Of Credential Creep
Whether referred to as “degree inflation” or “credential inflation,” the meaning remains the same: employers now require a higher standard of minimum credentialing for a job. The practice takes multiple forms, e.g., requiring a bachelor’s degree for a role that used to just need a high school diploma, or in some cases, requiring a Master’s/PhD when a bachelor’s is probably sufficient. Take a look at these Data Scientist job postings I saw while browsing LinkedIn. In particular, pay close attention to a) the sheer number of people applying to these jobs and b) the breakdown of applicants by education level:
The vast majority of them have Master’s degrees, and more often than not, several hundred of them have PhDs.
I’m Part of the Problem (Or at least a symptom)
Credential creep is personal for me. I myself am a perfect incarnation of its endemic impact, and it’s something I’ve thought a lot about over the last 3-4 years in particular. Nowadays, it feels like a Master’s degree is a part of the minimum job requirement in the data/ML/software space, with PhDs often being preferred.
An anecdotal example is when I interned at Amazon during grad school. I was able to look at all the internal documentation around credential expectations and standards for hiring across different roles (primarily intended for recruiters and hiring managers). I don’t remember the minutia, but the main takeaway was this: you’re probably not getting a data scientist role at Amazon without a Master’s degree, at a minimum (MS required). The standard very well may have changed since I was there, but a common theme among the larger tech companies is that the interview process is fairly standardized. If you’re trying to land a software engineer role at Google or Amazon or Microsoft, there’s essentially a 0% chance you’re getting an offer without going through at least 1 or 2 Leetcode-style live coding rounds. If you’re trying to get a more product-focused Data Scientist position in tech, you better believe you’ll be designing an experiment (A/B test) and live coding in SQL and/or Python.
But to be perfectly honest, I don’t think many roles in this space genuinely require that level of credentialing. Sure, an applied/research scientist at a company like OpenAI or Waymo or any large tech company might benefit highly from having someone in that position with actual research experience. Still, much of data science work is simple programming mixed with cross-functional collaboration, basic experimentation (that anyone, especially with any sort of STEM bachelor’s degree, can understand), and perhaps some model development. Understanding the math behind statistical machine learning and model development is probably the most difficult of all the things I mentioned. Still, you can gain plenty of exposure to everything you need to understand it in an undergraduate program (if you’ve taken calculus, linear algebra, and probability theory + a mathematical statistics course, you’ve pretty much touched on everything you need to know).
During my junior year of college, I felt fairly confident I was not going to get a job (at least, not a job I was interested in) without differentiating myself in some meaningful way. This outlook influenced both my decision to add a second major and to pursue a Master’s degree. It was not the only reason (I was also somewhat bored with just straight economics), but I’d be lying if I didn’t say it played a role in my decision. Thankfully, I genuinely enjoy school so this didn’t make me miserable, but not everyone feels that way. And that’s not even mentioning the cost of continuing education, both in terms of tuition as well as the opportunity cost of foregoing prime earning years to stay in school.
Credential Creep In Tech
This feels like a problem. While the reports on the state of the economy and labor market are generally positive right now, much of the job growth is in lower-wage positions in retail, hospitality, and government. Meanwhile, when it comes to tech, it’s not looking so great. The information sector, which saw a 6-month 0.8% decrease in employment, captures much of the white-collar work primarily focused on industries where technology, media, and information systems are central to job functions.
https://www.bls.gov/charts/employment-situation/employment-by-industry-monthly-changes.htm
But if we want to get even more granular (and depressing) for a second, according to the Federal Reserve, software positions (on Indeed) have declined by about 33% since 2020. Relative to a peak in openings in mid-2022, openings are down by closer to 70%.
And to make matters even worse, students only continue to major in computer science in increasing numbers. Over the past 5 years, the number of CS majors enrolled in four-year US colleges has increased by over 40%:
Where other white-collar work is concerned, it’s also looking a little grim. The WSJ recently reported that full-time, in-person MBA program applications have soared back to the “highest level in a decade.” For example, top business schools in the US, like Stanford’s Graduate School of Business, Harvard Business School, Booth School of Business at UChicago, Yale School of Management, and Pennsylvania’s Wharton School, all reported 20%+ year-over-year application increases.
