The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
I, !
.Ï, /¡11.
-..
CLEVELAND.
AÐDRESSES
iäloö"näñ, Ëiolil
Aprll 25, ì989
I
.
HOSKINS
Ytr
r
Èt'
Ìt
8i:
.-
c}
t.
!.¡
ii
ts,
I
Monetary Pol lcy:
The Importance
of ,Commlttlng to Prl ce Stabl
I I ty
H. Lee Hosklns, Presldent
Federal Reserve Bank of Cleveland
i
{
t
Natlonal Economlsts Club
l,lashlngton, D.C.
Aprll 25, 1989
Monetary Pol I cy: The Importance
Recent increases
commi
til
ng to pri ce stabi t i ty
ln prlce lndexes have caused more people to become
concerned about i nf I atl on . Ì.{any peopì
lt
of
e appeared
compl
acent about i nf I ati
on
of 6 percent lnflation has
aroused attention. For some tlme, I have been argulng for zero inflation. I
when
measured around
4 or 5 percent, but neys
favor an expllclt obJectlve of zero lnflation because
llkely nor deslrable for
any government
appreclably different from
ls
do not belleve
it
fs
to malntaln a stable lnflation rate
zero. Zero lnflatlon
the central bank unless there
I
some advantage
should be the objective of
to having inflation or
unless
the costs of going to zero lnflation are too great.
Crttlcs of thls zero lnflatlon pollcy argue that there could be some
short-run costs to worklng tovard zero lnfìatlon. They may be correct. But,
in
my
oplnlon, those costs would be overwhelmed by the long-term benefits of
having a stable
lf
prlce level. Furthernþre,
I think the short-run costs could
ls credlble. To be credlble, the
pollcy needs wlde support among the publlc and our polltlcal leaders. I
be reduced
thlnk our
a zero lnflatlon policy
vith hlgh and uncertain lnflation in the .l970s
success of western economles ln the lorr lnflation
unhappy experlence
as well as the remarkable
perlod of the 1980s should lend support to a policy of further reduclng
lnflatlon to zero.
Some
have also argued
that havlng a zero lnflatlon po'licy as an ongolng
proposltlon would be a bad ldea because
lt
would
llmlt the flextbility of the
Federal Reserve to deal with unforeseen au.n,, and economic emergencies. I
belleve the opposlte
ls true, and wlll return to thls
issue later.
-2-
I w'lll dlscuss two basic questions: l{hat ls a zero
inflatlon pollcy and why should we try to achleve it? Perhaps one of the
most often repeated obJectlons to a dlslnflation poltcy is that lt mlght
brlng about a recesslon. I wlll explain why I do not think re should allow
Thls afternoon
thls short-run
concern
to lnterfere wlth long-run
is a zero lnflatlon
l,lhat
A successful zero
inflatlon.
expllclt
l.lhen
I
pollcv?
lnflatlon pollcy
use the term "zero
commltment
poltcy.
to achleve prlce
would completely ellmlnate
lnflatlon pollcy,"
stablllty.
A zero
I
mean
a firm
and
inflatlon pollcy
would requlre a
transltlon perlod ln whlch
gradually.
price level would contlnue to rlse for the next few years,
The
we go
to zero lnflailon
but at a lower rate each year. Over a three- to five-year horlzon, the
target path would
become
a constant prlce level.
A successful zero lnflatton pollcy does not mean that actual price
lndexes would never
fluctuate.
The Federal Reserve cannot control the prlce
level over short time horlzons such as one quarter or
Temporary and unforeseen
even one year.
factors can cause the price level to deviate from
the deslred pollcy target
of no change ln the prlce leveì. In practlce,
non-pollcy aspects of the economy that affect lnflatlon have to be partlally
accommodated. The
prlce level mlght remaln sllghtly
above
or
below the
target path for a year or two, but durlng that tlme the publtc would know the
Fed's
goal.
Though
the prlce level mlght fluctuate sllghtly from year to
year, the publlc would expect to see a pollcy stance dlrected
returnlng the prlce level to the target
path. If
toward
ye have a successful zero
inflation pollcy, actual changes tn the prlce level should average zero
perlods of three to flve years.
over
-3-
to the prlce system. Under
the current strategy, there ls no expllclt goal for the price level. This
pollcy allows unforeseen changes tn the inflatlon rate to accumulate and
A multi-year target would provide an anchor
permanently lncrease the
of
thumb
for forecastlng
over the last 20 years
level. The result ls reflected ln our rules
lnflation. The rules that seem to have worked best
prlce
or so are rules that advlse us to adJust our inflation
rate forecast to equal the last lnflatlon rate reported.
