152 October 15, 1981 ADJUSTING THE CONSUMER PRICE INDEX INTRODUCTION Inflation is almost certainly the most serious problem now confronting the economy the Consumer Price Index (CPI). It is a measure, however, which seems seriously flawed and may be overs t ating the rise in the The most common measure of inflation is cost of living- during periods-of rising prices. The CPI measures the change in price of a uniform market basket of goods and and used as, a cost of living index which attempts to measure a con s tant standard of 1iving.l Although the CPI is a good indicator of pure price changes, it is not appropriate for many of its applications today of homeownership; its use of outdated buying patterns; failure to account for substitution in the market basket w hen consumers are faced with higher prices; and its limited applicability to specific subgroups I services over time. Unfortunately, it is often confused with, I Criticisms of the index include its treatment With more than one-half of federal expenditures adjusted for inflation either directly or indirectly by the CPI, and a soon to be indexed tax code, it becomes especially important to have as accurate a measure of inflation as possible. Each one percentage rise in the CPI now triggers an increase in fed e ral spending of about $2 billion.2 Thus, if biases in the index exaggerate the There are actually two CPIs frequently used price changes of urban wage earners, who comprise about 40 percent of the population households, or about 80 percent of the populati on.

Council of Economic Advisers and the Office of Management and Budget, the two indexes have never diverged by more than 0.5 index points 1980, the CPI-U was 256.2 and the CPI-W was 256.4.

Congressional Budget Office, Indexing with the Consumer Price In dex: Pro blems and Alternatives, June 1981, p. 2 The CPI-W reflects the The CPI-U is a somewhat broader index that covers all urban According to the In November 2 2 rate of inflation, beneficiaries of indexed programs will be overcompensated for increases in the cost of living. This places an unfair burden on taxpayers who finance these programs and adds to the difficulty of balancing the budget.

By one estimate, in fact, the federal budget would be 11 billion smaller today had the CPI been revised in 19

74. Thus, if fiscal responsibility is t.0 be achieved, the issue of indexing by the CPI must be addressed."

USES AND IMPLICATIONS OF THE CPI The CPI has three principal functions. It is used: 1) to analyze and develop economic policy; 2) to deflate other economic data and values; and 3) to serve as an escalator for wage payments contracts, and government expenditures.

The Formation of Economic Policy As an index of price change, the CPI attempts to measure the degree to which government economic policy ha s been successful in maintaining price stability, while also serving as a guidepost for future policy initiatives. Moreover, as the most commonly used indicator of.inflation, it helps form expectations .about inflation, and hence plays an essential role i n the long-term decision-making process of both businessmen and individuals. Any conceptual and statistical inaccuracies in the CPI may result in misleading economic signals that could prompt individuals in both the public and private sector to adopt inapp r opriate policies A Deflator A second important use of the CPI is to deflate other economic data. Economic series can be expressed in real dollars, i.e in a Even in the absence of any inaccuracies in the CPI, the whole concept of automatic indexation shoul d be re-evaluated, because in some cases it may result in gross inequities for example, benefited former Speaker of the House John W. McCormack by raising his initial annual pension of $50,000 (or 80 percent of his 62,000 salary) to $92,000 a year at the t ime of his death in 19

80. In contrast, the current Speaker of the House, Thomas P. (Tip) O'Neill, Jr receives only $79,125 a year. In addition, a four-star general who retired ten years ago would now receive $64,000 a year in retirement benefits, while an active four-star general would be paid only $54,000 annually. Though dramatic, these cases illustrate the problems with automatic indexing which may very well create disincentives for experienced federal or militaq personnel to remain actively involved T h e examples cited above are from: Clifford M. Hardin and Kenneth W. Chilton, Budget Control and Indexed Entitlements: Are They Compatible? (St. Louis Missouri: Center for Study of American Business, 1981), pp. 13, 15-16 The indexation of congressional pens i ons 3 dollars of constant purchasing power, by dividing the CPI into the current value of some other economic variable. comparisons over time in real terms without the obfuscating effects of inflation. The CPI is currently employed to deflate such series a s hourly and weekly earnings, retail sales, and some personal consumption expenditures used to calculate the gross national product. accuracy of the CPI is necessary This permits Because of its effect on these other indicators An Escalator Lastly, a major use of the CPI today is as an escalator to protect incomes against a decline in purchasing power due to inflation. In fact, the CPI was created during World War I to help adjust the wage payments of workers in shipbuilding when the cost of living had incr eased ffgenerally.ll The Bureau of Labor Statistics gathered survey information on workers' family expendi tures for about 145 commodities and services in thirty-two cities.

The index has since been expanded and refined to a point where it includes approxi mately 400 commodities and services in eighty-five areas. The CPI now is frequently used in collective bargaining agreements as an index for automatic cost-of-living adjustments known as COLAS). The first such COLA clause was negotiated in 1948 between th e United Auto Workers and the General Motors Corporation. of workers in the mid-l960s, these clauses became more and more popular As of January 1980, escalator clauses covered approxi- mately 58 percent of all workers in major bargaining units.4 Indexation is also widely used to adjust the wage payments of government workers at all levels An increasing proportion of federal expenditures in recent years has become explicitly indexed by the CPI to protect recipi ents from inflation. Indexing federal programs i s not something new. Military retirement benefits were adjusted as long ago as 1870 to keep pace with the growth of active-duty pay through a process known as I1recomputation.lf It was not until 1962, however that guidelines actually were prescribed for a d justing federal programs for inflation. In that year, legislation was enacted to index civil service retirement benefits to the CPI; this was extended to military retirement benefits the following year. In the 1970s, federal programs were indexed on a lar ge scale.

Social Security, the largest of the entitlement programs, adopted indexing in 1972 (to take effect in 1975) to adjust benefits for increases in the price level. Railroad Retirement, Supplemental Security Income, and Veterans Pensions soon followe d Social Security's lead. The growth of the relative importance of federally-indexed programs is illustrated by Table 1. From When rapid inflation began eroding the buying power 4 Phillip Cagan and Geoffrey H. Moore, The Consumer Price Index: Issues and A l ternatives (Washington, D.C American Enterprise Institute 1981 p 12. 'I 4 1962 through 1974, indexed programs never exceeded 6 percent of total outlays. Beginning in 1975, however, indexation assumed a growing role, with such programs comprising a third o f federal outlays. When other lluncontrollablesll are included, more than half of the federal budget in effect is on "automatic pilot.If5 This simply adds to the uncontrollability of the budget TABLE 1 Outlays for Major Programs Currently Indexed Billions o f Dollars Fiscal Year 1962 1963 1964 1965 1966 1967 1968 1969 1970. 1971 1972 1973 1974 1975 1976 1977 1978 1979 Total Indexed Programs 2.0 2.2 2.5 2.9 3.3 3.9 4.8 4.2 5.5 9.1 11.3 13.5 16.1 91.6 105.1 119.5 130.5 150.9 Indexed Programs as of Total Outlay s 1.9 2.0 2.1 2.4 2.4 2.5 2.7 2.3 2.8 4.3 4.9 5.5 6.0 28.1 28.7 29.7 29.0 30.5 NOTE: The table covers the following programs Disability, Military Retirement, Disabled Coal Miners, Food and Nutrition Assistance, Social Security, Railroad Retirement, Supplem e ntal Security Income, and Veterans Pensions. Outlays for these.programs are not included prior to their official indexation Civilian Retirement and SOURCE: Council of Economic Advisers and Office of Management and Budget Report on Indexing Federal Program s , January 1981, p. 13 One original justification for indexation was that it would remove much of the uncertainty for those dependent on federal Examples of uncontrollable items are interest on the national debt obligations entered into in the past but pay a ble in the future, and open-ended entitlement programs 5 programs by replacing what had been a sporadic practice of ad hoc increases. Indexing, it was argued, would assure benefit levels sufficient to maintain a constant standard of living, while sparing C ongress the burden of making repeated, time consuming adjustments. In fact, proponents of indexing maintained that it would be an effective way of llde-politicizinglt the issue of benefit increases. Political pressures had often led Congress to approve in c reases even greater than those that would have been provided by indexation. Between 1940 and 1974, when Social Security benefits were still adjusted on an irregular basis, the growth in benefits actually outpaced the CPI by 40 percent.6 Thus, indexing was also seen as a means of constraining costs in politically sensitive areas.

