TitleDescriptionKeywordsGeneratorRobots
20, September 2017
Association of Technical Market Analysts, BOMBAY STOCK EXCHANGE
Integrity • Intellect • Inspiration
Expand your Technical Analysis expertise.
Connect & network with recognized experts
  • Education
  • Guru Speak
  • Chart In Focus - Tom McClellan

Tom McClellan

Gold’s 13-1/2 Month Cycle: Right Translation
-Enable Images- in your email reader to see this chart o- use the -view in browser- link at the top.2 month cycle' width='600' height='317' />
September 07, 2017 

Gold has been pushing up to higher closing highs, which is getting the gold bugs all excited.  But we are now late in the 13-1/2 month cycle that is dominant in gold prices, and so we should expect a drop into the major cycle low due at the end of 2017.

But there is a lot more to the 13-1/2 month cycle than just when to expect the major lows.  For starters, there is a mid-cycle low that usually arrives around the mid-point of the whole cycle.  The mid-cycle low is usually not as punctual as the major cycle low, but it is still important for figuring out the message that gold prices have to convey. 

Just noting the existence of a mid-cycle low is not all there is.  How prices behave around that mid-cycle low conveys information about what lies ahead.  If the high for the cycle arrives before that mid-cycle low, that is known as “left translation”, and it carries a bearish message about what lies ahead.  It says that the price bottom at the major cycle low should take out the last cycle’s bottom. 

“Right translation” occurs when a higher high occurs after the mid-cycle low, and this says that gold prices are in a more bullish mode.  The prior cycle low should not get taken out at the next cycle low.  That is the condition we see right now, with gold prices pushing up to a higher high. 

So now that we have this message of a more bullish condition in gold, we have some assurance that the Dec. 22, 2016 low of 1130.70 for near-month gold futures should not get taken out.  That still leaves a big distance that gold could drop between now and the end of 2017 without violating this “rule”.  But the general point is that strength late in the cycle breeds strength again during the next cycle.  So once we get past the major cycle low at the end of 2017, we should be able to look forward to larger gains again for gold prices.  Until then, though, gold has some late-cycle corrective work to do, if only to get the whole world convinced that gold can never go up ever again.  Once that work is done, the setup will be complete for the next uptrend.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
5 Hits

Tom McClellan

Chart In Focus

Why Are Bond Yields Staying Low?

 

Chart In Focus

July 21, 2017

There is a 60-year cycle in bond yields which has existed since bonds first came about in the 1700s.  It says that bond yields should have ideally bottomed in 2010, and by now we should be well into a 30-year rise in yields lasting until 2040. 

But that is not how it has worked out.  Bond yields have stayed low 7 years beyond the ideal bottom date.  So is something wrong, or is the cycle broken?  Or perhaps is this just an normal sort of anomaly?

We have seen prior examples of bond yields bottoming a bit off schedule.  There was supposed to be a yield bottom in 1890, but bond yields waited until 1900 to finally bottom.  Wars, ship sinkings, and other human events have also skewed the cycle a bit.  So this current delay could just be another example of a delay in this cycle bottom’s arrival.

Or there could be a deeper explanation.  We tend to think of the Fed as the authoritative controller of interest rates, something which I debunked at this link.  A study of long-term historical data shows that another agent is more powerful than the Fed.  Sorry if that last sentence sounds like economic heresy.  But do you want to believe what your econ professor taught you, or what the data say?

The 60-year cycle also appears in global climate data.  One place where it shows up best is in a bit of data known as the Atlantic Multidecadal Oscillation (AMO).

Atlantic Multidecadal Oscillation and interest rates

According to this 60-year cycle, global average temperatures should have turned down by now, and made for a cooling planet like we had in the 1970s.  Cooler temperatures mean more freezing events, poorer crop yields, and thus higher prices and inflation rates which bring higher interest rates.  But temperatures have not turned down, which means that former Vice President Al Gore still has a complaint, and thus a business model. 

And it also means that interest rates have remained low (show as high readings on this inverted scale chart).  The movements of global temperatures, as modeled by the Atlantic Multidecadal Oscillation, tend to lead the movements of long term interest rates by about 6 years.  If we are going to see a rise in interest rates, we should expect to see a drop in temperatures beforehand. 

There is a lot of noise in the (monthly) temperature data, but this is a long enough chart to see the trend.  And we see that the AMO data tend to lead the interest rate trends by about 6 years.  The AMO data do not go back to before 1856 (reliable thermometers were still being perfected), and so we cannot see what it would have said about the little ice age in the late 1700s and early 1800s known as the Dalton Minimum.  

As for why this relationship between temperatures and interest rates works, my hypothesis is that it is because cooler temps are bad for crop yields, causing higher food prices which trickles into other price data, and thus pushing up interest rates.  I cannot fully explain the reason for a 6-year lag. 

As for the 1890 cycle low not arriving until 1900, the 1890s were pretty warm.  But then colder temps (and a quiet sunspot cycle) in the 1900s led to rising wheat prices, the creation of the Fed, and WWI.  This latest sunspot cycle has been the quietest one since that Dalton Miminum.

Cumulative Sunspots per cycle

Eventually that lower solar output that the planet is receiving should start to matter in terms of cooler temperatures, poorer crop yields, and thus higher inflation and interest rates.  Delaying an event that the cycle says should happen does not necessarily negate the cycle; it can instead postpone the work to be done, forcing the commodities markets to work extra hard to make up for lost time.  And that should lead interest rates to start the delayed rise toward a top in yields due in 2040.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
26 Hits

Tom McClellan

Chart In Focus

A Different Sort of Presidential Cycle

 

Chart In Focus

July 14, 2017

 

It is by now an overused phrase to say that we are in a different sort of presidency right now.  And befitting that theme, we are seeing a really different sort of behavior of the market relative to the Presidential Cycle Pattern.

This week’s chart shows a version of our Presidential Cycle Pattern that is constructed by averaging together the stock market’s performance only in periods when there is a new president from a different party than the last one.  We have found that the market’s personality differs quite a bit according to whether there is a new sheriff in town from a different group than the last one, versus a status quo type of president. 

