DEI Rollbacks Might Be a Breach of Fiduciary Duty (The Receipts)

According to the “As You Sow” report, a company that eliminates or substantially reduces its management diversity programs in response to political pressure, rather than financial analysis, is making a decision motivated by nonfinancial considerations.

Think about how wild it is to publicly announce you don’t want diversity, equity, and inclusion.

Not to quietly defund it. Not to let it wither in a reorg. To announce it. To put out a press release, to schedule the earnings call, to stand at a podium and tell the world that your company has decided, as a matter of strategy, that it would prefer a narrower pool of talent, a thinner set of perspectives, and a smaller share of the American consumer. Think about the confidence that takes. Think about what a person has to believe about the world, and about who is watching, before they decide that this is the announcement that will make them look competent.

I’m finna make you an evidence based argument that has nothing to do with politics or emotion. Straight data. Facts over feelings.

Here is the argument in one breath, and then I will spend the rest of this piece proving every joint in it. If the evidence links management diversity to financial performance, then dismantling diversity programs for political reasons might be considered a breach of fiduciary obligations. And if you’re claiming you care about fiduciary duty, that matters, because courts and state officials have spent the last four years arguing that putting ideology ahead of financial returns can breach fiduciary duty. That argument was built by the people running the rollback. That argument was built by conservative think tanks, by Republican attorneys general, by activist shareholders who sent threat letters to boardrooms. I am not inventing a new standard. I am picking up the standard they left on the table and turning it around.

Y’all know I was a professor of diversity, equity, and inclusion at a university, as well as having a master’s degree in human relations. I think I might not only be speaking from experience but also expertise.

Let me tell you what that expertise actually taught me, though, because it is not what you think. It did not teach me that the business case for diversity is the strongest case. It taught me that the business case for diversity is the weakest case that white institutions were willing to accept, and that the moment it stopped being convenient, they would abandon it and pretend they never signed. Two things can be true. The evidence can support the business case, and the business case can still be a trap. I am going to hold both of those at the same time for the next several thousand words, and I need you to hold them with me.

Now let me give you the history, because you cannot understand why corporate America folded in 2025 unless you understand what corporate America agreed to in 1990, and you cannot understand 1990 unless you go back to 1941.

In 1941, A. Philip Randolph threatened to march one hundred thousand Black workers on Washington to protest exclusion from defense industry jobs. Franklin Roosevelt, who needed those factories running and did not need those cameras, issued Executive Order 8802, banning discrimination in the defense industry and creating the Fair Employment Practice Committee. Understand the mechanism. Randolph did not win 8802 by demonstrating that integrated factories were more productive. Randolph won 8802 by making segregation more expensive than integration at a moment when the state could not afford the bill. That is the origin story. The first federal fair employment policy in American history was a hostage negotiation.

Truman desegregated the armed forces with Executive Order 9981 in 1948. Kennedy issued Executive Order 10925 in 1961, and that order is where the phrase “affirmative action” enters federal law. Johnson issued Executive Order 11246 in September 1965, which required federal contractors to refrain from discrimination and to take affirmative steps toward equal opportunity, and in 1967 Executive Order 11375 added sex to the covered categories. Then Nixon, of all people, expanded the Philadelphia Plan in 1969, imposing numerical goals and timetables on federally funded construction. People love to tell that story as Nixon the closet integrationist. Read the memos. The Philadelphia Plan drove a wedge between the building trades unions and the civil rights movement, two constituencies that were both aligned against him, and it cost him almost nothing. Crazy how the arithmetic works out.

And in 1966, quietly, with no march and no press conference, the EEOC created a form. The Employer Information Report, Component 1. Every private employer with a hundred or more workers had to tell the federal government how many people they employed by race, by ethnicity, by sex, across ten job categories. That form is the ledger. Everything I am about to argue depends on that form existing, and I need you to remember that, because it is going to come back at the end of this essay and it is going to come back with teeth.

Then came the turn.

