sign in sign up
<<
Home , Inflation, November 1973

This PDF is a selection from an out-of-print volume from the National Bureau of Economic Research

Volume Title: Explorations in Economic Research, Volume 2, number 2

Volume Author/Editor: NBER

Volume Publisher: NBER

Volume URL: http://www.nber.org/books/moor75-2

Publication Date: 1975

Chapter Title: Slowdowns, , and : Some Issues and Answers

Chapter Author: Geoffrey H. Moore

Chapter URL: http://www.nber.org/chapters/c7407

Chapter pages in book: (p. 1 - 42) 1

GEOFFREY H. MOORE National Bureau of Economic Research and Hoover Institution, Stanford University

Slowdowns, Recessions, and Inflation: Some Issues and Answers

ABSTRACT:According to the criteria used for many years by the National Bureau of Economic Research to identify business cycles, the United States entered uport its sixth since World War IIin )73. about the depth, duration, and scope of the decline persisted for many months thereafter, partly because some indicators such as real GNP stumped rapidly and continuously during 1974, while others, such as total employment and industrial produc- tion, did not. Sharp deterioration in virtually all major indicators after October 1974 clinched the matter. A system of monthly measurements designed to compare art incipient recession or slowdown as it develops with earlier recessionary periods is described and illustrated. An analysis of slowdowns and speedups in reveals that they have invariably been associated with reductions and advances in the of inflation. The factors responsible for the reductions appear, however, to have become less effective in recent years and to operate with a longer lag.

ACKNOWLEDGMENT I am indebted to Walter Ebanks for development of the of physical volume indicators and for other assistance during the couae of the project reported here. I am grateful also to charlotte Boschan, Otto Eckstein. Solomon Fabricant, Harold Hatcrow, John Meyer,nd Edward Smith for their useful comments and suggestions, to Felix Anderson and N. Irving Forman for the charts, to Mildred courtney and James Hayes for their patient handling of successive drafts of the manuscript, and to Jane Forman for editing the final one.

NOTE:This paper was prepared at the end of , when the latest data available for most economic indicators pertained to November. Rather than update the text and tables a new section VII has been added discussing more recent data. 125 G'1irey H.More

W INTRODUCTIONAND SUMMARY Despite all the talk during 1974 on whetherthe nation recession, rather little was or wasnot in a attention was paidto the question in the currentslowdown was whatrespect more serious or lessserious than those past, and by howmuch. This question in the especially for obviously ismore iniportant, policy-makingpurposes, than the much whether the slowdown debated pointof should be designateda recession. An more important questionis essential if answer to the be given to the a reasonably objective terminologicalone. answer is to Furthermore, theanswer to the question bearing upon about relativeseverity has another matter ofgreat concernthe a Previous experience rate of inflation. has somethingimportant to tell between slowdowns, us about therelation recessions, andinflation. Since theautumn of1 973 a systematic comparison of the slowdown withearlier ones has current been conductedon a continuing This isan extension of similar basis. studies madeduring the 1969-1970, as wellas during the recessions recession of 1 958.' In capsule of 1960-1961and 1957- form the principalfindings are: A slowdown in the rate ofeconomic growth 1973. In lateautumn of 1973 the began in thespring of the physical slowdown becamean actual decline volume ofaggregate economic in 1974 the decline activity. Until theautumn of was relatively modestby some more severe byothers. Beginning measures of activity,but in theautumn, all major indicators declinedsharply. physical Indicators showinga modest decline recessions duringthe first three in comparisonwith previous industrial production, quarters of 1974 the number of include theindex o they worked. persons employed,and the total A newcomposite index hours expressed in based entirely physical unitsalso showed upon indicators through October1 974. Indicators a relativelymodest decline earlier date showing relatively include the Shari) declinesat an sales in gross nationalproduct in constant dollars. constant dollars andretail Because of thediscrepancy upon physical units between and those measures of activitybased gates for obtained byadjusting changes, itwas unusually current dollaraggre- determine whether difficult, during the current most of 1974,to previous decline didor did not exceed recessions.However, in the the mildest of autumn of 1 974 developed inmost "leading further weakness other indicators," in measures of the employment, andin virtually all the decline physical volumeof aggregate widened. These activity. Thescope of year, to conclude developmentsmade it that the reasonable, latein the periodwas sufficiently similar in depthof I

400re Slowdowns, Recessions, and Inflation 127

decline, in duration and in scope, to warrant designating it a contraction, or recession. The business cycle peak date that seems best supported by tile in a evidence on the physical volume of economic activity is November 1973. .pect This date was initially selected on a provisional basis early in 1974. the Subsequent evidence generally confirms the original choice. Sonic meas- Lint, ures of activity, notably retail sales in constant dollars and the rate of of , reached their highs (lows in the case of unemployment) 'the before November 1973 while others, notably employment, reached their is to highs later, but the concensus centers on November or the fourth quarter of 1973. The selection of this date, even though the subsequent level of is a activity was influenced in part by the imposition and removal of the oil ion, embargo, is consistent with earlier practice in determining turning points in tion business cycles. The high rate of inflation that persisted during the 1973-1974 rent slowdown and recession had two unusual consequences. One is that it sis. brought about a continued rise in measures of activity, such as gross of national product, that are expressed in current or nominal dollars, even .7 while measures of physical activity were declining. A second is thatit created uncertainty about the accuracy of the procedures used to eliminate the effect of price changes oisales and inventories, and hence on the of measures of output derived in this manner. in 6. Since 1948 slowdowns in economic growth have invariably been of accompanied by reductions in the rate of inflation, while recoveries in Jut growth have been accompanied by a speedup in inflation. The factors cal responsible for the reductions appear, however, to have become less effective in recent years and to operate with a longer lag. This tendency 'us seems to be responsible, at least in part, for the persistence of high and of rising rates of inflation during the 1973-1974 slowdown and recession. irs Nevertheless, the peak rate of inflation appeared to have been reached in rs the autumn of 1974, and a decline in accord with historical experience ne seemed to have begun. an ai I liii COMPARISONS WITH EARLIER RECESSIONS e- The main idea behind the present study has been to measure the changes 0 in specific indicators, month by month, as new figures become available, of and compare them with corresponding measurements taken at comparable ss points of time in previous recessions or slowdowns. In this way the relative II severity of declines during the current period can be determined, and other similarities or differences in the character of the current and earlier periods e can be analyzed. Table 1 and Chart 1 show how this has been done for one broad index TABLE 1

Ranking of Seven BusinessMonths After Index of Five Coincident Periods of Indicators, Recession in Deflated Successive Months 19974: CyclePeak 1973Nov. 1969Nov. 1967Jan. Business Cycle 1960May Peak Date' 1957July 3953July Nov. 1948 2 Per Cent Change from Recess on sAverage, Five 43 76 5 8 1110 9 -2.1-1.9 (Feb)-0.7 (jan.) DeC( 12 -2.0 (Mar.) 79(Apr) -2.1-2.3 -fuR') 1.9(June) Ma) --0.3 -3.3-2.9ISc;)t) (Oct.)-2.5 (Aug.) (Nc)) -1.7- 1.6 -22-1.5 -2.7-2.5-2.5 +0.3 0.0 -0.7 Business Cycle Peakb -0.7 +1.9+1.5+0.9 08 -5 -35-2.72-15-1.5 -7.6-5.6-1.8-1.7 -2.2 -3.2 Q4i +6.2±4.6±2.4+3.1 -3.0-4.2-52 5,7r - 10,8-(2.1-1.6-11,0 -2.8 -32 Omitting -8.1 -6.6-Q 1967 and 1 -04 973 -109-11.5-11.3 -3Q -9.1--8,-8.3 -11.3 -123 ._9 -6.2-3 --2.46--(.8 -7.9-7 1 -8.7-7.8-83 0 with UltimateSeverity,Correlation Six 2 5 4(4)3(3) Ranking According to Size of Per Cent Declinec 2(2)1(1) 2(2)5(4.5) 3(3)1(1) 6(5)5(4.5) 7(6) Omitting Recessions +.53+.94 1973d 453 453 2(2)33)3(3) 1(1) 4(3)2(2) 4(4)5(4) 6(5)7(6) 6(5)7(6) +1.00 +89 867 2 3(2) 1(1)1(1) 4(3)4(3) 6.5(5.5) 7(6)5(4) 6.5(5.5) 6(5)7(6) 5(4)6(5) +71+75+89 10 9 2 3(2)2(2) 1(1>1(1) 4(3)3(2) 5(4)7(6) 6(5) 7(6)5(4) +1.00 +71+.71 1112 2 4(3) (3) Ranking of Six Recessions According to Ultimate Severity' (1) (2) (5) (6) (4) +82 NOTE: Dates for 1948-1969 are from the National Bureau's business cycle chronology exceptmanufacturingThe index for 1 of 967, five and which coincident is sales.inctuded indicatcs The as index an includes is published nonfarre as employment,Series 825 in Businessunemployment Conditions rate hnverted),Digest, U.S. industrial Department production, of Commerce. (3) (1) (2) (4) example of(5) a "minirecession." The November 1973 deflated . deflated.(6) °Abper rank centparenthesespeak ofNovember, changes1 ismeans provisional arethatomit and computed the the Decemberand x'rcenzage 1973-1974 is usedon the 1973) declineonly baseperiod foris ofis141.4 the and thethe purpose three-month smallesthence(1967 are =of amongtOO). testing comparable average the the wven centeredhypothesis with recessions the on ranks thatthe over business a of business ultimatethe corresponding cycle cycleseverity peak peak month. (bottom inte'aI; may The nave aline). base been for the current rank of 7 means that it is the largest. Ranks shownreached. in period (average or October, 5using'Business ranks1948-1949,nonfarmUltimate showncycle trough in severity employment,parentheses. -17.5. date. is measuredIn For real six GNP,by ranks the dcl. perthe centcorrelationThese decline per coefficient centfrom highdeclines shoudto low are: exceedin a 1969-1970, trerid-adiusted 0.83 to be -10.2; sigitilicant composite 1966-1967, at index based -3.0; on1960-196!. 12 -8.5: 1 953-1954 the trough is it the 13th month, August 1954. The trough of the the t0.05 967 level."min-recession" is in the 4th month, May 1967. physical volume aggregates (e.g.. industrial production. 1957-1958. -155; 953-1954, -1 7.2; CHART 1 Recession Patterns:Index of Coincident Indicators, Deflated

