How To Short The VIX Without Blowing Up Your Entire Portfolio: Backtesting A Hedged Short VXX Portfolio

By Michael Dampf:

For many years, shorting long volatility ETNs has been among the most crowded trades, and the time decay of the VXX has been the center of attention for volatility traders. The short volatility trade blew up however, when in February 2018 the VIX spiked more than 100% on a single day which finally resulted in the redemption of the infamous inverse volatility product XIV. Since then, "shorting the VIX" has become synonymous with picking up nickels in front of a steamroller. Once the dust settled however, investors were flocking towards the only inverse volatility ETN that survived the mini crash, the SVXY. As of today, the market capitalization has already exceeded 300 Million USD, even though by the time you are reading this it might just as well have joined its sister product XIV and gone to zero. Therefore, it has to be analyzed whether there is a way to effectively hedge a short volatility trade.

There are numerous ETNs out there that provide investors with either direct or inverse exposure to a rolling VIX futures position. The most famous ones among them are SVXY, VXX, TVIX, and UVXY. As both the UVXY and TVIX are leveraged ETNs, they are a story on their own which is why I will cover them in a separate article. The SVXY aims to provide investors with the inverse return of rolling VIX futures, which means that being long SVXY effectively equals a short bet on volatility. It has to be mentioned that it was restructured as a result of the 2018 February crash, which is why it now only provides investors with 0.5-times the inverse return of rolling VIX futures. The VXX, on the other hand, aims to provide investors with a long exposure to rolling VIX futures, so by being long VXX, the investor acquires a long volatility position. As a result, there are two ways to short volatility: long SVXY or short VXX. So what is the better trade?

On 99 out of 100 days it won't make a difference if one shorts the VXX or just goes long SVXY. What matters however, is the day where it will make a difference and unsurprisingly, one of these days happened in February 2018, when the VIX skyrocketed to unheard-of levels. Even though the VIX spiked 115% on the 5th of February, when markets closed, everything seemed alright: the VXX gained only 33.52%, while the SVXY lost -31.99%, which is basically in line with what they are supposed to do, as the ETNS track the VIX Futures index, not the VIX itself. Investors were due to wake up with a shock however: While the VXX gained a mere 4.88% between the close on the 5th and the open on the 6th, the SVXY lost -83.71% of its value. The movement of the VXX was again in line with the index it is supposed to track, which raises the question, what actually happened to the SVXY.

To uncover this extraordinary difference, it is important to shed some light on how the SVXY is structured: Since the ETN aims to provide the inverse of the return of VIX futures, the ETN provider (in that case ProShares) has to short the underlying futures. Once the futures gain more than 100% however, the provider faces a problem: They cannot pass on the loss to the investors, as the loss in any long position is limited by the amount invested. In order to avoid that risk, they start liquidating their position once the underlying value decreased by more than -75%. While that was not the case on the close of the 5th February, they still had to roll their short futures position, in order to keep up a constant one-month expiry. Rolling a short futures position effectively means buying back front-end futures and selling futures with the second closest expiration date. On that day, they had a massive position to cover, as at that point in time the combined market cap of inverse volatility ETNs exceeded 1bn USD. As a result, they were driving up prices, pushing the value of their ETN below the -75% barrier, where the liquidation process was set to be triggered. While Credit Suisse redeemed their inverse volatility ETN XIV as a result of this, ProShares decided to keep their product alive, and it started trading again with a loss of -83.71% on the next day. It is events like these that show why shorting the VXX for a short volatility trade is favorable to going long the SVXY. The difference between using the VXX and the SVXY for a short volatility trade over those two days is the difference between taking a hit of -40% and effectively blowing up your entire portfolio by losing more than 90%.

While the previous example demonstrates that shorting the VXX delivers better results than buying the SVXY, this does by no means imply that it is a smart idea to just open a margin account and use all the collateral to short the VXX. Not only is also a ~ 40% loss enough to end one's investment career, but Black Swan events happen and have the potential to leave the investor with a theoretically unlimited loss. Therefore, it is important to hedge the tail risk of the trade, even if that means sacrificing some of the profits for being able to sleep well at night.

