Tests of the ?I
can lead to sz
mates of a syrc
variable(s) of?
depend on ths
Students oft
Federal Resel-I
omy to whichi
ances are moc
watchers h a v g
p f exp1icitj.1
j o t hesis--&
historically ins
serve officials,r
of openness e n
condition of th
or incompetent
laws of econoni
In contrast. n
tion policy inst,
quately to the,
concerned a b a
moderately in - i
supply jumped
1970- 1974 perii
Estimates of.
stance of monel
if tests indicate1
rated over the!;
ics" have made
be attributable[
in the money s
certain vested i.
a nearly fully e
Tests of the Federal Reserve's
Re'action to the State of the
Economy: 1964-74
Thomas M. Havrilesky, Robert H. Sapp,
and Robert L. Schweitzer'
This article reports estimates of the systematic reaction of the Federal Reserve System, as reflected in a variable that it purports to control ([he Federal funds rate). to
the state of the economy, as reflected by the price level. the level of unemployment.
a measure of our country's international economlc position, and the rates of growth
of key monetary aggregates.
The "Laws of Economics"?
An entire generation of econonletric models is premised on the assumption that
either the money stock (more recently, the monelary base) or "the" interest rate is
given by the monetary authority. Unfortunately, while certain monetary aggregates
qnd/or ir~terestrates may be controlled by the monetary authority, they are likely
neither to remain fixed nor to be determined by processes that are random with
respect to the other variables of the model. such a s the price level and the level of
~ n e r n ~ ~ d y m eThus,
n t . the assumption that a monetary control variable 1s "given"
Reprinted. with additions and deletions. from the Sociul Scic,tice Qirurrer!~~.55. 4 (March. 1975), pp.
835-852. by pernlission of Southwest Social Science Association and the authors.
'The authors would like to thank James Dean. Neil d e Marchi. Richard Froyen, Edward Kane. a n d
Raymond Lomhra for helpful comments.
I
>For enample. cc
seen that one can]
Reserve completel!
changes in noncont
incomplete, defensi
"Federal Reserve '
Board o f Governor
'Karl Brunner. '
ing, 4 (Feh. 1972).
Journal o/ Ecotiorl~ia
Stability,'' Jolrrtrul
Brave N e w World
109-113.
4Adam S m ~ t l i S
.
467
Tests of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
can lead to serious bias in the estimation of rnacroecono~iiicn~odels.~
Reliable estimates of a systematic relationship between the state of the economy and the control
variable(s) of monetary policy should, therefore, be of interest to individuals who
depend on the predictions of these models.
Students of the Federal Reserve System have long noted the alacrity with which
Federal Reserve officials allude. ex post, to a long list of variables in the macroeconomy to which they purport to attend. Yet, there is little evidence that these assurances are more than public relations pronouncements. In fact, for over a decade Fed
watchers have been distressed by the reluctance of the m o n e t a r y a u t h ~to commit !
itself explicity, e x ante, 1 0 specific macroeconomic
. .- .. .... .goals and to a s p e c ~ f i c h t u r a l 1
h s t h e s i s in which these g o a l s u e contalngd.lThi,s lack of open commitment has
historically insulated the structural and policy-strategy conceptions of Fcderal Reserve officials from criticism and improvement (see footnote 6). In fact, this absence I
of openness enabled Federal Reserve officials to argue fallaciously that the unhappy
condition of the econom:. in the early 1970's was explained. not by their dereliction
or inconipetence (which is consequently hard to uncover). but by the failure of "the
laws of economic^."^
In contrast. many critics have contended that the recent shortconlings of stabilization policy instead stem from the failure of the monetary authority to respond adequately to the state of the economy. They contend that if the Fed had truly been
concerned about the rate of inflation, the money supply would have grown only
moderately in the early 1970's. Instead, the annual rate of growth of the money
supply jumped from 3.8 percent in the 1960's to approximately 7 percent in the
1970- 1974 period.
Estimates of a systematic relationship between the state of the economy and the
stance of monetary policy should shed additional light on this problem. For example,
if tests indicate that the monetary authority's reaction to price inflation has deterioI!
rated over the years, then in lieu of substantive evidence that the "laws of economics" have made i t necessary to "live with" more inflation, inflationary distress would
be attributable to the Federal Reserve's buckling under to pressure for rapid growth
in the money supply. The impetus for "easy money" stems fiom the antipathy of
certain vested interests and most political leaders to high and rising interest rates. In
a nearly fully employed economy there will be continuing pressure on the Federal
I
brve's
the
A '
'
1 ;,f9
-,
-
.
ral Reserve Sys11 funds rate), to
unemployment,
::rates of growth
4
1
I
issumption that
' interest rate is
: m y aggregates
,.they are likely
-e randoni with
ind the level of
iable is "given"
'{March. 1975). pp.
Edward Kane. and
1
>For example. consider the "reverse causation" problem discussed in Section I 1 of this book. I t can be
seen that one cannot reasonably deal with "reverse causation" simply by assuming that the Federal
Reserve completely oH'sets. through defensive operations, changes in the control-variable induced by
changes in noncontrolled factors suchas member bank borrw.ing. The true magnitude of these, probably /
incomplete. defensive reactions should be estimated. See Raymond E. Lornbra. and Raynlond G. Torto, ,
"Federal Reserve 'Defensive' Behavioc and the Reverse Causation Argument," Sin/! Ecotrottric Silrc11:
Board of Governors of the Federal Reserve System (Nov. 1972).
Crcclii rrtrll Btrttk'Karl Brunner, "The Ambi~uousRationality of hlonetary Policy," Jo1trt1ulo/.2lotrq1~.
i11g. 4 (Feb. 1972). pp. 3- 12: Jack 11. Guttentag. "The Strategy of Open Market Operations." Q~e~r~.ic,r/i.
Jour~~nlof
Ece~~orrie.,
75 (Feb. 1966). pp. 1-30: Thomas tlavrilesky, "A New Program for More \lonet.~ry
Stabilily." Jorrr~lulo/Poliiicirl Ecorrorr!~..80 (Jan.-Feb. 1972). pp. 171- 175. and ":\ Skeptic;tl View 01' the
Brave New World of Monetary Policy Experiment." Jo~rrrrulo/ Eco~ro~rlicIssret,s. 5 (Dec. 1971). pp.
109- 113.
"dam Smith. Slcper~rlo~r<~.
(Sew York: Random House. 1972). p. 231.
I
1
..
468
MONETARY POLICY
J
.
Reserve to soften the short-run effect on interest rates of chronic government deficits.
Increases in the gro\vth rate of the money supply in the short run keep interest rates
from rising far above the ceilings on rates that can legally be paid small savers
thereby protecting savings and loan associations from competition and deposit outflows, bolster the housing market (which is largely financed by the savings and loan
industry), and artificially stimulate the economy.
Unfortunately ivhile "easy money" is often politically acceptable in the short run
(e.g., during clection years), in the long run i t often has dire economic implications.
In an overheated economy, the moderating effect of rapid money supply growth on
interest rates tends to be only temporary. An increase in the growth rate of the
money supply is quickly transmitted to a higher rate of price inflation and this, in
turn, soon drives interest rates even higher. Periodic attempts to reduce inflation may
result in recession.*
Therefore, econometricians and economists are not the only ones who have become concerned about the actual, ex ante, stabilization commitments of the monetary authority. There was increasing unease in the popular media with the theme,
widely promoted in late 1973 by Arthur Burns, Chairman of the Federal Reserve
, System's Board of Governors, that the inflationary malaise was the bylproduct of
hard luck and administrative ineptitude, as evidenced by currency devaluations,
wheat deals. materials shortages, crop failures, and the machinations of supposedly
sinister sultans of oil. energy oligops, and commodity speculators. As dissatisfaction
with these a,d hoc explanations of inflation continued to mount and as the public
became aware that prices were accelerating long before the food and energy shortages, Congress and the electorate began to realize that a stricter control of money
supply growth would help restrain the inflationary proclivities of government. In
1975 by Congressional Resolution the Federal Reserve was asked to choose and
announce the specific money supply growth range that i t would seek for as long as a
year ahead.
Whether such announced growth ranges can be maintained under all economic
conditions in face of the chronic "easy money" pressures discussed above is debatable.** I t'they cannot. a return to the unsatisfactory monetary policy reactions of the
1964-1974 period is rather likely. Therefore. i t is instructive to estimate policy reactions of !he monetary authority during theLY64- 1974 decade when such monetary
disciplir!e was not maintained. Our procedure will allow us to test several hypotheses.
