The Sense and Cents of a College Education

October 21, 2018

by Steve Stofka

Should a young person invest money in a college education? Let’s look at the question from a financial perspective. Building a higher educational degree is as much an asset as building a house. Let me begin with the hard numbers.

Employment: A person is more likely to be employed. Here is a comparison of those with a four-year degree or higher and those with a high school diploma. The difference in rates is 2% – 3% during good times and as much as 6% during bad times.

UnemployRateCollVsHS

Is the unemployment rate enough to justify an investment of $50K or more in a four-year degree? Maybe not. During the worst part of the financial crisis, ninety percent of HS graduates were working. Why should a diligent person with good work skills spend time in college? Most college students take six years to complete a four-year degree. They must spend four to six years of study in addition to the loss of work experience and earnings in those years. The unemployment rate is not a decision closer.

Earnings: In 1980, when those of the Boomer generation were taking their place in the workforce, college grads earned 41% more than HS grads. Today, college grads earn 80% more. That gap of $567 per week totals almost $30,000 in a year and is less than the monthly payment on a $50,000 loan (Note #1). Can a person expect to earn that much additional when they first graduate? No, and that’s why many students struggle with their loan payments in the decade after they graduate.

MedWklyEarnCollVsHS

Maybe that earnings difference is a temporary trend. The debt is permanent. Should a young person take on a lot of debt only to find out the earnings difference between college and high school graduates was temporary? Unfortunately, that’s not the case. The big shift came in the 1980s when the gap in earnings grew from 41% to 72% in twelve years.

EarnDiffPctCollVsHS

There were several reasons for the explosive growth in that earnuings gap. Many Boomers had gone to college to avoid the Vietnam War draft. As they crowded into the workforce in the late 1970s and 1980s, they wanted more money for that education.

During the 1980s, the composition of jobs changed. Steel manufacturing went overseas to smaller and more nimble plants which could adjust their outputs more economically than the behemoth steel plants that dominated the U.S.

Automobile companies in Michigan closed their old plants. Chrysler needed a government bailout. The manufacturing capacity of Asia and Europe that had been crippled by World War 2 took several decades to recover. The U.S. began to import these cheaper products from overseas. As high-paying blue-collar jobs diminished, the advantage of white-collar workers grew.

As more companies turned to computers and the processing of information, they wanted a more educated workforce that could understand and execute the growing complexity of information. Manufacturing today relies on computer programs that require a set of skills that are more technical than the manufacturing jobs of the past.

A oft-repeated story is that the signing of NAFTA in 1993 and the admittance of China into the World Trade Organization were chiefly responsible for the growing gap between white collar and blue collar workers. I have told that story as well, but it is incorrect and incomplete. As the graph above shows, that gap has grown modestly in the past twenty-five years. The big shift happened in the 1980s when the first of today’s Millennials were in diapers and grade school.

When we adjust weekly earnings for inflation, we can better understand the evolution of this earnings gap. In the past forty years, high school graduates have seen no change in median weekly earnings. From 1980 to 2000, their earnings declined. The 25% growth in the earnings of college graduates came in two spurts: in the mid to late 1980s, and during the dot-com boom of the late 1990s.

EarnInflAdjCollVsHS

Since this trend has been in place for decades, college students can assume that it will likely stay in place for the following few decades. Like the mortgage on a home, the balance on a student loan doesn’t increase every year with inflation, but the earnings from that education do and they have increased more than inflation. The payoff to a four-year degree is the difference in earnings. That is the decision closer.

Notes:

  1. Using $50,000 loan for ten years at 6% interest rate at Bank Rate.

Lots of Changes

March 25, 2018

by Steve Stofka

What a week it was. A glance at the headlines would lead someone to believe that it was all about tariffs and an impending trade war between the U.S. and China. On Thursday and Friday, the Dow Jones Industrial Average lost more than 1000 points, or almost 5%. Was that all about tariffs? Hardly.

As expected, the Federal Reserve raised interest rates ¼% on Wednesday.  This put the Fed rate at 1.5% – 1.75%. Half of the members of the interest setting committee (FOMC) indicated that it might be necessary to raise interest rates four times this year. The market has been pricing in three interest rate increases for 2018. Until Thursday, a fourth increase had not been fully priced in.

