Unprecedented 200 Day Moving Average Swing: What Next?

When you take a step back and get a very long term perspective on the stock market, you realize that an apt analogy is a dog pulling on its leash. There is an overarching trend - which is the path set by the owner - but like a distracted puppy, the market can pull in one direction for short, intense spikes.

By watching the relative distance to its moving average we can tell where we are and how probable it is that we revert back to the very long term trend. This is what I referred to when I built a case for being bullish in March 2009: Another Reason We’ve Seen the Market Low. To my shock and horror, on November 20th 2008 the S&P 500 closed almost 40% below its 200 day moving average. It had never done that in the entire history of the index!

And in keeping with every other similar spike down, 60 days later, the index had recovered. In fact, we saw the sharpest rally to match the never before seen spike low.

We’ve now come full circle and are at the other end of the extreme. Last week, the S&P 500 closed 20.27% above its 200 day moving average. To give you an idea of how rare this is, here’s a chart of the S&P 500 relative to its long term moving average from 1950 onwards (click to see the larger version open in a new window):

Not only have we never seen this measure sink so low, we’ve also never seen it recover so fast! It took only 206 trading days for the S&P 500 to go from being 39.79% below its 200 day moving average to being 20.27% above it. The only other time we saw a similarly quick lurch from the abyss to the heavens was when bell bottoms and butterfly collars were the rage.

Back then, it took the S&P 500 only 185 trading days to go from being 28.8% below its 200 day moving average (on October 3rd 1974) to being 21.6% above its moving average (on June 26th 1975). That tidbit may be fascinating to know but I’m sure you’re wondering, so what happens when we are this far above the 200 day moving average?

To answer that, I went through the daily S&P 500 data and marked the appropriate dates where we approached or crossed the 20% Maginot line. Since such instances tend to clump together, I only considered the first date and ignored the repeats during the same month.

So for example, on May 5th 1975, the S&P 500 traded at 90.08 which was 19.66% above its 200 day moving average. Then a few days later on May 9th it closed slightly higher relative to its long term moving average (20.10%). I ignored all such repeats. I then calculated the forward 1 month, 3 month and 6 month returns (excluding dividends) from the first date. Here are the results:

The first thing you’ll notice is that there are very few examples: only 13 out of almost 15,000 trading days!

This is partially because of how I ignored repeat occurrences within a month. But a much more significant reason is that the market grinds out important tops slowly. Go back and look at the October 2007 top. It isn’t a spike. The S&P 500 petered out as it lost leadership and momentum. In contrast, bottoms are forged in the hellfire of fear and panic, leading to sharp spikes that pierce the chart like a needle.

As you can see, I also included a few ‘outliers’. These were times when the data was not as close to 20% but overall, it looked like a spike on the chart close enough to 20%. One of these outliers even caught the 1987 crash.

The data tells us that on average, the short term performance of the stock market when we are this high above the long term moving average is about -1%. That’s a return of approximately -13% annualized. The 3 month return is slightly worse while the 6 month return brings us back into the positive.

Not surprising. Especially when you consider everything else we’ve been looking at lately. In last week’s sentiment overview we saw how the option traders are extremely bullish. As well, yesterday we went over Lowry Research’s analysis of the market and their expectation for a correction.

Comments

[doubleguns:]

Nothing from the great depression, nothing to compare against. This sucker is not like the years in the chart above.

Closest thing is the 80's and the chart says up.

I'm not convinced that is right.

Great data, I am just not sure its in play here.

Sep 22 09:41 AM

[chap08:]

The other way of looking at the historical data is to ask - where were we in relation to the 200 day MA immediately prior to big crashes? The answer is that we are always close to, or below, the average before the real damage is done. This is true whether you are talking about tired markets rolling over (as in 2001 or 2008), or "sudden" crashes such as 1987. That's not to say the market doesn't revert to the mean - it does. But you usually get fair warning of a big crash, and this is not the indicator to warn you.

For the benefit of doubleguns above, this applies to 1929 and the 1930s too.

Sep 22 10:44 AM

[Roger Knights:]

A chart I saw yesterday had SPY over its upper Bollinger Band.

Sep 22 10:56 AM

[Larry House:]

I just don't think anything from the past can be predictive about what we are in. I do think it is a time for extreme caution. We do not have sufficient justification for the present market level or for the "all clear" that is so often sounded.

Sep 22 12:22 PM

[contracontr...:]

We will have to get used to markets overshooting on good/bad news more than they did in the past,as more market participants than ever have the ability and knowledge to change their positioning based on economic/corporate news,because of the internet,cheap trading costs,widely spread use of stop-loss orders..It's pretty certain markets will spend more time 20% above/below long term moving averages in the future,compared to the past 50 years.

Sep 22 12:58 PM

[jeremiah74:]

a. there is a long term upward bias in stocks because of inflation + equity risk premium. hence on average the 6month return is slightly positive.

b on the other hand the variance around the average is so wide that i dont think there is much to read into this apart from a.

Sep 22 01:13 PM

[aussiereader4:]

Just trading at the same level for a month or two will see it come back to the average. That is the nature of moving averages.

You can take any statistic and read what you want into it. All you have done is plot the history of a financial collapse and rebound rally using the 200 day moving average. Any collapse would look the same.

Sep 22 04:55 PM

[Audio Tactics:]

"1933 saw the S&P 500 59% rich to its 200-day" according to David Fry's post on 9/21/09.

Sep 22 08:40 PM