In 1978, in Regents of the University of California v. Bakke, Justice Powell wrote the controlling opinion, and Powell did something that reshaped the next fifty years. Powell rejected the remedial rationale. He would not accept “we are correcting societal discrimination” as a compelling state interest. What he accepted instead was diversity, framed as an educational benefit that accrues to the institution and to everyone in it. Read that again. The Supreme Court told America that repairing harm to Black people was not a good enough reason, but enriching the learning environment for everybody was. The justice frame died in 1978 and the benefit frame was born on the same day, in the same paragraph, by the same hand.

Corporate America learned that lesson faster than the academy did.

In 1987, the Hudson Institute published Workforce 2000 for the Department of Labor, projecting that white men would make up a shrinking share of net new entrants to the labor force. That report got misread across every boardroom in America as a demographic emergency. Then in March 1990, R. Roosevelt Thomas Jr. published an essay in the Harvard Business Review titled “From Affirmative Action to Affirming Diversity,” and that essay is the hinge. Thomas argued that affirmative action was a temporary intervention that had run its course and that what companies needed instead was to manage diversity as a competitive asset. Within a decade, the compliance officer became the diversity officer, the legal obligation became the strategic initiative, and the language of debt became the language of dividend.

Ellen Berrey documented this in The Enigma of Diversity. Lauren Edelman documented the mechanism in Working Law, showing how organizations construct symbolic structures that courts then accept as evidence of compliance, so that having a diversity policy becomes legally equivalent to not discriminating. Frank Dobbin and Alexandra Kalev spent thirty years showing that most of what companies actually built, the mandatory training, the grievance procedures, the diversity tests, did not increase management diversity at all, and sometimes decreased it.

So by 2020, when George Floyd was murdered and fifty billion dollars in corporate racial equity pledges hit the newswire in about ninety days, the whole apparatus was already built on a foundation of contingent value. Not owed. Not repair. Valuable. Useful. Good for the brand.

Derrick Bell told us exactly how that story ends. Interest convergence. Bell’s argument, laid out in the Harvard Law Review in 1980, is that Black people secure gains precisely and only when those gains converge with the perceived interests of white elites, and that when the convergence breaks, the gains are withdrawn without apology. Bell was writing about Brown. He could have been writing about the 2020 pledges and the 2025 press releases with the same pen.

Just admit what happened. The business case was never a floor. The business case was a lease. And in January 2025, the landlord showed up.

1. The fiduciary argument was built by the rollback crowd, and it points back at them. Conservative shareholder activists and state officials spent years arguing that boards which put ideology ahead of returns breach their duty. The federal workforce data associates management diversity with stronger performance and associates its absence with weakened performance and retention. Apply their standard honestly and the party exposed to a duty of care problem is the board that cut a program to please an administration without documenting any analysis.

2. Association is not causation, and the argument survives that concession. The correct claim is materiality, not proof. Green and Hand’s 2024 replication failure against McKinsey is real and should be cited by anyone who is serious. It does not touch the mandatory federal compliance filings, which are standardized, audited, longitudinal, and were never assembled by anyone selling diversity consulting.

3. The owners already voted, and they voted no. More than 98 percent at Costco, 97.3 percent at Apple, roughly 2 percent support at Goldman Sachs, under 1 percent at Levi Strauss. Boards that rolled back DEI overrode the near unanimous expressed judgment of their own shareholders. Target absorbed a demand shock its CEO acknowledged on an earnings call, slid from roughly $145 to $93 a share, posted eleven straight quarters of flat or falling comparable sales, and replaced its chief executive.

4. The business case was a lease, not a floor, and it was signed in 1978. Powell’s Bakke opinion rejected the remedial rationale and accepted only the benefit rationale. Thomas’s 1990 Harvard Business Review essay carried that logic into corporate practice. Derrick Bell’s interest convergence predicted the ending. When value stops converging with white institutional interest, the gain is withdrawn without apology.

5. The endgame is not the program, it is the instrument. On July 21, 2026, the EEOC voted 2 to 1 to propose rescinding the EEO1 through EEO6 reports, covering private employers, unions, state and local government, and public schools. The EEO1 is one of the only mandatory federal instruments that counts Black women as Black women, cross tabulated by job category. Killing it restores the DeGraffenreid condition by administrative procedure. The public hearing is August 11, 2026, and the comment period is open right now.