Normally a new president from a different party brings a market rise, at least until May of the first year, just on hope that everything is going to be better.  Then as investors realize that all of their hopes are not getting realized right away, that hope turns into disappointment, and the market declines during the summer and into autumn. 

What we are seeing this time is a rise, all right, but it is continuing now into July, and investors are not yet showing disappointment.  That’s a verbal story.  But if you look closely at the chart, you can see that the SP500 has been zigging and zagging all at the wrong times, according to the Pattern.  In other words, the correlation has been inverse.

But that statement is only true when one looks at a certain time frame.  The overall path of the SP50 has been higher like it was supposed to, but the minor pattern has been inverse. 

This revisits a point I wrote about in August 2010, under the headline of “Correlations May Not Be What They Seem”.  The point is that having a trend in the data makes an inverse correlation seem positive, both visually and quantitatively.  Here are a couple of charts from that article.  The first shows a perfect inverse correlation:

sine waves inverse

But if we add an artificial uptrend to that inversely correlated data, the calculation of a correlation coefficient flips to a strongly positive one:

sine waves in uptrend

The point in reviewing these principles is to see past the 8-month uptrend in the first chart, and notice the inverse relationship in the smaller movements, even as there is an overall uptrend.  That overall uptrend can disguise the inverse movement on the shorter time scale. 

With that point firmly in mind, we can see that the SP500’s movements have been backwards from what the schedule says.  There is your different sort of presidency, and different sort of price response.  The SP500 bottomed at the end of June, just when this Presidential Cycle Pattern said that a minor top was supposed to be seen.  And now the market is rising at a time when the Pattern says prices should be falling.

All of this inverse behavior would be concealed from those who look just at quarterly or monthly returns data.  You have to look at the chart closely to see the insight.  And the immediate message is that prices should continue higher for much of the rest of July.  A longer term chart shows that this autumn will get even more interesting, assuming that the inverse correlation of the minor patterns remains inverse.

presidential cycle first term presidents

If it does remain inverse, then the Pattern’s decline into late September should mean a strong advance for stock prices then.  But here is the caveat: If you ever count on a relationship remaining inverse, that’s when it can flip again and fool you.  So you must keep a watch on it, and notice if it “disinverts”.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
19 Hits

Tom McClellan

Chart In Focus

Phillips Curve Is Not Even Wrong

 

Chart In Focus

July 07, 2017

To paraphrase Wolfgang Pauli, the whole idea behind the Phillips Curve is “not even wrong”. 

A.W.H. Phillips studied the relationship between inflation and unemployment in the United Kingdom, and noticed that they were usually moving in opposite directions.  He therefore theorized that when unemployment is low (and it is hard to find workers), prices of things rise because employers have to pay more to hire qualified employees.  That led 2-3 generations of economists to undertake an effort to determine what is the exact tradeoff between inflation and unemployment, thinking that if a government or central bank changed one factor, there would be an equal and opposite reaction in the other factor.

The problem is that neither Dr. Phillips nor the generations who came after him understood the real relationship.  It is not one of opposition, but rather of lead-lag.

In the chart above, the plot of the CPI-U Inflation rate is shifted forward by 2 years (24 months) to reveal how the U.S. unemployment rate follows in the same footsteps.  That is important because the inflation rate bottomed in early 2015, and so the echo of that bottom is coming due right now for the unemployment rate.

This is not a perfect relationship, though.  Sometimes exogenous events put a thumb on the scale.  We saw that especially in 2008, when the commodities bubble sent the inflation rate up to an unnatural high, which was followed by a crash of equal magnitude to the downside.  Two years later, the unemployment rate did not exactly match those dance steps.  But after a few months, the relationship got back into step again. 

The CPI-U inflation rate rose from the low in April 2015 to its high in February 2017.  Add two years to those dates, and we get a low for unemployment due in roughly May 2017 and a high in February 2019.  But it does not appear to be as big of a rise as some of those we have seen in the past, so don’t worry too much.  But do brush up your résumé.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
28 Hits

Tom McClellan

Chart In Focus

Treasury-Bund Spread Gives Early Warning of the End

 

Chart In Focus

June 30, 2017

The 28-year record high spread between 10-year T-Notes and German 10-year “bunds” is finally starting to narrow just a bit.  This is a warning that the great bull market in stock prices from the 2009 low is in its last stages.  But it is not done yet.

Ever since June 2009, the yield on the US 10-year T-Note has been higher than its German counterpart.  It turns out that this is a pretty bullish condition, at least for as long as the spread between the two is rising.  As a bull market ages, though, this spread shows the wearing out by displaying a divergence relative to prices.  It can take many months for the divergence to finally matter, and bring a meaningful price decline.  We are just now starting to see those first signs of such a divergence.

This spread peaked in May 1999, 7 months ahead of the DJIA’s Dec. 1999 peak, and 10 months before the SP500’s March 2000 top. 

The lead time was even longer ahead of the 2007 top.  The Treasury-Bund spread had its highest reading in Oct. 2005, but started downward from a slightly lower top in June 2006.  That was still more than a year ahead of the stock market’s October 2007 top, so there was plenty of warning if one had listened to this indicator. 

Another peak in the Treasury-Bund spread arrived in April 2010, which was just a month ahead of the infamous Flash Crash, but 12 months ahead of a more significant top which came in April 2011. 

Now most recently, we have a peak in this spread in Dec. 2016, and only now in May and June 2017 is it really starting to decline.  So we are likely still a year or more away from a final price top for the stock market.  A top in mid-2018 fits well with the expectation offered by crude oil’s 10-year leading indication

It is important to understand when using these insights that this relationship between the Treasury-Bund spread and stock prices has really only worked for about the last 30 years.  Prior to the 1980s, it did not work at all.  Here is a longer term chart:

Treasury-Bund Spread 1956-2017

We should remember that in the decades after World War II, Germany was still a rebuilding country, divided in two by the postwar agreement with the Soviet Union, and definitely not a favorable risk in terms of its sovereign debt.  So its yields were understandably higher than those of U.S. debt.  Then in the late 1970s and the 1980s, the Federal Reserve was fighting inflation that began with the 1973-74 Arab Oil embargo, using higher interest rates.

