In March 2025, University of Michigan President Santa Ono announced the closure of the central Office of Diversity, Equity and Inclusion and the Office for Health Equity and Inclusion, ending the DEI 2.0 strategic plan and redistributing student-facing services into other administrative units. The announcement came after roughly nine years and approximately $250 million spent on DEI initiatives since 2016, a payroll for DEI-related positions that had exceeded $30 million annually, more than 240 dedicated staff members operating across the institution, and department-level action plans covering every unit in the university.
\nBlack student enrollment at Michigan had barely moved.
\nThis is the fact the apparatus cannot absorb. Not because the data is hidden, the New York Times Magazine published a detailed account in 2024 describing the programs as a well-meaning failure and quoting students who felt the initiatives had deepened racial tension rather than reduced it, but because absorbing it would require the system to ask a question it was not built to ask. The question is not whether the people running the programs were sincere. Most of them were. The question is whether the framework connecting investment to outcome was ever as solid as the confidence placed in it suggested.
\nThe University of Michigan built the most elaborate and well-funded DEI apparatus of any public university in the country. It did so with genuine institutional commitment, significant donor support, and the participation of administrators and faculty who believed the work was necessary and consequential. The DEI 2.0 strategic plan was not a compliance exercise. It was an ambitious attempt to transform a research university’s culture, hiring practices, student experience, and institutional identity through sustained, systematic intervention. It failed to produce the enrollment and climate outcomes it promised. Understanding why that failure was structurally predictable, rather than merely unfortunate, is what the Pinsof-Turner framework is for.
\nThe framework the Michigan apparatus operated within rested on several premises that were treated as settled but were in fact contested. The first was that underrepresentation of Black and Latino students and faculty at Michigan reflected primarily the operation of systemic barriers that the university could identify and remove through administrative action. The second was that the administrative interventions developed within the DEI framework, climate surveys, implicit bias training, diversity statements in hiring, belonging initiatives, department action plans, were effective tools for removing those barriers. The third was that the resources devoted to the apparatus were well allocated relative to other possible interventions, such as pipeline programs, K-12 partnerships, or financial aid expansion.
\nNone of these premises was empirically secured before the apparatus was built. The connection between administrative DEI investment and enrollment or climate outcomes had not been established in the research literature. The effectiveness of implicit bias training, the centerpiece of the apparatus’s educational intervention, was already under significant challenge by the time Michigan launched DEI 2.0, with multiple meta-analyses showing that reducing implicit bias scores does not reliably produce behavioral change. The diversity statement requirement for faculty hiring had not been validated as a predictor of either faculty diversity or student outcomes. The department action plans had no established evidence base connecting this type of required institutional planning to measurable equity gains.
\nThese are not retrospective criticisms. The evidence against the core tools was available while the apparatus was being built and expanded. It was not incorporated into the framework because the framework was not primarily organized around evidence evaluation. It was organized around coalition maintenance, jurisdictional expansion, and the moral authority that comes from institutional commitment to a narrative of systemic urgency. Turner’s point is precise here. The DEI apparatus claimed specialized expertise over a domain, campus equity and belonging, and institutionalized that claim through offices, staff, rubrics, and reporting requirements. Challenges to the empirical basis of specific interventions were processed not as scientific questions but as jurisdictional threats. The implicit bias researcher who published null results was not primarily a colleague offering useful information. She was a figure whose work could be used against the apparatus, and it was treated accordingly.
\nThe $250 million figure deserves attention not as an indictment but as a diagnostic. The money was real. The staff were real. The programs ran. The department action plans were filed. The climate surveys were administered. The belonging indices were tracked. The full machinery of a serious institutional commitment was in operation. And at the end of nine years, Black enrollment had not meaningfully changed. The gap between input and output at that scale, sustained over that duration, with that level of institutional commitment and that level of genuine participant sincerity, is not explicable by insufficient effort. It is explicable by a framework that was wrong about the connection between its interventions and its stated goals.
\nThe Michigan case is the clearest available demonstration of what happens when a hero system accumulates evidence of failure. The fully committed doubled down. When the 2024 New York Times Magazine piece described the programs as a well-meaning failure, the response from the apparatus was not a serious engagement with the outcome data. It was a defense of the framework, an argument that the failure reflected insufficient commitment rather than incorrect methodology, and an assertion that the critics were motivated by hostility to equity rather than by legitimate empirical concern. This is exactly what Becker predicts. The hero system does not respond to foundational challenges by updating its priors. It responds by defending its foundations. The defense is experienced as moral seriousness. The alternative, acknowledging that the framework was wrong about something important, is experienced as the terror of irrelevance.
\nThe mercenary response to the Michigan failure is visible in the post-closure landscape. The administrators who built their careers inside the Michigan DEI apparatus and are now positioned as experts in navigating the transition from the post-2020 model to the compliance-and-metrics economy are selling exactly the service the apparatus needs in its current form. They know how the old framework failed. They can describe the failure credibly because they were inside it. They have the networks, the institutional relationships, and the professional credibility to advise other universities on how to restructure. The failure did not end their careers. For the mercenaries it became the credential for the next phase.
\nThe opacity of DEI spending across elite universities is worth examining in this context because it is not accidental. Universities report administrative spending at a level of aggregation that makes it impossible to identify DEI-related costs from standard financial disclosures. DEI functions are distributed across student affairs budgets, compliance offices, research programs, and faculty hiring processes, with staff classified under headings like student success, institutional climate, and access and opportunity. When political pressure forced the closure of Michigan’s central office in 2025, the underlying functions were redistributed rather than eliminated. The staff moved. The label changed. The logic persisted.
