INEFFICIENT VS EFFICIENT MARKET HYPOTHESIS

"The experience of being proven completely wrong is salutary. 
No economist should be denied it and none are".   John K Galbraith.


David McMinn

Moon - Sun Finance

Background.  

Efficient Market Hypothesis (EMH) was originally proposed in the 1960s in a PhD by Eugene Fama, who believed that investors made well informed and intelligent decisions. Markets were considered to be efficient and rational in determining financial prices. At any given time, individual stocks were regarded to be priced at the correct level based on all known information. This was supposed to be ensured by the ready availability of ample information and by the vast number of rational investors avidly following each stock. Prices moved with the influx of new information. Free markets, so the hypothesis goes, could only be inefficient if investors ignored price sensitive data. Whoever used this data could make large profits and the market would readjust becoming efficient once again.

Random Walk. Economists commonly considered financial crises to occur as stray events. For example, Kindleberger (1996) viewed financial crises to be “random manifestations of mob psychology and mass hysteria rooted in the individual and collective psyche”. The timing of panics was widely deemed as being related to the chance occurrence of external events such as bankruptcies, interest rate rises and war, which change the perception of risk. According to this paradigm, investors by definition could not beat a random market, as prices could never be predicted from one moment to the next.

The random walk - efficient market theory reached its ascendancy in economics during the 1970's, but has since suffered severe set backs. Researchers have uncovered numerous stock market anomalies that contradicted the hypothesis. 

Over recent decades, the evidence is increasingly in favour of an Inefficient Market Hypothesis (IEH), which would more closely align with market reality. In the 1990's, behavioural finance discredited the theory of the 'rational investor' making informed market decisions. In the real world, investors are not rational and they make serious errors in their judgments, a view supported by a multitude of studies. More recently, there has arisen the 9/56 year cycle and Moon - Sun finance, which demolish the concept of random markets. 

The Moon - Sun Hypothesis

The 9/56 Year Panic Cycle. Major US and Western European financial crises tend to take place every 56 years in sequences, which, in turn, are interconnected in sub-cycles in multiples of 9 years. Remarkably, panics and crises fall with statistical significance in these 9/56 year patterns. Economists’ description of historical factors precipitating a panic seem logical - the scenario fits what actually happened - but it does not explain market timing. A more realistic approach is to consider that markets move in response to changing mass psychology between the extremes of optimism and pessimism. How people view particular financial opportunities change over time. A given set of circumstances may cause panic at one phase of the 9/56 year cycle, while similar circumstances in a different phase of the cycle may have no notable financial impact. It is cycles of upheavals in mass psychology that are postulated to determine the timing of panics and crises. 

Moon - Sun Cycles. Some external force was speculated to activate mob psychology during the crisis phase, causing millions of investors to react in the same distressed manner at the same time. Astrology was the first area examined, but all factors tested yielded negative results and could not linked to the markets, let alone a 9/56 year panic cycle. Traditional astrology was rejected as having relevance in financial patterns. On further analysis, numerous correlates could be established between the Moon, Sun and financial activity. This was extraordinary and completely contrary to what could have been expected from the random walk - efficient market theory. Clearly the market cannot be efficient or random, if there is a 9/56 year Moon - Sun cycle operating in financial history. It is the only business cycle known to the author that is actually statistically significant and based on definitive time units (9 and 56 years). 

There is much evidence to support a strong lunisolar influence in financial patterns. This shows up in the 9/56 year panic cycle, the timing of major financial crises, annual one day DJIA rises/falls, the biggest one day movements, DJIA highs, eclipse cycles (as distinct from eclipses) and so forth (McMinn, 2006). 

Lunar Nutation Cycle. Diagram 1 gives the ecliptical position of the lunar north (ascending) node at the time of major financial crises listed by Kindleberger (Appendix B, 1996) for the 1760-1940 period. The north node never appeared between 255 & 340 E⁰, a segment of 85⁰.

