Some Lags in Monetary Policy

8/12/2019 Some Lags in Monetary Policy

1/15

Carnegie Mellon University

Research Showcase

Tepper School of Business

3-1-1966

Some Lags in Monetary PolicyAllan H. MeltzerCarnegie Mellon University, [email protected]

Follow this and additional works at: hp://repository.cmu.edu/tepper

Part of the Economic Policy Commons, and the Industrial Organization Commons

is Conference Proceeding is brought to you for free and open access by Research Showcase. It has been accepted for inclusion in Tepper School of

Business by an authorized administrator of Research Showcase. For more information, please [email protected] .

Recommended CitationMeltzer, Allan H., "Some Lags in Monetary Policy" (1966).Tepper School of Business. Paper 631.hp://repository.cmu.edu/tepper/631
http://repository.cmu.edu/?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://repository.cmu.edu/tepper?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://repository.cmu.edu/tepper?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://network.bepress.com/hgg/discipline/1025?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://network.bepress.com/hgg/discipline/347?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPagesmailto:[email protected]://repository.cmu.edu/tepper/631?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPagesmailto:[email protected]://repository.cmu.edu/tepper/631?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://network.bepress.com/hgg/discipline/347?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://network.bepress.com/hgg/discipline/1025?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://repository.cmu.edu/tepper?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://repository.cmu.edu/tepper?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://repository.cmu.edu/?utm_source=repository.cmu.edu%2Ftepper%2F631&utm_medium=PDF&utm_campaign=PDFCoverPages


2/15

QJ cU^-y^,^,

Draft, February 1966Nob for Publicationor Quotation

SCKB CCMMENTS 09 LAGS IN MONETARY POLICYby Allan H Meitzer

Paper prepared for the meeting with Board ofGovernors,Federal ReserveSystems March 3 19660


3/15

LAGS XW MONETARY PQLICT

For almost two centuries economists have recognised that changes inthe quantity of money have a delayed or lagged effect on wages* output*pricfSp or gold fleam* Recently there have been a number of attempts tome&iure some of the lags and to suggest the importance of the lags for monetarypolicy* I will make no attempt to summarise each of the many studies or to;onment on the variety of procedures that have been used to generateparticular estimates* Instead, I will discuss sons of the main conclusionsand their relation to the problem of conducting monetary policy

The most widely accepted conclusion is that the response of pricesoutput* or balance of payments variables that are principal goals ofpolicy to changes in monetary policy (or for that matter the response tofiscalpolicy)is distributed over a long* period of time0 This is equivalentto saying that the response within the first month or quarter and even thefirst year* is generally thought to be a small fractionofthe eventual totalor to saying that there is a long*1delay between the time that monetaryaction is taken and the time that tuost of its effects are feltD

Suasnary ofthe vin?ir\~zOnce these rather general statements have beenmade,there are few

remaining areas of clearcut or widespread agreement One of the few isthe rather obvious point that the initial effect of monetary policy onshort-tem interest rates or bank reserve positions is extremely rapid*But between these very short-term effects which are measured in fractionsof a day or a week to the effect on gold flowspwhich Cagan argues takesplace over decadespthere are about as many estimates as there a n


4/15

estimate e The results appear to be sensitive to the statistical methodchosen, to the definitions used for the variables* and to the implicitchoices about causation*

Nevertheless, the opposing viewpoints can be grouped into two categoriesrepresented by those ho believe that most of the effect on output of changesin the money supply occur within a year and those who believe that most of theeffect is felt after a longer period of time If weignore*,temporarilymeasures of the lag between monetary policy action and the components ofspending such as purchases of newhousing,investment in plant and equipmentor inventory Investment, we can group the principal views about the lagbetween money and output as in Table 1>

TABLE 1Some Measures of the Length of the Lag in Monetary Policy

FriedmanFindings

At the peak 6 to 29 months -average 16 monthsAt the trough 4 to 22 months -average 12 months

Gm&vu&m^The lag is longand variable

Culbertson 3 6 months Lags are shortKareken andSolm At most 3 months This lagis short*butthis is an inappropriaway to measure the lag

