Sharpe Ratio Adjustments for Negative Returns

# Risk-adjusted returns ratio that does not reward high risk for negative returns


# Risk-adjusted returns ratio that does not reward high risk for negative returns












Think of Sharpe ratio, Treynor ratio, or anything where (excess) returns $r$ are divided by something that represents risk, $\sigma$:

$$\mathrm{performance} = \frac{r}{\sigma}$$

If the returns are negative (let's say for a short period), a performance indicator based on such a ratio is better (=less negative), the higher the risk (e.g. volatility) is.

First question: Does this make sense? I would not say the true performance is better (less poor), if the risk increases – even when the returns are negative.

Second question: Are there established (or even proposed) risk-adjusted return measures that penalize (or at least don't reward) high risk even for negative returns? I'd be quite happy to derive one, but I would expect that someone has already figured this out.

## Answer by Forgottenscience (score 7, accepted)

https://quant.stackexchange.com/a/32448

Yes, you are correct on both terms - it doesn't make much sense, and there exists a well-cited solution by C. Israelsen: "A refinement to the Sharpe ratio and information ratio." Journal of Asset Management 5.6 (2005): 423-427.

The adjustment he gives is to define $$SR_{adj} = \frac{r}{\sigma^{\frac{r}{abs(r)}}},$$

which solves the ranking problem during periods of negative (excess) return.

## Answer by Enrico Schumann (score 2)

https://quant.stackexchange.com/a/32449

1) In a certain, theoretical sense, it does make sense: suppose two portfolio managers delivered negative returns (-1%, say), and one had a higher volatility ("risk") than the other. Then the high-risk fund did better, in a way: despite higher risk, the portfolio manager succeeded in providing the same small loss as the low-risk manager.

2) I agree that this implies a perverse effect: maximise risk for a given negative return. In numerical portfolio optimisation, one may rather prefer a linear combination of risk and return instead of a ratio, or simply put a 'safeguard' into the objective function: when return turns negative, change the selection criterion (e.g. select only based on risk).

## Answer by amdopt (score 2)

https://quant.stackexchange.com/a/32450

- Does this make sense? Consider this: You are an investor. You have 2 investments. 1 high risk (hr) and the other low risk (lr). You expect the hr to be volatile and expect the opposite from lr. If the hr has a small loss and the lr has an equal small loss shouldn't the hr have a better ratio? It should. It performed better based on it's volatility and return potential.

- Yes. You might consider something like an Omega ratio. Omega does not assume a normalized distribution of returns and allows you to set a minimum acceptable return.

Using Omega to optimize a portfolio (i.e. assign weightings to each investment) is far more effective than other metrics in helping a manager achieve an expected return of a group of assets.