Indexing is being extended to a new area by the Economic Recovery Tax Act of 19

81. By it the tax system is indexed, as of 1985, by automatically adjusting individual tax brackets , the personal exemption, and the zero bracket rate to reflect inflation. The objective is to prevent further real tax increases that have been resulting from inflation-induced bracket creep as accurate a measure of inflation as possible. If the CPI does h ave an upward bias, the beneficiaries of federally-indexed programs are overcompensated for increases in the cost-of-living and taxpayers will receive real tax cuts once the tax code is explicitly indexed. By raising expenditures and lowering revenues, au t omatic indexation raises serious questions concerning sound fiscal policy management. Indeed, the very indexing which is supposed to measure inflation may actually contribute to it. example, a larger deficit may exacerbate inflationary pressures if increa s ed government borrowing crowds out more productive investment in the private sector. Inflation would also be fueled if the Fed is forced to monetize all or part of the additional deficit. Similarly, cost of living adjustments in the private sector based o n the CPI may artificially inflate labor costs and lead to a poor allocation of resources, thereby reducing the output of goods and services relative to the money supply actions drive up the CPI and contribute still further to the inflationary spiral This n ew legislation makes it especially important to have For These LIMITATIONS OF THE CPI IN ITS PRINCIPAL USES A Fixed Market Basket The CPI measures price changes by comparing current prices to base period prices, using base period consumption patterns for w eighting purposes. In other words, given a fixed market basket Council of Economic Advi.sers and Office of Management and Budget, Report on Indexing Federal Programs, January 1981, p. 6 6 of goods and services, it measures the relative costs of buying the same bundle of goods and services in subsequent time periods.

The market basket currently used is based on a 1972-73 survey As a fixed-weight index, the CPI tends to exaggerate the true cost of living because it fails to account for substitution as consum er preferences change As prices of various products change at different rates, consumers generally shift purchases from pro.ducts that have become relatively more expensive to those that have become relatively cheaper. Thus, during inflationary periods, t h e CPI usually overstates the rate of inflation by placing inordinate weight on items that have risen most rapidly in price. For example, in response to the escalating world oil prices, the proportion of gasoline in total personal consumption fell from 3.5 percent in 1972 to 2.8 percent in 1980.7 Despite the decline in gasoline consumption, the CPI continues to use the out-dated base period weight, which clearly biases the index upwards.

Moreover, because the CPI does not take into account substi tution, it is an inadequate measure of the cost needed to maintain a constant standard of living A true cost-of-living index would attempt to compare the cost to the consumer of reaching a certai n level of 'Isatisfaction or "utility, though such qualities admittedly would be hard to measure price changes, consumers can attain a given level of utility at less cost than they could if they were to buy the original market basket. Consider the followin g example By adjusting to relative Ilmagine a consumer who initially spends $2 on 1 pound of beef and 1 pound of pork, both of which cost $1 dollar [sic] per pound. If the price of pork then doubles but the price of beef remains the same the original baske t of purchases would cost 3.00 rather than $2.00 A fixed-weight index like the CPI would register a 50% increase in the trcost of living.Il However, when this person consumes one pound of beef and one pound of pork, additional amounts of pork and beef are w orth about the same to him We know this because in the original period he paid the same amount for the two meats Thus, although the consumer could spend his $3.00 on the original market basket, he could make himself even better off by purchasing, for exam p le 44 pound less of pork and pound more beef. That would mean that $3.00 is a higher expenditure than would be necessary to achieve his original level of satisfaction.8 The Bureau of Labor Statistics revises the CPI every 10 to 12 years. During periods of relative price stability, this is I 7 Congessional Budget Office, op. cit p. 46.

Council of Economic Advisers and Office of Management and Budget, op. cit pp. 28-29. 7 adequate, but when prices begin escalating rapidly, the index can easily become unrepresentative of actual consumption patterns.

The CPI could be made more precise by updating the consump tion patterns more frequently, thereby acquiring some of the characteristics of a chain-index. A chain-index measures pure price changes between adjacent periods using the earlier period's consumption pattern for weighting purposes. would complicate comparisons of price changes between non-adjacent time periods, it would make the CPI a more representative cost of l'iving index because it would consider cha n ges in consumption patterns resulting from changes in tastes, relative prices, or the introduction of new products. Yearly updates might prove to be a valuable improvement in the usefulness of the index Even though this The Homeownership Component Perhaps the most serious shortcoming of the CPI is its treatment of homeownership costs. accounts for nearly one-quarter of the index and is widely believed to overstate significantly the importance of housing in the CPI As a rough approximation of the magnitude o f this overweighting the CPI accords housing nearly five times as much importance as it does the residential rental component, even though only twice as many dwellings are owner-occupied. Even in the national income and product accounts, which measure the cost of current production for consumption, homeownership is weighted only two-and a-half times as heavily a rental units, also well below the factor of five in the CPI. The large weight attributed to housing often exaggerates the actual increase in the c ost of living The homeownership component Five separate elements are considered in measuring the cost of housing: the home purchase price, contracted mortgage interest cost, property taxes, property insurance, and maintenance costs.

