The employment level as a percentage of the US population continues to fall, even though the stock market has been rising.  That breaks a longstanding relationship shown in this week's chart.

Normally the employment level as a percentage of the population bottoms about a year after an important stock market bottom.  The last such price bottom was in September 2022, on a monthly basis.  That should have meant a bottom for the employment/population ratio around September 2023.  Instead it has kept falling.

The unemployment rate has done fine the whole time, ranging from the high 3s to the low 4s now.  So it is not a problem so much of people not able to find work.  Rather, you can blame the retiring Baby Boom generation who are deciding not to stay in the work force.

When World War 2 ended in August 1945, GIs came home, Rosie The Riveter gave up her factory job, and Americans got busy making babies, resulting in a big spike in births in 1946.  That continued until 1964, which coincidentally was one year after Ortho Pharmaceuticals introduced the first oral birth control pill.  As of the 2020 census, that 1946-64 birth year cohort totaled 73.5 million, or 22% of the total.  Now those Boomers are 62 to 80 years old, so it is understandable that they would be leaving the work force.  This effect is magnified because the 1946-64 Baby Boom generation has long been skewed to the more recent birth years, with 1960 being the peak birth year for the Boomers.  That 1960 cohort reaches "full retirement age" for Social Security at 66 years and 10 months, so more of them will soon be checking out of the work force.  Those who were born before 1960 have already been doing that, which is why we see the employment to population ratio falling in the first chart.

Much has already been written about how supporting these pensioners who are leaving the work force is going to create great strain on the economy, and the federal budget.  I won't belabor that here, but I will point out that we will get to repeat all of this again 30 years later when the "echo boom" generation of the kids of Baby Boomers will be reaching their retirement age. They peaked births in 1990.

Why this matters for stock market investors is that the Baby Boomers who are retiring are going to start drawing down their investment holdings at a greater rate, to support their preferred lifestyles in retirement.  That has not been much of a problem thus far, with the Fed now doing QE5 and keeping the stock market up.  Appreciation of stock prices has helped soften the effects of the selling of stocks that Boomers may have been doing.  A meaningful stock market downturn into the teeth of this demographic bulge would mean drawing down retirement accounts' stock holdings even more. 

That is not an insight one can trade on.  It is just important background information about what challenges lie ahead.

The NYSE's McClellan A-D Oscillator has been below zero since May 11, and so that means the companion McClellan Summation Index has been falling.  The context of how this is happening has an important message for what is next in the stock market.

This week's chart shows the Ratio-Adjusted Summation Index (RASI), which differs slightly from the classic version my parents created in 1969.  The RASI is calculated the same way, but using Advance-Decline (A-D) data that has been adjusted to factor out the changes in the number of issues traded over time.  See https://www.mcoscillator.com/learning_center/kb/market_data/ratio_adjusted_summation_index/ to learn more about that adjustment.

Adjusting the data this way makes for more useful long term comparisons.  And it allows us a special way to look at the strength of a new rally.  This involves the +500 level.

The RASI was above the +500 level at the beginning of 2026, which said that the market was strong and the bulls were in charge.  The market can still see a correction from that condition, but most of the time there will be a renewed advance.  We have seen that renewed advance after the ugly month of March 2026, when everyone was worried about the outcome of war with Iran.  That corrective period took the RASI down as low as -323, which is a very ordinary sort of level to see on a normal corrective pullback.

It is on the next rebound where the +500 level of the RASI comes into play again.  And on this latest attempt, the RASI stalled out at +470.  A failure at or around the +500 level says that the rally does not have enough "escape velocity" to keep the uptrend going, and it opens the door for a more meaningful pullback. 

I need to clarify that the +500 level is not a precise threshold for this usage.  It is not like +499 is bad and +501 is great.  It does not work that way.  There have been RASI +500 failures which did see the RASI go just barely above +500.  So readers need to not apply this interpretive criterion too strictly.  The higher that the RASI can climb up above +500, the more confident one can be in the message that there is more coming. 

I should also note that big exogenous events like Covid in 2020 can ruin a perfectly good bullish indication.  There were also a couple of instances in 2022 and 2023 when the RASI rose up above +500 and the rallies fizzled.  Nothing is perfect.

The point to take from seeing the RASI apparently failing now just below +500 is that the breadth data are telling us the rally is fizzling.  The market could change its mind, or it could become a more serious pullback.  For now we should take the message of warning that there is trouble, and it is appearing just as the market is supposed to be transitioning into the bearish period of annual seasonality lasting from May to October.

The daily Advance-Decline (A-D) Line for high yield corporate bonds is making a bearish divergence versus the SP500.  This is a concern because it conveys a message saying that there are liquidity problems.

Every day, thousands of corporate bonds are traded much like stocks, and FINRA publishes the data on advances and declines at https://www.finra.org/finra-data/fixed-income/market-activity.  High yield bonds are of special interest to me, because they tend to trade more like the stock market than like Treasury bonds.  That means they draw from the same sources of liquidity that stocks do, and so when their A-D data start doing poorly, it is a sign of liquidity troubles. The daily A-D Line changes every day by the difference between Advances and Declines.

High yield bonds are sometimes called "junk" bonds, and they are among the least deserving investment vehicles because of their higher default risk.  So if the HY Bond A-D Line is rising, that means liquidity is so plentiful that even the least deserving issues can get access to that liquidity.  When it starts to suffer, and especially in a way not reflected by what the more deserving SP500 is doing, that conveys a big warning that liquidity may be drying up.

The problem with any divergence like this is that it is not a "signal", it is just a "condition".  Nothing will tell us when that condition will decide to matter.  And sometimes a divergence can get "rehabilitated".  So the observation of a bearish divergence now between the HY Bond A-D Line and the SP500 is no guarantee of problems.  But it is a big warning that something is wrong.

Divergences can also be hard to read and interpret sometimes.  Some might notice that the all-time high for this HY Bond A-D Line was back on Feb. 20, 2026, and it is well below that level now.  But this is not a divergence I find very important.  The price decline in March 2026 took the SP500 down, and this HY Bond A-D Line went down with it.  That is very normal.  Coming out of the March 30, 2026 low, this A-D Line went up very fast, about as fast as it could travel.  That was a good sign for the strength of the rally.  The fact that the A-D Line in April could not make back all of the damage seen in March should not be held against it.  An A-D Line can never go up faster than a certain rate, even if all of the issues were to go up every day. 

