Front matter, What Happens During Business Cycles · Willard L. Thorp, American Economic...
Transcript of Front matter, What Happens During Business Cycles · Willard L. Thorp, American Economic...
This PDF is a selection from an out-of-print volume from the NationalBureau of Economic Research
Volume Title: What Happens During Business Cycles: A Progress Report
Volume Author/Editor: Wesley Clair Mitchell
Volume Publisher: NBER
Volume ISBN: 0-870-14088-4
Volume URL: http://www.nber.org/books/mitc51-1
Publication Date: 1951
Chapter Title: Front matter, What Happens During Business Cycles
Chapter Author: Wesley Clair Mitchell
Chapter URL: http://www.nber.org/chapters/c3163
Chapter pages in book: (p. -32 - 0)
NATIONAL BUREAU OF ECONOMIC RESEARCH
Studies in Business CyclesNo.5
What Happens
During Business Cycles
A PROGRESS REPORT
NATIONAL BUREAU OF ECONOMIC RESEARCH1965
OFFICERS
Frank W. Fetter, ChairmanArthur F. Burns, PresidentTheodore 0. Yntema, Vice-PresidentDonald B. Woodward, TreasurerWilliam J. Carson, Secretary
Geoffrey H. Moore, Director of ResearchHal B. Lary, Associate Director of ResearchVictor R. Fuchs, Associate Director of
ResearchDouglas H. Eldridge, Executive Director
DIRECTORS AT LARGE
Robert B. Anderson, New York CityWallace J. Campbell. Foundation for
Cooperative HousingErwin D. Canham, Christian Science MonitorSolomon Fabricant, New York UniversityMarion B. Folsom, Eastman Kodak CompanyCrawford H, Greenewalt, E. 1. du Pont de
Nemours & CompanyGabriel Flauge, Manufacturers Hanover Trust
CompanyA. J. Hayes, International Association of
MachinistsWalter W. Heller, University of MinnesotaAlbert J. Hettinger, Jr., Lazard Frères and
CompanyNicholas Kelley, Kelley Drye Neuhall
Maginnes & Warren
V. W. Bladen, TorontoFrancis M. Boddy, MinnesotaArthur F. Burns, ColumbiaLester V. Chandler, PrincetonMelvin G. de Chazeau, CornellFrank W. Fetter, NorthwesternR. A. Gordon, California
Percival F. Brundage, American Institute ofCertified Public Accountants
Nathaniel Goldfinger, American Federationof Labor and Congress of industrialOrganizations
Harold G. Haicrow, American FarmEconomic Association
Murray Shields, American ManagementAssociation
Moses AbramovitzGary S. BeckerWilliam H. Brown, Jr.Gerhard BryArthur F. BurnsPhillip CaganFrank G. DickinsonJames S. EarleyRichard A. EasterlinSolomon FabricantAlbert FishlowMilton FriedmanVictor R. Fuchs
H. W. Laidler, League for IndustrialDemocracy
Geoffrey H. Moore, National Bureau ofEconomic Research
Charles G. Mortimer, General FoodsCorporation
3. Wilson Newman, Dun & Bradstreet, Inc.George B. Roberts, Larchmont, New YorkHarry Scherman, Book-of-the-Month ClubBoris Shishkin, American Federation of Labor
and Congress of Industrial OrganizationsGeorge Soule, South Kent, ConnecticutGus Tyler, international Ladies' Garment
Workers' UnionJoseph H. Willits, Lan ghorne, PennsylvaniaDonald B. Woodward, A. W. Jones and
Company
Harold F. Williamson, Economic HistoryAssociation
Theodore 0. Ynrema, Committee forEconomic Development
H. G. GeorgiadisRaymond W. GoldsmithJack M. GuttentagChallis A. Hall, Jr.Daniel M. HollandThor HnltgrenF. Thomas JusterC. Harry KahnJohn W. KendrickIrving B. KravisHal B. LaryRobert E. LipseyRuth P. Mack
Jacob MincerIlse MintzGeoffrey H. MooreRoger F. MurrayRalph L. NelsonG. Warren NutterRichard T. SeldenLawrence H. SeltzerRobert P. ShayGeorge J. StiglerNorman B. TureHerbert B. WoolleyVictor Zarnowitz
DIRECTORS BY UNIVERSITY APPOINTMENTHarold M. Groves, i'VisconsinGottfried Haberler, HarvardMaurice W. Lee, North CarolinaLloyd G. Reynolds, YalePaul A. Samuelson, Massachusetts
Institute of TechnologyTheodore W. Schultz, Chicago
Willis J. Winn, PennsylvaniaDIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS
Willard L. Thorp, American EconomicAssociation
Allen lATallis, American StatisticalAssociation
DIRECTORS EMERITIShepard Morgan, Norfolk, Connecticut N. I. Stone, New York City
Jacob Viner, Princeton, New JerseyRESEARCH STAFF
Relation of the Directors to the Work and Publications
of the
National Bureau of Economic Research
1. The object of the National Bureau of Economic Research is toascertain and to present to the public important economic facts and theirinterpretation in a scientific and impartial manner. The Board of Directorsis charged with the responsibility of ensuring that the work of the NationalBureau is carried on in strict conformity with this object.
2. To this end the Board of Directors shall appoint one or moreDirectors of Research.
3. The Director or Directors of Research shall submit to the membersof the Board, or to its Executive Committee, for their formal adoption, allspecific proposals concerning researches to be instituted.
4. No report shall be published until the Director or Directors ofResearch shall have submitted to the Board a summary drawing attentionto the character of the data and their utilization in the report, the natureand treatment of the problems involved, the main conclusions and suchother information as in their opinion would serve to determine the suita-bility of the report for publication in accordance with the principles ofthe Bureau.
