Short-term forecasting of the russian ruble exchange rate using a time series model

This paper presents a comparative analysis of two time series models, ARIMA and ARIMAX, for the short-term forecasting of the Russian Ruble exchange rate. The relevance of this study is justified by the necessity to develop reliable forecasting tools for currency exchange rates within the context of contemporary economic challenges, including sanctions pressure, oil price fluctuations, and geopolitical instability. The research objective is to develop a time series model capable of performing short-term forecasts of the Ruble exchange rate, accounting for the current economic environment. The authors employ econometric analysis methods to construct two time series models where the dependent variable is the monthly average exchange rate of the RUB/USD currency pair. The ARIMA model does not account for the influence of exogenous variables, relying solely on the past values of the Ruble exchange rate, whereas the ARIMAX model incorporates exogenous variables to control for external factors. Utilizing statistical metrics, such as R² and MAPE, along with robustness checks via outlier testing and structural break analysis, the optimal model for short-term Ruble exchange rate forecasting was identified. The study results indicate that the ARIMA model is optimal for exchange rate forecasting. Based on the high explanatory power of both the ARIMAX and ARIMA models, the null hypothesis—that the Ruble exchange rate depends on its own past values—was confirmed. Consequently, the second hypothesis—that the Ruble’s dynamics are unpredictable and shaped by a multitude of qualitative factors difficult to incorporate into a model—was rejected.