The divergence which concerns me is the more recent one, following a HY Bond A-D Line peak on April 20.  It is the action since then which is telling the more concerning story about liquidity problems suddenly appearing.  Those problems have not yet affected the performance of the SP500, because illiquidity comes first for the weak.  Eventually it comes around for the big animals on the savannah, and that is the concern now with this apparent divergence. 

But let me say this again: This is not a trend change "signal".  It is just a warning of trouble, and one which could get rehabilitated.  It will bear watching in the days ahead, especially as the calendar now transitions into the worst 6 months of the year.

Note: This article appeared originally in the April 22, 2026 issue of The McClellan Market Report.

The war with Iran and the shutdown of oil shipments from the Persian Gulf caused a big spike in oil prices.  That means anyone with oil is now making a lot more profit from selling it.  Good for them.  But it has created a bit of a mystery - - Why are US drill rigs not responding to this price change?  Why are landowners not responding to the higher prices by drilling for oil?

This week's chart compares the Baker Hughes data on US rig counts to near month crude oil futures prices.  This seems like an obvious place to find a relationship, but the correlation is not that good.  So why not?

Digging deeper into the data provided an answer.  A primary point to note is that Baker Hughes (a company which supplies drill rigs and thus wants to keep track of its customers) only counts a rig once it is set up and actually in operation.  So new projects which have a rig traveling to get there, or just starting to get set up, do not count.  There is a lag time in the response to prices.

Further, it is not the near month price that is the deciding factor to order a drill rig.  A newly operating rig is not going to benefit from a temporary spike in oil prices.  Oil producers instead respond to price changes further out on the futures maturity spectrum.

So here is a chart that looks at the price of the oil contract 11 months out.

After some tinkering, I found that shifting those far-month prices forward by 13 weeks gets the best overall fit to the drilling rig data.  The war with Iran did not cause the far month oil futures contracts to spike as much as the near month ones, but there has been a spike.  And we are only now getting to the inflection point when that price response should start showing up in the rig count data. 

Using this 13-week offset gets a much better fit between the plots.  It does not necessarily predict the magnitude of the change in rig counts very well, but it gets the direction mostly right.  The importance to investors is that stocks of companies tied to operating drilling rigs should be expected to do better over at least the next 13 weeks.

The 14-day Choppiness Index for the SP500 hit a reading of 4.60 on April 17, 2026.  That's on a scale from 0 to 100, and it is the lowest reading in at least the last 30 years.  It might be the lowest ever, but I only checked back 30 years.

The math for the Choppiness Index was created by Australian commodities trader E.W. Driess, as a way to quantify how choppy (or not) the recent price action has been.  It shows a high reading when above 62, or a low reading below 38 (Driess was evidently a Fibonacci ratios fan).  The 14-day Choppiness Index can go well beyond those thresholds, as it has just done with the April 17 extreme low.

A high Choppiness Index reading says that the market has been trendless recently, and thus a trending move is likely to develop.  But such a reading will not tell you in which direction that trending move will go, just that some type of trending move is likely.  Conversely, a very low reading means that the recent data have been extremely linear, and thus a non-trending period is likely. 

That is not very shocking news to receive after we have just seen such a price move.  A move like that merits a pause for the bulls to catch their breath.  The Choppiness Index is just a way of quantifying that.  It will not say how long such a pause may last, nor what is going to happen after that pause.  It just says that a pause is likely. 

Sometimes a trend-ending low reading can mark an immediate trend change.  That is what we saw at the price low in April 2025.  The very low 14-day Choppiness Index reading of 18.9 on April 7, 2025 marked the end of that down move, and what unfolded next was a sharp reversal higher.  Very little time was spent in a "non-trending" market.  So it is possible for such readings to mark reversal points.  That is more likely when these low Choppiness Index readings coincide with bottoms.

And this is the important point to emphasize: The Choppiness Index says nothing about uptrends or downtrends, about tops or bottoms.  It is agnostic about direction, and only comments on how linear (or not) the recent action is.  Accordingly, the Choppiness Index cannot be used alone as a trading tool.  It is an informative indicator to employ alongside other tools.

On April 15, the SP500 and the Nasdaq Composite Index moved to new all-time highs, and added to those records on April 16.  The NYSE's A-D Line is still 1018 net advances away from equaling its own all-time high.

Some might see it as a bearish divergence to have the SP500 make a new all-time high while the A-D Line has not yet confirmed.  I do not see it that way, in part because the NYSE Composite Index is still 2.4% away from its own all-time high.  So in that sense, the A-D Line is actually doing better than prices.  It is just that the SP500 is doing even better than that.

And it is further worth noting that the NYSE breadth data have been averaging 738 net advances in the two weeks since the March 30 price low.  It is really hard to criticize that as being not strong enough.  It is a very steep rise in the A-D Line, which conveys the statement that liquidity is plentiful.

Divergences between the A-D Line and prices are a big deal for me.  And I should emphasize that bearish divergences are much more valid and important than bullish ones.  Sometimes the A-D Line can lag just a bit at turning up from an oversold bottom.

The fun thing about the A-D Line is that every stock gets an equal vote.  Some see this as a flaw, asserting that tiny little companies which do not matter much should not get the same vote as big stocks which drive the cap-weighted indices.  But this supposed flaw is actually a strength.

When liquidity starts to dry up, it hits the most vulnerable stocks first.  So if one looks only at the healthiest of stocks, one can miss the message that liquidity is starting to be a problem.  This is why I do not pay much attention to the A-D Line for the stocks which make up the SP500.  Those stocks are all on the varsity team, and they are not as vulnerable to liquidity problems as the more marginal companies' stocks.

I have also found that the NYSE's A-D Line works better at giving these messages than other A-D data.  The Nasdaq's A-D Line, for example, has a profound bearish bias, and always has.  It has never made a new high in its entire history.  It started going downward from the beginning of the data in 1972, and has kept going down.  This is because the Nasdaq exchange has looser listing standards, and so a company that is going to IPO then go broke is more likely to do that on the Nasdaq.  And every day it spends declining from its IPO price to zero will contribute to the Declines column.

I also find that the daily A-D Line for high yield corporate bonds can serve as an excellent indication of liquidity problems, or lack of problems.  Those corporate bonds are marginal investments, and they draw from the same pool of liquidity that the stock market does.  So when the HY A-D Line starts diverging, it is a great warning of trouble.  Right now, there is no sign of that trouble from that indicator.

