Moneary policymaking and inflaion expecaions: The experience of Lain America Luiz de Mello and Diego Moccero OECD Economics Deparmen Brazil/Souh America Desk 8h February 7
1999: new moneary policy regimes in Brazil, Chile, Colombia and Mexico. Inflaion argeing wih floaing exchange raes. Quesions: 1. How was moneary policy s persisence affeced?. Has moneary policy become more forward looking? 3. Are oher objecives of moneary policy being negleced? 4. How volailiy in key variables has been affeced? Wha was he role of shocks? 5. Is moneary policy anchoring inflaion expecaions? 6. Has ineres rae volailiy spread ino expeced inflaion volailiy? Answering hese quesions: convenional New Keynesian model (Q1- Q4), coinegraing echniques (Q5) and a mulivariae GARCH (Q6).
Convenional New Keynesian model π δe π (1 δ π λy = 1 ) 1 u π (1) y = μe y 1 ( 1 μ) y 1 φ( r Eπ 1) u y () r ρ γy ) τe = r 1 ( 1 ρ)( βeπ 1 u r (3)
Esimaion Resuls Brazil Chile Colombia Mexico 1 A B 1 1 1 1999:1 1999:7 1996:1 1999:1 1996:1 1999:1 1996:1 o 6: o 6: o 1999:9 o 6: o 1999:9 o 6: o 1998:1 1996:1 o 1998:1 1999:1 o 6: δ.49.5.54.48.51.6.5.49.5 (.35) (.54) (.83) (.167) (.76) (.84) (.88) (.66) (.78) λ... -. -.1.... μ (.8) (.3) (.4) (.14) (.13) (.1) (.) (.9) (.1).44.47.46.5.51.4.53.56.5 (.1) (.65) (.66) (.134) (.1) (.17) (.113) (.87) (.79) φ 1.85 1.59 1. -.48. -1.95-1.7 1.19 -.7 ρ (5.96) (1.376) (.168) (.751) (.35) (6.619) (1.774) (1.351) (1.78).3.66.61.3.63.9.56.1.11 (.99) (.71) (.57) (.547) (.85) (.169) (.9) (.9) (.97) β.54.14.19..11.4 -.1.9.15 γ (1.883) (.19) (.88) (.7) (.46) (.19) (.3) (.3) (.139).1.1.1 -. -..1..3.1 τ (.31) (.7) (.3) (.49) (.51) (.) (.4) (.1) (.6) 5.8.47. -.4 -.16 -.4 -.13.85 3.75 (.91) (.15) (.8) (1.138) (.77) (.4) (.35) (.491) (.63) 1. Expeced values are measured by one-period-ahead values in he relevan variables. Sandard errors are repored in parenheses. (), () and () denoe, respecively, saisical significance a he 1, 5 and 1% levels. Source: Daa available from he cenral banks of Brazil, Chile, Colombia and Mexico; and auhors esimaions.
Answer o quesions 1-3: Persisence has increased in Brazil, Chile and Colombia. Moneary policy has become more forward looking in Brazil and Chile. Moneary policy has became more counercyclical in Brazil bu less so in Colombia and Mexico.
Moneary response o an inflaionary shock Moneary regime 1 Moneary regime A. Brazil.. 1 3 4 5 6 7 8 9 1 11 1 1 3 4 5 6 7 8 9 1 11 1 13 -. -. -.4 -.4 -.6 VAR -.6 VAR Model -.8 -.8 B. Chile...15.15.1.1.5.5 -.5 1 3 4 5 6 7 8 9 1 11 1 -.5 1 3 4 5 6 7 8 9 1 11 1 -.1 -.15 VAR -.1 -.15 VAR Model -. -. -.5 -.5
Moneary response o an inflaionary shock C. Colombia Moneary regime 1 Moneary regime.3.3.5 VAR.5 VAR. Model. Model.15.15.1.1.5.5 -.5 1 3 4 5 6 7 8 9 1 11 1 -.5 1 3 4 5 6 7 8 9 1 11 1 -.1 -.1 D. Mexico.3.3.5. VAR.5. VAR Model.15.15.1.1.5.5 -.5 1 3 4 5 6 7 8 9 1 11 1 -.5 1 3 4 5 6 7 8 9 1 11 1 -.1 -.1
Counerfacual analysis (Q4) Wha role for shocks and moneary policy? H H H : σ X ( P1, S1) / σ X ( P, 1 S : σ X ( P, S ) / σ X ( P1, 1 1 S : σ X ( P, S1) / σ X ( P, 3 S ) ) )
Counerfacual analysis (Q4) Real ineres rae volailiy has been reduced in Brazil, Colombia and Mexico. Oupu volailiy has been reduced in Colombia.
Answer o quesion 4: The reducion in real ineres rae volailiy has o do wih shocks in Brazil, Colombia and Mexico. The reducion in oupu volailiy has o do wih shocks in Colombia.
