The Challenge of Retraining Workers for an Uncertain Future

By Adam Thierer of Mercatus. Excerpt:

"The reality is, most worker retraining plans are little better than a dice-roll on the professions and job needs of the future. As I noted in my last book as well as in a paper with Andrea O’Sullivan and Raymond Russell, concerns about automation, AI, and robots taking all our jobs have put worker retraining concerns back in the spotlight in a major way. That has led many scholars, pundits, and policymakers to suggest that more needs to be done to address the skills workers will need going forward.

That impulse is completely understandable. But it doesn’t mean we can magically predict the jobs of the future or what skills workers will need to fill them. It’s not that I am opposed to efforts to try to figure out answers to those questions, or perhaps even craft some programs to try to address them (although I agree with my colleague Matt Mitchell that many past worker training programs “seem indistinguishable from corporate welfare.”) But worker retraining or reskilling usually fails because it’s like trying to centrally plan the economy of the future. It’s a fool’s errand.

In my book, I pointed out that, when you look back at past predictions regarding the job needs of the future that we now live it, those predictions were off-the-mark. The fact is, an “expert” writing in the early 1980s about the job needs of the future didn’t even have the vocabulary to describe or understand the jobs of the technological era we now live in. Here’s how I put it in my book:

It’s also worth noting how difficult it is to predict future labor market trends. In early 2015, Glassdoor, an online jobs and recruiting site, published a report on the 25 highest paying jobs in demand today. Many of the job titles identified in the report probably weren’t considered a top priority 40 years ago, and some of these job descriptions wouldn’t even have made sense to an observer from the past. For example, some those hotly demanded jobs on Glassdoor’s list include[1] software architect (#3), software development manager (#4), solutions architect (#6), analytics manager (#8), IT manager (#9), data scientist (#15), security engineer (#16), quality assurance manager (#17), computer hardware engineer (#18), database administrator (#20), UX designer (#21), and software engineer (#23).
Looking back at reports from the 1970s and ’80s published by the US Bureau of Labor Statistics, the federal agency that monitors labor market trends, one finds no mention of these computing and information technology–related professions because they had not yet been created or even envisioned.[2] So, what will the most important and well-paying jobs be 30 to 40 years from now? If history is any guide, we probably can’t even imagine many of them right now.
Of course, as with previous periods of turbulent technological change, many of today’s jobs and business models will be rendered obsolete, and workers and businesses will need to adjust to new marketplace realities. That transition takes time, but as James Bessen points out in his book Learning by Doing, for technological revolutions to take hold and have a meaningful impact on economic growth and worker conditions, large numbers of ordinary workers must acquire new knowledge and skills. But “that is a slow and difficult process, and history suggests that it often requires social changes supported by accommodating institutions and culture.”[3] Luckily, however, history also suggests that, time and time again, society has adjusted to technological change and the standard of living for workers and average citizens alike improve at the same time.

Bessen’s point is really important, and too often forgotten in discussions about reskilling for the future. When I think about the sort of skills that I picked up the early 1980 as a teenager using a clunky old Commodore 128 computer, or that my own teenage kids pick up today just by tinkering with their gadgets (computers, smartphones, gaming consoles, etc), I think about how those skills were not centrally planned for by anyone. It was mostly just learning by doing. A lot of the coding skills people use today they learned by trial and error and without taking any course to do so.

In his book, Bessen uses the example of bank tellers to illustrate how convention wisdom about future trends is often wildly off the mark. With the rise of ATMs a few decades ago, many thought the days of bank tellers were numbered. But Bessen’s research shows that we have more bank tellers today than we did 40 years ago because once the ATMs could handle the menial tasks of counting and distributing money, the tellers were freed up to do other things.

I’m not saying we can just leave the future of workers to chance and hope everyone can learn on the fly like that. Some government programs will be needed, and many could even help. But let’s not kid ourselves into thinking that we somehow have a crystal ball that we can stare into and, like a technological Nostradamus, somehow divine the jobs and skills of a radically uncertain future.

Our better hope lies in creating an innovation culture that is open to new types of ideas, jobs, and entrepreneurialism. We might better serve the workers of the future by ensuring that they are not encumbered by mountains of accumulated red tape in the form of archaic rules, licenses, permitting schemes, and other obstacles to progress. My colleague Michael Farren also testified last year and offered some concrete near-term reform proposals to help bridge the skills gap by “revis[ing] the federal tax code to allow tax deductions for all forms of productivity-enhancing investments, including investment in training workers to perform new jobs,” and also addressing government aid programs “that might be lowering the supply of workers, thereby contributing to the lack of skilled workers available.”"