Diagnosing and Predicting Inflation in Poland with Application of Markov Switching Models and Future Inflation Indicator

Abstract

We investigate the inflation process in Poland during the economic transition period by applying Markov Switching Models to explain the inflation generating process. The analysis is limited to the period between March 1992 and October 2005, a period of relatively high inflation The application of non-linear models is a consequence of the transition period connected with high level of inflation. According to the Ball-Friedman hypothesis variation of inflation during such periods can be highly unstable. The results show that non-linear models significantly improve the description of inflation generating process during that period in Poland. In the analysis were included not only univariate Markov Models, but also such a model that incorporates leading information included in Future Inflation Indicator (FII). The most significant result of the research is that the model, where lagged values of FII are included as exogenous variables outperforms the univariate Markov Model in modeling inflation.

Keywords: Business Cycles, Inflation Cycles, Composite Leading Indicators, Markov Switching Models

JEL Classification: C5, E3