AllAboutTrends midday update shows how the S&P is following a classic structure: it dropped below a recent trendline of support, came back to retest from below, and has since fallen off. This is a bearish sign, at least with respect to the rise since mid-March. The bears are out quickly to say "it's the end!" while some more cautious pundits like Neely see a final and more vioent upwards thrust to come.
The market seems to be in its final run to an epic Triple Top back to 2000. Chris Martenson supplies a nice piece of analysis, expecting a 40% drop after we top. His chart shows this in the S&P:
Zoran Gayer, whose analysis updated Wave Theory with Chaos Theory, demonstrated how triple tops are good indicators pf trend changes. The Chaos Theory view, which is based on Thermodynamics, is that markets move in Thrusts and Plateaus, where a Plateau is the market on the edge of chaos, seeking order. A Thrust is order. All non-linear chaotic systems show this behavior, including weather and market economics. The formation in the S&P is a classic Plateau after a really strong thrust up in the '90s. The Plateau does not pre-determine the next Thrust, but Zoran found that after a triple touch of a trendline, the Plateau usually was showing exhaustion and the Thurst coming out of it would be the other way. In 2002-3, for example, we had a strong Thrust down, then a triple bottom in what he called the Iraqi War Triangle, followed by a Thrust up into 2007 that marked one of the largest asset bubbles of all time.
The timing of the triple top looks to be within the next few months. Seasonal patterns usually top around April or May, followed by a correction into September. Some pundits (like Neely) expect a sharp rise towards SP1600 in March, with a top in early April. Whichever it is, expect a bit of a throw-over at the end and then a really sharp "bifurcation" move down to mark the top.
The Plateau since 2000 marks a huge experiment with economic history, where all of us are the guinea pigs: central banks are trying to apply the lessons learned in the Great Depression to reflate the global economy. Problem is, they first created a terrible asset bubble in real estate, and after that corrected, they are now putting the global economy on central bank life support. As more fiscal borrowing is pumped out into the system, nominal GDP levels atrophy as the ever-increasing debt puts pressure on continuing this flow of monetary oxygen into an increasingly-moribund economy. As the system approaches the Rogoff Level of 90% debt-to-GDP, economies began to sag. Europe is already entering recession, as has Japan; and the US is skipping across the 0% GDP level, skating on very thin ice. Worries abound about China as well.
The Triple-Top Plateau therefore reflects markets at the edge of chaos on whether this huge central bank experiment will work.
Looking forward, history says this long secular bear market should last 16-20 years. The last one from 1966-1982 was 16 years; the prior one from 1929-49 was 20 years, and we saw a similar long plateau around Dow 100 from 1901-1921. The Depression of 1874-1896 was even longer, as was the one from 1837-1859. This puts the bottom out to between 2016 and 2020.
By now it is clear Apple has fallen off a cliff after a parabolic rise. History says the fall will at least retrace the parabolic rise. We saw this recently with Crude Oil rising over several years into a parabolical peak at $147 (cash price), then falling in six months to $31 before bottoming. We saw this in 2000 with a bunch of tech stocks, and in particular the leader at the time, Microsoft. Is AAPL, the leader in 2012, replicating MSFT, the leader in 2000? The charts are eerily similar (courtesy Bespoke):
History also says a bull trap will emerge on the way down, a major bounce where many think they have caught the bottom, but when it reverses, they instead get sliced & diced as if grabbing a falling knife. Same bulltrap happened to crude on its rapid drop. The AAPL false bottom may have already occurred around $500 (see chart). After bouncing at that level, it has hit an airpocket and now seems on a bottomless journey down. A scan of the punditry finds some expectation for a bottom at $410, and I expect that level to catch some bids, but a look at the chart (red line) shows no support level until AAPL gets to the start of the parabolic spike up, at around $350.
Apple is a quality company, and its recent miss seem to be more an anomaly of comparing a 14 week quarter a year ago with the 13 week quarter that recently ended than a sudden break in Apple's trajectory. Hence, at some point, a bottom will be found, and a buying opportunity discovered; the bigger question is whether, like Microsoft's lost decade, will Apple will grow quite strongly but never regain the recent stock highs? Maybe there is an app for that ...
Whne a trendline gets breahced, the market typicaly comes back to test it from below. Will it kiss it goodbye, or show the break to be a False Break? Chart courtesy SlopeofHope:
We have been following the AUD as an indicator of coming divergences in US markets, specifically that the correlation between the S&P and the AUD breaks.
Last night the Chinese GDP report caused a spike in the AUD. The report was in line with expectations, which suggests the AUD was a bit oversold.
Technically the AUD is in a triangle, and so is the S&P. We are expecting a downtick shortly, perhaps into the election, to be followed by a sharp rally.
In the S&P it could touch the 1550 area, a triple top back to 2000 and 2007. In the AUD it could breach $1.10.
The Euro is also at an appaent cusp point. Right now the Forex side of EWI is running a free week in case you wish to click through and learn more. Here is a taste of their perspective in this video about the AUD's triangle.
I have been following the Election Year pattern with a sereis of posts here and here. It typically bottoms aroudn October 10 and then rallies, usually signaling that the winner is now pretty well known. It seemed to take that path over a week ago, when Obama surged ahead, but the debate has thrown a monkey wrench in. Here is the chart a week ago:
I have been following the election year pattern and it remains on track, with a 2 - 3 day lag (see chart, courtesy PragCap). There is an interesting twist this election, but first a discussion of the general situation. The prognosis, per PragCap:
Looking ahead, this version of the Presidential Cycle Pattern says that we should expect to see a choppy uptrend continuing toward election day, perhaps with some significant “texture” along the way. The strong correlation up until now suggests that this pattern is working reliably. Once we see how the election turns out, we can then figure out which pattern to follow starting in November.
What is interesting is that this version of the chart has taken out the 2d terms and only focused on first term Presidents, a more true comparison to this election. The generic chart that we showed in the prior post combines first and second terms and the behavior is different in each term. Per PragCap:
As a rule, [new Presidents] all typically spend the first year in office “discovering” that conditions are even worse than we were told during the campaign, and that the “only solution” is some urgent package of tax changes, spending, regulations, etc. ...
When a first term president wins reelection to a second term, then he typically does not spend much time in the first year telling us all about what a lousy job his predecessor did. That lack of a persistent negative message seems to work out to bring slightly better market performance right after a reelection.
But with a new president from a different party, investors tend to get discouraged hearing that things are worse than expected, and so the market’s performance during the first few months with a new party president is on average slightly worse than if an incumbent wins reelection. Over time, though, it tends to average out to be just about the same.
This chart shows the comparison of first term vs second term Presidents:
What is different this time is the pattern has not worked well in the last three elections, perhaps because we hit a secular bull market peak in 1999 and have been in a secular bear through the past three elections. The pattern broke in 2000 and 2008 when we had market crashes; it also didn't hold in 2004.
This time there is a lot of can-kicking going on to push the problems to after the US election. A lot of commentary over the past week out of Eruope thinks the latest ECB moves can buy 3 months, and of course the fiscal cliff doesnt rear until after the lection. The one stinker is that we may hit the debt ceiling limit faster than expected, before the election.
"The market can remain irrational longer than you can remain solvent" - John Maynard Keynes
When AAPL ran up into an interim peak in March, I asked whether AAPL has peaked? I used a chart which showed two potential pivots: at 593-601, which formed the interim top; and 668-677. We have now come to 675 and fallen off a bit. Apple's next big announcement is coming, and often stocks run up on rumor and fall on news. That may be all this is; or, is this the peak?
If you expand this thumbnail of AAPL's incredible run up, courtesy SlopeofHope, you will see a classic parabolic rise pattern. We saw it in 2008 with oil, which shot up to $147 a barrel then fell like a drill bit straight down over the next six months to $32. We saw it with the NASDAQ from late 1998 to early 2000, where it too fell like a dead cat from April to October of that year. And of course in stock after stock and commodity after commodity.
AAPL has now become the highest market cap in history, at least in nominal dollars. In inflation adjusted dollars, Microsoft was worth more in 1999, and IBM was much higher in 1967. Still, a seminal achievement for AAPL, and a warning flag for investors.
I ran a comparison of Google's thrust to its peak vs. Apple in my prior post, supporting the argument of a peak based on analogy. Here is the comparison to Microsoft's thrust up during the dot-com bubble, which shows Apple may have more room to run:
Word of caution: parabolic thrusts end badly, but by their very nature (underlying stampede of the herd) are unpredictable of how high they will run and when they will end. Arguing from a few analogies, like Google and Microsoft, is very thin to base a trade upon. The dynamic for Google several years ago or Microsoft in 1999 were very different than today. This is not an argument for "this time it is different," just a reflection of the über momentum behind parabolic rises, and how the emotion of greed (or for money managers, fear of falling behind) can run for a longer time than a short can stay solvent. Particularly so in this environment, of a Global Scramble for Yield; money managers have to have AAPL to keep up with their brethren, and boy do they enjoy the outsized yields so far.
A little technical analysis may help. Here is the updated chart from the last report, with predicted levels of resistance, now shifted a bit to 671-688, a range we hit already:
This chart from Bespoke made the rounds a few weeks ago, and it is interesting to note that it continues to work as a predictor of this year's election rally. We have continued higher, and now may be in the final blow-off stage into early September.
It is mildly interesting to note that late August/early Sept tops happen in big turns years (1929, 1987 etc.); and also happen in relatively normal years. It is just that in an election year, the summer rally seems to last longer. The correction in Sept usually denotes uncertainty over the election. When it becomes clear who will win, the market rallies again. In 2008, the post-correction rally waited until the end; of course a lot was going on in Sept 2008.
The normal seasonal pattern is a rally commences in mid Oct to early November, and continues into late April or early May, with a mild correctiion in late Jan/early Feb. The market then has a June Swoon and Summer Rally, but essentially meanders sideways into Oct, until the seasonal rally takes off again.
The Election Year pattern shifts the swoon to later, usually after the Conventions, and starts the seasonal rally earlier, unless the election is a cliff-hanger (1980, 2000, 2008).
Of course, this is an average of many different years and circumstances, so don't take it to the bank. Just note that almost every election is said to be critical, that the country is in crisis, that This Time It's Different, and then it isn't.
This chart, courtesy Contrarian Advisor, provides a great view of where we are, and it predicts that the S&P gets to above 1500, if not to 1600, in fairly short order:
The big picture view is that since 2000 we have been in a large corrective triangle that is leading to a triple top in the S&P around the 1550 range, with tops in 2000, 2007 and 2012. The shape of the triangle is a Barrier Triangle, with a barrier around 1500-1600 and an expanding bottom below Sp800. The implication is a drop from 2013 - 2015 down to the lower trend line, meaning to new lows below the March 2009 Sp667 intraday low.
