When Equality Becomes Accounting: The Institutional Harm of DEI

Kenneth L. Marcus’s August 18, 2026 Wall Street Journal article, “Bean Counting, Antisemitism and the EEOC,” identifies a problem much larger than the seemingly technical question of whether employers should continue filing annual demographic reports with the Equal Employment Opportunity Commission. The issue is not really about paperwork. It is about what happens when a civil-rights system created to protect individuals from discrimination is gradually transformed into a system for managing demographic outcomes.

The distinction is fundamental.

Title VII of the Civil Rights Act of 1964 prohibits employers from discriminating against individuals because of race, color, religion, sex, or national origin. Marcus argues that the annual demographic reporting apparatus has increasingly encouraged employers to think differently: not in terms of whether John or Maria or David was treated fairly, but whether the organization has achieved the “right” numerical distribution of racial and sexual categories. The EEOC has now proposed rescinding certain annual race-and-sex reporting requirements, expressly questioning whether collecting this information is necessary to enforce Title VII and raising constitutional concerns about the existing regime.

This is precisely where the modern DEI project went wrong. Equality before the law became equality on a spreadsheet.

There is nothing inherently sinister about demographic information. Aggregate statistics can identify unusual patterns, inform investigations, and help regulators decide where enforcement resources might reasonably be directed. Marcus acknowledges as much. The problem begins when descriptive statistics quietly become normative targets. Once an employer measures the percentage of every racial and sexual category in every department, management inevitably begins asking whether the numbers are “correct.” From there, it is a very short distance to asking managers to change them.

At that moment, nondiscrimination has been replaced by demographic engineering.

The moral inversion is extraordinary. The civil-rights movement fought against the proposition that a person's opportunities should depend upon race. DEI, in many of its institutional forms, resurrected race as an administratively significant characteristic and then attempted to justify the practice because the intentions were different. But discrimination does not become nondiscrimination merely because the preferred and disfavored groups have changed.

A professional making a hiring, promotion, compensation, or admission decision should ask: Who is the best-qualified individual? Who performed? Who exercised judgment? Who demonstrated competence, character and responsibility? The DEI bureaucracy encourages another series of questions: Which group does this person represent? Is that group overrepresented or underrepresented? What will this decision do to our diversity statistics? What will our dashboard look like next quarter?

Those are not merely different questions. They arise from different moral systems.

The first treats the individual as the unit of value. The second treats the individual as an instrument for producing a desired collective outcome.

This is why the argument cannot be dismissed as a dispute between people who “favor diversity” and people who oppose it. Diversity in the ordinary meaning of the word is often enormously valuable. People with different experiences, cultures, intellectual traditions, and perspectives can improve organizations precisely because they are different individuals. What is objectionable is the bureaucratic substitution of visible demographic categories for genuine human diversity. A black engineer and another black engineer are not intellectually interchangeable. Neither are two women, two Jews, two Asians, two gays, two Christians, or two white men. To assume otherwise is not sophistication. It is stereotyping dressed in institutional language.

The consequences become especially apparent in the treatment of Jews.

Marcus points to the experience of Jewish employees at Stanford University, where a controversy arose over a DEI program that reportedly placed Jewish employees within a “whiteness accountability” framework. Stanford's own 2024 report on antisemitism and anti-Israeli bias subsequently discussed the episode and concluded more broadly that rigid DEI distinctions between “oppressed” and “oppressor” identities can reinforce negative stereotypes about Jews and Israelis. The report described concerns that Jewish identity had effectively been compressed into a racial classification associated with privilege, leaving little conceptual room for Jewish employees to discuss antisemitism as Jews.

This exposes one of DEI's deepest intellectual defects. Once humanity is organized along an oppressor-oppressed axis, every group must be assigned a location on the hierarchy. History, individual experience, religious identity, political persecution, immigration, economic circumstances and personal achievement become subordinate to the taxonomy.

Jews create a particular problem for that taxonomy because Jewish history does not fit neatly into it. A people repeatedly expelled, dispossessed, persecuted and murdered can nevertheless be classified as “privileged” when contemporary demographic categories require it. The classification then becomes more important than the person standing in front of you.

That is not inclusion. It is ideological reductionism.

The Columbia University case illustrates the difference between demographic theory and actual civil-rights enforcement. In 2025, Columbia agreed to establish a $21 million fund resolving EEOC charges concerning alleged antisemitic harassment of Jewish employees. The EEOC described it as the largest settlement for victims of antisemitism in the agency's history and its largest publicly announced discrimination or harassment settlement of any kind in nearly two decades.

Notice what makes the Columbia matter relevant. The remedy did not require a theory about proportional representation. It required allegations involving actual people, actual conduct, and actual injury. That is what civil-rights enforcement was designed to address.

The distinction should be obvious, yet modern institutions have repeatedly confused the two.

A demographic disparity is not itself proof of discrimination. If 70 percent of the engineers in a particular firm are Asian, 65 percent of nurses are women, 80 percent of professional basketball players are black, or a disproportionate share of plumbers are men, the numbers tell us something about the composition of those occupations. Standing alone, they tell us virtually nothing about whether a particular individual was unlawfully discriminated against.