Further, the Fed is projecting an unemployment rate below 4% by late 2018 and early 2019. The current rate is 4.1%. Many industries are already struggling to find qualified workers. Rarely does the unemployment rate dip below 4%, and each time, inflation has risen and the stock market has fallen – sometimes substantially.

CPIUnemploy

The downturn following the Korean War was short and shallow, but the other two periods of low unemployment were followed by steep corrections in the market.

On Thursday night, the White House tweety bird announced another change in the roster. Out with the old National Security Adviser, General H. R. McMaster. In with the new adviser, John Bolton, an old school war hawk who avoided military service in Vietnam by joining the National Guard. Bolton’s first instinct is war and regime change as a solution to global disputes. In choosing Mike Pompeo as his new Secretary of State and John Bolton as his new National Security Advisor, Trump has assembled a war cabinet. The market has still not priced in the heightened chances of conflict with North Korea or Iran. Nor has it recognized a greater likelihood of armed conflict with China in the South China Sea. That might come in the next few weeks.

On Thursday, Trump enacted tariffs on imported steel and aluminum from China as promised. Stronger action against China’s trade policies are overdue, as it has long violated the spirit, if not the letter, of the WTO global agreements. Car manufacturers wanting to set up a plant in China must have a Chinese business partner with a 25% stake and – surprise – access to industrial trade secrets. The national government heavily subsidizes key industries so that they can support their own industries and workers. They avoid labor and environmental regulations, and when caught, pledge to do better. They issue a national change in regulation, but the change is only published and enforced in a few local areas.

The theft of intellectual property is a hallmark of most developing nations like China. In the 18th and 19th century, the U.S. was notorious for copying products made by companies in England and France. Article 1, Section 8 of the Constitution added some promise of patent and copyright protection, but the laws instituted protected only U.S. citizens. A half century later, Charles Dickens was “one of the chief victims of American literary piracy” (Source). A foreign inventor had to establish citizenship or residency in the U.S. for two years to gain any patent protection. In 1887, the U.S. joined a 19th century version of the WTO called the Paris Convention. As China does today, the U.S. skirted international agreements for at least a decade (Patent history).

Older Chinese citizens may have watched patrolling U.S. naval ships from the shores of the Yangtze River. The nation remembers the century of U.S. gunboat diplomacy (Wikipedia article). Despite American free market rhetoric, Chinese leaders understand that mercantilism still retains a strong political influence in the trading policies of many developed countries, including the U.S.

When NAFTA was signed in the early 1990s, subsidies of American corn farmers enabled them to sell cheap corn to Mexico. Unable to compete, many farmers in northern Mexico went out of business. As farming jobs decreased in Mexico, many laborers journeyed north to the U.S. to pick crops so that they could support their families. The U.S. is partially responsible for creating the very environment that led to so much illegal immigration from Mexico.

Around the world, developed countries cry foul when another country subsidizes goods that are exported at a lower cost into their countries. Since 1963, the U.S. has imposed a protectionist tariff of 25% on imported light duty trucks, the so called “chicken tax”. Protected for over fifty years by this tariff, domestic truck manufacturers like Ford and Chevy had made few substantial changes to their work vans in the past few decades. In 2015, Ford finally made a substantial change to its F-150 pickup. Notice those Mercedes tall work vans on the road? They are built in Germany, disassembled to avoid the tariff, shipped to the U.S. and reassembled by U.S. workers. Ford uses the same process with its Transit Connect van.

Boeing imports parts from all over the world to build its Dreamliners. Chinese companies use southeast Asia as a manufacturing supply, then assemble and ship thousands of products to the U.S. and around the world. In the truly global manufacturing economy, a trade war is a threat to the profits of many large businesses. They have tuned their operations to the contradictory rules of international trade.

Business leaders understand the political strut of free trade. Each business wants free trade when it wants to compete in someone else’s market. Each business lobbies for more regulations, tariffs and barriers to protect its competitive position within its own market. Yes, it’s all lies, so it’s important that the rules underlying this game not change too much. Trade wars change the rules and that’s bad for business.

