The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Board of Governors of the Federal Reserve System
International Finance Discussion Papers
Number 591
September 1997
Money, Sticky Wages, and the Great Depression
Michael D. Bordo, Christopher J. Erceg, and Charles L. Evans
NOTE: International Finance Discussion Papers are preliminary materials circulated to
stimulate discussion and critical comment. References in publications to International Finance
Discussion Papers (other than an acknowledgement that the writer has had access to
unpublished material) should be cleared with the author or authors. IFDPs are available on
the WEB at www.bog.frb.fed.us.
Money, Sticky Wages, and the Great Depression
Michael D. Bordo, Christopher J. Erceg, and Charles L. Evans*
Abstract
This paper examines the ability of a simple stylized general equilibrium model that
incorporates nominal wage rigidity to explain the magnitude and persistence of the Great
Depression in the United States. The impulses to our analysis are money supply shocks. The
Taylor contracts model is surprisingly successful in accounting for the behavior of major
macroaggregates and real wages during the downturn phase of the Depression, i.e., from
1929:3 through mid-1933. Our analysis provides support for the hypothesis that a monetary
contraction operating through a sticky wage channel played a significant role in accounting
for the downturn, and also provides an interesting refinement to this explanation. In
particular, both the absolute severity of the Depression’s downturn and its relative severity
compared to the 1920-21 recession are likely attributable to the price decline having a much
larger unanticipated component during the Depression, as well as less flexible wage-setting
practices during this latter period. Another finding casts doubt on explanations for the
1933-36 recovery that rely heavily on the substantial remonetization that began in 1933.
Keywords: Dynamic General Equilibrium Model, Sticky Wages
* Bordo is a professor of economics at Rutgers University and NBER fellow
(bordo@rci.rutgers.edu), Erceg is a staff economist in the Division of International Finance of
the Federal Reserve Board (ercegc@frb.gov), and Evans is an assistant vice president and
economist in the Research Department of the Federal Reserve Bank of Chicago
(cevans@frbchi.org). The views in this paper are solely the responsibility of the authors and
should not be interpreted as reflecting the views of the Board of Governors of the Federal
Reserve System or of any other person associated with the Federal Reserve System. The
authors thank Ben Bernanke, Steve Cecchetti, Chris Hanes, Evan Koenig, Robert Kollmann,
Andrew Levin, Prakash Loungani, Sergio Rebelo, Anna Schwartz, John Taylor, and Michael
Woodford for comments on an earlier draft, as well as seminar participants at the Federal
Reserve Board, the Federal Reserve Bank of New York, and the NBER Monetary Economics
Group.
)NTRODUCTION
! FULL UNDERSTANDING OF THE FACTORS WHICH LED TO THE 'REAT $EPRESSION IN THE 5NITED 3TATES
CONTINUES TO ELUDE ECONOMISTS -ANY CONTRIBUTING FACTORS MAY HAVE BEEN IMPORTANT Õ
INCLUDING THE STOCK MARKET CRASH OF -ISHKIN AND 2OMER THE NONMONETARY
EdECTS OF BANKING PANICS "ERNANKE A DRAMATIC INCREASE IN WORLD TARIdS -ELTZER
AND #RUCINI AND +AHN ADHERENCE TO THE GOLD STANDARD %ICHENGREEN A AND 4EMIN
AND AN AUTONOMOUS DROP IN CONSUMPTION 4EMIN .EVERTHELESS &RIEDMAN AND
3CHWARTZÚS HYPOTHESIS THAT THE FAILURE OF 53 MONETARY POLICY TO OdSET BANK PANIC
INDUCED DECLINES IN THE MONEY SUPPLY WAS THE PRIMARY CAUSE OF THE 53 DOWNTURN BETWEEN
AND PROBABLY REMAINS THE MOST WIDELY SUBSCRIBED EXPLANATION
4HE RECENT LITERATURE WORKING WITHIN THE CONTEXT OF THE MONETARY EXPLANATION HAS FO
CUSED ON THE MECHANISM BY WHICH THE MONETARY COLLAPSE WAS TRANSMITTED TO THE REAL ECON
OMY 7HAT IS IN CONTENTION IS WHETHER THE REAL EdECTS OF THE MONETARY CONTRACTION WERE
ATTRIBUTABLE TO NOMINAL WAGE RIGIDITIES "ERNANKE AND #AREY TO INCREASED BANKRUPT
CIES ASSOCIATED WITH DEBT DEâATION &ISHER OR TO A SEVERE DISRUPTION IN THE PROCESS OF
çNANCIAL INTERMEDIATION DUE TO LARGE SCALE BANK FAILURES "ERNANKE
2ECENT EMPIRICAL RESEARCH HAS FOUND SUPPORT FOR THE HYPOTHESIS THAT MONETARY SHOCKS
OPERATING THROUGH A STICKY WAGE CHANNEL PLAYED AN IMPORTANT ROLE IN ACCOUNTING FOR THE
$EPRESSION IN THE 5NITED 3TATES AND ELSEWHERE )N AN IMPORTANT PAPER %ICHENGREEN AND
3OME
RECENT SURVEYS OF THE $EPRESSION PERIOD ARE CONTAINED IN PAPERS BY "ORDO "ORDO #HOUDHRI
AND 3CHWARTZ #ALOMIRIS %ICHENGREEN B AND 2OMER
3ACHS FOUND THAT REAL WAGES TENDED TO BE HIGHER AND INDUSTRIAL PRODUCTION LOWER
AMONG COUNTRIES THAT REMAINED ON THE GOLD STANDARD 4HESE AUTHORS INTERPRETED THEIR RESULTS
AS CONSISTENT WITH THE STICKY WAGE HYPOTHESIS )NSOFAR AS COUNTRIES THAT REMAINED ON GOLD
EXPERIENCED COMPARATIVELY LARGER PRICE DECLINES THE STICKY WAGE HYPOTHESIS PREDICTS THAT
THEY WOULD ALSO TEND TO EXPERIENCE LARGER INCREASES IN REAL WAGES AND CORRESPONDINGLY LARGER
OUTPUT CONTRACTIONS "ERNANKE AND "ERNANKE AND #AREY CORROBORATED THESE
PREDICTIONS EMPIRICALLY USING PANEL DATA ON A LARGER SET OF COUNTRIES THAN THE %ICHENGREEN
3ACHS STUDY "ERNANKE AND #AREY ALSO TRIED TO ASSESS THE EXTENT TO WHICH THE PRICE DECLINE
OPERATED THROUGH A STICKY WAGE CHANNEL OR THROUGH OTHER MECHANISMS SUCH AS DEBT DEâATION
4HEY CONCLUDED THAT ÝOUR EVIDENCE FAVORS THE VIEW THAT STICKY WAGES WERE THE DOMINANT
SOURCE OF NONNEUTRALITY P Þ
/UR PAPER FOCUSES ON THE ABILITY OF THE STICKY WAGE HYPOTHESIS TO EXPLAIN THE 'REAT $E
PRESSION IN THE 5NITED 3TATES OVER THE PERIOD 7HILE THE PREVIOUSLY MENTIONED CROSS
COUNTRY REGRESSION RESULTS SERVE AS AN IMPORTANT MOTIVATION FOR OUR ANALYSIS THOSE çNDINGS
DO NOT QUANTIFY THE EXTENT TO WHICH THE STICKY WAGE CHANNEL CAN ACCOUNT FOR THE MAGNITUDE
OF THE $EPRESSION AND ITS PERSISTENCE 4HROUGH SIMULATIONS OF A GENERAL EQUILIBRIUM MODEL
OUR PAPER PROVIDES A QUANTITATIVE ASSESSMENT OF THE EXTENT TO WHICH MONETARY SHOCKS WORK
ING THROUGH THE STICKY WAGE CHANNEL CAN ACCOUNT FOR THE SEVERITY OF THE $EPRESSION AND ITS
DURATION
4HE ECONOMY WE STUDY HAS A FAIRLY TYPICAL NEOCLASSICAL STRUCTURE WITH A REPRESENTATIVE
AGENT AND CAPITAL ACCUMULATION &OR SIMPLICITY MONEY IS ASSUMED TO BE VALUED BY AGENTS
THROUGH A MONEY IN THE UTILITY FUNCTION SPECIçCATION !S ALWAYS A CONTENTIOUS PART OF THE
ANALYSIS IS HOW TO INTRODUCE MONETARY NONNEUTRALITIES WE ASSUME THAT NOMINAL WAGES ARE
SET IN ADVANCE AND INVESTIGATE THE IMPLICATIONS OF TWO DIdERENT TYPES OF WAGE SETTING MECH
ANISMS 4HE çRST SPECIçCATION ASSUMES THAT WAGES ARE SET IN FORWARD LOOKING &ISCHER 'RAY
CONTRACTS EG 'RAY &ISCHER +ING #HO AND 0HANEUF 4HE ALTERNA
TIVE SPECIçCATION ASSUMES THAT WAGES ARE SET THROUGH 4AYLOR CONTRACTS 4AYLOR 4HE
ADVANTAGE OF THE LATTER SPECIçCATION IS THAT IT CAN POTENTIALLY ACCOMMODATE MORE PERSISTENT
LABOR MARKET DISEQUILIBRIUM ARISING FROM MONETARY DISTURBANCES AS THE NOMINAL WAGE SET IN
THE CURRENT PERIOD IS IN PART ÝANCHOREDÞ BY WAGE CONTRACTS IN THE PAST THAT ARE STILL IN EdECT
4HE MODEL IS SIMULATED OVER THE PERIOD BY INPUTTING ESTIMATES OF EXOGENOUS
INNOVATIONS TO THE MONEY GROWTH RATE - INTO THE STATE SPACE REPRESENTATION OF THE LIN
EARIZED MODEL -ODEL SIMULATIONS OF OUTPUT HOURS WORKED CONSUMPTION AND INVESTMENT
ARE COMPARED TO ACTUAL DATA
"OTH THE &ISCHER AND 4AYLOR CONTRACT VARIANTS OF OUR MODEL ARE SUCCESSFUL IN TRACKING
THE DOWNTURN IN OUTPUT AND HOURS WORKED BETWEEN AND EARLY &OR EXAMPLE
BOTH MODELS IMPLY AN OUTPUT DECLINE BETWEEN THE ONSET OF THE $EPRESSION IN AND
THE çRST HALF OF THAT IS WITHIN A COUPLE OF PERCENTAGE POINTS OF THE DROP THAT WAS
OBSERVED IN LOGARITHMIC PERCENTAGE TERMS 4HE 4AYLOR SPECIçCATION IS MORE SUCCESSFUL THAN
THE &ISCHER SPECIçCATION IN ACCOUNTING FOR THE PERSISTENCE OF THE DOWNTURN IN THE
PERIOD 4HE &ISCHER SPECIçCATION COUNTERFACTUALLY IMPLIES THAT THE STABILIZATION IN MONETARY
&OR
SIMPLICITY WE SUBSEQUENTLY REFER TO THESE CONTRACTS AS Ý&ISCHER CONTRACTSÞ
AGGREGATES IN SHOULD HAVE GENERATED A STRONG RECOVERY BY THE LATTER PART OF THAT YEAR
4HE 4AYLOR SPECIçCATION DOES BETTER BUT IT CANNOT ACCOUNT FOR THE CONTINUED DROP IN OUTPUT
THAT OCCURRED BETWEEN EARLY AND EARLY MOST OF WHICH WAS CONCENTRATED IN
THE $EPRESSION TROUGH %VEN SO THE 4AYLOR MODEL DRIVEN BY EXOGENOUS MONEY SUPPLY SHOCKS
CAN ACCOUNT FOR OF THE OUTPUT DECLINE OF THAT OCCURRED BETWEEN AND THE
çRST HALF OF 3IMILARLY THE 4AYLOR MODEL DOES WELL IN TRACKING THE BEHAVIOR OF HOURS
WORKED AND INVESTMENT THROUGH EARLY THOUGH IT MISSES SOME OF THE ADDITIONAL DECLINE
IN THESE VARIABLES THAT OCCURRED IN THE SUBSEQUENT YEAR )T PROVIDES A GOOD CHARACTERIZATION
OF CONSUMPTION BEHAVIOR THROUGH THE ENTIRE DOWNTURN PHASE OF THE $EPRESSION
&OUR KEY POINTS EMERGE FROM OUR ANALYSIS OF THE ABILITY OF THE STICKY WAGE MODEL TO
EXPLAIN THE DOWNTURN PHASE OF THE $EPRESSION &IRST THE MECHANISM THROUGH WHICH THE
STICKY WAGE MODEL PREDICTS THE MONETARY DECLINE WOULD AdECT THE MACROAGGREGATES SEEMS
CONSISTENT WITH THE DATA OUR MODEL DOES A GOOD JOB OF TRACKING WAGE BEHAVIOR OVER THE
PERIOD CORRECTLY IMPLYING A PRONOUNCED RISE IN THE REAL WAGE BETWEEN AND MID
FOLLOWED BY A GRADUAL DECLINE UNTIL MID 3ECOND OUR MODEL SIMULATIONS AND SENSITIVITY
