Forecasting Daily Transactions Volume for Financial Planning and Budgeting

Abstract

Appropriate transactions volume prediction is very crucial for making different Strategic and Tactical decisions at various management levels in any field. This paper analyzes the possible reasons behind daily transaction volume fluctuations for Bill Payment (BPay) and different factors influencing it. A forecasting solution around the Decomposition Concept of observed/explainable factors was developed, and a statistical Unobserved Components Model (UCM) was fit on the decomposed transactions volume line to forecast daily volumes for BPay in future time periods. Models were built for three main channels – Electronic, Paper Managed and Paper Unmanaged. The results show that error percentage (when forecasted volume was compared to actual volume) was less than 1% for Electronic and Paper Managed Channels whereas for Paper Unmanaged, the error was 3.6%.