?
К вопросу о повышении качества валютных продаж Банка России с использованием математического моделирования
This article considers the question of improving the quality of foreign exchange interventions conducted by the Central Bank of Russia in the domestic market. Quality is understood as the amount of rubles withdrawn from circulation resulting from the actual currency sale. The period under study is autumn of 2014, which was the most active period of currency sale. In addition to public data on interventions, the model uses information on the cost of Brent crude oil in the world market, as well as the official rate of the US dollar to the ruble set by the Bank of Russia. The purpose of the model-building is to study the possibilities of improving the quality of currency sale by taking into account short-term forecast for cost of oil futures. We regard the following as premises for studying this approach: the established fact that the US dollar rate to the ruble depends on the oil price, the relation between this price and budgetary framework of the Russian Federation, as well as the possibility of the CBR's return to thepolicy of supporting the ruble through regular currency interventions. Short-term forecast is constructed with the use of the neural network trend indicator, whose architecture contains basic axioms and principles of technical analysis. The values obtained in the course of study as well as actual known values allow us to estimate the possible volume of withdrawn rubles for a given volume of currency sale. Proceeding from the inverse relation between the dollar rate and the oil price, in case of short-term forecast for fall in oil prices, it is proposed to conduct currency sale transactions not on the current but on the following banking day. Study of the model has shown that in case of active interventions, the number of withdrawn rubles can be increased due to taking into account forecast for the world oil price.