Evolving the measurement of inflation
Inflation statistics are prominent, with concerns about the scale of rising prices. How do stats offices calculate the inflation rate and what problems are there?
The basic approach
There is a ‘basket’ of around 700 goods and services. That basket covers essential items such as bread and milk. It also includes expensive purchases like cars.
Each month, local collectors go into about 150 places, gathering more than 100,000 prices. There are central collection processes for larger chain stores.
Analysts then study how over 180,000 prices change each month. Those analysts calculate a weighted average for the change. For these weights, there are expenditure estimates in national accounts and household surveys. What goes into that ‘basket’ differs over time, reflective of changing habits.
Those changes then link together in a chain across time, producing the index.
The conceptual problems of inflation
There are two different ways to think about price changes:
- Across the economy: How are prices changing in economic interactions throughout the society?
- Experienced by households: How does each household bear rises and falls in prices?
These different approaches can lead to distinct measures. Domestic coverage means including only spending within a country. National coverage of an inflation index focuses on expenditure by residents.
There are questions of weighting. Do we seek spending shares on items by the average household? Should weights correspond to money spent: each item in proportion to total spending? The latter, ‘plutocratic’ weights often prevail.
Time is another factor. The index may capture price changes at the moment of acquisition. Another method involves distributing costs over the use of an item or service. A third route captures payments for the product, which can take some time with credit cards.
One average does not fit everyone. Each person has their own individual inflation rate. People differ in what they buy, in both quantity and quality.