From:http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/ICPEXT/0,,contentMDK:20118247~menuPK:299187~pagePK:60002244~piPK:62002388~theSitePK:270065,00.html 11/17/08 How are PPPs calculated? INCLUDEPICTURE "http://web.worldbank.org/shared/images/shim.gif" \* MERGEFORMATINET INCLUDEPICTURE "http://web.worldbank.org/shared/images/shim.gif" \* MERGEFORMATINET INCLUDEPICTURE "http://web.worldbank.org/shared/images/shim.gif" \* MERGEFORMATINET The overall PPP for GDP is built up from comparisons of item prices within smaller groups of expenditures that, in the terminology of the ICP, are referred to as basic headings. The requirement for basic headings are that: Value data, representing national expenditures, can be estimated for each basic heading. Basic headings be as homogeneous as possible from the point of view of the potential dispersion of the price ratios across countries of the goods and services belonging to each respective heading. Usually, the basic heading level is chosen on the basis that within a basic heading more detailed expenditure weights are not available. Since country groups will differ on the amount of expenditure detail that is available, the number of basic headings has varied between 150 and 250 for different regions and country groupings within the ICP. Within each basic heading, a country will provide prices for a selection of individual items from a set of written specifications developed by the ICP over the years. Within a basic heading price ratios of individual items, for example, a kilo of ?long grained rice in a plastic bag? are aggregated to produce price parities at the basic heading level. The price ratios and basic heading parities are all denominated in CCUs, national currency units per unit of the numeraire currency. For example, if country A is the numeraire country and its price for long grained rice is $2 per kilo, and the price in country B is 28 rupees per kilo, the price ratio will be 14 Rs/$. The input provided by the country statistical offices to the ICP, thus, are expenditures at the basic heading level and prices of items representing the corresponding basic heading for that country. This is the first and most basic step in moving from national data to international comparisons of real volumes and purchasing power parities Chapter 1 Introduction to the International Comparison Program Purpose of the Handbook The ICP 2003-2006 Handbook is at the center of a range of documentation prepared for the 2003-2006 round of the International Comparison Program (ICP). Although written for the current round, the Handbook relies heavily on experiences and operating procedures developed in earlier rounds of the program. Its purpose is to explain the methodology used to calculate Purchasing Power Parities (PPPs) from data collected in 2005. It is intended mainly for all those involved in, or with responsibilities for, the compilation of PPPs, including professional economists and statisticians working at all levels in National Statistical Offices and the various international and regional agencies involved in the program. Supplementary manuals (?Operational Manual for Regional Coordinators? and ?Manual for Price Collectors?) are also available to help National Statistical Offices and Regional Coordinating Offices collect price data and organize their work in a consistent and efficient manner. The Handbook is divided into 16 chapters, each dealing with a topic essential for understanding conceptual and methodological issues related to the program. This chapter provides background information about the program, a description of the relationships between PPPs and exchange rates, common fallacies about PPPs, uses of international volume and price data, calculation of PPPs, and integration of data collection and processing between ICP and CPI, etc. Chapter 2 describes the strategic framework adopted for the 2003-2006 round of the program, including its ownership, funding and governance. Chapter 3 defines the final expenditure components of Gross Domestic Product GDP and explains the prices used to value them. Chapter 4 addresses basic price concepts, representativity, comparability, brands, quality and methods of quality adjustment, etc. Chapter 5 explains how product lists (ICP basket of goods and services) are drawn up. Chapter 6 deals with the sampling survey procedures for household consumption goods and services. Chapter 7 contains guidelines for validation and editing of ICP data at national and regional levels. Chapter 8 deals with the price information needed for government services, especially on the data regarding compensation of government employees. Chapter 9 provides a detailed explanation of issues related to gross capital formation. Chapter 10 explains (a) the methods that may be used to estimate expenditure on dwellings as part of final consumption expenditure of households; (b) the information required on rents; and (c) the data on dwelling services that participating countries are required to submit for ICP 2003-2006. Chapter 11 describes the methods used to estimate regional purchasing power parities at the level of a basic heading. Chapter 12 explains how basic heading parities are to be aggregated to levels of the GDP within a region. Chapters 13 and 14 describe the Ring program to link the regional parities into a global comparison and the method used to compute the basic heading parities, respectively. Chapter 15 deals with inter-regional aggregation. The handbook will conclude, in Chapter 16, with material about publication, data access, and several appendices including a glossary. Background Information The biggest challenge in international comparisons of economic aggregates is conversion of national currency denominated values into a common currency. The literature in international economics supported by common