The Rebound that Stayed Flat

I’ve been working to draw a graph that compares employment trends since the end of the Great Recession with other important trends in the economy, and also with earlier periods. Here’s what I’ve come up with (click on the graph for a bigger pdf version, and click here for a spreadsheet with the graph and all its data):

Recent economic and employment trends

Using data from the invaluable online resource FRED (and with the help of an equally critical real-world resource, my RA Noam Bernstein), I’ve plotted the trends since 1995 in US GPD, total corporate investment in equipment, and total corporate profits from non-financial companies (and also for all companies, including financial ones). I set the January 1995 value for each of these equal to 100 to allow comparisons across them over the years.

I also plotted the US employment-population ratio, or percentage of working-age people who have jobs (the axis for this line is on the right-hand side of the graph).

The overall impression I get from this graph is one of divergence over time. There’s a steady, slow-growing black line in the middle. This is GDP growth, climbing along at a bit less than 3% per year. Then there are a couple volatile and quick-growing graphs that wind up well above GDP. These are profits (blue) and investment (green), both of which are about 2.5 times as high in mid-2011 as they were at the start of 1995. GDP, meanwhile, increases by only about 50% over that period.

And then there’s the employment ratio (red), which declines by almost five full percentage points over the same period, from 63% of the population to 58.1%.

When I look at this graph I see evidence of the computer age everywhere. After the recession of 2001 ended profits came roaring back and equipment investment eventually started ramping up sharply, but the employment ratio increased only between September of ’03 and December of ’06, the most frenzied time for both the construction and financial industries.

And since the Great Recession officially ended in June of 2009 GDP, equipment investment, and total corporate profits have rebounded, and are all now at their all-time highs (non-financial profits are near their historic high). The employment ratio, meanwhile, has only shrunk and is now at its lowest level since the early 1980s when women had not yet entered the workforce in significant numbers.

So current labor force woes are not because the economy isn’t growing, and they’re not because companies aren’t making money or spending money on equipment. They’re because these trends have become increasingly decoupled from hiring — from needing more human workers.

As computers race ahead, acquiring more and more skills in pattern matching, communication, perception, and so on I expect that this decoupling will continue, and maybe even accelerate. This doesn’t mean that companies are about to stop hiring altogether; there are still plenty of things that humans alone can do. But it means they’ll need to hire at an ever-lower rate, compared to how quickly they’re growing, making money, or buying equipment. Because America’s working age population will continue to grow for at least the next few decades, I predict that the employment ratio will not start to trend upward in the coming years. If anything, I think it’ll decrease.

Do you agree? If not, what trends do you see that will cause the employment ratio to increase? What force(s), in other words, will be more powerful than the technology improvements we’re going to see? Leave a comment, please, and let us know.

  • Graham Chastney

    Looking at the chart it strikes me that it’s probably too early to draw the hypothesis that you have. I’m doing this purely from the perspective of the chart that you have drawn.

    At the end of recession of 2001 there would appear to be a 2 year+ lag before employment started to lift. The data that you have only shows a similar period after the end of the 2009 recession.

  • mg

    How much of the decrease in labor force participation is because of baby boomers going into retirement?

  • Tonyli3s

    the graph shows the ratio of working age workers, not retirement age.

  • Tonyli3s

    is it possible that we may have a future where there simply is no work necessary in order to maintain our basic survival? i.e. every human being will have a group of robots attending to all of their needs(growing food, cleaning, repairs, medicine) and the humans are free to spend their time as they wish?

  • Michael Goldfarb

    Would be curious to see the chart going back to 1980 or even 1973. This may not be just the computer age but something more malign. A management mindset that views every recession as an excuse for pre-emptive downsizing in order to meet profit targets. The current picture may well have that mindset rather than technology as its explanation.

  • Peter C. Hart

    Could you separate the effects of technological change from outsourcing?

  • Anonymous

    This is good stuff. One thing to look at would be the overall aging of the population especially in the first world. The ratio of productive workforce to productive population might be throwing off the last statistic especially at the end of the baby boom. There might simply be more retired people, and thus a lower employment to population ratio.

  • CDW

    Excellent point. The rate of outsourcing over the past 20 years has been phenomenal.