-15 i1' I 03 6 9 i1' LLL-15 1. 12 15 1821 24 0 L1LL11_LI_LLkd' 3 6 9 12 iii 15 18 21 24

:

-15 I 21 24 0 I_jiiij.IjijI I II 3 69 12 Months otter peak 15 1821 of the physicalvolume of the economic commonly used activity. The monthly index includesmost of employment, aggregates: industrial the production, the rate is unemploymentrate (taken nonfarm recordedas an improvement inversely,so that a decline deflated by in a and viceversa), personal sales forconsumer income (manufacturers', expenditures,and total wholesalers',and retailers') business deflated byappropriate Slowdowns, Recessions, and Inflation 131

price indexes. They are cornbned in this index in such a way that each of the live components has an equal chance, on the average, to influence the movements of the index (otherwise one component that ordinarily moves in large swings, such as business sales, could swamp another one that typically moves in a narrower range, say nonfarm employment). 4 Each of the past recessions has been dated in earlier studies by the National Bureau and the chronology has been generally accepted as reasonably accurate, both as to the time the recessions began and ended and as to whether the designated periods were recessions and other periods were not.2 One marginal case is included in the table, namely 24 1966-1967. This has not been designated a recession because it was so mild and brief, but it is included in the table for comparative purposes as one episode that failed to meet the criteria for recessions. It was a period of

1 slowdown and earned the name "mini-recession" at the time. In some 4 other countries, such as West Germany, it was more serious. The date for the beginning of the current period, November 1973, was designated provisionally at the time this study was begun and did not imply any conclusion or forecast with regard to whether the ensuing period would be a recession or not, It was selected early in 1974 as a possible business cycle peak date on the basis of evidence then available, but it was recognized that subsequent evidence might either shift the date or support a conclusion of no recession. For example, it is now known that nonfarm 24 employment, one of the series that obviously has some bearing on the existence and dating of a recession, continued to rise slowly but rather steadily until October 1974. Nevertheless, the experiment began with the November 1973 peak date and, as will be seen, the evidence now available confirms this date as a reasonable choice. I The table then records the percentage changes in the index from the several peak dates. Note that the dates are not necessarily the same as the I peak dates in the index itself, although in this particular index the devia- tions are not great (the index peaked one to four mor.ths earlier at each turn except in 1967, where it did not decline at all, and in November 1973, where its peak coincided with the selected date). In the first month I or two not much can be said, because changes over one or two months in 4 any series are likely to be erratic and can be dominated by factors such as strikes or bad weather. Butitisclear that in each of the previous recessions, with the exception of 1966-1967, the index declined as time I went on. The average for the five recessions, omitting 1966-1 967, shows a drop of nearly 2'/2 per cent in three months and 6 per cent in six months. From November 1973 to February 1974 the index dropped rapidly, i.e., at about the average rate, as the energy crisis hit the . After February it held steady through July, then it began to decline again. The 3'/2 per cent decline in the eleven months from November 1973 to October 1974 was

C 3

I I I 132

less than half as large as theaverage decline in the five over the same interval,and smaller than prevwus r&'c'ri> The ranking of in any one ofthem. the severalrc t'ssiofls bottom panel further interesting in the tableshows result. The ranks haveusually remained a after the first monthor IWO. Moreover, the relativelystable ranks insucc scve monthS POSItIVelY correlatedwith the ultimate are be determined ranks. The lattercan of Course after the recessionis over and Only third and fourth ail upswing hasbegun. In months, in fact, theranks (or the the periods (now previous six including 1966-1967)were exactly the recessionary ranks. Thiswas rather fortuitous, same as theultimate however, since inthe next few the correlation months deteriorated. Nevertheless,there isa positive throughout (see theright-hand column correlation ordinarily in the table).The point one can getsome rough due is that recession from how to the ultimate severe it appears in severity ofa indications the first fewmonths. The are, of course, subjectto correction initial The currentperiod, on this as time goeson. basis, startedout with a relatively dine, and rankedfifth in the first sharp rk'- was clearly attributable, three months(through February at least in part,to the oil embargo. 1974). This were greater thanin any of the The declines 1949. Then the recessions except position improved, 1953-l954 and1948- and fromMay through current period rankedsecond, i.e., October the any of the other worse than recessions. However, 1966-1967 butmilder than October 1974, the percentage the eleventhmonth, declinesthrough 1970, one of were very close the mildestof the five to those for1969-. These results recessions since are compared World WarIi. measures of the with thosebased physical volume upon otheravailable current declines of economicactivity in Table in GNP andin retail sales, 2. The i.e.,. after both expressed for pricechanges, in realterms, recession periods. are larger than But this is in most ofthe earlier where the not true of theindustrial decline throughNovember production index, recession periodexcept the was smaller thanin any "mini-recession" of previous ment, i.e., thenumber 1 967.Nonfarrn in November on payrolls ofnonfarm employ- 1974 thanat the tentative establishments,was higher whereas it business cycle was lower afterthe first peak ayear earlier, (except, again, year in each ofthe previous the 1967 recessions basedupon a separate "mini-recession").Total civilian The survey of , employment, unemployment showed than in rate in November a similar ranking. several earlier (6.5 per cent)was at a higher preceding recessions, butthe increase level twelve months(1.8 in the rateduring the increase inany previous percentagepoints) was the rate, recession smaller thanits incidentally,is highly (except, again,1967). The ings of the correlated with increase in recessions; the the ultimate It appears, level is not. severity rank- then, thatthe measures ofactivity that are estimated interms Slowdowns, Recessions, and In1Iaton 133

