The Amazing Four Year Stock Market Cycle
Published May 13th, 2007 in Technical Analysis Tags: 4 year cycle, business cycle, cycles, Dow Jones, emh, four year stock market cycle, internet mania, investing, long term investment, market cycles, presidential cycle, stock market cycle.Is it possible to time the market using a system that is so simple, it only requires you to be able to count up to four? Can we invest for the long term using a system that only requires a few minutes of our attention every four years?
Not only is it possible, such a system has beaten the pants off the pros in the long term. Being so amazingly successful, it has garnered a name: the four year stock market cycle.
Many theories have been put forth to try and explain it. Some say it is due to the presidential cycle, some that it is due to the business cycle, some to astrology or other esoteric phenomena. While the reasons are up for grabs the results are quite clear. And they are the sort that makes EMH proponents pull their hair out in frustration. How can something so simple, so replicable, and so consistent exist decade after decade?
While academics debate it, you can use it to boost your long-term investment account. All you need to do is to watch for a low every four years. The start year is important, so I’ll give it to you: 2006. From that year, you can go back and forward in four year increments. Those years will be (in the future) or were (in the past) great times to invest in the stock market. Or to add to an existing investment portfolio.
Take the previous cycle: 2006. That was the last intermediate low. Before that, in 2002 we had the trough of the multi-year bear that resulted from the popping of the internet mania. The one before that? 1998 which was the trough from the Asian currency contagion that shook financial markets. Keep going and you’ll see that the four year cycle marks great buying opportunities in an uncanny way.
Of course, it doesn’t have a perfect track record. But it is a damn good one. Out of the last 27 four year cycles, only 5 of them have not been great buying opportunities. They were 1946 (flat), 1930 (ouch!), 1910, 1906 and 1902. You can keep going back in time but my chart (below) stops there:
Click to enlarge graph
For extra mojo, we can combine the 4 year cycle with the annual cycle. That is, we can take the best month within the year (historically October) that coincides with the four year cycle. But remember, the four year cycle is only a guide. No individual occurence has to follow the script to the letter. All we are interested in is putting the odds in our favour as we have detected them from historical observation.
The following graph shows the decennial performance of the Dow Jones Industrial Average. Think of it as putting the annual performance of every decade into a blender and mashing them together. We get a graph that shows the average performance of each year within a decade:
As you can see, year 4 has an unusual lift that the other years don’t. If you’re sharp, you’ll notice that years 7 and 8 also have some pretty good kick as well. This data is not showing the same thing as the four year cycle but it does present us with further proof that returns from the stock market are not random. If they were, then each year would show fairly similar performance.
4 year cycle, business cycle, cycles, Dow Jones, emh, four year stock market cycle, internet mania, investing, long term investment, market cycles, presidential cycle, stock market cycle
Enjoyed this? Don't miss the next one, grab the feed or
8 Responses to “The Amazing Four Year Stock Market Cycle”
- 1 Trackback on May 14th, 2007 at 9:32 am
- 2 Pingback on Jan 4th, 2008 at 12:52 am
- 3 Pingback on Jan 18th, 2008 at 6:20 pm
Leave a Reply
Search
Creative Commons License
Trader’s Narrative is powered by
1. The second graph in the article should show the distribution of gains averaged over groups of four years, not ten years, lest readers become concerned that the four-year period graph might invalidate the four-year cycle.
2. Even if the true averages of the performance of each year in a decade are actually equal, averages calculated from samples of decades usually aren’t. Thus, it requires a statistical analysis to justify concluding that unequal averages calculated from a sample of decades would stay unequal if calculated from an infinite number of decades. What I’m suggesting is that the extra work required in getting the ten-year period plot to prove anything would be better directed toward drawing the four-year cycle plot.
3. I’m not debunking the four year cycle. I would just appreciate an upgrade in the presentation of the second graph, so that it’s as clear and telling as the first.
Every year, the stock market has two corrections: a mild one in the spring and a more severe one in the fall.
Every decade, the stock market has two severe corrections (called recessions) caused by the inevitable movement of the economy and its business cycles. At the peak, overly strong demand pushes price levels high enough to throttle demand and cause excess production to produce excess inventories. It usually takes 12 to 18 months of economic recession to correct the excesses built up during the 4 to 6 year bull market.
For individual investors, it is easy to spot the turning points at the crests and troughs in the 4 to 6 year economic cycle. Simply calculate the 12 month moving average of the value of your favorite index fund. Switch into money-market funds (or the new bear ETFs) whenever the index is below its 12 month moving average.
The more widely followed 200 trading day moving average gets you in and out of the market sooner, but is a little less convenient to calculate (unless you use Stockcharts.com or a similar charting website). Right now, the Russell 2000 small cap index, my personal favorite, would have to slide down 8% to get below its 200 day MA.
John,
1. I don’t have that graph handy 2. yes, a reader from Sweden pointed that out also (via email)
Herb, thanks. Sometimes the simplest works best. According to Hulbert, one of the best timing newsletters only relies on a simple MA similar to what you suggest.
The Number 23
Pi
Add the digits of 23, minus one, 2 3-1=4 !!
Sorry to rain on the party…
5/27 (track record of 4-year cycle) => wrong almost 20% of the time
Care to sell some cheap puts?
very interesting.I am glad you have writen this as I have been looking back and seeing this and other pattern’s for quite some time.