On disrespecting authority, institutions, and independence

March 28th, 2026 — opinion

I. The Fence

The Pennsylvania Energy and Innovation Summit, held July 15, 2025, at Carnegie-Mellon, was convened by Senator Dave McCormick and featured President Trump. In the weeks preceding the summit, students circulated a petition opposing the event that gathered over 1,400 signatures. They painted the Fence, the University's century-old free speech landmark, with messages reading "CMU Community Says: Stop the Summit" and, later, "Protest the Summit," which they guarded for days.

On the morning of the summit, the Fence was repainted to read "No Rapists on Our Campus." President Farnam Jahanian ordered it painted over and took the Fence offline, the first in its hundred-year history that the tradition had been suspended.

The aftermath was swift and visceral. Students built a second forty-foot fence outside the president's office and adorned it with recreated messages from the Fence's history. The Instagram account @cmufence published the painters' response: "Shame on you for silencing us." Subsequent paintings accused Jahanian of supporting rapists. Even the president of the CMU Republicans called the decision "very uncharacteristic and concerning." Within a week, Jahanian reopened the Fence and issued an apology for failing to acknowledge the pain of sexual assault in his initial response. But trust, once broken, is not so easily repainted.

The Fence is not a public forum. It sits on private university property, is restricted to CMU affiliates by tradition and policy, and its use is governed by institutional norms: painting only between midnight and sunrise, covering the entire surface, no profanity. The First Amendment constrains the government, not Carnegie Mellon. So the students who cried censorship were not making a constitutional claim, and the administration that censored them did not violate anyone's legal rights. I state this because it clears the air: the debate over the Fence is not a debate about free speech law. It is a debate about institutional values, and those are far harder to adjudicate.

II. Selling out

The Students are right. Carnegie-Mellon sold out.

This is not a controversial claim if you study the sequence of events honestly. The administration denied students' request to designate the Fence as a peaceful protest space on the day of the summit. When the students painted it anyway, the university broke its own rules to silence them, twice, on the day the national spotlight arrived. Jahanian's letter framed the decision as a response to "anonymous, ad hominem messaging" that lacked "accountability and mutual respect." But the timing tells the real story.

Carnegie Mellon made a calculation: the institutional risk of embarrassing the executive branch, on camera, outweighed the institutional cost of betraying a hundred-year tradition. That is what selling out looks like.

The next logical question would be, "Was this the correct thing to do?"

I think it was. My answer lies in what a research university is as an institution, and in particular, in its relationship to the state. In my mind, this relationship is older and deeper than the Fence debate suggests.

III. Federal dependence

There is a fiction, popular among students and comfortable for administrators, that a university is an independent institution that happens to receive federal funding. This framing is comforting, but the relationship is backwards.

At Carnegie Mellon, over $650 million in annual research revenue flows through federal grants. At Johns Hopkins, federal sources account for roughly half of the university's $7 billion in total revenue. At MIT, 64 percent of campus-sponsored research comes from federal agencies. These are not supplementary funds. They are the operating budget.

So, the research university is an institution whose independence is made possible by federal funding, and is therefore conditioned on it.

This means that the decision to accept federal research money is itself a political act. It is not politically neutral to take money from the Department of Defense to build autonomous systems, or from the NIH to run clinical trials, or from the NSF to fund particle physics. It is a decision to enter into a relationship with the state, and relationships with the state come with the state's conditions. When a new administration arrives with different priorities and a willingness to use funding as leverage, the University discovers that its "independence" was always contingent.

This is not a critique, rather, it is a description. Federal funding has built the American research university into the most productive knowledge-generating institution in human history. But the exchange has a price, and the price is that you cannot take the government's money and then act as though the government has no claim on your behavior. You can negotiate the terms. You can push back at the margins. You can, as MIT's president put it, practice "engagement where possible, opposition where necessary, adaptation where sensible." What you cannot do is pretend that the relationship does not exist, or that your institution's political obligations ended when the grant was signed.

Consider the case of MIT, among the most vocal critics of Trump's funding cuts. Its president rejected the White House compact. It sued the NIH, DOE, and NSF over indirect cost caps and it framed each of these actions as a defense of academic independence. Yet MIT also operates Lincoln Laboratory, the Department of Defense's largest federally funded research and development center, running on over $1 billion annually, roughly 90 percent of it from the DoD, and freshly renewed under a $12.2 billion five-year contract with the Air Force. Lincoln Laboratory conducts mostly classified work, which is why it cannot be located on campus, though MIT faculty consult there and MIT students conduct thesis research there. An institution that genuinely believed academia should be protected from government entanglement would not house what is effectively a defense contractor inside its own structure.

