Trading Volatility ETFs: Ride the trend or exit early?

VTS Community,

Last week I tried to answer a question I've been getting A LOT lately and that is, why are we not exiting our current profitable trades early and taking the money?

If you missed that article specific to the Defensive Rotation Strategy and our current QLD position, check it out here

Today I'll do the same breakdown for our Tactical Volatility Strategy.  The general principle is the same but there are some key differences when talking about Volatility ETF trading specifically.

We've held our current SVXY position for a long time now:

Even though it has dipped a few percent in the last week, it's clearly still doing very well.  The question is, should we exit early or just stick to the rules of the strategy?

The Tactical Rotation Strategy is...  Tactical

The rules of the strategy dictate that we ONLY exit positions when the current Volatility environment changes to the point where the signals move to a different range.

As long as the trade dial is pointing to SVXY, we remain allocated

The first step in determining whether there's benefit to potentially exiting early is to check the historical data of all previous trends and see what would have worked best in those.

1)  The current streak is not the longest

While it has been 63 days allocated to SVXY which is a great streak and I'm happy we're in it, but it's actually only the 8th longest streak since January 2012.  Historically speaking anyway, it would not be unprecedented if it lasted a few more weeks, and maybe even longer.

2)  There will be "giveback"

If we look at the average return of the top 20 streaks, it's roughly 15% over an average of 62 days.  We're 63 days in now and up about 24% so we're currently well above the trend.

However, if we look at the column "peak return minus end return" we see that's an average of 9.41%.  This stat is showing how much profit was eventually given back from the high water mark down to the point the strategy got kicked out into safety to break the streak.

Historically, we can expect to lose about 10% from peak to eventual exit

3)  How will we know when to bail early?

If the average return of all the top streaks is about 15%, it may be a good starting point to consider just exiting every time a streak made it to 15% profit right?

Well if we did that, we would have been in cash a long time ago.

Now we don't yet know whether exiting at 15% would have actually been better.  It's possible we get kicked out violently and our current profit dips back to that range.  However, it's also entirely possible that our eventually profit on this streak will be much higher than 15%.

In the end, to maximize return it's pretty clear we will want to have a few of our streaks end in profit much higher than 15%, and hopefully double that or more.  Those few great periods, maybe it's only one per year, but they are hugely beneficial to the long-term results.

4)  I'll trust the signals

I personally don't like to make emotional decisions with my strategies.  They are all designed carefully to follow a specific set of rules that has worked very well in the past.  I'm certainly not going to start making gut instinct decisions now, 12 years into a successful track record.

5)  Volatility ETFs move differently than 2x Nasdaq

One comparison you may find interesting is vs the QLD action in our Defensive Rotation Strategy.  Here's the top 20 streaks for QLD positions within Defensive Rotation:

Same rate of return, but very different dynamic

Bear in mind, the long term rate of return of Defensive Rotation is basically the same as the Tactical Volatility.  They are both just a little over 30% a year long-term.  However, they way they get there is quite different.

A)  The average streak for Tactical Volatility is nearly twice as long as Defensive Rotation, 62 days vs just 35 days.  Now part of that is due to the fact that Defensive Rotation has an additional low end threshold, so sometimes it's getting kicked out into Bonds on the low end.  Tactical Volatility just holds SVXY the whole time, but that only accounts for a small part of the difference. 

The main reason is we want to set up our metrics to allow us to stay in Short Volatility trades longer than we do for stocks.  Remember, even when stocks start to "lose steam" and run on fumes, Volatility ETFs like VXX / UVXY can still continue lower due to the persistent decay in the VIX futures and we want to hang around for that action as well.  Just because stocks have peaked doesn't mean SVXY has...

B)  The average return for SVXY is about twice as much as QLD, 14.96% vs 7.72%.

This is also interesting because the Beta factor to S&P 500 of the QLD is 2.28, where as the SVXY is only 1.36.

This means using ALL days, the QLD moves a lot more than the SVXY does.  However, only using the BEST days during extended streaks, the SVXY actually outperforms the QLD.

C)  We can also expect more giveback with Volatility ETPs.  This is somewhat related to the above point with Beta.  It's beneficial during calm streaks that the SVXY moves more than the QLD, but it's a detriment whenever we see a shock down move.

In a market crash, the SVXY also moves faster than the QLD and this hurts the position and causes more giveback with Volatility ETFs.

* Side note:  One of the reasons we put the Tactical Volatility Strategy on hold in 2022 was because I was seeing excessive giveback, far higher than historical averages.  For that year, it was definitely safer to go with just the 2x Equity ETFs, rather than Volatility ETFs.

6)  Conclusion

The bottom line is, the strategy is designed to account for this giveback problem.  Mathematically we are better off riding the trends as long as they will last.  

Unfortunately that also means there will be times when in order to get kicked out into safety, we get kicked out hard.  It's really frustrating when it happens, but it's always important to focus on the big picture.

The best way to trade Volatility ETFs is with trend following.  We don't want to get in our own way by trying to apply arbitrary market calls on top of the data.

We are just better off trading the strategy like a cyborg...