VARI-X model for currency inflow and outflow forecasting with Eid Fitr effect in Indonesia

Autoregressive Integrated Moving Average (ARIMA) is one of the commonly used approaches to model univariate time series. This model can be extended to Vector ARIMA (VARIMA) for modeling multivariate time series. In many application, for instance in economic research, the AR part plays more important role not only for forecasting but also for interpretation. For this reason, the VARI model was employed in this research to forecast the currency inflow and outflow in Indonesia. This work also considered the effect of Eid Fitr as exogenous variables. The month when Eid Fitr happens as well as the months before and after the Eid Fitr happen were regarded as input variable. Such a model used in this work is called VARI-X. The result showed that the dynamic of currency inflow was influenced by the amount of outflow, conversely. Moreover, the empirical evidence also proved that Eid Fitr has a significant influence on both inflow and outflow. It can be concluded that people tend to spend more money in the month when the Eid Fitr happens and one month before, while in the month after Eid Fitr people tend to save their money in the Bank.