Commentary 400 Plt

7/21/2019 Commentary 400 Plt

1/20

Institut

C.D.HOWEInstitute

commentaryNO. 400

Price-Level Targeting:A Post Mortem?

Price-level targeting has convincing advantages, especially as a tool for

avoiding the worst consequences of economic downturns.

Then why havent central banks experimented with the regime?

Steve Ambler


2/20

$12.00

978-0-88806-921-40824-8001 (print);1703-0765 (online)

Es

s

en

tialP

olicy

Intellig

ence | Conseils indi

spensabl

essu

rles

poli

ti

qu

es

INSTITUT

C.D.

HOWE

I

NST

ITUTE

Finn Poschmann

Vice-President, Research

C N. 400

F 2014

M P

C.D. Howe Institute publications undergo rigorous external reviewby academics and independent experts drawn from the public andprivate sectors.

Te Institutes peer review process ensures the quality, integrity andobjectivity of its policy research. Te Institute will not publish anystudy that, in its view, fails to meet the standards of the review process.Te Institute requires that its authors publicly disclose any actual orpotential conflicts of interest of which they are aware.

In its mission to educate and foster debate on essential public policyissues, the C.D. Howe Institute provides nonpartisan policy adviceto interested parties on a non-exclusive basis. Te Institute will notendorse any political party, elected official, candidate for elected office,or interest group.

As a registered Canadian charity, the C.D. Howe Institute as a matterof course accepts donations from individuals, private and publicorganizations, charitable foundations and others, by way of general

and project support. Te Institute will not accept any donation thatstipulates a predetermined result or policy stance or otherwise inhibitsits independence, or that of its staff and authors, in pursuing scholarlyactivities or disseminating research results.

The Institutes Commitment to Quality

About TheAuthor

S A

is the David Dodge Chairin Monetary Policy at theC.D. Howe Institute,professor in the EconomicsDepartment at the Universitdu Qubec Montral,senior fellow of the RiminiCentre for Economic

Analysis, and a member ofthe Inter-University Centreon Risk, Economic Policiesand Employment (CIRPEE).


3/20

The Study In Brief

Recent research has shown that monetary policy based on price-level targeting has several advantages overthe traditional inflation targeting method, particularly in times of economic distress. Although severalcentral banks have been coping with the aftershocks of the 2008 financial crisis for prolonged periods,

none has adopted price-level targeting.

Tis Commentaryreviews some of the reasons for this in the Canadian and American contexts. Te relativemildness of Canadas 2008-2009 recession convinced the Bank of Canada that inflation-targeting canwork in troubled as well as tranquil times. Meanwhile, the severity of the US recession led the FederalReserve to explore several types of unconventional monetary policies, but not price-level targeting. Telatter requires a commitment to offset the effects of unexpected inflation on the price level and makesmonetary policy history-dependent.

Te Fed prefers to exercise discretion, and inflation targeting allows central banks to ignore past inflation

shocks and engage in fine-tuning of the business cycle. Te Bank of Canada shares the Feds predilectionfor discretion in this regard.

C.D. Howe Institute Commentaryis a periodic analysis of, and commentary on, current public policy issues. Michael Benedictand James Fleming edited the manuscript; Yang Zhao prepared it for publication. As with all Institute publications, the

views expressed here are those of the author and do not necessarily reflect the opinions of the Institutes members or Board

of Directors. Quotation with appropriate credit is permissible.

o order this publication please contact: the C.D. Howe Institute, 67 Yonge St., Suite 300, oronto, Ontario M5E 1J8. Te

full text of this publication is also available on the Institutes website at www.cdhowe.org.


4/20

2

Only Finland and Spain, upon joining the eurozone,have abandoned I, thereby relinquishing control oftheir monetary policy to the European Central Bank.Meanwhile, price-level targeting (henceforth P)is an alternative monetary policy framework thatoffers potential advantages over I.2P can, for

example, lead to greater stability both of inflationand of output. In troubled economic times such asthe recession in the aftermath of the 2008 financialcrisis, P can therefore increase the effectiveness ofmonetary policy, especially when the central bankspolicy rate is stuck at its lower bound of zero.

In spite of these advantages, and althoughseveral central bank policy rates have been at ornear their lower bound for prolonged periods sincethe financial crisis, no central bank has seriouslyconsidered adopting P except for the Bank ofCanada, and interest in P in academic circlesseems to have waned. Te Bank of Canada, whenit renewed its inflation target agreement withthe Government of Canada in 2006,3announcedits intention to study the costs and benefits ofswitching to P. Nevertheless, it decided not tomake the change when the agreement was againrenewed in 2011.

Tis Commentaryreviews some of the reasonswhy P has not been tried and why interest in

it has decreased. In Canada, the mildness of therecession contributed to policy inertia: CanadasI framework is still delivering acceptable results.As well, the Bank of Canada expressed doubtsconcerning how well the public would understandthe workings of a new monetary policy framework,

which is a necessary condition for its success.P is a regime that entails, above all, a commitmentto undo the effects of shocks on the price level. Itmakes current monetary policy conditional on pastshocks and requires that a central bank committo the future course of its monetary policy. Boththe US Federal Reserve and the Bank of Canadahave explicitly stated their preference for beingable to exercise discretion, and this is a decisivefactor explaining why they have not adopted P.Indeed, the Feds experiments with unconventionalmonetary policy since 2008 have had the effect ofincreasing its use of discretion.

The Oper ational Distinctionbetween IT and PT

Before discussing Ps advantages, it is useful toreview the main operational difference between Iand P. Te former, which has come to be thoughtof as conventional monetary policy, involves setting

I would like to thank Philippe Bergevin, Angelo Melino, Chris Ragan, Daniel Schwanen and anonymous reviewers for

comments on previous versions. All remaining errors and omissions are my own. Email: [email protected].

1 See Lim (2008) for a list of 28 countries that were inflation targeters at the time. In January 2012, the Federal Reserve

announced that it would explicitly target a rate of inflation of 2 percent, rather than periodically stating a desired

target range.

