Return to the Table of Contents.
A P R I L 1 9 9 7
by Thomas I. Palley
"THE one duty we owe to history is to rewrite it," Oscar Wilde said. In line with that aphorism, the Advisory Commission to Study the Consumer Price Index, chaired by Michael Boskin, has rewritten the economic history of the past twenty-five years. The Boskin Commission was formed at the behest of the Senate Finance Committee. Last December it released its long-awaited report on the CPI. The commission reported its conclusion that the CPI overstates the rate of cost-of-living inflation by 0.8 to 1.6 percentage points a year; its best estimate was that inflation is overstated by 1.1 percentage points. If cost-of-living inflation has been overstated, then the growth of the economy and real wages has been much higher than previously reported. The commission has thus solved the problem of stagnating wages, which is now revealed to be a mere fiction. Far from experiencing a "silent depression," the commission implicitly claims, American families have never had it so good.
From the archives
The political implications of this rewriting of history are obviously
enormous. The commission's findings are being used as a cloak for an
economic agenda that will injure lower- and middle-income households. By
"cooking the books" the commission threatens to undermine the dawning
consciousness that the U.S. economy is no longer delivering prosperity for
huge segments of society. Clearly, its findings need to be substantively
examined before being accepted.
Unfortunately, the majority of the American economics profession appears to have uncritically accepted the commission's finding that inflation is overstated; this implicitly makes a mockery of the profession's own intellectual accomplishments. If inflation, wages, and income have all been misstated, years of research have been conducted using incorrect data. Thus much of this research, which purportedly confirmed the profession's theoretical claims, is no longer valid.
The Boskin Commission cited four reasons for the CPI's mismeasurement of inflation: product-substitution bias, outlet-substitution bias, quality-change bias, and new-product bias. The CPI measures the cost of a given basket of goods in which the quantity of each good is based on the amount that the average consumer was buying on the date the index was calibrated. Product-substitution bias emerges because over time consumers alter the goods they buy, and consequently the bundle of goods that the CPI is tracking no longer represents consumer spending patterns. The Boskin Commission claims that consumers have been substituting low-inflation goods and that the CPI has been oversampling high-inflation goods.
Outlet-substitution bias arises because consumers have been changing the places where they shop, shifting from high-priced department stores to lower-priced discount stores. But the CPI oversamples department-store purchases and undersamples discount store purchases.
Quality-change bias reflects the fact that products tend to improve in quality over time, and thus consumers are effectively getting more product for their money. The CPI not only may fail to capture this improvement in quality but also may register quality improvements as price increases, since improving quality sometimes adds to cost.
New-product bias arises because new products are only gradually incorporated into the CPI basket of goods. But the prices of new products tend to fall rapidly soon after the products are introduced, and thus the price decline is not captured by the CPI.
The cornerstone of the Boskin Commission's findings is sample-selection bias in the construction of the CPI -- that is, the CPI basket of goods oversamples high-inflation goods and undersamples low-inflation goods and goods whose prices are falling. The commission is itself a delicious example of such bias: All its members were on record prior to the establishment of the commission as believing the CPI to be overstated. At the same time, the commission took no evidence from such well-known economists as Janet Norwood, a former head of the Bureau of Labor Statistics, and Dean Baker, of the Economic Policy Institute, who believe that the CPI provides a reasonable reading of inflation. In effect, the commission took account of all the evidence of overstatement of inflation by the CPI and downplayed the evidence of potential understatement.
The Bureau of Labor Statistics, which is responsible for maintaining the CPI, has long been aware of the different biases identified by the commission, and it carefully and scrupulously adjusts the CPI to root them out. Whereas the Boskin Commission conducted no new research, the bureau conducts frontier research on the CPI. The government's statistical agencies have a history of producing very fine statistical work. Attacking their statistics damages their credibility and clouds the nation's political and economic understanding.
Contesting the technical details of the commission's arguments, Dean Baker has pointed out that the CPI could possibly understate inflation. For example, the CPI fails to take account of the greater contributions for health-insurance premiums and doctor co-payments that have increasingly been forced on American households. It fails to take account of increased traffic congestion, which raises the cost of commuting.
But quibbling over what is included or excluded misses the real point: the Boskin Commission has fundamentally misunderstood the economic information contained in the CPI.
Consumers do alter their buying habits as new products come into being and lifestyles change. However, the most noticeable feature of current trends is the systematic substitution of cheaper products. This substitution is an outcome of the now widely recognized squeeze on ordinary household incomes. For example, suppose your child has just been accepted by Yale University, but you find you are unable to afford the tuition; instead she must attend Erewhon State College. The commission would point to this as an example of product-substitution bias and claim that the CPI is overstating inflation.
Exactly the same argument applies to the issue of outlet-substitution bias. One can also argue that the inferior service customers receive at discount stores represents a loss of quality and thus that discount shopping reduces quality-change bias.
The commission's reading of new-product bias is also indicative of its misunderstanding of the use to which the CPI is put. It is actually a good thing that the CPI does not immediately include new products, such as hand calculators and computers. Such products are often extremely expensive when they first come out, and beyond the range of ordinary households. These products are widely consumed only after their prices fall. The CPI has been used as a measure of the cost of living for an ordinary household, and as such it should include new products only after they have been widely adopted.
Aside from rewriting economic history, restating the CPI to show lower cost-of-living inflation would have profound consequences for American families. Many employment contracts have cost-of-living adjustment clauses that link wages to the CPI inflation rate; a lower rate would lower wage increases. The CPI inflation rate also serves as a benchmark for setting wages for workers who do not have contracts containing automatic COLAs. A lower CPI rate would give employers a reason to lower wage increases for these workers too.
The economic harm to working people would not be restricted to wage cuts. Income-tax brackets and the earned-income tax credit given to low-wage workers are both indexed to the CPI. Lowering the CPI inflation rate would therefore affect income-tax exemptions and push many middle-class families into higher tax brackets. Adopting the Boskin Commission's findings would be tantamount to imposing a tax hike that would particularly affect lower- and middle-income families.
Social Security payments are also indexed to the CPI, so lowering the CPI inflation rate would slow the growth of Social Security income. Social Security is the principal source of income for 66 percent of the elderly and the only source of income for 14 percent. The elderly are large consumers of health care, and costs have gone up faster in this sector than in the rest of the economy. Moreover, the elderly tend not to be consumers of those new products -- computers, for example -- that have decreased in price and slowed the growth of the CPI. From the standpoint of the elderly, the CPI underweights health-related spending and overweights new products, thereby significantly understating the inflation they experience.
Given the questionable intellectual foundations of the Boskin Commission's findings, the commission's high standing in Washington requires explanation. Both Democrats and Republicans have been keen to see its recommendations adopted, because they provide a potentially uncontroversial way to achieve deficit reduction. Raising taxes is unpopular, and little discretionary government spending is left to be cut. Restating the CPI as a measure of cost-of-living inflation offers an easy way to lower Social Security payments through reduced COLAs and raise tax revenues through reduced exemptions. The hope is that the CPI can be presented as an apolitical and boring technical issue that voters won't notice.
Revising the CPI would get the Republicans off the hook of deficit reduction, while simultaneously advancing the interests of business. This, however, would occur at the expense of working Americans and the elderly. Revising the CPI would get the Democrats off the same hook, but at the cost of another shameful desertion of the constituencies they claim to represent.
Illustration by J. C. Suarès
Copyright © 1997 by The Atlantic Monthly Company. All rights reserved.
The Atlantic Monthly; April, 1997; How to Rewrite Economic History; Volume 279, No. 4; pages 20 - 22.