ns of current dollars (fromreported sales, inventories, exports, imports, etc.), and then deflated by price indexes, presented a decidedly less favorable a picture relative to previous experience in the current period than did the le measures of physical activity obtained directly.Itisof some , re therefore, to examine a composite index of physical activity that is based entirely on measures that are not obtained by deflating data for price e changes. Walter Ebanks has constructed such an index using five compo- y nents, namely, index of industrial production, manhours of nonfarrn e employment, unemployment rate (inverted), railroad freight carloadings, S and tonnage of shipments by truck. The rriethod of construction is the same n as that used in the index shown in Table 1, so the average rate of change in this index is also 1per cent per month.6 1 Between November 1973 and September 1974 this index declined 4.9 per cent (see Table 2). This is somewhat smaller than its decline over the first 10 months of the 1969-1970 recession, and decidedly smaller than its decline during the recessions of 1960-1961, 1957-1958, 1953-1954, and 1948-1949. The above measurements represent only one way of comparing current economic changes with previous experience. They concentrate attention on their size relative to changes over corresponding intervals in previous recessions. They do not show readily how the current decline in any indicator to date compares with its total decline in previous recessions. This is of little interest early in a recessionary period, hut becomes more relevant as time goes on. Furthermore. the criteria used by the NBER in identifying business cycle contractions include not only their depth, but also how long the declines last and how widely diffused they are among different industries or other economic sectors. Table 3 gives a conspectus of such measurements back to 1920. The extraordinary depth, duration, and diffusion of the 1929-1 932 contraction stands out, as well as the severity of the 1920-1921 and 1937-1938 contractions. None of the recessions since World War II have approached these magnitudes. Among the milder recessions in the past half-century are those of 1926-1927, 1960-1961, and 1969-1970. The entries in the column headed November 1973 are not, of course, final, since the declines are recorded only as they stood at the time the table was constructed (December 1974), not as they may eventually become (see Section VII for a later version). The conflicting nature of the evidence on the current decline is nevertheless apparent once again. In terms of real GNP, the 2.7 per cent decline in the first three quarters of 1974 exceeded the total drop registered in several earlier recessions, in both duration and magnitude. On the other hand, the corresponding decline in industrial production was smaller than in any previous reces- sion. Nonfarm employment rose until September 1974; by November it TABLE 2 Ranking of Seven Periods of Recession After 9-I 2 Months, 1948-1974 Indica tOrP Months dfter Nov. Nov. Business Cycle Peak Jan. May Date 1957July 1953July 1945Nov. IndustrialRetailGNP sales product)on in dollars consta,t deflated Per Cent Change from Business Average, FiveRecessions Nonfarrn employm 12 9 -3.-4.3-2.7 (Oct.) (3rd q. (Nov.) +0.8+0,3 +1.3+ 1.6 -1.4 Cycle Peak +2.0-3.2 -0.8 3.4 Omftting 1967 and 1 973 -1.7 Unemploymentlotal employment rate,householdpayro/l survey survey 1212 +0.1+0.7 (No) (Nov.) -1.3-7.1 0.0 +2.3+3.6 - 1.3-1.6 -,y-3.0 -7.8 -4.2-6.2 -2-6.1-0.4 8 Index,Unemployment 5 COinc/dent ratechangelevel 1212 1.8/Novj6.5 (Nov.) 225.8 ±2.0 -0.9 7.1 -2.3 -2.8 -0.9 6.4 - 6.5 Index S phy'.jc V(ItJfl-<. Indpcatc)rcrtdicators, deflated 70 -3.3 (Oct.) --6.7 -4.2+1.9 --173-10.8 -10.9 +3.2 -72- 12.3 +2.6 -11.7 -8.7 2.6 - l3.3 12.5 Ranking According to Size of Per Cent Decline Omittingwith 1973UltimateSeverity,CorrelationRecessions Six GNPIndustrialRetail in constant sales, production deflateddollars 1112 9 365 4(4)2(2)5(4) 3(3)1(1) 2(2) 4(4)5(5) 7(6) 6(5)2(2) 4(3)3(3)1(1) .49 TotalNonfarm employment, employment. householdpayroll survey survey 1212 22 3.5(2.5) 3(2) 1(1)'(1) 3,5(2.5) 4(3) 6(5)6(5) 7(6)5(4) 5(4)7(6) .93+77 Unemployment rate,levelchange 12 25 2.5(2.5) 4(3) 1(1) 3(2)6(3) 7(6) 2.5(2.5) 6(5) 4(4)5(4) +33+77 Index,Index, 5 coincident5 physical volumeindicators, indicators deflated 1110 2 3(2)4(3) 1(1)1(1) 4(3)3(2) 7(6)5(4) 6(5) 7(6)5(4) +1.00 NOTE: For explanation of entries see notes to Table 1. (3) Ranking According to Ultimate(1) Severity (2) (4) (5) (6) TABLE 3 PeakSelected to Trough, Measures 1920 101974 of Duration, Depth, and Diffusion of Business Cycle Contractions 1920JulyJan. 1923JulyMay Nov.1926Oct. Mar.Peak1929Aug. and Trough 1937JuneMay 1945Oct.Feb.of Business Cycle Oct.1948Nov. 1953Aug. JulyContractions 1957Apr.July Feb.1960May 1969Nov. 1973 Nov. Business cycle chronology Indicator19211 18 14 13 43 1938 13 Duratic)n (months; 1954 1958 1961 1970Nov. na. Nonfjrm empioymenlNurifarmIndustrialGNP,IncreaseGNP, Masjmjpi constant employmentproduction currentdolljrs dollars na.11.gd n.a.14 5.5 14 6 n.j. 4.4 12 83 -31.6 363642 -10.8 12 69 -10.1 n.a 68 -5.2 12ii 6 -.4 121213 -4. 69 -2 2 12 6 -1 6 1512 Nonfarm inductric.s;CNP, urrent dollars na. n.a. +2.4d n.a. +22.o 25.2 +90 20.0 11 Diffusion (per cent)Depth" (per +3,4 22274. 4313j5 cent).9 +36 16ô.1 8 +3.8 147 10132 3 1 1426 17 8 +1.9-06 6 512 2 UnemploymentGNP,ndustrjjl wnstantMaximum rate; productiondollars Percentagewith declining -32.497 na. -17.995 -4.1-49 71 -7.p-2.0-3.0 100-53A-32.6-49.6 -32.4-13.297-16,2 n..i.- 38.3-11.9 na. 9Q -g.9-1.9-3.4 -10.0 -3.-1.9 -14.3 -3.9-2.6 82 -7.2-1.6-0.3 83 -8 -11 5 60 -4. U SOURCE: U.S.version1972, Department of pp. this 100-110. of Commerce U.S. Department table, see Solomon Fabricant, of Labor, BoardThe of Governors of the Recession of 1969-1970, Federalfl Reee System NationalVictor Bureau Zarnowitz, of ed., The BUSiness Cyc/e Today, NewEconomic York, NBER, Research for a (uller bSince 1946.PercengeAugustmaximum based change1974on figurechanges means NOTE:from is thein thatthe employment highest 18specific of thc for peak 30overany month na. indicates not available. industries had declines in employment. alter seasonalmonth or quarter in the series to the during the contraction and the increasessix-month are spans in 30 nonagricultural industries, trough month or quai'ter, over the inteal shown adlustment, over the six monthsfrom the May lowest to month to the highest. in centered on the fouh month ot the span. Hence Novemberabove. 1974, For Priorthe unemployment to 1948 based rate,on the percentage points. the 60 pr cent dated CN0'Entries decline.cyclical are1926-1928, based changes heon maximumdata1929-t933. in employment available figures through arein 41 annual averages (monthly industries. November 1974 and hence show only the data not avaiiablel for 1921, 1924. 1 928, decline in the series to the lowest point and 1933: increases, in percentage points, are reached to date. for 1919-1921 1923-1924. (eIfrey J.

was six-tenths ofI per cent below the September level, a small compared with its full decline in previous recessions. The drop unenipJ0. rate had reached a level exceeding the maximum levelreached earlier recessions, but the increase in the rate to November t from its previous low in was less than its totalrise in recession! any earlie, Finally, the relatively narrow scope of thecurrent decline through November 1974 is recorded in the fact that theproportion of industries with declines in employment had not exceeded60 per cent, whereas j most previous recessions this percentage had climbedto 80 or 90 Until the autumn of 1974 relatively few industries per cent hadexperienced tions in employment. In this respect, the situation reduc. resembled thatduring the 1967 mini-recession, when thepercentage of industries with declining employment (over six month spans)rose to 62 per cent, but no higher As of December 1974, therefore, theevidence concerning severity of the current decline in the relatie aggregate economic activitywas conflict. ing. Although unemployment hadrisen to recession levels, the increasein unemployment was smaller than in precedingrecessions. The decline industrial production and in thenumber of persons in modest. A composite employed also measure of activity based uponphysical units requirenoadjustmentforprice which changes,showeda declineap- proachingbut less thanthat in themildest of the Postwarrecessions Only those aggregatemeasures constructed from dollar price changes, such valuec deflatedfor as real GNP, showed declinesthat approachj severity of the worst of therecessions since 1948. the In view of the rapidity of price increases in1 974, and the being sure that theprices that are implicit in difficulty of the reported valuedata are the same as, or at leastare well represented by, available price indexes, those containedin the it seems possible thatthe deflated value were unduly depressed.Under more normal figures in output are usually conditions, when theswings far larger than inprices, errors in the process are of no great price deflation consequence. For example, reflected in salesare set by contract when the some months before, itmay make little difference to anestimate of the physical one is not sure about volume representedby the sales if the advance dating, provided pricesare not radically changing. But whenthey are changing the contract rapidly, knowledgeof the length of period and knowledgeof whether escalated may he of or not the contract price is critical importanceSimilarly physical volumeof inventories_required the estimation of the exceptionally difficult to derive estiniatesof outputis when pricesare changing rapidly valuing inventoriesmay also be changing. and methods of these The uncertaintysurrounding estimates is indicatedby the figures for1973 and early enornloijs revisions in theinventory rate of inventory 1974, whichmore than doubled the accumulation. estimat 139 nore Slowdowns, Recessions, and Inflation drop These may take years tountangle. But they point to the lent iieed for basic improvements in the datathemselves, as well as in our ways could eral of monitoring discrepancies whenthey arise. Such improvements its bring lasting benefits to our systemof economic intelligence and prevent rijer substantial inconsistencies of the kindthat developed during 1 974.

