Why It's Too Soon To Monkey With The CpiRobert Kuttner
Reading press accounts, you might think there was universal agreement that the government's consumer price index overstates inflation. This impression, however, is less the result of expert consensus than political opportunism by would-be budget cutters: If the CPI can be lowered by one percentage point, the deficit will be reduced by $634 billion over the next 10 years--a third of all savings needed to balance the budget. That's because the CPI governs annual cost-of-living adjustments in Social Security checks and other government benefits. What better way to cut benefits than through a seemingly technical adjustment? As an added bonus, a deflated CPI would trump the worry that wages have been flat. Cut the CPI, and presto--real wages are rising.
On June 26, the Senate Finance Committee appointed an expert panel on the CPI chaired by Michael J. Boskin, chief economic adviser in the Bush Administration. All five members had previously testified that the CPI overstates inflation. Not surprisingly, on Sept. 15, the panel's interim report concluded the index overstates inflation by one percentage point. But does it?
Janet L. Norwood, the respected former Commissioner of Labor Statistics, thinks not. Dean Baker, chief macroeconomist at the Economic Policy Institute (EPI), a Washington think tank, makes a convincing case that the CPI could understate inflation. Massachusetts Institute of Technology economist Zvi Grilliches, president of the American Economic Assn. and a Boskin panel member, thinks the whole debate is on shaky empirical ground.
STEAK VS. CHICKEN. Supposedly, the CPI overstates inflation for several reasons. One issue is quality improvements. Today's $3,000 computer is far superior to a 1990 $3,000 computer. In fact, the Bureau of Labor Statistics does adjust the CPI for quality changes in computers and a host of other products. Its new-car price index, adjusting for quality, shows a rise of only 150% since 1970--though the actual sticker price of a new car has nearly quadrupled. Doubtless, today's new cars are better--but four times better?
Second, since the CPI uses a fixed "basket" of goods, it misses what economists call substitution effects. If steak goes to $5 a pound, thrifty consumers switch to chicken. Critics say the CPI thus overstates what consumers actually spend. Yes, but then chicken isn't steak. If it's fair to adjust for quality gains, let's also count quality losses.
Third, the BLS waits years before including new products in the standard basket, thus missing some dramatic price reductions. The pocket calculator that first cost $1,000 now sells for $10. True, but as EPI's Baker observes, few people bought $1,000 calculators. Boskin et al further fault the CPI for underestimating the effect of retail discounting.
Baker concludes that the cumulative effects of all these factors may cause an overstatement of perhaps three- or four-tenths of a percentage point, not the full point projected by Boskin's panel. But, as Baker observes, that's not the end of the story. We also need to take into account the several areas in which the CPI likely understates inflation.
QUALITY OF LIFE. For example, the CPI omits most health-insurance costs, one of the most dramatic sources of consumer inflation. Here, too, quality changes cut both ways. Health technology is definitely more advanced. But with managed care, the consumer likely gets less face-time with the doctor, is often denied access to specialists, and increasingly gets the bum's rush out of the hospital. Similar quality deteriorations, uncounted by the CPI, are evident in cramped plane flights that now omit hot meals.
Nor does the CPI capture quality-of-life deteriorations that entail new costs. If you have to install alarm systems in your home and your car, that is a cost necessitated by deteriorating social conditions. Likewise, you may be spending more money on legal fees because the U.S. has become more litigious. The price of the "service" is beside the point. "I am worse off in direct proportion to the additional amount of money that I have to spend on my lawyer, regardless of the quantity or quality of legal services provided," Baker says. To demonstrate the absurdity of Boskin's proposed revisions, Baker applied them historically. The result puts the average worker in 1960 below that year's poverty line, not a believable finding.
These issues are subtle, intriguing, and worthy of dispassionate study. Accurate economic data are simply too important for a cheap political fix. The Boskin panel's final report, due next June 15, should resist the temptation to deliver technical cover for slashing benefits and instead tender a genuine piece of research.