I feature the NYSE's A-D Line in every issue of our twice monthly McClellan Market Report.  And I show the HY Bond A-D Line periodically there and in my Daily Edition when it has something useful to say.  If it as been a while since you were a subscriber, or if you are curious about learning more, check out our web site, www.mcoscillator.com.

Several analysts I respect have noted that the price structure in 2026 is looking a lot like that of 2025.  This week's chart highlights that point, but I had to do a little bit of tweaking to get the best pattern fit. 

Whenever I employ price pattern analogs, I strive to find the best overall fit of even the minor price wiggles.  That helps to validate (or not) that I have a good pattern comparison.  To get the best fit in comparing 2026 to 2025, I had to do a little bit of time shifting, so that there is not a perfect calendar alignment.  Doing this better aligns the lesser movements in the left hand of the chart.  The plots are shifted 15 trading days apart from a strict calendar alignment in order to get the better pattern match.

To help illustrate this point, here is the same chart with the two plots arrayed with a strict calendar year alignment in the chart:
We can still see a general similarity of the chart structures, but not as tight of a match of the smaller dance steps. 

I prefer to use a comparison with the best alignment of structures that I can get, even if that goes against strictly following calendar seasonality.  The presumption of analyzing any chart structure is that price structures are a manifestation of the unfolding of human emotions as they get applied to price movements.  Similar emotional scenarios result in similar price structures, or so goes the theory.  There is no perfect way to test that hypothesis. 

Anyone employing chart pattern analogs should also be aware that all of them will eventually break correlation.  In my experience, the moment when they break correlation is usually the moment when one is counting on the relationship most to continue working.  So I never give them my full confidence because of that inherent fickleness. 

Turning our attention back to the top chart, if the current year's pattern continues to match that of 2025, then we can expect a nice uptrend lasting through the summer of 2026.  One problem with that expectation is that the stock market typically moves sideways to lower from May to October, and especially in the second year of a presidential term.  So we will have to see which factor is stronger, i.e. the feeling of relief after the Iran War cease fire matching the relief after the April 2025 tariff crash, or the normal seasonal tendencies.  For the moment, the market is still matching the 2025 time-shifted pattern, and we are still in the bullish phase of normal annual seasonality.  The conflict does not arrive until May gets here.

When my parents wrote Patterns For Profit in 1970 to introduce the McClellan Oscillator and Summation Index to the world, they emphasized the point that the patterns in the structure of the indicators mattered more than the numerical levels.  One of the patterns that they highlighted was seeing a 3rd bottom in the Oscillator, higher than the prior two.  We just had one of these, conveying a big bullish message for the stock market.

The McClellan Oscillator measures acceleration in the Advance-Decline data.  Negative readings mean that there has been downward acceleration.  The observation that there are subsequent higher negative readings after the lowest one means that the downward acceleration is slowing.  That is a precursor to a change of price trend.  That is the setup, but not a signal by itself.

The confirmation this time came first with the Oscillator breaking its own declining tops line.  Then more confirmation came with the Oscillator going up above the zero line, signaling a switch to upward acceleration.  That does not necessarily mean upward movement for the A-D Line, since a change from downward to sideways constitutes upward acceleration in physics terms. 

The Oscillator can next give us additional confirmation of a bullish change by zooming up to a really high positive reading.  Very low negative readings are usually conclusive in their nature.  But a very high positive reading is a sign of strong initiation of an uptrend.  We have not yet gotten that piece of confirmation.  It will be something to look for in the days ahead.

We feature a chart of the McClellan Oscillator in every issue of our twice monthly McClellan Market Report, and frequently in our Daily Edition.  You can see a copy of the chart above every day at the "Breadth Data" page at our web site.  And you can read more free articles about the McClellan Oscillator here: https://www.mcoscillator.com/learning_center/kb/mcclellan_oscillator/

There is a fun magic trick which the McClellan Summation Index does when it crosses through its neutral level, which is the topic of this week's chart.  The moment of crossing down through neutral tends to mark at least a temporary bottom for the price decline which brought about that crossing.  This also happens crossing back up through zero to mark tops, but to a lesser extent.

The Summation Index chart shown above is the "classic" version, first created by my parents Sherman and Marian McClellan back in 1969.  It is different from the Ratio-Adjusted Summation Index (RASI) which we use for some other applications in that first, it does not adjust for the changing numbers of issues traded like the RASI does.

The classic version is also different in that its neutral level is at +1000, and there is a brief story about this.  When my parents first introduced the McClellan Oscillator and McClellan Summation Index to the world, there were no computers outside of large universities and the Pentagon.  So anyone wanting to use these tools had to do all of the math by hand.  The Summation Index changes each day by the value of the McClellan Oscillator, and some users would get confused about having to "add" a negative Oscillator reading to a negative Summation Index value.  So to help avoid this computational confusion, they artificially boosted the neutral level to +1000.  At that time in the early 1970s, the Summation Index had a total amplitude from highs to lows of only about 2000 points, so boosting the neutral level to +1000 meant that it would be very rare (and thus special) to see a negative Summation Index reading.

We thankfully have computers now to do our calculations for us, and the computers are not bothered nor confused by arithmetic operations.  But we have kept the +1000 neutral level for the classic version of the Summation Index out of custom, and also to prevent confusion for anyone who has been using it for years.  In reality, we could have the neutral level be any number we want and it would not make any difference.  But the principle of there being a neutral level is important for this week's lesson.

Crossing down through zero offers us this fun magic trick of marking a bottom for prices.  It may not be a permanent bottom, but it is a noticeable one.  This is not new; it has been working this way for decades, and it is just one of the special features of the Summation Index, and also of the McClellan Price Oscillator for price-based indices.  In fact, the Summation Index is mathematically like a Price Oscillator for the daily A-D Line.  The Price Oscillator is calculated as the difference between a 10% Trend and a 5% Trend of closing prices.  The Summation Index is calculated by summing McClellan Oscillator values, but it also can be arrived at by finding the difference between a 10% Trend and 5% Trend of daily A-D Line values.

This magic trick of marking price bottoms when crossing down through neutral is similar to the topic of a "rainbow convergence" of moving averages that I have written about before.  See the prior article linked below.  I cannot explain why this magic trick works, but it is a phenomenon with enough history to demonstrate that it does work even if one cannot explain it.