The latter three do not pose problems because they represent expenses that are incurred on a regular basis arise, however, from the use of current purchase and current financing costs Distortions Durable goods such as housing, cars, and appliances are purchased in one period bu t generally provide services over a number of periods. another commodity gives it a much greater weight than is appropri ate A true cost-of-living index should indicate the cost of consumption in each time period, i.e the price one would pay to ifconsume'l 'the shelter offered by that home. The price of a house, however, can also be viewed as an investment, one that could yield a homeowner a capital gain in later years when the house is resold. Economist Robert Gordon points out Far from being a source of h i gher prices, squeezed budgets, and falling living standards most Americans have found home ownership to be a source of wealth creation and one of the few spots in the family budget Treating the purchase of a house as simply 8 that is largely insulated fro m inflation Increases in home purchase prices for existing home owners are a source of higher wealth, and llleveragell (the small initial share of their down-payment equity) makes the value of their equity increase by a multiple of the percentage annual in c rease in house,prices. Because income is properly defined as consumption plus the change in one's wealth, higher home prices by this definition also raise individual incomes.y In other words, the price paid for a house represents both a consumption and an investment function, yet only the first belongs in the CPI. excluded from the index as changes in the prices of stocks and bonds already are. A properly constructed index thus would contain only the cost of "shelter servicesi1 provided by the house Change s in the investment value of a house should be The large weight accorded to mortgage interest costs presents another difficulty. total interest that borrowers contract to pay over the first half of the life of the mortgage (because mortgage contracts on av e rage remain in effect for only half of the originally scheduled dura- tion). Present expenditures on mortgage interest depend on both the purchase price of the home as well a the mortgage rate. By including the price of a home, this, again, ignores the di s tinction between the consumption and investment aspects of housing addition, the current treatment of mortgage payments disregards the fact that most.homeowners are not affected by rising home prices and interest rates because their payments have been det e r mined in the past when the size of this component was much smaller. One alternative to this approach is suggested by Cagan and Moore The weight is determined by including the In Tlhe weight of the mortgage interest could be based on the interest actuall y paid in the survey period of all homeowners, in which case the weight would reflect the distribution of mortgages by maturity in the survey sample. alternative would be lower than under the present method because in recent decades mortgage rates and hous e prices have been rising, and on the average relatively more has been spent on new mortgages than on old ones The weight of the mortgage interest under this In addition, the interest rate used in the index could be an average of rates currently in effect f or all outstanding mortgages. mined at previous dates but are still being paid. treatment would correspond to the way rents and some other contract prices are treated now. The mortgage These rates were largely deter- This I Robert J. Gordon, "The Consumer Price Index: Measuring Inflation and Causing It," The Public Interest 63 (Spring 1981 pp. 121-122. 9 component would then reflect a weighted average of actual interest payments being made on outstanding mortgages, rather than the current rate on new m0rtg ages.l"

This approach would not only improve the accuracy of the CPI, but would also largely reduce the inherent instability of the housing component. costs using the current rate and a fifteen-year moving average, along with the resulting impact on the CP I. As is evident from the table, fluctuations in mortgage interest costs as well as the CPI could have been markedly reduced had the fifteen-year moving average been employed during the eleven-year period from 1968 to 19