To move from disparity to discrimination requires evidence.

That requirement is not an inconvenience. It is the essence of justice.

Once statistical disparity itself becomes morally suspicious, however, organizations acquire powerful incentives to manufacture numerical balance. Hiring managers become demographic portfolio managers. Human-resources departments become compliance centers for racial and sexual allocation. Executives learn that the safest course is not necessarily to treat every applicant without regard to race or sex, but to produce numbers that will satisfy internal constituencies, consultants, boards, regulators and activists.

Eventually nobody needs to order a quota. The institution learns to impose one on itself.

This is what I have called institutional stupidity: not the absence of intelligent people, but the construction of systems in which individually intelligent people are rewarded for producing collectively irrational results. A hiring executive may understand perfectly well that race should be irrelevant to a candidate's competence. Nevertheless, if her bonus, performance evaluation, or promotion depends upon achieving a diversity target, the institutional incentive has already answered the ethical question for her.

DEI thus creates a particularly dangerous form of moral outsourcing. The manager can tell himself that he is not discriminating; he is merely complying with corporate objectives. The university administrator is merely following the framework. The consultant is merely applying the methodology. The board is merely reviewing the metrics. Responsibility disappears into procedure.

This is one reason compliance and ethics are not synonyms.

An organization can comply meticulously with an internal DEI policy and behave unethically. Conversely, an individual can refuse to manipulate hiring decisions by race or sex and thereby violate an internal corporate objective while acting ethically. Professional ethics requires judgment; bureaucracy requires conformity. Confusing the two is one of the great managerial errors of our time.

There is also a paternalistic dimension. Traditional paternalism says, “We know what is best for you.” Modern institutional DEI often says something more audacious: “We know what your identity means, which interests you possess, which historical category you occupy, and which outcomes should be produced on your behalf.”

That is precisely what Rational Paternalism rejects.

A professional may properly use superior knowledge to guide a client, employee or student when the guidance is rationally connected to that individual's interests and when the professional accepts responsibility for the advice. But demographic paternalism does something quite different. It substitutes assumptions about a collective for knowledge about the individual. It does not say, “I know something relevant that you do not know.” It says, “I know something about you because I know your category.”

That is not professional judgment. It is prejudice with an academic vocabulary.

The Objectivist objection is even more fundamental. Human beings are not fractions of racial, sexual, or ethnic collectives. They are individual moral agents. Achievement belongs to individuals. Responsibility belongs to individuals. Rights belong to individuals. Neither guilt nor entitlement can rationally be inherited merely by membership in a demographic category.

This does not require pretending that racism never existed, that discrimination never occurs, or that history is irrelevant. Quite the opposite. Genuine discrimination should be confronted vigorously precisely because it violates an individual's rights. An employer who refuses to hire someone because she is black should be held accountable. An employer who refuses to promote someone because he is Jewish should be held accountable. A university that tolerates actionable harassment because the victim happens to belong to an ideologically disfavored group should be held accountable.

But the remedy for discrimination cannot logically be another system of discrimination.

An important counterargument deserves acknowledgment. Demographic statistics can expose patterns that individual complaints might miss. If a large organization shows an extraordinary disparity combined with suspicious employment practices, the data may provide useful evidence and may properly trigger investigation. Abolishing every form of demographic information would therefore not automatically produce better civil-rights enforcement.

The correct distinction is between evidence and entitlement.

Statistics can be evidence. They should not determine an individual's entitlement to a job, promotion, or opportunity. They can generate questions. They cannot morally answer them. Once the numbers themselves become organizational objectives, the instrument used to detect discrimination becomes an instrument capable of producing it.

That is the real importance of the EEOC debate Marcus describes. The question is not whether government should stop caring about discrimination. It is whether government should return to caring about the person who was discriminated against.

That would not be a retreat from civil rights. It would restore their philosophical foundation.

For decades, corporate America, universities and government agencies built increasingly elaborate DEI infrastructures around the proposition that fairness can be measured by counting bodies according to race and sex. The result has been enormous expenditures, expanding bureaucracies, ideological conformity, racial categorization, resentment and, in some cases, the remarkable spectacle of institutions committed to “inclusion” failing to recognize discrimination against people who did not fit neatly into the approved hierarchy of victimhood.

The lesson should not be that diversity is harmful. The lesson is that collectivism is harmful.

A genuinely diverse society does not require racial accounting. It requires something both simpler and more demanding: equal rights, objective standards, freedom of association, intellectual pluralism, professional judgment and the willingness to evaluate people as individuals.

The EEOC was created to protect Americans against employment discrimination. If its renewed direction is toward investigating actual discrimination rather than administering demographic outcomes, that is not the abandonment of its mission.

It is a return to it.

Reference: Kenneth L. Marcus, “Bean Counting, Antisemitism and the EEOC,” The Wall Street Journal, August 18, 2026.