Forecasting Hierarchical Currency Circulation Data of Bank Indonesia Based on Two Levels ARIMAX Model
This study concern on forecasting monthly currency inflow and outflow data of Bank Indonesia, which are
hierarchical time series. Hierarchical time series forecasting needs special treatment to make the forecasts follow the
hierarchy rule. Firstly, base forecasts were calculated individually for all series by using two levels ARIMAX model with
Eid al-Fitr effect. The base forecasts were then revised by hierarchical methods including bottom-up, top-down and optimal
combination method. Finally, the performance of hierarchical methods were evaluated based on the out-of-sample root mean
squared error (RMSE) for all series. The results showed that the top-down method based on historical proportions and
optimal combination method perform best respectively for currency inflow and outflow data.
Keywords— Currency Circulation, Hierarchical Time Series, Bottom-Up, Top-Down, Optimal Combination.