The credential creep, unfortunately, makes a lot of sense. What’s that Milton Friedman quote again about inflation? Something along the lines of “too much money chases too few goods.” But instead of money, we have degrees, and instead of goods, we have jobs. I don’t hold it against myself or anyone else who has or wants to get an advanced degree to improve their job prospects. Acquiring more advanced degrees is one tried-and-true way of standing out in this increasingly competitive technical/white-collar job market (as I mentioned earlier, I’ve been there).
But it’s not just Tech
Now, while I’ve honed in on tech & knowledge work for the most part because this is what most directly affects me personally, this credential creep phenomenon doesn’t just affect the tech industry. It might even be worse in other sectors.
In a study called “Dismissed by Degrees,” Harvard Business School researchers found that over 60% of employers rejected candidates who did not have a college degree but were otherwise qualified for the role based on skills and/or experience. And in 2014, Burning Glass Technologies also published research highlighting this credential creep trend across the US labor market. They gathered data from millions of job postings and found that employers increasingly seek college-graduate-level talent for roles that historically have not required degrees. To quantify this effect, they looked at what they referred to as the credential gap: the difference between the percentage of job postings that require a bachelor’s degree and the percentage of people in these jobs with a degree. The more significant the credential gap for a particular occupation, the greater the indication that the job suffers from credential creep. In the table below, you can see how different occupational families stack up:
Some fare better/worse than others. And amongst certain occupational families, like Healthcare Practitioners, there was no evidence of credential creep whatsoever. This discrepancy likely has to do with the fact that jobs with their own standards and/or licensing requirements don’t need to rely on using standard academic credentials like a bachelor’s degree as a proxy for ability.
To be fair, there has been some acknowledgment of this problem. States have begun to remove the bachelor’s degree requirement in their hiring processes, and CNBC recently reported that “1 in 3 companies are ditching college degree requirements for salaried jobs”. But I can tell you that, to an extent, based on personal experience, it’s bullshit. For example, CNBC mentions Tesla as one of the technology companies moving towards “skills-based hiring.” Before I got the internship/co-op offer at Tesla, the recruiter made clear in all of our conversations that I must still be enrolled in school. I even had to get an official letter from UC Davis confirming that. If the company was primarily concerned with skills-based hiring, and I had cleared their bar for “skills” based on the interview rounds I passed, why was it critical I still be in school? If “skills-based hiring” truly is the goal, why not have an internship program that caters to individuals who want to see if they're qualified to do the work without spending vast sums of money on a college degree? Needless to say, every intern I met was enrolled in a college program of some type. Now, credit where credit is due: I met plenty of people at Tesla who didn’t have college degrees, but essentially all of them were technicians (which Tesla constantly, desperately needs more of since there is so much turnover. I met a guy on my first day who said he had worked at Tesla before, went to jail, and now that he was out, he could get a job at Tesla again).
But don’t just take my word for it. Harvard Business School and Burning Light Technologies reported on this credential inflation topic again this year, and this time together! Their takeaway: “Employers are still hiring the same people they were before.” AKA, people with degrees.
So why does any of this matter?
I implied this earlier, but there are major asymmetries in certain job markets; a mismatch between labor supplied and labor demanded. Plummeting openings for software development roles while there is sky-high enrollment in comp sci programs is but one example. This disconnect has a bunch of downstream effects:
supply side: comp sci yuppies (or any other white-collar, information-esque type majors) hit the job market without an offer prior to graduating, and a painful months-long cycle of applying and getting rejected ensues. Eventually, upon realizing that it’s not going to happen, they either decide to get another degree or apply to different roles anyway. Either way, this is costly: they have to spend more time, money, and effort on getting another degree or a job that is not directly associated with what they went to school for.
demand side: companies hiring for white-collar work have to deal with a disproportionate number of job applicants for their tech roles. This slows down hiring, which of course is costly, but then they also need to have some way of filtering the vast pool of applicants. How do they do that? Years of work experience sure, and yep, you guessed it, looking at what kind of degree(s) you have! And so the credential creep cycle continues, so on and so forth.Meanwhile, there are still significant vacancies in other industries and sectors. The inability to fill open vacancies in these industries inhibits the ability of the companies in those other industries to expand and grow, which, again, is costly.
So, what is the main driving factor contributing to this credential creep doom loop? What are some potential solutions for solving this problem? I’ll explore that in the next one.