The World Bank uses consumption PPP rates estimated from International Comparison Program (ICP) surveys to estimate international poverty. This section explains the basic ICP 2005 methodology for estimating private consumption PPP rates. Because PPPs for consumption are computed as a component of GDP, the sources and methods used to compile GDP PPP are similar to those used to compile consumption PPP.
Statistically, PPPs are expenditure-weighted averages of relative prices of a vast number of goods and services on which people spend their incomes. Calculation of PPPs requires two sets of data: (a) GDP expenditure broken down into 155 detailed components called “basic headings” (of which 110 are related to private consumption); and (b) national annual average prices of a sample of comparable items representing each of the basic headings.
The World Bank PPP conversion factors are result of two separate PPP programs. The first is the ICP 2005 round which is organized in five geographic areas: Africa, Asia-Pacific, Commonwealth of Independent States, Latin America, and Western Asia. Regional agencies took the lead in coordinating the work in the five regions. In parallel, the Statistical Office of the European Communities (Eurostat) and the Organisation for Economic Co-operation and Development (OECD) conducted its 2005 PPP program that included 46 countries. Eurostat covered 37 countries - the 25 EU member states, the EFTA countries (Iceland, Norway and Switzerland), and Bulgaria, Romania, Turkey, Croatia, Macedonia, Albania, Serbia, Montenegro and Bosnia-Herzegovina. The OECD part of the program included nine other countries—Australia, Canada, Israel, Japan, Republic of Korea, Mexico, New Zealand, Russia, and United States.
Data on basic headings and prices were collected within each country and aggregated at the regional level. Three broad aggregation processes required to compute regional PPPs are: (1) Averaging the individual price observations to form an annual national average price for each product in each country; (2) Averaging individual price ratios for products to obtain PPPs at the basic heading level between countries within a region; and (3) Averaging basic heading-PPPs to obtain aggregated PPPs for GDP and its major aggregates between countries within a region. For some of the stages, different techniques are used across regions. The key difference arises in compiling PPPs at the basic heading level, specifically: (a) the CPD (country-product-dummy) approach is used in Latin America, Africa, Asia, and Western Asia; and (b) the EKS (Elteto, Koves, and Szulc) approach is used in OECD/Eurostat and the Commonwealth of Independent States (CIS). For aggregation above the basic heading level, the EKS method is used in all regions other than in Africa, where the Iklé method was used.
Regional outputs must themselves be systematically matched with each other to produce global results. Two alternative methods are considered for linking regions in the ICP. One is to nominate “core” or “bridge” countries, which participate in two regional comparisons, providing a link between that pair of regions. The other is to select group of countries, a few from each region, which participate in a separate and parallel worldwide comparison organized specifically to provide a link between regions. With one exception, the second method was adopted for the 2005 ICP. This has become known as the “ring comparison” because of the way in which it works to determine the relativities between different global regions. The ring comparison is a much-reduced global ICP comparison in which transitive PPP relationships, at the basic heading level and above, are established between regions rather than between individual countries within a region. In total, special price collections are compiled for 18 ring countries, using a global product list to enable these relativities to be calculated. The one exception is the CIS, whose regional results are linked to the OECD/Eurostat region using the Russian Federation as a bridge country.
Additional information on the methodology and process of computation can be obtained from the Global Purchasing Power Parities and Real Expenditures: 2005 International Comparison Program, available at http://www.worldbank.org/data/icp.
For the purpose of comparing levels of poverty across countries, the World Bank uses estimates of consumption converted to US dollars using purchasing power parity (PPP) rates rather than exchange rates. PPP conversion allows national accounts aggregates in national currencies to be compared on the basis of their purchasing powers of the currencies in their respective domestic markets free from differences in price levels across countries, much the same way as constant price estimates do in a time series comparison of real values free from differences in prices over time.
Purchasing power parities are statistical estimates. Like all statistics, they are point estimates that fall within some margin of error of the unknown, true values. The error margins surrounding the PPPs depend on the reliability of the expenditure weights and the price data and how well the goods and services that are priced represent the consumption pattern and price levels of each participating economy. As with national accounts data generally, it is not possible to calculate precise error margins for PPPs or the real expenditure data derived from them.
