Years ago, an analyst named Richard Arms wondered what it meant when the stocks going up (or down) traded a greater share of the total volume than the other side.  So he created a ratio which now bears his name, the Arms Index, which looks at the A/D ratio versus the Up/Down Volume ratio.  The raw formula is as follows:

Advancing Issues / Declining Issues

-----------------------------------------

       Up Volume / Down Volume

It is neutral at 1.0, and readings below about 0.5 or above 2.0 show an overbought or oversold condition, respectively. 

I just spent the past weekend with Dick Arms at a conference of professional technical analysts.  He is very much alive and continuing to ply his craft as a technical analyst, even inventing new techniques that he spoke about to the group, including his work with Equivolume Charts.  He is a friendly and gracious man in person.  At my urging he also told the assembled group the story of how the Arms Index came about, including its name.  You can read more about Dick's work at http://www.armsinsider.com/. 

Dick first wrote about this indicator back in the 1960s, and had an article on it published in Barron's at the time.  It was picked up by many analysts as an interesting indicator, and listed as an index on quote systems like Reuters, Quotron, and Telerate.  Back then it was just referred to as a TRading INdex, and thus it was listed under the symbol "TRIN" which is still used on many quote systems to this day.

Before CNBC, there was a cable news TV network known as FNN which was based in Los Angeles.  John Bollinger was the Chief Market Analyst there, and he thought it was just wrong that Dick Arms' own indicator did not get listed bearing his name.  So Bollinger lobbied to have it listed instead on the FNN ticker as the "Arms Index" and the FNN management agreed.  Among well-educated technicians, it is known to this day as the Arms Index.

Many years ago, Peter Eliades of www.stockmarketcycles.com originated an innovation to the Arms Index calculation.  The daily values of the Arms Index are somewhat noisy, and so some smoothing can be helpful.  Such smoothing is often done as a simple or exponential moving average of daily closing Arms Index values.  But Peter had a different idea for how to smooth it, which he titled the "10-Day Open Arms" Index.  The formula looks at each of the 4 breadth items in the original formula separately, summing them over the past 10 trading days, and then re-using the original formula.  It looks like this:

             Last 10 day's Advances/Last 10 day's Declines

------------------------------------------------------------------------

 Last 10 day's Advancing Volume/Last 10 day's Declining Volume

It is still a 10-day smoothing, similar to a 10-day moving average, but any overweighting of a single day's value is combined with the other days in a slightly different way.  Each component's inputs from the past 10 days are summed together before being compared to the other components' 10-day totals.  It was an interesting twist on how to smooth a noisy indicator. 

And right now, that interesting twist is showing a pretty big overbought condition.  The chart above shows that the 10-day Open Arms Index is at a fairly high level on this inverted scaling, and such readings are reliably associated with meaningful tops for stock prices.  I have inverted the scaling for the 10-day Open Arms Index in this chart, to better compare extreme readings with the tops and bottoms for stock prices. 

When a high chart reading like the current one (i.e. a low raw reading) appears after a meaningful decline, it can sometimes be a signal of strong upward initiation for a new uptrend.  But when it appears late in an uptrend as it is doing now, it is more often a sign of conclusion for the advance. 

Taking this investigation a bit further, I had wondered a few years ago if looking at the Arms Index or the 10-day Open Arms Index on just "common stocks" would make for a better indicator than the composite version which looks at all issues.  A lot of technicians take the view that issues which are not real operating companies end up contaminating the data, and so they should be discarded.  I was already calculating the A-D and UV-DV data for common only issues myself, and so it was a matter of just adding a couple more columns to the spreadsheet. 

But rather than making for a better indicator, the results were somewhat disappointing.  Here is the Open 10-day Arms Index for just common stocks, once again inverted for better correlation to price action:

It shows that sometimes there are good high and low readings which correspond to price tops and bottoms, but at other times it seems really awful.  Its performance as an indicator is just not good enough to make me want to watch it and believe in its message. 

So what are we to conclude from this?  Dick Arms hit upon a great indicator when looking at all of the issues traded on the NYSE, and Peter Eliades made a nice addition to that work with his innovation.  Arms' intention was to look at the difference between what the advancing and declining stocks were doing in terms of how much volume each traded.  But it seems to me that this is not where the magic of his indicator lies.  Rather, the magic is the story that the Arms Index  tells about the difference in behavior, either advancing or declining, among the stocks which trade most of the volume versus the other issues which don't trade much volume.

The "common only" issues make up just 59% of the total NYSE issues by my count, and yet that group regularly trades more than 90% of the share volume.  Taking away the preferred stocks, bond funds, rights, warrants, and other issues that make up the other 41% of issues does not actually make for a better version of the Arms Index.  Instead, it degrades the useful message which comes in the composite version. 

Several months ago, I introduced the chart comparing the stock price pattern of Apple Corp (AAPL) in the current day to that of RCA in the late 1920s.  I later also drew a comparison to Microsoft (MSFT) before and after its top with the Internet bubble in 2000.  That latter one merits an update. 

Looking at AAPL's price pattern alone, the situation does not look all that bad.  It has broken out above the declining tops line which was in effect from the top last year, and then the price went on to drop back down to test the top of that broken downtrend line.  And since that test, there has been a pretty nice rally along with the rest of the stock market.  Maybe AAPL is finally getting back into sync with the seemingly endless bull market. 

Or maybe AAPL is just continuing to match the pattern laid down by MSFT 13 years ago.  The chart shows a failing pop up in MSFT at this equivalent point in the pattern analog, which was followed by another drop to even lower lows.  The two pops do not match each other perfectly, and indeed most of the chart shows imperfections in the way that AAPL is dancing out the same dance steps.  There is some of an effect I call "accordioning" of the bottoms and tops, some coming a bit earlier or later than MSFT's pattern said should have been the case.  So slight differences in timing are just a normal part of this relationship. 

We would know that AAPL is finally doing what all price pattern analogs do, eventually breaking the correlation to the prior pattern, if we see AAPL's share price make a more drastic deviation from the MSFT's prior pattern.  We have not seen that yet, and so the analog's continuation still gets the benefit of the doubt.

None of this is to say that Apple the company is the same now as Microsoft the company during the Internet bubble.  They are quite different.  But what is the same is the way that the investing public falls into and out of love with the tech darling du jour.  That investor behavior is what explains the chart pattern similarity, and that behavior is what technical analysts strive to take advantage of. 

Three weeks ago, I wrote about the indication of a liquidity problem which was being given by closed end bond funds.  Those are among the most liquidity-sensitive issues, and so when they are doing something different from the major market averages, it is a sign of potential liquidity problems which could come around and bite the rest of the market.

We did see the fulfillment of that prophecy with a halfway decent selloff for the major averages in mid-April.  But now the message from high-yield bonds and closed end bond funds is that liquidity has been restored, at least for the moment.  While the SP500 has made a pattern of lower lows and (thus far) lower highs, high-yield bond funds are defying that pattern and showing strength.

There is no perfect market index which captures the behavior of the high-yield bond universe.  I like to use HIO as a good proxy for that whole segment of the market.  It does a pretty good job of showing weakness ahead of liquidity problems which can bite into the overall market, and it also shows strength ahead of big rebounds in the SP500.  Right now, it is showing a pattern of higher lows and higher highs, and looking much stronger than the stock market. 

The chart I showed here 3 weeks ago was an A-D Line of closed end bond funds which are traded like stocks on the NYSE.  They are also quite liquidity sensitive, and so in the very rare instance when this A-D Line makes a lower high while prices make a higher high, it is rightly seen as a problematic message for the overall market.

That message of trouble for the market has now been resolved by seeing this A-D Line surge ahead to a higher high:

Whereas it was showing trouble for the market a month ago, now it is conveying a big "Never mind".  The higher high says that the previous liquidity problems have been resolved, at least for the moment.  This is a message which can change back again at any time, but for now the weaklings who are very liquidity-sensitive say that liquidity has been restored. 

Since bottoming in October 2012, inventory levels of copper have risen 190% in warehouses operated by the London Metals Exchange.  That's a huge and rapid increase, and it conveys a powerful message about the future for copper prices. 

Back in September 2012, spot copper prices topped out at $3.81/pound, and they have now fallen 18%.  In terms of big drops in copper prices, this one does not rank very high among the big drops in copper prices over the past few years.  But it is producing a huge and rapid rise in copper inventories.

It is normal for prices and inventory levels to generally move in opposite directions.  When copper producers don't like the market price and think that they can get a better one by waiting, they put their production into warehouse storage and wait for better times.  When prices rise up to or above a price level that the producers like, copper starts coming back out of inventory and onto the market.  So watching copper inventory levels can give us insights about where the producers think a fair price is. 

It was understandable that copper inventories would rise back in 2008, when the economy was grinding to a halt, and when copper prices plummetted from above $4/pound in July 2008 to $1.25/pound in December 2008.  And shortly after copper bottomed at the end of December 2008, copper inventory levels started coming back down again.

Now we are seeing an even more rapid rise in inventory levels, and it comes on just a small amount of drop in copper prices.  The first message to take from this is that copper producers don't think that $3.60 is a fair price.

