How Soon Will The Markets Recover?

Ouch.
Ouch.

As the S&P500 plummeted more than 2% today (thanks Greece!), I was reminded about a question I had about the market’s tendency to revert to the mean. Specifically, does the size of the drop have any relation to the duration before rebounding? Does a bigger dip take longer to recover, like a fighter who’s received a particularly strong blow to the temple? Or does a larger downward move snap back faster, like a rubber band?

I took a look at SPY (as a stand-in for the S&P 500) from its start in 1993 to June 26th 2015. I looked at all the instances where SPY closed down below the previous day’s close, which I’ll call the ‘signal day’. And then I determined the number of days it took for the close to be above the close of that signal day. The idea here is you might want to buy SPY (or perhaps a leveraged equivalent) at the close of this down day, with some idea of how long it’s going to take for your trade to be profitable.

I looked at a ten day window only. Any signal days that required more than 10 days to recover, I assigned an arbitrary value of 11 days to that signal. Is this valid? Yes in my opinion, because it reduces outliers from the averages, and also I see this as a short-term mean reversion idea. If it’s going to take longer than 11 days to recover, you’re probably better off closing the trade and waiting for another opportunity.

I expected to see some correlation between the depth of the drop and the length of the recovery, but ya know what? I don’t see it!

Screen Shot 2015-06-29 at 1.07.28 PM

The average recovery time hovers around 3 days, but note that this is skewed by outliers. The more important numbers to me are the median and mode. Across the board, for every percentile “bucket”, the median number of days to recover is “1”. That means, no matter how big the drop, you’ve got a greater than 50% chance that the next day’s close will be greater than your signal day. And with the average around 3 days, that might be a good place to set your maximum trade duration.

Seems like you could create a trading system around this, yes?

 

Update 06/30/15: The day after, the market closed slightly up. See? Told you so. (This one time, anyway…)

Leave a Reply

Your email address will not be published. Required fields are marked *