From 2012 through mid-2014, the VelocityShares Daily Inverse VIX Short-Term ETN (NASDAQ:XIV), an inverse VIX futures ETN, was a very profitable investment. It rose over 900% during that time. Since the end of that period, however, it's shed over 60% of those gains, returning to 2013 levels.
Any trading vehicle that dynamic is risky. Trading it profitably...and even just avoiding large losses...require a solid understanding of what drives its movements. Unfortunately, however, many retail investors trade XIV and other VIX-futures ETNs like the ProShares Short VIX Short-Term Futures ETF (NYSEARCA:SVXY), the iPath S&P 500 VIX ST Futures ETN (NYSEARCA:VXX) and the ProShares Ultra VIX Short-Term Futures (NYSEARCA:UVXY), based on an explanation of their value changes that's inaccurate.
The prevailing (but incorrect!) wisdom goes like this:
"The inverse volatility products (including XIV and SVXY) profit during contango by buying the cheaper front month, selling the more expensive second month and keeping the difference as profit. Since the term structure is in contango most of the time, this ongoing positive roll yield generates profits for the inverse volatility ETNs and drains value from the forward volatility ETNs VXX and UVXY."
You can find this explanation offered in many places, including articles on Seeking Alpha:
- 2 Words Of Caution For The VelocityShares Daily Inverse VIX Short-Term ETN: "XIV gains about 5% per month, thanks to the contango structure."
- Is XIV The Place To Be If Markets Stabilize And Volatility Declines? "XIV may very well gain 5% or more per month simply by rolling future contracts (buying back the ones that are about to expire and sell the next month").
- But Why? VXX And XIV Are Already Just Perfect: "What the XIV essentially does is short high and cover low on a daily basis, resulting in an uninterrupted series of profitable trades. So long as the contango is positive and high, that results in automatic increase in the XIV."
- It's also the thesis behind the charts used on the popular site vixcentral.com: "I tweet the contango because it is what predominantly determines in the long run the value of VXX. Yes, this is counterintuitive. The long-term value of VXX does not depend on the actual values of the futures or the VIX index. It depends on the relation between the first and second month futures."
Before diving into what's wrong with these explanations, I'd like to establish some background with a quick overview of futures and a description of how the VIX-futures ETNs are structured. I'll then describe the real reason these ETNs' value changes and look at why XIV was such a profitable investment from 2012 through mid-2014. I'll discuss the association between contango/backwardation and profitability in the VIX-futures ETNs, then end with some observations on XIV's dramatic rise and fall that may provide insight for trading these vehicles.
The terms contango and backwardation are used throughout, so let's explain those right away. When the futures term structure is in contango, the price of the front-month future, M1, is lower than the second month, M2. In backwardation, the reverse is true. Figures 1 and 2 below illustrate this:
Figure 1. Contango
Figure 2. Backwardation
Quick disclaimer here: I'm not a futures trader. What I know about the mechanics of this market comes from reading and from analyzing and trading VIX-futures ETNs.
A long position on a future is a contract to buy a commodity, a bale of cotton for example, on a certain future date at a certain price - the purchase (i.e., contract) price at which the future was obtained. If market price for that commodity remains low until that contract's delivery date, the purchaser pays for the benefit of having locked down the price ahead of time. The contract's seller, on the other side, has the short position and collects a holding fee for storing the commodity and for guaranteeing its price in advance. Using futures, both buyer and seller can establish pricing guarantees for their respective businesses.
At first, it may seem the benefit is all on the side of the seller, however, things can go wrong for the seller. Suppose the seller's warehouse is destroyed in a fire or hurricane, or workers go on strike. If the contract to deliver is for a time far in the future, the seller may not obtain the commodity until after having agreed to its selling price. In all cases, the seller is essentially placing a bet on being able to obtain and/or store the commodity at a low enough price prior to the delivery date to generate a profit. In other words, the buyer is paying the seller to assume a risk on the future price and availability of a commodity.
Since both short (seller) and long (buyer) positions can be bought and sold on a futures exchange, these contracts create opportunity for speculators to buy and sell futures without ever getting involved in the physical delivery of the commodity.
The VIX futures have a slight twist in that the VIX itself plays the role of physical commodity and the value of it is its spot price. Delivery is purely electronic. Instead of a bale of cotton appearing on the delivery date, the buyer receives (or pays) the difference between the VIX and the final settlement price of the future; the short position's account pays (or receives) the negative of that amount. In other words, to fulfill the contract at expiry, the seller "buys" a purely numeric value (the VIX) at its spot "price" and delivers that difference to the account that holds the long position. This is equivalent to what would happen in a physical delivery contract if, instead of delivering the physical commodity from a warehouse, the buyer and seller settled in cash for the difference between the contract price and the current spot price.
One last point: Futures are settled daily. At the end of each trading day, the exchange establishes a settlement price - usually based on the last few trades. All contract positions - short and long - are settled daily by crediting or debiting the difference between the prior contract price and that day's settlement price. This is equivalent to starting each trading day with a new contract that has the prior day's settle as its contract price.
