Does Inflation Targeting Anchor Long-Run Inflation Expectations?

Transcription

1 Does Inflaion Targeing Anchor Long-Run Inflaion Expecaions? Evidence from Long-Term Bond Yields in he Unied Saes, Unied Kingdom, and Sweden Refe S. Gürkaynak, Andrew T. Levin, and Eric T. Swanson Bilken Universiy, Federal Reserve Board, and Federal Reserve Bank of San Francisco

2 Summary Long-erm ineres raes in he U.S. move a grea deal - no jus uncondiionally, bu condiionally (in response o economic news) This behavior is inconsisen wih sandard macroeconomic models in which he seady sae is consan over ime and known by all agens Long-erm ineres raes in he U.K. and Sweden (boh inflaion argeers) generally do no respond o economic news - bu U.K. behaves similar o U.S. prior o Bank of England gaining independence from Treasury/Parliamen in All of our empirical evidence can be explained by changes in long-erm inflaion expecaions in he U.S. (and greaer anchoring in U.K., Sweden)

3 Ouline of Talk Review basic ideas from Gurkaynak, Sack, Swanson (25):. Macro Model Implicaions a. seady sae b. excess sensiiviy of long-erm ineres raes 2. Empirical Mehodology a. forward raes b. macroeconomic, moneary policy surprises This paper:. Compare U.S., U.K., Sweden a. compare inflaion argeers (U.K. pos-998, Sweden) o non-inflaion argeers (U.S., U.K. pre-997) 2. Compare sensiiviy of nominal raes, real raes, inflaion compensaion

4 A Benchmark Model (GSS 25) π π ε γ π μ π μ π y L A E = + ) ( ) ( y y E i y L A y E y ε π β μ μ + + = + + ) ( ) ( ) ( Clarida-Gali-Gerler New Keynesian specificaion: μ = Rudebusch Hybrid New Keynesian specificaion: μ =.3 Close he model wih a Taylor-ype Rule : [ ] i i c by a c i ε π = ) ( ) (

5 Figure : Impulse Response Funcions for Two New Keynesian Models percen.5 Ineres Rae Response o a percen Inflaion Shock Clarida Gali Gerler Rudebusch Ineres Rae Response o a percen Oupu Shock.5 percen Ineres Rae Response o a percen Ineres Rae Shock percen Time (quarers)

6 Figure : Ineres Rae Impulse Responses in Two Benchmark DSGE Macroeconomic Models percen Ineres Rae Response o percen Oupu Demand Shock CEE model (U.S.) ALLV model (Sweden) Ineres Rae Response o percen Inflaion Cos Push Shock percen Ineres Rae Response o percen Moneary Policy Shock percen Quarers

7 Empirical Mehodology Regress ineres rae responses in narrow (-day) windows on macroeconomic daa releases and moneary policy announcemens on hose same days Noe: ) use forward ineres raes (raher han erm raes) e.g., ( + fwd ) = ( + r ) 9o 9 ( + r9 ) 2) use surprise componen of macroeconomic daa release (or moneary policy announcemen)

8 Macroeconomic daa release (or moneary policy announcemen) ex ane forecas measure surprise componen of announcemen and financial marke response

9 Table : Responses of Forward Raes o Economic News (99-22) ---Ending yr. ahead Ending 5 yrs. ahead Ending yrs. ahead-- Coef. Sd. Err. Coef. Sd. Err. Coef. Sd. Err. Macroeconomic Daa Releases Capaciy Uilizaion.36.33***.26.57**.8.6 Consumer Confidence 2..4*** ***.97.54*** CPI (Core).67.42***.8.6***.9.66* Employmen Cos Index *** *** *** GDP (Advance) *** * ** Iniial Claims *** *** ** Leading Indicaors.95.34*** NAPM 3..5*** ***.53.63** New Home Sales.8.39***.65.54***.92.5* Non-farm Payrolls 5..57*** ***.88.97* PPI (Core) **.46.5*** Reail Sales *** **.93.92** Unemploymen Rae *** Moneary Policy Surprises.47.*** ** Huber-Whie sandard errors. *** indicaes significance a he % level, ** a he 5% level, and * a he % level. The esimaed coefficien indicaes he basispoin response of he one-year forward rae per sandard deviaion of he macroeconomic variable and per basis-poin surprise in moneary policy announcemen. Regresssions include consan erms ha are no shown in he able.

