Professional Documents
Culture Documents
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Ohio State University Press is collaborating with JSTOR to digitize, preserve and extend access to Journal of
Money, Credit and Banking.
http://www.jstor.org
ROBERT J. HODRICK
EDWARD C. PRESCOTT
PostwarU.S. BusinessCycles:
An EmpiricalInvestigation
We propose a procedurefor representinga time series as the sum of a smoothly varying trendcomponentand a cyclical component.We documentthe natureof the comovements of the cyclical componentsof a varietyof macroeconomictime series. We
find that these comovements are very differentthan the correspondingcomovements
of the slowly varying trendcomponents.
THE PURPOSE
OFTHISARTICLE
is to documentsomefeatures
of aggregateeconomicfluctuations
sometimesreferredto as businesscycles. The
investigationuses quarterly
datafromthe postwarU.S. economy.The fluctuations
studiedarethosethataretoo rapidto be accountedfor by slowlychangingdemographicandtechnologicalfactorsandchangesin the stocksof capitalthatproduce
seculargrowthin outputpercapita.
As Lucas(1981)hasemphasized,aggregateeconomicvariablesin capitalisteconomiesexperiencerepeatedfluctuations
abouttheirlong-termgrowthpaths.Priorto
Keynes'General Theory, the studyof theserapidfluctuations,
combinedwiththe
attemptto reconcilethe observationswith an equilibrium
theory,was regardedas
themainoutstanding
challengeof economicresearch.AlthoughtheKeynesianRevSupport of the National Science Foundationis acknowledged. We also acknowledge helpful comments by the participantsat the 1979 SummerWarwickWorkshopon Expectationand the money workshops at the Universities of Chicago and Virginia and at Carnegie-MbllonUniversity.In particular,we
thankRobertAvery, V.V. Chari, Lars Peter Hansen, CharlesR. Nelson, Thomas J. Sargent, KennethJ.
Singleton, and JohnH. Wood for comments. We also thankthe WhartonEconomic ForecastingAssociates for providingthe data.
This paperis substantiallythe same as our 1981 workingpaper.The only majorchange to the paperis
the additionof an Appendixof Tablesthatmirrorour originalsand containdataending in 1993. Since we
did not updatethe citations, we apologize to the many authorswho have used the Hodrick-Prescottfilter
and studied its propertiesin the interveningeighteen years since its original development.
ROBERT
J. HODRICK
is NomuraProfessorof InternationalFinance at the GraduateSchool
of Business, ColumbiaUniversity.EDWARD
C. PRESCOTT
is Regents'Professorat the University of Minnesotaand Advisorto the FederalReserveBank of Minneapolis.Both are research
associates of the National Bureau of EconomicResearch.
Journal of MoneyvCreditvand Banking, Vol. 29, No. 1 (February1997)
Copyright1997 by The Ohio State University Press
olutionredirected
effortawayfromthisquestionto theone of determining
the level
of outputat a pointin timein disequilibrium,
thefailureof theKeynesianTheoryin
the 1970shas causedmanyeconomiststo wantto returnto the studyof business
cycles as equilibrium
phenomena.In theirsearchfor an equilibrium
modelof the
businesscycle, moderneconomistshavebeen guidedby the insightsof Mitchell
(1913)andotherswhohaveusedtechniquesof analysisthatweredevelopedpriorto
thedevelopment
of moderncomputers.Thethesisof thispaperis thatthesearchfor
an equilibrium
modelof the businesscycle is only beginningandthatstudyingthe
comovementsof aggregateeconomicvariablesusingan efficient,easilyreplicable
techniquethatincorporates
our priorknowledgeaboutthe economywill provide
insights into the featuresof the economy that an equilibriumtheory should
incorporate.
