Панкратова Анастасия Александровна

The article presents methodology for making operational forecast of GDP of BRICS countries (Brazil, Russia, India, China and South Africa) based on the panel MIDAS method with lagged general correlating effects (LCCE). Balanced panel of data for 2013-2025 served as an empirical basis of the study. To solve the problem of asynchronous receipt of macroeconomic statistics, release pseudo-calendar was developed taking institutional features and hierarchy of publications of national statistical services into account. Comparative analysis of forecast accuracy (MAFE and RMSFE metrics) with basic AR(1) models and connective equations confirmed the effectiveness of PMIDAS approach with the account of heterogeneity. The empirical results indicate general reduction in forecasting errors. However, the revealed heterogeneity of estimates for individual countries indicates the necessity to take structural heterogeneity of the BRICS economies into account when specifying the model.