For example. as suegested by an ex-Governor of the Federal Reserve System, we
may be able to determine if: in the early 1970's. the monetary authority lost much of
the allti-inflationary militance that it is purported to ha1.e had under William
McChesney Martin in the 1960's. This, in turn, may clarify the allegations that the
once hallowed role of the Federal Reserve as an independent monetary trustee has
been lessened, perhap5 by too close association of its Chairman and Governors with
I
--
*The unccrtaintic!, lrom resulting accelt.r;~tionand decelcr;ition of money supply growth. a s discussed
by Wood. K;rsche and Ilumphrey in Section I a n d by Mayer, Kane and Friedman in Section IVC, may
le;ld
a n inetlicient, recession-prone economy.-Ed.
**See rhe opinions of Paul Samuelhon a n d Milron Friedman in Secrion IVC.-Ed.
Tests of the&
the executivet
sustain low in
quences.'
The Analp
Our statistics
to influence r
constraints 01
absence of stn
formance, fo:~
inspection ari
Our purpos
actions of tht I
the regimes o
We shall estn
authority. as 1
completely, co
measure of ou
key monetarj
The reactio
assumed utilit
given hypothei
a household's1
its budget an1
sells.
The coefficri
Federal Resen
tions of struct
ture of ,the ecn
unemploymen
certain monet.
JFor accounts.?
for easy money. SL'
Sanford Rose. "TI
4As discussed .1
monetary policym,
o f the economy. 7
interest rate and aby the Federal K
chosen, and henct
Optimal Reaction:
Economic Journub;
'ARk
POLICY
Tests of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
the executive branch which traditionally is often wont to stimulate the economy and
sustain low interest rates with inadequate regard for ultimate inflationary consequences.'
iovernment deficits.
rkeep interest rates
:paid srnall savers
In and deposit outle savings and loan
The Analytics of the Reaction Function
ile in the short run
tomic implications.
7 supply growth on
~ r o w t hrate of the
,flation and this, in
zduce inflation may
Our statistical work is not concerned with the potential strength of monetary policy
to influence the macroeconomy: nor are we directly concerned with institutiotlill
constraints on that strength (such as too close ties to the executive branch) and the
absence of strong incentives for Federal Reserve policmakers to improve their performance. for example by opening their views of the economic structure to critical
inspection and debate. Both of these issues deserve critical in~pection.~
Our purpose is to estimate the influence of the state of the economy on the policy
actions of the Federal Reserve in times of announced "ease" and "tigli tness" under
the regimes of two different chairmen and two different presidential administrations.
We shall estimate, by ordinary least squares, the policy reactions of the monetary
authority. as measured by a variable that is believed to have been closely. if not
completel>. controlled by the Fed, to the price level. the rate of unemployment. a
nieasuie of our country's international economic position. and the rates of growth of
key monetary aggregates.
The ieacrlon funcrion of the monetary authority can be derived by maximizing its
assumed utility function subject to its endowment of productive resources and a
given h~potheticalstructure of the economy. This procedure is analogous to deriving
a household's demand for a commodity by maximizins its utility function subject to
its budget and a given hypothetical structure of the markets in which i t buys and
sells.
The coefficients of the reaction function do not directly reveal the weight that the
Federal Reserve places on the explanatory variables: rather, they represent combinations of structural and utility parameters. Nevertheless. if w e e that the structure of-hanged
between periods over the time span of our estimation. all changes in the coefficients of the reaction functton. ~t statistically significant,
6 T G e c t changes in the monetary authority'
r e n d n terms of price inflation,
unemployment, the international position of
Ilar, and/or the growth rates
certain monetary aggregates. This is a rather strong assumption, and the reader
ones who have benents of the monelia with the theme,
le Federal Reserve
.the by-product of
ency devaluations,
[ions of supposedly
. As dissatisfaction
an'd as the public
1: and energy shortri control of money
of government. in
Eed to choose and
sek for as long as a
lnder all economic
;ed above is debaticy reactions of the
timate policy reacien such monetary
several hypotheses.
I'eserve System, we
hority lost much of
ad under William
illegations that the
metary trustee has
nd Governors with
fly growth. as discussed
Ln in Section IVC, may
'
SFor accounts of Martin's relative independence of and Burns' acquiescence to White House pressure
for easy money. see Sherman J. Maisel. Managing the Dollar (New York: Norton. 1973). pp. 113-123: a n d
Sanford Rose. "The Agony of the Federal Reserve." Forttrrle (July 1974). pp. 91-93. 180190.
bAs discussed above and pointed out in the Lombra and Torto reading earlier in this Section. the
monetary policymaker has a decidedly eclectic attitude toward competing hypotheses about the structure
of the economy. This lack of commitment makes it difficult for the Federal Reserve to choose between nn
interest rare and a monetary aggregate instrument. As a consequence considerable resources are allocated
by the Federal Reserve to devise an optimal control (reaction) strategy for an essentially arbirr;lrily
chosen. and hence probably suboptimal, control-variable (i~~strument).
See. Thomas Havrilesky "The
Optimal Reaction Function: Confluence of the Instrument Problem and the Target Problem" Solrrltern
~;ollolnirJolrrnc~l(forthcoming).
.>
;
?
\:,.,[,
/
//
I
470
MONETARY POLICY
Tests of the Ft
should keep in mind that changes in structural parameters could also account for
observed changes in the estimated reaction coefficients.
contended inn
actually does':
short policy cc
rate itself is id
monetary autk
The Emplrics of the Reaction Function
There have been several attempts to fit reaction functions to the data.' While it is not
our purpose to review critically this- literature, all articles of which we are aware
suffer from one of two shortcomings. First, previous researchers have used, as dependent variables, measures, such as the monetary aggregates. that the Federal Reserve
may not have attempted to control or may not have been able to control over
reasonably short policy control periods.* Second, previous researchers have specified
policy control periods that are quite long. This leads to a problem in ascribing oneway causality to the coefficients of the reaction function. For example, if the policy
period were one quarter, i t would be difficult to separate one flow of causation (from
. ... . _ .
the state of the economy to the control-variable) from the other, reverse. flow of
; .,/.',//
causation (from the control-variable to the state of the economy). In our research we
assume a control or policy period ofjust one month and use monthly averages of the
C
/
*+h Federal funds rate as the policy control-variable or instrument.
1 wznb
As-one
several
money
market variables. .(the
Federal funds rate, borrowed re- - -of
.. .- -. ....
...
-.- .
.. -..... . .. -.
serves.
.
free reserves, and t h e - ~ r e a s u bill
r ~ rate), the Federal funds rate is widely
regarded as-the,single vaGable among all possiblecandldates
. .- .
.-- .
.
..... .
that the Federal Reserve is.-,m_qstlikely actually to.have used a s a corlfrol-variable. (We are not suggest.,
ing. however, that the Federal funds rate actually is an optimal control-variable for
..>
monthly policy control periods; to the contrary. cogent arguments can be made that
a monetary aggregate is preferable in that it provides better inform::tion about what
the effective stance of monetary policy really is. Moreover, in the long run an interest
rate is not controllable as it is a market-determined, relative price.ls In addition, with
practically every turning point in monetary policy (as determined from reading the
minutes of the Federal Open Market Committee-FOMC-meetin~s in the 1960's)
the Federal funds rate changed more pronouncedly in the expected direction in the
weeks following the meeting than any other money market ~ a r i a b l eMoreover.
.~
it is
I
-
_
--
t
~
u$.,,,,?
-
-
,
'For exzmple. see James W. Christian. "A Further Analysis of the Objectives of American Monetary
Policy." Jountcrl of FInut~ce,23 (June 1968), pp. 465-477: Ann F. Friedlaender, "Macro Policy Goals and
Revealed Preference." Quarter!, Journal of Econotnics. 87 (Feh. 1973). pp. 24-43: Richard T. Froyen. "A
2 (July 1974): Thomas Havrilesky.
Test of the Endogeneity of .Monetary Policy." Jorrrt~ulof Ecotrot~~etrics.
"A Test ol Monetary Policy Action." Journul of Politico1 Econot)!l~.75 (June 1967). pp. 299-304: John A.
Wood. "A Model of Federal Reserve Beh;~vior." in George Horwich, ed.. Monetury Proccrs atrd Polic:~.:A
S ~ ~ n ~ p o s i l(Homewood.
nv
Ill.: R. D. Irwin. 1967). pp. 135- 166.