It ls qulte
dlfftcult to see any tendency of the prlce level or lnflatlon to retur:n to
some "normal"
It
or predlctable
should be clear
no lnherent reason vhy
that
average.
I
am concerned
about long-term lssues. There ls
thls zero lnflation pollcy
wouìd change any aspect of
hor the Federal Reserve operates ln the short run, between Federal
Market Committee meetlngs
or at the meetlngs. Knowing whether the current
policy stance ras approprlate would requlre Just as much analysls
Judgment
as
lt
pollcy, the
FOMC
sltuatlon. In February, after
members and nonvotlng
forecasts for economlc actlvlty and lnflatlon
of
and
does today.
Consider our current
monetary
Open
these forecasts lmplled
that the
GNP
chooslng a consensus
Presidents pooled their
ln 1989. The central tendency
def'lator would rlse between 4
and
4-l12 percent in 1989 (Q4/Q4). suppose that prevlously we had adopted
target path for the prlce level beginnlng wlth the
the
GNP
1988:Q4
a
actual level of
deflator and rlslng 4 percent in 1989, 3 percent ln 1990, and so on
untll there yas no further change ln the prlce level target ln 1993 or beyond.
Since I notr think that lnflatlon is lifeiy to rise somewhat above thls
path ln 1989, pollcy should be prepared to err on the slde of belng too
tight. How ttght? Judgment ls needed to say. Are we tlght enough today to
get back onto the hypothetlcal target path?
HoYever,
lt ls clear that economlsts
and the
I
thlnk economlsts dtsagree.
flnanclal markets do not
expect
-4-
the inflation rate to decline to 3 percent in
ln
I
my
example.
If
yield on long-term
percentage polnts lower than
If
or 2 percent ln 199ì, as
price stablìlty rithln 3 to 5 years,
markets expected
would expect the
1990
bonds
to
be much lower, perhaps
then
5
lt ls today.
a zero inflation poìlcy were adopted, the debates about pollcy would
be Just as
llveìy as they are now.
what level
of lnflatlon the current pollcy
Today, the ongolng debates are about both
stance
wlll achteve and what level
of lnflatlon the pollcy ought to achleve. Mtxlng these two lssues only
confuses the debate and creates uncertainty about the long-term outcome. l'lhy
not simply adopt a long-run goal and ellmlnate this source of uncertainty?
l.lhat
is
so bad about a
NOISE:
Inflatlon
llttle
lnflation?
adds "nolse,"
or dlstortlon, to nomlnal prlces.
People
ftnd lt dlfflcult to tell vhether a prlce ls htgh because of lnflatlon
or for
some
other reason, such as hlgh
complicates the calculatlons
dlfferent times.
inefflclencies.
It
that
demand
or
scarce
people must make
supply. Inflation
to compare prlces at
complicates prlce comparlsons and leads to
Perhaps the most important
resources assoctated wlth forecastlng
of
these
is the mlsallocation of
real interest rates.
SOCIALLY IIIEFFICIENT INSTITUTIONS:
Inflatlon creates potentlal
al losses for peopl e hol dl ng rcney and other assets denoml nated I n
dollars. To avold these prlvate losses, markets adJust by creatlng flnanclal
f I nancl
institutlons
and lnstruments
that would be unprofltable ln the absence of
inflatlon. An obvlous example is the flnancial "advice" lndustry that ls
concerned
wlth protectlng lnvestments from lnflatton
to the development of
new accountlng systems and
plannlng and contractlng perlods.
have evolved
COLA
rlsk.
Inflatlon has
led
the adoptlon of shorter
clauses and other forms of lndexlng
to protect varlous partles ln contracts
and government
-5programs. Large flrms have developed cash
paying the
lnflatlon tax on their
money
management techniques
to
avold
balances. Markets have been created
to trade futures contracts in foreign exchange and interest rates. There
also a brlef perlod ln whlch the financlal market traded futures ln the
government's reported
was
level of the Consumer Prlce Index. This market folded
as the lnflation rate declined
ln the 1980s.'
