Negative Sharpe ratio: meaning, cash and misleading rankings
A negative historical Sharpe ratio means the investment's average periodic return was below its benchmark, usually cash in the same currency. It does not necessarily mean a negative nominal return, and it is especially easy to misread as a ranking score.
The main Sharpe guide defines the measure. Here we examine what its negative sign does and does not establish, using a historical decade and a separate controlled risk example.
Key takeaways
A negative Sharpe ratio indicates a negative average excess return against the chosen benchmark, not necessarily a nominal loss.
With a negative numerator, a larger volatility denominator can make the score look less negative.
The ranking does not establish higher compounded wealth; the benchmark, sample and full return path matter.
Read the numerator before judging the score
For monthly observations, our annualised historical ratio is sqrt(12) × mean(r − f) ÷ sample standard deviation(r − f). The investment return is r, and the matching monthly Treasury-bill return is f.
Volatility is non-negative. Unless it is zero, the numerator determines the sign. A portfolio averaging 2% annually against 3% cash has a negative excess return even though its nominal average is positive. Inflation is another question; this calculation does not subtract it.
A zero denominator makes the ratio undefined. It does not create an infinitely good score or establish that the asset is safe in every respect.
A historical example: the US market in 2000–2009
We use 120 monthly observations from January 2000 through December 2009, with nominal US-dollar total returns for the value-weighted US market and monthly Treasury bills from the Kenneth French Data Library. This is a broad US-market series, not the S&P 500.
The annualised arithmetic investment mean was 1.018%, versus 2.720% for bills. Their difference was −1.702 percentage points. The sample standard deviation of monthly excess returns, multiplied by sqrt(12), was 16.649%. The annualised Sharpe ratio was therefore −0.102.
Despite the positive arithmetic mean, the market's CAGR was −0.38%, turning $10,000 into $9,626 before costs and tax. Arithmetic averaging and compounding differ, as the Sharpe-versus-CAGR example explains.
Cash dilution changes wealth and losses while leaving this ratio alone
Now form hypothetical monthly rebalanced mixes of that same historical market and the same bill benchmark. These are frictionless calculations, not observed fund returns.
Stock weight
Annualised excess mean
Excess volatility
Sharpe ratio
100%
−1.702%
16.649%
−0.102
50%
−0.851%
8.324%
−0.102
10%
−0.170%
1.665%
−0.102
The respective terminal values from $10,000 are $9,626, $11,646 and $12,882. The identical ratio conceals materially different wealth outcomes and investment exposures. The cash-scaling CSV records both.
For a constant stock weight w, the portfolio's monthly excess return is w × (r − f). Its mean and standard deviation both multiply by the same positive w, which cancels in the ratio. This identity applies to positive Sharpe ratios too. It assumes the cash holding actually earns the benchmark return and that fees, taxes and implementation frictions are absent.
At a 0% stock weight, the mix exactly equals the benchmark: excess return and excess volatility are both zero. The ratio is 0/0, not a measured improvement to zero Sharpe.
The “less negative is better” trap
Consider a separate hypothetical comparison with −2% annual arithmetic excess return for both investments. One has 10% volatility and a ratio of −0.20; the other has 20% volatility and a ratio of −0.10.
The second score is numerically higher because a larger denominator pulls a negative fraction towards zero. That arithmetic does not demonstrate that accepting twice the volatility improves the investor's position. The risk-ranking CSV reproduces the example. A geometric-growth comparison instead needs the full distribution of net wealth factors. Expected log growth differs from expected terminal wealth, and neither follows from this Sharpe ranking alone.
Method and limits
Definitions follow Sharpe's 1994 paper. Every sample standard deviation uses n − 1. Historical monthly inputs were matched against the primary source's September 2026 CRSP release; the input download preserves the existing January 1973–August 2026 cutoff.
A negative score describes this benchmark and sample, not a permanent property of equities. Another cash series or period changes the result; see risk-free-rate sensitivity. Calculations checked on 28 October 2026. This is descriptive analysis, not a recommendation to change exposure.
FAQ
Does a negative Sharpe ratio always mean the investment lost money?
No. It means the arithmetic average of the investment's periodic returns minus its benchmark's returns is negative. An investment can make a positive nominal return yet earn less than cash. Its compounded result is a separate calculation.
Why can increasing volatility make a negative Sharpe ratio look better?
Dividing the same negative excess return by a larger positive volatility produces a number closer to zero. For a -2% annual excess return, 10% volatility gives -0.20 and 20% gives -0.10. That numerical ranking alone does not establish a preferable investment.
Does adding cash improve a negative Sharpe ratio?
If cash exactly matches the benchmark and the portfolio maintains a fixed positive risky-asset weight, it scales both excess return and its standard deviation by that weight. The ratio stays unchanged before costs. At zero risky exposure both become zero, leaving the ratio undefined.
