How the VIX-index can predict winning days from losing days in the SP500
Updated: Sep 13, 2022
In Equivest we want to combine geology and quantitative analysis to maximize returns. Even the best portfolios suffer big losses during market downturns, especially if you invest in small cap exploration companies. In this post we will search for leading indicators that can help us reduce the market tail risk. When a storm hits the markets we often hear pundits talking about the VIX index and how it has increased to reflect an elevated level of fear in markets. To simplify, we can say that a high VIX means fear in the markets and a low VIX means smooth sailing. But is it also possible that the VIX has a predictive power that can be used in trading strategies?
We created an indicator based on the VIX-index that separate trading days into HighRisk and LowRisk days. Approximately 40% of days are HighRisk and 60% of days are LowRisk.
Looking at data from 2006 - 2022 staying long the SPY would have returned 309%. But the HighRisk days would give you -13% and the Low Risk days would returned 370%!
The indicator also does a particularly good job of avoiding the worst drawdowns such as the 2008 crash and the 2020 pandemic.
Can we do better than holding cash on the HighRisk days? Gold is known to be a safe haven investment during market volatility. Let's give it a try. It turns out that the gold price performs particularly well on HighRisk days, with a total return of 277%!
When rotating between SPY and GLD we are now able to achieve 650%, after trading costs, more than 2x the 309% return from holding the SPY. A bonus is also that we avoid some, but not all, of the worst drawdowns on the S&P500.
What indicators are used? The backtest uses 4 different indicators based on the VIX and the VVIX (volatility of VIX)
VIX 5 day moving average with 2 day lag
VIX 10 day moving average with 2 day lag
VIX 2 day trend, VIX (t=-1) - VIX (t=-2)
5 day correlation between VIX and VVIX with 2 day lag
We can share the actual parameters if you send us a message and also have something you want to share with us. If you can work with the data, you can figure them out for yourself, and maybe even produce better results.
Validating the backtest
The true performance of a strategy can only be measured by the out-of-sample performance. This strategy was made based on data from 2006-2019, meaning the out-of-sample data is the data from 2020-2022.
Unfortunately we can't conclude that the strategy has outperformed the markets, even after spectacularly dodging the bullet on the 2020 pandemic drawdown. It seems, however that the failure is primarily related to the gold price underperforming. I think most traders would agree it's been difficult to find a good SPY-hedge lately since both GLD and TLT have failed. Is it still a useful indicator?
We still have an indicator that does well at identifying the days with the highest returns on the SP500. The LowRisk days produce 18% return and the HighRisk days produce 7%. Meaning it mostly separates high return from low return. This can be useful.
Many traders use such regime indicators to refine other strategies. If you are going long on a strategy, it's nice to shave off some of the worst days on the SP500 and focus on those 60% best days.
We will certainly make sure to apply this filter to our trading strategies we generally want to reduce our exposure during the large market drawdowns. We are also excited to test this new tool on our discovery strategies.