October’s market slide created some pretty significant investor anxiety. While that has faded as the market has rebounded, Monday’s nearly 2% drop served as a reminder that we’re not necessarily out of the woods yet.
The defensive protocols in the new "Upgrading 2.0" have not kicked in yet. (You can read more about those changes in our January 2018 cover article, SMI’s Fund Upgrading Strategy Evolves: Introducing Upgrading 2.0.) As we noted in that article, we’ve purposely designed these defensive measures to not kick in quickly. They’re supposed to be slow, with the goal being that they trigger only when the market is showing significant signs of danger. In other words, we want these to turn on only during true bear markets, not during the much more common pull-backs and corrections (which are generally counterproductive to try to defend against).
What has that timing looked like in the past? Here’s what we saw in our research of the 2.0 system.
In 2000, the S&P 500 peaked on March 24. Upgrading 2.0 didn’t switch the first fund slot to cash until November 30. That was 251 days after the market peak, and the S&P 500 index was down -13.9% at that point.
In 2007, the S&P 500 peaked on October 9. Upgrading 2.0 didn’t switch the first fund slot to cash until June 30, 2008. That was 265 days after the peak, with the index down -18.2% at that point.
Both of those cases took pretty similar paths. But it doesn’t always play out that way. The twin corrections of 2015/2016 were the one instance in our twenty years of backtesting when the 2.0 protocols kicked in but it turned out to be a whipsaw — the market didn’t follow through to a full-fledged bear market, and it would have been better (in terms of overall returns) if the protocols hadn’t triggered.
In that case, the market peaked on July 20, 2015. After an initial 12% correction, the market re-tested that high again in early November before rolling over into a second official correction. Upgrading 2.0 would have switched the first fund slot to cash on Jan 29, 2016. That was 193 days after the peak, with the index down -8.8% at that point. (The low for that second correction was reached on February 11, just a couple of weeks later, at a level -14.0% below the prior July’s high.)
Naturally, we should expect each episode to unfold a little differently, even though the 2000 and 2008 examples look pretty similar. In one respect, this year has resembled the 2015/2016 scenario. This year we’ve seen a late January high followed by an official correction, then a rally to a slightly higher high in September, followed by a near correction in October (which may or may not be over yet).
In another respect, this year has resembled the 2000 and 2008 cases. If we hadn’t seen a new high in September, our November issue release date would have been 273 days after the prior high. That was by far the closest the 2.0 protocols had come to tipping the first fund slot to cash — right in line with the 251/265 day time frames of those earlier bear market episodes.
Of course, the market did set a new high in September, and 2.0 didn’t trigger last month. So at this point, we’re only 55 days after the peak, and the market is only down -7.3% from its September high. By those standards, we would seem to be quite early in the typical process.
In short, there’s no direct parallel to draw with any of the prior cases. All we really have at this point is a situation where — if you tilt your head just right and squint — you can sort of/kind of see the outlines of past 2.0 signals in this year’s market action.
Keep trusting the system
To sum all this up, the market’s ambiguous behavior this fall (at least to this point) makes it hard to glean much from these past examples. On the one hand, we saw the 2.0 protocols trigger faster in 2016 when we had a similar "double correction" setup resembling 2018’s dual periods of market weakness. That would seem to indicate that if the market were to weaken again from here, we could see 2.0 trigger rather quickly. On the other hand, if September’s new market high really did "reset the clock" and the market is able to keep its head above water here (or even go on to reclaim those highs), it could be many more months before a 2.0 signal might trigger.
Thankfully, we don’t have to figure that out in advance. We can just keep trusting the system and act on it whenever the signals come. Hopefully it’s helpful to see how the process has unfolded in the past. For me, at least, it’s comforting to see that the system has worked so well in the past despite the defensive protocols not triggering until several months after the market has peaked.
It just so happens that this time around, there seems to be evidence supporting both the possibility of a near-term trigger or one that could still be several months in the future.