  • Linda Lee

    What economists continue to ignore is not that computers do so much more but that consumers do so much more for themselves: pump their own gas, book their airplane tickets, print boarding passes, book hotels, design their own business cards, film their friend’s wedding, go online to design an addition to his or her house. We’re all so busy doing things for ourselves (yes, thanks to computers) we have eliminated hundreds of thousands of jobs, people who used to do this for a living.

  • Kirt Bhatt

    Your conclusion seems logical except for war or Government hiring to accelerates beyond any Republican’s dreaded scenario US economy is set on a path to lower hiring rates and high unemployment. In fact, Mr. Bush’s two wars were expected to delay or avoid this scenario. Unless the humanity miraculously moves away from Globalization this trend is likely to persist because “If all human beings are created equal in their abilities, there is no just reason to believe that Americans must get such good life while a worker in China must toil for a meager living. This dilemma of Globalization has only one answer i.e. all humans everywhere must have similar standard of living. Borders are bad for commerce but they are the only instrument in the hands of politicians to develop a separate country and a separate standard of living.

  • Anonymous

    Excellent point. There are entire industries that are simply gone forever.

    However, we do have almost countless infrastructure projects, which could employ millions of people at great wages. That is, if they aren’t outsourced to contractors who employ Mexicans and other Latinos at below market rates . . . .

  • Fselker

    There may be merit to the argument, but the graph is potentially misleading. You call the black line “slow growing” but it increases about 50% in the graph. Meanwhile the expanded axis on the right makes an 8% drop look dramatic. Also, that parameter is adjusted for population growth, while the others are not – they are total values.

  • John Revay

    Great graph, I think the computer effect of all big companies buying and implementing Big ERP packages…..

  • Domenic Manganelli

    The most important thing that is not on this graph is the cost of human capital. Is it now more expensive than ever to hire someone? Should all the costs associated with ones existence be paid for, and managed by, an employer?

  • Anonymous

    I think your premise is sound. The graphic is powerful. If the divergence or decoupling of employment from production and investment is exaggerated by the graph as some suggest It could be the trebling of the money supply helped push investment and profit lines up while a dramatic increase in the number of 3rd bedroom earners, people working unattached or self-employed from home has exploded. I suspect many of these shadow solo workers are not factored into the official employment numbers yet.

  • wren337

    Now imagine some far-distant future where humans lounge around in Roman togas and machines do all of the work. How would we get from here to there? Is life getting easier or harder for the average person? Who is reaping all of the rewards of this increased productivity?

  • Workin Dawg

    There are 80 million people in the baby boom generation and the very front edge of this group began turning 65 last year (people born in 1946). In about five years, we will have roughly 4 million people per year reaching what was once thought of as retirement age and this will go on for almost twenty years. As a society, we are completely unprepared (emotionally, economically, politically)for this many people crossing into “old age” at once. At some point, we will need millions of low wage “helpers” to assist the older boomers with everyday activities. Many of these “helpers” may well be other “old” people that are still in good health. Private assisted living facilities are simply too expensive and there are no public sector alternatives. A lot of our older citizens are going to “age in place” in their homes. I don’t see a technological substitution for the manual labor that is going to be needed. Both aspects of this should help to normalize the employment to population ratio (older people will drop out of the denominator as they are no longer “working age” and employment levels will rise as “helpers” are hired).

  • Dirk Cotton

    You can’t omit outsourcing from this analysis. Toward the end of my time with AOL, around 2005, there were far fewer programmers sitting around our offices. That wasn’t because we had fewer programmers, it was because the jobs were in Bangalore.

  • Cindy

    I’m curious as to what impact outsourcing has had on employment in the US. Though “evidence of the computer age” seems to have face validity, I don’t see any proof in your graphs of that being the source of lack of hiring in the US.

  • Cindy

    I’m curious as to what impact outsourcing has had on employment in the US. Though “evidence of the computer age” seems to have face validity, I don’t see any proof in your graphs of that being the source of lack of hiring in the US.

