Down With the Meritocracy

Affirmative action is over, all but ensuring that America’s elite colleges will become more white, insular, and privileged. These elite colleges are gateways to power in our modern government and economy, and this represents a massive loss for so many students that might not otherwise have access to that social capital. It is also a loss for the schools themselves — losing the ability to take race into account during college admissions will deprive administrators of an important tool for ensuring at least some level of diversity. Simply substituting parental income or geographic diversity is unlikely to make up for that loss — race, income, and geography intersect in complex ways and aren’t easily separable in America’s veritable caste system.

Given that these schools feed into many positions of power in government, business, and academics, and given that these schools are about to become even more insular with a higher proportion of the class coming from the upper class (and that number was already quite high; 2/3 Harvard students came from the richest 20% of families in 2017), I see two possible pathways forward.

1) Stop placing such a high premium on so few schools

A close friend of mine is completing his PHD in material sciences at the University of Illinois this summer before taking a job at NASA; he’s going to be a literal rocket scientist. He will be joining colleagues with degrees from Virginia Tech, Penn State, and Ohio. Yet the high level bureaucrats deciding their budget in Washington will largely come from Ivey League schools. Why is that? Is it important to have the “best of the best” in administration, but totally fine to use “the rest” to design and build the next rocket going to the moon? Or is that elite pedigree an imperfect measure actual competence in this world?

Nassim Taleb once observed that the tangibility of a field of study is inversely correlated to how much institution prestige matters in that field. Chemistry? State, technical, and regional research universities can all put you on a path to an influential career in the field. Economics? Better get in to one of the small band of prestige schools. Law, even more so; when was the last time we appointed someone with a state school law degree to the Supreme Court? 1962. And in government, while the President has a degree from the University Delaware, 41% of his mid to senior staff have degrees from Ivy League institutions. (Source) Across the liberal arts, a degree from a prestigious institution increases future earnings and presumably influence; for engineering and the sciences, school prestige does not matter at all to future earnings. (Source)

So are these humanities disciplines inherently more difficult, where only a narrow band of people with the right intelligence or training can do the job? Is economics or law harder to learn than rocket science? Or are we placing a bit too much importance on prestige in fields where results are more intangible? Are there really no University of Colorado law school graduates who would make excellent Supreme Court justices? Grinnell college political science majors who would make excellent presidential advisors?

As a narrow band of prestige schools, already insular and exclusive, begin to admit an even narrower band of society, perhaps we should stop placing so much importance on entrance to those schools.

2) Open up the admissions process at these schools

The dirty secret of Harvard admissions is that far more students would excel there than can be admitted. Tens of thousands of bright, motivated high school students across the country embody the ideals of scholarship, service, and innovation we want to cultivate, yet less than 2,000 can be admitted to Harvard undergrad every single year. So let’s make that less secret and the final selection process more transparent. Create a lottery that embodies the true racial, cultural, and ideological diversity of America.

A) Use a lottery

Currently, administrators winnow down that larger list of qualified applicants by ranking them on test scores, gpa, extra curricular, and interviews, and also factor race and geography into their rankings. Such rankings can be subjective or arbitrary (Harvard consistently gave Asian Americans lower personality scores), but they’re also predicated on the notion that a 17 year old with a 1440 SAT score (97th percentile) will be a worse student than one with a 1480 (99th percentile). Once minimum academic qualifications have been met, additional rankings like a few extra points on the SAT or a few more extra curriculars are just going to bias admissions towards kids with more time and support.

Instead, these schools should acknowledge that admissions can be a bit arbitrary and turn it into a formal lottery. Set a bar that still demands academic excellence, but is low enough that a meaningful number of students from every subset of society will qualify for it. This will draw bright students from every walk of life — by race, class, income, culture, and geography — with a clear, achievable standard. And then select, from all those tens of thousands of bright and hard working students, at random.

B) Weight that lottery by census block group clusters

Randomly, but not completely blindly. Simple random sampling in a cutoff sample would disproportionately select students along the same social lines already contributing to lack of diversity at elite schools. Instead, schools could use stratified random sampling by clusters of U.S. Census block groups to ensure a diverse student body representing every type of community and culture in the U.S.

I said at the top that substituting income and geography for race was imperfect, but sampling instead of ranking and using geographic clusters adds some nuance to that. America is a deeply segregated country: we had years of both explicit and implicit segregation along racial lines, and 21st century economics and self sorting have created geographic separation along economic, political, and social lines as well. This separation is often hyper localized (as anyone walking around an American city can notice the neighborhood changing completely in just a block or two), and the sampling should reflect that. Census block groups often cover only one or two city blocks, and are preferable to zip code, city, or state level weighting (The University of Michigan tried zip codes after affirmative action was abolished in the state and ended up accepting mostly kids from the college professor neighborhoods in predominantly Black zip codes).

College administrators can no longer look directly at race on college applications, and should be careful to avoid even proxies for race that could come under legal scrutiny in the decades ahead. However, creating a cluster analysis of census block groups based on household wealth, adult educational attainment, social mobility, employment, urbanity, and many other factors is perfectly defensible in creating a Freshman class that is representative of America. If, because America is deeply segregated on those measures by race, immigration status, and class as well and that happens to mirror the racial make-up of the country, that is a welcome but not legally challengeable side-effect.

Conduct a cluster analysis of all census blocks in the United States, and assign each census block to one of those clusters. Then, give each cluster a weight in proportion to their share of the U.S. population, and do any sampling for admittance to selective universities either by a) using those weights or b) using a sample of students stratified by those clusters.

This is an imperfect solution, but it will open up the hallways to power for a much more representative cohort of students, while giving elite a socially responsible pathway to diversity outside of affirmative action.

We admit a narrow segment of society to a small number of schools

Elite colleges were already insular, largely set up to reward the children of today’s upper classes with their own pathways to power in business, politics, and law. With the end of affirmative action, that insularity will only increase, but there’s no particularly good reason for it.

First, we place too much emphasis on institutional prestige in the humanities where results are harder to quantify. Looking at the sciences though, we know there are brilliant, qualified students graduating from all sorts of institutions. We should figure out ways to open up pathways for those students in business, government, and law and stop placing so much emphasis on institution name.

Second, we should open up prestige institution admissions and make it explicit that more students are qualified than can be admitted. Call it what it is, a lottery (whether by birth or by true randomness), and make it transparent and clear. A certain lucky percentage will get in, while the other bright and motivated high schools will still get fantastic educations at less prestigious institutions — educations that should not preclude them from prestige positions in society.