期刊名称：Central European Journal of Economic Modelling and Econometrics
印刷版ISSN：2080-0886
电子版ISSN：2080-119X
出版年度：2017
期号：1
页码：29-67
语种：English
出版社：Polska Akademia Nauk
摘要：The aim of this paper is to examine the problem of existing seasonalvolatility in total and disaggregated HICP for Baltic Region countries (Denmark,Estonia, Latvia, Finland, Germany, Lithuania, Poland and Sweden). Usingnonparametric tests, we found that in the case of m-o-m prices, includingfruit, vegetables, and total HICP, the homogeneity of variance during seasonsis rejected. Based on these findings, we propose an exponential smoothingmodel with periodic variance of error terms that capture the repetitiveseasonal variation (in conditional or unconditional second moments). In apseudo-real data experiment, the short-term forecasts (nowcasting) for theconsidered components of inflation were determined using different specificationsof considered models. The forecasting performance of the models was measuredusing one of the scoring rules for probabilistic forecasts called logarithmicscore. We found instead that while the periodic phenomenon in variancewas statistically significant, the models with a periodic phenomenon invariance of error terms do not significantly improve forecasting performancein disaggregated cases and in the case of total HICP. The simpler modelswith constant variance of error term have comparative forecasting (nowcasting)performance over the alternative model.