79. Moreover, it would resolve any problems that might arise if variable rate mortgages become popular Table 2 compares annual changes in mortgage interest Although the change discussed above would be an improvement over the current treatment of the housing component, still more can be do n e. The Bureau of Labor Statistics has developed several experimental indexes that attempt to estimate the cost of Ilconsum ing" a home. The most commonly discussed alternative, the CPI X-1, does this by allowing market rental rates for homes act as a prox y for the value of shelter services they provide TABLE 2 Alternative Mortgage Interest Costs in the CPI, 1968-1979 percentage change per year Mortgage Total CPI-U With fifteen December InteresE Costs Excluding year average to Current Fifteen-year Official m ortgage mortgage December rate average version interest interest 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 Average 1968-1979 16.3 19 2 15.9 -7.1 2.1 19.1 21.5 7.1 0.7 10.8 22.0 34.7 12.9 5.1 8.9 11.4 6.9 4.8 5.6 10.2 11.7 9.3 7.8 10.3 15 . 7 8.9 4.7 6.1 5.5 3.4 3.4 8.8 12.2 7 O 4.8 6.8 9.0 13.3 9.0 4.3 5.7 5.1 3.8 3.4 8.5 11.8 7.0 5.1 6.6 8.2 11.6 6.7 4.3 5.8 5.2 4.0 3.4 8.4 11.8 7 2 5.2 6.6 8.3 11.7 6.8 SOURCE: Phillip Cagan and Geoffrey H. Moore, The Consumer Price Index and Alternatives (Washington D.C American Enterprise Institute 1981 p. 39 Issues Cagan and Moore, op. cit pp. 37-38 10 Quality Chanqes One of the more cumbersome conceptual difficulties in calcu- lating the CPI is estimating the dollar value of quality changes.

For example , with each model change of an automobile a distinction must be made between the overall price increase and that arising from quality improvements to separate the effects of a price rise from that of a quality change, but the degree to which it succeeds i s questionable, given the difficulty of measuring the incremental value consumers place on quality improvements. For example, cleaner air and water and improved working conditions have been achieved on a large scale through government regulations. These be n efits, however, are attained only by devoting greater resources to the production of goods, thereby driving up prices and, consequently, the CPI. But the CPI ignores these ancillary changes. Thus, it appears that the same standard of living has become mor e expensive while, in fact, part of the higher costs actually may be buying what many regard as an improved overall quality of life.

The Treatment of Taxes and Government Services There are also questions about whether the CPI properly treats taxes and gov ernment services as part of the price of goods and services and are therefore incorporated in the CPI, whereas individual income taxes are excluded despite the fact that both are levied to finance govern- ment activities. The unequal treatment of taxes ca n lead to incongruities. An increase in excise taxes that is offset by an equal reduction in income taxes would inflate the CPI even if real disposable income remains the same.l If, for example, the government wished to pursue a policy of increased energy c onserva tion, it may decide to raise the sales tax on gasoline to discour- age its consumption and return the revenues to the public through reduced income taxes. This would accomplish its goal of conserva- tion by altering relative prices (making gasolin e relatively more expensive) without changing real incomes blind to this trade-off and would register an increase in the cost-of-living. In turn, this would trigger increased outlays of federal expenditures on the many indexed programs The Bureau of Labor S tatistics attempts Excise taxes are included Yet the CPI would be In addition, the government uses its power of taxation to supply goods and services provide benefits that affect the standard of living and, conse- quently, should be included in a comprehe n sive cost-of-living index the indication that more money is needed to preserve a given standard of living the value individuals place on the resulting change in the distri- bution of resources between the public and private sectors These goods and service s clearly By raising the CPI, an increase in excise taxes may give This may not be true, as it would depend on l1 This would probably alter the distribution of income. 11 Equity Considerations Automatic indexation has been opposed for reasons of equity Whe n there is a general decline in real, as well as efficiency national incomes, full indexation al.lows those members of society whose incomes are adjusted for inflation to preserve a constant standard of living at the expense of those whose wages are not es c alated. These disparities may arise for a variety of reasons e.g when there is a rise in the price of imported goods relative to domestically produced goods or when domestic labor productivity declines. Moreover, complete protection may conflict with publ i c policy objectives, e.g conserving gasoline to reduce dependence on foreign oil sources exclude imports from the CPI or to limit benefit increases to the CPI or a wage index (whatever figure is lower). The final deter- mination, of course, is contingent u pon the desirability of any resulting transfers between those financing and those receiving adjustments for inflation These concerns have led to proposals to Specialized Indexes A common criticism of the CPI is of its limited applicability to particular s ubgroups, such as the elderly and the poor. alleged inadequacy has prompted calls for special indexes that would more accurately reflect the cost of living of certain demographic groups and/or those benefiting from indexed programs.

