In a recent blog post we explored what typically happens after a huge 4-5% up day in the S&P 500, as was seen on December 26: How to trade after huge up days.

As always, bearish commentators were out to declare how unhealthy huge up days are. But history does not confirm that belief, huge up days have often been a sign that a major bottom was in and a new market advance starting. We are now 3 weeks later and the S&P 500 closed at 2610 yesterday, up another 5.8% since the close of that huge up day on December 26. Bond investors tend to wait a whole year to earn a 5.8% roi, so it’s always nice to do it in less than a month. That’s how knowing a little bit of price action history can help a stock investor.

Now the same investors and writers, who have of course missed the rally or are stuck in short positions, are out to tell that stocks always retest the lows after a rebound rally like this. But is that so? Or is it just their hope speaking?

How does a market behave after a bear market low? Does it often (or always) retest the lows? We don’t know yet if Dec 26 was the bear market low. The strong up day suggests that the market will be higher a year from now, but nothing is written in stone.
Whatever the odds of a further decline or a rally to new highs might be right now, there is only two main scenarios going forward from the current point. If Dec 26 is not the low of this bear market, then it is off course certain that the recent lows will be retested and broken. If on the other hand Dec 26 was the low, then will we see any kind of retest of that low? That’s a scenario we can test by studying the price action after past bear market lows.
One can argue about what is or isn’t a bear market. The criteria are not set in stone. Usually 20% decline is the threshold used to define a bear market, but there have been a few 19.9% declines which just managed to avoid the bear market stigma. So, I prefer to use an 18% or more decline to find the bear market lows I want to study.

So, here we have the historic examples (prior to 1950 the Dow Industrials is used, S&P 500 charts from 1950 onwards). I look for about 10% rally over the course of several weeks after the low and then we can see how much of a pullback or retest typically comes after that first rally:

1921 Bear market low. This was a severe bear market. First rally took the market up 10% in a few weeks but it never looked back and only chopped sideways for a few weeks before heading higher.

1929 Crash low. This one wasn’t the bear market low. But a 50% advance over 6 months could have been felt like a new bull market. The first rally after the November low was retraced about half by the next pullback, but nothing like a retest of the lows.

1932 Bear market low. First rally after the depression lows was retraced about 70% in the next decline. Those lows would never be revisited.

1938 Bear market low. First rally after the lows saw about 50% give-back and then continued higher.

1942 Bear market low. WW2 lows were followed by a steep advance. Only a few brief 2% pullbacks. Traders that were waiting to buy on a good pullback or retest of the lows never got much of a chance.

1946 Bear market low. Post WW2 bear market low saw several retests of the lows in the ensuing years, with the last one coming in 1949, which bottomed out just above the 1946 low.

1957 Bear market low. A mild 20% bear market. First rallies after the low were rather weak 5-6% advances alternating with pullbacks to retest the low. Just a choppy continuation of the prior bull market.

1962 Bear market low. Another mild bear market. The first significant rally was given back 80% in a near retest of the prior low.

1966 Bear market low. Another mild bear market. Here there was little or no give-back after the first significant rally and investors got no second chance to buy (or cover) near the low.

1970 Bear market low. Significant bear market. First rally after the low was retraced about 80% in a near retest of the low.

1974 Bear market low. Severe bear market. First rally off the lows was given back about 75% in the next decline.

1978 Bear market low. Mild bear market. Gave back a little over 50% of the gains of the first rally but no real retest of the lows.

1982 Bear market low. Very steep advance after the bottom and no pullbacks that allowed investors to buy (or cover) anywhere near the lows.

1987 Bear market low. More a crash than a bear market. Here we got a fair retest of the lows before advancing.

1990 Bear market low. Mild bear market. Gave back about 50% of the gains after the first significant rally.

1998 Bear market low. No significant pullback or retest of the low after this short bear market. Traders that had gone short were forced to cover at a loss and that helped to propel the market higher.

2002 Bear market low. Major bear market. First rally was given back 90%, so this can be seen as a proper retest.

2009 Bear market low. Severe bear market low. No pullback, much less a retest, worth talking about.

2011 Bear market low. Mild bear market. The first significant rally after the low was retraced about 60%. Traders who hoped for a retest are still waiting.

2018 Possible bear market low. We got a 10% rally of the recent low. But we can’t know yet if Dec 2018 will stand as the low of this move.

So, what to expect? There are two big challenges in trading the market at the current point. One is that we don’t (and can’t) know whether the late December low was the end of a bear market or just the first innings of a bigger decline. Stock markets and economies, just like earthquakes and a range of other natural phenomena, display a property known as “self-organized criticality“. Any Richter 4 earthquake could be a foreshock for a larger Richter 7 earthquake, but it could also be the main shock, one and done. In the same way any 10 or 20% drop in the stock market could be the foreshock in an ongoing bigger decline (see 2008), but it could also be a mild bear market that is over already. Known parameters like debt levels in the economy, interest rates, or p/e ratios do not allow us to predict what will be the case. It is unpredictable because we can’t tell with any certainty where the critical point is in a self-organizing complex system. Some shocks are the lead-in to bigger shocks, and some aren’t. And sometimes you get aftershocks, but not always…
The second challenge is that even if we could conclude (or hope) that the low is in, we then wouldn’t really know if that low will get retested or not. The average pullback after a first significant rally from a bear market low has been about 50% of the first rally gains. But sometimes you get an almost full retracement and sometimes there is no pullback at all. The S&P 500 is now at 2610, up from 2350 a few weeks ago. A 50% give-back would send the market back to 2480 if 2610 is the high of the first move (which we don’t know yet either). But it is equally possible that there is only a shallow pullback of a few %, which would e.g. take the S&P to 2550 before heading higher already. That can become highly uncomfortable for investors who wait for a pullback to buy or a chance to cover shorts. On the other hand the S&P could go for a retest of the lows and then you would feel pretty stupid if you bought at 2550 or even 2480.
The market never gives us easy edges in one way or another. The best approach is not to get too hung up with any possible outcome, bearish or bullish, and try to feel what the market IS doing from a neutral point of view. Being aware of the different outcomes that have happened in previous similar occasions gives us some sense of what we might expect now. People with strong opinions about upcoming bear or bull markets are usually only prepared for outcomes that confirm their belief. It is better to be prepared for all outcomes and have some sense of the odds for each of those possible outcomes. That’s what studying those charts of historic lows can be used for.
We may be in a bull market that goes on for another 10 years. Or we may be in the early stages of a 50% bear market and a major recession. There is no way to tell because of the self-organizing properties of an economy. That’s why I go with the flow until the market shows me that the flow has changed course, and then I go with the flow again…