Technological Unemployment in America

An interesting and very thorough study by Daron Acemoglu and Pascual Restrepo at MIT and Boston University respectively looks at how industrial robots are taking over jobs that were previously performed by humans and how this impacts the future of both wages and job creation.  For the study, the authors examined the effect of industrial robot usage over the period between 1990 and 2007 (with the year 2007 selected as the last year of the study to eliminate the impact of the Great Recession) on local labor markets in the United States with the data broken down into commuting zones.  Let's look at some of their conclusions, keeping in mind that a 2013 paper by Frey and Osborne concluded that 47 percent of U.S. workers' jobs are at risk of automation over the next two decades and that some of the impacts are felt by adding only one additional robot per thousand workers:

1.) Impact on Employment: Using the employment to population ratio between 1990 and 2007 and the increase in robots adoption of one more robot per thousand workers between 1993 and 2007, a commuting zone with a value of exposure to robots equals to the average for the entire United States experienced 0.37 percentage points lower employment to population ratio than commuting zones with no robot workers.  In other words, adding one more robot per thousand workers reduced employment by 6.2 workers.

2.) Impact on Wages: Since robots affect employment levels, they also impact the composition of the workforce which has an impact on wages.  Again, a commuting zone with a value of exposure to robots equal to the average for the United States experienced 0.73 percent lower wage growth than commuting zones with no robot workers.  In other words, adding one more robot per thousand workers reduced average yearly wages by about $200 in the affected commuting zones.

3.) Impact on Men vs. Women: The authors concluded that employment and wages fall for both men and women, however, the effects of adopting industrial robots on declining employment for men are about 1.5 to 2.0 times higher than they are for women.

4.) Impact on Different Industries:  Here is a graphic which shows the negative effects of robots on the employment to population ratio in different industries over the period between 1993 and 2007:

Here is a graphic which shows the negative effects of robots on the employment to population ratio for specific occupations over the period between 1993 and 2007:

As you can see, the greatest impact on the employment to population ratio is felt in the manufacturing industries followed by services and retail.  Just think about that the next time that you use a self-scanning and checkout system at your local retailer.

5.) Impact on Educational Levels:  Here is a graphic which shows the negative effects of robots on the employment levels for various educational levels over the period between 1993 and 2007:

The greatest negative impact of increased use of robot workers on employment is felt by male workers with a high school or some college education.

  

Here is a graphic which shows the negative effects of robots on wages for various educational levels over the period between 1993 and 2007:

The greatest negative impact of increased use of robot workers on wages is felt by males and female with less that a high school education.

To close this posting, let's look at how the employment to population ratio for American workers has dropped significantly over the past decade and a half:

Given that this study looks at the period only up to 2007 and the fact that U.S. industries and businesses have increasingly adopting the use of robots over the decade since 2007, one has to wonder how much of an impact technology has had on the rather stubborn "real" unemployment rate and dropping employment to population ratio since the Great Recession.  With the industrial robot workforce expected to quadruple by 2025, perhaps the Trump Administration's mantra that America's foreign trade partners are to blame for the country's loss of jobs is disingenuous and ignores the reality of technological advances that have made human workers redundant.  This situation is only likely to worsen given the advances in artificial intelligence.

The Federal Reserve's Failed Employment Mandate

Updated June 2016

The mainstream media has spent a great deal of energy questioning whether the Federal Reserve should continue to resume interest rate increases during 2016.  With the global economy looking rather shaky at best and the United States economy looking positively glowing when compared to its peers even though it is expanding at modest levels, the Fed is finding itself caught between a "monetary policy rock and hard place".  As we all know, one of the Fed's mandates as established by Congress is to maximize employment.  While the drop in unemployment from ten percent in October 2009 to its current level of five percent, it looks like the Fed has succeeded in meeting at least one part of its obligations, however, scratching slightly below the surface of the headline U-3 unemployment rate, we find that the Fed's long-term monetary policy experiment has been less than a resounding success when it comes to employment for America's prime-age workers.

Here is a graph from FRED showing the employment-to-population ratio since the late 1940s to help us put the current ratio into perspective:

Here is the same data for the period from the beginning of 2004 to the present:

The employment-to-population ratio is defined as the ratio of total civilian employment to the civilian non-institutional population.  It is used to track the pace of job creation and compare it to the rate of growth in the adult population.  As you can see, there was a precipitous drop in the civilian employment-to-population ratio during the Great Recession, from a pre-recession peak of 63.4 percent to a low of 58.2 percent in both 2010 and 2011.  Since its low point, it has only improved to its current level of 59.7 percent, an improvement of only 1.5 percentage points from its lows.  A 2015 analysis by the Congressional Research Service  suggests that demographics are playing a significant role in this stubborn decline for two reasons:

1.) retirement of older workers (i.e. the baby boom generation).

