Forecasting national income based on a modified Keynesian model

The paper considers several ideas that can be useful in a systematic approach to the analysis of both the dynamics of macro indicators and regression models based on such series. Approaches to accounting for seasonal fluctuations in econometric modeling are analyzed. Numerical experiments with models of different types of seasonality accounting are carried out, conclusions and recommendations are formulated. The best results in practice are obtained using adaptive modeling based on the multiplicative model of seasonality accounting. The possibilities of constructing forecast estimates are shown on the example of studying the dynamics of current trends in the Russian economy. A modification of the Keynes model for the open economy is proposed, on the basis of which it is possible to estimate future levels of national income. The results of forecasting based on the modified model are shown in comparison with the actual levels.

Keynes model, GDP forecasting, seasonality allocation, regression model, Holt-Winters model