Shorting The VIX: How To Protect Profits In A Volatility Spike
Mar. 30, 2015 5:00 PM • vxx
- The profitability of a volatility short is dependent on price movement of the underlying VIX futures contracts as well as the spread between contracts held by the ETF, known as contango.
- With volatility near 4-month lows, and near-month contango near 9%, a short position in VXX has been highly profitable in 2015.
- With mean reversion of the VIX ultimately favoring an increase in volatility, strategies to hedge and protect these profits in the event of a volatility spike are discussed.
Despite the S&P 500 dropping 2.2% last week for the largest weekly loss in a month, finally injecting some volatility into the markets, the past two months have largely been a snoozefest. A short position in any of the major volatility ETFs has returned a tidy profit, as the VIX settled near 4-month lows. This profit comes due to both the slumping volatility itself along with a strong assist from the contango currently present in the VIX futures market on which volatility ETFs are based. This is a recipe for easy money. However, the enemy of profitability is complacency. Even after the ugly turn to the markets last week, the VIX remains well below its 5-year average. With the tendency of the VIX to mean revert - that is, to approach an average value - there is significant upside to the VIX that threatens the profits of those betting on continued low volatility. I do not pretend to know where the markets and volatility will go, with rising geopolitical tensions and potential rate hikes facing off against a strong US economy. However, with the VIX futures curve and its associated contango remaining highly favorable for volatility shorts, this is no time to sit on the sidelines and wait for a better entry point. This article discusses strategies to capitalize on the large contango in the VIX futures market while hedging out risk should the VIX spike.
Unlike most equity trades in which share price is the sole determinant of profitability, volatility ETF profitability is determined by two components. First, there is the underlying movement of the volatility futures contract(s) that are held by the ETF. For the iPath S&P 500 VIX Short-Term Futures ETN (NYSEARCA:VXX ), the most popular of the volatility ETFs, that is the April and May 2015 VIX Futures contracts. When volatility increases, these contracts increase in value, as does the fund that holds them. This is obvious.
The second determinant is the spread between contracts. Typically, longer-term contracts trade at a premium to the front-month contract due to uncertainty regarding future volatility. This is known as contango. Funds such as VXX roll over from the front-month contract (April 2015) to the T+1 contract (May 2015) on a daily basis. When the VIX is in contango, as it usually is, the fund sells X shares of the front-month contract, but can only buy X-Y shares of the new contract, where Y is directly proportional to the size of the contango. As this rollover continues on a daily basis, day-after-day, month-after-month, gradually the ETF can buy fewer and fewer shares with its funds and its value decreases independent of movement of the underlying commodity.
Trading volatility ETFs is therefore about a balance between absolute value of the underlying VIX futures contract and the spread between the T+0 front-month and T+1 contracts. When markets are tranquil, volatility tends to be very slow which favors mean reversion and the volatility long positions. However, low volatility typically correlates to a large contango which favors the shorts. The opposite holds true during times of agitated markets.
Despite all of the hoopla, the front-month April 2015 VIX futures contract finished the week up just 0.3% at $16.22. This remains 20% below the five-year average VIX level of $20.15. Figure 1 shows the VIX front-month price for the past 1-year versus the 5-year average level.
Figure 1: 1-Year Performance of the VIX Front-Month Contract showing generally tranquil conditions outside of spikes in volatility last fall. [source: Yahoo Finance ]
Outside of a spike to the mid-$20s in October during the Russian debacle and a smaller peak in mid-February, the Front-Month VIX has remained below its five-year mean and has trended steadily downwards since the New Year. Overall, during this period, the net change in the front-month VIX contract is effectively flat.
As is to be expected, the VIX futures contango is elevated at these levels. Figure 2 below shows the VIX futures contract prices for the next 8 months and the associated contango versus the front-month contract. If this curve remained unchanged for the next eight months as contracts were rolled forward, VXX would lose around 19.7%, independent of movement of the underlying VIX contracts.
Figure 2: VIX Futures contracts for the next 8 months showing significant contango. [source: Yahoo Finance ]
VXX is currently rolling April contracts into May contracts. As of Friday's close, 56% of the ETF's funds remained in April 2015 contracts and 44% were in May 2015 contracts. The current % spread between April and May contracts is 9.2%. Historically, this is an elevated contango compared to a two-year average of 4.9%. Given that each contract is the front-month contract for around 20 days, VXX rolls over approximately 5% of its holdings per day. At the current contango, this equates to a daily contango-associated loss of 0.092 * 0.05, or 0.46%. While negligible on a daily basis, this can add up very quickly if the contango remains elevated.
