Base periods, chaining and volume measures are a family of techniques that official statisticians use to turn raw values into comparable series over time. A price index or a quantity index only makes sense when it is tied to a starting point, built with sensible weights and kept up to date as spending patterns change. The same tools allow national accountants to split a change in the value of output into a part caused by prices and a part caused by real activity.
The members of the family fit together in sequence. A base period and an index reference period fix the anchor. Chain linking joins short-period indices into a long series with fresh weights. Volume indices measure quantity change, and double deflation applies the idea to value added. These methods sit behind real GDP, consumer prices and industry output in every country that follows the System of National Accounts (SNA).
Base Period
The base period of an index is the period against which other periods are compared. In everyday use it is the year or month whose value is set equal to 100, so that a reading of 112 means a rise of 12 per cent since that time. Index manuals separate three ideas that the phrase can blur. The weight reference period supplies the quantities or expenditure shares used as weights. The price reference period supplies the prices in the denominator of each price relative. The index reference period is the one set to 100. In a simple fixed-basket index all three coincide, but in practice they often differ: a consumer index may use weights from a household survey two years old while being expressed relative to a more recent year.
Statistical agencies move the base from time to time, either to refresh the weights or simply to show recent movements more clearly. Changing weights alters the index; merely changing the reference period does not change any growth rate.
Index Reference Period
The index reference period is the period for which the index is given the value 100. Choosing it is a matter of presentation. A series can be re-referenced, often called rebasing, by dividing every value by the value in the new reference period:
Here is the old index in period and is the old index in the new reference period . Suppose an index stands at 125 in 2022 and 135 in 2024 on an old reference of 2015 = 100. Re-referenced to 2022 = 100, the 2024 value becomes . The rise between the two years is 8 per cent either way, which shows that re-referencing changes the level shown but not the movement. It should not be confused with the weight reference period, which actually affects the measured growth.
Chain-Linked Index
A chain-linked index is built by calculating an index for each pair of adjacent periods, using weights relevant to those periods, and then multiplying the links together. The method is called chain linking. Over periods 0 to the chained index is:
If volumes grow by 3 per cent in the first year and 2 per cent in the second, the chained index is , or 105.06 on the starting reference. Because the weights are refreshed each year, chained measures reduce the substitution bias discussed under index number bias and tests. National accounts in the United Kingdom and Europe publish annually chained volume measures (CVM) of GDP and its parts. The cost is non-additivity: chained components such as consumption and investment do not add up exactly to chained GDP except in the reference year, which is why tables expressed in chained currency units carry a residual line. The Chained Dollars section of nominal and real values covers the US presentation of the same idea.
Volume Index
A volume index measures the average change in the quantities of a set of goods or services between two periods, with each quantity weighted by a price. The Laspeyres form uses base-period prices and the Paasche form uses current-period prices:
Here is price, is quantity, and subscripts 0 and mark the base and current periods. Take two goods: good A goes from 10 units at 2 to 12 units at 3, and good B from 4 units at 5 to 5 units at 5. Then and . A volume index pairs with a price index so that the value ratio is split into the two: a Laspeyres volume index times a Paasche price index gives the value change exactly. The Laspeyres price index page shows the price side.
Double Deflation
Double deflation is the preferred way to estimate real value added for an industry. Instead of deflating value added directly, statisticians deflate gross output with an output price index and intermediate consumption with an input price index, and take the difference:
and are the output and input price indices, written as ratios to the base year. If output is 1,000 at current prices with , and intermediate consumption is 600 with , real output is 800, real inputs are 400 and real value added is 400. Single deflation, dividing nominal value added of 400 by the output price index, would give 320. The gap shows why double deflation matters when input and output prices move differently. Its weakness is volatility, since real value added is a small difference between two large numbers. Industry estimates of gross value added (GVA) rely on it wherever good input price data exist, and the implied GDP deflator follows from the result.
How the measures fit together
| Concept | Question it answers | Typical use |
|---|---|---|
| Base period | Compared with when? | Weights and price comparisons |
| Index reference period | Which period equals 100? | Presentation of published series |
| Chain-linked index | How are years joined with fresh weights? | GDP, consumer price index (CPI) |
| Volume index | How much did quantities change? | Real output and spending |
| Double deflation | What is real value added? | Industry output in the national accounts |
The formulas used inside each link are covered in price index formulas, and the treatment of new and improved products is covered in quality change in price indices.
Ref 1100001762