New Zealand s Involvement in the Joint OECD- Eurostat Purchasing Power Parities Programme

Transcription

1 New Zealand s Involvement in the Joint OECD- Eurostat Purchasing Power Parities Programme James Goodchild and Daniel Griffiths Inflation Measures Division Statistics New Zealand Aorangi House, 85 Molesworth Street PO Box 2922, Wellington james.goodchild@stats.govt.nz daniel.griffiths@stats.govt.nz Abstract New Zealand is one of 43 countries participating in the joint OECD-Eurostat Purchasing Power Parities (PPP) programme. The major use of PPPs is in making inter-country comparisons in real terms of gross domestic product (GDP) and its component expenditures. GDP is the aggregate used most frequently to represent the economic size of countries and, on a per capita basis, the economic wellbeing of their residents. Calculating PPPs is the first step in the process of converting the level of GDP and its major aggregates, expressed in national currencies, into a common currency to enable these comparisons to be made. PPPs are considered an important tool for meaningful inter-country comparisons. This paper discusses New Zealand s involvement in the PPP programme, the use and limitations of PPPs, and the calculation and measurement of PPPs. JEL Classification: E31, E23

2 Disclosure Statement Copyright Information obtained from Statistics New Zealand may be freely used, reproduced, or quoted unless otherwise specified. In all cases Statistics New Zealand must be acknowledged as the source. Liability While care has been used in processing, analysing and extracting information, Statistics New Zealand gives no warranty that the information supplied is free from error. Statistics New Zealand shall not be liable for any loss suffered through the use, directly or indirectly, of any information, product or service. 2

3 1. Introduction Purchasing Power Parity, or PPP, has received attention lately as a tool for inter-country comparison. PPPs are used in the calculation of a country s economic size, standard of living, comparative price level and labour productivity. A significant economic indicator of standards of living is GDP per capita measured in a single currency, namely the USD. PPPs present an alternative framework for comparison of standards of living by converting GDP per capita to volumes, which allows a more robust comparison. Comparisons of standards of living converted by current exchange rates face two criticisms. Firstly, a measure that uses currency values derived from the currency market suggests volatile movements in standards of living. That is, to use currency market valuation suggests that the wellbeing of a country s citizens fluctuates as much as currency markets. Such movements in standards of living are not in line with current experience. Secondly, currency markets include other factors that do not directly affect a country s wellbeing. Currency value given by international currency markets can be affected by other factors, such as currency speculation or trade restrictions, that cause currency value to deviate from economic fundamentals. The more a currency deviates from economic fundamentals the more erroneous it would be to make comparisons based on a measure derived from currency value in currency markets. PPPs are a measure of each country s price for the components of GDP given by the expenditure approach. In each country, prices of goods and services that represent the components of GDP are surveyed at a specific point in time. Price collection for all components is completed over a three-year period and reported to the OECD. Results are then compiled for a benchmark year, when total GDP is collected and deflated by PPP prices, and comparison between individual countries prices are made to the reference group (OECD 30). In essence, per capita GDP is converted to a per capita volume index using PPP price relatives, and then countries are ranked in terms of per capita income and assigned to income groups (ie high income, high middle income, low middle income and low income). Note that comparisons are made across countries at a point in time, that is, comparisons are spatial in dimension, not temporal. In this paper, section 2 outlines the history of the OECD-Eurostat PPP programme. Section 3 outlines the measurement of PPPs, and section 4 follows with the uses and limitations of PPPs. Section 5 concludes with New Zealand s involvement in the OECD PPP programme and includes a brief analysis of the 1999 benchmark year results pertaining to New Zealand

