Prepared by Peter Eliades' Stockmarket Cycles
The Cycles
This week is the 1,176th week since the December 9, 1974 major low in the Dow, a date which arguably began the great bull market that has continued into 1997. It may well be a coincidence but it attracted our attention as we wrote it down for this newsletter because the cycle we have been tracking as a possible resolution of the top of this great bull is 1,176 calendar days long.
In fact, as this section of the newsletter is being written on Thursday evening, June 26, the all-time high print of 7834.06 on the Dow Industrials has occurred on Friday June 20 at 2:52 p.m. EDT, just one hour and eight minutes before the beginning of our exact cycle date of June 23. It is, of course, far too early to attach any significance to that high on June 20, but if we get past the middle of July with no higher highs on the Dow, it will become far more significant.
There are many technical arguments that say the high of June 20 on the Dow will probably not be a final high. In the next section we will present a fascinating new study that argues it could well be a final high. In fact, because there is so much to write about in the "Technical Indicators" section, we are going to proceed directly to it.
Technical Indicators
Around two weeks ago, we made a remarkable discovery. We often experiment with conventional technical indicators by trying to look at them in a somewhat unconventional way. Welles Wilder's RSI indicator is one of a technician's most popular tools. It is, in effect, a normalized momentum oscillator. The two most common periods used are 9 periods and 14 periods. Often technicians have experimented with slightly longer or shorter periods. We decided to look at extremely long periods to determine if there were any patterns of significance. We used a daily chart and began by looking at RSI periods of 253 (the equivalent of one year on a daily chart) and longer. We won't go into the details of why we settled on a period of 316 for the RSI, but when we looked at a daily chart of the DJIA over the past 30 years in association with an RSI 316, we were struck by a remarkable observation. Since 1966 which is where our initial chart began, there had only been five time periods when RSI 316 reached a reading of 60 or higher. Look at the chart. The points labeled 1,3,5 on the left side of the chart and those labeled 1 and 3 on the right side of the chart accompany the only time periods in the past 40 plus years where upside price momentum was high enough to carry RSI 316 above 60. The more we studied the chart, the more we marveled at the similarity between points one through five in 1986-87 and points one to four in 1996-97. Let us give you the dates of both left hand points one to five and right hand points one to four.
L1 3-20-86 R1 2-13-96 L2 9-29-86 R2 7-23-96 L3 3-26-87 R3 1-21-97 L4 5-20-87 R4 4-11-97 L5 8-25-87
In all cases where there was more than one 60+ reading in a time period, we used the day of the extreme reading. At the intervening low points two and four, the day of the lowest reading in the time period was used. Left hand point five (L5) uses the actual date of the final price high, rather than the RSI high of the time period which occurred on August 17, 1987, just over one week earlier.
Because the left hand pattern so closely resembled the right hand pattern that was forming, we decided to measure the distance between points L1-L4 and point L5. The results were:
L1-L5.................................362 market days L2-L5.................................229 market days L3-L5.................................105 market days L4-L5..................................67 market days
The final part of the exercise was to add the corresponding left side distances between each of the first four points and point five to the corresponding right side points. If the pattern continued consistently, then we would be able to determine a potential time window for point R5. Here are the results.
R1 + 362 market days.......................= 7-21-97 R2 + 229...................................= 6-18-97 R3 + 105...................................= 6-20-97 R4 + 67....................................= 7-17-97
The comparable spacings within the two patterns is so consistent that the final time window for point five is a difference of only 22 market days.
If this were the only evidence that the market is currently in a rare configuration, it would be impressive but statistically not very significant. When we finished the above research over the past week or two, we became curious as to other occasions when RSI 316 may have reached the 60+ level. We had daily Dow prices in our computer back to 1915, so we ran the calculations back to that time period. Since 1915, guess when RSI 316 first hit 60+. Did you guess 1928-29? Good for you! It occurred on November 20, 1928. On November 28, it peaked out at 61.14, one of the highest readings of the 20th Century (the highest was 61.7). Just over nine months later, of course, the stock market topped prior to the greatest decline in its history, from 1929 to 1932. Before that decline, RSI 316 reached 60+ again for one day on January 2, 1929 and for several more days in late January and early February. Even though the 1928-29 pattern was quite different than the 1986-87 pattern, there were obvious similarities. The most important similarity is the fact that both the 1986-87 pattern and the 1928- 29 pattern led to the two greatest crashes in our history. The other similarity is that in both the 1928-29 case and the 1986-87 case, the first momentum peak above 60 occurred well before the final price peak. In the 1929 pattern, RSI 316 never made it back to 60+ at the final price top. In 1987, RSI 316 did make it back above 60 right around the final price high.
So is that it? Are we saying that the only two prior periods this century when RSI 316 reached 60+ were 1928-29 before the crash and 1986-87 before the crash? Unfortunately, they never make it that easy. The only other time period this century when RSI 316 reached 60+ was in 1954-55. The first time was November 23, 1954. The price high in that time zone occurred on April 9, 1956, some 16½ months after the initial 60+ reading. That April 9, 1956 high led to a moderate bear market decline of just over 20% and the high was not regained and surpassed until almost 2½ years later. That is the complete history of 60+ readings on RSI 316 for the century. So how do we delineate the 1996-97 readings of 60+ from 1928-29, 1954-55, and 1986-87? The odds are not good. Two of the previous three resolutions led to historic crashes. What are the distinguishing factors? Unfortunately there were distinguishing underlying valuation factors that put the current situation much more in the 1928-29 and 1986-87 category than the 1954-55 category. In late 1954 when RSI 316 first reached 60+ after the 1929 period, the market was still in a neutral valuation area with a P-E ratio around 13 and a dividend yield still above 4%. Both the 1929 and 1987 markets set overvaluation records on both a price to dividend and price to book value basis.