2 See Ambler (2009) for a detailed summary of some of the research that supports this conclusion.

3 See Bank of Canada (2006).

Twenty-nine central banks currently have explicit targets fortheir economies inflation rates and have adopted monetary

policies guided by some form of flexible inflation targeting(henceforth IT).1


5/20

3 Commentary 400

an inflation rate target and using the short-termnominal interest rate as the primary tool forachieving the target.4

Under I, unexpected deviations from theinflation target are forgotten. Te central bankmerely tries to bring inflation back to its targetedrate. For this reason, unexpected changes ininflation have a permanent effect on the pricelevel. In addition, I is flexible in permittingother objectives. Carney (2012, p. 6) defines theframework as follows:

Tus, under flexible I, the central bank seeks toreturn inflation to its medium-term target whilemitigating volatility in other dimensions of theeconomy that matter for economic welfare, suchas employment and financial stability. For mostshocks, these goals are complementary. However, forshocks that pose a trade-off between these differentobjectives, or that tilt the balance of risks in onedirection, the central bank can vary the horizon over

which inflation is returned to target.

In contrast, P involves setting a pre-announcedpath for the price level and using the same policyinstrument (the short-term nominal interest rate)to affect prices by influencing aggregate demand.Tis key distinction between I and P means thatthe latter approach involves offsetting the impactof inflation shocks on the price level. Under P,inflation shocks can have only a temporary effect onthe price level.

Tis crucial difference is illustrated in Figure 1below. Under I, inflation is gradually brought backto its target rate after a positive inflation shock,leaving the price level to follow a permanently

higher path. Under P, inflation is temporarilybrought below the target rate, and the price levelgradually returns to its initial target.

Tis operational distinction is related to theacademic literature on optimal monetary policy regarding discretion versus commitment. In theclassic reference, Kydland and Prescott (1977) arguethat a policymaker can achieve superior outcomesby committing to a future course of action.5Bycorrecting for the impact of shocks on the pricelevel, P introduces history dependence intothe monetary policy framework: past shocks affectcurrent policy, as if the central bank committed toits future policy on the basis of those shocks.

Under I, the central bank needs only to take

current economic conditions into account: bygonesare bygones. Average inflation only will equal thetarget by chance (if positive and negative inflationshocks offset each other). Under P, averageinflation will be equal to its target rate in the longrun by design.

For his part, Woodford (2001, 2012) notes thathistorical dependence is a hallmark of most optimalpolicy rules when agents are forward-looking. Itis the commitment to offset the impact of shocks

on the price level that gives P its advantages as astabilization tool, as also argued in the next section.Meanwhile, Vestin (2006) shows that P cansubstitute for a central banks explicit commitmentto its future actions.6

The Advantages of PT

P can, under the conditions outlined below, lead

4 It is important to note that both P and I are compatible with positive trend inflation. Te pre-announced path for the

price level can have a positive slope.

5 See Dotsey (2008) for a nontechnical introduction.

6 Te central bank in Vestins model can do just as well by aiming at a price-level target without committing to its future

actions as it can by aiming to control inflation while committing to future policies.


6/20

4

to less volatile fluctuations in both inflation and theprice level compared to I. 7

Consider an unexpected increase in inflation,resulting from an increase in spending or an

increase in the cost of firms inputs.8Under I, thecentral bank would raise its interest rate in orderto reduce demand, which dampens price increasesand gradually brings inflation back down to itstarget. Under P, the central bank would committo reducing inflation below the target rate in themedium term in order to bring the price level backto its target. As a result, future inflation under Pis expected to be lower than under I. Knowingthis, firms that fix their prices for several periods

would not raise them as much since they know thatthe general price level will not be as high in themedium run.9Because expected future inflation islower under P, current inflation is lower as a result.

Tis P expectational bonus means thatthe central bank need not work as hard to fightinflation. It requires only smaller interest ratehikes to bring the economy back to its long-runequilibrium, with the result that demand decreasesless. If the economy is continually hit by shocks that

affect inflation, both inflation volatility and outputvolatility will be lower under P than under I.

Rational expectations and credibility are keyfactors in making P a superior tool for economicstabilization. Since lower expected future inflationmoderates firms current price increases, it thereforereduces current inflation. Te theory of rationalexpectations holds that individuals forecasts of

future inflation are correct on average and theirforecast errors are unrelated to available information.However, for this to be the case they must understandhow the economy works and how the central

banks commitment to return prices to their targetpath means lower future inflation than under I.Finally, individuals must believe in the central bankscommitment to returning the price level to its targetpath. Tese assumptions seem strong, but evidencediscussed below seems to support them.10

Advantages at the Zero Lower Bound

When its policy rate hits a lower bound of zero, a

central bank can no longer use the rate to directlyincrease aggregate demand. In such a situation,P allows monetary policy to be much moreexpansionary than it could be under I. As notedby Boivin (2009), By forcing higher inflationthan under I and, thus, a lower real interestrate, P enables the central bank to respondmore aggressively to a deflationary environment.Consider an economy in which the policy rateis expected to remain steady for some time, with

inflation well below its target rate. Since thenominal interest rate cannot go any lower, thecentral banks ability to stimulate spending dependson its ability to affect the real interest rate byaffecting inflation expectations.

Under I, inflation will be expected to remainlow for as long as the economy remains in itsdepressed state, with the interest rate at its lower

7 P has other potential advantages that we ignore here, including increasing the predictability of future asset values

when assets are denominated in nominal terms. See Ambler (2009) for a more detailed summary of some of these other

advantages. Boivin (2009) provides a detailed analysis of the beneficial effects of P on price-level predictability. See Parkin

(2009) for a similar explanation of Ps benefits as a stabilization tool.

8 Te argument is completely symmetric in the case of a negative shock.

9 Tis effect is built into the New Keynesian Phillips curve, which is a central part of the macroeconomic models used as

forecasting tools by many central banks and in much current research. Tis curve states that current inflation depends on

expected future inflation and on the current output gap.

10 Te only historical experience with P was in Sweden in the 1930s. For this reason, the evidence comes from model

simulations and from laboratory experiments.


7/20

5 Commentary 400

Figure 1: Inflation vs. Price-level Targeting

Source: Kahn (2009).