That's where copper was hovering just as the big run up started in inventory levels.  The inventory rise makes a pretty emphatic statement that the producers think they can get a better price by waiting.

This does not mean that they have to be right.  But producers spend their time dealing with copper prices, figuring out how much to produce and when to sell.  So they are in perhaps a better position than some others are to know what a fair price is, and so the opinion that they are conveying with their inventory behavior is at least worth listening to. 

One other "smart money" group is the commercial copper futures traders, as tracked in the weekly Commitment of Traders (COT) Report published every Friday by the CFTC. 

As copper prices have been falling during 2013, the big money commercial traders have been trading the net short position that they held as a group back in January, and now they are at a decent sized net long position.  It is not yet as extreme of a big net long position as we saw at the major price lows back in January 2007 and December 2008, but they are getting there.

The upshot from both of these charts is that as copper prices fall, the seeds of an eventual bottom and big price rebound are being sown.  Neither chart is at a point where we could say that a price bottom is imminent.  But the rapid nature of the rise in inventories says it should be sometime this year rather than next year. 

The April 15 income tax filing deadline is just a few days away, so I thought it would be appropriate to take a look at what tax collections have been doing lately.  With the federal tax rate increases for 2013 now taking effect, it is not too surprising to see total federal tax receipts rising.  The top individual income tax rate rose back up to 39.6%, and the individual Social Security payroll contribution (AKA FICA) returned to 6.2% after being at 4.2% for 2011 and 2012.  That latter one means that most wage earners are seeing 2 additional percentage points of their wages claimed by Uncle Sam.  2% does not sound like much, but remember that a jump from paying 4.2% to 6.2% is actually a 48% increase. 

For the 12 months ending in March 2013, total federal receipts from all sources amounted to 16.2% of GDP.  That's well short of the 18% collection rate which guarantees a recession, but it is rising sharply from recent values.  The fact that it is still well short of federal spending at 21.9% of GDP is important, but that's a subject for another time. 

Why do I say that collecting taxes at a rate of 18% of GDP guarantees a recession?  I only make that assertion because the U.S. has ended up in a recession every single time in the past that it has gotten that high.  Some in Washington, DC think that it is different now, and that somehow we can avoid a recession with higher tax collections this time.  The 70 years of data which say otherwise do not seem to bother those people. 

So if we accept that 18% of GDP is a recession threshold, and the last 12 months show it only at 16.2%, then there's nothing to worry about, right?  Not so fast.  The data on federal receipts for the first 3 months of 2013 show a big jump in gross tax collections (unadjusted by GDP):

Total dollars collected for the first calendar quarter of 2013 equal $581 billion.  That's still less than the $888 billion that Uncle Sam spent from January to March, but the numbers are getting closer to each other.  That's good news for the deficit, but not such good news for the economy.  Taking too much money out of the economy in the form of taxes stifles private economic activity.  Those dollars don't get to do other things when Uncle Sam snatches them up. 

In 2012, the first 3 months saw $479 billion in total federal receipts.  So 2013 is already running 21% higher than the rate of gross federal receipts in 2012.  There is a huge seasonality factor in when those dollars get sent to Washington, DC, as the second chart shows, but so far 2013 appears to be following that seasonal pattern quite well, with a big overall bump in total collections.  April is understandably the biggest month, as people wait until the tax filing deadline to pay what they might owe for the prior year plus estimated taxes for the current year. 

Extrapolating forward, if the rest of 2013 continues to see federal tax receipts come in at a rate 21% higher than last year, and if GDP growth remains tepid, then trailing 12-month federal receipts will be up above that 18% recession threshold before the end of the year. 

It may be worth noting that during the late 1990s, federal tax receipts were able to stay up above 18% for quite a while before the recession was finally declared in late 2000.  That was an exceptional time, when the Baby Boom generation was still fully employed and contributing to GDP, and when a technology bubble and deregulation had companies hiring and expanding.  We don't have those conditions now, with Boomers now retiring, and with federal regulations increasing dramatically.  The high taxes back then did have their effect, though, causing the NYSE A-D Line to peak in 1998, killing the tech boom, and causing a change in 2000 for control of the White House. 

Some people already understand that setting federal tax collections too high has this destructive power.  Others have to be reminded about that every few years, a process which is now getting under way.

Shortly after the Flash Crash in 2010, I wrote a Chart In Focus article explaining that the problems that the stock market had just been through in the wake of the Flash Crash were not liquidity problems.  I cited as evidence then that we were still seeing strength in the Advance-Decline (A-D) Line for bond-related closed end funds (CEFs) traded on the NYSE.  We are seeing a different situation now, which is worthy of mention, but first we should review the terms I am using.

Only about 60% of the issues traded on the NYSE are real "common" stocks, things we would think of as the shares of real operating companies.  But they generally trade about 90-95% of the share volume every day.

The other types of NYSE issues are made up of preferred stocks (18%), foreign issues (11%), specialty issues (13%, including rights, warrants, and structured products), and bond CEFs (8%).  Of these, bond CEFs are the ones most often blamed for "contaminating" the A-D data of the overall NYSE, and so a lot of technicians prefer to use some purified version of A-D data to avoid that contamination.

I take the opposite view.  If anything, the bond CEFs improve the composite A-D data by virtue of their presence.  The reason for this is that the main virtue of A-D data is as an indicator of market liquidity.  It is possible to push the major large-cap indices higher during an illiquid period if the limited liquidity is channeled into the "right" stocks.  But to get the large majority of stocks to go up, there needs to be such prolific liquidity that there is enough to go around for everybody. 

When liquidity starts to dry up, the least deserving issues tend to get culled first from the herd.  That is what we are seeing now in the bond CEF A-D Line.  These liquidity-sensitive issues have already turned downward as a group, even though the SP500 was able to continue higher.  This is a message that there are now liquidity problems facing the market, even though the Fed is continuing to drop money from helicopters every month.  Perhaps $85 billion a month is just not enough. 

The 2007 example in this next chart helps us to understand why this warning sign is so important.  The bond CEF A-D Line peaked back in May 2007, even ahead of the overall A-D Line which peaked in June 2007, and ahead  of the final market price high in October 2007. 

The larger point is that these issues are more liquidity sensitive than others.  So if we see them suffering as a group, then the message is there is a liquidity problem which will likely come around to bite the rest of the market.  And when the market sees a meaningful dip for the SP500 without the bond CEFs getting hurt, the message is that it is likely a problem other than liquidity, e.g. geopolitics, investor mood, etc.  Those are easier problems for the market to get through than liquidity problems, which take longer to solve.

There is one notable anomaly in the middle of the chart which is worthy of mention.  Back in 2010, the impending end of the "Build America" bond program caused a lot of municipalities to rush some muni-bond issuances to market and take advantage of the free money from Uncle Sam.  That glut of new issuances then was more than the bond market could easily absorb, and so a lot of muni-bond prices fell as a result which took down the prices of some of the bond CEFs.  So that was an example of when the bond CEF A-D Line was "wrong" about an overall market liquidity problem.  We don't face a similar anomaly now.

It is hard to get one's mind around the idea that the stock market could be facing a liquidity problem when the Fed is throwing $85 billion a month at the banking system.  But that is the message here, from examining the actual behavior of those issues who are most sensitive to it. 

Under the standard valuation model for the stock market, earnings are all-important.  They drive valuation measures like the Price/Earnings (P/E) Ratio, and help fundamental analysts figure out how much a stock is worth.  Many analysts extrapolate that forward to the whole stock market, which can be a problematic exercise. 

As we near the end of March, and thus of the first quarter of 2013, the earnings reports are still trickling in for the 4th quarter of 2012.  The Commerce Department's Bureau of Economic Analysis (BEA) still has not even published Q4 figures for corporate profits.  That is an important point to note in light of this week's chart, which compares the movements of the DJIA to the BEA's measure of corporate profits as a percentage of GDP.  To access the data in this chart, see http://www.bea.gov/iTable/iTable.cfm?ReqID=9&step=1, and then proceed to Table 1.10, Gross Domestic Income by Type of Income.  Line 17 has corporate profits, and then you just need to divide that by quarterly GDP data which can be found at http://www.bea.gov/national/xls/gdplev.xls. 

The typical argument of a fundamental market analyst is that profits drive everything, and so waiting for word of what profits have done is an appropriate action before making decisions about how, when, and if to invest in the stock market.  But there are multiple problems with this premise.  The first is the presumption that earnings are predictive of future stock market action.  It is perhaps easier to provide justification for the opposite argument, i.e. that stock market movements are predictive of earnings data. 

This measure of profits as a % of GDP peaked in the 2nd quarter of 2007, which was ahead of the final stock market top in October 2007.  But we did not know that corporate profits had peaked in the 2nd quarter until late in the 4th quarter, when the BEA's data finally showed that corporate profits had turned down.  And given the see-saw nature of that quarterly data, it would have been easy to dismiss one quarter's downturn, and not believe that a real problem existed for earnings until well into 2008.  By then, the DJIA was already in free-fall.

You see, the 2007 example illustrates how the lag in earnings reporting creates a big problem.  The fundamental analysts can be right about earnings being important for the future of stock prices if we examine them only in retrospect, where they can be compared side-by-side with the market action of the day.  But when we factor in the reporting lag for such data, we find that waiting for the earnings message can mean getting data which is already almost 6 months in arrears by the time that it arrives for us to consider. 

