Strategy Performance by Market Regime

Designing VIX trading strategies like ours is made more complicated by the extremely bipolar nature of VIX products.

When things are good, they’re consistently good because of the existence of a consistent volatility risk premium, but when they turn sour, things can get out of hand quickly because of the “spiky” nature of volatility.

To illustrate, the graph below shows XIV (inverse VIX) from 03/2004, with losses exceeding -40% shaded red/black.

20140714.01

The four losses shown are -40%, -92%, -53%, and -74% respectively. Ouch.

All four of those losses represent major shifts in the market regime, and all are going to have a major impact (for better or for worse) on the backtested performance of a strategy, but those shifts tended to occur within very short periods of time.

That means that, when designing VIX trading strategies, it’s very easy to be fooled in to thinking a strategy is better than it actually is because it happened to be positioned correctly going in to one of these very brief, but very significant events.

I put little value in knowing a strategy did well during one of these bearish regimes (because it might just be good luck), but I do put value in knowing a strategy consistently did poorly during one of these regimes (because it might not just be bad luck).

To illustrate, of all of the 14 simple strategies we’ve tested previously on this blog, the two that have performed the worse during the bearish regimes are the First vs Second Month strategy (blue in the graph below) and the S&P 500 50/200-day Crossover strategy (grey). I’m assuming we traded the VIX ETPs XIV and VXX (see test assumptions).

20140714.02

Regardless of each strategy’s overall performance, their consistent inability to correctly play these bearish regimes correctly is a black eye and a reason I think not to trade either as a standalone strategy.

On the flip side, (as noted above) I don’t put much value in knowing a strategy did well during the bearish regimes because it might just be a product of luck, but for the sake of completeness, the two strategies that performed the best during the bearish regimes are DDN’s Momentum Rotation (blue) and TM’s RSI(2) strategies (grey).

20140714.03

In summary, it’s important to look at major market regimes when assessing VIX trading strategies. Good performance during bearish regimes isn’t necessarily a good thing, but consistently poor performance is probably a bad thing.

Good Trading,
Volatility Made Simple


Wonk notes:

In the analysis above, I ignored the period in late-2008 that is shaded black in our first graph. Because that massive crash in XIV (i.e. spike in volatility) came at the tail end of a long bearish period, many of the strategies we’ve talked about did well timing it. But the way in which the late-2008 crash unfolded was a unique event, and though there will be future crashes similar in magnitude, they will likely be dissimilar enough to make drawing any conclusions from the late-2008 crash futile.

Data prior to the launch of XIV and VXX has been simulated. We’re able to do this accurately using a combination of the indices and the futures data on which these ETPs are based. Read more about simulating data for VIX ETPs.

Posted in Strategy Mechanics.