- Approximately 20 year bull market
- Stock bubble and crash
- Recovery and financial panic
- Recovery and recessionary period
The analysis is timely because I believe we are currently at the peak of the second recovery and heading into what I call the recessionary period. My first article delved into technical analysis and Elliott Wave theory. This article will focus on the historical record and investigate how the cycle affects and is affected by with macro-economic and political circumstances.
(A quick note regarding Elliott Wave. My original article (I now believe) incorrectly classified the two recovery periods as peaks in a triangle pattern. It appears that in fact the final recovery period is the 1st leg of a motive wave with the recessionary period being the 2nd wave retracement. My reclassification is largely academic as the market's path remains the same).
This is certainly not the first article to identify long term stock cycles, however the vast majority of literature regarding cycles focuses on fixed time periods. While it is certainly compelling to see financial crashes occurring with regularity I don't believe that this is very useful as a tool for determining when one should be in or out of the market. I believe a more accurate rendering of the stock price cycle can be made using milestones based on the character of market peaks and valleys. The analysis in this article might convince you that both the amplitude and time (x and y axis on a chart) are variable and non-deterministic even while the cycle itself appears inevitable.
The Charts
I first discovered this pattern by scanning through historical stock charts with my own charting software. Many charting platforms do not include historical data and long term charts mask the pattern because of the extreme swings on non-logarithmic charts. Even on logarithmic charts the pattern is difficult to see because the amplitude and timing of the pattern varies with each period. Finally, prior to 1950 the best record is the Dow Jones Industrial Average which is a poorer reflection of stock market levels than broad market indexes such as the S&P 500.
I've managed to identify four instances of the pattern as demonstrated in the following chart:
I've drawn in the period from 1885 through 1896 because my data does not go that far and I needed to extrapolate. Also the period from 1914-1949 did not fit cleanly in a screenshot so I penciled in the tail end of the 1946-1949 recessionary period. Otherwise, each chart begins where the previous chart left off, and obviously, the current period is yet to play out.
The first observation of these charts is that they are indeed wildly different, especially in regard to amplitude. It would be difficult to identify 1929/1937/1946 as comparable with 2000/2007/2011 based on the look of a chart alone. Let's therefore look at each chart and attempt to identify some of the variables that affected the cycle.
1885-1914
This cycle saw an incredible blossoming of technology and productivity that began with the Victorian era railroad boom and culminated with the beginning of World War I. I mark 1885 as the beginning because I've identified the panic of 1873 as the panic leading into this cycle, and the period from 1873 to 1885 as the prior "setback" period.
The stock bubble during this period was a dramatic double top in 1899 and 1901 which I believe was affected by the turbulence of the time. In 1898 the US fought the Spanish American war. From 1899-1902 the US was engaged in the Philippine-American war while the British fought the second Boer war. As we will observe in future cycles, armed conflict tends to reduce the amplitude of stock waves. The most dramatic waves will be observed as those that occur during times of peace (the so called peace dividend). Had peace been prevalent during this period then I believe the stock market top would have been at the same level as the 1907 and 1909 recoveries.
The stock market crash of 1901 saw the plummeting in value of railroad stock, the dot.com of the late 1800s. Then 1903-1906 saw a rapid recovery, no doubt enhanced by the ending of the previously mentioned wars. Much like today, an Earthquake was responsible for taking the head off of the rally but this only revealed underlying weaknesses that culminated in the financial panic of 1907. Much like the panic of 2007, the financial system ended up on the verge of collapse, rescued only be the collaboration of government and Wall Street.
This article (The Remarkable Efficiency of the Dollar-Sterling Gold Standard 1890-1906) had the same perfect timing as Ben Bernanke's "Great Moderation" speech of 2006. In the context of a 20 year bull market and recent recovery a monetary theorist can be forgiven I suppose. As it turns out of course, monetary expansion had reached it's tipping point and signaled the end of the cycle.
The recovery into 2009 was swift but (as I predict will occur for the 2011 market) a new high was narrowly missed and a 25% bear market correction took place. The stock market then went "dead" until 1914. A World War was certainly good reason for the stock market to drop suddenly. It is interesting to observe that even under these conditions the market found support at the level of the previous financial panic. 1914 marked the beginning of the next cycle, a cycle dramatically affected by two world wars.
