In his latest note, Gluskin-Sheff’s David Rosenberg gets a bit reflective, as the market continues to batter away at his bearish standpoint:
I’m not sure if you all had a chance to read Dave Parkinson’s column in
yesterday’s Globe & Mail (page B13 — Hibernation). When he interviewed me,
he asked what it was like to be “alone.” I told him that “alone” was not the
same as being “lonely” … it’s actually a comfortable place for me to be. I told
Mr. Parkinson that the last time anyone asked me that was back in the fall of
2007 — actually, after a day at the Fed when I was viewed as being a nutcase for
calling for a recession seeing as the S&P 500 had just hit a record high.
Indeed, the S&P 500 did hit 1,565 on October 9, 2007, a double-top after rising
to 1,553 on July 19 of that year. It was a failed test, with 20/20 hindsight. That
got me thinking, are double-tops danger signs? Well, have a look at what
happened as the parabolic upmove ended in March 24, 2000 — the S&P 500
rallied to 1,527 and then corrected to back and do a retest of the highs on
September 1 when it stalled out at 1,520.
In the cycle before that one, we saw the same thing occur in 1990 — a surge to
367.4 on June 4 followed by a brief corrective phase and then a retest of that
high on July 16 (at 368.9).
The market did the exact same thing in 1987 too — a frenetic runup to the
336.8 high on August 25, a brief respite that got the bulls refreshed, and then a
sudden move back above 328 on October 5 … and the rest was history.
Go back to the end of the cyclical bull-run in 1980. On November 28, the S&P
500 was sitting at 140.5, it then slipped back to only then retest above 138 on
January 6 … and that was all she wrote.
So what has happened this time? We hit an interim peak at 1,150 back on
January 19, 2010. We had a brief corrective phase and now the S&P 500 has
matched its January 19 high. In other words, we could be at an inflection point,
but keep in mind that there are countless examples of the market blowing
through interim highs as well. We only know in hindsight where the double-tops
So I asked Walter Murphy (www.wminsights.com), one of my favourite technical
strategists who was schooled by the likes of the legendary Bob Farrell, what
clues we should be looking for to assess whether or not this is a classic double-
top or just a pause that refreshes. Here is what he sent back to me:
I usually use a “weight of the evidence” approach. I first want to find
the “internal” peak or trough and then look to find the divergences.
Most people look at breadth; so do I. But breadth oscillators are
more important than the A-D line itself. That said, breadth was way
out of character in 2000, so it wasn’t much help.
In 2007, both the A-D line and oscillators did not confirm October’s
high relative to July’s high. The A-D line made new bear market lows
in 3/09, but the McClellan Summation Index & S&P’s Short Range
Oscillator (which combines both breadth and the DJIA) did not.
Momentum is another indicator to watch. The weekly Coppock
Curve peaked in early 1999 and in late 2006. It bottomed in 12/08.
Sentiment often behaves very much like momentum. The 10-week II
bull/bear bear ratio peaked in 7/99 and 2/2007 and 7/2007; it
bottomed in 12/08.
I also look at the indexes — if only a few big cap indexes are making
a new high or low, that is a sign of a change. If they all do it together
that’s a sign of a trend change.
Finally, I look at the Bullish Percentage Index (a P&F indicator). Like
breadth, it was out of character in 2000, but the fact is that it
peaked well before the market peak. In 2007, it peaked in Feb. It
bottomed in October 2008.
So using that history, I would make the case that the internal peak
prior to 2000 was in 1999; in 2007, it occurred no later than July.
At the recent bear market low, the internal low was October 2008.
Currently, the A-D line is making new highs, but the McClellan and
SRO are not. The former peaked last September and the latter
peaked last July. Both are in uptrends, so there is still a chance that
they can challenge their earlier peaks.
The weekly Coppock Curve peaked last June.
The bull/bear ratio peaked in January, so I tend to consider that a
Virtually all of the popular indexes (except the DJIA and S&P) are
making new highs. The troops are leading the generals.
The BPI peaked in September.
So, I think that, until proven otherwise, the internal peak was
recorded last summer. However, with breadth oscillators in a new
uptrend and with most averages at new highs, the divergences
aren’t full in place yet.
Hope that helps,