Yearly Volatility

This post considers yearly volatility. This is the histogram of the yearly log-returns of GE stock price, compared with the normal distribution with the same mean and standard deviation:



Since the standard deviation of GE daily log returns is 3.848, I should expect a standard deviation of the GE monthly log returns to be close to 3.848 times the square root of 1/256, that is 0.242. In fact, the mean is 0.090, the standard deviation is 0.219, the skewness is -0.450 and the kurtosis is 3.490.

I generate 21 day mean, historical volatility, skewness and kurtosis as usual by calculating these quantities for a sample of 21 consecutive days and sliding this sample over the entire range of available data.



I calculate the same quantities for a log-normal random walk with the same standard deviation and mean as the GE data.



The mean and historical volatility histograms of the random walk are, of course, in good agreement with the theoretical distributions (the "chi distribution" is in fact the histogram of the historical volatility of a log-normal random walk 10 times larger):



The histograms of the GE data are pretty much in agreement as well:



Categories: stock options, volatility
Technorati Tags: stock options, volatility