5. A copy of any manuscript proposed for publication shall also besubmitted to each member of the Board. For each manuscript to be sosubmitted a special committee shall be appointed by the President, or athis designation by the Executive Director, consisting of three Directorsselected as nearly as may be one from each general division of the Board.The names of the special manuscript committee shall be stated to eachDirector when the summary and report described in paragraph (4) aresent to him. It shall be the duty of each member of the committee toread the manuscript. If each member of the special committee signifies hisapproval within thirty days, the manuscript may be published. If eachmember of the special committee has not signified his approval withinthirty days of the transmittal of the report and manuscript, the Directorof Research shall then notify each member of the Board, requesting ap-proval or disapproval of publication, and thirty additional days shall begranted for this purpose. The manuscript shall then not be published un-less at least a majority of the entire Board and a two-thirds majority ofthose members of the Board who shall have voted on the proposal withinthe time fixed for the receipt of votes on the publication proposed shallhave approved.
6. No manuscript may be published, though approved by each mem-ber of the special committee, until forty-five days have elapsed from thetransmittal of the summary and report. The interval is allowed for thereceipt of any memorandum of dissent or reservation, together with abrief statement of his reasons, that any member may wish to express; andsuch memorandum of dissent or reservation shall be published with themanuscript if he so desires. Publication does not, however, imply thateach member of the Board has read the manuscript, or that either membersof the Board in general, or of the special committee, have passed upon itsvalidity in every detail.
7. A copy of this resolution shall, unless otherwise determined by theBoard, be printed in each copy of every National Bureau
adopUd October 25, and February 6, and February 24, 1941)
Studies in Business Cycles
1 Business Cycles: The Problem and its Setting(originally published as No. 10 in the General Series)By Wesley C. Mitchell
2 Measuring Business CyclesBy Arthur F. Burns and Wesley C. Mitchell
3 American Transportation in Prosperity andDepressionBy Thor Flultgren
4 Inventories and Business Cycles, with SpecialReference to Manufacturers' InventoriesBy Moses Abramovitz
5 What Happens during Business Cycles: AProgress ReportBy Wesley C. Mitchell
What HappensDuring Business Cycles
A Progress Report
WESLEY C. MITCHELL
National Bureau of Economic Research, Inc.
DISTRIBUTED BY
Columbia University Press, New York and London
Copyright, 1951, by
National Bureau of Economic Research, Inc.
261 Madison Avenue, New York i6, N.Y.
All Rights Reserved
Printed in the United States of America
introduction
Shortly before his death Wesley Mitchell put in my care thecompleted parts of the "progress report" he was preparing onhis long and elaborate investigation of "what happens duringbusiness cycles". This book is substantially the document heleft behind. I have felt free to make numerous changes of de-tail, but I have not interfered with the design, nor attempted tocomplete the narrative. The work of a major scientist, even ifnot half done, deserves a life of its own, unencumbered by thehand or voice of another. So it is especially when, as in thepresent case, the fragment has well defined contours, balance,and direction. But for the guidance of students who may takeup the book for the first time, I shall put down a few remarksabout Mitchell's objectives and what he accomplished.
I
Business cycles are not merely fluctuations in aggregate eco-nomic activity. The critical feature that distinguishes themfrom the commercial convulsions of earlier centuries or fromthe seasonal and other short-term variations of our own ageis that the fluctuations are widely diffused over the economy—its industry, its commercial dealings, and its tangles of finance.The economy of the western world is a system of closely inter-related parts. He who would understand business cycles mustmaster the workings of an economic system organized largelyin a network of free enterprises searching for profit. The prob-lem of how business cycles come about is therefore inseparablefrom the problem of how a capitalist economy functions.
This conception governs Mitchell's posthumous book, as itdoes his earlier writings. Mitchell was not content to focusanalysis on the fluctuations of one or two great variables, such
vii
Viii INTRODUCTION
as production or employment. His concern was with businesscycles and he therefore sought to interpret the system of busi-ness as a whole—the formation of firms and their disappearance,prices as well as output, the employment of labor and otherresources, the flow of incomes to and from the public, costs andprofits, savings and investments, the merchandising of securi-ties as well as commodities, the money supply, its turnover, andthe fiscal operations of government. Not only that, but hesought to penetrate the facade of business aggregates and tracethe detailed processes—psychological, institutional, and tech-nological—by which they are fashioned and linked together.
Thus Mitchell took as his scientific province a terrain as far-flung and intricate as Walras' and Marshall's. But he exploredmore fully than his predecessors the obstacles to the mutualadjustment of economic quantities in a disturbed environment."Time . . is the centre of the chief difficulty of almost everyeconomic problem." 'Pursuing this Marshallian theme throughuncharted jungles of statistics, Mitchell detected systematicdifferences in the rates of movement of economic variables,and arrived at an early stage of his scientific work at the con-ception that our economic system of interdependent partsgenerates a cyclical path instead of moving toward an equilib-rium position. This fateful twist aside, Mitchell's economic out-look was thoroughly Marshallian. Had he lived to finish thisbook, he would have inscribed on its title page Marshall'smotto: "The many in the one, the one in the many."
The hypothesis that each stage of the business situation tendsto develop out of the preceding stage and to grow into thenext in a cyclical pattern poses two major questions: Doeseconomic life actually proceed in recurrent fluctuations hav-ing similar characteristics? If so, by what processes are con-tinuous and repetitive movements of this character broughtabout? In a search for definite and dependable answers, Mit-chell examined "facts on a wholesale scale", as had Darwinbefore him in a related field, and Lyell before Darwin. "Mysuccess as a man of science", wrote Darwin, "has been deter-
Alfred Marshall, Preface to the first edition of his Principles.