The creatlon and maintenance of these lnstltuttons is an optimal
response to an uncertaln inflation pollcy. It is socially lnefficient
compared
used
to a world ln whlch there ls a stable prlce level.
to protect us from lnflation could be better devoted to creatlng
products and servlces that people deslre
The costs incurred
when
in an inflation-free
INFLATION
in minlmizing private losses from inflatlon are greatest
IS
COSTLY l^lHEll
IT
INTERACTS l.lITH THE TAX SYSTEM:
economlc
this
stlll
A recent study estimated
source alone
Any positive
dlstortions òssociated wlth the existlng
tax system because taxes on the return to capital are not
inflation.
may
at low and stable lnflatlon rates.
rate of inflation magnlfles
of
envlronment.
lnflation ls very high or unpredictable. But lmportant costs
be present even
from
The resources
adJusted for
that the cost of l0 percent lnflatlon
ls (as a rough order of magnitude) about 0.7 percent
gross natlonal product (GÌ{P) each year, and possibly as high as 2
to
3
percent of GtlP each year.2
In addltlon, lnflatlon
causes
other economic distortions by lnteracting
wlth the current tax system to alter the allocation of capital across sectors
of the economy, the debt/equity mix chosen by firms, and the choice of asset
llfe.
For example, mortgage interest payments are tax-deductlble; thus,
inflatlon ralses the value of the deductton relatlve to the real cost of
houslng, givlng people an incentive to over-lnvest in houslng. A siml lar
effect leads flrms to acquire more debt rather than to lssue new stock when
-6faced with a need
for
new
funds.
These costs
of inflatlon
could be reduced
in the tax system. But, in the absence of tax
changes, the distortlons represent important costs even to low, stable,
ful ly anti ci pated I nfl atlon.
by approprlate changes
I}¡FLATION REDUCES ECONOMIC GROHTH: l.lhen
inflation, I
efflclency
am
economy
ls malnly
pollcy,
such as the
pollcies
can also
determlned by
rate
growth.
The long-term growth
of technologlcal lnnovation.
rate of
an
monetary
But government
economlc growth.
a large set of countrles, rrith very dlfferent instltutlons
and economic condltlons, supports
reduced by
lmpedes economfc
factors havlng llttle to do wlth
affect the rate of
Evldence from
is
I talk about the cost of
talklng about the ways that inflation
and reduces economlc
and
lnflatlon
this conclusion: long-term economtc growth
and by greater
varlatlon in the growth rates of
the
supply.' Thls negatlve effect of lnflation on long-term growth may
reflect a varlety of causes, includlng adverse effects of inflatlon on
rnoney
capital formation (elther dlrectly or through lnteractlons with the tax
system), the use of scarce resources to form socially inefficient
institutlons,
inflation.
and the
I'lhlchever
dlstortion introduced 'lnto the price
of
these channels
system by
ls most lmportant, the evidence
indlcates that the reductlon ln the economic growth rate assoclated wlth
htgher lnflatlon
ls large and pervaslve.
Hhat are the costs
havlng zero
of
adopting a zero lnfìation poìlcy
lnflatlon as an ongolng pollcy
-- the costs of
and the one-time costs
of gettlng
to zero lnflatlon?
The most often-repeated obJectlon
cannot get
to a zero inflatlon pollcy ls that
to zero lnflatlon wlthout lnduclng a recesslon.
The next
we
-7often-mentloned obJectlon to a zero
inflation policy is the notion that
zero lnflatlon policy would requlre the Federal Reserve to glve up
lmportant degree of
flexibility -- flexlbllity that is
unforeseen events such as
needed
financlal crlses, supply shocks,
ZERO INFLATI0N POLICY AND RECESSION: He
don't
to
like droughts, strikes,
poltticaì dlsturbances ln other countrles;
many smal
I
Even
dtsturbances
if
respond to
understand recessions
by
and wars; by economic and
and by the aggregate effects of
to indlvlduals, to flrms,
we ellmlnated
an
and wars.
completely, but we believe they can be caused by pollcy actlons;
macroeconomlc dlsturbances
a
all effects of policy,
and
to industries.
we yould
still
have
recesslons and expanslons. Recesslons occur because some economlc events are
unpredlctable. They cause past declslons to be wrong, declsions about where
to invest,
what
Conslder
to produce, what to
to sell.
for a moment the analogy between recesslons and earthquakes.
