A price index is a statistical measure of the average change in the prices of a defined set of goods or services over time, expressed relative to a base period that is usually set equal to 100. National statistical offices, central banks and international organisations compile price indices to measure inflation, to adjust wages, pensions and contracts, and to separate changes in value into changes in price and changes in volume.
The best-known example is the Consumer Price Index (CPI), but the same methods are used for producer prices, trade prices, house prices and national accounts deflators. International manuals on consumer and producer price indices, produced jointly by bodies including the IMF, the International Labour Organization, the OECD, Eurostat and the World Bank, set out recommended practice.
Main types of price index
Price indices differ by the transactions they cover and the prices they record:
- Consumer price indices track the prices households pay for a basket of goods and services, including taxes.
- The Producer Price Index (PPI) measures prices received by domestic producers or paid for their inputs, usually excluding taxes on products.
- Import and export price indices measure prices in international trade and feed into the terms of trade.
- A house price index tracks the price of dwellings, which requires special methods because no two houses are identical.
- Implicit deflators, such as the GDP deflator, are derived by dividing a nominal aggregate by its volume estimate.
Weights and formulas
A price index averages many individual price changes, so it needs weights showing how important each item is, usually its share of expenditure. Weights for a CPI come mainly from household budget surveys and national accounts data. The three classic formulas differ in whose quantities are used as weights. The Laspeyres index uses base-period quantities:
The Paasche index uses current-period quantities, and the Fisher index is the geometric mean of the two:
- and are the prices of each item in the base period and the current period.
- and are the quantities bought in the base period and the current period.
- means the values are added up across all items in the basket.
As a worked example, take a basket of two goods. In the base period apples cost 2.00 and 10 are bought, and bread costs 3.00 with 5 bought. In the current period apples cost 2.50 with 8 bought, and bread costs 3.30 with 6 bought. The Laspeyres index is . The Paasche index is . The Fisher index is . Laspeyres is highest because it ignores the shift away from apples, whose price rose most. More detail on these and related formulas is in price index formulas and the entry on the Laspeyres price index.
Chain-linking and quality change
A fixed basket becomes outdated as spending patterns change and new products appear. Most statistical offices therefore update weights regularly, often every year, and link the successive short-period indices together into a chain index; the method is explained in base periods, chaining and volume measures.
A second challenge is that products change. If a new phone model costs more but performs better, part of the price rise reflects improved quality rather than inflation. Statisticians use methods such as overlap pricing, option costs and hedonic adjustment to remove the quality element, as described in quality change in price indices.
Uses of price indices
Price indices are central to economic policy and to everyday contracts. Central banks set inflation targets in terms of a consumer price index. Governments index tax thresholds, benefits and pensions to it, and some bonds pay coupons linked to it. In national accounts, indices are used as deflators to turn nominal values into volume measures, the basis of real GDP and the distinction between nominal and real values.
Limitations and bias
No price index is perfect. A fixed-weight index can overstate the rise in living costs because it ignores substitution towards goods that have become relatively cheaper. Other sources of bias include new goods entering the basket late, shifts to cheaper outlets and imperfect quality adjustment. The choice of formula at the lowest level of aggregation, where prices are combined without weights, can also matter. These issues and the axiomatic tests used to judge formulas are covered in index number bias and tests. Treatment of owner-occupied housing is a further point of difference between indices and between countries.
External links
Ref 1100001827