The doom & gloom crowd remains strident in calls for an imminent P3 wave down, despite the inexorable rise of the market. Their position is getting harder to support, if it has even been very supportable for a while. After a big drop as we had in 2008, the bounce usually goes 50-62%, but seldom beyond; if it gets about 70.7%, and especially 78.6%, it almost always goes back to the old highs. We hit the 62% retracement a while ago, and odds then switched to running up to a new high.
An alternative view to the Big Triangle is that the 2000-2009 drop is a classic Flat correction. This makes the wave up either the start of a new bull market - which should exceed the 2007 highs - or a corrective X wave with another down pattern to follow. Since that structure will be horizontal, given how this correction has so far unfolded, it also suggests the X wave will go back to the 1550 range.
X waves are not well understood in wave analysis, since they seem like a "plug" of a failed structure, but they are a very standard feature in markets, especially in extended corrective structures of a sort that don't break as triangles. So far the correction since 2000 has gone 12 years, so a wave analysts should look for a triangle or a series of horizontal corrective patterns connected by X waves.
X waves tend to be simpler than the corrective patterns they connect, and therefore are usually zigzags - sharp rises, as we have seen since 2009. This one seems to be resolving as a large double zigzag, where the first zigzag rose into 2011, and after the big May to Oct correction, we started the second zigzag, which we are still in. Simple fibonacci analysis says it should go to the 1550 range: using the weekly numbers above, the first wave rose 680 points, and the second zigzag should run 61.8% to 100% of the first one. Using 61.8% of 680 points, we get a target of 1544.
Usually stocks run up in an election year - and even more so in the year before - and then correct hard afterwards. Why is not hard to figure: the incumbent tries to get re-elected and uses whatever ways he can to pump the economy, then takes a huge hit in the first year to give his party a shot at another recovery into the next election. This is the Four-Year Cycle, the most predictable stock market cycle because it is driven by actions in the real economy. At a simple level, one could just stay long and pull out before 2013, when a whole bucket of negative events happen, like the repeal of tax breaks that are pumping the economy up.
Technical analysis supports this fundamentals view right now.
Continuing the name-that-tune game of market analogies, The Big Picture gives us the current market overlayed on the generic four-year cycle driven by Presidential re-election policies. It supports the view that we are in a large sideways market since 2000, and should hit a triple top of the 2000 and 2007 peaks, both around Sp1550:
Of the many technical tools and analogies, this one is the easiest to decipher, since it is based on pump-priming for re-election. The Obama budget indeed continues a huge spending splurge with large deficits, a formula for re-election. Maybe the budget is DOA in the Republican House, but don't under-estimate their craven nature and culpability in reckless spending. The Obama budget gives them plenty of room to announce phony "cuts" against a huge baseline increase. Party on!
I have oft commented on the many market analogies proffered, such as to 1929, 1937, and 1987. MarketThoughts compares this market on an inflation-adjusted basis to other long bear markets (secular bears, as distinct from cyclical bears), including 1966-82, 1929-49 and 1901-1920, and shows that we are so far fairly normal, albeit we dropped farther during the 2007-09 Big Drop:
The implication from the averages is that a drop is due, although their analysis concludes that we have more upside before that drop, in a break from the average pattern. Worth reading.
I see lots of pundits trying to pick the top. Fool's errand. Better to watch for a bifurcation off the top before betting on a change of trend; otherwise ride it up. The recent bifurcation carried down into a trading plateau, but not below, meaning no break of trend. We would need a sharp drop below 1275 that stays below to confirm a change.
The market is in a channel and is touching the top trendline. The recent trading range was right on the mid-channel between the upper and lower trendlines. A bounce off the midline confirms a continuation of trend. You can see this in the following chart comparing the Dow with the Trannies (courtesy EWTrends). The Dow remains in the upper half of the channel while the Trannies have broken down, a divergence in Dow Theory, which would require a reversal back up in the Trannies to support a continued bull market:
Typically a trend ends with a false break, which in this case would be a thrust above that quickly (relative to the length of the trend) falls back below. Right now the trendline is above where we are, both on a closing and intraday basis, at 1328 Wednesday and climbing 1.5 pts/day. The STU is doing a good job of watching this closely and positing potential turn levels.
I see a lot of the pundit-sphere pointing to a 78% retrace level (of the drop from Oct 2007 to Mar 2009) as the next key level. The exact 78.6% level depends on counting from intraday highs/lows or daily, or (for more precision) bifurcation points, but a reasonable level to use is 1375. Before we get there, we have another typical turn level ahead, 1352, where the recent wave C from the July 2010 low to now gets around 61.8% of the prior wave A from March 2009 to April 2010. A typical relationship in a sharp or "zigzag" correction is for waves A and C either to be equal or to relate by 61.8%.
If this is a wave 2, the 78% level is really the point of no return. Theoretically a wave 2 can go 99%, but the odds say once past 78% then almost always this was not a wave 2. I can make a stronger case that once a wave 2 goes much past 62%, it is unlikely to be a wave 2, and statistically once past the repetition level of 70.7%, it rarely is a wave 2. The prime exception is a wave 2 that retraces a really sharp drop of the sort we had back in 2008 (call it a "crash"). The role or psychology of a wave 2 is to make the vast majority of market participants believe the sharp drop was an anomaly and the market is back to the prior trend. The sharper the fall, the further the comeback.
Psychology may also come into play on the two-year anniversary of the Mar 6 low. Interesting is that we bottomed intraday at 667, and are close to the double, at 1334. Whether enough astrology or Gann traders are still out there, this might cause a stutter.
Finally a lot of pundits have been comparing this market to April 2010, where the market went into a three-week trading range with a couple of sharp drops that were recovered before we bifurcated down. The high was in the middle of this range, and the eventual bifurcation down that led to the Flash Crash occurred about a week after the top. The implication is we have begin a three week or so plateau punctuated by a final high (1330? 1334?) that will be followed by another sharp and deep drop. One of the best discussions of this comes from AllAboutTrends:
While a slight pop to say 1330 may mark the end, or a stutter around 1334, more likely we bump along the upper trendline up to 1350-1375 before a serious pause occurs. Given the trend slope of 1.5 pts a day, this points to a turn around 20 or more trading days out, or around the vernal equinox. We might test the midline again before then, but don't get too excited.
It's Groundhog Day and the USD is once again giving a wake-up call. AUD popped above parity and the USD is heading into a key support level. Phoenix Capital Research raises the Alert Level with this chart, which shows the mighty Buck once again hitting a key support level:
If it breaks below, and does not quickly come back above (a false break), it raises the awful spectre of a Dollar collapse. The report also sees a possible head & shoulders pattern which suggests a 50% further drop:
Now, H&S patterns are oft seen and seldom realized, because the underlying requirements that need to be met are more than appear in the pattern:
The EWT recently laid out the theory behind the head & shoulders pattern (EWI makes it available for free, in case you would like to peruse an EWT on point to today). Edwards & Magee's book says that the first shoulder should rise on heavy volume, but the second shoulder should have "decidedly less volume" than either the head or the first shoulder. ... The EWT notes as well that the whole formation should come after an "extensive trend". ... Many purported head & shoulders patterns fail these two tests. The weak volume indicates exhaustion of trend and a coming rotation to the other direction.
The current weakness is driven by oil dynamics, specifically the risk of spreading turmoil in the Arab world. Oil and the USD are pretty well locked into an inverse relationship (Oil up. Dollar down), although it often is not clear which is the driver. It could be said that after the US went off gold in 1971, it backed the Dollar with an Oil Standard by making a deal with the Saudis in 1973: you trade your oil in Dollars, and we will defend your oil fields. The US gets a great seiniorage benefit from this, as the rest of the word has to maintain high Dollar reserves just to pay for oil.
Tonight's STU notes that at the intraday low today (DX76.88) the Dollar Index has retraced 78.6% of its prior rise from Nov4 (75.63 on election day!) to Nov30 (81.44), a level that almost always is the extreme for a wave 2 counter-trend move. The percent of Dollar Bulls also fell to 7%, in the range seen at previous bottom (3%-7%) since the April top at 88.71. They expect a rally "now".
The trendline in the first chart stands at 76.08, below this 78% level, suggesting that we may break the 78% level and fall to test the trendline. Wave theory at that point no longer supports a reversal. Watch this development closely.
It is commonly asserted that the Fed sets interest rates. Instead, history shows the Fed follows the market - and gets the economy into trouble when it diverges too long from what the market is saying. Tony Caldaro has an excellent piece on the Fed and rates in the 'Oughts. He shows how the Fed sets the Fed Funds rate (red line) based on movement in the 1-year Treasury Bill rate (black line):
In 2003 the Fed held rates too low compared with the 1-yr, and the Greenspan Bubble of excessively easy credit was on. In 2006 the Fed held rates too high, and the Housing Bubble was burst. The signal for being too low or too high appears to be 50 bp - when the 1-yr is 50bp or more out of alignment, the Fed Funds rate needs to be changed.
Given this correlation, wouldn't it be more consistent policy if the Fed Funds rate were simply set by the market, and not the Fed? Or the Fed behaved by rule and not discretion?
Right now the Fed Funds rate and the 1-yr are very close. No short term rate increase is on the horizon. The 1-yr would need to crest 0.75% to spark a possible change. With a commodities-based inflation emerging due to QE2, this may happen sooner than investors, and the Fed, think.
Great stuff by Tony!
There is a general trader view to let the election come, then fade the market. This is no more profound than "buy the rumor, sell the news." More likely, the market stays a bit longer, waiting for the Fed's QE2 announcement the day after, since there is a well-known FOMC pattern: the market usually rises into a FOMC pronouncement, then falls. The question du jour is, does the market follow a simlar pattern during mid-term elections?
Bespoke provides a wealth of data for elections and markets, worthy of baseball-statistician-turned-political-poll-watcher Nate Silver and his 538.com. Stocks normally do well on election day and the days after: a gain of 53 bp (0.53%) since 1970. Before 1970 the market was closed on election day, and looking at the day after as a surrogate, markets have been up 28 bp since 1900, vs. the average of 3 bp for all trading days. Over election week the mid-term election gain has been almost 1%:
-----
The fundamentalist point of view is that QE2 will not do much to help the economy, especially when the expected Tea Party victory will turn fiscal policy towards austerity. Trader's Narrative reports that the word from the pits in Chicago is that the average response from the primary dealers is to expect $930B of QE, but very few expect anything beyond $500B to be announced, setting up the market for a missed expectation and a drop Wed.