1914-1949
This period saw two world wars and a depression that severely warped both the amplitude and timeline of the natural cycle. It also saw an amazing bull period driven by the widespread adoption of mechanization, the modernization of industry and dramatic advances in transportation and communication. While the shape of the 1920s bull market is dramatic, the overall growth is roughly equivalent to the growth seen in the 1949-1969 bull market and the 1982-2000 bull markets. World War I in this case I believe caused the growth to be compressed into a shorter time period and thus take on a parabolic shape. IF we were to instead stretch the 1920s bull market over the full 15 year period it might have looked like this:
The stock market drop was dramatic. I don't have a good explanation for why the 1930 crash went as low as it did. Elliott Wave theorists would say that this was the end of a big old wave and they might be right. What I can say though is that in 1933-1934 the dollar was devalued by 40% which I believe once again affected the amplitude of the chart. Had the dollar not been devalued the chart might have looked like this:
Finally, war broke out in 1937. The US did not enter the war until 1942 (the bottom of the "V"). I believe the effect of World War II was once again to delay the cycle. Sans WWII the chart might have looked like this:
This chart is now looking considerably more like those of the other periods including our current cycle. Still noticeably different is that the valley out of the panic is higher than the valley out of the stock crash. Also, the 1946 recovery peak is still higher than the 1937 recovery peak. I believe this could probably be attributed to further easing by the government during the course of the war. Possibly Roosevelt era economic policies had an effect on the amplitude of peaks and valleys as well (socialist structure reducing oscillation?). Regardless, what I believe is important is understanding that the cycle conforms to the pattern and that in the future one should be ready to adjust trading timing if dramatic conditions are in place.
The end of World War I saw the 1946 recovery and then slip into a recessionary period that lasted until 1949. This is the first cycle we're observing where the end of a war ushers in a recessionary period. We will see this again in the 1970s and I believe in the coming years. The end of a war however provides an economic springboard for the next bull run.
1949-1982
The post war boom was witnessed by a rather jagged 20 year bull market. The bull market was once again held back for three years by the Korean War. The end of the Korean War ushered in the first motive wave that saw the market rise 100% in less than 3 years. From 1957 through 1962 the market continued upwards but more cautiously. I believe this was due to escalating tensions with the Soviet Union. The growth of the entire bull market might have been even greater than it was had the cold war not been an ongoing drag on the US and world economies. After the Cuban Missile crisis, with tensions easing, the market once again rocketed up 100% into the stock bubble of 1969.
Once again in 1970 the dollar was devalued. This time the effect was paradoxically the opposite. Inflation took root which inflated the nominal value of the stock market and the amplitude of the 1973 top. Inflation continued to drag on the market until the early 80s.
Finally, the end of the Vietnam war in 1975 ushered in a temporary recovery and then brutal economic period (technically not a recessionary period but stagflation probably was worse). The spring board of an ending war once again set the stage however for a powerful bull run.
1982-current
Readers of this article will be most familiar with this period but perhaps not be aware that the overall growth of this bull run was almost twice as much as the prior periods. A quick look at the chart shows that the market accelerated beginning in 1994. This coincides with the introduction of the Internet but I believe the real effect was the peace dividend obtained from the fall of the Soviet Union in the early 90s. This period demonstrates the potential for a bull market that is unrestrained by armed conflict. The 90s bull market was roughly equivalent in growth to the 1920s bull market which was also a period without armed conflict. I believe that a 20 year cycle that is unconstrained by armed conflict could produce 1000% aggregate returns.
As one can follow from my logic, I believe we are closing in on the end of the second recovery period. While quantitative easing has been in effect, we have not had a devaluation in the currency. However, the Afghan/Iraq war I believe will dampen the top of the recovery much as the Vietnam war dampened the top of the 1977 recovery (relative to the 1973 recovery but not relative to the 1969 top which was pre-devaluation). It is worth noting however that the 2007 peak was higher than the 2000 peak which one would not expect given the level of military commitment at the time so it is conceivable that the market could progress to a level that would challenge the prior top (assuming one took a liberal Elliott Wave count).
Evidence From the 1800s
I haven't been able to find raw data for stock market levels from the 1800s. All I have been able to find so far is this chart from the Elliott Wave International:
The panic of 1857 followed with expected results and finally a very small "tepid recovery" shows up as a blip next to the vertical line.
The similarity between this period and the 1914-1949 period is interesting. In both periods the initial crash was much steeper than the secondary crash. Unique to the 1812-1860 period is the length of time that this cycle took. Almost 50 years passed from one point to another with 20 years separating the two peaks (however the boom period is almost precisely 20 years and so maintains that pattern). It is a bit horrifying to realize that the stock market level on this chart began and ended at the same level. That represents a 50 year period of zero returns!
Unseen on the American stock market chart is the London market crash of 1825 which I believe represents the "breather" for this period.
The following period from 1860 to 1885 is less distinct. We certainly can see the tremendous period of growth that followed the previous cycle but must keep in mind that the first few years here were during the American Civil War which was an inflationary period. The panic of 1873 was a definite stock market bubble, the second of three eventual railroad bubbles although it was accompanied by a demonetization of silver which reduced the money supply and prompted the eventual debate that we would see in the next cycle.
It is difficult to divine our pattern in this period from so small a chart. Part of my identification process is looking at the "shape" of the waves which requires a close up view.
While the period prior to 1810 is drawn in by hand, again we can make out a pattern that looks distinctly like an episode of the modern pattern from 1780-1812. Woe be to those who bought at the top of the South Sea Bubble (the spike in 1720) who paid a price level that wouldn't be seen again until 1870!