Only in the late 1980s did things settle down, and with the fall of the Berlin Wall in 1989, Germany came to be seen as a legitimate financial and industrial power, and a good risk in terms of its sovereign debt.  And that is when this leading indication for stock prices began to work. 

I was stationed with the US Army in Germany at the time the Berlin Wall came down, and the two Germanys reunited.  It was obvious that it was a transformational moment in history, but I could not have known then that it would mean this change in how the yield spread would work as an indicator.

If we ever see a time when the yield on German bonds is higher than that of U.S. 10-year T-Notes, making a dip below zero in these charts, that will mark a great buying opportunity for the stock market, if the past 30 years’ experience is any guide.  And a few months from now, we should see a major price top for the stock market, which the current divergence is just now starting to foretell.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
35 Hits

Tom McClellan

Chart In Focus

Narrow Range for McClellan Oscillator

 

Chart In Focus

June 23, 2017

A quiet market is one of the hallmarks of a price top, when no one seems to care enough about risk to move the market very much in either direction.  The NYSE’s McClellan A-D Oscillator has recently been displaying some of that quietness, trading only a few points above and below zero until just the past couple of days.  That quietness in Oscillator readings is telling us something about that very complacency I was talking about.

So to look at it more quantitatively, this week’s chart looks at the 15-day high-low range of Oscillator values. It’s calculation method disregards at what point level that range occurs; it is just looking at the highest minus the lowest readings over the past 15 trading days.  What we can see in this chart is that the high readings occur near inflection points, usually as prices are turning up. This makes sense. Think about a nice oversold (very low) Oscillator reading, which is then followed a few days later by a crossing of zero to a nice high reading. That is going to increase the magnitude of the readings for this 15-day range indicator.

But it is the low readings for the 15-day range that I find even more interesting, because they tell us more about what is going to happen, as opposed to what just did happen. These low readings, below around 120 points, seem to give pretty good signals that a price top is at hand.  This is similar to several other types of indicators that portray market calmness, such as Bollinger Bandwidth (AKA standard deviation), VIX, and Average True Range. Calmness of prices and investor complacency go together, and usually appear together at market tops.

A low reading like what we have just seen does not necessarily have to lead to a big price decline.  But it does say that traders are feeling complacent, and thus that those who are going to be buying have likely done so, and thus there is little immediate pressure left to help push prices higher.  Waiting for a fearful event, with its associated volatility and oversold readings, will get a better short term entry point. 

This is a newfangled way of looking at McClellan Oscillator interpretation.  For some of the more longstanding methods, check out our Learning Center chapter on the McClellan Oscillator.  And to see a chart of the McClellan Oscillator updated every trading day, visit our Market Breadth Data page.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
45 Hits

Tom McClellan

Chart In Focus

Correlation Between VIX and SP500

 

Chart In Focus

June 15, 2017

When the normally inverse correlation between the VIX Index and the SP500 gets crazy, if offers us a great message.  That is the point behind this week’s chart, which is based on a great observation by Jesse Felder of www.TheFelderReport.com

Jesse first wrote about it in a Tweet here back on March 3, 2017, and that same day it was featured in a MarketWatch article.  I did my own investigation, which revealed that this is indeed a really cool insight. 

What Jesse did, and what I have replicated here, is to calculate a 10-day Pearson’s Correlation Coefficient between the VIX and the SP500.  You can do this incredibly easy in any spreadsheet program, or even more easily as Jesse did at www.stockcharts.com.  Just call up a chart of $SPX, choose as your indicator “Correlation” from the list, and set “$VIX,10” in the parameters window.  It is that easy.  Then you can adjust the period under observation as you might wish. 

What we see is that most of the time, the correlation hangs around down near -1.00, meaning that they have a strongly negative correlation.  In other words, if the SP500 goes up, the VIX usually goes down, and vice versa.  That’s what is normal.  But the instances of abnormal behavior contain the really interesting information. 

Here is a regression chart showing the one-day SP500 change versus the one-day VIX change:

VIX SP500 scatterplot

Each dot represents one day’s combination of the SP500 change and the VIX change.  You can see that most of the dots line up close to the linear regression line, and that makes complete sense.  It is not a perfectly inverse correlation, but it is a very strong one.  Over this entire study period since January 2014, the correlation for their daily percentage changes is -0.83, which is pretty close to a perfectly inverse correlation.

But if you calculate the correlation coefficient using the raw VIX and SP500 indices rather than their daily changes, then the math is different.  I don’t want to get too deep into the statistics, but I want to make the point that there are differences between running correlation analyses of the raw indices and those of their daily changes.

What’s more, real statisticians will tell you that Pearson’s Correlation Coefficient is the wrong statistical tool to use for a time series anyway.  There are other more suitable tools for analyzing the strength of relationships between two contemporaneous time series.  But they are a whole lot harder to use and to program, and so a lot of technicians just use what is easier.  And sometimes what is easier can sometimes be good enough. 

The chart of the 10-day correlation between VIX and SP500 is good enough to tell us when there is a moment of strange behavior between the two data sets.  And those moments of strange behavior just happen to be pretty good at marking tops for stock prices.  The higher that the Correlation Coefficient goes, especially when it gets above zero, the more important the message.  And that usually means a more significant the price top.  But meaningful tops can be found when the Correlation Coefficient gets up to a level shy of the zero line. 

Ten days seems like a good period for this purpose, but others may work as well.  Darshan Dorsey asserts that a 22-day correlation coefficient works even better. 

On June 7, 2017, the 10-day correlation went up as far as -0.17.  That was not quite to zero, but big price tops have been found on lesser readings.  Usually the corrective mode suggested by one of these readings lasts until this 10-day correlation gets back down closer to -1.0, or perhaps longer.  So there is still a lot of room for a correction to do its job from here before we can say that the correlation has returned to “normal”.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
38 Hits

Tom McClellan

Chart In Focus

The Unexplainable 4-Year Rerun

 

Chart In Focus

June 08, 2017

President Obama and President Trump are entirely different types of leaders.  No one would contest that.  And the second term of a presidential term is a lot different from a first term in the way that a president interacts with the public, with Wall Street, with Congress, and with the economy.  So there should not be any stock market similarity between Trump’s first term and Obama’s second term. 