2%

2%

Price Level (Year zero = 100)Inflation (Percent)

Years

Years Years

Inflation (Percent) Price Level (Year zero = 100)

Price-Level Targeting

Years

4

3

2

1

0

112

110

108

106

104

102

1001 2 3 4 5

Inflation Targeting

4

3

2

1

0

112

110

108

106

104

102

100

1 2 3 4 5

1 2 3 4 5 1 2 3 4 5

bound. Firms and households expect that sooneror later inflation will increase towards its targetedlevel.11Under P, they know that inflation willhave to rise above its trend rate in order for theprice level to get back to its target path. However

long the horizon for the price level returning toits target path, average expected inflation over thathorizon will be equal to the target or trend level ofinflation.

11 Tis reasoning neglects the possibility of multiple long-run equilibria under I, as discussed by Benhabib, Schmitt Groh

and Uribe (2001). Tey show that a negative economic shock may propel the economy towards this low-inflation steady

state, where it can remain stuck for a very long time. Ambler and Lam (2013) demonstrate that a P regime does not suffer

from this problem.


8/20

6

Comparing the two situations, expected inflationwill be higher under P than under I over mosthorizons. Tis phenomenon is also present whenthe policy rate is not at the lower bound. Te

divergence in real interest rates between P and Iincreases dramatically the longer the interest rate isexpected to remain at the lower bound. In turn, realex ante interest rates at different horizons will belower under P.

Meanwhile, interest-sensitive components ofspending will be higher under a P regime. Byundertaking to return the price level to its targetpath, the central bank is promising to makemonetary policy less restrictive to compensate for

having been too restrictive (since it cannot reduceits policy rate below zero) while at the lower bound.Tis is a clear example of the history-dependentnature of monetary policy under P, followingWoodfords (2012) analysis.12

Once again, credibility and commitment areimportant to achieving an effective policy. Becauseinflation is likely to exceed the target rate fora period under P, the central bank could verywell be tempted to renege on its commitment to

inflate. If inflation is costly, the bank could increaseeconomic welfare by breaking its commitment tohigher inflation. Te longer the economy remainsstuck at the lower bound the higher inflation mustbe to get back to the target path, and the greater thetemptation to renege. Tis is a classic example ofthe so-called time-inconsistency problem. Kydlandand Prescott (1977) showed that optimal policyunder a P commitment is inherently subjectto time inconsistency. Reneging on promised

future policies can be beneficial, but if everyoneunderstands this, the announced policies will not bebelieved without a credible way to commit to them.

Expanding the Toolkit:Unconventional MonetaryPolicy

Because of Ps theoretical advantages, especiallyat the lower bound, P has garnered widespreadsupport among academic economists andpolicymakers.13In aNew York imesop-ed piece,Romer (2011) succinctly summarized the threemain alternative policy instruments available to a

central bank such as the Fed, including P:14

Te Fed could engage in much more aggressivequantitative easing, both in size and in scope, tofurther lower long-term interest rates and valueof the dollar. It could more effectively convey tomarkets its intentions for the funds rate, which

would also lower long-term rates. And it could set aprice-level target, which, unlike an inflation target,calls for Fed policy to take past years price changesinto account. Tat would lead the Fed to counteract

some of the extremely low inflation during therecession with a more expansionary policy and lowerreal rates for a while.

Forward guidance and quantitative easing(henceforth QE) were both tried by the Fed and theBank of Canada post-2008;15P was not. Te maindistinction between the first two and P is thatboth the former can be implemented in ways thatdo not hinder the central banks ability to exercise

12 Amano and Ambler (2009) establish that P is more effective than I both in keeping the central banks policy rate from

hitting the zero lower bound and in reducing the length of episodes in which the policy rate is at zero.

13 See for example Krugman (1998) in the context of Japan, as well as Bernanke (2003), Mankiw (2008), Hall and Woodford

(2009), Evans (2010) and Romer (2011).

14 See Williams (2011, 2012) for a more detailed discussion of quantitative easing and forward guidance.

15 Te Bank of Canada merely announced its willingness to engage in massive purchases of securities if the circumstances

warranted. It did not actually do so.


9/20

7 Commentary 400

discretion, whereas P is much less compatiblewith freedom to act.

Forward Guidance

Forward guidance involves announcing the futurepath of the central banks policy rate over a fairlylong horizon. When the Bank of Canada loweredits overnight rate to 0.25 percent in April 2009 (therate considered by the Bank to be its effective lowerbound), it also announced a path for the policy ratethrough the first half of 2010.16

With monetary policy now operating at the effectivelower bound for the overnight policy rate, it isappropriate to provide more explicit guidance thanis usual regarding its future path so as to influencerates at longer maturities. Conditional on theoutlook for inflation, the target overnight rate can beexpected to remain at its current level until the endof the second quarter of 2010 in order to achievethe inflation target. Te Bank will continue toprovide such guidance in its scheduled interest rateannouncements as long as the overnight rate is atthe effective lower bound.

By announcing its intentions to keep the short-term interest rate at its lower bound for an extendedperiod of time, thereby influencing expectationsfor the short-term nominal interest rate, the Bankalso affected interest rates at longer horizons.Its commitment to keep interest rates low wasconditional, and the main condition was theoutlook for inflation. However, the Bank didnot specify a precise inflation rate or an expectedinflation rate that would actually trigger a move

away from the lower bound.17

For its part, the Fed has used forward guidancerepeatedly since the 2008 financial crisis.18InDecember 2008, the Federal Open MarketCommittee (FOMC) announced that economic

conditions are likely to warrant exceptionally lowlevels of the federal funds rate for some time. Telength of the time period was not specified. TeFOMC used similar language in each of its periodicstatements until August 2011 when the time periodwas made more explicit, and it noted that economicconditions are likely to warrant exceptionally lowlevels for the federal funds rate at least throughmid-2013. In January 2012, the period wasextended until late 2014. More recently, on

September 13, 2012, it again extended the periodat least through mid-2015.