'If thc monelary instrument is not controlled but rexponds lo the economy in an unconstrained fashion.
we will 'nave a "reverse causation" prohlem in econo;netric models that attemp! to estimate the erect of
that instrument on the economy. See footnote 2.-Ed.
8Much of the previous literature on the reaction function~confirsesthe issue of what the optit~itrlcontrol
variable would he for measuring the real thrust and appropriateness of monetary policy with the issue of
~.
estimating how the Federal Reserve u c t u ~ l ! reacted.
9The changes from ease to tightness or vice versa occurred in May 1960. June 1962. November 1966.
December 196;l. July 1968. December 1968. January 1970. August 1971. Septemher 1972 and March 1974.
Stephen H. Barnett. "The Monetary Indicators at Turning Points." Fintrttcitrl Anu!~,.o.>Jour~tul.26 (Sept.Oct. 1970). pp. 29-32.
What Are 'F
and lndica
After the cred
of the Board.)
some Fed spA
include the be
and the mon?
have been att:
most of the li!
garded becaue
refers to a var:
may be used i
cal superiorit)
donp_t_.seernat
-B64 1974 pel
short-run
inst:
-.-... . .
. - ..
. Nor have tk
mediate large.
Hypotheticall.
the optimal rt
tion of that ir
kinds of hypo
so superior to
to that sector
stances, the o
"track" on) th
the theoretics
-
See the artii
"The Federal
stem from its lac1
6. Further discu!
commitment is fr
' T h e instrum
Instrument Varin
723-729. Also. M
TARY POLICY
47 1
rests of the Federal Reserve's Reaction to the State of the Economy: 7964-1974
;Id also account for
f
contended in numerous Federal Reserve publications that the monetary authority
actually doer try to erect close control over the Federal funds rate during relatively
short policy control periods, such as a month or a week; therefore. even though the
rate itself is influenced from the demand side, we assume that i t is controlled by the
monetary authority.*
khta.' While it is not
lhich we are aware
lnve used, as denenhe Federal Reserve
de to control over
dlers have specifiec'
n in ascribing onermple, if the policy
(of causation (from
IT. reverse. flow of
.In our research we
hly averages of the
-"
&;
1
What Are ~ o n e t a r y~ a r ~ e tInstruments,
s,
and Indicators?
irate, borrowed re~ n d srate is widely
a t the Federal Reire are not suggestantrol-variable for
bcan be made that
n:: tion about what
5ng run an interest
I* In addition, with
Llfrom reading the
ings in the 1960's)
?d direction in the
le.9 Moreover, i t is
h
lf American Monetary
I
acro Policy Goals and
hchard T. Froyen. "A
3; Thomas Havrilesky.
. p p 299-301: John A.
I Process atrd Po/iqr: A
After the credit crunch of 1966, and especially after Arthur Burns became Chairman
of the Board of Governors. attention in the financial press and in the writings of
some Fed spokesmen seems to have focused on monetary aggregates (defined to
include the bank credit proxy and various measures of reserves. the monetary base
and the money stock). Labels such a s guides, instruments, indicators, and targets
have bzen attached to these monetary aggregates. For purposes of this paper and in
most of the literature on the theory of monetary policy, the word "guide" is disregarded because i t has little consistent. analytic meaning. The label "instrument"
refers to a variable controlled by the policymaker over the policy control period and ,
may be used interchangeably with the label, control-variable. Despite their theoreti-/
cal superiority, for control period5zch a s a m ~ t or
h aquarter.m.onetary..aggregates
d ---o not
.--.seem
..
actually to have been used as instruments
. by
.
.the Federal Reserve for the
.
. ..
'to. . as
. strategic
.
3 6 4 - 1 9 7 4 period covered by this study. They a r e n e v e r referred
short-run
instruments.in
Federal
Reserve
publications.*
-......., . ..
Nor have the-monetary-aggregates;-in our opinion, likely been used as true intermediate target-variables for shorter control periods such as a month or a quarter.
Hypothetically. an optimal stabilization strategy under uncertainty would formulate
the optimal reaction of monetary policy (the optimal instrument and optimal reaction of that instrument) to information about the state of the economy; in certain
kinds of hypothetical structures information from one sector of the economy may be
so superior to information from all other sectors that certain variables, endogenous
'to that sector, are controllable by monetary policy action. Under these circumstances, the optimal stabilization strategy may be to compensate for shocks to (to
"track" on) these particular variables. In such cases, these variables are referred to in
the theoretical literature as intermediate target-variables.'O
!
I
t,
lnconstralned fashlon.
estimate the effect of
pat the oprit~rnlcontrol
dicy wlth the issue of
962. November 1966,
,972 and March 1974.
rrs Jo~trtrcrl.26 (Sept.-
!
t
See the article by Lonlbra and Torto in this Section.-Ed.
**The Federal Reserve's failure to choose a monetary aggregate instrument in its short-strategy may
stem from its lack of commitment to specific hypothesis about the structure of the economy. See footnote
6. Further discussion of the instrument and intermediate target problems and evidence of this lack of
commitment is found in the articles by Davis and Lombra and Torto in this Section-Ed.
) T h e instrunlent problem is well uridersrood. See John Kareken and Neil Wallace. "The Monetary
Instrument Variable Choice: How Important?" Journal of Money, Credit and Banking. 4 (Aug. 1972), pp.
723-729. Also. William Poole. "Optimal Choice of Monetary Policy Instruments in a Simple Stochastic
.
JJ'~'
I
J
/
'
.
Tests of the FA
MONETARY POLICY
i
.
i
I
,
5
[,'
\,,
r
i
,
,
.
:,
I
Despite their semantic confusion. articles by Federal Reserve econon~istshave
placed the role of monetary aggregates in reasonable perspective; in several widely
read articles it can be discovered that the monetary aggregates are used neither as
instruments, as intermediate target-variables (in the rigorous sense described above),
nor as indicators (in the sense made known by Brunner)."
Federal Reserve economists have very carefully explained that projected growth in
certain monetary aggregates (such as Reserves to Support Private Deposits, RPD's)
is based on quarterly projections of the growth of aggregate income generated by the
FRB-MIT-Penn econometric model. The quarterly income projections are then
placed on a monthly basis. Then, simulations of the growth rates of various monetary
aggregates are made for alternative Federal funds rates. For a given Federal funds
rate there will be a growth rate for each monetary aggregate.'*
Thus, it is clear that in making the short-run projections the Federal Reserve
assumes that the Federal funds rate will remain at a level predetermined by its
stabilization policy decisions. Subsequent gaps between projected growth and actual
growth of a monetary aggregate then simply tend to reflect 'how far the monetary
authority's initial income projection is o f . When the growth of a monetary aggregate
is far out of line for several months. Federal Reserve economists sometimes indicate
that corrective adjustments in the monetary policy instrument mr?).be applied. that
is, that the aggregates may be serving as intermediate target-variables in formal
strategic models.,Ho~e~~,.~n~~alit~III~~ide~swings.in.the~gr~owth.rates
of the monetarget-vari.
that
.
.
their
role
as
intermediate
-..
.
.
... . ~. aEpear to be so tolerated
tary aggregates
~..
ables m u s t b e d o u k d . Whether or not the growth of certain monetary aggregates
historically constrained monetary policy actions over the 1964-1974 period is an
e m G c a l question. Nevertheless i t should be clear that. despite recent improvement
in policy strategy, monetary aggregates are not yet always perceived and used by the
Macro Modci." Q~tarter!,Jour~ic~l
of E c o ~ ~ o ~ ~84
i i c('May
s.
1970). pp. 197-216. The label "instrument" is
sometimes confused with the label, "intermediate target". e.g., Harold Shapiro. and Robert Holbrook.
"The Choice of Optimal Intermediate Targets." Aniericcr~iEcono~~rrc
Review. 60 (Supplement. May 1970).
pp. 40-46.For further clarification of the target problem. see Karl Brunner. Targels and Indica~orsof
Monelurj* P ~ ~ I (San
i c ~ ~Francisco: Chandler. 1969): Thomas Havfilesky. "Finding the Optimal Monetary
S trategy with Information Constraints," Jo~rrnalof Fitio~rc-e.27 (Dec. 1972). pp. 1045- 1056; and Thomas
Havrilesky, "The Optimal Reaction Function; Confluence of the Instrument Problem and the Target
Problem" Sortlhcrn Econotnic Jour~lal(forthcoming).