INTRODUCTION 1X
mined. . . by complex and diversified mental qualities and con-ditions. Of these, the most important have been—the love ofscience—unbounded patience in long reflecting over any sub-ject—industry in observing and collecting facts—and a fairshare of invention as well as of common sense."2 These, too,were the sources of Mitchell's scientific strength. In his quartoon Business Cycles, published in 1913, he anchored a theory offluctuations to an array of empirical observations unprece-dentedly full for its time. But Mitchell was not content withthis achievement. World War I had ushered in a new era ofeconomic statistics, able theorists were elaborating new hy-potheses, and statistical analysts were rapidly fashioning newdevices for disentangling economic movements. Eager to ex-ploit the new materials for research, Mitchell launched in 1922a fresh investigation of business cycles.
II
The science of economic fluctuations is only beginning to passinto an inductive stage. Even today the descriptions of busi-ness cycles by economists often resemble the descriptions ofplant life by writers of antiquity, who commonly relied on"casual observations, no experiments and much speculativethinking".8 If later botanists often "could not identify theplants by the descriptions", so it has also been in economics.As long as investigators worked by themselves, they could notvery well "collect the masses of raw data pertinent to the studyof cyclical behavior, segregate the cyclical components frommovements of other sorts, and assemble the findings to form arealistic model of business cycles by which explanations couldbe judged".4 In recent decades the organization of scientific
'Charles Darwin, "Autobiography", in Life and Letters, edited by FrancisDarwin (D. Appleton & Co., 1888), Vol. I, pp. 68, 85-6.8 William Crocker, "Botany of the Future", Science, Oct. 28, 1938, pp. 387,388.
4Mitchell, infra, p. 4. All other page references, unless otherwise indicated.are to the text.
X INTRODUCTION
institutes has greatly enlarged the possibilities of empiricalresearch in economics. Mitchell made the most of the oppor-tunity afforded by the resources of the National Bureau. Tak-ing his own and others' explanations of business cycles as"guides to research, not objects of research" (p. 5), he delveddeeply into the facts of cyclical behavior and the relationshipsamong them. The wish to contribute to economic policy wasstrong in Mitchell. Stronger still was his conviction that intelli-gent control of business cycles depends upon sound theoreticalunderstanding, which requires tolerably full and accurateknowledge of what the business cycles of experience have beenlike.
Business Cycles: The Problem and its Setting, the first majorinstalment of Mitchell's investigation, was published in 1927.The second appeared in 1946 under the title Measuring Busi-ness Cycles. In the meantime numerous studies of specialaspects of cyclical fluctuations were prepared by the Bureau'sstaff, and a small group was steadily engaged in analyzing thecyclical behavior of economic processes.5 It was Mitchell'shope to integrate the findings of his collaborators with hisown and other investigators' results; that is, to develop a modelof business cycles from carefully screened observations, to useit in explaining how the cycles of experience are typicallypropagated, and then press on to account for the outstandingdifferences among them.6 But he would have fallen short ofthe, goal even if he had lived to complete the present book.Many of the needed materials—especially for foreign countries—were not in shape for use, and the subject of business-cycledifferences required systematic investigations yet to be under-taken. As it stands, Mitchell's report barely covers the firstthree of the seven parts he had planned. Part I sets out his aims,methods, and materials. Part II deals with the great variety ofcyclical movements characteristic of individual economic ac-
See the list of publications on business cycles at the end of this volume.
For a fuller account, see "Wesley Mitchell and the National Bureau", in theBureau's Twenty-Ninth Annual Report.
INTRODUCTION xi
tivities. Part III, not fully completed, shows how the cyclicalmovements of different parts of the American economy fittogether into business cycles, and paves the way for analyzingthe processes of expansion, recession, contraction, and revival,to which the last four parts were to be devoted.
Thus the book is a 'progress report', both in the sense inwhich Mitchell intended the phrase and in the poignant senseforced by his death. Yet no existing publication elucidates sofully or so authoritatively what happens during business cyclesas Mitchell's fragment. The accent of the book is on character-istic behavior, formalized in the concept of a 'typical cycle'."The only normal condition" of business, as Mitchell once ex-pressed it, "is a state of change";7 but some states of change are'normal' and others 'abnormal', and Mitchell's 'typical cycle'is designed to take account of such differences. Hence, thisconcept is similar in some respects to the classicists' 'normal'.The role of each is to segregate the effects of complex causes:both are devices of abstraction: both are tools for analyzingnew, concrete situations. Mitchell was keenly concerned aboutthe wide variations among the business cycles of experienceand eager to press investigations of them. But he deemed itessential, as a first step, to lay bare the typical characteristicsof the alternating waves of prosperity and depression that haveswept the economic world in modem times. In this emphasishe conformed to the usual practice of business-cycle theorists.He broke with tradition, however, by extracting what is 'typi-cal' or 'aberrant' from mass observations, and thus substitutingfact and measure, as well as may be, for the impressionisticjudgments that have ruled business-cycle literature.
III
Mitchell begins his survey of what happens during businesscycles by illustrating the varieties of behavior characteristicof economic activities in the United. States. Some of the figures
Bwines: Cycles: The Problem and Its Setting, p. 376.
Xii INTRODUCTION
in his introductory chart merely confirm common knowledge.For example, commodity prices generally rise and fall with thetides in production; business failures increase during contrac-tions of aggregate economic activity and diminish during ex-pansions; the output of durables fluctuates more widely thanthe output of perishables; and prices are more stable at retailthan at wholesale. It is less generally known, however, thatcrop production moves rather independently of business cycles,or that production typically fluctuates over a much widerrange than prices, that the liabilities of business failures usuallyturn down months before economic recovery becomes generaland turn up months before recession, that both durables andperishables experience their most vigorous decline well beforethe end of contraction, and that retail prices characteristicallymove later as well as less than wholesale prices.
