Kurtosis and Fat-Tail Risk in Financial Return Distributions
Summary
The document explains kurtosis as a measure of how heavy or light a return distribution’s tails are relative to a normal distribution. It distinguishes ordinary kurtosis from excess kurtosis, for which the normal distribution is the zero baseline, and classifies distributions as mesokurtic, platykurtic, or leptokurtic. A central clarification is that kurtosis describes tail behavior, not simply the height or sharpness of a distribution’s peak; because deviations are raised to the fourth power, extreme observations can strongly affect the measure.
Examples compare normal, uniform, and Laplace distributions, then discuss asset returns and the use of heavier-tailed models such as Student’s t when estimating risk. The article also compares standardized S&P 500 ETF and Bitcoin returns to show that higher volatility does not necessarily imply higher kurtosis. It argues that tail risk should inform performance and risk evaluation, while the examples do not establish that any particular distribution will reliably predict future extremes.
Key ideas
- Kurtosis measures tail heaviness relative to a normal distribution rather than peak height.
- Excess kurtosis subtracts the normal distribution’s kurtosis, making zero the normal baseline.
- Positive excess kurtosis indicates heavier tails, while negative excess kurtosis indicates lighter tails.
- Extreme observations can have a large effect because deviations are raised to the fourth power.
- Volatility and kurtosis are distinct, so a more volatile asset need not have greater relative tail heaviness.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.