Noam Scheiber argues that college-educated people, especially younger ones, have moved left on economic issues in part because their economic situations have worsened in various ways. Eric Levitz counters that the real mechanism is that college-educated people have become more female and nonwhite over time and that their cultural liberalism caused them to associate with Democrats, which subsequently led to them adopting Democrat-aligned economic views.

On the way to reaching his conclusion about demographics and culture, Levitz also argues that Scheiber is wrong that the economic situations of college-educated people have worsened.

In this part of the argument, Levitz does a good job of pointing out holes in some of the data Scheiber uses. Most significantly, Scheiber uses a dataset that shows that, among self-reported underemployed college graduates, the share working low-wage jobs has increased. But this same dataset shows that the percentage of self-reported underemployed college graduates fell during this period and that the percent of all college graduates who report being underemployed in a low-wage job is essentially unchanged in the last 30 years.

I don’t really find this dataset all that compelling because “jobs that require a college degree” is a moving target as employers ratchet up degree requirements when the supply of college graduates grows. But it is the data Scheiber used and Levitz is right that it does not really support Scheiber’s point.

However, the main alternative measure offered by Levitz to prove that college graduates are doing well does not really his preferred conclusion either. Levitz offers the following graph to show that college graduates have higher inflation-adjusted wages today than they did 25 years ago:

What this graph actually indicates is that wages, when measured this way, are basically unchanged over the last 25 years despite the fact that GDP per capita increased 42 percent over the same period. As with the underemployment figures, I don’t really like this measure, but it is Levitz’s chosen measure and it actually looks really bad for both college and high school graduates.

Relative Positions

What I think both writers overlook — and what I think a lot of economic discourse overlooks — are the changes in the relative positions of college-educated individuals. People do not just care about whether their inflation-adjusted income or consumption is the same as some comparator from decades ago. They also care about where they are in relation to others in their current day. More concretely, what a college degree supposedly promises is not just a certain level of income, but also a position in the upper-middle class strata of society.

By rapidly increasing educational attainment in this country, we have mechanically pushed down the relative social position that is obtainable through a given educational outcome. The frustrations caused by this spill over into a lot of economic discourse though usually in ways that are not coherently or clearly articulated. “A high school degree used to get you a middle class life,” “I was told that if you get a college degree, you can live the American Dream,” and similar kinds of sentiments all get at the point that the relative position of each educational rung keeps sliding backwards, generating outcomes below what people were led to expect by prior generations.

One way to quantify this is to look at the evolution of earnings among individuals between the ages of 25 and 34 in each educational group. Rather than report those numbers in absolute inflation-adjusted dollar terms, we can report them in relative terms by seeing where those earnings place individuals in the overall earnings distribution.

The graph below takes the median earnings of young people in every educational group and then expresses those earnings in terms of where that placed them in the overall earnings distribution (which includes older workers) in their time period.

So, for instance, young adults with Bachelor’s degrees went from earning at the 64th percent in 1964 to earning at the 55th percentile in 2024. Median young high school grads went from earning at the 49th percentile (almost literally middle class) to earning at the 29th percentile.

The next graph plots the same data but in a slightly less abstract way. Rather than expressing the median earnings of each education group in terms of their percentile in the overall earnings distribution, this one expresses the median earnings of each education group as a percent of the median earnings of the overall distribution.

So what we see here is that, in 1964, young individuals with a Bachelor’s degree were earning 29 percent more than the median earner. By 2024, they were earning only 10 percent more than the median earner. Young individuals with a high school degree went from earning just 3 percent less than the overall median of the society to earning 32 percent less than the overall median of the society.

One natural reaction to these figures is to say that they are, in some objective sense, economically irrelevant. By definition, there are only 100 percentiles. We can’t all be in the top 10, top 20, or top 50 of them. Thus, as greater shares of people graduate high school and college than before, it is mathematically inevitably that the relative outcomes of the individuals in each educational group will decline. But if the absolute outcome at each percentile is going up due to broad income growth, everyone is still better off in terms of absolute income and consumption. Thus, a college graduate sitting at the 50th percentile of the income distribution should not lament that prior generations of college grads were more squarely in the upper middle class. They should be happy that they have more than prior generations of median earners and realize that their proper comparator is not to college graduates from the 1960s or 1990s but to high school graduates from the 1960s or 1990s. After all, that is more accurately who they would have been had they been alive back then.

The problem with this reaction is that it is violently at odds with the way our society generally presents and understands education. We live in a very inegalitarian society, but we tell a story about why that is the case (and why it is fair and just) that is rooted in differential educational attainment. Yet the life advice that comes out of this story doesn’t actually work in aggregate because pushing up educational attainment doesn’t vault everyone to the upper middle class, but instead mechanically pushes down the relative outcome of every educational group. It’s not hard to see how this sort of expectation-betrayal would frustrate people and how it would make people feel like they are on a treadmill that they can’t get ahead of.

To my mind, the most sensible reaction to understanding how all of this works is to recognize that trying to win the inequality game by getting to the top through educational attainment is a bit of a fool’s game and that it makes far more sense to focus on cutting down class differences through socialist and social democratic policies. Don’t focus on getting yourself to the 90th percentile, a feat that gets harder and harder to achieve every year, especially through the mechanism of education. Just bring the 90th percentile closer to you through egalitarian policies.

More generally, the problem with this kind of reaction is that it doesn’t really defeat the point someone like Scheiber is making. Class is a hierarchical concept that describes someone’s relative position in an economic system. We can talk about how lower class Americans are actually richer than 12th century kings or whatever other kinds of absolute comparisons people sometimes make, but saying such things does not make lower class Americans any less lower class nor does it make the 12th century king any less upper class.

College-educated Americans are objectively tumbling down the class hierarchy of America. In the graph above, we see the median young college graduate is earning at the 55th percentile, which means that just under half of them earn even less than that. This is pretty squarely within any definition of the working class and historically this slice of the class hierarchy has been quite a potent force in labor movements and leftist politics, which did not draw exclusively or even mostly from the poorest of the poor.

Now I am not saying this describes all of the psychological mechanisms involved in how people come to political opinions. I don’t pretend to fully understand how people process these kinds of educational and economic developments, consciously or subconsciously. But it does make sense conceptually and, absent some reversal in mass college education, the class position of college-educated people is destined to keep getting worse going forward.