ugh tries S in IN THE CONTEXT OF ent. IHI! THE CURRENT SLOWDOWN Juc- GROWTH CYCIES ring In recent years some of the businesscycle studies at the National Bureau !ng have been directed toward implementing aconcept of the business cycle er. that differs in important respects fromthe one employed above. The new concept, termed the growth cycle, isperhaps more pertinent to themilder industrial Ct- type of economic fluctuationthat the United States and other fl countries have been experiencing sinceWorld War II. Ilse Mintz began the in work by developing a growth cyclechronology for West Germany, 1950 as 1967, and more recently for theUnited States, 1948-1970. Still more Icil recently, Philip A. Klein constructed asimilar chronology for the United ap- Kingdom, l95D-1972. The research oninternational economic indicators flS. begun by the NBER last year will makeextensive use of the gtowtll cycle or concept for all the industrial countriesincluded in the study, by applying a the standard set of techniques to comparabledata for each country. Growth cycles are alternating periodsof slow and rapid economic of growth. They differ in two major respectsfrom the business cycles hereto- he fore identified by the NBER. First, aslowdown may or may not encompass decline in aggregate he a business cyclecontraction, i.e., a period of actual prior to an actual es economic activity. Second, a slowdown maystart Dr. gs downturn in aggregate economicactivity, and end after the upturn. on Mintz's study for the United Statesillustrates both types of difference. Her and es growth cycle chronology includeseight slowdowns between 1948 le 1970, whereas there are only fivebusiness cycle contractions (or reces- slowdowns encompass the five recessions,starting 11 sions). Five of the eight the same time. The fly a few months earlierin each case but ending at about of other three slowdownsin 1951-1952,1962-1963, and 1966-1967 entail a sustained decline in IS interrupted expansion phases but did not of growth for periods of a year or e activity, merely a markedly slower rate chronologies and how IS more. Chart 2 showsthe relation between these two of the indexes used in this study. of they fit the experience recorded by two 1973, bringing to an end theperiod g The ninth slowdown began early in peak date of rapid growth that begantoward the close of 1970. The y basis, the first quarter. Like the d tentatively is , or on a quarterly earlier slowdowns, this one beganwell before any decline in aggregate CHART 2 Business Cydes and Growth Cycles, 1948i 974

usiness Cycle ChrOrsalOQY pT Fl PT p,, J Ape May Feb 4o?to P '404 DeflAted led nee

60 140 '20 '00 30 -. '0 60 o 70 - 60 50

I 1 'S48'50 '52 '54 '56 '58 '60 '62 64 '66 '68 '70 30 '72 ' p Grow?6 Cycle ChronOl0y prp 7 p Oc' ,n Jr Mo? tW3 1 pIP I saMar FeFee PT P 1 30p '- r MIr Jar, Oct !M, p T '40a Mar 000 Defloied eodnnde* 43 1

.1 601-- -1160 -

-s Defloled coincidetitindee

s -M23

ae - Js0 60S

Rato -'440 948 I LaJ '53 '52 '54 '56 IS I I '58 '60 -P--I I '62 '64 '66 II 430 '66 '70 '72 NOTE:In uppor parid. shdd '74 cScie a'eas arebusiness c 0'.rJ0?%T1Both pafleI certlain the same essntraUic,ny. In knr-rp,nsl shad'd indc'c-s. acs'ab arc' grsaih

activity became apparent, antedating business cyclepeak of by &ghtmonths the The current November1973 discussed designated slowdowncan be compared a5ove, the same with the eight technique usedabove to earlier ones by Table 4 illustrates compare business the methodfor the cycle contractions. October and same series used November 1974 in Table 2.Data for current slowdown, represent the1 9th and respectively.By that 20th monthsof the P resented in thetable time all of were more the indicatorsrep- (I in any ofthe three depressed thanat the corresponding milder P recessions. With slowdowns,i.e., those date respect to the that didnot encompass five slowdownsthat did encompass reces- pr 141 Slowdowns, Recessions, and IniItion

sions, the current slowdown appeared mildwhen measured by industrial production or employment or by the rise in unemploymenthut serious when measured by real GNP, deflated retail sales, orthe level of un- employment. 60 Since tables like Table 4 could be and wereconstructed as soon as a tentative date was established for the start of the currentslowdown, and 20 position of the 0 updated month by month thereafter, the comparative 0,0 current slowdown could he monitoredcontinuously. Since the peak dates 00 0 u of the slowdowns precede the business cycle peaks,by intervals ranging 0 from 3 to 8 months, earlier recognition of the relativeseverity of a current 0 slowdown may be possible. Also, the wider range of experiencecovered by the growth cycle chronology is an advantage.For example, it became clear early in 1974 that the 1973-1974 slowdown was moreserious than the three minor slowdowns of 1951-1952, 1962-1963,and 1966-1967. On the other hand, for many months thereafter,the evidence was conflict- ing as to the position of the current slowdown amongthe five previous slowdowns that encompassed recessions. Measuresof activity based upon physical units registered relatively slight declines,whereas measures based upon dollar values deflated for pricechanges registered substantial de- dines. The possible reasons forthis anomaly have been discussed above. It should be noted that none of the previousslowdowns lasted more than 20 months, and most were within the rangeof a year to a year and a half in length. Hence, the 20 month period fromMarch 1973 to November 1 974 is at the long end of the range. In mostof the previous slowdowns the 0 various measures of aggregate economic activityby the 20th month had begun to register upturns, but such a developmenthad not become apparent by the 20th month of the currentslowdown. This is another way to judge the relative severity of a currentslowdown.

[WI LEADING INDICATORSIN 1973-1974 The preceding discussion has concentrated upon measurespertaining to the "real" economy. It was here, of course, thatthe evidence of recession could first be observed. Aggregates expressed incurrent dollars, such as GNP, total business sales, and personal incomecontinued to rise during 1974, though at a somewhat slower pacethan in 1973. General indexes of prices, , and unit labor costs rose at anunprecedented pace. The divergence between the current dollar aggregatesand measures of the physical volume of activity is one feature thatdistinguishes the current decline from most previous recessions, atleast prior to 1969. In most previous recessions both nominal andreal aggregates have declined at TARLE4 Ranking of Nine Months Slowdow0PeriodsCurrent of Economic GrowthWithout Slowdowns Recession5 Slowdown After 18-20 Months, 1948-1974 Indicator PeakAfter March1973 1966June April1962 1951June Peak Dates 1969Mar. G roy. ih Slowdownc1960Feb. 1957Feb. With Rec,scjo0c1953Mar. 1948July A "era go. D IndustrialRetailGNP salesproduction in COnStant defl,teddollars 201918 -7.6-1 (Oct.).2(Q3) +4.0 Per Cent Change +4.8+6.5 + 11 3 +6.2 from Growth +0.7 Cycle Peak42.3 -0.6 - 1.2 4 2 SlowdownsRecess on '5' it h Nonfarm employmentpayroll surve, 20 -1.4 (Nov.) +6.0 9.O +123 -6.6-04 + 0.0 , -5.0-1.5 0.0 87 '-1.1-1.4 UnemploymentIota 1 employment rate,levelhousehold su'ev 20 20 6.52.5 (Nov3.0 (Nov.) (Nov) +3.5 2.53.4 +3.2+5 0 +03 +1,)-0.1 -1.0-3.0 -2.3-2.5 -2.6-06 -1 +15 5 Index,IndexUnemployment 5 5 CoinCident rate,changendicatc,rc deflated 1920 '-0.3 (Qct) 1.5 (Nov 3.8 0.0 -0.1'-6.5 -0.5 2.6 +2.41-5,8 6.5 +2.8 6.7 -2.7 -L7 3 -0.5 6 I I, volume ndicatr,rs 18 -0.2 (Sept.) +100+2.8 +66'-92 + 14.2+72 3 -4,1+0.1 -14.9 - 13.6 - 12.9 --26 2 -102 -5+24 1 0.2 r Correlation Ultimate with Ranking According to Size of Per Cent Declineb SlowdownSeverity, 8 5C IndustrialGNPRetail in constant sales, production deflateddollars 201918 8.569 3(3)4(4) 2(2)3(3)1(1) 2(2)1(1) 9(8)7(7)6(6) 5.5(5.5) 5(4)5(5) 7)7)8(7)8(81 8.5(8)5.5(5.5) 7(6) 6(5)2(2)4(4 +28+74 69 TotalNonfarm employment employmenthouseholdpayroll survey survey 20 43.5 2(2)1(1) 3.5(3) 313) 2(2)1(1) 5(4)5(4) 6(5) 9(8)7(6) 9(8)7(6) 8(7)8(7) +81 88 UnemploymentNOTE: rate.rate,levelchange Peakdatesfor 1948-1 969 are from use Min, "Dating United StatesGrowth Cycles," 20 57.5 2(2)3(3) 4(4)2(21 111)1(1) Exp!oration in Economic Research, Summer 1974. 6)515(5) 7.5(7) 4)4) 9(8) 8(7)3(3)Table 8, and peOain othe concepto) 7(6)6(6) +45+83 Index,Index,5A 5 rankcoincident 5 physical ofdecline,Per 1 means cent Ranks changesindicators,that the shown percentage are deflated computed in parentheses decline on is the the omit base smallest the ot 1973-1974the (Or three-month rise the period largest) average and among deflated deviation cycles:' The March 1973 peak is tentative. 19 4 2(2) 3(3)hence are comparable withcentered the ranks on theof ultimate growth cycleseverity peak (bottom month. the1(1) nine slowdowns over the correspondflg interval; a rate of 9 means 6(5) 5(415)4) 9(8)8(7) 0.05 level. 8(7)9(8) line). 7(6) that 1 is the largesS +.83 dultimate nonfarmUsingseverity1957-1958, ranks semployment,volume measured shown -15.5; indicators in by parentheses. real 1953-1the GNP,per cent954, etc.). Fordecline -17.2; Theseeight from1948-1949, ranks per high cent the to correlation declineslow -17.5. in a are: 18 4 3(3) (2) coefticient should exceed 0.64 to 1966-1970.be 2(2)significant -3.0: at th (flc1962-1trend-adjusted 963, -2.8: composite 1951 -1952. index -3.8;based 1969-1970. on 12 physical volume aggregatesRanking e.g., According to Ultimate Severityd1(1) (3) 6)5) (5) (4) (6) (7) -10.2; 1960-1961 (81industrial produsiion. -65; I.

