Thursday, January 5, 2012

American Unemployment - Who Has Suffered The Most?

While the headline jobless claims and overall unemployment statistics are showing modest improvement in America's overall employment picture, it is interesting to look behind the sterility of the numbers and examine just who is suffering the most and who is benefiting the most during this jobless "recovery".

In their paper entitled "Employment Patterns During the Recovery: Who Are Getting the Jobs and Why?", Ausegul Sahin and Jonathan Willis of the Federal Reserve Banks of New York and Kansas City respectively, examine the distribution of employment losses and gains during the most recent recession measured by educational level, gender and age.  An analysis of this data shows a fairly clear relationship between the amount of education an individual has, their gender and their age and their chances of losing and regaining employment in this very difficult market.  Please note that the data in this study is current to the end of August 2011.

During the Great Contraction from December 2007 to January 2010, the American economy lost 7.8 million jobs; on the flip side, employment of workers aged 16 and older increased by a very paltry 1.1 million over the 19 month period from January 2010 to August 2011.  This employment gain data can be broken down further by gender, age and by level of educational attainment as shown in this table:

Looking at educational level, we can see immediately that the vast majority of the job gains were experienced by older males with at least some college education; those with a degree experienced an increase in employment of 1.1 million and those with a least some college experienced an employment gain of 345,000.  On the other hand, those with high school or less experienced total employment contraction of 642,000.  Note that there is also a very strong gender bias in the job gains; men gained 1,514,000 jobs while women saw a net drop of 314,000.  Worker age also played a very strong role in employment gains with just about all of the gains experienced by workers aged 55 and over; this group saw employment gains of 1,430,000 with workers aged 25 to 54, who make up the vast bulk of America's workforce, seeing a drop of 512,000 jobs.

Let's examine each of the three factors, educational level, gender and age to see how they affect employment gains and losses.

Educational Attainment:

The level of educational attainment has a strong impact on the level of job growth since the "end" of the last recession.  Looking back a decade, it is becoming apparent that employment gains in the recovery phase of an economic cycle are showing a trend toward a more educated workforce.  This really shouldn't be all that surprising considering the "gutting" of America's manufacturing sector over the past three decades.  Here is an interesting graph showing that America's manufacturing sector has seen its share of the jobs market drop from 30 percent in 1950 to around 10 percent in 2007 as jobs moved overseas:

As I noted above, employment gains during the 19 months starting in January 2011 were concentrated among workers with at least some college education.  Here is a graph showing the average monthly growth in employment from January 2010 to August 2011 for all four groups in black, the peak to trough job losses from December 2007 to January 2010 in grey and the peak to peak job gains from March 2001 to December 2007 in blue:

Peak to trough employment (grey bars) fell the most for high school graduates with no college.  These workers saw a drop in employment of 110,000 jobs per month over the period from December 2007 to January 2010 while those with a Bachelor's degree or higher saw modest monthly employment gains of 10,000 jobs.  Even as the economy turned north (black bars) between January 2010 and August 2011, high school graduates with no college still saw the number of jobs shrink by 30,000 per month in contrast to those with a Bachelor's degree or higher who saw employment gains of 58,000 per month.

Over the peak to peak period from March 2001 to December 2007 (blue bars), by far the greatest employment gains were experienced by Americans with a Bachelor's degree or higher.  These Americans saw average monthly employment gains of 89,000.  For workers with some college, the growth with still at a reasonable 28,000 per month and was practically non-existent for those with high school or less.

Here is an interesting graph showing how employment changed for lower education industries including manufacturing, construction and transportation:

Note how these sectors were hit very hard during the latest recession with job losses (grey bars) ranging from just over 20,000 per month to 90,000 per month and how job recovery (black bars) has been modest at best with job additions ranging from 15,000 to 21,000 per month, a rate that is far, far below the jobs lost.

To summarize these observations, American workers that have attained at least some college are the only group that has seen employment growth over most of the past decade.  On the other hand, American workers with a high school diploma or less have seen the weakest job growth and have suffered from the highest level of employment contraction as the Great Recession took hold and during the aftermath.

Employment Patterns by Gender:

During the recent economic bust and boom, males have experienced the greatest fluctuation in employment.  The decline in employment for men accounted for 70 percent of total job losses during the contraction phase and the gain in employment for men accounted for 90 percent of the total job gains during the recovery phase.

Statistics show that employment opportunities for men have grown by an average of 88,000 per month (black bars) since January 2010 in contrast to women who have seen their job opportunities increase by only 9,000 per month as shown on this graph:

During the down side of the cycle, between December 2007 and January 2010, men saw their employment prospects decline by 242,000 per month (grey bars) in contrast to women that saw their prospects decline by "only" 106,000 per month.

