n 1997, a commission appointed by the Senate Finance Committee concluded that the annual Consumer Price Index computed by the Bureau of Labor Statistics was probably 1.1 percent too high, and perhaps even more.

The Boskin commission's findings, named after its chairman, Michael Boskin, a Stanford economist who served as President George Bush's chief economic adviser, made headlines in the financial news media, and many took its conclusions as gospel.

But perhaps they shouldn't have. At the time, many members of the Senate Finance Committee were cheered by the findings. To them, it meant the federal government could justify reducing Social Security benefits, because payments are tied to increases in the price index. So are countless wage contracts, as are the federal income-tax brackets.

If it was right, it also meant that both the economy and productivity were growing faster than reported, a boon to arguments on Wall Street and in Washington that the nation was embarked on a new economic revolution.

But a new, comprehensive study sponsored by the National Academy of Sciences, and partly commissioned by the Bureau of Labor Statistics itself, raises serious questions about just how much the Boskin commission or anyone else knows about the true rate of inflation.

The study, undertaken by a panel of 13 economists from a wide variety of institutions and led by Charles Schultze of the Brookings Institution, restores one's faith in cool-headed economic analysis and measured use of economic theory.

"It is the report that the Boskin commission should have done," says Joel Popkin, a Washington economic consultant.

The main contention of the Boskin report was that the Labor Department's statistical experts did not fully take into account the improved quality of many products. You may pay $1 for a razor, but if you get twice as many shaves as before, you are really getting more for your money. In effect, you are paying 50 cents for that razor.

The same is true with automobiles. If they are more durable or have automatic window locks, say, buyers are getting a little more for their money. The labor bureau basically agrees with the theory and makes many such adjustments. But Mr. Boskin's group and many other economists contend that the adjustments are not adequate and that reported inflation has been much too high.

The problem with that argument, says the Schultze panel, is that such estimates are difficult to make and easy to exaggerate. It points out many holes in the Boskin analysis. For one thing, in some cases, it is likely the Bureau of Labor Statistics over-adjusts for quality, meaning inflation is too low, not too high. For another, the Schultze panel cites evidence that the Boskin commission may have overstated quality in important areas.

For example, Jack Triplett, a former bureau economist now with Brookings, says the Boskin commission's analysis of quality increases for cars was simply misinformed. Two government economists, Brent Moulton and Karen Moses, seriously criticize the work on housing and medical care.

One of the striking oversights of the Boskin report, I believe, was that it paid almost no attention to areas in which product quality deteriorated, like airline service and education. And in the years since, consumer surveys consistently have found declining satisfaction for both goods and services.

This Christmas season brought the point home for me. My new printer broke down within a year, my laptop within 18 months. My new answering machine garbles messages.

One assumption of the Boskin commission was rejected outright by the Schultze panel. How do you take account of the benefits to consumer well-being of new products?

Some ingenious economists believe you can assume the product already exists but is sold at a price so high that no one will buy it. They argue that the difference between the imagined price and the actual price of the new product, once introduced, should be included as a reduction in the price index. But such an estimate is difficult to make credibly, and including benefits from new goods as a reduction of inflation, the panel concluded, is itself theoretically questionable.

The most important contribution of the panel, however, was to clarify the central assumptions about the uses of the C.P.I. and other price indexes.

The simplest way to measure prices is to compute the changes for a fixed basket of goods and services. This is what probably most Americans think the consumer price index does. But in the early 1960's, a government commission urged the Bureau of Labor Statistics to adopt a broader concept: the index should reflect the cost of maintaining a given "standard of living." Thus, if products improve, the standard of living improves, or it costs less to maintain the old standard of living.

Certainly quality improvements have to be taken into account. But so must many other factors that affect the standard of living, including the quality of public goods, like transportation and the environment.

The Schultze panel concludes that such an index is inherently too ambiguous to develop in an objective way. How does one account for traffic congestion, for example? A car is worth less if traffic deteriorates. If the crime rate rises, people may have to buy more security devices. Should that be translated into a fall in living standards?

The panel also cites evidence that consumers need constant improvements just to remain as satisfied as they were. How does one incorporate rising expectations into an index that measures the standard of living?

The Schultze panel therefore recommends a conditional cost of living index that is relatively unambiguous, including quality adjustments but deliberately leaving out issues like congestion, environmental degradation and consumer expectations.

The panel makes no precise estimate of whether the C.P.I. overstates or understands the true rate of inflation. But given the questions the panel raises about the commission's central assumptions and its analysis of specific product areas, the claim that the C.P.I. has been and may remain seriously overstated is not nearly as clear-cut as many economists have assumed. And reported inflation has already been markedly reduced through recent changes introduced by government statisticians themselves.

I hope this report is put between hard covers soon. It is my business book of the year.