The technical point of view is that the market is nearing a top of sorts, even from the bulls such as Carl Futia. The ever bearish STU sees the S&P in the final throes of an ending diagonal, which requires one more sharp move up - perhaps the election day rally. This structure limits the rally to below Sp1209 (or wave 3 would be shorter than wave 5). Their alt count is a running triangle (which has a large second leg) that also means a sharp rally to come. (The current waves do not meet Neely's definition of a running triangle, since leg D failed to finish above leg B.) Normally the thrust out of a running triangle goes at least the widest part of the triangle, which would target 1214.
Under Fractal Finance we have just had a double top at 1196, and a triple top would normally be expected before a change of trend. This would be consistent with an election rally that holds below 1200 and fades. If instead we break and hold above 1196, Fractal Finance would remain bullish, as it has been since mid-Sept.
Another technical issue is an increasing divergence among the leading sectors, typical of a top. Tech stocks have been leading, and quite strongly, having run from a one-year low in relative strength (not price) to a one-year high in just two months (chart courtesy Bespoke):
The fomer leading sector of the Hope Rally, however, financial stocks, have lagged badly in the same period, and are on the verge of breaking to a new one-year low as well as breaking below a support level at 42.5 (chart courtesy EWTrends):
The irony of this divergence is that QE2 is largely designed to shore up the banks, especially in the face of a pending foreclosure-gate disaster. The banks manage the primary dealers of Treasuries, and would be beneficiaries of a concerted program of QE. They could use it to continue to swap toxic stuff out and good stuff in. For some reason the expectations of QE are not rallying the prime beneficiaries. The risk of a second wave of bank issues is trumping the QE bailout.
In Fractal Finance, when a break comes out of a trading range, then reverses back in quickly (relative to the length of the range), it is a False Break. Yesterday the Dollar continued an upwards break, but today fell back into last week's congestion zone. The S&P did the same in reverse. Under wave theory, what looked like the start of five-wave impulses in both indexes turned into three wave zigzag (sharp) corrections, as you can see from this chart by Principal Analysis:
The S&P had done a triple top in the e-mini futures but only a double top in the cash index, so a third retest is underway. If we break above the trading range, under Fractal Finance this is very bullish.
Another bullish indicator is the Golden Cross, which now is inevitable:
According to Traders Narrative, the Dow had the Golden Cross on Oct 1, and the Naz is also about there. We had a Death Cross back in June, and it proved to be of little importance. In general I do not give much credence to these crosses, but they are interesting to the extent they create momentum among traders who do follow them.
The bears are running out of things to hang on to. It is close to the kickoff of the seasonal strong period (Nov-May) when bears should go into hibernation. The bears can point to the S&P triple top, and to the reversal today not retracing all of the break yesterday. This bearish wave count is best put in this chart from the site which sounds like a Euro luxury car, gi61et:
Here is a chart of same not-quite-all-the-way retrace in the Dollar Index (courtesy Contrarian Advisor):
For those of you who like to fade Prechter, tonight's STU became short-term bullish, or perhaps more precisely, short-term on the sidelines. They read the false breaks in both the Dollar Index and the Dow/S&P as corrective waves, and therefore expect a final wave to come to retest or even break recent levels (below 76.14 in the DX). It would take another reversal back up above the 78.36 high in the Dollar Index to change this. Yet they remain agog at the bullish extremes in stocks - levels above both the April high and the Oct 2007 high. Some sort of stock correction is overdue, although it appears it will wait for the Dollar to bottom.
Those sentiment extremes came a couple of weeks before the actual top. This would suggest any clarity will await both the Nov2 mid-term election results and the Nov3 expected launch of QE2. The Repub's taking back the House seems in the bag, with around 55 seats likely to switch. Another 44 seats are contested ("99 seats on the wall" was the quip by Politico.) If the 44 split evenly, the power reversal will be historic. The Senate seems too close to call, with a gain to 49 seats likely but getting to 51 seats (necessary for control) problematic. As a consequence, uncertainty as to the future political climate remains an issue. The launch of QE2 is pretty well assumed, but the scale and timing are uncertain, and may be influenced by the Q3 GDP report on Oct29.
Neely expects an imminent strong up move in stocks to complete wave 3 of the final five-wave C. Target would be 1200. After that comes the overdue correction, but not an end to the upwards trend. The interesting site TradeYourWayOut has a bullish view which points to a top in March 2011, but first they expect the overdue correction:
There is a high expectation that the Repubs take at least one house of Congress, blocking any further expansion of the Obama agenda, and leading to gridlock for the next two years. There is a general view that gridlock is good for stocks, so a post today that looked at the data caught my eye. PragCap reported on a Fidelity analysis that shows NO positive correlation between mid-term gridlock and stocks:
Large-cap stock returns during post-midterm election years have been about the same, whether or not the outcome resulted in gridlock or harmony (i.e. one party controls both houses of Congress and the White House). Small-cap stocks, meanwhile, outperformed in years of political harmony compared to gridlock.
PragCap goes on to comment that gridlock may be bad for the economy in the circumstances we are in. The next chart shows how the stimulus at least buoyed GDP during the past 18 months, and (as I have commented) once past Peak Stimulus the shrinking stimulus is actually a drag on GDP growth that makes a coming double-dip more likely (the black line is the drag from less stimulus plus less inventory rebuilding):
There is hope that the private sector can more than overcome the drag, and expectations have been that the Q3 GDP report (which is released right before the mid-term elections) would be at least flat to Q2.
In August, however, the trade deficit widened, and it acts as an additional drag on GDP growth: the change in exports less imports is a major component of GDP growth, and a wider trade deficit means it could come in more negative relative to the prior quarter. There had been an import surge in Q2 which was largely based on slightly higher oil prices plus restocking for holiday sales, so expectations were for this component to decrease between Q2 and Q3, which would have pushed Q3 GDP growth up on this component.
Now, the NYT reports that Macroeconomic Advisors has reduced Q3 GDP expectations to 1.2%, a decline from Q2's 1.6%. Calculated Risk had expected a 2% Q3, but calculates that the trade deficit will end up about the same as Q2, meaning it will not improve GDP growth, and now estimates 1.5% GDP in Q3.
Ironically, a poor GDP report boosts the odds of QE2, which should boost stocks until they run up enough to account for the purported benefits of QE. If gridlock results in less fiscal stimulus and less effort to fix serious structural problems (unemployment, exports, bank lending), the normal rise in stocks between the mid-terms and the next election may be less than expected.
The WSJ ran an update to its July piece on the 1937 analogy. The chart shows we would be at the secondary top before the big drop if the analogy holds:
The 1937 has been the most robust of all these historical analogies, and was pushed by EWI as far back as March 2009 when they noticed that the end of the big drop in 2008 and beginning of the Hope Rally was a fractal with a similar bounce in 1937.
Like all analogies, they are curiosities unless backed by similar causal factors. Pattern recognition without causation is not predictive. The original WSJ piece, which I discuss in 1937 all Over Again, pointed to a number of similarities, including the potential double-dip. Those similarities still obtain, and the new piece adds two more:
It hasn't escaped traders' attention that something dramatic tends to happen at the turn of a month. Last year we saw turn points around the beginning of Sept, Oct and Nov, and more recently had the strong move up off Jul2 and the second move up off the beginning of Sept. Take a look at the stats since 1990:
The last session of the month tends to be down - note how the bars skew to the negative side:
In contrast the first session of the new months tends ot be positive 2 out of 3 times:
Since June 2009, the last session has closed lower 12 out of 15 times. The last six have been down, as have 10 of the past 12 months. On this basis, odds say to short before the open on Sep30 and go long on Oct 1.
Coterminously with this stats report, the STU today puts forth a series of time relationships that point to a top next week. At 1158 on Oct 5, the wave off the July bottom and the one off the Sep bottom will be identical in distance from the bottom and time (26 days). The whole discussion in this issue is pretty interesting, and is likely to be expanded upon in the monthly EWFF coming out Friday.
My back of the envelope Fractal Finance work also points to a turn date next week, based on the times between key bifurcation points, which are not the tops or bottoms but the points after those times when the market bifurcates out of a prior trading range and runs in a new move. The April 26 top, for example, didn't bifurcate down until the first week of May; and the Feb 5 bottom didn't bifurcate up until a week later.
Watch for a Dollar Index bottom as a signal for an imminent turn in equities. The buck continues to free fall and has very full bullish supporters, so some sort of low is due. Bill McLaren sees the next turn date in the Dollar as Oct 5.
The AUD hit 97c today is is approaching parity. The Japanese are trying to stem the rise of the Yen vs the Dollar, and the USD may be nearing key support. EvilSpeculator sees the USD in a black hole downward without any specific technical indicator of an immediate bottom, but is watching daily for the signs of it. This chart from ContrarianAdvisor shows the upper channel that may be the next support level for the Dollar Index. If so, we have a pretty serious move down before that Oct 5 bottom.
There are ... lies, damned lies, and statistics - Mark Twain, quoting Disraeli
September may be the cruelest month, but this time the joke is on the bears. This is the best Sept since 1939. I have been providing the plethora of historical analogies for consideration, and here is yet another: the WSJ looks at Septembers in mid-term election years, and creates a composite from seven such years in which future market conditions were uncertain.
There is a lot of speculation that after the election the market will rise, but the WSJ notes how money managers look forward, and once it became relatively clear what the mid-terms would bring (gridlock!) they looked beyond that issue to expected macro conditions.
The WSJ provides a companion chart of the actual data beyond the blended chart, and that too is interesting in that the average hides a dramatic skew in the six month period before and after:
1938 still is looking like a close comp in both economic worries (of a double dip) and the rise before the mid-terms. If this pre-election rally holds, it suggest the seasonal strong period, which normally begins in Nov, has already started. If patterns hold, it suggests a continued bull run into at least Jan 2011 if not the normal 'sell in May' high in Q2. If the 1938 analogy obtains, we should see the next deep drop begin on Q2 2011.
September almost always has a correction from its peak early in the month. It is well known to be the worst month for stocks (worse than October, with all of its memorable crashes), but it often starts well, as this one has - the third best since 1901:
The Big Picture provides more September stats on what happens after such a start: nearly every Sep has a correction averaging 4.5%. Only four exceptions since 1917. A priceless comment points out that this is an "almost useless factoid" since it is derived by comparing the highest close in the first part of the month with the lowest in the latter part. A random walk will say this happens almost always (over 90% of the time).