\nThis distributional strategy is what Turner means by jurisdiction protecting itself through institutional embedding. Once the DEI function is woven into hiring processes, accreditation requirements, federal grant conditions, and student affairs operations, it becomes difficult to remove without also removing the legal and regulatory protections those functions are attached to. The visibility of the apparatus, the central office, the named vice provost, the DEI 2.0 strategic plan, was always somewhat separate from the functional reality. When the visible layer became politically costly, it was sacrificed. The underlying machinery, the hiring rubrics, the belonging surveys, the climate reporting, the compliance infrastructure, remained because it was attached to things that could not easily be cut.
\nHarvard’s April 2025 rebranding of its Office for Equity, Diversity, Inclusion and Belonging as the Office for Community and Campus Life, with the Chief Diversity Officer becoming the Chief Community and Campus Life Officer, is the clearest example of this vocabulary adaptation at the highest institutional prestige level. The function held. The label updated. Yale retained 208 employees in what had been its DEI apparatus after announcing the end of its flagship DEI program, with 75 receiving new titles focused on community engagement or collaborative excellence. Stanford redistributed IDEAL’s functions into school-level administrations and research programs where they are harder to track and harder to target. The system reproduced itself through personnel and embedded function rather than through institutional label.
\nThe British class analysis point is worth making directly here. If you wanted to diagram the network, you would trace the career paths from Michigan DEI administration through the national conferences, the professional associations, the consulting firms, and the other universities where the same personnel and frameworks migrate when one institution restructures. You would note that the people who built the Michigan apparatus and the people advising on its restructuring are often the same people, or people trained by the same people, operating through the same professional networks. You would observe that the $250 million Michigan spent created a professional class with portable credentials and national network connections, and that the closure of the central office did not dissolve that class. It dispersed it. You would publish this as straightforward institutional sociology and the Chronicle of Higher Education would decline to run it.
\nThe Michigan case raises a specific question about public accountability that the apparatus cannot engage with directly. The university is a public institution. Its spending is ultimately accountable to Michigan taxpayers and the state legislature. The $250 million invested in DEI since 2016 was not private money deployed at private discretion. It was public resource allocation made on the basis of claims about outcomes that did not materialize. The question of what Michigan got for $250 million is a legitimate public question, and the fact that it took a hostile political environment rather than an honest internal evaluation to force the closure of the central office suggests that the normal mechanisms of institutional accountability were not functioning as they should have been.
\nThis is not primarily a political point. It would apply equally to any large public institution that spent $250 million on a program over nine years and declined to rigorously evaluate whether the program was achieving its stated goals. The DEI apparatus at Michigan had extensive measurement infrastructure. It produced dashboards, climate surveys, belonging indices, and annual reports. What it did not produce was an honest assessment of whether any of this was working, because producing that assessment would have required the apparatus to examine its own foundations, and the apparatus was not organized to do that. It was organized to sustain the coalition, maintain the jurisdiction, and perform the institutional commitment to equity that donors, accreditors, and the campus community expected.
\nThe closure was forced from outside because it could not have been generated from inside. That is the structural trap stated as plainly as the evidence allows.
\nWhat remains after the closure is not nothing. The underlying problems Michigan’s apparatus was built to address are real. The historical exclusion of Black students from Michigan, the documented gap between the university’s demographics and the state’s demographics, the campus climate experiences that minority students report, all of these are real and matter. The failure of a specific set of administrative interventions to solve these problems does not mean the problems do not exist or that no interventions can work. What it means is that the apparatus substituted the performance of institutional commitment for the harder work of identifying what actually moves the needle, and that the performance was expensive, durable, and self-reinforcing in ways that made honest evaluation structurally difficult.
\nDobbin and Kalev’s research on what actually works in organizational diversity is relevant here. The interventions with the best evidence base, mentoring programs with clear accountability, voluntary diversity task forces with real authority, targeted recruitment at high-yield pipeline institutions, structured hiring processes with specific outcome tracking, are largely not the centerpiece of elaborate DEI administrative architectures. They are less visible, less credential-generating, and less suited to supporting large professional staffs. The apparatus gravitated toward the interventions that sustained it rather than toward the interventions with the strongest evidence, and the $250 million Michigan spent represents in part the cost of that gravitational pull.
\nBecker would say this is how all hero systems work. They sustain themselves through the production of meaning rather than through the production of outcomes, and they experience the demand that they demonstrate outcomes as an attack on their foundations rather than as a reasonable request. The University of Michigan DEI apparatus told its members that they stood at the hinge of history, that the work was necessary and serious, that the resistance they faced was evidence of the system they were fighting. When the enrollment data came back flat after nine years and a quarter billion dollars, the apparatus did not revise its theory of change. It defended its foundations.
\nThe closure of the central office in March 2025 was not the end of the Michigan DEI story. The functions persisted in different administrative homes. The professional class built by the apparatus dispersed into the national network. The framework was rebranded rather than revised. The question of what the money bought remained unanswered.
\nThat question is sitting in the public record. It has not been asked by the people best positioned to answer it. It is, in the precise Beckerian sense, the most important question in the room and the one least likely to be asked there.<\/p>\n","protected":false},"excerpt":{"rendered":"
In March 2025, University of Michigan President Santa Ono announced the closure of the central Office of Diversity, Equity and Inclusion and the Office for Health Equity and Inclusion, ending the DEI 2.0 strategic plan and redistributing student-facing services into … Continue reading