historical u.s. money growth, inflation, and inflation ... growth, inflation, and inflation...

1 With chain-weights, price indexesand quantity indexes are calcu-lated separately for componentsof GDP and therefore the differ-ence between nominal GDPand real GDP growth is onlyapproximately equal to thechange in the GDP price index.With fixed-weights, the GDPdeflator is defined as the ratioof nominal to real GDP andhence the gap between nomi-nal and real GDP growth ratesis precisely equal to the growthrate in the GDP deflator.

FEDERAL RESERVE BANK OF ST. LOU IS

13

NOVEMBER/DECEMBER 1998

Historical U.S.Money Growth,Inflation, andInflationCredibility*William G. Dewald

INTRODUCTION

Although many forces affect individualprices in the short run, the historicalrecord shows that in the long run

changes in the general level of prices, i.e.,inflation, have been linked systematicallyto changes in the quantity of money. TheFederal Reserve uses as its principal mone-tary policy target an overnight inter-bankinterest rate, the federal funds rate, whichit manipulates by open market operationsthat change its portfolio of governmentsecurities, which in turn influences mone-tary growth. Economists both inside andoutside the Federal Reserve monitor a widerange of indicators so as to judge theappropriateness of a monetary policy targetrelative to the goals of achieving a stableprice level and sustained real growth. Formany years presidents of the Federal ReserveBank of St. Louis and many of its economistshave called attention to research showingthat long-term growth of monetary aggre-gates is among the more important of theseindicators. They also have championed thepreeminence of price level stability as amonetary policy goal to provide the bestenvironment for sustained economic growthin a market economy.

The historical data reveal a consistentcorrelation between long-term growth ratesin broad monetary aggregates, spending,and inflation in the United States, but not

between such nominal variables and realoutput. Data from the bond market showthat, despite inflation being at its lowest levelin decades, the Fed has not regained fullythe inflation credibility that it lost in the1960s and 1970s.

NOMINAL AND REALGROWTH, AND INFLATION

The financial press and the public oftenseem to believe that the way to containinflation is to pursue policies that reducereal economic growth. The view that itnecessarily takes lower real growth, oreven a recession, to slow inflation is animproper reading of historical data. It failsto differentiate between the short run andthe long run. Sustained real output growthdepends on increases in the supply of laborand capital, and increases in the productivityof such inputs. Growth in demand for outputcertainly influences what is produced, butfundamental scarcities limit the aggregateamount of how much can be produced ona sustained basis. Furthermore, as the leveland variability of inflation increase, pricesignals become fuzzier and decisions aredistorted, which would tend to decrease realgross domestic product (GDP). Thus, oneshould not expect an increase in demandgrowth to increase real growth on a sustainedbasis, but if at all, only in the short run. Anexamination of the historical recordsupports this proposition.

Figure 1 is a plot of percentage changesover 1959-1997 in nominal GDP, real GDP,and the GDP price index. The chart includesyear-over-year and 10-year moving averages.The four-quarter changes—the fine lines—remove the high frequency noise from thedata. The 10-year changes—the heavylines—remove the business cycle fluctua-tions as well. The gap between nominaland real GDP growth rates approximatesinflation, as measured by percentage changesin the GDP price index, which are shownin the bottom panel of Figure 1.1

William G. Dewald is the retiring Research Director of the Federal Reserve Bank of St. Louis. Gilberto Espinoza provided research assistance.*Earlier versions of this paper were presented at the Morgan Stanley Dean Witter Interest Rate Conference, Little Rock, Arkansas, May 6,1998; the Conference on the Conduct of Monetary Policy, Institute of Economics, Academia Sinica, Taipei, Taiwan, June 12, 1998; and theInstitute of Developing Economies, Tokyo, Japan, June 17, 1998.


14


Figure 1 reveals several regular patterns:

• Inflation trended up through about 1980, and downsince then.

• Annual growth rates in nominal GDP and real GDP went up and down together.

• Annual growth rates in nominal and real GDP were also more volatile than annual inflation rates.

