This post tackles whether those who work with their hands earn a premium and have brighter job prospects than those who work with machines. The headline result is that those who work exclusively with their hands and those who work exclusively with machines make more money than those who work with their hands and machines together. The job prospects, however, of those who work with their hands and those who work with machines are not significantly correlated. In other words, job prospects are not related to whether or not a person works with her hands or with machines.

I used O*Net, a repository of wage and employment information, to classify whether a person works with their hands, works with machines, or works with their hands and machines together. I first searched for “hand” and extracted wage and employment information for those occupations that involved using hands. Examples of occupations that require workers to use their hands include dentists, explosives workers, and fine artists. I then searched for “machines” and extracted wage and employment information for those occupations that involved using machines. Examples of occupations that require workers to use machines include cardiovascular technologists, computer systems engineers, and geoscientists. Finally, I extracted wage and employment data for those who use both their hands and machines together on the job. Examples of occupations that require both hands and machines include farmworkers, construction workers, and automotive repairers.

Turning first to whether or not those who use their hands or those who use machines have brighter job prospects than those who use both hands and machines together, I run a regression of projected job openings between 2018 and 2028 on “hand” and on “machine.” The excluded category is “hand and machine.” Therefore, the interpretation of the regression coefficient on “hand” and on “machine” is relative to “hand and machine.” I also control for the number of people currently working in each occupation in 2018. The coefficients for both “hand” and “machine” are highly insignificant which means there is no significant relationship between “hand,” “machine,” and job prospects.

Next, I turn to whether or not those who use their hands and those who use machines earn a premium compared with those who work with both their hands and machines together. I ran a regression of median annual wage in 2019 on “hand” and on “machine.” Once again, the excluded category is “hand and machine.” The interpretation of the regression coefficients on “hand” and on “machine” are relative to “hand and machine.” I also control for the number of people employed in each occupation in 2018. The coefficient on “hand” suggests that those who use their hands earn $25,220 more than those who use “hand and machine.” The coefficient on “machine” suggests those who use machines earn $16,512 more than those who use “hand and machine.” Both of these statistics are highly significant, which says that it is unlikely to be due to chance.

The unresolved questions is, “why do those who work exclusively with their hands or exclusively with machines make more than those who use their hands and machines together?” One potential explanation is that those who use their hands exclusively are in occupations that are difficult to automate. Therefore, working with hands earns a premium. Those who use machines exclusively are in occupations that require technical knowledge to operate sophisticated machinery. Finally, those who use both hands and machines are in occupations where workers are easily substituted–there is neither cache in the handiwork nor specific knowledge required to operate the machines.

Nevertheless, the interpretation of the results presented in this post suggests that further research is required to explain why those who use their hands and machines make less than those who exclusively work with their hands and those who work exclusively with machines. I posit an explanation, but further work would help shed more light on the interpretation of these results.