Te Fed announcements all stated that it waslikely that it would keep the federal funds rateat exceptionally low levels, without defining thecircumstances that would trigger a move awayfrom these low levels. Tis allowed the Fed to useits discretion to decide that current circumstanceswarranted an increase, irrespective of previouspronouncements. If forward guidance were instead

explicitly made contingent on economic conditions,it would make the Fed more predictable andcredible but would reduce its discretion.

QE

QE involves expanding the size of the central banksbalance sheet by engaging in large purchases ofassets on the open market. Open market operationsare a traditional part of a central banks toolkit.Yet QE as practised by the Fed after 2008 wasunconventional for three reasons.

16 See Bank of Canada (2009, 2009b).

17 Melino (2011) concludes that the Banks conditional commitment helped the Canadian economy to exit the recession more

quickly, but notes the vagueness of the notion of what constituted a substantive change in the inflation outlook.

18 See Tornton (2012).


10/20

8

1. Te increases in its balance sheet wereunprecedented in size.19

2. Te Fed expanded the types of assets purchased

to include longer-term government debt and alsoprivate-sector assets, both fixed-income securitiesand equities.

3. QE came to be associated with exit strategies,the idea that with economic recovery, increasinginflation and policy rates rising from their lowerbound, the massive increase in the Feds balancesheet would be unwound. Te exact trigger andthe final level and composition of the Fedsbalance have never been specified in detail, onceagain giving much leeway for discretion.

Te Fed aggressively expanded its balance sheet inthe wake of the Lehman Brothers bankruptcy inSeptember 2008. It has engaged in two subsequentQE rounds. Buying assets other than governmentsecurities and the extension of loans to non-bankinstitutions such as AIG, constituted an expansionof the Feds discretionary powers.

QE was accompanied by other measuresimplemented by the Fed alone or in conjunctionwith the US reasury. Te discretionary nature ofthese measures is highlighted by both White (2010)and Cochrane (2012). White judged that theFeds actions were not only discretionary but alsoquestioned whether they were all legal under theFederal Reserve Act. Cochrane (2012) judged thatthe exercise of such discretion should bring intoquestion the Feds independence from elected andaccountable officials.

Bank and other bailouts (Bear Stearns, AIG,the forced takeover of Merrill Lynch, etc.) also

constituted an unexpected transfer of wealth

from taxpayers to bank stakeholders (particularlysenior bondholders). Tis increased expectationsthat certain banks were too big to fail, raisingimportant issues of moral hazard and potentially

encouraging managers and shareholders to takerisks knowing that taxpayers would bear thedownside risks.20

Te overall effectiveness of the Feds QE policiesremains a matter for debate.21Tey may havedecreased yields on assets with longer maturitiesrelative to those with shorter maturities. Tey mayhave reduced deflationary pressures by increasingthe size of the monetary base, but evaluating themonetary bases impact on inflation and prices is

made difficult by the unprecedented decrease inthe velocity of circulation of base money duringthis period. Tis, in turn, may be related to theuncertainty of the post-tapering level of the Fedsbalance sheet, but such a view is also controversial.22

Te relative robustness of Canadas financialsector during the crisis, combined with the mild-ness of the Canadian recession, meant that theBank of Canada needed not travel very far downthe road of QE. After the first onset of the financial

crisis in 2007, it announced a term purchase andresale agreements program to provide liquidity tothe financial system by purchasing eligible securitiesto be resold at a fixed price from eligible financialinstitutions.

Yet the policy was limited in scope. In fact, theBank of Canadas balance sheet increased by about50 percent as a share of GDP from 2007 to 2009,before falling back to its previous level. Tis is incontrast to the Fed, whose balance sheet has more

than doubled as a share of GDP since 2007.23

It

19 See Chart 2 of Santor and Suchanek (2013).

20 Roberts (2010) contains a cogent discussion of the moral hazard problems that were exacerbated by the Feds bailouts.

21 See Santor and Suchanek (2013) for a summary of some of the literature.

22 See Woodford (2011) for one point of view.

23 See Santor and Suchanek (2013).


11/20

9 Commentary 400

is, of course, impossible to know what the Bankof Canada would have done in the face of a morehighly distressed financial sector, but in 2008 itannounced principles to guide its future liquidity

interventions. Tey were intended to limit the scopeand duration of intervention while minimizing itsdistortions and mitigating moral hazard.24Tis mayindicate a greater willingness on the part of theBank of Canada compared to the Fed to tie itsown hands.

Objections to PT

Canada

With its intensive research effort to study theadvantages of P, the Bank of Canada showedmore willingness ex ante to entertain the possibilityof P than any other central bank. Despite theseverity of the Great Recession in many countries,particularly the United States, Canadas was themildest in the past 30 years.25Canadas financialsystem and banks also weathered the recessionmuch better than their American counterparts. Te

Bank of Canada itself made the case that Canadasmonetary policy framework performed quite wellduring the crisis and that moving to a completelydifferent monetary framework was not warranted,especially given the uncertainty involved in such atransition.

Te Bank of Canadas own case against adoptingP is laid out most clearly in its background paperto the renewal of its Inflation-Control argetAgreement with the federal government (Bank

of Canada, 2011). Te document claims thatPs benefits in terms of economic welfare would

be modest, although it acknowledges that thesebenefits would be enhanced once the costs and risksof the nominal interest rate hitting its zero lowerbound are incorporated (p. 14). Te key phrase in

the Banks document is probably the following(p. 14): However, these models assume that agentsare forward-looking, fully conversant with theimplications of PL (price-level targeting) and trustpolicymakers to live up to their commitments. Onthe basis of its own studies and those of outsideacademics, the Bank doubted P would havesufficient credibility, especially immediately uponits implementation, and whether individuals wouldunderstand the workings of such a monetary policy

framework and form their inflationary expectationsaccordingly.

Te Bank reached this conclusion despite someof the evidence produced by its own researchprogram. Using experimental evidence, Amano,Engle-Warnick and Shukayev (2011) showed thatindividuals would, in fact, be quick to adapt to aregime switch from I to P and to base theirinflation forecasts on the knowledge that the pricelevel reverts to its target path.