ANALYSIS SUGGEST THAT IN ADDITION TO SLUGGISH WAGE ADJUSTMENT A LARGELY UNANTICIPATED PRICE
DECLINE PLAYED A SIGNIçCANT ROLE IN ACCOUNTING FOR THE SEVERITY OF THE $EPRESSION /UR RESULTS
INDICATE THAT THE $EPRESSION WOULD HAVE BEEN CONSIDERABLY LESS SEVERE IN ITS REAL EdECT
EVEN WITH AN EQUALLY LARGE PRICE DECLINE HAD A GREATER COMPONENT OF THE PRICE DECLINE BEEN
ANTICIPATED OR IF NOMINAL WAGES HAD ADJUSTED MORE RAPIDLY TO THE REDUCTION IN LABOR HOURS
4HIRD OUR COMPARATIVE ANALYSIS OF THE RECESSION FURTHER HIGHLIGHTS THE LIKELY JOINT
IMPORTANCE OF THE UNANTICIPATED CHARACTER OF THE PRICE DECLINE AND EXTREMELY SLUGGISH NOMINAL
WAGE ADJUSTMENT DURING THE PERIOD )N THE RECESSION OUTPUT FELL ONLY HALF AS
MUCH IN PERCENTAGE TERMS AS THE 'REAT $EPRESSION PERIOD DESPITE SIMILAR SIZED CUMULATIVE
DECLINES IN THE PRICE LEVEL 4HE LITERATURE HAS SUGGESTED THAT A MUCH LARGER COMPONENT OF THE
PRICE DECLINE WAS ANTICIPATED DURING THE EARLIER PERIOD AND THAT WAGES ADJUSTED MORE RAPIDLY
TO AGGREGATE DEMAND TWO FACTORS THAT COULD HELP EXPLAIN THE DIdERENT EXPERIENCES OF THE TWO
PERIODS /UR SIMULATIONS INDICATE THAT THE OUTPUT DIdERENCES ACROSS THE TWO PERIODS CAN BE
RATIONALIZED IF MUCH OF THE DEâATION WERE ANTICIPATED AND NOMINAL WAGES ADJUSTED
MORE RAPIDLY IN THIS EARLIER PERIOD
&OURTH OUR ANALYSIS OF THE PERIOD CASTS DOUBT ON 2OMERÚS EXPLANATION
OF THE RECOVERY PHASE 2OMER ARGUES THAT THE RECOVERY WAS LARGELY ATTRIBUTABLE TO
THE SUBSTANTIAL AND SUSTAINED INCREASE IN MONETARY AGGREGATES THAT BEGAN IN /UR
MODEL AND 2OMERÚS HYPOTHESIS BOTH IMPLY A STRONG RECOVERY IN ECONOMIC ACTIVITY DUE TO
THE STIMULATIVE EdECT OF THE REMONETIZATION AND ITS AdECT ON THE PRICE LEVEL (OWEVER THE
MECHANISM THROUGH WHICH OUR MODEL GENERATES A RECOVERY Õ NAMELY A FALL IN REAL WAGES Õ
IS BADLY AT ODDS WITH THE DATA )N FACT REAL WAGES ROSE SHARPLY BEGINNING IN MID AS
NOMINAL WAGES ROSE MORE THAN PRICES 4HE LARGE RISE IN NOMINAL WAGES APPEARS TO HAVE BEEN
LARGELY EXOGENOUS IN NATURE REâECTING THE IMPLEMENTATION OF VARIOUS LEGISLATION INCLUDING
THE .ATIONAL )NDUSTRIAL 2ECOVERY !CT .)2! .)2! BASICALLY MANDATED WAGE SCHEDULES FOR
VARIOUS INDUSTRIES 4HUS WHILE THE MONETARY EXPANSION OF PRESUMABLY MITIGATED
THE ADVERSE EdECTS OF THE NOMINAL WAGE INCREASES WE CONCLUDE THAT THE MONETARY EXPANSION
WAS NOT NEARLY STRONG ENOUGH TO LEAD THE 53 ECONOMY OUT OF THE $EPRESSION 'IVEN THAT
OTHER FACTORS INCLUDING THOSE EMPHASIZED BY "ERNANKE PROBABLY CONSPIRED WITH THE
REAL WAGE INCREASE IN ORDER TO SLOW THE PACE OF THE ECONOMIC RECOVERY IT REMAINS FOR FURTHER
RESEARCH TO PROVIDE A CONVINCING ACCOUNT OF THE FACTORS THAT PRODUCED THE ECONOMIC RECOVERY
FROM THE 'REAT $EPRESSION
4HE REMAINDER OF THIS PAPER IS ORGANIZED AS FOLLOWS 3ECTION DISCUSSES THE BASIC MODEL
AND PRESENTS THE TWO ALTERNATIVE WAGE CONTRACT STRUCTURES 3ECTION EXAMINES THE IMPULSE
RESPONSE FUNCTIONS FROM A MONETARY DISTURBANCE WHILE SECTION COMPARES SIMULATIONS OF
EACH MODELÚS IMPLICATIONS FOR MAJOR MACROAGGREGATES TO THE CORRESPONDING DATA 3ECTION
ATTEMPTS TO ACCOUNT FOR DIdERENCES IN THE RESPONSE OF OUTPUT TO SIMILAR SIZED PRICE DECLINES
THAT OCCURRED IN THE RECESSION AND IN THE $EPRESSION 3ECTION CONCLUDES THE PAPER
4HE BASIC MODEL WITH WAGE CONTRACTS
4HE MODEL USED TO INVESTIGATE THE HYPOTHESIS THAT MONEY OPERATING THROUGH A STICKY WAGE
CHANNEL CAUSED THE 'REAT $EPRESSION HAS A FAIRLY STANDARD NEOCLASSICAL STRUCTURE ! REPRE
SENTATIVE AGENT MAKES CONSUMPTIONINVESTMENT DECISIONS AND PORTFOLIO CHOICES IN A MANNER
CONSISTENT WITH UTILITY MAXIMIZATION SUBJECT TO AN INçNITE HORIZON BUDGET CONSTRAINT 4HE
PRIMARY DEPARTURE FROM A ÝTYPICALÞ NEOCLASSICAL SPECIçCATION IS THAT OUR MODEL ALLOWS FOR
MONETARY NONNEUTRALITY BY DROPPING THE ASSUMPTION OF CONTINUOUS LABOR MARKET CLEARING
)NSTEAD WAGES ARE DETERMINED BY &ISCHER OR 4AYLOR CONTRACTS )N THE SHORT RUN Õ THE PERIOD
BEFORE LABOR CONTRACTS CAN BE RENEGOTIATED Õ HOUSEHOLDS ARE ASSUMED TO SUPPLY ALL LABOR
DEMANDED AT THE GIVEN WAGE RATE SO THAT THE QUANTITY OF LABOR ACTUALLY HIRED IS DETERMINED
BY THE DEMAND FOR LABOR SCHEDULE OF THE REPRESENTATIVE çRM
/UR RATIONALE IN CONSIDERING TWO DIdERENT CONTRACT STRUCTURES IS THAT THEY PROVIDE TWO
RATHER EXTREME PERSPECTIVES ON THE STICKY WAGE TRANSMISSION MECHANISM )N PARTICULAR IT IS
WELL KNOWN THAT THE &ISCHER CONTRACT SPECIçCATION BASICALLY IMPLIES THAT MONETARY INNOVA
TIONS ONLY HAVE REAL EdECTS OVER THE CONTRACT INTERVAL WHICH IS ASSUMED TO LAST FOUR QUARTERS
IN OUR ANALYSIS "Y CONTRAST THE 4AYLOR WAGE CONTRACT SPECIçCATION CAN BE PARAMETERIZED TO
ALLOW MONETARY INNOVATIONS TO HAVE EdECTS THAT PERSIST WELL BEYOND THE CONTRACT INTERVAL
(OUSEHOLDS
4HE REPRESENTATIVE HOUSEHOLD SEEKS TO MAXIMIZE A UTILITY FUNCTIONAL OF THE FORM
8
n 4 B
S
S
,
/
S
S
S
WHERE n IS THE DISCOUNT FACTOR B IS REAL CONSUMPTION , IS NOMINAL CASH BALANCES AT THE END
OF PERIOD S AND / IS THE PRICE LEVEL 4HE PERIOD UTILITY FUNCTION IS GIVEN BY
4 B
S
,
P KM B
/
S
S
S
` P KM
,
/
S
S
4O FOCUS ATTENTION ON THE ROLE OF NOMINAL WAGE CONTRACTS IN DETERMINING LABOR MOVEMENTS
WE ABSTRACT FROM THE HOUSEHOLDÚS LABOR SUPPLY DECISION BY OMITTING LABOR FROM THE UTILITY
FUNCTION !S WE DISCUSS BELOW WE ASSUME THAT THE HOUSEHOLD HAS A TIME INVARIANT TARGET
VALUE OF LABOR HOURS DENOTED BY + AND THAT THE PROCESS OF WAGE ADJUSTMENT IMPLIED BY
EITHER OF THE TWO CONTRACT SPECIçCATIONS THAT WE EXAMINE IS CONSISTENT WITH THE HOUSEHOLD
SUPPLYING THIS QUANTITY OF LABOR IN THE LONG RUN !T ANY POINT IN TIME HOWEVER LABOR HOURS
WILL VARY FROM + DEPENDING UPON VARIATIONS IN THE DEMAND FOR LABOR
(OUSEHOLDS ACCUMULATE NOMINAL ASSETS ACCORDING TO THE FOLLOWING LAW OF MOTION
! ! `
S
S
1 ` ! `
S
S
6+
S
)*
S
S
S
{
S
7 ` , ` , `
S
S
S
/B
S
S
/(
S
S
WHERE ) IS THE RENTAL PRICE OF CAPITAL IN NOMINAL TERMS * IS THE CAPITAL SUPPLIED TO çRMS
6 IS THE NOMINAL WAGE RATE + IS TOTAL HOURS WORKED { IS NOMINAL çRM PROçTS 1 IS THE
NOMINAL INTEREST RATE ON BONDS ! IS NOMINAL BOND HOLDINGS 7 IS LUMP SUM CASH TRANSFERS
FROM THE GOVERNMENT AND ( IS GROSS REAL INVESTMENT 4HE çRST TERM IN PARENTHESES IN
EQUATION IS HOUSEHOLD NOMINAL INCOME (OUSEHOLD INCOME CONSISTS OF INTEREST INCOME
ON BONDS WAGE INCOME INCOME FROM CAPITAL çRM PROçTS AND LUMP SUM CASH TRANSFERS
FROM THE GOVERNMENT 4HE SECOND TERM IN PARENTHESES IS HOUSEHOLD NOMINAL EXPENDITURES
WHICH CONSISTS OF SPENDING ON CONSUMPTION AND INVESTMENT GOODS AND ON ACCUMULATING CASH
BALANCES
(OUSEHOLDS PURCHASE INVESTMENT GOODS IN ORDER TO INCREASE THEIR STOCK OF PHYSICAL CAPITAL
WHICH THEY IN TURN RENT TO çRMS
* ` p*
S
S
7E
(
S
HAVE EXAMINED VERSIONS OF OUR ECONOMY WHICH MODEL ENDOGENOUS LABOR SUPPLY CHOICES BY THE HOUSE
HOLD 4HOSE RESULTS ARE QUALITATIVELY SIMILAR TO OUR çNDINGS REPORTED HERE "Y ABSTRACTING FROM THE HOUSE
HOLDÚS ENDOGENOUS LABOR SUPPLY DECISION HOWEVER IT IS CLEAR THAT WE ARE NOT ATTEMPTING TO EXPLAIN THE 'REAT
$EPRESSION WITH LARGE INTERTEMPORAL LABOR SUPPLY MOVEMENTS
4HE HOUSEHOLDÚS OPTIMIZATION PROBLEM IS TO CHOOSE B , ! * AND w TO MAXIMIZE
S
S
S
S
S
SUBJECT TO CONSTRAINTS AND
8
S
n 4 B
S
S
,
/
w
S
S
S
1 ` ! `
S
, ` , `
S
S
S
!
S
6+
S
/B
S
)*
S
S
{
S
S
7`
S
/ * ` ` p/ *
S
S
S
S
S
3EVERAL OBSERVATIONS ABOUT THE HOUSEHOLDÚS DECISION PROBLEM ARE IN ORDER &IRST BECAUSE
HOUSEHOLD LABOR SUPPLY IS DETERMINED BY THE LABOR DEMAND OF çRMS WAGE DETERMINATION
IS DISCUSSED BELOW
+ IS NOT A CHOICE VARIABLE 3ECOND THE çRST ORDER CONDITIONS IMPLY A
S
DEMAND FOR MONEY FUNCTION OF THE DOUBLE LOG FORM
KM
P
,
`KM
/
` P
S
S
KM B ` KM
S
1
S
1
S
4HUS THE MODEL IMPLIES THAT HOUSEHOLD CONSUMPTION NOT OUTPUT IS AN ARGUMENT OF THE
MONEY DEMAND FUNCTION !CCORDINGLY POSITIVE MONEY SUPPLY SHOCKS TEND TO HAVE SMALLER
EdECTS ON THE PRICE LEVEL TO THE EXTENT THAT THEY INDUCE A RISE IN REAL MONEY DEMAND THROUGH
THEIR STIMULATIVE EdECT ON CONSUMPTION 4HIRD THE MODEL ABSTRACTS FROM çSCAL POLICY CONSID
ERATIONS THE ONLY ROLE OF GOVERNMENT IS TO DISTRIBUTE LUMP SUM CASH TRANSFERS TO HOUSEHOLDS
THAT IT çNANCES BY PRINTING MONEY
&IRMS
&IRMS RENT LABOR AND CAPITAL SERVICES FROM HOUSEHOLDS AT THE PREVAILING NOMINAL WAGE 6
S
AND CAPITAL RENTAL PRICE ) (ENCE çRMS SOLVE A STATIC OPTIMIZATION PROBLEM OF THE FORM
S
L@W { / % * + ` 6 + ` ) *
S
S
S
S
S
S
S
S
4HE PRODUCTION FUNCTION % * + IS ASSUMED TO BE #OBB $OUGLAS
% * + * +`
t
S
S
S
S
t
4HE çRST ORDER CONDITIONS TO THIS PROBLEM GIVE THE DEMAND FOR LABOR AND CAPITAL SCHEDULES
OF A REPRESENTATIVE çRM
% * +
6
/
% * +
)
/
S
S
S
S
S
S
S
S
7AGE DETERMINATION
4HE MODEL IS COMPLETED BY DESCRIBING THE STOCHASTIC PROCESSES GOVERNING THE EVOLUTION OF
WAGES AND THE SUPPLY OF MONEY 4HE WAGE CONTRACT STRUCTURES DISCUSSED BELOW BOTH IMPLY
THAT MONETARY SHOCKS AdECT REAL WAGES IN THE SHORT RUN AND HENCE THE LABOR DEMAND AND
OUTPUT SUPPLIED BY çRMS BECAUSE PRICES BASICALLY RESPOND IMMEDIATELY TO SUCH SHOCKS WHILE
WAGES TAKE SOME TIME TO FULLY ADJUST
&ISCHER CONTRACT SPECIçCATION
)N OUR SET UP &ISCHER WAGE CONTRACTS ARE ARRANGEMENTS IN WHICH THE NOMINAL WAGE PREVAILING
J PERIODS IN THE FUTURE IS SET IN THE CURRENT PERIOD !T THE CONTRACTED WAGE HOUSEHOLDS EXPECT
TO WORK A TIME INVARIANT QUANTITY OF HOURS WHICH WE DENOTE BY + ! NATURAL CHOICE FOR +
IS THE AVERAGE SHARE OF THE TIME ENDOWMENT SPENT WORKING TYPICALLY CALIBRATED TO BE IN
THE RANGE OF THIRTY PERCENT 'IVEN THAT LABOR HOURS ARE DEMAND DETERMINED THE CONTRACTED
