Allan Meltzer and the Search for a Nominal Anchor 1

Transcription

1 Allan Melzer and he Search for a Nominal Anchor 1 James Bullard Presiden and CEO, Federal Reserve Bank of S. Louis Melzer s Conribuions o Moneary Economics and Public Policy Philadelphia, Pa. Jan. 4, 2018 Inroducion Allan Melzer was an ousanding moneary economis wih a long and disinguished career. 2 He has been an absolue fixure on he cenral banking and moneary economics scene during my enire career, so much so ha I have a hard ime picuring ha scene wihou him. He was also a grea friend of he S. Louis Fed. Melzer published seven papers in he S. Louis Fed s Review. The firs was published in 1969 and was iled Conrolling Money. 3 The las was published in 2005 and was a precursor o Volume 2 of his monumenal A Hisory of he Federal Reserve. 4 I will use his las Review publicaion as a saring poin for my commens. The paper, iled Origins of he Grea Inflaion, discusses he run-up o he exraordinary inflaion of he 1970s and early 1980s in he U.S. from a policymaker perspecive. Melzer presens in he syle laer carried on in he Hisory. He quoes he views of individual Federal Open Marke Commiee (FOMC) members, along wih hose of he saff, ha describe wha hey hough hey were doing in adoping paricular moneary policy acions a paricular imes. He presens his maerial almos deadpan, wih lile judgmen, o ry o discern wha he proagoniss were hinking as he Grea Inflaion developed. And indeed, hey were hinking and saying a wide variey of hings. As he narraive 1 Any opinions expressed here are my own and do no necessarily reflec hose of he Federal Open Marke Commiee. 2 Taylor (2017). 3 Melzer (1969). 4 Melzer (2003, 2005, 2009a, 2009b). 1

2 progresses, however, i becomes quie apparen ha here was lile in he way of a clear framework for moneary policy during hese years. Melzer saes ha he Commiee had no common baseline even for fundamenal quesions, such as he causes of inflaion. Ideas for he provision of a nominal anchor cenered on keeping fiscal deficis low, closely regulaing he growh of he money supply, or keeping ineres raes sufficienly high. None of hese were widely acceped, and all had deracors. Some FOMC members eschewed macroeconomic heory alogeher. He noes ha hen-board Governor Sherman Maisel ofen exhored he Commiee o spend more ime on he developmen of a coheren model ha could be used o beer guide decision-making, bu wih lile effec. As a resul of he heoreical incoherence, acual decision-making was perhaps more open o influence by oher facors, such as domesic poliics, which Melzer discusses in some deail. From his siuaion some five decades ago as described by Melzer, we come o he presen. Is he siuaion oday very differen? The debaes in he moneary heory lieraure and he fiscal heory of he price level lieraure roll onward. Cerainly, much progress has been made, bu complicaed argumens abound and clear resoluion seems disan. Ye oday, despie he remaining heoreical incoherence, inflaion conrol has been in place in he U.S. for abou wo decades and his despie he global financial crisis. If here is a problem wih inflaion oday, i is because inflaion has been lower han promised, no higher. How is i ha a nominal anchor for he U.S. has been found wihou clear resoluion o he puzzles posed in he academic lieraure? Inflaion Targeing The answer o his quesion is surely inflaion argeing. Inflaion argeing does no provide a comprehensive accoun of inflaion dynamics or of he mehods used o aain a given inflaion goal over he medium erm. As implemened in he U.S., i is really only a credible saemen of a goal and a promise o ry o achieve he goal in he medium erm via all means a he disposal of he FOMC. Tha is no enough o saisfy any modern noion of wha consiues a macroeconomic heory. Neverheless, i has succeeded in providing he U.S. wih he nominal anchor ha so eluded us during he 1970s and early 1980s. Inflaion argeing came upon he global cenral banking scene in he 1990s. Early adopers included he Reserve Bank of New Zealand and he Bank of England, bu many oher cenral banks around he world have followed since ha ime. The European Cenral Bank was conceived and esablished wih a formal inflaion arge. The Bank of Japan and he Federal Reserve did no join he club formally unil fairly recenly, wih he U.S. adoping an inflaion 2

