Transcript π t - Seðlabanki Íslands

Seðlabanki Íslands
Is inflation control more difficult
in very small open economies?
The Challenges of Globalisation for Small Open
Economies with Independent Currencies
31 May – 1 June 2007
Thórarinn G. Pétursson
Central Bank of Iceland
General consensus
• Keiko & Hunt (2006)
– “… inflation stabilization is more challenging in Iceland
than in other industrial countries primarily because of the
relative magnitudes of the economic shocks”
– “... Inflation-output variability trade-off faced by the
monetary authorities in Iceland is much less favorable
than those faced by the monetary authorities in many
other industrial countries, even other small inflationtargeting countries like New Zealand”
– “Iceland and New Zealand are both subject to to bigger
shocks than the other large countries. The key difference
between Iceland and New Zealand is that the inflation
and exchange rate shocks in Iceland in Iceland are
considerably larger than those in New Zealand”
• Hunt (2007)
– “Macroeconomic volatility in Iceland is already much
higher than in all other industrial countries. To a large
extend this simply reflects the small size of the economy”
Country sample
• Want observations on “very small” countries that are
informative for the more developed countries such as ICE…
• … without extending the analysis to all the very small
Caribbean islands that are very poor and undeveloped
• The full sample therefore includes 34 countries
– Median population 9.5m
– Median GDP per capita 29.8 thousand PPP adjusted US$
– GDP about 50% of world GDP (15% of world population)
• Population distribution of the 34 countries
–
–
–
–
–
–
–
POP size < 1m: 3 countries
POP size 1-5m: 7 countries
POP size 5-10m: 8 countries
POP size 10-20m: 6 countries
POP size 20-50m: 3 countries
POP size 50-100m: 5 countries
POP size > 100m: 2 countries
Inflation more volatile in small countries...
• Standard deviation of
inflation 1985-2005
• Inflation measured as
annualised quarterly
changes in seasonally
adjusted consumer
prices
• Countries ranked
according to PPP
adjusted GDP in 2006
INFVOL
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
Rank correlation
with size = -0.56
Median = 0.03
T UR = 0.20
ISR = 0.23
0.02
0.04
0.06
0.08
0.10
0.12
0.14
... and harder to predict
• Standard deviation of
inflation forecast errors
from 1995-05
• Estimated from a 40
quarters rolling window
VAR model including
inflation, import price
inflation, output gap
and interest rate
INFE
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
Rank correlation
with size = -0.53
Median = 0.02
T UR = 0.15
0.02
0.04
0.06
0.08
0.10
What can explain this?
• A simple New-Keynesian Phillips curve gives some
clues
πt = απt-1 + (1 – α)Etπt+1 + φyt + λxt + εt
•
•
•
•
πt: inflation and Etπt+1: expected inflation
yt: the output gap
xt: exchange rate shocks
εt: other supply shocks
• Key factors therefore seem to be
– Persistence of inflation shocks and variability of inflation
expectations
– Variability of real shocks and exposure to idiosyncratic
shocks
– Variability of exchange rate shocks and its pass-through
to prices
– Variability of other shocks
– Credibility of monetary policy and its transparency
Smaller economies are more open…
IMS
OPEN
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.0
Rank correlation
with size = -0.63
Median = 84.7
40.0
80.0
120.0
160.0
200.0
240.0
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.0
Rank correlation
with size = -0.63
Median = 66.4
40.0
80.0
120.0
160.0
... more exposed to idiosyncratic shocks...
• Measured as the
contemporenous
correlation of the
cyclical components of
domestic and world
GDP
INTER
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
-0.40
-0.20
0.00
Rank correlation
with size = 0.59
Median = 0.40
0.20
0.40
0.60
0.80
1.00
… and with more volatile real economy
CVOL
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
YVOL
Rank correlation
with size = -0.63
Median = 0.01
0.01
0.02
0.03
0.04
0.05
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
Rank correlation
with size = -0.46
Median = 0.01
0.01
0.02
0.03
0.04
Interest rates seem more volatile...
