Work is rapidly becoming nonroutine

Some researchers at the Federal Reserve Bank of New York wondered if something structural is going on in the US workforce, in particular, if some underlying change in the nature of work was keeping unemployment high. And they sliced the census data in a different way, trying to find if work is becoming less routine. And it is.

Stefania Albanesi, Victoria Gregory, Christina Patterson, and Ayşegül Şahin, Is Job Polarization Holding Back the Labor Market?

To study job polarization, it’s useful to put occupations into different task groups, with each occupation categorized as either routine or nonroutine. An occupation is routine if its main tasks require following explicit instructions and obeying well-defined rules. These tend to be middle-skilled jobs. If the job involves flexibility, problem solving, or creativity, it’s considered nonroutine. Job polarization occurs when employment moves to nonroutine occupations, a category that contains the highest- and lowest-skilled jobs. The chart below shows the employment share of routine and nonroutine jobs in the United States starting from 1976. The share of routine occupations declined from 60 percent in 1976 to 40 percent in 2012, resulting in job polarization.


It’s useful to further classify occupations into cognitive and manual categories based on the amount of mental or physical activity required of the job: cognitive jobs require mostly mental activity; manual jobs, mostly physical.

For instance, nonroutine cognitive jobs are generally high-skilled and include management, technical, and healthcare workers, such as doctors and engineers. Manual nonroutine occupations are generally low-skilled and include service and protection workers, such as waiters and security guards. Cognitive-routine jobs include sales and office workers, such as sales agents and office assistants. Manual routine occupations include construction workers, mechanics, and machine assemblers.

As the chart below shows, the share of routine jobs has been steadily decreasing for the cognitive and manual categories since the 1980s, and the decline accelerated during the 1981-82 and 2007-09 recessions. Most of the rise in the employment share of nonroutine jobs reflects the increase in cognitive nonroutine occupations.

cognitive v manual

These are large transitions: a 20% increase — from 40% to 60% — in the number of jobs involving non-routine work in the 40 years since 1975. It’s fairly linear, so we could extrapolate, I guess, that by 2025 an additional 10% of jobs will swing toward non-routine.

The researchers found, however, that the ongoing unemployment in the US is not being caused by this ‘job polarization’. Both nonroutine and routine occupations are having difficulties finding work. So, while this trend is important, the difference in cognitive skills among the routine occupations is not the root cause of unemployment: it is not a structural issue, since even the unemployed in non-routine occupations are having difficulties finding work.

Still, the authors don’t touch on a related issue. In a better economy of the near future, where the demand for non-routine work would be increasing above the available levels, we might see a structural issue, and that would be solved by retraining those currently working in routine occupations so they could move into the more cognitive and less manual jobs. Less construction workers and more waiters, for example.

The larger questions are the shift to different styles of work as more work is unprogrammed, and no longer following the orthodoxies of tightly scripted business processes. I’m addressing some of that in an upcoming report on the shift away from business processes based on the adoption of social network-based cooperative work patterns.