AN ESTIMATED NEW KEYNESIAN POLICY MODEL FOR THE CZECH REPUBLIC

1
AN ESTIMATED NEW KEYNESIAN POLICY MODEL FOR THE
CZECH REPUBLIC

• Aleš Melecký
• Department of Economics
• Technical University of Ostrava

2
Outline of the Presentation

• Motivation
• Model
• Data and Estimation Method
• Estimation Results
• Preview
• Impulse Response Analysis
• From inflation to exchange rate targeting
• Conclusions

3
Motivation

• Regional monetary integrations, including the
  Czech Republics integration into EMU
• From IT to ERT
• Description of Czech economy and identification
  of shocks
• Identification of transmision channels

4
The employed models features

• Two blocks
• a small open economy (the Czech Republic)
• the rest of the world (euro area economy)
• New Keynesian policy model with rational
  expectations
• Relaxed cross-equation coefficient restrictions
• Empirically determined lag structure of the
  transmission mechanism

5
Model Equations Domestic Block

• Phillips curve inflation dynamics
• IS curve output gap dynamics
• MP reaction function interest rate dynamics
• UIP equation exchange rate dynamics

6
Model Equations Foreign Block
7
Data

• Quarterly data series
• 1995Q1 to 2007Q4 for the Czech Republic
• 1981Q1 to 2007Q4 for the euro area
• Data Sources
• Datastream, Czech National Bank statistics, Fagan
  et al. (2001)(extrapolation back to 1981 for the
  euro area using growth rates)

8
Data

• Output gap
• Deviation of the log GDP from its trend,
  estimated by the HP-filter and multiplied by 100
• Inflation
• Annualized percentage changes in CPI for the
  Czech Republic and harmonized CPI for the euro
  area
• Interest rate
• Three-month PRIBOR for the Czech Republic
• Three-month EURIBOR for the euro area
• Real exchange rate CZK/EUR
• Cross exchange rate using synthetic USD/EUR rate
  and CZK/USD rate, times CPIs ratio, in logs,
  linearly detrended, and multiplied by 100

9
Estimation method

• Choose GMM over MLE-based methods to ensure
  higher robustness of estimates against possible
  misspecification
• Up to three lags of variables in the system used
  as instruments
• HAC estimated using the Bartlett kernel with the
  variable Newey-West bandwidth selection
• Pre-whitening of moment conditions applied
• BIC used for lag-length selection

10
Estimation Results
11
Estimation Results Phillips curve

• Inflation process more forward- than
  backward-looking both in CR and EA
• Sensitivity of inflation to demand pressures,
  almost three times higher in CR than EA
• Significant estimate of implies that Czech
  firms incorporate expected changes in excess
  demand into current prices (1 for 10)
• ER pass-through a 10 increase in CZK/EUR
  expected to result in a 1.2 increase in Czech
  inflation
• CR is estimated to face more than three times
  larger supply shocks than EA
• Data fit lower for CR (adj.R2 0.25) than for EA
  (adj.R2 0.83) -- less observations and the
  transition in CR

12
Estimation Results IS curve

• Output gap rigidity apparent in both countries
  (rho_ylt0)
• Real interest rate elasticity fairly similar in
  CR and EA
• Interest rate transmission channel about
  two-times longer in EA than CR (6 vs. 3 periods)
• Exchange rate transmission channel (2-period lag)
  faster than IR transmission channel in CR
• Exchange rate elasticity estimated to be smaller
  than interest rate elasticity
• a 10 increase in EA demand estimated to result
  in a 4 increase in Czech output
• Data fit of OE-IS curve for CR (adj.R2 0.89)
  slightly better than CE-IS curve for EA (adj.R2
  0.83)
• IS shocks size appears to be marginally higher
  in CR than EA

13
Estimation Results MP rule

• CNB and ECB smooth their interest rates, where
  ECB policy rate shows slightly more inertia
• ECB reaction to expected inflation appears to be
  higher than that of CNB
• ECB appears to place less weigh on output gap in
  its reaction function than CNB
• ECB a more conservative central banker than CNB
• MP shocks in CR more than four times larger than
  in EA, i.e. ECB applies significantly less
  discretion than CNB
• Both MP reaction functions fit data well adj.R2
  for CR 0.84 and for EA 0.98

