How much further would the S&P500 have to fall before we worried about a bear market?

An average price of 4220 or lower in February could signal potential bad news

Market jitters are making themselves felt across the world as I write this post. I pubished a piece today on Unherd speculating that bullish sentiment would be hard to prop up in 2022 because interest rates are set to rise and the inflated profit margins that corporates have been availing of due to lockdowns are likely to decline.

Any time I write a piece like this I’m always asked: “do you think we have entered a bear market?” My response is no different from anyone who has worked around markets for a while: “I have no idea, we will only know after the fact”.

This answer has always frustrated me. Not because I think that there some way to predict the future in the near-term — there is not — but because investors should have some way, however approximate, of gauging their concern of an upcoming bear market. ‘Sophisticated’ investors are known to laugh at day traders and chartists — with their arbitrary moving averages and their Bollinger bands — but in reality we all should have some sort of way to approimately gauge our level of concern.

In what follows I will try to mix two things that some will say should not be mixed: namely, chartism and actual long-term data analysis. Please do not expect magic beans by the end of this post. But I think the result does give us some sort of guiding posts as to what we should be looking at during the present tubulance.

I have kept the methodology as simple as possible. I take monthly stock market data going back to 1871. I then calculate the quarterly price change. I then turn this quarterly price change in a three period moving average to further smooth it out.

Looking at this data I have determined that a reasonable proxy for a bear market using this metric is a reading of -14% or less. This captures all of what I would consider bear markets in the dataset. We can then turn to the data and see the price declines one month and two months prior to these bear markets. Here they are in two histogram charts (removing all sub -14% readings, obviously).

Ignoring a few outliers, we can say a few things from this.

• We would want our metric to be reading negative two months before we would expect a bear market.

• We would want our metric to be reading less than -8% one month prior to expecting a bear market.

But this is just a minimum. In reality, we would a stronger than this to get scared. Remember the two charts only show us the minimum threshold needed to fear a bear market; they do not tell us the number of times this minimum threshold has been broken.

In reality, the minimum threshold is broken all the time. For example, we have a total of 97 readings of -8% or below but only 18 of them predicted a bear market (success rate = 18.6%). Meanwhile, readings of -8% or lower make up just over 68% of our total correct one month trailing bear market signal.

To avoid being Chicken Littles we therefore want to take a conservative approach. Let’s say that we would start to get worried if our metric started showing a -12% reading or greater on a one month basis. This strikes the right balance. There have been a total of 15 readings of -12% or under and 8 of them have called a bear market (success rate = 54%). Meanwhile, readings of -12% or lower make up around 42% of our total correct one month trailing bear market signals.

Well, what would this look like moving forward?

It would mean that if we see monthly prices for the S&P500 fall below 4220 we should get genuinely concerned that we have entered a bear market.

“Disater!” you might say, “the S&P500 is already down from 4796 to 4410; another hit like that and we’re in a bear market.” Not quite. Recall that we are using a monthly measure. That is, the monthly average price of the S&P500 — not its price on a given day. If we were using daily data it would be too noisy to get any signal.

So, the correct approach would be to take a monthly moving average of the daily price. If this crosses 4220 in February, we should get worried. But as the chart below shows, this would mean larger price declines moving forward — and, mopst importantly, ones that did not reverse.

Would I short the market based on this price signal? No. But would I be open to the possibility of an imminent bear market if I saw it? Yes. That is all we can really ask of these sorts of metrics.

Right now, however, fears of an imminent bear market are overblown.