The T-Theory Confidence Indicator just made its 2nd lower low after making its 2nd lower high. Can you say downtrend in risk-taking? Remember there are negative multi-year and multi-month divergences that don't occur often and preceded 20%+ SPX declines in a pattern...year 2000, 2004, 2007, 2011 and ?2014?. (Edit: 2004 was ~10% not 20% but price went sideways for 17 months.)
My consumer discretionary spending indicator was looking for a significant low in mid-August. Typically, the downtrends span 5+ calendar weeks and 8-10%+ in points. However, a couple have been smaller like the most recent one in April 2014 at 4-5% (10%+ in other indexes). We have a similar situation now in that some indexes like RUT and Europe have suffered greatly for 4-5+ weeks while Dow is just starting calendar week 4 and SPX calendar week 3 approximately 4% from their respective tops. Upon further review, technically, this week is the 27th week since the spending signal and that is the most common number of weeks to a bottom, but 26 and 28 have occurred. 20 weeks also occurred a few times during the last bear market, but I was projecting the longer 27-28 weeks in this bull market case. So, all that suggests it is slightly possible that 1916 was a significant bottom but more likely that we go a little lower this week or next.
My study of all the 3-4%+ downtrends in the last few years suggests that most damage is done in the first 3-5 days after a top (or often after a failure just below the top on a retest days later) with day 5 or 6 most commonly ending the short-term downtrend sometimes leading to another rally to new highs and sometimes leading to a slightly lower low a few days or so later. Last Friday when SPX hit 1916, it was completing trading day 6 after the 1991 top or day 3 after the 1985 top...take your pick but right on the recent historical average. And, by the way, the point range for those drops were typically 50-80pts and this one was 69-75pts depending on whether you use 1991 or 1985. Once again, that fits the pattern. And, I should also mention that it seems more common than not to break some recent-month low but not expected previous tops/support. In this case, that means SPX would be likely to break the June 1925 low but not the 1897/1900 April/May highs that most people seem ready to buy. And, one more thing, the bounce that occurred before the final low in those samples (assuming their was a lower low made) was almost always 20-30pts and not much if any bigger. So, combining that with the spending signal above, I tend to think SPX will probably bounce to 1936-1950 and then make a lower low this week or early next in the 1901-1915 range probably bounceing near OPEX max pain. Anything lower than that would probably be extremely short-lived but very worrisome if not short-lived (such as hanging around 1900 for a day or so).
If that scenario plays out, Tony Caldaro's call for 1 more high at 1991-2019 or so could play out. But, my discretionary spending data is not signaling for a 10%+ rally (absence of a signal does not necessarily mean it won't happen) and I've outlined 5-10 serious intermediate-to-long-term signals that suggest a 10-20% correction is likely any month now. So, caution is advised. Another thing that makes another test of the highs likely is that most bull market tops don't just suddenly end after a near-parabolic rise without a retest, the previous instances of Fed QE closure supported rallies all the way up until a month or two before and there are a lot of articles about 20% corrections out there with some world and US indices needing relief rallies after 4-6 weeks of downtrending and bouts of bad world news. Still, being bullish could be like playing with fire in the coming weeks so don't overstay your welcome.
Ultimately within the next 2-3 months, I would not be surprised by a breakage of the 200dSMA at 1860-1870 followed by a test of previous tops/bottoms at 1730-1740. All bull markets I could find back to year 1900 that exceeded their previous bull market high by 10%+ all failed about 1 year later and then retraced at least 50% of that breakout and usually 60-90% (sometimes more). 1576-->1991=415pts*50%=208pts and 1991-208=1783 minimum. And 2-3 months would be typical but 2-3 weeks is also possible with so many negative divergences building after a long bull run, bad seasonality coming, no 10% corrections in a long time, lack of volume, some ending triangles forming, round number 2000 psychology, Fed easing, world hot spots, economic sluggishness with few corporate/Fed bullets left etc etc.
In fact, in defense of a possible 2-3 week cascade downward, what I'm about to say might appear to make me wishy-washy, but my discretionary spending data signaled the mid-August bottom by .01 points (94.51 versus 94.50) and then spending dropped further to 93.90 the next month and then slightly lower still for a few months. I don't think my tools are precise enough for me to feel 100% comfortable with a signal triggered by .01. I believe in the signals but it's entirely possible that the real significant low might not come until the 2nd or 3rd week of September rather than the 2nd or 3rd week of August. If so, all of my analysis above could still be right on track but the retest I am expecting (probably after a lower low this week) may just be a 50-80% retracement rather than a near test of the highs. Essentially, even if I adjust the spending bottom to mid-September, we're only talking a difference of 1-3% between the 2 scenarios (SPX 1991-2020 for a new high versus 1950-1990 for a failed rally). I don't know which scenario is more likely but I'll be prepared to abandon my scenarios if I see SPX linger near 1900 or 2020 too long. Good fortune.
Thought provoking stuff S2 and well researched. Thanks for sharing.
ReplyDelete