sense indicates that cross-country comparisons that rely on market exchange rates are prone to yield misleading results due to frequent movements in exchange rates out of line with domestic inflation. To overcome the deficiencies of exchange rates, the ICP was established in 1968 by the United Nations and the University of Pennsylvania to facilitate inter-country comparisons of Gross Domestic Product (GDP) and its components based on PPP. A PPP is defined as the number of currency units required to purchase the amount of goods and services equivalent to what can be bought with one unit of the currency of the base country, for example the U.S. dollar. PPP conversion rates allow users to compare real economic outputs across countries at a common set of average international prices. They are analogous to time-series indices, which measure changes in national output over time at constant prices. PPPs are averages of price ratios between countries. Each country participating in ICP provides national average prices for more than 1000 closely specified products. The characteristics of each product are specified very carefully to ensure that countries measure similar products for the sake of comparison. Product prices are collected on average 4 times a year in many markets and outlets?rural and urban, formal and informal ? to obtain the national annual average prices. PPPs are computed by obtaining average ratios of average prices between countries for comparable items. Uses of PPP Data PPPs are widely used both in academic research and various applications. International organizations, such as the OECD , United Nations, World Bank, IMF and Eurostat, have always been major users of PPPs. PPPs have been used in multiple academic papers dealing with international aspects of pricing, growth or structural convergence. There is also a growing demand for PPPs from a variety of other national users, such as government agencies, public enterprises, private firms, banks, journalists and individuals who conduct studies and make policy analyses requiring comparisons of GDP and average price levels between different countries. A growing area of using PPPs is poverty comparisons. Other uses include: Comparisons of GDP, GDP per head, GDP per hour worked, size of economies based on PPP conversion to a common currency; Comparisons of relative price levels; Inter-temporal analysis of relative GDP per capita or relative prices; Analysis of price convergence; Cost of living index across countries. The 2003-2006 round of the ICP covers around 147 countries in all regions of the world, including OECD and European Union members. PPPs generated by ICP are based on a global survey of prices. It is a highly complex operation whose structure and organization is explained in Chapter 2. Comparisons of Price and Volumes The ICP works with GDP measured from the side of final expenditures: household and government consumption, gross capital formation plus net exports. These flows are defined in the ICP in the same way as the internationally agreed System of National Accounts (SNA), thus ensuring that the GDPs of different countries valued in different currencies are conceptually consistent. The change in the value of the GDP for a single country between two time periods can be decomposed into two components: Volume change (or volume index); Price change (or price index). In order to compare the volumes of the GDPs, or other expenditure aggregates, it is necessary to express them in common terms by removing effect of price change. In international comparisons, however, the ratio of the values of the GDPs in two different countries splits into the following components: Volume ratio (or volume index ); Price ratio (or PPP). PPP can be further decomposed into the following: Exchange rate; Price level . The exchange rate simply converts the GDPs into the same currency units. Even when valued in the same currency unit, the ratios of GDPs in different countries have to be split into their volume and price components. The GDP for a single country in two time periods is valued in exactly the same currency unit in both periods, but the change between the two periods has to be split into its volume and price components to isolate the volume change. It is a known fact that, after converting expenditure flows in different countries into a common currency, say the dollar, using exchange rates, the dollar prices are not the same in different countries. Not only are the prices of individual goods and services different but there are systematic differences in price levels between countries. Price levels are usually lower in poorer countries. It is necessary to adjust for these price level differences to get at the volume comparisons. The task of the ICP is to measure these differences in price levels by comparing the prices of the same products in different countries. The objective is to construct price indices using the same approach as is used to measure price inflation within a country. PPPs, Exchange Rates, Volumes and Prices The inter-relationships between the PPPs, exchange rates, volume indices, and price indices for GDP or other expenditure aggregates such as household consumption can be summarized as follows: GDP or expenditure ratio = volume index x PPP (as per definition of PPP) GDP or expenditure ratio = volume index x exchange rate x price level index PPP = exchange rate x price level index Price levels would be the same in different countries only if the exchange rate equaled the PPP. Exchange rates depend on many factors and all the empirical evidence strongly suggests that in general they deviate from the PPPs and are often volatile which makes them difficult to be used in international comparisons. PPPs have to be calculated in order to obtain volume comparisons. PPPs are different from price indices in