3 arge in January Neverheless, in my opinion he U.S. has followed a de faco inflaion argeing policy from 1995 onwards. As of ha ime, U.S. inflaion had declined o abou 2 percen and has no deviaed subsanially from ha arge over he las 22 years, a leas no on he scale observed during he 1970s and early 1980s. U.S. policymakers were well aware of he rends in he inernaional policy debae oward inflaion argeing during his era, and many members of he Commiee were sympaheic o he argumens supporing inflaion argeing. Generally speaking, he evidence from around he globe is ha inflaion argeing has been successful in hose counries ha have implemened i. Inflaion raes are generally lower oday han in he pre-inflaion argeing era, and inflaion raes are ofen mainained relaively close o he announced inflaion arge. The variance of inflaion has generally been lower han in he pre-inflaion argeing era. In addiion, he variance of inflaion expecaions has also declined, generally speaking, because financial marke paricipans have ended o view many cenral bank inflaion arges as relaively credible commimens. These developmens are all in line wih wha one migh expec based on heories ha compare a noisy and purely discreionary moneary policy like he one Melzer described for he U.S. during he run-up o he Grea Inflaion, as opposed o a policy based more on a commimen by he cenral bank as o wha i inends o achieve. Wha would Allan Melzer say if he were here? My reading of his views is ha he would sress ha he ac of naming a credible inflaion arge recognizes some fundamenal ruhs ruhs ha were no recognized in he run-up o he Grea Inflaion. In paricular, i assigns he responsibiliy for he inflaion rae o he cenral bank. I srongly suggess ha inflaion is a moneary phenomenon. Bu i sops shor of requiring an implemener of he policy o commi o any paricular macroeconomic heory beyond hese preceps. I would add my own view a his poin, wihou rying o implicae Allan Melzer, as I am no sure ha he would sress i as much as I do. My poin is his: The ac of naming an explici inflaion arge recognizes he imporance of macroeconomic expecaions and fuure policy expecaions as being paramoun in he moneary policy arena. I is he expecaions of fuure policy ha can be unruly and lead o changes in economic behavior ha can, in urn, change he naure of he macroeconomic equilibrium. If expecaions are he problem, hen a naural way o help ie hem down is o credibly commi o an inflaion arge. This approach has worked admirably during he las wo decades. One could ie privae secor expecaions down sill furher by credibly commiing o a 3

4 moneary policy rule, such as a Taylor-ype rule. 5 The Taylor-ype rule has he inflaion arge as a criical componen, bu also specifies how he cenral bank should respond o various shocks o he economy. Would here be addiional gains, on op of hose already achieved hrough he naming of a credible inflaion arge iself, by a cenral bank commimen o such a rule? Allan Melzer hough so, and esified o ha effec before Congress in Ohers have been opposed, including my long-ime colleague and now ougoing Fed Chair Jane Yellen. 7 Two Indicaions of Moneary Policy Qualiy How can we decide if here are addiional gains ha would accrue o he U.S. economy via moneary policymakers more explicily adoping a Taylor-ype policy rule? Much has been wrien on his issue. Some of he issues addressed in he lieraure include he idea ha he Taylor-ype policy rule is a more model-dependen objec and ha is policy prescripions will be valid only in cerain model environmens. A very real issue, herefore, is he subsanial model uncerainy ha characerizes oday s policy landscape. Bu insead of rehashing hese argumens, I wish o go in a differen direcion in he remainder of my remarks. 8 I wan o urn now o make an assessmen of wheher inflaion argeing, as implemened implicily in he U.S. since 1995 (and explicily since 2012), has led U.S. policymakers o adop somehing we can view as close o opimal moneary policy. If recen moneary policy can be viewed as close o opimal, hen aemps o furher pin down expecaions of fuure policy acions may be less desirable. If recen moneary policy is viewed as less close o opimal, hen furher moneary policy commimen may confer imporan benefis o he economy. I will proceed by considering wo examples. In hese examples, he evidence on wheher recen moneary policy is close o opimal or no is somewha mixed, so he resuls here are broadly inconclusive. Neverheless, I hink i is insrucive o work hrough hese examples. To make an assessmen of his ype requires a model. Since here are many models o choose from, we could simply sop here and say we do no know. Bu in he spiri of rying o undersand a lile more abou he effecs of inflaion argeing, I will use a very simple version of a New Keynesian model. Since mos of you know he ideas behind his model very well, I 5 Taylor (1993). 6 Melzer (2015). 7 Yellen (2015). 8 For more on my views concerning he Taylor rule, see Bullard (2017). 4

