The Old Boys' Club vs Upward Mobility
The old boys’ network dominated high-status jobs in the 20th century. Is it getting weaker?
When I saw that Michelman, Price, and Zimmerman’s recent NBER paper “Old Boys' Clubs and Upward Mobility Among the Educational Elite” included both lots of data about social mobility and a chance to bash Harvard, I knew I had to write about it.
The researchers examine data from the 1930s on membership in Harvard’s exclusive “final clubs” and later economic success. They summarize their findings as follows:
First, students from high-status backgrounds are more likely to join exclusive campus clubs than low-status students, but less likely to succeed in the classroom. Second, the labor market premium for club membership is much larger than the premium for academic success. Third, exposure to high-status college peers pushes high-status students towards high-status paths in their social and professional lives, but does not affect students from less privileged backgrounds, and thus reinforces rather than reduces inequality.
For proponents of upward mobility, this is disheartening, but perhaps not too surprising. The real showstopper is the effect size:
Members of selective final clubs in the lowest academic rank group earn 26% more than non-members in the top two academic ranks, and are 3.2 times more likely to have topcoded earnings.
In other words, membership in a final club is way, way more important than doing well in class. You might expect that the effect seen here is mostly the result of a correlation between family background and final club membership, but the authors put this firmly to rest:
Brothers who are members of selective final clubs earn 44% more than brothers who are not, and are 2.6 times as likely to report topcoded incomes. Specifications comparing final club members to “near-missers” who join the social organizations that feed into final clubs but are not selected for final club membership yield similar conclusions.
This does admittedly still leave open the possibility that final clubs are really good at selecting soon-to-be successful people, but the huge effect sizes seen here make it hard to believe that this is the whole story (also, there’s the fact that final club members are demonstrably less academically competent than their peers). All told, this paper is very compelling evidence that at the Harvard of the 1930s, membership in the old boys' network was much more important than academic success in determining future income. Combined with the fact that final clubs are disproportionately white, male, and upper-class (21% of private feeder students join selective final clubs, compared to 2% of other students), this is a dismal look for social mobility.
Although the paper’s data is limited to Harvard, it seems likely to me that a similar phenomenon occurred at other elite schools; Yale’s secret societies and Princeton’s eating clubs are well known for connecting their aristocratic membership to upper-crust opportunities. In fact, both candidates in the 2004 presidential election were alumni of Yale’s Skull and Bones.
It seems clear that in the early twentieth century, membership in the old boys' network dominated all other factors in determining future socioeconomic status. The obvious question: is this still true today?
The Crimson seems to think so. Harvard’s student newspaper beat me to the punch on bashing Harvard, publishing an editorial on the NBER article that is essentially an extended mea culpa about the institution’s elitism and failure to propel less-privileged students to post-graduation success. The piece concludes:
Still, Harvard markets itself as an instrument of social mobility: a ladder that propels students to higher status post-college lives. To move forward productively, we must begin to interrogate such falsities; to be honest about how stratifications of wealth and class affect students’ capacity to wield the Harvard experience as a tool for social ascension; and to ultimately explore the modes through which we can level the playing field at Harvard.
I am very sympathetic to this argument; I felt uncomfortable and out of place just reading the linked article about the cocktail parties and formal dinners that make up the Porcellian’s punch process. There’s no denying that students from wealthy families and august private high schools have a significant leg up in the world. That being said, I like to think that progress has been made since the 1930s. In a previous post, I claimed that a dearth of role models and information was now at least as important as a lack of connections in keeping ambitious, less-advantaged students out of top jobs. Is that true?
The NBER paper notes that differences in career choice, academic performance, and campus participation persist between private and public school Harvard students through much of the 20th century. Without modern data on final club membership or income differences between feeder and non-feeder students, the authors make no comment on how the premium of final club participation has changed in a modern context. I’m not certain either, but I think there are some good reasons to believe that the old boys' network is less important than it once was.
For one thing, the well-known Dale and Kreuger paper found that, conditional on acceptance, attending a more selective university in 1973 did not raise the income of wealthy students later in life. For students with families in the bottom decile of income, however, the study did find significant positive effects on future earnings. As I wrote about previously, I think Dale and Kreuger may undersell the value of elite education in general. That being said, it's pretty hard to square this data with a narrative where club membership at a prestigious university raises the earning potential of wealthy students by 50% or more.