Thisstudyshouldbe viewedas documenting
somesystematicdeviationsfromthe
restrictions
uponobservations
impliedby neoclassicalgrowththeory.l Ourstatistical
approachdoes not utilizestandardtime seriesanalysis.Ourpriorknowledgeconcerningtheprocessesgenerating
thedatais notof thevarietythatpermitsusto specify
a probability
modelas required
forapplication
of thatanalysis.Weproceedin a more
cautiousmannerthatrequiresonly priorknowledgethatcanbe supported
by economictheory.Themaintained
hypothesis,basedupongrowththeoryconsiderations,
is thatthegrowthcomponent
of aggregateeconomictimeseriesvariessmoothlyover
time. The sensein whichit variessmoothlyis madeexplicitin section1.
Wefindthatthe natureof thecomovements
of thecyclicalcomponents
of macroeconomictimeseriesareverydifferentfromthecomovements
of theslowlyvarying
componentsof the corresponding
variables.Growthis characterized
by roughly
proportional
growthin (percapita)output,investment,consumption,
capitalstock
andproductivity
(outputperhour),andlittlechangein thehoursof employment
per
capitaor household.In contrast,thecyclicalvariations
in outputariseprincipally
as
the resultof changesin cyclical hoursof employmentand not as the resultof
changesin cyclicalproductivity
or capitalstocks.In thecase of the cyclicalcapital
stocks in both durableand nondurablemanufacturing
industries,the correlation
withcyclicaloutputis evennegative.Anotherdifferenceis in thevariability
of componentsof aggregatedemand.Cyclicalconsumption
variesonly one-halfandinvestmentthreetimesas muchas doescyclicaloutput.
Section2 presentsour findingsregardingthe comovementsof theseserieswith
the cyclicalcomponentof realGNP,as well as anexamination
of thecyclicalcomponentsof prices, interestrates,andnominalandrealmoneybalances.Section3
examinesthe serialcorrelation
propertiesof a numberof the series.
Severalresearchers,
usingalternative
methods,haveaddedandareaddingto our
knowledgeof aggregateeconomicfluctuations.2
Ourview is thatno one approach
dominatesall the othersandthatit is best to examinethe datafroma numberof
differentperspectives.We do think our approachdocumentssome interesting
regularities.
1. Lucas(1980)intexprets
theworkof Mitchell(1913)in a similarlight.
2. ExamplesincludeLitterman
andSargent(1979),NelsonandPlosser(1980),Neftci(1978),Sargent
andSims(1977),Sims(1980,a, b), andSingleton(1980).
: 3
1. DECOMPOSITION
PROCEDURE
The observedtime seriesareviewedas the sumof cyclicalandgrowthcomponents.Actually,thereis also a seasonalcomponent,butas the dataareseasonally
adjusted,thiscomponenthas alreadybeenremovedby thosepreparing
the dataseries. If growthaccountingprovidedestimatesof the growthcomponentwitherrors
thatweresmallrelativeto the cyclicalcomponent,computingthe cyclicalcomponentwouldbejusta matterof calculatingthedifferencebetweentheobservedvalue
andthe growthcomponent.Growththeoryaccounting(cf. Denison1974),in spite
of its considerable
success,is farfromadequateforprovidingsuchnumbers.If our
priorknowledgeweresufficientlystrongso thatwe couldmodelthegrowthcomponentas a deterministic
component,possiblyconditionalon exogenousdata,plus a
stochasticprocessandthe cyclicalcomponentas someotherstochasticprocess,estimatingthe cyclicalcomponentwouldbe an exercisein moderntimeseriesanalysis. Ourpriorknowledgeis not of this variety,so thesepowerfulmethodsarenot
applicable.Ourpriorknowledgeis thatthe growthcomponentvaries"smoothly"
overtime.
Ourconceptualframeworkis thata given time seriesYtis the sumof a growth
componentgt anda cyclicalcomponentct:
Yt=gtict
fort= 1, . . . ,T.
(1)
Ourmeasureof the smoothnessof the {gt} pathis the sum of the squaresof its
seconddifference.The ct aredeviationsfromgt andourconceptualframework
is
thatoverlongtimeperiods,theiraverageis nearzero.Theseconsiderations
leadto
the followingprogramming
problemfordetermining
the growthcomponents:
T
(2)
4 : MONEY, CREDIT,ANDBANKING
The dataanalyzed,with the exceptionof the interestrates,are in naturallogarithmsso thechangein thegrowthcomponent,gt - gt_l, corresponds
to a growth
rate.
