Chaotic Behavior of Financial Time Series-An Empirical Assessment
Ravindran Ramasamy, Mohd Hanif Mohd Helmi

Abstract
Financial Time Series often exhibit either chaotic or persistent or mean reversing behaviour. This behaviour could be quantified through Chaotic Exponent which ranges from zero to one. The rescaled range technique developed for hydrology by Hurst is applied in financial time series to estimate chaotic exponents which determines whether financial time series behaviour is purely chaotic white noise or any pattern exists. We have computed Chaotic Exponent coefficient for nine shares traded in Kula Lumpur Stock Exchange, nine popular stock market indices and nine selected exchange rates. As per theory pure random time series which behaves in a chaotic form cannot be forecasted, but somewhat skewed or non-normal financial time series could be forecasted. The forecasts could be used in several financial decisions like pricing of derivatives and very useful in hedging decisions. Our results indicate that the chaotic exponent of the selected financial time series is not consistent. They are either mean reversing or persistent, occasionally they show pure randomness. This proves that the forecasting of financial time series is relevant for hedging decisions.

Full Text: PDF