Diagram 1   NORTH NODE ECLIPTICAL POSITION & FINANCIAL CRISES

Annual One Day Falls are the biggest one day % fall in the year commencing March 1. AOD falls (=>4.50%) for the DJIA correlated exceptionally well with lunar phase for the period 1910 to 2005. Of the 29 AOD falls in Diagram 2, 28 appeared approximately in two diagonally opposite 90^o segments - first quarter to full Moon & third quarter to new Moon (significant p < 10^-5). Only the 1930 AOD fall deviated from this observation. Interestingly, pre 1910 AOD falls and AOD falls less than 4.50% did not show this lunar phase effect. The Sun's position on the ecliptical circle was also highly relevant. Of the 29 major AOD falls (=>4.50%) only three happened in the 4.5 months beginning November 1 (significant p < .01), with none occurring after 1928. (NB: The AOD fall for 2001 was taken as September 11, the day of the World Trade Center attack.)

Diagram 2             LUNAR PHASE & DJIA AOD FALLS

Annual One Day Rises. Of the 32 AOD rises (=>4.00%) during 1897-2000, 23 had lunar phase in the 200-020 A^o half angular circle, a findings that was only marginally significant (p < .05). However, all 13 AOD rises in the three months to November 15 occurred with lunar phase between 190 & 015 A^o. There were no exceptions (significant p < .01).

The 1929 & 1987 Panics. One of the most remarkable Moon - Sun parallels occurred between the great panics of 1929 & 1987. Intervals of precisely 717 and 718 lunar  months appeared between the record highs, the October highs, the black days, the recoveries and several other major shifts in investor sentiment. In 1929, there were 56 days between the record high and the black day, while the comparable interval in 1987 was 55 days. Of course such regular intervals cannot occur in a random walk - EMH and one of the theories must be completely wrong - the EMH or the Moon - Sun hypothesis.

October Panics & Lunar Phase. There are two types of October panics - those that occur a few days prior to a new Moon and those taking place around the full Moon. This has been a consistent trend for more than 200 years.

6 major October panics were listed by Kindleberger (Appendix B, 1996). 

There have been 10 DJIA AOD falls (=>3.60%) since 1896 that took place in October. (NB: The annual one day (AOD) rise or fall is the greatest % one day movement in the year commencing March 1.)

Combining these two lists gives 12 events, all of which have lunar phase between:
*    150 & 205 A^o, 1847, 1897, 1907, 1927, 1937, 1989. (Within a few days of a full Moon.)
*    315 & 350 A^o, 1799, 1857, 1903, 1929, 1987, 1997. (A few days prior to a new Moon.)

No exceptions. Strangely, 8 of the 12 events happened in a year ended in 7, where as 1.2 could have been expected by chance.

DJIA Peaks & Troughs. More recent research has shown that market peaks and troughs of DJIA bear markets since 1896, can be correlated closely with lunar phase. If you obtain a complete listing of peaks and troughs, no significance will be realised. However, if you take the ecliptical position of the Moon, correlates can be produced for lunar phase with the DJIA peaks and troughs. If you take the eclipiticla position of the Sun, the significance can be achieved for DJIA peaks. (Lunar Phase Significance: DJIA Peaks & Troughs).

Moon - Sun Influence. The 9/56 year cycle must work on several levels before it is reflected in financial trends. The Moon and Sun are hypothesised to emit scientific forces (presumably tidal), which influence the physiological cycles of the general population and thereby impact upon peoples’ feelings of well being. The collective mood fluctuates through cycles of optimism - crisis - fear, in harmony with the Moon and Sun. These mood swings must filter through the financial structures and fashions prevailing in a particular era before showing up in patterns of financial prices and indices.

Evidence to support the Moon - Sun hypothesis is only statistically significant in relation to large populations. It is impossible to foresee how one person will behave during acute market events. Even so, the prospect of predicting when millions of investors are likely to react adversely on extreme days is becoming increasingly promising. If Moon - Sun cycles can be unraveled to predict financial trends accurately, it will be curious to see how the main players react. According to EMH, this new information would be fully exploited by rational investors and the Moon - Sun anomalies would disappear from financial patterns. Only time will tell.