"A monetary indicator has a special theoretical meaning: when the structural hypothesis ofthe policymaker is unspecified. an indicator may be used. nevertheless. to reflect the thrust of policy ac~ionsunambiguously and independently of the state of the economy. See Brunner. Targos a ~ i dI~tdica~ors.
I2Stephen H. Axilrod. "Monetary Aggregates and Money Market Conditions in Open Market Policy."
and Richard G . Davis. "Short-Run Targets for Open Market Operations," Open M u r k ~ vPoliq. rrnd Operaling Procedrrres-Staff Studies. Board of Governors of the Federal Reserve System. 1971. Stephen H.
Axilrod and Darwin L. Beck. "The Role of Projections and Data Evaluations with Monetary Aggregates
as Policy Targets," and James L. Pierce and Thomas D. Thomson. "Some Issues in Controlling the Stock
.
11, Federal Reserve Bank of Boston. 1972.
of Monetary Aggregates," Cunrrolling M o n e r a ~ Aggrcgares
I
monetary autlc
con trol-theorei
N
a
Fitting the4
As mentioned;
causality to th,
action). The r;
FOMC. At the
meeting, are d:
ade forecasts d
heeded by mci:
meeting of t h ~
level of the Feil
dent on the sta
the previous ni
changes in polr
lag between a i
because of lag:
quarterly, data
"tight" to "ea$
Board of Cove
The explana:
ables with whit
sale price inde?
of the internati
(FX), and, for.:
of three moner
money supply
observations a):
month.
Explanatory
(by the Fed) n
constant term 1
ance and, therl
addition, the u
estimates d o nc
We d o not ii
operations by 1
monetary aggr
etc. (not to int
'For fairly com
Torto earlier in th
the Federal Reser
variables.-Ed.
9RY POLICY
Tests of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
economists have
in several widely
-e used neither as
described above),
monetary authori~yas instruments, or intermediate targets in the proper short-run
control-theoretic sense outlined above.*
er Polic) and Oper-
Fitting the Reaction Function
As mentioned earlier, we assume that monthly data permit us to attribute unilateral
causality to the reaction function (from the state of the economy to Federal Reserve
action): The monthly specification also conforms to the monthly meetings of the
FOMC. At these meetings data on the state of the economy. compiled since the last
meeting, are discussed and policy decisions are rendered. Throughout the past decade forecasts of future economic activity are assumed not to have been systen~atically
heeded by monetary policymakers. Policy actions for the four weeks fi)llowing a
meeting of the FOMC are carried out by its Account Manager. Thus. the average
level of the Federal funds rate in any calendar month can be viewed as being dependent on the state of the economy, as vietved and interpreted by the pol~cymakers.in
the previous month. If the Chairman initiates or the Account Managcr anticipates
changes in policy before the actual meeting of the FOMC, the assumed one-month
lag between a change in the state of the economy and policy action is still reasonable
because of lags in reporting data. A final ad\.antage of using monthly. rather than
quarterly, data is that it allows researchers to pinpoint and estimate shifts from
"tisht" to "easy" policy and shifts associated \vith changes in the membership of the
Board of Governors and in presidential administrations.
The explanatory variables in the reaction function (the state-of-the economy variables with which policymakers are assumed to have been concerned) are: the wholesale price index (P), the unemployment rate (U), an historically sensitive barometer
of the international position of the dollar, the exchange rate of the Deutsche Mark
(FX), and, for all periods after November 1966. the annualized rate of growth of one
of three monetary aggregates: the adjusted bank credit proxy (BCP), the narrow
money supply (M), and reserves available to support private deposits (RPD). All
observations are monthly averages and enter the reaction function with a lag of one
month.
Explanatory variables were not entered as differences between actual and desirid
(by the Fed) magnitudes because, in addition to its being unknown, subtracting a
constant term from an observation vector v:hich enters linearly, would add no variance and, therefore, would not affect the absolute magnitude of o u r estimates. In
addition, the use of monthly data lagged only one period obviously means that our
estimates d o not reflect the entire (distributed lag) reaction of the Federal Reserve.
We d o not introduce a measure of the need for "even-keeling" or other defensive
operations by the monetary authority because even-keeling refers to adjustments in
monetary aggregates such as reserves, the monetary base. Federal Reserve Credit.
etc. (not to interest rate adjustments) by the central bank in the face of Treasury
1971. Stephen H.
metary Aggregates
ntrolling the Stock
< of Boston. 1972.
'For fairly complete discussion of the monetary polic)maker'i strategy see the reading by Lombra and
Torto earlier in this Section. There it is made clear that aggregates are n& u x d as instruments and that
[he Federal Reserve is [a,concerned about interest rate volatility to use [hem as intermediate urge[variables.--Ed.
ojected growth in
Deposits. RPD's)
generated by the
ections are then
;arious monetary
-n Federal funds
Federal Reserve
?termined by its
.o\vtli and actual
a r the monetary
netary azgregate
netimes indicate
be applied. that
iables in formal
ies of the monehate
target-vari..-.- . ..
....
-tary aggregates
74 period is an
nt improvement
and used by the
~
be1 "instrument" is
! Robert Holbrook,
lement. May 1970).
rr and Indicarors of
Opt~malMonetary
1056; and Thomas
-m and the Target
rhesis of the policydicy actions unamrdicarors.
en Market Policy.''
.I
473
I
/+,..! : 4 y ,
1
'
/
Tests of the'
MONETARY POLICY
funding and refunding. We assume that the Federal funds rateis controlled and,
r
therefore. insulated from the i n f l u e n r e ~ r r f ~ r ~ a c t i v i & + a n & t h edisturbances.
-
----
How the Federal Reserve Behaved
The Federal Reserve under Martin:
Tightness before the Crunch, January
1964 to November 1966
Estimation begins with the first full calendar year of Lyndon B. Johnson's presidency. At this time William McChesney Martin was Chairman of the Board of
Governors. In July 1962. according to the minutes of the FOMC, the Federal Reserve shifted from a policy of ease to a policy of tight money. Tight money policy
seems to have persisted until shortly after the credit crunc
In all equations that follow the figures in parenthesis a
the coefficeint of determination, F is the F-statistic, D-W
statistic, N is the number of observations, and the t subscript ;&fen to the time
period.
(,
;)
f
",.
.',
.,
7
.-.'
.. .
C J
'
1
,.
oG
\Ci b.lp
,
':
.
,:
?
,
.
t,o
strengthen
Federal Funds = - 80.988 - 0.0365 U , - I
Rate,
( - 10.342) ( - 0.2603)
R2 =.98 Fk = 427.527 D-W
+
(1)
The results suggest that the Federal Reserve under Chairman Martin showed little
concern for unen~ploymenteven though the,unemployment rate did not fall below 4
\ .lf",
percent until January 1966 and did not go below 3.7 percent until September 1966.
The coefficient on the unemployment variable is statistically insignificant. The coefficient on the price level and exchange rate variables are both positive and significant
at the .0] level. A one-point rise in the wholesale price index (P) appears to have
induced a .56 percentage point rise in the Federal funds rate and a one-cent rise in
&
the U.S. dollar price of the Deuhche Mark seems to have induced a I percentage
Hp\
Q @/ point rise in the Federal funds rate. The R2 and F-statistics attribute high explana8'
tory pou.er to the equation and the D-W test statistic indicates an absence of autoC
correlation in the residuals.
These results lend support to the widely held belief that the Federal Reserve under
Martin was strongly anti-inflationary13and that the tight money period of the early
1960': was, in fact, partly brought on by the international weakness of the dollar.
The Martin Era: Ease affer the Crunch,
December 1966 to November 1967 and
July 1968 to December 1968
The credit crunch of 1966 is alleged to have been instructive to monetary policymakers of the difficulties of singlehandedly attempting to cool the economy and
1JMaisel, Matlc~gillg/he Dollr~r,pp. 60-69.
rates thatwt
policy of m a
proxy was fis
By Decem
ing. that th:
diverted by :I
quite sizable'
unexpected 3
shifted b a c k
a tactical rns
arrival of th
sumed.
Because tk
the Decembr
ease. Decem,'
Federal
Funds
,+
0.5577 P , 1 256 F X , - ,
( 1 1.348)
5 5 9 3 ) _r
= 1.759 N = 34
t
-_9_4h_t
=
Rate,
The shift t+
The coefficie
icant at the .C
October 196(
unemployme
centage poinl
rate coefficie
p ~ i c elevel va
crunch perio
-Federal fund
lishing a pati
a steady ani
maintain i n *
William Mc(
"When the c
gether the resul
"Because tht
equations fall sr
the Cochrane-0
the R2 rises to .
