Marcin Błażejowski

Modelling and Forecasting Business Cycle in CEE Countries using a Threshold Approach

We propose to apply a time-series-based nonlinear mechanism in the threshold autoregression (TAR) form in order to examine business cycles in Central and Eastern European economies and compare them to the entire EU business cycle. The threshold variables, such as consumer price index, short and long interest rates, unemployment rate and an exchange rate vs. the U.S. Dollar, have been considered. The purpose of the paper is to model and to predict business cycles in Central and East European (CEE) economies (the EU Member States) and compare them to business cycles of the entire EU28 area and Eurozone EU19. We found that the exogenous mechanism played an important role in diagnosing the phases of business cycles in CEE economies, which is in line with the entire EU economic area. The results of business cycle forecasting using bootstrap technique are quite promising, while bootstrap confidence intervals are used for diagnosis.

Automatyczna procedura budowy specyfikacji zgodnego dynamicznego modelu ekonometrycznego w oprogramowaniu Gretl

In the last years we can observe intensive development of automatic model selection procedures. Best known are PcGets and RETINA. Such intensive work encourage to work on a new procedures. The concept of Congruent Modelling, formulated by Prof. Zygmunt Zielinski, is a very good framework for such development, including programming work, as well as many theoretical considerations. In the paper we present our concept of algorithm for automatic congruent modelling procedure and propose its implementation in Gretl.