Is the current business cycle different? Does how we measure matter?
DATA AND METHODOLOGY
The U.S. NIPA are constructed by the Bureau of Economic Analysis (BEA) of the U.S. Department of Commerce. From time to time, the base period for expressing real magnitudes has been updated because relative prices and expenditure patterns in the U.S. economy change. Development of new products and services and changes in relative prices affect economic behavior. In order to perform meaningful analysis of an economy, one must consider the structure of the economy at the time in question.
Data on Gross National Product (GNP) and its components expressed in prices and weights of base years prior to 1987 come from the Survey of Current Business;(2) data on the NIPA expressed in terms of 1987 prices and weights were downloaded from the Data Resources Inc. U.S. central data bank as was the Conference Board's index of consumer confidence. The data are available from the author upon request.
Since 1957, seven recession troughs have occurred, including that of March 1991. The recoveries following six of them (all but the July 1980 trough) continued for more than the seven quarters considered here. For each of these recessions and recoveries (with the exception of the recovery following the 1980 recession), we examine the changes in GNP and some of its principal subaggregates expressed in terms of 1987 prices and expenditure weights and the prices and weights of the BEA base year closest to the event.(3 4 5)
As shown in Table 1A, when base years close to the dates of the recessions are used to evaluate the data, in terms of GNP decline, at 1.6 percent, the 1990-91 recession was three-fifths as deep as the 2.6 percent average GNP fall of the preceding six. The recession was mild compared to the preceding three, but far deeper than the 1969-70 recession and slightly deeper than the 1960-61 downturn. None of the subaggregates(6) of GNP behaved in a fashion markedly different from its behavior in at least one previous recession. Expenditures on consumer services fell during the last recession, but the decline was small.
Table 1A Percent Change in Real GNP and its Major Components During Recessions 57:Q3 60:Q2 69:Q4 73:Q4 80:Q1 81:Q3 90:Q3 to to to to to to to 58:Q2 61:Q1 70:Q4 75:Q1 80:Q3 82:Q4 91:Q1 GNP -3.4 -1.4 -0.1(1) -4.9 -2.3 -3.2 -1.6 Consumption -0.4 -0.5 1.5 -0.4 -0.9 2.3 -1.4 Durables -9.8 -8.6 -7.1 -9.1 -6.9 2.5 -5.0 Nondurables -0.9 0.1 2.8 -1.0 -1.0 1.8 -1.6 Services 3.2 1.9 2.9 3.1 0.7 2.6 -0.4 Investment -21.0 -15.1 -5.2 -34.3 -16.0 -27.1 -11.1 Structures -4.5 -1.8 NA NA NA NA -10.0 Producer Durables -17.4 -10.5 NA NA NA NA -4.8 Exports -10.8 2.9 3.1 1.6 -3.8 -14.1 1.4 Imports 3.0 -7.2 0.5 -11.5 -12.9 -7.1 -5.8 Government 5.1 3.1 -1.6 2.6 0.5 4.7 2.1 Federal 4.1 2.4 -9.7 0.5 2.0 10.2 3.7 State, Local 6.1 3.9 5.5 3.8 -0.4 0.8 1.0 Base Year 1958 1958 1972 1972 1982 1982 1987 NA: Not Available 1 The path of real output during this recession was unusual. GNP in 1970:Q4 was 0.1 percent lower than in 1969:Q4. It fell 0.4 percent between 1969:Q4 and 1970:Q2. In 1970:Q3, however, real GNP rose above previous peak output (achieved in 1969:Q3) only to fall during the following quarter.
The situation is different if the changes in real GNP and its major components are expressed in terms of 1987 prices and expenditure weights, however. These data are presented in Table 1B.(7) GNP changes expressed in base years close to the recessions and in 1987 prices and weights are shown in Figure 1.
Expressing changes in GNP and its major components in terms of 1987 prices and expenditure weights does not make the 1990-91 recession appear "unusual" relative to previous downturns. It does, however, change many of the figures that pertain to prior recessions. Six of the seventy numbers that describe changes in GNP and/or its components during earlier recessions change sign with the change in base year. (The 1960-61 recession is transformed to a period of positive growth.) In another eleven cases, the sign of the number is the same in Tables 1A and 1B, but the number in one of the tables is at least twice as great in absolute value as the corresponding figure in the other table. One can obtain a very distorted view of past recessions through the lens of 1987 prices and weights.
