How do massive stock market rallies end?
Now that I am convinced we’ve moved into totally unjustifiable extremes of complacency in risky assets, I am having a look at some historic stock market breaks and how they have unfolded. In that light, the current setup is rather ominous.
Note: All charts are courtesy of Bloomberg – kudos to them for keeping such long data sets on the major US market indices.
Disclaimer: I’m an FX strategist and not an equity strategist. But I also fancy myself as a bit of a chartist and I like to look at historic patterns, particularly when the past might provide a historic parallel for the present situation. In this case, I’m interested in what many historic major equity chart tops look like in a technical sense now that if feels like we’ve entered into a blow-off territory technically. Somewhat to my surprise, I found that many major market tops had remarkably similar traits as the one we have just posted.
What is in this post?
A look at big stock market tops in the S&P500 that resemble the current situation (with the assumption that the current situation risks proving a major top), including the all-time classic 1929, but also 1969, 1973, and 1987. The resemblance is not only technical, but also due to the fact that in all instances, bond yields were rising.
The major market tops I have left out include 2007 and 2000 because they bore little resemblance to the current setup technically – though they provide very similar lessons once the top was in place – i.e., in the nature of the subsequent sell-off. In 2000, almost all focus was on bubble tech stock indices rather than the broader S&P500 (in fact – while tech crashed, does anyone realise there was a massive rally in neglected value stocks for most of 2000 and even 2001?) As well, the 2007 was unlike any of the scenarios discussed here because bond yields were falling steeply once the top was forming.
What is “the setup” or analog?
One thing that all of the episodes shown here have in common: there was a flourish of buying at the top of an already major market advance in the S&P500 index, a flourish that actually sees the index breaking above the upper bound of some kind of ascending formation, whether wedge or channel. And then that advance is quickly reversed.
Also – bond yields were generally rising – whether viciously like in 1987 or more gently as in the current episode (though today’s equivalent is any hint of a shift towards the tighter side in expectations of Fed accommodation in addition to the modest rise in actual yields – thus I’m rushing to get this out ahead of the FOMC meeting later today, which is a key event risk in that regard).
Am I calling a top?
I am not calling a top – I am suggesting that there are scary parallels in the current market setup relative to significant past market tops.
What to do?
The lesson that every single major equity top in the past offers is that one does not need to “pick a top”. The more interesting setup is what follows and confirms the top: first you get the initial sell-off that takes out the highs of the previous major wave of trend – in Elliott Parlance, an A-Wave. The size of the initial A-wave from major tops has varied tremendously over time, but even more interesting than the initial sell-off wave from the top is the “throwback” or B-wave that sees the bulls making another go for it or nervous bears covering their initial success and provides the best trading opportunities in terms of risk/reward.
After the initial sell-off, almost regardless of its magnitude, in every single case there has been a significant throwback B-wave, and that wave is almost always very close to the golden 0.618 Fibo (61.8 percent) of the A-wave’s magnitude.
To take an example – not shown below, but perhaps the most dramatic, even in 2000 after the blowout initial tech stock decline in the spring of 2000, when the Nasdaq 100 Index lost a mind-boggling 1,919 of its total 4,816 points and traded at 2,897 at its spring-time low, or nearly 40 percent off the highs from March, patient bears then saw a massive throwback B-wave rally into the summer that took it back to 4,089, six points above a perfect 4.083 retracement. That was the high until an attempt through that level in September failed and then the market went on to lose over 70 percent of its value by about 12 months later.
1929 and 1987 were relatively concentrated scenarios, on the other hand, I suspect due to the involvement of huge amounts of leverage in both cases – 10 percent margin in 1929’s case and poorly understood new S&P500 futures in the latter case. IN both instances, there was a steep run higher with very little correction and then a blow-off top. Immediately after the blow-off, an A-wave of some size removed the last spring of the blow-off, a perfect B-wave retracement formed of almost exactly 61.8 percent, and then a brutal and violent C-wave. Again, the two keys from a trading perspective are the top of the B-wave at golden Fibo’s and then the break of the A-wave low (the latter is shown with a flat yellow line in the charts below.)