Since vast resources ar e now devoted to support the elderly, it is essential to determine the applicability of the CPI as an escalator for this group. The large weight accorded housing in the CPI tends to exaggerate the cost of living for older people because only a very small p roportion of them are in the market seeking housing. Alvin Rabushka and Bruce Jacobs of the Hoover Institution calculate that 70 percent of the elderly live in their own homes and over 80 percent of these make no mortgage payments whats0ever.l" Older peop l e, on the other hand, consume a relatively greater proportion of medical services than others the cost of which has in recent times been rising faster than the rate of inflation. On the surface, this factor would appear to lead to an underestimation of th e rise of the cost of living for the elderly. But because of increased publicly financed medical coverage available to this group through programs such as Medicaid and Medicare, the medical component has generally been overstated.lY In any case, it is clea r that the consumption patterns for the elderly are not the same as for other age groups of the poor. This imprecision arises by including the expendi This Similarly, the CPI may inaccurately gauge the buying patterns l2 Alvin Rabushka and Bruce Jacobs Agi n g and Public Policy: Rethinking Issues and Programs Hoover Institution Reprint Series No. 33, p. 1 la .The CPI does not take into consideration an offset for medical expenses covered by insurance, but this is an average for everyone in the sample. 12 ture s of high-income families with low-income families. Higher income groups receive greater weight because they spend more including goods beyond the affordability of low-income house holds. As a result, this acts to overweigh luxury goods and underweigh nece ssities, and is therefore a poor tool in determin ing eligibility and benefit levels for the various income security programs.

A word of caution regarding the construction of specialized indexes is that there is no guarantee that these indexes will give re sults any different from a general index or that 'they will reduce the costs of indexing. In fact a more representative index may actually be more costly if it turns out that the CPI had understated the rate of price increase for certain groups.

Nevertheless, the tremendous importance of indexing based on the CPI requires a closer evaluation of specialized indexes.

ALTERNATIVES TO THE CPI CPI x-1 A "rental equivalencei1 (CPI X-1) index now being developed by the Bureau of Labor Statistics attempts to circ umvent some of the more serious problems connected with the housing component. This approach tries to distinguish between the consumption and investment aspects of housing by using.market rents as a proxy for the shelter services of a similar owner-occupi e d home. In essence, it seeks to capture the cost a household incurs by llconsumingll the housing services itself rather than renting to someone else. To be accurate, of course, this index must take into account differences in location, size, and quality. The CPI X-1's treatment of homeownership is the single most palliative change that could be desired because it reduces the excessive weight now accorded housing. This indeed, however, would still retain the remaining shortcomings of the CPI.

PCE "Chain Index"

Many economists suggest replacing the CPI with the Personal Consumption Expenditure (PCE chain index" of the National Income and Product Accounts. This index would be preferable to the CPI for several reasons 1) its coverage is somewhat broader for it inc l udes all goods and services currently produced for con'sumption, approximately two-thirds of GNP; 2) it employs the rental equivalency approach used in computing the CPI X-1; and 3 it uses current consumption patterns rather than those determined in 1972-

73. The principal drawback to the PCE chain index is its omission of used items, such as cars and appliances, that also comprise a part of a consumer's living cost.

According to estimates for the Congressional Budget Office CBO federal expenditures in 19 81 would be lower by 11 billion 13 had the PCE chain index replaced the CPI in 1974.14 new index at this time, however, may not result in immediately reduced outlays. This paradox stems from the effect of interest rates on the CPI. Just as the index overs t ates inflation when interest rates are rising, it understates when they are declining As a consequence, some economists argue, it may be useful to delay switching to an alternative index to allow the government to recoup past losses from overindexation. T h is objection could be overcome by instituting a one-time adjustment in benefit increases to make them more accurately reflect what benefits would have been had a more precise measure been used all along. CBOIs calculations reveal that such a change would r educe 1981 increases from about 11.2 percent to about 3 percent.lS Once this is accomplished, a more accurate index, such as the CPI X-1 or the PCE chain index, could be permanently installed Changing a Usinq the Lower of a Wage or Price Index The index c o uld also be changed, some maintain, to assure that everyone bears some of the burden of an economic slowdown or surging inflation of living adjustments to the lower of the CPI or a wage index. This could preclude recipients of indexed programs from gainin g relative to active workers during periods when real wages are falling. Wages ordinarily rise by the rate of increase in prices plus a premium for productivity advances, though this was not the case on two occasions in the past decade (1974-75 and 1980 In these years, those receiving indexed benefits actually did better than active workers, whose wages or salaries were not automatical ly adjusted for inflation. Over time, however, a change in this method would reduce the real level of benefits, unless a If catch upif period were allowed for recipients to regain their original benefits in real terms. This could be achieved by delaying the switch back to a price index from a wage index until benefits attained their previous level in real dollars.