Earthquakes occur when the plates
understand what
ls
have no reason
of the earth shlft.
l{e don't compìetely
gotng on lnslde the earth, but sclentists belleve that
this shifting may occur in
way
buy and nhat
many smal
to thlnk that if
I
quakes
government
or wlth a few large ones.
l,le
geologlsts suddenly discovered
to delay the next earthquake that lt would be good to do so. In fact,
slnce the plates have to shift by the
causlng a much ïorse earthquake
lf
we
same amount anyway, we may
Just
trled to prevent the small
ones.
be
I thlnk the same ls true of recesslons. Shlfts are occurring ln the
economy
that economlsts and pollcymakers do not completely understand --
changing technology and the changlng tastes
Shifts occur
whlch are consldered
sptlls, etc. If
accommodated
we
let
to
of
consumers and lnvestors.
be uncontrollable
--
droughts, oll
market forces operate, these changes
or corrected ln a natural
and gradual
fashion.
wlll be
Market forces
a
-8uork best in a stable pollcy environment. l.lithout doubt there
will
always
be
dtfflculttes, but it ls to our long-term advantage to alìow for
some shift in the economlc "plates" as the world changes.
Perhaps thls earthquake analogy seems to be extreme, but lt ts no more
extreme than the idea that ronetary pollcy can or should be used to ellmlnate
short-term
the buslness cycle. Our attempt at fine tuning the
pollcy during the 1970s probably
made
want a varlable and uncertaln policy
matters
economy
with monetary
rorse. Certainly,
we don't
to exacerbate the business cycle. It is
important to understand that recesslons can occur even under an ideal pollcy.
FLEXIBILITY AND DISCRETION: The most important obJectlon
macroeconomlsts
pol i cy'
that
s
to an ongoing pollcy of zero lnflatlon relates to
preconrni
tment
to
an obJectl
pollcymakers need flexiblllty
ve.
The argument typl cal
to deal wlth
financial crlses, h,ars, and supply shocks.
presence
of a rule
There
of
the
ly espoused
unforeseen events such
These people argue
that
Is
as
the
rrould prevent pollcymakers from actlng senslbly.
ls no reason why the Federal
policy of prlce stabflity and
to the economy. The Federal
sttll
be
Reserve
Reserve could not have an overrlding
free to
accommodate temporary shocks
could, as lt does now, allow the prlce
to supply shocks like last year's drought, an
oll prlce shock, or a shlft in the terms of trade. As long as these
surprlses lnvolve temporary factors, thelr effects on the price level wlll be
level to fluctuate ln
sel
f-correctl
response
ng.
Conslder the drought
1989
in
response
for a ¡noment.
The
prlce indexes rose in
1988 and
to the shortfaìl in food production. As productlon levels
return to normal ln
1989 and 1990, we expect
food prlces. There was
ltttle
need
offsetting
downward pressures on
for a pollcy reactlon ln response to
drought. An easler rcnetary poltcy would not have replaced the lost
the
food
productlon, and a ttghter monetary poltcy was not needed to brlng the prlce
-9-
level back tnto llne wìth the pre-drought expectations. Looklng back over
1988, the Fed tightened, but not
in
response
to the drought.
I thlnk the Fed
did a good Job of looking past the short-term effects of the drought in lts
analysls of the underlying lnflatlon
trend. Policy tightened, not because of
the drought, but because there was evidence of rising inflation and inflation
expectatlons across a yide range of markets.
In general, we can never be sure whether changes in the price level òre
caused by
policy or by some factor that
may have been
either permanent or
temporary. Governments and businesses employ economists to sort out
matters. But the uncertalnty remains. Overall,
it
seems
such
optlmal to allow
ltttle lmmedlate reactlon from pollcy.
Time and the accumulation of evidence wlll usually tndicate whether a mlld or
these prlce variatlons to occur wlth
a strong pollcy
response
ras appropriate. As long as the long-run pollcy ls
credible, pollcy actions should ìargely be antlclpated and the short-term
market adJustments should have
little effect on long-term economic
performance.
l'lhlìe
some
people argue
that commlttlng to zero tnflatlon
Fed's fìexibllity to deal wlth unexpected crlses,
more
in
llkely to be true.
emergency
dealtng wlth crises,
long-run
if
pollcy. It is
flexible lf lt
had an
ls,
the Fed started from the base of a predlctable
a misconception to thlnk that the Fed would be less
overrldlng goal of prlce stablìity.
in 1987.
lnflation expectatlons were rislng
tlghten
its
and
Before the crash, lnflatlon
the Federal Reserve had begun to
pollcy stance. Confronted by the stock market crash, the Federal
Reserve acted very
Immedlateìy
to affect the economy
the actlons would be ¡rore effectlve in
Consider the stock market crash
and
I I imlt the
I thlnk the opposlte ls
The Fed would have more pobrer
sltuatlons, that
wl I
flexlbly ln provldlng llquldlty to the economy.
after the crash and substantlal pollcy easlng, surveys
showed a
-t
0_
decline ln prlce expectatlons at both short and long horizons. But before
it
rlsing.
long,
became apparent
The
crlsis
that the
economy was sound and
had passed and
was
policy ttghtened. Clearìy, there
to thlnk that a long-run jnflation obJectlve would
reason
the Federaì Reserve to act differently in
stlll
ls
no
have constralned
this lnstance.
lon
Conc I us
inflation pollcy would lmprove economic performance over the long
A zero
haul.