Each of the three investigators used different measures of mcney andoutput, a point which is important because the purpose of the Kareken andSolow measure was toshearthat Friedman*s estimate was not only inaccurate


5/15

3

but fallacious o They ccncludad that his result was based wholly on hiscomparison of peaks and trough? in the level of output with peaks andtroughs in the rate of change of money0 As noted in Table 1 they didnot think that their own procedure was a useful way to measure the lagcIt was designed to show only that a comparison of turning points in moneywith turning points in output would shear results substantially differentfrom Friedman*80

The conclusion that the lag is long is reached in most of the largeand growing list of studies that purport to measure the lag in monetarypolicy by estimating how long it takes the demand for money to return toequilibrium once it has been displaced These studies are relevantto the discussion* if we interpret the results as saying that monetarypolicy has had its full effect when equilibrium isrestored.Prices,output and interest rates have then adjusted to positions at which thepublic is willing to hold the nominal quantity of momy at the new levelsof interestrates,prices and outputc On this interpretation it takesabout four or five years before most of the effect of monetary policy hasbeen achieved0

The estimates of a four to five year or longer lag* and the relatedconclusion that monetary policy has a small effect in the first year?depend on the use of measurement procedures and a line of reasoningthat I find difficult to accept0 Virtually all of the studies I have seenforce output and interest rates to affect the amount of money demanded withan identical lag and include as part of the lag the adjustment of pricesto changes in output0 In fact prices may not begin to adjust untillong after output* interest rates, and the quantity of money demanded have


6/15

completed a substantial part of the initial adjustment to monetary policyoInshort,these studies do not measure the lag that is of principal interestfor policymaking so that even if the results are unaffected by thestatistical properties of the estimating methods we cannot conclude thatthe lag in monetary policy is as long as these studies suggest* Qr$ toput the same point in another way9we cannot conclude on the basis of thesestudies that the initial effects of monetary policy on output and interestrates are as small or as long delayed as the studies suggest*

One cannot conclude from Table 1 that we have the usual case ofFriedman and one or two allies against almost everyone else0 Having shownthat Friedman*s findings of a long and variable lag are incorrect* Karekenand Solow go on to show that the lag is long and variable0 These conclusionsare based on their preferredprocedure,the investigation of the lags inadjustment of various components of output* Monetary policy is viewed ashaving an immediate impact on net reserves and ehort-tena interest rates0These, in tum9affect the componente of spending either directly or byhanging intervening variables sudi as new orders for capital equipment0Their tentative estimates of the average length of time that it takes forsome of the effects to become evident is show in Table 20


7/15

5

TABIB 2Kareken andSaLorsEstimates of the Timing of Responses to Monetary Policy

1)Variable Affected

Business Equipment(asstndng the fullinmediate effecton new orders*)

Inventory Investment

Total Bank ReservesInterest Rates onBank loans

2 )Proportion of the TotalEffect17ft30-31*5*

50-55o7*57o4-65o2f.

35-WsS52-565567*183-869-95&

25*

(3)Number of Months requiredto achieve the proportionin (2)369215369215126

* Their estimate of the effect on new orders suggest that 45$ of theadjustment takes place in the first quarter after the policy changeand 90$ occurs within one yearp The percentages in coltxan 2)wouldhave to be adjusted to obtain the correct distribution of the responseto monetary policy

Kareken and Solo* introduce a large number of qualifications to thesefindings9so there is little reason to discuss the particular results indetail, The main points to be noted are 1)that banks fully adjust theirreserves to changes in monetary policy within one quarterTrfiilemost otheradjustments take substantially longer and 2)that the estimates of the lagsare not constants but depend on the way in which they are measured and on


8/15

6

assumptions about the way in which the economy operates* Kareken and Soloware quite explicit on the latter point* They note repeatedly that theestimate of the lag varies with the particular assumptions made about thedeterminants of inventoxy investment capitalexpenditure,etc*