bore Slowdowns, Recessions, and Inflation 145

"fles Commerce and the National Bureau have likewise experimented with 'ices various deflated indexes. All of them showed little or no increase after real midI 973 and declined during 1974. One such index is analyzed in Table and 5, after the manner used in Table 1. (?for Through this deflated leading index declined more sharply than in the mini-recession of 1967, but less sharply than in the 1969-1970 any recession or any o the earlier ones. Alter July the picture changed rapidly d j for the worse. The September 1974 index was lower, relative to its level at lars. the tentative business cycle peak in November 1973, than in any three of wifl the five previous recessions. The October 1974 index was lower than four e in of the five and equal to the decline in the 1948-1949 recession. Further- 5ig more, in all the earlier recessions except 1969-1970 the index hadalready the begun to rise (see Chart 3). an The deterioration in the relative ranking of this index in Autumn 1 974 the was unusually sharp. In most earlier recessions the ranking after thefirst uc- two or three months was maintained quite steadily throughout, although in n. 1969-1 970 there was a similar deterioration when the General Motors tly strike took place (Autumn 1 970). The causes of the sudden shift need further art study: the unprecedented increase in interest rates earlier in the year, the r,s sharp drop in the growth of the supply during the summer, the shift in presidential administrations, the continued rapid rise in prices, and the accompanying deterioration in consumer and business confidence are a! among the candidates for such a review. Butwhatever the causes, the effects g were registered in virtually all the leadingindicators. Indeed, by September 1974 nearly all leading indicators expressed in If current dollars had begun to decline. At the same time, the downturns in the nonmonetary leading indicators that had started much earlier con- tinued. Table 6 contains the record of when the highs in the twelve leading d indicators were reached. Relative to the designated business cycle peak of November 1973, which is based on the behavior of physical volume or constant dollar data(seeSectionVI,allfour of the nonmonetary leaders and two of the current dollar leaders exhibited leads. The only leaders that did not give early warnings in this period are those expressed in current dollars, but for them, of course, the appropriate comparison is with the peak in the current dollar volume of economic activity. At this writing, that stilllies in the future.

IVl DATING THE BUSINESS CYCLE PEAK In view of (a) the sharp andextended declines in leading indicators expressed in physical units or in constant dollars, (b) thewidespread and substantial declines in various measures of the physicalvolume of aggre-

L 'I TABLE 5 IndexRanking of ofTwelve Seven leading Periods of Recession in Indicators, Deflated Successive Months, 1948-1974: BusinessMonthsCyclePeakAfter 1973Nov. Nov. Jan.Business Cycle Peak May Datea July 1967 1960 1957 1953 1948Nov. Average, Five 2 1 Recessions 5 a 76 8 10 9 -1.5 (Dec.) 12 -1.9-2.1 (Feb.) (Jan.) -l.7(Apr.)-1.6 (Mar.) -4.6-2.5-3.2-2.2 (Aug.) (July)(June)(May) -9.6-7.8 (Sept.) (Oct.) -1.0 (Nov) -3.0-2.5 -5.2-4,1 --.1-5.8 -5.8 -p.5-6.4 Per Cent Change from -1.7-1.3 -1.4-0.8 Business Cycle Peakb -2.9-0.3 -4.5-1.2 -0.1+0.2-1,2-1.8 -5.8-4.6-3.0-2.4 -10.2 -8.9-7.2-5.1 -5.2 +2.6+1.9+ 1.5+1.7 -4.2-6.l-6.5 -10.7-12,-12.0-12,3 -8.2'--v.7 +± 1.71.0 - 1.6-2.7 -6.3-8.3 -7.4-7.4 -6.7-7.4 -5.8-7.4 -1.5 -3.2-4,4-5.6 -7.1-5.1 -12.4 -9.6 - (3.5 -13,4 1967 arid 1973 Omitting --3,3 1.0 -10.4- 12.6-82-8.4 -4.6 -8.5-7,7-6.3 -8.9-9.0 -7.3-8.2 -6.5 3 -3.2 -8.2 -5,5 with UltimateSeverity,CorrelationRecessions Six 1 6.5 3(3) Ranking According to Size of Per Cent Declinec 5(5) 2(2) 1(1) 4(4) 7(6) Omitting 1973d +14 432 23 4(3) 1,5(1)d 2(1)d2(2) 3(2)3(2)1(1) 5(4) (4) 6(5) 7(6) -4-1.00+1.00 ±94 765 2 3.5(2.5) 3(2)4(3) 1(1)d1(fld 3.5(2.5) 4(3)3(2) 6(5)5(5) 5(4) 7(6) +.89+93+94 98 22 5(4)3(2) 1(1)a1(1)d 4(3) 7(6)7i6)6(5) 3(3)3(2)5(4) 6(5)7(6) +54+89 NOTE: 111012 afterequipment;Includes average deflatedbuilding workweek, permits,change inprivate manufacturing; consumer housing; instalment initialnet business claims credit; forformation; ratio, unemployment sensitive deflated materials insurance; inventory prices change;deflated to WPI; ratio, new ratio, orders,price to durableunit prices labor to cost, 6.5 5 4(4) (5.5) 1(1)dl(l)' (fld 2(2) (2) 5(5) (4) 3(3) ;(3) deflated contracts and orders,CPI. These plant andare the same 12 series as in themanufacturing; deflated corporate profits 6.5(6) (5.5) dollars are deflated by an +70-v.83±71 peak is provisionalDates for 1948-1969 and is used are only from fcr the the National purpose Bureau's of testing business the hypothesis cycle chronology that apublished businessappropriate except index forcycle 1967. priceof peakleading index.which may indicators is have included been published as an in Business Conditions Digest (series 8111, except that here the seven series that are in current Ranking of Six Recessions According to Ultimate Severityc (1)1 (2) (4) example of a "mini-recession.'' The' November 1973reached.(5) (6) parenthesesdA"Using5Per November,rank Cent ranks of 1changes means omitshown arid theDecember thatare in 1973-1974 parentheses.thecomputed percentage 19731 onperiod istheFor decline baseandsix ranks is henceol the the thesma three-monthare correlation lest comparable among average ccfficient the with seven centeredthe should recessions ranks onexceed of the ultimateover business 0.83 the severitycorrespondingto hecycle (bottompeak interval; line), a rank of 7 112.1 11967 = 100). (3) significant at the 0.05 level. month. The base for the current permod (average br October, means that it is the largest. Ranks shown in rides based on 12 physical volume aggregates e.g., indL.strlal proauction. "UltimatetBusiness severity1948-49,nonfarm cycle is troughmeasured employment,- 17.5. date, by In the 1953-1954 realper centGNP, decline the dcl. trough These from high per In cent low declinesin a trend-adlusted are: 1969-1970. composite -10.2; 1966-1967. '-30; 1960-1961. is in the 1 3tlm n'iorith, August 1954. The trough o) the 1967 "mini-recession ' -8.5; 1957-1958, -'15.5: 19,3-1954, -t 7.2; is in the 4th moeti. MJv 1967.

