The Trend is the Cycle: Job Polarization and Jobless Recoveries is an academic paper, but very approachable and well worth reading. The key finding is the job market is becoming more polarized, with job growth becoming concentrated at the high and low wage ends of the spectrum.
The study authors crunched through a mix of government data, breaking jobs into 3 broad categories:
1. Non-routine cognitive jobs: These are managerial, professional and technical workers - in other words highly educated and skilled knowledge workers. These jobs tend to be higher paying.
2. Routine cognitive and manual jobs: These are jobs that are accomplished by following defined processes and procedures. Many clerical and manufacturing jobs fit in this category. These jobs tend to be good paying, middle class jobs that haven't required high levels of education.
3. Non-routine manual: These are service jobs and include such things as bartenders, home health aides and others that are manual in nature, but require flexibility, problem-solving or human interaction skills. These jobs tend to be relatively low paying.
As the chart below (from the study) shows, over the last 3 decades the percent of jobs that are non-routine cognitive and manual have grown, while the percent of routine jobs have fallen.
The good news is the number of high paying knowledge jobs has rapidly grown over the last 3 decades, moving from 29% of jobs to 39%. The bad news is routine jobs - the backbone of the middle class - are going away due to automation and foreign competition. Their share of jobs has fallen from 58% to 44%.
The other bit of bad news is we will likely continue to lose routine jobs. This hollowing out of middle class jobs is one reason median household incomes have fallen over the last decade. It's also one of the reasons unemployment and under-employment are so high.
The well known VC Mark Andreesen says these shifts are resulting in two types of jobs: (1) jobs where people tell computers what to do; and (2) jobs where computers tell people what to do. Needless to say, the jobs where computers tell people what to do are not going to be high paying.
We think the future of jobs is more nuanced than this, but it's a great sound byte.
We'll have more on this in the coming weeks. For those interested in this topic, The Polarization of Job Opportunities in the U.S. Labor Market is another approachable academic paper.