Major Dow Theory Non-Confirmation
Special Announcement: Launching of Dow Theory Chartist Blog
I am pleased to announce the launch of the Dow Theory Chartist Blog, courtesy of Google Blogger. The Blog, which is located at: http://dowtheorychartist.blogspot.com/, will contain weekly updates and analysis based on the themes covered in the newsletter. New postings will be made over the weekend, or as circumstances warrant.
If you desire more frequent, informal commentary you may subscribe to my discussion group hosted by Yahoo! Groups by sending an email request to: firstname.lastname@example.org. Members of the group are notified immediately when a new posting has been made on the Blog and are also provided intraweek updates.
A major development since the last newsletter is the continuing non-confirmation between the DJIA and Dow Transports. The non-confirmation has gone on for the better part of 7 months, which is the longest of any that I am aware of . Every non-confirmation should be considered a warning of a potential trend reversal, but should also be considered within the context of an overall Dow Theory analysis. In this issue, I will analyze the current non-confirmation using those criteria. Please keep in mind the purpose of this endeavor is to establish what I believe is the most probable outcome. Since there are no certainties, it is always prudent to wait for a confirmed signal before taking action.
First, let me briefly review how the present non-confirmation developed. (Please refer to the charts on page 2.)
- Both averages were in strong, confirmed uptrends from March 2003 until January 2004. (The actual Dow theory buy signal was given in early May 2003.)
- In February 2004 a non-confirmation warning was given when the DJIA made a new high for the trend but the Transports diverged and made a lower high (it had peaked the previous month.)
- In March 2004 the Transports followed the lower high with a lower low that was confirmed by the DJIA - and a Dow theory sell signal was given.
- Since the March sell signal, the DJIA has continued making a series of lower highs and lower lows (a downtrend) while the Transports have made a series of higher lows and higher highs (an uptrend.) Thus, as it stands now, the current trends are unconfirmed.
Based on historic evidence, one of the averages will reverse course and confirm the other. The question is: which one? Arguments for both the bullish and bearish case are sound, but of course only one can be right. Let's examine some evidence for each case.
The Bullish Case
The Transports are an early cyclical group while the more diversified DJIA is comprised of relatively later cycle groups . Therefore, it would be expected that the Transports lead the DJIA into the early or accelerating stages of an expansion. It intuitively seems more bullish for the Transports to be outperforming the DJIA than the other way around. Furthermore, it would seem unlikely the DJIA would peak during a period of relative out- performance by the Transports.
The following chart plots the ratio of the Transports / DJIA since 1980 - a period that included 4 major market peaks. The vertical lines mark the corresponding 4 peaks in the ratio itself and the explanations follow below.
- The ratio peaked in August 1985 and the DJIA peaked 2 years later in August 1987.
- The ratio peaked in August 1989 and the DJIA peaked 1 year later in July 1990.
- The ratio peaked in November 1993 and the DJIA peaked 2 months later in January 1994.
- The ratio peaked in October 1997 and the DJIA peaked 9 months later in July 1998. (Note the 2000 peak in the DJIA was never led by a discernable peak in the ratio, which accentuated the extreme non-confirmation of that period.)
The results support my hypothesis. The DJIA never peaked before the ratio at any of the major tops during this period. In fact, in two of the instances the DJIA peaked a year or more after the ratio peaked. Even the 1997-1998 instance had a 9 month lead time. The shortest interval came in 1993-1994, which also happened to be followed by the shallowest correction among the four. Clearly, the sector rotation model was an effective predictor of major tops during this period.
Currently, the ratio is making new highs within an uptrend than began in 2000 (it leads at bottoms too - but that's for another time.) If the trend of past four cyclical bull market peaks holds true, the DJIA should peak months or even years after the ratio peaks. Therefore, the DJIA should reverse course and at least exceed its February 2004 high before peaking.
A final point has to do with the meaning of the non-confirmation to the current trend. Since a strong trend would normally include confirmed movements by both averages, the lack of confirmations for the past 7 months implies a weak trend. The direction of the underlying trend during that time as measured by the largest US and world indices has been unequivocally down. It follows that the downtrend has exhibited underlying weakness based on the non-confirmations. And a weak trend usually turns out to be corrective.
The Bearish Case
Picking up on the final point made above, it is also true that a non-confirmation often occurs at least one lower degree than the underlying trend it is warning about. In other words, a non-confirmation of a short-term high often forewarns of an intermediate-term reversal, and so forth. For instance, the short-term non-confirmation in February 2004 warned of an impending intermediate trend reversal; whereas the intermediate-term non-confirmation in 1999-2000 warned of a primary trend reversal. (For the purpose of this example, I am defining an intermediate-term non-confirmation as one that either includes 2 or more consecutive non-confirmed short-term highs or lows or is an intermediate trend itself.)