• Recessions—marked by the shaded bars—occurred more frequently from the late 1960s through the early 1980s when inflation was high and rising than since the early eighties when it trended down.

• Inflation typically accelerated before cyclical peaks, but then decelerated beginning in recessions and extending into

the early phase of recoveries.

• In recent years inflation has been its lowest and most stable since the late 1950s and early 1960s, and, atypically, it has continued to decelerate in the seventh year of the expansion.

The year-over-year movements in nominaland real GDP are matched closely in the shortrun, an observation seemingly suggestingthat policies to increase nominal GDP growthwould increase real GDP growth, too. Thatshort–run relationship, however, does nothold up in the long run. From the 1960sthrough the early 1980s the increase in 10-yearaverage nominal GDP growth was associatedwith a matching increase in 10-year averageinflation, but, if anything, a decrease in 10-yearaverage real GDP growth. Thus, increasedaverage nominal GDP growth in the longrun was not associated with increased realGDP growth, but only with inflation.

International evidence supports thisfinding that inflation harms long-run growth.

Figure 1

Inflation and Real Growth (Year/Year and 10-Year Averages)Percent

25

20

15

10

5

0

-5

-10

25

20

15

10

5

0

-5

-10

1959 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97

Nominal GDP

Real GDP

61

1959 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 9761

GDP Inflation

NOTE: The bold lines are centered 10 year moving averages of the respective series.


15

2 See Barro (1996) andEijffinger, Schaling, andHoeberichts (1998).

3 See Bruno and Easterly(1996).

4 See Friedman and Schwartz(1963).

5 Despite M2’s imperfections as a cyclical indicator, theConference Board’s monthlyLeading Indicators Indexincludes M2 relative to theprice level as one of its 10components.


Studies of other countries have identified asmall negative effect of even moderate inflationon real growth.2 Small differences amountto a lot over long periods because of com-pounding. Thus, it may not be an accidentof history that the most highly industrializedeconomies with the highest per capita incometoday have had comparatively low inflationover extended periods. With respect to coun-tries that have experienced inflation of 40percent a year or more, the evidence isunambiguous: High inflation reduces real growth.3

High inflation also has been linked tocyclical instability. There was a deep reces-sion in 1981 and 1982. This recession wasassociated with a genuinely restrictive mone-tary policy and interest rates at unprecedentedlevels. The rate of unemployment built up tomore than 10 percent and inflation fell farmore sharply than most forecasters hadexpected. Despite some relapse in the late1980s, inflation has trended down since theearly 1980s, and real GDP growth has averagedsomewhat less than it did in the 1960s, butthis is because of lower productivity growthand not because of recessions and unemploy-ment. In fact, the U.S. economy has per-formed very well relative to its potential andbetter than ever in terms of cyclical stability.The 29-quarter expansion from 1991:Q1through 1998:Q2 had not yet lasted as longas the record 34-quarter expansion of the1960s. However, as inflation decreased fromthe end of the recession in 1982:Q4 through1998:Q2, there were 63 expansion quartersand only three contraction quarters, anunprecedented era of cyclical stability inU.S. history. It surpassed the record of1961:Q1 through 1973:Q4, which included47 positive growth quarters and four con-traction quarters. The record was not tooshabby in either case, but there was a differ-ence. The 1960s and early 1970s were aperiod of accelerating inflation, which laid afoundation for the instabilities that followed.The 1980s and so far the 1990s have beena period of decelerating inflation, which haslain a foundation for stable price level cred-ibility and efficient resource utilization.The next figures bring monetary growthinto the picture.

MONETARY GROWTHAND INFLATION

M2 is a measure of money that MiltonFriedman and Anna Schwartz trace in theirMonetary History of the United States.4 It isa broad measure made up of assets havinga common characteristic: Each is eitherissued by the monetary authorities, forexample, currency and coin, or is an oblig-ation of a depository institution legallyconvertible into such standard monetaryunits. M2 assets can be divided into M1and non-M1 categories. M1 componentscan be used to make payments directly(currency, travelers checks, and checkingaccounts). Non-M1 components, whichcan be readily turned into M1 assets,include savings deposits, money-marketmutual fund balances, and short-term timedeposits. Such non-M1 components of M2have become increasingly accessible todepositors for payments in recent years.M2, as a broad monetary aggregate, repre-sents the essence of “liquidity,” i.e., a waystation between income receipts andexpenditures for both households andnon-financial businesses, and, as such, avariable that would be expected to berelated to total national spending incurrent dollar terms, i.e., nominal GDP.

