How many college grads don’t have “college jobs”?

The New York Fed on underemployment

  • The number of “college jobs” – those mostly held by people with college degrees – is not keeping up with the number of college graduates.
  • Objective data from an impartial source (the Federal Reserve) on underemployment of college grads casts doubt on the policy case for increasing college attendance for the US as a whole. “Policy”: government regulation and funding of education.
  • Though college attendance has been declining since 2011, declines are driven by older students. High school graduates before COVID were becoming more likely to attend college, despite it becoming a riskier investment of time and money.
  • Any policy argument for boosting college enrollment needs to address this underemployment and at the very least assess the chances of future economy-wide growth in demand for college degree holders.

The Fed on Underemployment

The Federal Reserve in New York tracks the proportion of college graduates who hold a “college job”.  Fed researchers identify these jobs by surveying people in a job category to see if over half of their coworkers have a degree, and then estimate how many college grads work in each of the categories.

  • About a third of all college grads are underemployed – they hold positions that do not generally require the degree (“non-college jobs”).
  • The Fed has tracked these statistics for thirty years.  In 1990, the proportion of grads with non-college jobs was 34%.  In 2020 during the COVID recession: 33%.
    • This ratio includes those with only a Bachelor’s degrees as well as those with additional graduate degrees.
  • The proportion of underemployed recent grads (aged 27 and below) stood at 40% at the end of 2020, moderately higher than the ratio for all college grads (33% at that time). Again, this proportion has barely changed since 1990.
  • Of the third who are underemployed:
    • The economic situation of grads without college jobs has deteriorated. The share of these grads with pay above $45k annually has fallen, from 55% in 1990 to 45% in 2020.
    • This indicates that the proportion of college grads whose career situation can be viewed as objectively unsatisfactory has risen.
    • The 45% share holding non-college jobs with pay above $45k earn an income roughly equivalent to college grads’ median pay ($57k, per 2018 USD Census data).
    • The trend growth in underemployed grads making less money is the most troublesome part of this.

Federal Reserve on College graduate underemployment

$45k wage level is real/inflation/adjusted; % of all college grads, including those with graduate & professional degrees.The 19% is not a static population. Some will earn more than $45k in one year and then fall under that benchmark in another — and vice versa.

The cost of college has risen between 1990 and 2020 by any metric. The proportion of graduates who end up working jobs paying about the same as a median high school graduate ($35k in 2018) or a bit more has gone up.  It’s a bad squeeze.  Now, not all of these 19% in the chart above are unhappy; some will be doing fulfilling jobs they enjoy.  But many of them will be questioning the pay-off from college.  And the trend growth is concerning.

The chart above forces the question: if more and more college degree holders are holding “non-college” jobs that don’t pay well, shouldn’t fewer students go to college? In fact, individuals have been making that cost-benefit assessment in the last decade and enrolling less. There are 1.5 million fewer people enrolled in college in 2018 than there were in 2011 (“peak college”) and National Student Clearing House data is displaying continuing declines. If you read the trade press, this is supposed to represent a tragedy. It’s better to perceive the decline as individuals rationally responding to economic information and perceptions.


Significantly, this decline is occurring among older attendees, who are generally more sensitive to economic trends when deciding to enroll. But the likelihood of a high school grad enrolling in college directly or soon after has been increasing, at least until COVID.

Undergrads, age <25, enrolled in college (full- plus part-time) as % of the age group

Undergrads, age <25, enrolled in college (full- plus part-time) as % of the age group

Among the many motivations for this climbing enrollment, one economic driver is surely a poorly performing job pool for workers with just a high school degree. If fewer college grads are able to follow college with a classic post-college trajectory, the picture is even worse for those without college. So on an individual basis, going to college often makes sense. But that does not justify a policy – which is supposed to help the entire nation – encouraging college attendance.

How many people are these trends affecting?

Let’s step back and look at the numbers of people affected by these trends. 60 million people with a Bachelor’s degree were in the workforce in 2018, including those with graduate degrees. 19% of those – 11 million – worked in non-college jobs earnings the same or a little more than a typical high school degree holder.

While those 11 million will have profited from their college program in many ways, the economic results are not among them. College had little impact on their earnings.

Other college grads did land college jobs, which will often have non-monetary benefits, but are not earning much more money than the typical high school grad. That group comes to a total of 7 million. An example would be journalists – most went to college but are not included in the 11 million people above despite steep declines in pay throughout the industry.

To flesh out the logic here, college grads have traits besides their degree credentials that help them outearn high school graduates, as a group. In fact, many high school graduates also possess such skills and traits, with many significantly outearning the $35k average income for their group (2018 US Census Bureau). (A factlet for you: more high school degreeholders earn over $100k than holders of professional degrees. Of course, there are many more people with just the diploma.) So it’s fair to say that these college graduates have not benefitted financially much at all from attending college because their non-educational traits would have allowed them to be relatively successful with just a high school education. They may have more satisfying or enjoyable jobs overall, but the financial returns have been small.

There are also 6 million college grads who hold non-college jobs who earn over $45k. They earn roughly on par with their college classmates in jobs where many of their coworkers at similar levels haven’t finished college.

Putting these three categories together – college grads with non-college jobs or working at college jobs but who aren’t earning significantly more than High School diploma holders – forms a group of 24 million individuals, 40% of all workers with at least a Bachelor’s.

Though these 24 million often gained personally from the college experience, overall they are not receiving an unambiguous, direct monetary benefit. Many would be where they are now financially without college. Their “college ROI” has been minor.

When data and messaging collide

The Fed tracking of underemployment is significant. College attendance is risky and the financial rewards from completing the degree – to say nothing of those dropping out and never finishing – are marginal for literally tens of millions. But governments and higher ed as an industry hammer away with the message that as many people should go to college as possible, despite it not being supported by labor market numbers. So why the discrepancy? Our next installment on college ROI will look at the statistical confusion underlying the “go to college” messaging:

  • The fact that some people should do something does not mean everyone should do it.
  • The fact that on average a group does better is not the sole criteria for success. The range of outcomes is important, too. (Variance!)

What’s a better framework? Two words: “supply” and “demand”.

Fed and US Census data sets are blended in this last sections, so numbers are rounded to avoid false precision.

Read this post and others at our CTAS Higher Ed Business blog on Substack.