Forecasting Bank Customer Dynamics in Cameroon using ARIMAX Modelling Approach

Cameroon's banking sector has expanded, increasing the need for reliable forecasts of customer growth to support planning and resource allocation. This study applies quantitative time-series analysis to the reported annual number of commercial-bank borrowers per 1,000 adults in Cameroon for 1973–2022. An autoregressive integrated moving average model with exogenous variables (ARIMAX) was estimated using LogGDP and inflation as external regressors. The Augmented Dickey-Fuller results classify the customer series and inflation as stationary at level, while LogGDP is reported as integrated of order one. Descriptive results indicate substantial variation in the customer series and marked volatility in inflation. The fitted model shows that the first lag of the dependent variable is positive and statistically significant (p = 0.009), indicating strong historical persistence in the reported number of borrowers. By contrast, the coefficients of LogGDP and inflation are not statistically significant within the selected specification. Residual diagnostic results indicate that autocorrelation remains at early lags, suggesting that the model does not fully capture the temporal structure of the series. The study therefore finds that recent historical values provide useful information for modelling short-term movements in formal credit participation, but the current specification should be regarded as preliminary. Verification of the annual series, evaluation of alternative lag structures, and formal out-of-sample forecast assessment are required before the model can be used as a validated forecasting system.