David McMinn

Moon Sun Finance

Most economic theories regard business cycles as occurring in variable waves of progressive expansion and contraction ad infinitum. Many analysts are skeptical about using such cycles in their investment decisions, due to their poor predictability. However, financial crises, which are the most dramatic phase of the business cycle, tend to occur every 56 years in sequences. These sequences in turn are interconnected by sub-cycles based on multiples of 9 years to produce a 9/56 year grid pattern. The financial crises typically involve a period of over optimism, greed and rampant speculation, followed by crisis, fear and economic slump. Each crisis is unique and cannot be compared with other crises in a particular sequence.

Cycle theorists assume a fairly regular repetition of crises every so many years. The concept of numerous sub-cycles lasting for several decades and operating simultaneously is alien to cycle paradigms. Even the 56 year sequences, which may persist for centuries, should fade out over very long time frames. Important crisis events may also skip a beat in a particular sequence, only to reappear 112 years later and the whole cycle must change progressively never repeating itself exactly.

THE PRINCIPAL 56 YEAR SEQUENCES

Funk (1932) was the first to propose three 56 year sequences for the US economy. For convenience, all sequences in this cycle were numbered from 01 to 56 with (1761, 1817, 1873, 1929, 1985) being designated Sequence 01. Hence, Funk's three sequences were denoted as 01, 21 and 41 (see subsequently). The sequence year is taken as commencing on March 1, as this seems to be the year of best fit. No definition exists of what constituted a 'major' crisis in economic history. Thus Appendix 1 gives a list by Kindleberger (Appendix B, 1996) of what he considered were 'major' financial crises. This source was chosen given his pre-eminence as an historical economist and because he presented the most comprehensive listing of major crises over recent centuries. The list provided an external reference independent of this cycle study. Years in which these major crises occurred have been given in BOLD throughout the text. The Funk's three 56 year sequences are presented as follows.

Astonishingly in the 100 years to 1930, Funk’s three sequences contained 6 years, in which occurred five of the worst panics in US history (1837, 1857, 1873, 1893 & 1929 - the exception being the crisis of 1913). Given this highly accurate trend, additional 56 year sequences were postulated to exist in recent centuries. References on US and Western European economic history were gleaned for such patterns with the ensuing sequences being regarded as the most significant.

Within the 56 year sequences, crises often take place in approximately the same month of the year. The crisis month is given wherever possible in the text.

Sq 01 - 1873 (USA - September), 1929 (USA - October), 1985 (US$ - September).

Sq 03 - 1763 (Continent - September), 1931 (UK - September), 1987 (World - October).
                        Also 1819 (USA - April/May), 1931 (Austria/Germany - May/June).

Sq 05 - No trend.

Sq 09 - 1713 (England - Jan - April 1714), 1769 (France - February 1770), 1825 (England - 
                        December), 1881   (France - January 1882). 1993 (USA - February 1994)

Sq 12 - 1716 (England - Jan - March 1717), 1772 (Amsterdam - January 1773).
                         Also 1772 (England - June), 1884 (USA - May), 1940 (USA - May).

Sq 21 - 1837 (USA - May), 1893 (USA - May - July).

Sq 32 - 1792 (USA - March), 1848 (France - March), 1904 (France - February).

Sq 41 - 1745 (England - Dec 6), 1857 (England - Oct; Continent - Nov).

Sq 48 - Insufficient data.

Sq 50 - 1810 (England - December 1811), 1922 (Germany - January 1923).

Sq 52 - 1924 (France - March), 1980 (USA - March).

For some sequences, available data was insufficient to indicate possible trends.

SUB-CYCLES IN MULTIPLES OF 9 YEARS

36 Year Sub-Cycles. The 10 principal sequences may be linked in 36 ysc Series 1 & 2 based on multiples of 9 years (9, 18, 36, 54 years) (see Table 1). These sub-cycles operate for comparatively short periods, where as the 56 year sequences may persist for several centuries. Between 1760 and 1940, 34 years appeared in the 36 ysc Series 1 & 2, of which 15 were major financial crisis years according to Kindleberger (1996). (This figure does not include the crises of 1828 and 1864 that occurred early - ie: pre March.) Using a chi-squared test, the probability was assessed at p < 10^-4 and thus very unlikely to occur by chance. For the period 1940 to 1990, a number of currency crises were included in Kindleberger’s listing of crisis years from 1958 onwards, none of which appeared within the 36 year sub-cycles. Even including these currency speculations, 17 of Kindleberger’s 42 crisis years (1760-1987) fell in the 36 year sub-cycles, which gave p < .001 and was thus still significant.

Sequence 52 was the most important 56 year sequence in seismic history. All events happened in the 9 months ending May 31.

     
Crucially, three sequences (Sqs 34, 43 & 52) in Table 3 experienced numerous record events. 
*    Sq 34 - The biggest northern Californian quake (San Francisco. mag 8.25. April 18, 1906).  
*    Sq 34 – Record New Mexico quakes happened on July 16 and November 15 in 1906 (both at 5.8 mag).   
*    Sq 34 - Equal first rank Arizona quake on January 25, 1906 (Flagstaff. mag 6.2).
*    Sq 43 - Record quake for Nevada (Pleasant Valley. mag 7.7. Oct 3, 1915). 
*    Sq 43 - Record quake for Baja California (Volcano Lake. mag 7.1. Nov 21, 1915).
*    Sq 52 - Record quake for western USA. (Great Cascadia. mag 9.0. Jan 26, 1700).
*    Sq 52 - Record US volcanic eruption (ex Alaska) (Mt St Helens, May 18, 1980).

The 1857 Fort Tejon event (mag 8.25), the record quake for southern California, was the notable exception and fell outside the three 56 year sequences.

Moderate Californian Earthquakes. The US Geological Survey listed 42 Californian moderate earthquakes (=> 6.5 and =< 6.9 mag) for the 1800 to 2004 period. Of this figure, 17 occurred in an 18/56 year pattern (see Table 4), in contrast to the 5.3 that could have been expected by chance. Unexpectedly, the moderate events fell in a completely different section of the 9/56 year grid compared with the major earthquakes in Table 3.

 
It was presumed that the major and moderate Californian earthquakes would occur in the same sector of the 9/56 year grid, but this was not observed. The major quakes also happened preferentially in a 9/56 grid, where as the moderate events gave greatest significance in an 18/56 year grid. The disparity between major and moderate quakes in the 9/56 year cycle remains enigmatic and was unexpected.

CONCLUSIONS