Over the past three days, the iPath S&P 500 VIX Short-Term Futures ETN (VXX) has seen a bit of an uplift as overall levels of volatility have inched upwards once again.
In this piece, I will dig into the drivers of this movement, as well as reiterate my call for shorting the VIX at these levels.
At a very high level, it’s important to understand and isolate the key drivers of major movements in the instruments we trade. When it comes to VXX, it is holding futures which settle off of the reported levels of the VIX on certain dates, which means that its performance is essentially tracking changes in the VIX over short time periods.
The VIX itself is often called the “fear gauge”, and for good reason: it has a high inverse correlation to the changes in the S&P 500 over most time periods. In other words, as the markets fall (and thus, investors become more fearful), the VIX tends to rise.
And no surprise, since the S&P 500 has been falling over the past few sessions, the VXX has been along for the ride.
What this essentially means for short-term traders is that if you want to time a trade in VXX, it makes a lot of sense to understand where the S&P 500 is likely headed over the next few trading sessions.
In the chart above, I have marked up my technical view of the S&P 500. Over the past week, the market attempted to breach through the resistance level established during the selloff in early March. Prior support tends to lead to future resistance, and this key level which was tested in March held back two attempts of the market to climb higher.
The fact that the market approached this resistance level over the past week on waning momentum as seen by the shaded area of the MACD was potentially a strong warning sign that the odds of further upside, at least in the short term, were declining. This should have, at minimum, put traders on notice that the rally of the last two months was due for a pause.
At present, we have seen the MACD turn over into bearish territory – the last time this happened was a few days into the historic selloff which started in late February. Based on this switch into bearish momentum, coupled with the prior failure to breach resistance, I am currently bearish the S&P 500. I believe that in the short term we are going to likely find support at the prior level of support established in mid-April, but until then, I believe the downside is more likely than the upside in the S&P 500.
The key tie-in here to VXX is that what this essentially means is that if I were a short-term trader, I would be looking to buy the instrument, because as the market falls, the VIX tends to rise. The VIX is a calculation of implied volatility based on a basket of options in the S&P 500, and since investors tend to be long and therefore hedging with puts, as the market falls, put options tend to be bid up and the VIX tends to rise. So, over the next few days (or even a handful of weeks), VXX is likely to increase in line with a rise in the VIX. However, as we’ll discuss in a bit, my investment time horizon is not the next days to weeks and I am long-term bearish this instrument.
While the VIX may surge over the next few days to weeks, it’s important to note that the odds don’t favor volatility hitting the same levels established early on in this selloff. The reason I say this is that a study of past crises tends to show one key feature: volatility tends to peak during the first part of a crisis, even if the market continues to sell off over months or even years.
As a key data point which demonstrates this concept, here is the 2007-2009 bear market shown in conjunction with the S&P 500.
If you look closely, you’ll notice that the VIX itself peaked around October 2008 – but the market continued selling of for nearly six months before the ultimate low was reached. Volatility remained elevated during this time period, but the VIX itself never reached the highs established during the front part of the crisis. In fact, the market sold off an additional 30% over the next few months, while the VIX reduced by nearly half over the same time period.
What this analysis essentially says is that the VIX is unlikely to achieve the historic highs seen earlier this year. We could march through more examples of market selloffs with a comparison to the VIX, but the key message is the same: the VIX tends to peak at the start of a selloff and generally declines through the remainder of the selloff. This basically means that if you are in the volatility bull crowd, while you may see a few weeks of gains due to an additional wave of selling in the market, moderate your expectations and don’t look for new highs in the index.
We’ve shown two key things up to this point – first off, the VIX is likely to rise in the short term, benefiting long traders in VXX, and secondly, VIX bulls should moderate expectations and not look for the record-breaking movements seen in the first part of this year. So, all in all, the tone to this point is actually quite bullish for the VIX. And yet, as seen in the Seeking Alpha ratings score, I am strongly bearish VXX. Why is this?
Put simply, my time horizon is different than that of short-term traders. For example, in the first chart I showed in this article, VXX was performing quite admirably across the past few days. However, a longer-term chart makes the case that VXX has on net been a losing proposition for traders over time.
I’d encourage you to look very closely at the above chart. This is the long-term performance of VXX. It shows the return which investors would have earned had they been holding this instrument through time. Does it look like an attractive long-term hold?
This year, we have seen the greatest run-up in the history of the VIX. Do you see even a blip on this chart showing this epic rally in VXX? I don’t. In fact, if you zoom in to frame up the current VXX rally in perspective, you’ll see a sobering story: the greatest rally in the history of the VIX only erased about 1-2 years of losses in the instrument, and that too only briefly.
