I wanted to comment on some things I think that Yascha Mounk and his guest miss in their conversation about why Americans are disillusioned with higher education.
1. Most Americans see college degrees as instrumental.
Relatively few Americans pursue college degree through a love of learning. Most see it as a means to an end with the end being a better job and higher lifetime earnings. Even if college still pays economically, disillusionment can grow if the perceived return is overstated or increasingly uncertain.
2. The “college income premium” has always been exaggerated.
There is a small premium that results from having a college education but is not huge and it’s getting smaller. The explanations for the premium usually rely on mean incomes. When you discount the top 5% of income earners, the picture looks considerably different. A disproportionate share of the very highest earners are in professions that require postgraduate credentials. When those outcomes are included in estimates of the “college premium” they blur an important distinction: the returns to a bachelor’s degree alone versus the returns to extended, credential-dependent educational pathways.
| Measure | Typical Headline | Without top 5% |
| Earnings premium | +70–100% | ~+5–25% (marginal/median) |
| Lifetime gain | $500k–$1M+ | ~$70k–$160k median |
| IRR | often unstated | ~8–20% depending on assumptions |
| Payback period | ignored | ~8–10 years |
That top 5% include physicians, lawyers working for big firms, and MBAs working for major banks whose incomes distort the results. Median, mode, and standard deviation are all revealing.
3. What matters is what you study.
Some degrees have a very high return on investment. Others may be zero or negative.
4. The benefits are backloaded.
To obtain a college education you defer earnings for several years. That means that a college education actually has negative benefits at first with the positive benefits appearing later.
How confident are you about what the economic conditions will be in 20 years?
5. Offshoring
Over the last twenty years in particular a significant number of jobs that require a college education have been offshored, many to South Asia. The positions that have proven particularly vulnerable to offshoring have been engineers, accountants, analysts, and backoffice professionals, particularly in customer support and HR. Basically, anything that is codifiable, deliverable digitally, and modular, i.e. it can be divided into pieces.
There are hundreds of millions of people in South Asia who speak English and have college educations. We don’t know how many jobs have been offshored because we don’t keep track of them.
6. AI
Very much the same positions vulnerable to offshoring are increasingly vulnerable to AI.
The bottom line is that Americans are making rational decisions based on the facts on the ground.
On the tangent on AI and job prospects; I thought this post on Volokh Conspiracy was interesting (https://reason.com/volokh/2026/04/28/apparent-surge-in-self-represented-litigation-using-ai/)
I have a feeling the field of law is going to start feeling like what the software field has been feeling for the past 4 months. A sudden surge in capability and productivity leads to a strain in all types of “infrastructure choke points”. And far from getting rid of jobs, it requires a surge of workers to address them.
In law, that would be judges, judges clerks, possibly civil litigators. Could the use of AI in law usher the biggest change in the practice of law since the merger of equity and law?