The European Comparisons Programme (ECP) - in which Eurostat closely co-operates with the Organisation of Economic Cooperation and Development (OECD) - is aimed at making annual comparisons of Purchasing Power Parities (PPP) and volume comparisons of the main National Accounts (NA) aggregates; GDP and its components of final consumption expenditure. ECP is built to deliver justifiable results at higher levels of aggregation.
GDP level, deflated by PPP is most often the focus of the international comparison of prices and volumes. The real per capita GDP (converted by PPP into common currency) is one of the most important criterion for allocation of EU structural funds.
Purchasing Power Parities (PPPs) are currency conversion rates that are applied to convert economic indicators in national currency to an artificial common currency, called Purchasing Power Standard (PPS), which equalises the purchasing power of different national currencies.
In this article, as distinct from standard PPP publications, four different types of data are presented at the same time. In addition to nowcast, preliminary and final data stemming from the regular PPP dissemination calendar, revised data series are put forward.
The data presented in this paper are final for 2001 and 2002 and preliminary results for 2003.
Preliminary results for every given year are published 12 months after the end of the reference year. At that time, the results of the two consumer price surveys for the reference year as well as preliminary data for all other input data are available.
Final data are published 24 months after the end of the reference year. Collected prices do not change between the preliminary data and the final data but all other input data and particularly GDP expenditure weights that are used in the aggregation of PPPs are subject to revisions. 24 months after the end of the reference year other input data have stabilized sufficiently for the calculation of the final PPPs.
The data referring to the years from 1995 to 2000 are revised. The revision of PPPs was done at the most detailed level for all of the 31 countries from 1999 to 2000 and only for the EU15 and the EFTA countries from 1995 to 1998. For the other countries these years were retrapolated using 1999 revised PPP and implicit price deflator of GDP.
The years 1991 to 1994 for the EU15 and the EFTA countries are retrapolations using the revised figures for 1995 and implicit price deflators.
Table 1 presents the PPPs for 1991-2003 at GDP level, EU-15=1.
In order to eliminate the influence of the different exchange rates and to provide reasonable comparisons of both the volume of GDP and the price levels between the countries, PPPs are used to convert the GDP of 31 European countries participating in the ECP exercise to so called Purchasing Power Standard (PPS).
Table 2 presents the Per capita volume indices of GDP in PPS for 1991-2003, EU-15=100.
Yet there are significant developments for the individual countries if one looks over the whole period 1991 to 2003. For the United Kingdom an upward tendency is visible, while Germany and Italy show slight downward tendencies. France, on the other hand, remains at about 13 to 15 % points above the EU average in all years. However, all four countries belong to the same group in all years.
For Greece and Spain a certain catching-up with the average of the EU can be observed between 1991 and 2003. The same can be said about a number of New Member States, namely the three Baltic States, Hungary, Slovenia and Slovakia.
Table 3 presents the Comparative price level indices at GDP level for 1991-2003, EU-15=100.
Comparative Price level Indices (PLIs) are obtained as the ratio of PPP to the exchange rate for each country and they present a comparison of the price levels with respect to the EU average. A value over 100 indicates a higher general price level than the EU average; a value under 100 indicates a lower general price level. In a country with an index above 100, the use of exchange rates would overestimate the volumes; while the opposite holds if the PLI is lower than 100.
PLIs also indicate the degree to which a country's exchange rate reflects its general price level vis-a-vis the general price level of the reference country: a value over 100 indicates that the exchange rate understates the general price level; a value under 100 indicates that the exchange rate overstates the general price level.
Both volume indices and comparative price level indices are not intended to rank countries strictly but they only provide an indication to the comparative order of magnitude.
In 2002 final results, within the group of highest price level /≥125%/ are 2 EFTA countries (Switzerland and Norway) and 1 EU member state (Denmark). At the same time, in Portugal (73%) and Greece (75%) is observed the lowest price level in EU. Among the Candidate countries, Bulgaria (33%), Romania (35%) and Turkey (47%) have the lowest comparative price level idices.
The 2003 preliminary results do not indicate any changes within the groups.
Following the 1991-2003 results, in 1995 the EU and EFTA members have the highest comparative price level of all years before or after.
Table 4 presents the
Comparative Price Level Indices of Households Final Consumption Expenditure (HFCE) for 1991-2003, EU-15=100.
In 2002 the PLIs for HFCE were between 43 (Slovakia) and 131 (Denmark) among the EU Member States. When considering all participating countries in the ECP the range was wider: from 39 (Romania) to 145 (Norway).
The New Member States and the three Candidate Countries all have PLIs well below the EU-15 average; the highest index being that of Cyprus (88% in 2002). For Bulgaria the HFCE comparative price level index is 40% while the comparative price level index at GDP level is 33%. In comparison, Luxembourg has 99% comparative price level index of HFCE which represent 10% less than the comparative price level index at GDP level.