One other myth that we hear frequently is that corporate profits now are higher than at any time in history.  But this week's chart reveals that this myth is just not true.  They are admittedly high, especially in comparison to the readings for the past 40 years.  But from the beginning of the BEA's data in 1947 up through 1966, corporate profits averaged 10.3% of GDP, which is actually higher than the most recent reading of 9.8% for Q3 of 2012.  So one could simultaneously argue that the most recent level is "high" compare to recent years, and also argue that it is still low compared to historic levels.  Which argument you hear from someone will tell you what that speaker's bias is. 

One point that we can say history does support is that when corporate profits as a percentage of GDP get up this high, the Federal Reserve usually steps in with tightening monetary policy, acting like the overspeed governor on an engine by retarding the throttle to keep the RPMs from getting too high.  Thus far, the current group of Fed officials have not indicated in their recent comments that they see this current situation as being similar to past instances when tighter policy came about.  But if corporate profits remain at a high percentage of GDP, then the Fed's decision makers are going to have a harder time continuing to justify the free-money policy.

It should not be surprising that a chart of gold prices measured in dollars should look a lot like gold prices in some other currency.  After all, gold prices tend to jump around more than the currencies do.  But there can be subtle differences in the chart patterns when priced in different currencies, and there can be information in those subtle differences.

This week's chart compares the dollar price of spot gold to the price measured in euros, a topic I featured here back in 2010.  The key insight which every gold trader or investor should understand is that most of the time these two will move together, but when they disagree, it is usually the euro price that tells the truer story about where both are headed.

Most often, this disagreement will take the form of a divergent top or bottom structure.  Just recently, we saw the dollar price of gold make a flat bottom, but the euro price of gold made a higher low, which disagreed with the dollar price's flat bottom.  The disagreement can also take the form of different behavior relative to good old fashioned trend lines.

The dollar price of gold is still operating underneath a declining tops line.  So based on that one single criterion, gold is still in a downtrend.  But the equivalent line drawn across the same points on the euro price plot has already been broken.  The euro price plot has already announced that the downtrend is over, and now it is already making a higher high.

We saw a similar type of trend line disagreement back in May 2012, shown at the left end of the chart.  The euro price of gold broke its downtrend line ahead of the breaking of the equivalent line on the dollar price chart, and we saw in subsequent weeks that the euro price was correct then about where both plots were headed.

We don't get these sorts of divergences at every important turn.  So the lack of a divergence is not necessarily a sign that the existing trend is going to continue.  But when we do see these divergences, they are worth paying attention to, and the euro price is usually right about where gold is headed.

Wholly apart from the technical lesson about the behavior of gold prices, there are a lot of people in the world this week who just saw an increase in the marginal utility of gold bullion as a transportable way to store one's wealth.  When depositors in Cypriot banks were confronted with the possibility that their bank deposits might have a levy of 7-10% assessed on them, plus the reality that they cannot access those deposits while the banks are all shut down, then suddenly gold under the mattress starts to look a whole lot better.  Gold pays zero interest, but that's still better than -7%. 

And while governments can still take action to seize gold holdings, it is somewhat harder to do than to seize bank holdings.  Thus, for a lot of people, gold this week suddenly became a more useful and thus more valuable item. 

One of the more interesting intermarket relationships these days is the strong correlation we see between crude oil prices and both the Canadian dollar and Mexican peso.  That correlation makes some sense, since each of those countries is a major oil exporter, and so the fortunes if their economies and thus their currencies are linked to what the price of oil does. 

Right now, it is even more interesting, because of the mixed message we are seeing in the chart above.  Most of the time, all three lines in this chart dance together, like birds flying in formation.  But if you look closely at the most recent data, there is a big disagreement between the Mexican peso and the Canadian dollar about where the group is headed.  The peso has swooped up to a multi-month high, while the Canadian dollar has turned downward and appears poised to drop to a multi-year low.

For its part, the price of crude oil seems to be taking more of a middle of the road path, not wanting to go along with either currency, but rather it is seemingly taking a compromise position in the middle of its price range of the past 2 years.

So with such a disagreement, who is right?  Unfortunately, history shows that neither currency has a better track record of being "right" when they disagree.  Unlike the case I wrote about last week, where GE's share price seems to know better than the Dow where both are headed, it does not work that way for these two currencies.  One can find plenty of instances in history to justify either argument about which one tends to be "right". 

Making this situation all the more interesting now is the situation we see in the Commitment of Traders (COT) Report data.  For years, I have used the COT data on both the Canadian dollar and the Mexican peso as surrogate sentiment indicators for crude oil.  The COT data for crude oil futures themselves can sometimes convey confusing messages, so having other sources to turn to is a useful way to get perhaps cleaner insights. 

From each currency's perspective on its own, the COT data make sense.  With the peso up at a new multi-month high, commercial traders are holding onto a big net short position.  And with the Canadian dollar falling to a multi-month low, commercial traders are betting that the Canadian dollar will rebound.  They are now net long the Canadian dollar futures to an even greater degree than what we saw when both oil prices and the Canadian dollar put in price lows back in 2011.

So if the relationship of oil prices to these currencies is going to continue being well correlated in the future as it has been in the past, then the two currencies cannot both be right about where they are headed.  Something is going to have to break, and somebody is likely to be proven "wrong". 

Even looking at the "cross rate" between the Peso and the Canadian dollar does not help us much about figuring out where oil is going.

We can see when looking at that cross rate that its turning points sometimes mark tops for oil prices, sometimes bottoms, and sometimes neither.   Sometimes the charts just don't give us the answers that we want them to give.  But if we are open minded, then perhaps we can find answers to questions we were not even asking. 

Perhaps the best point to take from this strange and divergent behavior between these two currencies is that perhaps we are nearing a point when it ought to reverse.  That would mean the steep swoop upward in that cross rate could turn around, and history shows that such a reversal could be a sharp one.  To take advantage of that potential reversal, a currency trader would have to short the peso and go long the Canadian dollar, but that is a sophisticated trade that is difficult to put on and to manage in case the position suddenly goes against you.  I never recommend any trades to anyone, because I have no way of evaluating what a person's risk tolerance and trade suitability may be.  But some traders may be set up to handle it, and to make that decision for themselves. 

As for oil itself, that potential currency turmoil could mark the start of a move that takes oil up or down out of its sideways structure.  So we can prepare for a trending move, and to plan on figuring out as it happens which direction that trend is going. 

The financial press have gotten all excited about the new all-time high in the DJIA.  But a fascinating non-confirmation in this instance comes from the share price of General Electric (GE). 

My late friend Larry Katz introduced me to the phenomenon of GE as a bellwether for the Dow several years ago.  He noticed that the two were strongly correlated, which is not surprising since GE is a major industrial and financial company and is a component of the DJIA.  But Larry noticed that the correlation is not nearly as interesting as is the fact that when the two disagree it is usually GE that tells the truer story. 

But that statement is not always true.  At the end of 2012, the DJIA was leaping ahead to higher highs while GE was refusing to confirm.  That divergence said that the uptrend was problematic, and that problematic message lasted until mid-January when GE evidently decided that it should go ahead and get on board with the uptrend.  Nothing works perfectly every time.

Spring forward to March 2013, and we see that GE has not yet exceeded its Feb. 19, 2013 closing high even though the DJIA has risen almost 300 points higher.  Is this another failure like late 2012, or is it instead a more standard legitimate sign of a bearish divergence like most of the rest of the instances in the history of these two price series? 

One other price series which is showing a similar divergence now is the price of copper.  Before 2007, copper was properly known as the metal with a PhD in economics because of its strong correlation to bond yields and GDP.  Since 2007, however, copper has acted much more like a financial asset, with the fascinating property of making divergent top conditions versus the SP500 at interesting places.  We are seeing just such a divergence now. 

By itself, this divergence between copper and the SP500 could be set aside as an anomaly.  But when we combine it with the divergence between GE and the DJIA, it becomes much more of an attention-getting situation, calling attention to the problematic nature of the current multi-year high in the major averages.  When these situations appear, there is usually a scary-feeling dip which ensues, and which helps cure the problematic price anomaly.  That remains on the agenda for March.

Since its creation in 1999, the movements of the euro currency versus the dollar have shown an increasingly positive correlation to the movements of the U.S. stock market.  Exactly why it works this way is interesting to ponder but not particularly relevant for analysis.  The fact of the correlation carries more importance than any reasons why it should or should not work that way. 

For doing comparisons of the euro to the SP500, I like to go one step farther, and look at the euro/yen cross rate.  This is the exchange rate which would apply if you wanted to exchange some of your euros for Japanese yen, or vice versa.  Doing this seems to produce an even better correlation, as it incorporates the yen's recent role as the beneficiary of the so-called "risk off" trade in addition to the euro's role as "risk on" target. 

The fun part of this relationship does not lie in the finding that there is a good correlation most of the time.  Instead, the real utility comes when we see a disagreement between the euro/yen cross rate and the SP500.  When divergences appear, it is usually the euro/yen cross rate that knows the true story about where both are headed. 

