Polarization and the Falling Middle

Last week’s post was all about how technology might be biased towards higher educated workers. Since these workers already made a lot, new technologies have tended to drive inequality up. However, the reason behind the suffering wages of middle-income workers was probably less clear. After all, technology made those with a college education more productive. But, technology didn’t make those with less education any less productive. Yet, middle-income workers’ wages have fallen over the last four decades.

The question is: if technology has helped some workers, is it possible it has hurt others? Some economists think that the answer is “yes,” with the workers right in the middle being hurt. So, why did this damage happen? It turns out, that computers and other machines are good at tasks that can be broken down to a series of readily anticipated contingencies. A robot on an assembly line can put a screw into a hinge as long as the hinge is always in the same place, always facing the same direction, etc. Machines are good at the routine.

Here’s the issue — the routine jobs that machines have replaced aren’t necessarily low-wage jobs. Consider janitorial work, which stereo-typically is low paying. It would be very hard to program a machine to deal with moving furniture, navigating stairs and hallways, and knowing how to deal with unanticipated events (like a college dorm completely covered in pistachios?!).

On the other hand, middle-income jobs can find themselves in the cross hairs. For example, a student of mine mentioned that their mother was a court reporter. This job is solidly middle-income, and often doesn’t require a college degree. But, voice recognition software is improving, and the student’s mother felt her job was threatened. Indeed, Utah has used digital recording to replace some of these workers. The only thing that might save the day for these workers are the aspects of the job that are non-routine — like Boston accents.

So what does any of this have to do with the word “polarization?” Did I just want an excuse to post a picture of a polar bear? Maybe. But that’s only one reason. It turns out, technology has tended to replace jobs in the middle. And the jobs that have replaced those middle-income jobs have tended to be either high or low wage. The jobs have moved to the poles…hence the term polarization. The net result has been less demand for jobs that used to be middle class — and subsequently lower pay.

Skeptical? Let’s look at the data. The below combines data from two economists far better than me — David H. Autor and David Dorn — to classify jobs based on how routine they are. The figure classifies jobs by how much they paid in the 1970s. From left to right, it moves from the lowest paying to highest paying. Turns out the jobs that were more routine than average (positive numbers) were right in the middle.

Figure 1. RelativeRoutineness” of Jobs Based on Pay in the 1970s

Note: Authors’ calculation from Current Population Survey March Supplement and Author and Dorn (2013). Calculation based on men’s Earnings.

And, to be sure, the occupations in the middle have also suffered the largest wage losses. The figure below shows wage growth for the same occupation bins. Not surprisingly, wage growth has been negative for jobs in the middle, and positive for both low- and high-paying occupations. Pay has experienced polarization. The jobs that are the least routine are at the poles and are the jobs that have experience wage growth.

Figure 2. Wage Growth between 1975-1979 and 2013-2017 by Occupational Pay in the 1970s

Authors’ calculation from Current Population Survey March Supplement. Calculation based on Men’s Earnings.

So, now we have a more complete picture of the role of technology. Technology has tended to be biased towards those with more education, pushing their wages up. Technology has tended to replace jobs that are routine, which have tended to be middle-income. Hence, the rise of the top and the fall in the middle — polarization.

Before moving on to talk specifically about women’s wages, I want to spend the next few posts talking about a few other reasons middle-income workers have been losing out. So, if you like what you’ve read so far get pumped! Because the next few weeks will include exciting topics like globalization, unionization, and even some celebration! OK, celebration is a bit much…but those other things will be there.

Leave a Reply