For VIX bulls that are using VXX as your chosen instrument, I would encourage you to look closely at this chart. If you are trading for anything but the shortest of swings, this should give warning that something is at work in this instrument which is punishing long-term returns. This effect is so noteworthy that the popular ETF.com has labeled this (and other volatility ETPs) as “vast” erasers of investor capital.
So, why is this? What is driving these long-term returns, and why would it result in me putting a “very bearish” rating on VXX despite thinking that the VIX is going to rise in the short term? Let’s dive in and see.
In this section, I will take a deep dive into territory I have covered before when discussing VXX: roll yield. I recently received a comment from a reader pointing out that I at times am repetitive in my discussion of this topic. My rebuttal is actually pretty straightforward: I give as deep of an explanation of this topic as space allows, because I’ve been frequently receiving comments or messages to the effect of investors looking to better understand this concept or not having understood it at all. And secondly, as we’ll discuss in a bit, roll yield is almost the entire driver of long-run returns of this instrument, which means that if you have a long-term horizon, it is critical to monitor and study this concept. In fact, I would argue that it is impossible to truly understand VXX’s movements without a deep grasp of the concepts discussed here. If you understand roll yield and its various nuances and implications, then feel free to end the read here, but if you’re curious to know exactly what VXX declines over almost all lengthy time periods, this section is for you.
Let’s start from the beginning. VXX is an instrument which seeks to give exposure to VIX futures. The basic problem here is that the VIX itself isn’t a directly investable instrument – it is a calculation of implied volatility on a basket of S&P 500 options. What this basically means is that if you look at the quote for the VIX, it actually represents an annualized volatility figure based on the implied volatility of an options model using current trading prices for S&P 500 futures.
Since VXX is unable to directly track the VIX itself, it invests in VIX futures. VIX futures settle off of what the VIX is reported at during a certain window. What this basically means is that when you are trading VXX, you are holding futures contracts which ultimately settle at some point in the future.
I’ve used this analogy before, but I think it’s a really important one to internalize when it comes to the VIX in particular. If you were to draw out in your mind how VIX futures would normally price in relation to each other, what would you draw? Are futures contracts which settle, say, a year from now, high or lower than today’s VIX level?
The basic idea here is that since the VIX itself is the “fear gauge”, I would generally only see it increasing in value when times are fearful. Since by definition most of the time things are normal and therefore not fearful, the value of the VIX should generally be pretty low. And we do see this in the data with the VIX mostly staying in the vicinity of 15-20.
Over the past 30 years, about 75% of all months have seen the VIX end under 25 or so. However, the 25% of the time in which the VIX is not in this range sees the VIX rise substantially – with moves of several standard deviations from the mean not unheard of. In other words, when times are normal, the VIX is generally pretty tame in the 15-20 range. However, when times are fearful, substantial tail-risk exists with very large movements from the mean observed.
This basic model would imply that if you were pricing a futures curve during normal times, you would generally expect the futures which price closest to today to be priced lower than those further out. The idea here is that during normal / non-fearful times, the odds of something scary happening at some point in the future increases the further out you look. This would necessitate a higher-priced futures curve as investors both speculate and hedge on the future potential for VIX spikes. This arrangement of increasing futures prices is what is known as “contango”, and it has described the VIX futures market in 87% of all days for the past decade. Said another way, the market is almost always normal, and when times are normal, VIX futures price in the pattern of contango.
And this is where things start leaving the theoretical space and move to the practical application as it relates to VXX. Since VXX is giving exposure to futures in contango, we have a bit of problem if the markets remain normal: these futures contracts will converge towards the spot level of the VIX during a typical month.
The basic reasoning here is that since VIX futures are generally in contango due to times being normal times, and since this contango is reflective of both a hedge and a speculation that volatility will be higher in the future, if times remain normal, this premium or differential between the futures contract and the spot level of the VIX will decline through time. This past sentence was a bit of a wordy one, but I’d encourage you to read through this until it clicks because this is exactly why sites like ETF.com note that VXX is a “vast” eraser of wealth. Futures contracts converge towards spot prices, which means that the return you get from holding a futures contracts is a function of both the underlying change in price for the instrument and the degree of convergence pricing into the curve.
To connect the dots here, let’s loop back to VXX’s methodology. It is following the S&P 500 Short-Term VIX Index. This index gives exposure to the front two VIX futures contracts and continuously rolls this exposure forward so that the weighted-average holding of VIX futures is about 30 days into the future.
Since VIX futures are almost always in contango, VXX is almost always rolling its exposure into higher-priced futures than the spot level of the VIX. Despite popular belief, this in and of itself results in no gains or losses for investors. A popular myth widely repeated by the media is that selling lower-priced futures and buying higher-priced futures in a contango market results in losses.