Two 2001 papers by academics from the University of Michigan firmly supported a lunar phase effect in market activity. The market tends to rise on a new Moon and fall on a full Moon with statistical significance and applies to most world markets (Kathy Yuan et al, 2001 & Ilia Dichev & Troy James, 2001). Similar correlates were made during the 1970's for the DJIA (eg: Matlock (1977); Guarino (1978)), but such early findings were ignored by mainstream economics. An outline of my own research is presented on this web site and firmly supports the Moon - Sun hypothesis. The first paper on the 9/56 year panic cycle was presented by McMinn (1986) at an economics conference in Melbourne, Australia. 

9/56 year cycles may also be relevant in cycles of other phenomena, not only market activity. Curiously, the timing of major historic earthquakes can fall selectively in these patterns, probably due to lunisolar tides triggering a build up of geological stress. These cycles may also have some relevance to the weather, although no evidence can be offered to support such a proposition.

The Decennial Cycle 

The decennial cycle is another anomaly that shows up in US stock market patterns. Under this scenario, the US market bottoms in a year ended in ‘2’ and then progressively rises to a peak in a year ended in a ‘6’ or ‘7’ and experiences a crisis and slump. The market rises to another peak in a 9 or 0 ended year, followed by another market collapse. During the 2000’s, the market has been following the decennial cycle according to plan. It hit a bear market low in October 2002 and had been rising ever since.

The decennial cycle can be used effectively for stock market speculating. According to R W Miller of Triple Screen Trading, “if one were out of the market at the beginning of the ‘0’ year, entered the S&P 500 on June 30 of the ‘2’ year, then were out from August through October of the ‘7’ year, and finally re-entered until the end of the ‘9’ year, the value of $1 invested in 1900 would be worth $6,660.86 in 2002 versus just $148.41 were you instead fully invested over the entire period of time. An awareness of the 10-year cycle would have produced 44.9 times the return”. An investor obviously would have done very well over the long term, by playing the market according to the decennial cycle. This should not arise if markets were random or efficient.

Other Anomalies 

Some of the findings in Moon - Sun cycles are quite remarkable, but numerous other anomalies arise, which further discredit the EMH.
Sunny Day Effect  - weather (David Hirshleifer & Tyler Shumway)
Seasonal Affective Disorder - solar photoperiod (Kamstra, Kramer & Levi)
Daylight Savings Anomaly - solar photoperiod (Kamstra, Kramer & Levi)
Temperature - warm/cold weather (Cao & Wei)
Geomagnetic Storms - sunspots (Krivelyova & Robotti).
Numerous other anomalies were listed by Russell & Torbey (2002).

Inefficient Market Hypothesis

Much of EMH is untestable, unverifiable and non-science and thus can never be confirmed or negated through rigorous assessment. This resulted in protract disputes between those academics who supported the EMH and those who did not. According to Roll (1997), "EMH (is) one of the most controversial and well-studied propositions in all the social sciences. It is disarmingly simple to state, has far-reaching consequences for academic pursuits and business practice and yet is surprisingly resilient to empirical proof or refutation. Even after three decades of research and literally thousands of journal articles, economists have not yet reached a consensus about whether markets - particularly financial markets - are efficient or not".

In contrast, the Moon - Sun hypothesis is scientifically testable and thus competing views relating the Moon and Sun to earthly experiences may be confirmed or negated. Such findings are reproducible in subsequent studies and have a high degree of predictability, both of which are scientific criteria. Numerous hypotheses can be tested and, based on the findings, can be expanded upon or rejected. Thus it will be possible to develop a valid scientific theory based on market behaviour, arising from Moon - Sun cycles. 

Clearly, EMH will have to be revised in the light of the numerous inefficient anomalies. Hence the proposal for the Inefficient Market Hypothesis (IMH) and the Non Random - Inefficient Market Theory. There are three forms of the IMH hypothesis:

The "Weak" form regards there being only a limited correlation between the stock market and Moon - Sun cycles. The weak form would be of limited use in making accurate predictions or profitable trading, but would be of interest in scientific terms.

The "Semistrong" form asserts that there is a general trend for the markets to follow Moon - Sun cycles. 