Students who will take the trouble to ponder these facts arenot likely to leave Mitchell's chart quickly. They will noticethat orders for investment goods tend to lead the tides inaggregate activity, that private construction is more closely re-lated to business cycles than public construction, that callmoney rates or even commercial paper rates greatly overstatethe fluctuations in the rates of interest at which bank customersordinarily borrow, that interest rates in New York tend tomove before and more widely than in the interior, that thenumber of business failures lags behind the liabilities, that bondprices tend to lead stock prices which themselves lead theturns in aggregate activity, that bank deposits appear to becomparatively steady during depressions, that imports con-form closely to business cycles while exports do not, thatgrocery sales fail to show the regular response to businesscycles characteristic of retail trade at large, etc. And if thereader looks beyond the large processes that have dominatedtheoretical literature, he will see how peculiar the cyclical be-havior of smaller sectors of activity can be. For example, cattleslaughter tends to move with the tides in aggregate activitywhile hog slaughter moves inversely; the• dollar volume ofresidential construction contracts fluctuates less, not more, than
INTRODUCTION Xlii
the physical volume; cotton stocks held at mills run parallelwith mill production, while stocks in public storage move in-versely.
Thus business cycles are complex phenomena—far more sothan has been commonly supposed. The sales of a large firmmay be dominated by the tides in aggregate activity; the for-tunes of a small firm are rather at the mercy of personal factorsand conditions peculiar to the trade or locality. Some activities,like local transit or net gold movements between the UnitedStates and Great Britain, are apparently free from cyclicalfluctuations. Others, notably farming, undergo cyclical move-ments, but they have little or no relation in time to businesscycles. And these irregular responses, passed over lightly bytheoretical writers, accord with reason:We cannot expect any activity to respond regularly to business cyclesunless it is subject to man's control within the periods occupied bycyclical phases, and unless this control is swayed, consciously or not,by short-period economic considerations. The domination of harvestsby weather, the 'migratory property' of petroleum underground, themixed motives of governments in undertaking construction work, thelong-range planning that weighs with many men in a position to set'administered prices', the time-consuming negotiations that preventprompt adjustments of certain other prices and many wage rates, theexistence of long-term contracts, the years required to complete somelarge undertakings—these are concrete examples of the multifariousobstacles that interfere with prompt and regular response to thecyclical tides (p. 95).
The processes that fail to bear the imprint of business cyclesare nevertheless a minority. Almost nine-tenths of Mitchell'sbasic sample of approximately 800 time series fluctuate in sym-pathy with the tides in aggregate activity, but the movementsof this imposing majority are far from uniform. Between thecyclical recalcitrants like farming and the cyclical regulars likefactory employment, there is a continuous gradation. Coal andiron production conform more closely to the tides in aggregateactivity than the production of textiles or gasoline. The pricesof industrial commodities do not conform as well as their pro-duction, while the opposite relation rules in farming. Employ-ment conforms better than wage rates, bank loans than invest-
Xiv INTRODUCTION
ments, open-market interest rates than customer rates, stockprices than bond prices, etc. Some conforming processes moveearly in the cyclical procession; for example, orders for invest-ment goods. Others, like interest rates, are laggards.
Of course, most processes respond to the tides in aggregateactivity by rising during expansions and declining during con-tractions, though they may do so with a lead or lag. But busi-ness cycles also generate countercyclical movements:
Brisk business increases the domestic demand for textile goods and sodiminishes the exports of raw cotton; it increases the sale of freshmilk and so restricts the production of butter; it increases the volumeof coin and paper money held by the public and stimulates borrowingfrom the banks, thereby enlarging demand liabilities and tending toimpair reserve ratios; it leads department stores to carry larger stocksof merchandise and lowers the piles of iron ore at blast furnaces; itactivates share transactions on stock exchanges and discourages trans-actions in bonds. The declines in this list, and many others, are ascharacteristic a feature of business cycles as the advances (p. 66).
However, the processes that run counter to business cyclesdo so, by and large, with less regularity than those that re-spond positively. An expansion of money incomes stimulatesa general increase in buying, and this influence may obscurethe concomitant impulse to shift demand away from inferiorarticles to goods of higher quality. As it turns out, purchases ofstaples such as pork, flour, coffee, and potatoes frequently de-cline during expansion, but their inverted response is less regu-lar than the positive response of more costly articles.In general, influences that tend to repress an activity in expansionencounter more opposition than influences favoring an increase, andwhen repressing influences win out, their victories are less regularfrom cycle to cycle than the victories won by influences that pushforward. Mutatis mutandis, the like holds true in contraction (p. 96).
Large as are the variations in the cyclical timing of economicprocesses, the differences in amplitude of fluctuation are moreimpressive still. In high grade bond yields, for example, thecyclical wanderings are confined to a narrow range; the totalrise and fall is typically only about 10 percent of their averagevalue during a business cycle. The amplitude of the overall
INTRODUCTION XV
index of wholesale prices, excluding war episodes, is nearlytwice as large; the amplitude of factory employment four orfive times as large, of private construction contracts over tentimes and of machine tool orders over twenty times as large.On the other hand, stocks of industrial equipment are remark-ably steady, expanding usually during contractions as well asexpansions of business cycles. The proportions among eco-nomic quantities keep changing so systematically over a busi-ness cycle that thevery essence of the phenomenon is omitted unless the chart of businesscycles contains numerous lines that indicate the wide differencesamong the rates at which, and also some of the differences in thetimes at which, various elements in the economy expand and contract.For, unless these divergencies in cyclical behavior are pictured by fitsymbols, we have no suggestion of the basic business-cycle problem:how an economic system of interrelated parts develops internal stressesduring expansions, stresses that bring on recessions, and how theuneven contractions of its varied parts pave the way for revivals(p. 295).