Wage Growth – Not

August 12, 2017

Ratios are important in baseball, finance and cooking, in economics, chemistry and physics, and yes, even love. If I love her a lot and she kinda likes me a little, that’s not a good ratio. I learned that in fourth grade.

Each week I usually turn to one or more ratios to help me understand some behavior. This week I’ll look at a ratio to help explain a trend that is puzzling economists. The unemployment rate is low. The law of supply and demand states that when there is more demand than a supply for something, the price of that something will increase. Clearly there is more demand for labor than the supply. I would expect to see that wage growth, the price of labor, would be strong. It’s not. Why not?

I’ll take a look at an unemployment ratio. There are several rates of unemployment and there is no “real” rate of unemployment, as some non-economists might argue at the Thanksgiving dinner table. The rates vary by the types of people who are counted as un-employed or under-employed. The headline rate that the Bureau of Labor Statistics (BLS) publishes each month is the narrowest rate and is called the U-3 rate. It counts only those unemployed people who have actively searched for work in the past month. In the same monthly labor report, the BLS publishes several wider measures of unemployment, U-4 and U-5, that include unemployed people who have actively searched for a job in the past 12 months. U-6 is the widest measure of unemployment because it includes people who are under-employed, those who want full-time work but can only find part-time jobs. Included in this category would be a person working 32 hours a week who wants but can’t find a 40 hour per week job.

The ratio that helps me understand the underlying trends in the labor market is the ratio of this widest measure of unemployment to the narrowest measure. This is the ratio of U-6/U-3. In the chart below, this ratio remained in a narrow range for 15 years. Unemployment levels grew or shrank in tandem for each group. By 2013, the ratio touched new heights, climbing above 1.9 then crossing 2 in 2014. The two groups were diverging. The U-3 rate, the denominator in the ratio, was improving much quicker than the U-6 rate that included involuntary part-time workers.

U-6-U-3Ratio

What would it take to bring this ratio down to 1.85? About 1.5 million fewer involuntary part time workers. What does that involve? Let’s say that those involuntary part-time workers would like an average of 15 more hours per week of work. That is more than 20 million more hours of work per week, which seems like a lot but is less than a half percent of the approximately 6.1 billion hours worked per week in the 2nd quarter of 2017.  These tiny percentages play a significant role in how an economy feels to the average person.

Let’s turn to a ratio I’ve used before – GDP per hour worked. I don’t expect this to be a precise measurement but it reveals long term trends in productivity. In the chart below, GDP per hour has flatlined since the end of the recession.

GDPPerHour201706

There are two ways to increase GDP per hour: 1) productivity gains, or more GDP per hour worked, and 2) reduce the number of hours worked more than the reduction in GDP. Door #1 is good growth. Door #2 is the what happens during recessions. GDP per hour rises because hours are severely reduced. I would prefer slow steady growth because the alternative is painful. Periods of no growth can be wrestled out of their torpor by a recession, a too common pattern. There were two consecutive periods of flat growth followed by recession in the 1970s and from the mid-2000s to the present day.

GDPPerHour1971-1984

The economy can withstand two years of flat growth without a recession as it did in the early 1990s. It is the long periods of flat growth that are most worrisome. In the early 1970s and late 2000s, the lack of growth lasted three years and were followed by hard recessions. The lack of growth in the late 1970s led to the worst recession since the 1930s Depression. GDP per hour growth has been flat for eight years now and I am afraid that the correction may be hard as well. Maybe it will be different this time. I hope so.

///////////////////

Participation Rate

Some commentators have noted the relatively low Civilian Labor Force Participation Rate (CLFPR). This is the number of people who are working or looking for work divided by the population aged 16 and older. (BLS). The rate reached a high of 67% in 2000 and has declined since then. For the past few years, the rate has stabilized at just under 63%.

A graph of the rate doesn’t give me a lot of information. Starting in the 1960s, the rate rose slowly as more women came into the workforce and the large boomer generation came into their prime working years. So I divide that rate by the unemployment rate to look for long term cyclic trends. Notice that this ratio peaks then begins a downward slide as recessions take hold.

CLFPR-UI1947-2017

In mid-2014 this ratio finally broke above a long term baseline average and has been rising since. Today’s readings are nearly at the peak levels of early 2007.

