Title: Forecasting unemployment rate in Poland with dynamic model averaging and internet searches
Authors: Krzysztof Drachal
Addresses: Faculty of Economic Sciences, University of Warsaw, Poland
Abstract: The aim of this research is to estimate dynamic model averaging (DMA) model for unemployment rate in Poland. One can find multiple potential factors influencing unemployment rate. They can be significantly affecting unemployment rate only in certain periods. Therefore, a method incorporating time-varying parameters as well as the model uncertainty itself seems desirable. Additional aim of this research is to incorporate the Google search data into the econometric model. In this research, DMA is not able to significantly beat ARIMA model in case of forecast accuracy. Despite DMA success in other fields, for unemployment forecasting, this method seems vulnerable.
Keywords: Bayesian models; dynamic model averaging; DMA; macroeconomic time-series; unemployment forecasting; Google search volume index; Poland.
Global Business and Economics Review, 2020 Vol.23 No.4, pp.368 - 389
Received: 03 May 2019
Accepted: 23 Oct 2019
Published online: 28 Oct 2020 *