5 have relegaed he deails o an appendix. The Case of he Disappearing Phillips Curve The firs example is he case of he disappearing Phillips curve. Here we will begin wih he empirical evidence, which is nealy summarized in Figure 1, adaped from he laes annual repor of he Bank for Inernaional Selemens (BIS). 9 The figure shows he coefficien on a measure of resource slack (unemploymen) in a regression of price inflaion on resource uilizaion, using he auhors preferred specificaion. The coefficien is esimaed repeaedly in rolling 15-year samples, and he poin esimaes, along wih 90 percen confidence bands, are ploed in he figure. The sample runs from he 1980s o he presen. The daa are for a panel of G-7 economies, and he poin esimae is a weighed average across economies. The main idea of he figure is ha he slope of he Phillips curve was once negaive in he 1980s bu has been drifing oward zero in he inflaion argeing era since The coefficien has no been differen from zero in recen years hence he disappearing Phillips curve ha has been widely discussed in financial markes and in moneary policy circles. The empirical phenomenon documened in Figure 1 can be relaed o he idea ha moneary auhoriies have moved closer o implemening opimal moneary policy during he inflaion argeing era. 10 The deails of his argumen can be found in he appendix, bu I will describe he approach here. I begin wih a sandard, wo-equaion, linearized New Keynesian model. I hen assume ha he moneary policymaker wishes o sabilize a quadraic funcion of inflaion gaps and oupu gaps over an infinie horizon. I give he policymaker a Taylor-ype linear feedback policy rule o work wih. I allow he policymaker o choose jus one parameer in his feedback rule in order o minimize he loss funcion. Tha single parameer is he coefficien on he inflaion gap. We could hink of his as represening he weigh placed on inflaion sabilizaion versus oupu sabilizaion in he Taylor-ype rule. In his simple exercise, he soluion o he policymaker problem is o se he value of he coefficien on he inflaion gap o a very large value echnically, infiniy. In his siuaion, opimal moneary policy would call for very low olerance of deviaions of inflaion from arge. The cenral bank reacs very aggressively o keep inflaion under conrol. In his simple model, we can also wrie ou explicily he value of he regression coefficien in a 9 Bank for Inernaional Selemens (2017). Also see Blanchard (2016). 10 Gillizer and Simon (2015) relae he flaening of he Phillips curve in Ausralia o inflaion argeing. Blanchard (2017) poins o inflaion argeing as one of he possible explanaions for he disappearing Phillips curve. 5

