Correlating the Federal Reserve Board's latest pronouncement on inflation with the experience of filling up at the pump or stocking the family larder brings to mind Mark Twain's adage regarding the three varieties of lies: lies, damned lies and statistics.

"Readings on core inflation have improved modestly in recent months. However, a sustained moderation in inflation pressures has yet to be convincingly demonstrated," the Fed said in its Aug. 7 statement.

For those in need of a translation, the Fed says if you exclude energy and food, the cost of living has improved modestly. But, there's still a chance prices could spike.

Like many who are paying $2.75 for a gallon of gas and $3.50 for a gallon of milk, you may think excluding energy and food when measuring inflation is disingenuous. But some economists say it is a more accurate long-term measure of inflation because energy and food prices fluctuate more than the prices of other goods. Excluding them "provides a sharper signal of the underlying trend," said Global Insight economist Kenneth Beauchemin.

The debate over core inflation and headline inflation, a measure that includes food and energy prices, highlights the ongoing examination of how accurately the Bureau of Labor Statistics' consumer price index measures inflation. The federal agency has made a number of changes over the last decade in the way it calculates the CPI.

The changes in large part were driven by a 1996 report that concluded that the bureau was overstating inflation by more than 1 percent annually.

Economists generally agree the changes have improved the index's accuracy. But there is a vocal minority who persuasively argue the CPI understates inflation. The minority includes a few conspiracy theorists who say the government has a vested interest in understating inflation.

"When they keep inflation low, it makes economic growth look higher. It's government artifice at it's best," said Jayme Wiggins of Intrepid Capital Funds in Jacksonville, Fla.

Moreover, since Social Security recipients receive annual cost-of-living adjustments based on one measure of the CPI, lower inflation puts less of a burden on the federal budget. (The 1996 commission estimated that if the upward bias of the CPI were left uncorrected, it would have contributed about $148 billion to the deficit in 2006.)

Calculating the CPI starts with something called the consumer expenditure survey. Each month, the bureau collects price information for more than 200 items at thousands of urban locations across the country. Prices in rural areas aren't included in the survey. It uses the data to compile a variety of regional and national CPIs. The most widely quoted is the CPI-U, the national measure of inflation based on prices paid by urban consumers.

That figure, released monthly, by no means accurately measures inflation as experienced by individual consumers, particularly those who live in rural areas. If you walk to work, subsist on a food item not included in the survey or don't switch to cheaper alternatives when prices go up, the inflation you experience will be different from the CPI.

"Everybody's consumption patterns are going to be different than the average in the CPI," said University of Pittsburgh business school professor Josephine Olson.

Critics of the Bureau of Labor Statistics have more fundamental problems. One of their biggest is with the way the agency measures the cost of housing. For homeowners, the bureau uses something called the owners equivalent rent, basically what homeowners would pay if they rented their home. When home prices go up faster than rental rates, as they have in recent years, the CPI understates inflation for homeowners, critics contend.

They also have a problem with some of the adjustments the bureau made following the 1996 report, whose authors cited two big reasons why inflation was being overstated.

Previously, the Bureau of Labor Statistics assumed that if the price of beef went up, consumers would keep buying beef instead of switching to cheaper meats. Moreover, the agency wasn't taking into account quality improvements. If the price of a car went up because antilock brakes were included as standard equipment, the consumer would be paying more, but for a higher standard of living than was previously reflected in the CPI.

So, to reflect the expectation that consumers will respond rationally to price increases and to keep the CPI based on a consistent standard of living, the Bureau of Labor Statistics now adjusts prices downward. Critics have a problem with that.

In an October 2004 piece, PIMCO's Bill Gross estimated inflation would have been 0.5 percent to 1.1 percent higher each year since 1987 if the adjustments had not been made. He cited a Wall Street Journal article that stated that although personal computer prices had fallen 8 percent a year over the previous decade, the bureau estimated they had dropped by 25 percent a year since 1997, reflecting the fact that consumers were buying more powerful computers with more features.

"The CPI as calculated may not be a conspiracy, but it's definitely a con job foisted on an unwitting public," Mr. Gross told clients in October 2004.

There's no denying that measuring inflation is complex and difficult. It's commendable that the Bureau of Labor Statistics continues making adjustments to do a better job. But the agency's efforts will never persuade every consumer that inflation is as under control as their government says it is.