4 2. History of the OECD-Eurostat Purchasing Power Parities Programme The OECD-Eurostat Purchasing Power Parities (PPP) programme is a subset, or regional implementation, of a larger exercise called the International Comparison Program (ICP). However, the origins of international price and volume comparisons of GDP can be traced back to the experimental comparisons undertaken by the Organisation of European Economic Cooperation (OEEC) in the 1950s. This experiment consisted of two studies using different approaches. The first was a comparison made from the expenditure side using mainly price data. Initially, the study covered France, Germany, Italy, the United Kingdom and the United States 1, and was subsequently expanded to include Belgium, Denmark, the Netherlands and Norway 2. The second study was a comparison of the United Kingdom and the United States 3 made from the production side using mainly quantity data. In both studies, purchasing power equivalents rather than exchange rates were used to express GDPs and their components in a common currency and price level. In the late 1960s the ICP was initiated to continue and extend this research. Because the OEEC studies had demonstrated that comparisons made from the expenditure side demand less data than those made from the production side, the ICP would focus on expenditure side comparisons. While the ICP began as a research project, its ultimate goal was to undertake worldwide PPP-based comparisons of GDP on a regular basis. Responsibility for the project was shared by the United Nations Statistics Division (UNSD) and the University of Pennsylvania, with the university taking the lead role. It was organised as a cooperative effort involving many institutions and representatives of the countries participating in the project. The project s initial task was to develop a methodology for a comprehensive system of international comparisons based on PPPs. Experience for developing this methodology was gained from the two OEEC studies; comparisons of centrally planned economies organised by the Council for Mutual Economic Assistance; comparisons carried out in the early to late 1960s in Latin America; and comparisons between centrally planned and market economies started by Statistics Austria during the 1960s. The next task for the ICP was to test the methodology by making actual comparisons using PPPs. The first three rounds, or phases, of the ICP (1970, 1973 and 1975) were essentially experimental. Comparisons were made for a small group of countries that were representative of different income levels, social systems and geographical regions. During phase I, comparisons of GDP were made for ten countries for 1970; six of the 10 countries also provided data for 1967 and for these six countries, comparisons were made for 1967 as well. In phase II, the 10 phase I countries were joined by six others, and comparisons covering both 1970 and 1973 were undertaken for all 16 countries. Phase III comparisons provided results for 34 countries, with 1975 as the reference year. A number of major developments followed the completion of phase III. Firstly, the ICP became a regular part of the work programme of the UNSD, with the University of Pennsylvania continuing 1 See Gilbert M, Kravis I (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 as quoted in the OECD PPP Manual (2003). 2 See Gilbert M, and Associates (1958). Comparative National Products and Price Levels: A Study of Western Europe and the United States, OEEC, Paris as quoted in the OECD PPP Manual (2003). 3 See Paige D, Bombach G (1959). A Comparison of National Output and Productivity of the United Kingdom and the United States, OEEC, Paris as quoted in the OECD PPP Manual (2003). 4

5 to advise on methodological issues. Secondly, Eurostat began to play an increasingly important role in the organisation of comparisons for the European Union (EU) 4, as well as providing technical and financial assistance for regional comparisons in Africa and encouraging the OECD to become involved in the project. The third, and perhaps most significant, development was the regionalisation of the ICP. It became clear after phase III that as the number of participating countries increased, centralised organisation of the project was no longer feasible, especially as no single international body was in a position to undertake the task. Operational advantages had been recognised in grouping countries according to their geographical proximity, and the phase III comparisons for the EU in particular had demonstrated that a comparison tailored to meet the specific requirements of a region need not jeopardise the larger comparison. In fact, Eurostat had organised the comparison for the EU using a different list of representative items and a different aggregation method, while at the same time retaining the basic methodological rules and classification system. As a result, the price and expenditure data could still be used for the world comparison. The following three phases of the ICP (phases IV, V and VI) continued under a regional structure. This placed a greater share of the work on the regional organisations of the United Nations, namely the Economic Commission for Europe (ECE), the Economic Commission for Latin America and the Caribbean (ECLAC), the Economic and Social Commission for Asia and the Pacific (ESCAP), and the Economic and Social Commission for Western Asia (ESCWA). The following table summarises the number of countries involved in each phase of the ICP. Regional grouping Phase I 1970 Phase II 1973 Phase III 1975 Phase IV 1980 Phase V 1985 Phase VI 1993 Africa Middle East Asia Oceania North America Caribbean South America Europe Total Results published / Note that phase VI was started but not completed, although reports were published for the regional comparisons undertaken in Africa, the Middle East, Asia and Europe. After phase VI failed to produce a world comparison, the ICP underwent a thorough independent review initiated by the United Nations Statistical Commission (UNSC). The review identified a number of shortcomings in the management, resourcing and operation of the exercise, and the 4 At this time it was called the European Community (EC) however, it became the European Union (EU) with the signing of the Maastricht Treaty in

