On one level, the affirmative-action debate is a moral one, dealing with the appropriateness of discriminating by race to benefit underrepresented minority groups. On another, it’s a practical one, regarding the thorny empirical question of how preferences improve or worsen life for those affected.
A new academic working paper, published through the National Bureau of Economic Research, provides a long-run look at how state affirmative-action bans affect outcomes such as employment and educational attainment among various racial and ethnic groups. It’s an interesting preview of the future, given that the Supreme Court effectively barred the practice nationwide, including at private schools that receive federal funds, last year.
The upshot is that while racial preferences can make a difference, it’s less of a difference than you may expect.
The researchers looked at census data on Americans between ages 25 and 51 from 2000 to 2021. The idea is that some of these individuals hit their college years before, and some after, affirmative-action bans went into effect for public colleges in California (1998), Texas (1997), Florida (2001), and Washington (1999).
The authors use two control groups. They compare trends for blacks and Hispanics born in these states with those of whites from the same states (essentially measuring whether racial gaps widened or closed). They also compare trends in “ban” states with trends in states that didn’t change their affirmative-action laws. They break out the findings by race, sex, state, outcome being studied, and control group used.
The results are something of a mishmash: they show lots of statistically insignificant results, some harms to minority groups’ outcomes, and even some improvements. But a few patterns stand out.
One is that ending affirmative action has “at most small effects on the educational attainment, earnings, and employment of men of any race/ethnicity,” as the authors themselves put it. It may even improve job-market outcomes for black men.
Another is that Hispanic women appear to be the clearest beneficiaries of affirmative action, and thus bear the heaviest burden from its removal: they saw their college attainment decline by as much as 4 percentage points coupled with declines in employment and earnings when the practice is banned. This is surprising, given that preferences tend to be stronger for blacks than for Hispanics when measured using college admissions data.
Still another is that the state-by-state results vary, suggesting that the underlying dynamics play out differently in different places. Presumably, there’s variation in how aggressively state colleges were using affirmative action before the bans, and in whether those colleges complied with, evaded, or ignored the bans.
Why is the picture so complicated? Here are some theories.
For starters, only a small share of colleges are selective enough to use affirmative action to begin with. Students not admitted to those schools often end up getting degrees at less selective colleges, so we would not necessarily expect large numbers of people to end up with less education. When explicit racial preferences are banned, they are often replaced by facially race-neutral efforts to improve diversity, minimizing the effects of a policy change even at selective schools. And preferences can sometimes create “mismatch” effects, wherein their beneficiaries struggle among peers with stronger academic qualifications, and thus wind up not benefiting from preferential admissions—potentially a reason that African-Americans, despite receiving the largest preferences, don’t seem to reap the most benefits.
It would be hard to characterize this study as a clear “win” for either side of the racial-preferences debate. It doesn’t show that the sky falls when affirmative action ends. It also doesn’t show that affirmative action harms its intended beneficiaries on balance, as more aggressive proponents of mismatch theory have sometimes argued, or even that it measurably harms whites.
What it does show is that, while affirmative action can be a truly big deal at elite colleges—with preferences sometimes equivalent to hundreds of SAT points—its society-wide effects are often far more muted.
A final word: the Supreme Court’s ruling last year in Students for Fair Admissions v. Harvard will have a broader impact than the state laws studied here, and the effects could differ. But whether they’re worried or hopeful, observers should keep their expectations in check.
Photo: Andy Sacks / The Image Bank via Getty Images