The Sharpe ratio measures how much return an investment earned above a risk-free benchmark, usually a short-term Treasury bill, for each unit of volatility it carried. William Sharpe introduced it in 1966 as the "reward-to-variability ratio" for judging mutual funds; the measure caught on, and the name that stuck is his own. A higher ratio means more excess return per unit of risk, which puts a calm fund and a volatile one on the same scale. It only works when every ratio in a comparison covers the same period, uses the same bill rate and is calculated the same way. The stocks hub covers the fund basics behind it.
How is the Sharpe ratio calculated?
Subtract the bill rate from each period's return, then divide the average of those excess returns by their standard deviation.
That is the historic, or ex post, version in Sharpe's 1994 restatement: the average differential return over the period divided by how much it varied. The bill rate should match the return interval and the dates, so monthly returns pair with the 1-month bill yield over the same months. What is the risk-free rate explains that choice, and T-bill rates today shows current yields. Because the bill rate barely moves from one month to the next, the standard deviation of excess returns comes out close to the plain standard deviation of returns.
Sharpe recommends measuring over short periods, such as months, and then annualizing. If monthly excess returns are independent of one another, their average grows in line with the number of months and their standard deviation with its square root, so the monthly ratio times √12, about 3.46, gives the annual figure. A monthly Sharpe ratio of 0.25 becomes about 0.87 a year.
Fact sheets rarely print the ratio, but they supply the inputs for a shortcut: the annualized return minus the average bill yield, divided by the annualized standard deviation. It does not match the textbook method exactly. Run both ways on the same 36 months of US stock market returns to September 2026 from Kenneth French's data library, the shortcut gives 1.41 and the textbook method 1.31, so compare only ratios calculated the same way.
How do you compare two funds with the Sharpe ratio?
Work out each fund's excess return over the same bill rate, divide it by that fund's own volatility, and prefer the higher ratio only if the periods match.
Worked example: Vanguard's growth and value ETFs, three years to September 30, 2026
Both funds track large-company US indexes from Morningstar and charge 0.03% a year. Returns are annualized at net asset value, and standard deviation is Vanguard's figure from the last 36 monthly returns. The bill rate is the 4.58% average 1-month Treasury yield over the same three years in FRED's daily series.
| Morningstar Growth ETF (VUG) | Morningstar Value ETF (VTV) | |
|---|---|---|
| Annualized return | 26.30% | 18.75% |
| Minus the 1-month bill average | 4.58% | 4.58% |
| Excess return | 21.72 points | 14.17 points |
| Standard deviation | 16.84% | 11.31% |
| Sharpe ratio | 21.72 ÷ 16.84 = 1.29 | 14.17 ÷ 11.31 = 1.25 |
The growth fund earned 7.55 points a year more, but it also swung about half as much again, so per unit of volatility the two finished nearly level.
A higher ratio has a concrete meaning. Sharpe's 1994 article shows that an investor can combine the higher-ratio fund with Treasury bills to match the other fund's volatility and expect more return. Here, in the ratio's arithmetic, about two-thirds in the growth fund and one-third in bills would have matched the value fund's 11.31% volatility and earned roughly 19.2% a year, against 18.75%. Vanguard's S&P 500 ETF, which holds both kinds of company, scored 1.47 on the same basis: 22.85% minus 4.58%, divided by 12.43%. The ratio judges each fund alone, though, and ignores how two funds would behave together, a limit Sharpe himself points out. Growth vs value investing explains what separates the two styles.
How is the Sortino ratio different?
The Sortino ratio divides return above a target by downside deviation, which counts only returns below that target, so large gains no longer count as risk.
Downside deviation sets every period above the target to zero, squares the shortfalls, averages them over all periods with the zeros included, and takes the square root. The target can be zero, the bill rate or a return a pension plan needs to earn. Frank Sortino developed the measure in the early 1980s. A paper by the commodity trading adviser Red Rock Capital, hosted by CME Group, warns that it is often calculated wrongly by dropping the zero periods, which hides how often an investment fell short.
Worked example: Same Sharpe ratio, different Sortino ratios
Two hypothetical funds over five years, with a target return of 0%.
| Year | Fund X | Fund Y |
|---|---|---|
| Years 1 to 4, each | −2% | 6% |
| Year 5 | 18% | −14% |
| Average | 2% | 2% |
| Standard deviation | 8.94% | 8.94% |
| Downside deviation | 1.79% | 6.26% |
| Sortino ratio | 2 ÷ 1.79 = 1.12 | 2 ÷ 6.26 = 0.32 |
The funds share an average and a standard deviation, so their Sharpe ratios match at any bill rate. Fund X loses a little often and gains a lot once. Fund Y gains steadily and then takes one large loss, the pattern of strategies that sell options. Only the Sortino ratio tells them apart.
How does the Sharpe ratio compare with similar measures?
The Sharpe ratio scales excess return by total volatility; its close relatives swap in downside risk, beta or a benchmark index.
| Measure | Return above | Divided by | Best suited to |
|---|---|---|---|
| Sharpe ratio | Treasury bills | Standard deviation of excess returns | A whole portfolio, or a fund held on its own |
| Sortino ratio | A chosen target | Downside deviation below the target | Returns with lopsided gains or losses |
| Treynor ratio | Treasury bills | Beta | One holding inside a diversified portfolio |
| Information ratio | A benchmark index | Tracking error | An active manager judged against an index |
| Alpha | The return beta predicts | Nothing, since it is a return gap | Performance after market exposure |
Sharpe's 1994 article treats the information ratio as the same calculation with a benchmark index in place of bills. The beta and alpha entries cover the market-based measures, and drawdown covers the peak-to-trough losses that volatility can understate.
What are the pitfalls of the Sharpe ratio?
The ratio assumes that an average and a volatility tell the whole story, so lopsided returns, smoothed prices, a chosen window or a negative result can all mislead.
- Non-normal returns. Sharpe built the ratio on the mean-variance framework, which treats the average and standard deviation as enough to judge an investment. A strategy that collects small gains and occasionally takes a big loss, like Fund Y, can post a high ratio until the loss arrives, and the Red Rock paper notes that standard deviation understates risk for that kind of return pattern.
- Smoothed returns. The √12 rule assumes months are independent. Andrew Lo showed in 2002 that serial correlation, which hedge funds holding hard-to-price assets often show, can overstate an annual Sharpe ratio by as much as 65%. One mortgage-backed securities fund in his sample dropped from 4.03 to 2.44 once that was corrected.
- Period choice. Sharpe stresses that the ratio is not independent of the period measured, and the table below shows how far the window alone moves it for one market.
- Negative ratios. When excess return is negative, more volatility makes the ratio look better: a fund 2 points behind bills scores −0.40 with 5% volatility and −0.10 with 20%. Rankings below zero mean little.
- Mismatched inputs. A ratio from monthly data and one from daily data, or one using the 3-month bill and one using zero, are not comparable. Check how a source calculated its figure before comparing it with another.
| US stock market, monthly returns | Annualized Sharpe ratio |
|---|---|
| 10 years to September 2019 | 0.99 |
| 10 years to September 2026 | 0.82 |
| 10 years to September 2009 | −0.03 |
| July 1926 to September 2026 | 0.45 |
| Calendar year 2022 | −0.90 |
The figures use the value-weighted return of US companies listed on the major US exchanges, from Kenneth French's data library, minus the 1-month Treasury bill rate. The same market earned nearly a point of excess return per unit of volatility in one decade and nothing in another, so a fund's three-year ratio says as much about those three years as about the fund. The bitcoin vs S&P 500 study shows the same effect across four windows for two very different assets.