A new NBER paper released last month examines how minimum wage increases in the U.S. affect employment in low-skilled automatable jobs. Consistent with the majority of papers on the subject, it finds that low-skilled workers (high school diploma equivalent or less) holding such jobs become unemployed as a result of minimum wage increases.

As the authors explain, many papers examining the effects of increasing the minimum wage focus on teenagers and restaurant workers. However, this study aims to shed light on how other subgroups are disproportionately affected by minimum wage increases:

...the perspective we adopt in this paper suggests there may be subgroups of workers among those groups not usually considered in the minimum wage literature who may be adversely affected by minimum wages, because they tend to be employed in automatable jobs.

My colleague Sam Bowman has previously explained what elasticities mean in the context of the minimum wage debate: “an elasticity of -0.5 means that a 1% rise in the minimum wage is associated with a 0.5% fall in employment for the affected group.” Although the paper measures the effects of a minimum wage increase in a different way, the corresponding elasticities can easily be calculated.

Overall, the estimated elasticity for low skilled automatable jobs is -0.08: a small but negative effect on employment. However, the effects are markedly more pronounced for older workers (age >40) in automatable jobs and their young counterparts (age ≤25). In the case of automatable service industry jobs, low-skilled young people saw an elasticity of -0.28, with the figure for older workers being -0.22. Older manufacturing workers faced an elasticity of -0.19. Younger manufacturing workers fared even worse, with an elasticity of -0.53.

The authors also find that women in automatable jobs are more likely than men to become unemployed as a result of minimum wage increases. For men, overall elasticity was -0.09. For women, it was -0.13.

As for those low-skilled workers that managed to hold onto their automatable jobs post-minimum wage hike, the paper finds a decrease in the amount of hours worked:

...a $1 increase in the minimum wage generates a 0.23 decrease in hours worked for low-skilled individuals who held an automatable job in the previous period. The decline is negative and statistically significant in manufacturing, transport, wholesale, retail, and services (sometimes only at the 10-percent level).

Perhaps the most interesting finding is that the effects of minimum wage increases on automatable jobs seem to be getting larger over time. The Bureau of Labor Statistics dataset used by the authors runs from 1980-2015, but using data from 1995-2016, the authors find a noticeably higher disemployment effect (an overall elasticity of -0.27). This is hardly surprising; as technology advances, the cost of substituting labour for capital goes down. But there is now empirical evidence to support the claim that the disemployment effects of the minimum wage are getting worse over time.

Of course, this is just one paper in a sea of economic literature on the effects of the minimum wage. But it’s worth remembering that even if the minimum wage doesn’t kill jobs, or lower overall hours worked among the least well-off, it may still hurt the poor in other ways.