10 Figure 3: Response of Forward Raes o Macroeconomic Surprises Response of forward raes Years ahead Cons. Confidence Response of forward raes Years ahead Reail Sales Response of forward raes Years ahead NAPM (ISM) Response of forward raes.5.5 Response of forward raes Response of forward raes Years ahead Iniial Claims Years ahead Nonfarm Payrolls Years ahead Unemploymen Response of forward raes Years ahead Emp. Cos Ind. Response of forward raes Years ahead Core CPI Response of forward raes Years ahead Core PPI

11 Table : Responses of Forward Raes o Economic News (99-22) ---Ending yr. ahead Ending 5 yrs. ahead Ending yrs. ahead-- Coef. Sd. Err. Coef. Sd. Err. Coef. Sd. Err. Macroeconomic Daa Releases Capaciy Uilizaion.36.33***.26.57**.8.6 Consumer Confidence 2..4*** ***.97.54*** CPI (Core).67.42***.8.6***.9.66* Employmen Cos Index *** *** *** GDP (Advance) *** * ** Iniial Claims *** *** ** Leading Indicaors.95.34*** NAPM 3..5*** ***.53.63** New Home Sales.8.39***.65.54***.92.5* Non-farm Payrolls 5..57*** ***.88.97* PPI (Core) **.46.5*** Reail Sales *** **.93.92** Unemploymen Rae *** Moneary Policy Surprises.47.*** ** Huber-Whie sandard errors. *** indicaes significance a he % level, ** a he 5% level, and * a he % level. The esimaed coefficien indicaes he basispoin response of he one-year forward rae per sandard deviaion of he macroeconomic variable and per basis-poin surprise in moneary policy announcemen. Regresssions include consan erms ha are no shown in he able.

12 Figure 4: Response of Forward Raes o Moneary Policy Surprises Response of forward raes Years ahead

13 Figure 5: Ten-year-ahead Forward Rae and Moneary Policy Surprises Jan Response of year ahead forward rae 2 Apr Jul92 Dec9 Dec98 Aug9 Apr94 Feb94 Apr92 Sep96 Oc99 May98 Sep94Jul94 Feb9 Oc9Feb95 May Nov98 Feb92 Mar95 Jul96 Feb97 Jun May2 Jan9 Dec94 Sep92 May97 Jan96 Mar93 Mar2 Jul98 Aug95 Dec99 Jul9 Oc92 May96 Sep97 Dec Nov96 Feb98 May9Dec Nov9 Nov Oc9 Mar9 Jan2 Dec92 Aug98 Nov92 Oc Aug99 Aug92 Nov93 Sep95 Nov95 Nov99 Mar Dec9 Aug96 Jun99 Oc98 Jul9 Oc Sep9 Nov97 Aug Mar Feb9 Jul97 Aug97 Dec97 Dec96 Feb93 May95 Mar98 Aug Sep2 Aug93 Feb99 Apr92Mar97 Dec95 May93 Dec2 Jul9 Mar96 Jul93 Mar9 Jan May99 Jul95 Sep93 Nov Mar99 Mar94May92 May9 Dec9 Nov2 Feb Aug2 May Jun2 Sep98Jun Aug9 Nov9 Dec93 Nov94 Aug94 May Moneary policy surprise

14 Table 2: Response of STRIPS-based Forward Raes o Economic News ---Ending 5 yrs. Ahead Ending yrs. Ahead--- Coef. Sd. Err. Coef. Sd. Err. Macroeconomic Daa Releases Capaciy Uilizaion 2.5.9**.37.79* Consumer Confidence **.6.63* CPI (Core) **.75.6* Employmen Cos Index ** *** GDP (Advance) ** * Iniial Claims *** Leading Indicaors NAPM *** *** New Home Sales * Non-farm Payrolls ***.93.8* PPI (Core) ** Reail Sales *** 2.5.4* Unemploymen Rae **..9 Moneary Policy Surprises ***

15 Summary of Empirical Observaions Far-ahead forward raes: ) exhibi economically and saisically significan responses o economic news 2) move in same direcion as oupu and inflaion surprises 3) move in opposie direcion o moneary policy surprises

16 Possible Explanaions According o Fisher s Equaion: i * = r * + π * The N-year-ahead forward rae saisfies (for large enough N): fwd = NoN r * + π * + ρ +

17 Possible Explanaions According o Fisher s Equaion: i * = r * + π * The N-year-ahead forward rae saisfies (for large enough N): fwd = NoN r * + π * + ρ + Noe: Changes in r* alone are no sufficien o explain empirical findings