Just recently, we saw a peak in the euro/yen cross rate back on Feb. 5, and it has now broken its 3-month rising bottoms line which is a pretty clear indication of trend change.  But the SP500 had carried on for a while as if nothing was wrong, and only recently started running into some turbulence. 

That divergence, with the euro/yen cross making a lower high and the SP500 making a higher high, was a big sign of the trouble that is just now coming around to the stock market.  And even though the SP500 was knocked down pretty hard in February, it appears to be trying to get back up off the mat.

The euro/yen cross rate does not offer any hope that the SP500's recovery effort will be successful.  The stock market might still ignore that message, as there is no guarantee that the euro/yen cross rate has to be "right" this time.  But there is considerable historical evidence that when they disagree, it is more often the euro/yen cross rate that knows the real story.

For this week's chart, I wanted to share a bit of analysis that I have been writing a lot about in our Daily Edition.  Subscribers to that publication have been watching this situation in gold prices unfold in real time, and now that the price objective has been met it becomes a really nice lesson about one element of classical bar chart analysis. 

Gold prices reached a peak back in October 2012, and then started a decline which has now taken off more than $200 per ounce from its price high.  Partway down, gold prices reversed in November to form what would become a (mostly) symmetrical triangle.  Triangles are indecision patterns, and I say that it was mostly symmetrical to mean that the slope of the declining tops line was roughly the symmetrical inverse of the slope of the rising bottoms line.

Triangles like this can be useful because they give us a measuring objective for the price move once we see a breakout from within the boundaries of the triangle.  The height of the triangle is measured as shown by the solid pink link, and then that vertical distance is applied to measure from wherever the breakout point comes, either upward or downward.  So in the case of that triangle, the roughly 100 point height gets subtracted from the point where prices broke down out of the triangle in December 2012 to produce a downside measuring objective of around $1585/oz.

One could calculate these numbers with more exacting precision, but to do so could create the misleading impression that prices will live up to that precision.  Real life does not work that way, and so I have found that using approximations is much better, just to keep expectations in check.  The idea is that you are getting an "objective", which is not the same thing as a guarantee.  And prices don't have to stop just because they reach an objective; they can go on well beyond it. 

Partway down from the breakdown out of that first triangle, gold prices paused to form what turned out to be another triangle structure.  It was even more interesting that the formation of that second triangle involved two separate retests of the underside of the broken (solid) uptrend line from that first triangle.  We can now see that prices broke down once again out of that second triangle, and so the blue lines are meant to show what the downside objective would be from using that second triangle. 

Interestingly, that second triangle carried an almost identical message, calling for a downside objective just a little bit lower than the $1585 objective from the first (pink) one. 

Now we see that both objectives have been met.  I had been officially bearish for my "current opinion" on gold as published in our Daily Edition, but now with the downside objective reached and an obvious oversold condition evident, I exited to a "neutral" stance.  I don't think it is time to catch the falling knife and bet on a big gold rebound, but at the same time I am out of arguments for why gold should continue lower.  I will be updating that opinion in the Daily Edition as time moves forward. 

There is one last point about using classic bar chart analysis techniques.  Most of the classic Edwards and Magee style chart analysis ideas come from the days when they were developed for stock price charts.  I am talking about chart structures like head and shoulders, rounding bottom, flags and pennants, triangles, triple tops, etc.  It can be risky to just transplant those analysis techniques to a different market, especially commodities, and to then just assume that they will work the same. 

The psychology of commodities markets, the bond market, currencies, etc. is much different from how traders think about stock prices movements.  People don't buy stocks to consume them, like they do with oil or wheat; the only use for a stock is to (hopefully) earn dividends and capital gains from owning them.  Accordingly, chart pattern analysis does not necessarily work the same in other markets, because the psychology which drives the patterns is different. 

I ignore out of hand all head and shoulders structures outside of stock price charts because I have found that they are just not that reliable.  One also cannot do the volume analysis piece as easily, which Edwards and Magee asserted was essential for validating or refuting a head and shoulders structure. 

By looking at a lot of charts, I have satisfied myself that triangles "work" for gold prices as they did in this case.  Anyone wanting to employ such charting techniques outside the stock market should put himself through a similar process of validating that they work in the other market before just assuming that they do.

The genuine old saying is that "A fool and his money are soon parted."  The more modern version is, "A fool and his money are SOME PARTY!"  That latter one seems to be the theme for the stock market in recent years, as the Fed's efforts to avoid a repeat of the 1930s by printing more money have at least kept the stock market aloft.

The Fed's efforts have thus far not done much to keep the unemployment rate from repeating what happened in the 1930s.  But at least those who are still employed are contented with their 401Ks.  But the risk for investors is that a stock market whose only support agent is the Fed printing money is one that is big danger of trouble should the printing presses ever slow down.

The Fed's first effort at "quantitative easing", known as QE1, added money to the banking system through "Permanent Open Market Operations", or POMOs.  That effort ended in March 2010, and the stock market did not survive long after being weaned off the easy money.  The illiquidity which resulted from shutting off that flow of new money resulted in the Flash Crash in May 2010.  The stock market really did not find its footing again until the Fed announced QE2 in July 2010, and started pushing that money into the banking system beginning in August 2010.

The Fed tried to wean the banking system off that extra money again in June 2011, paring its purchases of Treasury securities back to just reinvestments of the interest payments it earned.  The market did not like that, and the SP500 dropped 17% in just 3 weeks during July and August 2011. 

In response to that illiquidity event, the Fed decided to get imaginative, and it started "Operation Twist" in lieu of a pure QE effort.  The idea was to change the yield curve by selling shorter term debt instruments and buying up longer ones, thereby dropping yields on long term Treasuries and by association on mortgage rates.  On alternating days, the Fed would buy and sell securities in amounts ranging from $1 billion up to $8 billion.  Overall this helped the stock market, but the habit of separating the purchases and sales actually caused some sloshing of the money supply in interesting ways. 

I was able to create an interesting indicator that looked at the amounts and dates of the Fed's announced purchases and sales, and use it to model what was ahead for the stock market.  This next chart shows how I did that:

The purchases and sales seemed to show up a couple of days later in the movements of the stock market.  The lower indicator shows total net purchases or sales over the prior 4 days, and it is shifted forward by 2 trading days to illustrate how the stock market follows those movements.  The Fed has even been nice enough to publish a schedule of its intended purchases and sales up to a month ahead of time. 

What is interesting right now is that the last of the sales was done on Dec. 20, 2012, and since then it has been only purchases, which means that the cash has been flowing in only one direction - - into the banking system.  The response has been an abnormally steady and linear uptrend for the SP500.

That's all fine for investors for as long as the Fed keeps the party going.  But now that the market is relying so much on these fund flows, any interruption or even a hint of an interruption in those flows would leave the market overbought and at risk of a big drop.  As investors who have gone all in to ride this nice linear uptrend start to get twitchy, we could see a rush for the exits much like what we saw in July and August 2011.  In fact, I am looking for just such an event. 

So far, Fed officials have not given any hint of an interruption in the money flows, and in fact they have stated that it should continue until the unemployment rate drops to 6.5%.  So the market's built-in presumption is that the party will go on indefinitely.  That sets up the very big risk of a sudden change in attitudes any time there is a hint of a change at the Fed.  Smart investors already know this, and will be inching their way over toward the exits, so that they can jump out ahead of the rush.

The news this week out of Europe has to do with upcoming elections in Italy, and what that might mean for the future of the euro.  But few news stories are covering the really important development in Europe, which is the shrinking size of the European Central Bank (ECB) balance sheet.

The U.S. Federal Reserve has been doing its part to inflate the banking system, most recently with its announced program to purchase $40 billion per month of mortgage backed securities (MBS).  But the Fed's total balance sheet size of $3.02 trillion pales in comparison to the size of the ECB's €26 trillion balance sheet, which equates to $34.6 trillion as of December 2012.  While the Fed's holdings have increased 6.7% since the September 2012 Jackson Hole conference, the shrinkage of the ECB's balance sheet over that same time period has more than made up for the Fed's money printing. 

The reason why this is important is because there is a very strong correlation between the combined size of the ECB and Fed balance sheets and the movements of the world's stock markets.  When the balance sheets are growing, that is overwhelmingly a bullish factor for stock prices. 

When balance sheets stop growing or start shrinking, it is a little bit more complicated.  Past episodes of shrinking balance sheets in 2008, 2010, and 2011 all were associated with big drops in the SP500 and other indices.  As 2013 gets underway, the SP500 is continuing higher and challenging its 2007 all-time high, but it is doing so in an environment when the combined balance sheet size has not been rising.

This creates a bit of an analytical quandary.  Is this stock market strength with no balance sheet rise a sign that the stock market and the world economy are finally able to proceed on their own, and without central bank stimulus?  Or is this just another example of what we saw in early 2010, when the stock market continued higher for a few months after the faucet was turned off? 

In that 2010 example, what resulted was the Flash Crash in May 2010, and a lower bottom in July 2010, after which the stock market started upward again when the Fed announced its QE2 program, and when the ECB jumped in with similar monetary stimulus.  So to bet on the hypothesis that the stock market can now fly on its own is to say "it's different this time", which is one of the more dangerous phrases for investors to ever utter.

