DEBATE HEATS UP ON ACCURACY OF CONSUMER PRICE INDEX.

Byline: Jeff Gelles The Philadelphia Inquirer 

If the American economy were a living, breathing organism, Sabina Bloom would be like something out of ``Fantastic Voyage'': a tiny scientist trying to help comprehend the beast by examining it with precision instruments.

Day after day, for 25 to 30 hours a week, Bloom travels around the Philadelphia area to check prices for the federal Bureau of Labor Statistics. Her information helps build the bureau's monthly Consumer Price Index for All Urban Consumers, one of the most-watched numbers anywhere.

Lucky Charms cereal. Suzuki motorbikes. Pup-Peroni dog snacks. Outpatient knee surgery. Bloom tracks them all and many, many more, as she collects data for the bureau's massive Point of Purchase Survey, which tracks about 95,000 items at 22,000 retail outlets around the country.

The numbers she writes down - and how federal economists crunch them - are at the heart of a debate over the accuracy of the price index. The debate intensified last month when a Senate-appointed commission of five economists estimated that the CPI was overstating inflation by more than 1 percentage point a year.

If ever there were a statistic that mattered in real life, the CPI is it.

Social Security payments and federal tax-bracket cutoff points have for years risen with the CPI to insulate taxpayers and the elderly from inflation. In the private sector, the index is a benchmark in wage and salary negotiations.

There's an element of folksy familiarity as Bloom, 38, makes her rounds. She is called ``the wing lady'' at one roadside eatery where she checks the tab for three dozen chicken wings, a house specialty. She jokes with an employee at a motorcycle dealership as she cajoles him into checking prices.

This is not just shopping. The prices she tracks are chosen through a carefully randomized procedure meant to ensure that the index is statistically reliable, a procedure that soon comes into play when she prices a woman's jogging suit at a discount store.

The $29.88 suit she priced on her last visit is gone. In its place is another, not quite the same, selling for $29.99. Here is an example of the central puzzle of the consumer-price-index debate. It boils down to this: Is the 11-cent price increase really an 11-cent price increase?

The answer is: Not necessarily.

Questions have long been raised about the CPI, many of them by the Bureau of Labor Statistics itself. The biggest problems, just about everyone agrees, are built into a central concept of the index: that it measures price changes in a fixed market basket of goods, something arguably quite different from measuring the true cost of living.

Think of it this way: Nobody buys all the same things at the same places year in and year out. And when they do buy the same thing, the new product may be different - subtly, like Bloom's jogging suit, or vastly - from the old one.

The BLS says it carefully examines such changes and may not count attendant price shifts in the CPI. But the Senate commission, headed by Stanford University's Michael J. Boskin, concluded that the bureau was not handling those problems adequately.

A true cost-of-living index, the commission's report said, should adjust for the changes in what people buy and where. It should recognize that increased quality, new products and more choices add value to what consumers get for their money.

Taken together, the report said, the problems cause CPI increases to be overstated year after year.

How much? The Boskin commission's estimate is an upward bias of 1.1 percentage points a year.

There is no questioning the allure of the conclusions.

If the CPI has been overstating price increases by 1 percentage point a year, a correction could save the U.S. Treasury a trillion dollars by 2008, the report said. In fiscal 2006, the deficit would drop by $134.9 billion, thanks to a $90.5 billion cut in Social Security and other payments and $44.5 billion in extra tax revenue.

Boskin has called making adjustments to the CPI ``a no-brainer.'' Zvi Griliches, a Harvard economist who also served on the commission, is a bit more reserved.

``These numbers are not cast in stone. They are our best estimate. As we said, there is some danger of uncertainty,'' Griliches said.

Griliches rejects as misguided the concerns raised by some economists that a downward adjustment could harm people receiving Social Security or other government payments. ``A lot of the people who are getting these things are in the middle or upper end of the income distribution, and a lot of the people who are paying for it are in the lower middle classes,'' he said. ``The money that is being given is being taken from somebody else. The question is: Is this right?''

Griliches also rejected the notion that quality changes tend to help the affluent more than the poor.

``Tires that don't blow out so often or beepers that the elderly can use for emergencies are not exactly playthings of the rich,'' he said.

CAPTION(S):

Photo

Photo: Sabina Bloom checks on motorcycle prices in a survey for the CPI.

Knight-Ridder Tribune Photo Service