RSVOL
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
Rank correlation
with size = -0.34
Median = 0.01
T UR = 0.29
0.01
0.02
0.03
0.04
0.05
0.06
0.07
... and monetary policy less predictable
• Standard deviation of
interest rate forecast
errors from 1995-05
• Estimated from a 40
quarters rolling window
VAR model including
inflation, import price
inflation, output gap
and interest rate
• An alternative measure
using a forwardlooking Taylor rule
gives very similar
results
RSE
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
Rank correlation
with size = -0.34
Median = 0.01
T UR = 0.46
CHI = 0.09
0.01
0.02
0.03
0.04
0.05
0.06
But inflation seems less persistent...
• Estimate a unilateral
inflation equation
allowing for a break in
mean inflation of
unknow timing
πt = (α + βDt) + γπt-1
+ ΣiφiΔπt-i + εt
• Inflation persistence is
given by γ
PERS
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00
Median = 0.46
Rank correlation
with size = 0.25
0.25
0.50
0.75
1.00
… and exchange rates more stable
EERVOL
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
Rank correlation
with size = 0.22
T UR = 0.12
Median = 0.03
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
But exchange rate noise is greater…
• A signal extraction
approach suggested
by Durlauf and Hall
• Gives a lower bound
for the variance of
model noise in rational
expectations models
• Closely related to the
variance of the
exchange rate risk
premium
EXNOISE
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
0.00 0.05
Rank correlation
with size = -0.43
Median = 0.11
0.10 0.15 0.20
0.25 0.30 0.35
0.40 0.45
… and the pass-through to prices greater
• Estimate a VAR model
with domestic and
foreign inflation,
exchange rate
changes, the output
gap and short-term
interest rate
• Generate accumulative
impulse responses for
a 1% shock to the
exchange rate using
the generalised
approach of Pesaran
and Shin
PT
USA
JAP
GER
UK
FRA
IT A
CAN
ESP
AUS
T UR
HOL
BEL
SWE
AUT
SWI
GRE
CZE
NOR
POR
CHI
DEN
IRE
HUN
FIN
ISR
NZ
SLK
LIT
SLO
LAT
EST
CYP
ICE
MAL
Rank correlation
with size = -0.30
Median = 0.20
Not included
0.00
0.20
0.40
0.60
0.80
1.00
Cross-country inflation volatility
• What explains the cross-country variation of
inflation volatility (INFVOL)?
• Potential explanations
– Country size: SIZE
– Country openness: OPEN
– Real economy fluctuations: YVOL
– Importance of idiosyncratic shocks: INTER
– Inflation persistence: PERS
– Monetary policy predictability: RSE
– Exchange rate pass-through: PT
– Exchange rate noise: EXNOISE
Cross-country inflation volatility
INFVOL = -0.004 – 0.025OPEN + 0.279RSE
(0.3)
(2.0)
(4.2)
+ 0.182PT + 0.230EXNOISE
(6.3)
(3.3)
N = 33, R2(adj) = 0.72, SE = 0.028
• Other variables not significant
• Robustness with respect to country sample
– PT and EXNOISE robust to excluding any country
– OPEN significant at the 10% level, except when
excluding ISR (p = 0.60) and MAL (p = 0.13)
– RSE significant at the 1% level, except when excluding
TUR (p = 0.56)
Interpretation
• So the general view seems correct
– Inflation control is more difficult in small open economies
to the extend that they have noisier exchange rates and a
larger pass-through to inflation
• But there is hope
– Adopt a credible and transparent inflation-fighting
monetary policy that can reduce the pass-through, make
monetary policy more predictable and contribute to
reducing other sources of volatility
– Reform other macro policy to give better support to
inflation fighting credibility
– These policy reforms can also contribute to reducing
FOREX risk premia
• Structural issues
– More open to international trade
– Improve functioning of FOREX market
– Join a larger currency area?