OVER the past two decades, central bankers in many countries have earned praise for taming inflation. In America, the rate of increase of the consumer-price index (CPI) has fallen steadily, save for an upturn during the first Gulf war. Granted, the prices of assets—first shares, before they crashed; now, residential property—have risen dramatically. But consumer-price inflation has clearly been brought under control. This has generally been good news for consumers. But has everyone benefited to the same extent?

Lately, there have been echoes of the days when inflation roared. Petrol now costs well over $2 a gallon in America and has touched £1 a litre (around $7 per American gallon) in Britain, where fuel taxes are much higher. Although inflation generally remains low, higher petrol prices cause more pain to some people than to others. If you drive a gas-guzzling SUV, the chances are that your personal cost of living is rising much faster than the official price index. If you walk or cycle to work, rising petrol prices make little difference to you: your personal inflation rate, other things being equal, is less than the official one.

Roughly speaking, official price indices measure the price of the goods and services consumed by the “average” household. Official inflation rates are often used to determine annual increases in social-welfare benefits and old-age pensions, or as reference points for wage bargaining. But because everyone's spending differs from the average in some way, so does everyone's personal inflation rate. Do these variations matter?

In a recent paper*, Bart Hobijn and David Lagakos, economists at the Federal Reserve Bank of New York, try to answer this question. The authors delve into the details of America's consumer-price index (CPI) from 1987 to 2001 and examine the extent of “inflation inequality” among different social groups. Even though inflation has been low, the prices of education and health care, for example, have risen much faster than those of manufactured goods, such as cars and clothing. In part, this is because labour-intensive industries, such as schools and hospitals, do not make productivity gains as easily as manufacturing. Service industries have also benefited less from the price-lowering, competitive effects of globalisation and internet shopping.

Has that made for wide inflation gaps among Americans? By analysing various demographic slices, the authors find that differences between urban and rural consumers, or between black and white Americans, are much narrower than they seem to have been in earlier decades. But other gaps persist. Families with young children, according to the study, have faced much lower inflation than other households, thanks mostly to lower health-care bills and to not having to pay university fees. Inflation in education averaged 6.6% during the period from 1985 to 2000, while inflation in health care ran at 5.6%.

During the period studied, inflation rates were most widely distributed in 1991, with a standard deviation of 1.8 percentage points—implying that roughly one-third of Americans had personal inflation rates at least this far away from the CPI measure. By 1996, when inflation was lower, the standard deviation had fallen to only 0.2 points.

The most important question is whether the old face a different inflation rate from the young, because America's state pensions are indexed to the overall CPI. The authors say that, over the period studied, old people faced a CPI inflation rate that was around 0.2 to 0.4 percentage points per year higher than the average, which implies that the real value of their pensions has dwindled over time. As you would expect, this is mostly attributable to higher health-care inflation, and old people's heavier use of it. They spend, on average, twice as much of their income on health care as younger people do. This result seems specific to America. A recent study of British spending patterns found that the old faced lower inflation than the young, thanks in part to the provision of tax-financed health care by the state.

It is inevitable that any index will fit the spending patterns of some people better than others. One way of dealing with the differences might be to use different indices for different demographic groups. But there are sure to be differences within groups as well as between them. In any case, Messrs Hobijn and Lagakos find that most variations iron out over time. Households that face higher-than-average inflation rates one year tend to have lower-than-average rates the next year, and vice versa.

In the same way, booming house prices hurt first-time buyers much more than those already on the property ladder. More troubling is the thought that official consumer-price indices might overstate or understate inflation across the board. One line of economic research over the past decade has focused on whether official indices overstate inflation. The consensus is that they do, by failing to adjust sufficiently for improvements in the quality of goods and services, or the availability of newer, more innovative ones. In other words, because a $500 television today is much better than a $500 set of a few years ago it is, in effect, cheaper. This means that inflation is actually lower, for all consumers, than a simple price index would suggest.

America has made more progress than most other countries in adjusting its main index, the CPI, for such changes. The effect of this is likely to be to save the government money, by reducing the rate at which pensions and benefits increase. Even so, a better, quality-adjusted CPI will still reflect only the average inflation faced by consumers. Pensioners could still be suffering a slightly higher inflation rate than younger Americans.

* “Inflation Inequality in the United States”. Federal Reserve Bank of New York, Staff Report no. 173, October 2003

Copyright © 2004 The Economist Newspaper and The Economist Group. All rights reserved.