Analysis of Composit Index Using the Time Series Model

The leading index gives useful information for the prediction of an economic event and the coincident index provides a summary of the state of actual economic activity. It has been observed that there exist regular and short time lags, usually 3 or 4 months, between the leading index and the coincident index. In this paper, therefore, we evaluate the index of leading indicators as a tool for prediction. From this standpoint, the relationships between the leading indicators and the coincident index are analyzed by both the regression model and the time series model (transfer function model). And the predictive performances of each model are compared.