And yet the pattern correlation to four years earlier which began during Obama’s second term persists even now, with Trump in office.  That is the point of this week’s chart, and it is a relationship I have shown before

I confess that I had figured that the relationship would break correlation by now, but it seems to want to persist, for reasons of its own.  This is where the news followers’ heads explode.  Some people think that it is the news that drives the stock market.  But if the news is wholly different, and yet the behavior is the same, then perhaps that hypothesis about the news driving the market needs to be revisited. 

If this strong correlation continues, then we can look forward to a late-June price bottom, followed by another surge to a higher high in July.  Given how strong this correlation has been, would you want to bet against it continuing?

Tom McClellan
Editor, The McClellan Market Report

Continue reading
42 Hits

Tom McClellan

Chart In Focus

When Not To Go Short Volatility

 

Chart In Focus

June 02, 2017

The VIX is a supposed “volatility index”, but it does not really measure actual volatility.  Instead, it measures what options traders think about volatility.  All of the various investment vehicles that have popped up in recent years that are tied to the VIX have enabled traders to go long or short “volatility” with relative ease compared to a few decades ago. 

And the short volatility trade has been among the most profitable, especially since the Fed, ECB, and BOJ started various flavors of quantitative easing in 2009.  Betting on increased volatility has only worked a small fraction of the time. 

VXX and XIV are ETNs that allow investors to pretty easily go long or short volatility.  VXX bets on a rising VIX, or perhaps I should more precisely say that it bets on rising prices for VIX futures, since that is what it actually invests in.  And XIV is a short VIX ETN which tracks the overall stock market very closely.  So as stock prices rise, the VIX typically falls, and thus XIV goes up. 

What’s more, XIV gets a further boost from the contango in VIX futures.  It goes short at a (usually) higher priced contract 3 months out, and then covers when that contract is in its final month before expiration.  Usually that works out well for XIV investors who get to harvest that “contango”.  Here is what the current term structure in VIX futures looks like:

VIX futures maturities

So as long as the spot VIX Index remains low, and VIX futures retain their nice, steep contango, it is a gold mine to own XIV and harvest that contango.  That is what has led some analysts to proclaim that XIV is a great solution as a permanent part of one’s portfolio.  But there are times when XIV turns out not to be such a good investment, times when there is very little opportunity left to harvest. 

This week’s chart helps us to see when those times are, and the key is to look at the price level of the highest priced VIX futures contract.  When that falls to a low level, there is little opportunity left to harvest that contango, and there is also arguably too much optimism.

The price of the highest VIX futures contract is shown in the top chart on an inverted scale, the better to correlate with price action.  When that price gets “below” 18 (high readings on the chart), that tends to mark a topping condition for XIV.  In other words, there is little opportunity left. 

Here is a longer term chart of that inverted scale price history of the highest VIX futures contract, versus the SP500:

Highest priced VIX futures contract

Clearly the better opportunities to invest in the stock market, or to short volatility, come when the VIX futures are at really high prices (low chart readings).  And when the highest VIX futures contract’s price falls into the teens, there is not much opportunity left to be short volatility, at least not profitably. 

This is not to say that the market has to go down, nor that volatility has to shoot up right now.  All it says is that risk/reward is now no longer in favor of those who have enjoyed and profited from shorting volatility in the recent months.  Someday, the market will present us with a better opportunity.

Tom McClellan
Editor, The McClellan Market Report

 

Continue reading
52 Hits

Tom McClellan

Chart In Focus

Junk Bonds Don’t Confirm Higher Highs

 

Chart In Focus

May 25, 2017

The SP500 has rebounded from the May 17 one-day panic to push to a higher high.  But high-yield bond ETFs like HIO are not confirming that higher high, and that’s a problem.

High yield bonds typically move in sync with the stock market rather than with T-Bonds.  They are all about liquidity and default risk, much more so than inflation and other interest rates.  So when liquidity starts to dry up, that condition often shows up first in the high-yield bond market.  Eventually those liquidity problems come around to bite the rest of the stock market.  So it is important to pay attention to these divergences.

Sometimes the divergences are bullish, as we saw when HIO made a pattern of higher highs in March and April 2017 while the SP500 was making lower highs.  That foretold the stock market strength which eventually materialized.

The message does not always work, though.  Right after the November 2016 elections, it got a bit screwy.  Lots of things were screwy then.  So don’t assume that it is always perfect.  The same point applies for all technical analysis techniques and “rules”. 

Right now, this divergence between HIO and the SP500 fits well with my expectation for a 2-3 week dip into a low due in June 2017, which should be followed by a strong new uptrend.  Such a selloff could serve the useful purpose of scaring out the weak hands.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
58 Hits

Tom McClellan

Chart In Focus

Will Labor Shortage Kill Housing Boom?

 

Chart In Focus

May 20, 2017

We know by the message from lumber prices that the next 12 months should be a positive period for all sorts of housing related data.  New home sales, for example, tends to follow in the footsteps of lumber price movements with a lag time of about 1 year.  So because lumber prices have been trending strongly higher, that should mean higher numbers of new home sales.

This is especially true with all of the “echo-boom” generation getting into their late 20s, and starting to look at buying versus renting.  The peak birth year of the echo boom was 1990, and those kids are now 26-27.  Zillow says that the average age of a first time home buyer is 33 years, so that pig is still moving through the python.

But there are new worries that those homes won’t get built for those kids to buy, because there may not be enough trained labor to build those homes.  Data from the Bureau of Labor Statistics (BLS) shows some tightening in labor rates for the construction industry.  And when the overall unemployment rate is at 4.4% (yes, I know about the problems with those numbers), it is harder to attract workers out of other industries to come and work construction. 

Here is total employment in the construction sector, all types, seasonally adjusted:

Construction employees

It has had a big overall rise, just as total population has in the U.S.  And there are obvious big swings during recessionary periods. 

Looking deeper, here is the unemployment rate for construction workers:

Construction unemployment

There is an obvious seasonal fluctuation, as winter gets in the way of construction work.  Spring 2017 is seeing the jobless rate drop as normal.  But the 12-month moving average shows that the overall rate is already down to the same low level it reached in 2006.  This suggests that there is not a whole lot of slack remaining in the market for construction workers.

