Pickup and Letdown

May 8, 2016

Based on ISM’s monthly survey of Purchasing Managers, the CWPI blends both service and manufacturing indexes and gives additional weight to a few components, new orders and employment.  Last month we were looking for an upward bend in the CWPI, to confirm a periodic U-shaped pattern that has marked this recovery. This month’s reading did swing up from the winter’s trough and we would expect to see further improvement in the coming few months to confirm the pattern. A break in this pattern would indicate some concern about a recession in the following six months. What is a break in the pattern? An extended trough or a continued decline toward the contraction zone below 50.

Since the services sectors constitutes most of the economy in the U.S., new orders and employment in services are key indicators of this survey.  A sluggish winter pulled down a composite of the two but a turn around in April has brought this back to the five year average.

Rising oil prices have certainly been a major contributor to the surge in the prices component of the manufacturing sector survey. The BLS monthly labor report (below) indicates some labor cost increases as well.  Each month the ISM publishes selected comments from their respondents.  An employer in the construction industry noted a severe shortage of non-skilled labor, a phenomenon we haven’t seen since 2006, at the height of the housing bubble.

Last week the BEA released a first estimate of almost zero growth in first quarter GDP, confirming expectations.  Oddly enough, the harsh winter of 2015 provided an even lower comparison point so that this year’s year over year growth, while still anemic, is almost 2%.



April’s employment data from the BLS was a bit disheartening.  Earlier this week, the private payroll processor ADP reported job growth of 150,000 in April and lowered expectations for the BLS report released on Friday.  While the BLS estimate of private job growth was slightly better, the loss of about 10,000 government jobs, not included in the ADP estimate, left the total estimate of jobs gained at 160,000. The loss of government jobs is slight compared to the total of 22 million employed at all levels of government but this is the fourth time in the past eight months that government employment has declined.

A three month average of job growth is still above 200,000, a benchmark of labor market health that shows job growth that is more than the average 1% population growth  With a base of 145 million employees in the U.S, a similar 1% growth rate in employment would equal 1.5 million jobs gained each year, or about 125,000 per month.  To account for statistical sampling errors, the churn of businesses opening and closing, labor analysts add another 25,000 to get a total of 150,000 minimum monthly job gains just to keep up with population growth.  The 200,000 mark then shows real economic growth.  In March 2016, the growth of the work force minus the growth in population was 1.2%, indicating continued real labor market gains.

Job growth in the core work force aged 25 -54 remains above 1%, another good sign.  It last dipped briefly below 1% in October.  This core group of workers buys homes, cars, and other durable goods at a faster pace than other age groups; when this powerhouse of the economy weakens, the economy suffers. In the chart below, there is an almost seven year period, from June 2007 through January 2014 where growth in this core work force group was less than 1%.  From January 2008 through January 2012, growth was actually negative.  The official length of the recession was 17 months, from December 2007 through June 2009.  For the core work force, the heart of the economic engine, the recession lasted much longer.

In 2005, a BLS economist estimated that the core work force would number over 105 million in 2014.  In December 2014, the actual number was 96 million, a shortage of 9 million workers, or almost 10% of the workforce.  In April 2016, the number was almost 98 million, still far less than expectations.

Some economists and pundits mistakenly compare this recovery from a financial crisis with recoveries  from economic downturns in the late 20th century.  For an accurate comparison, we must look to a previous financial, not economic, crisis – the Great Depression of the 1930s.

The unemployment rate in April remained the same, but more than a half million people dropped out of the labor force, reversing a six month trend of declines.  It is puzzling that more people came back into the labor force during the winter even as GDP growth slowed.

Average hourly earnings increased for the second month in a row, upping the year over year increase above 2.5%.  For the past ten years, inflation-adjusted weekly earnings of production and non-supervisory workers have grown an anemic .75% per year.  In the sluggish winter of January and February 2015, earnings growth notched  a recovery high of 3%, leading some economists and market watchers to opine that lowered oil costs, on the decline since the summer of 2014, would finally spur worker’s pay growth in this long, subdued recovery.  A year later, earnings growth is about 1.2%, a historically kind of OK level, but one which causes much head scratching among economists at the Federal Reserve.  When will worker’s earnings begin to recover?