To test our
percent level th,
period.
hTARY P O L I C Y
le_i_s~.controlledand,
b t h e r disturbances.
B. Johnson's presiin of the Board of
IC, the Federal ReTight money policy
946-
-k2
refers to
gurbin-Watson test
: refers to the time
475
Tests of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
strengthen the dollar. Calls
--- for fiscal restraint resounded about the naticln_a!!d,-~---.-=n-ed
bj: the c~cl_l_t
crunch-and t a l ~ ~ ~ : a h o u s i n $ ~ c.r ibmught--&)ut-bys..i.s. .- .-.... . .
interest
rates that we&;_t&q_h.igh." the Federal Reserve \vas purported to have shifted to n
policy of monetary ease in December 1966. (In addition. in June 1966 the bank credit
proxy was first mentioned in the FOMC directive.)
By December 1967 i t seemed apparent that fiscal restraint would not be forthcoming. that the Johnson Administration would not raise taxes to pay for resources
diverted by the warfare-welfare state, and that the Vietnam War deficit wouM be
quite sizable. The shift to tightness in December 1967 was short-lived because of the
unexpected passage of the tax surcharge bill in mid-1968. The monetary authority
shifted back to a policy ofease in June 1968. It turned out that the return to ease was i
a tactical mistake because the effect of the surchage \vas overestimated. With the ;
arrival of the Nixon administration in January 1969 a policy of tigl~tnesswas resumed.
Because the entire, post-crunch, Martin-Johnson period was one of ease, except for
the December 1967 to June 1968 interlude, we consolidated the two subperiods of
ease. December 1966 to November 1967 and July 1968 to December 1968.14
I
.lh
,1256 FX, (3.593) ,
,
,
(1)
1:
Funds
Rate,
=
-20.247 - 1 . 7 1 4. ,~- ,
( - .884) ( - 2.185)slb
R2
fartin showed little
lid not fall below 4
il September 1966.
nificant. The coefttve and significant
') appears to have
a one-cent rise in
I a 1 % percentage
lute high explanaI absence of auto-
-
ral Reserve under
eriod of the early
less of the dollar.
tary policymakeconomy and
=
.70 F
=
+
us<
. 3 5 8 9 ~ , _, . 1 7 6 5 F X , _ ,+ .3968BCP,.,
(3.153)
( - ,2412)
(.?009)
7.569 D-W
=
.52 N
=
18
(2)
The shift to ease is_quitemark.ed when-mmpared.to_the. pre-credit crunch r r ~ u l t s . ' ,~
The coefficient on the rate of unemployment is now negative and statistically significant at the .01 level. The unemployment rate rose rather steadily from 3.5 percent in
October 1966 to 4.3 percent in October 1967. For every one percent increase in the '
unemployment rate, the Federal funds rate would fall by approximately 1.71 percentage points. The shift to ease is further corroborated by the fact that the exchange :
rate
is no longer [email protected] i g n i f i ~ ~ n t - 5thatthecoefficient
d
on the ,
pf~~clevel variable, while still positive a n d ~ s i g n ~ f i c a n t , ~ ~ d ~ c I i n e ~ ~tfh~e m
i r e~-~ ~ 6 _ f o r
c~u-nchperiod to approximately .36; this indicates a .36 percentage point rise in the 1
Federal funds rate for every one point increase in the whole?Leprice_ index. Establishing a pattern that was to be repeated during later periods of monetary tightness,
a steady -anti-inflationary monetary policy w i s apphrently politically digcult to
inaintain in the face of rising unemployment and rising interest rates (even under
Willianl h4cChesney Martin). Despite the explicit mention given to the bank credit
'
I
' W h e n the entire. post-crunch. pre-Nixon period. December 1966 to December 1968. is lumped together [he result is similar to that reported here except that the coefficient on U , - ) is so~newhatsmaller.
"Because the R1 and F-statistics are lower than the previous result. the explanatory power of the
equations fall somewhat. The D-W statistic indicates autocorrelation in the residuals.'When corrected by
the Cochrine-Orcutl iterative technique. the signs and significance of the coetRcients are unchanged and
the R2 rises to .97 with a D-W of 2.42.
To test our pooling procedure, we used Chow Tests. The results of the tests led us to reject at the .95
percent level the null hypothesis that the small subperiods obey the same relation over the whole sample
period.
-
476
MONETARY POLICY
proxy in FOMC directives during [his period,l"lie statistical results indicate that i t
did not elicit a systematic response from the n1onet;iry authority.
The Martin Era: Nixon's Game Plan I,
January 1969 to January 1970 (and
December 1967 to June 1968)
Richard Nixon was elected partly on the promise that he would cool an overheated
economy. The Federal ~ e s e r v ehad begun the cooling process during the first half of
1968 only to move prematurely and mistakenly to a policy of ease because of the
widespread overestimation of the depressive effect of the tax surcharge. Because this
period of tightness in 1968 was so short (seven months). we consolidate i t with the
tight money period that began when Nixon assumed the presidency."
Federal
Funds
Rate,
=
+ .9767 U, _
10.570) (1.524)
- 73.671
(-
R2
=
,93 F
,+
(1
.8622 P, _
1.946)
46.329 D-W
=
=
,
- 4.987 FX, _
I .990)
(-
,782 S
=
, + .7979 BCP, _ ,
(.6500)
20
(3)
That the period is one of tightness is suggested by the positive and significant (at
the .99 level) sign on the coefficient of the price level variable. The coefficient indicates a .86 per~eptage~po~~n_t_~i~e_~in-the
Federal
. .. . . .~. .funds
..
rale for every single
point rise
.in_thc.~u~holesale
price.index; this is the large? value that we estimate for this coefficient in any subperiod .of the entire 1964-1974
..
period.I8 I n terms of the coefficient
CI
(.;:\
alone, one can infer that monetary policy was at its tightest during the first year of
'
. the Nixon administration.
The coefficients on the unemployment and exchange rate variables are both statistically signiEcant at the .90 level, but possess the "_wrongwsigns. This suggests, somewhat anomalously, that a rise in the Federal funds rate was associated with a rise in
i
I
. I,,,,'~
unemployment
c
and a decline in the Federal funds rate was associated with a rise in
'
. -the price of foreign exchange during the year. Both the rate of unemployment and
i ,.'I; . . the excha~gerate rose over the period (see note 23).
P ;I . : 2 i
These phenomena can be explained by a steady rise in the desired level of uneni-. .
'
1'
'
.: .
,
;
\
" .-
+
..-
'6Maisel. Monaging rlre Dollar. pp. 85. 230.
I7When the January 1969 to January 1970 period is fit separately the results are:
FF,
=
- 1 1 1.067
( - 4.930)
+
,3519 U, (.4041)
1
+
1.381 PI
(4.514)
- 1
- 1.103 FX, .1
(-
2.982)
+
1.6897 BCP, (1.1981)
R2 = .85 F = 11.175 D-W = 1.4880 N = 13.
Compared t o the combined lime periods. autocorrelation in [he residuals does not appear to be a problem.
Moreover, all signs are unchanged and only the coeficierlt of the unemployment variable loses significance.
I8The fact that interest rates should rise wirh anricipated price inflation is of little concern here as the
Federal funds raw is a conlrol variable and thus IS insula~edfrom market iorces.
Tests of the
F&!
ployment from.
against the Den
that was belou
raised since a
assumed desirec
i t did over the-]
desired rate oft;:
could have indu
rate was continu
i t was below the;
lowered to take.h
rose. the Federa
rate rose even fa
These interprc
press during 196:
the economy ani
rate goals.
Finally. as in ti
statistically signit
The Federal ti
Nixon's Game
February 197l
Onceagain, in 15
ploymen t became:
ma1 advisors were
direction of tight;
unemployment ra.
1970 to 5.6 percer
soon become the I
Chairman of the'l
the wide gap beta
of inflation. Unfo
meant that the NI:
persist with a tighl
Consequently, "gr
havior.
True to form,
spending to pump
would surely have
with an easier mo
I9This led us to intrc
proxy. The results wer,
'OLeonard Silk, Nix,
rARY POLICY
r~ltsindicate that i t
roo1 an overheated
i n g the first half of
sse because of the
:large. Because this
iolidate i t with the
7cy."
