Forecasting Macroeconomic Indicators of the Russian Economy Using Dimension Reduction Models

Anton A. Skrobotov – Researcher of the Russian Presidential Academy of National Economy and Public Administration, Candidate of Economic Sciences (Moscow, Russia). E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Aleksey V. Tsarev – Younger Researcher of the Russian Presidential Academy of National Economy and Public Administration (Moscow, Russia). E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

This study is devoted to the analysis of forecast power of dimension reduction models for Russian macroeconomic time series. Various approaches and methods for modeling and forecasting high-dimensional macroeconomic time series are reviewed and discussed. The approaches considered are applied to Russian data on 29 time series for the period from January 2000 to June 2019. A comparative analysis of the results indicates that in half of the cases, the random forest model is the best in terms of predictive power, and its use on average improves the quality of the constructed forecasts by 5%, 25%, and 30% for the short, medium, and long-term periods, respectively.

The article was written on the basis of the RANEPA state assignment research programme.

Key words: forecasting, forecast evaluation, Russian macroeconomic time series, LASSO, random forest.