History Says Volatility is Not Over

By in
History Says Volatility is Not Over

We have 2 “Data” items about US equity market volatility today:

#1: At the end of June we showed you this chart, an analysis of seasonal patterns in the CBOE VIX Index (expected near-term S&P 500 volatility). The starting point is June 30th at 100, and the lines reflect the daily moves in the VIX through July and August by decade. We wrote that the most recent decade (2010s) saw the VIX fall early in July, which meant the first 2-3 weeks of 2021’s July should be fine. After that, however, volatility has always risen going into early August and then again into late August.

That pattern is playing out, almost to the day (S&P all time high on July 12th), so here’s what this historical analysis says will happen next:

  • Today’s VIX close of 22.5 is +42.4 percent higher than June 30th’s close of 15.8.
  • The only vaguely analogous years are 1996 (+47 percent increase over the first 12 trading days of July) and 2002 (+44 percent). On average, the VIX is usually flat to its June close at this point in July.
  • In 1996 and 2002, the VIX did not hit its near-term high on the 12th trading day in July (i.e., the same timeframe as now). It took 5 more trading days to hit the July/August VIX peak in 1996 and 14 more days in 2002.

The bottom line here is that history says we have 1-3 more weeks of volatility ahead of us, and likely lower stock prices as well. We’re working in some unusual territory, far higher than the historical seasonal patterns, but this assessment also fits with the typical seasonal cadence of volatility:

  • When we first published this data, we told you to circle August 6th on your calendar. Every decade shows an average VIX high 26 trading days into Q3. That works out to that 8/6/21 date. It also squares to the day with the 2002 experience.
  • It’s also worth noting that the analysis shows a second VIX peak in late August, 43 trading days into the quarter. That works out to 8/25/21 this year.

Before we wrap up this discussion, a brief reminder of why this is not just market astrology:

  • July and August are peak summer vacation months, so there is less natural liquidity in US equity markets over this period.
  • We believe in the old saw that says, “the prices you see on the screen today are the market’s best guess about economic conditions and corporate earnings power 6 months in the future”.
  • That puts us right at the end of 2021 in terms of what markets are focused on today. What will Holiday 2021 look like in terms of consumer spending? How will the pandemic develop during the northern hemisphere’s winter months? And how will those 2 factors play off each other?

Takeaway: history says today was the start of a more volatile period for equity markets, and this should last through August. We hope we’re wrong on this. Perhaps today’s selloff was fierce enough to reset asset prices. But the history is clear enough – be careful over the next few weeks, if only because real liquidity will remain abnormally low.

#2: Our playbook for an equity market crash. To be clear, we’re not calling for one. But … We’ve been doing this long enough to know wonky things happen in July/August. “Plan your work, then work your plan” is always a solid approach.

Our magic number is a down +5 percent day for the S&P 500. That is a 5 standard deviation move, so solidly in the “highly unusual” camp. Some color here:

  • Since 1957, the index on average gains a very small amount each day (about 0.03 percent) and has a 1 percent standard deviation around that mean.
  • The statistically minded reader will know that a 5-sigma move should only occur once every 3.5 million days (article below), but it happens far more frequently in stock markets because returns have fatter tails than a classic normal distribution.
  • The last 10 years (call that 2,400 trading days) have seen the S&P 500 move by +5 percent in a day on 5 separate occasions: August 8th, 2011 (-6.7 pct, and in line with the VIX analysis above in terms of peak August volatility), March 9th, 2020 (-7.6 pct), March 12th, 2020 (-9.5 pct), March 16th, 2020 (-12.0 pct) and March 18th, 2020 (-5.2 pct).

Our rule of thumb is to buy the second and every subsequent down 5 percent days, but not the first. Real volatility – the sort that creates good entry levels for a +1-year holding period – tends to cluster rather than being “one and done”.

Takeaway: we’ll come back to you with more of the history on this buy signal should the need arise (that first down 5 percent day), but we wanted to put this idea in front today. A wise old trader once told us “Markets crash not from overbought levels but oversold ones”. Today’s equity market action certainly puts us in oversold territory.


Scientific American article about 5 standard deviation events: https://blogs.scientificamerican.com/observations/five-sigmawhats-that/