Let's see this in action.
Figure 3 below shows the observed VXX price movement versus predicted price movement if VXX tracked the front-month VIX contract perfectly over just the past six months.
Figure 3: Observed VXX performance vs. predicted performance showing significant underperformance of the ETF. [source: Yahoo Finance ]
While VXX tracked the futures contract well for the first month or so of the period, the predicted and observed prices begin to diverge thereafter. By January 2015, VXX was trading at $31 versus a predicted price of $37, and by March that spread had increased to $8. Overall, during this time, VXX declined 7%, while the front-month VIX contract on which the ETF is based gained 20%, a massive underperformance attributable to contango. If we add up all of those miniscule contango-associated losses due to rollover, we get Figure 4.
Over just the last six months, losses from the rollover process alone total 27.2%, or an estimated 54% annualized.
While this seems like easy money, it is necessary to remember that there are two components to trading VIX ETFs: Rollover and underlying volatility. Friday's VIX front-month close of $16.25 is 19% below the five-year average for the front-month contract. Historically, when the front-month VIX is within +/-5% of this level, the contract gains an average of 6% over the next 90 days. At these levels, VXX's performance is negative over the same time period - thanks to Contango - but experiences significant volatility.
Between the initiation of the fund in 2009 and present, there have been 185 occurrences when the front-month VIX is within 5% of $16.25. Figure 5 shows the average performance of the VXX versus the maximum and 95th percentile performances on a daily basis for these 185 occurrences.
Figure 5: Historical 90-day VXX Performance when VIX front-month was within +/-5% of current value illustrating significant risk to an exposed short position. [source: Yahoo Finance ]
While the VXX continues to trend lower even at historically low volatility levels that favor mean reversion of the underlying contracts, there is significant risk to holding a long-time short position. Over the last five years - a time dominated by comparatively tranquil markets with relatively low volatility - a short VXX position has lost as much as 182% over a 90-day period, a loss which would likely trigger a margin call in many accounts. Even the 95th percentile is greater than a 50% loss, which erases any profits brought about by rollover losses, and then some.
For those who are risk averse, but would still like to capitalize on contango in order to isolate the consistent profits from contango-associated losses
while protecting against losses from a rally from the underlying VIX futures contracts on which volatility ETFs are based, I propose the following strategy.
With the funds acquired from the short sale of VXX, an inverse ETF can be purchased that roughly tracks VXX but holds its funds in fixed securities and is not susceptible to rollover losses. In this way, the risks associated with fluctuations in volatility can be at least partially hedged against, while leaving the paired position entirely exposed to rollover losses.
In order to determine the best inverse ETF to use for this hedged position, we must calculate its beta relative to VXX. In finance, beta is traditionally thought of as a measure of the volatility of a security or portfolio in comparison to the market as a whole. A stock with a beta of 1 indicates that a stock's price movement will mimic that of the market - if the S&P 500 gains 5%, the stock will gain 5%; if the market is flat, the stock will be flat; and if the market falls 5%, the stock will fall 5%. A stock with a beta of 2 is more volatile than the market - a tech stock, for example - and will gain or lose twice that of the S&P 500 or whatever index is used as the benchmark. A beta of 0.5 is comparatively less volatile - a utilities stock, for example - and will gain or lose half of the market's performance.
While beta is traditionally applied to the stock market as a whole, it is a relatively simple calculation and can be used to compare any two equities or funds against each other. Equation 1 below shows the equation used to calculate beta:
Beta = Covariance (Daily % Chg stock for which beta is being calculated, Daily % Chg underlying index)/Variance (Daily $ Chg Underlying index)
In this case, we are looking for an inverse ETF that most closely tracks the performance of VXX, the underlying index in this case. That is, what ETF has a beta of 1.0 compared to VXX?
Figure 6 below shows the betas of several of the more volatile 1x, 2x, and 3x ETFs versus VXX, as well as their respective R-Squared values. All beta values are calculated over a period of 2 years. All of these ETFs hold stocks or options which do not require rollover or, if they do, carry minimal to no contango-associated losses.
Of the 1x ETFs, the ProShares Short QQQ ETF (NYSEARCA:PSQ ), which tracks the inverse of the Nasdaq 100, had the highest beta at 0.19, along with the highest correlation with an R-Squared of 0.6. This means that if the funds acquired from shorting VXX were used to purchase PSQ, the underlying movement of the volatility ETF would only be hedged by roughly 20%.