Some specific examples may help clarify this point:
For an example of a short-term non-confirmation preceding an intermediate trend reversal, refer back to annotation 2 on the charts from page 2. For an example of a minor non-confirmation preceding a short-term reversal, see my Blog posting from September 24.
As for intermediate-term non-confirmations, there are numerous examples of them from 1999 through 2003. Below are charts detailing the massive one from 1999-2000.
After both averages peaked at new all-time highs in March 1999, the DJIA continued making higher highs while the Transports made lower highs until both broke down in March 2000, hence giving a primary degree sell signal. Upon closer inspection, an intermediate low was made in October 1999 that was confirmed by both. That was the only confirmed reaction high or low during the entire period, as previously mentioned in the footnote on page 1. The major sell signal was given when that low was violated 5 months later. However, the warning was building for over 12 months. This particular example most closely resembles the current non-confirmation, except now it is the Transports that are in an uptrend and the DJIA a downtrend. In addition to both containing intermediate trends that completely diverge, they each developed at mature stages of a primary uptrend. The 1999-2000 one came during the primary uptrend that began in August 1998. The ultimate peak occurred about 18 months later. The current one has developed during a primary uptrend that places its origin in March 2003. The Transports are the one now making new highs, and it has been 18 months from the last primary trough. From this perspective, the current non-confirmation can be considered a major warning. I'll come back to this example later.
There were three other significant non-confirmations between 2000 and 2003. The first two preceded intermediate trend reversals in the direction of the primary trend and the last one led to the primary trend reversal of 2003. The first one occurred between the peaks of August/September 2000 and January 2001, which led to the steep declines of February and March 2001. Note, in this instance, it was the Transports making a higher high and the DJIA a lower high. Apparently, the sector rotation model doesn't always hold true.
The next one was between the peaks of February and May 2001, which preceded the sharp downturn into September. Here the Transports reassumed their traditional leadership role.
The final example is the divergence between the October 2002 and March 2003 lows, which led to the primary uptrend that has been in place since. Here again, as in the early 2001 example, the Transports gave a false signal based on the sector rotation model.
To summarize some key observations taken from the above examples:
- Every non-confirmation in the observed period of 1999 thru 2003 led to a trend reversal of at least the same degree.
- The Transports did not always lead in the direction of the ensuing trend.
- The one example of a complete trend divergence (like the current one) eventually led to a primary trend reversal.
Applying those observations to the current situation, it can be hypothesized that:
- The non-confirmation will lead to a trend reversal.
- It is irrelevant which average is diverging.
- The major divergence in trend implies a larger, primary degree reversal is imminent. Since the primary trend is up, that reversal will be down.
Weighing the Evidence
I truly believe both are very compelling cases. Whichever way it turns out, a precedent will be broken. It reminds me of the old cliché: "when in doubt, stay out." From an investment advisory standpoint, that is what I recommend. But, then I did make an earlier promise to speculate on what I think is the most probable outcome. So here it goes.
As a trader, I always give larger degree trends precedence over smaller ones. For example, I trade a short-term trend in the same direction as the intermediate trend. To do otherwise is usually counterproductive. In the same vein, I give the larger trend the benefit of doubt when a point of indecision is reached at the smaller one. Since my analysis has reached a point of indecision as to the future direction of the intermediate trend, I will look to the larger trend(s) for an answer.
The current intermediate trend is down, except for the persistent non-confirmation by the Transports. The primary trend, as I define it, is up. Keep in mind the primary trend is subordinate to the larger secular trend. And that secular trend was confirmed down when the DJIA and Transports made lower primary degree lows in 2002 as compared to 1998. It may be a relatively weak signal, as discussed in prior writings, but it is nevertheless down until proven otherwise. So it is not as simple as just looking at the direction of the primary trend because the probability of a reversal of that trend must also be considered. For example, the primary trend in 1999 was up but the intermediate non-confirmation was resolved to the downside.
Fortunately there is another method in the Dow theory arsenal to help measure the progress of the primary trend: stages. As I've mentioned in previous newsletters, and expounded on in the April 2004 issue, I use Elliott Wave Theory (EWT) to supplement the technical application of Dow theory stages.