6 regression of he inflaion gap on he oupu gap, which is he heoreical counerpar of he slope of he Phillips curve in he BIS regression of Figure 1. The Taylor-ype rule coefficien on he inflaion gap appears in his formula. As his value ends o infiniy he policymaker is following somehing closer and closer o opimal policy and a he same ime he slope of he Phillips curve is ending oward zero. This resul is one way o sae he idea ha cenral banks have become beer and beer a inflaion argeing, and ha his success has driven he Phillips curve slope o zero. The empirical evidence in Figure 1 can herefore be inerpreed as a signaure of opimal moneary policy in observable daa. Figure 1 is saying, in effec, ha policymakers have already jumped on he Taylor rule bandwagon during he las wo decades. Price Level Targeing The second example is price level argeing. I have so far argued ha inflaion argeing has conferred considerable benefis on he economy. And ye, wihin he sandard New Keynesian model, opimal moneary policy is ofen characerized as price level argeing or is close cousin, nominal income argeing. In recen years, FOMC members (including me) have discussed boh price level and nominal income argeing as a possible fuure of U.S. moneary policy. A move in his direcion would require considerable debae and reflecion, bu migh also confer subsanial benefis o he U.S. economy if i could be implemened effecively. If we simply ake as given ha price level argeing is opimal policy wihin he New Keynesian consruc, a signaure in he U.S. daa of opimal moneary policy would be wheher he price level in he U.S. follows a prescribed price level pah. If i does, hen he FOMC has been de faco price level argeing even if he Commiee has no officially said ha i has been doing so. The hallmark of price level argeing is ha periods of below-arge inflaion are averaged ou wih oher periods of above-arge inflaion in such a way ha he economy remains on a price level pah consisen wih a given inflaion rae. In he price level argeing world, he saring dae maers. I have already argued ha 1995 is he poin a which he Volcker-era inflaion sabilizaion came o full fruiion, and ha from ha poin onward, he FOMC aemped o mainain a 2 percen inflaion arge. Bu did he Commiee do even more, implicily aemping o keep he U.S. economy on a 2 percen price level pah? In he fall of 2012, I argued ha he FOMC had kep he U.S. on a 2 percen price level pah since 1995, and ha his was an ousanding achievemen given he global financial crisis during 6

7 he inervening years. 11 Again, a leas as of 2012, he acual implemenaion of U.S. moneary policy could be viewed as opimal. By iself, his would sugges ha here would be lile o be gained from addiional moneary policy commimens by he cenral bank. However, as Figure 2 illusraes, in he las five years, he U.S. has fallen off he 2 percen price level pah esablished in The deviaion from he pah is now fairly subsanial, abou 4.6 percen on he low side. Recen FOMC forecass do no seem o anicipae enough inflaion o reurn he economy o he 2 percen pah. Therefore, we will have o conclude ha moneary policy has no been opimal from his perspecive. Summary and Conclusion I have ruminaed on Allan Melzer s excellen career by casing i in erms of a search for a nominal anchor for he U.S. I ook as a saring poin Melzer s las aricle for he S. Louis Fed s Review, Origins of he Grea Inflaion. Tha aricle, a precursor o he much larger work, A Hisory of he Federal Reserve, Volume 2, suggess ha during he period of he run-up o he Grea Inflaion, here was no coheren moneary policy framework. This lack of coherence may have lef moneary policy suscepible o influences oher han hose from moneary science, such as poliics. Today, some five decades laer, here remains a grea deal of heoreical and empirical uncerainy abou he effecs of moneary policy on he economy. And ye, despie his uncerainy, he U.S. and many oher counries have been able o achieve low and sable inflaion and, if anyhing, have faced a problem of inflaion being lower han promised, insead of higher han promised as in he 1970s and early 1980s. How was he nominal anchor found, and wha was i? My answer is ha i has been inflaion argeing, eiher praciced implicily or explicily. This may no have been he answer many economiss were expecing during he earlier porions of he poswar era ideas hen revolved around consan money growh, low fiscal deficis or he level of shor-erm nominal ineres raes. Inflaion argeing has worked well because i deals more direcly wih he coordinaion of macroeconomic expecaions han oher approaches. By commiing o an inflaion arge, inflaion has generally been kep lower and less variable, and inflaion expecaions have also been less variable. This has been a major achievemen of U.S. moneary policy, and one o which Allan Melzer made an ousized conribuion. If he naming of a credible inflaion arge coordinaes expecaions and helps inform he 11 Bullard (2012a, 2012b). 7