IvSo much for the varieties of cyclical behavior that come to thesurface once the lid is lifted from aggregate activity. Whatsort of whole do the parts make up? When the individualpieces are put together it appears that every month some activi-ties reach cyclical peaks and others decline to their troughs; sothat expansion and contraction run side by side all the time.But the peaks tend to come in bunches and likewise the troughs.Hence, when troughs gain on the peaks, expansions grow morenumerous and in time dominate the economy. Their supremacyis short lived, however, and gradually gives way to the en-croachments of contraction. The business cycle of experienceis the alternating succession of these sustained majorities—firstof individual expansions, next of contractions, then of expan-sions once again, and so on.Business cycles Consist flOt only of roughly synchronous expansionsin many activities, followed by roughly synchronous contractions ina slightly smaller number; they consist also of numerous contractionswhile expansion is dominant, and numerous expansions while contrac-tion is dominant (P. 79).
Cha
ract
eris
tic D
irect
ion
of T
wen
ty-s
ix 'C
ompr
ehen
sive
' Ser
ies d
urin
g a
Bus
ines
s Cyc
le
Bon
ds so
ld, N
. Y. S
tock
Exc
hang
eR
.R. b
ond
pric
esB
usin
ess f
ailu
res,
liabi
litie
s, m
v.C
omm
on st
ock
pric
esSh
ares
sold
, N. Y
. Sto
ck E
xcha
nge
Cor
pora
te se
curit
y is
sues
Con
stru
ctio
n co
ntra
cts,
valu
eD
epos
its a
ctiv
ityB
ank
clea
rings
or d
ebits
, N.Y
.C.
Inco
rpor
atio
ns, n
o.B
ank
clea
rings
or d
ebits
, out
side
N.Y
.C.
Ban
k cl
earin
gs o
r deb
its, t
otal
Impo
rts, v
alue
Indu
stria
l pro
duct
ion,
tota
lFu
el &
ele
ctric
ity p
rodu
ctio
nPi
g iro
n pr
oduc
tion
R.R
. fre
ight
ton
mile
sFa
ctor
y em
ploy
men
tFa
ctor
y pa
yrol
lsIn
com
e pa
ymen
ts, t
otal
Cor
pora
te p
rofit
sB
usin
ess f
ailu
res,
no.,
mv.
Dep
artm
ent s
tore
sale
s, de
flate
dW
hole
sale
trad
e sa
les,
valu
eW
hole
sale
com
mod
ity p
rices
R.R
. bon
d yi
elds
-E x
p a
n s
i o n
T)-o
ugh
Firs
t Mid
dle
Last
toto
toth
irdfir
stm
iddl
ela
stto
third
third
third
peak
+ ++
++
++
++
++
++
++
++
++.
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++.
++
++
++
+—
—+
+
Last
third to
troug
h+ + + + + + +
—+
—+
—+
Serie
s
Con
tract
ion
Peak
Firs
tM
iddl
eto
toto
first
mid
dle
last
third
third
third
——
+—
++
——
+
+
± + + +
+ + +
No.
of
Bus
ines
sC
ycle
sC
over
ed
19 14 16 16 8 7 16 18 19 14 14 16 5 5 16 9 6 S 4 4 16 4 3
11 19
% o
f Con
form
-in
g M
ovem
ents
durin
g Sp
an o
fSt
ages
in W
hich
Serie
s Is S
aid
toR
ise
Fail
8679
6574
8610
094
8294
8810
075
8675
9488
100
8984
8010
079
100
9394
7510
010
010
010
010
010
010
089
100
100
100
100
100
5010
010
075
8810
075
100
100
8.91
7465
conc
erni
ng th
eir s
ign
from
Tab
le 3
1, S
ec. A
. A p
lus d
enot
es ri
se, a
min
us d
enot
es fa
fl.Th
e tw
o se
ries o
n fa
ilure
s are
inve
rted
here
. Bon
d pr
ices
are
trea
ted
as th
e
+in
verte
d re
plic
a of
bon
d yi
elds
; see
Tab
Je 3
1, n
ote
e,in
the
seco
nd se
gmen
t of c
ontra
ctio
n.
INTRODUCTION XVII
According as the expansions or contractions of individual ac-tivities dominate, the aggregate activity of the economy surgesforward or recedes. And when economic crosscurrents are ator near their maximum, the direction of aggregate activity isreversed: it begins to rise if it has been falling, or to fall if ithas been rising.
The turmoil that goes on within the cycles in aggregate ac-tivity has a systematic core. A highly simplified picture of thesystem is afforded by the accompanying table, which conden-ses Mitchell's analysis of "comprehensive series" in Chapter 10.The table shows directions of movement during a typical busi-ness cycle—here divided into eight segments, four each forexpansion and contraction. Of course, each segment includesseveral months, and the table is therefore insensitive to minordifferences in timing, such as the short lag in income payments.Further, it hides many crosscurrents that would appear in lesscomprehensive series, and omits certain business factors ofwhich we should take account—especially wage rates, inven-tories, banking, and governmental finance. But with all itsfaults, the table gives an effective glimpse of the typical roundof developments that constitute a business cycle.8
Let us then take our stand at the bottom of a depression andwatch events as they unfold. Production characteristically risesin the first segment of expansion; so does employment 'andmoney income; and so do commodity prices, imports, domestictrade, security transactions. Indeed, every series moves upwardexcept bond yields and bankruptcies. In the second stage thebroad advance continues, though it is checked at one point—the bond market where trading begins to decline. Bond pricesjoin bond sales in the next stage; in other words, long-terminterest rates—which fell during the first half of expansion—begin to rise. In the final stretch of expansion, declines becomefairly general in the financial sector. Share trading and stock
This and the three following paragraphs are adapted from the NationalBureau's Thirtieth Annual Report.
Xviii INTRODUCTION
prices move downward; the liabilities of business failures,which hitherto have been receding, move up again; securityissues and construction contracts drop; the turnover of bankdeposits slackens; and bank debits in New York City, thoughnot as yet in the interior, become smaller.