In general, PLIs for HFCE are less dispersed in the EU than in the whole set of 31 countries. This is confirmed by the coefficient of variation, which is a measure of dispersion, of different country groups. The coefficient was 29% for the EU25, 15% for the EU15 and 19% for the New Member States in 2002. For the whole set of countries the price dispersion was noticeable higher, showing 38% in 2002.
From the now revised time series it can be seen that the PLIs are moving closer to each others - general price levels of all participating countries have been converging over the range of years displayed in Chart 1.
For both, GDP and HFCE, the New Member States have witnessed increase of their relative price levels over the period from 1995 to 2003, and they are getting closer to the EU15 average.
In order to achieve continuity and consistency within the time period before and after the first accession of the 10 candidate countries in 2004 the volume indices for 1995-2003 in
Table 5 are expressed in relation to the
enlarged European Union (ЕС-25=100)3.Although the indices are available in Eurostat for all the participating countries they are not intended to rank countries strictly. They only provide an indication of the comparative order of magnitude of the per capita GDP volume in one country in relation to others. It is therefore preferable to use these indices for dividing countries into groups of a comparable level as shown in Table 6. To even underline this point, the countries are shown within groups in alphabetical order of the codes.
Table 6. GROUPING OF THE ECP COUNTRIES ACCORDING TO GDP PER CAPITA IN PPS, EU25=100, 1995 - 2003
| |
126 % and more
of EU25 average |
101% - 126%
of EU25 average |
76 % - 100%
of EU25 average |
51% - 75%
of EU25 average |
50% and less
of EU25 average |
| 1995 |
AT,CH, LU, NO |
BE, DE, DK, FL, FR, IS, IT, NL, SE, UK |
CY, ES, IE |
CZ, EL, PT, SL |
BG, EE, HU, LT, LV, PL, SK, TR |
| 1996 |
AT,CH, DK, IS, LU, NO |
BE, DE, FL, FR, IE, IT, NL, SE, UK |
CY, ES |
CZ, EL, PT, SL |
BG, EE, HU, LT, LV, PL, SK, TR |
| 1997 |
AT, CH, DK, IS, LU, NO |
BE, DE, FL, FR, IE, IT, NL, SE, UK |
CY, ES |
CZ, EL, HU, PT, SL |
BG, EE, LT, LV, PL, SK, TR |
| 1998 |
CH, IS, LU, NO |
AT, BE, DE, DK, FL, FR, IE, IT, NL, SE, UK |
CY, ES |
CZ, EL, HU, PT, SL |
BG, EE, LT, LV, PL, RO, SK, TR |
| 1999 |
AT,CH, DK, IS, LU, NO |
BE, DE, FL, FR, IE, IT, NL, SE, UK |
CY, ES, MT, PT |
CZ, EL, HU, SL |
BG, EE, LT, LV, PL, RO, SK, TR |
| 2000 |
AT,CH, DK, IE, IS, LU, NO |
BE, DE, FL, FR, IT, NL, SE, UK |
CY, ES, MT, PT |
CZ, EL, HU, SL |
BG, EE, LT, LV, PL, RO, SK, TR |
| 2001 |
CH, DK, IE, LU, NO |
AT, BE, DE, FL, FR, IS, IT, NL, SE, UK |
CY, ES, PT |
CZ, EL, HU, MT,SL |
BG, EE, LT, LV, PL, RO, SK, TR |
| 2002 |
CH, IE, LU, NO |
AT, BE, DE, DK, FL, FR, IS, IT, NL, SE, UK |
CY, EL, ES, PT, SL |
CZ, HU, MT, SK, SL |
BG, EE, LT, LV, PL, RO, TR |
| 2003p |
CH, IE, LU, NO |
AT, BE, DE, DK, FL, FR, IS, IT, NL, SE, UK |
CY, EL, ES, SL |
CZ, HU, MT, PT, SK |
BG, EE, LT, LV, PL, RO, TR |
| p - Preliminary data. |
| Notes: |
| * For 1995 -
1998 the data for the 10 New Member States are based on backwards extrapolations.
Malta is for these years not included in EU25. |
| ** The purchasing
power parities (PPPs) measure the price ratios in the partner countries
and are applied to convert the economic indicators to an artificial common
currency called Purchasing power standard (PPS), which equalizes
the purchasing power of different national currencies, i.e. 1 PPS buys
the same given average volume of goods and services in all countries,
whereas different amounts of national currency units are needed to buy
this volume of goods and services, depending on the national price level. |
| Source of data for grouping:
EUROSTAT: CIRCA/29.11.2004 |
In 2003, based on the updated nowcast PPPs, the grouping of the participant countries was the following:
Group I (
126% of the EU25 average): Switzerland, Ireland, Luxembourg, Norway;
Group II (
101% and <126% of the EU25 average): Austria, Belgium, Germany, Denmark, Finland, France, Iceland, Italy, the Netherlands, Sweden and the United Kingdom;
Group III (
76% and <100% of the EU25 average, i.e. close to the EU25 average): Cyprus, Spain, Greece and Slovenia;
Group IV (
51% and <75% of the EU25 average): Czech Republic, Hungary, Malta, Portugal and Slovakia;
Group V (<50% of the EU25 average): Bulgaria, Estonia, Lithuania, Latvia, Poland, Romania and Turkey.