Of the 2x and 3x leveraged ETFs, the Direxion Daily Russia Bear 3X Shares ETF (NYSEARCA:RUSS ) had the highest beta at 0.85, nearly matching VXX. However, this fund has a very poor R-Squared of just 0.2, meaning that despite the high beta, it correlated poorly with VXX. The Direxion Daily S&P 500 3X Shares ETF (NYSEARCA:SPXS ), on the other hand, had a lower beta of 0.55, but a very good correlation of 0.72, making it a much better hedge against VXX, even if it hedges only 55% of the position. This is reflected below in Figure 7, a scatterchart that compares the performance of VXX against both RUSS and SPXS.
Figure 7: Comparison between SPXS and RUSS vs. VXX illustrating the better correlation between SPXS and VXX despite the higher beta for RUSS [source: Yahoo Finance ]
The chart shows that while SPXS is flatter and therefore has a smaller beta, its points have a tighter, more linear grouping and therefore carries a stronger correlation to VXX.
Let's say we decided to go with SPXS as our hedge. Figure 8 below shows the performance of a hypothetical portfolio in which a short position equal to 100% of the portfolio value is established and the funds from this short sale are used to purchase an equal value of SPXS.
Figure 8: Performance of a $1,000 portfolio that is short VXX and hedged with a long position in SPXS over the last 7 months. [source: Yahoo Finance ]
This chart illustrates two important points. First, the hedged position did its job into the market swoon of late September/early October, hedging out 50% of the loss when VXX spiked. However, subsequently, when the markets again became tranquil, SPXS significantly underperformed versus VXX with the hedged position ultimately becoming a losing one. This can be attributed to leverage-induced underperformance of 3x ETFs, a phenomenon that I have discussed at length in a previous article . A trader employing this hedged strategy must be very nimble, and a little bit lucky. A trader would need to buy the hedged SPXS long only when the VIX dips sufficiency low to very strongly favor mean reversion. I do not believe we are there yet, and would probably wait until the VIX front-month dips towards $13-$14. In order to maximize profitability, one would then need to close the position on a price spike when the trend favors mean reversion to the downside. While this can be left to the individual, I would consider selling some or all of the position once the VIX reaches $20, based on historical trends.
An alternative strategy would be to short the corresponding long ETF instead of buying the inverse ETF. The corresponding long ETF to SPXS is the Direxion Daily S&P 500 Bull 3X Shares ETF (NYSEARCA:SPXL ). Shorting this product has the advantage that, over the longer term, the leverage-induced decay works for the position rather than against it. A trader can therefore relax his/her order entry and exit points on VIX troughs and spikes, respectively. On the other hand, it carries two significant disadvantages. First, it is an additional short position meaning that the funds from the VXX short cannot be used to finance this position, resulting in the net profit effectively being halved as other funds must be diverted to create the paired-trade. Secondly, as it is a short position, should the market become very tranquil for an extended period of time, losses from this position can theoretically exceed 100%, overwhelming profits from VXX. Personally, the risks outweigh the benefits here.
A second alternative would be to purchase the iPath S&P 500 VIX Mid-Term Futures ETF (NYSEARCA:VXZ ). This is another volatility ETF, but unlike VXX, it holds VIX futures contracts for months T+3 through T+6 instead of just the near month and T+1. Because it is a volatility ETF, it has a very high Rsq to VXX of 0.83, but a beta of just 0.45 due to the decreased volatility of medium-term futures contracts. And lastly, while the spread between months 4 and 7 is just 4.3%, and a much smaller fraction of the fund is rolled over on a daily basis, this ETF maintains exposure to contango, unlike the other ETFs discussed. Further evaluation is required. I will be composing a follow-up to this article discussing advantages and disadvantages to this strategy.
In conclusion, with relatively tranquil markets and an historically elevated contango, a short VXX position has returned consistent profits since the New Year. However, with the VIX slumping, there are significant risks to holding a VXX short position long term should volatility spike. Inverse ETFs with betas comparable to VXX are identified as potential hedges. However, thanks to leverage-induced decay, none of these make for good long-term holds. However, by buying when the VIX is suppressed and selling on a spike, a significant proportion of risk can be hedged out. I am currently short a VXX position equal to 10% of my portfolio with no hedges. Should the front-month VIX futures contract dip below $13, I will consider adding a long position in SPXS or a comparable ETF.
Disclosure: The author is short VXX.
The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.