Where Did the Jobs Go?

What Obama and Romney Can Learn from the Recession


The current economic debate is missing a compelling theory of growth. (Credit: Flickr, The U.S. Army & davelawrence8.)

September 03, 2012 | Roger Pielke Jr

US unemployment rates remain high and a major issue in the US presidential campaign. This post seeks to explain with simple math where the jobs went as a first step in understanding how unemployment might be reduced. The math is simple, and it helps to show how most debate over economy and employment miss what actually matters most – and that is innovation policy. 

In December, 2007 the US economy employed about 138 million people. In February, 2012 that number had shrunk to about 133 million (data). Where did those jobs go? Two recent studies provide some important insight to this question.

 Last month, Georgetown University's Center on Education and the Workforce published a study looking at job losses from 12/2007 to 2/2012 (hence my use of those dates) by economic sector and educational attainment (here in PDF).  The Figure below shows their overall results.


The graph shows that for those who attended some college, or with more post-secondary education, employment had recovered to December 2007 levels by February 2012. The balance of “lost” jobs was among those with secondary (i.e., high school) education or less.

A longer view, beyond the time frame of the recent recession, shows that all job growth in the US since 1989 has come from those with some post-secondary education, with a 14% drop in those employed with a secondary degree or less. Thus, the recent recession amplified trends that were already in place and part of the ongoing structural change in the composition of the economy.

The Georgetown report also looked at job losses by economic sector by educational attainment. This is shown in the figure below, which shows the dramatic loss of jobs by those with the least education in construction, IT, manufacturing and financial services. Together, these four sectors account for the loss of 5.6 million jobs from December 2007 to February 2012. Within those sectors 3.7 million, about two thirds, of the job losses were among those with the least education, secondary or less.

A big part of the loss of jobs of course is the nature of the recession itself, originating in the finance industry and the housing bubble – hence the large loss of jobs in construction and financial services and the bubble burst.

The loss of low-skill jobs in manufacturing and IT (and other sectors in the economy) is arguably the consequence of industries in these sectors trying to maintain productivity growth to remain competitive by shedding the lowest skilled jobs at a rate faster (slower) than the decline (subsequent rebound) in output. These numbers are borne out by looking at productivity data.

According to the BLS, in quarterly data from 2000 to 2006 labor productivity gains (in the non-farm sectors) over the previous 13 quarters (i.e., the period from Dec ’07 to Feb ’12) ranged from 10.4% to a 16.5%. During the 13 quarters examined in the Georgetown study, those rates of productivity gain over 13 quarters dipped to 5.0% to 8.9%.

Despite the recession, productivity gains were positive. These gains occurred because productivity reflects output per hour and over the period from December 2007 to February 2012, output increased by less than 1% representing the economic recovery, but hours worked decreased by almost 6%, which is why jobs have not recovered as much as the economy.

In a paper prepared for the Jackson Hole Economic Policy Symposium and released a few days ago (here in PDF) Edward Lazear and James Spletzer look at much of the same data as found in the Georgetown study and conclude:

The recession of 2007-09 witnessed high rates of unemployment that have been slow to recede. This has led many to conclude that structural changes have occurred in the labor market and that the economy will not return to the low rates of unemployment that prevailed in the recent past. Is this true? The question is important because central banks may be able to reduce unemployment that is cyclic in nature, but not that which is structural. An analysis of labor market data suggests that there are no structural changes that can explain movements in unemployment rates over recent years. Neither industrial nor demographic shifts nor a mismatch of skills with job vacancies is behind the increased rates of unemployment.

Lazear and Speltzer appear to get so caught up in an academic debate over “structural” versus “cyclical” changes that they focus on this debate rather than the more important questions of policies and consequences.  Their inability to clearly define “structural” changes in the economy is a clear indication that any debate over “structural” vs. “cyclical” is unlikely to be resolved empirically, and thus can only provide so much insight to policy questions. Their argument appears to be that “cyclical” implies nudging by central bankers, and not much else, whereas “structural” implies other government actions. It is not difficult to see how such a framing might be easily captured in Left vs. Right politics of the day.

However, economic policy is more complicated than either/or. The recent recession showed that the long-term change in the US economy to higher-skilled employment received a one-time shock, with the economy shedding a large number of unskilled jobs. This shock was in the same direction of long-term trends in the economy with respect to low-skilled employment, especially in manufacturing.  If the economic experience of the past several decades is any indication, then these jobs are not coming back, and nor should we want them to, as they are associated with low pay and low productivity.

