The BLS Thinks These Jobs Will Grow a Lot. I Think They’re Wrong.

by Andrew McAfee on February 23, 2012

In a post at The Atlantic, Jordan Weissmann draws a couple interesting graphs using data from the BLS’s recently-released projections of job growth to 2020. Weissmann concentrates on those jobs that require only a high-school diploma or less, and points out the surprising stat that 63% of all US jobs created (12.8 million total) will be in this category. Here are the ones the BLS thinks will grow most, arranged in descending order of volume:

From "The Future of Work for High School Grads," by Jordan Weissmann

My first question when I look at this graph is not ‘what do these jobs pay?’ but instead ‘which of these will be deeply affected by automation and other kinds of technological progress — more deeply than the BLS is currently thinking?’ This of course is an impossible question to answer precisely, but after spending some time looking at the BLS’s projections methodology and its factors affecting occupational utilization, I think it’s likely underestimating the broad impact of technology.

The BLS is certainly not alone in this. As we wrote in Race Against the Machine, the pace and scope of digital progress is so quick it’s catching a lot of us by surprise. I certainly wasn’t expecting self-driving cars and Jeopardy! champion supercomputers any time soon, yet here they are.

We also highlighted how much less labor-intensive sectors like retail have become, so when I see ‘Retail salespersons’ as the job with the greatest projected growth in this category I get skeptical. Same thing with ‘Office clerks,’ ‘Customer service reps.,’ ‘Accounting clerks,’ and ‘Cashiers.’ Thanks to technology, I predict that our economy will need a lot less of each of these per unit of GPD, which implies that GDP is going to need to grow really fast — faster than projected, in all probability — to generate the kinds of jobs growth projected in the chart above.

And if over the next few years we see breakthroughs in robotics similar to the ones we’ve seen in AI and machine learning over the past few years, a lot of the other jobs could grow less than projected above. Are heath care, fast food, janitorial, groundskeeping, and construction robots going to leave the labs and enter the workforce before 2020? Every time I watch another set of robot videos, I get more convinced that they are.

Because a lot of lower-education jobs are physical, they’ve to date been largely isolated from the effects of automation. I think this isolation might well be coming to an end. Do you agree? Leave a comment, please, and let us know.

And if you were wondering what these jobs pay, here’s another graph drawn by Weissmann, with jobs listed in same order as in the graph above:

From "The Future of Work for High School Grads," by Jordan Weissmann

Yuck. Many of these don’t look like very good jobs, do they?


Gordon Haff February 23, 2012 at 2:29 pm

And I’m guessing that the category with the highest salary (sales reps) covers a lot of jobs that do, in fact, typically need a college degree even if they all don’t.

Ted Kupper February 23, 2012 at 5:27 pm

Hey Andrew, just a heads up that the robot videos link is broken. 

Ted Kupper February 23, 2012 at 5:37 pm

And to answer the question, I definitely agree. 

Clerks, retail salespeople, and cashiers clearly are at risk, as are customer service reps. There should be many fewer of these jobs per unit of GDP in the next 8 years. 

I’m more skeptical that health care, groundskeeping, janitorial and construction workers are apt to be replaced in the 8 year timeframe you posited. These jobs might be possible to automate by then, but to be cost-effective, especially given how human labor prices will plummet, I think wil take longer. 

Fast food is a toss up. The technology should be there, but we’ll see if McDonald’s can figure out how to fire 150,000 workers (or whatever the number is) without causing riots.

You didn’t mention security guards, TAs, truck drivers, freight movers, or medical assistants and secretaries. All these are also, IMO, vulnerable to automation or technologically-enabled outsourcing. 

I think childcare and home health workers are fairly safe, if only because culture will continue to lead people to pick humans for these tasks until very advanced robotics are possible.

Edward February 23, 2012 at 7:22 pm

Andrew, excellent blog. I’ll be attending Penn for med school next year, and as such I’m forced to look at a fairly long time horizon before income starts to actually accrue substantially (4 years med school + perhaps 7 years of residency). Just curious – do you predict that the next 20 years will witness the rise of fully autonomous (or minimally controlled) robot surgeons or telesurgery from low-cost countries? Love to hear your thoughts. Thanks, and keep up the stellar work.

Sean Enhance February 24, 2012 at 6:03 am

We also highlighted how much less labor-intensive sectors like retail have become, so when I see ‘Retail salespersons’ as the job with the greatest projected growth in this category I get skeptical.  And you are so right to be so, in fact I really don’t believe that this sectors will have the growth they predict, not saying that they will not grow, but not as much as the predictions.

John jake March 20, 2012 at 4:38 am
Anonymous May 31, 2012 at 9:25 am

 People get attracted with everything that is free. A free e-course will push people to give their email address. Louis Vuitton Handbags Outlet a seven-part course every few days will keep their anticipation with your company. The e-course indirectly sells your product too. A free e-book is also an attractive method to exchange the email addresses of people. Make the e-book as relevant and as informative as possible. Just make sure that you have links to make your product/service visible. Free give-away and sign-up bonuses are valuable means to get on the lead. Visitors get freebies and bonuses by just signing up and exchanging their e-mails with you.

Anonymous September 24, 2012 at 7:15 am

Retail will need a lot *less* people. This will, like always, happen first where labour is most expensive. Here in Stavanger/Norway, entry-level-wages are at $25/hour, and already some shops have started staffing only half the lines with a human, staffing the rest with a machine that lets the customer scan his goods himself. (to name but one, among many, examples)

Quesolosa November 2, 2014 at 6:21 pm

McDonald’s won’t be firing anyone. They’ll just slow the rate of replacement as they purchase more automation. They have a high turnover rate to begin with so there won’t be any riots.

Comments on this entry are closed.

Previous post:

Next post: