| Popis: |
Beauty Contest Approach Measuring Output Gaps in Baltic States The thesis estimates the potential output and the output gap for the economy of the Baltic states using several different techniques. The research begins with the output gap estimation using two commonly adapted univariate statistical filters: Hodrick-Prescott filter and Christiano-Fitzgerald filter. The analysis then proceeds to a more complex setting, introducing the “Beauty contest” concept, focusing on the model specification rather than on a prior selection of the methodology (conventional model “horse-race”). This presents a methodological approach of univariate unobserved components model based on structural time series and Kalman filter moving on to a bivariate setting that uses additional macroeconomic variables helping to improve the estimation results. The macroeconomic variables, which could accommodate specific cycles (demand, supply, financial, external, fiscal), are selected for the bivariate model based on introduced “beauty contest” selection criteria. The comparison between all methods used indicates that the introduction of additional variables improves the feature of output gap estimates, resulting in smoother, more well-defined and visible cycles with distinctive extrema. |
