Tabell’s Market Letter – October 11, 1991

Tabell’s Market Letter – October 11, 1991

Tabell's Market Letter - October 11, 1991
View Text Version (OCR)

TABELL'S MARKET LETTER 5 VAUGHN DRIVE, CN 5209, PRINCETON. NEW JERSEY 08543-5209 MEMBER NATIONAL ASSOCIATION OF SECURITIES DEALERS. INC (609) 987 -2300 October 11, 1991 One of the reasons for the market',s continued trendlessness over the past couple of weeks may be, it has been suggested by ' e analysts, G.qncem CR- the part .of moey agers 9yerthe.possibiHty of a seriou decline during the month ofOctober. Those- managers, it is reasoned, -recan, from legend at least, 'October, 1929, and they certainly remember, from bitter experienCe, Meltdown Monday', October 19, 1987. Even more recent, of course, is the UAL fiasco of Friday, October 13, 1989, which produced a 190-point drop in the Dow and no fewer than 1600 declining issues. While this may constitute powerful folklore, it may well be a misleading interpretstion of the historical record. A study of the Dow's changes for October, since it was first computed in 1897, fails to show any particular downward bias. The diastrous October of 1987 is not even the worst month in market histol)'. The Dow fell 23.2 in that month, but the record for the worst month since 1897 belongs to September, 1931 when the average fell 30.7. Slightly larger drops were also posted in March 1938 and April 1932. In addition, the 20.3 fall of October 1929 was slightly exceeded by the 21.7 decline in May 1940 in response to the fall of France. A look at the overall record for the 93 Octobers since 1897 actually shows more advancing cases than declining ones, 50 rises versus 46 falls. The average percentsge change for the month is, indeed, slightly negative, -I.12. However, three other months also negative average percentsge changes, all of them greater than the October figure. Two-month perioda ending in October do show the worst seasonal pattern for all such perioda, but this is largely due to the effect of September, of which more later. There is, perhaps, some cause for apprehension in the fact that large one-day drops seem to be concentrated in October, or, at least in the autumn in general. October 19, 1987 was, of course, the worst single day in stock-market histol)', and October 28 and 29, 1929 hold second and third place. The DnA has declined 5 or more on 52 separate occasions since 1926, and 15 of these declines, a rather large proportion, have token place in October. Nine such drops have occurred in September and seven in November. Some concern, therefore, about the general fall season may perhaps be appropriate. What is, in fact, especially interesting about the month of October is the fact that it is sandwiched between two months which do, in fact, possess especially significant and interesting seasonal patterns. As we have often noted in this letter, there does indeed exist a month with a particular downward bias, but that month is not OcU;ber, but September. The overall market, as we all know, possesses a slight but identifiable upward bias. That bias becomes more and more evident when one examines perioda of as long as a month. Of — —1131 market months-since-1897, 643 have-seen-the-Dow-move-upward,-and-only-488-have-seen-it-decline.-There-is,–therefore,a-51-T— chance that any given market month will be an upward one. The 94 Septembers on record through last month, though, have produced only 36 advances and 58 declines. Taking all of the 1131 montha, the average change for a given market month is plus 0.56 . For September, the average change has been minus 1.33 . Application of standard test of ststistical significance are applied to this phenomenon, one finds that the tendency in the direction of a declining September is the most significant, slightly edging out the oft- remarked phenomenon of the year-lnd rally. The seasonal pattern which may be of the greatest current practical use, though, is the one which involves November, since that month, in fact, lies ahead of us. By itself, November appears to be quite an ordinal)' month, having produced 56 advances and 37 declines m the DnA with an average rise for the month of 0.74. What is of interest, however, is the fact that November is the only month of the year which seems to possess some predictive value. We have no idea why this should be, but it is a manifeststion we discovered a few years ago, and we remain intrigued by it. We noted above that some 57 of all one-month periods were upward ones. We can of course extend this measurement to longer perioda of time, and the upward expectstion naturally increases. It is 60.9 for six-months and 62.3 for all twelve-month perioda. Let us contrast this with the record for all perioda following an upward November. For all period lengths from one month to one year, such periods have shown a greater expectstion of being upward than would be expected from looking at all periods of similar length. This difference is most striking for intervals of five to ten months in length and is most significant for eight-month periods. Taking eight months as an example, we find that, since 1897, 696 produced rising markets and 433 produced falling ones, thus producing an upward expectstion of 61.6. When we look at eight-month perioda following an upward November, however, we discover that 41 such perioda were upward ones, and only 15 were downward. The percentsge of upward periods, then, was 73.2, significantly higher than the expected 61. 6 figure. Similar results are seen when one looks at the average percentsge change for all eight-month periods following an up November. The norma! expectation for any such period would be a change of 5.08. However the average percentsge change for the 41 eight-month perioda after an upward November is almost twice as great, 9.05. November turned out to be a-good predictor in the most recent case, that of 1990. -It was an upward month with the-Dow posting a loo-point rally from 2442 to 2559. The eight months following this instance of an upward November ended, of course, on July 31st, 1991. On that day the Dow closed at 3024.82, 18.7 above its November close, an advance of approximately three times the norma! expectstion. As far as patterns for the fall season are concerned, then, there does exist some justification for looking for a sharp one-day decline this month, a decline which, we suppose, might manifest itself if a downside breakout from the eight-month-old trading range tokes place. Let us not forget, though, to look at the market's performance next month for a possible clue as to the sort of market climate we might expect in the first half of 1992. Dow Jones Industrials 2973.84 AN1HONY W. TABELL DELAAELD,HARVEY,TABELL S & P 500 380.00 Cumulative Index (10/10/91) 6297.40 No statement or expression 01 opInion or any other matter herein contained IS, or IS to be deemed to be, directly or indirectly, an offer or the soliCitation of an oHer to buy or sell any security referred to or mentioned The matter IS presented merely lor the convenience of the subscriber While we beheve the sources of our Information to be reliable, we In no way represent or guarantee the accuracy thereof nor of the statements made herein Any act!Qn to be taken by the subSCriber should be based on hiS own InvesttgatlOn and informatIOn Delafield, Harvey, Tabelllnc, as a corporatron and Its ofhcers or employees, may now have, or may later take posItions or trades In respect to any secUrities mentioned In thiS or any future Issue, and such poSitron may be different from any views now or hereafter expressed In thiS or any other Issue Delafield, Harvey, Tabelllnc, which IS registered wrth the SEC as an Investment adVisor, may give adVice to rts Investment adVISory and other customers Independently of any statements made In thiS or In any other Issue Further Information on any security mentioned herein IS available on request

Download PDF