2.) younger workers pursuing or extending their education.

As a result, one would expect that if younger and older workers were subtracted from the employment-to-population ratio, the ratio for prime-age workers between the ages of 25 and 54 would show that there is little remaining impact of the Great Recession, particularly given the Fed's massive interference in what passes for the free market.  With that in mind, here is the data from FRED showing what has happened to the employment-to-population ratio for prime age workers since the beginning of 2004:

The ratio for prime-age workers hit a high of 80.3 percent at the beginning of 2007 and fell to a low of 74.8 percent in late 2009 and 2010.  Since then, it has recovered to its current level of 77.7 percent, an improvement of 2.9 percentage points, however, it is still 2.6 percentage points below its pre-Great Recession peak.  

Let's do a bit of an analysis of the data:

Low point - 74.8 percent in November 2010

Current level - 77.7 percent in April 2016

Total Change - 2.9 percentage points

Time elapsed - 66 months

Average monthly improvement - 0.044 percent per month

As long as there is no economic slowdown or sudden economic growth spurt, it will take another 59 months for the prime-age employment-to-population ratio to return to its pre-Great Recession level.

This begs the question - how successful has nearly seven and a half years of Federal Reserve monetary policy really been when it comes to meeting at least one of its congressionally-mandated obligations?  It would be my guess that millions of prime-age American workers would unequivocally state that the Fed and its long monetary policy have been an unqualified failure and that additional interest rate increases should be off the table.

Prime-Age Workers in America - The Employment-to-Population Dilemma

While the headline U-3 unemployment rate has shown improvements since the Great Recession, in large part, unemployment has fallen because job-seekers are giving up hope and leaving the labor force.

According to David Cooper at the Economic Policy Institute, the best measure of labor market health is the prime-age (i.e. 25 to 54 years of age) employment-to-population ratio or EPOP.  This measure provides us with the percentage of prime-age workers that are employed.  By using a specific age range, we eliminate the problem of changes in the size of the workforce that are related to younger adults leaving the workforce to attend school or older adults retiring from the workforce.

Mr. Cooper goes on to look at the changes in the difference in the prime-age employment-to-population ratio for each state from the third quarter of 2007 when the Great Recession was just around the corner to the third quarter of 2014.  Here is a bar graph showing the results:

You will notice that, as a whole, the United States saw the prime-age EPOP ratio drop by 3 percentage points between 2007 and 2014.  Some states saw a significant decline in the percentage of prime-age adults that were working over the seven year time span; Georgia, Kentucky and New Mexico saw their EPOP ratio drop by 7 percentage points or more.  In addition, Arkansas, Rhode Island, Alabama, Nevada and Hawaii saw their EPOP ratio drop by 5 percentage points or more.  Even the states with relatively healthy oil-driven economies like Texas and North Dakota saw small declines, seeing drops of 0.7 and 1.6 percentage points respectively.

If the economy was in normal mode, one would expect that the prime-age employment-to-population ratio would increase.  Looking at the graphic, you will note that there was an increase in the share of prime-age adults that were working in only two states; Oklahoma and Michigan.  Unfortunately, in both cases, the population of 25 to 54 year olds had decreased in both states; by 2.8 percent in Oklahoma and by 12.5 percent in Michigan over the seven year period.  In other words, the only two states that saw an improvement in the percentage of adults that were working between 2007 and 2014 were because the number of adults in both states declined.  It had absolutely nothing to do with job creation.

Let's switch gears for a moment and take a quick look at the employment-to-population ratios for all OECD nations for workers ranging from 15 to 64, 15 to 24 and 25 to 54 years of age to help us put the American situation into perspective:

  

If you look across the line showing the data for the United States, between 2007 and 2014 you'll notice that the employment-to-population rate dropped by 4.4 percentage points for those aged 15 to 64, by 6.6 percentage points for those aged 15 to 24 and by 4 percentage points for those aged 25 to 54.  By comparison, over the same timeframe, on average, OECD nations saw their employment-to-population rate drop by only 1.2 percentage points for those aged 15 to 64, by 3.5 percentage points for those aged 15 to 24 and by only 1.4 percentage points for those aged 25 to 54.  This tells us that the United States economy hasn't done nearly as well at creating jobs, particularly for its prime-age adults, as its OECD peers.  Some nations, including Korea, Japan, Israel, Austria, Chile and Germany (among others) have actually seen their prime-age employment-to-population ratio rise over the years between 2007 and 2013, a sign of a healthy job market.

Obviously, the rate of job creation in the United States simply isn't keeping up with the growing demand for those jobs by Americans who fall between the ages of 25 and 54.  While the employment-to-population ratio doesn't provide us with any indication of the quality of jobs being created, it is apparent that, despite the best efforts of central bankers, growth in the number of prime-age workers is outstripping the economy's ability to create employment of any kind.