Breathers and Volatility
The concept of the breather is a recent addition to my thesis. I discovered the breather by accident as I was trying to identify the stock market bubble for the 1885-1914 period. I found the stock market crash of 1893 in Wikipedia and was confused. While there was clearly a stock market crash in 1899-1901, from the account of the Wikipedia entry the 1893 crash appeared to have been much more violent. It then struck me that this crash occurred about a decade before the bubble eventually burst, quite similar to the crash of 1987 which was considerably more horrifying than the 2000 crash both in terms of economic effect and the effect on individual investors.
I then began to look at the other periods and sure enough discovered severe corrections that occurred about midway through each bull period. "Breather" is probably a terrible term because in fact these corrections share the trait that they are severe and they are sudden. A close look at the 1987 crash or the 1962 crash reveals that there was very little time to get out and only the subtlest of warnings that it was about to happen. The rapidity of these collapses left investors startled.
Thus, part of my thesis is that one should expect a severe correction about a decade into a bull market. It would be nice to be out of the market when this happens but the real lesson is to buy the market at the bottom since the bull market can resume quite rapidly (although it can take a year as in the case of 1987). (It should also be noted that the 1950s/1960s bull market was peppered with numerous harsh corrections as was the 1890s bull market. It might be said that these markets were breathing heavily).
While my pattern I believe is repeatable it is predictive in regards to timing and amplitude in only a few places. The two decade bull market is a relatively stable variable given the absence of a major war. The magnitude of the bull market ranges from 400% to 700%. I think it is safe to assume that if the market crashes after a 200% gain that one has experienced a breather. The valley out of the initial crash is variable however. In some instances it returns to what is approximately the mid-point of the bull wave but in other instances it retraces the entire wave (1929, 1873). The amplitude of the secondary peak is highly variable as is the nadir. The third peak is generally lower than the secondary peak and the final retracement is quite reliably at the mid-point of that final rally.
In the famous 1987 documentary "Trader", Paul Tudor Jones states that he found incredible correlation between the 1929 bull market and 1986-87 bull market. He famously predicted a 1988 crash (which occurred earlier than his initial prediction although he did catch it) based on this correlation along with Elliott Wave theory. What he was dead wrong about though was his prediction as to the depth of the crash. He had expected a crash of the same magnitude as the 1929 crash. He had also expected similar economic results, namely a 5 year depression.
The correlation between these two bull periods is in fact quite evident and makes an interesting point, that bull markets absent war time conditions will look similar! Crashes will also look similar. My own analysis of market tops has found incredible similarities right down to the minor trend patterns. Yet this is not predictive as the market is subtly different at each market top. The thesis is therefore helpful as a kind of road map through those tops and bottoms. Using this thesis one would be on alert for a market crash as was Tudor Jones, however knowing that a breather was expected rather than an end point, and knowing the effects of war time conditions, one would not have expected as severe a crash as 1929 and would have been perhaps readier to buy the bottom.
Hard Money
One area that we should explore is the role of monetary policy and it's possible impacts on the cycle. Here in 2011 the call for "hard money" in the form of a gold standard has become vogue. Gold and silver prices have rocketed. Have similar situations occurred in past cycles? (It is interesting to note that the 1896 election turned on whether the country would retain a gold standard or a silver standard. The silver standard at the time was essentially a stand-in for fiat money. Fiat money itself would have been unacceptable at the time. The gold standard won the day. The price of silver did not seem to be affected by the politics).
This graphic shows the price of silver during the first three cycles. The red dot indicates the bubble peak. The green dot indicates the price at the bottom of the financial panic. The purple dot indicates the price at the peak of the final recovery heading into the recessionary period. I picked silver because the price of gold has been fixed for so many years. Silver by comparison has been relatively free to fluctuate with only small periods of price fixing.
There are few conclusions to be drawn. Certainly there are some patterns. In the first two charts the price of silver declined during the bull period. This also occurred in the 1982-200 bull period (not shown). The exception is the 1950-1969 bull period although the price was relatively flat. While a dollar index would show a declining dollar, that would be relative to world currencies. From a hard money perspective it would appear that the dollar strengthens during a bull period (as one would expect as rates should rise).
Next we see that in each circumstance silver comes alive heading into the first recovery. Gains during this period are roughly equivalent to gains in the stock market. This leap therefore I believe is probably demonstrating a rising price of all assets, most likely representing loose money heading into the financial panic. We'll investigate whether this is indeed happening.
Finally we cannot really draw a conclusion regarding the price of silver heading into the final recovery or recessionary period. The 1970s Hunt Brother's corner represents a dramatic turn upwards in the price of silver. We can also see that silver exploded heading into the World War I period but this would be expected in war time so we cannot attribute this to cycles. In fact, the 1946-1949 period saw a decline. My conclusion therefore is that the price of silver (and by proxy rates and inflation) is not determined by the stock cycle, although it does appear that if a silver bubble is going to appear that the end of a cycle would be a good place to look for one.
This concludes part 1. In a future post I will dig deeper into other economic variables and how they may or may not be affecting the pattern.
0 comments:
Post a Comment