Choosing Risk-Adjusted Performance Measures
How Sharpe, Treynor, Jensen’s alpha, information, and Sortino measures connect returns to different kinds of risk—and how to interpret them cautiously.
Match the Measure to the Risk
Why the risk measure matters
Return alone does not show how much risk a portfolio took to earn it. Risk-adjusted measures compare return with a chosen measure of risk, so they help answer different questions about performance. The right measure depends on the risk you want to evaluate and the comparison you are making. Use returns measured over the same period and expressed on a consistent basis.
The table distinguishes the measures by what they compare with return and when they are suited to use. Here, portfolio return is , the risk-free rate is , portfolio standard deviation is , portfolio is , market return is , benchmark return is , and target return is . () is the volatility of active returns relative to the chosen benchmark. Downside deviation () captures downside variation relative to the selected target.
| Measure | Risk or comparison | Formula | Suited to |
|---|---|---|---|
| Sharpe ratio | Total volatility | (R_p − r_f) / σ_p | Comparing complete portfolios, especially those not fully diversified |
| Treynor ratio | Systematic, market-related risk | (R_p − r_f) / β_p | Comparing diversified portfolios where unsystematic risk is relatively small |
| Jensen’s alpha | Return above or below the CAPM prediction | R_p − [r_f + β_p(R_m − r_f)] | Assessing performance relative to a market-risk model |
| Information ratio | Benchmark-relative risk (tracking error) | (R_p − R_b) / TE | Evaluating a portfolio relative to a chosen benchmark |
| Sortino ratio | Downside variation relative to a target | (R_p − T) / DD | Focusing on shortfalls rather than all return variation |
The Sharpe ratio uses total volatility, treating upside and downside variation alike. The Treynor ratio uses instead, so it focuses on market-related risk. is different in form: it is not a ratio, but the portfolio’s return minus the return predicted by the Capital Asset Pricing Model (CAPM). The information ratio compares active return—portfolio return above the benchmark—with . The Sortino ratio compares return above a chosen target with downside deviation.
Read Ratios Through Their Denominators
A higher return need not mean a higher Sharpe ratio
The Sharpe ratio is excess return divided by total volatility:
In the source example, Portfolio A returns 12%, with a 4% risk-free rate and 16% volatility. Portfolio B returns 10%, with the same risk-free rate and 8% volatility.
For A, subtract the risk-free rate to find excess return, then divide by volatility:
For B:
A earned the higher raw return, but B earned more excess return per unit of total volatility. The denominator changes the comparison: B’s volatility is low enough relative to its excess return to produce the higher Sharpe ratio. This result compares performance on that measure; it does not by itself establish which portfolio is better for every investor.
Interpret Results with Care
Read each result in context
A higher Sharpe or Sortino ratio indicates more excess return for the risk each measures. A higher Treynor ratio indicates more excess return per unit of , and a higher information ratio indicates more active return per unit of . Positive indicates that the portfolio returned more than CAPM predicts; negative alpha indicates less. These measures are not interchangeable, so a higher value on one does not automatically mean stronger performance on every other risk dimension.
Each measure also depends on choices or assumptions that matter to its interpretation:
Standard deviation counts upside and downside variation alike. It may not describe unusually asymmetric returns.
depends on the market benchmark and period analyzed, and does not capture every risk.
depends on the benchmark chosen for the comparison.
Downside deviation depends on the selected target.
Ratios estimated from short histories can mislead, and returns must be comparable for the comparison to be meaningful. Consider the reliability of the data, the underlying return history, portfolio goals and exposures, fees, and the investment objective. Treat these measures as tools for examining performance—not as standalone rankings or proof that one portfolio suits every investor.
: A measure that compares return with a selected measure of risk.