Still, the size of the drop at 4.5% average is pretty big. Does the strong start mean a bigger drop? Perhaps some reader can find stats on how likely that range is within the early and late parts of a month (or, more generally, across any three week period).
The thrust up must be faster than the triangle. Neely says twice as fast, meaning it completes in half the time, which means it should end today or tomorrow. So it could be done ... but we now hit a really nuanced area between orthodox wave theory and NEoWave. The STU has called this a thrust out of a running triangle. Neely notes that a running triangle indicates powerful forces (in this case bullish), and can propel the thrust to spike much higher than that - going usually 1.6x the triangle width (32 pts, targeting 1138), and at times 2.6x (44 pts, targeting 1150) - happy predictions for the head & shoulders crowd, and emboldening to the bulls.
But Neely has more precise requirements for such a running triangle, including that the trendlines should slope (in this case upwards). You can see such a pattern from 1987-1994, an upwards sloping running triangle that was followed by a very strong thrust, the extended fifth wave of the dot-com mania. In this case the running triangle is a barrier triangle - it hit a flat barrier rather than sloping upwards. This makes it merely an "irregular" triangle to Neely, which implies a thrust that is only "slightly" greater than from a normal triangle. Getting to 1127 fits that, as would a final attempt to hit 1130 later today or tomorrow. The market is about to overlap, however, which suggests a top is in, or we are seeing an ending diagonal (terminal pattern to Neely) form of the fifth wave thrust.
The S&P has twice run into the 1130 area and fallen back, and is getting close to a third test. We have been in a trading range between 1040 and 1130 since last May (after the flash crash) with a momentary spike down to 1011. A triple top will signal a reversal down.
In looking at the market in a Fractal Finance fashion, triple tops or bottoms almost always characterize the end of a trading range. The fourth or B wave triangle is a classic example, and is a continuation pattern of the larger trend. The occurrence of a triple top or bottom at a different position creates a reversal pattern. The classic head & shoulders is a variant of this phenomenon, as is the diamond pattern. Many have spotted the potential for a head & shoulders type of triple top combining the Jan high (1150), the April high (1220) and a retest at 1150, if we go that far; but the current pattern instead suggests resistance at 1125-30 and then a drop. Recently we had a triple bottom at 1040 which at the time I commented should signal a reversal up, which it did.
Although the triple top/bottom per se has not been a part of orthodox wave theory, the triple top/bottom was a major topic in tonight's STU from EWI. They point out how it has marked major tops and bottoms since 2000:
The most dramatic of these is the 2002 bottom, which Zoran Gayer called the Iraqi War Triangle:
(As a comment on Zoran's wave count, he had not yet matured his thinking, so posited a wave 2 triangle and expected a drop. Wave 2's do not have triangles, so shortly after this he advanced his theory to encompass the triple bottom pattern as a trend reversal, not a continuation.)
According to the STU, triplets typically come after a long move, such as the runup to 2000 or the drop into 2002. Sometimes these triplets form a diamond pattern, a common technical reversal pattern that is more general than the head & shoulders.
In Fractal Finance this is a way of describig the core fractal of the market, of a Thrust (the long move) and a Plateau (the trading range around the top or bottom). The plateau forms a range that could be drawn as a box, and often has a False Break along the way, which is spike outside the box that is quickly reversed back into the box - the "head" in the head & shoulders or the middle of the diamond in that pattern.
This thrust & plateau can be contrasted with the V shaped reversals of the sort we saw in March 2009, but that can be understood as part of a larger plateau on a bigger time scale. Time scales can confuse. The V shaped reaction is the False Break at a larger scale. The bottoming process in 2008-9 formed a plateau at a higher level (around Dow 8K), then shot below and reversed back up. It wasn't until after the July 2009 bottom that we finally bifurcated out of that range.
At an even larger scale, the market has so far formed a double top around Sp1550 as part of a large trading range from 1999 to now. That range is the plateau following the very powerful thrust up from 1982. One implication to keep in mind is that Fractal Finance would suggest we do not go to appreciable new lows before running up to test the old highs around 1500 one more time; indeed, more likely we hit and bounce off the bottom of the 2000-2010 plateau at around Dow7500. This third top is not required but the triple top is commonplace. Timing is not imminent, given the seven year gap between the first two peaks, but may be what happens if we have a powerful rally into 2012. This is a different perspective than orthodox wave theory, and would suggest the ultimate bottom is after 2012.
In the near term, the Dow has come right to the downward trendline that connects the April high and the August high, and sits at the 78% retrace, about as far as it should go under the STU wave count. The S&P has shot above the trendline, apparently to close a gap. A reversal back below the gap (1110) would make an exhaustion gap and could signal a reversal. Any appreciable move higher would make the STU have to reconsider their wave count.
From a Fractal Finance point of view, however, the most likely scenario is a retest back to the 1127 area. Interesting is that Carl Futia, who uses a variant of the box or plateau approach, expects the market to run into the 1127 area and hit resistance. He then expects a break down followed by a move up towards 1300. Again, Fractal Finance would instead expect the reversal to continue down.
All in all, the next few days should provide a good test of Fractal Finance against orthodox wave theory as well as the bullish Carl Futia's box approach. The timing for the triple top and bifurcation down under Fractal Finance remains the autumnal equinox around Sep 23.
We are in the second half of the last preseason game between Fractal Finance and Wave Theory. So far Fractals are up 3-0. On Monday I posed the test: fractals predicted a bottom and a run up to at least 1074 if not a triple test at 1125-1130; waves predicted a fast drop "early" in the week from a nested 1-2 formation. Fractals won the first half of that game as the bounce came (off a triple bottom) and the nested 1-2 formation was blown up.
While the nested 1-2 is gone, an alternative wave structure is stil intact: that the drop from Aug9 to Aug27 (or 30) was a complete five-wave down, a wave 1 down, and we are now in wave 2 - a single 1-2 not a nested 1-2. I discussed how the drop counts better as a three-wave pattern down, since the final waves (iv) and (v) are lost in the clutter along the triple bottom at 1040. The impulse waves (1 and 3) are also about the same length (60 pts) while wave 5 is a stub, which means no wave extended; a pattern characteristic of a three-wave zigzag not an impulse. (Under Neely rules, one leg of an impulse HAS to extend.) We have also blown past the normal wave 2 retrace of 50-62% and stand around 2/3 back up. This wave structure is well captured by AllAboutTrends, which expects a pullback next week:
The next test is at 1111, the 78% retrace. Theoretically a wave 2 can go back 99% and not break the count, but it seldom goes beyond 62%, and almost always when it breaches 78% it keeps going.
All in all, the STU is relying on a Hail Mary pass to save their count.
Tonight's STU fights back to resurrect their count, with two interesting observatioons:
We might therefore see a bit of a pullback as the trendline acts as resistance. Whether it will mark the point of a bearish reversal to new lows we shall see, but I think it unlikely. The STU has two bearish-reversal scenarios to consider:
If they get their B wave we may have the preseason game continue into some sort of overtime before it gets resolved. If you want to follow this a bit closer, EWI is running a Free Week promo next week, centered on their Euro and Asian STU services.
The Fractal Finance view does not need to to belabor minute details of wave structures to make its prediction, but I know a lot of readers would like it translated into patterns they are more familiar with. Let me offer it below the fold.
Continue reading "Fractal Finance Gets its Big Test Next Week" »
It is well known that September is the poorest month for sticks, worse than October. It gets less press because a lot of the well-known crashes occur in October (eg. 1987). The seasonal strong period kicks off in Nov, but at times starts in Oct, after one of those crashes; and hence Oct overall is slightly positive for stocks.
I have been intrigued by continued comments in the STU that we are in a seasonal positive period. They are using it to help explain why their dreaded nested 1-2 setup is not acting like a coiled spring with a sharp drop down. The implication is when the momentary positive seasonality passes, the market drops.
I scanned for data, and it comes back predicting the first three days of this month will be positive. So far so good for Wed and Thurs; what about Friday? A summary of findings below the fold, including a look at the First Friday in September - tomorrow!
Continue reading "September is the Cruelest Month For Stocks, But Kicks Off Very Positive UPDATED" »
The PE ratio has dropped rapidly in the past year, from 23 to 15 (see chart, courtesy WSJ). The WSJ analysis is this reflects increasing uncertainty ahead in earnings estimates. As I recently noted, 2011 estimates have dropped considerably in the past month. On top of that, the skew in estimates has widened, with the spread from lowest to highest increasing from $12 (between $85 to $97 next year across the S&P) to $15 (between $80 and $95). The WSJ concludes this measures an increase in uncertainty, and uncertainty begets a backing off of bullishness in stocks, which turns up in the PE ratio.
It is interesting to look at the WSJ chart above for the past decade: it is as if the great credit bubble of Greenspan short-circuited a normal bottoming process, which is still ahead of us. Similarly, the great inflation under Jimmy Carter kept PEs skimming across the bottom until Volcker finally got it under control and the great bull market of the past 25 years was on. Both these examples support the argument that economic uncertainty depresses the appetite for equities.
SeekingAlpha published a counterpoint to this conclusion: that the sharp drop is merely an artifact of PE being backwards looking:
Corporate earnings have improved very dramatically since the depths of the recession. A higher E with a flat-to-lower P means a lower P/E Ratio. Mystery solved.
This argument proves too much, since it also means as earnings rise in a bull market, PE's should be depressed. A quick look at the chart shows the opposite: PE ratios dropped rapidly in periods of uncertainty, such as 2002 and 1974 (and to the right of the chart, in 1938 and 1930-32), but rose during periods of economic (and earnings) growth.
I still believe the better analytical tool is to look at the first derivative - the rate of change - of earnings forecasts. As they drop, the market tends to drop; and with the PE ratio looking back 12 months, the drop in P should show up in a rapid drop in PE, especially as we step forward each quarter. Using this metric, right now we have rolled over from growth in forward earnings forecasts, which had reached elevated levels (to respect the recovery, such as it was):
The author of this chart has the following observation on earnings revisions and the S&P (emphasis added):
[A]nalysts were forecasting a near 20% decline in earnings at the market’s trough. Today, expectations are for 22% growth in the year ahead. The average annual gain on the S&P when earnings growth estimates have been below 4.2% has been a positive 17.2% return, consistent with last year’s monster bear market rally. Regrettably, the average annual performance of the S&P when earnings growth estimates have been above 14.2% has been a decline of 4.6%.
