CEU eTD Collection (2022); Sivak, Tomas: Comparison of Inflation Forecasting Models: The Case of Slovakia

CEU Electronic Theses and Dissertations, 2022
Author Sivak, Tomas
Title Comparison of Inflation Forecasting Models: The Case of Slovakia
Summary I perform forecasting accuracy tests comparing six different inflation-forecasting models for Slovakia. Using an RMSE objective function, I find that on average the performance of the models is in the following order (from best to worst): i) the average of all individual methods, ii) the stationary VAR model, iii) the nonstationary VAR model, iv) the disaggregated ARMA model, v) the aggregated ARMA model, and vi) the factor model. Another contribution of this exercise is to show that univariate ARMA forecasts of aggregate inflation can be improved by forecasting the components of HICP inflation separately and then taking the appropriate weighted average of the component forecasts. While there is a large literature documenting the improved performance of combined forecasts, this is usually in the context of averaging across models rather than data components.
Supervisor Lieli, Robert
Department Economics MA
Full texthttps://www.etd.ceu.edu/2022/sivak_tomas.pdf

Visit the CEU Library.

© 2007-2021, Central European University