Melihovs Phillips

Phillips Curve Estimation for Latvia : 

Phillips Curve Estimation for Latvia Aleksejs Meļihovs, Anna Zasova 2007

The motivation: 

The motivation Improvement of understanding of the inflation formation mechanism; Testing what type of Phillips Curve model is better for the Latvian reality; Understanding what type of inflation expectations is more important for Latvian inflation dynamics.

Estimated Models: 

Estimated Models Three types of Phillips Curve models have been estimated: Traditional; New Keynesian and Hybrid. Every type of Phillips Curve models has been modelled for two cases: closed and open economy.

Phillips Curves – open economy case: 

Phillips Curves – open economy case Traditional Phillips Curve New Phillips Curve Hybrid Phillips Curve П –inflation E – expectation operator Y – business cycle X – external factors vector

Variables: 

Variables Inflation: core inflation (CPI excluding administrative, energy and non-processed food prices) Business cycle (two alternative indicators): GDP gap (HP filter), unemployment gap (HP filter). Exogenous variables for open economy models: producer prices in main foreign trade partner countries, nominal effective exchange rate.

Why core inflation not CPI?: 

Why core inflation not CPI? National Banks tend to focus on core inflation because: it excludes highly volatile factors; measures of core inflation capture the permanent, or long run trend in CPI; core inflation is generally associated with expectations and demand pressure components of measured inflation and excludes supply shocks.

Core inflation vs CPI inflation: 

Core inflation vs CPI inflation

Why business cycles not marginal production costs and foreign producer prices not import prices?: 

Why business cycles not marginal production costs and foreign producer prices not import prices? Too short time series for MPC: data is available only from 1999; unemployment rate and GDP data – from 1995; for such short time series three years make a big difference! The same story is for import price statistics in Latvia: as Latvia is a price taker, foreign producer prices are viewed as second best for import price approximation.

Business Cycles : 

Business Cycles Russian crisis Rapid growth Recovery after the Russian crisis Recovery after the Banking crisis Banking crisis A sharp fall in oil transit

Exogenous Factors: 

Exogenous Factors q-o-q, % Russian crisis Re-pegging form SDR basket to Euro Correction in world financial markets

Estimation Technique: 

Estimation Technique GMM; Instrumental variables* are: core inflation, GDP gap, unemployment gap, NEER and foreign producer prices (lagged from -1 to -3); Expectations are replaced with realizations. * Lagged from -1 to -3.

Estimated results for Traditional Phillips Curve: 

Estimated results for Traditional Phillips Curve

Estimated results for New Phillips Curve: 

Estimated results for New Phillips Curve

Estimated results for Hybrid Phillips Curve: 

Estimated results for Hybrid Phillips Curve

Hybrid Phillips Curve – structural model with coefficient normalization: 

Hybrid Phillips Curve – structural model with coefficient normalization

Conclusions (I of II): 

Conclusions (I of II) In Latvian inflation dynamics the weights of lagged and future expected inflation are found to be roughly equal (similar to, e.g., Hungary); This suggests that inflation in Latvia has more inertia, as compared to the Euro area or USA (according to GG 2001 in Euro area and USA lagged inflation matters much less for their inflation dynamics);

Conclusions (II of II): 

Conclusions (II of II) At the same time, price adjustments in Latvia are estimated to take place more often than in the Euro area and USA – approximately once in 6 months on average. According to GG 2001 in Euro area this period is about 3 years and in USA – approximately 1.5 year.