These adverse developments soon engulf the economic sys-tem as a whole, and the next stage of the business cycle is thefirst stage of contraction. Production, employment, commodityprices, personal incomes, business profits—indeed, practicallyevery process represented in the table declines. Of course, theliabilities of business failures continue to rise, which merelyattests the sweep of depression. Long-term interest rates alsomaintain their rise. But in the next stage the downward driftof bond prices ceases; that is, the rise in long-term interest ratesis arrested. By the middle of contraction, bond sales join theupward movement of bond prices. More important still, theliabilities of business failures begin declining, which signifiesthat the liquidation of distressed business firms has passed itsworst phase. These favorable developments are reinforced inthe following stage. Share trading and prices revive; businessincorporations, security issues, and construction contractsmove upward; money begins to turn over more rapidly; eventotal money payments expand. Before long the expansionspreads to production, employment, prices, money incomes,and domestic trade. But this is already the initial stage of gen-eral expansion—the point at which our hurried trip around thebusiness cycle started.
Of course, this recital delineates characteristic movementsduring business cycles, not invariant sequences. That the de-scription fits imperfectly individual business cycles is apparentfrom the conformity percentages in the table. Yet these per-centages also suggest that the deviations from type are not sonumerous as to destroy the value of a generalized sketch. Andif this much is accepted, an important conclusion immediatelyfollows, notwithstanding the omissions of the table; namely,that the check to the dominant movement of business activity,whether it be expansion or contraction, is typically felt espe-
INTRODUCTION XIX
cially early in financial processes and activities preparatory toinvestment expenditure.
The contraction phase of business cycles is not, however, theprecise counterpart of expansion. This is clear from the tableand becomes clearer still when numerical values are attached toits signs and intervals. The arrays of individual turning pointsat business-cycle troughs "are more dispersed and skewedtoward leads" than are the arrays at peaks. Expansions of aggre-gate activity average longer than contractions. They are alsomore vigorous, so that the trough from which a given expan-sion starts is ordinarily above the level from which the preced-ing expansion started. In the first segment of expansion the rateof improvement "is more rapid than at any other stage of thecycle". A "sharp and general retardation" of the advance oc-curs in the next segment. In the third, while "reacceleration isthe rule", the advance "does not regain the speed" it had at thebeginning of expansion. In the final stage of expansion "thebusiness tide . . . becomes fuller of eddies". Contractions fol-low a different pattern. "The fall accelerates somewhat in thesecond segment of contraction, whereas the rise is much re-tarded in the second segment of expansion." The next stage"brings a moderate retardation" of the decline, whereas it"brought a moderate reacceleration" of the advance. The clos-ing stages of expansion and contraction are similar "in that therate of change becomes slower; but this retardation is muchmore marked at the end of contraction than at the end of ex-pansion".9Thus the notions often suggested by the picturesque phrasing belovedof writers upon 'booms and busts'—that prosperity grows at a dizzierpace the longer it lasts, and that slumps gather momentum as theyproceed—are wrong if our measures are right. Scarcely less misleadingare the implications of the mathematical constructions often used torepresent business cycles. A set of straight lines sloping upward torepresent expansion, connected at a sharp peak with downward slopingstraight lines to represent contraction, misrepresents the facts. .
Sine curves are not less objectionable. . . . What our observationssuggest is that the shapes of business cycles are phenomena sui generis(p. 300).
°See pp. 75, 299-305.
XX INTRODUCTION
V
These, then, are some of the broad results that emerge fromMitchell's examination of the cyclical process of the Americaneconomy. The full range of the book, its suggestions for fur-ther research, and its exemplary scientific care await the reader.Economists anxious to wield a simple formula of the causes ofbusiness cycles or the means of controlling them will not findMitchell's fragment to their liking. Those willing to take con-clusions on faith may chafe at its patient elaboration of evi-dence. But men who seek so earnestly to understand how oureconomic organization works that they insist on judging evi-dence for themselves are more likely to lament that too muchdetail has been suppressed. Scholars will respect Mitchell'spronouncement that his report on findings, after many yearsof research, is "ill proportioned, tentative, and subject tochange as the investigation proceeds" (p.5).
This book is not easy and everyone will save time by a care-ful reading of Part I which, besides outlining aims and meth-ods, provides the modicum of technical vocabulary requiredfor comprehending what follows. Economic theorists arelikely to find especially suggestive Chapter 7, which sets outthe facts and inquires into the causes of the changing propor-tions among economic quantities in the course of a businesscycle; also Part III, which centers on the consensus of fluctua-tions in leading sectors of the economy. Chapter 8 is a usefulreminder to all that, despite their persistent traits, businesscycles are changing phenomena; and that just as each newmember of a group has traits of his own, which cannot be in-ferred from knowledge of the 'average man', so each businesssituation must be judged in the light of its own circumstancesas well as according to historical patterns. The bulk of thischapter is devoted to technical problems in the decompositionof time series, and only specialists will want to study it fully.Readers pressed for time might move lightly through Chapters5 and 6 also, except for the closing sections which will repaycareful reading.
INTRODUCTION Xxi
The modern theory of employment, which for a time pushedaside both value and business-cycle theory, is now slowly beingfitted into older economic knowledge. The younger econo-mists are rediscovering that cost-price relations play a signifi-cant role in shaping the national income and its movements,that the 'consumption function' itself moves cyclically, thatinvestment is not an autonomous variable, that price inflationdoes not wait for full employment, and that both investmentand consumption are heterogeneous aggregates that cannotbe understood without separate analysis of their parts. If ourharassed generation can win the opportunity to pursue thearts of peace, the fruit and example of Mitchell's work willhave their quiet but decisive part over the years in bringing thetheory of fluctuations into ever closer contact with the ebband flow of experience.