The growthrateof labor'sproductivity
has variedconsiderably
overthisperiod
(see McCarthy1978).In the 1947-53 period,the annualgrowthratewas4.20 percent, in the 1953-68 period,2.61 percent,in the 1968-73 period,only 1.41 percent, andin the subsequent
periodit wasevensmaller.Partof thesechangescanbe
accountedfor by a changingcapital-labor
ratioandchangingcompositionof the
laborforce.But, as shownby McCarthy,
a sizableandvariableunexplained
componentremains,even aftercorrectingfor cyclicalfactors.The assumptions
thatthe
growthratehas been constantoverourthirty-year
sampleperiod,1950-79, is not
tenable.Toproceedas if it werewouldresultin errorsin modelingthegrowthcomponentandtheseerrorsarelikelyto be nontrivial
relativeto thecyclicalcomponent.
Forthisreason,an infinitevalueforthe smoothnessparameter
was notselected.
Thefollowingprobability
modelis usefulforbringingto bearpriorknowledgein
the selectionof the smoothingparameterA. If the cyclicalcomponentsand the
seconddifferencesof the growthcomponentswere identicallyandindependently
distributed,
normalvariableswithmeanszeroandvariancescr2 andcr2 (whichthey
arenot),the conditionalexpectationof thegt, giventhe observations,
wouldbe the
solutionto program(2) when< = (rl/Cr2
As thisprobability
modelhasa statespacerepresentation,
efficientKalmanfiltering techniquescanbe usedto computethesegt.4By exploitingthe recursivestructure, one need not inverta (T + 2) by (T + 2) matrix(T is the numberof
observations
in the sample)as wouldbe necessaryif one solvedthelinearfirst-order
conditionsof program(2) to determinethe gt. The largestmatrixthatis inverted
usingthe Kalmanfilteringcomputational
approachis 2 by 2. If T is large,this is
importantbecauseinvertinglargematricesis costly and therecan be numerical
roundingproblemswhenimplemented
on computers.Kalmanfilteringcanbe performedwithcomputerpackagesthatarewidelyavailable.
Ourpriorview is thata 5 percentcyclicalcomponentis moderately
large,as is a
one-eighthof 1 percentchangein the growthratein a quarter.Thisled us to select
<
= 5/(1/8) = 40 or A = 1,600as a valueforthesmoothingparameter.
Oneissue
is, how sensitivearethe resultsto the valueof A thatis selected?To explorethis
issue, variousothervaluesof A weretried.Table1 containsthe (sample)standard
deviationsandautocorrelations
of cyclicalrealGNPfor the selectedvaluesof the
smoothingparameter
as well as statisticsto testforthepresenceof a unitrootin the
cyclicalcomponents.5
Thesenumberschangelittleif Ais reducedby a factorof four
4. This minimizationhas two elements, gOand gO- g_i, which are treatedas unknownparameters
with diffuse priors. The Kalmansmoothingtechnique(see Pagan 1980) was used to compute efficiently
the conditionalexpectationsof the g,, given the observedy,. The posteriormeans of gOand gO- g_ 1 are
the generalizedleast squaresestimates. The conditionalexpectationof the g, for t 2 1 are linearfunctions
of these parametersand the observations.
5. The tests for the presence of a unit root are augmentedDickey-Fullertests in which the change in
the cyclical componentis regressedon a constant, the level of the cyclical component,and six lags of the
TABLE
STANDARD
THE
DEVIATION
SMOOTHING
AND
SERIAL
PARAMETER;
CORRELATIONS
SAMPLE
PERIOD:
A = 400
Standard
A = 1600
1.56%
Deviations
GNPFOR
OF CYCLICAL
DIFFERENT
VALUES
OF
1950.1-1979.2
A = 6400
1.80%
A = infinity
2.03%
3.12%
Autocorrelations
Order
2
3
4
.48
.15
- .14
Order
-.32
Order
6
7
8
- .39
- .42
- .44
Order
Order
10
Order
Order
Order
Order
Unit-Root
.84
.57
.27
- .01
-.20
- .30
- .38
-.44
-.44
-.41
- 4.47
.80
Order
-.41
.36
Test
5.02
.87
.65
.41
.17
.00
- .11
- .20
-.27
-.31
-.32
- 3.57
.94
.84
.73
.61
.52
.44
.38
.31
.25
.20
- 1.15
gjt =
WitYji,
(3)
i=l
gjt
=
i=
change
Wl
in the cyclical
component.