Figure 2 plots growth rates in M2, nominalGDP, and the GDP price index. It reveals someregular short-term patterns in the year-over-year data:5

• M2 and nominal GDP growth rates slow before and during the initial stages of a recession.

• M2 growth turned down many more times than the number of cyclical peaks.

• M2 growth turned up during each recession and early recovery except during the most recent instance when it continued to slow.

• M2 and nominal GDP have been growing at similar rates between 1995 and 1998.


16


This figure reveals the major reasonwhy M2 has been discredited as an indicatorof the stance of monetary policy in recentyears—in the short run, movements in themonetary aggregates, nominal GDP, andinflation sometimes appear to be unrelated.For example, whereas M2 growth sloweddramatically between 1992 and 94, nominalGDP growth accelerated. That discrepancyproduced the largest and most persistentdeviation between the growth rates in M2and nominal GDP in many years. Thisdeviation has led the Federal Open MarketCommittee (FOMC) and the public to placeless emphasis on the money supply targets.

Nevertheless, giving up on the aggre-gates might be a mistake. The reason is thatthere has been a close long-term fit betweenM2 growth and nominal GDP growth and,in turn, inflation. Figure 2 shows the gen-eral upward trend in 10-year average M2growth, nominal GDP growth, and inflationduring the 1960s and 1970s, and the gen-eral downward trend in these 10-year

averages during the 1980s and so far in the1990s. Such a longer-term historical rela-tionship is presumably a reason why M2 isone of the monetary variables for which theFederal Reserve continues to announce atarget range in the Congressional Humphrey-Hawkins hearings twice a year. In hisHumphrey-Hawkins testimony in February1998, and again in July, Chairman AlanGreenspan noted that M2 growth might beback on track as an indicator of nominalGDP growth and inflation, after it appearedto have been off track earlier in the expansion.

Observations about M2, nominal GDPgrowth, and inflation over the long runsupport Milton Friedman’s dictum that“inflation is always and everywhere a mon-etary phenomenon.” The looseness of theshort-term association supports his dictumthat “lags are long and variable.” Figures3 and 4 makes these points with datagoing back to 1875.

Figure 3 shows that short-run, year-over-year changes in these historical series

Figure 2

M2 and Nominal GDP Growth, and Inflation(Year/Year and 10-Year Averages)Percent

25

20

15

10

5

0

-5

-10

25

20

15

10

5

0

-5

-10

1959 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97

Nominal GDP

M2

61

1959 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 9761

GDP Inflation

NOTE: The bold lines are centered 10–year moving averages of the respective series.


17

are very noisy. Yet, even on a year-over-year basis, the association of large movementsin M2 with large movements in nominalgross national product (GNP) and inflationis apparent, (for example, the contraction ofmonetary growth and nominal GDP growthin the early 1930s and the associated defla-tion). In less turbulent times such as recentdecades, however, there is no clearly discern-able systematic short-run association betweenbroad money growth, nominal GNP growth,and inflation. Of course, a change in theprice level over a year or two is not reallywhat is meant by inflation unless it is sub-stantial enough to change the price level a lot.

Figure 4 shows that over the past 35years the long upward and downward cyclein M2 and nominal GNP growth rates andinflation is only one of a series of comparablelong cycles in U.S. history. Following aperiod of low M2 growth and deflation inthe 1870s and 1880s, there have been fourlong inflation-disinflation cycles. They aremarked on the figure by troughs in centered10-year average inflation in 1893, 1909,1928, and 1962. In 1998, it is not knownyet whether the last 10-year average plottedwas a trough.

