November 03, 2014

Volatility of volatility, and a new method of hedging.

Today I logged into a paper trading account I had played around with to test a couple ideas.  I had opened up two short positions.  The first in $UVXY (2x long volatility), and one in $SVXY (1x short volatility).  The premise was to make money off of volatility decay.

Sure enough, since April, the experiment had netted about 8% in profits.  However, as a big fan of understanding draw downs and risk, I decided that I wanted to see how it got there.  My trading platform doesn't show graphs, etc (I use Interactive Brokers, if there's a way to show your account returns as a graph, let me know).  So I recreated the positions in Google finance.

What I found was that the max draw down even in this short period of time was massive.  I think something like 15%.  Further, while I expected returns to be fairly consistent and steady, they were the exact opposite.  In fact, if both instruments were tracking perfectly, the general assumption would be that I would stay flat from period A to B. 

In order for two short positions to suffer big losses even when one is 2x leveraged, at least one of the positions had to be outperforming it's expectations.  This is something that volatility decay preachers don't understand.  There's no such thing as free money.  There are, in fact, a few 3x ETFs that you can short as pairs for fairly consistent returns (or was the case a year ago when I last checked), but for most ETFs, they have periods of under and over-performance.  Especially long-underlying or long-market.  In a way, $SVXY is the latter, and in a way $UVXY is the former. 

As a tangeant, if you ever want to exploit a poorly tracking 3x ETF, stick to shorting those that short the market, and the more volatile the sector better.  Historically those trades have been the most favorable.

In any case, I found a raging clue to the wild swings of my strategy when I overlayed a comparison with $SPY.  They were almost completely inverse correlated.  Interestingly enough, the volatility of the two did not result in a falling price overall, but rising.  When I noticed these two things I stopped caring about why they were so inversely correlated, and decided that a strategy could be made of this to create a low-risk low-volatility long term trade.

It was a matter of ratios between the three instruments.  I didn't try modifying the ratio of long vs short VIX, instead I just kept increasing $SPY's influence until I had a nice steady line moving up.  Very few dips, and at the end, as the market made that big V-shape reversal, the volatility of volatility hedged $SPY strategy's returns flew upwards.  It's likely the result of $UVXY collapsing while $SPY rocketed up.  See the pic below.


It must be noted that I used $SPXL, the 3x leveraged $SPY.  While it inflates returns, it also demonstrates just how limited the risk was during this period.

You'll notice above and in the title, "volatility of volatility".  This is the key to understanding why shorting two opposite ETFs resulted in an inverse relationship to $SPY. 

You see, the periods between dips are boring for volatility ETFs.  Long volatility continually degrades, short volatility $SVXY grows.  As it grows, it's impact becomes larger, such that if this strategy had a volatility free period for a very long time with no rebalance, I expect it's returns would be mostly flat as it would lose money on the hedge as it gains from the long-market.

Because of this, returns when shorting both long and short volatility would run inverse to the market.  These consistent moves result in compounding, the risk you take when shorting paired ETFs.  Even up above you can still see how they move apart.  I've minimized it without taking on a lot of extra risk from having too much market exposure though.

During compounding, the returns of the volatility ETFs diverge.  When the market dips, volatility spikes, and the ETFs converge.  This convergence results in volatility decay, resulting in a gain.  And it happens as the market falls.  Once again, an inverse relationship.

I've not gone back and looked at this strategy from a longer term perspective because I believe that our QE bull market makes for poor back testing.  However, it can show 1 of 2 unfavorable markets for this strategy, so I will eventually take a look.  The second unfavorable would be a protracted bear market where volatility stalls out and the strategy may be weak. 

Still, it's clear there's some potential here, and I look forward to seeing how it performs over the next 6 months, with and without a mid-year rebalance.  If still promising, I will likely integrate it into my portfolio alongside the permanent portfolio.


Share This Article
Share on StockTwits

0 comments:

Post a Comment