You could argue that at some point the sharp drop will bottom, and maybe we are about to bottom. The PE is still relatively high at 15; no major secular bull market began with PEs as high as they are now. PEs were below 10 in 1933, 1938, 1949, 1974 and 1982, all kickoffs to a bull (although whether 1974 fits is debatable). A further drop below 10 is probably required.
We might see an interesting phenomenon if corporations continue to take advantage of low rates to borrow: they might turn around and repurchase stock, as HP just announced. This would improve the P by reducing the number of shares, spiking the PE ratio. Related to that is the interesting speculation on ZH today that QE2 may not help the economy, but it could spur a wave of M&A: borrow cheap and consolidate your sector. This should also spike P, by putting an acquisition premium into stocks. Would this be enough to skew the past, and find a bottom with PEs as high as 15? It would take (in effect) more than a 50% premium in P across the S&P to create this effect - unlikely to say the least.
After a crash as we saw in 2008, the punditry launched rationales for various crash analogies - 1929, 1987, Japan, etc. A veritable fleet of analogies was floated into the blogosphere. By themselves these comparisons are intellectually interesting and little more - it is not sound to extrapolate from a one-time sort of event. Market patterns reflect underlying psychology, so when a pattern is correlated to similar exogenous conditions, the odds of some level of predictability rise. This is why the 1937 comparison appeals to me, given the high-level similarities:
This time around we avoided some crucial errors, especially the Fed intervention in 2008, which prevented the sort of freezing-up of the banking sector that we saw in 1931 after the Bank of the United States failed in Dec 1930; this may be a primary reason why the 1929 analogy failed to fit. There are also significant differences in the 1937 analogy, which is why it has been a bit remarkable that Summer 2010 is tracking closely to 1937 (with update posts here and here).
In terms of timing, starting with 2000=1929, the tighter comparison would have been 2008=1937, not 2010=1937. Remarkably, the pattern we saw in the summer of 2008 seems also to match very closely as well to the market in the past few months. The exogenous events are also, oddly enough, lining up, especially the upcoming election:
The STU provided this 2008 chart on Friday, which I am reprinting with permission, and accompanied it with some interesting analysis:
In 2008 the peak of the summer rally happened on Aug11 with a subsequent bottom on Aug20 and a secondary top on Sep2. The "bifurcation" out of the August trading range didn't happen until Sep19, when we fell off a cliff. This time the summer peak has been Aug9, and the recent bottom on Aug25, both fairly close to the 2008 dates
You can see in the STU chart how they draw a 'plateau' around the period of Aug11-Sep19, showing the trading range. The bifurcation occurred on a second drop below the range; the first was a false break that was rapidly retraced into the range.
The wave pattern has been an ABC flat, which has subdivided enough to have ended, or might have a small subdivision left. While the wave structure points to a drop Monday or at least early this week - and that remains the STU's primary wave count and prediction - the 2008 pattern suggests instead we meander upwards through the upcoming week (Sep3) and perhaps even a few weeks beyond.
We just had a fifth Hindenburg Omen. A flurry of Omens does not increase the odds of crash after the second, but does strengthen the Fractal Finance case that market entropy (disorder) is increasing. (An Omen, if you recall, is when both new highs and new lows are high on the same day, plus other criteria). Entropy increases as a trading range extends, and leads to a sharper thrust out of the trading range. This means a sharp bifurcation is coming once the divergence in positions is worked out. That working-out shows up as a continued trading range.
In the Fractal Finance view, we have been in a trading range (a "plateau") since May25 which has ranged from 1040 to 1130, with the momentary break of July 2 a False Break that was quickly corrected. The mid-point of the range is 1075. Plateaus are sideways moves like flats and triangles. and like triangles, often end near the mid-point. The recent drop off the Aug9 high also ran into a small plateau from 1074-1087. Thus a suitable end point of this whole trading range is 1074, the bottom of the recent plateau and the midpoint of the whole move. Neely as well often sees a retrace in a corrective move to the mid-point of a "box" built around the trading range. The 2008 experience, however, says the small degree plateaus are but a part of a large pattern, from Aug9 to sometime in Sep. While the 2008 pattern suggests we go higher than the mid-point, even retest the 1130 range at the top of the box, note that the bifurcation happened off the mid-point of the range (see chart and look at Sep19).
We can now try our fourth test of Fractal Finance vs Wave Theory. So far Fractal Finance has succeeded in each test.
The Bear's bear, David Rosenberg, opines that it is not earnings but changes in earnings expectations that drives the market:
In Ahead of the Tape this morning, the WSJ supported this point of view. Q2 earnings were strong across the S&P: up 38%, well above the expected 27% YoY increase. Yet the market is dropping - and began to fall just as earnings forecasts began to be revised downward for 2011 (see chart). 2011 earnings estimates started in January at 25% growth, were reduced to 20% in April, and now are hovering at under 15%. The change of view:[T]he peak in earnings forecasts coincided with the peak in the market [in April].
And guess what? The forecast peak in the last cycle was in October 2007, again right when the S&P 500 was hitting its highs. Before that, earnings estimates were starting to get cut in August 2000, just ahead of the peak in the market.
[T]allies with the outlook for broader economic growth, which has weakened considerably.
Even the 15% prediction may be revised downward, as it is nearly triple the expected GDP growth for 2011, which itself is being revised downward. Goldman this morning is urging analysts to cut GDP estimates even further, to the range of 1.5% in 2H10 and early 2011. Yet as recently as the last week of June, before the poor GDP report, the WSJ was touting only modest decrease in earnings expectations:
At the time earnings were expected to rise in the second half of 2010. No longer.
We may be at "Peak Earnings":
If the macro economic data continues to disappoint to the downside (as JP Morgan suggests) then it’s clear that analysts have been fooled into believing that the stimulus and 80% rally are self sustaining.
Clearly analysts have not yet fully embraced the economic slowdown. Why not? Using an expectations model from The Pragmatic Capitalist, they may have overshot mean reversion on expectations too much to the positive, and are due for a double dip there, too:
As the Hindenburg Omen meme has soared across the blogosphere, the reaction is coming in from statistical skeptics. One stats maven shows that the Omen is only modestly predictive (r-square of 45%), about the same as a more understandable alternative: how much is the market over-valued?
A more prosaic criticism comes from the odd numerology in the Omen, such as the days low and high above 2.2%, or the second confirmatory Omen to occur within 36 days, and then predicting that within 41 days, half the time, a crash occurs. Why these particular numbers? Because the Omen is a backwards fit to historical data, which is pattern matching of the worst sort; a pattern must be based on underlying causality to be predictive.
Nonetheless, the creator of the Omen has exited the market, expecting it to do its worst. He had expected to exit by Sep anyway, as we are in the seasonal bad period.
The Omen has also emboldened a bull, who says if indeed we get the wash-out bottom in Sep/Oct, it would be the four-year low point, pointing to good upside in the last two years of the Presidential cycle. Yes, that cycle is the most steady of any, but even then should not be blindly followed. The typical pattern is driven by stimulus prior to re-election, but with the likely loss of control of Congress, Obama may be hard pressed to pass any more stimulus. Maybe the Repubs will push a tax cut, giving Obama the stimulus anyway? Ironies would abound.
One of the scariest technical indicators is the Hindenburg Omen: it predicts a crash, or at least a serious drop. It happens when the market skews with high numbers of both new lows and new highs. Under Fractal Finance principles, it reflects a high degree of chaos - distribution from one position to another - right before the market finds its new direction.
You can read the requirements here, but suffice to say that when it occurs we also see a flurry of bearish blog posts. As now.
It takes two observations to make the Omen. While the Omen goes back to the '70s, the classic treatment on it came from Robert McHugh back in 2005, who added two more tests, including having two confirmed Hindenburg sitings within 36 days, which confirms the Omen.
The Double Omen has occurred before every stock market crash, but not every Double Omen is followed by a crash. A warning, not a prediction. But one with high odds: over 75% predictive.
We just confirmed a Double Omen:
The Double Omen preceded the 1987 crash, the 1989 mini-crash, the 1998 crash, the 2000 top, the drop before 9/11, and the plunge into the 2002 low. The last Double Omen was in June 2008 (crash!).
A "crash" can be mild (>5%) to severe (>15%). The crash occurs within four months, with over half within 41 days. Odds are 77% will result in at least a mild crash (>5%). The odds from the previous 27 confirmed double Omens are:
The methodological debate right now is whether the low volume and high degree of program trading 'bots have caused a false signal. Maybe the Flash Crash was also a consequence of high-frequency-trading and the anomalous behavior of these 'bots. Or maybe take warning: we have a high fractal correlation with the 2008 drop right now.
In my recent post, The Market is Due for a Bounce, I mentioned how Fractal Finance required a bounce, not a final wave 4-5 down as allowed by orthodox wave theory. Of course the bounce happened; and then Fractal Finance required it to be a false break and settle back into the trading range of 1074-1087, which it has now done. So far so good. So what is next?
I have been using EWP for charts and count, as they are clear and simply stated. Their most recent chart notes how declines have been sharp and bounces sloppy, a sign of a down trend. Last night they expected a decline, and got it today. They see the odds favoring a deep drop, but await confirmation, which is a break below where we got to so far today.
The Wed STU had an interesting take: that the bounce yesterday created a nested 1-2 with the break out of the trading range on Tues being wave (ii) off the recent drop, and the drop & rebound Wed being the inside waves i-ii. The Tues bounce filled the gaps mentioned in my Bounce is Due post, and retraced a little over 50% of the wave (i) drop, both suitable indicators of an end to a wave 2 bounce. The wave structure broke as a flat, which completed. If this is a nested 1-2 pattern, the market should have had a sharp break down. Today started that way, but has stalled within the prior trading range, as Fractal Finance predicts from a false break.
The STU allows for a more complex pattern to develop, with a target of the 62% retrace around Sp1106 by Friday (tomorrow) or possibly into Monday. Given the five-wave up pattern on Tues, a zigzag could have been developing, but the deep drop today obviated that. If the market does continue back up tomorrow, a likely wave pattern is a larger flat, where the first wave back was wave A. The next chart shows this, courtesy Market Thoughts.
In Fractal Finance, the base fractal is a Thrust and Plateau, as shown in the final chart. In comparing this to wave theory, the sideways corrections (flats, triangles and double- or triple-corrections) form the 'plateaus' of disorder. The impulsions, impulses and zigzags, mark the 'thrusts' of order. Sometimes it helps to re-envision wave patterns that are given traditional labels. The recent ending diagonal, for example, entered a plateau around the turn of the month, and it was the sharp break below it that marked a change of trend - the bifurcation point.