6 recommendation to the UNSC was that a new round should not be sanctioned until the shortcomings had been addressed. The UNSC responded by asking the World Bank, the de facto global coordinator of the ICP since 1993, to propose a strategy for a comprehensive solution. The World Bank, in consultation with other interested agencies, drew up an implementation plan for a new round of the ICP. The plan involved mobilising funds from a variety of sources and establishing a governance infrastructure to provide effective management and coordination for central, regional and national levels. It also involved providing complete, clearly written documentation on technical and procedural guidelines and standards. This allowed countries to participate in either a full comparison covering GDP, or a partial comparison covering final consumption expenditure. Where possible, using regular national statistical programmes to obtain price and national accounts data for the ICP and linking participation in the ICP with national statistical capacity building was encouraged. The UNSC considered the implementation plan and agreed to a new round, which began in The European Comparison Programme One of the regional comparison programmes that resulted from the regionalisation of the ICP was the European Comparison Programme (ECP). The Economic Commission for Europe (ECE) is responsible for the ECP and publishes the results of its comparisons, but the actual comparisons are organised by other agencies. This is because the ECP is a combination of independent comparisons undertaken for different groups of countries. The Eurostat-OECD PPP programme is central to the ECP and brings with it coverage beyond Europe through the inclusion of non-european OECD member countries. This is where New Zealand s involvement lies. The Eurostat-OECD PPP Programme Although EU member states were involved in a comparison for 1970, the first official comparison organised by Eurostat was for It covered the nine countries that were EU member states at that time. Eurostat subsequently organised comparisons every five years, in 1980, 1985 and These comparisons were for EU member states and countries that were in line for EU membership, such as Greece, Portugal and Spain. However, countries falling into neither of these two categories also participated, such as Israel in 1980, Austria in 1980, 1985 and 1990 and Switzerland in Encouraged by Eurostat, the OECD started organising comparisons for OECD member countries that were not already included in Eurostat comparisons in the early 1980s. Thereafter the OECD worked closely with Eurostat to effect comparisons for 1985 and When put together, the Eurostat and OECD comparisons covered 22 OECD member countries in 1985 and all 24 OECD member countries in During this time formal agreements between the two organisations were made to establish the Eurostat-OECD PPP programme that New Zealand contributes to. Eurostat and the OECD agreed to coordinate the data collections in two different groups of countries with the object of combining the datasets of the two groups in a single comparison. The coverage of the Eurostat-OECD comparisons is shown below, highlighting New Zealand s involvement. 6

7 Summary of Combined Eurostat-OECD Comparisons Countries Coordinated by Eurostat EU member states Non-EU OECD countries Other countries Countries Coordinated by the OECD Non-EU OECD countries New Zealand Other countries Total The table below shows all the countries that are currently participating in the OECD-Eurostat programme. Eurostat-OECD Participating Countries Eurostat OECD Northern group Central group Southern group Pacific countries Non-OECD countries Finland Denmark Estonia Iceland Ireland Latvia Lithuania Norway Sweden UK Austria Belgium Czech Republic Germany Hungary Luxembourg Netherlands Poland Slovak Republic Slovenia Switzerland Italy Bulgaria Cyprus France Greece Malta Portugal Romania Spain Turkey Australia Canada Japan Korea Mexico New Zealand United States Israel Croatia Macedonia Russian Federation Ukraine 7