a CHART 3 Recession Patterns: Index of LeadingIndicators, Deflated

Nov. 1948 July 1953 15r 15

10- 10

-5 -5--

-151t I i i I i i I i I I -151 I I i I 0 3 6 9 12 15 18 21 24 0 3 6 9 12 15 18 21 - July 1957 41 > 4,

- 10 E5 0---____--

-155L - a I -15 0 3 6 9 12 15 18 21 24 0 3 6 9 12 1518 21 24 Nov. 1969

-'5b -15 \ 0 3 6 9 12 15 0 3 6 9 12 15 MOflIh otter peak 18 2124 gate economicactivity, and (C) the upturn in either absence of any firmindication of an group of indicatorsit seems the currentperiod is reasonable to condude that comparable in depth,duration and diffusion previous businesscycle recessions. with The Precedingreview has indeed revealed somecontradictions in these respects between measuresex- TABLE 6Chronology of Peaks in Selected Leading Indicators, 1972-1974 Lead(-) or Lag(+), in Months, at November 1973 Series' Date of Peak Peak

Indicators Measured inNonmorietary Units New building permits, private housing (29) 11 Average workweek. niIg. (1) -9 business formation (1 2) March 1973 -8 2' Initial claims, unemployment insurance (inverIed(5) 4 Indicators Measured in Current Dollar Units Index of stock prices, S&P 500 (19) -10 Change in consumer instalment (113) March 1973 -8 Industrial materials price index (23) April 1974 +5 Contracts and orders, plant and equip- nient (10) July 1974 +8 New orders, durable goods industries (6) August 1 974 + 9 Corporate profits after taxes (16) Q3 1974k +9 Change in book value, mfg. and trade 4 inventories (31) October 1974 -4-11

NOTE:The indicators are the twelve selected in 1966 in Indicators of 8ui,rieS Expansions and Contractions, Geoffrey H. Moore and Julius Shiskin. New York, National Bureau of Economic Research, 1967, with two exceptions. tnitiat claims was substituted several years ago for nonagricultural placements because of administrative changes affecting the latter (see 8usines Conditions Digest, ). Also, because of its ambiguity in relation to the monetary- nonmonetary classification used in this table, the ratio of price to unit labor cost in manufacturing iv omitted. It reached its highest value to date in November 1974. Number in parentheses is the series number in Business Conditions Digest. U.S. Department of Commerce tnsonthly). bLatest available figure.

pressed in physical units and measures expressed in constantdollars. Nevertheless, no important measure of physical activity hasfailed to exhibit a decline. Where these declines have been brief, as in the caseof 4 employment, they seem likely to be extended into the comingmonths. Hence, it is appropriate to reconsider the tentative datefor the business cycle peak that was selected early in 1974 and to determine whetherthat n month, i.e., November 1973, or some other is supported bythe presently available evidence. To that end we have brought together inTable 7 and h Chart 4 eleven comprehensive measures of the physical volumeof activity collection does d and marked the dates when they reached their peaks. [his not, of course, include any measures that areexpressed in current dollars,

I Ii ISO ('nflrey

for the reason that none of the current dollarcounterparts of these with the exception of retail sales, has at this writing series yet macliedd peak. would serve no useful purpose to allow lhis substantial (liVerger tween nominal and real measures to influence the choice of be. a peak datein the business cycle. A similar, though less extreme,divergence Occurred at the 1 969 peak, but rarely at earlier business cycleturns. Surely is a substantial divergence between the "real" and the Vl)ere there "flOflliflal"meas. ures of aggregate economic activity, few would hesitateto say that "real" was of more concern and should be the the decisive criterionIt Was this consideration that led Solomon Fabricant in hisanalysis of the peak to opt for as the date instead 1969 of a laterPoint. Indeed, GNP in current dollars did not decline at all in 19691970ii tii5 had been the sole criterion (assonìe have advocated in the past), would have been designated. no recession The evidence in Table 7 and Chart 4 points clearly toNoveniber 1973 the appropriate choice for the peakdate. Retail sales reached its highvelI before then, and totalfinalsales(i.e., real GNP less thechange in inventories) reached its high in the thirdquarter; eniploynient rise well into 1974. The bulk of the continueJ to I highs, however,came in J973 fourth quarter, or in November. November 1973 appears then to used to identify previous business meet the criteria cycle peaks in theNational Bureau's chronology, namely the datewhen aggregate economic activity reach its

TABLE 7 Chronology of Peaks in ElevenMeasures of the Physical Volume of Aggregate Economic Activity,1972-1974 Seriesa Date of Peak Retail sales, inconstant dollars (59) March 1973 Final sales inconstant dollars (273) Unemployment rate (43) August 1973k October 1973t GNP in constant dollars(205) Disposable personal November I 973b income in constant dollars(225) Index of industrialproduction (47) November 1973 Index of five November 1973 coincident indicatorsrleflated (825) Index of live physical Novrniljer 1973 volume indicatorsn Total civilian November 1973 eniployrnt11thoijsehol1 survey (842) Septeniljer 1974 Nonfarm employnientpayroll survey (4 I) Manhours in nonfarni October 1974 establishments (48) October 1974 Number in parenthesesis the series numIr (monthly). in BUrgfleSS ond,r,'inç Digest U S Departmentof commerce 5Mjd..month of quarter. 'Date of trough. d5 text (or explanationof contej- CHART 4Selected Measures of the Physical Volumeof Aggregate Economic Activity, 1971-1974 Cs, It 3 e-

31 Ret.ad sales, deflated in l8illl 1967 dotlarsf at re Final sales, in constant dollars 0 (Aasnoal rate. billion dollars) S. I \ 78 he 76 74 as Si 72 unemoloyrrant rote I innetted) 9 (Percent) 6J

6.5 S n 84( 7.o ON P ve constant dollars 82 (Annoal rate. bilhon doPers) 80f as 78 cli 7b( . 620 14 Ditosable personal income irt constant dollars In 600)An,uaal rate, billion dollars) to 50 th lode,, of ,nct,striat produCtion 1961=100) 12

11 40 ts lode,, of live coincideot indicators. dellated II 35(19671001 30 to 25 Indeu of five physical voloete ind,catO,S (1967= 100)

7 5 Total eseployrnent, hoosetrold survey (M,llionrtl 3

.Nontarr,r emoloynrenl. payroll survey (Millions) 3E

150 Manhours in nonlarm einploynrent (Annual rate, billion rnanlroors) 145 140 JILITI.T.I '35 1971 1972 1 lv

decline. highest point and began anextended, substantial, and widespread e hesitant and True, for 1 0 or 11 months afterNovember 1 973 the decline was prolonged the not fully evident in all measuresof aggregate activity. This period during which it wasdifficult to determine confidently thatsuch a

I

()Ore Slowdowns, Recessions, a no Inflation 153

973 IVU GROWTH RATES AND INFlATION RATES le to Ci ng As noted in Section II, the United States has experienced eight periods of d ir slowdown in economic growth since 1948; a ninth began in 1973. Five of the the eight periods encompassed recessions while the other three periods a ks were milder affairsdeclines in employment and output were less wide- r to spread, total activity continued to grow hut at a slower rate, and un- S in employment scarcely rose at alI)2 The rate of inflation slowed perceptibly but or remained very low during each of the eight periods of slower growth, rger and rose perceptibly, sometimes drastically, at other times. In short, the conditions that produced slower economic growth also reduced the rate of be inflation, and the rate of inflation was not reduced otherwise. en - Let us see specifically how the record supports this finding. Table 8 shows what happened to output, the unemployment rate. and the rate of an- its inflation--measured by the consumer price indexduring the five periods that embraced recessions. These periods each lasted from a year to about a 74 year and a halt. Gross national product in real terms, i.e., alter allowing for nt. price changes, declined at annual rates ranging from about half of 1 per cent es. to 2½ per cent. Unemployment rose by 2 to 5 percentage points to levels an between 6 to 8 per cent. Each percentage point nowadays represents en approximately 900,000 persons, so these are not inconsequential numbers. ent The inflation rate declined to much lower levels than when the slow- Id downs started. In two instances, 1949 and 1954, the rate became negative, ses that is, the dropped for a short time. In another two instances, ig 1958 and 1961, the inflation rate dropped to zero, that is, the price level became stable. In the last instance, 1970, the rate was cut in half but did as not fall below 3 per cent. Hence, recessions have invariably been accom- ly panied by a reduction in the rate of inflation, but have unfortunately also ed been accompanied by substantial increases in unemployment arid a reduc- er tion in output. ok Table 9 extends the record to the three slowdowns without recession. all These periods also lasted from a year to a year and one-half, but real GNP )il continued to grow, at rates ranging from 2 to 3½ per cent. Despite the of slowdown, unemployment roselittle or not at all. That is,the rise in at employment accompanying the rise in output was just about sufficient to ig keep up with the growth in the labor force. In all three instances there was at some reduction in the rate of inflation, though in 1962-1963, when the rate Il - was already at a low 2 per cent, it did not dropmuch below that. Finally, Table 10 looks at the other side of the coinwhat happened when the economy grew at a rapid rate. In each of these eight periods real GNP grew at rates of 4'/2 per cent per year or more: the average for all .

