Measuring Historical Volatility
This paper by Louis H. Ederington (University of Oklahoma) and Wei Guan (University of South Florida St. Petersburg) proposes the adjusted mean absolute deviation as an alternative method for measuring historical volatility and provides tests of how well this method forecasts volatility, comparing it to traditional methods (standard deviation and GARCH).
The paper was first released in August 2004 and then published in Journal of Applied Finance, Vol. 16, No. 1, Spring/Summer 2006.
These free download links may no longer work:
The University of Oklahoma
Financial Management Association – Journal of Applied Finance