18 Possible Explanaions According o Fisher s Equaion: i * = r * + π * The N-year-ahead forward rae saisfies (for large enough N): fwd = NoN r * + π * + ρ + Noe: Changes in r* alone are no sufficien o explain empirical findings Changes in ρ alone are no sufficien o explain empirical findings

19 Possible Explanaions According o Fisher s Equaion: i * = r * + π * The N-year-ahead forward rae saisfies (for large enough N): fwd = NoN r * + π * + ρ + Noe: Changes in r* alone are no sufficien o explain empirical findings Changes in ρ alone are no sufficien o explain empirical findings Bu changes in π* alone are sufficien o explain all empirical findings

20 Model wih Time-varying π* π = μ E π + + ( μ) A ( L) π + γ y + π ε π y = E y + + ( μ) Ay ( L) y β ( i Eπ + μ ) + ε y * i i = ( c) π + a( π π ) + by + ci + ε π = π + θ( π π ) + ε * * * π*

21 Impulse Responses for Rudebusch Model wih Time-varying π* (Perfec Informaion)

22 Model wih Time-varying π* and Imperfec Informaion π = μ E π + + ( μ) A ( L) π + γ y + π ε π y = E y + + ( μ) Ay ( L) y β ( i Eπ + μ ) + ε y * i i = ( c) π + a( π π ) + by + ci + ε π = π + θ( π π ) + ε * * * π* ˆ π = ˆ π + θ( π ˆ π ) κ( i iˆ ) * * *

23 Expeced Impulse Responses for Rudebusch Model wih Time-varying π* (Imperfec Informaion)

24 Acual Impulse Responses for Rudebusch Model wih Time-varying π* (Imperfec Informaion)

25 Summary of Analysis for he U.S. Forward nominal ineres raes respond significanly o economic news Changes in r* alone no sufficien o explain empirical findings Changes in risk premium alone have a hard ime explaining empirical findings (e.g., cyclicaliy goes wrong way) Changes in π* alone are sufficien o explain all empirical findings Imperfec informaion in privae secor no necessary, bu increases magniude, inuiiveness of effecs of changes in π*

26 Comparing he U.S., U.K., and Sweden (GLS 26). U.K.: inflaion argeer since Ocober 992 (hough Bank of England no independen pre-997) Sweden: inflaion argeer since January 993 U.S.: informal commimen o price sabiliy 2. U.S., U.K., and Sweden have all issued inflaion-indexed bonds - compare forward nominal raes, real raes, and inflaion compensaion U.S.: TIPS since January 997 (forward raes since 998) U.K.: inflaion-indexed bonds since a leas mid-98s Sweden: inflaion-indexed bonds since 994 (forward raes since 996) Relaively shor sample, bu daily bond yield daa, frequen release of imporan macro saisics yield abou observaions per counry

27 Table : U.S. Forward Rae Responses o Economic News (998-25) -year Forward -year Forward -year Forward -year Nominal Rae Real Rae Inflaion Compensaion Nominal Rae ending in yrs ending in yrs ending in yrs Capaciy Uilizaion.57** (2.76).98 (.68).46 (.74).52 (.94) Consumer Confidence.46** (2.89).62 (.99).9 (.47).43 (.94) Core Consumer Price Index.98 (.5).8 (.74) -.24 (-.57).42* (2.3) real GDP (advance) 2.7* (2.5) 2.7* (2.22).3 (.52).77* (2.6) Iniial Jobless Claims -.3** (-3.79) -.73* (-2.53) -.2 (-.32) -.5* (-2.) NAPM/ISM Manufacuring 2.28** (2.65) 2.98** (4.27).5** (3.52).48* (2.55) New Home Sales.53 (.36).5* (2.38) -.27 (-.98).42** (3.47) Nonfarm Payrolls 4.44** (7.3).84* (2.26).32** (3.46).52 (.83) Reail Sales (ex auos).69** (2.68).44 (.7).56 (.48).88 (.45) Unemploymen Rae -.97 (-.47).78 (.79).89 (.36) -. (-.5) Moneary Policy.23* (2.2) -.2 (-.43). (.3) -.2 (-.43) # Observaions R Join es p-value.**.**.5**.2**