It should be clear, however, that whether the current string of 60+ readings which began in February 1996 are resolved as they were in 1928-29 and 1986-87 or 1954-55, the price top is quickly becoming overdue. If we measure from the first reading of 60+ in each era to the final price top, we come up with:
1928-29..............................287 calendar days 1954-55..............................503 calendar days 1986-87..............................523 calendar days 1996-97......................................?
The longest prior period from first 60+ reading to final price high would lead to a top by July 18 at the latest.
If these potential dates of resolution are starting to sound familiar, it is because they are appearing in the same time window as the potential top anticipated by the Coppock Curve "Killer Wave." In our May 16 newsletter, we used the prior Coppock Curve Killer Wave signals to attempt to get some hints as to when the final top might occur. Without repeating the details of the analysis, the average resolution of a market peak measured from the top of the Coppock Curve preceding the peak was 390 calendar days. The longest resolution took 469 calendar days from the prior Coppock Curve peak. That would imply a final market peak between June 25 and September l2, 1997 depending on whether the April 1996 Coppock peak (333.98) or the May 1996 peak (333.21) is used.
Finally, remember the 1,176 calendar day cycle. It was due June 23 and based on the prior longest resolution (1185 calendar days between 4-27-81 and 7- 25-84) could resolve as late as July 2.
Here are three completely unrelated technical tools each overlapping each other in terms of a time window for a final price high. The 1176 calendar-day cycle suggests that if the top occurs within a time window that sees all three tools end up to be correct, then the final top will occur before July 2, 1997. All three tools also agree the top could have been seen already.
In our last issue we discussed the McClellan Oscillator and Summation Index and the potential significance of a 3,000 plus reading on the Summation Index. In that issue, we noted that a reading above 1,000 on our adjusted Summation Index corresponded well with readings above 3,000 on the normal McClellan Summation. After writing that newsletter, we spent a day going over all the data since 1940 and revised our Summation Index taking it all the way back to 1940. It turned out that on our new adjusted Summation Index, readings above 1,140 were required in order to equate with normal Summation Index readings above 3,000. The correlation is virtually perfect between adjusted Summation readings above 1,140 and normal Summation readings above 3,000. It proves our point made in the last letter that a move above 3,000 in this time period would not have significance unless the adjusted McClellan also moved above the equivalent adjusted reading. As we noted, rather than the 1,000+ required with our original adjusted Summation discussed in the last newsletter, the number required now is 1,140+. On June 20, our adjusted McClellan reached a reading of 1100.4, and has since turned down, but not significantly.
Our contention is that, although there is little argument that recent McClellan Summation readings were high, they have not reached the area of prior high momentum readings that virtually guaranteed a higher market over the following year. Here is one example. On January 4, 1977, the adjusted Summation Index closed at 1138.8 and the normal Summation Index closed at 2750.1. On an adjusted basis, it was higher than our reading on June 20, one week ago. Over the next 14 months, the Dow was down over 25%. The relatively high McClellan reading did not help the Dow over the next year.
Monday, June 30 marks the end of another month. Unless the Dow is down over 400 points on Monday, it will be an up month. Only once this century has the Dow Industrials been up more than 68 months out of the past 100 months. That occurred in December 1961. Just short of six months later, the Dow was down over 28% from the December 1961 close. There is also another first that will almost certainly occur on Monday, June 30. The century old records of 30 months up out of 40 and 69 months up out of 100 will be reached in the same month. It is a record extreme based on these two extreme readings being reached in the same month.
Let us put these current readings in perspective for you. The only times this century the market has been up more than 67 out of 100 months prior to the past year were October 1961 through February 1962 (those were all readings of 68 per 100 with the exception of December 1961, as noted above, which reached 69 per 100), and January, February and April 1966.
Similarly the only times this century the market has been up more than 29 out of 40 months prior to the past year were February 1937 and April through August 1961.
Now for the first time this century we will see both records reached in the same month, come Monday June 30, 1997. In association with the other evidence presented in this technical section, it should make it clear the market risk is enormous.
Here is the finishing touch for the technical section. Our colleague, Bob Nurock writes a private advisory via e-mail. In his June 16 report, he wrote, "...sentiment has begun to show significant deterioration. Our Premium Ratio on Index Options reached its lowest and most bearish level since OEX options began trading in 1985. Its previous low reading was the week of August 14, 1987, leading to the lowest three-week average of the indicator model the following week, after which the market topped out." Those statistics fit in perfectly with the technical material just presented and add to the bearish outlook for the market.
On June 20, the Dow reached its 28th record high for the calendar year and its 141st record high for the past three calendar years. The 141 is a record for a three-year period, approached only by the 138 seen in 1927, 1928, and 1929.
Market Projections
On our daily telephone update of June 10, we said, "The reason we believe it is possible to reach the upper projections over the next two weeks is that the OTC Composite has a nominal 20-week projection calling for a move to between 1460 and 1493." On that day, the Index had closed at 1401.60. On June 25, the OTC Composite reached a high of 1458.85, meeting the projections within .08%. All outstanding upside projections have now been met, although we cannot rule out higher highs generated by what we call "loop projections."
June 27, 1997 Peter Eliades
Stockmarket Cycles
P.O. Box 6873, Santa Rosa, California
Added to the WWW 07-05-97
Last updated on 07-06-97
Hosted by:
One Crossroads Place
610 West Maple Ave, Suite WWW
Independence, MO 64050
(816) 252-4080
sysop@kcmo.com