Because M2 growth tracks all previousinflation-disinflation cycles, it goes a longway to avert the suspicion that the relation-ship between monetary growth and inflationis spurious. Monetary historians such asMilton Friedman and Anna Schwartz haverecognized that the mere association ofmonetary growth with inflation does notestablish the direction of causality. To con-firm that monetary growth causes inflation,they cite the evidence that the long-termrelationship between monetary growth andinflation has remained much the samethroughout history, including periods whenwe know that monetary growth resultedfrom supply-side factors. For example, whenmonetary growth accelerated in the 1890s,as engineering advances increased goldoutput, there was an associated inflation.Gold was then a standard into which cur-rencies could be converted. When monetarygrowth collapsed in the early 1930s becauseof bank failures, there was an associateddeflation.

The historical record also includesepisodes when demand pressures led theFed to support monetary increases. Inboth World Wars I and II, Fed policies to


Figure 3

M2, Nominal GNP Growth, and Inflation (Annually)Percent

30

20

10

0

-10

-20

-301875 95

Nominal GNP

M2

GNP Inflation

85 95 1905 15 25 35 45 55 65 75 85


18


help the government finance its debt stim-ulated monetary growth. What followedwere substantial increases in inflation.Nevertheless, even in wartime, there isreason to think that the Fed could havekept a damper on inflation. When federaldeficits rose in the 1980s, but the rate ofmonetary growth fell, inflation did notrise. It fell. The historical evidence is thatwhen the Fed has held interest rates downin the face of demand pressures by stimu-lating monetary growth, inflation hasaccelerated. However, in periods such asthe 1980s, when monetary growth has notaccelerated, inflation has not accelerated.

Every major acceleration in M2growth has been associated with a majoracceleration in inflation. Likewise, everymajor deceleration in M2 growth has beenassociated with a major deceleration ininflation. Accordingly, policy makersmight be making a serious mistake if thenoisy short-term movements in M2 andinflation persuaded them that money doesnot matter anymore. At a minimum,policy makers and the public might bewise to monitor monetary growth, mindfulthat inflationary demand pressures do not

cause money growth unless the monetaryauthorities passively allow that to happen.Since the long run consists of an accumu-lation of short runs, it follows thatsustained shifts in M2 growth are worthnoting when formulating monetary policy.Keeping longer-term average M2 growthand nominal GDP growth in the neighbor-hood of longer-term real growth remains apractical guide for achieving a stable pricelevel environment.

INFLATION AND INTERESTRATES

Readers might be surprised that mone-tary policy has been discussed to this pointwithout much reference to interest rates.This approach was not an oversight.

Interest rates compensate lenders forgiving up current purchasing power andtaking some risk. One risk is that borrowersmight default. Another is that what theypay back might have less purchasing powerthan what was lent.

Despite the conventional wisdom tothe contrary, interest rates often have notbeen a good measure of the thrust of mon-

Figure 4

M2, Nominal GNP Growth, and Inflation(10-Year Averages)Percent

20

15

10

5

0

-5

-101875 95

Nominal GNP

M2

GNP Inflation

85 95 1905 15 25 35 45 55 65 75 85


19


etary policy on demand growth and infla-tion.6 Because increases in expected inflationwould tend to raise rates, rising nominalinterest rates do not necessarily signal ananti-inflationary (tighter) monetary policy.Correspondingly, falling nominal interestrates are not necessarily a measure of amore inflationary (easier) monetary policy.

Nominal interest rates are highly sensi-tive to inflation and inflationary expectations.High inflation expectations lead lenders todemand compensation for the expecteddepreciation in the purchasing power ofthe money they lend, and borrowers areforced to add an inflation premium to theinterest rates they pay.

Apart from default and inflation riskpremiums, real (inflation adjusted) interestrates depend largely on underlying realfactors such as domestic saving and invest-ment and international capital flows, noton monetary growth and inflation. Thus,regardless of monetary growth and inflation,higher real interest rates generally reflectincreased investment opportunities ordecreased saving. That real interest ratesreflect underlying real factors is anotherreason why interest rates are not a reliablemeasure of the stance of monetary policy.

