whole regio whether expressed in Asian dollars or United States dollars, is the same. However, in each sub-aggregate group of the GDP the purchasing power of the Asian dollar is determined on the basis of ESCAP average prices. In the ESCAP study, the results of the quantity comparisons are expressed relative to the average ESCAP level.

The ranking of the countries according to the level of real GDP per capita indicates that in 1985 its level in Japan and Hong Kong was 4-5 times as high as the Asian average, while Bangladesh and India reached about one third of the average of me 10 ESCAP countries analysed. The selected data below present the relative position of Nepal within ESCAP, since its results are not included in the tables with global results.

4. EEC COMPARISON

A comparison including the 12 member countries of the European Economic Community (EEC) was conducted by EUROSTAT with reference year 1985. The EEC comparison included Belgium, Denmark, France, Germany (Federal Republic of), Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain and the United Kingdom, all of which participated in previous rounds of the ICP as well.

TABLE F. 1985 PER CAPITA GDP COMPARISON IN ASIA
(Average of the 10 countries = 100)

GDP rank	Country name	GDP per capita	Household consumption	Government consumption	Capital formation a
1	Japan	525.3	486.2	218.4	698.9
2	Hong Kong	454.2	483.1	202.1	451.8
3	Iran (Islamic Rep. of)	204.1	213.4	212.9	179.5
4	Korea, Rep. of	176.8	164.3	153.2	212.7
5	Thailand	116.2	118.8	272.9	68.7
6	Sri Lanka	81.4	74.8	158.4	76.8
7	Philippines	78.5	93.8	139.4	25.8
8	Pakistan	58.8	63.4	185.4	14.4
9	Bangladesh	36.5	43.7	82.4	7.3
10	India	33.1	36.6	45.7	21.6
	Nepal b		33.1	55.9

^a Including net exports.
^b Nepal provided data only for the comparison of consumption.

In the EEC comparison a classification system consisting of 258 basic headings was used. The ample specification list, as well as other technical documents elaborated by EUROSTAT, served as an example for developing similar documents for the purposes of other regional comparisons, including the 1985 OECD comparison. The total number of products included in the specification lists for various aggregates was close to 3,000 items. For price collection, the surveys were carried out between autumn 1984 and spring 1986. For the first time in the ICP, instead of concentrating on the reference year, price surveys were split over a longer period to distribute the workload more equally. This experiment of EUROSTAT was further explored later on and became known as the "rolling benchmark year approach".

Price collection in the EEC countries was the responsibility of the national statistical institutions. Data collection in the EEC countries was largely financed by EUROSTAT. In the calculation of parities at the basic heading level the EKS method was used, which enabled EUROSTAT to give more importance in the computations to items that were more representative in a given country. For multilateral aggregation in 1985, EUROSTAT used the G-K method. Results are expressed at average prices of the 12 countries using the Purchasing Power Standard (PPS) as the reference monetary unit. In order to avoid taking the currency of one country as reference, in the Community framework the ECU was chosen. Thus, the GDP of the Community expressed in PPS as the numeraire is equal to the total GDP of the Community in ECU. As mentioned elsewhere in this report, the choice of the numeraire or of the reference unit is a matter of convention and does not affect the results of the comparison.

The comparison results for the EEC regarding the price level index show that in 1985 three groups of countries could be distinguished in relation to the Community average. Denmark, Germany and France had a price level well above the Community average; in five countries, the Netherlands, Belgium, Luxembourg, the United Kingdom and Ireland, it was around the average; and in Italy, Greece, Spain and Portugal it remained well below it.

The results illustrate the divergence of exchange rates and PPPs and its consequences on the ratios in real terms among countries. The report by EUROSTAT mentions two extreme cases as examples of this phenomenon. Denmark's share of total Community GDP equals 2.31 per cent if converted by exchange rates of ECU but falls to 1.85 per cent if calculated in real terms. On the other hand, Portugal's share amounts to 0.82 per cent on the basis of values in ECU but increases to 1.59 per cent expressed in real terms based on PPP conversion. In 1985 the per capita volume index for real GDP was highest in Luxembourg (127.8 per cent) compared to the EEC average, two and a half times as high as in Greece and Portugal at the other end of the ranking.

Unlike other regions in the ICP framework, the EEC comparison results computed by EUROSTAT are used in the Community not only in economic analysis and fore- but also as one of the criteria for the distribution of Community revenue and funds.

The results of the EEC comparison were incorporated into the OECD comparison, the European Comparison Programme and the global ICP for 1985, whereas the requirements of the "fixity principle" in respect of the 12 Community countries have been taken into consideration to preserve the uniformity of the EEC data.