.
+
,7979 BCP, _
(.6500)
,and significant (at
he coefficient indiry single point rise
,a-t .f6-r.-t7ii s..c-6G
fii
i -of thecoefficient
~g the first year of
les are both statis]is suggests, somehted with a rise in
~ t e dwith a rise in
nemployment and
-ed level of unem-
-I
8 9 7 BCP, 981)
,
lpear lo be a problem.
variable loses signife concern here as the
Tests of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
ployment from, say 3.5 to 5 percent and a steady rise in the desired exchange rate
against the Deutsche Mark during the period. In this way an unemployment rate
that was below the desired rate would have caused the Federal funds rate to be
mired since a higher Federal funds rate would take the economy closer to the Fed's
assumed desired rate of unemployment. As the actual rate of unemploynlent rose. as
it did over the period. the Federal funds rate would have been raised only if the
desired rate of unemployment rose even faster. In this way, higher unemployment
could have induced a higher Federal funds rate. Likewise. if the desired exchange
rate was continuously raised. a rise in the actual price of foreign exchange. as long as
it was below the desired rate. would have caused the Federal funds rate to have been
lowered to take the monetary authority closer to its goal. As the actual exchange rate
rose, the Federal funds rate would have been lowered only if the desired exchange
rate rose even faster.
These interpretations are supported by widespread discussions. in the financial
press during 1968 and especially 1969. of purposely imposing a coolinz-of period on
the economy and adopting a "more realistic" set of unemployment and exchange
rate goals.
Finally. as in the previous results. the coefficient on the bank credit proxy was not
statistically significant.I9
-
The Federal Reserve under Burns:
Nixon's Game Plan 11, Pre-freeze Ease,
February 1970 to July 1971
'
Once-again, in I 9 7 0 ! . h _ e . e . K e _ c ~ n e I a ~ f r ~ ~ . m s tand
~ aunemtes.
pLoyment.besame-unpala~ahl~-t~fh-e~ad.m~nj.ktra~tin. Even in 1969 formal and informal advisors were already suggesting that economic policy had shifted too far in the
direction of.tightness.'O By 1970, while the rate of price inflation had slowed, the
unemployment rate went rather steadily upward (rising from 4.2 percent in February
1970 to 5.6 percent in June 1971). "Gradualism" in the war against inflation would
soon become the order of the day. As early as February 1970 i t was favored by the
Chairman of the Board of Governors, Arthur Burns. I t was argued at the time that
the wide gap between actual and potential output would eventually reduce the rate
of inflation. Unfortunately. price and wage setters were learning that "gradualism"
meant that the Nixon administration felt it could not allow the monetary authority to .
persist with a tight money policy long enouzh to dampen their inflationary practices.
Consequently, "gradualism" was predicated on an invalid model of wage-price behavior.
True to form, early in 1971 President Nixon ordered increases in government
spending to pump up the economy. Increases in government spending and borrowing
would surely have meant hizher interest rates u less the Federal Reserve con~plied
with an easier monetary policy and comply it d ~ dThe
.
old game plan was scuttled;
L --,r-:
rL1
2
IqThisled us to introduce the growth of the narrowly defined money supply in place of the bank credit
proxy. The results were quite similar to those reported above.
loLeonard Silk, N i x o ~ ~ o ~ ) t(New
i c s York: Praeger, 1973). pp. 3- 19.
477
n I,
-
478
MONETARY POLICY
and sentiment for anti-inflationary relentlessness in'the conduct of monetary policy
waned.*'
Monetary ease was soon reflected in the growth rate of the money supply. During
the 1960's money grew at the annual rate of 3.8 percent-in the 1970- 1974 period it
grew a t a 7 percent rate. This quantum leap soon caused adherents to be attracted to
the argument that the rate of growth of the money supply should be a constraint on
monetary policy actions.
T h e results for the February 1970 to July 1971 period are:
J
Federal
Funds
Rate,
=
33.676
- 1.2596 U, _
(2.0248) ( - 2.5475)
R2
=
.93 F
=
,(-
.61965 P, 2.5585)
41.424 D-W
,+
=
1.7359 FX, (2.5373)
1.266 N
=
, + .I7305 BCP, ,
-
(. 14339)
18
(4)
Introducing a money supply constraint in place of the bank credit constraint yields:
Federal
Fun&
Kate,
=
,
32.821 - 1.3635 U, - - ,5806 P,
(2.2040) ( - 3.2445)
( - 2.724)
R2 = .94 F =47.791
'.,
'
.,
I'
\
.
,
.
.
D-W
_, +
=
5‘'
1.6289 FX, _
(2.5773)
1.352 N
=
,+
18
-79498 M, _
,
.( 1.3697)
(5)
The Federal funds rate fell steadil? during this period. The shift to ease is marked
notably by the negative and significant sign on the unemployinent rate. The rate of
unenlplogment rose from 4.2 percent in February 1970 to 5.6 percent in June 1971.
T h e second equation indicates that a 1.36 percentage point decrease in the Federal
funds rate occurred for each one percent rise in the uneinployment rate.
T h e continual weakening position of the dollar abroad seenls to have been of
monlent to the monetary authorit).. This entire period. especially the spring of 1971,
featured growing pressure on the dollar and a rush of dollars into Deutsche Marks
and other currencies (while government officials continued to insist on the strength of
the dol!ar). T h e coefficient on the exchange rate variable is positive and statistically
significant. In the second equation a one-cent rise in the price of the Deutsche Mark
induced a 1.63 percentage point rise i n the Federal funds rate."
The bank credit proxy in the first equation is once again insignificant. but when
the growth rate of the narrow moneq supply is introduced in its place in the second
equation, the coefficient becomes statistically significant. A one percent increase in
2'Maisel. ,Monuging !he Dolkrr, pp. 265-268. Silk claims the shift of G a m e Plan II occurred a(ier the
1970 elections. in which the Republican party fared rather badly. Silk. Niuotro~~iics.
p. 68. I t is ofien
suggested that Richard Nixon was moved to stimulate the economy by memories o f the 1959-1960
recession which thwarted his earlier bid for the presidency.
T h e statistical result seems to contradict former Governor Maisel's contention that international
considerntions affected F O M C decisions onl) a few times during his tenure. Maisel, h f ~ t i ~ g i nfhr
g Dollrr,
p . 224.
Tests of the F&
the annual r a t a
Federal fundsrn
money supply TI
Federal Reserv
avowing a stric
Finally, once:
price level varih
was rising at a m
the Federal Res
tion was desirbl
short-run Phillip
level. Our conjec
desired inflation1
our coniecturei
from the monetr
man, while t h e ?
in the (reaction)
end statistically!
Finally the R2
W statistic fallsia
The Burns EL
Policy, Augu:
By August 197 1
boiling again. T h
a break with tht
reaction function
Federal
Funds
Rare,
=
Federal
Funds
Rate,
- 34
(- 1
=
-
:
(-
''Silk. N i x o ~ ~ o n i
='There are. of co,
the explanatory varis
lions our results (esr
fashion.
479
iCTARY P O L I C Y
Tests of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
: of monetary policy
the annual rate of growth of the narrow money supply induced the Fed to raise the
Federal funds rate by .79 percentage points. presumably to keep the growth of the
money supply within preconceived limits. This result obtains despite statements from
Federal Reserve officials during the anxious months of the 1970 liquidity crisis disavowing a strict money supply constraint."
6' ' "
*')
r,'
Finally, once again we seem to have an a.nomalous r e s u l ~ t h ecoefficient on the
price level variable is negative and statistically significant. The wholesale price index
was rising at an annual rate of about 4 percent during this period. This su!gests that c k
the ~ e d e r a lReserve may have implicitly believed that a positive rate of price infla,
. :
tion was desirable in order to reduce unemployment and move leftwdrd along a
.
short-run Phillips curve by validating, rather than resisting. increases in the price
level. Our conjecture is that there occurred over the period a ~ r a d u a ls h ~ f from
t
zero
I
desired inflation to a policy of encouraging an increase in the price level. W w h e - r
b4,Ti f
0.~?rco~&_re_jsf1.ue.o~.~.of..2*
!he.~vera!! statistical _resu!ts.suggest;l.jrastic c h a w
. . .
71from the ______
monetary
of earlier
periods.- For example, when Martin was Chairt
_ . poiicy
.. _
.. __
I r l "
I
man, while the shift to a policy of ease in 1967 and 1968 was marked by a reduction
in the (reaction) coefficient of the price level variable, the coefficient was still positive
and statistically significant.
