A study soon to be published in the American Economic Review provides the first rigorous estimate of how much the often-criticized method of measuring the cost of housing has overstated the Consumer Price Index (CPI). It shows that the CPI has on average been as much as 25 percent higher than it would have been if housing costs were correctly measured.
The overstatement has occurred because the index includes changes in the expected costs of owning a home-- current interest rates and house prices --but it does not include changes in the expected appreciation of house values that correspond to the expected cost changes. The study, by Ann Dougherty and Robert Van Order, corrects the index for the true costs of consuming housing services. It is not opportunism but greater accuracy that is the reason for changing the index.
The imprecision in the current index severely hampers efforts to reduce federal expenditures. For example, more than one-third of federal outlays are explicitly tied to the CPI. The rationale for this linkage is obvious. The inflation rate alone should not determine whether the expenditures on a particular program increase or decrease. Indeed, with perfect indexation inflation would have no effect on the real value of expenditures. However, because of the index's bias, the inflation-adjusted expenditures on these programs increase rather than just keeping pace with inflation. From the Dougherty and Van Order study, one can infer that about one-fourth of the increase in these indexed expenditures --$6 billion to $7 billion in next year's budget--will be caused solely by the bias. In other words, about one dollar in six of the budget cuts enacted this year will be replaced in next year's budget by the CPI's overcompensation for inflation. It is no wonder that there is a good deal of criticism of indexed programs. However, it is the index rather than the need to index that is at fault.
Considering that if we had had the prescience to change the index in 1968, the reduction in federal expenditures would have been sufficient to eliminate the current federal deficit as well as cut taxes by $60 billion, an obvious question arises: Why not change the index?
There are three possible reasons for not doing so. First, there are technically difficult concerns involved with such a change. This, however, is an issue that has been studied to death. There is no perfect measure of the CPI. There will never be one.
Another reason for not changing the index is that it never seems like the right time. This is because just as the measure of the housing component causes the CPI to be overstated when rates are increasing, it also has the opposite effect when they are falling. Thus, when interest rates are thought to be high and about to decline it is less desirable to change the index. The present measure declines more rapidly than does the corrected one, and, correspondingly, the present index makes the incumbent administration appear to be a better inflation fighter.
One did not have to be Alfred Kahn to realize that the zero rate of inflation implied by last July's CPI was wrong. Nevertheless, at the time, the announcement certainly did not hurt the last administration in the election polls. Unfortunately for the last administration, interest rates did not continue on a downward spiral, so that by Election Day the CPI was once again overstating rather than understating the inflation rate. No study can predict the politically auspicious moment for a change in definition, any more than interest rate movements can be precisely predicted. However, the Dougherty and Van Order study can and does address what appears to be the fundamental obstacle to change-- the issue raised by organized labor.
In 1977, the last time BLS proposed changing the index, labor representatives argued that the bureau "simply could not make a credible case in support of the user-cost approach." If they could not then, they can now.
Dougherty and Van Order present a theoretically compelling case for a user-cost approach, as well as estimates of how much the current index varies from their theoretical ideal. Moreover, while Congress can mandate that federal expenditures should be linked to a different index, this has nothing to do with the index used in specific contract negotiations. Hence, a congressional decision to use a different index for federal expenditures would not affect private negotiations as long as the present index was also published.
With an administration committed to a monetarist attack on inflation, and a Federal Reserve that is letting interest rates move freely, it is clear that there is a willingness to tolerate high interest rates as a way to lower inflation. However, control over the budget becomes unnecessarily difficult to achieve when one of the chief means of lowering inflationary pressures-- high interest rates--is erroneously depicted in the CPI. It is time to use a new index for federal expenditures.