What it means
Kurtosis describes the shape of a distribution's tails rather than its average or its spread. Excess kurtosis simply rebases that measure so that a normal distribution scores zero, which makes it easy to see at a glance whether data is more extreme-prone than the textbook assumption.
The business relevance is risk modelling. Many standard tools, including simple value-at-risk calculations and option pricing models, assume returns follow a normal distribution, and if the real data has positive excess kurtosis those tools will systematically understate the chance of a large loss.
Interpretation follows three broad categories. A positive value, called leptokurtic, indicates fat tails and a sharper peak; a negative value, called platykurtic, indicates thin tails and outcomes clustered more evenly; and a value near zero suggests the normal assumption is reasonable.
Financial data almost always shows positive excess kurtosis. Daily equity returns, credit losses and commodity prices all produce more very large moves than a normal curve implies, which is why practitioners supplement standard models with stress tests and scenario analysis.
The nuance is that the measure is fragile in small samples. Because it depends on deviations raised to the fourth power, a single outlier can dominate the result, so a figure calculated from twelve monthly observations should be treated as a rough indication rather than a reliable statistic.
In practice
Real-world examples.
Example
A risk team at an asset manager calculates that a fund's daily returns show excess kurtosis of 4.2. It concludes that the standard value-at-risk figure understates tail risk and adds a stress test based on the three worst historical days.
Example
An insurance actuary reviewing storm claims finds strongly positive excess kurtosis in annual loss data. Most years are quiet and a few are catastrophic, so the actuary prices reinsurance on the tail rather than on the average year.
Example
A treasury team modelling daily cash balances finds excess kurtosis close to zero. Balances vary but rarely swing violently, so the team is comfortable using a normal assumption to size its overdraft buffer.
Formula
Calculation
Kurtosis = (sum of (each value - mean) to the fourth power / n) / (variance squared).
Excess kurtosis = kurtosis - 3.
Take nine monthly returns, expressed in percentage points: -9, -1, 0, 0, 1, 1, 1, 1 and 6.
Mean = (-9 - 1 + 0 + 0 + 1 + 1 + 1 + 1 + 6) / 9 = 0 / 9 = 0.
Because the mean is zero, each deviation is just the value itself.
Sum of squared deviations = 81 + 1 + 0 + 0 + 1 + 1 + 1 + 1 + 36 = 122.
Variance = 122 / 9 = 13.5556, so standard deviation is about 3.68 percentage points.
Sum of fourth powers = 6,561 + 1 + 0 + 0 + 1 + 1 + 1 + 1 + 1,296 = 7,862.
Fourth moment = 7,862 / 9 = 873.5556.
Variance squared = 13.5556 x 13.5556 = 183.754.
Kurtosis = 873.5556 / 183.754 = 4.75.
Excess kurtosis = 4.75 - 3 = 1.75.
A positive result of 1.75 says these returns have noticeably fatter tails than a normal distribution, which is unsurprising given that two of the nine months account for almost all of the variation.Case study
Seen in the real world.
The following is an illustrative, fictional case. Marlow Ridge Capital, an invented multi-strategy investment firm, reported to clients that its flagship fund had an annualised standard deviation of 6%, which sounded reassuringly modest next to equity market volatility. The risk committee, however, asked for the shape of the return distribution rather than just its width.
The analysis showed excess kurtosis of about 5.8 on monthly returns. Nearly all months fell within a narrow band of plus or minus 1%, but three months in six years had produced losses beyond -7%. The smooth average concealed a return profile that delivered small steady gains punctuated by rare severe drops, a pattern typical of strategies that sell insurance-like exposure.
Marlow Ridge changed two things. It began reporting the worst three months alongside the standard deviation in client materials, and it cut position sizes in the strategy that generated most of the tail exposure. Investor questions about volatility became far easier to answer honestly once the tail behaviour was disclosed rather than averaged away.
Watch out
Common mistakes.
- Confusing kurtosis with skewness. Kurtosis measures the weight of both tails together, while skewness measures whether the distribution leans to one side.
- Assuming high kurtosis means high volatility. A series can have a low standard deviation and still produce rare extreme moves, which is exactly what fat tails describe.
- Calculating it from a handful of observations and treating the answer as precise. Fourth powers make the measure extremely sensitive to a single outlier in a small sample.
Questions
People also ask.
Why subtract 3?
Because the kurtosis of a normal distribution is exactly 3, so subtracting it sets the normal case to zero and makes the number easy to interpret.
What counts as a high excess kurtosis?
There is no universal threshold, but values above roughly 1 suggest meaningfully fatter tails than normal, and daily financial returns often exceed 3.
Can excess kurtosis be negative?
Yes, a negative value means thinner tails than a normal distribution, with outcomes spread more evenly and extreme results rarer.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%