External Reserves, ARIMA, SARIMA, SARIMA-X, ARDL, Statistical Loss Functions
This paper proposes three short-term forecasting models for the adjusted external reserves using the seasonal autoregressive integrated moving average (SARIMA), seasonal autoregressive integrated moving average with an exogenous input (SARIMA-X) and an autoregressive distributed lag (ARDL) processes. The performances of the proposed models are compared with the existing model obtained using an autoregressive integrated moving average (ARIMA) process using the pseudo-out-of-sample forecasting procedure over July 2013 to May 2014. The results show that SARIMA model outperformed the other models in three to six months forecast horizon, whereas ARDL model performs better in one to two months forecast horizon. Therefore, in forecasting external reserves in longer horizon, the paper concludes that seasonality should be accounted for by using the SARIMA model.
CBN Journal of Applied Statistics
Sani, Doguwa I. and Alade, Sarah O.
"On Time Series Modeling of Nigeria’s External Reserves,"
CBN Journal of Applied Statistics (JAS): Vol. 6
, Article 1.
Available at: https://dc.cbn.gov.ng/jas/vol6/iss1/1