A reader sent me a link to a CNBC article  on food insecurity in the U.S. The problem is widespread and not always confined to those who fall below the poverty benchmark. Contrary to some perceptions, food insecurity is especially prevalent in rural areas, where food costs can be 50% higher than urban centers.  How does the government determine who is food insecure? The USDA publishes a guide with a history of the project, the guidelines and questions.  To point out the highlights, I’ll include the page links within the document. The guidelines have not been revised since this 1998 revision.

In surveys conducted by the Census Bureau, respondents are asked a series of questions.  The answers help determine the degree of household food insecurity.  The USDA repeatedly emphasizes that it is household, not individual, insecurity that they are measuring.  The ranking scale ranges from 0, no insecurity, to 10, severe insecurity and hunger. An informative graph of the scale, the categories and characteristics is helpful.

In 1995, a low .8 percent were ranked with severe food insecurity (page 14) . To be considered food insecure, a household must rank above 2.3 (household without children), or above 2 .8 (with children) on the scale.  Above that are varying degrees of insecurity and whether it is accompanied by hunger. (Table)

The USDA admits that measuring a complex issue like this one can provoke accusations that the measure either exaggerates or understates the number of households.  What are they measuring?  Page 6 contains a formal definition, while page 8 includes a list of conditions that the survey questions are trying to assess, and that a condition arose because of financial limitations like “toward the end of the month we don’t have enough money to eat well.”

Page 9 describes the rather ugly pattern of progressively worse food insecurity and hunger.  At first a household will buy cheaper foods that fill the belly.  Then the parents may cut back a little but spare the kids the sensation of hunger.  In its most severe stage, all the family members go hungry in a particular day.

Those of you wanting additional information or resources can click here.



Almost a month ago the giant aluminum manufacturer Alcoa kicked off the first quarter earnings season.  87% of companies in the SP500 have reported so far and FactSet calculates a 7% decline in earnings.  They note “the first quarter marks the first time the index has seen four consecutive quarters of year-over-year declines in earnings since Q4 2008 through Q3 2009.”  Automobile manufacturers have been particularly strong while the Energy, Materials and  Financial sectors declined.  Although the energy sector gets the headlines, there has also been a dramatic decrease in the mining sector.  The BLS reports almost 200,000 mining jobs lost since September 2014.

The bottom line for long term investors: the economic data supports an allocation that favors equities.  The continued decline in corporate earnings should caution an investor not to go too heavily toward the equity side of the stock/bond mix.


(Edited May 11th in response to a reader’s request to clarify a few points.)


January 10, 2016

Happy New Year!

Wait, get rid of the exclamation point.

Happy New Year.

After this week!  What are you kidding me?!  Get rid of the Happy.

New Year.

Ok, that’s better.  The New Year was not so happy when the market started its first day of trading last Monday.  For the tenth consecutive month, manufacturing activity in China contracted, which weighed down commodities (DBC down over 4%), energy stocks (XLE down 7%), emerging markets (EEM down more than 8%) and the broader market, which was down 6%.  Even stocks (Johnson and Johnson, Coca-Cola) regarded as relatively safe dividend paying equities suffered losses of more than 3 or 4%.  Investors and traders were re-pricing future profits and dividends.

December’s powerful employment report buoyed the mood for a short time on Friday morning but traders soon turned their attention again to China and the broader market fell about 1% by day’s end.

Given the decline in stocks, one would guess that the price of bonds, hard hit during the past few weeks, had showed some strong gains.  TLT, a popular ETF for long term Treasuries, gained more than 2% during the week but remains range bound since last August.  Treasuries are a safe haven for risk averse money, but the prospect of rising interest rates mutes the attractiveness of long term bonds.