NOMINAL WAGE SET AT TIME S TO APPLY IN PERIOD S
PRODUCT OF LABOR OF A REPRESENTATIVE çRM AT S
J EQUALS THE EXPECTED VALUE MARGINAL
J EVALUATED AT THE TIME INVARIANT HOURS
WORKED +
6
S
J
$
S
N
/
S
J
%
S
J
O
*
S
J
+
4HUS THE NOMINAL WAGE IS SET SO THAT THE REAL WAGE THAT IS EXPECTED TO PREVAIL WHEN THE
CONTRACT TAKES EdECT IS EXPECTED TO ALLOW HOURS WORKED TO EQUAL ITS LONG RUN AVERAGE LEVEL
!LTHOUGH A MONETARY SHOCK CAN CAUSE HOURS WORKED TO DEVIATE FROM ITS LONG RUN LEVEL DURING
THE SUBSEQUENT J ` PERIODS NOMINAL WAGES FULLY ADJUST TO THE NEW POST SHOCK PRICE LEVEL
AFTER J PERIODS -ONEY IS VIRTUALLY NEUTRAL AFTER J PERIODS EXCEPT FOR THE SMALL EdECT OF
CAPITAL STOCK ADJUSTMENT THAT OCCURS ALONG THE TRANSITION
4HE APPROXIMATE NEUTRALITY OF MONETARY SHOCKS BEYOND THE LENGTH OF THE CONTRACT PERIOD
IMPOSES A STRONG A PRIORI CONSTRAINT ON THE ABILITY OF MONETARY SHOCKS TO EXPLAIN BUSINESS
CYCLE âUCTUATIONS AT LEAST IF REASONABLE ASSUMPTIONS ARE MADE ABOUT THE LENGTH OF THE CON
TRACT INTERVAL 7E ASSUME THAT WAGE CONTRACTS LAST FOUR QUARTERS 2ATHER THAN ATTEMPTING TO
4HIS
ASSUMES CERTAINTY EQUIVALENCE
DERIVE GREATER PERSISTENCE IN THE REAL EdECTS OF MONETARY INNOVATIONS BY STRETCHING THE LENGTH
OF THE CONTRACT INTERVAL BEYOND A YEAR Õ WHICH WOULD SEEM TO BE VERY DIbCULT TO JUSTIFY
EMPIRICALLY GIVEN THE GENERALLY WEAK POWER OF LABOR UNIONS AT THE ONSET OF THE $EPRESSION
Õ WE CONSIDER ALSO THE ALTERNATIVE STAGGERED OVERLAPPING WAGE CONTRACT SCHEME SUGGESTED
BY 4AYLOR
4AYLOR OVERLAPPING WAGE CONTRACTS
)N THE 4AYLOR CONTRACT SPECIçCATION THE LABOR FORCE MAY BE REGARDED AS DIVIDED INTO EQUAL
SIZED COHORTS OF WORKERS WITH THE NUMBER OF COHORTS EQUAL TO THE LENGTH OF THE CONTRACT
PERIOD 7ITH THE CONTRACT INTERVAL SET EQUAL TO FOUR QUARTERS THIS MEANS THAT THERE ARE FOUR
COHORTS OF WORKERS $URING THE BEGINNING OF EACH QUARTER ONE OF THE FOUR COHORTS AGREES TO
A NOMINAL WAGE THAT IT WILL RECEIVE FOR THE FOLLOWING YEAR &OR EXAMPLE COHORT ! AGREES TO
A WAGE W DURING THE WINTER AND THIS PROCESS IS REPEATED BY THE OTHER COHORTS DURING THE
S
SUBSEQUENT THREE QUARTERS OF THE YEAR &OLLOWING 4AYLOR THE CONTRACT WAGE W DEPENDS
S
ON THE GEOMETRIC AVERAGE WAGE 6 OF ALL COHORTS IN THE ECONOMY THAT WILL PREVAIL OVER THE
S
FOUR QUARTER LIFE OF THE CONTRACT AS WELL AS ON THE EVOLUTION OF HOURS WORKED 3PECIçCALLY
THE CONTRACT WAGE OF THE COHORT RENEGOTIATING ITS WAGE AT TIME S IS GIVEN BY
r
KMW KM6
S
S
s
o+ ` +
S
$
S
KM6S
o+ ` +S
o+ ` +
S
KM6
S
KM6
S
o+ ` +
S
WHERE o AND 6 DENOTES THE GEOMETRIC AVERAGE WAGE OF ALL COHORTS IN THE ECONOMY AT
S
TIME S
6 W W ` W ` W `
S
S
S
S
S
A
WHERE WE TAKE FOR ALL H 7ITH REPEATED SUBSTITUTION OF A INTO
H
H
YIELDS
KMW $
S
S
KMW `
S
KMW
!S IN 4AYLOR
S
KMW
`
S
o
0
J
r
KMW
`
S
+ ` +
S
s
J
KMW
S
KMW
S
TWO FEATURES OF EQUATION ARE NOTEWORTHY &IRST CONTRACT WAGES
HAVE AN INERTIAL OR BACKWARD LOOKING COMPONENT THAT IS PAST CONTRACT WAGES APPEAR IN THE
EQUATION 4HIS INERTIAL COMPONENT REâECTS IN PART THAT THE CURRENT CONTRACT WAGE DEPENDS
ON THE CURRENT AVERAGE WAGE ACROSS ALL COHORTS WHICH ITSELF IS AN AVERAGE OF CONTRACT WAGES
SET IN THE PAST THAT ARE STILL IN EdECT 4HIS INERTIAL COMPONENT OF WAGES TENDS TO ÝANCHORÞ
THE CURRENT WAGE 3ECOND THE ONLY CHANNEL FORCING NOMINAL WAGES TO ADJUST ARE CURRENT AND
FUTURE EXPECTED LABOR HOURS !CCORDINGLY THE PARAMETER o IS OF CRUCIAL SIGNIçCANCE GIVEN THE
STRUCTURE OF DEPENDENCE BETWEEN THE CURRENT CONTRACT WAGE AND PAST AND FUTURE WAGE RATES
IT CONTROLS HOW QUICKLY WAGES ADJUST TO CHANGING LABOR MARKET CONDITIONS ,OW ABSOLUTE
VALUES OF o IMPLY COMPARATIVELY SLOW ADJUSTMENT OF NOMINAL WAGES Õ AND AS SHOWN IN THE
NEXT SECTION ALLOW THE MODEL TO ACCOUNT FOR MUCH MORE PERSISTENT REAL EdECTS OF MONETARY
SHOCKS THAN THE &ISCHER SPECIçCATION
-ONEY STOCK EVOLUTION
4HE MODEL ECONOMY ABSTRACTS FROM A BANKING SECTOR AND THE ENDOGENOUS CREATION OF MONEY
4HUS THE STOCK OF MONEY IS ASSUMED TO BE EXOGENOUSLY DETERMINED 4HE GROWTH RATE OF THE
STOCK OF MONEY IS ASSUMED TO FOLLOW A çRST ORDER AUTOREGRESSION OF THE FORM
F F
S
F
S
|F
S
}
S
KM, ` KM, `
S
S
WHERE IS INDEPENDENT AND IDENTICALLY DISTRIBUTED IID -
S
4HE ASSUMPTION THAT THE MONEY GROWTH PROCESS EVOLVED EXOGENOUSLY OVER THE $EPRESSION
PERIOD SEEMS REASONABLE GIVEN THE FOCUS OF THE PAPER Õ TO DERIVE THE CONSEQUENCES OF THE
MONETARY DECLINE AND SUBSEQUENT EXPANSION FOR REAL ACTIVITY IRRESPECTIVE OF THE SOURCES
OF THE MONETARY DECLINE )N PARTICULAR THE MECHANISM THROUGH WHICH MONEY IS POSITED TO
AdECT REAL ACTIVITY IN THE MODEL THE STICKY WAGE CHANNEL MEANS THAT EQUAL SIZED SHOCKS TO
THE QUANTITY OF MONEY THAT MIGHT HAVE BEEN GENERATED FROM VERY DIdERENT SOURCES SHOULD
HAVE SIMILAR EdECTS ON REAL ACTIVITY 3OME CANDIDATE ALTERNATIVE SOURCES INCLUDE THE EdECT
OF &EDERAL 2ESERVE OPEN MARKET OPERATIONS ON THE BASE OR THE EdECT OF BANK PANICS ON THE
MONEY MULTIPLIER /UR FORMULATION IS IN ACCORDANCE WITH THE RECENT LITERATURE INVESTIGATING
THE STICKY WAGE CHANNEL 4HIS LITERATURE HAS MAINLY EMPHASIZED HOW VARIOUS SHOCKS WERE
IMPORTANT BECAUSE OF THEIR EdECT ON MONEY AND THE PRICE LEVEL AND NOT BECAUSE THEY HAD A
LARGE DIRECT EdECT ON ACTIVITY "ERNANKE AND #AREY
#ALIBRATION AND 3OLUTION -ETHOD
4HE MODELÚS FREE PARAMETERS INCLUDE THE TECHNOLOGY PARAMETERS t AND p THE UTILITY FUNCTION
PARAMETERS n AND P AND THE MONEY SUPPLY EVOLUTION PARAMETERS | AND } 4HE MODEL IS
ASSUMED TO HOLD AT A QUARTERLY FREQUENCY
4HE CAPITAL SHARE PARAMETER t IS SET EQUAL TO AND THE QUARTERLY DEPRECIATION RATE TO
IN THE RANGE OF TYPICAL ESTIMATES OF THESE PARAMETERS DERIVED FROM POSTWAR DATA 4HE
DISCOUNT FACTOR n IS SET TO IMPLYING AN ANNUAL REAL INTEREST RATE OF ABOUT FOUR PERCENT
4HE SHARE PARAMETER P IS DETERMINED FROM THE HOUSEHOLDÚS STEADY STATE MONEY DEMAND
FUNCTION UNDER THE ASSUMPTION THAT VELOCITY EQUALS ONE )N THIS CASE EQUATION IMPLIES
1
P
` P
1
WHERE 1 IS THE STEADY STATE VALUE OF THE NOMINAL INTEREST RATE "ECAUSE IT SEEMS REASONABLE
TO ROUGHLY DEçNE THE STEADY STATE VALUE OF A VARIABLE AS ITS AVERAGE VALUE OVER A ÝTYPICALÞ
PERIOD THE STEADY STATE VALUE OF 1 IS TAKEN TO BE ITS AVERAGE VALUE OVER THE PERIOD
5SING THE INTEREST RATE ON THREE MONTH COMMERCIAL PAPER TAKEN FROM THE "ALKE 'ORDON DATA
APPENDIX OF 4HE !MERICAN "USINESS #YCLE AS A PROXY FOR 1 THE AVERAGE VALUE IS
/N
THE OTHER HAND THE FAILURE OF OUR MODEL TO ACCOUNT FOR THE DURATION OF THE $EPRESSION PERIOD MIGHT
ARGUE FOR INCORPORATING AN ENDOGENOUS THEORY OF MONEY SUPPLY DETERMINATION SO THAT VARIOUS SHOCKS THAT
ÝINDIRECTLYÞ AdECTED REAL ACTIVITY THROUGH THEIR EdECT ON MONEY AND THE PRICE LEVEL MIGHT ALSO EXERT DIRECT
EdECTS OF VARYING MAGNITUDE &OR EXAMPLE BANK PANIC SHOCKS THAT INDUCED DECLINES IN THE MONEY MULTIPLIER
AND HENCE AdECTED REAL ACTIVITY THROUGH THE CHANNEL OUTLINED IN OUR MODEL MIGHT HAVE AN ADDITIONAL EdECT
TO THE EXTENT THAT BANK CLOSURES ASSOCIATED WITH THE PANICS RAISED THE COST OF çNANCIAL INTERMEDIATION
AT A QUARTERLY RATE IMPLYING THAT P 3IMILARLY THE MONEY SUPPLY PARAMETERS | AND
} ARE ESTIMATED BY USING OBSERVATIONS ON THE GROWTH OF - OVER THE SAME PERIOD &ITTING A
çRST ORDER AUTOREGRESSION TO - GROWTH IMPLIES THAT | AND THAT } WHERE } IS
IN UNITS OF QUARTERLY RATES OF CHANGE ! DISCUSSION OF THE IMPLICATIONS OF OUR SPECIçCATION OF
THE MONEY SUPPLY RULE AND SENSITIVITY OF OUR RESULTS TO THESE PARAMETER CHOICES IS PROVIDED
BELOW
4HE APPROXIMATE DYNAMIC BEHAVIOR OF THE MODEL IS OBTAINED BY çRST LINEARIZING THE
APPROPRIATE NONLINEAR STOCHASTIC DIdERENCE EQUATIONS AROUND THE STEADY STATE VALUES OF THE
STATE VARIABLES AND THEN APPLYING THE METHOD OF "LANCHARD AND +AHN 4HE EQUATIONS
DETERMINING THE EVOLUTION OF THE STATE VARIABLES INCLUDE THE WAGE EQUATION FOR THE
&ISCHER CONTRACT CASE OR FOR THE 4AYLOR CONTRACT CASE AND THE HOUSEHOLDÚS çRST ORDER
EQUATIONS FOR REAL BALANCES AND CAPITAL SUBJECT TO THE AGGREGATE RESOURCE CONSTRAINT
)MPULSE 2ESPONSE &UNCTIONS
&IGURES PLOT THE MODELÚS IMPULSE RESPONSE FUNCTIONS )2& TO A ONE STANDARD DEVIATION
POSITIVE INNOVATION IN THE GROWTH RATE OF MONEY FOR THE &ISCHER WAGE CONTRACTS MODEL AND TWO
PARAMETERIZATIONS OF THE 4AYLOR CONTRACTS MODEL )N ONE 4AYLOR CONTRACTS PARAMETERIZATION
NOMINAL WAGES RESPOND RELATIVELY SLOWLY TO DEVIATIONS IN LABOR HOURS FROM THEIR AVERAGE LEVEL
4HE
ANNUALIZED NOMINAL RATE OF INTEREST IS /UR DISCOUNT FACTOR IMPLIES A ANNUALIZED REAL RATE
OF INTEREST )NâATION VARIED OVER THIS PERIOD BETWEEN SLIGHTLY POSITIVE AND NEGATIVE RATES
4HE BASIC RESULTS DISCUSSED BELOW ARE NOT PARTICULARLY SENSITIVE TO SUBSTITUTING THE GROWTH RATE OF - FOR
- 4HE - DATA ARE TAKEN FROM &RIEDMAN AND 3CHWARTZ
"ECAUSE THE MONEY STOCK GROWS AT A NONZERO RATE ON AVERAGE THE STATE VARIABLES REPRESENTING NOMINAL
MAGNITUDES Õ NAMELY THE NOMINAL WAGE AND PRICE LEVEL Õ ARE SCALED BY THE STOCK OF MONEY SO THAT THE
SYSTEM IS STATIONARY IN THE TRANSFORMED VARIABLES
o ` AND ANOTHER WAGES RESPOND MUCH MORE RAPIDLY o ` )N ALL OF THESE
çGURES THE INNOVATION OCCURS IN TIME PERIOD
&IGURE PLOTS THE )2& OF THE PRICE LEVEL FOR THE &ISCHER AND 4AYLOR CONTRACT PARAMETER
IZATIONS /N IMPACT THE MODEL PREDICTS THAT A POSITIVE INNOVATION IN THE STOCK OF MONEY
INCREASES REAL MONEY DEMAND BY LESS THAN THE INCREASE IN THE MONEY SUPPLY 4HIS INNOVATION