28 Table 3 U.K. Forward Rae Responses o Economic News, pre-bank of England Independence (993-April 997) Average Earnings real GDP (preliminary) Manufacuring Producion Producer Price Index core Reail Price Index Reail Sales Moneary Policy -year Nominal Rae 3.23** (3.33).75 (.68).76 (.88) 2.3** (3.2) 2.39** (3.9) 2.7** (2.98).67** (5.73) -year Forward Nominal Rae ending in yrs.69 (.82) 2.36* (2.4) -.27 (-.2) 2.98** (2.95) 3.2** (3.7).32 (.3) -.54** (-3.9) -year Forward Real Rae ending in yrs.55 (.72).55 (.7) -.6 (-.76).76* (2.3).62 (.88).5 (.7).6 (.27) -year Forward Inflaion Compensaion ending in yrs.5 (.2).8* (2.2).33 (.29) 2.22* (2.6) 2.6** (3.8) -.9 (-.24) -.6** (-5.99) # Observaions R Join es p-value.**.**.3**.**

29 Table 4 U.K. Forward Rae Responses o Economic News, pos-bank of England Independence (July ) Average Earnings real GDP (preliminary) Manufacuring Producion Producer Price Index core Reail Price Index Reail Sales Moneary Policy -year Nominal Rae.8** (4.2) 2.4** (4.2).26** (3.9).2 (.55) 2.6** (4.83).58** (3.92).72** (5.96) -year Forward Nominal Rae ending in yrs -.38 (-.5) -.53 (-.47).59 (.8).22 (.58) -.89 (-.84) -.8* (-2.5) -.2 (-.92) -year Forward Real Rae ending in yrs -.2 (-.53) -.3 (-.9).63** (2.69).44 (.88) -.3 (-.53).9 (.34). (.2) -year Forward Inflaion Compensaion ending in yrs -.26 (-.94) -.49 (-.54) -.4 (-.9) -.22 (-.63) -.76 (-.95) -.8** (-2.75) -.3 (-.) # Observaions R Join es p-value.**

30 Table 5 Swedish Forward Rae Responses o Domesic Economic News (996-25) Consumer Price Index core Consumer Price Index real GDP (preliminary) Indusrial Producion Producer Price Index Reail Sales Unemploymen Moneary Policy -year Nominal Rae.94* (2.55) 2.72** (4.26).79 (.7) -.4 (-.24).63 (.83) -.49 (-.72) -.26 (-.67).72** (3.62) -year Forward Nominal Rae ending in yrs. (.25) -.68 (-.7).72 (.2) -.7 (-.3) -.47 (-.4).26 (.49) -.42 (-.93).25 (.56) -year Forward Real Rae ending in yrs.6 (.64) -.36 (-.96).43 (.94) -.4 (-.4) -.23 (-.) -.39 (-.93) -.37 (-.57).3 (.63) -year Forward Inflaion Compensaion ending in yrs.85 (.3) -.33 (-.37).29 (.45) -.67 (-.55) -.24 (-.76).65 (.2) -.4 (-.).23 (.48) # Observaions R Join es p-value.**

31 (a) GDP Surprises 2 Unied Saes 2 Unied Kingdom ( ) response in basis poins Sweden 2 Unied Kingdom (998 25) response in basis poins surprise in sandard deviaions surprise in sandard deviaions

32 (b) Inflaion Surprises response in basis poins Unied Saes Unied Kingdom ( ) Sweden 2 Unied Kingdom (998 25) response in basis poins surprise in sandard deviaions surprise in sandard deviaions

33 (c) Moneary Policy Surprises Unied Saes Unied Kingdom ( ) response in basis poins Sweden Unied Kingdom (998 25) response in basis poins surprise in basis poins 2 2 surprise in basis poins

34 (a) Far-Ahead Forward Nominal Raes 5 3 Sweden Unied Kingdom Unied Saes percen

35 (b) Far-Ahead Forward Inflaion Compensaion percen

36 Long-Term Inflaion Expecaions from SPF

37 Conclusions ) In he U.S., far-ahead forward nominal raes and inflaion compensaion respond srongly o economic news 2) All of our empirical findings are consisen wih changes in π* 3) Sandard macro models impose an incorrec resricion: consan π* 4) A credible inflaion arge seems o reduce long-erm ineres rae sensiiviy, as in he U.K., Sweden Caveas: Forward raes in he U.K., Sweden are sill fairly volaile uncondiionally - bu perhaps less so han in he U.S. Inflaion argeing is no a silver bulle credibiliy, commimen required - even hen, some uncondiional volailiy remains