In this regard, technological change inthe 1990s, coupled with the long expansion,may have increased the return to capitalinvestment in the U.S. economy, and hencethe demand for capital relative to historicalexperience, which would tend to increasereal interest rates. In such circumstances,there is a monetary policy risk in under-estimating the upward pressures on realinterest rates that result from an increasein real investment demand. Any attemptto attenuate such pressures by stimulatingmonetary growth would risk a build up ofinflationary pressures.

Fundamentally, monetary policy is tighteror easier not in terms of whether nominalor real interest rates are rising or falling,but in terms of whether inflationary pressuresare falling or rising. As the historical figureshave demonstrated, inflation in a longer-term sense is associated with high monetarygrowth. Figure 5 shows that increases anddecreases in inflation trends are reflected inmajor increases and decreases in nominalinterest rate levels.

Figure 5 plots the federal funds rate,the 10-year Treasury bond rate, and annualchanges in the Consumer Price Index. Wheninflation held in the range of 11–

2 to 2 percent

6 The faultiness of interest ratesas measures of monetary policyin a non-inflationary environ-ment was evaluated in Dewald(1963).

Figure 5

Inflation and Selected Interest RatesPercent

20

18

16

14

12

10

8

6

4

2

01959 97

Ten-YearTreasury Bond

CPI Inflation

FederalFunds Rate

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95


20

during the late 1950s and early 1960s, 10-yearTreasury bonds yielded about 4 percent. Fromthe mid-1960s through the early 1980s, infla-tion trended up and so did both short- andlong-term nominal interest rates. Since then,inflation has trended down and so have bothshort- and long-term nominal interest rates.Thus, the events of recent decades tend toconfirm that high inflation is associated withhigh nominal interest rates and low inflationwith low nominal interest rates. As a corollary,Federal Reserve policies that increase thegrowth of the monetary aggregates, andthereby inflation, would in due course alsoincrease nominal interest rates, despite themyopic view that expansionary monetary poli-cies lower nominal interest rates. FederalReserve policies cannot lower nominal interestrates permanently except by actions that lowerinflation.

INFLATION CREDIBILITYAND INTEREST RATES

Given the propensity to save, averagereal (inflation-adjusted) interest rates would

tend to rise with an increase in trend realGDP growth. The reason is that measuredreal GDP growth is associated with increasedreal rates of return on investment. Averagenominal interest rates tend to deviate fromthe real rate of return on investment by anamount that reflects expectations of inflationand inflation risks. The greater the gapbetween nominal interest rates and real ratesof return, the lower the Fed’s credibility isfor keeping inflation low. Thus, the differ-ence between nominal interest rates and trendreal growth provides a crude measure ofinflation expectations in the bond market,i.e., inflation credibility.7

Real GDP growth (averaged over 10 yearsto remove business cycle movements) drifteddown from about 4 percent during the1950s and early 1960s to about 2 percentduring the late 1970s and early 1980s. Itthen rose back up to about 2.5 percent sofar during the 1990s. Five-year averageinflation drifted up from about 1 to 2 per-cent during the 1950s and early 1960s tonearly 10 percent in 1980, then back downto about 3 percent during the 1990s. The


7 The analysis rests on theassumption that the real rate ofinterest equals the rate of growthof real GDP, when both seriesare averaged over a moderatelylong period of time. This condi-tion arises in theoretical modelsin which consumers are Ricardianand the rate of time preferenceis zero. The condition that thereal interest rate is equal to thereal output growth rate arises intheory since real GDP growth isacting as a proxy for an equilib-rium rate of return on invest-ment. It would be appropriatelyexpressed in per capita terms.Per capita GDP growth hasslowed more than overall GDPgrowth over the period plottedon Figure 6. Therefore, currentinflation credibility would havefallen even more relative to itslevel in the late 1950s andearly 1960s than indicated onFigure 6, if per capita real GDPgrowth had been used to proxythe equilibrium rate of returnon investment.

