Description

My idea is relatively simple and straightforward, but it's not as "tangible" as other ideas.  Nevertheless, I think the risk/reward ratio of the trade is favorable.

My trade idea is selling 1 UXX1 (November) future vs buying 1 UXZ1 (December) future.  Based on closing prices last night, you can do this for a debit of 0.25.  Likely, I think the opportunity will exist over the next 2-4 months to sell this for more than 0.50, and if you hold the spread until shortly before expiration, I think it's possible to get better levels.

UXX1 and UXZ1 should both track the level of the CBOE Volatility Index (VIX).  The futures are designed to reflect investors' consensus view of future (30-day) expected stock market volatility.  This is an important aspect of VIX futures that many often overlook - they are a pure play on implied volatility and not on realized volatility.  Again, this means that we're more concerned with investor's expectations of the 30 day volatility for November vs December.  I'm going to come back to this later in the writeup because an argument to consider is that investors could anticipate December being a less volatile month than "normal", and thus maybe there is a reason for the relatively flat part of the term structure.

First, let's look at the VIX futures curve and its shape:

(chart not working, sorry - so pls use the data below, paste in excel, and put in chart form)

In case the chart doesn't come out well, here is the data:

As you can see, the shape of the curve is upward sloping in this diagram.  If the term structure is unchanged in a few months, we'd expect the VIX futures to "roll" down the curve and the slope between Nov and Dec would increase to the current levels in the front end of the curve (ie around 1 pts).

Interestingly, the VXX has an important, albeit indirect, impact on this trade.  As the futures roll down the curve, the VXX ETN is hedged by selling front month futures and buying second month futures (this is an oversimplification, but approximates the logic).  There are other write-ups on VIC that go into more detail on the VXX so I won't spend much time on it, but the impact on this trade is that if unable to take the trade off before November became the front month contract, there is a very good chance that the VXX hedging impact would drive the spread between Nov and Dec to levels wider than 0.25 to exit the trade.  Above, I mentioned that a potential argument against this trade is that someone could argue December very well could be a "less-volatile-than-normal" month,  and thus the curve expectations are correct to price this part of the term structure this way.  That logic is not altogether incorrect, but the effect of the VXX is so powerful that it's likely the front end of the curve could be driven more by technical factors (VXX hedging) as the time to settlement decreases.

The main risk to the trade is that the VIX curve inverts and the front month futures are higher than the back end futures.  Could this happen?  Absolutely - although not very often.  If you were worried about this happening, you could buy a forward put to hedge yourself (but I think this overcomplicates things and there are a lot of assumptions that need to be made) as the market would likely be a lot lower (inverted volatility curves tend to happen in weak markets).  I am comfortable with this risk because even when the market declines, although the volatility curve shifts up the vertical axis, and perhaps the percentage differences between the contracts flatten out, I think the magnitude of the spread between the two contracts will still be wider.  This effect will help reduce, and hopefully entirely offset the inversion.  But to be fair, it's impossible to know and depending on your assumptions, that would change your expected value for this trade.

As a pure hypothetical example, if one sold a spread of selling front month at 19, and buying second month at 20, he'd pay 1.  Let's then assume the SPX declined and volatility increased such that the second month traded at 40.  If the curve shape stayed the same (despite the move up on the y-axis), one would expect the second month to trade at 38, and the spread could theoretically be unwound at 2. 

However, in a market decline, we'd expect the slope of the curve to flatten.  For this example, let's say rather than a 5% difference between the contracts, the difference decreased to 2.5%.  With second month at 40, the front month would trade 1 point less, or 39, for a difference of 1 pts and a trade PNL that was also flat.

Again, I don't want to ignore the beta to the SPX of this trade completely, because it's possible that the futures invert and the spread one paid 0.25 for could be trading a negative spreads (ie this is not floored at zero).  While you could hedge some of this risk out with the forward put hedge as mentioned above, I tend to trade this kind of idea in smaller size so if it goes against me, I can easily trade out (personally, I would stop myself out of the trade at a 0 spread because if that happens, I think that's a good indication that the macro environment has changed and the expected value of the trade isn't as large) and move on.  For reference, mm202 noted in his VXX writeup that the backwardization after the Flash Crash lasted only a few days.  (I haven't verified separately, but that doesn't disagree with my recollection.)

Hopefully, in that situation, there would be more dislocations in the market that one could earn a higher return elsewhere despite unwinding this trade.

Catalyst

Passage of time and rolling down the VIX futures curve.