Tere is other evidence to suggest that having tolearn about P when it is first implemented doesnot overturn its advantages. Gaspar, Smets andVestin (2007) used an adaptive learning model tosimulate the transition to P. Tey concluded thatP retains its advantage over I even if agentsmust gradually learn how the new regime works.Tey used the European Central Banks forecastingmodel, very similar in structure to the Bank ofCanadas oEM model, so their conclusions

would also apply in the Canadian case (subject tothe validity of the modelling framework).

24 See Longworth (2010) for details.

25 See Bank of Canada (2011).


12/20

1 0

Meanwhile, the Bank of Canadas study byKryvtsov, Shukayev and Ueberfeldt (2008) wassomewhat more pessimistic. Tey concludedthat a temporary loss in credibility following the

announcement of a transition to P could lead toshort-run costs that would not be made up bylater gains.

Te Bank of Canadas own bottom line concerningP, as expressed in its 2011 report, was a messageof prudence:

Given the current state of knowledge, the potentialbenefits of PL (Price level targeting) in increasinglong- term certainty about the price level andproviding greater short-term macroeconomicstability, relative to the current I framework, do notclearly outweigh the costs and risks associated withreal-world expectations and credibility falling shortof the model ideal.

Te Bank of Canada viewed the mildness of the2008-2009 recession as confirmation that its Iregime was delivering good results and a rationalefor not fixing an unbroken policy framework.Ragan (2011) echoes these arguments, noting that

P could be confusing because of the need forinflation expectations to be variable, in contrastto the simplicity of I where inflation at all butthe shortest horizons will be anchored to theinflation target.

Some authors have argued that the benefits ofCanada moving to P would be small because thehistorical performance of the price level has beenclose to that under a P regime. Figure 2 in Melino(2011) shows that Canadas CPI has drifted verylittle from constant 2 percent inflation betweenJanuary 1996 and November 2010, whether this isdue to good monetary policy or historical accident(the particular sequence of positive and negativeshocks to inflation in Canada over the period).

Te choice of a start date may also be importantin comparing the impact of an I-to-P shift.Boivin (2009, Figure 2) shows that projectingthe price level path forward from 1992 leads to a

substantial and persistent divergence of Canadasactual CPI from this path.

Although this is not explicit in its backgrounddocument (Bank of Canada, 2011), the Bank

of Canada has clearly been concerned aboutpreserving its discretionary powers. Governor MarkCarney (2009) made this quite clear:

Te design of monetary policy frameworks dependsin part on the trade-off between flexibility andcredibility. Tis, in turn, is a function of both theextent to which (inflexible) rules enhance credibilityand the ability of central banks to exercise thediscretion required to deploy any flexibility in acredible manner.

Te Bank of Canada believes that discretion is auseful feature of its monetary policy framework.Although it does not explicitly acknowledge this,Ps greater necessity for commitment inevitablyremoves much of this discretion. As noted above,Vestin (2006) shows that targeting the price levelinstead of inflation can substitute for commitment:it obliges the central bank to take account of pastrates of inflation rather than allowing it to ignore

past deviations from its target, just as if it hadexplicitly committed to its policy upon observingpast inflation.

Te theme of flexibility was also at the heart of a2012 Carney speech:

We did so because, in a complex and continuouslyevolving world that no one can predict withcertainty, policymakers need a robust framework;one that remains appropriate no matter thecircumstances. Inflation targeting is disciplined but

flexible. It allows central banks to deliver what isexpected while dealing with the unexpected.

Tis quote highlights what McCallum (2004) viewsas a crucial distinction between the way centralbankers view discretion and the way it is viewed byacademics:

From the academic point of view, by contrast, themain issue is not about flexibility or its absence.


13/20

1 1 Commentary 400

Instead, the central issue is whether monetarypolicy is conducted in a period-by-period fashion that is, as a sequence of unrelated decisions orinstead in a rule-based manner that views policy

as an ongoing process. o explore the nature ofthis distinction, let us suppose that in either casepolicy is conducted so as to be optimal or best in relation to current economic conditions. Tefirst way of proceeding is for the central bank torespond optimally to todays conditions, treating pastconditions (and expectations formed in the past) asunalterable and therefore irrelevant. Also, the centralbank recognizes that tomorrow it will do the same;it will optimize anew, treating todays conditionsand expectations as irrelevant for decisions taken

tomorrow. Tis is the standard way, developedby engineers and applied mathematicians, ofconducting optimal control analysis. It represents,to academic monetary economists, policymakingunder a regime of discretion.

Discretionary policymaking takes as irrelevanttodays expectations when making tomorrowsdecisions. Central bankers that value discretionwill not attempt to influence those expectationsby announcing that inflation could and should

be different from its long-run target rate in themedium run in order to unwind the effects ofprevious shocks on the price level. P is a policybased upon influencing todays expectations offuture inflation. Since the Bank of Canada hasnever followed what could be termed a rule-basedpolicy (from an academic point of view), it is notsurprising that it would be reluctant to undertake aradical change in its monetary policy framework.

Te Bank of Canada has shown concern about

inflation expectations, but its concern is limitedto making sure that inflation expectations inthe medium term are well-anchored and equalto its target, irrespective of current economicconditions. Te BanksMonetary Policy Reportscontain guidance on the Banks time horizonfor inflation to return to its target. However, itis impossible to verify ex post whether the timehorizon is realistic and credible since the return to

the target is contingent on an absence of furthereconomic shocks, a condition that is never realized.Committing to an unverifiable time horizon meansnot committing to anything concrete.

Because of Canadas positive historicalperformance under an I regime, the Bank ofCanada has acquired credibility over time. Inflationexpectations in Canada are strongly anchored at2 percent for all but the shortest horizons, whichindicates that markets think that the Bank ofCanada is able to hit its inflation target with onlytemporary deviations.

Meanwhile, the game theory literature showsthat reputation can act as a commitment device.

In other words, the Banks solid reputation may beenough to put a brake on its exercise of discretion.Tis may be a topic worthy of further study.