One of the most important market cycles that we have tracked over the years has been the 9-month cycle, which is sometimes also referred to as the 40-week cycle.  During the 1980s and 1990s, it had a fairly reliable period of about 185 trading days.

That period suddenly changed in 2007, and since then it has been running at between 156 and 168 trading days, or about 8 months.  I do not have a good explanation for why that change would occur, just as I don't have a good explanation for why a 9-month cycle would have existed in the first place.

My assumption has always been that it was a "natural" cycle, meaning that it was tuned to some harmonic frequency of how the human mind works, especially in a crowd phenomenon like the stock market.  This would be differentiated from man-made cycles such as the 4-year Presidential Cycle which is tied to the U.S. election frequency.  If there were a 3-month cycle, then we could conclude it was somehow tied to quarterly earnings or quarterly tax deposits. 

I never came up with a good explanation for some man-made cause for a 9-month cycle, which is why I thought that it was something "natural" to the market.  Having studied the theory of vibrations of mechanical structures years ago, I understood that every solid body has a natural frequency at which it vibrates.  If you take a yard-stick and hold one end on the edge of your desk, then you can give the free end a bump and then see and hear the "thwang" of it vibrating at its natural frequency.  If you shorten or lengthen the portion of it that is sticking out into the air, you can change that natural frequency. 

Similarly, in a fluid medium like water or motor oil, there are waves that travel at a natural frequency for that medium.  So if you had an experimental wave pool and substituted motor oil for water, you would see a change in the natural frequency of how waves travel.  Changing the viscosity of the fluid makes for a change in natural frequency.

If we take that same principle and apply it to the fluid medium of the financial markets, then by observing that a cycle has changed its period in a meaningful way, then we can conclude that the "viscosity of money" must have changed in some way.  What that change might have been is a different question.  This change in period from 9 months down to 8 months began in 2007, which just happens to be when the SEC eliminated the "uptick rule" for shorting stocks.  It is hard to see how that might have changed the viscosity of money.  Other changes have been under way for years, such as the shift to decimalization which started back in 2000, and the change to 3-day stock settlements instead of 5 which started in 2004. 

Interestingly, the 9-month cycle period survived both of those changes, and lasted until 2007.  If those structural changes were really responsible for changing the viscosity of money and thus changing the cycle period, then I would think they should have mattered right away. 

Whatever the explanation, we have enough evidence now in the 5+ years since the change in period to say that the old 9-month or 185 trading day period is no longer operative.  And for the first few years of the new period, a couple of other attributes of the old 9-month cycle also survived with the new period.

The first of those attributes is to have a mid-period bottom about halfway between the major cycle lows.  Here is a chart from 1996-1999 which illustrates this pretty well:

The mid-cycle lows tended to be not as significant as the big end of cycle lows, but still noticeable.  We can see evidence in the main chart above that this tendency has continued, at least up until recently.

The other attribute of the old 9-month cycle which survived for a while was that this cycle did not invert.  Other cycles in the market can sometimes invert, which is when a top appears instead of a bottom at the end of the cycle.  In studying decades of examples of 9-month cycle lows, I could not find a single instance of a genuine inversion.  Instead, the 9-month cycle had an interesting personality quirk, which was that about every 6-8 years, it would undergo a "phase shift", where it stopped beating on the old schedule and started up again on some new one.  Here is an example of that behavior:

What is happening to the market right now does not seem to fit the old model at all.  We have already seen one true cycle inversion top back in early 2011,  and as of right now (Feb. 1, 2013) it appears that another inversion is arriving.  2011 also saw the last of the real mid-cycle lows, with the last two major cycles not seeing a proper looking mid-cycle low.  Maybe we are in the middle of a newfangled phase-shift event, or maybe some other explanation will be found for this different behavior we are observing. 

Other than noting that a market top appears to be arriving now at the projected end of the current cycle, my point in bringing all of this up is that once-trusted market phenomena can change.  The 9-month cycle had a great track record for decades of working in fairly predictable ways, but it is not working in those ways any more.  Maybe it will come back, but probably not.  So anyone who looks at market history to develop an hypothesis about what the "rules of physics" are for the market needs to understand that those rules can change.  My observation has also been that such rules are most likely to change at the moment when one decides to count on them most to continue working.

Back on September 26, 2012, we published a chart in our twice monthly McClellan Market Report newsletter comparing the price pattern in Apple Corp (Nasdaq: AAPL) to that of RCA from the 1920s and 1930s.  I had taken a chart previously published by Global Financial Data of RCA's price behavior, and graphically overlaid a chart of Apple's recent price action, being careful to equalize time spans.  The point of the comparison was that Apple's price behavior during its big advance over the past 3 years looked an awful lot like the pattern RCA had shown during its own big advance in the 1920s.  The implication was that the topping behavior of that tech darling from years ago might be playing itself out again in the pattern of Apple's price behavior.

I updated that chart in my Nov. 8, 2012 Chart In Focus article, which is linked below, after Apple's initial decline seemed to have validated the chart pattern analog, at least up through that date.  I subsequently attained the raw monthly data for RCA with the help of the nice folks at Global Financial Data, which allowed me to create the chart above.  It has been featured in the pages of our twice monthly newsletter, and it appears to be continuing to work even now.  They were also the subject of a Dec. 12, 2012 interview I did on CNBC.

Apple's earnings disappointment for Q4 has sent its share price plunging on January 24, 2013, and that sharp drop is exactly in keeping with what RCA's price pattern said should be happening now.

It is fascinating to watch the commentary about Apple in the financial media.  Fundamental analysts are lowering their earnings outlooks and price targets.  TV news anchors are asking whether Apple as a company has finally run out of mojo 15 months after Steve Jobs passed away.  And all of this bearish talk about Apple is arising at a time when RCA's price pattern suggests that Apple should be nearing a meaningful bottom.

Following its drop in the 1929 crash, RCA actually saw a multi-month rebound effort in 1930.  That rebound eventually rolled over, taking RCA's share price down to much lower levels during the early 1930s, but the rebound in early 1930 was an impressive one while it was going on.  Given Apple's price pattern resemblance to RCA's behavior thus far, it would not be too surprising to see a similar bounce in Apple in the months ahead, especially given how bearish so many investors seem to be turning lately about Apple's stock.

In a similar vein, Apple's price pattern also shows a nice resemblance to another tech darling from the past, as shown in this next chart which has also been featured in our newsletter as well as our Daily Edition:

Microsoft's share price during the run up to the 2000 top of the Internet bubble looks an awful lot like Apple's price behavior since 2010.  One big exception to the pattern correlation is noted in the chart.  MSFT got knocked off track in April 1999 when the Justice Department's antitrust action against the company brought some bad news for the share price.  Following that dip, MSFT got back on track and continued higher.  Apple's share price skipped making a similar dip at that point on the chart, although the other landmarks on MSFT's price pattern did see equivalent behavior in AAPL's price plot.

If Apple's share price continues to follow the path set down by MSFT in 2000, then we can expect a continued decline toward a bottom ideally due Feb. 19.  But before you go counting your chickens about that coming true, remember that all price pattern analogs that I have ever studied have broken correlation eventually.  And they seem to have a habit of breaking correlation right at the moment when one is counting on them most to keep going (how do they know?).  Some kind of break in correlation is almost assured, since the RCA and MSFT patterns seem to disagree about what happens after that initial bounce from AAPL's presumptive February 19 low.

Remember also that in all of this analysis, I am just comparing the stocks' price patterns, which are statements about the behavior of investors.  I would never make the statement that Apple the company is the same as Microsoft or RCA.  They are very different from each other in real ways.  But from the viewpoints of investors in each era, each of these companies was seen as the tech darling of its time, and so investors seem to have responded in similar and predictable ways in each case in terms of the share price behavior.  

It is time to review some ancient stock market jargon, in order to understand the principle of the disappearing market volume.

A "lot" is 100 shares of stock.  In the era of specialists controlling trading on the floor of the NYSE, trading was done much more easily in "round lots", meaning multiples of 100 shares which the specialist could apportion out to other traders, much like how an air traffic controller directs multiple inbound and outbound aircraft in the same airspace.  In the days before computers and calculators, putting transactions in round lots made things easier; figuring out how much money had to change hands for a stock at $93/share is much easier if the quantity is 100 shares.  That's easy math: $9300.  It gets harder if the multiple is different than 100.

"Odd lots" means anything other than a round lot, and in the old days, odd lot orders would go to the "odd lot desk" in order to get aggregated with others into round lots, or otherwise married up with a corresponding opposing order or acceptance.

In the days of paper order slips, order ledgers, and share certificate clerks running paper back and forth across lower Manhattan, these were important functions.  Now that computers move bits around and actual certificates stay in place (or are imaginary), these ideas seem antiquated.  Why should it matter whether a trade is 200 shares or 43?  The computer does not care, and it does not have trouble doing the math of things that get multiplied by some number other than 100.

When you step up to the big leagues of trading, a "block" is any trade of 10,000 shares or more that is done all at once.  Block trades used to be how the big institutions would move money around without bothering to mix with the commoners who traded round lots or (ick!) odd lots.

