A Story About Market Crashes

Submitted by Nick Colas of DataTrek Research

Since 1958 the S&P 500 has declined by +5% in a day some 22 times. In 82% of those events, stocks were already oversold. Outside the Financial Crisis, the average 3-month return after a crash is +8.5%, with only 1 significant (11%) drawdown. From the first 5% down day of the Financial Crisis (which had 12 5% days), 1-year returns were still essentially flat. Bottom line: markets right now are vulnerable to a crash. Be ready to at least stick a toe in the water if that happens.

* * *

“Markets don’t crash when they’re overbought… they crash when they’re already oversold.” I first heard that bit of wisdom in 2000 from a trader at SAC Capital who went by the name of “Spider”. Yes, he was a wiry sort of fellow, but his nom du guerre came from his adeptness at trading SPY.

Given current market conditions, we’ll dedicate Story Time Thursday to his quip and analyze both how true it is and whether we can make money buying crashes.

First, we need to define what a “Crash” even is and for that we’ll use 1-day return data for the S&P 500 back to its 1958 inception:

• On average the S&P rises by 0.03% a day.

• The standard deviation of those daily returns is 0.98%, which incidentally is why Jessica uses 1% days as our basic measure of US equity market volatility.

• In the scientific community an observation of 5-standard deviations is one widely accepted benchmark of statistical robustness. It was, for example, the metric used to announce the discovery of the Higgs Boson particle.

• That makes a 5% (4.89%, if you want to be precise) 1-day decline in the S&P a legitimately unusual observation, along with any 5% (4.95%) 1-day upside move.

This math explains why statistically minded investors think of equity returns as having “fat tails” rather than those defined by a normal distribution:

• The odds of a 5-standard deviation move are about 1 in 3.5 million.

• And yet, since 1958 (15,647 trading days) there have been a total of 39 days with +5% moves: 17 positive, and 22 negative.

The bottom line is that a 5% decline is a good definition of a “Crash”, so let’s look at all the instances where that has occurred and what was happening around/after those events:

#1: October 19 1987, a 20.5% decline:

• US stocks had been on a tear from January – September 1987, up 33% YTD through the end of the month and hitting fresh all-time highs in the process.

• The S&P 500 then hit a wall in October, declining by 12.2% from the 1st through 16th.

• If you bought the close and held for 3 months you saw a 12.1% gain.

Conclusion: this is undoubtedly the source of the “crash from oversold” aphorism, and buying the drop netted a very good short-term return.

#2: September 2008 – January 2009:

• During the depths of the Financial Crisis the S&P 500 had a +5% drop on 12 days, the largest cluster in the 1958 – present data.

• The first instance was on September 29, 2008 with an 8.8% decline. Month-to-date returns prior to that day were -5.7% and -17.7% YTD.

• So yes – that first “Crash” of the Financial Crisis came during already rough market conditions. But, unlike buying the close of the 1987 Crash, holding 3 months after the September 2008 crash delivered a -21.4% return. Holding for a year did get you to essentially break even (-4% price return, +3% dividend payout), however.

• The reason for that disappointing “buy the crash return”: 11 other 5% crashes over the next 4 months. The last one wasn’t until January 20, 2009.

Conclusion: crashes can cluster, and it can take a year (and a lot of volatility) to recoup a “buy the crash” purchase.

#3: Finally, here is a list of the other 9 crashes back to 1958 and the details around them:

• October 26, 1987: -8.3%, from obviously still oversold conditions post Black Monday. The 3-month future return from the close was +9.6%.

• October 27, 1997: -6.9%, NOT in a clearly oversold market since the S&P was only down 0.6% MTD. The 3-month future return: +10.5%.

• August 31, 1998: -6.8%, from an oversold S&P 500 down 8.3% MTD. The 3-month future return: +21.6%.

• January 8, 1988: -6.8%, but NOT from clearly oversold conditions (MTD +4.8%, 1-month +12.3%). The 3-month future return: +9.4%.

• May 28, 1962: -6.7%, from an oversold market down 7.1% MTD. The 3-month future return: +5.9%.

• August 8, 2011: -6.7%, from clearly oversold conditions (-7.2% MTD, 1-month -10.4%). The 3-month future return: +14.0%.

• October 13, 1989: -6.1%, NOT from an already oversold market (+1.8% MTD, 1-month +1.9%). The 3-month future return: +1.0%.

• April 14, 2000: -5.8%, from a market JUST entering oversold from the start of the dot com bubble bursting (-3.8% MTD, but 1-month +6.0%). The 3-month future return: +11.3%.

• October 16, 1987: -5.2%, from oversold conditions as noted in Point #1 above. Because of Black Monday, the forward 3-month return from this close was -10.8%.

Now, let’s aggregate this data and draw some conclusions:

• In 18 of the 22 cases here (82%), the S&P 500 did drop by 5% in a day (our “Crash” definition) after already experiencing significant losses.

• Trading myth confirmed: markets crash when already under significant duress.

• Outside of the 2008 – 2009 Financial Crisis, if you bought the close of a down 5% day you made an average of 8.46% over the next 3 calendar months with 90% of those instances yielding positive returns. The only exception, but still notable, was October 16th 1987.

• Looking just at the 2008 – 2009 experience, buying the first down 5% move on September 29th was not a great idea, but if you had the fortitude to stick with it you were at least whole in a year.

So, let’s move this discussion to the present day: what if we get a 5% crash day as a result of concerns about COVID-19’s effect on the global economy? Two final thoughts:

• History says buy that close and the data is crystal clear on that point. Crashes are opportunities to make solid 3-month returns with little risk of further cataclysmic drawdowns.

• If you think COVID-19 bears closer resemblance to the 2008 Financial Crisis than a “garden variety” crash, then history says to buy the first down 5% close in small size and wait for more to add to positions.

Bottom line: we’re going into a Friday-Monday sequence that will remind many old hands of 1987 and 2008 – 2009, so let’s be prepared for the possibility of a down 5% day and stay clear headed about what to do next.