After their introduction, the prices of many new products and services fall substantially. Substituting later, lower, prices of these goods and services for earlier, higher, prices reduces the relative importance of these goods and services (and of the fall in output of these goods and services) in historical data, reducing the measured decline in GNP. The decline in output during the 1969-70 recession may have been too small for the price changes to make a difference in the measured fall in GNP. Only in 1980 do the 1987 base-year data show a larger decline in GNP than do the figures based on prices and weights of earlier years. The explanation of this anomaly may lie in the credit controls that were largely responsible for that recession. It is possible that the particular mix of expenditures affected by the controls makes the recession deeper when it is measured in terms of 1987 prices and weights.
Table 1B Percent Change in Real GNP and its Major Components ($1987) During Recessions 57:Q3 60:Q2 69:Q4 73:Q4 80:Q1 81:Q3 90:Q3 to to to to to to to 58:Q2 61:Q1 70:Q4 75:Q1 80:Q3 82:Q4 91:Q1 GNP -2.6 0.4 -0.1 -4.4 -2.7 -2.9 -1.6 Consumption 0.6 0.1 1.7 -0.8 -1.0 2.1 -1.4 Durables -5.2 -9.3 -8.2 -10.2 -7.6 0.4 -5.0 Nondurables -0.9 0.5 2.0 -3.2 -1.6 1.5 -1.6 Services 3.4 1.8 3.6 3.1 0.6 2.8 -0.4 Investment -16.2 -8.8 -5.8 -30.5 -16.3 -21.9 -11.1 Structures -4.8 -0.4 NA NA NA NA -10.0 Producer Durables -17.2 -11.0 NA NA NA NA -4.8 Exports -13.0 4.0 2.9 3.3 -3.6 -12.9 1.4 Imports 7.5 -6.9 1.5 -10.4 -12.0 -2.4 -5.8 Government 1.7 4.2 -1.9 2.7 -0.4 3.4 2.1 Federal -1.4 2.5 -8.3 1.7 1.6 6.2 3.7 State, Local 6.4 6.2 4.4 3.4 -1.8 1.4 1.0 NA: Not Available
Consumption of services generally holds up better in terms of 1987 dollars than in dollars of earlier base years; consumption of goods tends to be weaker. Like consumption of services, exports and imports are stronger in terms of 1987 prices and weights than when calculated on the bases of earlier prices and weights. On the other hand, using the most recent base year tends to weaken government expenditures during earlier downturns.
Expressed in terms of base years closest to the individual recessions, the behavior of the U.S. economy during the 1990-91 downturn was not out of line with its behavior during other recessions, but the data in Table 2A show that its behavior during the current recovery is different from that during past expansions. First and foremost, between 1991:Q1 and 1992:Q4, GNP growth was less than three-fifths as great as it was during the first seven quarters of the next slowest expansion, that following the 1973-75 recession.(8) Some observers have attributed the weakness of the current recovery to the shallowness of the 1990-91 recession, reasoning that past recoveries were stronger because far more capacity was idled by the recessions that preceded these recoveries; with less slack in the economy after the 1990-91 recession, a slower recovery was to be expected. The 1969-70 recession, which was followed by 10.0 percent growth during the first seven quarters of recovery, was much shallower than the 1990-91 downturn. The 1960-61 recession, which was followed by growth of 11.5 percent during the first seven quarters of expansion, was also shallower than the 1990-91 recession in terms of GNP decline.(9) Turning to the components of GNP, across the board, growth in consumer expenditures, which account for two-thirds of GNP, was much weaker than during past recoveries. There are several demographic and economic factors that may help explain the relative weakness of consumer spending during the current expansion:
Table 2A Percent Change in Real GNP and its Major Components During First Seven Quarters after Recession Troughs 58:Q2 61:Q1 70:Q4 75:Q1 82:Q4 91:Q4 to to to to to to 60:Q1 62:Q4 72:Q3 76:Q4 84:Q3 92:Q4 GNP 11.5 11.6 10.0 9.0 11.4 5.1 Consumption 9.2 8.9 10.0 9.9 8.6 4.8 Durables 22.7 21.8 31.7 20.7 23.8 12.6 Nondurables 7.2 6.0 5.8 8.2 6.6 3.2 Services 7.1 8.2 7.9 8.1 6.3 3.8 Investment 42.7 29.3 26.0 27.8 62.5 14.3 Structures 13.5 8.8 NA NA 24.4 7.8 Producer Durables 17.9 19.4 NA NA 34.1 12.3 Exports 13.4 7.1 11.1 7.6 12.7 5.4 Imports 14.6 18.3 13.5 23.1 43.1 11.1 Government 0.3 11.2 0.6 0.6 3.7 -0.2 Federal -4.1 16.2 -6.0 0.6 2.1 -5.1 State, Local 6.2 5.7 5.7 0.5 4.9 3.3 Quarters(1) 2 1 1 4 1 5 Base Year 1958 1958 1972 1972 1982 1982 NA: Not Available 1 Number of quarters between trough of the recession and the surpassing of the previous peak GNP.