I wrote here in February about how the Investors Intelligence bull-bear spread had gone to an extremely high level.  It has since come back down again, as the war with Iran has stock market analysts rethinking their optimism.

The Investors Intelligence survey of investment advisors and newsletter writers (including me) comes out every week.  Like most sentiment surveys, it responds to what prices are doing.  There is nothing like an uptrend to get people excited, and selloffs tend to make people turn bearish.  That is not true for all survey respondents, obviously, and some of us work to trade against what the crowd is doing. 

The remarkable point about the Investors Intelligence survey data is just how precisely it matches the price action, most of the time.  The chart above shows the SP500 on a detrended plot, meaning it is a comparison of where the index value is versus its 200-day simple moving average.  Making this simple adjustment results in both plots staying close together most of the time.

Occasionally, though, the sentiment data overshoot what prices have done.  I call that a manifestation of "pure" sentiment as opposed to price-induced sentiment.  When we see these sentiment overshoots, the message is that the crowd has gone too far in either its enthusiasm or its pessimism, and the stock market usually responds by moving away from that overshoot.  That movement tends to continue until sentiment oversteers to get back in step with what prices are doing, and we have seen that happen just now. 

Three weeks ago, the bull-bear spread was at +40.8 percentage points.  The latest data show that it is down to +18.5, which is a really big reappraisal of the market's prospects.  A war breaking out can do that.  When sentiment lurches like this, it is usually a great opportunity to bet against the crowd.

On a separate note, I recently did an interview with Lars Von Thienen of the Foundation for the Study of Cycles, https://cycles.org.  We discussed the sunspot cycle and its impacts on the financial markets, a topic I addressed here back on Feb. 26, 2026.  You can find a link for that 46-minute interview at our Home page, https://www.mcoscillator.com/.

The war against Iran has pulled down the major US indices, and it has hit some overseas markets even harder.  One encouraging sign, though, is that there is no bearish divergence apparent yet between EEM and the SP500.

EEM is the "emerging markets" ETF sponsored by iShares.  It supposedly invests in stocks in new markets that are just getting going, although its two largest holdings are Taiwan Semiconductor (TSMC) and Samsumg.  One could argue that both Taiwan and South Korea are already past the "emerging" stage.  TSMC alone accounts for a 13% weighting in the fund.

What I find interesting about EEM is that it pretty reliably shows us a bearish divergence versus the SP500 at important price tops.  It is not showing us one right now.  The chart above shows us that there were big divergences at each of what I would call the major price tops of the last 5 years.  And that includes just before the 2020 Covid Crash at the left end of the chart. 

Divergences are important, although one big problem is that they can persist for a long time before they finally decide to matter.  A divergence is "a condition, not a signal".  It will not tell us when it is going to matter.  And divergences can sometimes get "rehabilitated".  That is possible.

EEM shows divergences, much like the NYSE A-D Line does, when liquidity starts to get tight.  Emerging markets stocks are arguably of lower quality than the big "blue chip" stocks which make up the major averages.  If there is plenty of money to go around, then even the less deserving ones can get some.  When liquidity starts to dry up, the weak get hurt first.

We do not have a divergence now, and that is perhaps even more important of a message than if we did have a divergence.  The lack of a divergence carries the message that the recent price top is likely not the end point for the uptrend.  There can be stumbles along the way, and the start of a war which shuts down shipping from the Persian Gulf is arguably a big reason for a stumble.  But the message is that liquidity was strong going into that event, and so prices will likely recover.  Whether the EEM can recover to a new high remains to be seen, and so we could be seeing the commencement of building a divergence but we do not have it yet.

The shutdown of shipping through the Strait of Hormuz into the Persian Gulf has understandably brought a big spike in the price of the near month crude oil futures contract.  The contract for April 2026 delivery is now above $90.  But if you go out 11 months to the March 2027 contract, the price is still at $66.50.

The condition of having the near month contract price above the far month contracts is known as "backwardation".  The opposite condition is called "contango".  Financial futures like T-Bonds and SP500 futures almost always trade in contango, with the more distant contracts at a higher price, thanks to something called the "time value of money".  With the right portfolio engineering, a trader could go long the distant month and short the near month, thereby having a neutral position relative to the market, but collecting "interest" on the difference in contract premiums.

In crude oil futures, we have seen persistent backwardation ever since the 2020 Covid Crash.  Anyone with some available supply of crude oil therefore has a big incentive to sell it now and collect the higher price, versus holding onto it for future delivery.  Years ago, there were several instances when oil futures went into contango, and so oil producers had an incentive to buy storage tanks to hold their current production.  When contango has gotten really extreme, some traders have even rented crude oil tanker ships to act as floating storage in order to take advantage of the difference in pricing in the futures market. 

There are several messages to take from the current condition of backwardation.  The first is that crude oil consumers like oil refineries are desperate to secure any oil which can be delivered, so that they can keep operating.  With the unknown outcome of Iran's announced closure of the Strait of Hormuz, buyers are in a panic. 

The big spread to the distant month contract prices also tells us that crude oil futures traders do not see the current panicked condition as being permanent.  The distant month contracts have come up a little bit since the US-led invasion of Iran began, but that rise is not nearly as big as the near month. 

This creates the big spread you see in the chart above, and these episodes of very large backwardation are great markers of tops for oil prices.  Normally a spread of about $7/barrel is enough to mark a top, although it obviously can get to a much bigger spread like what we are seeing today. 

The high backwardation spread we are seeing now is a topping condition, but it is not a "signal" to say that a top is in.  Oversold or overbought conditions always have the potential get even more extended.  But this current spread is an interesting bit of history, and a sign that this big spike in near month crude oil prices should not last very long.

Every president wants to have improving jobs numbers, and they all do what they think is necessary and appropriate to achieve that. This includes President Trump, but he is fighting against a more powerful force with a long track record.

It turns out that the sunspot cycle has a very strong correlation to the U.S. unemployment rate (U-3), after one adjustment.  This week's chart compares those two sets of data, with the plot of sunspot numbers shifted forward by 3 years.  The bottom of each sunspot cycle coincides nicely with a bottom for the unemployment rate 3 years later.  And peaks in unemployment tend to echo peaks in the sunspot cycle, with that same 3-year lag time.

The peak month for this current sunspot cycle was in August 2024, so if we count forward exactly 3 years from that we get August 2027 as a projected top for the unemployment rate.  It does not always work out exactly to the month, but that is a pretty good rule for planning purposes.

