Is Technology Biased?

A few weeks back, I pointed out that over the last forty years, middle-income workers have seen their wages grow slowly. In fact, the typical working man has seen their real wage decline since the mid-1970s. And while middle-income women have seen some gains, their higher-income counterparts are doing a heck of a lot better. (And, yes, “heck of a lot” is a technical, Economic term.) The net result has been rising income inequality. This post is the first of several that will try to help explain this phenomenon. And the post takes a time-honored approach — blame technology! So, how could rising income inequality be caused by technology?

The answer is pretty simple. You see, one of the key inputs into how much employers will pay their workers is those workers’ productivity. How much stuff do the workers produce? The more workers are able to produce, the more the employer will be willing to pay them. Other factors matter too of course, as the primer video on labor demand and supply provided below explains (if you don’t want more detail on labor demand and supply, feel free to skip). But, productivity matters a lot, and if only some workers get more productive, that will lead to income inequality.

Which gets us to technology. One of the good things technology does is to make us more productive. I mean, sure, sometimes we spend a bit too much time at work watching Golden Retrievers at a pool party on YouTube. But, then again, while I’m watching that cuteness I can do a statistical analysis on my other monitor, make a pretty graph in Excel, and post it on the internet. All things that were impossible in the 1970s.

Here’s the rub. Many economists believe that one reason for rising income inequality is that new technologies have improved the productivity mainly of more educated workers. And these workers already made more money than others.

To illustrate, the figure below uses data maintained by the Department of Labor to break occupations into three educational requirement groups — low, medium, and high. It then identifies how many individual technologies (e.g., mapping software, word processors, statistical packages, etc.) the job uses. It also identifies how many of those technologies are “Hot,” which means they are posted as required for many new jobs. The point of the figure is pretty simple: jobs that require more education use a whole lot more new and “Hot” technology.

Figure 1. Number of Technologies and “Hot” Technologies used by Jobs by Level of Education Required

Note: Calculations based on O*Net 23.1 Database. This database comes from the Department of Labor and can be found here:

Since wages are driven by technology, and new technologies are used more by highly educated workers, the figure below should not surprise you. It shows how the earnings of various college-educated workers have changed relative to high-school educated ones. The figure shows two things worth noticing. First, all college educated workers have seen wage gains relative to high-school educated ones. Second, the larger increases in income have gone to the most educated workers. Hence, income inequality is rising.

Figure 2. Ratio of Annual Income for High School Educated to College Educated Workers, 1975-2017

Note: The figure shows the ratio of median earnings for full-time full-year working men aged 25-54 with a formal degree to those with a high school degree. Data come from The Current Population Survey March Supplement.

So, one explanation for our first stylized fact is that technology has benefited the most educated. Indeed, education is now a gateway to the upper class in a way that it simply was not in the past. In 1975, only 43 percent of workers in the 80th percentile (higher income than 80 out of 100 people) of income had at least an Associate’s Degree. Today, that number is 69 percent.

I would be lying if I said that this explanation of technology “biased” towards the more educated was the only driver of inequality. After all, this explanation explains why some workers are doing well — those with more education. But, so far, I have not explained why some workers — those in the middle — are actually losing out. For that explanation, we are going to need to learn about something that sounds cool, but has damaged the middle-class: polarization (pun intended). Tune in next week.

2 thoughts on “Is Technology Biased?

  1. Any technology can have a mixture of two effects on a labor market: it can enhance the productivity and hence the demand for a particular skill set, or it can act as a substitute for a particular skill set. The most at risk of the latter outcome are the lower-paid workers whose comparative advantage is in the performance of repetitive tasks that are most susceptible to substitution by technology.

Leave a Reply