Arthur F. BumsAugust 1950
Acknowledgments
My task as editor has been greatly lightened by MillardHastay, who helped me verify factual details and took chargeof the preparation of the appendices. Geoffrey H. Moore alsoput his knowledge at my disposal. Both Moore and Hastayhad worked very closely with Wesley Mitchell and he re-cords in the text his indebtedness to them on special points.It is only proper to add here that Geoffrey Moore participatedover a long period in the planning and execution of the sta-tistical compilations of this book, and that Millard Hastay'sparticipation was especially extensive in the closing stages ofMitchell's work. A committee of the Board of Directors, con-sisting of Gottfried Haberler, C. Reinold Noyes, and GeorgeSoule, read the manuscript with critical care. So also didMoses Abramovitz, Daniel Creamer, and Ruth Mack of theresearch staff. Of the many others who have made a contribu-tion to this book, either by serving at one time or another asMitchell's assistants or aiding me at the editing stage, I wishto mention particularly Martha Anderson, Cicely Applebaum,Florence Cohen, Dorothy Cook, Sally Edwards, HarryEisenpress, Frances Goldberg, H. Irving Forman, SimonKuznets, Karl Laubenstein, Sophie Sakowitz, Regina S. Sands,Julius Shiskin, Johanna Stern, and the late Denis Volkenau.
A.F.B.
CONTENTS
PAGE
INTRODUCTION, by Arthur F. Burns vii
Part IAims, Methods, and Materials
Chapter 1THE TASK ESSAYED BY THE NATIONAL BUREAU 3
Chapter 2OUR BASIC CONCEPT 6
Chapter 3OUR METHODS OF MEASURING CYCLICAL BEHAVIOR.. 9
I Specific Cycles 9II Relation of Specific to Business Cycles 10
III Reference Dates 10IV Reference-Cycle Bases and Relatives 13
V Treatment of Secular Movements 13VI Reference-Cycle Patterns 14VII Average Rates of Rise or Fall per Month 15
VIII Indexes of Conformity 15IX Our Use of Averages 18X Tests of Consiience 22XI Our Sample of Time Series 23
xxiii
XXIV CONTENTS
Part IIVarieties of Cyclical Behavior
Chapter 4 PAGE
A SAMPLE OF REFERENCE-CYCLE PATTERNS 29
Chapter 5VARIETIES OF CYCLICAL TIMING .. 51
I Our Measures of Cyclical Timing.. 51II Irregular Timing . 56
III Inverted Timing . .. 61
IV Neutral Timing .. .... 66
V Positive Timing .. 67
VI Leads and Lags 68VII Summary, Doubts, and the Next Step. 75
Chapter 6DEGREES OF CONFORMITY TO BUSINESS CYCLES 80
I Numerical Values of the Conformity Indexes.. 80II The Problem of Bias in the Indexes 83
III What Activities Conform Well and What Iii. 91IV Factors Influencing Degree of Conformity.. 95
Chapter 7VARIETIES OF REFERENCE-CYCLE AMPLITUDE 100
I Range and Distribution of Amplitudes 100II The Problem of Bias 101
IlL Amplitudes in Various Sectors of the Economy.. 104IV Factors That Influence Amplitudes 105
A Relations between specific- and reference-cycle amplitudes 105
B Relations between specific and businesscycles 110
C Cyclical fluctuations in consumers' purchases 114D Cyclical fluctuations in producers' purchases 120
CONTENTS xxvPAGE
E Cyclical fluctuations in employment ofresources 128
1 Land 1282 Labor 1303 Capital 137
F Cyclical amplitudes of prices and production 170
V The Problem of Comparing and Combining Ref-erence-Cycle Amplitudes 175
A Amplitudes of comprehensive series andtheir components 175
B Amplitudes of paired series and full samples 177C The amplitudes of business cycles 182
Chapter 8CYCLE-BY-CYCLE VARIABILITY IN CYCLICAL BEHAVIOR.. 185
I Problems Raised by Cycle-by-Cycle Variability 185
II Cycle-by-Cycle Variability within Series 186A Components of average deviations 186B Contribution of irregular movements 189C Contribution of secular movements 196D Contribution of cyclical movements 199E Statistical tests 200F Decomposition of average deviations 204G Estimates of components 207H Some conclusions 209I Average deviations of reference-cycle
amplitudes 216III Cycle-by-Cycle Variability among Series 218
A Influence of average amplitudes 219B Influence of regularity in cyclical timing. . 220C Statistical tests 224
IV The Representative Value of Cyclical Patterns. . 227
Technical Note I 231Technical Note II 236Technical Note III 246
XXVI CONTENTS
Part III•The Consensus of Cyclical Behavior
Chapter 9 PAGE
THE AIM OF PART III 25!