8 3. Measurement and calculation of PPPs Measurement The measurement of PPPs is a detailed exercise of international price comparison for about 3,000 goods and services. PPPs are price relatives in national currencies for the same good or service in different countries. Calculation occurs at three levels: product level; product group level; and, finally, weights (from national expenditure) are introduced at the aggregation level. Each country participating in the programme is required to provide a set of national average annual prices and a breakdown of national expenditure. There are approximately 3,000 goods and services for which prices are measured as a sample of all goods and services covered by GDP. These can be aggregated into three groups of products and services: individual consumption of households; gross fixed capital formation; and government consumption. Each of these will be discussed individually. For the 1999 benchmark year, the final product list covered about 2,500 consumer goods and services, 34 occupations in government, health and education services, 186 types of equipment goods and 20 construction projects. The goods and services chosen are subject to specific criteria. As PPPs are spatial comparisons, item specification becomes a much bigger issue than for temporal comparisons. When we undertake a spatial comparison, identical goods need to be priced in each country. To achieve this, goods and services selected for pricing by Eurostat and the OECD (in collaboration with countries cooperating in the programme) need to be representative, comparable and consistent. For a good or service to be representative, that item must represent goods and services that are classed under that particular basic heading as part of final expenditure on GDP. Comparable items must be of the same quality in each country. Prices must be consistent with methods of valuation used to estimate the expenditures on GDP. In addition, all prices should be market transaction prices paid by purchasers. As well as this, for calculation purposes, countries are asked to specify which items priced are representative of their particular country. This occurs because, even though a wide range of items are priced, some items will be more readily available in some countries than others. To make useful comparisons, all items are priced, but one or more items are designated as representative of the basic heading for each particular country. The price of individual consumption by households is measured by a wide range of products and services purchased by households. Surveys are undertaken to gather price information on all goods outlined by the internationally recognised Classification of Individual Consumption by Purpose (COICOP). These surveys are staggered over a three-year collection period and cover: food, drinks and tobacco; personal appearance; house and garden; transport, restaurants and hotels; services; and furniture and health. Some surveys are self-explanatory, although the items in other surveys might require further explanation. Items surveyed for personal appearance include goods such as clothing, footwear and jewellery. The house and garden survey covers non-durables such as whiteware, audio and video equipment, goods used in the maintenance of the house and garden and other miscellaneous household expenses. Transport, restaurants and hotels relates to goods and services such as the purchase, running and maintenance of motor vehicles, dining out and accommodation. Services covers repairs, transport services, education and energy usage. Gross fixed capital is defined as the sum of gross fixed capital formation, change in inventories and acquisitions, less disposal of valuables in the System of National Accounts (SNA 93) and European System of Accounts (ESA 95). For the PPP programme, this is broken down into three expenditure categories: machinery and equipment; construction; and other products. 8

9 Machinery is measured by collecting purchase prices for machinery and equipment goods, such as truck tractor units, heavy machinery, industrial equipment, laboratory equipment and office equipment. Construction is the cost of building a factory building, sewer main, asphalt road or residential building. Other products are defined as plantation, orchard and vineyard development; changes in breeding stocks; land improvements; computer software used in production; mineral exploration; and other intangible assets. Government services are measured by the services provided to households by general government. This is disaggregated into individual consumption expenditure by government and expenditures on collective services. Broadly speaking, individual services are housing, health, recreation and culture, education and social protection; and collective services are general public services, defence, public order and safety, economic affairs, environment protection, and housing and community amenities. Of the 36 basic headings set out by the SNA 93, prices are collected only for nine basic headings to represent government services. The prices collected are used as representative of the basic headings that were not priced. Information for the basic headings that are priced is collected through employee compensation data for a number of occupations. Calculation When price data has been collected, PPPs are calculated at different levels. Firstly, price relatives are calculated at the product level. Items priced are then grouped together to arrive at a simple (unweighted) average price for products. Secondly, products are grouped together to derive simple (unweighted) average product group prices for basic headings. Finally, expenditure for the basic headings is applied to derive aggregates (ie GDP, PHC). The calculation of PPPs uses the Èltetö-Köves-Szulc (EKS) method at the product group level and aggregation level. The EKS method is used to ensure that price relatives for countries are averaged in a way that preserves transitivity. Transitivity holds as long as any chain of indirect comparison between countries will obtain the same result as a direct comparison. That is, to compare the relative price levels for apples between New Zealand and the United Kingdom directly would yield the same result as deriving the relative price level for apples by using the relative price level between New Zealand and Australia, Australia and Korea, and Korea and the United Kingdom (an indirect comparison). This is an important quality for PPPs to have. Another desirable quality of the EKS approach is that comparisons are based on representative products of countries. In calculation, the EKS method uses Laspeyres- and Paasche-type and Fisher indexes. The use of indexes differs somewhat between temporal comparison and spatial comparison. Whereas for temporal comparison an index would have a base and current period, in spatial comparison an index has a reference country and a partner country. The Laspeyres-type index is calculated using the relative price between the representative goods of the base country and the partner country. The Paasche-type index is calculated using the relative price between the representative goods of the partner country with the base country. Another difference between the Laspeyres and Paasche is that calculation for basic headings is given by the geometric mean, not the traditional arithmetic average. Calculation of EKS PPPs for a basic heading (product group level) is undertaken in six steps. The first three steps are the calculation of matrices of Laspeyres- and Paasche-type and Fisher PPPs (the Fisher index is the geometric mean of the Laspeyres- and Paasche-type indexes). For many basic headings, some country Fisher PPPs will be missing that need to be completed. This is done by calculating the geometric mean of all the indirect Fisher PPPs connecting the 9

