Abstract. This paper examines the relationship between lunar phase and the seasonal timing of DJIA peaks. When peaks at the beginning of a bear market (≥ -20%) occur at around the same time of year, they will usually have similar lunar phase. This was a reasonably consistent trend since 1890 and numerous examples have been presented in the paper. The only notable exceptions were the 8 bear markets commencing between September 26 and December 10. Why these patterns arise in financial history remained unknown. Given the role played by lunar phase and seasonality, the answer involved the changing angles between the Moon and the Sun in the heavens. Much more research is required to decipher this enigma.
Keywords. Dow Jones Industrial Average, bear markets, peaks, seasonality,
Major DJIA peaks occurring at the same time of year often had similar market outcomes (McMinn, 2010b). This concept was expanded upon and the highs at the beginning of a DJIA bear market were assessed in relation to lunar phase. Those peaks forming around the same time of year usually have similar lunar phase. It has been a persistent trend over the past 125 years and many examples have been given in the text. Only the season between September 26 and December 10 did not exhibit this phenomenon. These findings may also apply to the beginning of DJIA corrections, although more research is necessary before any firm conclusions can be drawn. How lunisolar cycles actually drive market activity is completely unknown. Seasonal and lunar phase effects would imply that the heavenly positions of the Moon and Sun are important in solving the problem. Alas nothing more can be stated.
An extensive coverage of DJIA bear markets since 1900 was sourced from Bespoke Investment Group (2008) and has been reproduced in Appendix 1. The author inserted additional bear markets for the 1890-1899 period, as well as four DJIA corrections that registered falls of between -18.5% and -19.9%. (A bear market is commonly defined as stock market fall of over -20%, preceded by a rise of over +20%.) Extreme volatility resulted in 8 bear markets in the five years to 1933, which would distort any findings based on data including this period. To overcome the problem, only the key bear market for the early 1930s (Sep 3, 1929 to Jul 8, 1932) was included in the listing. DJIA closing values have been used throughout the paper.
Annual one day (AOD) rises and falls are exceedingly important in financial patterns, a finding strongly supported by numerous correlates (McMinn, 2000, 2004, 2009). An AOD rise or an AOD fall is taken as the biggest one day percentage rise or fall in the year commencing March 1. They represent the biggest one day shifts in market sentiment during a given year. Moon Sun data was timed at noon (New York) on the relevant trading day. Degrees on the ecliptic circle have been given as E°, while the angular degrees between the Moon and Sun (lunar phase) were abbreviated to Ao. Thus, 00 Ao denotes a new Moon, 90 Ao a first quarter Moon, 180 Ao a full Moon and 270 Ao a third quarter Moon. For some panics, the AOD fall recorded the actual low (eg: 1987, 1997 and 1998). However, the low after the panic was called the post-crash low in this paper. The term season has been taken as a given period during the solar year and did not apply to the traditional seasons – spring, summer, autumn and winter.
The layout of this paper is as follows. In Section 2, the peaks at the beginning of a DJIA bear market were arranged in increasing order by month – day (year ignored). The listing was then assessed in relation to lunar phase. Those peaks taking place at around the same time of year usually had similar lunar phase. In Section 3.0, the start of DJIA corrections may also show seasonal and lunar phase effects. Correction highs in 1950, 1960, 1980, 1997 and 2011 aligned in the same patterns established for corresponding bear markets. Section 4.0 concluded that the timing of major seasonal DJIA highs may be intimately linked to lunar phase. How lunisolar cycles actually influenced financial activity remained a mystery.
2.0 DJIA Peaks and Lunar Phase
Lunar phase cannot be correlated with the full listing of the 33 peaks at the start of a bear market. However, excellent relationships could still be realized if these events were listed by month – day (year ignored), as shown in Appendix 2. All 14 highs between February 10 and September 3 had lunar phase between 270 A° and 080 A°, a range of 170 degrees. Even better outcomes could be achieved by breaking the seasons into small subsets.
2.1 August 1 – September 10. There were five tops during this season and all had lunar phase around the
new Moon or the full Moon (see Table 1).
All five biggest one day percentage falls in DJIA history (≥ -8.70%)
occurred after the new Moon peaks in 1899, 1929 and 1987.
The remarkable parallels between the
1929, 1987 and 1997 October panics have been commented upon by Carolan (1998) and McMinn. The lunar phase similarities for the highs,
panics, recoveries and lows for these events were summarized in Table 2. (The 1997 correction
involved a decline of -13.2%.) For these panics,
lunar phase occurred in the following ranges:
The AOD falls from 310 Ao to 325 A°.
The AOD rises from 330 Ao to 350 A°.
The post-crash lows from 135 to 175 A°.
The 1895 and 1899 highs occurred on September 4 and September 5 respectively with the former happening around the full Moon and the latter around a new Moon. The ensuing AOD falls took place at about the same time of year on December 20 and December 18 respectively, with the AOD rises occurring a few days later. The 1897 peak was anomalous as it did not align with trends in Table 2.
2.2 September 11 to September 25. Two highs happened on September 12, 1939 and September 21, 1976, with lunar phase at 349 A° and 330 A° respectively. The 1939 high was followed by panic in May 1940, although no panic happened after the 1976 peak.
2.3 September 26 to December 10. This season contained 8 DJIA peaks, but there was no lunar phase clustering. Why this anomaly arises in financial cycles is puzzling.
The peaks of September 30, 1912 and October 9, 2007 formed at about the same time of year, but there was no common lunar phase (see Table 3). Even so, the DJIA top for 1912 was debatable as it was recorded at 94.15 on September 30. This was only marginally higher than the equal second rank highs at 94.12 on both October 3 and October 8. If the 1912 high had have been recorded on the latter date, it would have aligned very closely with lunar phase on October 9, 2007.