Growth in the core work force aged 25 – 54 remains strong, up over 1% from last year.  The number of people not in the labor force dropped by 277,000 from last month, a welcome sign.  However, we need to put aside the politics and look at this in a long term perspective.  For the past twenty years, through good times and bad, the number of people dropping out of the workforce each year has grown.

This demographic trend is more powerful than who is President, or which party runs the Congress.  Depending on our political preferences, we can attribute this 20 year trend to Clinton, Bush, Obama, Democrats or Republicans. The job of the good folks running for President this year will be to convince voters that their policies and prescriptions can overcome this trend.  Our job, as voters, is to believe them.

The Bureau of Labor Statistics recently released a report of a ten year comparison of the reasons why people have left the work force.  Based on this BLS analysis, a Bloomberg writer who had not done their homework mistakenly reported that there has been a dramatic increase in the number of 20-24 year olds who had retired.

This “statistic” points out a flaw in BLS and Census Bureau data. BLS data is partially based on the Current Population Survey, or CPS.  Interviewers are not allowed to follow up and challenge the responder.  Both the BLS and the Census Bureau have been aware of the problem for at least a decade but I don’t think anyone has proposed a solution that doesn’t present its own challenges.

Looking at Chart 3 of the BLS report, the percentage of retired 20-24 year olds was .2% in 2004, .6% in 2014. The number of retired 16-19 year olds was .2% in 2004 and .2% in 2014. Do we really believe that there are almost 200,000 retired 16-19 year olds in this country? See page 16 for the BLS discussion of this problem.

Now, let’s put ourselves in a similar situation.  We are 22 and have recently graduated from college and are having trouble finding a job that actually uses our education.  Because of this, we are staying at our parent’s home.  We answer our parent’s landline phone (Census Bureau is not allowed to call cell phones).  Somebody from the Census Bureau starts asking us questions.  In response to the question why we are not working, we are presented with several choices, one of which is that we are retired (see pages 16 and 17)  Sarcastically we answer that yeh, we are retired.  The questioner can probably tell by our tone of voice that we are being sarcastic but is required to simply record our response.  How valid is that response?

Understand that problems of self-reporting and questionnaire design underlie all of the data from the monthly Household Survey, including the unemployment rate. This gives those with strong political views an opportunity to claim that government statistics are part of a conspiracy.  Claims of conspiracies can not be disproved, which is why they are so persistent throughout human history.

Each year some research firms predict a global recession. Ominosity is the state of sounding ominous and this year is no different. Adam Hayes, a CFA writing at Investopedia, gives some good reasons  that he believes such a widespread recession is possible. All of these risks are present to some degree.

What makes me less convinced of a global recession is the strength of the U.S. economy.  Just as China “saved” the world during the financial crisis, the U.S. may play the role of the cavalry in this coming year. Let’s look at some key data from the recent ISM Purchasing Manager’s index.  This is the new orders and employment components of the services sectors which comprise 85% of the U.S. economy.  Growth remains strong.

Recessions are preceded by a drop in new orders and by a decline in employment.  When payroll growth less population growth is above 1%, as it is today, a recession is unlikely.

Let’s climb into our time machines and go forward just 11 months.  It is now December 2016 and the IMF has enough data to make a post-facto determination that the entire world’s economy went into recession in March 2016.  We look at the SP500 index.  Holy shit!  We climb into our time machines, go back to January 2016 and sell all of our stocks.  Missed that cliff.

What about bonds? Should we sell all those?  Darn it, we forgot to check interest rates and bond prices when we were ahead in the future.  Back into the time machine.  Go forward again.  U.S. Fed halted rate increases in March 2016, then lowered them a 1/4 point.  The ECB had kept interest rates negative in the Eurozone and bond prices have stayed relatively flat.  OK, cool.  We get back in our time machines and go back to January 2016 and decide to hold onto our bond index funds.  The interest on those is better than what we would get on a savings account and we know that we won’t suffer any capital losses on our investment during the year.

