In partnership with:
Eurex
Swiss Financial Services

The Cycles of Financial Markets | David Gurwitz, Charles Nenner Research Center | #85

1 comment
  • Slav D July 9, 2015 at 8:16 am / Reply

    On a amplitude(#of bars of various time length ie weekly,quarterly,daily) vs frequency domain, at what cycle frequency do youe make the cut off in your selection of relevant cycles? I am no engineer but of little that i know in signal processing, Fourier transforms the time domain into frequency domain to help distinguish repeatable cycles of various amplitudes, and i am guessing amplitude would be the # of time bars of selected time length . One could also plot the bars in say individual stocks float turnover, and distinguish cycles that way. So for amplitude or Y axis you would have # of shares traded and for x axis a frequency of sinusoidal waves. Does Nenner always use Fourier to go from time to frequency domain or do you look at cycles in other variables as well? say volatility/volume, or sentiment/price etc? For trend-followers it may be nice to know when in time price range expansion/compression occurs and if repeatable cycles of price range are present and forecastable. Since even equity curves move up/down can we use cyclical analysis to predict manager/strategy performance? I believe that kinda cycle research would be extremely interesting.

    I hope these questions makes some sense Niels.

Post a comment