The Importance of the Lag for PolicyThe fact that economists attempt to measure the lags and argue about

their length indicates that there is now much more agreement that monetarypolicy has some affect on economic activity than there was at one time*However, the lag studies open a new question ofinterest,whether or notthe existence of a long delay in the effect of monetary policy on outputor its components improves or worsens the success of efforts to promote fullemployment and price stability by monetary policy* On this issue there aretwo divergent views which I will state simply, at the risk of removing muchof their content*

One view is that a l^ng lag allows ample time for the correction oferrors o Since long lags mean relatively small initialeffects,errors dueto misinformation^ changing conditions^ or misinterpretation of prevailingconditions can be largely or completely offset by policiee that push theeconomy in the opposite direction* Since monetary policy operations generallyaim for small or marginaleffects,long lags help to avoid serious errors*This argument would be more persuasive if the lag were constant or nearlyso* By improving forecasts and methods offorecasting,most of the problemscreated by a long lag could be avoided If the lag remained constant* However^the evidence suggests that the lag is not constant, butvariable,so we cannot accept the measures of length without keeping in mind that the lengthis subject to change o For example if the Kareken and Solow estimates of


9/15

7

the response in inventory investment is adjusted for the variability of thelag* their finding that on the average 24$ of the adjustment occurs in thsfirst quarter is changed to a range of 14-33$* and their average of 67$in the first year becomes a range of 45 to BOf>0

Those who have generally taken the second view* that long lags increasethe problem of discretionary monetary policy9frequently start from findingssuch as those I have just cited *hich suggest that the lag is highly variableoThe essential point in this position is that frequent changes in the directionof policy cannot be expected to offset one another Instead the changesmay and often will increase the variability ofincome,cumulate fora time and thus increase the extent to which there is inflation or deflation*,or produce other undesirable effects

To pursue this line of discussion in more detail I will briefly discusssome reasons why the lag is variable and comment on the findings about itslength* I will then offer some suggestions designed to reduce the variabilityand the length of the lag and thereby increase the effectiveness of monetarypolicy

The Effect of Policy on Ths LagBefore turning to thosetopice,let me note that we have equated the

lags*o rdelays in the effect of monetary policy with the time that elapsesbefore a substantial part of the effect of monetary policy is observableHost recent discussions of lags in either monetary or fiscal policy havestarted by assuming that there are two other lags The first is calledthe recognition lag* the length of time it takes for policy makers to


10/15

e

recognise that changes in policy a n desirable* The second, or action lag,represents the delay between the recognition of need for action and the tineaction is takenQ In the Kareken and Solow study^ these two lags are groupedwider the name inside lag*, and the lags we discussed a moment ago arecalled outside lags** I will not dwell on this distinction because itappears to be unimportant* My studies with Brumer suggest that the lagbetween the recognition of the need for action and the time that the Federal Resertakes action frequently is so small as to be inconslquential, at leastin the postwar years* The more important question for policy is whether theFederalReserve,having recognised the need to takeaction,has moved in theappropriate direction* This raises the question of how the Federal Reservedecides whether it has moved to a somewhat easier or somewhat tighter positionand reopens the question ofho r tighter and easier are measured0

To make the discussion less complicated I will equate monetary policyactions with changes in the monetarybase,defined as bank reserves pluscurrency held by the public By this choice I have Included in policy actionall acts of CQomission, for example decisions to buy or sell securitiesin the openmarket,as well as decisions that permit the monetary baee tochange because of gold movements changes in the Treasurybalance,or anyof the so-called non-controlled factors that affects the else of the base*This is not to suggest that theOpenMarket Committee attempts to controlthebase,only that whatever is chosen as the current target of monetarypolicy interestrates,bank borrowing^ freereserves,or total reserves monetary policy will affect the monetarybase* And since the base is themoet important determinant of the money supply and bank credit and an important


11/15

9

deteradnant of market interest ratee9the direction of changes in the baseindicate the direction in which the money supply* bank credit and interestrates will be moved by policy actiono

