Abstract
This article deals with the issue of to what extent local maxima of one time series can be used to predict the turning points of other series of interest. It is argued that the theory of high-level crossings is the appropriate framework to handle such issues. The article then defines precisely the notion of a turning point and introduces a test for investigating the predictive value of the U.S. leading-indicator series.

This publication has 0 references indexed in Scilit: