No, White Women Are Not The Biggest Beneficiaries of Affirmative Action

The idea was always baseless and it will remain baseless because it's not true

Most people are opposed to affirmative action on the basis of race. Since most people are opposed to racism, and affirmative action by race is a form of racial discrimination, that should probably be taken for granted. Strangely, a common refrain to White people in particular being opposed to affirmative action is the claim that their opposition is hypocritical because White women have been the most benefitted by affirmative action historically.

You can see this claim made in many outlets. Here are some examples:

USA Today: “White women benefit most from affirmative action. So why do they oppose it?”

Time: “Affirmative Action Has Helped White Women More Than Anyone”

Teen Vogue: “Affirmative Action Benefits White Women Most”

Vox: “White women benefit most from affirmative action—and are among its fiercest opponents.”

The African American Policy Forum: “According to the United States Labor Department, the primary beneficiaries of affirmative action are white women.”

Or, a few days ago, live on CNN (the claim appears 48 seconds in):

---

Says Who?

If you’re familiar with how affirmative action has worked and how it currently works, you probably think these claims don’t make much sense. But, since lots of people are forwarding the claim publicly, we should take the accusations seriously and check their sources. Let’s start with Vox.

Vox began by citing a 2006 paper by Kimberlé Crenshaw. In her article, Crenshaw wrote that “the primary beneficiaries of affirmative action have been Euro-American women” and she provided no citations to support her claim. Excellent start.

The next citation was a 1995 California Senate Government Organization Committee report that found there were 57,250 White women holding managerial jobs, compared to 10,500 African American women, 19,000 Hispanic women, and 24,600 Asian women. That’s all the evidence in this report; there’s no indication that anyone received these jobs due to affirmative action, and these numbers don’t even seem suspicious.

Per the Census, California was 57.2% White in 1990 and 46.7% White in 2000, whereas the figure 57,250 figure there is 57250 / (57250 + 10500 + 19000 + 24600) = 51.41% of the female managerial jobs, or basically exactly the representation you would expect from their population numbers! At the same time, Blacks were 7.4% of the state in 1990 and 6.4% in 2000, but they were 9.43% of the female managerial job holders in 1995. Hispanics were 25.8% of the state in 1990 and 32.4% in 2000, while constituting 17.06% of the female managerial job holders, and Asians (plus Pacific Islanders) were 9.6% in 1990 and 11.1% in 2000, while holding 22.09% of the female managerial positions. In other words, Whites were proportionally represented, Hispanics were underrepresented, and Blacks and Asians were overrepresented. This speaks against White women being the primary beneficiaries of affirmative action since they do tend to be more qualified than Blacks and Hispanics but they’re still only proportionally represented, whereas it could speak for affirmative action benefitting Blacks, and maybe Asians.

Regardless of who’s overrepresented in what ways, this evidence so far is indeterminate since we don’t have access to estimates from a counterfactual world without affirmative action. There are definitely ways to estimate the impacts of affirmative action—like estimating the dynamic effects of federal contractor status on employment shares, as we’ll see—but they just weren’t cited.

The next citation from Vox is a 2015 report out of New York that showed White women were disproportionately represented among business owners. At best, that’s all it showed. There’s no evidence that disproportionality was due to affirmative action. If anything, we should expect there to be totally benign disproportionality because Blacks and Hispanics average lower rates of education and wealth, and many of New York’s Asians are poor immigrants, so the native Whites should be disproportionately entrepreneurial given the baseline includes groups that tend to not do much for benign reasons (Blacks, Hispanics) or groups penalized by virtue of things that will disappear with time, like being English second language learners (Asians).¹

Vox’s next citation is a Department of Labor report that I don’t think they actually read. I suspect this for two reasons. Firstly, they linked all the other citations but this one went unlinked. Secondly, they wrote that the report “found that 6 million women overall had advances at their job that would not have been possible without affirmative action” and this was somehow considered relevant when there was no evidence of anything favoring White women. The report simply has nothing to do with White women being recipients of affirmative action. Vox cited the report—but did not link it—because other papers and their citations mentioned it and incorrect interpretations of its conclusions. Similarly, Vox’s final citation was a report from the American Association of University Women showing that women were receiving the majority of bachelor’s, master’s, and doctoral degrees as of 2009—that’s right, nothing about White women.

Vox seemed to get the issue with these last two citations, because they wrote “To be clear, these numbers include women of all races; however, breaking down affirmative action beneficiaries by race and gender seems to be rare in reported data.” OK then! The citations are thus irrelevant. If the idea is to show that there’s sex-based affirmative action, then the citations don’t even do that, because they don’t provide anything like a counterfactual estimate.

To be sure, there are some papers that estimate the impacts of affirmative action on employment demographics, but they are generally ambiguous. For example, Kurtulus used data from over 100,000 private sector firms that obtained and did not obtain federal contracts and thus either did or did not fall under an affirmative action mandate, and found… well, just look:²

But maybe USA Today, Time, Teen Vogue, and the African American Policy Forum cited work that Vox missed.

Teen Vogue starts off by citing the same information-free article by Kimberlé Crenshaw and then goes on to talk about how women have made advances in educational attainment and employment, and that there are a variety of cross-sectional differences by race. They cited exactly no counterfactual estimates.

Time cited more citation-free claims that White women have disproportionately benefitted from affirmative action, irrelevant studies on callback rates by race and resources on cross-sectional rather than counterfactual differences in earnings and wealth by race. They also cited a resource they described as claiming that “After IBM established its own affirmative-action program, the numbers of women in management positions more than tripled in less than 10 years. Data from subsequent years show that the number of executives in color at IBM also grew, but not nearly at the same rate.” The linked resource included one paragraph on IBM that didn’t support those claims. Finally, Time cited a resource that claimed federal contractor status raised female employment 15.2% versus 2.2% at companies that weren’t federal contractors, but this is irrelevant and, as my citation above shows, a causal analysis provides limited support for their causal-inference-inappropriately-ripped-from-cross-sectional-data.

USA Today said White women “expanded their share of senior leadership jobs at twice the rate of women of color” between 2020 and 2022, but provided no counterfactual.

The African American Policy Forum alluded to some report from the United States Labor Department that showed White women were the primary beneficiaries of affirmative action, and they described it how a few of these other outlets did, so we can be reasonably certain this is yet another example of losing the game of telephone and misreporting about a report that doesn’t say what people want it to.

---

Journalistic Telephone

Tim Wise claims White women were the majority beneficiaries of affirmative action based on a book by Ellis Cose, who misinterpreted the same 1995 Department of Labor report that Vox heard about from someone—the report they didn’t read before misreporting on it. Robin DiAngelo and Wendy Leo Moore fell prey to the same dynamic: they heard someone say something that wasn’t supported and believed it, and Politico cited them. Al Jazeera published the same style of misunderstanding, since they cited the Teen Vogue article that I’ve shown contained no relevant information to the claim that White women have benefitted most from affirmative action.

A lot of reporting is like this. Journalists hear something from someone, trust them, and report it. The initial ‘someone’ had low enough epistemic standards to contort an observation like ‘White women are overrepresented in X’ so that it would be followed by the unjustified line ‘thus showing they received affirmative action’, and though the first claim might be true, the latter claim is never justified. Through reporting, it simply becomes common knowledge even though it never becomes true knowledge, all because some journalist said it.

After reading the claim said by one journalist, some other journalist cites that journalist, thinking they’re trustworthy, and the claim now has a life of its own. It’s simply out there, and it may no longer require a citation to claim, because everyone knows that. Through a corrupt game of telephone, journalists (and academics, to be fair!) give birth to popular beliefs that are totally unjustified.

Affirmative action benefits its targets—in academia, in the labor force, in grant-giving, and more—and as we know because we can read the court cases and see the job listings, those targets are not and have not primarily been, White women.

1. And maybe personality differences, too!

2. Here is the result for males:

This link leads to other citations on affirmative action among employers.

After writing this article, I was also linked to this, which previously dealt with this topic.