It is also important to understand that sometimes outside events come along which disrupt this nice correlation.  Covid was an obvious one, and we cannot really blame this sunspot model for not predicting the 2020 spike in unemployment.  There was also an exogenous spike in 1954, after the end of the Korean War brought a slowdown in government defense spending and a rise in unemployment.

Perhaps the biggest example of an exogenous event was when the Fed in the mid-2000s kept interest rates too low for too long, fueling the housing bubble and its eventual collapse.  That led to the "Great Financial Crisis (GFC)" in 2008, which did not fit the timing suggested by the sunspot data.  But by the late 2010s, the two data plots were back into correlation again.

Whenever I show a comparison of data like this, I always get asked what factor could explain the relationship.  Wanting to know that is a natural human tendency.  We all want to know the "why".  It is not always necessary, though, to get after the "why" if the "is" can be established well enough.  We have good data on unemployment rates going back to 1947.  Over that period, it is a well-established correlation, except for that handful of exogenous events I mentioned above.  It is pretty clear that something is going on here, even if we cannot explain what it is.

Some analysts have hypothesized that sunspots affect the weather, which affects agriculture and thus somehow flows through to unemployment.  But the data on sunspots and agricultural production do not match up as well as these data.  Another theory is that the charged particles emitted during higher sunspot numbers affect the wiring in human brains, causing us to collectively change our moods.  That is an interesting hypothesis, but I know of now way to construct a satisfactory experiment to test that one.

I have personally seen enough evidence of this relationship working for long enough that I can accept its validity even if I cannot explain it.  Others may need more evidence.  I invite you to ask yourself, for how much more time would it need to keep "working" before you could accept that it is valid?

The last sunspot cycle bottom was 2018-2020, and the period is about 11 years.  So we can figure on the next sunspot minimum being due 2029-31, and thus the low for the unemployment rate 3 years later in 2032-35.  We do not have the full data for February 2026 yet, but there are already several days in February with zero sunspots.  It can be argued that this cycle is quieting down faster than normal.  If so, that could matter for improvements in jobs data, but not until the 3-year lag time goes by.

The Federal Reserve is now doing QE5, although we are not supposed to officially call it that yet.  All four previous rounds of quantitative easing (QE) have been unquestionably bullish for the stock market.  But the same cannot be said for T-Bond prices.

During all 4 previous QE episodes, T-Bond prices have seen a dramatic drop, which also means a rise for long term yields.  So if QE5 is going to unfold in the same way, then we can look forward to higher long term yields on T-Bonds, and presumably also on home mortgages and other long term debt.

One factor which is different this time is that bond prices have not just seen a big rally leading up to the start of QE5.  All four previous episodes saw QE starting as bond prices were at a spike top.  So there was a lot of room each prior time for bond prices to give back the big gains they had just made.  This time there has only been a small rise in T-Bond prices.

Why would QE be bad for bonds?  This does not make intuitive sense.  After all, if the Fed is going to step in as a new incremental buyer of Treasury debt, then that additional demand should mean upward pressure on prices, and yet we see it works the opposite way.

Part of the answer is that the Fed's purchases of Treasury debt during prior rounds of QE have tended to be in shorter maturity instruments, T-Bills and T-Notes.  So they are not adding very much direct additional buying pressure at the long end of the yield curve.

QE also acts as an economic stimulus, which in theory drives up demand for longer term credit from businesses looking to finance expansion.  That puts upward pressure on longer term rates. 

That is all great in theory, but what is still unknown is if it can work this same way for a 5th time without having the big run up first in bond prices. 

Recent data from the Investors Intelligence weekly survey of investment advisors and newsletter writers showed a very high bull-bear spread.  That spread is simply the numerical difference between the percentage of respondents classified as bullish versus those who are bearish.  High readings show extreme confidence, which every card-carrying contrarian knows is a sign of a top for stock prices.

But before you go placing any sell orders just based on this one indication, a few caveats are important to keep in mind.  The first is that any overbought reading on any indicator constitutes just a "condition, not a signal".  Sometimes an overbought condition can go on to get even more overbought.  It is also worth noting that in many cases, the stock market has been able to continue trending higher in spite of such a reading, especially in a strong bull market. 

That point is worth pausing to reflect upon.  If the stock market can ignore an overbought condition and keep trending higher, that is a sign you are in a strong trend.  And that is really useful information to find out.  Very low readings are much more reliable in terms of marking focused bottoming events.

A more concrete indication that we are seeing right now is the lack of a divergence versus prices.  This week's chart looks back 10 years, and there have been a few really noteworthy price tops over that period, plus a whole lot of lesser tops which did not matter much.  The really big price tops saw a divergent lower top on this bull-bear spread versus prices.  The one big exception to that "rule" was in 2020, when Covid took everyone by surprise.  So to see no divergence now means that we can have pretty high confidence that the market now is NOT at a really important top.  It may still be making a lesser top, but not a major one.

One point about these data from Investors Intelligence is that they seem to track with prices more precisely than other sentiment survey data that I have researched.  I noticed this point years ago, and sought a way to track that even better.  What I came up with was using a detrended plot of the SP500 instead of the raw price data.  To do that, I calculated how far the SP500 was away from its 200-day simple moving average.  This chart shows that comparison:

This chart uses a shorter lookback period, and compares that same bull-bear spread to that detrended plot of the SP500.  With a little bit of scaling adjustment, we see that the two plots are almost identical most of the time.  In other words, sentiment responds to whatever prices are doing.  A rising market gets people more bullish, and vice versa.  Occasionally, though, we see a big disagreement when the bull-bear spread seems to overshoot what prices have been doing.  In other words, survey respondents are getting more fearful or more confident than prices say that they should have done.  That usually marks a turning point for prices.

In the current moment, that turning point has taken the form of having the SP500 lagging behind what other parts of the market are doing, causing the detrended price plot to fall.  Sentiment should soon catch up (i.e. down) to what prices have been doing.

One last point to notice in that lower chart is that the zero levels on the two Y-axes are offset.  This was done in order to get the best fit of the two plots, and it reveals something.  There is about a 15 percentage point bullish bias in the Investors Intelligence survey data versus how prices behave.  I would argue that this bullish bias is appropriate, because in the long run the stock market does tend to go up.  It is just interesting to quantify what that bias amounts to in these data.