14
Estimation Results - UIP

• The estimated UIP equation implies that the
  standard deviation of the exchange rate shock is
  5.86
• ER shock is thus the largest shock in the system
• Also, estimated to be significantly positively
  correlated over time

15
Impulse Response Analysis

• How domestic (Czech) economic variables respond
  to individual structural shocks, both domestic
  and external
• The estimated system of equations with rational
  expectations solved into an VAR form using QZ
  algorithm of Sims (2002)
• The reduced-form used to generate impulse
  responses of domestic variables to selected
  shocks

16
Impulse Response Analysis

• Domestic IS shock hits the Czech economy
• IS shock ?output gap ?inflation CNB reacts
  and ?interest rate ?interest rate differential
  between CR and EA CZK appreciates against EUR
  the IR hike and CZK appreciation bring the output
  gap and inflation back to their steady-states

17
Impulse Response Analysis

• Domestic AS shock hits the Czech economy
• AS shock ?inflation CNB ?interest rate CZK
  appreciates against EUR ?output gap into
  negative values IR and output gap return to
  steady states, CZK depreciates against EUR and
  settles around its steady state

18
Impulse Response Analysis

• Domestic MP shock hits the Czech economy
• MP shock CNB ?interest rate ? inflation ?
  real interest rate differential between CR and EA
  CZK strongly appreciates against EUR output
  gap contracts significantly as a result of lower
  external price competitiveness

19
Impulse Response Analysis

• Foreign IS shock hits the Czech economy
• Foreign IS shock ?foreign demand ?Czech
  output gap ?inflation ?EA inflation CNB
  ECB ?interest rate positive real IR
  differential for CZK CZK appreciates
  inflation ? faster in CR real IR peaks sooner
  higher IR in EA EUR appreciates against CZK
  before settling at its steady state

20
Impulse Response Analysis

• Foreign AS shock hits the Czech economy
• Foreign AS shock ECB ?interest rate ?real
  interest rate CZK depreciates against EUR ?
  Czech net exports and output gap ?Czech
  inflation CNB ?interest rate ?real interest
  rate differential CZK depreciation slows down
  different intensity of CNB ECB reactions and
  lengths of the monetary transmission swings of
  the Czech variables before reaching steady states

21
Impulse Response Analysis

• Foreign MP shock hits the Czech economy
• Foreign MP shock ? EA real interest rate
  depreciation of CZK against EUR ?net exports
  Czech output gap positive ? inflation Czech
  output gap peaks quickly returns to steady
  state rapid adjustment causes a period of
  deflation CNB ? interest rate inflation
  returns back to its steady state

22
Impulse Response Analysis

• Exchange rate shock hits the Czech economy
• Exchange rate shock short CZK appreciation
  against EUR negative Czech output gap ?
  inflation Czech output gap expected to adjust
  back fast positive output gap adjustment
  short-lived ? inflation initially CNB ?
  interest rate after several quarters CNB
  ?interest rate to smooth out ER response which
  returns to zero in about 13 quarters

23
From inflation to exchange rate targeting

• Estimating CBs loss function
• Optimalization of inflation targeting and
  movement to exchange rate targeting
• More in our article From Inflation to Exchange
  Rate Targeting Estimating the Stabilization
  Effects for a Small Open Economy(Ales Melecky
  and Martin Melecky)

24
Conclusions

• Model Estimation
• Pricing of firms in the Czech Republic shows
  higher rigidity than pricing of firms in EA
• ECB appears to be a more conservative central
  banker with higher MP rate inertia compared with
  the CNB
• The interest rate transmission channel of MP is
  significantly shorter in CR than EA
• A foreign demand shock has the highest impact on
  the Czech economy, higher than an ER shock or any
  domestic shock
• Out of the domestic shocks, the supply shock
  appears to be the most influential

25

• Thank you!
• ales.melecky.ekf_at_vsb.cz