t-statistic
of
(5 percent)
-2.89
(4)
Yjti
_00
on the
or -3.50
(l
One
rejects
level
of the cyclical
the presence
of a unit
component
root
is more
in the cyclical
negative
percent).
than
component
if the
the critical
value
(AX
:,
\
0|41 \t-01
AW
W
A T\
l As
-1
IV,,
-2
-3
-5
1950
1952
1954
1956
1958
1960
19621
1964
1966
YEAR
1968
1970
1972
1974
t976
1978
FIG. 1.
where
(s)
.8941l'llyj,l < m.
,=-r
cjt= aj +
E
i= -2
>iGNPt_i
(6)
Thefirstset of variablesstudiedaretherealaggregatedemandcomponents.The
resultsaresummarized
in Tables2 and3. Theseriesthatvarytheleastareconsumption of services,consumption
of nondurables
andstateandlocal government
purchasesof goodsandservices.Eachof thesehasstandard
deviationless thanthe 1.8
percentvalueforrealoutput.Theinvestmentcomponents,includingconsumerdurableexpenditures,are aboutthreetimes as variableas output.Covariabilities
of
consumption
andinvestmentwithoutputaremuchstrongerthanthecovariability
of
government
expenditures
withoutput.
TABLE
AGGREGATE
SAMPLE
DEMAND
PERIOD:
1950.
COMPONENTS:
STANDARD
DEVIATIONS
AND
CORRELATIONS
WITHGNP
1-1979.2
Average
Standard Deviations
Whole
Real
GNP
Total
Consumption
Services
Nondurables
Durables
Total
Invest.
Fixed
Residential
Nonresidential
Equipment
Structures
Total
Government
Federal
State
and
Local
1.8
1.3
.7
1.2
5.6
5.1
10.7
4.9
5.8
4.5
4.8
8.7
1.3
First Half
1.7
1.2
.7
1.0
6.1
4.2
8.5
4.4
5.6
3.8
6.5
11.6
1.6
Second Half
1.9
1.4
.6
1.3
5.0
5.9
12.4
5.3
5.9
5.1
2.2
4.2
1.0
opferReenalt
in Percents
First Half
Whole
.739
.615
.714
.574
.714
.436
.684
.707
.512
.258
.266
-.170
.503
.441
.575
.298
.454
.123
.554
.642
.225
.353
.377
-.408
Second Half
GNP
.917
.781
.808
.884
.884
.637
.777
.760
.698
.152
.125
.131
61.7
26.8
26.5
8.4
14.2
4.4
9.7
6.0
3.7
22.6
10.8
11.8
Factors of Production
TABLE 3
OFASSOCIATION
WITHGNP ANDMEASURE
STRENGTH
DEMAND
COMPONENTS:
AGGREGATE
OFSTABILITY
1950.1-1979.2
SAMPLE
PERIOD:
R2 for Regression
Correlationwith
Real Output
Squared
Total Consumption
Services
Nondurables
Durables
Total Invest. Fixed
Residential
Nonresidential
Equipment
Structures
Total Government
Federal
State and Local
.546
.378
.510
.329
.509
.190
.468
.500
.262
.067
.071
.029
cJ, = aJ +
2
E
fij,GNP,+
Stability Measure
.620
.424
.589
.415
.552
.441
.602
.631
.367
.119
.129
.095
.922
.877
.968
.829
.785
.809
.831
.908
.834
.509
.482
.298
TABLE
OF PRODUCTION:
FACTORS
SAMPLE
PERIOD:
STANDARD
DEVIATIONS
AND
Standard Deviations
Whole
Real
GNP
Capital
WITHGNP
CORRELATIONS
19SO.1-1979.2
First Half
in Percents
Whole
Second Half
1.8
1.7
1.9
1.7
1.2
.7
2.0
1.4
1.0
.7
1.8
.5
1.2
1.l
First Half
Second Half
Stocks
Employees
1.4
2.0
1.4
.7
2.1
.6
1.6
Productivity
1.0
1.0
Inventory
Capital
Stock
Durables
Capital
Stock
Nondurables
Hours
Work
.5
Week
.507
-.210
-.236
.853
.820
.773
.100
.309
-.274
-.297
.824
.800
.732
.361
.686
-.178
-.185
.896
.854
.831
-.231
TABLE
OF PRODUCTION:
FACTORS
SAMPLE
PERIOD:
1950.
