Comparation of Alterantive Policy Rules in a Structural Model of the Czech Republic

Hledík, Tibor

The main goal of this thesis has been a study of alternative policy rules in a small structural model calibrated to capture the Czech economy. After the overview of the historic development of economic theory and structural modeling we have specified a small open economy model that has served as a main technical tool for the analysis. The model represents a framework, where forward-looking model-consistent expectations are formed with respect to the development of the exchange rate and interest rates. Inflation expectations are forward looking too with some nominal rigidities in inflation dynamics. The model's structure is relatively simple. The IS curve captures the dynamics of real GDP, that exhibits real rigidity, motivated by habit formation or investment adjustment costs. In our specification the real GDP is a function of (the deviation of) real XR, real IR and foreign demand (from corresponding equilibrium levels). The Phillips-curve is based on the F-M type wage setting behavior, therefore it enables to consider domestic prices, that are modeled as mark-ups over wages. CPI inflation then consists of domestic, imported and administered inflation, including the effect of any indirect taxes changes. The exchange rate is modeled by the UIP arbitrage condition. Exchange rate expectations are forward-looking, but with some inertia in expectation formation. Interest rates with one year maturity are also modeled as an arbitrage condition on the money market, they are fully model-consistently forward looking. The model is closed by a Taylor-type forward-looking policy rule. The interest rate exhibits some inertia and feeds back from deviation of inflation from target and output from its equilibrium. The specification (parameterization) of the rule is general enough to examine CPI and domestic inflation targeting. The model specification has been followed by empirical work leading towards the implementation of the previously specified model on Czech data. Based on the sources of the Czech Statistical Office, Czech National Bank, Consensus Economics Inc., we first processed the data by executing seasonal adjustment and other transformations necessary for being consistent with the definition of model variables. The database has been created by an automatic MATLAB based routine, therefore the calculations were relatively easy to update. The database being completed, we have set up a Kalman-filter for determining equilibrium values for the real interest rate, exchange rate and output. At the same time through Kalman filtering we identified all model residuals. We paid special attention to the decomposition of the output gap and discussing In order to assess the overall dynamic properties of the model and judge how well the model fits the data, we conducted several exercises. First we decomposed some of the important endogenous variables of the model to shocks to see, whether the identified shocks are in line with our intuition and episodes of the recent Czech economic history. We found, that the shocks are not in contrast with some of the clearly distinguishable episodes. After the shock decomposition we run in-sample simulations to see, how well the model is able to fit the reality two years ahead. We found the overall results quite encouraging. We were able to fit quite well the output gap as well as MP inflation. Domestic inflation has been slightly more inertial in model simulations than in reality, but even in this case the results were acceptable. The model was not able to fit the 2001-2 appreciation of the nominal XR footnote{Understandably it neither forecasted well the fast fall in inflation after the appreciation period.}, which is not a big surprise. The model calibration part of the thesis concludes, that the model fits the data and economic story reasonably well.

guest :: login Digital Repository
		Search		Submit		Help		About