Sometimes, a picture is indeed worth a thousand words.
SVXY Total Return Price data by YCharts
Ever since the financial crisis, investors has been asking for ways to hedge market volatility. When there is demand there is supply, and a whole series of ETFs based on the volatility index (VIX) has come into being. With these ETFs came the opportunity for investors to make incredible amounts of money. In this article, I would like to discuss one that has the potential to make billionaires. It is called the ProShares Short VIX Short-Term Futures ETF (SVXY). There is a ETN incarnation of the same called the VelocityShares Daily Inverse VIX Short-Term ETN (XIV). While I restrict this article to SVXY only, the same conclusions can be made about XIV, as the underlying assets are identical for these two.
The above chart shows the performance of SVXY since inception. In less than 2 years, this ETF has grown close to 5x. Can it keep up this performance? To examine that, let's first study how this ETF is constructed. This ETF is designed to return, every day, the negative of the daily return of the most popular volatility ETF, the iPath S&P 500 VIX Short-Term Futures ETN (VXX). To model the SVXY, I started by modeling the VXX.
VXX invests in a combination of the two front month VIX futures. I had earlier written in an article titled An ETF And An ETN That May Return 10-30% By August End that explains how the VXX is constructed.
The VXX ETF ... keeps exposure of one month in the two front month futures on the volatility index (VIX). It will invest in the combination of the two forward month futures such that its weighted average exposure is a month. This is how it works.
Today was the expiration of the July VIX futures, so as of end of day today the VXX ETF is fully invested in the August futures. Over the next 5 weeks it will sell a portion of its August futures every day and buy September futures on VIX, to maintain the one month average exposure. Since September futures are currently trading at around 7% higher than the August futures, each time it does this it will lose a little bit of money, known as the roll yield (loss). In addition, the August futures is currently about 12% higher than the spot VIX, so that will erode as well, as it must match spot VIX on expiration.
For readers interested in knowing how this works out in practice, I recommend this site.
Let's first model the VXX and in turn the SVXY in a situation where VIX is unchanged. We can then add back the impact of a VIX change. Consider a 28 day VIX expiration cycle, where VIX starts at 19 (the historical average), and remains unchanged every day. At the start of the cycle, the first month future is modeled as 12% higher than VIX, and the second month future is modeled 8% higher than the first month future. (These are historical averages ever since VIX futures started trading.) At the end of the cycle,the first month future converges to the value of VIX, and the second month future is 12% higher than VIX, as it is about to become the go forward first month future. The model assumes that the VIX futures decay at a constant rate from start to finish. This is how such a model would perform.
As the above chart shows, under this circumstance, VXX would decay by 9%, and SVXY would increase by about 10%. (One interesting feature of this model - the starting VIX price doesn't matter. You can plug in 10 or 30 and the returns will still be identical to the one at 19.) If VIX remained unchanged throughout the year (which is unlikely), this would result in a 200%+ annual return for SVXY. Of course, in reality VIX would move around. So, I next modeled the impact of a change in VIX.
Monthly VIX changes have been roughly in between -45% to +165% historically, so I modeled SVXY for that range, in a fixed interval of 10%. As the above chart shows, SVXY is extremely sensitive to VIX changes, however, the impact is more pronounced when VIX drops than when VIX rises.
But what is a realistic monthly change in VIX? Next, I drew the distribution of monthly VIX changes using data going back to 1990.
Clearly, there is a long tail. However, the 95% confidence intervals are -40% to +43%. This means the 95% confidence interval for monthly change in SVXY is roughly -25% to +80%. The average VIX monthly change is about 2%, so the average expected monthly change in SVXY is around 8% per my model.
A expected monthly return of 8% is the stuff that makes not millions, but billions. It translates to around 150% return for a year. A mere 10 bucks invested at that rate of return would turn into a billion in about 20 years. However, if only life were so easy. Note that there is a chance that SVXY drops by up to 25% in a given month. A few months like that, and investors would be facing extreme drawdowns, which are very difficult to recover from. I decided to check what has happened in reality.
VIX futures started trading in 2004. Using data from CBOE, and using the algorithm presented in the prospectus of the volatility ETFs, I created synthetic values for SVXY going back to where the futures just started trading. Then, I measured annual return for each trading day, and averaged that to get to average annual returns. Finally, I calculated maximum drawdowns. Then I repeated the same but ignoring 2007-2009, the financial crisis years, as these are (hopefully) an outlier that we will not again see in decades. These are the results for simulated SVXY.
So, there you have it. In a rather turbulent period, SVXY has still returned, on average, about 80% per year. In order to become a billionaire over the usual 30 years of investing life, an investor would have to put in 30 dollars into it in a retirement account, and wake up a billionaire by the time of retirement.
This sounds way too easy, so what's the catch? Well, look at the max drawdown figure. It is a whooping 90%. The investor would need an iron stomach to weather that, and I believe most people would be unable to do that. In fact, at about a 20% drawdown, I suspect people will be itching to sell, and then buy back when the ETF rallies, thus eroding all the gains. For a volatile ETF such as SVXY, people would for sure try to time the market, and in the process lose all the gains. With great upside comes great drawdown, as Spiderman would have said.
However, for those that have a strong stomach and a thousand dollars to spare, I recommend putting in a thousand dollars in SVXY and put that in a lockbox. Don't look at it for 20 years. I repeat, do not look at it, and ignore all the market gyrations.
I suspect when you open the lockbox in 20 years you will be extremely happy.