10 countries for which PPPs are missing. Stage five is completing the matrix of Fisher PPPs using the EKS method. This takes the geometric mean of the direct Fisher PPP and all indirect Fisher PPPs to ensure transitivity. This gives a matrix of EKS PPPs. The geometric mean of all the country EKS PPPs is then taken, to give a vector standardised EKS PPP. The calculation of aggregates makes use of the PPPs calculated at basic headings and the weighting information from national expenditures. This allows for the calculation of weighted Laspeyres- and Paasche-type indexes. There are five steps in the aggregation process, which is very similar in process to the calculation of PPPs for basic headings. The first three steps are the calculation of Laspeyres- and Paasche-type and Fisher indexes. Steps four and five are the calculation of EKS PPPs and the standardising of EKS PPPs, respectively. 4. Uses and Limitations of PPPs PPPs are used to generate volume measures with which to compare levels of economic performance, economic growth and overall productivity, or as measures of price convergence and competitiveness. These measures are utilised by many users, such as universities, research institutes, journalists and, most notably, international organisations. Among these international organisations are the International Monetary Fund, the World Bank, United Nations, OECD and European Union (EU). The EU uses PPPs in determining the allocation of Structural Funds. In New Zealand, the Reserve Bank has used PPPs in conjunction with Uncovered Interest Rate Parity (UIP) to derive an equilibrium exchange rate 5. The most commonly used PPP statistics are the comparison of GDP and GDP per capita using PPPs as a common currency. PPPs represent a way of converting GDP in domestic currencies to a common unit of measurement. This conversion can also be undertaken by converting all domestic GDPs into US Dollars. The associated problem with this is that converting GDP into US Dollars uses market exchange rates. Market exchange rates are a measure of the relative cost of (traded) goods and services between countries, but also represent levels of financial flows of capital, relative levels of interest rates and trade restrictions. Also, market determined exchange rates can be volatile. This suggests that GDP in a common currency would also exhibit volatility and makes market exchange rates ill suited for currency conversion of standards of living and the size of economies. PPPs are used to compare the relative size of economies. This comparison is done by the size of an economy relative to the overall size of the OECD group. An individual country s GDP (measured in PPP currency ) is then compared to the total GDP of the European group of 30 countries (OECD 30). In the 1999 benchmark year, New Zealand contributed 0.28 percent of the total OECD GDP measured by PPPs 6. This can be compared to the share of total GDP calculated by market exchange rates (0.22 percent). A significant difference can be seen between the two approaches, and the discrepancies are larger for lower-income countries. An interesting outcome of the 1999 benchmark year results is that the 15 European Union nations are virtually equivalent in size to the United States of America when compared using PPPs. GDP per capita is most often used as an indicator of standards of living. Although criticised as an incomplete measure of economic welfare, GDP per capita remains the prevalent measure of relative standards of living and economic performance. It provides an indication of the output for every person in a country, that is, when in a common unit (PPPs), GDP per capita is the volume of goods and services available to everyone living in the country. Assuming that more goods and services are preferred and provide higher utility, then higher GDP per capita is preferred 5 See Stephens D (2004). 6 In comparison, Australia contributed 1.85 percent by PPPs and 1.56 percent by exchange rate-conversion. 10