Finally the R2 and F-statistics indicate a reasonably good statistical fit and the D;
W statistic falls in the inconclusive range.
m e y supply. During
:1970- 1974 period i t
nts to be attracted to
d be a constraint on
,
+ ,17305 BCP, _
',
,
(. 14339)
a
r;
Lit constraint yields:
,
rt to ease is marked
n t rate. The rate of
rcent in June 1971.
m e in the Federal
nt rate.
s to have been of
the spring of 197 1,
o Deutsche Marks
I on the strength of
ve and statistically
he Deutsche Mark
nificant, but when
lace in the second
lercent increase in
I1 occurred after the
iirs, p. 68. 11 is often
.ies'of the 1959- 1960
I
ion that international
Mirclirging 111eDollor,
The Burns Era: Nixon's New Economic
Policy, August 1971 to September 1972.
By August 1971 government deficits and easy money had the domestic econbmy
boiling again. The move to a policy of wage and price controls dramatically signaled
a break with the previous policy of gradualism (Game Plan 11). Estimates of the
reaction function, using three different monetary aggregates as constraints are:
Federal
Funds
Rate,
=
- 34.759
- ,0360 U , (-1.5216)(-.0373)
R 2 = .81
Federal
Funds
Rate,
=
- 37.459
I
+
F . = 9.326
- ,0410 U , _ ,
( - 1.625) ( - ,0659)
R 2 = .81
,6840 P I _ I - 1.3053 FX, _ - .0277 BCP, (3.326)
( - 5.5958)
( - ,0240)
F
=
+
D-W
=
1.21 N
=
,7053 P , - , - 1.3165 F X , _ I - .3011 M,
(3.141)
(-5.764)
( - .4978)
9.689
D-W
=
1.21 N
(6)
14
I
(7)
= 14
2'Silk. Ni.rorruniirs. p. 105.
T h e r e are, of course. finite san~pleproblems nssociated w ~ t h:,I1 of our periods. In addition. some of
the explanatory variables. for certain periods, are rather Iiighly correlated. 7hcrcli)re. under these conditions our results (especinlly the perverse signs we occ,lsionally get) >hould be intcrpre~edin a gclarded
fashion.
MONETARY POLICY
Federal
Funds
Rate,
37.131 .- .0417 U, _
(-1.651)(-.0442)
= -
R2 = .82 F
vb
,+
,
,
,
.7049 P, - - 1.303 FX, - - 1.1661 RPD, (3.501)
( - 5.785)
( - .6578)
(8)
=. 9.939
D-W
=
1.30 N
=
-
i
Complete reliance on the freeze to hold prices down is not apparent because the
coefficient on the price level varjable is positive and significant. although,not a s great
in magnitude as during the original Game Plan 1 period of 1969. As the price level
rose, so too did the Federal funds rate. The results suggest that the monetary authoriry responded to the administration's signal and took firm action to restrict price
inflation.
However, there seems to have been no concern for the prevailing level of unem\<
ployment in this period. the coefficient of the unemployment variable is statistically
insignificant, suggesting that the Federal Reserve was willing to,live with the 5.5 to 6
percent rate of unemployment that obtained. In fact. the unemployment rate stayed
at the 5.8 to 6 percent level from August 1971 until June 1972.
Once again, the sign on the coefficient of the foreign exchange variable is significant but negative in sign. Again, this suggests that the central bank had gradually
moved to even higher desired levels of exchange rates during this era. T h e result is
consistent with the administration's policy of letting the dollar float during this period of actual and impending devaluations of the dollar and upward revaluation of
several foreign currencies.
Finally. neither the rates of growth of bank credit, the money supply, nor reserves
available to support private deposits (RPD's) prove to have been statistically significant constraints on monetary policy during the period. even though RPD's were
announced to have become an "official" constraint on monetary policy in February
1972.
,-..t
Y
J
The Federal funds rate lurched continually upward from September 1972 to September 1973 and leveled out late in 1973. Many interpreters would label this a sign of
tightness. The abandonment of cuntmls and this indication of tight money suggested
to many observers a return to the old Game Plan I. Our results with respect to the
coefficient of the price level variable would lead us to disagree.
Federal
Funds
Rate,
=
- 6.3175 - 1.6929 U , _ , + .09838 P I - ,
( - 1.1470) ( - 2.4571)
(4.5887)
R2 = .97
F
=
95.88
D-W
=
+
,29339 FX,_,-.75752 BCP,-,
(5.6079)
( - .63544)
1.80 N
=
Federal
Funds
=
-::
( - ;.
Rate,
14
The Burns Era: Phase O(ut), a Return to
the Old Game Plan? September 1972 to
February 1974
I,'.[ *{
d
Tests of the F
17
(9)
Federal
Funds
Rate,
=
-43
( - .6
As price conr
of price inflatia
the period, it d
periods. In the>
index generated
funds rate- t h d ~
the pre-freeze
Once more, b~
and the Fed's pr.
did not rise quic
weak reaction, ti
growth. In an er
virtually guaran
In contrast to:
significant sign (
equations the ce
tightness since ' i
September of IS
Federal funds :;
desired rate of u
ment decreased.
Another elemc
and significant s:
a wave of selling
in the period prc
summer of 1973.
bank took steps;
was temporarily
letting interest I
'$Rose, "Agony o
rate movements (in
losses) also helps acc
T A R Y POLICY
1.1661 RPD,-I
( - ,6578)
(8)
I-
Federal
Funds
Rate,
=
- 6.2691
iling level of unemriable is statistically
ive with tlie 5.5 to 6
loyment rate stayed
e variable is signifibank had gradually
11s era. The result is
loat during this peward revaluation of
supply. nor reserves
1 statistically signifbough RPD's were
'policy in February
e r 1972 to Septemlabel this a sign of
11money suggested
with respect to the
R2
Federal
Funds
Rate,
- 1.7 107 U, -
+
( - 1.0975) ( - 2.353)
I
pparent because the
:]though not as great
9. As the price level
1e monetary authorion to restrict price
481
~~~t~of the Federal Reserve's Reaction to the State of the Economy: 1964-1974
=
- 4.9535
=
.97
F
=
93.157
--1.8633 U,-,
( - .89241) ( - 2.6850)
RZ = .97 F
=
.09584 P, _
(4.4398)
+
D-W
1
=
+
,29508 FX, _ - ,12627 M, _
(5.3131)
( - .I8808)
1.992 N
=
17
(10)
.I0144 P,-I + ,26497 FX,_, + ,68347 KPD,(4.7861)
(4.4593)
(I ,063 1 )
101.254 D-W
=
2.18
N
=
17
( 1 1)
AS price controls were reduced (the shift to Phase 111) in December 1972. the
of price inflation accelerated. Even though the Federal funds rate rose steadily
the period, i t did not respond as strongly to increases in the price level as in
periods. In the second and third equations a one-point rise in the wholesale
index generated, over the period, a mere .09 percentage point increase in the
funds rate-the lowest value of this coefficient for a n r period of our studv.
the pre-freeze expanslonary period.
Once more, because of political antagonisms toward high and rising interest rates
and the Fed's practice of smoothing changes in interest rates.l5 the Federal funds rate
did not rise quickly enough as the demand for money increased. As a result of this
weak reaction, there' occurred in 1972 and 1973 a rapid acceleration of money supply
growth. In an economy recovering from the doldrums of 1970 and early 197 1 . this
J
virtually guaranteed enormous inflationary pressures in 1973 and 1974.
In contrast to the wesk reaction to price inflation, there is a sizable negative and
significant sign on the coefficient of the unemployment variable. In two of the three !->'
J
equations the coefficient has a greater absolute value than in any period of ease or
tightness since 1964. Unemployment rates fell rather steadily from 5.6 percent in
September of 1972 to 4.5 percent in October of 1973. The negative reaction of the
Federal funds rate suggests, as with all results under the Nixon administration, a
desired rate of unemploylnent in the areas of 5.5 percent. such that, when unemployment decreased, the Federal funds rate was raised.
Another element of tightness in the reaction function for this period is the positive P
and significant sign of the coefficient of the foreign exchange variable. Early in 1973
a wave of selling buffeted the U.S. dollar and the dollar was devalued. Nevertheless,
in the period proceeding the devaluation and during the continued weakness of the
summer of 1973, in accordance with the Smithsonian Agreement of 1971. our central
bank took steps to bolster the exchange rate because i t was believed that the dollar
was temporarily undervalued. The statistical results indicate that these steps included
letting interest rates rise to narrow the gap between foreign and domestic rates.