We take all the cash we have from selling our stocks and put them entirely in bonds.  Wait, could we make a better return in gold, or real estate?  Back in the time machine and back to the future!  But now we notice that the SP500 index in December 2016 is different.  So is the intermediate bond index.  What’s going on?  The future has changed.  Could it be the Higgs boson causing an abnormality in space-time?  Maybe there’s something wrong with our time machine.  But where can we find a mechanic who can diagnose and repair a time machine?  Suddenly the thought occurs to us that a lot of other investors have gone into the future in their time machines, then have returned to the present and bought and sold.  That, in turn, has changed the future.

The time machine is called the human brain.  Each day traders around the world make decisions based on their analytical and imaginative journeys into the future.  The Efficient Markets Hypothesis (EMH) formulated by Eugene Fama and others postulates that all those journeys and decisions essentially distill all the information available on any particular day.  Therefore, it is impossible to beat a broader market index of those decisions.

Behavioral Finance rests on the judgment that human beings are driven by fear and greed which causes investors to make mistakes in their appraisals of the future.  An understanding of the patterns of these inclinations can help someone take advantage of opportunities when there is a higher likelihood of asset mispricing.

Each year we read of those prognosticators who got it right.  Their time machine is working, we think, and we go with their predictions for the coming year.  Sometimes they get it right a second year.  Sometimes they don’t.  Abby Cohen is a famous example but there are many whose time machines work well for a while.  If I could figure a way to fix time machines, I could make a fortune.

Spring is springing

May 10, 2015


The dollar’s appreciation against the euro and other currencies in the first quarter of this year caused a natural slowdown in exports, which has hurt manufacturing businesses in this country.  U.S. products are simply more expensive to customers in other countries because dollars are more expensive in other currencies. The PMI manufacturing survey showed a decline in employment for the month.  The non-manufacturing sector, which is most of the economy, rallied in April.  As I noted last month, the CWPI should have bottomed out in March-April, reaching the trough in a wave-like series that has been characteristic of this composite index during the past six years of recovery.  Any change to this pattern – a continuing decline rather than just a trough – would be cause for concern.

April’s resurgence in the non-manufacturing sector more than offset the weakness in manufacturing. In fact, there was a slight gain in the CWPI from March’s reading.

Employment and new orders in the non-manufacturing sector are two key components of the composite index and leading indicators of movement in the index.  They have been on the rise since the beginning of the year.  While the decline in the overall index lasted 5 – 6 months, this leading indicator declined for only 3 months, signalling a probable rebound in the spring. Now we get some confirmation of the rebound.



Released at the end of the week a few days after the PMI surveys, the monthly employment report from the BLS confirmed a renewal in job growth after rather poor job gains in March.  April’s estimated job gains were over 200K, spurring a relief rally in the stock market on Friday.  Gains were strong enough to signal that the economy was on a growth track but not so strong that the Fed would be in any rush to raise interest rates before September.

March’s job gains were revised even lower to below 100K, but the story was that the severe winter weather was responsible for most of that dip.  As the chart below shows, there was no dip in year over year growth because the winter of 2014 was bad as well.  Growth has been above 2% since September of last year.

During the 2000s, the economy generated plus 2% employment growth for a short three month stretch in early 2006, just before the peak of the housing boom.  The past eight months of plus 2% growth hearkens back to the strong growth of the good ole ’90s.  Like the 90s, Fed chair Janet Yellen warned this week that asset prices are high, recalling former Fed chair Alan Greenspan’s 1996 comment about “irrational exuberance.” Prices rose for another four years in the late 90s after Greenspan’s warning so clairvoyance and timing are not to be assumed simply because the chair of the Federal Reserve expresses an opinion.  However, history is a teacher of sorts.  When Greenspan made that comment in December 1996, the SP500 was just under 600.  Six years later, in late 2002, after the bursting of the dot-com bubble, a mild recession, the horror of 9-11 and the lead up to the Iraq war, the SP500 almost touched those 1996 levels.  An investor who had pulled all their money out of the stock market in early 1997 and put it in a bond index fund would have earned a handsome return.  Of course, our clairvoyance and timing are perfect when we look backward in time.