5. OECD COMPARISON

Twenty-two OECD countries were covered in the OECD 1985 comparison. In addition to the 12 EEC mem- the countries covered were Austria, Australia, Canada, Finland, Japan, New Zealand, Norway, Sweden, Turkey and the United States of America; only two OECD member countries, Iceland and Switzerland, were not included. All countries had previous experience with the ICP except for Australia, New Zealand, Sweden and Turkey, which participated in the programme for the first time in 1985.

As before, the PPPs and associated estimates of real expenditure on GDP based on price and expenditure data for 1985 were calculated in collaboration with EUROSTAT. In respect to the comparison of the 12 Community countries and Austria, the coordination responsibilities rested with EUROSTAT; the OECD organized the comparison for the remaining 9 countries. The actual calculations for the entire OECD group were carried out by EUROSTAT. The results for the 12 Community countries, calculated at average EEC prices, remained unchanged in the OECD framework; only real values for Be EEC as a whole and the results for the other 10 OECD countries were shown at average OECD prices-that is, the average prices of the 22 countries. As a consequence of the "fixity principle", the linked data are not additively consistent, so linked components do not sum to linked totals.

The classification scheme of the 1985 OECD comparison consisted of 202 basic headings. The difference between the EEC and OECD classifications was confined to the number of expenditure categories for Private Final Consumption, consisting of 160 basic headings in the OECD comparison to which the 216 EEC headings aggregated exactly. In the OECD comparison the product specification lists developed by EUROSTAT were used. These lists were updated after consultations with the countries concerned and enlarged by the OECD to include commodities representative of the 10 additional OECD countries. The specification lists for 1985 covered some 2,500 consumer goods and services, 30 occupations in govern education and health services, 250 types of machinery and equipment, and 18 construction projects. As in the procedure followed in the EEC comparison, items were priced in five surveys conducted between autumn 1984 and spring 1986. Rents, insurance and medical services not covered in these surveys were subjects of special inquiries in mid- 1986. Prices provided for machinery and equipment, as well as construction, were mid-year prices for 1985, and in many countries price collection was done by private consultants. Price data collection for consumption was the responsibility of national statistical services. In some cases the price surveys were carried out in the capital cities and, where needed, the individual price quotations adjusted by EUROSTAT to national averages and for 1985.

Expenditure data were collected according to the detailed breakdown for 1985 or the latest year available, which might be 1984 or even 1983. This expenditure breakdown was then applied to national data on GDP and its main components available for 1985.

In the calculations for 1985, the EKS method was used to obtain the basic heading parities and the G-K method was used to aggregate the PPPs and real values. In this procedure Austria was treated as the EEC countries and therefore is PPPs and real values were initially calculated at average EEC prices. However, in the final tables Aus- data are shown at average OECD prices. The OECD comparison used the United States dollar as the common unit of currency.

In the report on the OECD comparison, the presentation of the detailed results used two expenditure classifications, those of the ICP and the SNA, having 56 and 57 expenditure categories, respectively. Data according to the ICP classification offer a better comparative measure of the actual volume of goods and services consumed by households of different countries. On the other hand, re suits following the SNA classification provide better comparative estimates of the total volume of services financed by government. The OECD report also contains a brief explanation of the methods applied.

Regarding the results of the 22 OECD countries in 1985, the indices of real value of the GDP per capita, with the United States level as 100, show that the two countries closest to the United States level were Canada (92.5) and Norway (84.4), while Luxembourg, the country with the highest comparable figure in the EEC comparison, ranked fourth among the OECD member States. Turkey, the twenty-second in the OECD ranking, reached about one fifth of the United States level.

Comparison results of the 22 OECD countries are fully covered in the globally processed data for 1985.

6. EUROPEAN COMPARISON PROGRAMME

The European Comparison Programme (ECP) for 1985, which covered 20 European countries, was implemented by means of cooperation among three interna- organizations, the European Communities, the OECD and the Economic Commission for Europe (ECE) of the United Nations, as well as the Austrian Central Statistical Office and the national statistical offices of the participating countries. The major novelty of this comparison was that it extended the comparison results available from the common EEC-OECD study to three Eastern European countries, Hungary, Poland and Yugoslavia, which, together with Austria, formed the so-called Group 11 region. The ECP was built from the following three principal components:

1. The EEC comparison, covering its 12 member countries, namely: Belgium, Denmark, France, Germany (Federal Republic of), Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain and the United Kingdom
2. A portion of the OECD comparison, covering the 12 members of the European Community, plus 5 European OECD members-Austria, as well as Finland, Norway, Sweden and Turkey;
3. The so-called Group II comparison, in which Hungary, Poland and Yugoslavia carried out a set of binary comparisons with Austria in an effort coordinated by the Austrian Central Statistical Office and the United Nations Economic Commission for Europe (ECE).