At this point* we come to the first lag of any consequence? the lag ofthe money supply bank credit or interest rates behind monetary policy actionsThe year I960 provides an excellent illustration of the way in which the choiceof indicators and the length of the lag are related o The Open Market Conmitteerecognised the need for a more expansive monetary policy as early as themeeting of March1st,i0e0at least two months before the cyclical peak* Thedecision *to supply reserves to the banking system soaewhat more readily1* wasreviewed at each subsequent meeting during the yearo Free reserves roee bymore than $1 billion in the course of the year> but the growth rate of thebase slowed until late in the year and the economy headed into recession*

The evidence for the period since 1919 suggests that I960 is not aunique example During mild cycles of the type that we have experienced inthe postwar years * the growth rate of the base is highest* on the average*near the peak of an expansion and lowest* on the average* near the bottom ofthe recession Measured by the growth rate of the base* monetary policyfeeds the expansion and increases the severity of the contraction

The argument can be translated into a statement about the length of thelag by noting two paints0 First* the length of the lag depends on how the lagis measured* which is another way of saying that it depends on the indicatorof policy that is chosen0 Second* the use of relatively poor indicators ofthe effect on output and prices has caused policy to move in the directionopposite to the direction that is desired and desirable0 This delays thereturn to higher levels of employment or increases inflationary pressures**


12/15


13/15

11

Let me make clear that I am not suggesting a monistic explanation ofthe variability of the lago Fiscal policy changes* major industrial strikesand a variety of other forces are operating Federal Reserve cannot do verymuch to eliminate these sources of variation* and I have ignored themfor that reason The Federal Reserve can eliminate one important sourceof variability by paying much less attention to money market indicators andmuch mors attention to indicators of the effect of monetary policy onoutput* employment and prices

Some Suggested ChangesMjrfirst suggestion for reducing the length and variability of the

lag is to obtain a better measure or indicator of the effect of monetarypolicy on economic activity0 Our oun tentative conclusion is that if fullemployment and price stability are the main goale of monetary policy* themost useful indicator is a weighted combination of changes in the monetarybase* the reserve requirement ratios and the rediscount rate Whether ornot this is the best indicator* it is a better indicator and would reducethe length of the lag by reducing the time between turning points in theeconomy and the timing of changes in the direction or sise of policyoperations

My second suggestion is that the rate of change of the monetary basebe made more unifcm The reason for this suggestion is that reducingvariations in the base will reduce variations in the money supply and inthe lag between changes in money and their effect on output and pricesVJhilemany of the changes in the base are smoothed or eliminated in the


14/15

12

moneymaxfcetbefore they can affect output, others remain* Acceleration anddeceleration of output generally have a delayed affect on prices and are asource of one of the more difficult policy problems the choice of adirection for policy when output is falling and prices are rising*

Much cf the variability in policy could be eliminated by allowing thebanks to make adjustments in their reserve positions through the Federalfunds market when there are temporary changes in float or incurrency,depositredistribution between different types ofbanks,etc*

Let me make clear that I am not suggesting the establishment of a rule*The Open Market Committee has an excellent record in judging turning pointsin economic activity but its choice of money market variables as the indicatorof monetary policy often leads it to believe that its policy is one ofrestraint on the economy when it is in fact only restraining the money market*For reasons suggestedabove,frequent changes in the direction or sise ofchanges in the base, contribute to the variability of thelag,the uncertaintyabout the timing of the effect of monetarypolicy,variations in the rate ofchange in output and similar problems*

A third suggestion designed to make the effect of monetary policy morecertain and more regular is the elimination of regulation Q* While time doesnot psrait me to go into the problem atlength,it is clear that changes inRegulation Q have had pronounced effects on the growth rate of the moneysupply and have increased the variability of monetary policy and hence thevariability of its effect on output and prices*

Whilenjysuggestions do not provide either a measure or a means ofmeasuring the variouslags,they suggest scms ways in which the length


15/15

.13

and the variability of the lag in output and in prices can be reduced.By removing these sourees of variability in prices andoutput,a contributioncan be made toward reducing the size of fluctuations in output and improvingthe performance of the economy.

Documents

Some Lags in Monetary Policy