Meanwhile, job openings continue to climb.

Construction job openings

The 12-month moving average (MA) is back up to where it peaked in early 2007.  Back then the housing bubble was choking itself off with all of the speculative buying, condo-flipping, and overleveraging with junk mortgages.  Now it seems that the genuine demand for workers to build the houses that the echo-boomers need is exhausting the available supply of trained construction workers. 

This has all sorts of implications for the housing economy, and anecdotal reports show that twenty-somethings are having a hard time finding available stock of housing to buy.  That is likely to mean continued price increases for starter homes, although not necessarily for other parts of the housing market.  Whether it has implications for Congress addressing immigration laws to allow more construction workers to come in from Mexico and Central America is a wholly separate question.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
64 Hits

Tom McClellan

Chart In Focus

China Is Pulling U.S. Bond Yields Higher

 

Chart In Focus

May 12, 2017

It is not a surprise that U.S. Treasury yields are correlated with those of other countries.  And so it should not be too much of a surprise that when China’s 10-year sovereign bond yield is zooming upward, that the U.S. 10-year T-Note yield should follow. 

China’s 10-year yield is getting pressure from shorter maturities.  Their 5-year sovereign yield just pushed up above their 10-year yield for the first time since records began.  The moves are being credited to Beijing’s efforts to stop the slide in their currency.  Maybe China’s central bankers are believing former Fed Chairman Ben Bernanke’s 2007 statement, that “the yield curve could be inverted for a considerable period without significant implications for the economy as a whole, yes--- possibly for some banks, but not for the economy as a whole."  Yeah, that turned out well. 

Whatever the cause, it appears to be having an effect on U.S. yields, as this week’s chart shows.  But while the chart shows what is currently happening, I like to have an idea about what is going to be happening.  And it turns out that these same data have a tell about that point.  When the two yields get too close together, or too far apart, that offers us information about what comes next.

China-US 10-year yield spread

Right now we are seeing a comparatively high spread between the two.  That condition typically means U.S. T-Note yields have further to rise.  Conversely, a very low spread, below the lower band, says that U.S. yields should fall in the weeks that follow. 

But that predictive effect really only works on what the U.S. yields are going to do.  It does not necessarily tell us what the Chinese 10-year yields will do.  For that, there is a perhaps even more interesting relationship:

China 10-year yield vs. copper prices

It turns out that the Chinese 10-year yield and copper prices are very closely correlated.  That’s not much of a surprise.  But the fun part is that when the two disagree, it is copper that usually ends up being right about where both are headed. 

That is important because right now, copper prices are trending downward while the Chinese 10-year yield is trending upward.  If copper is correct as usual, then the Chinese 10-year yield should not have much further to trend upward.  Getting the Chinese yield to put in a top should eventually mean a top also for the U.S. 10-year yield, but not for at least a few more weeks.

Tom McClellan
Editor, The McClellan Market Report

 

Continue reading
60 Hits

Tom McClellan

Chart In Focus

High-Yield Bond A-D Line

 

Chart In Focus

April 27, 2017

Junk bonds are the canaries in the stock market’s coal mine. 

If you want to know ahead of time that trouble is coming for the stock market, then one of the best places to look is the high-yield (or junk) bond market.  The movements of prices among these bonds correlates much more closely to the stock market than to T-Bonds.  More importantly, when liquidity gets tight, the junk bonds are the first to be sold by traders wanting to lessen their portfolio risk. 

We can see the importance of this message in this week’s chart, which features A-D data from FINRA TRACE.  For those who like the full spelling of acronyms, that means “Financial INdustry Regulatory Authority Trade Reporting and Compliance Engine”.  FINRA tracks the price changes on a total of 7876 individual bonds, and breaks down the Advance-Decline statistics into categories of Investment Grade, High Yield, and Convertible bonds.  The chart above features the A-D data for the High Yield bonds.

This A-D Line arguably does a better job than the composite NYSE A-D Line at doing what we want an A-D Line to do, which is to show us divergences at important times.  That is the whole reason behind ever looking at breadth data of any type.  We want it to give us an answer which is different from what prices are saying, but only at the right moments. 

A lot of analysts mistakenly assert that if one is interested in the stock market, then one should only look at A-D data from the stock market.  And to take that point further, they assert that one should filter out all of the contaminants such as preferred stocks, rights, warrants, bond closed end funds, and other detritus which together are making the stock market less pure.  I debunked that point in a March 24, 2017 article

Just recently, the overall NYSE A-D Line moved to a new all-time high, saying that liquidity is plentiful and it should lift the overall stock market.  The same message comes from this High Yield Bond A-D Line, which has also pushed ahead to a new all-time high.  The message is that liquidity is so plentiful that even junk bonds can go higher.  And history shows that such plentiful liquidity is also beneficial for the overall stock market.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
71 Hits

Tom McClellan

Chart In Focus

Is Iron Ore Weighing Down Stock Market?

 

Chart In Focus

April 20, 2017

Some U.S. stock market investors are getting worried about the price of iron ore in China.  This week’s chart helps to show why.

One analyst who noticed this relationship was Alastair Williamson of Stock Board Asset, who published this Tweet on April 18, 2017:

StockBoardAsset tweet

It is definitely an intriguing chart, and a relationship I had not explored before.  I have come across a large number of interesting intermarket relationships like this one, and it is always fun to find (or be shown) a new one.  But not all of them have merit.  So what I’d like to do is use this one to show you how I typically like to contemplate a new relationship that I encounter.

So my first question is about whether the relationship is a durable one.  The answer, it turns out in this case, is no.  The two have only recently fallen into this apparent correlation.  Here is a longer term look:

Iron ore prices versus SP500

Prior to early 2015, there did not seem to be any relationship at all.  Since that point, they do seem to be correlating.  But perhaps what we have is more of a leading indication relationship.  Here is that same chart, with the iron ore futures price plot shifted forward:

Iron ore and SP500 forward offset

I played around with different offset periods, and got the best looking alignment of the recent data with a 29 trading day forward offset of the iron ore price data.  I do not have an explanation for that period; it is just what seems to work the best.  And this adjustment still leaves the period before early 2015 showing hardly any correlation at all.