STRENGTH
OF ASSOCIATION
WITH
GNP
AND
MEASURE
OF STABILITY
1-1979.2
R2 for Regression
Correlation with
Real Output
Squared
Capital
Inventory
Stock
Durables
Capital
Stock
Nondurables
Hours
Week
Employees
Average
fij,GNP,+I
t=-2
Stability Measure
Stocks
Capital
Work
2
cjt = aj +
Product
of Labor
.257
.044
.056
.728
.672
.600
.010
.622
.235
.129
.838
.700
.801
.453
.828
.782
.740
.954
.513
.935
.773
10 : MONEY,CRED1T,
ANDBANKING
TABLE 6
MONETARY
ANDPRICEVARIABLES:
STANDARD
DEVIATIONS
ANDCORRELATIONS
WITHGNP
SAMPLE
PERIOD:
1950.1-1979.2
su
Whole
Real GNP
M1
Nominal Value
Velocity
Real Value
M2
Nominal
Velocity
Real Value
InterestRates
Short
Long
Price Indexes
GNP Deflator
CPI
ndardDeviations in Percents
First Half
1.8
1.7
1.9
.9
1.6
1.5
.8
2.0
1.2
1,0
1.0
1. 1
1.9
1.8
2.4
1.4
.9
27
.24
.06
1.0
1.3
1.0
1.3
Second Half
Whole
First Half
Second Half
1.7
.661
.614
.565
.675
.801
.079
.649
.415
.865
1.3
1.2
2.1
.480
.529
.432
.175
.818
221
.665
.131
.828
510
.738
.640
.255
175
- .239
-.316
.490
.223
814
799
19
1.1
1.3
negativein thesecondhalfwiththecorrelation
fortheentireperiodbeingsmalland
negative.
3. SERIAL CORRELATIONPROPERTIESOF DATA SERIES
A sixth-order
autoregressive
processwas fit to a numberof the serieswhichdisplayedreasonablestablecomovementswithrealoutput.Figure2 presentsplotsof
the unitimpulseresponsefunctionsforGNPandnineotherseriesfortheestimated
TABLE 7
MONEYANDPECEVARIABLES:
STRENGTH
OFASSOCIATION
WITHGNP ANDMEASURE
OFSTABILITY
SAMPLE
PERIOD:
1950.1-1979.2
22
CoITelationwith
Real Output
Squared
for Regession
CJf= atJ +
2
E ISJIGNPr+I
t= -2
Stability Measure
M1
Nominal Value
ve OClty
Real Value
M2
. ..
.
.
> omlna va ue
Velocity
Real Value
InterestRates
Short
Long
Price Index
GNP Deflator
CPI
.
.437
.378
.319
.445
.408
.495
.378
.281
.678
.230
.280
.187
.371
.376
.428
.749
.650
.684
.260
.037
.506
.381
.748
.724
.261
.330
.567
.481
*057
.010
_
-
o
o
CSJJ
o
CONSUMPTION
GNP
o
o
V--
CM
310
20
11O
20
20
10
FIXEDINVESTMENT
HOUR
1b
110
1'o
do
3l0
0-
uz
EMPLOYEE
;
uz
o
o
o
WORKWEEK
\
o
uz
'
'-o
31
20
110
'
Po
o
-
of
M1
d
CM
w
210
'
- /
.NTORY
INVE
/-
UZ
cr.
