Working out automation
The future of work is unclear. Here are five important, overlooked questions about technological change and jobs.
We don’t know whether automation, algorithms and artificial intelligence will ultimately create more jobs than they destroy. Opinions are all over the map. One widely cited study predicted 47 per cent of jobs would be automated, and technological change has, in fact, contributed to declining employment in recent years. Some are already preparing for a world without work.
But automation has been going on for centuries, and jobs still exist: that’s because automation replaces some kinds of human labour while boosting demand for others.
Furthermore, job upheaval today is relatively modest. The mix of jobs in the economy is changing more slowly in recent decades than in the 1940s and ’50s.
Today, economists worry that the labour market isn’t dynamic enough; many measures of fluidity and dynamism, like migration and job turnover, have been declining for decades.
But this uncertainty should not distract us from other questions about automation that we’re sure to face regardless of whether automation adds to or subtracts from the total number of jobs. Here are five important, overlooked questions about automation and jobs:
Will workers be able to transition?
The pain from automation arises not only from how many jobs are eliminated, but also from whether workers in automated jobs can transition to other work. On Indeed’s website, we have data on how some workers in threatened occupations are seeking new opportunities, such as retail workers looking at customer service. But transitions may be harder than in the past. Job churn has slowed in recent decades, as firms both hire and fire less than they used to, and because people move less than before. The labour market may be changing less today than in the 1940s and 50s, but today’s slower employment growth and lower mobility could make transitions more drawn out and painful.
Who will bear the burden?
Regardless of how many jobs are eliminated by automation, the pain will be uneven. The less-educated are far more likely to work in “routine” jobs, which are more susceptible to automation, than workers with a tertiary degree. Men are more likely to work in routine jobs than women are. And the geographic divide is stark: just a third of jobs in metro Washington, D.C. and San Jose, California are routine, versus half or more in much of inland California and many smaller southern and Midwestern metros. These regional differences line up with the partisan divide: counties that voted more strongly for Donald Trump in 2016 have a higher share of routine jobs and therefore are more likely to be affected.
How will the supply of labour be affected?
Automation might affect labour supply, not only labour demand. Just as past technological innovations, like washing machines and kitchen appliances, reduced the time needed to do household work and contributed to the entry of women into paid employment, future technological advances related to automation might also shift how much people are willing and able to work. For instance, autonomous vehicles might turn commuting into productive work time. Or, autonomous vehicles could chauffeur kids to school and activities, freeing up parents to work more hours. Alternatively, automation could boost productivity and lower consumer prices, possibly reducing labour supply since people will need to work less to afford the same items. It’s far from clear which of these effects will win out.
How will automation affect wages?
The pace of automation depends on prices, not just technological feasibility. Just because a robot or algorithm can perform a task as competently as a human doesn’t mean that humans will be replaced. Automation depends on the cost of the technology relative to the cost of human labour. In today’s tight labour market, for instance, rising wages and worker shortages might encourage automation and boost productivity. At the same time, automation that replaces workers in some sectors could push them into the labour supply for other sectors, potentially depressing wages, slowing productivity and aggravating inequality. Again, it’s not clear which force will be stronger.
How will searching for jobs change?
Artificial intelligence has the potential to predict better matches between job seekers and open positions. Automated screenings and tests can potentially remove human biases that disadvantage certain candidates. However, algorithms might also reinforce human prejudices if the algorithms are trained on biased data sets. Plus, algorithms might be differentially applied to certain groups: One expert warns of a future where “the privileged … are processed more by people, the masses by machines.” Finally, people might be skittish about automated hiring. A recent survey found people less enthusiastic about algorithms evaluating job candidates than about self-driving cars or robot elder caregiver, which could slow down their adoption. We don’t need to wait to discover whether automation creates more jobs than it destroys to start answering these questions and acting on the answers. Making job transitions easier, focusing on those most at risk of job loss and thinking about labour supply, wages and job searches are essential for navigating these technologies, whether or not automation ultimately adds to or subtracts from overall employment.
Jed Kolko is the chief economist at Indeed. Copyright 2019 Harvard Business Review/ Distributed by NY Times Syndicate
To join the conversation, please log in. Don't have an account? Register
Join the conversation, you are commenting as Logout