Modelling and Forecasting Foreign Debt Using ARIMA Model: The Zambian Case from 2022 to 2035

Abstract

The study sought to model and forecast Zambian Government foreign debt from 2022 to 2035 using Autoregressive Integrated Moving Average Model. The secondary data of time series during the period of 1973 to 2021 on Zambia’s foreign debt are used as the basis of forecasting for the next 15 years by using ARIMA (Autogressive Integrated Moving Average) Model. The ARIMA (1, 1, 2) model was used due to its accuracy, mathematical soundness, and flexibility, thanks to the inclusion of AR and MA terms over a regression analysis. The results showed that ARIMA (1, 1, 2) is an adequate model which best fits foreign debt time series datadue to the smaller deviations in the mean absolute percentage error and mean square error and its errors are smaller than Moving Average (MA), Generalized Autoregressive Conditional Heteroskedasticity Model (GARCH), Simple Exponential Smoothing (SES), Error, Trend and seasonal model (ETS), Double Exponential Smoothing (DES), Seasonal Autoregressive Integrated Moving Average (SARIMA), Vector Autoregressive Model (VAR), Vector Error Correction Model (VECM), Threshold Autoregressive model (TAR), Triple Exponential Smoothing (TES), hybrid ARIMA, and Artificial Neural Network Model (ANN). The results show that the value of the Zambia’s foreign debt is predicted to keep increasing from 2022 to 2035 amounted to USD 80.5862 billion. The results also show that compared to government debt in 1973, within 49 years, Zambia’s foreign debt is predicted to rise by 209.34%.