CLFPR-UI

Some pundits use the CLFPR as a harbinger of doom that includes: 1) too many people are depending on government benefits and don’t want to work; 2) there is a shrinking pool of workers to pay for all these benefit programs; 3) thus, the moral and economic character of the nation is crumbling. During the 1950s and 1960s, when the participation rate was lower than today, our parent’s generation managed to pay off the huge debts incurred by World War 2. It is true that benefit programs were much less than those of today.

In “Men Without Work” Nick Eberstadt documents a long term decline in the percentage of prime age (25 – 54) males who are working.  Some interesting notes on shifting demographics: foreign born men of prime working age are more likely to be working or looking for work than U.S. born males. According to the Census Bureau’s time use surveys, less than 5% of non-working men are taking care of children.

In 2004 the participation rate for white prime age males first fell below those of prime age Hispanic males and has remained below since then.  In 1979, 10% of black males aged 30-34 were in jail.  In 2009, the percentage was 25%.

So why should I care about participation rates and wage growth? Policies initiated in the 1930s and 1960s instituted a system of inter-generational transfer programs.  In simple terms, younger generations provide for their elders. Current Social Security and Medicare benefits are paid in whole or in part by current taxes. We are bound together in a social compact that is not protected by an ironclad law.  Beneficiaries are not guaranteed payments.

For 40 years, from 1975-2008, the number of workers per beneficiary remained steady at about 3.3 (SSA fact sheet). In 2008, the financial crisis and the retirement of the first wave of the Boomer generation marked the beginning of a decline in this ratio to the current level of 2.7.

In their annual reports, both the SSA and the Congressional Budget Office note the swiftly changing ratio.  Within twelve years, the ratio is projected to be about 2.3.  In 2010, benefits paid first exceeded taxes collected and, in 2016, the gap had grown to 7% (CBO report) and will continue to get larger.

Policy makers should be alert to changes in the participation rates of various age and ethnic groups because the social contract is built partly on those participation rates.  As with so many trends, the causes are diffuse and not easily identifiable.  Economic and policy factors play a part.  Cultural trends may contribute to the problem as well.

Congress has a well-established record of not acting until there is an emergency, a habit they are unlikely to change.  Fixing blame wins more votes than finding solutions, but  I’m sure it will all work out somehow, won’t it?

 

The Unemployment Delinquency Cycle

June 4, 2017

I’m scratching my head. No, it’s not dandruff. The BLS released their estimate of job gains in May and it was 100,000 less than the ADP estimate of private payroll growth. We’d all like to see these two monthly estimates track each other closely, which they tend to do. In an economy with 146 million workers, a 100,000 jobs is only 7/100ths of a percent, but this discrepancy comes just two months after a HYUUUGE spread of 200,000 job gains in the March estimates.

A simple solution to multiple surveys? I average them. The result is 191,000 job gains in May, close to that healthy growth threshold of 200,000. In the chart below I’ve shown the average of the two estimates for the past five years and highlighted the downward trend of the peaks. Reasons include a decline in oil and gas industry jobs, and a natural feature of a mature recovery.

PayemsADPAvg

We saw the same pattern of declining job gains from the early part of 2006 through late 2007 before the average dipped below zero. Boosted by a hot housing market in the early part of the decade, construction employment began to cool in 2006.

PayemsADPAvg2002-2010

Some areas of the country are particularly hot. Denver’s 2.1% unemployment rate is absurdly low as is the state’s rate of 2.3%. Both are at historic lows, less than the go-go years of the dot-com boom. Colorado’s rate is the lowest among the 50 states (BLS). While income inequality has been rising in other hot metro areas like San Francisco, it has fallen in the Denver metro area.

There is a downside to strong growth. Back in “ye olden days,” like the 1970s and 1980s, I was introduced to a rule of thumb. It stuck with me because it seemed too simple. Here’s the rule: whenever the unemployment rate gets below 5% in an area, the price of some key component of  the economy is rising much faster than its long term average.   Lower unemployment leads to a mispricing of some asset.

