How the rising minimum wage could lead to more automation

The minimum wage is rising rapidly. In 2015 4 per cent of employees aged 25 and over were on the legal minimum hourly wage of £6.70. Since then, George Osborne’s “national living wage” (NLW) policy has raised the wage floor to £7.50 and it is planned to reach 60 per cent of median wages in 2020 (£8.56 under current forecasts of wage growth).

Because these increases have been, and will continue to be, much greater than for wages in general, the fraction of employees aged 25+ subject to the minimum looks set to treble in just five years, reaching 12 per cent in 2020.

There is a case for a higher minimum wage to help low-wage people—especially since evidence up to 2015 suggests we hadn’t reached the point where a higher minimum starts affecting the employment prospects of those it’s designed to help. But clearly if we raised the minimum indefinitely we would eventually reach such a point: firms appear to be able to adjust to a minimum of £7 per hour without shedding jobs, but no one supposes the same would be true at £100 per hour.

The challenge is that we don’t know at what point inbetween the negative impacts on jobs (and/or hours of work) would become larger than we’re willing to tolerate. So as we push the minimum up into uncharted territory we need to do so in a careful manner that allows us to respond to new evidence on its effects.

New IFS research uncovers another reason why previous evidence on the employment effects of the minimum wage is of limited value when looking to the future: the jobs affected by a higher minimum are quite different from those affected by it before. Many of the very lowest paid jobs are personal service and care jobs. The slightly higher-paid roles that will now be brought within the minimum wage net include a lot more retail cashiers and receptionists, to give two examples. The latter jobs are more easily automated than the former. As a result, in many cases it would now be easier than it was before for firms to respond to a rising cost of labour by switching to machines.

To investigate the possibility of automation more comprehensively, the research used a proxy for “ease of automation” that has been adopted in academic research, known as the “routine task intensity” (RTI) of each occupation. It then looked at how RTI varies for people at different wage levels. The logic behind using RTI is that current technologies are relatively effective at executing “routine” tasks—such as calculating supermarket bills—and less effective at tasks that involve forward planning, responding to unusual situations or human interaction. As a result, we would expect jobs involving a lot of routine work to be more easily automated than other jobs. While RTI measures are only proxies for “automatability,” research has found that they are both strong predictors of past technological adoption and capable of explaining labour market trends in a number of advanced economies.

The research reveals that, as the minimum wage is raised, it increasingly affects workers in occupations that are easier to automate. It turns out that the workers set to be brought within the minimum wage net in 2020 are more than twice as likely to be in the top 10 per cent most routine occupations as those already paid the minimum in 2015 (11 per cent compared with 5 per cent).

This suggests that minimum wage employers in 2020 would tend to find it easier to respond to a higher minimum wage by substituting workers for machines. However, this does not mean planned rises in the minimum wage will definitely have adverse consequences for low-skilled workers. The reality is more uncertain—we do not know what the employment effects will be—but also more nuanced than that. While technology can replace certain jobs, it can also create new jobs that are complementary to that technology (e.g. people to service machinery). If automation takes place, it would matter greatly whether low-skilled workers are able to move from jobs that are automated to jobs that are created, or whether the new jobs would go to others such as those with more advanced skills.

There is much more to learn here, and the recent and upcoming policy changes in the UK offer an opportunity to do so. As new evidence comes in, policymakers seeking to help low-wage workers should respond accordingly.