The "Strong" form considers there to be intimate links between Moon - Sun cycles and market activity. Surprisingly, there is support for the strong version, especially during times of extreme market behaviour (McMinn, 2006). This was supported in relation to lunar phase in relation to DJIA AOD falls (1910 to 2005) and DJIA historic market peaks.

In Conclusion

How does EMH stand up in the face of the 9/56 year panic cycle, Moon - Sun influences and the many other anomalies? Not very well unfortunately, as such anomalies devastate the EMH paradigm.
*       Investors do not behave rationally in their decisions.  
*       Financial markets function with mathematical structure and are non random.
*       Free markets allocate financial resources and determine prices very inefficiently.
Clearly there is a contradiction and one of them has to be completely invalid - the Moon - Sun hypothesis or economic orthodoxy. Free markets are not efficient nor are they random, despite what the EMH would have us believe. IMH is far more reflective of the market and the real world.  

The main theme to emerge is the need for much more research. Numerous questions remain unanswered. How relevant are 9/56 year patterns and Moon - Sun cycles in recent decades, especially with the emerging global economy of the 21^st century? What role do tidal harmonics and the Fibonacci - Lucas numbers play in the markets? How important is tidal resonance in solving the 9/56 year enigma? Some of this research may have already been done. If excellent correlates were produced, the results may never be published due to the potential profits to be made. Further questions arise in relation to natural phenomenon. Do major world earthquakes and volcanic eruptions fall within the 9/56 year cycles and are they activated by Moon - Sun tidal triggers? Is weather similarly influenced?

The Moon - Sun finance firmly endorses those analysts who consider past performance to be a strong indicator of future market trends. There are reasonably regular financial patterns rather than randomness. Thus, various technical analytical theories may have some validity. The problem is to decipher such patterns, which are complex and hard to decode. Even so, there may be emerging a simple unified theory based on the Moon - Sun influences in market activity. This would reduce the complexity of market cycles to a few basic principles, which could be of immense benefit in understanding many financial phenomena. It also offers the potential to make accurate predictions years in advance. 

IMH reduces the whole concept of free markets to an absurdity. Trillions of dollars in financial assets are traded every day worldwide. This is supported by a massive investment in research, trading infrastructure, communications and so forth. All this activity for a world financial market that is largely based on the heavenly positions of the Moon and Sun. Free markets are inefficient and at times ridiculously inefficient, as witnessed by the numerous manias and panics of recent centuries. A more realistic evaluation is essential if economic theory is to progress. This will take years given the rigidities within academia. Alas good conservatives are always slow learners. By definition, they cannot cope with new ideas, not even good new ideas.

© Copyright. 2005-2007 David McMinn. All rights reserved.

References

Dichev, Ilia & James, Troy. Lunar Cycle Effects In Stock Returns. University of Michigan Business School working paper. 2001.
Funk, J M. The 56 Year Cycle in American Business Activity. Privately published. 1932.
Guarino, Frank J. Relationships of Stock Market Fluctuations to The Lunar Cycle. A thesis presented to the Faculty of the Graduate School, Pace University. American Federation of Astrologers. 1978.
Kindleberger, C P. Manias, Panics & Crashes. John Wiley & Sons. Published 1978. Revised 1989. Revised 1996.
Roll, W. Market Efficiency: Stock Market Behaviour in Theory and Practice. Edited by Lo, Andrew W. 1997.
Matlock, Clifford Charles. Man & Cosmos: A Theory of Endeavor Rhythms. Development Cycles Research Project. 1977.
McMinn, David. The 56 Year Cycles & Financial Crises. 15th Conference of Economists. The Economic Society of Australia. 25-29 August, 1986.
McMinn, David. Market Timing By The Moon & The Sun. Twin Palms Publishing. First published: 2002. Revised: 2006.
McMinn, David. Market Timing By The Number 56. Twin Palms Publishing. First published 2002, Revised 2006.
Yuan, Kathy, Zheng, Lu & Zhu, Qiaoqiao. Are Investors Moonstruck? Lunar Phase & Stock Returns. University of Michigan Business School working paper. 2001.