The NYSE's daily A-D Line just made a new all-time high on Feb. 4, 2026, which is a statement that liquidity is plentiful.  But just a day later, we have gotten the 3rd Hindenburg Omen within 6 trading days.  This is a message that for all of that supposedly plentiful liquidity, the market has some serious problems.

The late Jim Miekka created the Hindenburg Omen signal back in 1995.  He intended it as an improvement on the late Gerald Appel's "Split Market Sell Signal", which occurred any day that saw both New Highs and New Lows on the NYSE exceed 45 issues.  Miekka saw the problem of not adjusting that for the increased number of issues traded.  And he wanted to add some other filtering rules to get better signals.  An important one is that the market has to be in an uptrend to give a signal.  Miekka thought it was not very useful to get such a warning once stock prices are already in a downtrend.  The criteria for a Hindenburg Omen signal are listed in the chart above. 

If you do an Internet search, you will likely find other sets of criteria at some technical analysis web sites, which is unfortunate.  And because Miekka has passed away, we cannot get him to correct the record, so that task is left to others.  The criteria I use are those which Miekka personally told to Greg Morris, who recorded them in his 2006 book, "The Complete Guide To Market Breadth Indicators". 

The basic idea for the Hindenburg Omen signal is that during a normal uptrend, there should be more stocks making New Highs than making New Lows.  That is the normal condition.  If you get a condition where the uptrend is still underway, but the numbers of New Lows start perking up, then that is a sign of trouble.  What constitutes "perking up"?  That's a matter for opinion, and perhaps back-testing, but Miekka set his threshold at both NH and NL being greater than 2.8% of Advances plus Declines on the same day.  Other analysts might choose a different criterion, which people are perfectly allowed to do.  But for the sake of consistency, and so as to avoid confusion, I stick with Miekka's criteria.

A single signal is interesting, but the message gets more compelling when we see clusters of multiple signals in a short time frame.  We had a grouping of 5 signals from Oct. 29 to Nov. 13, 2025, but the market shrugged.  Now we have 3 more (so far) and would have had a 4th on Feb. 4, 2026 except that the NYSE's McClellan A-D Oscillator was just barely positive that day.

I have found that using a 6-month lookback period to count cumulative Hindenburg Omen signals is useful.  On that basis, we are now up to 8 signals, which is a pretty high reading.  Some noteworthy market tops have come from big readings like this.  But we must also note that there have been other times when we got a bunch of H.O. signals and nothing much happened.  That is possible.  A Hindenburg Omen signal, or a bunch of them, is a warning but not a guarantee of trouble.  It says "pay extra attention".

The last such cluster appeared at the end of 2024, right after the presidential election.  It did not tell us exactly what trouble was brewing, but it did appear just ahead of the stock market's violent drop on tariff worries courtesy of President Trump.  A lesser cluster appeared in early 2022, ahead of that year's bear market. 

It is noteworthy that there was a cluster of 10 signals in 2013, and the uptrend just powered on through.  It is important to remember that the Fed back then was doing QE3, and QE can paper over lots of problems.  The Fed now is doing yet another round, QE5, although not as vigorously as some past QE episodes.  So it is possible that the Fed will paper over the trouble again this time.

Now that the Fed has cut short term rates several times, we are seeing a further steepening of the yield curve.  It is tough to portray the entirety of changes in the whole yield curve over time, and so I am summarizing it in this week's chart by showing just the spread between 10-year and 3-month Treasury yields. 

The 10y-3m spread does a really cool trick, which this week's chart illustrates.  Changes in that 10y-3m spread draw us a roadmap for what small cap outperformance or underperformance is going to look like.

The green plot in this chart is the relative strength ratio of the Russell 2000 vs. the Russell 1000.  It is calculated by simply taking the numerical value of the Russell 2000 Index, and dividing it by the Russell 1000.  The line goes up when small caps are outperforming on a relative basis, and it goes downward when large caps are doing better.  And the cool part of this trick is that the green plot makes those movements corresponding to what the 10y-3m spread was doing 15 months prior.  We are getting the answers ahead of time.

They are not perfect answers, though.  The nice predictive relationship can break down when the Fed is putting a thumb too heavily on the scale with actions other than interest rate policy.  QE3 in 2013 caused a temporary inversion of the relationship.  And in 2020 the Fed started QE4 in response to the Covid shutdowns, which also disrupted the nice correlation for a while.

The Fed is doing QE5 now, although (shhhh!) we are not supposed to call it that yet.  But the magnitude of the Fed's recent purchases is a lot smaller than what we saw in prior rounds of QE, so the potential for skewing the leading indication is likely much less now.

The Russell 2000/1000 relative strength line bottomed in July 2025, and has been rising since then as small caps have started outpeforming again after a really long period of underperformance.  Seeing what the 10y-3m spread has been doing, its message is that we should expect small caps to continue outperforming for at least the next 15 months.  We do not yet know the end point of that projected small cap outperformance, because we have not seen a topping out of the 10y-3m spread. 

When technical analysts learn about volume, they are taught that it is useful for helping to confirm (or not) what prices are doing.  This includes hopefully seeing expanding volume to confirm a price breakout.  And a head and shoulders structure should ideally have the heaviest volume on the left shoulder or the head.

For big ETFs like QQQ, it works differently.  As a gross generalization, QQQ volume works as a fairly pure inverse sentiment indication.  High volume is a sign of a bottoming condition, while a low volume day is a marker of a top.

Low volume days can be problematic, though, because sometimes volume is low due to holiday-light trading.  So one must employ at least a mental filter when evaluating low volume readings.

This week we saw a pair of high volume days, brought about due to the market's reaction to President Trump's announcement about imposing tariffs on European countries who won't go along with plans for the US to annex Greenland.  That news understandably got traders excited, although it was pretty quickly walked back when President Trump announced in Davos that he had arrived at a "framework of a deal".  Prices rebounded on the news.

But the sentiment message had already been posted.  The Jan. 20 selloff brought a big volume spike in QQQ, high enough to be a decent bottom marker, although clearly not the highest volume ever.

This phenomenon "works" because during selloffs, some traders turn to QQQ versus individual stocks for its greater trading liquidity.  It also serves as a shorting vehicle for those wishing to hedge overall portfolio risk but without turning to the futures market.  Similarly, when everyone is feeling happy and complacent, they don't trade QQQ as much and so a low volume day can be a marker of a price top. 