TABLE 8 Periods Growth Rates, Unemployment, and Inflation During Periods of Slow Growth and Recession High Growth andRecession of Slow Low Months Dura.tion, in Per CentGrowthin RealRate ofGNP Low Unemployment Rate in Per Cent High Change High Inflation Rate in Per Cent Low (CPI) 7/48 (1) 10/49 (2) (3)15 (4) (5) 3 (6) 8 (7) (8) 14 (9) Change (10) 2/573/53 8/54 1517 -1.0-0.5 42 6 +4±5 4 -1-4 -18 -3 3/692 60 2/615/58 12 -1.5-2.5 3D 78 +2+4 0 -4 NOTE: For eplaninon of figures see notes following 11/70 20 Table JO -0.5 6 ±3 7 4) 0 TABLE 9 Growth Rates, Unemployment, and Inflation During Periods of Slow Growth Without Recession High Slow GrowthPeriodsRecession ofWithout Low MonthsDura-tion, in Per CentGrowthinRate Real of GNP Low Unemployment Rate in Per Cent High Change High Inflation Rate in Per Cent Low (CPI) Change 6/51 (1) (2) (3) (4) (5) 3 (6) 4 (7) (8)14 (9) (10) 4,62 6/523/63 1112 2.03.5 5 6 +1 2 1 1 15 NOTE6/66 For explanation of fi1ures see notes following Table 10. 10/67 16 3.0 4 4 0 4 2 21 TABLE 10 of RapidPeriods Growth Rates, Unemployment, Dura- Per CentRate of and Inflation During Periods of Rapid Growth Low (1) Growth High (2) Months tion, (3) fl Growthin RealGNP (4) High (5) Unemployment Rate in Per Cent Low Change Low Inflation Rate in Per Cent High (CPI) Change 10/49 6/51 20 11.5 8 (6) 3 (7) (8) (9) (10) 8/546/52 2/573/53 30 9 4.58.0 46 43 1 14 14 2 18 3 2/615/58 4/622/60 1421 7.56.5 78 5 2 0 24 25 1 0/67 3/63 6/66 39 6.0 4 6 45 23 I 42 32 11/70 3,733169 28 17 7.04.5 53 2 2 13 3 Notes to Tab$es 8, 9, and 10 Cols. I and 2: recessionofSummerGrowth aggregate cycle 1 encompassed974. economicchronology It is based activity is byupon from the (e.g., specific Isaslowdowns Mintz,real cycle GNP, "Dating in peaks industrialTable United and 8, troughsfromproduction, States peak inGrowth a to nonfarmtrend.adjusted Ircugh, Cycles employment,are: 11/48-10149, composite etc.). The753-8/54, Explorations in Economic Research, National Bureau of Economic Research, index of nine comprehensive measures of the physical volume March 1973 peak (Table 10) is tentative. The periods of 7/57-4/48, 5/60-2/61, and 11/69-11.70. Colt.Col.Col, 5 and4:3: 6: 1962-1963.SpecificPerInterval cent cyclebetween change 1966-1967 peaks atdates seasonallyand inaretroughs cols. exceptions; adjusted in the seasonally annualthe increases rate adjusted from in theunemployment quarterunemployment containing rate associatedrate the were monthly notwith datethe growth in col. cycle turns in 1 and 2. sufticiently large or sustained enough to be identified as cyclical. 1 to the quarter containing the monthly date in col. 2. cols. I and 2. rhe entries (or 1951-1952. ColsCol.Ccl. 7: Sand 10: 9: Change,Change,theSpecificThe latest in entries cyclepercentage inhigh percentage peaks isfor tentative 1962-1963 andpoints, points,troughs (see between are test). between in an the exception; the seasonally thefigures figures the in adjusted cols.decline in colt. 5 andrate in 8 theand 6.of ratechange 9. was in not the sufficiently consumer pricelarge or indexsustained measured enough over sixto bemonth aentifred spans as dated cyclical. at end The of entryspan. for

00tt TABIF iiLeads (-) and Lags (+) of Unemployment and the Inflation Rate During Growth Cycles (number of months)

The Growth Unemployment Inflation Rate )ds ol Cycle Rate (CM) High Low Low High High Low (4)b (5)h (6)b WIth (1 ) (2) (3)b r (see Ytoa 7/48 6 6 gthe 10/49 0 8 ia[jon 6/51 1 4 6/52 7 +8 +3 +7 slow 3/53 8/54 +1 +3 e,or 2/57 0 4 ade 5/58 +2 +5 the 2/60 0 4 SOme 2/61 +3 +4 nt. 4/62 1-3 +1 owth 3/63 +2 0 6/66 +5 -2 r,'als -6 flin 10/67 0 3/69 +2 +13 +12 +21 med 11/70 +7 wer 3''3 +4 egan Medianleadorlag +2 ±2 -2 e See notes to cots. 1and 2 following Table 10. change in the CPI and ubt. bnyal inmonthsbetween the specific cycle turns in the unemployment rate or rate of the growth cycle turns in cots. 1 and 2. Highs in the unemployment rate arecompared with tows in the growth tion end of span. cycle,and viceversa.The rate of change in the CPI is measured over six month spans dated at 'S Ifi 'Assuming that the 13 per cent annual rate in October 1974 (overpreceding six months) represents a cyclical peak. s of h of was negative 0.9 per cent. Unemployment rose from 4.6 per centin October 1973 to 6.5 per cent in November 1974, that is,by nearly 2 percentage 144 points. The rate of inflation, measured in the same way asin our tables, has the leveled off.It reached 12.6 per cent in , 11.9 inJuly, 12.1 in the August, 12.5 in September, 13.2 in October.(These are annual rates, up. seasonally adjusted, over the preceding six months.) od If the 13 per cent rate in October proves to be thecyclical peak. the 19 e. month lag in the downturn of the inflation rateafter the start of the rly economic slowdown (March 1973) has beenunusually long. But it must be ext noted that the lag of 21 months at the previous upturnalso was unusually of long. In other words, the latest upswing in thegrowth cycle (28 months, the to March 1973) was of about the sameduration as the a

()0r TABLE 12The 1973-1975 Contraction Compared with Three Preceding Contractions bei- Sin 1973- --1 929-1 933 ord 1975 Initial tely Based on 1957- 1937- Decline Ion Available1958 1938 Cone- (lOS Figures Full Full Full sponding to Datea Decline Decline Decline to(1) Indicators (1) (2) (3) (4) (5)

Duration (months) Business cycle chronology na. 9 13 43 na. GNP, current dollars 6 9 42 12 he GNP, constant dollars 12 6 6 36 12 on Industrial production 15 14 12 36 15 Nonlarm employment 73 4 14 11 43 4 Unemployment rate 16 1 6 11 60r 18' '3 t C Depthd (per cent) 'f- GNP, current dollars -2.6 -16.2 -49.6 ir) -12.3 GNP, constant dollars -5.0 -3.9 -13.2 -32.6 -6.8 ly Industrial production -13.5 -14.3 -32.4 -53.4 -28.3 y Nonfarm employment -2.9 -4.3 --10.8 -31.6 -3.2 U nemployn-tent rate; e low 4.3 3.7 11.0 3.2e 3.2r e High 8.2 7.5 20.0 25.2 12.be f Increase 3.9 3.8 9.0 22.0 9.4

Diffusion Nonlarm industries; Maximum percentage with declining 85 88 97 100 100 employ ment Nov. '74Sept. '57Dec. '37June '33Aug. '30

NOTE:na. indicates not available or not applicable. The intervals 1mm peak to the lowest point reached to date are: GNP in constant dollars, Q4 73-Q4 74; industrial production, November 1973-; rionfarm eniployment, October 1974-February 1975; unemployment rate, October 1973-February 1975. 5No decline. 'Based upon annual averages br 1929 (low) and t933 (high). dpercentage change from the series' peak month or quarter to its trough monthor quarter, over the intervals shown above. 'In lieu of monthly data an estimate of the approximate unemployment rate 18 months after the 1929 low (3.2 per cent) was obtained by averaging the annual figures for 1930 and 193 t (8.9 and 16.3 espective)y). 'For 1957-1958 and 1973-1975, based on changes in employment over six-month spans in 30 nonfarm industries, centered on the fourth month of the span. Hence the interval covered runs from three months beforeto three months after the month shown on the bottom tine. For 1929-1933 and 1937-1938. based on cyclical changes in employment in 41ndustries. August t930 is the date selected to correspond with November 1974 )col. 1). since both are 12 months after the boxiness cycle peak (August 1929 and November 1973, respectively).