For 18 months, growth in the core work force, those aged 25 – 54, has been positive.  This age group is critical to the structural health of an economy because they spend a larger percentage of their employment income than older people do.

Construction employment could be better.  Another 400,000 jobs would bring employment in this sector to the recession levels of the early 2000s before the housing sector got overheated.

In the graph below, we can see that construction jobs as a percent of the total work force are at historically low levels.

Every year more workers drop out of the labor force due to retirement, or other reasons.  The population grows by about 3 million; 2 million drop out of the labor force.

The civilian labor force (CLF) consists of those who are employed or unemployed (and actively looking for work).  The particpation rate is that labor force divided by the number of people who can legally work, those who are 16 and over who are not in some institution that prevents them from working.  (BLS FAQ)  That participation rate remains historically low, dropping from 65% five years ago to under 63% for the past year.

That lowered rate partially reflects an aging population, and fewer women in the work force relative to the surge of women entering the work force during the boomer “swell.”  A simpler way of looking at things shows relatively stable numbers for the past five years:  those who can work but don’t, as a percentage of those who are working.  The population changes much more than the number of employed, and the percentage of those who are not working is rock steady at about 66%.  This percentage is important for money flows, the vitality of economic growth and policy decisions.  Those who are not working must get an income flow from their own resources or the resources of those who are working, or a combination of the two.

The late 90s was more than just a dot-com boom.  It was a working boom where the number of people not working was at historically low levels compared to the number of people working.  The end of the dot-com era and the decline in manufacturing jobs that began in the early 2000s, when China was admitted to the WTO, marked the end of this unusual period in U.S. history.  Former Secretary of Labor Robert Reich (Clinton administration) sometimes uses this unusual period as a benchmark to measure today’s environment.

Not only was this non-working/working ratio low, but GDP growth was rather high in the 1990s, in the range of 3 – 5%.

Let’s look at GDP growth from a slightly different perspective.  Real GDP is the country’s output adjusted for inflation.  Real GDP per capita is real GDP divided by the total population in the country.  Real GDP per employee is output per person working.  As GDP falls during a recession, so too do the number of employees, evening out the data in this series.  A 65 year chart reveals some long term growth trends.  In the chart below, I have identified those periods called secular bear markets when the stock market declines significantly from a previous period of growth.  I have used Doug Short’s graph  to identify these broader market trends.  Ideally, one would like to accumulate savings during secular bear markets when asset prices are falling and tap those savings toward the end of a secular bull market, when asset prices are at their height.

In the chart above note the periods (circled in green) of slower growth during the 1968-82 secular bear market and the last few years of the 2000-2009 secular bear market.  After a brief upsurge at the end of this past recession, we have continued the trend of slower economic growth that started in 2004.  A rising tide raises all boats and the tide in this case is the easy monetary policy of the Federal Reserve which buoys stock prices.  In the long run, however, stock prices rise and fall with the expectations of future profits.  Contrary to previous bull markets, this market is not supported by structural growth in the economy, and that lack of support increases the probability of a secular bear market in the next several years, just at the time when the boomer generation will be selling stocks to generate income in their retirement.

Earthquakes in some regions of the world are inevitable.  In the aftermath of the tragedy in Nepal, we were reminded that risky building practices and regulatory corruption can go on for decades.  There is no doubt that there will be  horrific damage and loss of life when the inevitable happens yet the risky practices continue.  The fault lines in our economy are slower per employee GDP growth and a greater burden on those employees to pay for programs for those who are not working. The worth of each program, who has paid what and who deserves what is immaterial to this particular discussion.  Growth and income flow do matter. Asset prices are rising on shaky growth foundations that will crack when the fault lines slip.  Well, maybe the inevitable won’t happen.