Let’s turn to the other component of this credit cycle: loan delinquency.  The institutions who loan money expect that a certain percentage of borrowers will default. Lenders include the cost of those defaults when they calculate interest rates and loan service fees. The non-defaulting borrowers pay for the defaulters. During recessions, the delinquency rate on consumer loans usually rises above 4%. When unemployment is low and growth is strong, the delinquency rate goes below 3%.  Lower delinquency leads to a mispricing of credit risk.

Let’s review these two mispricings. The price of an asset is a price on some future flow of use or income that will come from the asset.  The interest rate on a loan is the price of money and the price of risk.  Let’s put these two mispricing together and we have another rule of thumb: as the difference, or spread, between the unemployment rate and the delinquency rate on consumer loans gets closer to zero, the more likely that the economy is overheating. A rising spread indicates a coming recession because unemployment responds faster than the delinquency rate to economic decline and increases at a faster rate. The spread changes direction and grows.

UnemployDelinquencySpread

Here’s the process. As the unemployment rate decreases, lending terms and loan criteria become more favorable. When we buy stuff on credit, we commit a portion of our future income stream to a creditor. When an economy begins to decline and unemployment increases, some income streams become a trickle or stop altogether. A loan payment is missed, then another, and those in more fragile economic circumstances default on their loans.

As the delinquency rate rises, lending policies begin to tighten again, making it more difficult to qualify for loans. Many businesses depend on the flow of credit, so this tightening causes a decline in sales, which causes businesses to lay off a few more people, which further increases both the unemployment rate and the delinquency rate. This reinforces the downward trend.

The NBER is the official arbiter of the beginning and end of recessions but often doesn’t set these dates until several years later.  This change in the direction of the spread is a timely indicator of trouble ahead. An understanding of the credit cycle is crucial to an understanding of the business cycle, which influences the prices of our non-cash assets.

Next week I’ll take a look at the cycle of asset pricing.

Wage and Industrial Growth

July 6, 2014

This week I’ll take a look at the monthly employment report, update the CWPI and introduce a surprising medium term trading indicator.

***********************

Employment

On Wednesday, the private payroll processor ADP gave an early forecast that this month’s labor report from the BLS would be robust, near the tippy-top of estimates of job gains that ranged from 200K to 290K.  The BLS reported $288K i net job gains, including 26K government jobs added. 17,000 of those jobs were in education at the local level.  Rising sales and property tax revenues have enabled many city and county governments to replace education jobs that were lost during the recession.

Job gains may be even better than the headline data shows.  ADP reports that the large majority of hiring is coming from small and medium sized firms.  The headline number of job gains each month comes from the BLS Establishment Survey, which underestimates job growth in really small firms.  The Household Survey estimated about 400K job gains this past month.  Usually, the Establishment Survey is thought to be the more reliable estimate but in this case, I would give a bit of a bump up toward the Household Survey estimate and guesstimate that job gains were closer to 330K this past month.  The BLS also revised April and May’s job gains upward.

The unemployment rate decreased .2% to 6.1% and the y-o-y decline in the rate has accelerated.

Excellent news, but let’s dig a bit deeper. The BLS tracks several unemployment rates.  The headline rate is the U-3 rate.  The U-4 rate includes both the unemployed who have looked for work in the past month, and those who have not, referred to as discouraged workers.  The trend in discouraged workers has been drifting down, although it is still above the normal range of .2 to .3% of the work force.

I would be a whole lot more optimistic about the labor market if the employment rate of the core work force aged 25 – 54 were higher.

Slowly and inexorably the employment level of this core has been rising in the past few years but the emphasis is on the word slowly.

The number of workers who usually work part time seems to have reached a high plateau, close to 18% of the Civilian Labor Force (CLF).  The CLF includes most people over the age of 16.  June’s Household Survey shows a historic jump of 800,000 additional part time jobs added in the past month.

A closer look at the BLS data makes me doubt that number. The unseasonally adjusted number of part timers shows only a 400,000 gain, leading me to question any seasonal adjustment that doubles that gain.  Secondly, the BLS did not seasonally adjust last month’s tally of part time workers, leading me to guess that June’s figure includes two months of seasonal adjustment.

