Fat-Tailed Distributions
Textbooks teach that returns form a pretty bell shape (a normal distribution). But the real market produces extreme values far more often than that bell shape. Why is this mismatch dangerous?
The gap between the normal-distribution assumption and reality
Many financial theories (e.g., modern portfolio theory) assume that returns follow a normal distribution. In a normal distribution, events far from the mean should occur only very rarely, at a predictable probability.
But when you gather actual stock returns, both ends (the tails) of the bell shape are far chunkier than theory predicts. In other words, extreme events like crashes or surges appear more often than a normal distribution predicts.
A distribution with such fat tails is called 'fat tails' or a 'fat-tailed distribution.'
Measuring tail thickness with kurtosis
How fat the tails are is measured with a statistic called kurtosis.
- A normal distribution has a kurtosis of 3. - If kurtosis is greater than 3 (excess kurtosis > 0), it means the tails are fatter than a normal distribution, and such a distribution is called 'leptokurtic.'
Stock returns mostly have a kurtosis greater than 3. The middle is more peaked than a normal distribution (more ordinary days), and at the same time the tails are fatter (more extreme days too). The market's character of 'calm in normal times, occasionally explosive' comes from here.
Actual stock returns not only have high kurtosis but often have negative (-) skewness. It means big declines come slightly more often and more forcefully than big rises. Many of the top daily-move events for the S&P 500 were declines, not rises.
The lesson fat tails give investing
Acknowledging fat tails changes a few things.
First, you can't rely on standard deviation alone. Standard deviation summarizes risk well when assuming a normal distribution, but if the tails are fat, the 'normal-times volatility' may look low while the actual worst case is far worse.
Second, you have to treat 'events that should almost never happen in theory' as things that can naturally occur. A 6-sigma event (once in millions of years in a normal distribution) happens several times even over decades in financial markets.
Third, you have to look together at measures that directly view the 'tails'—the maximum drawdown (MDD), tail risk, and conditional VaR—to see the true risk.
常见问题
Q. Is an asset with high kurtosis unconditionally risky?
High kurtosis merely means 'extreme values appear often'; whether that extreme is up or down is told by skewness. That said, most risky assets have a fat left (loss) tail and negative skewness, so high kurtosis is often connected to 'the possibility of large losses.' You have to look at kurtosis and skewness together.
Q. If I know about fat tails, can I avoid crashes?
No. Fat tails don't tell you 'when' an extreme event will come. They only tell you 'that it will come' and 'that it comes more often than you think,' letting you set a bearable weight and a plan to endure in advance. It's a tool for preparation, not prediction.
相关页面
📋 结果基于历史数据计算,过去的收益不代表未来的收益。
📋 本服务旨在帮助理解投资、供教育之用,并非投资建议。