2.4 December 11 to January 15. There were three highs during this season in 1961, 1973 and 2000, all of which had lunar phase between 75 A° and 95 A° (see Table 4). The correction commencing on January 5, 1960 also had lunar phase within this range. There were no parallels for the timing of the ensuing panic or market low.
2.5 January 16 to February 9. The three bear markets commencing in this season had lunar phase between 235 A° and 295 A° (see Table 5). The 1934 high was followed by a stock market panic on July 26 and the 1966 high by a US financial crisis in August.
2.6 February 10 to April 28. Lunar phase ranged within 60 degrees for the peaks in 1937 (332 A°), 1956 (307 A°) and 1981 (274 A°) and within 45 degrees for the highs in 1892 (076 A), 1923 (034 A°) and 2002 (062 A°) (see Table 6). This and August 1 - September 10 were the only seasons that gave two clusters widely spaced apart in the lunar phase circle. The 1980 correction commenced on February 13 with lunar phase at 324 A°, which was close to that recorded for the 1937 and 1956 peaks. Lunar phase at the start of the 2010 correction (156 A°) did not align with other peaks in this season.
2.7 April 29 – June 30. All four peaks in this season had lunar phase between 320 A° and 015 A°, a range of 55 degrees (see Table 7). The 1950 and 2011 correction highs also had lunar phase within this range.
The highs in 1901, 1946 and 2001 showed up in the month to June 20 (see Table 8) and were each followed by an initial panic in early September with another major one day fall 6 days later. The associated major one day rises took place in the month to October 15.
For 2008 and 2011, the tops occurred on May 2 and April 30 respectively, a little earlier in the year than in 1901, 1946 and 2001. On the day of these two peaks, lunar phase was at 321 A°, followed by two major one day falls 6 days apart in early August and mid-October. Importantly, the AOD rise in 2008 occurred four days after the first one day fall, whereas the AOD fall showed the same timing in 2011 (see Table 10).
2.8 July 1 – July 31. The July 16, 1990 and July 17, 1998 highs had lunar phase at 286 A° and 284 A° respectively (see Table 11). The panics occurred in August and the bear market slumps were both around -20%. Lunar phase did not align closely for the AOD falls or the post-crash lows.
The link with seasonality and lunar phase may also apply to the beginning of DJIA corrections. There is some evidence to support this speculation and examples have been given in Table 12 for selected corrections (≥ -12.5 ≤ -18.5%) since 1950. The correction highs in 1950, 1960, 1980, 1997 and 2011 exhibited similar seasonal and lunar phase effects as established for historical bear markets. However, this was not repeatable for corrections starting in 1953, 1967 and 2010 and thus there was no overall consistency. Table 12 did not present a complete list of all historic corrections over the past 65 years and a more detailed study is essential before any firm conclusions can be reached.
4.0 Discussion and Conclusions
Lunar phase could not be correlated with the full listing of the 33 peaks at the beginning of DJIA bear markets as presented in Appendix 1. Excellent links between lunar phase and these DJIA peaks could only be accomplished by breaking the sample down into small seasonal subsets. The DJIA highs occurring at about the same time of year usually had very similar lunar phase, with the findings being summarized in Table 13.
The relationships between seasonality, lunar phase and the beginning of DJIA bear markets held up very well over the past 125 years. The only major anomaly was the September 26 – December 10 season, when no lunar phase effect was observed. All other seasons experienced a clustering of lunar phase in restricted ranges. The findings supported the Moon Sun hypothesis, which viewed financial markets as being structured mathematically in time and moving in tune with lunisolar cycles (McMinn, 2004, 2010, 2013). They also offer clues for the design of follow up studies, which are essential to expand upon the concepts presented in the paper.
Lunar phase can be an indicator of an impending bear market. When significant peaks do form, it is worthwhile to check its lunar phase to see if it aligns closely with historical peaks in Appendix 2. Using this method, the 2011 panic was predicted, with two large one day falls 6 days apart, However, it happened earlier than expected – in early August rather than September – October. This approach is primitive and much more research is required before lunisolar cycles can be upgraded into a useful forecasting tool.
The commencement of bear markets can be linked directly to Moon Sun cycles, given the key roles played by lunar phase and seasonality. The relationships are very interesting, but they cannot explain how the Moon and Sun could possibly influence trading activity. Obviously, the changing angles between the Moon, the Sun and the spring equinox point (000 E°) were vitally important in solving the mystery. That is all that can be stated. Even so, there may be emerging a simple theory based on Moon Sun tidal harmonics, which would reduce the complexity of market cycles to a few basic principles. Such a paradigm shift would offer the potential to make accurate financial forecasts years in advance. Even when this breakthrough is achieved, it probably will not be published given the potential profits to be made.
Carolan, C. (1998). Autumn panics: a calendar phenomenon The Market Technician. Journal of the Society of Technical Analysts. p 12-18. Issue 32. July.
McMinn, D. (2000). Lunar Phase & US Crises. The Australian Technical Analysts Association Journal. p 20-31. January/February.
McMinn, D. (2004). Market
Timing By The Moon & The Sun. Twin Palms Publishing. 153p.
McMinn, D. (2010b). DJIA Peaks, Seasonality and Market Outcomes. Market Technician. Journal of the Society of Technical Analysts. Issue 67. p 10-13. October.
McMinn, D. (2013). Fibonacci – Lucas Numbers, Moon-Sun
Cycles and Financial Timing. Market Technician. Journal of the Society of
Technical Analysts. Issue 75. p 9-13.