The S&P 500 Is Trading According to Plan - Why That’s Cause for Concern

High correlations with prior environments point to new risks in 2020.

The big S&P 500 Index is trading almost exactly as expected right now. And that could be a problem for your portfolio.

A couple of weeks ago, the prognosis looked good: Looking at statistically similar market environments, the S&P 500 looked on track for higher ground in May. And sure enough, we’ve seen stocks push higher in the intervening sessions, narrowing the current drawdown to around 12% shy of all-time highs.

This year's post-crash trading environment continues to have incredibly high correlations with two prior crisis environments: 1929 and 1987.

Those two market regimes are unique in that they were the only other two substantial drops that came suddenly from records in the S&P 500 (or its predecessors).

And the relationship continues to hold incredibly well, even now:


The chart above has been an incredibly useful tool for navigating the COVID-19 crash; it shows the blended 1929 and 1987 post-peak drawdowns versus the 2020 drop.

The prior periods are scaled in 2020 trading days to account for the record speed of this year’s selloff. And while history doesn’t repeat, it clearly does rhyme. That’s not surprising – active research in behavioral finance suggests that market participants react in similar ways to extreme conditions in the absence of external information.

Pick your measure of statistical similarity – correlation, cointegration, etc. – and the similarities between these three timeframes has been extraordinary.

But it’s also problematic.

The fact that completely unrelated periods continue to have such strong correlations implies that the market is currently still reacting to the February crash, not the particularities of the 2020 pandemic.

In other words, folks seem to be looking at the stock market’s recent strength as a signal that the economy is still okay, rather than the other way around.

At some point, we’re going to see the current relationship decouple from our prior analogs. Looking at where we are in the rebound process, that disconnection is likely to happen sooner rather than later:


While 1929 and 1987 had extremely similar initial price reactions for quite a while, they were economically very different events. Once that idiosyncratic data started being priced into the market, the relationship ended. We’re right at that stage in the process now.

That’s a problem for two reasons. First, it means that we lose a market blueprint that’s been extremely effective for the past three months now. And second, the early new economic data that could come into play from here are ugly.

The asterisk on any market forecasts for 2020 has to be the Federal Reserve.

This Fed has been “unprecedented in its unprecedentedness”, and it’s not hard to imagine direct purchases of equities going forward. Unfortunately, that’s probably what it would take to avert new lows in the face of worse negative real GDP growth than 2008, and little clarity on when things get “back to normal”.

Long-term market uncertainty looks elevated until there’s more specific data on the impact of coronavirus quarantines on the economy and when this might finally be over. That’s important to at least keep in mind from a position sizing standpoint here.

Screen Shot 2020-05-21 at 11.00.38 AM

But, at least in the immediate-term, the most important takeaways for your portfolio are still that the near-term trend is pointing up and to the right, and relative strength yields 50% higher forward win rates in crisis investing environments.

The uptrend in the S&P 500 looks constructive, and nearby trendline support provides an early warning for when that may change.

In the meantime, I’ll continue watching this space for new changes in the data.