--
'
"Rose. "Agony o r the Federal Reserve," argues that [he Fed's propensity to stabilize upward interest
rate movements (in order t o protect the bond positions o f the government securities dealers from u p i t a l
losses) also helps account for the present inflation.
MONETARY POLlCY
Finally, in all the equations the coefficient on the monetary ;iferegate was insignificant, suggesting that despite the upsurge of ballyhocj there was little systematic
concern with the growth rate of monetary aggregates in this period (the only period
in which the coefficient of one of these variables was significant and positive was the
February 1970 to July 1971 interlude just after Arthur Burns became Chairman of
the Board of Go\.ernors). On balance. the results for this subperiod suggest an apparent return to a posture of tight money mitigated by an extremely mild reaction to the
enormous price inflation and rapid money supply growth that took place.
A Return to "The Old Time Religion":
1974- 1975
In March 1974 the Federal funds rate moved very sharply upward but by early fall
a small downward rebound occurred. This overall rise was widely interpreted as a
spiritual reconciliation of the. Nixon-(Ford)-Burns regime with the "old time religion" of a stable monetary policy. While the direction of the move in the Federal
funds rate was surely appropriate, the magnitude and persistence of the increase in
face of the recession that it caused was rather excessive. especially when contrasted
to the preceedin: years of excessive monetary ease. .
Statistical estimation is not feasible because of the lack of sufficient data points.
Nevertheless. since 1974 the Federal funds rate appears to have been reacting positively but a good deal more strongly to increases in the price level and negatively but
considerably less strongly to the rising rate of unemployment than in any earlier
period of monetary tightness under either Martin's or Burns' Chairmanship.
The adoption of a preannounced desired monetary growth range in 1975 was
widely lauded. Nevertheless. there is skepticism regarding \vhether the Federal Reserve can sustain monetary restraint during periods of economic boom. If our impression from the preceding tests is correct. monetary stabilitv in future periods will
be. once again. challenged by pressures for a moderating influence .on interest rates
during periods of prosperity.
Summary of Statistical Results
-
The statistical findinss for the 1 9 6 4 1974 period indicate the Federal Reserve System
under the chairmanship of Willianl ILZcChesney Martin reacted more strongly to
price inilation than i t did under the chairmanship of Arthur Burns in periods of tight
money. A similar result holds for periods of easy monetar! policy. This tends to
corroborate the belief that the Federal Reserve lost much of its ;inti-inflationary
nli1it:inc.e ~vhenBurns was chairman.?" Moreover, i t is consistent with the near doubling of the rate of growth of the money supply that occurred under Burns by 1974.
IGrwo o f the previous \~udicso f the rc;ic~it,n hnclion rel:~leto the present research. When he uses the
Treasury hill rate as the dependent variable. Chri5ti;ln's resul~s;111:1chrrasc~nablesigns to the explanatory
v3ri;lhlcs. yet onlj the coellicienr on thc prtce lcvel var~ahleis s i g n ~ f i c ; ~Tor
n ~ ~hc.1952-1964 period.
Uniortuna~ely.Christian uses qui~r~erly
data.
Frirdlaencler's study emplclys sirnuistion experiments l o eslirnate impact rnulripliers. This allows her to
Tests of the F e t
During the%
ployment rate ii
consistent with!:
mately 5 percel-:
to 6 percent lewi
The internatit!
Federal Reserve
Burns was chair
there was relua
periods the mai
encourage the d
197 1 and early I ;
Despite the tn
cerning monetar,
have systematic&
one half of the c t
icant Federal Ri
supply, the bank.
Overall, the re
responded to the
its implicit man&
us that these respc
in terms of price
years has increasi.
after short bouts,
inflation. Our con;
a much publicizel
and price setterst
behavior to the
unavoidable mate
simply reflects de
responsibility to s
fi
--
separate estimated red
during the Kennedy-.
policymakers than unc
ETARY POLICY
;:.aggregate was insig::was little systematic
xiod (the only period
?,andpositive was the
hecame Chairman of
'iod suggest an appar.y mild reaction to the
[took place.
ward but by early fall
.dely interpreted as a
h the "old time reli:move in the Federal
rce of the increase in
ally when contrasted
ufficient data points.
e been reacting posiel and negatively but
than in any earlier
'hairmanship.
range in 1975 was
ther the Federal Relic boom. If our imn future periods will
nce on interest rates
era1 Reserve System
d more strongly to
IS in periods of tight
~licy.This tends to
its anti-inflationary
with the near dourder Burns by 1974.
!arch. When he uses the
signs to the explanatory
:the 1952- 1964 period.
bliers. This allows her ro
During the Johnson administration the statistical results suggest a desired unemployment rate in the area of 3.5 to 4 percent. Under the Nixon presidency results are
consistent with the hypothesis that the desired rate of unemployment rose to approximately 5 percent during 1969, stayed at this level in 1970 and 197 I , but was at the 5.5
to 6 percent level during the 1972- 1974 period.
The international position of the dollar appears to have induced tightness by the
Federal Reserve while Martin was chairman in the 1964-1966 period and while
Burns was chairman during episodes of international pressure on the dollar when
there was reluctance to let i t float, as in early 1971 and early 1973. In all other
periods the monetary authority either seemed to ignore the exchange rate or to
encourage the value of the dollar to fall to more realistic levels, such as during late
1971 and early 1972.
Despite the tremendous eruption of Federal Reserve policy pronounLements concerning monetary aggregates, widespread arguments favoring their control seem to
have systematically influenced monetary policy actions only during the first year and
one half of the chairmanship of Arthur Burns: in all other periods statistically significant Federal Reserve reactions do not attach to the growth rate of the money
supply, the bank credit proxy or reserves to support private deposits.
Overall, the results indicate that in the 1964-1974 decade the Federal Reserve
responded to the state of the economy in a manner only very roughly consistent with
its implicit mandate from Congress to stabilize the economy. Casual observation tells
us that these responses have not resulted in an adequate performance of the economy
in terms of price inflation. Our results suggest that the monetary authority in recent
years has increasingly resigned itself, first, to "living with" higher inflation and, then,
after short bouts of severe monetary stringency, to reactlng less strongly to price
inflation. Our conjecture is that such failure and fatalism are not the consequences of
a much publicized breakdown of "the laws of economics," (except insofar as wage
and price setters may have invalidated standard economic models by adapting their
behavior to the Fed's failures). Nor are they consequences of special factors such as
unavoidable materials shortages and crop failures. Rather, the poor performance
simply reflects deterioration over the 1964-1974 period of the Federal Reserve's
responsibility to sustain monetary stability.
-separate estimated reduced form coefficients and utility weights in her regressions. Her results suggest that
during the Kennedy-Johnson years price stability was more highly regarded by monetary and fiscal
policymakers than unemployment, a resulr that is not inconsisrent with ours.
MONETARY POLICY
6. Monetary Policy in an Open
Economy
'I
/
I
The final reading in the previous section by Havrilesky, Sapp, and
Schweitzer indicated that the international position of the dollar had a
significant statistical influence upon monetary policy during the
1964- 1974 period. No longer can monetary economists regard lightly
the effects of the growth of international economic interdependence on
monetary policy. Together with the international oil crisis, recent
changes in the exchange-rate system have greatly influenced the effectiveness of traditional monetary and fiscal policy.
The opening article in this section by Donald L. Kohn explores the
implications of the growth of international interdependence, the move
to flexible exchange rates, and the reduction in supply of a key commndity such as oil for monetary and fiscal policy in the count~iesinitiating such actions and on other countries in the world. The second
article by Donald S. Kemp presents the monetary theory of the halance of payments. This theory incorporates some of the fundamentals
of money supply and money demand theory developed in Section 111
of this book. Comparison of the new monetary balance-of-payments
theory with older theories is the theme of the third and final reading in
this section by one of the originators of the monetary approach, Harry
G . Johnson.
Interc
Rate I
Econc
Donald t
Over the past 20,
increasing degret
tity of goods and
and capital flows
levels. As a resul
fited greatly in t
inputs.
The rise in int
tional economy
conditions. That
been increased v~
reduction of nat
nomic activity. P
bility in their ex
internal economi
The growth of
cations for the
traditional mone
Reprinted from th
permission o f the pu!