What is the Predicative Ability for Unemployment Rate of Different Methodologies for Output Gap Estimation?

The submitted diploma thesis is dedicated to the calculation of the estimation of output gap by three estimating techniques (Hodrick-Prescott filter, Kalman filtering, and Cobb-Douglas production function) for the Czech Republic during the years 1995-2009, and, consequently, to their usefulness for testing the Okun's law as a suitable predicative tool. In other words, which of the mentioned methods is the most appropriate one for an accurate prediction of the rate of unemployment. To achieve this, the Box-Jenkins methodology is used for modelling and designing the cyclical unemployment forecasts and dynamic version of the Okun's law. The results indicate that, for the given period, there is relatively strong and unstable relationship between changes in output gap and changes in cyclical unemployment in the Czech Republic. The most suitable technique for the estimating of the future development of unemployment gap seems to be the Cobb-Douglas production function approach. This is namely based on the quantity of gained forecast errors and Granger causality