8 macroeconomic equilibrium, hen perhaps furher coordinaion can be achieved by being even more explici abou he fuure acions of moneary policymakers. One mehod of doing his would be for he cenral bank o commi o using a Taylor-ype moneary policy rule. Wheher his would confer added benefis or no would depend on wheher one hinks here are in fac addiional gains o be had for he economy by more ighly buoning down fuure expecaions of moneary policy. Such quesions canno really be answered wihou he assisance of a macroeconomic model. Bu if we go ahead wih a popular and simple version of he New Keynesian model, we can consider wo examples of wha opimal policy would look like and how i compares o acual moneary policy oucomes during he inflaion argeing era. The disappearing Phillips curve example suggess ha acual U.S. moneary policy has ended o be closer and closer o opimal in he las wo decades, and ha his has, in fac, killed off he Phillips curve correlaion previously apparen in he daa. The price level argeing example suggess U.S. moneary policy was close o opimal beween 1995 and 2012, bu has since fallen away somewha. Wheher we can do beer han hese wo examples sugges, or wheher he available gains have largely accrued already via he commimen o inflaion argeing implicily made in he U.S. in he 1990s, remains an imporan quesion for fuure research. I am sure ha research will coninue o be influenced by he enormous conribuions of Allan Melzer. 8

9 References Bank for Inernaional Selemens, 87 h Annual Repor. Basel, Swizerland, June 25, 2017; hps:// Blanchard, Olivier. The US Phillips Curve: Back o he 60s? Peerson Insiue for Inernaional Economics Policy Brief PB16-1, January 2016; hps://piie.com/publicaions/pb/pb16-1.pdf. Blanchard, Olivier. Should We Rejec he Naural Rae Hypohesis? Peerson Insiue for Inernaional Economics Working Paper 17-14, November 2017; hps://piie.com/sysem/files/documens/wp17-14.pdf. Bullard, James. A Singular Achievemen of Recen Moneary Policy. Invied lecure delivered a he Universiy of Nore Dame, Sep. 20, 2012a; hps:// epember2012final.pdf?la=en. Bullard, James. Price Level Targeing: The Fed Has I Abou Righ. Remarks delivered a he Economic Club of Memphis, Oc. 4, 2012b; hps:// fmemphisoc42012final.pdf?la=en. Bullard, James. The Policy Rule Debae: A Simpler Soluion. Presiden s Message in he Federal Reserve Bank of S. Louis Regional Economis, Firs Quarer 2017, p. 3; hps:// Gillizer, Chrisian and Simon, John. Inflaion Targeing: A Vicim of Is Own Success. Inernaional Journal of Cenral Banking, Sepember 2015, 11(Supplemen 1), pp ; hp:// Melzer, Allan H. Conrolling Money. Federal Reserve Bank of S. Louis Review, May 1969, pp ; hps://files.slouisfed.org/files/hdocs/publicaions/review/69/05/conrolling_may196 9.pdf. Melzer, Allan H. A Hisory of he Federal Reserve, Volume 1: Chicago, IL: Universiy of Chicago Press, Melzer, Allan H. Origins of he Grea Inflaion. Federal Reserve Bank of S. Louis Review, 9

