Eric Hanushek
Lifelong Learning and Dynamic Adjustments to Changes in the Labour Market
Keynote Speech excerpts, PILLARS Inaugural Conference
When the effects of automation, robots, and AI are discussed, the focus so far tends to be on costs and harm. Less is said about what can be done to lessen potential harm, which leads to the idea of lifelong learning.
Here vocational education plays a key role, but the views on it around the world differ widely: from being eliminated in the US (and resurrected by Trump) to much more reliance on it in Europe (with Germany often cited as an example of successful application of vocational education leading to employment).
Strikingly, a study led by Eric and Ludger Woessmann with their colleagues Guido Schwerdt and Lei Zhang, showed that the returns to vocational education taper off in the later years of the working life, as shown below:
The reason could be skill obsolescence, not keeping up with technological developments, or maybe different retirement planning.
But when it comes to discussing the related issue of lifelong learning, it is often more slogan than reality, with details seldom provided. Encouraging adult education is one of the ideas being floated, but usually accompanied with many as yet not properly answered questions: Should employers be incentivised to be more active in training their employees? What if employers countered with why should they retrain a middle-aged employee when they could more easily hire a younger one? So, should incentives be provided to employees instead? Or should training be government provided?
In view of this, Eric and his colleagues looked at it from a different standpoint. Starting from the fact that while the net aggregate effect of robots, trade and AI disruption may be zero or even positive, this is not true for individuals. Their proposal turns worker displacement on its head, by taking a cohort of young workers and focusing on adjustments before plant closures occur.
Knowing that the world will change for displaced workers, do they invest in enhancing their human capital? Do they get a payoff of that investment? Should they pursue other adjustments before closures? Who exactly should adjust?
To find out, Eric and his colleagues examined integrated employment histories (such as those available in Germany), with unique person and establishment identifiers, taking a sample that included people who completed first apprenticeship or started their first job between 1995 and 2010, mainly from large firms, observing plant closures 5+ years after employment entry, and dividing their “universe” among early adjusters, late adjusters, and controls (“stayers”).
Given that the terminology can be somewhat confusing, a clarification: with “Late adjusters” is meant people who do adjust, instead of jumping ship at least a year ahead of plant closure; those that jump ship early would be the “early adjusters”, and in Eric’s study are not being taken into account as people reacting to the disruption itself; they are likely simply bad job matches. Late adjusters, in contrast, take the hint and start investing in reorienting themselves occupationally before the plant closure approaches.
The researchers then make a comparison between stayers vs. late adjusters.
Late adjusters experience a huge loss of income immediately, but then they climb up afterwards and end up overtaking stayers in income. Adjusters, in other words, do better over time.
In terms of second apprenticeships, late adjusters do a lot more in terms of getting new training:
Late adjusters are also much more likely to switch occupations with plant closures:
Late adjusters are much more likely to switch regions, compared to control:
How about external disruptions, such as the often-mentioned robots, AI and trade? Well, robots are mainly employed in the automobile industry, so the authors won’t focus on them. As to AI, it is a bit early days yet. Thus, their focus is on trade exposure:
By being in “risky” industries (more exposed to trade), late adjusters are led to make more changes to their occupation compared to those in low exposure industries.
The upshot is that late adjusters make larger adjustments (occupational and regional in nature), a trend that is amplified when being in risky/threatened industries (trade exposed). They also invest more, in such things as second apprenticeships or going back to university. Their reward is that they recover pre-event earnings after disruption and major changes, and then some.
Despite these findings, more research is necessary (isn’t it always?) to further refine strategies to help workers cope with disruptive occupational changes.