Design of the VIX-Futures ETNs
The VIX-futures ETNs XIV, SVXY, VXX and UVXY represent a mixture of M1 and M2 contracts. This mixture is rebalanced daily to maintain the equivalent of synthetic future contracts with an expiration date that's always one month in the future. This rebalancing progressively weights M2 higher and M1 lower until, when M1 expires, all contracts have been moved (rolled) to what was M2 on the day, it becomes the new M1 (since the old one has just expired).
This daily roll from M1 to M2 to maintain a constant one-month expiration date appears to be the source of much confusion. It should be clear from the previous section that profit and loss occur via the nightly settlement. The daily roll occurs after that profit or loss has already accrued and does not, in itself, change the ETN's value any more than trading two fives for a ten changes one's cash total.
Sources of Profit and Loss in the VIX-Futures ETNs
The daily change in ETN value that occurs via the nightly settlement is the one-day change in contract price of M1 multiplied by the number of M1 contracts plus the one-day change in contract price of M2 multiplied by the number of M2 contracts on that day. To see what might influence these price changes, consider what the VIX futures prices represent. For volatility futures, there's no storage cost, because the commodity is virtual. It's a measurement produced, published and maintained by the CBOE. That means the difference between these futures and spot VIX represent the current risk premium to hedge against a rise in VIX prior to the future's expiry. Buyers are transferring that risk to sellers and sellers charge that premium for carrying that risk. This difference in value, which is constantly being readjusted as market conditions change, represents the current premium longs need to pay to induce short positions - it's the going market cost of volatility insurance.
If the VIX subsequently spikes up above the contract price and remains high, the risk premium may have been an inadequate compensation and the inverse ETN's value declines to reflect that loss. Conversely, for buyers, their hedge paid off, sheltering them from at least some market downside. If the VIX remains low, however, sellers pocket the risk premium as profit.
Contango, Backwardation and the Mythical "Roll Yield"
One reason the conventional (but incorrect) wisdom summarized at the start of this article seems so plausible is that it appears to be supported by evidence. XIV and SVXY do indeed tend to go up during periods of contango and down during periods of backwardation while VXX and UVXY move inversely. Let's look more closely at that association.
When the term structure is in contango, a positive risk premium is in effect. This is the normal state of affairs because without that premium, sellers would have no inducement to de-risk the chance of volatility spikes for the buyers of these contracts. If contango were not the normal state of affairs, there would be no market. Contango is the result, not the cause, of the ongoing risk premium that's required to make the market in volatility futures.
When the VIX spikes above the values of both M1 and M2, the term structure shifts into backwardation. M1 (and usually M2) contracts rise in price to reflect the higher estimate of future VIX based on this spike, but typically remain below spot since prior VIX (which was lower before the spike) is also factored into the market's estimate of future VIX. The inverse-VIX ETNs lose value when this happens while volatility hedges represented by the long side of these contracts pay off. Backwardation is not the cause of these valuation changes; it's only their visible manifestation.
Contango and backwardation are thus lagging indicators. They show what's already happened, not what's about to happen.
The "roll yield" that's so commonly described as a profit-loss mechanism in the volatility ETNs simply doesn't exist. As described previously, profit and loss happen during settlement and are due to changes in the prices of both M1 and M2 as the forward estimate of future VIX adjusts to changing market conditions. The rebalancing, or roll, happens after this valuation change.
Where did the misconception regarding a roll-yield profit/loss come from? The confusion may have arisen because the term "roll yield" is commonly used in futures trading - but with a different meaning. It's used in the context of buying (or selling) a future, letting it expire, then buying (rolling into) a new future with the same duration. That's different from how VIX-futures ETNs work. And even in the strategy of rolling an expiring future, the term roll yield is just a bookkeeping convenience. The past profit or loss from an expired contract is not changed by subsequently opening a new contract. Each new contract is an independent bet (or hedge) on the future value of VIX. Its net profit or loss still comes from paying or receiving a risk premium and from the difference between past-estimated and future-realized values of the VIX.
The Spectacular Rise and Fall of XIV
If it's not due to contango and it's not due to roll yield, why did XIV rise so vigorously from 2012 through mid 2014? Take a look at the chart in Figure 3, where this period of steady rise on plots of both the VIX and XIV is highlighted. The period of steadily rising XIV is notable for exactly corresponding to the period in which the VIX declined steadily and relatively smoothly to its lowest point since the 2009 crisis. After mid-2014, the VIX began ratcheting back up - slowly at first, then more aggressively, in an upward trend that started in the latter half of 2015. After spiking in August 2015, the VIX failed to return to its earlier lows. This is the point at which XIV finally breaks down. In between mid-2014 and mid-2015, both XIV and the VIX became much more volatile than before. Although XIV put in a new high during that time, once the upward trend in the VIX took hold, XIV began its precipitous fall.
Figure 3. XIV and the VIX (aligned)
In summary: There is no yield from rolling VIX futures, and contango does not determine the profitability for the VIX-futures ETNs. It seems almost too obvious, yet the bottom line here is that changes in the market's expectation for the future value of the VIX are what matter for the VIX-futures ETNs. Those expectations can be altered by changes in the trend and behavior of the VIX itself.
This was a long discussion. I hope it shed light on how these volatility products work and their drivers for profit and loss.
Disclosure: I/we have no positions in any stocks mentioned, but may initiate a long position in XIV, VXX, OR UVXY over the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.