Fat tails and tail risk

What fat tails mean

Fat tails means that very large price moves happen more often than a normal (bell-curve) distribution would predict. Tail risk is the risk of loss from those rare, extreme moves.

A distribution describes how often returns of each size occur. The tails are its far ends: big losses on the left and big gains on the right. In a "fat tail distribution," those ends hold more probability than the bell curve allows.

Precise definition

For experienced readers:

• A distribution is fat-tailed (leptokurtic) when its kurtosis is higher than that of the normal distribution. In other words, it has positive excess kurtosis.
• Heavy-tailed distributions have tails that decay more slowly than an exponential. Some follow a power law. For very heavy tails, the variance may not even be finite, so standard deviation stops being a reliable measure.
• Tail risk usually refers to the left tail: losses beyond what a model such as Value at Risk (VaR) with normal assumptions would flag.

Some details matter in practice:

• Skew and kurtosis are different. Returns can be both skewed and fat-tailed. For example, sharp drops can be larger than sharp rises.
• Volatility clustering. Calm periods followed by turbulent ones can make a distribution look fat-tailed overall, even if each period alone looks closer to normal.
• Unstable estimates. Tail estimates from a short price history are unreliable, because the extreme events are by definition rare in the sample.
• Correlations change in a crisis. During extreme moves, assets that usually move independently can move together (see correlation coefficient).

Worked example (hypothetical)

A trader measures the typical daily range of a currency pair from recent months. They size a leveraged position so that an ordinary bad day costs only a small part of the account. A normal model treats a move many times larger than the typical day as practically impossible.

Then an unexpected central bank decision hits. The price jumps by many multiples of its usual range within minutes, and the market gaps straight past the trader's stop-loss. The order fills at a much worse price. The loss is far larger than planned, and margin may drop to the stop-out level.

The model was not "unlucky." It underestimated how often such moves occur. Leveraged CFD and forex trading can lose money fast, and gaps like this are one of the main reasons.

To check how a price move translates into account money for your own position size, use the tools on the calculators page.

Where you see the term

• Platforms. Trading platforms do not usually label anything "fat tails." Strategy tester reports in MetaTrader do show maximal drawdown and runs of consecutive losses, which reflect tail outcomes.
• Risk reports. Banks and funds discuss VaR, expected shortfall (the average loss beyond VaR) and stress tests. Stress tests exist because normal-based VaR misses fat tails.
• News. Headlines use "tail risk," "tail hedge" and "black swan risk" around crashes, surprise policy decisions, broken currency pegs and a flash crash.

Related terms

• Black swan: an extreme event that almost nobody expected.
• Annualized volatility and average daily range: measures of typical movement, not of the tails.
• Financial contagion and flight to quality: how extreme moves spread across markets.
• Circuit breaker: exchange trading halts during extreme moves.
• Exposure and risk management: how position size relates to possible losses.
• More terms are in the trading glossary.

Frequent questions

What is fat tails in simple words?

Big market moves happen more often than the bell-curve model says they should.

What is the tail risk meaning?

It is the risk of loss from rare, extreme price moves at the far ends of the return distribution.

Is black swan risk the same as tail risk?

Not quite. Tail risk covers any extreme move. A black swan is an extreme event that was not anticipated at all.

Does a stop-loss remove tail risk?

No. In a gap or a fast market, a stop-loss can fill at a much worse price than the level you set.

Can a model measure fat tails reliably?

Only roughly. Extreme events are rare, so estimates based on limited price history carry large uncertainty.