In theory, improving productivity sounds great, but today there is the persistent matter of high unemployment in the near term, particularly among the low skilled part of the workforce. Lazear and Speltzer argue that the data suggest that,

There is no evidence that the recession resulted in a long-lasting skills gap that would require retraining experienced workers to work in different industries.

They reach this conclusion by looking at supply of workers and demand for employees, and not surprisingly, finding a pretty good balance.  Over the longer term it may indeed be the case that the economy organically reverts to a lower rate of employment. The important questions for policy however are whether specific government actions can make that longer term shorter (and, of course, avoid making the longer term longer).

I think that government policy is important here, and beyond the typical debate over stimulus vs. austerity or cycles vs. structures. Because the economy is non-stationary in the sense that sectors and skills are undergoing constant change, there is always the need for training and retraining within the workforce. Because all employment growth since at least 1989 has come from those with at least some post-secondary education, this indicates that there are both supply and demand reasons to focus on retraining -- these are the people who innovate, which leads to new industries, new sectors and ultimately new jobs in a vibrant, growing economy.

Of course, the persistent unemployment will not be easily remedied simply by offering training to the millions of low-skilled workers currently unemployed. Nor will it be addressed by short-term stimulus. What is needed is a more comprehensive approach to innovation in the economy, of which constant training and retraining of workers must be a part.

The larger conclusion is that missing from much of ongoing debate over jobs is any explicit, underlying theory of economic growth, and how policies contribute to it. Academic economics has too long relied on invisible hands and academic either/ors to guide policy debate. The way that unemployment goes down is through economic growth. Both Romney and Obama appeal to economic growth, but I have no idea how either candidate thinks that they can make it happen.

Where did the jobs go? They went to a no-go zone in current debate over economics and that is a place called innovation policy.


  • The increase in productivity with concurrent loss of jobs, especially those requiring minimum skill and training, is exactly what is expected as automation and information technology make replacing humans with machines (particularly with computers) more efficient and less costly.

    This trend is bound to increase with further ‘innovation’. This is part of the predicted old picture of eventual growing technological unemployment.

    It may not be ‘fixable’ – except by continuously reducing the work-week, increasing ‘welfare’ of various kinds, and ‘ending’ population growth. This hardly fits with anybody’s ‘cure’ for the jobs problem!

    There is NO reason to believe that innovation will – or even CAN – produce more new, high-demand products that will use more workers, rather than more automated machines, to increase the global employment picture!.

    By Len Ornstein on 2012 09 04

    Reply to this comment / Quote and reply

  • 1) For some reason, Roger, your email (@gmail) was placed in this comment section. You might want to get that changed!

    2) Regarding the above: innovation is a factor, even an important or most important factor. There are, however, lots of other factors, many of which are also very important.

    Health care is a huge issue. Not having a national program of health care means loss of employment deals a double blow to the newly unemployed.

    FIRE is also an issue - I’d not be surprised at all if Americans pay the highest percentage of income towards housing in the first world

    By c1ue on 2012 09 04

    Reply to this comment / Quote and reply

  • The President and eco-greens have said that the “green” economy was going to generate jobs.  These would be higher tech, of course, and require more training.  So, did, in fact, the increase in college+ degree’d jobs increase as a result of “green” activity?  And at the expense of blue-collar work?

    Where were the higher tech jobs?

    Breakdown of jobs by job-type: what percent of jobs overall are we talking about?  Perhaps the uptick in two is not material.  (Except for those with the jobs.)

    By Doug Proctor on 2012 09 04

    Reply to this comment / Quote and reply

  • The long term trends of falling relative wages and rising unemployment among for those without much education is indeed troubling, but I think the ‘solution’ if one exists, surely involves changing the social and familial expectations which make opting out of formal education acceptable to far too many young people.  While it will never be possible to eliminate completely the imprudent choices which lead to poor education and subsequent low lifetime earnings, social expectations of a pursuit of education, and indeed, expectations of more responsible choices by young people in all matters, seems to me the only approach which will have much long term impact.  The structural change in the economy is real enough… low skill jobs are disappearing….  but recognizing the problem as structural does not in itself provide much guidance.  The root cause of the problem the growing disconnect between the requirements for gainful employment and social/cultural norms which are lax with respect to responsible behaviors among young people. 

    There will always be a small fraction of people for whom higher education is (intellectually) out of reach, but a poor educational background, irresponsible behaviors in general, and low lifetime earnings are more often the product of low expectations, not a lack of intellectual ability or opportunity.  The change that is needed lies in the realm of personal and social values.  Neither central bank policies nor government programs for the unemployed will change people’s values.