Getting an indication like this week's pair of high volume days does not tell us much about how far prices will move on a rebound.  That can vary a lot.  It just says that a bottom-worthy sentiment indication has announced itself.

We have just seen a new all-time high in the NYSE's daily Advance-Decline (A-D) Line.  It confirms the new highs in the major indices, and that is a really good thing.

Years ago, I undertook a study to see what it means to have a new high in the A-D Line.  In order to increase my sample size, I specified that it had to be just a new 3-year high instead of all time, and then I looked at what the max drawdown was in the SP500 over the succeeding 3 months.  The recent decades' results are shown in this week's lead chart. 

The short version is that if you see a new A-D Line high, you have pretty good assurance that the biggest drawdown you are likely to see over the next 3 months is limited to about 10%.  But the big caveat is that you can throw this rule out the window if the Federal Reserve or Congress puts a thumb on the scale. 

After the bottom of the Great Financial Crisis (GFC) in March 2009, the Fed decided to throw money at the banking system with its first ever round of "Quantitative Easing" or QE1.  They ended that suddenly in April 2010, and what followed came to be known as the "Flash Crash" on May 6-7, 2010, when liquidity dried up all at once and bids just disappeared.  Some blue chip stocks traded for pennies, and those trades had to get wiped out by the stock exchanges.

The Fed honchos saw that they had not fully fixed the liquidity problems with QE1, and so they started QE2 in August 2010, and everything was great.  Or at least it was great until the Fed once again ended QE2 very suddenly in June 2011, and we saw a 19% drop. 

The FOMC learned its lesson, and when they decided to end QE3 in late 2014, they wisely "tapered" their slowdown of purchases of Treasury and mortgage debt.  That allowed the banking system and the financial markets to adjust more slowly to the withdrawal symptoms.  We still saw a market correction in 2015 because of ending QE3, but it was more muted.

We also had a big exception to the rule about new A-D Line highs back in 2020, when the arrival of Covid into the US led to a government shutdown.  Remember "2 weeks to flatten the curve"?  That brought a big drawdown in April 2020, which I named the "Covid Crash" and others took up that label.  That was a genuine exception about the supposed assurance of limited drawdowns from a new A-D Line high.

This principle has also worked further back in time.  Here are a couple more charts looking back further in history, so that we can see that the chances of a big drawdown are pretty limited for at least 3 months after a new A-D Line high. Note that the big time gaps between the red drawdown bars reflect periods when there were not any new 3-year A-D Line highs.

There was one notable exception in October 1978, when the SP500 saw a 13.5% drawdown.  There was a dollar crisis then, thanks in part to rising inflation rates.  It all came to a climax on Oct. 24, 1978, when President Carter announced a tougher anti-inflation program to stabilize prices and the currency.  The SP500 was at 97.49 that day, and continued to fall further to a closing low of 92.49 on Nov. 14, 1978 before starting to rise again. 

The key point to all of this historical analysis is that seeing a new A-D Line high is bullish news, and it offers us "some" assurance that we won't see a big ugly bear market for a while.  The handful of exceptions help to illustrate how there are no guarantees in the financial markets, especially when folks in Washington, DC decide to try to "help".

We are still in the "best 6 months of the year", due to last until early May 2026, and so the current uptrend ought to continue.  But we should also expect a meaningful dip this month, based on the DJIA's Annual Seasonal Pattern (ASP) shown in this week's chart.

I created this ASP by chopping up the daily data into 1-year chunks, restating their value for each year to a starting value of 1.00, then averaging them together as the percentage change from that starting value rather than the numerical value of the index.  Doing that revaluation allows for more accurate averaging than using the raw numerical values, because with raw values the later years with higher index numbers would have a bigger weight. It puts each year on an equal footing.

I also leave out the entire year of 2020, because the Covid Crash was such an abnormal event and a very large one, such that it skews the data.  The point of creating such an average is to depict what "normal" looks like, and the market's response to the Covid shutdowns and then the Fed's monstrous QE was definitely abnormal.  Throwing it out is like discarding the highest and lowest judges' scores in platform diving or figure skating competitions.

One big factor affecting the use of the ASP this year is that there have been several big news events which have pushed the stock market around in ways not reflected in the ASP.  We have not had one of those events for a while, and the market is behaving itself pretty well in tracking the ASP.  But that could change at any moment.  One might argue that the stock market was a lot more sensitive to news events early in President Trump's current term, but stock traders are getting more accustomed to tumultuous news out of Washington DC, and thus they are less reactive to it now.

The month of January typically sees a mid-month dip, with a top ideally due Jan. 10, and then a bottom ideally due Jan. 23.  After that bottom, the market on average gets back to trending higher.  And it could get a boost from the QE5 which the Fed has now restarted, although they are not calling it that yet.  Since the last FOMC meeting, the Fed's Treasury holdings are up by $46 billion.  They are still rolling off mortgage backed securities (MBS), but the net effect of both factors is still a $32 billion rise in the Fed's balance sheet.  That is a bullish tailwind for the stock market.

Whether that QE can paper over the normal seasonal tendency for a January dip is what remains to be seen.

The Annual Seasonal Pattern is one of the indicators I feature regularly in our twice monthly McClellan Market Report and our Daily Edition. https://www.mcoscillator.com/market_reports/

The University of Michigan has been conducting its "Survey Of Consumers" since 1955. The recent data have been among the lowest of their published readings, meaning that survey respondents are not feeling all that great about the supposed 4.3% rate of US GDP growth.  Perhaps people are more attuned to indications like manufacturing employment being in a decline since May 2024.  And thus far, the low gasoline prices have not gotten any traction in the hearts of consumers.

Lows in consumer sentiment measures tend to be bad for whoever is in the White House, and to some extent this applies to Congress as well.  The UMich survey data reached what was then an all-time low in 1980, which doomed Jimmy Carter's reelection chances and led to Ronald Reagan's victory.  A similar low reading in 2008 saw a change from Republican control under Bush 43 to President Obama in a decisive popular vote and electoral college victory.  There was another very low reading in 2022, which led to a change in Congress, and eventually to Donald Trump winning another term in 2024, even though the survey numbers had recovered some by the time of that election.

So the Republicans currently in control of the White House and both houses of Congress will want to see these consumer sentiment numbers start to rebound.  That is the point for politicians to worry about.