As an aside, I am often asked about whether I have investigated the Pearson’s Correlation Coefficient for a particular relationship.  A lot of technical analysts like to employ that tool, because it is very easy to get a result in Excel or other programs.  But statisticians know that Pearson’s is not a tool that is designed for time series data.  It is better for analyzing attributes of a population.  And there are better tools for time series data, but they are harder to use. 

Pearson’s Correlation Coefficient can also get fooled by trends in the data.  See more about that topic in this article from 2010: Correlations May Not Be What They Seem.

The last part of the answer about correlation coefficients is that I usually do not need a number to tell me whether data are correlated.  It is usually pretty obvious from looking at the chart.  The eye can get fooled, it’s true, and so one should be mindful of that when doing any chart analysis.  But if there is something there in a relationship, or if there is not, it is usually pretty obvious. 

Coming back to the relationship under discussion, this difference in behavior before and after 2015 leaves us with a difficult question: Is this a durable relationship, or just a spurious correlation that the stock market happens to have fallen into just recently?  That matters a lot as we contemplate the recent sharp drop in iron ore prices.  If this is a relationship which can be believed, then that sharp drop says bad things about the future for the stock market. 

To contemplate that point further, here is a chart that zooms in on that 29TD forward offset relationship:

Iron ore prices and SP500 forward offset

The correlation between the SP500 and this forward-shifted plot of iron ore prices which seems to have worked pretty well in 2015 and 2016 seems to now be breaking down.  So on that basis, I am not worried about the stock market repeating the recent plunge in iron ore prices.  But I do plan to keep watching it.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
64 Hits

Tom McClellan

Chart In Focus

Gold Resolves Some Bearish Divergences

 

Chart In Focus

April 15, 2017

A week ago, it was not looking good for the gold bulls.  The dollar price of gold had not yet made a higher high, even though the Japanese yen had already pushed to a higher high.  When divergences like that happen, it is typically bearish news for both gold and the yen.

But what looked like a bearish divergence then has now been resolved in favor of the bullish case.  The price of gold has now joined the yen in making higher highs. 

This is an important point for all chartists to understand: just because you see what looks like a divergence, that does not mean it has to persist.  Divergences do matter, and they deserve our attention, but they can resolve themselves so you have to keep watching and pay attention.

A similar divergence was also showing in the comparison between the dollar price of gold and the price measured in euros.

Gold priced in euros

A week ago, the dollar price of gold seemed to have stalled at a downtrend line, even though the euro price had already broken the equivalent long before.  And the price of gold as measured in euros had not yet made a higher high to confirm the dollar price’s higher high.  That is a problematic sign, and I like to say that whenever the two disagree, it is usually the euro price that ends up being right about where both are headed.  So it was troubling last week when we seemed to have a divergence.

Now that divergence has been made moot by the price of gold in both currencies moving higher.  Remember that all divergences in real time are only potential divergences.  One cannot call them “for sure” divergences until much later.  Apparent divergences are worth noting, but not worth panicking about absent more proof.  They can get resolved, as these examples illustrate.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
59 Hits

Tom McClellan

Chart In Focus

AAII Survey Readings Nearing A Bottom Indication

 

Chart In Focus

April 07, 2017

 

 

The bullish percentage in the AAII survey has been falling during 2017, and is getting close to a low enough reading to mark a good bottoming indication for prices.  That fits well with my expectation from other sources for a low this month.  But before you go interpreting the AAII numbers, it is best to understand some of their quirkiness. 

The raw data come from responses by AAII members on that organization’s web site, and the data are published weekly at http://www.aaii.com/files/surveys/sentiment.xls.  One problem that this survey method creates is that there is not a consistent pool of respondents week to week.  And they are not randomly selected either; the pool consists of people who decide that they want to participate.  That can create some inconsistencies. 

The next point to understand is that price tops usually do not come at the same time as the highest readings for the bullish percentage.  Instead, divergences at price tops are the conditions to look for.  We just had a pretty classic looking divergence at the March top, but one could have called a divergence before then and seen prices continue up into that top. 

So marking a top can be difficult, while very low readings for the bullish percentage are better indicators of price bottoms.  But even there, inconsistencies can arise.  Back in August 2015, we saw the lowest bullish reading in 2 years, and rather than marking a bottom it coincided with a top just ahead of the China-fueled minicrash later that month. 

In a similar way, a high reading for the bearish percentage is usually a great marker for a price bottom.  But they too can be inconsistent.  That same August 2015 instance saw a high percentage of bears, just as prices were topping.  That’s not how one wants a contrary sentiment indication to work.

AAII bearish percentage

We just saw a high reading again for the bearish percentage reported on March 9, the same week that saw the top for the SP500.  That’s not how a sentiment indication is supposed to work.  The crowd is supposed to be wrong more reliably. 

Since that high reading for the bearish percentage, the crowd backed off from it, and just now the data are reestablishing a more proper high bearishness reading that fits better with what prices are doing.  That should work well with the price bottom which is due next week, as discussed in our latest McClellan Market Report

The point to take away is that the AAII numbers can be a useful insight to keep in mind along with other indications.  But they can also bring some inconsistencies that can burn you if you count too much on this single measure of the crowd sentiment.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
70 Hits

Tom McClellan

Chart In Focus

Lumber and Eurodollars: A Curious Intermarket Relationship

 

Chart In Focus

 

Intermarket relationships are fun.  They often reveal surprising relationships, and those relationships are all the more fun when they offer us insights from which we can make money.

This week, we look at the important data from the Commitment of Traders (COT) Report, which I have previously shown have importance for the stock market.  See this, and this, and also this

You might wonder, who cares about COT data for eurodollar futures.  Well, you would, if you knew the great leading indication it gives for the stock market a year later.  But it also has importance in other ways. 

This week’s chart shows that the commercial traders’ net position in eurodollar futures is very well correlated to the movements of lumber futures prices.  I have previously shown how lumber futures prices give a great leading indication for interest rateshousing starts, and other important economic series. 

