Building a regression model to determine the share of the US dollar in the structure of international reserves

Abstract

The article examines the share of Russian foreign currency deposits in the broad currency M2, as well as the volume of gold and foreign exchange reserves (hereinafter gold reserves) of the Russian Federation in the period from 2012 to 2022. The sample is presented by quarterly data, the basis for the study was the official reports of the Bank of Russia and the IMF. In addition, based on the data obtained, a forecast is made to determine the share of the US dollar in the structure of international reserves for the first quarter of 2023, which makes the regression model relevant and significant.
The aim of the work is to build a qualitative and adequate model that allows using it to predict the share of the US dollar in the structure of international reserves.
In the course of the study, such methods as regression and statistical analysis, grouping and clustering of data were used.
The study revealed the presence of multicollinearity in the model. In addition, the regressor “X2” (the volume of gold and foreign exchange reserves) had little effect on the endogenous variable, as a result of which it was excluded from the sample. After adjusting the model, a 95% forecast was made for the share of the US dollar in the structure of international reserves for the first quarter of 2023.
The topic of the work is very relevant and is due to the shifts taking place in the international financial system in recent years. The rampant rate of de-dollarization on the part of developing countries radically changes the structure of international reserves and forces US to look for analogues to the US dollar.