So in the future, I will refer back to this post in the event an A-wave correction forms from here – see chart 1 below for what makes an A-wave in the present case.
Chart 1: The current S&P500 setup
We see two interesting things about the current setup – first that the rally has become so aggressive despite this being a multi-year bull market that it has broken the ascending wedge-like formation that has characterised the market for some time. And bond yields bottomed last summer and started coming higher again from November until recently, where we also have talk of a potentially tighter Fed.
From here, for the idea that we have a top on our hands to gain traction, the focus needs to be on whether we get an A-wave that fully retraces the blow-off top – something like a move to 1,500, though anything through 1,525 begins to break up the current bullish trend. From that A-wave low, whether it is 1,500 or 1,485 (Feb low) or lower, we’ll then look for a 61.8 percent throwback as the critical level for bearish counter-attack. And beyond that, we await the break of wherever the A-wave bottoms out – as shown with the yellow lines in all of these charts except for this one as there is no A-wave yet. The break of the A-wave low touches off a much dreaded C-wave. It’s the C-wave where the bulk of the percentage downside occurs. Stay tuned.
Chart 2: 1987
We can see that there was a similar sprint higher in the late summer of 1987 in an already very well established bull trend. (Yes – the market did a similar sprint higher in the spring of 1986 – but remember that the mid-1980’s equity market rally was off the inflation-adjusted 28-year low in 1982 and bond yields were collapsing in a positive growth environment in early 1986 – in 1987, by contrast, bond yields were ripping higher in March – September ahead of the crash.)
In the 1987 market – the upper bound of an ascending wedge (even using a logarithmic axis, it should be noted) was also broken in a blow-off, and things quickly destabilised from there – at the third week after the break – about where we are in 2013 in terms of blowing through an upper bound of a major formation. Then we had three weeks of selling that took out the blow-off, three weeks of throwback rallying attempts to a near-perfect 0.618 Fibo, and then an extraordinary three weeks of C-wave selling that was aggravated by portfolio insurance futures selling. Talk about a compressed bear market.
Chart 3: 1973
A less compelling setup, perhaps, but interesting to note the sharp rally again through the semi-wedge-like formation and to note that bond yields were generally rising from late 1971 lows and continued to rise through August of 1973 ahead of this top. The throwback B-wave fell 1.4 points short of a perfect 61.8% retracement before a brutal sell-off took the index into a C-wave and eventually the 1974 lows.
Chart 4: 1968
The 1968 top was interesting for its ending in a three-week intense buying spree (see 1987 and possibly today) that took it beyond the salient bounds of the upside of the technical formation as well... Oh, and the equity market was topping just as bond yields were spiking higher from summer lows and had moved to a multi-decade high at the time of the market peak…
Chart 5: 1929
Not a real parallel technically, but once again note the very steep three-week sprint higher (!) followed by three weeks of doji-like compressed-range uncertainty and then a symmetric three-week full sharp retracement of that last sprint higher. Armageddon quickly followed the formation of the A-wave after a mere one week as widespread margin use (only 10 percent was required at the time) wiped out the market within a few weeks once the A-wave low was violated. I couldn’t find good data on bond yields in the 20s, but low granularity data I did find seems to suggest they were rising (any readers care to link me to some good 80-year old bond data series?), though I suspect this was a classic case of mania that simply failed as the market simply couldn’t find net new buyers and my understanding was that companies saw activity drying up very sharply in the summer before the October crash.
Chart 6: Bond yields through 200 years – from Societe Generale
What a perspective – this appears to be for 10-year rates. Again, it’s important to point out that bond yields were moving higher in every single scenario outlined above – they were also rising in 2000, by the way. John Hussman of Hussman funds reminds us of the challenge posed by higher bond yields to the current rally almost weekly as he is also worried about rising bond yields. Two ways to look at this chart – bond yields must rise eventually, or real bond yields can continue to go more negative (if this were a chart of real rates – the recent lows would be even more exaggerated), so that bond holders see more and more of their savings confiscated through negative real rates. And how is either scenario good for equities?