The sharp decline in male employment was likely related to the sectors of the economy that experienced the greatest contraction during the latest recession (i.e. construction, manufacturing and trades, transportation and utilities) since these sectors of the economy hire more males than females.  Since the construction and manufacturing sectors have hardly recovered since the Great Recession, how is it that men have experienced the greatest employment gains?    For two reasons; first, because men made up a larger fraction of the overall jobless pool (56 percent of the total unemployed) and second, because men's behaviour is different than that of women since men tend to be more attached to the workforce than women.  The authors suggest that men are more likely to accept jobs that pay poorly when faced with the alternative of unemployment in very tight job markets.  As well, women are generally more likely to stop working when their wages decline since men are less sensitive to declines in wages.  In addition, men with lower levels of education may be willing to accept lower paying jobs since the jobs market in America has swung in favour of more highly educated workers.  Employers may well be looking at the increased pool of rather desperate unemployed workers and making a decision to hire more highly educated workers than they normally would since the labour of these out-of-work males may be available at bargain basement prices.  

Employment Patterns by Age:

I was quite surprised to note that employment growth since the so-called end of the Great Recession has been mainly among those aged 55 and older.  To me, that seems almost counter-intuitive. Employment gains for Americans 55 years of age and older has increased by 75,000 per month (black bars) since January 2010, in contrast, employment levels for Americans between the ages of 25 and 54 have actually declined by an average of 27,000 per month over the same period of time.  This is somewhat surprising since the younger age group is considered to be the prime age for employment.  Here is a graph showing job gains and losses for each of the three age groups:

Employment of prime-age workers showed the sharpest decline during the down side of the cycle between December 2007 and January 2010 with job losses averaging 255,000 per month (grey bars) compared to job gains of 48,000 per month for workers 55 years old and older.  How can this be explained?  Over the past decade, demographic changes have come into play as baby boomers age.  The overall share of prime-age workers (54 and under) has dropped from 57 percent to 52 percent and the share of older workers has increased from 27 percent to 32 percent.  This demographic shift alone accounts for 88 percent of the total monthly employment increase for older workers.  In contrast, the shrinking number of prime-age workers accounts for only about half of the total decline in employment for that age group.  Even during the recovery, the prime-age group has lost an average of 16,000 jobs per month when comparing the size of their cohort to the overall population (i.e. correcting for the shrinking size of their cohort).  This indicates that the recession's impact on employment was (and still is) very severe.

Another factor that enters the employment pattern by age is the shift in retirement age.  Declines in the value of real estate and other investments are resulting in more Americans working later in life.  There are increased numbers of Americans working well into their 70s; as time passes and the demographic patterns shift toward an older and older population, it is expected that Americans over the age of 55 will make up an increasing proportion of the workforce.

In conclusion, in spite of recent improvements in jobless claims and unemployment statistics, this study shows that the recovery has not treated all members of American society equally.  Employment growth has been concentrated in jobs where more highly educated workers are required.  As well, men are more likely to take jobs that they may not feel are suitable (i.e. don’t pay as well) over the longer term because they are often more attached to the workforce than women; this has improved the job situation for the hairier gender at the expense of women.  Older workers have also been net beneficiaries of the "recovery" as the work force shifts toward older workers.  From this data, it is apparent that the current sluggish recovery in America's job market is not related to a mismatch between workers and jobs.  Rather, the demand for the most sought-after workers has simply not kept pace with growth in the size of the overall population.   In closing, these two  graphs showing the declining labor force participation rate since the 1940s for both men and the more recent decline for women pretty much explains why the American employment issue is stubbornly intransigent and on trend with past economic behaviour:


  1. Employed doesn't mean living wage. Where are those statistics.

  2. anon, here you go. it's worse than you thought

  3. A "living wage" is merely an arbritrary classification...

    ... What is a "living wage"? Enough to purchase 3 new apple devices in one year? Do you think your grand-parents wage (adjusted for inflation) could afford these modern luxuries?

    What was life like before two cars in the drive-way, McMansion houses, computers, HD tvs, PVR, & SmartPhones?

    1. I think everyone understands what's meant by a living wage. I can put food out that I buy in a supermarket, I can have a roof, pay for utilities (not cable), I have a car because you can't go to work without one (for the vast majority, I can have some savings accumulate because I'm sure you'll bash the person who can't pay the bills once they're laid off. I don't think anybody here thinks a living wage is a McMansion 4iPads and a Lexus. Get over yourself.

  4. As a matter of fact, I did a posting on changes to wages in America here:

    For median wage earners, their annual real increase in hourly pay averages out to 0.27 percent (not compounded) annually over a thirty year period from 1980 to 2010.