1. Fewer people are entering the high-spending household-formation years in the early 1990s than did so during the past two decades.
Table 2B Percent Change in Real GNP and its Major Components During First Seven Quarters after Recession Troughs 58:Q2 61:Q1 70:Q4 75:Q1 82:Q4 91:Q4 to to to to to to 60:Q1 62:Q4 72:Q3 76:Q4 84:Q3 92:Q4 GNP 9.5 8.4 9.0 8.5 10.6 5.1 Consumption 7.8 7.9 9.2 9.3 8.5 4.8 Durables 13.1 17.2 29.6 21.6 24.0 12.6 Nondurables 5.6 4.9 5.5 7.9 6.6 3.2 Services 9.0 8.9 8.3 7.8 5.3 3.8 Investment 40.3 17.9 27.8 26.7 53.5 14.3 Structures 23.0 9.7 NA NA 27.2 7.8 Producer Durables 17.5 15.7 NA NA 28.2 12.3 Exports 16.1 6.9 9.8 7.3 10.8 5.4 Imports 11.9 16.7 14.0 20.0 41.4 11.1 Government -2.1 6.4 -2.4 -0.5 4.1 -0.2 Federal -7.3 8.1 -8.8 -1.5 4.0 -5.1 State, Local 4.9 4.5 3.2 0.2 4.1 3.3 Quarters(1) 2 2 1 3 2 5 NA: Not Available 1 Number of quarters between trough of the recession and the surpassing of the previous peak GNP. 2 Evaluated in terms of 1987 prices and expenditure weights, there was no decline in real GNP between 1960:Q2 and 1961:Q1.
2. Consumer confidence has risen proportionately less during this recovery than it did during the first seven quarters of the recoveries of the 1970s and 1980s.(10)
3. An increasing proportion of adult children has returned to or remained in the parental home after finishing school. The joint household requires fewer goods than two separate households.
Fixed investment has also been weaker than during earlier recoveries, but the demand for investment goods is a derived demand. As long as the demand for consumer goods and services is weak, the demand for producer durables and nonresidential structures used to produce them can be expected to be soft. The same factors that may be restraining the growth in output of consumer goods and services may account for softness in the growth of investment in residential structures relative to that of past recoveries. In addition, in spite of cuts in federal purchases during the 1958-60 and 1970-72 expansions, in each of the earlier recoveries, government spending rose, but, because of cuts in defense spending, overall government purchases of goods and services fell slightly during the first seven quarters of the current expansion.
The situation is somewhat different if the changes in real GNP and its major components are expressed in terms of 1987 prices and expenditure weights, however. These data are presented in Table 2B.(11) GNP changes expressed in base years close to the recoveries and in 1987 prices and weights are shown in Figure 2.
As prices of newly introduced goods and services fall, the outputs of these goods and services tend to rise more than they would absent the decline in price. With the substitution of later, lower, prices of those goods and services for earlier, higher, prices, the importance of these goods and services in GNP falls. The growth of GNP also declines. Thus, it is not surprising that using 1987 as the base year for all of the data reduces GNP growth during past expansions.(12) Further, it causes four of the sixty-one numbers in Table 2A pertaining to earlier recoveries to change sign. For all but the recovery from the 1957-58 recession, shifting the base year changes the number of quarters between the trough of the downturn and the surpassing of previous peak output. The distortions introduced to the data by using 1987 prices and weights may be less severe during expansions than during recessions, but using prices and expenditure weights of a period that is not close to the expansion can lead to a mistaken analysis of the recovery.
Expressing the data in terms of 1987 prices and weights does not change the major difference between the current recovery and those of the preceding three and a half decades: consumption is growing far more slowly coming out of the 1990-91 recession than it did in previous expansions. In terms of 1987 prices and weights, the growth of investment remains weaker than it had been in past recoveries. The decline in government purchases during the current recovery loses its uniqueness, however. The 1987-based data show declines in federal expenditures sufficient to cause general government purchases to fall during three of the preceding five recoveries.