inter-temporal comparisons in the sense that they are not ratios that measure the percentage by which one price level exceeds another. Their dimensions are different as they measure ratios of currencies. Their values depend on the units in which the currencies themselves are measured. However, it can be seen from the third equation above that PPPs can easily be transformed into price level indices by dividing them by the corresponding exchange rates. It is purely a matter of convenience whether to calculate PPPs or price level indices from the price data collected in different countries in the ICP. In order to calculate price level indices, the prices in different countries would first have to be converted into a common currency unit using market exchange rates. It is less trouble to calculate PPPs as they can be calculated directly from the prices in different currencies. It has long been customary in ICP work to calculate PPPs first and derive the price level indices from them subsequently, rather than vice versa. Price level indices have all the properties of regular price indices from inter-temporal comparison. Moreover, it is important to note that PPPs are transformed into price level indices simply by dividing by a scalar, the exchange rate. PPPs, thus, behave like regular price indices and have all the properties of price indices. Two Special Cases In the special case where different countries use the same unit of currency, for example, the countries of the European Monetary Union, the exchange rate is unity between every pair of countries so that the PPPs are the price level indices. Thus, the PPPs show relative price levels in the different member countries directly. Another special case would be a hypothetical situation in which exchange rates were such that the price levels in all the countries compared were the same. The price level indices would all be unity and the PPPs and the exchange rates would coincide. In this hypothetical situation, exchange rates could be used to make volume comparisons between countries. However, as already stated, there is overwhelming evidence to demonstrate that price levels often vary substantially in different countries. This is why the ICP is needed. Two Common Fallacies Notwithstanding the previous paragraph, many users of the data assume that GDPs converted at exchange rates provide volume comparisons directly. This fallacy leads to faulty analysis and inappropriate policy recommendations. Because exchange rate converted GDPs are expressed in the same currency unit, such as the dollar, it often seems to be tacitly assumed that they must be valued at the same price levels. However, no one would commit the same fallacy when comparing the current dollar values of US GDP, for example, in two different time periods. Yet, the evidence clearly shows that the differences in price levels between developed and developing countries, for example, can be much greater than the changes in the price levels between successive time periods in the same country. Another common fallacy is that PPPs are calculated because they provide estimates of underlying equilibrium exchange rates. There are no assumptions whatsoever about exchange rates, or their determination, underlying the ICP other than that, in general, exchange rates are different from PPPs. Some Uses of International Volume and Price Data Volume comparisons of GDP or other expenditures aggregates are needed for several purposes: Living standards in different countries may be compared by comparing per capita volumes of household consumption, or components of household consumption. Comparisons of poverty may be made using selected consumption data. Aggregate productivity levels may be compared by comparing per capita volumes for GDP as a whole. Gaps in living standards or productivity levels may be tracked over time to see whether inequalities in the distribution of world income are increasing or decreasing, and how fast. Policies governing loans and other forms of aid or assistance to poorer countries require information about comparative living standards and productivity levels that cannot be obtained without PPPs. Budgetary contributions to international agencies often take account of differences in the volume of GDP, or per capita GDP, in the member countries. Companies or other agencies operating in several countries are interested in the relative price levels in those countries. Tourists and others travelling from one country to another are also very interested in the relative price levels. There is demand for the price indices as well as the volume indices. The Calculation of PPPs The concept of a PPP is clear and simple at the level of a single good or service. For example, consider salt. If the price of a kilo of a salt in country A is PA units of currency and the price in country B is PB units of currency, the PPPAB for salt is defined as the ratio PB /PA. The ratio is usually normalized by setting PA equal to one, so that the PPP can be expressed as a certain number of units of currency B per unit of currency A. If a given amount of A?s currency is converted into B?s currency at the salt PPP rate of PB / PA it must, by definition of the PPP, purchase the same quantity of salt in B as it can in A: hence, the name ?purchasing power parity?. However, the measurement of PPPs runs into traditional index number problems of the kind encountered in inter-temporal price indices as soon as more than one good or service is involved. In practice, the relative prices of different kinds of good and services vary from country to country because demand and supply conditions differ between countries. The greater the variation in relative prices between a pair of countries, the greater the variation in the individual PPPs for different goods and services. In order to obtain a comprehensive