They just voted to delete the only federal form that counts Black women as Black women, and the entire coverage of that vote fit into a handful of trade publications written for compliance officers. That is the void: the receipts exist, the rulemaking is public, the comment period is open right now, and almost nobody is translating any of it for the people it will actually cost.

I’m fighting to fill a critical void left by the retreat of public education media. I document and teach the histories, legal frameworks, and cultural knowledge that are being systematically erased or distorted. With no corporate backing or wealthy sponsors, this work depends entirely on readers like you. As a Black educator and researcher my work depends entirely on a community of readers, not corporate sponsors. If everyone reading this became a paid subscriber, we could build a full-time digital sanctuary: a new, independent source of PBS-depth reporting and curriculum, centered on Black expertise. But right now, less than 1% of my followers are paid subscribers.

Link to As You Sow Article Here.. Click here

https://static1.squarespace.com/static/59a706d4f5e2319b70240ef9/t/6a615681aba7f2405edeeec4/1784764033335/AsYouSow2026_Diversity+Benefit_Final.pdf

History of affirmative action and the corporate diversity turn

• Katznelson, Ira. When Affirmative Action Was White. W.W. Norton, 2005.

• Anderson, Terry H. The Pursuit of Fairness: A History of Affirmative Action. Oxford University Press, 2004.

• Berrey, Ellen. The Enigma of Diversity: The Language of Race and the Limits of Racial Justice. University of Chicago Press, 2015.

• Edelman, Lauren B. Working Law: Courts, Corporations, and Symbolic Civil Rights. University of Chicago Press, 2016.

• Dobbin, Frank, and Alexandra Kalev. Getting to Diversity: What Works and What Doesn’t. Harvard University Press, 2022.

• Dobbin, Frank, and Alexandra Kalev. “Why Diversity Programs Fail.” Harvard Business Review, July/August 2016.

• Thomas, R. Roosevelt, Jr. “From Affirmative Action to Affirming Diversity.” Harvard Business Review, March/April 1990.

• Johnston, William B., and Arnold E. Packer. Workforce 2000. Hudson Institute for the U.S. Department of Labor, 1987.

• Skrentny, John David. The Ironies of Affirmative Action. University of Chicago Press, 1996.

The empirical debate on diversity and firm performance

• Herring, Cedric. “Does Diversity Pay? Race, Gender, and the Business Case for Diversity.” American Sociological Review 74, no. 2 (2009): 208.

• Green, Jeremiah, and John R. M. Hand. “McKinsey’s Diversity Matters/Delivers/Wins Results Revisited.” Econ Journal Watch 21, no. 1 (2024).

• Bourveau, Thomas, Rachel Flam, and Anh Le. “Behind the Curtain of Workforce Diversity: Evidence from EEO-1 Reports.” Management Science 72, no. 2 (2026): 1699.

• As You Sow and Whistle Stop Capital. Workplace Diversity and Financial Performance: An Analysis of Equal Employment Opportunity (EEO-1) Data. November 2022.

• Columbia Business School. “Beyond the EEO Curtain: Strategic Data Sharing.” Research brief, 2023.

• Georgeac, Oriane A. M., and Aneeta Rattan. “The Business Case for Diversity Backfires.” Journal of Personality and Social Psychology 124, no. 1 (2023): 69.

Education

• Gershenson, Seth, Cassandra M. D. Hart, Joshua Hyman, Constance A. Lindsay, and Nicholas W. Papageorge. “The Long-Run Impacts of Same-Race Teachers.” American Economic Journal: Economic Policy, 2022.

• Fultz, Michael. “The Displacement of Black Educators Post-Brown.” History of Education Quarterly 44, no. 1 (2004).

• Tillman, Linda C. “(Un)Intended Consequences? The Impact of the Brown v. Board of Education Decision on the Employment Status of Black Educators.” Education and Urban Society 36, no. 3 (2004).