The starting point is the Boskin Commission in the US. This group of dignified economists recently published a report commissioned by the Senate's Finance Committee which claimed that consumer price inflation is overstated by 1.1 per cent a year.
There were three main components of this upward bias the commission discovered in measured inflation. One, the consumer price index does not take account of the fact that when the price of one good goes up people will switch to cheaper goods. If fish becomes dearer, people eat more chicken. Unlike Britain's retail price index, whose components are updated once a year to take account of this substitution and changing tastes, the US CPI does not. The Boskin Commission estimated that this adds 0.4 per cent a year to the measured price level.
Secondly, far more shopping is done at discount stores, whereas the official price index takes no account of discounts. This adds about 0.1 per cent a year.
Thirdly, as the quality of products improves, a given price pays for more economic well-being. Although the price of a car has gone up, so has the amount of "car services" the consumer gains - better fuel economy, faster speeds, electric windows and so on.
New products also improve consumer living standards and price indexes typically miss the early falls in price and improvements in standards linked to items like home computers or mobile phones. Although the American statisticians do take account of quality changes, the commission said this effect added another 0.6 per cent a year to measured consumer price inflation.
Like a large stone lobbed into a millpond, the Boskin report has sent ripples of excitement through the world of economics. Last March the Bank of England published a working paper estimating that the "plausible range of bias" in the UK's RPI was 0.35-0.8 per cent. Following the Boskin report, both the Treasury and the Bank of England are revisiting the UK figures to see if the upward bias here might be greater than they had thought.
Why the excitement? Surely it is a good thing if inflation has actually been lower than we thought all along? The answer is that if measured inflation has these inherent flaws, having an inflation target might be the wrong monetary policy. This is not just a matter of setting the target at the wrong number - say 2.5 per cent when it ought to be 3.5 per cent. There is a conceptual problem, too.
This was eloquently expressed by Alan Greenspan, the Federal Reserve chairman, in a speech in early December*. It is worth quoting him at length. The problem ties in with the chairman's interest in the increasing weightlessness of the economy.
"One factor that will continue to complicate the task [monetary policy] is the increasing difficulty of pinning down the notion of what constitutes a stable general price level. When industrial product was the centrepiece of the economy during the first two-thirds of this century, our overall price indexes served us well. Pricing a pound of electrolytic copper presented few definitional problems. The price of a ton of cold rolled steel, or a linear yard of cotton broad woven fabrics, could be reasonably compared over a period of years.
"But as the century draws to a close, the simple notion of price has turned decidedly ambiguous. What is the price of a unit of software or a legal opinion? How does one evaluate the price change of a cataract operation over a 10-year period when the nature of the procedure and its impact on the patient changes so radically?"
One could take this line of argument further. At the moment computer software companies charge a price per software package. But most experts think this cannot be sustained when software can be so freely copied around the Internet. A charge per use is expected to become the pricing model. In theory, this would present no problems to the statisticians. But it is hard to know what the price of a programme is now when some people pay $50 for a package of which they might use a tiny part and some nothing for repeated use of more computer intelligence.
Mr Greenspan went on to argue that a general sense of the purchasing power of money over time remains, so measurement procedures will probably improve. The difficulty is what to make of the CPI or RPI at a time when the range of goods and services available increases dramatically or there are sharp improvements in quality.
Before getting too carried away with this, it is worth noting that the Boskin Report has some vociferous critics. In an article in the current issue of the New York Review of Books, Jeff Madrick points out that much of the report's evidence is anecdotal rather than analytical. The figures are based on conjecture - he quotes Zvi Griliches, a Harvard economist and member of the commission, describing them as "squishy".
In addition, he notes, the Bureau of Labour Statistics, which compiles the CPI, already makes quite a big adjustment for quality change. Without its quality adjustments, reported inflation in 1995 would have been 4.7 per cent rather than the actual 2.5 per cent. It might be considered stretching things to make another 0.6 per cent adjustment down.
The goods for which the quality issue is most relevant - new electronics products - account for only about 2 per cent of consumer purchases. Besides, some quality changes work in the opposite direction. Some consumer goods, such as cheap household electrical items like toasters and kettles, have become shoddier and more disposable.
Even so, Mr Greenspan's point highlights a real dilemma for policymakers. Some officials believe an inflation target is on its way to becoming unworkable - just as Gordon Brown has embraced it wholeheartedly.
* Available at http:/www.bog.frb.fed.us