This is simply not true, because a futures contract in one month is a materially different instrument than that in another month. To say that you gain a return from buying or selling one futures contract and buying or selling another contract would be similar to saying that you instantly earn money from buying stock A at $50 and selling stock B at $75. It just doesn’t make sense. The only way you could bridge the prices between futures contracts would be physical storage and delivery, and this is impossible in the VIX market (and indeed for most market participants trading most futures contracts). If this concept doesn’t quite make sense, I would encourage you to get out a pen and paper and write out every transaction you would do from buying one futures contract and selling another – as you’ll quickly see, no profits or losses result from these transactions because these are separate contracts.
The real problem of rolling exposure in a contango market isn’t that you’re selling out of a lower-priced contract and buying into another priced contract. The problem is that futures prices “converge to the spot price”. In other words, if you are trading a futures contract one year out, over the next year, the difference in price between this futures contract and the spot price will shrink to be basically zero because this futures contract becomes the spot underlying commodity or instrument at expiry. In other words, over time, the distance between futures and spot becomes zero (with few exceptions), which means that the relative position of futures contracts in relation to spot is very important to monitor.
If you recall from earlier in the article, I mentioned that the data shows that 87% of all days for the past decade have seen the front two VIX futures contracts in contango. What this means from a rolling perspective is that 87% of the time, VXX is holding exposure in futures prices which are higher than the spot level of the VIX. And as just mentioned, since futures converge towards spot, this means that VXX is losing money about 87% of the time as the futures contracts it holds shrink in relation to the VIX.
Okay, this is all theory, but does the data actually support this? Yes, it does.
In the chart above, I have taken the average difference of the front two VIX futures contracts by trade date for the past decade. The trade date calendar starts after expiry of the front contract, so this essentially shows the typical lifecycle of the futures which VXX holds.
As you can see, on average, the front contract starts about 8% above the spot level of the VIX and then ends the month about 2-3% above the spot VIX. This means that, on average, the front-month VIX futures declines by 5-6% during a typical month. The second-month contract tends to start at about 15% above the spot price and end the month at about 11% above the spot VIX, for a change of about 4% during the month.
VXX is holding these two futures contracts. It starts the month about 100% in the “M1 to Spot” line and ends the month about 100% in the “M2 to Spot” line. In other words, on average, VXX is losing money from roll yield because it holding futures in contango and futures converge to spot.
And the real problem here is that if you recall from earlier, the VIX is almost always around 15-20, which means that it really doesn’t go anywhere when observed over lengthy periods of time. Sure, you’ll have a few periods in which the VIX explodes to the upside, but over the long term, it reliably reverts to this range.
What this means is that over long periods of time, most of VXX’s return is going to be coming from roll yield – it is holding futures (the price of which don’t move very much through time), and these futures are in contango. Since futures converge to the spot price, VXX has been losing money from roll yield on about 87% of days for the past decade.
To numerically see this, here is the long-run return of this index provided directly by S&P Global.
And investment in VXX would have declined at annualized rate of about half over the past decade. Said another way, on average, if you hold exposure in VXX for a year or longer, the odds of you ever recuperating your full investment become vanishingly remote. For example, this year we have seen the largest run-up in the history of the VIX… and investors who parked capital in VXX two years ago are still sitting on losses.
This is underlying tendency and concept around which I am short VXX through put options and put spreads. VXX tracks VIX futures, and VIX futures are almost always in contango. Since the VIX really doesn’t trend anywhere when seen over lengthy periods of time, most of the returns investors will receive are tied to futures prices rolling down to spot levels.
And this key thesis is why I’m strongly bearish VXX. Since the long-run returns of this note are strongly negative, as long as the VIX is elevated, fresh short positions are generally warranted – provided your investment horizon is long term.
In terms of actual mechanics of the trade, I suggest trading put options (especially put spreads). The reason why I can’t advocate a direct short in VXX is that volatility is volatile. In other words, a short position can be substantially wiped out due to the fact that when you trade VXX, you earn a percent return of a percentage change. For example, if you were short the VIX at 12 and it went up to 24, you’d be down 100% on your investment.
For me, I am holding put spreads in VIX in 2021 with a few other volatility spreads all the way out to 2022. The reason why I favor put spreads is that it allows you to strip out a good degree of implied volatility on the trade, since you will be buying one option and selling another to offset it.
Ultimately, I may take heat on this position over the next few weeks as the market continues to sell off – but I am confident in the established theory and data of roll yield, and believe that taking long-term short positions on VXX is the best way to play this ETP.
Over the next few weeks, we are likely to see a surge in volatility as the S&P 500 shoots for lower territory. The natural state of VIX futures is contango, which means that on average, investors will lose from roll yield. Since the long-run returns of VXX are negative due to roll yield, as the VIX pops, short long-dated positions should be added.
Disclosure: I am/we are short VXX. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.