A review of the contemporary mainstream literature on exchange rate modelling clearly indicates that the rational expectations hypothesis (RE) is almost invariably taken as a point of reference in empirical investigations. This paper tests the RE hypothesis for the Polish foreign exchange market within the Roman Frydman and Michael Goldberg model that builds on the hypothesis of imperfect knowledge economics (IKE). The employed modelling strategy consists in the formulation of assumptions about the persistence of nominal rate, prices and interest rates and of the verification of competing scenarios congruent with RE and IKE. As a result of the analysis, the RE hypothesis is rejected in favour of the IKE alternative.
The aim of this paper is to examine the problem of existing seasonal volatility in total and disaggregated HICP for Baltic Region countries (Denmark, Estonia, Latvia, Finland, Germany, Lithuania, Poland and Sweden). Using nonparametric tests, we found that in the case of m-o-m prices, including fruit, vegetables, and total HICP, the homogeneity of variance during seasons is rejected. Based on these findings, we propose an exponential smoothing model with periodic variance of error terms that capture the repetitive seasonal variation (in conditional or unconditional second moments). In a pseudo-real data experiment, the short-term forecasts (nowcasting) for the considered components of inflation were determined using different specifications of considered models. The forecasting performance of the models was measured using one of the scoring rules for probabilistic forecasts called logarithmic score. We found instead that while the periodic phenomenon in variance was statistically significant, the models with a periodic phenomenon in variance of error terms do not significantly improve forecasting performance in disaggregated cases and in the case of total HICP. The simpler models with constant variance of error term have comparative forecasting (nowcasting) performance over the alternative model.
The paper investigates Bayesian approach to estimate generalized true random-effects models (GTRE). The analysis shows that under suitably defined priors for transient and persistent inefficiency terms the posterior characteristics of such models are well approximated using simple Gibbs sampling. No model re-parameterization is required. The proposed modification not only allows us to make more reasonable (less informative) assumptions as regards prior transient and persistent inefficiency distribution but also appears to be more reliable in handling especially noisy datasets. Empirical application furthers the research into stochastic frontier analysis using GTRE models by examining the relationship between inefficiency terms in GTRE, true random-effects, generalized stochastic frontier and a standard stochastic frontier model.
en
pl