LEADS TO AN IMMEDIATE INCREASE IN THE PRICE LEVEL RELATIVE TO ITS BASELINE PATH &OLLOWING THIS
INITIAL JUMP THE çGURE INDICATES THAT THE PRICE LEVEL ACTUALLY DECLINES SOMEWHAT IN THE SHORT
TO MEDIUM TERM 4HE PRICE LEVEL DECLINE REâECTS THE RISING PROçLE OF CONSUMPTION AND ITS
INâUENCE ON REAL MONEY DEMAND FROM EQUATION .EVERTHELESS PRICES CONTINUE TO BE
HIGHER THAN BEFORE THE MONEY INNOVATION
4HE HIGHLY FRONT LOADED RESPONSE OF THE PRICE LEVEL IMPLIES THAT THE PRICE LEVEL INCREASES
ASSOCIATED WITH POSITIVE MONEY INNOVATIONS ARE LARGELY UNANTICIPATED 4HIS IMPLICATION IS
NOT PARTICULARLY SENSITIVE TO OUR CHOICE OF THE MONEY GROWTH PERSISTENCE PARAMETER | IN
ORDER FOR MONEY GROWTH INNOVATIONS TO LEAD TO A PERSISTENT CHANGE IN THE RATE OF GROWTH
OF THE PRICE LEVEL | MUST BE IN THE RANGE OR HIGHER ABOUT THREE TIMES OUR ESTIMATE
4HE çGURE ALSO SHOWS THAT THE PRICE LEVEL )2&S DO NOT VARY SUBSTANTIALLY ACROSS THE THREE
SPECIçCATIONS THOUGH THE PRICE LEVEL EdECTS OF MONETARY INNOVATIONS ARE SMALLEST IN THE NEAR
TO MEDIUM TERM IN THE 4AYLOR SPECIçCATION WITH LOW o ` )N THIS CONTRACT SPECIçCATION
A ONE STANDARD DEVIATION MONETARY INNOVATION HAS THE LARGEST EdECTS ON CONSUMPTION AND
REAL MONEY DEMAND
&IGURE PLOTS THE )2& OF THE NOMINAL WAGE )N THE &ISCHER CONTRACT CASE NOMINAL WAGES
DO NOT RESPOND AT ALL DURING THE çRST FOUR QUARTERS FOLLOWING A SHOCK BUT THEN JUMP TO A LEVEL
CONSISTENT WITH STEADY STATE LABOR HOURS 4HE 4AYLOR CONTRACT SPECIçCATIONS ARE CONSISTENT
WITH A GRADUAL ADJUSTMENT OF NOMINAL WAGES AND THE SPEED OF ADJUSTMENT VARIES DIRECTLY
WITH THE ABSOLUTE VALUE OF THE PARAMETER o
&IGURES AND PLOT THE )2&S OF THE REAL WAGE AND HOURS WORKED RESPECTIVELY )N THE
&ISCHER CASE REAL WAGES REMAIN SUBSTANTIALLY BELOW THEIR EQUILIBRIUM LEVEL UNTIL FOUR QUARTERS
AFTER THE SHOCK AT WHICH POINT THE REAL WAGE JUMPS TO A LEVEL SLIGHTLY ABOVE ITS STEADY STATE
LEVEL DUE TO THE ABOVE STEADY STATE CAPITAL STOCK LEVEL #ORRESPONDINGLY WHEN REAL WAGES
ARE UNUSUALLY LOW HOURS WORKED ARE MUCH HIGHER THAN THEIR STEADY STATE LEVEL HOURS WORKED
FALL TO THEIR STEADY STATE LEVEL WHEN REAL WAGES INCREASE 4HE BEHAVIOR OF HOURS WORKED CLEARLY
REâECTS THE ASSUMPTION THAT LABOR HOURS ARE DETERMINED BY A BASICALLY STATIC LABOR DEMAND
SCHEDULE
&OR THE 4AYLOR SPECIçCATIONS THE )2&S OF THE REAL WAGE AND HOURS WORKED ARE QUALITATIVELY
SIMILAR TO THOSE OF THE &ISCHER MODEL EXCEPT THAT THE ADJUSTMENTS TO THE STEADY STATE ARE
RELATIVELY SMOOTH -OREOVER REAL WAGES ADJUST MUCH MORE SLOWLY AS THE ABSOLUTE VALUE OF
o DECLINES IMPLYING THAT HOURS WORKED REMAIN ABOVE THEIR STEADY STATE LEVEL FOR A LONGER
PERIOD #LEARLY THE )2&S INDICATE THAT THE 4AYLOR MODEL WITH LOW ABSOLUTE VALUES OF o
PRODUCE MORE PERSISTENT MONETARY NONNEUTRALITIES AND THIS IS HOLDING THE CONTRACT LENGTH
çXED
&IGURES THROUGH PLOT THE )2&S OF OUTPUT CONSUMPTION AND INVESTMENT RESPECTIVELY
3INCE THE CAPITAL STOCK ADJUSTMENT TENDS TO BE QUITE MODEST THE OUTPUT )2&S BASICALLY
MIRROR THE CORRESPONDING HOURS WORKED )2&S &OR CONSUMPTION THE PEAK RESPONSE TENDS TO
BE CONSIDERABLY SMALLER THAN THAT OF OUTPUT REâECTING THE TRANSIENT NATURE OF THE SHOCKÚS
IMPACT ON INCOME 4HE ABSOLUTE MAGNITUDE OF THE CONSUMPTION RESPONSE IS LARGEST IN THE
4AYLOR MODEL WITH A LOW ABSOLUTE VALUE OF o SINCE THE EdECTS ON HOUSEHOLD INCOME ARE LARGEST
IN THIS CASE 4HE STRONG RESPONSE OF INVESTMENT IN EACH OF THE CASES REâECTS THE HOUSEHOLDÚS
DESIRE TO SMOOTH CONSUMPTION BY INVESTING IN PRODUCTIVE CAPITAL
#OMPARISONS OF -ODEL 3IMULATIONS WITH $ATA OVER THE
PERIOD
)N THIS SECTION TIME SERIES PLOTS OF OUTPUT CONSUMPTION INVESTMENT AND HOURS WORKED
GENERATED FROM MODEL SIMULATIONS ARE COMPARED WITH CORRESPONDING DATA 4HE OBJECTIVE IS TO
PROVIDE A QUANTITATIVE ASSESSMENT OF THE ABILITY OF THE STICKY WAGE MODEL INCLUDING BOTH THE
&ISCHER AND 4AYLOR CONTRACT SPECIçCATIONS TO EXPLAIN THE BEHAVIOR OF MAJOR MACROAGGREGATES
DURING THE 'REAT $EPRESSION PERIOD
$ATA
4HE DATA ON OUTPUT CONSUMPTION AND INVESTMENT USED IN THE COMPARISONS ARE PRIMARILY
BASED ON THE QUARTERLY NATIONAL INCOME ACCOUNTS DATA CONSTRUCTED BY "ALKE AND 'ORDON
3OME EdORT IS REQUIRED IN ORDER TO çT ACTUAL NATIONAL INCOME ACCOUNTS DATA INTO
THE MUCH SIMPLER ACCOUNTING IDENTITY IMPOSED IN THE MODEL BECAUSE THE MODEL DOES NOT
SPECIFY EITHER AN EXTERNAL OR A GOVERNMENT SECTOR /UTPUT IS IDENTIçED AS REAL '$0 FROM
THE .)0! 4HE CONSUMPTION COMPONENT IN THE MODEL IS EQUATED WITH PRIVATE SPENDING ON
NONDURABLE GOODS AND SERVICES PLUS A MEASURE OF PUBLIC CONSUMPTION SPENDING 4HE PUBLIC
CONSUMPTION SERIES IS DERIVED FROM SUBTRACTING A MEASURE OF GOVERNMENT INVESTMENT SPEND
ING FROM GOVERNMENT PURCHASES GOVERNMENT INVESTMENT SPENDING IS BASED ON INTERPOLATING
ANNUAL DATA THAT APPEARS IN THE "%! PUBLICATION &IXED 2EPRODUCIBLE 4ANGIBLE 7EALTH IN THE
5NITED 3TATES &INALLY INVESTMENT IS DEçNED AS A RESIDUAL BETWEEN OUTPUT AND OUR MEASURE OF
CONSUMPTION !LTHOUGH NET EXPORTS ARE INCLUDED IN THIS MEASURE OF INVESTMENT THIS SEEMS
REASONABLE GIVEN THAT THE MODEL DEçNES INVESTMENT AS FOREGONE CONSUMPTION 4HE HOURS
WORKED SERIES IS DERIVED FROM MULTIPLYING TOTAL EMPLOYMENT BY THE AVERAGE LENGTH OF THE
WORKWEEK IN THE MANUFACTURING SECTOR 4HE EMPLOYMENT DATA ARE QUARTERLY AND FROM THE
"UREAU OF ,ABOR 3TATISTICS !GGREGATE WORKWEEK DATA ARE NOT AVAILABLE QUARTERLY SO WE USE
THE AVERAGE HOURS WORKED PER WEEK IN THE MANUFACTURING SECTOR FROM THE .ATIONAL )NDUSTRIAL
#ONFERENCE "OARD "ENEY !LL OF THESE DATA ARE DEçNED IN PER CAPITA TERMS
/UR COMPARISONS ALSO INVOLVE NOMINAL AND REAL WAGES 4HE NOMINAL WAGE SERIES REPRESENTS
AN AVERAGE OF PAYROLL WAGES IN MANUFACTURING INDUSTRIES !S WITH THE WORKWEEK DATA
NOMINAL WAGE DATA ARE NOT AVAILABLE AT AN AGGREGATE LEVEL OUTSIDE OF MANUFACTURING 4HESE
DATA WERE COMPILED BY THE .)#" AND ARE AVAILABLE FOR THE PERIOD TO "ENEY
/UR REAL WAGE SERIES IS DERIVED BY DEâATING THIS NOMINAL WAGE SERIES BY THE '$0
DEâATOR FROM "ALKE AND 'ORDON
)N
OTHER WORDS OUTPUT EQUALS CONSUMPTION PLUS INVESTMENT
RESULTS REPORTED BELOW ARE ESSENTIALLY UNCHANGED WHEN WE CONSIDERED AN ALTERNATIVE MEASURE OF
4HE
4HE DOWNTURN PHASE OF THE 'REAT $EPRESSION
4HE MODEL IS SIMULATED OVER THE PERIOD BY INPUTTING ESTIMATES OF MONEY GROWTH
INNOVATIONS OVER THAT PERIOD INTO THE STATE SPACE REPRESENTATION OF THE MODEL 4HE MONEY
GROWTH INNOVATIONS ARE DERIVED FROM THE MONEY SUPPLY EVOLUTION EQUATION USING DATA ON
- GROWTH AND THE PARAMETER ESTIMATES OF THE - GROWTH PROCESS EQUATION 4HE
SIMULATIONS ALSO ASSUME THAT ALL MODEL VARIABLES INCLUDING THE GROWTH RATE OF - WERE AT
THEIR STEADY STATE LEVEL IN "ECAUSE THE SIMULATED DATA ARE REPRESENTED IN ÝPERCENT
DEVIATION FROM STEADY STATEÞ FORM HISTORICAL DATA ARE REPRESENTED IN A COMPARABLE FORM BY
CONSTRUCTING EACH SERIES AS A LOGARITHMIC PERCENTAGE DEVIATION FROM ITS LEVEL
%STIMATES OF THE INNOVATIONS TO MONEY GROWTH ARE PLOTTED IN &IGURE )T IS EVIDENT FROM
THE çGURE THAT OUR PARAMETERIZATION IMPLIES A SERIES OF LARGE NEGATIVE INNOVATIONS TO MONEY
GROWTH BEGINNING IN )N FACT THESE INNOVATIONS ARE VERY CLOSE TO THE ACTUAL RATE OF
MONEY GROWTH GIVEN THE LOW VALUE OF THE AUTOREGRESSIVE PARAMETER /UR MODEL IMPLIES THAT
THESE NEGATIVE MONEY GROWTH INNOVATIONS SHOULD HAVE LED TO A SERIES OF LARGELY UNANTICIPATED
DECLINES IN THE PRICE LEVEL AS SHOWN BELOW
&IGURES DISPLAY THE EdECTS OF THE PRICE DECLINE ON OUTPUT HOURS WORKED CONSUMP
TION AND INVESTMENT IMPLIED BY THE &ISCHER WAGE CONTRACT AND ALSO THE DATA FOR THESE
VARIABLES )T IS CLEAR FROM THE SIMULATED DATA THAT THE NEGATIVE SHOCKS TO MONEY GROWTH
OVER THE PERIOD WERE SUbCIENTLY LARGE TO ACCOUNT FOR A MASSIVE CONTRACTION IN ALL OF
THESE MACROAGGREGATES 'IVEN ITS SIMPLE STRUCTURE THE MODEL DOES REASONABLY WELL FROM A
NOMINAL WAGES THAT CONTROLS FOR CHANGING INDUSTRY COMPOSITION 4HIS DATA WAS PROVIDED BY #HRIS (ANES SEE
(ANES A B
QUANTITATIVE PERSPECTIVE IN çTTING THE OBSERVED PATH OF OUTPUT AND HOURS WORKED UNTIL EARLY
&OR EXAMPLE THE MODEL IMPLIES A PERCENT DECLINE IN OUTPUT RELATIVE TO ITS LEVEL
BY WHICH IS VERY CLOSE TO THE PERCENT DECLINE THAT WAS OBSERVED
(OWEVER THE MODEL PERFORMS POORLY ALONG SEVERAL DIMENSIONS "ECAUSE THE MODEL DOES
NOT ALLOW MONETARY SHOCKS TO AdECT HOURS WORKED BEYOND THE ASSUMED LENGTH OF WAGE CON
TRACTS ONE YEAR
THE PATH OF HOURS BASICALLY TRACES THE PATH OF MONEY INNOVATIONS EXCEPT
FOR A SHORT LAG SEE &IGURE !S A RESULT THE MODEL INCORRECTLY IMPLIES THAT THE STABILIZA
TION OF MONETARY AGGREGATES IN THE SPRING OF SHOULD HAVE LED TO A SUBSTANTIAL RISE IN
HOURS WORKED !LSO THE SUSTAINED REMONETIZATION THAT OCCURRED DURING THE PERIOD
SHOULD HAVE LED TO AN EXTREMELY STRONG EXPANSION IN HOURS WORKED BEGINNING IN MID
"ECAUSE THE PATH OF OUTPUT IS LARGELY DETERMINED BY THE PATH OF HOURS WORKED THE MODEL IS
FAR Od THE MARK IN ITS PREDICTIONS OF A SUBSTANTIAL RECOVERY IN OUTPUT IN AND A MAJOR
BOOM BY LATE &INALLY IT IS EVIDENT FROM &IGURE THAT THE MODEL SUBSTANTIALLY UN
DERPREDICTS THE MAGNITUDE OF THE FALL IN CONSUMPTION THAT OCCURRED DURING THE $EPRESSION
DOWNTURN 4HIS REâECTS THE &ISCHER MODELÚS IMPLICATION THAT MONETARY SHOCKS HAVE A VERY
TRANSIENT EdECT ON PERMANENT INCOME
)T IS CLEAR THAT ACCOUNTING FOR THE PERSISTENCE OF THE DOWNTURN IN OUTPUT HOURS CONSUMP
TION AND INVESTMENT WITHIN A PURELY MONETARY FRAMEWORK REQUIRES THAT MONETARY SHOCKS
