Definition
Downside deviation measures the size of returns below a stated target by taking a lower partial second moment. A common sample convention uses the square root of the average squared shortfall over all n observations, assigning zero to observations above target. Other denominators and target choices exist and must be named.
How it is calculated
Downside deviation relative to target T = square root of the average squared negative shortfall min(0, R_i minus T) across n observations.
| Symbol | Meaning | Unit |
|---|---|---|
| DD | downside deviation | return per period |
| T | target return | return per period |
| R | observed return | return per period |
| n | number of observations | count |
Interpretation
Only misses below the target contribute to this measure. A calm stream above target does not add downside variation, while repeated shortfalls do. The target is part of the risk definition rather than a formatting choice.
Assumptions and conventions
-
Target, frequency and denominator convention are fixed.
-
The lower partial moment is the intended risk measure.
Return per period, same as R and T.
Limitations and common misreadings
-
Downside deviation is not automatically a sample standard deviation.
-
Changing the target changes the result.
References
- Frank A. Sortino, Lee N. Price. (1994). Performance Measurement in a Downside Risk Framework. The Journal of Investing. Whole article. doi:10.3905/joi.3.3.59. Source