    By Steve Fitzpatrick on 2012 09 04

    Reply to this comment / Quote and reply

  • It is a shame that the higher ed layer isnt broken up by STEM vs other. Looking at the industry sectors that are increasing higher ed jobs, eg energy, you could deduce that STEM is worthwhile. This leads on to other questions.
    What percent of the population is really capable of using STEM higher ed if available?
    Can society cope with a high percent of graduate unemployment? (assuming that a highly educated un/under-employed person can be dangerous. refer work by Turchin at U Connecticut. corelates grad unemployment with revolutions)

    By Richard Hill on 2012 09 04

    Reply to this comment / Quote and reply

  • I’d still like to see more thorough analysis of the jobs/education data.  A couple of thoughts:

    IME, PhDs/MS’s generally don’t come from low income families.  Do the education data reflect parental income biases? 

    I’d like to know what percentage of people are actually employed at their level of education.  I’ve read a few newspaper articles about people with grad degrees, even in STEM fields, working at Target and Wal*Mart – granted, these are anecdotes.  But it’s worth more work.

    Are we getting higher productivity because we’ve turfed the least-motivated demographic in the job market? 

    How do we determine what counts as employment that is consistent with a person’s level of education?  The physicists on NPR’s 13.7 blog complain regularly that their students all go to the financial industry – I can’t imagine that PhD’s in physics are working at the level of their degrees in the financial industry. 

    I wonder also about the effect of NSF funding for faculty research.  How many jobs are there for astrophysics PhDs?  Not many, judging by the number of applicants for each job.  In other words, I’m suggesting that the funding levels available for graduate assistantships – especially PhDs – far outstrip the number of jobs available to the graduates once they complete their degrees.  These students then take the place of jobs that were formerly held by MS students, who now take BS (ahem) jobs, etc.  Yes, their level of education ensures they’ll be employed.  But not at the level of their education.

    Is there a hidden comment on our education system in this data?  Having hired and worked with many Bachelors and Masters grads, I’m consistently surprised by the lack of basic skills – graduates that can’t even sort data in an Excel spreadsheet.  Is the BA/BS the new GED/HS diploma?  Does the dial on the Marshal stack go to eleven now?

    By jim on 2012 09 15

    Reply to this comment / Quote and reply

  • Jim On “These students then take the place of jobs that were formerly held by MS students, who now take BS (ahem) jobs, etc. “

    I got a B.Chem in 1969 from a large, competitive university.  Throughout my career it became obvious the PhD’s I was working with were becoming less and less educated.  I took one of the first Computer Science classes offered.  On the first day the prof pulled out a big paperback book, “Programming in Fortran” and said: ” you’ll need to pick this up at the bookstore.  You’ll need to learn Fortran to do the homework assignments.”  Sort of on the side we had to take something like 25% of the credits in non-tech electives such as English Lit, History, Sociology, etc.
    Eventually it became obvious from the education that my kids got that standards have gone down the tube.

    By George M on 2012 10 04

    Reply to this comment / Quote and reply

  • This shock was in the same direction of long-term trends in the economy with respect to low-skilled employment, especially in manufacturing.  If the economic experience of the past several decades is any indication. http://www.bau-tech.ch

    By Stellen für Handwerker on 2012 12 13

    Reply to this comment / Quote and reply

  • From where I’m from, we do have job fluctuations. It’s an interesting debate about skills vs post secondary education.

    By Graham on 2013 02 11

    Reply to this comment / Quote and reply

  • spielsuchttherapie logoBook of Ra connected has been planned decided on your favored diversion and methodologies with 5 reels and 9 pay-lines. A space machine is there and it has wild scramble in addition to hazard aspects that will permit entertainers illuminate and gathers past predeterminations of the Pharoah on top of their compartment. This makes the gambling joint sizzling sultry on the web.

    By Anonymous on 2013 02 13

    Reply to this comment / Quote and reply

  • Read here Interesting post and thanks for sharing. Some things in here I have not thought about before.Read more Thanks for making such a cool post which is really very well written.will be referring a lot of friends about this.Keep blogging   click here

    By Vivek on 2013 02 16

    Reply to this comment / Quote and reply

Submit a comment

About Roger Pielke Jr

I am a professor of environmental studies at the Center for Science and Technology Policy Research at the University of Colorado at Boulder. I also have appointments as a Research Fellow, Risk Frontiers, Macquarie University; Visiting Senior Fellow, Mackinder Programme, London School of Economics; and Senior Visiting Fellow at the Consortium for Science, Policy and Outcomes of Arizona State University. I am also a Senior Fellow of The Breakthrough Institute, a progressive think tank.

Click here to view his recent articles.