EXERT CONSIDERABLY MORE PERSISTENT REAL EdECTS THAN IS AdORDED BY THE &ISCHER CONTRACT SPECI
çCATION 4HE )2&S SUGGEST THAT THE 4AYLOR MODEL WITH SUbCIENTLY SLUGGISH WAGE ADJUSTMENT
IE A LOW ABSOLUTE VALUE OF o MAY PROVIDE A BETTER ACCOUNT OF THE BEHAVIOR OF THESE
MACROAGGREGATES AFTER EARLY !CCORDINGLY WE SIMULATE THE 4AYLOR MODEL BELOW CHOOS
ING o ` WHICH IMPLIES )2&S THAT ARE CONSIDERABLY MORE PERSISTENT THAN IN THE &ISCHER
SPECIçCATION REFER AGAIN TO &IGURES 3OME JUSTIçCATION FOR THIS PARAMETER CHOICE IS
PROVIDED BELOW
&IGURE COMPARES THE DATA ON OUTPUT BASED ON THIS PARAMETERIZATION OF THE 4AYLOR
MODEL WITH THE CORRESPONDING SIMULATED SERIES OVER THE PERIOD REFER TO THE
LINE MARKED ÝWITHOUT -$ SHOCKSÞ IN &IGURES WHICH DISTINGUISH THESE SIMULATIONS FROM
A VARIANT THAT ALLOWS FOR MONEY DEMAND SHOCKS CONSIDERED BELOW 4HE MODEL TRACKS THE
OUTPUT DOWNTURN EXTREMELY WELL THROUGH EARLY 4HE MODEL DOES MUCH BETTER THAN THE
&ISCHER SPECIçCATION IN ACCOUNTING FOR THE PERSISTENCE OF THE OUTPUT DOWNTURN BETWEEN EARLY
AND MID 4HE 4AYLOR SPECIçCATION CANNOT HOWEVER ACCOUNT FOR THE CONTINUED
OUTPUT DECLINE THAT OCCURRED BETWEEN EARLY AND MID IN THIS CASE THE MODEL
PREDICTS A SLIGHT RECOVERY IN OUTPUT OVER THAT PERIOD $ESPITE THIS SHORTCOMING THE MODEL
STILL APPEARS TO DO REASONABLY WELL IN ACCOUNTING FOR THE MAGNITUDE OF THE OUTPUT DECLINE
EXPERIENCED AT THE $EPRESSION TROUGH IMPLYING AN OUTPUT DECLINE OF PERCENT IN THE çRST
HALF OF FROM ITS LEVEL COMPARED TO THE PERCENT DECLINE OBSERVED
! MUCH MORE SERIOUS SHORTCOMING OF THE MODEL IS ITS SIGNIçCANT OVERSTATEMENT OF THE
RECOVERYÚS STRENGTH BEGINNING IN LATE OR EARLY )N PARTICULAR IT PREDICTS THAT
THE REMONETIZATION BEGUN IN LATE SHOULD HAVE ALLOWED OUTPUT TO RECOVER RAPIDLY TO ITS
PRE $EPRESSION LEVEL AND SURPASS IT BY THE SECOND HALF OF
!S SEEN IN &IGURE THE MODEL PROVIDES A FAIRLY GOOD CHARACTERIZATION OF THE BEHAVIOR
OF HOURS WORKED DURING MOST OF THE DOWNTURN PHASE OF THE 'REAT $EPRESSION (OWEVER THE
MODEL SIGNIçCANTLY UNDERPREDICTS THE CONTRACTION IN HOURS WORKED THAT OCCURRED BETWEEN
EARLY AND EARLY 4HE MOST IMPORTANT FAILURE OF THE MODEL IS ITS IMPLICATION OF A
RAPID RECOVERY IN HOURS WORKED FOLLOWING THE REMONETIZATION THAT BEGAN IN LATE !S
WITH THE OUTPUT IMPLICATIONS THE MODEL PREDICTS THAT HOURS WORKED SHOULD HAVE RETURNED TO
THEIR PRE $EPRESSION LEVEL BY MID (OWEVER ACTUAL HOURS WORKED RECOVERED VERY SLOWLY
BETWEEN AND AND REMAINED ABOUT PERCENT BELOW ITS LEVEL AT THE END OF
4HE MODEL BASICALLY CAPTURES THE OBSERVED PATH OF CONSUMPTION &IGURE
BOTH WITH
RESPECT TO ITS TIMING AND MAGNITUDE (OWEVER IT UNDERPREDICTS THE MAGNITUDE OF THE DECLINE
TO A MODEST DEGREE 4HE MODEL IMPLIES A DECLINE IN CONSUMPTION OF ABOUT PERCENT AT
THE TROUGH COMPARED TO THE PERCENT DECLINE OBSERVED 2ELATIVE TO THE &ISCHER MODEL
THE 4AYLOR MODELÚS PREDICTION OF A MUCH MORE RAPID AND DEEP DECLINE IN CONSUMPTION OVER
REâECTS THE MORE PERSISTENT MONETARY NONNEUTRALITIES IN THIS CASE
4HE MODELÚS ABILITY TO TRACK INVESTMENT Õ SHOWN IN &IGURE Õ IS SIMILAR TO ITS PERFOR
MANCE IN ACCOUNTING FOR OUTPUT AND HOURS WORKED 4HE MODEL DOES WELL THROUGH EARLY
BUT IMPLIES A SLIGHT RECOVERY OF INVESTMENT OVER THE SUBSEQUENT YEAR IN CONTRAST TO ACTUAL
INVESTMENT SPENDING WHICH DECLINED SOMEWHAT FURTHER UNTIL REACHING A TROUGH IN
!GAIN A MUCH MORE SERIOUS SHORTCOMING OF THE MODEL IS ITS PREDICTION OF A MAJOR BOOM IN
INVESTMENT SPENDING IN THE SECOND HALF OF
&IGURE INDICATES THAT OUR MODEL DOES WELL IN ACCOUNTING FOR THE OBSERVED DECLINE IN
THE PRICE LEVEL THROUGH THE EARLY PART OF THOUGH IT DOES UNDERSTATE THE CONTINUED SHARP
PRICE LEVEL DECLINES THAT OCCURRED OVER THE FOLLOWING YEAR 4HE FAILURE OF OUR SIMPLE MODEL
TO ACCOUNT FULLY FOR THE OBSERVED PRICE LEVEL DECLINE OVER THAT PERIOD MIGHT IN PART REâECT
SHOCKS TO THE MONEY DEMAND FUNCTION THAT CAUSED THE PRICE LEVEL TO DECLINE BY MORE THAN
WOULD BE IMPLIED BY THE BEHAVIOR OF THE MONEY STOCK ALONE
)NCORPORATING MONEY DEMAND SHOCKS INTO THE ANALYSIS
'IVEN THE CRUCIAL ROLE OF PRICE LEVEL CHANGES ON REAL ACTIVITY IN OUR MODEL IT SEEMS IMPORTANT
TO INVESTIGATE WHETHER THE MODELÚS FAILURE TO ACCOUNT FULLY FOR THE PRICE DECLINE MIGHT ALSO
WEAKEN ITS ABILITY TO ACCOUNT FOR THE CONTINUED DOWNTURN IN MAJOR MACROAGGREGATES THAT
OCCURRED BETWEEN EARLY AND EARLY TO MID !CCORDINGLY WE MODIçED THE MODEL
SLIGHTLY TO ALLOW FOR MONEY DEMAND SHOCKS } AS WELL AS MONEY SUPPLY SHOCKS 4HE PREFERENCE
S
SPECIçCATION WAS REFORMULATED AS
4 B
S
,
P KM B
/
S
S
S
` P } KM
S
,
/
S
S
&IGURES ALSO DISPLAY THE SIMULATIONS OF THE MODEL OVER THE SAMPLE PERIOD
WHEN BOTH MONEY SUPPLY AND MONEY DEMAND SHOCKS ARE INPUT INTO THE STATE SPACE
REPRESENTATION 4HE MONEY SUPPLY SHOCKS ARE THE SAME AS DISCUSSED ABOVE 4HE MONEY
DEMAND SHOCKS ARE MEASURED TO MATCH EXACTLY THE MODELÚS IMPLIED PRICE LEVEL PATH WITH
THE ACTUAL PRICE LEVEL DATA 4HE SIMULATIONS LABELLED ÝW MONEY DEMAND -$ SHOCKSÞ
CORRESPOND TO THIS CASE
)T IS EVIDENT FROM THE çGURES THAT ALLOWING FOR MONEY DEMAND SHOCKS DOES ACCOUNT SLIGHTLY
BETTER FOR THE MAGNITUDE OF THE DOWNTURN IN MACROAGGREGATES PARTICULARLY IN THE EARLY
EARLY PERIOD (OWEVER THE EdECTS OF THESE CHANGES IN THE SPECIçCATION ARE NOT
PARTICULARLY LARGE &OR EXAMPLE THE MODEL WITH MONEY DEMAND SHOCKS IMPLIES AN OUTPUT
DECLINE OF ABOUT PERCENT IN 1 FROM ITS PRE $EPRESSION LEVEL COMPARED TO THE
PERCENT DECLINE IMPLIED BY THE MODEL THAT DOES NOT ALLOW FOR MONEY DEMAND SHOCKS
3UMMARIZING OUR RESULTS THUS FAR AND TEMPORARILY FOCUSING ON THE DOWNTURN PHASE OF THE
$EPRESSION THE STICKY WAGE MODEL SEEMS TO PROVIDE A GOOD CHARACTERIZATION OF THE BEHAVIOR
OF MAJOR MACROAGGREGATES THROUGH EARLY 4HE MODEL DOES HOWEVER SEEM TO UNDERSTATE
THE SUBSEQUENT DECLINE IN REAL ACTIVITY THAT OCCURRED BETWEEN EARLY AND EARLY EVEN
IN OUR PREFERRED SPECIçCATION THAT EXACTLY MATCHES THE PRICE LEVEL DECLINE 4HE INABILITY OF A
PURELY MONETARY MODEL TO EXPLAIN THE DOWNTURN FULLY MAY REâECT THE FACT THAT OTHER ÝSUPPLY
SIDEÞ FACTORS INCLUDING THE EdECTS OF BANK PANICS AND DEBT DEâATION CONTRIBUTED NOTICEABLY
TO THE OUTPUT DECLINE PARTICULARLY IN THE PERIOD (OWEVER IT IS STILL INTERESTING THAT
THE STICKY WAGE MODEL APPEARS TO ACCOUNT FOR ROUGHLY THREE QUARTERS OF THE OBSERVED OUTPUT
DECLINE OF PERCENT AT THE $EPRESSION TROUGH -OREOVER çGURES AND WHICH EXAMINE
IDENTIFY A TIME SERIES FOR }T WE ASSUME THAT }T IS A RANDOM WALK AND ITS INNOVATIONS ARE UNCORRELATED
WITH THE INNOVATIONS IN THE MONEY GROWTH PROCESS 4HE MONEY DEMAND SHOCK INNOVATIONS ARE IDENTIçED
BY COMPARING THE STATE SPACE REPRESENTATION FOR a0T-ODEL AND THE ACTUAL DATA FOR a0T$ATA )N PARTICULAR
a0T-ODEL `a0T$ATA DEPENDS UPON CURRENT AND LAGGED ENDOGENOUS STATE VARIABLES LAGGED MONEY GROWTH RATES
THE MONEY GROWTH INNOVATION AND THE MONEY DEMAND INNOVATION 3INCE WE HAVE ALREADY IDENTIçED A TIME
SERIES FOR THE MONEY GROWTH INNOVATIONS THIS DETERMINES THE MONEY DEMAND INNOVATION 4HE ASSUMPTION
THAT THE MONEY SUPPLY AND DEMAND INNOVATIONS ARE UNCORRELATED MAY APPEAR EXTREME "UT ALLOWING FOR EVEN
SIGNIçCANT CROSS CORRELATION AND AUTOCORRELATION IN THE SHOCK PROCESSES EG ALLOWING MONEY SUPPLY SHOCKS
TO ACCOMMODATE CHANGES IN MONEY DEMAND HAS VERY LITTLE EdECT ON THE BEHAVIOR OF THE REAL AND NOMINAL
VARIABLES DEPICTED IN THE çGURES
4O
THE MODELÚS IMPLICATIONS FOR REAL AND NOMINAL WAGES PROVIDE IMPORTANT EVIDENCE IN SUPPORT
OF THE STICKY WAGE CHANNEL AS A CAUSE OF THE 'REAT $EPRESSION 4HE çGURES SUGGEST THAT THE
MECHANISM IN THE MODEL THAT ACCOUNTS FOR THE DOWNTURN IN THE MACROAGGREGATES Õ NAMELY
A RISE IN REAL WAGES AS NOMINAL WAGES FELL MORE SLOWLY THAN THE PRICE LEVEL Õ IS BROADLY
CONSISTENT WITH THE DATA AT LEAST OVER THE DOWNTURN PHASE OF THE $EPRESSION
4HE MODEL ACCOUNTS EXTREMELY WELL FOR THE RISE IN REAL WAGES THAT OCCURRED THROUGH MID
AND THE MODEST DECLINE IN REAL WAGES OVER THE SUBSEQUENT YEAR 4HE MODELÚS ABILITY
TO ACCOUNT FOR PERSISTENT REAL EdECTS ON THE MACROAGGREGATES IS NOT ACHIEVED BY ASSUMING
ÝEXCESSIVELY SLUGGISHÞ NOMINAL WAGE BEHAVIOR /UR PARAMETERIZATION OF o ` çTS NOMINAL
WAGES WELL OVER THE PERIOD AND IN FACT IS THE VALUE OF o THAT ALLOWS THE MODEL
TO çT OBSERVED NOMINAL WAGE BEHAVIOR AS CLOSELY AS POSSIBLE IN A LEAST SQUARES SENSE
4HE RECOVERY PHASE OF THE 'REAT $EPRESSION
4URNING TO THE RECOVERY PHASE OF THE $EPRESSION IT IS CLEAR THAT THE STICKY WAGE MODEL
PROVIDES A POOR çT 4HE MODEL IMPLIES THAT THE REMONETIZATION THAT BEGAN IN SHOULD
HAVE GENERATED A MUCH MORE RAPID RECOVERY THAN WHAT ACTUALLY OCCURRED )T IS TEMPTING TO TRY
TO PROVIDE A BETTER çT OF THE RECOVERY PHASE OF THE $EPRESSION BY FORCING MONETARY INNOVATIONS
TO HAVE MORE PERSISTENT EdECTS ON REAL ACTIVITY 4HIS WOULD BE THE CASE IF NOMINAL WAGES
RESPONDED EVEN LESS TO ECONOMIC ACTIVITY THROUGH LOWER ABSOLUTE VALUES OF o !S SHOWN IN
&IGURES
AND ARE DERIVED FROM THE 4AYLOR MODEL THAT INCLUDES MONEY DEMAND SHOCKS AND HENCE çTS
THE PRICE LEVEL IN çGURE EXACTLY 4HE VERSION OF THE MODEL THAT DOES NOT INCLUDE MONEY DEMAND SHOCKS
HAS QUITE SIMILAR IMPLICATIONS FOR REAL WAGE BEHAVIOR AS IS SHOWN IN çGURE 4HIS SIMILARITY ACCOUNTS FOR
WHY THE TWO SPECIçCATIONS HAVE SIMILAR IMPLICATIONS FOR OTHER REAL VARIABLES THOUGH THE MODEL WITH MONEY
DEMAND SHOCKS çTS OBSERVED REAL WAGE BEHAVIOR SLIGHTLY BETTER