  • Neal Narkis

    I was somewhat surprised, yet pleased with the results displayed in the graph. The results verified what I have felt for many years. I think that wren337 and Tonyli3s were right on target by basically saying that as a society, we have finally achieved the productivity to free us from the burden of labor to spend time on more noble pursuits (or perhaps more selfish pursuits – golf anyone?) The questions wren337 also raise are perhaps even more important than the results. How does society as a whole, finally reap the rewards for this apparent need for less labor? As a baby boomer, we were told as kids that one day automation would free us from the burden of labor and would give us more free time for cultural activities, charitable work, or just to play. I have called it the “great technology lie”. Your graphic shows what I have believed for quite some time – that we have had massive productivity gains thanks to advancing technology and automation, and that the need for labor to provide for the needs of society has been greatly reduced. But is our work week any shorter? Do we have more free time? Is our stress level any lower? Why is it that studies have shown that as a society we are sleep deprived? I believe that all of this productivity gain has gone to the increasing gap between the richest 1% and the other 99% – thus the “great technology lie” – society doesn’t reap the rewards, only the top 1% do. I can only hope that I will be around long enough to see all of society share in these productivity gains. As wren337 puts it, the big question is how do we get from here to the point where we sit around in Roman togas and let the machines do the work?

  • Anonymous


    On my blog, I calculated the number of “human brain equivalents (HBEs)” added each year by computers. By my calculations, in 2015, only 1 million HBEs will be added worldwide. Compared to the human population of 7+ billion, this is trivial. But by 2025, I calculate that 1 billion HBEs will be added each year. And by 2033, I calculate that 1 *trillion* HBEs will be added each year.

    So by my calculations, it should not be possible to see the impacts of computers at present, but in the next 10-20 years, I expect to see huge impacts.

  • ezracolbert

    the excel file is nice, and you deserve a lot of credit for that
    however, you loose points with the black background white text webpage
    you also loose points cause I’m pretty sure that very similar versions have been floating around the liberal net for quite a while

    on the technical graphics scoring, you get a mandatory 1 point deduction because the top to bottom order in the key doesn’t follow the top to bottom order of the data, eg in the key the two blue lines should be on top, cause in the data the visual impression is that the two blue lines are on top (go reread tukey or those guys, NOT tufte, who is a blowhard)

    you get credit for mixing colors and line width; people like P Krugman don’t seem to get that one
    You also get credit for putting the dates underneath the tick marks; people often put the dates btween the ticks, which another mandatatory one point deduction

    I’m not sure how you should be handling data with diff scales and volitility; the use of a arbitrary point as 100% is std, but highly suspicious – why that point ? what happens if you use a diff point ?
    And, you need to indicate the ~ decadal variation in the diff data sets – clearly there is a lot more in profiits then GDP

    also, the trouble with Y axis in % of start time is that you don’t know the absolute value of the numbers – in this case, the two blue and the one green line – it would be informative to know that the absolute values are not to discrepant (eg, hypothetically, if the absolute vaue of the two blue lines differed by 10fold, then the smaller one is irrelevant)

  • Anonymous

    Until the late 19th century, capital was always more scarce and thus more valuable than labor. During the 20th century their relative importance changed dramatically but the pendulum started moving back towards capital. Presently, there is ample amounts of capital but it is very concentrated and more free than ever to move around the globe. This combined with technology is bound, I am afraid, to take us back to what humankind experienced in the 19th century. The only force that can slow it down is a willful government and that explains the heavy push since the ’80′s to demonize government. In sum, unless we limit the influence of money in elections, we’re bound towards a 19th century social structure.

  • Anonymous

    I’m a bit late to the party but, I agree with your point. The retirement of the large Baby Boomer demographic and extended lifespans may have some effect on the data. However, the long term trend is for increased automation/decreased human employment. There are important jobs that need doing (Social Workers, Child Care Providers, Elder Care Providers) but, or society does not value these professions and they are poorly paid.

    A new form of societal organization may be needed to avert a tremendous amount of human misery and economic collapse.

  • per wiklund

    This trend of increasing unemployment ratio seems to be continuing. The solution I think must be to “democratize capital”, that is spreading ownership of companies profiting from automated production of goods and services. In this way people do not have to make a living from offering their (obsolete) skills on the work market.
    In fact, in the eighties, the ruling Social Democratic party in Sweden tried to introduce so called “löntagarfonder” (employee funds), which was a scheme for transferring profits from private companies to Swedish citizens. The proposal was voted down as a dangerous socialist experiment.

    Maybe it is time to take up this idea again?