DEBATING THE CPI: HOW REAL IS IT?
Byline: Joanna Ramey
WASHINGTON — Federal Reserve Chairman Alan Greenspan may never meet Cecilia Rohan. But when the nation’s top financial dog questions the validity of the Consumer Price Index — the government’s key measure of inflation — he’s questioning, at least in part, her decision that a long-sleeve 100 percent polyester dress is comparable to a short-sleeve dress of 70 percent triacetate and 30 percent rayon.
Rohan is one of 350 CPI data collectors for the Bureau of Labor Statistics who fan out to roughly 21,000 stores a month to take stock of price changes.
Following strict statistical orders, these inflation gumshoes also take note of year-to-year changes in quality and technology, the value of which Greenspan said is so difficult to judge that it leads to overstating the true rate of inflation by 0.5 percent to 1.5 percent.
The CPI measuring February’s price changes is due out today, coming as it always does on the heels of the Producer Price Index, which covers wholesale prices. The government has been tracking consumer prices since 1913, refining the process through the years.
Meanwhile, Greenspan is not alone in his doubts about the CPI.
“It’s in the eye of the beholder,” said Irwin Cohen, chairman, Deloit & Touche’s trade, retail and distribution group, of the CPI’s subjective nature of weighing how changes in a product’s composition and quality should be reflected in its price performance from month to month and year to year.
“There is probably some bias upward in the CPI because of the adjustment for quality. Quality is very judgmental,” said Ira Silver, chief economist, J.C. Penney Co. “I don’t think the situation is that clear. What makes for quality adjustments in apparel?”
Take the polyester dress, which Rohan tracked down recently at a national department store in a Washington suburb. (A confidentiality agreement between retailers and the BLS prevents disclosure of the store’s name.)
Rohan was charged with finding an unlined, long-sleeved, 100 percent polyester, dry-clean-only Jones New York dress like the one she priced at $180 in the same store during March 1994. She knew she wouldn’t likely find the exact dress. The object was to find one with the same qualities. If a match couldn’t be found, BLS statisticians gave her some leeway, allowing her to make substitutions only on brand name, man-made fiber and sleeves. No sample would be taken if, even with the substitutions, a match couldn’t be found.
With a blue notebook in the crook of her left arm, Rohan began combing the better-dress department, feeling material, checking labels. She pulled a $164 David Warren dress of 100 percent polyester off a rack, then quickly put it back because it was lined. A $295 Ellen Tracy red dress met all the criteria, save for its 100 percent wool content.
After another pass around the carousels, Rohan settled on the closest match: an unlined black Jones New York dress with half sleeves, 70 percent triacetate and 30 percent polyester for $150.
Rohan noted on her data sheet these details, along with some extra value-added characteristics like the tailored collar, button placket and belt, which in the final number-crunching might influence the dress’s price movement. For example, the triacetate — which the BLS has deemed, in the consumers’ mind, to be a lesser quality than polyester — might register a price dip, which could be offset by the price-enhancing tailored detailing.
But this was just one of 480 dresses nationwide the BLS would be pricing for the March index, to be released in April and whose individual traits and prices will be assigned a statistical weight and fed into a computerized pricing model. In pricing apparel, information will be culled from 2,073 sales points — stores and catalogs — over the course of the month, 22.5 percent of which are discount stores, 61 percent full-priced department and specialty stores and 16.5 percent mail order catalogs.
“We think the price of a dress is dependent on all kinds of things,” says Charles Fortuna, a BLS apparel analyst, ticking off a litany of factors plugged into a complex model of variables that judge how much prices swing. “It’s dependent on the style, the fibers, whether the dress is a one-piece, a two-piece, whether it comes with a belt or not, whether it’s a solid color or multicolor, whether it comes from a discount store or full-service department store, or from a store in a small town or a big town.”
The BLS has five main categories of women’s apparel it tracks each month: dresses; coats and jackets; separates and sportswear; suits, and underwear, nightwear, hosiery and accessories. The agency decides how much weight to give each item, according to statistical models based on consumer buying data provided by the Census Bureau. The BLS also plugs into the formula Census Bureau surveys reflecting changes in where people shop, such as shift in recent years to discount and off-price stores.
Moreover, in analyzing month-to-month changes, the statistical model makes complex seasonal adjustments in an effort to wash out of the comparison seasonal factors that influence pricing.
To further insure the index reflects what’s happening in the marketplace, the agency quizzes individual stores as to what items are hot and then focuses on certain ones to price.
The current debate over how well the CPI reflects inflation has gotten widespread attention because of its use as a benchmark for cost-of-living increases in federal programs, like Social Security.
Greenspan has argued the government could save $150 billion during the next five years by revising the index downward as a whole.
But Greenspan — who has primarily poked holes in the CPI’s ability to capture price swings in real estate and rapidly evolving products like electronics — has stopped short of advocating a wholesale revamping of the index. He has asked for a commission to be set up that would determine how the CPI findings should be translated into cost-of-living adjustments.
While at the same time agreeing with Greenspan that the CPI might be a bit out of kilter, economists largely applaud the BLS’s diligence and ability to dog prices in the marketplace. They also note it would be no easy task to make deep changes in the CPI’s complicated statistical protocol and sheer volume of products involved. Besides, if the index has an upward bias, it is consistently so, which still makes it a statistically valid measure, they add.
“If the CPI has some kind of bias, I don’t think the bias has changed that much. That is the critical thing,” Silver said. “I think they do a very good job, given that prices and merchandise change that much. I think it broadly reflects what happens with prices.”
“In any kind of large survey, there is bound to be flaws,” said Carl Steidtmann, director of research, Management Horizons, a division of Price Waterhouse. “No index is perfect. It’s as good as anything around.”
In addition to looking to the index as a measure for wage and rent increases, retailers commonly use the CPI to weigh their stores’ price performance against a national standard. Many retailers also use a BLS index derived from the CPI for calculating taxes based on inventories, if they don’t opt to use their own in-house price surveys.
If there are changes to be made to the CPI, Cohen, of Deloit & Touche, said it would help if the index were calculated according to average monthly price swings, instead of the incremental month-to-month and year-over-year method now employed. He said this would help retailers better forecast their taxes over the year, based on the Internal Revenue Service’s Last-In-First-Out or First-In-First-Out model.
Currently, Cohen said stores must estimate taxes during the year, “then they’ll hold their breath” until the yearend CPI is released in January before determining their final tax bill.
“What I want to do is take out that surprise,” Cohen said. “January would only be one-twelfth of the computation.”
Silver said it’s important that the BLS remain open to market changes and constantly hone its pricing models. He wonders if statisticians have adequately accounted for retailers’ weekend promotional pricing and how they are accounting for the quality in apparel made from the new wrinkle-free fabrics.
Like other changes in the marketplace, the impact of wrinkle-free fabric on apparel prices is being studied and eventually incorporated into the CPI, said the BLS’s Fortuna.
“We don’t have a good estimate of value across the board right now,” he said. “But we believe we’ll have a good estimate for some things, like men’s pants and shirts, this spring.” — Fairchild News Service