Back in those old days, it was tremendous fun to compare the number of block trades to the odd lot trades, and even to look at odd lot shorts.  That way, an analyst could evaluate what the big guys were doing versus what the little (presumably clueless) traders were doing.  It was fun while it lasted.

The advent of computerized trading algorithms changed all of that.  Nowadays, you can have a computer break up a block trade into multiple odd lots, thereby disguising what the big institution is doing and hopefully preventing other traders from sniffing out a big move and trying to front run it.  So the whole idea of little guys trading in odd lots and big guys trading in blocks has been thrown out the window.  That helps to explain why the total volume of block trades is now down to a small fraction of what it used to be.

But that is not the whole explanation, which is (finally!) what this week's chart is all about.  The migration out of block trades has not just been to smaller denominations of share volume, but off the exchange entirely.  If a major institution wants to move big money into or out of a stock, it does not necessarily need to trade the actual shares.  One option controls 100 shares of a company's stock, and so 100 call options does the equivalent of what a block trade meant in the old days.

I had a question posed to me this week by a colleague named Greg Adams, who is the Chief Investment Strategist for Allen, Mooney, & Barnes, a money management firm.  Greg had noticed that total daily volume has been shrinking in a lot of the stocks that he and his firm are accustomed to following, and the upshot of this is that his ability to trade full blocks of shares like he used to has gone away.  Greg hypothesized that the major institutions which formerly traded the blocks in the old days had migrated away, and that the trading may still be taking place but in the options market instead of the regular stock market.

The chart above supports that hypothesis in a big way, or at least it did up until the past couple of years.  The big drop in block trades from 2006-2008 matched a corresponding rise in call option trading, as if the activity was migrating from one part of the market to another.  For big institutions, the attraction of options is obvious: Why buy the stock when you can rent the potential price appreciation for a small option premium, and in the process conceal from other traders what you and your gigantic portfolio are doing?

So the idea that trading was migrating from blocks to options made sense for a while, but it makes less sense lately.  Both the total call volume and the blocks volume are now near multi-year lows.  The chart above shows 21-day moving averages, smoothing out the fluctuations over the period of about a month.

Interestingly, both block volume and call option volume were on the rise in the late 1990s, when the Internet bubble and the Baby Boom population reaching peak earning years worked together to bring more money to the market.  During the 2000s, they seem to have moved in opposite directions, as the rising tide was no longer lifting all boats, and it became a fight over dwindling market share.

Now both markets are evidently shrinking, as the Baby Boom generation starts to enter retirement mode in earnest, and as the stock market's movements are driven to a larger extent by whatever the Fed wants to do to manipulate things.  That shrinkage in blocks and in call options matches a drop in overall volume over the past couple of years.

In terms of both share volume and dollar volume, trading activity has been shrinking pretty dramatically since the second half of 2011.  We calculate dollar volume ourselves, using data published by the NYSE on the composite trading of all its listed stocks (not just floor volume).  We multiply the closing price by total daily volume, and then smooth it with a 21-day simple moving average for display in this chart.  That's not quite perfectly accurate, since not all of a stock's trading is at the closing price.  But it is a good enough approximation.

Share volume has been shrinking for a lot longer, but the explanation for that is understandable.  Companies are not splitting their stock now as much as they did in years past, following the model of Berkshire Hathaway and just letting the price continue to go up as the stock appreciates.  That's why looking at dollar volume helps us see that up until 2011, trading activity was remaining about the same as it had been.

But over the past year and a half, there is a distinct change in all of these volume measures.  Part of that may be due to the implementation of the Volcker Rule, prohibiting banks from doing short term proprietary trading for their own accounts.  The collapse of MF Global (and Bear Stearns, and Lehman) also takes away some of the prior levels of trading activity.

The point to take from this for individual traders is that we are in an environment of shrinking volume all across the market, and so that should be taken into account when interpreting volume figures for technical analysis purposes.  It also increases the likelihood of a Flash Crash type illiquidity event, if there is not as big of a background level of trading to absorb any ripples or turbulence.  As with the Flash Crash, which unfolded right after the end of QE1, we should see those illiquidity events more frequently once the Fed starts to pull back on all the extra money it is throwing at the banking system.

The big monthly Employment Situation Report just released this week showed growth in nonfarm payrolls, but no progress in reducing the unemployment rate.  There was similarly no improvement in the data series shown in this week's chart, which examines the percentage of the overall population which is currently "employed".

This data series is never going to get to 100%, because newborn babies and 95 year olds do not typically have jobs.  Since 1948 when the Labor Department first started publishing this data, it has ranged from a low of 54.9% to a high of 64.7%, and the average has been 59.2%.  So the reading of 58.6% for December 2012 is just a hair under that long term average.

The composition of the work force has changed a lot since 1948 when this data began.  More women are now working outside the home than in the days of Ozzie and Harriet.  But now more Baby Boomers are reaching retirement age and leaving the active work force.  Since the economic slowdown began in 2007, there have been other workers leaving the work force for reasons other than retirement, and the frustrating point for the politicians is that their efforts so far have not done much to bring those work force numbers back up.

The point in this week's chart is that the Civilian Employment-Population Ratio is remaining stubbornly low despite the stock market making back almost all of what it lost in the 2007-09 bear market.  There is usually about a 1-year lag between important bottoms in the SP500 and corresponding bottoms for this measure of employment.  Over the four decades preceding this most recent recession, that lag has ranged from as low as 9 months to as high as 14 months.

We did not really see a normal sort of bottom in this Employment-Population Ratio this time.  It did stop going down, but it still has not started going back up again like it has done after past stock market recoveries.  The implication is that if we see another stock market downturn, we could see civilian employment levels go down even further.

One other interesting observation about this data relates to the recent discussions in Congress about what the ideal tax rates would be for reducing the deficit.  Those who argue for higher tax rates have noted that the higher rates which were in effect during the 1990s led to budget surpluses, and thus they must be a good thing.  But what those proponents of higher tax marginal rates evidently fail to consider is that the percentage of the population which was working during the 1990s was at an all-time high, well above the long term average levels, so there were more workers versus non-workers then, and thus a broader tax base.

This was partly because the Baby Boom generation was in its peak earning years.  Boomers were born between 1946 and 1964, so at the peak for this ratio in 2000 they were 36 to 54 years old.  Now at the end of 2012, the boomers are 48 to 64 years old.  As a group, they are now at an age when they are less interested in starting new businesses, and more interested in keeping what they have as well as playing with their grandchildren.  So Boomers as a group do not represent the same sort of engine of economic growth that they did in the 1990s.

Coming up behind them is an "echo boom" generation, which saw its peak year for births in 1990.  As those youngsters reach their own peak earning and entrepreneurial years, we should expect a renewed economic boom.  But right now, those people born in 1990 are just 22 years old, and so most of them are not quite ready to start the next great corporation, and they are certainly not ready to buy up all of the Baby Boomers' stock portfolios and McMansions.

From a policy standpoint, unless we can collectively find a way to get more of the total population back into the work force, we cannot hope to have the tax base needed to fund lofty federal spending goals. 

Last week, I looked at the yield curve as modeled by the spread between the 1-year and 10-year yields on Treasury Notes.  That's not the entirety of all maturities on the entire curve, but it does give us a simple graphical representation of what the slope of the yield curve looks like.  The two main points of that article were that an inverted yield curve is always bad for the stock market and for the economy, and also to note that just because the yield curve is not inverted, that does not necessarily mean that the stock market is immune to having trouble.

This week's chart gives us a good insight as to why that is.  For this week's chart above, I have flipped around the yield spread plot so that we are looking at the 10-year yield minus the 1-year, whereas last week I showed it as the 1-year minus the 10-year.  The result is the same, just upside down.

One other adjustment in this week's chart is that I have shifted forward that yield spread by 22 months to reveal that the movements of the DJIA seem to follow in the footsteps of this yield spread model.  It is not a perfect fit, just a really good one.

We can see at the left end of the chart that the 1987 peak and crash followed similar movements 22 months earlier in this yield spread.  And the 1990 bear market came on schedule, helped along by Saddam Hussein deciding to invade Kuwait that year.  Later in the 1990s, the relationship got a little bit off track, as the Internet bubble caused stock prices to be skewed higher than was modeled by yields.  But the end of the Internet bubble in 2000 was followed by stock prices working hard to get back on track with what this model showed.

We saw a similar skew during the next bubble, which was in real estate prices during the mid-2000s.  That bubble once again caused the stock market to remain aloft longer than it should have, but then to correct extra hard during 2008 in order to get back on track.  The rise up from the 2009 bottom has been right on schedule according to this 22-month leading indication.

There are a couple of points about monetary policy which jump out from the realization that the stock market works this way.  The first is that there is a big lag between when the Fed first starts cutting short term rates, and when that move finally has an effect on the stock market.  It is similar to how the release of water from a dam takes some time before water levels rise downstream.

The second point is that the Fed's current efforts to buy up longer term Treasury and mortgage debt in an effort to stimulate the economy is a misguided policy.  What the Fed is doing is artificially suppressing longer term rates, thereby flattening the yield curve and pushing down this spread.  That's not what helps the stock market go up, and thus it is not what helps the economy improve.