The US Case

In contrast to Canada, the US post-2008 sufferedits severest recession since the Great Depression.It was also of a relatively new type. Gorton (2010)characterized the interbank market panic as a 21st

century bank run, with banks and other financialinstitutions suddenly refusing to renew short-termfinancing to one another, as opposed to previousbank runs caused by depositors simultaneouslyattempting to withdraw funds from charteredbanks. Given the severity of the latest US crisis,it is not surprising that a large number ofpolicymakers and academics suggested P as apossible response, as noted in Section 3. Most of theproposals for moving toward P promoted it not asa new and permanent monetary policy frameworkbut rather as a policy contingent on the Feds targetfor the federal funds rate being stuck at the zerolower bound.

Te possibility of reneging was actually built intosome of the proposals to announce a target for theprice level. Evans (2010) proposed a contingentprice-level path to be implemented at the zerolower bound, but which would disappear once a


14/20

1 2

more normal policy rate was reestablished.He wrote:

I consider price-level targeting a policy option that

is only appropriate for the unusual situation of aliquidity trap. In more usual times, the Fed wouldaddress lower- than-desirable levels of employmentand inflation by adjusting the federal funds rate.However, as discussed earlier, we are currentlyconstrained from doing so by the zero bound.

Tis suggests that once the federal funds rate wasno longer at its zero bound, the targeted price-levelpath could be abandoned before the price levelregained its pre-announced path. Evanss proposal

seems to ignore the fact that all of the benefits fromP come from its impact on inflation expectations,that this impact hinges critically on credibility, andbuilding the possibility of abandoning the targetedpath into the initial announcement is a sure recipefor robbing the proposal of its beneficial effects.Tis argument is applicable more generally. Anyimplementation of P understood to be a possiblytemporary policy to address particular economiccircumstances would fail due a lack of credibility.

Evanss proposal also involved establishing aprice-level target path that retroactively correctedfor price-level drift. Since any price-level pathinvolves picking an initial level, he proposed atarget path that extended from December 2007.If his proposal had been adopted, it would haveundermined expectations that, under I, the centralbank would let bygones are bygones. Implementinga monetary policy framework whose success hingeson commitment by breaking a previous implicitpromise is not a good way to establish credibility.

In the US context, any policy that entails aperiod of inflation higher than the Feds announcedtarget of 2 percent26would also encounter problemsfrom the inflation hawks on the Feds Open

Market Committee. (Te Fed publishes the FOMCdeliberations, whereas the Bank of CanadasGoverning Council meets behind closed doors anddissenting views are masked by the consensus view

that accompanies the Banks announcement of itsovernight rate.)

Proposals to switch to P when the economy isat the zero lower bound promise higher inflationthan the target in the short to medium run, withoutalso committing to undoing positive shocks toinflation in the future. In this respect, inflationhawks may have a valid point. Teir point of view iswell summarized by Kocherlakota (2011):

Moreover, I believe that the FOMC could onlyhave systematically lowered the unemployment ratefurther by generating inflation rates over a multiyearperiod that were higher than its communicatedobjective of 2 percent. Such an outcome couldpotentially lead the public to lose faith in thecredibility of the FOMCs communicated objectiveand thereby increase the probability that the FOMC

would lose control of inflation. As I discussed earlier,this scenario would require a policy response that

would generate substantial losses of employment.

Hetzel (2012) contends that discretion helpsachieve consensus within a monetary policycommittee such as the FOMC:

Another political economy advantage of thelanguage of discretion is how the focus on individualpolicy actions facilitates the ability of the FOMCchairman to achieve consensus within the FOMC.

Te chairman would achieve consensus onlywith difficulty over an articulation of policy as asystematic set of procedures for responding to the

economy in a way designed to trade off betweenconflicting objectives. Te language of discretionallows the chairman to avoid divisive issues about

what variables the central bank controls and how itexercises that control.

26 As noted in the introduction, this target was first announced explicitly in January 2012. See Federal Reserve Board (2012).


15/20

1 3 Commentary 400

Remarks made by Ben Bernanke (2003b) whenhe was a governor of the Federal Reserve Bank ofNew York, soon after leaving academia, defined hisconcept of constrained discretion:

First, through its words and (especially) its actions,the central bank must establish a strong commitment to

keeping inflation low and stable. Second, subject tothe condition that inflation be kept low and stable,and to the extent possible given our uncertaintiesabout the structure of the economy and the effectsof policy, monetary policy should strive to limit cyclicalswings in resource utilization(emphasis in theoriginal).

Tis passage seems to be more about describing theobjective function of the central bank. He equatesunfettered discretion with the absence of anobjective function. He also equates commitmentwith following a rule independent of economicconditions such as Friedmans (1960) k-percentrule that specifies a constant of base money growth.Tis sets up a strawman version of commitmentthat is more in line with the way central bankers, asopposed to academics, view commitment, followingMcCallums (2004) distinction outlined in theprevious sub-section. For Bernanke, having anobjective function is the only constraint on policythat is desirable, and the central bank should beable to respond to current economic conditions,unconstrained by past promises or by the necessityto offset the effects of past shocks.27

Conclusions

P has convincing advantages, especially as a tool

for avoiding the worst consequences of economicdownturns. Switching to a new monetary policyframework would necessitate a public learning

curve, but this would be likely to diminish ratherthan overturn Ps advantages. Nevertheless,central banks (especially the Bank of Canada) havebeen intrigued by Ps potential, but had practical

reasons for not adopting it during the last recession.Suddenly promising to boost inflation above targetin a recession without having demonstrated awillingness to do the opposite in booms with higherinflation would rob such promises of much of theirneeded credibility.

However, the main reason that central bankshave not experimented with P is that they aretoo wedded to the ability to exercise discretionin the conduct of monetary policy. P acts as a

substitute for commitment in the sense of Kydlandand Prescott (1977). I is a regime that allowsbygones to be forgotten and allows the central bankto decide on its optimal policy without regard topast economic conditions. Central bankers haveresisted strongly academic studies demonstratingthe superiority of rules-based approaches tomonetary policy as imposing too much rigidityon the monetary policy process and robbing themof the flexibility to react to circumstances not

accounted for in their forecasting models or inmechanical rules.28Te financial crisis was a specialcircumstance par excellence.