This proposal would reduce outlays for indexed programs only if wages continue to lag behind prices nothing to correct any of the other deficiencies in the CPI One way of doing this would be to limit cost Moreover, it would do Capping the CPI Some experts suggest reta i ning the CPI as the official barometer of inflation but imposing a discretionary cap on benefit increases federal pay adjustments. The advantage of this approach over automatic indexing is that it adds flexibility and permits the consideration of changing economic conditions. Unfortunately This would treat indexed programs just like current l4 Congessional Budget Office, op. cit p. 75 l5 Ibid D. 76. 14 such a cap may be arbitrary and subject to shifts in political winds, while also failing to correct for t he flaws in the CPI.

Comparisons of Alternative Indexes Table 3 illustrates the difference had the various indexes Both the CPI X-1 and PCE chain been used over the past decade. index would have been more stable than the CPI and would have grown at a less rapid rate TABLE 3 Comparison of Percent Changes in Alternative Indexes Average PCE Hourly Year CPI u CPI x-1 Chain Earnings 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1970-80 3.5 3.4 8.3 12.2 7.4 5.1 6.7 9.0 12.7 12.5 116.6 3.7 3.3 8.0 11.1 6.8 5. 2 6.3 7.8 10.6 10.9 103.0 3.8 3.7 7.6 11.0 6.4 4.9 6.3 8.0 9.9 10.4 99.8 6.9 7.7 6.6 8.3 6.2 7.7 7.7 8.0 7.8 8.8 109.6 NOTE: Percent changes, annual rates, fourth quarter to fourth quarter.

Percent change over ten-year period, 1970:4 to 1980:4.

SOURCE: Congressional Budget Office, Indexing with the Consumer Price Index Problems and Alternatives, June 1981, pp. 56 and 58 The index measuring average hourly earnings appears to indicate that the worker lost ground to inflation in the 1970s.

This is somewhat misleading because the index excludes other forms of compensation such as fringe benefits and employer contri butions to various social insurance and private benefit plans which have increased substantially over the last ten years.

CONCLUSION In a recent a rticle in the Wall Street Journal, Lindley Clark, Jr. asserted The Consumer Price Index is a dandy index of inflation when there isn t much inflation to measure. Iri le Lindley H. Clark, Jr The CPI Does Fine If There's No Inflation,"

Wail Street Journal, April 14, 1981, p. 31. 15 Although this may be too harsh an assessment, it does indicate that the various shortcomings of the CPI may make it.inappropriate for many of its present applications. Recognizing that there is no perfect measure, policymakers ma y have to settle for either a version of the CPI or some alternative index. The final new form, of course, would be set by the intended specific purpose of the index Probably the most feasible solution to this problem is simply to improve the present index . twofold better indicators of changes in the cost of Jiving, they are not nearly as well known as the CPI. Thus, retaining the CPI in a modified form may be desirable if problems associated with a switch to a lesser known alternative are to be avoided. Se c ond, most of the distortion.in the CPI can be eliminated by employing a rental equivalence measure of homeownership and by updating consumer buying patterns more often The advantages.would be First, although the CPI X-1 and the,PCE chain index are Many of the suggestions to improve the CPI outlined in this paper have been around for at least twenty years, when they were made by a committee headed by former University of Chicago economist George Stigler. Since then, despite the wealth of evidence in support of changing the CPI, the Bureau of Labor Statistics has done suprisingly little to upgrade the index.

Robert Gordon According to IJt is striking that the BLS spent $50 million during 1972-77 to revise the CPI without curing any of its major defects It see ms clear in retrospect that the BLS spent its revision money on the wrong things, improving the number of outlets covered or the number of consumers surveyed rather than investing money in more rent data on single-family homes or on performance data for n ewly introduced models and products.

With billions of dollars in cost of living adjustments hinging upon each percentage point change in the CPI, it is imperative that its flaws either be corrected or a new index adopted for adjustment purposes.

Peter Germanis Policy Analyst Gordon, op. cit p. 134.