The 2017 NBER paper “Mobility Report Cards: The Role of Colleges in Intergenerational Mobility” gives some more recent data. The study finds that the Ivy League does a poor job of admitting lower-income students: 3% of Harvard undergraduates in the 1980-82 birth cohorts came from the lowest income quintile of families, compared with more than 70% from the top quintile. Once admitted, however, disadvantaged students seem to have access to many of the same resources as everyone else; at “Ivy Plus” universities, 13% of lowest-quintile students reach the top 1% of income—a stratospheric increase in upper-tail social mobility that far exceeds that found anywhere else. When one recalls that in the 1930s only 2% of public school students even managed to join one of Harvard’s final clubs (let alone ascend to the lofty ranks of the 1%), it starts to seem very likely that things have changed at elite universities.
My own experience as a public school student participating in recruiting at both tech and consulting firms lines up with this observed trend. I got the sense that many of the tech companies I interviewed with weren’t overly concerned with what university I was at, let alone what fraternal organizations I was part of. At McKinsey, I think it’s likely the Yale name helped me land an interview, but once I was inside the process I felt like I was given a fair shake—as far as I can tell, they were truly interested in assessing my ability, not who I knew.1 I actually decided to apply to the firm because a Midwestern public school friend told me that they had been very fair to him. I admittedly don’t have much personal experience with finance recruiting, but I did recently see a friend-of-a-friend try and fail to leverage close personal connections into a spot at a leading investment bank.
Clearly, top firms are not a perfect meritocracy. Still, it seems like the old boys' network has been declining in strength. What might be causing this trend? I don’t have very much data here, so I am going to speculate wildly.
The increasing importance of math and other hard technical skills. This is my number one guess. Since the 1930s, measurable technical skills have been increasing in economic importance. The original NBER paper cites a 2012 study by Philippon and Reshef that finds increasing skill intensity in finance, beginning in the late 1970s—just as the gap between feeder and non-feeder Harvard students pursuing finance careers begins to close.
Source: Old Boys' Clubs and Upward Mobility Among the Educational Elite
With the rise of the software industry, schools like Yale and Harvard are sending 10% or more of graduates to work in high-paying, highly technical fields like programming and data science. Increasing returns to technical skill make it hard to rely on the old boys' network to find qualified applicants.
Social and political pressure. I think most readers would agree that elitism and “old boys' networks” are tolerated less than they once were. Easily obtainable hard evidence of this trend doesn’t come to mind, but as an experiment I looked at the Google word usage chart for “inequality”, “nepotism”, and “old boys' club” vs the more neutral “finance” (all scaled to fit on the same graph):
Source: Google Books Ngram Viewer
It seems like something is going on here. Investment banks and consulting firms are certainly taking steps to separate themselves from the “old boys' club” image.
More low-income students are going to elite schools. The “Mobility Report Cards” study I cited earlier (in part to shame Harvard for not admitting more low-income students) also shows that over time, the fraction of less-wealthy students at Harvard is trending up. From 2000-2005 to 2006-2011, this proportion of students from families in the bottom 60% of income increased by about 125%, from 16.1% to 20.9%. The corresponding decreasing share of students from very wealthy backgrounds no doubt reduced the power of the old boys' network.
For fans of social mobility like myself, this is good news. If we can identify factors that have been weakening the hold of the old boys' network on top firms, we have a recipe for speeding the process along. Assuming my above speculation about causation is at all correct, it suggests there are at least two dimensions along which policy interventions might increase upwards mobility. One is a focus on math education; STEM skills seem to give lower-income students the ability to compete effectively for top jobs, even against well-networked peers. The second is an emphasis on broader access to elite universities for students from disadvantaged backgrounds. The “Mobility Report Cards” study found that schools ranging from the University of California system to the Ivy League have a huge impact on the capacity of enrollees to move from the bottom quintile of the income distribution to the top.
History has proven that the old boys' network is not infallible; with the right policy, it may be defeated entirely.
I didn’t get the job; this is not a paid advertisement.