PPP covering a wide range of different goods and services, such as the consumption goods and services purchased by households, it is necessary to take some kind of average of the individual PPPs. They also have to be weighted in order to reflect the relative importance of different kinds of goods and services. Relative quantities vary from country to country in response to variations in relative prices. Consumption patterns vary greatly between countries for a variety of reasons: differing relative prices, differing climates, tastes, and cultures, differing income levels, etc. Some goods and services may not be consumed at all in some countries. Thus, the average PPP for a set of individual products will vary, depending on which country?s consumption pattern is used to provide the weights. Aggregate PPPs The PPP for total GDP, or total consumption, looked at from the perspective of the residents of country A will not be the same as that viewed from the perspective of country B. Two different PPPs, and hence two different international price indices, can be calculated from the same set of price observations depending on which country?s expenditure patterns are used as weights. These are the international equivalents of the familiar inter-temporal Laspeyres and Paasche indices. However, the gap between the international Laspeyres and Paasche price indices can be very much greater than the index number spread encountered in inter-temporal price indices because patterns of consumption can vary so much between countries. If equal importance is attached to both countries there is no reason to prefer either the Laspeyres or Paasche type indices. In this situation, the Fisher index, i.e., the geometric average of the two indices, is commonly used. Another problem when many countries are included in the program is that binary PPPs calculated separately between pairs of countries are not transitive . Either the binary PPPs must be adjusted to make them transitive by using the EKS method, or multilateral methods must be used instead to calculate transitive PPPs for the entire set of countries simultaneously. PPPs for Basic Headings The calculation of PPPs has to proceed in two stages because detailed expenditure data are only available down to a certain level of disaggregation. The same problem occurs in CPIs . The smallest aggregate for which expenditure data are available is called a basic heading. The estimation of the elementary PPP for a basic heading has to be made purely on the basis of the price observations as there is no information on the corresponding quantities and expenditures within the heading. However, the underlying index number problems do not disappear because the expenditure data are not available. The solution adopted in the ICP is to introduce an implicit weighting system by distinguishing between representative and unrepresentative products. The formulae used to calculate the elementary PPPs from the raw price data are described in Chapter 10. However, the basic ideas can be summarized as follows. First, prices are collected for a limited number of individual products within each basic heading, the numbers, typically varying between 5 and 15, for practical resource reasons. If the estimation of the elementary PPP is approached from the perspective of country A, the ideal choice of products would be those that are most representative of the expenditures within the basic heading in country A. The representative products are simply those that are known, or believed, to account for the largest shares of the expenditures. In principle, products can be ranked by order of their expenditure shares. The n most representative products are the n products that account for the largest possible share of total expenditures. A representative product is therefore one that it is believed would carry a relatively large expenditure weight if the expenditure weights were known. In practice, representivity will inevitably vary from country to country. The products that are most representative of country A need not be the most representative for country B, nor C, nor D, etc. Consider the elementary PPP between A and B. The representative products of the two countries may overlap but they will not be the same, especially if the countries have different standards of living, climates, customs, etc. One option is to calculate the PPP as an average of the price ratios for country A?s representative products. This is a rough and ready attempt to estimate what the Laspeyres index for the basic heading would be for country A if the expenditure data were available. Alternatively, an average can be taken of the price ratios for B?s representative products which provides a rough estimate of the Paasche index. The index based on A?s representative products will tend to greater than that based on B?s products for the same reason that Laspeyres tend to exceed Paasche: namely, the likely negative correlation between the price ratios and the (unknown) quantity ratios. As there is no reason to prefer one estimate to the other assuming equal importance is attached to both countries, the solution adopted is to take a geometric average of the two, as with the aggregate indices. The resulting index may be expected to provide a rough estimate of the Fisher index. When there are many pairs of countries, the separate binary indices will not be transitive. In order to impose transitivity, the EKS method may be used. Using this methodology, the sets of products used to estimate elementary PPP for a given basic heading are not the same for different pairs of countries. As the objective is to calculate elementary PPPs between every pair of countries, the overall ICP list of products to be priced within a basic heading has to include representative products for all the participating countries. Procedures