Chapter 10THE EVIDENCE OF COMPREHENSiVE SERiES. 255
I A Statistical Summary 255
II Scope and Adequacy of Comprehensive Series.. 255
III Typical Behavior of Comprehensive Series duringBusiness Cycles 277
A Cyclical timing 277B Conformity to business cycles. 281C Reference-cycle amplitudes . . 282D Reference-cycle patterns 288E Rates of change during successive segments
of reference cycles 296F Cycle-by-cycle differences 305
Appendix ASUPPLEMENT OF REFERENCE-CYCLE MEASURES 311
Appendix BSOURCES OF DATA 329
Iiidex 363
List of Tables
TABLE PAGE
1 Tentative Reference Dates of Business Cycles inthe United States 12
2 Specimen Indexes of Conformity 17
3 Basic Sample of 794 Monthly or Quarterly Seriesfor the United States Used in Later Tables 24
4 Distribution of 794 Monthly or Quarterly SeriesAccording to Their Characteristic CyclicalTiming 52
5 Economic Character of Series That Lead or Lagat Reference-Cycle Troughs and Peaks 70
6 Two Summaries of the Numerical Values of Con-formity Indexes of 794 Monthly or QuarterlySeries 81
7 Series with Business-Cycle Conformity Indexes of±100, 0, and Intermediate Values Classified by theNumber of Business Cycles They Cover 85
8 Groups of Series Ranked According to TheirIndexes of Conformity to Business Cycles 92
9 Average Conformity of Groups of Series Com-pared with Conformity of Inclusive Indexes orAggregates 97
10 Summaries of Average Reference-Cycle Ampli-tudes of 794 Monthly or Quarterly Series 102
11 Groups of Series Ranked According to TheirAverage Reference-Cycle Amplitudes 106
12 Average Reference-Cycle Amplitudes at Succes-sive Links in the Shoe-Leather-Hide Chain ofSupply 125
13 Reference-Cycle Timing, Conformity, and Am-plitude of 'Comprehensive' Series on Employment
xxvii
XXVIII LIST OF TABLES
TABLE PAGE
and Related Factors in Four Business Cycles, 1921-1938 134
14 Range of Variation of Balance-Sheet Ratios amongMinor Industrial Divisions, 1937 139
15 Reference-Cycle Bases, Timing, Conformity, andAmplitude of Kuznets' Estimates of CapitalFormation, Savings, and Related Magnitudes, atCurrent Prices, Four Business Cycles, 1921-1938 154
16 Reference-Cycle Timing, Conformity, and Am-plitude of Selected Series Relating to Investment 159
17 Reference-Cycle Conformity and Amplitude ofSeries on Prices and Production 173
18 Analysis of Differences between Mean Reference-Cycle Amplitudes of Full Samples and PairedSeries in Table 17 179
19 Effects of Smoothing Four Monthly Series uponTheir Average Reference-Cycle Patterns and theAverage Deviations from Them 191
20 Mean Average Deviations from Average Refer-ence-Cycle Patterns, by Groups of Series 202
21 Mean Average Deviations from Average Refer-ence-Cycle Patterns of Fñll Sample, and Theo-retical Expectations concerning Stage-by-StageChanges in Their Components-. 204
22 Estimates of the Secular, Cyclical, and IrregularComponents in Mean Average Deviations fromAverage Reference-Cycle Patterns of Full Sample 209
23 Ratios of Mean Average Deviations at ReferencePeaks to Mean Average Deviations at ReferenceTroughs, by Groups of Series 212
24 Average Deviations from Average Ref erence-Cycle Amplitudes of Seven Series ComputedDirectly and Indirectly 217
25' Mean Average Deviations from Average Refer-ence-Cycle Patterns of Three Groups of SeriesHaving Virtually Identical Reference-Cycle Am-plitudes and Conformity 223
LIST OF TABLES XXIX
TABLE PAGE
26 Summary of Relations among Average Deviationsof Reference-Cycle Patterns, Average Ampli-tudes, and Regularity of Cyclical Timing in 29Groups of Series 226
A Weights Implied by Macaulay's 43-Term Ap-proximately 5th-Degree Parabolic GraduationFormula 233
B Variance and Lagged Covariances of a SmoothedSeries Derived by Graduating a Purely RandomSeries of Unit Variance with Macaulay's 43-TermFormula 233
C Variances of Original and Smoothed Time Seriesand of Averages of Successive Items from Them 234
D Percentage Reduction of Standard Deviation Dueto Averaging 235
27 Changes in the Reference-Cycle Bases of Pig IronProduction Before and After Adjustment forSecular Trend 239
28 Average Deviations from Average Ref erence-Cycle Patterns of Nine Series Before and AfterAdjustment for Secular Trend 242
29 Effect of Shift from T-T to P-P Analysis uponAverage Deviations from Average Ref erence-Cycle Patterns of Seven Series 244
30 Illustrations of Relations between Standings atStages IX and I of Adjacent Reference Cycles inShares Traded on New York Stock Exchange... 245
31 Typical Behavior of Comprehensive AmericanSeries with Respect to Business Cycles 256
32 Summary of Typical Movements of 34 Monthlyor Quarterly Comprehensive Series in SuccessiveReference-Cycle Stages 279
33 Summaries of Numerical Values of ConformityIndexes of 34 Comprehensive Series and of FullSample of Monthly or Quarterly Series 283
XXX LIST OF TABLES
TABLE PAGE
34 Summaries of Reference-Cycle Amplitudes of 34Comprehensive Series and of Full Sample ofMonthly or Quarterly Series 284
35 Array of Average Reference-Cycle Amplitudesof Comprehensive Series 286
36 Stage-by-Stage Distribution of Average Refer-ence-Cycle Standings of 34 Monthly or QuarterlyComprehensive Series 289
37 Comprehensive Series Having Very Low, Middle,and Very High Average Standings at the NineStages of Reference Cycles 292
38 Stages in Which Comprehensive Series Haxe Ex-treme and Middle Standings 294
39 Segment-by-Segment Distribution of AverageRates of Change in 34 Monthly or QuarterlyComprehensive Series 298
40 Change from Segment to Segment Of ReferenceCycles in the Typical Direction and Rate of Move-ment of 34 Monthly or Quarterly ComprehensiveSeries 301
41 Three Sets of Observations upon Consensus in theMovements of 34 Monthly or Quarterly Compre-hensive Series during Successive Segments of Ref-erence Cycles 307
42 Reference-Cycle Measures of Series in Chart 1.. 312
43 Supplement to Table 31 and Chart 7 326
List of Charts
CHART PAGE
1 Average Reference-Cycle Patterns of a Sample ofSeries 3 2
2 Percentage Distribution of 794 Series Accordingto Their Characteristic Cyclical Timing 54
3 Percentage Distribution of 794 Series Accordingto the Numerical Value of Their Indexes of Con-formity to Business Cycles 82
4 Percentage Distribution of 184 Series Covering 10or More Cycles According to the Numerical Valueof Their Indexes of Conformity to Business Cycles 87
5 Percentage Distribution of 794 Series Accordingto the Numerical Value of Their Average Refer-ence-Cycle Amplitudes 101
6 Average Reference-Cycle Patterns of Riw andSmoothed Data of Four Series 190
7 Average Reference-Cycle Patterns of AmericanComprehensiveSeries .... 264
xxxi