CJURABLE
CAPITAL
--
UZ
-
20
30
UZ
o
o
o
1'0
20
r\
CM
3'0 b '
1o
CM
o
cM -
VELOCI5M1
D
o
30
10
20
3|0
ANDBANKING
12 : MONEY,CRED1T,
Ct=
Zia,_i
i=o
be the invertiblemoving averagerepresentation,parameter0, equals the value of the unit responsefunction in period i. One must take care in interpretingthe responsepattern.Two moving averageprocesses
can be observationallyequivalent (same autocovariancesfunction) yet have very differentresponsepatterns. We chose the invertiblerepresentationbecause it is unique. It is just one way to representthe serial
correlationpropertiesof a covariance stationarystochastic process. Others are the spectrum, the autoregressive representation,and the autocovariancefunction.
TABLEA1
STANDARD
DEVIATION
OF THE
SMOOTHING
Standard
Deviations
AND
SERIAL
PARAMETER.
CORRELATIONS
SAMPLE
PERIOD:
OF CYCLICAL
GNPFOR
1947.1-1993.4
A = 400
Autocorrelations
Order 1
Order2
Order3
Order4
OrderS
Order6
Order7
Order 8
Order9
Order 10
Unit-Root Test
TABLE
A = 1600
1.47%
DIFFERENT
A = 6400
1.80%
A = - infinity
2.14%
.86
.64
.39
.16
.81
.53
.22
-.03
-.21
-.32
-.39
-.43
-.40
-.35
-6.52
VALUES
44.94%
.9o
.96
.73
.53
.34
.18
.02
-.05
-.27
-.30
-.37
-.40
-.40
.91
.86
.80
.74
.69
.63
.58
.52
.47
-2.34
09
.19
26
28
98
-5.91
A2
RealGNP
Total Consumption
Services
Nondurable
Durables
Total Invest. Fixed
Residential
Nonresidential
Equipment
Structures
Total Government
Federal
State and Local
GNP
Average
opferReenalt
Whole
First Half
Second Half
Whole
First Half
Second Half
GNP
1.8
1.2
0.7
1.2
5.5
5.5
10.9
5.1
6.1
4.8
3.9
6.9
1.5
1.8
0.9
0.7
1.0
5.4
4.5
9.1
4.6
5.8
3.8
5.4
9.5
1.9
1.8
1.4
0.8
1.3
5.6
6.4
12.6
5.6
6.4
5.6
1.2
1.9
1.1
.719
.685
.707
.457
.732
.462
.746
.798
.469
.350
.348
- .216
.511
.544
.558
.112
.470
.755
.659
.715
.397
.515
.540
- .453
.875
.810
.827
.787
.927
.745
.820
.871
.528
-.012
- .164
.015
61.7
31.2
24.5
6.9
15.2
5.1
10.1
6.1
4.0
21.6
10.7
10.8
TABLE A3
AGGREGATE
DEMAND
COMPONENTS:
STRENGTH
OFASSOCIATION
WITHGNP ANDMEASURE
OFSTABILITY
SAMPLE
PERIOD:
1947. 1-1993.4
R2 for Regression
Correlationwith
Real Output
Squared
Total Consumption
Services
Nondurables
Durables
Total Invest. Fixed
Residential
Nonresidential
Equipment
Structures
Total Government
Federal
State and Local
.517
.469
.500
.209
.536
.213
.557
.637
.220
.123
.121
.047
Cj, = 0tj +
2
22
jiGNP,+
Stability Measure
.571
.512
.520
.324
.580
.482
.662
.702
.396
.229
.224
.080
.808
.873
.872
.669
.796
.731
.929
.955
.792
.500
.436
.200
TABLEA4
WITHGNP
FACTORSOF PRODUCTION:STANDARDDEVIATIONSAND CORRELATIONS
SAMPLEPERIOD:1947.1-1993.4
