My colleague Brad Hershbein and I have a new blog post at the Upjohn Institute website on our recent surprising discovery: the percentage return to getting a college degree, in terms of higher earnings, versus getting only a high school degree, is much lower for individuals who grew up in a low-income household, than for individuals who grew up in a non-low income household.
Why is this finding surprising? It is not surprising that one’s family income background affects a person’s earnings. But what is surprising is that this effect is much greater for someone who goes on to complete college versus someone who only gets a high school degree.
Adding to the surprise is that this pattern in the returns to college is not evident for blacks versus whites. One might expect that since blacks, on average, grow up poorer than whites, the percentage return to college versus high school should be smaller for blacks. In a forthcoming working paper by Brad and me, along with our colleague Marta Lachowska, we find that the percentage return to college for nonwhites is no less than that for whites. (This paper is entitled “The Merits of Universal Scholarships: Benefit-Cost Evidence from the Kalamazoo Promise”. It should be available within the next few weeks.)
This new blog post further explores this new finding of lower returns to college for individuals who grew up poor. One discovery in examining this data is that this pattern mostly occurs because of how parental income background is associated with the earnings of college graduates. For individuals who only get a high school degree, “having the right parents” does not have huge effects on the career earnings path. But having the right parents matters a lot more to career earnings for those who go on to get a college degree.
In addition, when we look at career earnings in more detail with age, the extra return to college for those with higher-income parents, versus those with lower-income parents, seems to increase as a person ages and gets on in their career.
One hypothesis that some have suggested to explain these findings is that this pattern might occur if individuals from higher-income backgrounds tend to go on to even higher education, getting MDs or MBAs or other advanced degrees at a higher rate than those from lower-income backgrounds. But this cannot explain our results, because we still see similar differential returns to college if we restrict our sample to individuals who only get a college degree, and no higher degree.
We also briefly explore how these findings vary by race. What we find here is also surprising: this pattern of differential returns to college by family income background is more pronounced for whites than for blacks.
One final surprising aspect of this new discovery is why on earth this hasn’t been discovered before? As a researcher, I am very surprised when an analysis of well-known data sets such as the Panel Survey of Income Dynamics yields dramatic new findings. My best hypothesis is that these patterns become more evident if one observes later career earnings, so perhaps these patterns were not as evident when the PSID had fewer years of coverage. In addition, perhaps no one was looking for this pattern precisely because it is not evident in the many databases that are available to compare the return to college for blacks versus whites.
Brad and I are actively engaged in further exploring these research findings. People have suggested these patterns might be due to: neighborhood/local area/region of family background; other family background characteristics; high school test scores or high school quality; quality of college attended; choices of majors and occupations; choice of neighborhood or area to live in after college. We plan to explore these hypotheses and more over the next several months. Stay tuned!