The point for investors to worry about is that as consumer confidence numbers start to rebound from very low levels like this, it tends to be extremely bullish for the stock market.  Usually that rebound occurs from a simultaneous low in stock prices along with the consumer sentiment data.  It is a rather unusual condition to see extremely low survey readings like this happening with stock prices at new all-time highs.

Part of the reason why low consumer sentiment are bullish is that such readings usually arise from a condition of seeing a bad economy, which the Federal Reserve then steps in to do something about.  We have had 3 rate cuts in 2025, and now the Fed has started up QE5 (although they are not calling it that yet).  Those are bullish conditions for both the stock market and the actual economy.  It should lead to a rebound in the consumer confidence data, and that should help keep the bull market going.

Back in September I reported on how the data on US trucking from Cass Information Systems was not looking good for the stock market.  Since then the message has gotten a little bit worse, as seen in this week's chart.

Cass tracks trucking data in a lot of different ways, and publishes their own indices.  This week's chart shows the 12-month percentage changes in their Shipments and Expenditures indices.  Both measure trucking activity, but in different ways.  And what is important for our purposes is that these trucking data are well correlated to the stock market.

So it is NOT good news for the stock market to see that both of these indices are down year-over-year.  That is a condition reliably associated with bear markets for stock prices.  A couple of points are worth noting, though.

The first point is that Covid really screwed up the trucking industry, first by sending everyone home for "2 weeks to flatten the curve".  Then while everyone was home, they all decided to order a bunch of stuff online, resulting in big delays at US ports getting container ships in, and getting trucks to haul off those containers.  It was quite the bust-boom-bust cycle for truckers, and arguably the Cass Shipments Index (red line) still has not recovered.  The green Expenditures Index growth rate did go up above the zero level, but it is back down to below zero, showing shrinkage versus a year ago.  Having both of these at or below zero is not a good condition for the stock market.

EXCEPT!!  Point 2 is that when the Fed is doing QE, as they just started up again in December, the stock market can ignore that condition.  We saw in 2012-14 in the middle of the chart that during QE3 both of these indices were showing zero growth, and the SP500 just kept on trending higher.  It turns out that Fed money printing smooths over a lot of problems, for the stock market at least.

We know from how things turned out with QE1 through QE4 that while QE is happening, it is very bullish for stock prices irrespective of what the actual economy might be doing.  So earnings matter much less, and trucking data too.  We also know that shutting off the QE fountain tends to lead to bear markets, but that is a problem for another day.

Gold prices are making a repeating stair-step pattern of movements as gold trends higher.  Each part of this sequence sees a multi-month up move after breaking out to higher highs, then another multi-month consolidation of those gains. Right now gold is in a consolidation phase.

This week's chart shows our Proportional Price Oscillator, which is a variation on the classic McClellan Price Oscillator.  The Price Oscillator which my parents developed back in 1969 is similar to the McClellan A-D Oscillator, which measures the difference between two EMAs (10% Trend and 5% Trend) of the daily A-D difference.  The math is the same for the Price Oscillator, but using closing prices in place of that A-D difference.

Sherman and Marian McClellan were the first to do this mathematical trick of finding the difference between two moving averages, as opposed to just looking at each moving average on its own.  In 1977, the late Gerald Appel adapted this technique to create what he called Moving Average Convergence and Divergence, or MACD.

The Proportional Price Oscillator (PPO) goes one step further, dividing the Price Oscillator by the closing price (and then multiplying by a constant to get to normal sized values).  Doing this helps to normalize what would otherwise be expanding amplitudes of Price Oscillator values owing to higher price levels over time.  The PPO is thus better for longer term chart evaluations, especially if there has been a big change in price levels.  I was the one who innovated this change to the Price Oscillator back in the 1990s, and since then platforms like StockCharts.com have added PPO to their arsenals of indicators you can use.  By adjusting the Price Oscillator for higher price levels, the PPO is therefore equivalent to the Ratio-Adjusted Summation Index (RASI) for A-D data.

I am including the PPO for gold futures in this week's chart in order to make an important point about gold's stair-step pattern of surges and consolidations.  What we are seeing during the consolidation phases is that the PPO works its way back down to near the zero neutral level, which helps reset the gold market to a nice equilibrium state, the better to support the launch of the next ascending phase.

That is important right now because gold's PPO has only made it partway down to neutral, which implies that there is more consolidative work to do.  And this sideways period has only been underway for 2 months, since the Oct. 20, 2025 all-time price high.  Prior consolidation phases have lasted 3-5 months.  There is no mandate that it has to happen exactly the same way this time, but this pattern seems to be repeating nicely so it would be strange if we got a different behavior this time.

Along the way, it is quite possible that we can see gold prices make an incrementally higher price high, as it tries to convince everyone that the next breakout is starting.  Then after a few months, when traders finally give up hoping for a legitimate breakout move, the real one can start.

Years ending in 5 are up years almost without exception through the entire history of the DJIA.  The two exceptions were 2005, when the DJIA fell 0.6%, and 2015 with a 2.2% drop.  Thus far in 2025, the stock market is back to its normal year 5 behavior, with a 12.5% gain through Dec. 4, 2025.  It would have to be a really bad rest of the month to spoil this year 5's record.

 This week's chart shows the Decennial Pattern for the DJIA, created by chopping the price history up into 10-year chunks of time, and averaging them together on a percentage gain basis.  It is similar in that respect to the Annual Seasonal Pattern and the 4-year Presidential Cycle Pattern.  Putting the data together this way allows us to see what typically happens over the course of each decade.

Coming up is the more problematic year 6 of the decade, which is more flat on average.  But that flat part does not start for a while, and we still have more of the uptrend seen in year 5s that continues into year 6.

This next chart shows the same comparison, but zoomed in to see the data better.  I have also adjusted the scaling so that we can get more of each of the plots on the same chart. 

The year 5 bullish effect was interrupted earlier this year when the stock market had a bit of indigestion after hearing of President Trump's tariff plans.  Investors eventually got over that, and the patterns have matched up nicely since about May 2025.  The Decennial Pattern shows that this uptrend should continue all the way to a top due ideally April 6, 2026.  That is when the sideways period of the year 6 is supposed to start.

Real life approximates the average pattern, but never duplicates it exactly.  There is no logical reason why years of the decade should matter and manifest similar behaviors, but this is something which has been going on for decades.  So there must be something to it, even if it does not make logical sense.