Evaluating a leading indicator: an application-the term spread

This paper analyzes the procedures that have previously been used to evaluate indicators. These methods determine whether the indicator correctly classifies periods when there was (not) a recession. These approaches do not show whether or not an indicator signaled a turn or failed to predict it. This paper then presents a new approach and applies it to the term spread series. The results are mixed because the indicator predicts every recession but also generates a large number of false signals. This result may explain why economists do not always place great weight on this series.