Moore CHART 5Diffusion Indexes of Prices, 1970-1975

Per cent rising Percent rising GNp 100 lstrjal d the L PIoy Feb 25 23 I

957- 0 -100

e far 75 )rt of 50 2 Purchased moterial's prices (PIJAPM} Mar how. 440 hers, 25

year leo 0 orre- 75 Feb itial 3. Wholesale prices, manufOctured goods 62 than 50 farm 25 IV 0 920 tOO

II 75 ow Anhicipol 150 ose 50 4. Retailers selling prices the 25

ta ,i,t,I!.I,lJfl,, JiflsLknL1sa, 0 ely 1970 1971 1972 1973 1974 1975 vas MOTE: The spans oser whk h price changes arc measured, the nuniber of itemsi uverrxt. and the sources are: rial (1)9 months, 13 materials, Bureau of Labor Statistics; 12) I month, approximately 250 resIx)rident nsernbers of the Natio,ia( Ass.icjatiori ,,1 Purc liasing Maliagerielit; (3)6 ruoiiIti, 22 indexes, Bureau of Labor Statislix s; 14)4 are quarters, approximately 250 respondents, Dun & Bradstreet, Inc., Survey of Businessmen's Expectations. The a per cent rising is plotted in the final month or quarter of the span. of an of peak levels during 1 974, with especially rapid declines in the second half of 71 the year.1's 4. One of the concomitants of a softening of -supply pressures is e that increases in prices become less widespread and reductions more 'el widespread. Evidence of this type of development, measured in terms of e diffusion indexes., began to show up during 1974 (see Chart 5)15 The first to show it was the diffusion index of spot maiket prices of industrial of materials, which reached its high (100 per cent rising) in February 1974, ly and fell steadily to 23 per cent rising in February 1 975. At about the same 0 time the diffusion index based on reports of purchased materials prices by 5, members of the National Association of Purchasing Management declined from a high of 97 per cent in March 1974 to 50 per cent in February 1975. The last reduction of comparable magnitude was in 1970. In the autumn of 1974 reductions of similar size began to appear in the diffusion index for manufactured goods prices at wholesale. The Dun and Bradstreet indexes

t $ 164

for manufacturers', wholesalers', andretailers' prices, 1)0thactual anticipated, also started to drop. InflationWdS becoming less and generj Finally, there is some evidence fromthe cornpre1e;1sj.e themselves. the measure of the rate of prfte lndeses inflation used jSection vi, the percentage change in the consumer price indexOver the Preceding si months, seasonally adjusted at annualrate, reached its high October 1974, at 13.2 to datein per cent. The succeeding figuresare: 1974, 12.7; December 1974, .2; ,12.1; and 1975, 10.5. The month to monthchanges have declined Fehruat, more sharply they usually do, from as a high of 1.3 per centper month in September to 0.9 in October and Augustand November, 0.7 inDecember, January, and 0.6 in February.'In addition, the 0.6 rate of increaseir wholesale price index forindustrial the (Ieclined sharply August 1974. The high point in the alter six-month changeat an annual 36.9 per cent in August; it has rate declined Continuouslysince then cent in February 1975, a decline that to 8.4per was accomplished in the time it took to rise from aboutone.thi,d the correspondinglevel in early trend continues, the historical 1973i association betweeneconomic slowdo.5 recessions, and the rate oI inflationwill have repeated itself onceagain.

NOTES

See Economic IndicatorAnalysis Dc'ring 1969-1972 Nations Economic Growth PaulA. David and Melvin and Householdse Inc., 1974); also, Reder, eels. (New Masurjng Recessions York: AcademicPrtsi Busin055 Cycle Indicators eel. (New York: NBFR,1961). Geoffrey H. Moore The NBER chronology of business cyclepeaks and troughs has recently been reexamined and revised inthe study of intensiseh indicators undertakenby Victor Charlotte Boschan forthe Bureau of Zarnossjt7 and EcommAnalysis U.S Department The revised dateshave not been of Commerce the time the work used in thisreport because they began, but itis unlikely that were not available at difference in the they would make results. Only fourof the ten peak any appreciable changed one troughby three months and trough dates,I948_1970 are month each (July (August to May1954) and three to August 1957,May to April 1960, peaks by one lithe peak November to December in the index wereused as the staing 1969 might or might point, thdeclines in earlier peris not be sharper thantho higher level, but shown in the table.They would sla would reach less farinto the bema covering the first ten recession period. In fact, months of decline computations decline in substantially after the index's the same own peak Place the curler' One of the useful position as in thetable properties of this indexis that it movement isIper cent per month is sü constructed thatits asera change in each This IS achievedby standardizing of the fivecompone5 in such the rates ci average rate of changeof I lr cent. a way as to producean index sith ar a particular change Beçaij50 of Ihis in the index is propeitv it is easy tosee sshei above or belossaverage. r

165 Morre Inflation ....., Slowdowns, Recessions and

governed in part by Cttjaj The reason for this is that the levelof the unemploynleflt rate is and S. has been higher. in gen eral factors unrelated to the stateof Prosperity. In recent years, the rate of the large increase in the proportiunof female ' both prosperity and rerCcsinfl, because higher unemployment rates thanadult nier whethcr the I, and younger workers, who have less affected by namely, labor is tight or easy. The changein the rate over short periods us Employment? by Geoffrey H. eding si such shifts in labor force conhI)ocit ion.See flow Full is Full Domestic Affairs Study - date in Moore, American Enterprise Institutefor Public Policy Research, OVem1 No. 14, luly 1973, pp. 27-28. of the Physkal For further description and analysis seeWalter W. Ebanks, "A New Index February , . Volume of Economic Activity" reproduced on the carrie harply,as The longer historical perspectiveof Table 3 can, of course, be in each recession. gust and plan as in Table 2, i.e., in terms ofdeclines during the first 12 months declines are small relative to those in1921, 1929, or 0.6 in Even on this basis the current declines in industrial productionduuing the first 12 se in the 1937. For example, the percentage 1923, -11.9; months following business cycle peakdates are: lan. 1920, -28.0; May rply alter -22.9; May 1937, -32.5; Feb.1945, -35.0; Nov. 1948, Oct. 1926, -5.2; Aug. 1929, 1969, -7.1, Nov. rate was -8.0; May 191,0, -1.6; Nov. -6.2; luly 1953, -7.8; July 1957, figures are not a'ailable o 8.4 per 1973, -3.9. In the sameorder, starting with 1929 because nonfarm employment are: -9.6,-9.7, -6.3, -4.2, one-third earlier, the percentage declines in .3, _39, -1.3, -1.3, +0.7. 3. If this ot a survey thatwould For example, considerationmight be given to development of realized prices from the sameenterprises that WClownS, obtain a representative collection orders, and inventories. Anotherpossibility would be again. supply figures on aggregate sales. and set of alternative estimates,data on sa!es, orders, to collect, and utilize in a would flag problems inventories in physical units. Astatistical monitoring that of might also serve auseful purpose. affecting the current interpretation United Business CvcJe (New York:NBER, 1966) and "Dating use Mintz, Dating Post-scar 1974; Philip A Explorationsiv' Economic Ree,irh. Summer States Growth Cycles," NBER (in preparatioul). Klein, "Postwar GrowthCycles in the ," in The Business Cycle Today,Victor Zarnowitz, ed. eho!dc j 10 "The 'Recession' of 1969-170," (Nesv York: NBER, 19721, p.124. mb Press, creating a supply constraint at abusiness cycle peak Moore The most recent case of an event of 1959. when the steel strike curtailedoutput in the second half was in 1959-1960, and compensating for Anticipation of the strike stimulatedoutput before it occurred, tensively afterwards. Where the peakwould have ri output stimulated output Witz and unanticipated losses carried not taken place is difficult to say,hut the poststrike surge )mmerce. come if the strike had determined to be in May 1960. the economy to a new high,and the peak was ailable at these periods, which t draw uponin what follosvs, see be preciable For a careful analysis of Mintz, "Dating United StatesGrowth Cycles." 1970, are documenting this tendency in termsof wholesale prices, A report by Philip Cagan s by one Behavior of Wholesale Prices inthe 1920's and entitled "Changes in the Recession r 1969). Research, Winter 1975. Post-World War II" appeared inExplorations in Economic periods correlation coefficients betsveenthe lags in the unemployment 14,The simple and partial 11 are a from a CPI rate Ic), and time It), forthe 17 observations in Table utaticins rate (u), the lags in the current as follows: = +74 ri,- +.60 average r, = +.57 rn,, = +24 rates of = +60 r,,,., = -f,32 with an coefficient should exceed .48,and At the 05 level of significancethe simple correlation whether that there is some linkage the partial coefficient shouldexceed .50. The results suggest