The STEM Crisis is Real

Recently Robert Charette penned a blog post on IEEE Spectrum titled “The STEM Crisis is a Myth,” which makes the case that rather than having a shortage of STEM (science, technology, engineering and mathematics) workers, the United States has a surplus. Unfortunately his analysis is flawed. Indeed, the post is strikingly similar to a report published by the Economic Policy Institute (EPI) (which Charette cites) in April, and about which the Information Technology and Innovation Foundation published a comprehensive rebuttal.

First, like EPI, Charette argues that STEM graduates are not obtaining STEM jobs:

In the United States, you don’t need a STEM degree to get a STEM job, and if you do get a degree, you won’t necessarily work in that field after you graduate. If there is in fact a STEM worker shortage, wouldn’t you expect more people with STEM degrees to be filling those jobs?

Yes, it’s true: some STEM graduates don’t obtain STEM jobs. But Charette is missing a baseline for comparison of this figure. Sure, some STEM graduates work in fields outside their major, but is this more or less than graduates in general? In fact, STEM gradates are significantly more likely than the average graduate to work in the field of their major: 81.2 percent of employed STEM graduates consider their job related to their major, compared to just 72.5 percent of employed graduates with any major. Moreover, some people get STEM degrees and realize that a career in STEM is not for them, not because of money or job prospects, but because they prefer to be doing other things.

In addition, Charette argues that “even in the computer and IT industry, … not everyone who wants a job can find one. A recent study by EPI … found that more than a third of recent computer science graduates aren’t working in their chosen major; of that group, almost a third say the reason is that there are no jobs available.”

What Charette ignores is that, almost unequivocally, computer science graduates have superior labor market outcomes compared to the average graduate. Looking at both employed and unemployed graduates, 73 percent of computer science graduates are employed in a job related to their major; just 61 percent of all graduates are. Likewise 5.4 percent of computer science graduates reported that a computer science job was unavailable; but over 8 percent of graduates of any major reported that they were unable to find jobs in their field of study.

Oddly, Charette also makes the contradictory argument that STEM jobs are being filled by non-STEM graduates, and thus no need for STEM graduates:

If many STEM jobs can be filled by people who don’t have STEM degrees, then why the big push to get more students to pursue STEM?

The reason is simple: there are too few STEM graduates to fill STEM positions,  and thus employers are forced to draw from the pool of non-STEM graduates. This is a symptom of the crisis, not a solution.

Charette also criticizes a 2011 study that projected growth in science and engineering occupations:

The Georgetown study did not fully account for the Great Recession. It projected a downturn in 2009 but then a steady increase in jobs beginning in 2010 and a return to normal by the year 2018. In fact, though, more than 370 000 science and engineering jobs in the United States were lost in 2011, according to the Bureau of Labor Statistics.

We were unable to find numbers in the Bureau of Labor Statistics (BLS) database to support this claim. Take a look yourself. These kinds of jobs were not lower in 2011 than in 2010.  In fact, comparing these two more detailed tables, we find that the United States added somewhere between 60,000 and 256,000 science and engineering jobs in 2011, depending on what is counted as science and engineering.

Finally, Charette argues, like the EPI report, that slow wage growth in IT indicates there is a surplus of computer graduates (he extrapolates IT to STEM as a whole):

The price of labor has not risen, as you would expect it to do if STEM workers were scarce. In computing and IT, wages have generally been stagnant for the past decade.

Is IT wage growth flat? The reality is significantly more nuanced than Charette makes it seem. High-skill IT occupations, such as software development, saw wage growth, while lower-skill occupations, such as software programming have seen wage declines.

But even if wage growth has not been as high as one might expect in an industry with shortages, what Charette overlooks is the global nature of this labor market and the tech industry generally.  STEM occupations face global competition, which puts some limits on wage growth. If enterprises in the United States increase wages too much, then they lose competitive advantage to enterprises in other economies. For example, computer programming (and occupations like it) can be more easily outsourced to lower-cost foreign nations such as India, and this depresses the rate at which programmer wages can rise. Meanwhile, American workers maintain an advantage in relatively high skilled and high paying occupations such as software development, which is why these sorts of occupations have seen salary increases. This “globalization” effect is not true with occupations like truck driving or nursing, where labor cost increases are more easily absorbed.

This post by Robert Atkinson was featured on IEEE Spectrum’s homepage debate on STEM education.