Initially, the three components were conducted separately, but the methodology has been coordinated to facilitate the linking of the comparisons in the second stage of the work. Since producing results for the EEC and the European OECD countries has been the responsibility of the relevant agencies, the organizers of the ECP focused mainly on the Group 11 comparison and on the linking of the above three components that was achieved by the dual participation of Austria in two comparisons simultaneously.

The three Eastern European countries had accumulated considerable experience from their participation in previous phases of the ICP. In addition, the direct bilateral comparison and the frequent consultations with Austria offered an excellent possibility for checking and improving the comparability of data that positively contributed to the high quality of this regional comparison. Final processing of the ECP was conducted by EUROSTAT, and the report of the European comparison was produced at the ECE of the United Nations. The regional report covers results for the 20 European countries at the level of 53 aggregated categories.

The detailed classification used in the Group 11 comparison included a breakdown of the GDP with 288 basic headings. This classification had to be harmonized with the ones used in the EEC and OECD comparisons to permit linking. For each binary comparison between Austria and the Eastern European countries, more than 1,000 representative items were selected and priced. A peculiarity of the Group II comparison is that it uses quality adjust- to the prices collected to express quality differ- in the commodities, enhancing both the representativeness and comparability of the specifications used. Since the Group II comparison was carried out on a bilateral basis, the PPPs for Hungary, Poland and Yugoslavia at the basic heading level were obtained from the direct bilateral comparisons with Austria as the 'base country, and in the aggregation those PPPs were weighted by the expenditures of the two respective countries (Austria and each of its three partners) by using the Fisher index formula. In contrast, in the EEC and OECD exercises a multilateral approach was applied using the G-K method for aggregation. In the linking process, the fixity arrangement was required for the 12 Community countries and Austria to keep their results unchanged in the broader ECP context, and the method of generating overall ECP results was chain-linking carried out at each level of the aggregation. As a consequence, the ECP results have limited additivity; that means the relative shares, of real values cannot be compared throughout all 20 countries. As a matter of fact, regarding the actual numerical results, the lack of additivity had only a minor effect on the 1985 ECP data.

The 1985 European comparison revealed that 63 per cent of the estimated total real GDP of the 20 European countries was produced in just four countries, Germany (Federal Republic of), Italy, the United Kingdom and France, where about half of the population of these countries live on 28 per cent of the total territory involved. in respect of differences in the level of per capita real GDP, the group of 20 European countries looks relatively ho- as compared to other regions of the ICP. Yet the fourfold difference between the highest level-two countries, Luxembourg and Norway, sharing the top of the ranking-and that of the twentieth-ranked country (Turkey) is quite significant. For the countries included, the per capita GDP indices were expressed in comparison with the Austrian level. It should be mentioned here that for presenting results of the ECP the "international Schilling" (AS-1), based on the currency of Austria, was chosen as numeraire. This means that the AS-1 value of the GDP of Austria is the same as at national prices, at the total GDP level only. The per capita GDP levels of seven countries, Austria, France, Finland, the Netherlands, the United Kingdom, Italy and Belgium, were very close to each other and to the average of the 20 European countries in 1985. In the context of the global comparisons this means about 65-70 per cent of the United States level or about $US 11,000 at international prices.

The three Eastern European countries, Hungary, Yugoslavia and Poland, in that order, ranked 17, 18 and 19 among the 20 European countries, with per capita real GDP levels reaching 47.2, 44.2 and 37.1 per cent of the Austrian level, respectively.

An interesting feature of the European Comparison Programme for 1985 was that the so-called similarity indices among the price structures and the quantity structures of the participating countries were calculated. The greatest similarities in respect to price structure were found between France and each of the following countries: Belgium, Austria, the Netherlands and Luxembourg; the price structure similarities in 1985 were the lowest between Portugal and the Netherlands and between Poland and each of the following countries: Belgium, Austria and France.

In the European comparison a strong correlation was observed between the exchange rate/PPP ratio and the relative level of per capita GDP. In the most developed countries the ratio of exchange rate to PPP was close to 1, while exchange rates compared to PPPs at the GDP level were 2-2.6 times as high in Hungary, Portugal, Turkey and Yugoslavia.

The results of the ECP have been integrated unchanged into the global comparison of 56 countries for 1985.