In my short post Is Volatility A Risk?, I urged that any definition of risk:
should be evaluated by its usefulness. Any single definition is incomplete and insufficient for making an investment decision.
Here’s a specific case.
How The Sharpe Ratio Broke Investors’ Brains (Link)
Institutional Investor
This is a good overview the shortcomings of Sharpe ratio, most of which should be well-understood by anyone who has computed a standard deviation.
I’ll expand on some of the less obvious points:
- If you annualize Sharpes from monthlies you risk overstating it if the strategy returns are serially correlated.Why? Because you are understating the vol which you can no longer assume scales at the square root of time. This is a complicated issue because auto-correlation, while easy to compute, is itself subject to variation.
- Pardon my yawn, but apparently option sellers game the Sharpe ratio fetish by selling nickels in front of a steam roller. If the image of straw allocators investing on the basis of a single measure keeps you up at night then, sure, sound the alarm. Skewness can hide within vol.
A quick demo:
a) Bet $1 on a fair coin
b) Bet $.33 on heads on a coin that costs 9-1 if tails but has 90% of coming up heads (still a fair coin).
These bets have the same vol ($.33 creates risk or vol parity weighting) but the payoff shape is materially different. - There are popular alt ratios like Sortino, Calmar, and Omega which try to correct for skewness by penalizing drawdowns and giving hall passes to upside volatility. These are not panaceas since they correlate strongly to Sharpes. This reinforces the idea that you can’t compress the nature of any strategy into a single number. (I feel the tendency to pretend that anybody evaluates investments so naively is a straw man drubbing of allocators signaling no deeper handle of the problem than an influencer who read Taleb on a cross-country flight. Like do you even know an allocator?)
- A point the article didn’t mention: you can have high Sharpe strategies that cannot generate high returns. Like investing in T-bills. If the cost of levering the strategies is prohibitive then Sharpe would yet again not be the only number you can look at.
Ok, I’m done suspending my disbelief that anyone uses a single metric in isolation to decide anything of importance. The post is worthy reading for new investors who just discovered Sharpe before they run out and impale themselves on it. I hope my additions made it a touch more interesting for the initiated.