P has disappeared recently from academicdiscussions and from central bank workingpaper series, and it has all but disappeared fromdiscussions on the blogosphere. Te main alternativemonetary policy framework discussed activelytoday is nominal GDP targeting (NGDP), whichreplaces a target price level path with a target path

for nominal income.29

Nominal income becomesthe long-run anchor for monetary policy ratherthan the price level.

27 Carney (2013) also seems to conflate discretion with the absence of a clearly defined objective function.

28 Tere are no doubt political economy considerations involved as well. Discretion means more power and less accountability.

29 See Christensen (2011), Sumner (2011) and Beckworth (2010) for summaries of the ideas behind NGDP.


16/20

1 4

30 NGDP has the added advantage of ensuring an appropriate response to a supply shock that moves real output and the

price level in opposite directions. A P rule would lead central banks to expansionary monetary policy in booms and

contractionary policy in downturns, while NGDP would not force the central bank to magnify changes in real output to

offset the changes in prices.

Under NGDP, deviations of nominal incomefrom the target path are the only measure of theneed for tightening or loosening, so it is potentiallysimpler.30In contrast, under P the need for

tightening or loosening of monetary policy ismeasured not only by deviations of the price levelfrom target but also on measures of the output gap,which is not directly observable. NGDP witha level target is a framework in which bygones

are not bygones and deviations from target mustbe corrected. In this respect, it is a rules-basedapproach to monetary policy similar to P. imewill tell if NGDP gains enough traction to

overcome central bankers reluctance to acceptinglimits on their discretion. Te active discussionconcerning NGDP will keep alive the debate overrules versus discretion.


17/20

1 5 Commentary 400

References

Amano, Robert, and Steve Ambler. 2009. Price-Levelargeting and the Zero Lower Bound. Draft, Bank

of Canada and UQAM.Amano, Robert, Jim Engle-Warnick, and Malik

Shukayev. 2011. Price-Level argeting andInflation Expectations: Experimental Evidence.

Working paper 1118. Ottawa. Bank of Canada.

Ambler, Steve. 2009. Price-Level argeting andStabilization Policy: A Survey.Journal of EconomicSurveys, 23: 974997. Spring.

Ambler, Steve, and Jean-Paul Lam. 2013. Inflationargeting, Price-Level argeting, the Zero Lower

Bound, and Indeterminacy. Draft. Universit duQubec Montral.

Bank of Canada. 2006.Inflation arget RenewalAgreement Background Document. Ottawa.

. 2009. Bank of Canada LowersOvernight Rate arget by 1/4 Percentage Pointto 1/4 Percent and, Conditional on the InflationOutlook, Commits to Hold Current Policy Rateuntil the End of the Second Quarter of 2010. Pressrelease. April 21. Available at:

http://www.bankofcanada.ca/2009/04/publications/press-releases/fad-press-release-2009-04-21.

. 2009b.Monetary Policy Report. April.

. 2011. Renewal of the Inflation-Controlarget: Background Information November 2011.

Beckworth, David. 2010. arget the Cause not theSymptom.Macro and Other Market Musings(blog).

Available at:http://macromarketmusings.blogspot.com/2010/03/target-cause-not-symptom.html.

Benhabib, Jess, Stephanie Schmitt-Groh and MartnUribe. 2001. Te Perils of aylor Rules.Journal of

Economic Teory, 96: 4069. January.

Bernanke, Ben. 2003. Some Toughts on MonetaryPolicy in Japan. Washington, D.C. Federal ReserveBoard. May 31.

. 2003b. Constrained Discretion andMonetary Policy. Remarks by Governor Bernanke

Before the Money Marketeers of New YorkUniversity. February 3. Available at:

www.federalreserve.gov/boarddocs/speeches/2003/20030203/default.htm.

Boivin, Jean. 2009. Getting it Right When You Might BeWrong: Te Choice Between Price-Level and Inflation

argeting. Commentary 297. oronto: C.D. HoweInstitute. September.

Carney, Mark. 2009. Some Considerations on UsingMonetary Policy to Stabilize Economic Activity.

Remarks delivered at Jackson Hole Symposium.August 22. Available at:http://www.bankofcanada.ca/2009/08/publications/speeches/some-considerations-using-monetary-policy.

. 2012. A Monetary Policy Frameworkfor All Seasons. U.S. Monetary Policy Forum. New

York City. February 24. Available at:http://www.bankofcanada.ca/2012/02/publications/speeches/monetary-policy-framework-all-seasons.

. 2013. Monetary Policy After the Fall.Eric J. Haonson Memorial Lecture. University of

Alberta. Edmonton. May 1. Available at:http://www.bankofcanada.ca/2013/05/publications/speeches/monetary-policy-after-the-fall.

Christensen, Lars. 2011. Market Monetarism: TeSecond Monetarist Counter-Revolution.Te Market

Monetarist(blog). September 13. Available at:http://thefaintofheart.files.wordpress.com/2011/09/market-monetarism-13092011.pdf.

Cochrane, John. 2012. Te Federal Reserve: FromCentral Bank to Central Planner. Wall Street

Journal. August 31. Available at:http://online.wsj.com/article/SB10000872396390444812704577609384030304936.html.

Dotsey, Michael. 2008. Commitment Versus DiscretionIn Monetary Policy. Federal Reserve Bank ofPhiladelphia Business Review, Q4, 18.


18/20

1 6

Evans, Charles. 2010. Monetary Policy: ools for Non-raditional imes. Federal Reserve Bank of ChicagoAnnual Report, 713.

Federal Reserve Board. 2012. Press Release: Longer

Run Goals and Policy Strategy. January 25.Available at:http://www.federalreserve.gov/newsevents/press/monetary/20120125c.htm.

Gaspar, Vtor, Frank Smets and David Vestin. 2007. Isime Ripe for Price Level Path Stability? Workingpaper 0818. European Central Bank.

Gorton, Gary. 2010. Slapped by the Invisible Hand: TePanic of 2007. Oxford University Press.