Lumber prices are unique in that they reflect multiple economic inputs, and so lumber prices sit at the intersection of all of those factors.  “Timber” refers to logs that are standing as trees, or freshly cut.  To turn “timber” into “lumber”, one must add costs of transportation, electricity (for milling), labor, cost of capital, storage, and shipping once again.  So lumber prices reflect more than just the cost of trees, or the demand for 2x4s. 

The value of the Commitment of Traders (COT) Report data for eurodollar futures is something I have written about before concerning the stock market.  In that usage, the eurodollar COT data give a 1-year leading indication for how the SP500 will move. 

But this is the first time I have shown a comparison to lumber prices.  This one has the two plots shown on a coincident basis, i.e. no time offset.  They show a very close correlation, and the importance of that point is when the eurodollar COT data get to an extreme value, that usually marks a turning point for lumber prices. 

That is important right now because we are looking at the largest net long position for the commercials (as a % of total open interest) in the entire 31-year history of the COT Report data.  This suggests there is not much further to go for either that data, or for lumber prices.  And as a stock market leading indication, it says to be looking for a peak about a year from now, but a really fun bullish time leading up to that peak.

A peak for lumber prices now might not seem that important to most people.  Hardly anyone trades lumber futures, and indeed total open interest in lumber futures typically ranges from around 3500 to 6000 contracts.  It is a tiny futures market (eMini SP500 futures have around 2.8 million contracts, for comparison).

But lumber prices are terribly important as an economic indicator.  They tell us about what interest rates and housing data are going to do over the next year.  Here is what I mean:

Lumber versus HGX Index

Lumber versus 2-year T-Note yield

So if lumber is indeed topping out now, then we can look for growth in housing to top out about a year from now, and the same for short term interest rates.  But we don’t know yet that lumber is topping now, nor that the eurodollar COT indicator is going to turn down now.  We just know that both are stretched.  So for that forecast of a year from now, we need to watch how both the eurodollar COT data and lumber prices behave over the next few months.  That will be a topic we cover and update in our twice monthly McClellan Market Report and our Daily Edition.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
66 Hits

Tom McClellan

Chart In Focus

Why Don’t We Use Just Common-Only A-D Numbers?

 

Chart In Focus

March 24, 2017

For as long as there have been Advance-Decline (A-D) data that people have been interested in following, there have been criticisms of that very A-D data for including “the wrong sorts” of issues.  Back in 1962, Joe Granville and Richard Russell both pointed to the big divergence between the NYSE A-D Line and the major averages like the DJIA.  That divergence preceded a 27% decline in the DJIA, so in that moment the A-D Line suddenly became much more interesting to a lot of people.

But critics noted then that the NYSE-listed issues contained utilities and insurance company stocks which were “interest rate sensitive”, and which were supposedly contaminating the data.  The same criticism persists today, but now it is leveled against preferred stocks, bond related closed end funds (CEFs), and other issues that trade like stocks on the NYSE.  Those other issues make up about 40% of the issues, although they trade only a small fraction of the volume.

Many analysts assert that one should preferentially follow the A-D numbers for “common stocks”, sometimes referred to as “operating companies only”.  This distinction supposedly filters out those damned contaminants.  So everyone would supposedly be happier and better off if they just followed the right sorts of data, and ignored those “others”.

The problem is that the purified A-D data are actually not always better.  The chart above shows that when the Common Only A-D Line disagrees with the NYSE Composite Index, it is usually the price index that ends up being right.  This is a big problem.  The whole reason for hiring an A-D Line to work for you is to give you a different answer from what the price indices are saying.  Having an index give you the same message that prices give you is useless.  And having data that give you the wrong answer at a pivotal time is worse than useless. 

The one type of issue on the NYSE that is most often blamed for contaminating the A-D data is the closed end bond funds.  They only make up about 7% of the listed issues, and hardly trade any of the volume, but they are continually trotted out as the “usual suspects” for messing up the A-D data.

This is problematic both from the standpoint of raw prejudice, and more importantly because it is just not true.  These issues tend to be the better canaries in the coal mine, warning of trouble ahead of when such warnings come from the common-only A-D data or other indications. 

This next chart shows an A-D Line for the Bond CEFs, those supposed contaminants of the A-D data.  The point is that this Bond CEF A-D Line often tends to be a better indicator of liquidity for the stock market than the data for the actual stocks.  This point is counterintuitive, we know, but that is what the data show.  So what are you going to believe: your preconceived notions, or the actual data?

Bond CEF A-D Line

What about just following the A-D data for the components of a particular index?  This one fails the usefulness test as well.  Here is a chart of the DJIA compared to an A-D Line made up of just the 30 Dow components:

DJIA Stocks' A-D Line

At the left side of the chart, the DJIA topped out on Oct. 9, 2007, but its A-D Line kept making higher highs until Dec. 10, 2007.  Not only was there no bearish divergence, the A-D Line was wrongly saying things were fine as a bear market was starting.  And the same thing happened again in 2011 at the right end of the chart. 

The whole point of watching A-D data is to get a different and better answer about what the future holds.  We want to see divergences, especially at price tops, to tell us that trouble is coming.  But when the DJIA itself makes a lower high and it’s A-D Line makes a higher high, that is a message one is better off not listening to.  Instead, one should turn to better “experts”, who know what lies ahead. 

I once believed, as others still do, that it was better to purify the A-D data, to get the “better” answer.  That is why I went to the trouble of writing all the code (with help from others, thanks R.N. and crew) in order to gather and arrange these data, in pursuit of those supposedly better answers.  But then I tested that hypothesis, and I found it wanting.  So if one is going to follow the scientific method, a reevaluation of the hypothesis is required whenever the data do not support one’s hypothesis. 

It turns out that the parts of the A-D data which so many people criticize for being “impure” actually seem to have the better answers a lot of the time.  It is as though the chaff may be more valuable than the wheat.   Think of this next time you hear someone tell you that you should only use data for the common stocks.  That is a prejudiced viewpoint, stemming from assumtions rather than examinations, and which the data do not support.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
54 Hits

Tom McClellan

Chart In Focus

Huge Imbalance in Crude Oil Positions

 

Chart In Focus

March 10, 2017

There is a giant wall of short positions held by the smart-money “commercial” traders in crude oil futures, and it is going to lead oil prices to come crashing down. 

