Skip to main content
aifinhub
Risk & Portfolio Construction Formula

Kurtosis Formula

Kurtosis is the fourth standardized moment: the average of deviations from the mean raised to the fourth power, divided by the variance squared. It measures tail heaviness. A normal distribution has kurtosis 3, so excess kurtosis (kurtosis minus 3) above zero signals fatter tails and more frequent extreme moves than a normal model would predict.

By AI Fin Hub Research · AI Fin Hub Team
Best Next MoveCalculators

Returns Distribution Analyzer

Paste a returns CSV. Histogram, normal-overlay, QQ plot, skewness, excess kurtosis, Jarque-Bera test, tail-weight index. See why Sharpe alone misleads.

CalculatorOpen ->

On This Page

Formula

Copy the exact expression or work through it step by step below.

Kurt = ( (1/n) x sum_i (R_i - mean_R)^4 ) / sigma^4 Excess kurtosis = Kurt - 3 where sigma = population standard deviation (divisor n)

Variables

R_i

Period return

Individual return observation.

(R_i - mean_R)^4

Fourth-power deviation

Deviation from the mean raised to the fourth power. Because the exponent is even, all terms are positive and large deviations dominate enormously, so kurtosis is driven almost entirely by the tails.

sigma^4

Variance squared

The square of the variance, which standardizes the fourth moment so the measure is unitless and comparable across assets of different volatility.

Kurt - 3

Excess kurtosis

Kurtosis relative to the normal distribution's value of 3. Positive excess kurtosis (leptokurtic) means fatter tails; negative (platykurtic) means thinner tails. Asset returns are almost always leptokurtic.

Step By Step

  1. 1

    Compute the mean and population standard deviation of the returns.

    Returns -10%, 0%, 0%, +10% have mean 0%.

  2. 2

    Raise each deviation from the mean to the fourth power and average them.

    Fourth powers: 0.0001, 0, 0, 0.0001; mean = 0.00005.

  3. 3

    Square the variance to get sigma to the fourth power.

    Variance = (0.01 + 0 + 0 + 0.01)/4 = 0.005; sigma^4 = 0.005^2 = 0.000025.

  4. 4

    Divide the mean fourth-power deviation by sigma to the fourth, then subtract 3 for excess kurtosis.

    0.00005 / 0.000025 = 2.0; excess = 2.0 - 3 = -1.0.

Worked Example

Kurtosis of a symmetric four-return sample

Returns

-10%, 0%, 0%, +10%

Mean

0%

Deviations equal the returns. Fourth powers: (-0.10)^4 = 0.0001, 0, 0, (0.10)^4 = 0.0001. Sum = 0.0002; mean = 0.00005. Variance = (0.01 + 0 + 0 + 0.01)/4 = 0.02/4 = 0.005; sigma^4 = 0.005^2 = 0.000025. Kurtosis = 0.00005 / 0.000025 = 2.0. Excess kurtosis = 2.0 - 3 = -1.0.

Kurtosis of 2.0, or excess kurtosis of -1.0. This small, evenly spread sample is platykurtic (thinner-tailed than normal) because the values are bunched rather than tail-heavy. Real return series behave oppositely: daily equity returns commonly show kurtosis well above 3, meaning crashes and spikes happen far more often than a normal model assumes.

Common Variations

Sample (bias-corrected) kurtosis: spreadsheet KURT functions apply a small-sample adjustment and return excess kurtosis directly.
Robust kurtosis: quantile-based measures (such as the Crow-Siddiqui or Moors definitions) resist the extreme sensitivity of the fourth moment to outliers.
Jarque-Bera test: combines skewness and excess kurtosis into a single statistic for testing normality of returns.

Try These Tools

Run the numbers next

Sources & References

Related Content

Keep the topic connected

Planning estimates only — not financial, tax, or investment advice.