Figure 6

Long-Term Yields, Real Growth, and Bond Market Inflation CredibilityPercent

14

12

10

8

6

4

2

0

10

8

6

4

2

0

-2

Five-Year Treasury Yield Five-Year Average

Real GDP Growth Ten-Year Average

1965 1970 1975 1980 1985 1990 19951960

1965 1970 1975 1980 1985 1990 19951960

CPI Inflation Five-Year Average

Inflation Credibility


21


five-year average of the five-year Treasurysecurity yield rose from 2 percent duringthe 1950s to 12 percent during the early1980s; but it then fell back to about 6 per-cent in the 1990s. Since bond yields rosewhen inflation accelerated, but real GDPgrowth slowed, the influence of inflationoutweighed the influence of real GDPgrowth on bond yields. Correspondingly,when inflation decelerated, bond yields felleven though real GDP remained stable.

The difference between the five-yearaverage of the five-year Treasury securityyield and the 10-year average of real GDPgrowth is an estimate of the bond market’sfive-year inflation forecast, adjusted forinflation risk. It is the height of the shadedarea in the lower panel of Figure 6. Thismeasure of inflation credibility roughly laggedinflation, indicating that bond yields havenot been very forward looking in forecastinginflation. The measure of inflation credibilityhovered close to zero during the 1950s andearly 1960s, which was credibly a zero-infla-tion-expectations period. It under forecastinflation from the late 1960s until the early1980s when inflation was rising. It overforecast inflation in the 1980s and so far inthe 1990s as inflation has fallen. It peaked

at about 10 percent in the early 1980s, butfell to about 4 percent in recent years.

The bond market inflation forecast (orinflation premium) over the past five yearsrepresents a substantial gain in credibilitycompared with the early 1980s, but a sub-stantial loss compared with the 1950s andearly 1960s. In that earlier period, actualinflation was about 2 percent, but the bondmarket forecast a rate close to zero. In recentyears, inflation has averaged about 3 percent,but the bond market has forecast about 4percent inflation inclusive of an inflationrisk premium. Thus, despite recent inflationbeing the lowest and most stable indecades, bond markets have seemingly notyet been convinced that inflation is downto stay. If the inflation premium wereeliminated, bond yields could fall to matchtrend real growth, as was the pattern in the1950s and early 1960s. That is about 3percent, which is considerably lower thanthe approximately 5 to 51–

2 percent bondyields observed in mid 1998.

Double-digit inflation and inflationaryexpectations are what explain the all timepeak in security yields in October 1981 asplotted in Figure 7. Since then, the entireyield curve has shifted down by roughly

Figure 7

Government Securities Yield CurvePercent

18

16

14

12

10

8

6

4

2

0305 10 15 20 25

October 2, 1981

November 7, 1994

October 15, 1993

January 3, 1959

November 10, 1998

10 percentage points, undoubtedly a reflec-tion of the decline in inflation and inflationaryexpectations. Although markets do notexpect double-digit inflation today, they donot expect price stability either. Duringthe 1950s and early 1960s inflation waslow and generally expected to stay low, acondition that was reflected in long-termrates hovering in the 3 to 4 percent rangeas represented by the January 3, 1959,yield curve on the figure. Despite the his-torical record of an unstable price level inthe short run, there really was widespreadexpectation of longer-term price stabilityuntil inflation took off in the mid-1960s.In fact, never before the 1960s had the U.S.federal government borrowed long term atmore than a 4 1–

4 percent rate. During the expansion that began in

1991, the yield curve touched a cyclicallow on October 15, 1993. It then shiftedup to a cyclical peak on November 7, 1994.Three-month bill rates had increased from3 percent to 5.4 percent and 30-year bondrates, from 5.8 to 8.2 percent. The latterwas presumably an illustration of increasesin long-term interest rates indicative ofrising inflationary expectations in the bondmarket. Although inflation, in fact, did notincrease much during the 1990s expansion,bond markets may well have been anticipatinga repeat of the experience of inflationaccelerating as had typically occurred in thepast. Historically, monetary policy often haslagged behind market interest rates inexpansions and thereby added to, ratherthan damped, inflationary pressures. Bycomparison, the record during the 1990sexpansion has been very good: An extendedperiod of positive real growth withinflation held in check. Yet, with bondrates still above the real growth trend, thebond markets seemingly continue toreflect the fear that inflation will rise again.

HOW TO GET AND KEEPINFLATION CREDIBILITY

What could the Federal Reserve do toenhance its inflation credibility, and therebyallow long-term interest rates to stay lowand prospectively fall further? Most impor-

tant, the Fed should continue to keep inflationlow by limiting the rate of monetary growth.A practical goal would be to get back tothe low inflation and low interest rates ofthe late 1950s and early 1960s. One wayto persuade markets that low inflation ishere to stay is for the FOMC to focus moresharply on the desired outcome for inflationby following several other countries thathave legislated specific low inflation targetsfor their central banks. This list includesAustralia, Canada, New Zealand, and theUnited Kingdom, as well as Portugal,Spain, and Sweden. Whether or not suchefforts are directly responsible, the fact isthat these countries have had considerablesuccess in bringing inflation down andkeeping it down.

A second proposal comes from econo-mists who have argued that credibilitywould be enhanced if there were anannounced policy rule (with respect to the federal funds rate or monetary growth)and the Fed acted on the basis of that rule.The advantage of a rule is that markets wouldknow in advance how the Fed would reactto deviations of nominal spending, inflation,or other variables from specified targets.

A third proposal made by Dewald(1988) is that federal budget offices basetheir budget projections over a 5- to 10- year horizon not on their own inflationassumptions, but on longer-term inflationforecasts from the Federal Reserve. Sincethe Fed has the power to influence inflationover the long term, why not relieve thebudget offices of the responsibility formaking an independent assessment offuture inflation as they make their budgetprojections? Not only could the budgetoffices benefit, but also every business,state and local government, and householdcould benefit from having confidence thatthe Fed would act to keep inflation as lowas it had forecast. Lars Svensson (1996)has proposed that the Fed make its ownannounced inflation forecasts an explicitpolicy target. By using a forecast as aguide to policy, the Fed would be focusedon this objective, but not blind to otherthings going on in the economy that influ-ence inflation.



22

An environment of credible price sta-bility has a high payoff in a market economy.The historical evidence examined in thisarticle supports the conclusion that risksof starting another costly inflation-disinfla-tion cycle could be avoided by monitoringM2 monetary growth and maintaining asufficiently tight monetary policy to keep adamper on inflation. Having achieved thelowest and most stable inflation environmentin many decades, the Federal Reserve hasan unusual opportunity to persuade mar-kets that it will continue to keep inflationlow and, in principle, eliminate it. Anenvironment of credible price stabilitywould allow the economy to functionunfettered by inflationary distortions—which is all that can be reasonably expectedof monetary policy, but precisely whatshould be expected of it.

REFERENCESBarro, Robert. “Inflation and Growth,” this Review, (May/June 1996),

pp. 153-69.

Bruno, Michael, and William Easterly. “Inflation and Growth: In Searchof a Stable Relationship,” this Review (May/June 1996), pp. 139-46.

Dewald, William G. “Monetarism is Dead: Long Live the QuantityTheory,” this Review, (July/August 1988), pp. 3-18.

_______. “Free Reserves, Total Reserves, and Monetary Control,”The Journal of Political Economy, LXXI, 1963, pp. 141-53.

Eijffinger, Sylvester, Eric Schaling, and Marco Hoeberichts. “CentralBank Independence: A Sensitivity Analysis,” European Journal ofPolitical Economy, Vol. 14 (1998), pp. 73-88.

Friedman, Milton, and Anna J. Schwartz. A Monetary History of theUnited States, 1867-1960, Princeton University Press, 1963.

Svensson, Lars. “Commentary: How Should Monetary Policy Respondto Shocks While Maintaining Long-Run Price Stability?—ConceptualIssues,” Achieving Price Stability, A Symposium Sponsored By theFederal Reserve Bank of Kansas City, August 29-31, 1996, pp. 209-27.



23



24

historical u.s. money growth, inflation, and inflation ... growth, inflation, and inflation...

Documents