In 2003 the per capita GDP volume indices varied from 41% (Latvia) to 215% (Luxembourg) within the EU25. For the three EU Candidate Countries the respective indices are as follows: Bulgaria 30%, Romania 30% and Turkey 28%. The per capita GDP is for Luxembourg by far the highest among the participants. One of the reasons for this is the large share of cross-border workers in total employment: while contributing to GDP, they are not taken into consideration as part of the resident population which is used to calculate GDP per head.
GDP per capita volume indices are mainly constructed to compare countries within a given year and they should not be over-interpreted when compared over time, particularly when it comes to small changes from year to year. When compared over a longer time period the per capita GDP indices show a relatively stable pattern. Moving from a group to another is relatively rare. Over the period from 1995 to 2002, only Ireland has significantly changed its grouping, moving from 99% of the EU average in 1995 to the highest group (>126%) with an index of 134 in 2002
In summary, two major observations can be made concerning the revision and resulting series. Firstly, the level of GDP per head in PPS has been revised consistently for the years under revision, reflecting better now, in individual years, countries' relative position to the EU average and to other countries than in the results before (e.g. Denmark, Germany, France, Sweden, Iceland and Switzerland). Secondly, new data show much more logical development over time, even if over time comparison is not the main purpose for which PPPs have been developed. Formerly observed "jumps", have been removed successfully (e.g. Belgium, Denmark, France and the United Kingdom). It can be concluded that the revision has well addressed the criticism raised with regard to the reliability of PPPs and related economic indicators, particularly when it comes to the development over time in comparison with the National Accounts.
Regular annual PPP revisions at Eurostat
PPPs are established on an annual basis, therefore only annual revisions apply. According to the regular publication calendar:
T+5 months: now-cast of PPP for the year T, based on the first GDP and main aggregates estimates for T (available T+4 months) 4 and extrapolated prices of T-1;
T +12 months: preliminary PPP for the year T, based on the first estimate of detailed GDP expenditure weights (available T+9 months), the first consumer price survey of T; and first estimates of other input data.
T+24 months: final PPP for the year T, based on revised detailed GDP expenditure weights (available T+21 months), both consumer price surveys for T and revised data for all other input data.
The weights underlying the PPP calculations, however, are not always fully consistent with the latest aggregated GDP data to which the PPP are applied. This is due to a built-in feature of the NA compilation and revision process, where more aggregated data may be revised more often than the detailed expenditure weights, which are usually revised only once a year and submitted to Eurostat with delays of T+9, T+21 months etc. To obtain full consistency the PPP calculation should be re-run after each ESA95 data delivery to Eurostat. Obviously, this is not feasible and thus, establishing and revising PPP in line with the NA delivery calendar for detailed expenditure weights and applying these PPP always to the latest available aggregated data is a compromise between up-to-dateness of data, stability of data and workload.
Use of Purchasing Power Parities
PPP is a concept that is not immediately and easily understood. This has generated misunderstandings as sometimes PPPs are used for purposes for which they are not suited. In essence PPPs are price comparison in space and therefore the most recommended applications are spatial ones - i.e. comparisons of PPP converted indicators across countries at a given point in time. The most common usages of PPPs are to generate comparable volume measures in per capita terms across countries and to calculate general price level indices. They can be also used in other areas such as labour productivity comparisons at GDP level and in measuring the relative size of economies.
The differences in values of GDP between countries, even when revalued in a common currency using exchange rates, do not only correspond to a "volume of goods and services" component but also to a "level of prices" component, which can sometimes assume sizeable proportions. Exchange rates are determined by many factors, which reflect demand and supply on the currency markets, such as international trade and interest rate differentials. In other words, exchange rates usually reflect other elements than price differences alone. To obtain a pure comparison of volumes, it is essential to use special conversion rates (spatial deflators) which remove the effect of price level differences between countries. Purchasing Power Parities (PPPs) are such currency conversion rates that convert economic indicators expressed in national currencies to an artificial common currency, called Purchasing Power Standard (PPS). This conversion equalises the purchasing power of different national currencies.
Despite being designed for spatial comparisons, PPPs and related economic indicators can be used for inter-temporal comparisons but with certain limitations and with necessary care. In addition, the sampling of items and price collection for PPPs are not designed to capture the pure price change over time as in the case of consumer price indices but price differentials over space. In effect, GDPs converted using PPP should be understood more like current price volume series.
Finally, it needs to be underlined that PPP-based indices cannot be used to establish a strict ranking of countries because PPPs are statistical constructs rather than precise measures. They provide only an indication of the relative order of magnitude in a country in relation to others in the comparison. Therefore, these indicators are best used to assign countries to groups as, for example, in all Eurostat and OECD publications on PPP. In the Table below the uses of PPP-based data are divided into three groups, "recommended uses", "uses with limitations" and "non-recommended uses" in a summary form.
Uses and limitations of PPP-based data