This next chart takes a longer look at this relationship, looking all the way back to 1955:

The fascinating point which jumps out from this longer look is that the lead-lag relationship has changed over the decades.  For the current time frame, I needed a 22-month forward offset to get the patterns to line up.  But you can see that the further back in time we go, the more that this 22-month offset seems to be too much.  Back in the 1950s and 1960s, a forward offset of about 12 months worked better.  For some reason, the lag time of interest rate changes having their effect on the stock market has been lengthening.

I am at a loss to provide an econometric reason as to what has changed in the economy to explain such a lengthening.  One would think that with the easier ways we have now to move money around and take advantage of opportunities, that cycles would shrink.  It certainly works that way with fluid flows, like my river analogy above.  If you switch to a lower viscosity fluid, it will flow faster and more readily.  So reducing the viscosity of money should make for shorter lags instead of longer ones, if my thinking is correct.  But the chart shows that my reasoning is not correct in terms of the stock market's response to interest rate changes.  I'm okay with that, and can just follow what the chart says instead of what my mind thinks it should do.

Looking ahead, the flattening of the yield curve during the past 2 years (thanks, Dr. Bernanke!) means that stock prices will be under downward pressure. 

With the Fed keeping its foot on the neck of the interest rate market, and effectively keeping short term rates at zero, the net result is that we have a fairly steep yield curve.  What that term means is that if you were to plot all of the different maturities for a type of interest bearing asset like T-Bonds, then a steep yield curve would be sloped from lower left to upper right, as in the small chart.  An "inverted" yield curve happens when short term rates go up higher than long term rates, and it is a reliable sign that trouble is coming for the stock market and the economy.

Fed Chairman Ben Bernanke made a comment back in February 2007 that will now go down as being among the worst quotes ever by a Fed chairman.  Referring to the inverted yield curve then, he said that he did not foresee it as putting "tremendous pressure" on the banking sector.  I bet he wishes he could take that one back.

We do not have an inverted yield curve now, with short term rates at near zero, and longer term rates above that.  To get to an inverted yield curve now, we would need to see either the Fed raising short term rates, or the 30-year T-Bond zooming up to around a 240 price level.  Neither is very likely, but as this week's big chart shows us, that does not necessarily immunize the stock market from trouble.

It is difficult to portray changes over time in all of the different maturities which make up the Treasury yield curve.  So I am modeling the overall steepness or inversion of that curve by just measuring the yield spread between 1-year and 10-year T-Notes.  The rise since the absolute bottom for this spread back in 2010 means that the yield curve is flattening somewhat, but still is a long way from being inverted.  We can see that whenever this yield spread does go above zero, it nearly always means an important long term top for the stock market.

This is partly because the higher short term rates associated with an inverted yield curve do two harmful things.  First, they make it more expensive for businesses to borrow needed capital, which therefore puts a squeeze on earnings and capital formation.  Second, they attract money away from riskier investments in the stock market so that investors can take advantage of the higher short term yields elsewhere.  That pushes stock prices down.

But the whole point of this week's chart is that the stock market is not necessarily immune from a big decline just because the yield curve is still "normal".  The circled instances show periods when the stock market still managed to make an important decline.  The one at the left end of the chart was the 1962 bear market, which saw a 27% drop for the DJIA.  Part of that drop was that President Kennedy was putting pressure on the steel companies, as a favor to his union supporters, which included sending FBI agents to the homes of steel company CEOs late at night to notify them of the president's displeasure at their intended price increases for steel.  Wall Street saw that as the government attacking business, and investors fled.

The second circled example was in 1978, when the Fed was actually raising short term rates but not fast enough to keep up with long term rates that were being driven higher by runaway inflation.

The third circled example was the crash of 1987.  In September 1987, the month before the crash, the 1-year yield was at 7.67% while the 10-year yield was at 9.42%, creating a negative 1-10 spread of -1.75 percentage points.  But that did not stop the stock market from undergoing the worst crash since 1929, helped out by Alan Greenspan taking office in August 1987 and wanting to "mark his territory" as the new dog in town.  He surely succeeded.

So the point is that even though the Fed says it is doing "everything possible" to avert another big financial crisis and associated stock market decline, there are nevertheless some precedents for the stock market getting into trouble anyway.

For years we have used mutual fund assets at Rydex and other fund families as a sentiment indicator.  The proliferation of ETFs has had a big effect on the old-fashioned style of mutual funds, but it brings along with it the opportunity to look at investor sentiment in a whole new way.

This week's chart looks at a comparison between the share price of SPY, the leading ETF tied to the SP500 Index, and the total number of shares outstanding.  That latter number vacillates with investor mood, waxing and waning as the level of the SP500 goes up or down.

The way that ETFs work is that when more investors want to own shares, they bid up the share price and it goes above the net asset value (NAV) of the ETF's holdings.  When that happens, the sponsoring firm issues more shares and thereby helps to drive the share price back down closer to the NAV.  When investors throw up their hands and wail in fear, selling their ETF shares in quantities exceeding the market demand, then the sponsoring firm redeems shares and nudges the share price back toward the NAV.  So the number of shares outstanding can be used as a gauge of investor interest in the stock market, especially for the bigger ETFs like SPY which command the greatest interest.

But one big problem is that there are no static measures of where "high" and "low" are for ETF share levels.  We can only make inferences about that from looking at outstanding share levels on a relative basis.  To help in that effort, I have added 50-1 Bollinger Bands to the chart above, so that we can see how excursions above or below those limits are associated with meaningful tops and bottoms for stock prices.

The 5.7% rise in the share price of SPY since the Nov. 15 bottom has brought about an almost equal rise of 5.6% in SPY shares outstanding.  Thus far, the shares outstanding has not yet poked its head up above the upper 50-1 Bollinger Band, but it is getting very close.  The 50-1 identification convention means that use a 50-day simple moving average of outstanding shares.  Then the upper and lower bands are each set one standard deviation above and below that moving average.  Most charting software programs default to a standard 20-2 convention, but I find that the 50-1 convention works well for a lot of sentiment related indicators, including the VIX.  It seems to work nicely for SPY shares outstanding, as this week's chart shows.

You can track this data yourself by going each day to the SPDR web site and recording the published value each day for total shares outstanding.  That's kind of a big pain, and so if you are a subscriber to eSignal or QCharts, you can access the data under the symbol $SPY.SO.

In just a few more days, the upper band will have dipped lower, and the number of shares outstanding may have popped up higher by enough to move above the upper 50-1 Bollinger Band, giving an overbought condition befitting a meaningful top.

The share price of Apple Corp surprised a lot of people this week with a one-day drop of 37 points, or -6.43%.  Part of why that was a surprise is that over the past year, the standard deviation of Apple's daily price change has been 1.8%, and so this was a down move which was greater than 3 standard deviations.

Those who have taken a course in statistics may recall "The 68-95-99.7% Rule".  Simply stated, the rule indicates that for a normal distribution:

• 68% of all observations should fall within 1 standard deviation of the mean
• 95% of all observations should fall within 2 standard deviations of the mean
• 99.7% of all observations should fall within 3 standard deviations of the mean

So if Apple's daily share price change followed the rules for a "normal" distribution, and if one were to know that the standard deviation of daily returns is 1.8%, then there should be a 99.7% chance that any single day's change would be within +5.4% and -5.4%.  Saying it another way, there should be only a 0.3% chance of a daily percent change being greater than 5.4% up or down.  That's roughly one instance out of every 333 times, and so in a year with only 252 trading days, it should almost never happen.

But just in 2012, we have already seen 5 different days when Apple's share price rose or fell by more than 5.4%, something which the whole notion of a "normal" distribution says just should not happen.  This is just one example of what some market analysts are referring to when they talk about the concept of "fat tails".

The "tails" of a bell curve distribution should see diminishing numbers of observations the farther you get away from the mean value.  But when a distribution has "fat tails", that means there are more observations than theory would suggest out near the tails.  To help see what that means for Apple's share price movements, here is a chart showing where they have fallen over the past 24 years:

What we find is a whole lot of observations well outside of that standard deviation value of 1.8% that I mentioned above.  Standard deviation of daily returns is often used as a measure of the inherent volatility of a stock price or other data, but as most traders know, volatility is not constant.  And that leads to the final point, which is that the figure for the 1-year standard deviation of daily price changes varies dramatically over time. 

Since 1988, Apple's 1-year standard deviation has been as low as 1.3% and as high as 4.5%.  In fact, the past 3 years when Apple's product innovations have brought so much of a big gain in its shares has seen unusually low volatility in comparison to the rest of its history.  And if Apple is starting to transition into a corrective period as the comparison to RCA's share price pattern suggests is coming, then we can figure on volatility increasing, and on having big up and down days becoming much more common occurrences.  We have been updating that comparison of Apple's share price now to RCA's price pattern from the 1920s and 1930s with subscribers to our twice monthly McClellan Market Report and our Daily Edition.