çGURE ASSUMING THAT o ` ALLOWS THE MODEL TO PROVIDE A SOMEWHAT BETTER ACCOUNT
OF THE BEHAVIOR OF OUTPUT DURING THE SLOW RECOVERY 4HE MODEL ALSO ACCOUNTS SOMEWHAT
BETTER FOR THE DEPTH OF THE DOWNTURN IN AND EARLY #ORRESPONDING SIMULATIONS FOR
CONSUMPTION HOURS AND INVESTMENT ALSO INDICATE A MODEST IMPROVEMENT IN çT THESE RESULTS
ARE NOT DISPLAYED
"UT AN EXAMINATION OF WAGE BEHAVIOR OVER THE RECOVERY PHASE OF THE $EPRESSION SUGGESTS
THAT IT IS NOT REASONABLE TO ATTEMPT TO RESCUE A STICKY WAGE EXPLANATION FOR THE RECOVERY
BY CONSIDERING LOWER ABSOLUTE VALUES OF o 4HIS IS BECAUSE THE STICKY WAGE MODEL FAILS
DRAMATICALLY ON TWO IMPORTANT FEATURES OF THE RECOVERY PHASE &IRST IN THE ABSENCE OF OTHER
SHOCKS THE STICKY WAGE MODEL PREDICTS THAT THE RECOVERY SHOULD HAVE OCCURRED AS DECLINING
REAL WAGES ENCOURAGED AN EXPANSION IN LABOR HOURS WORKED "UT AS çGURE SHOWS REAL
WAGES ROSE SHARPLY BEGINNING IN INCREASING BY PERCENT BETWEEN AND
!S INDICATED BY COMPARING çGURES AND REAL WAGES ROSE DESPITE A MARKED INCREASE IN THE
PRICE LEVEL ASSOCIATED WITH REMONETIZATION AS NOMINAL WAGES ROSE MUCH MORE SHARPLY THAN
PRICES 3ECOND THE MODELÚS IMPLICATION OF A STRONG RECOVERY IN HOURS WORKED LEADING THE
RECOVERY IS ALSO INCONSISTENT WITH THE DATA 2ETURNING TO çGURE OBSERVED HOURS WORKED
RECOVERED MUCH MORE TEPIDLY THAN PREDICTED BY THE MODEL Õ AND MUCH MORE SLOWLY THAN
TOTAL OUTPUT AS SEEN BY THE BEHAVIOR OF ACTUAL OUTPUT AND HOURS WORKED IN çGURES AND
(OURS WORKED REMAINED PERCENT BELOW ITS PRE $EPRESSION LEVEL AS LATE AS THE END OF
WHILE OUTPUT HAD RECOVERED TO WITHIN PERCENT OF ITS PRE $EPRESSION LEVEL
4HE BEHAVIOR OF REAL WAGES AND HOURS WORKED DURING THE RECOVERY PHASE OF THE $EPRESSION
SUGGESTS THAT THE RAPID RUN UP IN WAGES BEGINNING IN MID PLAYED A KEY ROLE IN DAMPENING
THE PACE OF THE RECOVERY 4HE RISE IN WAGES IS DIbCULT TO EXPLAIN GIVEN THE HIGH UNEMPLOYMENT
RATES OF THE PERIOD AS HAS BEEN NOTED BY OTHERS 4HE SHARP INCREASES IN NOMINAL WAGES SEEM
TO BE ATTRIBUTABLE TO VARIOUS LEGISLATION PARTICULARLY THE .)2! THAT BECAME EdECTIVE IN
MID 7EINSTEIN 4HIS LEGISLATION ESSENTIALLY SET WAGE SCHEDULES FOR A BROAD SET OF
INDUSTRIES 4HE FACT THAT MOST OF THE NOMINAL WAGE INCREASE OCCURRED OVER A THREE MONTH
PERIOD IN Õ NOMINAL WAGES ROSE BY PERCENT BETWEEN *ULY AND /CTOBER Õ SUGGESTS
THE çAT NATURE OF THE WAGE JUMP THOUGH INCREASED UNION POWER BEGINNING IN THE EARLY S
MAY HAVE ALSO CONTRIBUTED TO UPWARD PRESSURE ON NOMINAL WAGES /THER FACTORS IN ADDITION
TO THE REAL WAGE HIKES PROBABLY DELAYED THE RECOVERY &OR EXAMPLE "ERNANKE HAS
EMPHASIZED THAT WIDESPREAD BANK FAILURES DURING THE PERIOD MAY HAVE EXERTED A STRONG
CONTRACTIONARY IMPACT BEYOND THEIR DIRECT EdECTS ON MONETARY AGGREGATES BY RAISING THE
COST OF CREDIT INTERMEDIATION
4HUS WHILE THE SUSTAINED MONETARY EXPANSION THAT BEGAN IN UNDOUBTEDLY HELPED
AMELIORATE SOME OF THE CONTRACTIONARY EdECTS OF THE NOMINAL WAGE HIKES AND OTHER FACTORS WE
CONCLUDE THAT THE MONETARY EXPANSION WAS NOT NEARLY STRONG ENOUGH TO EXPLAIN THE RECOVERY
!LTHOUGH PROVIDING AN ACCOUNT OF THE KEY FACTORS THAT CONTRIBUTED TO THE RECOVERY IS BEYOND
THE SCOPE OF THIS PAPER IT IS INTERESTING TO NOTE THAT RECENT LITERATURE HAS SUGGESTED THAT
PRODUCTIVITY IMPROVEMENTS MAY HAVE PLAYED A SUBSTANTIAL ROLE IN THE RECOVERY 4HERE IS
SOME INDUSTRY EVIDENCE THAT PRODUCTIVITY INCREASED MARKEDLY IN THE RECOVERY PHASE OF THE
$EPRESSION BECAUSE OF SHAKE OUTS OF INEbCIENT PRODUCERS &OR EXAMPLE "RESNAHAN AND 2Ad
çND THAT AUTOMOBILE PRODUCTION BY HAD RETURNED TO ITS LEVEL ALTHOUGH
THERE WAS A CONSIDERABLE CHANGE IN THE MIX OF PRODUCERS -OST OF THE INCREASE IN OUTPUT
FROM VERY DEPRESSED LEVELS IN OCCURRED THROUGH AN EXPANSION OF ESTABLISHMENTS THAT
HAD RELATIVELY HIGH PRODUCTIVITY LEVELS IN AND TO A LESSER EXTENT BY ENTRY OF NEW çRMS
THAT HAD A SIMILARLY HIGH LEVEL OF EbCIENCY MEASURED BY OUTPUT PER WORKER
4HIS INDUSTRY LEVEL EVIDENCE IS SUPPORTED AT A MORE AGGREGATE LEVEL IN RECENT WORK BY
#ECCHETTI AND +ARRAS 4HESE AUTHORS USING VARIOUS SETS OF RESTRICTIONS TO IDENTIFY
THEIR VECTOR AUTOREGRESSIONS çND THAT SUPPLY SHOCKS ACCOUNTED FOR A LARGE FRACTION OF OUTPUT
GROWTH DURING THE RECOVERY PERIOD OF THE $EPRESSION 7HILE THIS EVIDENCE IS SUGGESTIVE THAT
SUPPLY SHOCKS MAY HAVE PLAYED AN IMPORTANT ROLE IN OdSETTING THE OTHER CONTRACTIONARY EdECTS
OF THE REAL WAGE HIKES PROVIDING A FULLER EXPLANATION FOR THE DELAYED ECONOMIC RECOVERY OF
THE S CLEARLY REMAINS AN IMPORTANT TOPIC FOR FUTURE RESEARCH
#ONTRASTING THE 'REAT $EPRESSION AND THE RECESSION
/UR RESULTS ABOVE SUGGEST THAT A RELATIVELY SIMPLE MODEL DRIVEN BY MONETARY SHOCKS OPERATING
THROUGH A STICKY WAGE CHANNEL CAN ACCOUNT SURPRISINGLY WELL FOR THE BEHAVIOR BOTH OF THE
EXPENDITURE COMPONENTS OF '$0 AND OF WAGES OVER THE DOWNTURN AND BOTTOMING OUT PHASES
OF THE 'REAT $EPRESSION (OWEVER IT IS USEFUL TO DRAW ATTENTION TO TWO
FACTORS THAT PLAY AN IMPORTANT ROLE IN THE MODELÚS ABILITY TO ACCOUNT FOR THE SEVERITY OF THE
$EPRESSION 4HESE FACTORS INCLUDE THE UNANTICIPATED NATURE OF THE PRICE LEVEL DECLINE
AND SUbCIENTLY SLUGGISH WAGE ADJUSTMENT THAT ALLOWS MONETARY SHOCKS TO EXERT EXTREMELY
PERSISTENT REAL EdECTS /UR ANALYSIS BELOW HOWEVER SUGGESTS THAT THESE TWO FACTORS WERE NOT
AS IMPORTANT IN THE EARLY S
/UR MODELÚS IMPLICATION THAT THE PRICE LEVEL DECLINE WAS LARGELY UNANTICIPATED Õ A FEATURE
EVIDENT IN THE HIGHLY FRONT LOADED )2&S OF THE PRICE LEVEL FROM MONETARY SHOCKS Õ TENDS TO
MAXIMIZE THE REAL EdECTS OF A GIVEN SIZED DECLINE IN THE PRICE LEVEL 3PECIçCATIONS IN WHICH A
LARGER FRACTION OF THE PRICE DECLINE IS ANTICIPATED PRODUCE SMALLER REAL EdECTS &OR EXAMPLE IF
A FUTURE PRICE DECLINE IS ANTICIPATED SEVERAL QUARTERS IN ADVANCE NOMINAL WAGES WOULD BEGIN
ADJUSTING DOWNWARD SOONER IN RESPONSE TO EXPECTED FUTURE REDUCTIONS IN THE QUANTITY OF LABOR
HOURS THAT çRMS WILL DEMAND #ONSEQUENTLY REAL WAGES WILL NOT RISE AS MUCH WHEN THE PRICE
LEVEL DECLINE IS ANTICIPATED
4O ASSESS THE QUANTITATIVE SIGNIçCANCE OF THIS çGURE COMPARES THE BEHAVIOR OF THE
SIMULATED PATH OF OUTPUT DERIVED FROM THE 4AYLOR MODEL ABOVE IN &IGURE WITH THE OUTPUT
PATH DERIVED UNDER THE ASSUMPTION THAT THE PRICE DECLINE WAS FULLY ANTICIPATED AS OF
4HE MODELÚS IMPLICATIONS FROM THIS PERFECT FORESIGHT SIMULATION ARE OBTAINED BY SOLVING THE
4AYLOR CONTRACTS COMPONENT OF THE MODEL ASSUMING THAT AGENTS ALSO HAVE PERFECT FORESIGHT
ABOUT THE ACTUAL PATH OF THE CAPITAL STOCK 3PECIçCALLY THE AVERAGE WAGE 6 IS DEçNED
S
BY EQUATION A AND IS A GEOMETRIC AVERAGE OF THE CONTRACT WAGES FW W ` W ` W ` G
S
3UBSTITUTING THE CONTRACT WAGES INTO
S
S
S
AND LOG LINEARIZING YIELDS A LINEAR RELATIONSHIP
BETWEEN THE LOGARITHMS OF W W ` W ` W ` / * AND + 3OLVING THIS RELATIONSHIP FOR +
S
S
S
S
S
S
S
S
THE ONLY UNKNOWN VARIABLES ARE THE CONTRACT WAGES SINCE / AND * ARE ASSUMED TO BE KNOWN
S
S
UNDER THE PERFECT FORESIGHT ASSUMPTIONS 3UBSTITUTING THIS SOLUTION FOR + INTO THE 4AYLOR
S
CONTRACTS EQUATION YIELDS A SINGLE DYNAMIC EQUATION FOR THE CONTRACT WAGE W IN TERMS
S
OF PAST AND FUTURE CONTRACT WAGES AS WELL AS CURRENT AND FUTURE PRICE LEVELS AND CAPITAL
STOCKS 'IVEN THE SOLUTION FOR W THE AVERAGE WAGE 6 AND LABOR HOURS + ARE DETERMINED
S
S
S
THE PRODUCTION FUNCTION DETERMINES OUTPUT
)T IS EVIDENT THAT THESE ÝPOLAR OPPOSITEÞ ASSUMPTIONS ABOUT THE DEGREE TO WHICH THE
$EPRESSIONÚS PRICE DECLINE WAS ANTICIPATED LEAD TO SUBSTANTIAL DIdERENCES IN THE PREDICTED
OUTPUT DECLINE 4HE MODEL SUGGESTS THAT IF THE FUTURE PRICE LEVEL DECLINE DURING THE $EPRESSION
BECAME FULLY ANTICIPATED IN OUTPUT WOULD HAVE CONTRACTED ONLY ABOUT PERCENT EVEN
IF THE CAPITAL STOCK DECLINED AS SHARPLY AS ACTUALLY OCCURRED COMPARED TO PERCENT IN THE
MODEL IN WHICH THE PRICE DECLINE WAS UNANTICIPATED 4O THE EXTENT THAT PERFECT FORESIGHT
ABOUT DEâATION WOULD HAVE TRULY AMELIORATED THE DATAÚS FALL IN THE CAPITAL STOCK THE OUTPUT
DECLINE IMPLIED BY THE PERFECT FORESIGHT SIMULATIONS WOULD BE SOMEWHAT SMALLER
7E INTERPRET THE EXISTING EMPIRICAL EVIDENCE AS OdERING SUPPORT FOR THE HYPOTHESIS THAT
AT LEAST A SUBSTANTIAL COMPONENT OF THE PRICE DECLINE DURING THE $EPRESSION WAS UNANTICI
PATED )N PARTICULAR THE MAGNITUDE OF THE PRICE DECLINE WAS UNPRECEDENTED IN THE PEACETIME
SIMULATE THE EVOLUTION OF XT FROM A POINT IN TIME INITIAL VALUES OF THE PAST CONTRACT WAGES AND
FUTURE PATHS OF THE PRICE LEVEL AND CAPITAL STOCK ARE REQUIRED 7HEN WE SAY THAT THE PRICE DECLINE WAS FULLY
ANTICIPATED AS OF WE ASSUME THAT THE INITIAL CONTRACT WAGES FXT XT` XT` XT` G
EQUAL THEIR STEADY STATE VALUES AND ANY INçNITE SUMS OF DISCOUNTED CURRENT AND FUTURE PRICE LEVELS AND
CAPITAL STOCKS ARE WELL APPROXIMATED BY THEIR çVE YEAR çNITE DISCOUNTED SUMS OF FUTURE VALUES /N THIS LATTER
ASSUMPTION OUR CALCULATIONS WERE QUALITATIVELY ROBUST TO TRUNCATING THE RELEVANT INçNITE SUMS AT THREE OR
FOUR YEAR HORIZONS
4HESE SIMULATIONS ARE CONDUCTED USING THE SAME WAGE ADJUSTMENT PARAMETER o `
4O
HISTORY OF THE 5NITED 3TATES AND IT FOLLOWED AN EIGHT YEAR PERIOD OF PRICE STABILITY &RIEDMAN
AND 3CHWARTZ 4HIS SUGGESTS THAT A LARGE FRACTION OF THE DECLINE WAS UNANTICIPATED
(AMILTON PROVIDES MORE FORMAL EVIDENCE THAT A LARGE COMPONENT OF THE DEâATION WAS
UNANTICIPATED ESPECIALLY IN THE çRST TWO YEARS OF THE DOWNTURN (AMILTON EXAMINES THE TIME