A Dow Theory Stages Perspective
In applying this methodology, I am working under the assumption that the current cyclical (primary) bull market began in March 2003. I understand there are alternative interpretations that place its origin in October 2002. Suffice it to say there are numerous factors, much too involved to get into here, that lead me to prefer the former. My preferred counts are shown below, followed by an explanation.
I am assuming a basic ABC corrective wave for the structure of the primary bull market since I believe it is a secondary reaction of the larger, secular downtrend. Under Dow theory it is recognized that a weak primary trend often ends after the 2nd stage rather than completing the typical 3 stages. Under my methodology, I assume the former when I believe the trend is a secondary reaction (i.e., corrective.) If it becomes evident that an impulsive 5 wave (3 stage) structure is developing instead, then the secular downtrend assumption will need to be modified. It is also likely that the ultimate structure will be more complex than what is shown here, but I prefer a simple count until it becomes invalidated. I believe a basic approach is usually all that is required to capture the essence of the stages theory.
The 1st stage of the bull market, labeled as A, ran from March 2003 until January and February 2004 for the Transports and DJIA, respectively. Wave B for the Transports was an ABC correction that completed in March 2004. Wave B for the Industrials may have completed in August, although that has not yet been confirmed. The trend from the February high to the August low did contain 5 overlapping waves, a common pattern for a B wave. If that is the case, the DJIA is now in the early stages of its C wave.
One of the additional benefits of augmenting Dow theory analysis with EWT is the ability to project approximate price targets. The most common length for a C wave is to be equal to A. Sometimes a weak trend has a C wave that is only 61.8% of A. These are not precise targets, only estimates. If C equals A for both the DJIA and Transports, then the projected targets are 13120 and 3915, respectively - both would be all-time highs. If C equals only 61.8% of A, then the targets are 11846 for the DJIA and 3467 for the Transports - still an all-time high for the DJIA but not the Transports. Either one is in essence implying a test of the all-time highs.
The analysis of the stages of the current primary bull market tips the scales in favor of the bullish resolution to the current intermediate-term non-confirmation. The calculated targets of the next stage imply new all-time highs are likely, which contradicts the secular bear market assumption. However, I am cautiously maintaining that assumption because that is what Dow theory has last signaled. Under Dow theory, a trend is assumed to remain in place until proven otherwise. Thus, I think it highly possible that this primary trend will fall short of the projected targets; i.e., truncate. If new highs are in fact made, then the bear market assumption is invalidated and the trend will probably run the full 3 stages (5 waves.)
Since the Transports have been advancing in their C wave since March while the DJIA may have only begun theirs in August, it is very likely that the Transports will peak first. That would cause a peak in the Transports / DJIA ratio discussed in "The Bullish Case" section above. Once that peak occurs, the top in the DJIA will likely be registered within months. Such a sequence of events would be consistent with the sector rotation model and the past results of the aforementioned ratio study.
Does this mean the points made in the bearish case are invalidated? Not at all. One of the details I mentioned a few times was the "interruption" that occurred during the extended 1999-2000 non-confirmation. The averages were out of gear for awhile, and then appeared to link up, only to subsequently diverge again. I think it is very possible another elongated period of structural non-confirmation with occasional interruptions will ensue. The fact that the Transports are so far ahead of the DJIA adds to the likelihood of this. The non-confirmation of the past 7 months has not been meaningless. On the contrary, I think it, along with what will likely follow, is warning of a commensurately large reversal on the horizon. I believe that will be the case even if new highs are made in the meantime. With that in mind, I continue to recommend caution but also recognize we may be entering a tradable market phase. I do not believe we will be in an investable market phase until valuations reach levels seen at previous secular bottoms, but that's also for another time.
I'll leave you with 2 charts of the DJIA and Transports covering the final phase of the bull market leading into the 2000 top and the trading range that has ensued since. In essence, it looks very likely the top of the ranges will be tested in the not too distant future. However, those tests may only punctuate what has been a long, complex topping process of a very large degree.
 The 1999 non-confirmation that began in March of that year was interrupted by a confirmed lower low in October. The DJIA then went on to make its historic high in early 2000, which was not confirmed by the Transports.
 According to Sam Stovall's popular sector rotation model, Transportation is an early expansion group while industrials and capital goods are middle expansion. While the DJIA components have certainly expanded well beyond the industrial sector, the average composition clearly gravitates around middle expansion.
 There were two others that I did not specify. The first was the upside non-confirmation in September 2000 that led to an intermediate downtrend; and the other the downside non-confirmation in March 2001 that led to the spring 2001 rally. In all there were 6 between 2000 and 2003, including the large one that began in 1999. They all lead to reversals of some degree.