withRealOutput
Correlations
SecondHalf
Whole
FirstHalf
Standard
Deviations
in Percents
FirstHalf
SecondHalf
Whole
RealGNP
CapitalStocks
Inventory
CapitalStockDurables
CapitalStockNondurables
Hours
WorkWeek
Employees
Productivity
TABLE
1.8
1.8
1.8
2.1
1.2
1.0
1.8
1.1
1.5
0.9
2.4
1.1
1.0
1.9
1.1
1.6
1.0
1.8
1.2
0.9
1.7
1.0
1.5
0.8
.510
.510
- .055
.883
.778
.828
.239
.547
.387
- .125
.860
.778
.808
.151
.475
.619
.021
.911
.783
.850
.360
AS
GNPAND
MEASURE OF STABILITY
R2 for Regression
CoITelationwith
Real Output
c,, - sx, +
'
Squared
Capital Stocks
Inventory
Capital Stock Durables
Capital Stock Nondurables
Hours
Work Week
Employees
Average Productof Labor
J t=-2
,,GNP,+I
Stability
Measure
'
.801
.967
.874
.992
.994
.989
.933
.373
.728
.356
.869
.764
.858
.465
.260
.260
.003
.779
.605
.685
.057
TABLE A6
WITH GNP
ANDCORRELATIONS
DEVIATIONS
STANDARD
MONETARY
ANDPRICEVARIABLES:
1947.1-1993.4
SAMPLE
PERIOD:
Real GNP
M1
Nominal Value
Velocity
Real Value
M2
Nominal
Velocity
Real Value
InterestRates
Short
Long
Price Indexes
GNP Deflator
CPI
Deviations
inPercents
Standard
Second
Hif
Whole
FirstHif
1.8
1.8
1.8
wole
-
with RealOuWut
colTelations
Second
Hif
FirstHif
-
2.1
2.7
2.7
1.3
2.1
1.6
2.7
3.1
3.4
.368
.328
.347
.542
.680
.219
.318
.104
.436
1.8
2.5
2.4
1.4
2.5
1.8
2.2
2.6
2.9
.337
.404
.319
.324
.672
.058
.357
.151
.49l
1.1
0.6
0.6
0.2
1.5
0.8
.324
.032
.335
.228
.358
-.020
1.0
1.6
1.0
1.4
1.0
1.7
.327
.247
-.635
-.585
-.156
-.222
TABLE
A7
MONEY
AND
PRICE
VARIABLES:
STRENGTH
OF ASSOCIATION
WITH
GNP
AND
MEASURE
OF STABILITY
SAMPLE
PERIOD:
1947.1-1993.4
CoIIelation with
Real Output
Squared
R2forRegression
2
cJ,= (xJ+ E J,GNP,+l
i=-2
StabilityMeasure
.783
M1
Nominal
Value
.135
.229
Velocity
.108
.280
.747
Real
.120
.270
.738
Value
M2
Nominal
Value
.114
.291
.782
Velocity
.163
.377
.755
Real
Value
.102
.321
.707
Interest
Rates
Short
.105
.336
.701
Long
.001
.191
.701
.024
.199
.430
.049
.248
.485
Price
Index
GNP
Deflator
CPI
LITERATURE
CITED
16
Sims,Christopher
A. "Macroeconomics
andReality."Econometrica48 (1980a),1-48.
. "Comparison
of Interwar
andPostwarBusinessCycles:Monetarism
Reconsidered."
AmericanEconomicReview 70 (1980b),250-57.
Singleton,KennethJ. "RealandNominalFactorsin theCyclicalBehaviorof InterestRates,
Output,andMoney."Carnegie-Mellon
UniversityWorking
Paper,1980.
Stigler,StephenM. "Mathematical
Statisticsin the EarlyStates."Annals of Statistics 6
(1978),239-65.
Whittaker,
EdmundT. "Ona New Methodof Graduations."
Proceedings of the Edinburgh
MathematicalSociety 41 (1923),63-75.