11 (monotonicity). The strength of PPP GDP per capita is its ability to provide a volume measure. Using PPPs and not exchange rates ensures that domestic GDP per capita is converted with an appropriate methodology, deflating each component of GDP by the relevant prices. GDP per capita for individual countries is represented by an index of the country s position relative to the OECD average (OECD 30=100). This allows the ranking of countries relative to one another in a comparable way. To summarise, the GDP per capita results countries are grouped into four income brackets according to their position relative to the OECD average. The high income group are the countries with an index value of greater than 120, that is, they have a per capita GDP at least 20 percent higher than the OECD average. Among this group (based on 1999 benchmark year results) are Luxembourg, the US, Norway, and Switzerland. The next group of countries are those with index values of 100 to 120, which are grouped as the high middle income group. Inclusions among this group include Canada, Ireland, Japan, Australia, Germany, United Kingdom and France. The next group is the low middle income group (50 to 100) and includes countries such as Israel, Spain, New Zealand, Greece and Korea. The last group is the low income countries (below 50), including Poland, Mexico, Russian Federation and Turkey. Another useful PPP statistic is labour productivity. Measures of productivity are constructed in the same way as GDP per capita, but they use employment data to construct an index of GDP per person employed. Comparison of per capita GDP and labour productivity can now be made. This comparison provides insight into possible further gains to productivity and competitiveness, and gives perspective to a country s growth relative to its current level of income and productivity. As, by definition, labour productivity is a volume measure of output to volume of labour input, PPPs as a volume measure are indispensable. Differences between the measures of GDP per capita and labour productivity can point to differences in labour utilisation and demographic structures 7. Comparative price levels are important in the study of the purchasing power of currencies. This can be measured by the ratio between PPPs and market exchange rates. If this ratio is equal to one, then it can be said that, on average, one unit of currency can purchase the same amount of goods and services in the foreign country under consideration as in the domestic country. When PPP exceeds exchange rates, it can be said that one unit of the currency under consideration buys less domestically than on other markets and vice versa 8. It is strongly emphasised that this ratio is not a measure of over- or undervaluation of a currency, but a measure of the purchasing power of currency. Pertinent to the study of development, comparative price levels have been used in the study of the relationship between price level and per capita GDP. There exists a positive correlation between comparative price levels and GDP per capita. For countries with low income per capita, exchange rates often exceed PPP rates, indicating low price levels, and lead to an understatement of the size and importance of these economies if market exchange rates are used. The low comparative price level could be indicative of the importance of non-traded goods that can be obtained relatively cheaply in developing countries. The two most universally accepted uses for PPPs are for inter-country comparisons of GDP statistics and inter-country comparisons of price levels. PPPs have also been used for other purposes, however there are limitations and there are areas where PPPs should not be used at all. Limitations in the use of PPPs arise for a number of reasons: precision, difficulties with 7 The OECD also notes that differences could arise from possible statistical issues concerning the measurement of labour input. 8 See Schreyer P, Koechlin F (2002). 11

12 temporal comparisons (ie changing sample and changing price structures), and low level comparisons and comparisons of disaggregated data. PPPs are a useful instrument but with limited precision. Small differences between countries should not be over-interpreted. When using PPPs to compare GDP per capita across countries, results may change due to revisions in GDP and population data. In this context, country comparisons are best made by grouping countries with similar results. The example below shows the middle income group of OECD countries ranked by GDP per capita. You can see that New Zealand and Spain are very close together, with New Zealand ranking slightly ahead. Due to the limited precision of these comparisons, it would be unwise to draw any conclusions about differences in the economic wellbeing of these two populations. You can conclude, however that both New Zealand and Spain are in the low-middle income group Results, GDP per Capita: Middle Income Group Iceland Canada Netherlands Ireland Austria Japan Belgium Germany Australia Sweden Italy United Kingdom Finland France New Zealand Spain Portugal Greece Korea Czech Republic Hungary Because PPP samples are devised to maximise the comparability of items across countries, they tend to be small and change over time. Samples for temporal price indexes, on the other hand, tend to be larger and are devised to maximise comparability over time within a country, but not to deliver international comparability of items. Therefore, when PPP samples change over time, it makes little sense to compare prices over time. Temporal comparisons are also hindered by changing price structures. Comparing, for example, New Zealand s per capita GDP relative to that of the OECD as a whole between 1996 and 1999, it would make economic sense if the difference was equal to New Zealand s rate of GDP growth relative to the OECD as a whole between 1996 and 1999 divided by the relative population growth. However, this is not the case. The 1996 result uses fixed PPPs (fixed 1996 international prices), whereas the 1999 result comes from the 1999 survey and therefore reflects current 1999 international prices. Therefore, one source of the differences are changes in price structures. 12