10 March/April 2005, 87(2, Par 2), pp ; hps://files.slouisfed.org/files/hdocs/publicaions/review/05/03/par2/melzer.pdf. Melzer, Allan H. A Hisory of he Federal Reserve, Volume 2, Book 1: Chicago, IL: Universiy of Chicago Press, 2009a. Melzer, Allan H. A Hisory of he Federal Reserve, Volume 2, Book 2: Chicago, IL: Universiy of Chicago Press, 2009b. Melzer, Allan H. Tesimony before he U.S. Senae Commiee on Banking, Housing, and Urban Affairs hearing on Federal Reserve Accounabiliy and Reform, March 3, 2015; hps:// B4C4-0524B9ECDAFD. Taylor, John B. Discreion versus Policy Rules in Pracice. Carnegie-Rocheser Series on Public Policy, December 1993, 39(1), pp ; hps://doi.org/ / (93)90009-l. Taylor, John B. Making he Rules and Breaking he Mould (Allan Melzer: ). Cenral Banking, aricle published on June 19, 2017; hps:// banks/economics/ /making-he-rules-and-breaking-he-mould-allan-melzer Woodford, Michael. Ineres and Prices. Princeon, NJ: Princeon Universiy Press, Yellen, Jane L. Leer from Chair Yellen o House Speaker Ryan and House Minoriy Leader Pelosi. Nov. 16, 2015; hps:// pdf. 10

11 Appendix Consider he linearized equaions of he sandard New Keynesian model: 12 where y is he oupu gap, 1 y = Ey i ( ρ ε ) Eπ σ + π = κy + βeπ + u, π is he inflaion gap, and ( ), ρ + ε is he naural real rae; ε and u are a naural rae and a cos-push shock, respecively. The wo shocks are i.i.d. and have 2 2 variance σ ε and σ u, respecively. Moneary policy is he choice of he nominal ineres rae se according o a sandard Taylor-ype linear feedback policy rule: i = ρ+ φπ + φ y. π y Assuming ha he Taylor principle is saisfied, he raional expecaions equilibrium indicaes ha he evoluion of he oupu gap and inflaion can be represened as linear funcions of he shocks, and ha his evoluion also depends on he policy parameers in he Taylor-ype policy rule: ε φπ u y =, σ + φ + κφ y ( ) κε + σ + φy u π =. σ + φ + κφ Opimal policy can be viewed in a variey of ways, bu le s suppose we view consrained opimal moneary policy as he choice of he Taylor rule coefficien on he inflaion gap, keeping he Taylor rule coefficien on he oupu gap fixed. This migh be inerpreed as he policymaker s degree of aenion o deviaions of inflaion from arge, relaive o he degree of aenion o oupu gap developmens in he Taylor-ype rule. The policymaker hen chooses his coefficien opimally o minimize a loss funcion of he form: 2 2 ( β ) β ( απ + y ) 1, = 0 where α is posiive and represens he relaive weigh on he desirabiliy of inflaion sabilizaion compared o oupu sabilizaion in he eyes of he policymaker. The value of his y π π 12 For a exbook reamen of he sandard New Keynesian model, see Woodford (2003). 11

12 parameer could be high or low, encompassing policymakers who migh prefer more or less oupu sabilizaion versus inflaion sabilizaion. I urns ou ha regardless of he value of α he soluion o his problem indicaes ha opimal policy requires a large coefficien on he inflaion gap, ha is, φπ. Wha are he implicaions of his finding for he slope of he Phillips curve in such a model? The slope of he Phillips curve ha is, he coefficien on he oupu gap in a regression of he inflaion gap on he oupu gap is given by ( π, y) ( ) ( y) Cov κσ φ σ + φ σ γ = =. Var y 2 2 ε π u σε + φπ σu We can ake he limi of he righ-hand side of his expression o deduce ha, under opimal moneary policy as described above, he Phillips curve becomes fla: lim γ 0. φ π = 12

13 Figure 1. Slope of he Phillips Curve Noe: Rolling 15-year window esimaes and confidence bands from a panel of G-7 economies. Source: Bank for Inernaional Selemens (2017). Figure 2. The U.S. Price Level Pah Noe: De faco price level argeing occurred from he mid-1990s o 2012, bu he acual price level has since fallen off he price level pah. Sources: Bureau of Economic Analysis and auhor s calculaions. Las observaion: November