Each week, the CFTC reports on the numbers of long and short positions held by futures traders.  They are broken down into 3 separate groups:

Commercials – Those engaged in the business related to that commodity.  They are the big money, and thus presumably the smart money.  Think Cargill for grains, or Goldman Sachs for financial futures.

Non-Commercials – Large speculators.  Think hedge funds.

Non-Reportables – Those whose positions are too small in number for the CFTC to bother tabulating them individually. 

This week shows the commercial traders net short position, expressed in numbers of contracts.  They just reached an all-time (since 1986) record for the number of contracts that they are net short, i.e. short positions minus longs.  Every futures contract is simultaneously a long and a short position, with the two sides of that contract held by different parties.  The short side is the one that has to deliver the product, and the long side wants to take delivery, or at least that’s the design.  Speculators also play in the futures markets, never intending to take or make delivery. 

In the crude oil market, the commercial traders are often the producers, using futures markets for their original intended purpose, which is to be able to lock in prices now for sales of future production.  So the direct message of seeing the commercial traders reach an all-time net short position is that the smart producers think that recent prices have been a great deal to lock in for their future production.  If oil prices were going to rise, then locking in now would not be a great deal.  But if oil prices are about to fall, then smart traders would want to lock in prices before that happens.  This seems to explain what we have just seen.  And understand that the speculators, large and small, have taken the opposite site of that big imbalance in positions.

The last time that the commercial traders came even close to this big of a net short position was back in June 2014, just before crude oil prices were cut in half.  I cannot forecast the magnitude of the coming move down, but the message here is that it should be substantial.

The notion of a down move for oil prices is confirmed by the spread between near and far month futures prices, known as “contango”.

Crude oil contango

That spread recently narrowed to its smallest point in over 2 years, a sign that near month crude oil prices were topping.  When there is a big contango, then a trader could buy cheap oil in the spot market, rent a place to store it, and sell a more expensive futures contract.  Then all he has to do is wait a few months to deliver at that higher price.  As long as the spread is big enough to cover the cost of storage, then there is a lot of money to be made.

But when contango gets really small, up close to zero, then the cost of storage eats up all of the profit margin from that game, and traders look to dump that supply on the market and stop paying rent.  This is the sort of condition we are entering now, with a glut of oil coming onto the spot market. 

The point is that prices for crude oil are likely to fall for a while, and we will probably know that they are done falling only when we see the commercial traders covering their shorts in a big way.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
65 Hits

Tom McClellan

Chart In Focus

Higher Drug Related Fatal Crashes After Marijuana Legalization

 

Chart In Focus

March 03, 2017

I don’t smoke pot, and I generally don’t like hanging around with people who do.  This applies especially to those who are driving in the lane beside me, doing 70 MPH on the Interstate. 

At the same time, I don’t really appreciate having the government tell me what I can or cannot put into my body.  Tell that other fellow, but don’t tell me. 

I live in Washington State, which since the Nov. 2012 election has been undertaking a great social experiment thanks to a popular referendum which legalized recreational (vs. medicinal) marijuana use.  This week, I found myself wondering what the effects of that change have been on vehicle accidents. 

Thankfully, the Washington Traffic Safety Commission (WTSC) has been compiling such data since 2008, and they are the source for the data in this week’s chart.  I have not yet checked for data from Colorado, or other states in the process of legalizing such marijuana use.  So that is a problem for analyzing this big social experiment.  So is the small number of years since the change in law, which makes for sample size issues.  These are admittedly imperfect data (as all data are).

With those caveats placed on the table, let’s take a look at what these limited data do say. 

Since 2012, fatal crashes involving drunk (alcohol) drivers have continued to trend gently downward.  It is not a fast enough downtrend to suit me, but it is a useful baseline for comparison, especially in comparison to other types of intoxicants. 

Also since 2012, fatal crashes involving stoned drivers have been trending upward, breaking a downtrend which preceded the change in the law.  The data for “drugs” includes the data for “marijuana”, and the chart makes it pretty clear that the new upward trend since 2012 is the result of higher marijuana-related crashes.

Let me circle back at this point and note that we are still talking about relatively small changes in total numbers.  And I should further note that it feels rather morbid to talk dryly about car accidents in which real people died, and to use statistics to point fingers.  I get both points about how creepy and how statistically dubious this is, given the limited data. 

And I want to further stress that these stats are only for crashes that were fatal.  The Washington Traffic Safety Commission does not offer statistics on marijuana involvement in crashes which are less than fatal.  That’s a problem with this analysis, and not one I can remedy with the available data.  We also cannot adequately rule out other factors, such as changes in population density, changes in law enforcement priorities, effects of changing road conditions, etc.

One technique that we can employ with the limited data is to compare the rise in marijuana and overall drug crashes with those due to other causes, or as Claude Rains would have said in Casablanca, “the usual suspects”.  So here in this next chart, I present some of the other factors included in that data series from the WTSC:

Speeding, young driver, distracted crashes in WA 2008-16

The fatal crashes associated with speeding and with young drivers have continued to generally trend downward (hurray!!!).  The “distracted” crashes have seen an upward movement since the 2012 change in marijuana laws, but that may be reasonably explained by the increased penetration of smart phones into society. 

The point of this comparison is to provide a background of other “usual suspects” to compare against the observed change of trend in the “marijuana” and “all drugs” data.  If the uptick in those series were part of some background change in circumstances, such as larger numbers of drivers being on the road, or greater general road safety problems, then we should expect to see the uptick in crashes appear across all types of crashes.  But when we see a general continuing downtrend in crash statistics except for “distracted” and “drugs”, then it is reasonable to conclude that the change in drug use law has resulted in a change in accident rate outcomes. 

Now go forth and draw your own conclusions.  But don’t bother to pass me that doobie.  And put down your phone when driving, please, especially when you are passing me.

Tom McClellan
Editor, The McClellan Market Report

Continue reading
90 Hits

Recent Replies on Forums

Conclave 2016 Size chart