Is moneary policy anchoring inflaion expecaions? (Q5) Brazil Chile 14 7 4.5 8 1 Expeced Inflaion (lef scale) Inflaion arge (lef scale) Ineres rae (righ scale) 5 4 Expeced Inflaion (lef scale) Inflaion arge (lef scale) Ineres rae (righ scale) 7 6 1 3 3.5 5 8 1 3 4 6 19.5 3 4 17 1 Jul-1 Jul- Jul-3 Jul-4 Jul-5 15 1.5 Sep-1 Sep- Sep-3 Sep-4 Sep-5
Is moneary policy anchoring inflaion expecaions? (Q5) (con.) Colombia Mexico 7 8 8.5 6.5 6 7.5 7.5 6.5 Expeced Inflaion (lef scale) Inflaion arge (lef scale) Ineres rae (righ scale) 18 16 14 5.5 7 5.5 4.5 1 1 5 Expeced Inflaion (lef scale) 6.5 3.5 8 4.5 Inflaion arge (lef scale).5 6 Ineres rae (righ scale) 4 Sep-3 Sep-4 Sep-5 6 1.5 Nov- Nov-1 Nov- Nov-3 Nov-4 Nov-5 4
Co-inegraion analysis Brazil: Chile: E r π 1 =.74.4 r 1.5 π 1 = 15.5.49E π 1.1 π 1 r = E 1.69 5.53 π 1 Mexico: E π r 1 =.1.4 r.96 π 1 =.18 1.87E π 1.88 π 1
Answer o quesion 5: There are long-erm relaionships beween he ineres rae, expeced inflaion and he inflaion arge, suggesing ha moneary policy has been conduced in a forward-looking manner and helped anchor inflaion expecaions.
Moneary policy and inflaion expecaions: Are here volailiy spillovers? (Q6) r = a1 a11r 1 a1 ( Eπ 1 π ) a e a y 1 1 13 1 14 1 ε1 ( Eπ 1 π ) = a a r a E a e a y 1 1 ( π 1 π ) 1 1 1 3 1 4 1 ε (4) (5) h11, = c1 b ε 1, 1 b11b1ε 1, 1ε, 1 b ε, 1 h 11 1 = h = c b b ε 1 ( b b b b ) ε ε b b ε 1, 1, 11 1, 1 1 1 11 1, 1, 1 1, 1 h, = c3 b ε 1, 1 b1bε 1, 1ε, 1 b ε, 1 1 (6) (7) (8)
MGARCH Resuls Moneary reacion funcion and inflaion expecaions: M-GARCH analysis of volailiy spillover effecs Brazil Chile Colombia Mexico r ( E π ) π 1 1 r ( E π ) π 1 1 r ( E π ) π 1 1 r ( E π π ) r 1.79 -.43.6.6 -.11 -.58 -.3.4 (4.1) (-6.19) (9.64) (.75) (-.84) (-.89) (-.3) -1.35 ( Eπ 1 π ) 1 1 -.3.86.19.4 -.19.16 1.68 -.1 1 1 (-1.1) (33.6) (.36) (3.96) (-.91) (.87) (4.6) (-1.14) e 1.1.3 -.1. (.81) (6.1) (-.58) (.5) y 1.7.16.4.9.1.1.13.4 (4.4) (6.15) (.15) (3.46) (6.63) (.58) (.3) (3.) ECT (-1) -..7 (-.18) (4.95) Variance equaions C 11.41.7..84 (4.15) (.48) (.6) (8.87) C 1.48. -.6.1 (9.37) (.1) (-.8) (.5) C.1.11.7.7 (.35) (1.4) (.1) (.5) B 11.48.89 1.3.9 (1.43) (3.63) (6.) (.4) B 1.7.1.3.11 (.54) (1.4) (.5) (.4) B 1-1.8.7 1.8 -.18 (-5.33) (1.15) (.66) (-4.47) B -.59.13 -.7.15 (-.3) (.73) (-1.71) (1.4) Diagnosic ess Log-likelihood -4.83-177.1-111.33-78.13 AIC -3.4-8.97-7.86-4.96 SBC -.74-8.51-7.58-4.68 Normaliy es (KS).11.11.11.6.9.14.7.7 LB(6) 3.59 6.4 6.1 1.9.84 3.14 8.44 4.3 ARCH(6) 1.4 1. 4.41 8.7 5.83 1.16 5.4 7.55 The numbers in parenheses are saisics. AIC and SBC are he Akaike and Schwarz Bayesian informaion crieria, respecively. KS is he Kolmogorov-Smirnov es of he null hypohesis of normaliy. LB (6) and ARCH(6) are he Ljung-Box es for auocorrelaion and he Lagrange Muliplier (LM) es for heeroscedasiciy up o 6 lags, respecively, perf on he univariae sandardised residuals. In he case of Chile, ( Eπ 1 π ) refers o expeced inflaion. ECT = error-correcion erm. (), () and () denoe saisical 1 significance a he 1%, 5% and 1% levels, respecively. The samples are: 1:7 o 6:1 for Brazil, 1:9 o 6:1 for Chile, 3:9 o 5:1 for Colombia and :11 o 6:1 for Mexico.
Answer o quesion 6: Greaer volailiy in he moneary sance has lead o higher volailiy in expeced inflaion in Brazil, Colombia and Mexico. No volailiy spillover effec was deeced in he case of Chile
Moneary policymaking and inflaion expecaions: The experience of Lain America Luiz de Mello and Diego Moccero OECD Economics Deparmen Brazil/Souh America Desk 8h February 7