The ICP includes economies ranging from city-states to large and diverse countries such as China, India, Brazil, and Russia, which collectively account for over 40 percent of the world’s population and include many people living in remote, rural locations. These and similar countries have to produce national average prices for goods and services that are comparable with other economies in their region. The accuracy of the PPPs for these countries depends upon the extent to which the selected goods and services are representative of their entire economy and on their ability to provide nationally representative average prices. The need to measure prices for internationally comparable goods and services means that they are more likely to reflect consumption patterns of urban areas. If this “urban bias” is consistent between countries, and if the urban to rural price differentials are similar across countries, any bias will tend to cancel out in the estimation of PPPs; if not, results for some countries may be biased, up or down, depending on the extent of over or under representation or urban and rural areas.
PPPs should not be used as indicators of the under- or overvaluation of currencies, nor should they be interpreted as equilibrium exchange rates. The PPPs cover all of GDP valued at purchaser’s prices, which includes both traded and non-traded goods. Exchange rates, unlike PPPs, reflect the demand for currencies as a medium of exchange, speculative investments, or official reserves.
The 2005 PPPs are based on a new methodology designed to overcome problems encountered in previous rounds of the ICP. Therefore, users should be cautious about making comparisons to previous estimates of PPP-based GDP and components.
Time series of PPP conversion factors are obtained through extrapolating benchmark year data to other years. It has been established that the values extrapolated from one base year to another can be very different from those obtained from the base year survey and that the difference is larger for countries that have experienced larger shifts in relative prices from one benchmark to another. Therefore, as far as possible, extrapolations beyond benchmark years should be done sparingly and with caution.
For additional information on the data issues and limitations, see Global Purchasing Power Parities and Real Expenditures: 2005 International Comparison Program, available at http://www.worldbank.org/data/icp.
The PPP data used by The World Bank for comparing poverty come from three sources:
- Latest round of benchmark surveys;
- Regression estimates for non-benchmark countries and
- Extrapolations of benchmark year data to household survey years.
The latest data for 146 benchmark countries are estimated from the ICP 2005 round and OECD-Eurostat 2005 round surveys. Data for 41 countries are estimated from regressions.
The GDP expenditure breakdown into 155 basic headings and its various sub-aggregates are provided by countries. For countries where data for some of the required basic heading levels are not available, estimates are made by the local national accounts experts using alternate data sources. In some countries where expenditure-based GDP are not available, the percentage distribution of an economically comparable country from the same region is applied to production-based GDP aggregates to impute the basic heading details.
For the 2005 ICP, a new approach has been adopted for collection of national level prices for basic heading items. First, different “product clusters,” groups of products with shared broad characteristics such as “fresh whole milk” are defined, from which more specific products are specified. The classification of individual consumption by purpose (COICOP) is used as the starting point because most countries use it, or a compatible variant of it, in compiling their national accounts or their consumer price index (CPI). It is mapped to the OECD/Eurostat PPP classification to identify the products making up each ICP basic heading. The broad characteristics of each group of products are identified using the US Bureau of Labor Statistics “checklist” from its CPI as a starting point to develop a series of “Structured Product Descriptions” (SPDs), which show the key characteristics of each product to be priced (for example, type, variety, seasonal availability, quantity, packaging and pricing basis). An SPD identifies those characteristics that are price determining.
Each region goes through an independent exercise to define the specifications of the products to be priced. Once an SPD is set up for the product cluster, individual products are identified by selecting the specific characteristics for each product to be included (at least potentially) in the pricing list. Each selected product is given a detailed product specification. The number of products specified under each basic heading vary significantly from one basic heading to another and even for the same basic heading in different regions, because of the number and diversity of price-determining characteristics. Developing the final product specifications is a lengthy, iterative process. Chapter 5 of the ICP 2003-2006 Handbook describes in detail the preparation of the SPDs and the steps taken to derive product specifications within a region.
A basic concept guiding the price collection is that the prices should be consistent with those underlying each country’s national accounts expenditure estimates. The prices collected should include all non-deductible taxes, and include tips or gratuities where significant. And the prices recorded must be annual national average prices consistent with the values recorded in the national accounts.
National annual average prices for all collected items along with GDP expenditures for the basic headings are sent to regional coordinators for processing. The regional coordinators run the data through an editing process to establish consistency. They then compute PPPs and PPP- measured values for each of the basic headings and various sub-aggregates leading up to GDP measured in PPP.