13 PPPs are an aggregate economic measure. Therefore, lower-level economic comparisons using PPPs are not recommended. Cross-country productivity comparisons by industry, for example, should not be undertaken unless industry-specific PPPs are available. The OECD has advocated a traffic light approach in considering the appropriate use of PPPs 9. A summary appears below: Areas where PPPs are well placed: Volume comparisons of GDP: a. GDP per capita b. GDP per hour worked c. Size of economies. Comparisons of relative price levels. Areas where PPPs may be used but with limitations: Cost of living index across countries. Time series analysis of relative GDP per capita. Analysis of price convergence. Areas where the use of PPPs is not recommended: As an indicator for the over- or under-valuation of a currency. As a precision tool to establish rankings between countries. As a measure to generate output and productivity comparisons by industry (unless there are industry-specific PPPs). As a way of constructing national growth rates. As a volume measure of exports or imports across countries. 5. New Zealand s Commitment to the OECD-Eurostat PPP Programme New Zealand s involvement in the PPP programme began in the early 1980s, with price data supplied on an as requested basis. Historically, New Zealand has not contributed actively to the PPP programme; requirements were met by a combination of using existing data and some supplementary pricing. In the late 1990s Statistics New Zealand reassessed its commitment, and a more concerted effort was made to be involved in the data collection process, provision of prices, index outputs, methods discussions and meetings. November 2001 saw Statistics New Zealand s first attendance at an OECD PPP member countries conference in Paris. Staff also attended conferences in Mexico in 2002, and subsequently hosted the Non-European OECD countries PPP Conference in February Going forward, Statistics New Zealand is increasing its commitment to the PPP programme. A much more active role in data supply, attending conferences and input into pricing specifications 9 See Schreyer P (2004). 13

14 has already been achieved. There still exist areas for improvement. At present, work is being finalised for the 2002 benchmark year results. Full results are due to be released by the end of Turning now to the PPP results for the 1999 benchmark year, figure 1 presents a comparison of GDP per capita converted by PPPs and exchange rates. The results are telling of the difference between conversion using PPPs and exchange rates. As an example, in examining Japan, exchange rate-based GDP per capita suggests that Japan is 68 percent higher than Canada, whereas by PPPs, this difference reduces to only 6 percent. This is a reflection of the higher relative price in Japan. For New Zealand, it can be clearly seen that using exchange rate-converted GDP per capita understates the true volume of goods and services per person. Using PPP-based conversion, New Zealand has a GDP per capita index value of 83, compared with 64 using exchange rate conversion. This represents a 30 percent improvement. Such a result would strongly suggest that New Zealand has a low comparative price level. This is supported by figure 2. As discussed earlier, precise rankings should be avoided and instead countries should be assigned to income groups, as high income, high middle income, low middle income and low income. It can then be seen that New Zealand falls into the low middle income group of countries which have index values of between 50 and 100. Figure 1 Final Expenditure on GDP Per Capita for Selected Countries 1999 (OECD30 = 100) Based on PPPs Based on exchange rates United States Canada Japan Australia Country United Kingdom EURO 12 New Zealand Korea Mexico Indexes of final expenditure on GDP Figure 2 presents GDP per capita and comparative price levels. Any country with a comparative price level of less than 100 will have a greater GDP per capita converted using PPPs than exchange rates, and vice versa. Japan, as previously discussed, had a large divergence between its GDP per capita converted by PPP and exchange rates. This can be explained by the very high price level prevalent in Japan, given by a comparative price level of

15 Figure 2 Comparative Price Levels and Indexes of PPP-Based GDP Per Capita 1999 (OECD30 = 100) PPP-based GDP per capita Comparative price levels PPP-based GDP per capita Comparative price levels United States Canada Japan Australia United Kingdom EURO 12 New Zealand Korea Mexico 0 Country As with GDP per capita groupings, countries can also be classified into price level groups. Corresponding roughly to income groupings, countries are divided into high price level (110 and above), medium high price level (between 90 and 109), medium low price level (between 60 and 89) and low price level (less than 60) groupings. New Zealands comparative price level of 77 places it in the medium low price level group. 15

16 References Ahmad S (2003). Historical Overview of the International Comparison Program, World Bank ICP paper. OECD (2003). Revised draft outline of the manual for the Eurostat-OECD PPP Programme. Schreyer P, Koechlin F (2002). Purchasing power parities measurement and uses, OECD Statistics Brief March, No. 3. Schreyer P (2004). The PPP Programme and the Uses of PPPs, OECD, presentation to Statistics New Zealand staff and public, February, Wellington. Stephens D (2004). The equilibrium exchange rate according to PPP and UIP, Reserve Bank of New Zealand, Discussion Paper 2004/03. 16