That same survey shows a one month loss of more than 500,000 full time jobs lost (Table A-9 BLS Employment Situation).  The year-over-year percent change in full time workers is 1.8%.  As you can see in the graph below this is in the respectable range.  The unseasonally adjusted y-o-y gains is close to the seasonally adjusted gain, leading me to believe that the losses, if any, have been overstated due to month-to-month fluctuations in seasonal adjustments.

However, if you are selling a newsletter that says the stock market is grossly overvalued and the end is coming, then you would want to highlight the change in June’s seasonally adjusted numbers, to wit:  500,000 full time jobs lost;  800,000 part time jobs gained.

While the Civilian Participation Rate has steadied, it is rather low.  The Participation Rate is the number of people working or looking for work as a percent of most of the population above 16. Below is a chart showing the declining participation rate and the unemployment rate.

Now let’s divide the Participation Rate by the Unemployment Rate and we see that this ratio is still below the 34 year average.

                                                                                      
********************

Wage Growth

Each month the BLS reports average weekly earnings as part of the labor report. Year-over-year inflation adjusted wage growth is flat but has probably declined below zero.

An investor would have done very well for themselves if they had paid attention to this one indicator.  (There is a week lag between the end of the month price of SP500 and the release of the employment report for that month but it is close enough for this medium to long term analysis.)

The SP500 has gained almost 50% since the first quarter of 2006.  An investor going in and out of the market when inflation-adjusted wage growth crossed firmly above and below 0% would have made 134% during that same period.  “Ah, ha!  The crystal ball that will give me a glimpse into the future!” The problem with any one indicator is that it may work for a period of time.  This one has worked extremely well for the past eight years.  This series which includes all employees goes back only to March 2006.  The series that includes only Production and Non-Supervisory employees goes back to 1964.  The two series closely track each other.  I have left the CPI adjustment out of both series to show the comparison.

However, an investor using this strategy in the mid-1990s would have been out of the market during a 33% rise.  She would have been in the market during half of the 2000-2002 downturn and been mostly out of the market during an almost 50% rise from 2003-2005.  In approximately twenty years, she would have made half as much as simply staying in the market.

The ups and downs of wage growth may not be a reliable indicator of the market’s direction but it does indicate positive and negative economic pressures.  Poor wage growth in the mid-2000s probably fueled speculation in real estate and the stock market.

From the mid-1980s to the mid-1990s, a decade of negative inflation adjusted wage growth exerted downward pressure on labor income, which naturally led to a stratospheric increase in household debt.

The stock market quintupled as inflation adjusted wages stagnated.  During this period an investor would have been better to do the opposite: buy when wage growth fell below zero, sell when it crossed above.  As long as workers were willing and able to borrow to make up for the lack of wage growth, company profits could continue to grow and it is profits that ultimately drive stock market valuations.

Wage growth ultimately influences retail sales which impacts GDP growth.  The difference between the growth in retail sales and wage growth roughly tracks changes in GDP.

If retail sales growth is more than wage growth for a number of years, the imbalance has to eventually correct.  We are in a period of little wage growth and modest sales growth which means that GDP growth is likely to remain modest as well.

**********************

Constant Weighted Purchasing Index (CWPI)

Purchasing Managers surveyed by the Institute for Supply Management continue to report strong growth.  The CWPI index, based on both the Manufacturing and Services surveys, continues to rise as expected.

A composite of new orders and employment in the services sector remains strong.  February’s dip below 50 was an anomaly caused by the severe winter weather which coincided with inventory adjustments.

We see that this is a cyclic indicator, responding to the push and tug of new orders, employment, deliveries and inventories.  If the pattern continues, we would expect a decline in activity in the several months before the Christmas shopping season, a cycle that we have not seen since 2006.

The CWPI generates buy and sell signals when the index crosses firmly above and below 50 and has generated only 8 trades, or 16 separate transactions, in the past 17 years.  It is suited more to the long term investor who simply wants to avoid a majority of the pain of a severe downturn in the market.  Because it charts a composite of economic activity, it will not generate a signal in response to political events like the budget disagreement in July 2011 that led to an almost 20% drop in the market.  A strategy based on the CWPI gained 180% over the past 17 years as the market gained about 110%.

************************

Takeaways

Strong employment report but wage growth is flat and declining on a year over year basis.  CWPI indicator continues to rise up from the winter doldrums and should peak in two months.