Monetary
pollcy can add certainty and stability to the prlce level
over longer horl zons
flexlbillty
wl
thout gi vi ng up short-term fl
of
and the effectiveness
Federal Reserve has long-term
I
inflation
exi bl I I
emergency measures
ty.
Short term
are enhanced lf the
credibllity.
also believe that the costs associated wtth lmplementing a zero
inflatlon pollcy are greater today than they would have been if the policy
had been adopted two years ago when
decllnlng.
The short-term costs
of
start from a lower inflatlon rate
reduce those costs
goal
inflation expectations yere lower and
reduclng inflation
and
if
will
be ìower
lf
we
the policy is credible. The way to
is to begin today vith a flrm
commitment
to a long-run
.
l,le know
that
successfully.
If
no one can predict buslness cycles systematlcally and
the cycle was predlctable, we would knor how to
rþney over the cycle so
manage
that inflatlon is zero on average. Hlstory lndlcates
that, in the lnterest of prolonging economlc expanslon, monetary pollcy has
fostered higher inflatlon and possibly lowered the potentlal growth of the
economy.
Many people have argued
growth.
that price stablllty would restraln
I thlnk Just the opposlte ls true.
economlc
An economy operates rnre
efftclently vith a stable currency and much of our recent economlc
due
to the success
we have had
success ls
ln reduclng lnflatlon slnce the 1970s.
-l
l^le have been though
l-
a dl sappol ntl ng coupl e of years i n
whÍ
ch we have
allowed lnflation and lnflation expectatlons to rlse by about 2 percentage
points.
The economy would be more
contlnued the deceleratlon
there.
l.le should adopt
lnflatton
efficiently organized today if
of inflatlon all
the way to zero and held
lt
a pol icy that, on average, reduces the rate of
by one percent per year
for the next five years.
in lnflation vould be zero, allowlng
more productive ends.
we had
By 1994 the trend
us to dlrect our economlc resources to
-12-
Footnote
s
For a dlscusslon of these costs, see Milton Friedman, "The Resource Cost
Irredeemable Paper Money," Journal of Pollttcal Economy, voì. 94, no.
(June .l986),
of
pp. 642-647.
3
2.
See Stanley
Fischer, "Towards an Understanding
of the Costs of Inflation:
II," in Karl Brunner and Allan H. Meìtzer, eds., The Costs and
Consequences of Inflation, Carnegie-Rochester Conference Series on Public
Policy, Amsterdam: North-Holland Publishing Company, vol. 15, (.l981), pp.
5-4r
3.
.
See the empirical study of long-run growth rates in 47 countries by Roger
Kormendi and Phillp G. Megulre, "Macroeconomlc Determlnants of Growth:
Cross-Country Evldence," Journal of l.lonetary Economics, vol . 16, no. 2,
(September 1985), pp. l4l-63. Theoretlcal mode'ls that reach these
concluslons lnclude Robert J. Barro, "A Capital Market in an Equilibrium
Business Cycle Model," Econometrlca, vol.48, no. 6 (September 1980), pp.
1393-ì417; Angelo Mascaro and Allan H. Meltzer, "Long and Short-term
Interest Rates ln a Risky !'lorld," Journaì of Monetary Economlcs, vol .12,
no.4 (November 1983), pp.485-5.l8; and Alan C. Stockman, "Anticipated
Inflatlon and the Capital Stock in a Cash-in-Advance Economy," Journa'l of
Monetary Economics, vol.8, no.3 (November l98l), pp. 387-393. In
addition to the empirical evldence cited above, recent simulations by
Marianne Baxter suggest that these effects may be quite large. See
"Approximating Suboptimal Dynamlc Equi I ibria: An Euler Equation
Approach," Horking Paper No. 139, Rochester Center for Economic Research,
Unlversity of Rochester, Aprll 1988.
@FedFRASER