Apparently stable increments in finance data: could arch effects be the cause ?

Abstract

The observed high kurtosis of stock market returns and other variables of a speculative nature has aroused interest in stable law distributions. This paper makes the point that most historical findings that returns indeed follow stable laws may have been caused by conditional heteroskedasticity. This presumption is enhanced by Monte Carlo simulations.

Keywords:

• Stable distributions
• conditional heteroskedasticity