The benefits of using VIX Options vs. SPX Options as portfolio hedging tools are a topic that I will cover in a different article, but I still want to provide a brief overview of this thesis. The outperformance of VIX vs SPX options mainly stems from the unique volatility surface of the VIX options. Implied volatility is a major price driver, as it is an important input variable of the Black-Scholes model, with higher implied volatility pushing the price of an option up and vice versa. While for SPX options (like for most equity options), implied volatility tends to decrease when getting closer to expiry, the opposite is true for VIX options, which usually encounter an increase in implied volatility when getting closer to expiry.

As the table demonstrates, hedging the short VXX trade with a long VIX call yields better results than hedging it with SPX options. Not only is the annualized return higher, but the maximum drawdown is lower, the sharpe ratio is higher and it lowers correlation much more significantly. Before examining the results of the VIX call hedged portfolio more closely, it is important to mention the brief spike and consecutive drop at the beginning of 2018. Unsurprisingly, this is a result of the volatility spike on February 5th, which was already discussed earlier. What happened here to the VIX Call options hedged portfolio is quite remarkable: As VIX futures spiked massively at that day, the option that was previously deep out of the money (delta < 25) was suddenly deep in the money (delta > 75), which made its value shoot up more than 10-fold! As the value (and therefore the allocation to the option) already went up the days before, this had a massive impact on the value of the portfolio and more than compensated for the loss, which resulted from the short VXX position. Over the course of the next days, however, implied volatility of VIX options collapsed, which finally resulted in an overall loss for February 2018. As the portfolio simulation does not contain any take profit events, the VIX calls hedged portfolio could not capitalize from that massive spike, even though it should be noted that active management would have enabled an investor to take at least a part of the profits generated by that massive move. This event also impacted the performance metrics, which is why they should be taken with a grain of salt - excluding this event would yield an annualized volatility of 9% for the VIX option hedged strategy and consequently also a higher sharpe ratio.

Taking this event into consideration, hedging a short VXX portfolio with VIX calls certainly is the most attractive option, as it provides the investor with a relatively smooth equity curve and shorter underwater periods. This becomes particularly obvious during the 2008 financial crisis, where the naked VXX portfolio lost out massively, but also the SPX put hedged portfolio lost in value. What is more, the SPX put hedge is so ineffective, that even an unhedged VXX portfolio recovers more quickly from the financial crisis than the hedged portfolio, as the premiums of the SPX puts drag down overall performance and decrease the speed of recovery.

In a last step, I want to compare the VIX call option hedged portfolio with a long position in the S&P 500. Looking at the chart above, it becomes evident that the VIX call hedged portfolio also outperforms the S&P 500, even though the events from February 2018 almost wiped-out all of the alpha generated since 2008. Moreover, it also has to be mentioned that the short VXX VIX calls hedged portfolio only uses 27.5% of the available equity in the portfolio, whereas the S&P 500 portfolio uses all of its available equity.

Taking all these factors into consideration, hedging a short VXX portfolio with VIX calls can provide an investor with attractive returns. The potential convex pay off of VIX options, combined with the time decay of the VXX result in a portfolio that allows traders to sleep well at night, even if they are exposed to a short VXX position. In order to make the most of this trade, profits should be taken in moments of extremely high volatility. Moreover, different option allocation sizes, as well as potentially dynamic allocation, and more frequent portfolio rebalancing should be tested in order to increase the return in decrease the drawdown of the portfolio.

In order to compare different short VIX portfolios, 3 different portfolios were constructed. Apart from the hedging tool used (or the lack of it), the set up was the same for every portfolio: Every third Wednesday of the month, 2.5% of the available capital is allocated to options with ~ 60 days till expiration. The options that were previously in the portfolio are rolled so that the DTE never goes below ~ 30 days. Moreover, 25% of the collateral is used to establish a short position in the VXX. Since data for the VXX is only available from 2009 onwards, the SPVXSTR index is used for 2008. The SPVXSTR index simulates the return of a rolling long front end VIX futures position, which is why it presents an almost perfect proxy for the VXX performance. The delta of the options is based on the results of a research article I will upload later this month, which conducts an in-depth comparison of VIX and SPX options.

See also 'True Tech' And 'Best-Managed' Firms: How Do They Relate on seekingalpha.com

The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.