LEADING INDICATORS FROM FOURIER
by William T. Taylor
The purpose of this article is to expand upon my previous work July 1984 Technical Analysis of Stocks
& Commodities magazine) into an area not normally discussed as part of Spectral Analysis, that is, the
detection and measurement of phase (lead or lag) relationships of cyclical components from two or more
time series. In certain cases where there are dominant cyclical components of the same length in two or
more time series, it can be important to know if there is a lead or lag relationship on which to anticipate
changes in direction of one price based on the knowledge of, for example, the leading nature of another
I will begin with a discussion of the basic nature of cycles followed by a brief explanation of Spectral
Analysis. Emphasis is on how to interpret the results rather than the mathematics behind the procedure.
This topic is discussed at some length in the earlier article (mentioned above) and the reader is invited to
look there first for a better understanding of how this analytical tool is put to use.