Often a plateau contains a false break below, followed by a similar one above. This happened over the past two weeks. A bifurcation does not occur until the thrust out of a plateau completely leaves its range. The test comes quickly. A false break needs to return back to the range in a much smaller period than the length of the plateau. (The market is fractal, and a plateau can hours, days, or even decades, as we saw from 1966-1982, or are seeing from 1998 to date.) So far this week, the plateau which began 7 trading days ago was re-entered within 2 days, meaning it has been a false break.
Typically a bifurcation occurs at a triple-top or bottom, which wave theory usually would count as the ends of waves 3 - 5 - 2/B. Sometimes these triplets form a head & shoulders where 5 is higher (at a top, or lower at a bottom) than 3 or 2/B; but often they all hit around the same resistance zone. In either event they are distribution patterns, with positions rotating from smart to dumb money. The chaos theory, a plateau reflects a market seeking order - the pattern of distribution, as the relative buying/selling pressure equalizes as the direction of the market becomes vague. Hence a final run up tomorrow to around the same 1100 level a third time would be within the theory.
More likely, however, is a the market melts down from here. The drop today ticked to the level of the prior false break of Monday. Typically a second false break is the precursor to a bifurcation. Hence tomorrow is another test of Fractal Finance:
The Big Picture snapshots NY fashion and notes how the hemlines are below the knees. The NYT sees it as a backlash against the miniskirt. The hemline indicator goes back to 1926:
The "science" behind this is that stocks are a refection of social mood, and it also shows up most vividly in fashion, music and movies. They are concurrent indicators, and can confirm an imminent change of trend. When skirts get long, movies get dark and music goes in the studio, watch out.
This month's Elliott Wave Theorist doubles down on music as an indictor with a deep dive into the popularity of The Beatles, their musical transformations and the 1960s stock market. Even if the stock market angle is not interesting, the dissection of Beatles history is a magical history tour for Beatles fans and well worth the read. You can access it for free here.
Looking at this chart, reprinted with permission, a couple of items jump out:
The concept of social mood is captured in this album cover, quickly pulled, that reflected the darker mood of the late 1966 bear market drop. It presaged themes that would dominant the bear market '70s: The Exorcist, anti-heroes, heavy metal, disco dancing in dark venues. Good thing it was pulled, because The Beatles brand was an ebullient bull market brand.
EWT's conclusion: a phenom like The Beatles occurs as a combination of talent, luck and timing. Once they become the symbol of popular mood, the manic stage is on. People pour their hearts into the symbol. Music seems to have a peculiar affinity for this, with similar manias around Sinatra ('40s), Elvis ('50s) and Michael Jackson ('80s). When the mood shifts, as in the end of the '60s bull market, the act no longer fits. The Beatles broke up, Elvis went to Vegas, and Michael Jackson became known as an oddball.
Implication for investors: the hemline indicator is but one of a lot of indications of a hunkering down for a long bad patch. TV is flooded with shows on magic, the occult, and darker themes. The better movies explore darker side of human nature, such as The Dark Knight. Bull market acts are gone, and the state of the music industry is largely doom & gloom.
The death cross is interesting not because it has much predictive power but because it has been widely perceived as predictive, and can spook the herd. Yet I wonder if in the age of the Internet is the herd so easily moved? Whereas in the past (about a decade ago!) investors would be heavily influenced by brokers, or simply rely on mutual fund managers, today there is a great flood of commentary pummeling the attentive. In the past week there has been a stampede of death cross comments, most of which pointing out its self-congratulately character: when the cross occurs, the trend is already pretty clear, so what does it add?
The last Death Cross occurred in Dec 2007, right at the start of the big drop; and the last Golden Cross happened in June 2009, shortly after the Hope Rally kicked off, confirming a longer trend up. In both cases it would have given guidance to ride the trend. Many clever contrarians who thought the Hope Rally would end at its first peak in June 2009 were shown up fairly quickly; and the great mass of fervent fundamentalists who dismissed the warning sign in late 2007 lost an Ark's worth of wealth over the next year. With recent 20:20 hindsight, the Crosses look pretty useful.
This hasn't stopped the fundamentalist sheep dogs trying to keep the herd away from technical analysis. The WSJ proffered a debunking a few days ago, using an interview with Pierre Lapointe, global macro strategist at Brockhouse Cooper, who said:
The death cross IS nonsense. They’re no better than a flip of a coin to predict future returns. Check out these odds: Since 1970, only 10 of the 21 occurrences actually resulted in a market pullback a month after the death cross. Three months later, the market was down only 43% of the time.
This morning they had a slightly less breezy market note:
Such a death cross has also marked every other bear market since 1972, notes Ed Yardeni, president of Yardeni Research. And the market's 50-day average is currently less than 1% above this tipping point.
Dow Jones ran a more considered piece that look into the stats, concluding:
There have been nine occasions when the 50-day moving average crossed the 200-day moving average since 1998, and five could be considered false signals. One might say they seem pretty good at marking contrarian turning points.
The particular market analyst went on to note the bigger head & shoulders pattern shown in my prior post, and added that breaching the neckline (which we have done) with the death cross pretty well confirms a deeper drop. Overall the point was that a Death Cross in itself doesn't say much (other than that the herd might be spooked) but in concert with other indicators can give guidance. I agree with that - none of these indicators in isolation is worth that much, but the bundle of indicators can predict market direction. What I get amused by in this case is combining the self-congratulatory Death Cross with the often-seen and seldom-realized head & shoulders: dumb and dumberer so to speak.
But if they spook the herd, they can be acted upon. We should get a really good test in a couple of days: will a Death Cross spur a bearish dump? Or, if the Internet has leveled the paying field between "experts" and investors, we might see the herd fading the experts, in effect betting for a bounce just when the "experts" say they should be spooked.
Over at The Big Picture, from the Chart Store we have some deeper stats on Death Crosses. Two charts are presented, one where the 200 DMA is still rising (as it was in Dec 2007) and other where it is falling and the 50 DMA catches up (as we see now):
If we return to Pierre's comments on the Death Cross, he does add that there have been nearly two dozen death crosses over the past four decades, and the S&P has declined by 0.4% on average one month after such a formation. After six months, on the other hand, the market on average had gained nearly 5%. Here is how he continued:
Fundamentals is the only way to go… until risk moves investors’ attention away from them. And this is where we stand right now. Risk is dominating. The global economy is in much better shape than it was in 2007 before the credit crisis and global recession: home prices are much lower and less of threat; mortgage rates are at all-time lows; and companies are back in hiring mode (albeit at a slow pace). But investors are parked on the sidelines.
Funny comments! The world is in better shape after the greatest drop since the 1930s in the global economy? Such is the wisdom from fundamentals. The corner is always about to be turned.
In bull markets, investors rely on fundamentals - why not? everything is going up. In bear markets, they turn to technicals - especially when they get tired of being told to buy the dips because happy days are almost here again. At a real bottom they don't listen to the siren calls of brokers at all, and that is the time to buy back in.
Mother Market is a perverse lover, always trying to fool. As we begin to understand that Mother Market is a nonlinear chaotic system, we can appreciate that fooling the herd is her essential nature. Chaotic systems always seek order, and occasionally find it. When they find it, markets react crisply and quickly - think Flash Crash. As they seek it, they live on the edge of chaos and jink and jive precisely to fool their followers: as too many take a position on one side, the market will jink to the other.
The whole field of technical analysis comes from the classical world of linear thinking: patterns that emerge and behave consistently. Markets are fractal in nature, and behave with a constantly shifting underlying order which we can barely discern if ever understand. The fractal nature of markets was first recognized by Mandelbrot (the picture is the visual representation of the fractal Mandelbrot Set), and initiated furtive investigations of fractal finance. The herd instead believes in the Random Walk theory and follows the Efficient Market Hypotheses. If the market is fractal, it is non-random and potentially predictable via technical analysis. The leading Elliott Wave practitioners recognized the fractal nature of wave patterns a decade ago, but have not pushed the theory into the world of Chaos Theory. Zoran Gayer was making strides in this direction when he unfortunately passed away.
Hank Wernicki of Elliott Fractals is going where no one has gone before, constantly testing fractal patterns as predictors of future market moves. He posts often in comments to this site, and exposes his work in realtime for success and criticism. I deeply respect that process. One of the great improvements in human knowledge from the Internet is to treat the community as a test bed in itself. Companies are being launched today using rapid iteration from realtime feedback and deep analytics to figure out what customers respond to. Hank is pushing the envelope on the usefulness of market fractals by field testing his fractals as he goes, and improving their accuracy and predictability.
Over the weekend Hank got endorsed by Bob Prechter's EWT:
He tells me his site was overwhelmed Monday, so I held off a day to publish this. Everybody loves Hank! Kudos!Fractal analysis: After watching Henry (Hank) Wernicki for a year or so, I think he is onto something. Wernicki visually scans past short term patterns in the S&P for ones that look like “now,” and on that basis he attempts to forecast the market’s next move, usually a few hours or days ahead. When the market stops copying itself, his opinion is stopped out and he moves on to another picture. Hank makes bold calls, so if you are a day trader, you might want to check out his work. You can find him at www.elliottfractals.com.
For the past two months the market has begun to swing between panic and greed. Robert McHugh reports that we have had eight 90% down days and five 90% up days - very unusual, and not conducive to a continued rise. This sort of whipsaw action is more typical of market tops and bottoms, where divergences emerge - and we are seeing extreme divergences. Friday's STU reports that Friday was another odd day, with volume the lowest in two months (before the 90% up/down days started) and yet the lowest TRIN in 58 years and the highest TICK since 1987. A record low TRIN indicates that huge buying pressure was need to move the market up just a little bit. The STU concludes:
it's got to be one for the record books of bearish setups
Bloomberg Businessweek will run a cover story on the market bears, which normally would be considered bullish: when a story finally hits the popular press, the trend it celebrates is usually just about over. Most of us know this as the Sports Illustrated Effect: a quarterback or pitcher who gets on the cover lauding success one week typically has a disastrous outing the next time. It is clear, however, the point of the story is to roast not toast the bears. Prechter is not the main course (Roubini is), but is grilled well-done:
One of the most famous of these extremist outsiders is Robert Prechter. In the 1970s he revived an old system of measuring investor psychology called the Elliott Wave Principle and used it to advise subscribers of his investment newsletter on Oct. 5, 1987, that they liquidate their stock holdings. Two weeks later, the market crashed. Prechter was hailed as a genius—though the label didn't stick. In 1993, The Wall Street Journal ran a page-one story headlined "Robert Prechter Sees His 3600 on the Dow—But 6 Years Late." He stayed pessimistic through the 1990s, and in 2002 said the Dow Jones industrial average would fall below 1000. It surged 25 percent the following year and kept going up until 2007.
Ironically, the dismissive tone towards bearish advisors turns the cover story indicator from bullish to bearish. Since the market is flat or up 80% of the time, random darts and rally monkeys will seem more prescient than bearish pundits. Fear unfolds faster than greed, so the timing of bearish predictions needs to be much closer to the event. This raises the stakes for bears preserving their hides. Fortunately for Prechter's reputation, despite his Lost Decade from 1994-2004, he predicted the 2007 top and the 2009 bottom with pretty close timing. After two premature calls for the end of the Hope Rally, he made a correct call right before the April top. Still, he tends to intersperse a longer term bearish view with the short term calls, as this analysis of Prechters record demonstrates, where the 4/2/09 quote they attribute to him, for example, misses his call a few days earlier of the coming Hope Rally; and the report simply ignores his correct 2007 call. Acting on that one call before a 50% fall would have cured a lot of prior misses, although his complete miss of the dot-com bubble is simply inexcusable.
So what about now? The bullish Thurs and Fri took out the STU's über-bearish triple nested 1-2, but did so on volume 40% below average, which gives no comfort to the bullish case. The simplest way to count the market right now is to make the end of wave 1 down the May25 low (although Daneric makes a case for a leading diagonal down that ended on Jun8, which leads to similar predictions as the Big Tease count). The May25 first wave creates either:
The June Swoon count will emerge or die pretty shortly. It might start with a pop on Monday then a drop. A bounce above Jun3 levels kills it, and a fall below the recent low at Es1075/Sp1080/Dow10108 pretty well confirms it.
My take: given a month-long wave 1 down to Mar25, the bounce in the purported June Swoon wave 2 is a bit short and fast: a 38% retrace that took place in a week. Normally a wave 2 would go back 50-62% of the drop, and take more time. So while you might want to be prepared for a reversal this week, expect a slower and longer retrace. We head into a triple witching Friday (quarterly close out of options), so also expect the whipsaw market to continue.
As it spikes up or down it also measures the buying or selling pressure it takes to move the market. As Dick Arms interprets:
The Index seems to work most effectively as an overbought/oversold indicator. When the indicator drops to extremely overbought levels, it is foretelling a selling opportunity. When it rises to extremely oversold levels, a buying opportunity is approaching.
The recent spike above 13 shows it took a lot of selling to move the market down. This can signal exhaustion of a trend and a buying opportunity if it comes at a bottom, but can be used as an early warning signal if it comes in the middle. Bespoke smoothes the TRIN to a 10-day average to take out noise, and came up with this stunning warning shot across the bow: the TRIN spiked in March 2007, signaling the coming credit crisis:
It may be signaling the second wave of that crisis is ahead.
Stock market patterns reveal market psychology. Wouldn't it be better for market psychology to predict market patterns? We are living through a great example right now with the potential Leading Diagonal pattern.
The market is a nonlinear chaotic system which at times follows order (impulses) or (more often) lives at the edge of chaos as it seeks order (corrections). Impulses typically stretch as the psychology is one of piling on. We saw this in 1987 and again recently with the Flash Crash; and to the upside in the 1984-6 and 1995-7 bull runs. The impulse patterns require one of their segments to stretch or extend, reflecting this piling-on psychology.
At times impulses instead compress and form what are called Diagonals or (at the end of a wave) Terminal patterns. This compression comes from a market psychology of anticipating a major event or policy change. In Mar08 we saw an Ending Diagonal (ED) as the market anticipated a bailout of Bear Stearns and slowed the piling on. Another recent ED happened right before the GM bailout in Nov08. I discussed ED's at some length a few months ago.
A leading diagonal (LD) is more rare and indicates hesitation in the kick-off to a strong move. The chart from EWI shows a bullish LD. They give trader tips on diagonals as well.
In a diagonal, unlike normal impulses, wave 4 overlaps wave 1. In addition, the wave ends converge in a wedge. (I explored the expanding ED, and concluded it was not a proper structure; it doesn't fit the psychology of compression.) Also, wave 1 is longer than 3, and 3 is longer than 5. It compresses.
The LD starts a wave like a nested 1-2 structure while the ED ends a wave like a triangle. The LD pattern breaks as a 53535 structure, while the ED is 33333. This avoids confusion as to where one is in a larger wave structure. (The site I cited in the prior post argues an LD can be 33333, which I think is unsound.)
When one sees nested 1-2s, as now, the antennae should perk up for an LD. When technical analysts call out a Diamond Formation, the antennae should perk up for an LD. Whether it is nested 1-2s or an LD pattern cannot be confirmed until the fifth wave:We had smaller degree nested 1-2s off Apr26 and then a stunning drop - confirmation.
We have larger degree nested 1-2s into Friday and then overlapping waves - caution flags. Today is more of the same.
How can market psychology inform us as to which it is early enough to place bets in the market? In contrast with the prior ED examples where we fell hard and were momentarily saved by government bailouts, this time we were rising until belief in bailouts began to fade. The fall off April 26 coincides with the Greek Tragedy in Europe and a whole series of ever-increasing bailouts. This weekend the G20 met to discuss continued global stimulus, and the Europeans are backing off. This does not end the matter, as a major G20 meeting is upcoming in Toronto in two weeks.
My conjecture is the current overlapping waves down reflect the fifth leg of an LD. Today the market is so far in a trading range as well. The tell from market psychology: we are in a wait-and-see period for the next and possibly final bailout attempt.
if we are in an LD, it should drift down but no farther than Sp970-1000, shorter than the third leg from Sp1174 to Sp1041. After that the retrace should head back at least 50% into the 1095-1107 area. The 1106 and 1115 gaps may remain unfilled. Then we start the next Big Down. This could all look like a June Swoon followed by a Summer Rally and then yet another August Peak before an October Crash.
The SlopeofHope sees the glitter of a Diamond Formation in today's market, a rare and bearish pattern. Like a head & shoulders, it shows distribution of stock and signals a change in trend. The diamond is not a recognized wave pattern, and you can read in an EWT from 2005 how to interpret it under wave theory.
It is interpreted by the STU tonight as a nested 1-2 pattern, and they give strong support to a stock distribution story: we had an all-time record of buying climaxes last week (at least back the past 20 years of the Investors Intelligence database). The STU notes that we had 1019 buying climaxes last week, whereas the week leading to the March 2009 bottom had exactly the opposite, 1010 selling climaxes. A buying (selling) climax is when a stock makes a 52-week high (low) and then closes the next week lower (higher), and signals rotation away from the trend. This chart from EWP shows the setup:
If so, tomorrow should drop pretty fast below the recent lows around Sp1182 as we kick off wave iii of (iii). Indeed, it sets up a modest-risk, high-gain short, since if right, it drops hard and a lot, but if wrong, the stop can be confidently placed at the prior wave top (Cash Sp1209). Even if the nested 1-2 setup fails tomorrow, the broader indications of stock distribution signal that the top is very near.
EvilSpeculator throws some TA cold water on this by noting that the VIX closed back within the 2.0 Bollinger Band (two standard deviations) below its 20DMA, giving almost giving a short-term buy signal (a reader adds that: to confirm, it now needs a higher close (sell) or lower close (buy) than the close of the day back inside the 2.0 Bollinger Band). If the STU is right and we do drop tomorrow, the aborted VIX signal will have failed, which says the "countervailing trend is strong" in the words of Friday's STU. That would be consistent with a 3 of 3 down.
I have commented recently that the STU tends to find nested 1-2 patterns too frequently, sometimes with inside waves larger than outside. This one does not suffer from that, and both down waves can count as "5s" while the upwaves are "3s". Friday's drop still lacks an extended wave, but we have passed beyond that issue for the moment: either we get the hard wave 3 of 3 drop tomorrow or the nested 1-2 count is simply off. The other concern I have with the STU count is that the 2nd waves both retraced more than 62%, indeed 78% in the S&P and 89% in the Dow today. While allowed, generally wave 2's don't go past 62%, so it raises a red flag that some other count is on.
I have seen some comment that we are back in the ending diagonal, but this pattern no longer fits that, and if a new ED is forming, it is premature to call it. Instead, as suggested today by EWTrends, we may be in for an extended sideways period, as we saw back in November. Wave Principle suggests one way to count this is as a complex correction wave 4 or a B wave before the final wave up:
The imperfect drop Friday gets counted as a "3" and today as another "3" for waves A and B of a second flat or maybe a triangle. Or, the wave (ii) is extending sideways out a bit, and the imperfect drop would be an X wave. In either case, we have a sideways move.
If we do rush down, an interesting take on conforming the anomalies is from Market Thoughts: the Friday drop can be counted as a leading diagonal, a variant on the ending diagonal which has a compressed 5-3-5-3-5 impulsive structure without an extension.
I suppose it is still theoretically possible that last Thurs is the top - the drop from which started before the "Sue Goldman" story broke, indicating it was driven by technicals not news. (I find the market in the short run almost always technical, with occasional jinks and ives from news that do not last very long - as the bounce from Sue Goldman clearly shows!)
While the break was expected, it is not expected to be more than a pullback before a higher high. It is not clear that it is over, but most likely so. The pattern to look for is whether a break is a false break, meaning goes below the lower trendline that has marked low points all the way up (as this one did) and then comes back above and stays above. MarketPulse shows the break of the trendline, and then coming above:
The Dow settled in at a 78% retrace of the recent high (the close at 11117 sits right there), although went almost all the way back intraday. The S&P closed a gap at 1203 and sits right at 78% as well, although it too went a bit higher intraday. If this were an impulse down, the first drop would be a wave 1 and the bounce back a wave 2. Waves 2 can go back 99% but seldom break 62%, and almost always when they break 78% go to new highs. Thus both major indices sit at the point where any move higher essentially de-confirms a top.
On the other hand, this whole break could be merely a correction, a wave B or minor wave 4, before a continued advance. As such, the 78% rule does not apply. Besides the obvious move to new highs, watch whether it stays above the trendline, which also de-confirms the break down.
If the top is still ahead, what could flag that it is near? One indicator is to watch junk bonds, which act like an intermediary between bonds and equities, and are sensitive to changes in underlying market conditions - specifically, a change in junk indicates a change in riskier bets. Doug Short has a nice piece today from Chris Kimble which shows how they tend to bottom or top ahead of the S&P:
Another indicator is copper, my favorite "Cunary in the coal mine". Copper has been rising in price as well as inventories, indicating real demand. In 2008 copper prices plunged and inventories rose - no surprise. As prices rose, inventories got worked off, and inventories turned back up, but prices kept rising. Usually copper rises early in a recovery, so this is a good sign; but much of it is attributed to Chinese stockpiling. As suppy/demand has rebalanced, prices have begun tailing off, an indicator that Chinese tightening is having an effect. They key to watch is a continued rise in prices, which either reflect a huge "bubble echo" in commodities (more on that in a later post), or a real recovery. (Funny, the prior link speculates about Peak Copper driving prices up even as inventories stay tight, which to me is yet another indicator that we are in a Bubble Echo in commodities - when oil prices peaked i 2008, Peak Oil was the explanation, not bubbleconomics).) If instead prices keep dropping with inventories, this indicates slackening demand, and the Cunary chirping "top ahead!"
Since much of the the Cunary is driven by China, a third indicator is the SSEC, the Shanghai composite. Emerging markets play a similar role to junk bonds in that they are a riskier place to invest and are sensitive to trend changes. Trendlines runs a series of posts on the SSEC and is worth checking out. Their current view is a drop in the SSEE to complete a triangle, then a thrust higher. We all remember how it bottomed first, in Nov2008, before our Mar2009 bottom, and led the way up. Its secondary top after the 2007 peak was back in August 2009. If it thrusts higher out of the triangle as Trendlines suggests, we might see it beat the Aug2009 high. The top after that could give a three or so month early warning of the top in the US.
ContrarianAdvisor, befitting its name, takes the opposite view: the SSEC has just broken below the 50 DMA and the lower trendline of a wedge, and is within a tenth of a percent of the lows for 2010. If we stay below the wedge, the SSEC would signal that it is now in the next leg down, and chirp a warning for US stocks. The potential fall suggested by this pattern is 30%.
While a fall by the SSEC itself this is not worth much as an indicator, look to see if it is confirmed across other emerging economies, especially the BRIC or BRICKS (BRIC plus Korea and Singapore). It may be the wedge with three tops so far is presaging a long topping action in the US, the triple top that may emerge from Jan19, Apr15 and whenever the final top comes between May and August.
More ominously is if the wedge breakdown is due to expectations of China letting the Yuan float again. The drop in the SSEC has been dramatic - yesterday was nearly a 5% fall.
Dow spent most of the day above 11K, but fell below right before the close - and then shot up to close a little above. S&P got within a point but couldn't break 1200. Naz100 is still below 2K. A scattering of commentators poo-pooed the round number levels, but they are widely watched and inspire the media to crow about higher highs ahead. Tonight's STU, in contrast, takes the round-number levels quite seriously, as they have tended to be levels of resistance for further advances. The Dow hovered around 100 for over a decade, and after almost touching 1K in 1966 and slightly breaching it in 1968 and 1973, the Dow did not stay above that level until the mid-'80s, teasing us for 20 years. Suffice to say, they don't make hats for Dow 11715.
Surrounding 11K are a number of warning signs. These do not mean THE top is in, but raise the odds of a pullback. The STU lists a bunch tonight. I find the most interesting to be the "anemic" volume (their word) of less than 1B shares on the NYSE for five of the seven trading days in April. Volume should pick up as we approach Options Friday, but so might downside volume. We seem on track for an interim top around April 16, a Martin Armstrong Day (inverted), or Options Friday on April 16. First let's see if the Naz100 and S&P can join the round-number club.
Volume has been a puzzle throughout the Hope Rally, diminishing as the Rally meanders on. This is not typical of a bull market, which should be gaining momentum. One index I watch is the $NYAD, which shows the NYSE advance-decline, and can be looked at in cumulative or moving-average mode. Usually it drops ahead of a market top - it turned down in June 2007 for example (see chart). But it is continuing upwards, to levels of overboughtness.
Mole over at Evil Speculator sent me a really fascinating analysis of the TRIN and what it is telling us right now. I strongly urge you to peruse his Sunday missive, The Flow of Investor Sentiment. He starts by explaining that the TRIN is the advance/decline number divided by advance/decline volume, and is used to indicate (among other things) overboughtness:
His TRIN chart with loads of annotations shows that the the 10 DMA of the TRIN moves up in bulls and down in bears - the opposite of what the oversold/overbought crowd expects of the TRIN itself, but this is a moving average - but after turning sharply down since Dec has now begun to reverse back up.
The divergence indicates a trend change coming.
Mole goes on to explain what he thinks is happening to investor psychology, and concludes it *might* support the current STU view, of a coming B wave pullback before the final C wave to end the Hope Rally.
The title of today's post comes from Tony Caldaro, who commented on my pencast Are We In A Bull Market comparing his bull market view with Prechter's bull-oney reply. Ironically, Tony got the phrase from an old Prechter post. In The Yelnick Challenge, Tony rises to the occasion and deepens his argument for a new bull. He gives four requirements for March 2009 being a major bottom, and the Hope Rally the kick-off to a new bull market to, yes, higher highs soon. To counter the bull-oney charge, he comments on the two points I made in the pencast, about lack of alternation between the Jun09 and jan10 corrections, and the lack of impulses in his motive waves (the three upwards move off Mar09, Jul09 and Feb10):
Another issue remains the lack of increasing volume. While at a major turn, volume usually increases (both before and after the turn) as positions rotate in and out, and then increases at the kickoff of the new direction. We saw that around March 2009, which marked it as an important low - confirmed as events have turned out. Yet the continued lag of volume in the whole Hope Rally since the turn is a major concern. Typically a new bull market brings increasing volume, especially in the leg Tony thinks we are now in, wave 3 of III, which normally shows gaps in the direction of trend as investors finally recognize the new trend is here and pile on. The always-interesting site of Doug Short has a chart from a weekly guestblogger on point:
Tony has an explanation for this in his post which I recommend you read, about when different sets of investors recognize the change. He thinks the public is not there yet, and presumably that would explain the lack of the more normal markers of a wave 3 of 3. Yet, objectively, why is it then a wave 3?
Where I think Tony is spot on is his characterization of March 2009 as a major bottom. This doesn't mean we are in a new bull, but may mean it stands as a low point for a while. Here is my view of the three choices we are in - the first two make Mar09 a major low for a while:
What I respect about Tony is his adding of more rigor to wave analysis. He added to orthodox wave theory what Neely did - more precise rules and guidelines - but kept within the frameworks of Elliott Wave theory rather than pushing to a new wave approach. Just a glance at the myriad of wave counts by the plethora of blogger wave sites shows how orthodox wave theory has too high a degree of discretion and subjectivity. As you may have noticed rather than join the wave amateur hour with my own chart fiddling, I prefer to comment on broader issues using wave theory as a perspective. I applaud Tony for pushing towards an objective approach even as I doubt his current count.
In this regard, he also gives us tests of whether his bull market view is correct. In his Weekend Update, he notes that wave 1 of III (from Jul09) went 281 S&P pts, and wave 3 of III normally would go some Fib relationship of wave 1:
A short 3 (61.8%) would end very close to an Sp1214 end point arrived out from a Sideways Bear perspective. (I will post on this Sunday evening, after the Masters.) If we turn around there, it does not confirm his bull count, but it does causes him a methodological problem. Usually one wave of an impulse extends, typically 3 or 5. Indeed, under Neely rules, one segment must extend. And 3 cannot be the shortest. If 3 is shorter than 1, wave 5 has no room to extend, since it has to be shorter than 3. So 1 would be the candidate extended wave, making his wave III as a whole an odd duck which started stronger than it finished.
A normal 3 at Sp1326 would be close to another Fib relationship at Sp1333, calculated as 161.8% of wave I (the wave from Mar9 to Jun11, 2009). The normal 3 would blast past the 62% retrace level of Sp1229, and past the 2/3 retrace at 1275. It would be close to another relationship, the 70.7% level at 1307. If we get there, the odds of his New Bull increase, and the odds of Prechter's Big Bad Bear decrease. Prechter's P2 count is a wave 2, which rarely retraces more than the range of 62%-67%.
An extended 3 at Sp1500 would pretty well confirm Tony's view. It would have blown by the 78% retrace at 1378. If a wave 2 seldom goes beyond 2/3, when it goes beyond 78% the odds are de minimus of it being the right wave structure. Under wave rules, waves 2 can retrace as much as 99%; but Tony's extended 3 would be followed by a 4th and 5th waves, which would very likely exceed the 2000 and 2007 highs in the S&P. Hence the Big Bad Bear would be dead. Even if a Sideways Bear remained, it would have been a mistake not to have played the rise! Better to make money in practice than be right in theory.
A final perspective on this bull or bulloney discussion comes from Glenn Neely, who is in the Sideways Bear camp. He has been increasing his profile, publishing a series of perspectives in SafeHaven and issuing some audio commentaries. His sees us as in the least predictive part of the wave structure, and uses this chart to explain why. More as well in this audio commentary.
We may now have a clean wave structure for the choppy rise from July 2009 to January 2010. This makes predictions of the next phase of the market much more reliable: most likely we continue up to the 62% to 2/3 retracement, or the range of Sp1250 +/- 25 pts. This being the third zigzag since Mar 2009, it will be the end of the Hope Rally, as a triple zigzag is the max seen. Timing is least clear, but if it runs about the length in time of the first zigzag, or three months, it should end in May, or at the outside, August.
The new wave structure comes from the current EWFF, which has gone back to the future with more wave discussion and less socioeconomic context. They introduce a new wave form, first discussed in April 2006, that previously was seen in forex markets, not equity markets: the skewed triangle, an upwards sloping triangle which shares some (but not all) characteristics of Neely's running triangle. It skewers upwards due to the strong force of the larger trend. (A running flat has a similar skewed shape.) When in a B wave, as this one is, waves A and C are expected to be very close Fibonacci multiples of each other - and in this case the C wave almost precisely equals the A wave. Neat.
More in this pencast discussion. Comments welcome on the pencast format for discussing waves.
In the short run, when the Euro trading desks open overnight, we should see a final spike up to end the A wave of this final zigzag. If so, we get a more serious B wave pullback that might fool the bears into emerging from momentary disarray and hibernation. Carl Futia has us already at the top of his latest box, and expects a drop tomorrow (see chart). (Carl's box suggests a sideways move to the end of the box.) The STU in contrast is the more bullish, expecting a pop up towards Sp1200/Dow11K before the drop, but as always, they have an alt count as well. Neely too today went long. Carl may be momentarily the bear, but his target for the final run up is Sp1320, a 280 pt rise from the Feb low, matching the 290 rise of the first zigzag and the 280 rise of the second.
Recent Comments