have to be established to ensure that each country?s representative products are identified and included on the overall list of products for that basic heading. These procedures are explained in Chapter 5. It should be noted that the methodology just described is not the only one endorsed by the ICP. Other methods have been proposed and used in the past, in particular the Country Product Dummy (CPD) method. The methodology is not immutable. It has to be kept under review and evaluated in the light of ongoing research in this area. Comparability The individual products whose prices are compared between countries must obviously be the same. The prices in the numerators and denominators of the ratios that are fed into the calculation of the elementary PPPs must refer to exactly the same product. As the price collectors cannot observe the products that price collectors in other countries are pricing, each price collector is given a very precise or tight specification of the product, including a picture where possible. The specification is so precise that it is hoped that the same product can easily be identified in different countries. Specifying all the relevant price determining characteristics of a product requires a great deal of careful preparatory work. The procedures used in ICP to draw up detailed specifications of the products to be priced are described in Chapter 5. Role of PPPs in the Statistical System PPPs should be integrated into a country?s statistical system as closely as possible. They should not be viewed as ad hoc special purpose statistics that are outside the mainstream of a country?s statistics. The concepts, definitions and classifications used in the ICP are consistent with those used more widely in other economic statistics. As the SNA occupies a central place in the statistical systems of most countries, a large measure of consistency is achieved by the ICP adopting the same concepts and classifications as the SNA. In any case, the PPPs have to be consistent with the expenditure data drawn from the national accounts, because the weights for the calculation of PPPs come from the national accounts. The PPPs also have to match national accounts aggregates used to obtain measures of real GDP and consumption. On the other hand, it is not possible to ensure complete consistency between PPPs and the temporal price indices in the national accounts. These are implicitly related because changes in PPPs over time are dependent on the relative rates of inflation in different countries. However, the system of weights used to calculate PPPs are influenced by data from more than one country, whereas the weights used to calculate price indices within a country are not. The need for transitivity within the ICP also imposes constraints on the types of index formulae used to calculate a set of multilateral PPPs that do not apply to the price indices within a single country. For these and other kinds of reasons, the index formulae used to measure price movements within a country and those used to calculate multilateral PPPs are generally not fully consistent with each other. This can create discrepancies between the two kinds of data which may be difficult to explain. It is impossible to avoid these problems, but care should be taken to try to minimize them. Integration of Data Collection and Processing between ICP and CPIs One of the specific objectives of the ICP is to strengthen or reinforce CPI capability within countries, especially those with limited statistical resources. In many countries, consumer price indices (CPI) are not intended to cover the whole of household consumption as defined in the SNA. It is common for CPIs to exclude certain types of households and certain types of goods and services, whereas PPPs cover all households and all consumption goods and services. Secondly, existing CPI programs run continuously with a fairly steady work flow, whereas ICP data are collected intermittently. In principle, PPPs refer to a point of time, or single period of time, although the process of data collection is usually staggered in practice to avoid bottlenecks in both the collection and the processing of the data. Notwithstanding these differences in coverage, the price surveys and data processing procedures for CPIs and the household consumption PPPs in the ICP should be aligned with each other as closely as possible in order to exploit economies of scale and obtain the maximum benefits for both. Price data collected for CPI purposes should be used to the maximum extent possible for ICP purposes. Additional price surveys required for ICP should exploit the existing CPI survey infrastructure as much as possible. The computer software developed for data collection and processing for the ICP, namely the Tool Pack, is intended to meet the needs of staff working in both areas OECD -Organization for Economic Cooperation and Development; Eurostat- Statistical Office of the European Union. Measure of differences in real expenditure across countries Measure of overall price level in a country relative to its exchange rate (PPP/Exchange rate) Transitivity: The property of indices when the price or quantity relationship among any two of three countries is the same, whether derived from an original-country comparison between them or from the comparison of each country with any third country. In the case of three countries, where PPP is a price or quantity index and j, k and i are countries, the transitivity test is satisfied if PPPjk = PPPji /PPPki. When this test is satisfied, there is a unique cardinal scaling of countries with respect to relative quantities and prices. CPI: A measure of the average change in consumer prices over time in a fixed market basket of goods and services
Want to see the other 13 page(s) in How are PPPs calculated.doc?JOIN TODAY FOR FREE!