SERIES BEHAVIOR OF COMMODITY FUTURES PRICES AND CONCLUDES THAT THE MARKET BELIEVED THAT
PRICES WOULD REMAIN STABLE OR RISE DURING MOST OF THE $EPRESSIONÚS DOWNTURN /F COURSE
THE SPECIçC IMPLICATION OF OUR MODEL THAT THE PRICE DECLINE WAS COMPLETELY UNANTICIPATED
IS SUBJECT TO DEBATE SEE #ECCHETTI 3TILL THE IMPORTANT CONCLUSION TO BE DRAWN IS
THAT THE REAL EdECTS OF THE PRICE DECLINE PREDICTED BY OUR MODEL DEPEND DIRECTLY ON THE EX
TENT TO WHICH THE PRICE DECLINE WAS UNANTICIPATED 4HUS IF THE PRICE DECLINE WAS LARGELY
UNANTICIPATED THE REAL EdECTS WOULD BE LARGE AND CLOSE TO THOSE SHOWN IN OUR SIMULATIONS
(OWEVER IF THE PRICE DECLINE WERE LARGELY FORESEEN THE STICKY WAGE MODEL WOULD HAVE MUCH
MORE DIbCULTY EXPLAINING THE MAGNITUDE OF THE DOWNTURN WITHOUT REFERENCE TO OTHER FACTORS
! SECOND IMPORTANT FACTOR IS THAT WAGE ADJUSTMENT IN THE MODEL MUST BE SUbCIENTLY
SLUGGISH FOR MONETARY POLICY TO HAVE HIGHLY PERSISTENT EdECTS 4HIS IS EVIDENT FROM OUR
COMPARISON OF THE &ISCHER AND 4AYLOR CONTRACT SPECIçCATIONS ABOVE )N PARTICULAR IN ORDER
TO ACCOUNT WELL FOR THE DEPTH OF THE DOWNTURN MONETARY SHOCKS MUST HAVE REAL EdECTS THAT
ARE SOMEWHAT LONGER THAN THE ONE YEAR PERIOD IMPLIED BY OUR SPECIçCATION OF THE &ISCHER
CONTRACT MODEL 4HIS HOLDS DESPITE THE FACT THAT THE MODEL BASICALLY ALLOWS CONTRACTIONARY
MONETARY SHOCKS TO EXERT MAXIMAL EdECTS ON OUTPUT FOR A GIVEN LEVEL OF WAGE SLUGGISHNESS
!LSO
SEE %VANS AND 7ACHTEL ON THIS POINT
AN IMPLICATION THAT THE PRICE LEVEL DECLINE WAS UNEXPECTED (OWEVER THE FACT THAT OUR
PARAMETERIZATION OF WAGE ADJUSTMENT çTS THE BEHAVIOR OF BOTH REAL AND NOMINAL WAGES QUITE
WELL PROVIDES SUPPORT FOR OUR SPECIçCATION
A PUZZLING INTERPRETATION
7HILE THE UNANTICIPATED CHARACTER OF THE PRICE DECLINE AND SLUGGISH WAGE ADJUSTMENT APPEAR
IMPORTANT IN ACCOUNTING FOR THE SEVERITY OF THE $EPRESSION THESE SAME FACTORS DO NOT APPEAR
TO EXPLAIN WHY THE OUTPUT DECLINE DURING THE RECESSION WAS MUCH LESS SEVERE THAN
DURING THE 'REAT $EPRESSION !FTER ALL PRICES DECLINED BY ROUGHLY THE SAME CUMULATIVE
MAGNITUDE DURING THESE TWO EPISODES
&IGURES AND COMPARE THE PRICE DECLINE THAT OCCURRED DURING THE PERIOD
WITH THE DECLINE THAT OCCURRED DURING THE 'REAT $EPRESSION 4HE PRICE LEVEL IN THESE çGURES
IS REPRESENTED AS A PERCENTAGE DEVIATION FROM ITS AND LEVEL RESPECTIVELY 4HE
MOST STRIKING FEATURE OF THE çGURE IS THAT THE PRICE DECLINE DURING THE RECESSION WAS
INITIALLY MUCH SHARPER THAN WHAT OCCURRED DURING THE EARLY DOWNTURN PHASE OF THE 'REAT
$EPRESSION )N PARTICULAR THE PRICE LEVEL FELL PERCENT BETWEEN WHEN THE PRICE LEVEL
REACHED AN INTERWAR PEAK AND "ETWEEN AND THE PRICE LEVEL FELL BY ONLY
PERCENT &RIEDMAN AND 3CHWARTZ NOTE THAT THE PRICE DECLINE IN WAS THE
STEEPEST PRICE DECLINE EXPERIENCED IN THE 5NITED 3TATES SINCE AT LEAST THE EARLY AFTERMATH OF
THE #IVIL 7AR AND PERHAPS WAS THE LARGEST THAT EVER OCCURRED IN THE 5NITED 3TATES
&IGURE INDICATES THAT OUTPUT CONTRACTED SHARPLY BEGINNING IN AND FELL BY
PERCENT OVER THE SUBSEQUENT YEAR Õ SLIGHTLY LARGER THAN THE PERCENT DECLINE THAT OCCURRED
DURING THE çRST YEAR OF THE $EPRESSION çGURE (OWEVER THE SUBSEQUENT BEHAVIOR OF
OUTPUT SHOWS A MAJOR DIdERENCE BETWEEN THE TWO PERIODS /UTPUT REBOUNDED STRONGLY FROM
ITS CYCLICAL TROUGH IN THE DOWNTURN BEGINNING IN REACHING ITS PRE RECESSION
LEVEL BY MID IN STRONG CONTRAST TO THE PERSISTENT OUTPUT DECLINES THAT OCCURRED IN
)F THE PRICE LEVEL DECLINE THAT OCCURRED IN WAS IN FACT LARGELY UNANTICIPATED AND
WAGES ADJUSTED AS SLUGGISHLY AS IN OUR PARAMETERIZATION OF THE 'REAT $EPRESSION PERIOD
THEN THE PRICE DECLINE OF Õ BASED ON OUR EARLIER RESULTS Õ SHOULD HAVE GENERATED A
MASSIVE AND PERSISTENT DEPRESSION 4HIS INTUITION IS CONçRMED IN &IGURE WHICH INDICATES
THAT OUTPUT SHOULD HAVE CONTRACTED BY ALMOST PERCENT IN LOGARITHMIC TERMS OR BY ALMOST
ONE THIRD FROM ITS LEVEL IN 4HE SIMULATION IN &IGURE IS DERIVED USING ESSENTIALLY
THE SAME PARAMETERIZATION OF THE 4AYLOR CONTRACTS MODEL THAT WE CONSIDERED FOR THE
PERIOD 4HE ONLY DIdERENCE IN SPECIçCATION IS OUR INCLUSION OF BOTH MONEY SUPPLY AND MONEY
DEMAND SHOCKS AND THESE SHOCKS WERE COMPUTED TO çT THE OBSERVED BEHAVIOR OF MONEY AND
THE PRICE LEVEL
4HIS IS THE SAME PROCEDURE THAT UNDERLIES THE ANALYSIS RELATING TO &IGURE 7E ADD MONEY DEMAND
INNOVATIONS BECAUSE THIS PERIOD WAS CHARACTERIZED BY A STRONG RISE IN MONEY VELOCITY IE THE OBSERVED -
CONTRACTION OF AROUND PERCENT OVER THE PERIOD WAS FAR TOO SMALL TO ACCOUNT FOR THE PERCENT
PRICE DECLINE THAT OCCURRED OVER THIS PERIOD 4HE CONSTRUCTION OF THIS SIMULATION CONFORMS TO THE OBJECTIVE
OF ASSESSING HOW THE PRICE DECLINE WOULD HAVE AdECTED REAL ACTIVITY IF IT HAD IN FACT BEEN UNANTICIPATED
#ONTRASTING WITH
4HERE IS EVIDENCE THAT THERE WERE CONSIDERABLE DIdERENCES BETWEEN THE TWO PERIODS IN THE
EXTENT TO WHICH THE PRICE DECLINES WERE ANTICIPATED AND IN THE SENSITIVITY OF NOMINAL WAGES
TO THE DOWNTURN IN ACTIVITY 4HESE FACTORS MAY HAVE PLAYED A LARGE ROLE IN ACCOUNTING FOR
PRONOUNCED DIdERENCES IN THE BEHAVIOR OF REAL ACTIVITY
&IRST IT SEEMS LIKELY THAT THE DISINâATION EXPERIENCED IN HAD A MUCH LARGER
ANTICIPATED COMPONENT 0REVIOUS MAJOR WARS INCLUDING THE 7AR OF AND THE #IVIL 7AR
WERE FOLLOWED BY LARGE DEâATIONS 4HE MONETARY AUTHORITIES WERE COMMITTED TO A RETURN TO
GOLD CONVERTIBILITY AT THE ORIGINAL PARITY FOLLOWING WARTIME SUSPENSIONS OF GOLD CONVERTIBILITY
"ORDO AND +YDLAND 7ORLD 7AR ) SET THE STAGE FOR A SIMILAR EXPERIENCE DURING THIS
PERIOD THE 53 PRICE LEVEL MORE THAN DOUBLED BETWEEN AND EARLY )T WAS CLEAR
TO CONTEMPORARIES OF THE PERIOD THAT A LARGE DEâATION WAS REQUIRED TO RESTORE THE REAL PRICE
OF GOLD TO ITS PREWAR LEVEL AT THE ORIGINAL PARITY )T ALSO SEEMS THAT MOST OBSERVERS WOULD
HAVE EXPECTED CORRECTLY THAT THE 5NITED 3TATES HAD A STRONG COMMITMENT TO RESTORING
THE PREWAR RELATIONSHIP BETWEEN PRICES AND THE PRICE OF GOLD !S NOTED BY &RIEDMAN AND
3CHWARTZ THE âEXIBLE EXCHANGE RATES THAT PREVAILED DURING THE EARLY INTERWAR PERIOD
ÝWERE REGARDED AS A TRANSITORY EXPEDIENT PENDING A RETURN TO GOLD AND MONETARY AUTHORITIES
EVERYWHERE SOUGHT TO FACILITATE A RETURN TO çXED PARITIES P Þ !CCORDINGLY THE MAJOR
SOURCE OF UNCERTAINTY ABOUT THE DEâATION IN THE PERIOD IMMEDIATELY PRECEDING THE MONETARY
TIGHTENING PROBABLY REGARDED ITS TIMING /NCE THE &EDERAL 2ESERVE RAISED THE DISCOUNT RATE
BY BASIS POINTS BETWEEN .OVEMBER AND &EBRUARY THERE COULD HAVE BEEN LITTLE
DOUBT ABOUT THE &EDERAL 2ESERVEÚS COMMITMENT TO DEâATE OR THAT A LARGE DISINâATION WOULD
ENSUE
! SECOND CONSIDERATION THAT MAY ACCOUNT FOR THE LESS PERSISTENT REAL EdECTS OF THE PRICE
DECLINE IN THE RECESSION COMPARED TO THE 'REAT $EPRESSION PERIOD IS THAT WAGE SETTING
APPEARS TO HAVE BEEN MUCH MORE RESPONSIVE TO THE DOWNTURN IN ECONOMIC CONDITIONS DURING
THE EARLIER PERIOD %CONOMISTS WRITING DURING THE DEPRESSION PERIOD NOTED THE PRONOUNCED
DIdERENCES IN WAGE ADJUSTMENT ACROSS THE TWO EPISODES A SUBJECT EXAMINED MORE RECENTLY
IN WORK BY /Ú"RIEN /Ú"RIEN NOT ONLY PROVIDES STRONG EVIDENCE THAT WAGES
BECAME LESS RESPONSIVE TO AGGREGATE DEMAND BETWEEN THE RECESSION AND THE LATE
S BUT THAT THE CHANGE WAS RECOGNIZED BY CONTEMPORARIES OF THE PERIOD AND REâECTED A
CONSCIOUS CHANGE IN WAGE SETTING PRACTICES AT THE çRM LEVEL DESIGNED TO MOLLIFY THE ADVERSE
EdECTS OF WAGE DECLINES ON HOUSEHOLD PURCHASING POWER /Ú"RIEN OBSERVES THAT THE VIEW
BECAME WIDELY HELD DURING THE INTERWAR PERIOD THAT RAPID WAGE ADJUSTMENT DURING THE
RECESSION CONTRIBUTED MARKEDLY TO THE RECESSIONÚS SEVERITY (E ARGUES THAT THIS PHILOSOPHY
ENCOURAGED COLLUSIVE BEHAVIOR ON THE PART OF MAJOR EMPLOYERS TO NOT CUT NOMINAL WAGES
FOR A SIGNIçCANT TIME PERIOD AFTER THE ONSET OF THE $EPRESSION /Ú"RIEN NOTES THAT CERTAIN
MAJOR EMPLOYERS SUCH AS 'ENERAL -OTORS AND )NTERNATIONAL (ARVESTER LEFT NOMINAL WAGE
RATES UNCHANGED UNTIL THE FALL OF TWO YEARS INTO THE DOWNTURN &INALLY THE CONTEXT OF
THE TWO PERIODS SUGGESTS A CHANGE IN WAGE SETTING BEHAVIOR 4HE EXTREME PRICE âUCTUATIONS
DURING THE PERIOD WOULD HAVE ENCOURAGED OR PERHAPS FORCED GREAT WAGE âEXIBILITY
4HE
&EBRUARY INCREASE IN THE DISCOUNT RATE WAS THE ÝSHARPEST SINGLE RISE IN THE ENTIRE HISTORY OF THE
3YSTEMÞ &RIEDMAN AND 3CHWARTZ P
4HE 'REAT $EPRESSION HOWEVER FOLLOWED ALMOST A DECADE OF PRICE STABILITY
REVISITED
&IGURES AND ASSESS WHETHER ALTERNATIVE ASSUMPTIONS ABOUT PRICE EXPECTATIONS AND WAGE
ADJUSTMENT DURING THE PERIOD MAY HELP ACCOUNT FOR THE MUCH SMALLER OUTPUT DECLINE
THAT OCCURRED &IGURE PERFORMS A SIMULATION OF THE MODEL UNDER THE ASSUMPTION THAT THE
PRICE DECLINE BECAME FULLY ANTICIPATED AS OF THAT IS THE QUARTER AFTER THE PRICE LEVEL
PEAKED AND SHORTLY AFTER THE MONETARY TIGHTENING BEGAN !S IN THE CORRESPONDING PERFECT
FORESIGHT SIMULATION RUN OVER THE $EPRESSION PERIOD IN &IGURE THE çGURE IS BASED ON
SIMULATING THE 4AYLOR WAGE SETTING COMPONENT OF THE MODEL ONLY TAKING THE CAPITAL STOCKÚS
ACTUAL PATH AS KNOWN 4HE çGURE INDICATES THAT IF THE PRICE DECLINE BECAME FULLY ANTICIPATED
BY THE TIME PRICES ACTUALLY BEGAN TO FALL THE OUTPUT EdECTS WOULD HAVE ONLY BEEN ABOUT HALF
AS LARGE AS IMPLIED BY THE SIMULATION IN çGURE Õ THOUGH STILL SOMEWHAT LARGER THAN WHAT
ACTUALLY TOOK PLACE
&IGURE REPEATS THE SIMULATION IN çGURE EXCEPT THAT WAGE ADJUSTMENT IS ASSUMED
TO OCCUR CONSIDERABLY MORE RAPIDLY THAN IN THE PARAMETERIZATION OF THE PERIOD )N
