Modelling of Consolidated Budget Tax Revenues of the Russian Federation

This study presents a comprehensive analysis dealing with factors affecting the tax system and determining the consolidated budget tax revenues of the Russian Federation. All models are constructed using monthly data during the period from January 2010 to September 2015. All nominal indicators are adjusted to September of 2015 real prices with the help of consumer price index. The research includes the study of the methodology and predictive properties of the naive, naive seasonal, structural models, autoregressive and simple moving average models, as well as autoregressive model with distributed lags in order to select the optimal one for forecasting budget tax revenues and to be able to obtain accurate forecasts regardless economic conditions. The study finds that the total tax revenue predictions based on the structural model elaborated with the help of the factors forming the tax bases of the main types of taxes are more accurate than those which are based on the autoregressive moving average model and autoregressive model with distributed lags, considering the period of economic recession. During the period of structural changes in the economy primitive naive forecasts are also close enough to the real values of the tax revenues, while in the immutability of the existing trends the most accurate predictions are obtained on the basis of autoregressive model with distributed lags. The results of testing the significance of differences between the forecasts obtained indicates that only construction of naive, naive seasonal predictions, and the forecasts on the basis of structural model pays off. The necessity of constructing other models is caused only by the increase of accuracy required and the aim of the research.