The current recovery -- but not the 1990-91 recession -- has been different from those of other business cycles since the mid-1950s. The chief difference has been weaker consumption growth. A smaller number of people in the household-formation years, the perception of poor job prospects, and an increase in the number of adult children living with parents may account for some of the weakness in consumption expenditures. Fixed investment has also lagged during the current recovery. The weakness in consumption expenditures and the reduced rate of household formation are likely to account in large measure for the softness in investment expenditures. In addition, a decline in federal defense purchases has occurred sufficient to cause a fall in overall government spending during the current recovery.
The data presented here demonstrate that a very distorted view of economic history can emerge if the set of prices and expenditure weights used to analyze movements in the real economy does not pertain to a time close to the period being analyzed. If one is studying the economy of a particular time, it is necessary to do so in the context of the mix of goods and services available at that time and the relative prices at which they are available. What does it mean to use 1987 prices and expenditure weights to examine the 1960s or the 1970s? Many things that were commonplace in 1987 -- such as personal computers, fax machines, and compact discs to name just three -- did not even exist two or three decades ago. Conversely, many everyday items of earlier years -- such as twenty-one inch black and white television sets and adding machines -- have virtually disappeared. Thus, it is important to update the expenditure weights and prices that underlie the national accounts frequently. One must pay attention to the numbers. How we measure does matter.
1 The data pertaining to the 1990-91 recession and the subsequent recovery will be revised by BEA in the future.
2 Survey of Current Business, August 1965, pp. 26-27; July 1966, p. 11; October 1982, p. 42; February 1986, p. 21; March 1986, p. 59; July 1986, p. 26; and July 1987, p. 20.
3 The base year chosen is not the one used by BEA at the time of the recession or recovery, which may have pertained to a year more than a decade prior to the event considered. In addition, since it takes considerable resources to gather and assimilate the data needed to update the base year for the NIPA, the last few base-year changes have taken about four years to introduce. Thus, the NIPA were first published on a 1982 base in the December 1985 issue of the Survey of Current Business and first published on a 1987 base in the November 1991 issue of the Survey.
4 It would be interesting to examine in addition the changes in real GNP and its components expressed in terms of the base years used by BEA at the times of the cyclical downturns and recoveries to get the view presented to contemporaneous observers of the business cycles. The NIPA are revised several times, however. In the cases of some past recoveries, the last revisions occurred after the base year was changed. For these expansions, there are no data published using old base years comparable to those used in this study with respect to the degree of revision they have undergone.
5 BEA sometimes revises historical NIPA data beyond the annual revisions of the preceding three years' quarterly figures and changes of base year. Data based on 1972 prices and expenditure patterns come from the benchmark revisions of the NIPA carried out in connection with the revision of the input-output table for the U.S. economy.
6 The choice of GNP components presented is governed by the series appearing in the Survey of Current Business.
7 For the sake of consistency, whenever data expressed in base years earlier than 1987 were unavailable, the designation "NA" is retained in Table 1B even though such data are available in terms of 1987 prices and expenditure weights.
8 The recovery from the 1980 recession lasted only four quarters, during which real GNP rose 3.3 percent from its recession trough. During the recovery from the 1980 recession, it took two quarters for the economy to surpass its previous peak output.
9 Moreover, the average time taken in the previous recoveries (either including or excluding the recovery from the 1980 recession) to surpass previous peak GNP was two quarters. During the current economic expansion, it took five quarters to surpass the prerecession peak GNP.
10 This index was first published in 1970.
11 For the sake of consistency, whenever data expressed in base years earlier than 1987 were unavailable, the designation "NA" is retained in Table 2B even though such data are available in terms of 1987 prices and expenditure weights.
12 Expansion during the four quarters of the aborted recovery from the 1980 recession remains 3.3 percent in terms of 1987 prices and weights, however. In addition, the period between the trough of the 1980 recession and the surpassing of previous peak GNP remains two quarters.
Michael Ulan is an international economist with the U.S. Department of State, Washington, DC. The views expressed are the authors's and not necessarily those of the Department of State or the United States Government. The author wishes to thank Robin King, William G. Dewald and G. Paul Balabanis for helpful comments on an earlier draft of this paper.
|Printer friendly Cite/link Email Feedback|
|Date:||Apr 1, 1994|
|Previous Article:||Challenges to American competitiveness: a Japanese perspective.|
|Next Article:||Ockam's Razor, economics and business client.|