PARTICULAR WE SET o ` IMPLYING THAT THE EdECT OF PRICE LEVEL INNOVATIONS ON OUTPUT
HAVE A HALF LIFE OF ABOUT HALF AS LONG AS IN OUR PARAMETERIZATION 4HE çGURE IMPLIES
THAT OUTPUT WOULD HAVE ONLY DECLINED BY ABOUT PERCENT GIVEN THE PRICE LEVEL DECLINE
THAT ACTUALLY OCCURRED REASONABLY CLOSE TO THE MAGNITUDE OBSERVED
4HESE SIMULATIONS MAKE IT MORE APPARENT WHY A QUANTITATIVE GENERAL EQUILIBRIUM MODEL
IS USEFUL FOR EVALUATING THE STICKY WAGE HYPOTHESIS /UR RESULTS SUGGEST THAT IT IS NOT ENOUGH
TO ARGUE THAT THE MASSIVE PRICE DECLINE THAT OCCURRED DURING THE 'REAT $EPRESSION PERIOD
CAUSED THE DECLINE VIA A STICKY WAGE CHANNEL 2ATHER OUR RESULTS SUGGEST THAT THE UNANTICI
PATED CHARACTER OF THE PRICE DECLINE PROBABLY PLAYED AN IMPORTANT ROLE IN EXPLAINING WHY THE
DOWNTURN WAS SO LARGE &URTHERMORE IT IS EVIDENT THAT WHILE THIS HYPOTHESIS IS APPEALING
INSOFAR AS IT CAN MATCH THE DOWNTURN IN MAJOR MACROAGGREGATES AND THE BEHAVIOR OF REAL
WAGES IT DOES REQUIRE THAT MONETARY SHOCKS HAVE HIGHLY PERSISTENT OUTPUT EdECTS .EVERTHE
LESS THIS DEGREE OF PERSISTENCE IS CONSISTENT WITH THE ENDOGENOUS BEHAVIOR OF WAGES DURING
THE PERIOD &INALLY A COMPARISON WITH THE RECESSION FURTHER HIGHLIGHTS THE
PERIOD AS AN OUTLIER EXPERIENCE A SIMILAR SIZED PRICE DECLINE HAD MUCH SMALLER OUTPUT EdECTS
DURING THE EARLIER DOWNTURN PROBABLY BECAUSE A MUCH LARGER FRACTION OF THE PRICE DECLINE
WAS ANTICIPATED IN
#ONCLUSIONS
/UR PAPER HAS FOCUSED ON THE ABILITY OF A PARTICULAR HYPOTHESIS ABOUT HOW MONETARY SHOCKS
AdECT REAL ACTIVITY Õ NAMELY THROUGH A STICKY WAGE CHANNEL Õ TO EXPLAIN THE MAGNITUDE AND
PERSISTENCE OF THE $EPRESSION IN THE 5NITED 3TATES /UR RESULTS SUGGEST THAT MONEY OPERATING
THROUGH A STICKY WAGE CHANNEL PLAYED A QUANTITATIVELY SIGNIçCANT ROLE IN ACCOUNTING FOR THE
BEHAVIOR OF MAJOR MACROAGGREGATES OVER THE DOWNTURN PHASE OF THE $EPRESSION 4HE VERSION
OF THE MODEL WITH 4AYLOR CONTRACTS IS SURPRISINGLY SUCCESSFUL IN ACCOUNTING FOR THE BEHAVIOR
OF OUTPUT CONSUMPTION HOURS WORKED AND INVESTMENT )N ADDITION THE MECHANISM THROUGH
WHICH THE STICKY WAGE MODEL PREDICTS THAT A MONETARY DECLINE WOULD AdECT THESE VARIABLES
SEEMS CONSISTENT WITH THE DATA 4HIS IS ESPECIALLY REâECTED IN THE MODELÚS ABILITY TO ACCOUNT
FOR THE BEHAVIOR OF REAL WAGES OVER THE DOWNTURN
7HILE IT IS OF COURSE LIKELY THAT OTHER FACTORS DISCUSSED IN THE INTRODUCTION PLAYED A ROLE
IN THE DOWNTURN THE QUANTITATIVE SUCCESS OF THE SIMPLE MODEL WE CONSIDER IS INTRIGUING
-OREOVER THE STICKY WAGE MODELÚS ABILITY TO ACCOUNT FOR THE RISE IN REAL WAGES IN ADDITION TO
THE BEHAVIOR OF THE OTHER MACROAGGREGATES MAY PROVIDE AN IMPORTANT BASIS FOR DISCRIMINATING
BETWEEN ALTERNATIVE HYPOTHESES ABOUT THE CAUSES OF THE DOWNTURN 4HIS IS BECAUSE IT IS NOT
APPARENT THAT OTHER HYPOTHESES WOULD YIELD PREDICTIONS CONSISTENT WITH OBSERVED REAL WAGE
BEHAVIOR EVEN IF THEY COULD PROVIDE A SATISFACTORY ACCOUNT OF THE OTHER MACROAGGREGATES
/UR MODEL SIMULATIONS AND SENSITIVITY ANALYSIS SUGGEST THAT BOTH THE ABSOLUTE SEVERITY OF
THE 'REAT $EPRESSIONÚS DOWNTURN AND ITS RELATIVE SEVERITY COMPARED TO THE RECESSION
WERE IN LARGE PART ATTRIBUTABLE TO THE PRICE DECLINE HAVING A MUCH LARGER UNANTICIPATED COM
PONENT DURING THE $EPRESSION AND ALSO TO MUCH LESS âEXIBLE WAGE SETTING PRACTICES 4HESE
RESULTS SEEM TO BE AN IMPORTANT REçNEMENT TO THE STICKY WAGE EXPLANATION FOR THE DOWNTURN
4HE çNDINGS ARE IN LINE WITH BOTH (AMILTONÚS RESULTS THAT SHOW A SUBSTANTIAL COMPO
NENT OF THE PRICE DECLINE TO BE UNANTICIPATED AND WITH /Ú"RIENÚS WORK THAT INDICATES
WAGE SETTING BECAME MUCH LESS âEXIBLE DURING THE COURSE OF THE S
7ITHIN THE CONTEXT OF OUR STICKY WAGE MODEL HOWEVER OUR RESULTS CAST SERIOUS DOUBTS ON
EXPLANATIONS OF THE RECOVERY PHASE OF THE $EPRESSION WHICH RELY CRITICALLY ON THE SUBSTANTIAL
REMONETIZATION THAT BEGAN IN 4HE MOST SERIOUS EVIDENCE AGAINST THIS EXPLANATION IS
ITS COUNTERFACTUAL PREDICTION THAT REMONETIZATION SHOULD HAVE REDUCED REAL WAGES STIMULATED
LABOR DEMAND AND ALLOWED OUTPUT TO RECOVER "Y CONTRAST THE DATA INDICATE THAT REAL WAGES
ROSE SUBSTANTIALLY BEGINNING IN MID AS NOMINAL WAGES ROSE EVEN MORE RAPIDLY THAN
PRICES AND THAT HOURS WORKED RECOVERED MUCH MORE SLOWLY THAN OUTPUT 7HILE THERE IS SOME
INDUSTRY AND AGGREGATE EVIDENCE THAT POSITIVE SUPPLY SHOCKS PLAYED A ROLE IN ALLOWING THE
ECONOMY TO GRADUALLY RECOVER DESPITE SOME LARGE EXOGENOUS INCREASES IN REAL WAGES DUE TO
LEGISLATIVE PROGRAMS PROVIDING A CONVINCING ACCOUNT OF THE FACTORS MOST RESPONSIBLE FOR THE
RECOVERY CONTINUES TO BE AN IMPORTANT TOPIC FOR FURTHER RESEARCH
2EFERENCES
"ALKE .ATHAN 3 AND 2OBERT * 'ORDON !PPENDIX " (ISTORICAL $ATA IN 2OBERT
* 'ORDON ED 4HE !MERICAN "USINESS #YCLE #ONTINUITY AND #HANGE #HICAGO
5NIVERSITY OF #HICAGO 0RESS
"ENEY - !DA 7AGES (OURS AND %MPLOYMENT IN THE 5NITED 3TATES .EW
9ORK .ATIONAL )NDUSTRIAL #ONFERENCE "OARD )NC
"ERNANKE "EN Ý.ONMONETARY EdECTS OF THE &INANCIAL #RISIS IN THE 0ROPAGATION OF THE
'REAT $EPRESSIONÞ !MERICAN %CONOMIC 2EVIEW *UNE
"ERNANKE "EN Ý4HE -ACROECONOMICS OF THE 'REAT $EPRESSION ! #OMPARATIVE !P
PROACHÞ *OURNAL OF -ONEY #REDIT AND "ANKING
"ERNANKE "EN
AND +EVIN #AREY Ý.OMINAL 7AGE 3TICKINESS AND !GGREGATE SUPPLY
IN THE 'REAT $EPRESSIONÞ 1UARTERLY *OURNAL OF %CONOMICS
"LANCHARD /LIVIER * AND #HARLES +AHN Ý4HE 3OLUTION OF ,INEAR $IdERENCE %QUATIONS
5NDER 2ATIONAL %XPECTATIONS Þ %CONOMETRICA
"ORDO -ICHAEL $ Ý4HE #ONTRIBUTION OF Ù! -ONETARY (ISTORY OF THE 5NITED 3TATES
Ú TO -ONETARY (ISTORY Þ IN -ONEY (ISTORY AND )NTERNATIONAL &INANCE %S
SAYS IN (ONOR OF !NNA * 3CHWARTZ %DITED BY -$ "ORDO #HICAGO 5NIVERSITY OF
#HICAGO 0RESS
"ORDO -ICHAEL $ AND &INN +YDLAND Ý4HE 'OLD 3TANDARD AS A 2ULE !N %SSAY IN
%XPLORATIONÞ %XPLORATIONS IN %CONOMIC (ISTORY
"ORDO -ICHAEL $ %HSAN #HOUDHRI AND !NNA * 3CHWARTZ Ý#OULD 3TABLE -ONEY
(AVE !VERTED THE 'REAT $EPRESSIONÞ %CONOMIC )NQUIRY
"RESNAHAN 4IMOTHY & AND $ANIEL -' 2Ad Ý)NTERINDUSTRY (ETEROGENEITY AND THE
'REAT $EPRESSION 4HE !MERICAN -OTOR 6EHICLE )NDUSTRY Þ *OURNAL OF %CO
NOMIC (ISTORY
#ALOMIRIS #HARLES Ý&INANCIAL &ACTORS IN THE 'REAT $EPRESSIONÞ *OURNAL OF %CONOMIC
0ERSPECTIVES 3PRING
#ECCHETTI 3TEPHEN ' Ý0RICES AND THE 'REAT $EPRESSION 7AS THE $EâATION OF
2EALLY 5NANTICIPATED Þ !MERICAN %CONOMIC 2EVIEW
#ECCHETTI 3TEPHEN ' AND 'EORGIOS +ARRAS Ý3OURCES OF /UTPUT &LUCTUATIONS DURING
THE )NTERWAR 0ERIOD &URTHER %VIDENCE ON #AUSES OF THE 'REAT $EPRESSIONÞ 2EVIEW OF
%CONOMICS AND 3TATISTICS
#HO *ANG /K AND ,OUIS 0HANEUF Ý! "USINESS #YCLE -ODEL WITH .OMINAL 7AGE
#ONTRACTS AND 'OVERNMENTÞ )NSTITUTE FOR %MPIRICAL -ACROECONOMICS $ISCUSSION PAPER
&EBRUARY
#RUCINI -ARIO AND *AMES +AHN Ý4ARIdS AND !GGREGATE %CONOMIC !CTIVITY ,ESSONS
FROM THE 'REAT $EPRESSIONÞ *OURNAL OF -ONETARY %CONOMICS
%ICHENGREEN "ARRY 'OLDEN &ETTERS 4HE 'OLD 3TANDARD AND THE 'REAT $EPRESSION
.EW 9ORK /XFORD 5NIVERSITY 0RESS A
%ICHENGREEN "ARRY Ý4HE /RIGINS AND .ATURE OF THE 'REAT 3LUMP 2EVISITEDÞ %CO
NOMIC (ISTORY 2EVIEW -AY B
%ICHENGREEN "ARRY AND *EdREY 3ACHS Ý%XCHANGE 2ATES AND %CONOMIC 2ECOVERY IN
THE SÞ *OURNAL OF %CONOMIC (ISTORY
%VANS -ARTIN AND 0AUL 7ACHTEL Ý7ERE 0RICE #HANGES $URING THE 'REAT $EPRESSION
!NTICIPATED %VIDENCE FROM .OMINAL )NTEREST 2ATESÞ *OURNAL OF -ONETARY %CONOMICS
&ISCHER 3TANLEY Ý,ONG TERM #ONTACTS 2ATIONAL %XPECTATIONS AND THE /PTIMAL -ONEY
3UPPLY 2ULEÞ *OURNAL OF 0OLITICAL %CONOMY
&ISHER )RVING Ý4HE $EBT $EâATION 4HEORY OF THE 'REAT $EPRESSIONÞ %CONOMETRICA
&RIEDMAN -ILTON AND !NNA * 3CHWARTZ ! -ONETARY (ISTORY OF THE 5NITED 3TATES
0RINCETON 0RINCETON 5NIVERSITY 0RESS
'RAY *O !NNA Ý7AGE )NDEXATION ! -ACROECONOMIC !PPROACHÞ *OURNAL OF -ONE
TARY %CONOMICS
(AMILTON *AMES $ Ý7AS THE $EâATION $URING THE 'REAT $EPRESSION !NTICIPATED
%VIDENCE FROM THE #OMMODITY &UTURES -ARKETÞ !MERICAN %CONOMIC 2EVIEW
(ANES #HRISTOPHER Ý"ARGAINING 0OWER AND .OMINAL 7AGE 2IGIDITY IN THE $OWNTURN
OF AND Þ 5NPUBLISHED MANUSCRIPT 5NIVERSITY OF 0ENNSYLVANIA A
(ANES #HRISTOPHER Ý#HANGES IN THE #YCLICAL "EHAVIOR OF 2EAL 7AGES Þ
5NPUBLISHED MANUSCRIPT 5NIVERSITY OF 0ENNSYLVANIA B
+ING 2OBERT ' Ý-ONEY AND "USINESS #YCLESÞ 5NPUBLISHED MANUSCRIPT
-ELTZER !LLAN ( Ý-ONETARY AND /THER %XPLANATIONS OF THE 3TART OF THE 'REAT $E
PRESSIONÞ *OURNAL OF -ONETARY %CONOMICS
-ISHKIN &REDERIC 3 Ý4HE (OUSEHOLD "ALANCE 3HEET AND THE 'REAT $EPRESSIONÞ *OUR
NAL OF %CONOMIC (ISTORY $ECEMBER
/Ú"RIEN !NTHONY 0 Ý! "EHAVIORAL %XPLANATION FOR .OMINAL 7AGE 2IGIDITY DURING
THE 'REAT $EPRESSIONÞ 1UARTERLY *OURNAL OF %CONOMICS
2OMER #HRISTINA Ý4HE 'REAT #RASH AND THE /NSET OF THE 'REAT $EPRESSIONÞ 1UAR
TERLY *OURNAL OF %CONOMICS !UGUST
2OMER #HRISTINA Ý7HAT %NDED THE 'REAT $EPRESSION Þ *OURNAL OF %CONOMIC (ISTORY
2OMER #HRISTINA Ý4HE .ATION IN $EPRESSIONÞ *OURNAL OF %CONOMIC 0ERSPECTIVES
3PRING
4AYLOR *OHN " Ý!GGREGATE $YNAMICS AND 3TAGGERED #ONTRACTSÞ *OURNAL OF 0OLITICAL
%CONOMY
4EMIN 0ETER $ID -ONETARY &ORCES #AUSE THE 'REAT $EPRESSION .EW 9ORK 77
.ORTON
4EMIN 0ETER ,ESSONS FROM THE 'REAT $EPRESSION #AMBRIDGE -! -)4 0RESS
5NITED 3TATES $EPARTMENT OF #OMMERCE "UREAU OF %CONOMIC !NALYSIS &IXED 2EPRO
DUCIBLE 4ANGIBLE 7EALTH IN THE 5NITED 3TATES
7EINSTEIN -ICHAEL - Ý3OME -ACROECONOMIC )MPACTS OF THE .ATIONAL )NDUSTRIAL 2E
COVERY !CT Þ )N +ARL "RUNNER ED
"OSTON -ARTINUS .IGHOd
4HE 'REAT $EPRESSION 2EVISITED
@FedFRASER