Tail risk
Also called fat-tail risk, extreme-loss risk
The risk of rare, extreme losses far larger than normal swings, from the 'tail' of the return distribution. Value-at-risk can miss it, which is why Basel rules for bank trading books moved to expected shortfall.

On this page· 5 sections
Tail risk is about the rare day or week whose loss is far bigger than anything in the usual range of results.
The name comes from a picture of returns as a curve. Ordinary outcomes bunch in the middle, and the extreme gains and losses sit in thin strips at either end, the tails.
How it works
Everyday risk measures describe the middle of the curve. Standard deviation assumes a symmetrical return distribution and treats upside and downside deviations alike[3]. The Sharpe ratio is built on mean and variance, so it is designed for normal return distributions and is not suitable for asymmetric ones[2]. A measure that describes the middle can say little about the tail.
A history can mislead for the same reason. In Kidd’s CFA Institute example, the Japanese equity market had ten years of only positive annual returns from 1980 to 1989, along with 36 months of negative returns, the worst of them −12%. In 1990 the market fell 39%, a loss the small sample of negative months would not have predicted[3].
Regulators treat the tail as a separate problem. The Basel Committee’s summary lists among the issues with the earlier, VaR-based approach an insufficient measurement of tail risks[1]. Its revised framework uses expected shortfall, a measure of the average of all potential losses exceeding the VaR at a given confidence level, which makes up for VaR’s shortcomings in capturing the risk of extreme losses, that is, tail risk[1]. The expected shortfall for capital is calibrated to a period of stress[4], and banks that use internal models must run stress tests that include low-probability events[4].
Example
Say a strategy gains 0.2% on 98 days out of every 100 and loses 5% on the other 2.
- Total: 98 × 0.2% − 2 × 5% = 19.6% − 10% = +9.6%, an average of 0.096% a day.
- The standard deviation of the daily results is about 0.73%.
- A −5% day sits about seven standard deviations below the average, and costs 25 times what a normal day earns.
- The two bad days erase 10 of the 19.6 points the other 98 days earned, about half.
The average day looks calm. The two days in the tail decide the result.
Watch out
A calm record says little about the tail. In the Japanese example the sample’s worst month was −12% and the next year’s loss was 39%[3].
Averages hide it. A strategy can make small, steady gains and still carry a rare large loss, as in the example.
A measure that ignores losses beyond its threshold ignores the tail. That is the gap the Basel Committee describes for VaR[1].
Sources
- International bodyBank for International Settlements (Basel Committee on Banking Supervision)The market risk framework: In brief (revised market risk framework, January 2019)Checked November 9, 2026
- Professional bodyCFA InstituteThe Sharpe Ratio and the Information Ratio (Deborah Kidd, CFA, Investment Performance Measurement, 2011)Checked November 9, 2026
- Professional bodyCFA InstituteThe Sortino Ratio: Is Downside Risk the Only Risk that Matters? (Deborah Kidd, CFA, Investment Performance Measurement, 2012)Checked November 9, 2026
- International bodyBank for International Settlements (Basel Committee on Banking Supervision)Minimum capital requirements for market risk (BCBS d457, January 2019)Checked November 9, 2026
