If it can be resolved into simple sinusoids, it might be possible to predict stock market behavior, even if only with partial success. But I am sure that there is a straight line component with a tilt of about +10% which would account for the fact that stock markets generally rise as time goes on.
Yes, FT will allow you roll out a spectrum of energy (frequencies). You should then find useful contrasts as you wade across markets of the world. Let’s wait for your paper.
Further, I would run FT on both equity and bond paths, and attempt to use the frequencies to study financial and economic shocks. FT is underused. So, well done for picking it up.
What Fred has written above is true. But for volatile market and fast fluctuations, it is better to use wavelet analysis instead.
@FredAmoya @RCMittal. Thanks for your comments. I think it would be useful to run autocorrelation analyses as well, in order to find any recurring "motifs" in the stock market behavior over time.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs