Stocks quoted in this article:
Well, it took a lot of work, but we finally got there. By there I mean that the CBOE Volatility Index (VIX) finally closed 20% above its 10-day simple moving average, the criteria I use to officially declare VIX overbought. We hit it intraday at other times, but haven't closed there since late July.
There's absolutely nothing magical about a 20% violation. Especially when you consider there's nothing magical about VIX to begin with. The Chicago Board Options Exchange (CBOE) just changed its methodology a week ago, so in a sense it's all apples to oranges anyway. I use 20% for the very scientific reason that it's a round number and gives us a respectable number of data points to examine.
At 10% or 15%, you'll get more "signal" -- and, of course, more "noise." At 30%, you basically get serious implosions. And so on. At 20%, you get about three to four signals per year. This is actually the fourth one in 2014. Here's a rundown of the violations since 2009:
I included both the July 17 and July 31 occurrences for now, but ultimately I'm going to eliminate the July 31 event, as it includes too much overlapping data. It's also not quite three months in either case. So basically, unless the SPDR S&P 500 ETF Trust (SPY) closes above 195.71 on Friday, it will mark the first time in two-and-a-half years that going long when the VIX gets officially overbought and holding for three months hasn't worked.
Having said that, the average returns of the strategy are 4.23%, and the median returns are 4.49% -- not enormously different from a randomly timed three-month hold (average 3.67%, median 4.52%). That includes the July 17 numbers, but not the July 31 numbers. The trade has worked much better over the one-month horizon: 2.13% average vs. 1.31% random average, and 3.93% median vs. 1.83% random median.
The one-month trade has worked seven straight times now (again, not counting July 31). But, it was pretty choppy before that, with five misses in eight tries. And, two of those misses were pretty bad -- a 6.16% hit in July/August 2011, and a 9.55% miss in April/May 2010 (the "flash crash").
Anecdotally, it does feel more like 2011 than anything else of recent times. SPY dropped about 12% from mid-July into early August, and then meandered for two months before ultimately bottoming on Oct. 4. The ultimate drop was about 21% over the course of nearly three months. VIX incidentally got something like 70% above its 10-day simple moving average at its peak. So, overbought can beget really overbought at times.
Could we see something similar this year? I have no idea. If we do, though, it means ultimately declining to something like 160 in SPY in November/December. It's more likely we see a whoosh and turnaround way before. I'm just throwing the precedent out there.
It's also worth noting that the end of 2011 turned into a really great buying opportunity -- it just took a lot of angst to get there first.
Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research.