Hall, Robert and Susan Woodford. 2009. Te FedNeeds to Make a Policy Statement. Financial Crisisand Recession(blog). Available at:http://woodwardhall.wordpress.com/2009/04/13/the-fed-needs-to-make-a-policy-statement.

Hetzel, Robert. 2012. Central Bank Accountability andIndependence: Are Tey Inconsistent?Journal of

Macroeconomics, 34, 616625. September.

Kahn, George. 2009. Beyond Inflation argeting:Should Central Banks arget the Price Level?Federal Reserve Bank of Kansas City Economic

Review.Tird quarter, 3564.

Kocherlakota, Narayana. 2011. Making MonetaryPolicy. Speech in Sidney, Montana. October 13.

Available at:http://www.minneapolisfed.org/news_events/pres/speech_display.cfm?id=4768.

Krugman, Paul. 1998. Its Baaack: Japans Slump andthe Return of the Liquidity rap. Brookings Paperson Economic Activity, 29, 137206.

Kryvtsov, Oleksiy, Malik Shukayev, and Alexander

Ueberfeldt. 2008. Adopting Price-Level argetingunder Imperfect Credibility. Working Paper20083. Ottawa: Bank of Canada.

Kydland, Finn, and Edward Prescott. 1977. RulesRather than Discretion: Te Inconsistency ofOptimal Plans.Journal of Political Economy85,473491. June.

Lim, Guay. 2008. Inflation argeting.Insights.University of Melbourne. November. Available at:http://insights.unimelb.edu.au/vol4/11_Lim.html.

Longworth, David. 2010. Bank of Canada Liquidity

Facilities Past, Present, and Future. Remarks atthe C.D. Howe Institute. oronto. February 17.

Mankiw, Gregory. 2008. Te Next Round ofAmmunition.Greg Mankiws Blog. December 16.Available at:http://gregmankiw.blogspot.ca/2008/12/next-round-of-ammunition.html.

Melino, Angelo. 2011.Moving Monetary Policy Forward:Why Small Steps and a Lower Inflation arget

Make Sense for the Bank of Canada. Commentary 319.

oronto: C.D. Howe Institute. January.

McCallum, Bennett. 2004. Misconceptions RegardingRules versus Discretion for Monetary Policy. Cato

Journal, 23, 365372. Winter.

Parkin, Michael. 2009. What is the Ideal Monetary PolicyRegime? Improving the Bank of Canadas Inflation-

targeting Program. Commentary 279. oronto: C.D.Howe Institute. January.

Ragan, Christopher. 2011. Te Roads Not aken:Why the Bank of Canada Stayed with Inflation

argeting. E-Brief. oronto: C.D. Howe Institute.November 9.

Roberts, Russell. 2010. Gambling with other PeoplesMoney: How Perverted Incentives Caused theFinancial Crisis. Mercatus Center. George MasonUniversity. April 28. Available at:http://mercatus.org/publication/gambling-other-peoples-money.

Romer, Christina. 2011. Dear Ben: Its ime for yourVolcker Moment.New York imes. October 29.

Available at:http://www.nytimes.com/2011/10/30/business/economy/ben-bernanke-needs-a-volcker-moment.html?pagewanted= all&_r=0.


19/20

Support the Institute

For more information on supporting the C.D. Howe Institutes vital policy work, through charitable giving ormembership, please go to www.cdhowe.org or call 416-865-1904. Learn more about the Institutes activities andhow to make a donation at the same time. You will receive a tax receipt for your gift.

A Reputation for Independent, Nonpartisan Rese arch

Te C.D. Howe Institutes reputation for independent, reasoned and relevant public policy research of thehighest quality is its chief asset, and underpins the credibility and effectiveness of its work. Independence andnonpartisanship are core Institute values that inform its approach to research, guide the actions of its professionalstaff and limit the types of financial contributions that the Institute will accept.

For our full Independence and Nonpartisanship Policy go to www.cdhowe.org.

Recent C.D. Howe Institute Publications

January 2014 Laurin, Alexandre, and William B.P. Robson. Equipping Canadians for Success: A Shadow Budgetfor 2014.C.D. Howe Institute Commentary 399.

January 2014 Dachis, Benjamin. Full Trottle: Reforming Canadas Aviation Policy.C.D. Howe InstituteCommentary 398.

January 2014 Dachis, Benjamin, and William B.P. Robson. Baffling Budgets: Canadas Cities Need BetterFinancial Reporting.C.D. Howe Institute Commentary 397.

January 2014 Schwanen, Daniel. radable Services: Canadas Overlooked Success Story. C.D. Howe InstituteE-Brief.

December 2013 Herman, Lawrence L. Who Else Benefits from CEA? Some Implications of Most-FavouredNation reatment. C.D. Howe Institute E-Brief.

December 2013 Blomqvist, ke, Colin Busby, and Don Husereau. Capturing Value from New Health echnologies inLean imes.C.D. Howe Institute Commentary 396.

December 2013 Guay, Richard, and Laurence Allaire Jean. Long-erm Returns: A Reality Check for Pension Funds

and Retirement Savings.C.D. Howe Institute Commentary 395.November 2013 Richards, John. Diplomacy, rade and Aid: Searching for Synergies.C.D. Howe Institute

Commentary 394.

November 2013 Blomqvist, ke, Boris Kralj, and Jasmin Kantarevic. Accountability and Access to Medical Care:Lessons from the Use of Capitation Payments in Ontario. C.D. Howe Institute E-Brief.

November 2013 Neill, Christine. What you Dont Know Cant Help You: Lessons of Behavioural Economics forax-Based Student Aid. C.D. Howe Institute Commentary 393.

November 2013 Dachis, Benjamin, and William B.P. Robson. Equipping Canadian Workers: BusinessInvestment Loses a Step against Competitors Abroad. C.D. Howe Institute E-Brief.

October 2013 Blomqvist, ke, and Colin Busby. Paying Hospital-Based Doctors: Fee for Whose Service?C.D. Howe Institute Commentary 392.


20/20

C.D.HOWE

Institute

67YongeStreet,Suite300,

Toronto,Ontario

M5E1J8

Commentary 400 Plt

Documents

Transcript of Commentary 400 Plt