In the 1955 movie Rebel Without a Cause, the lead character Jim Stark (James Dean) gets involved in a "Chickie Run", in which teenaged boys with more testosterone than brains race stolen cars toward a cliff.  The idea is to jump out of the car just before it goes flying off the cliff, and whoever jumps out first is deemed to be the "chicken" (coward), and subjected to the derision of his peers.  That scene ends badly for one of the characters whose jacket gets caught on the door handle preventing him from jumping out.

Congress is now fully involved with its own Chickie Run, racing toward the Dec. 31 deadline that the press has dubbed "The Fiscal Cliff", after which mandatory cuts in certain military programs and other spending items will take effect unless Congress and the President act to strike a deal to close the budget deficit.  Some of the players seem intent on getting the other side to jump first, evidently favoring the political victory that they think they would get more than any actual solution to the problem.  Others are voicing the opinion that actually going over the cliff, and letting everyone see the effects of the mandatory tax hikes and spending cuts, would actually be a good thing.  That's about like Jim Stark concluding that it is better to be dead in a car wreck than to be considered a "chicken". 

Over the past 12 months, total federal receipts have amounted to 15.4% of GDP, while total expenditures have been 22.7%.  That's clearly a huge gap, that is going to need some bold action to get those numbers to come back together.  Proponents of big government programs insist that revenues must be increased (AKA tax rates must go up), because they don't want to shrink that expenditures number.  On the other side, a favorite slogan is that "we don't have a revenue problem, we have a spending problem".   

This week's chart offers us some evidence in support of the idea that more revenues from higher tax rates won't actually help.  It shows a comparison between total federal receipts as a percentage of private GDP (factoring out government spending from the GDP numbers) versus real GDP growth.  There is a fairly strong inverse correlation between the two, which means that if federal receipts go up (as a fraction of GDP), then GDP growth goes down. 

Let me say that again, just for emphasis: raising tax receipts (as a percentage of private GDP) is associated with slower or even negative real GDP growth. 

There was one period when that inverse correlation did not really work, and that was during the 1990s when President Clinton was in office, and when higher tax rates and total receipts did not seem to hurt the economy.  That point has led some pundits to take the myopic view that we just need to go back to Clinton's tax rate policies and everything will be fine. 

The problem with that logic is that President Clinton's tax rates were not the only factor affecting the economy during the 1990s.  Clinton rode a huge technology boom, that not only increased economic productivity but also led to huge investments in startup companies.  That decade also saw the Baby Boom generation in its peak entrepreneurial years, whereas now those aging Boomers are starting to slide into retirement rather than starting new companies. 

Furthermore, the corrosive effects of having federal tax receipts up too high really was being felt in the 1990s, even if it did not show up in the GDP numbers.  The total number of issues listed on the Nasdaq actually peaked all the way back in December 1996, more than 3 years ahead of the final price high for the Nasdaq Comp.

What was happening at the end of the 1990s is that the tax bite was so high that it put a real drag on the economy, stifling the innovation and business building efforts that had been in effect during the early 1990s.  Putting the brakes on the economy by having taxes up too high took too much money out of the real economy, and pushed the economy in the recession which was finally evident starting in early 2000. 

Turning back to the lead chart above, the inverse correlation between federal receipts and GDP growth was restored beginning in about 2001, and it remains in effect now.  So if someone proposes that the total federal receipts line should be bent upward by increasing the tax bite, then it is reasonable to conclude that the real GDP growth rate line would see itself bent downward in response.  In other words, the Congressional Budget Office was correct in its admonition that higher tax rates (from the expiration of the Bush era tax cuts) would push the U.S. economy into a recession once again.

For our roles as investors, we have to be ready for whatever Congress does.  The proof offered here that raising tax receipts would be a bad idea is not sufficient to keep Congress from going ahead with it.  Then again, having a Chickie Run in stolen cars is a bad idea, but having something be a bad idea is often not enough of an impediment to keep some people from proceeding with it. 

For more about "private GDP", see my October 2011 Special Report, "It's the PRIVATE Economy That Matters".  

When chartists talk about "magic" properties of indicators, it is with some degree of reverence for the mysterious ability to show us amazing insights that defy the presumption of randomness.  They touch upon the romantic nature of staring at lines and dots on a white background (for those who grew up like I did charting on graph paper) and finding those pearls of information.

The Summation Index that my parents developed 43 years ago carries at least a couple of magical properties.  I was assuredly no help in its development, since at age 8, I was more interested in Hot Wheels cars than breadth statistics, and I had not mastered the algebra skills needed for the manual calculation of exponential moving average values.  This was 1969, when hand-held calculators were still a long way off, let alone personal computers, and so manual calculations and plotting on graph paper was how one did it, if it was going to get done.

The McClellan Oscillator first came about when my parents wanted to dig deeper into the indicators used by the late P.N. (Pete) Haurlan, who was an actual rocket scientist and who published the Trade Levels Report newsletter.  He was the guy who introduced the use of what we now call exponential moving averages to the tracking of stock prices.  It was a piece of math borrowed from his other field of rocketry and satellite tracking.  Read more about that interesting story here.

Pete Haurlan liked to track individual exponential moving averages of the daily breadth numbers (advances minus declines).  But he did not identify which EMAs he used to calculate his various "Haurlan Index" values.  When Sherman and Marian McClellan started looking at breadth data on their own, they could not find a single EMA that worked really well.  Their key insight was to look at the difference between the 10% Trend (19-day EMA) and 5% Trend (39-day EMA), and that numerical difference between the EMAs came to be known as the McClellan Oscillator.  Coincidentally, on the other side of the country, and without communication between them, an analyst named Gerald Appel started playing around with the difference between two moving averages in 1969, and his work in that area came to be known as Moving Average Convergence and Divergence (MACD).

My mother Marian (who passed away in 2003) had been a math major in college, and so she had done integral calculus at a time when it was not commonly taught like it is today.  When she looked at the plot of the McClellan Oscillator, she naturally wondered what it would mean to integrate the "area under the curve", like students do regularly in calculus.  That's how the Summation Index was born, as an indicator which changes each day by the value of the Oscillator.

One of the problems that my parents encountered when trying to teach people to use the indicators was if you had a negative McClellan Oscillator value and you had to add it to a negative Summation Index value, then people got confused about what to do with all of the + and - signs.  That led to some calculation errors.  At that time, they noticed that the Summation Index had a total amplitude of about 2000 points, which was smaller than its amplitude today because of the smaller number of issues traded then.  So they just arbitrarily moved all of the Summation Index values upward by 1000 points so that having to deal with the manual calculation problems brought by negative values would be a rare event, and thus an unusual one.  In the 1960s and 1970s, the Summation Index almost never got below 0 except in really extreme conditions.

We have maintained that +1000 neutral level convention to this day for the "classic" version of the Summation Index, so as not to cause confusion.  For the more modern version that we call the Ratio Adjusted Summation Index (RASI), we use the more natural zero level as the neutral level. The RASI factors out changes in the number of issues traded, making it better for longer term comparisons and analysis. 

The idea of a neutral level is an important one, as this week's chart illustrates (yes, I'm finally getting to this week's chart).  When the Summation Index crosses through its +1000 neutral level, it usually reveals one of the magic tricks that it possesses.  Crossing down through neutral usually marks a bottom of some type.  It may not necessarily be THE bottom, but it is usually a least a bottom of some significance.

At the left end of the chart above are two examples of this principle.  The first dip below +1000 marked a preliminary bottom that was followed by a failing bounce.  The next dip below +1000 was a couple of days early for marking the August 2011 price bottom, but the effect was still there.

When the Summation Index crossed down through +1000 back in May 2012, the price bottom marked by that event was a bit early for the final bottom, but it was a bottom nonetheless.  The effect of passing through +1000 is not permanent, but it is there.

Now we are seeing another downward crossing through the +1000 neutral level, just as the market seems to be reaching the conclusion that taxes are going higher, and earnings are going to fall, and money will dry up, and the sky is falling, and we're all going to die!!!  But my suspicion is that the Summation Index is going to do another bit of magic, and mark at least a temporary bottom with this crossing down through +1000.

One of the agenda items that is as yet unfulfilled for this bull market is to see a Summation Index failure which would indicate the end of the up move.  We typically quantify that using the new-fangled version of the Summation Index, the RASI that I mentioned above.  Important tops for stock prices usually seem to have a relationship to a failure at or below the +500 level on the RASI.

To end an uptrend of significance, there is usually a failure of the RASI to climb up above +500, a level which currently equates to around +2500 on the classic version of the Summation Index.  When the RASI dips down to a nice oversold level and then is able to climb up above +500, it promises us higher price highs.  It is a message of adequate liquidity to continue, even as a shorter term overbought condition might necessitate a brief pause to refresh. 

It is at the moment when the RASI cannot climb back up +500 that its other big magic trick becomes evident.  A few good examples are highlighted in the second chart.  Thus far, we have not seen such a failure yet.  It is really not typical for the stock market to see an important top at a really high Summation Index level, like the raw value high of +3689 seen back on Sep. 21, 2012 (RASI equivalent: +888).  A much more normal resolution is to see a dip down toward neutral like what we have just seen, and then a failing attempt for the RASI to climb back up above +500.  If we see that unfold in the next few weeks, with a failure by the RASI to climb back up above +500, then we'll know that the typical weakness of the first year of a new presidential term is playing out according to the normal script.