This is not hypocrisy as MIT has never pretended it doesn't work with the defense department. Instead, this is the Fence's tension writ large, institutionalized. MIT's objection to funding cuts is not that government shouldn't have a role in shaping academic research. It is that the government changed the terms of a partnership MIT has participated in and profited from for decades. The dissonance is not between independence and dependence, it is in MIT's belief that the relationship should be governed by negotiated terms and the government's demonstration that it is governed by power.

The dependency is not only political. It is financial, and the financial arrangement is harder for universities to defend. For decades, universities have negotiated indirect cost rates, overhead charged on top of every research dollar, that range from 50 to 70 percent at top institutions. Harvard's negotiated rate is 69 percent. MIT's is 59 percent. Carnegie Mellon's is 53 percent. For every direct dollar awarded to a researcher, the university collects an additional 50 to 70 cents for facilities, administration, compliance, and infrastructure that might not even be directly related to the funded work.

Trump's efforts to cap these rates at 15 percent were blocked by federal courts as arbitrary and improper. But the political impulse behind the cap, the suspicion that universities have been padding overhead to subsidize operations that have nothing to do with the funded research, is not without some basis and is not a new accusation: Stanford faced a scandal over inflated overhead costs in the 1990s.

The NIH's own justification for the cap included a revealing data point: of the 72 universities surveyed, 67 were willing to accept research grants from private funders with zero indirect cost coverage. Harvard required a minimum of just 15 percent, the very same rate it called devastating when the government proposed it. Only three schools in the sample refused to accept rates below their negotiated federal rate.

If these overhead charges reflect the true cost of conducting research, it is difficult to explain why universities will do the same work for free when a foundation asks. The negotiated rate starts to look less like a cost recovery mechanism and more like a price set to what the federal government will bear. In other words, when a university president says "we are being robbed," the frank version of that sentence looks more like: we are being robbed by an administration that is also, not entirely without reason, accusing us of having robbed them first.

The students who painted the Fence were operating as though Carnegie Mellon is a sovereign entity that can say whatever it wants to whomever it wants without consequence. This is the privilege of an institution that does not depend on the person it is insulting. Carnegie Mellon is not that institution. No major research university is.

IV. Institutional stewardship

Farnam is doing a good job. This is an uncomfortable sentence to write, but it is true, and the discomfort it produces is precisely the point.

Public data makes the case more clearly than any argument can. In the eighteen months since Trump's second inauguration, Trump has frozen, cut, or clawed back federal funding from universities across the country.

Johns Hopkins saw $800 million in USAID grants terminated overnight and conducted the largest layoffs in its history: 2,222 positions. Columbia had $400 million in grants canceled before settling with the administration for $200 million, plus a $21 million EEOC fund. Roughly 180 researchers were fired as a consequence. Harvard ran its first deficit since 2020, borrowed $750 million, and faces an estimated $368 million per year in endowment taxes. Northwestern had $790 million frozen and eliminated 425 positions. Cornell saw over $1 billion in federal funds frozen, with $250 million in stop-work orders and $80 million in unreimbursed expenses. Even MIT, which posted a $426 million surplus in FY25, has imposed a hiring freeze on every position since February 2025 and projects a $300 million annual budget shortfall starting in FY27.

Carnegie Mellon, by contrast, broke even. It was not targeted by the Trump administration. It imposed no university-wide hiring freeze. Its research revenue grew 24 percent since FY20, exceeding $650 million. It made $33 million in proactive cuts before any deficit materialized. Its 75 layoffs were confined to SEI, its Defense contracting arm. It carries no endowment tax liability.

One may attribute this to luck, or to Carnegie-Mellon's smaller federal footprint relative to a Johns Hopkins or MIT. Both are partially true, but the pattern is hard to ignore. Universities that confronted the administration most visibly suffered the most severe financial consequences. The university whose president painted over the Fence on the day the president visited has emerged, by every material metric, in better shape than nearly all of its peers.

There is also a less charitable reading of Harvard's defiance that requires consideration. Harvard's endowment is $56.9 billion. MIT's is 27.4 billion. These are institutions that can absorb a $2.2 billion funding freeze and fight a multi-year legal battle without existential risk. When Harvard sued the Trump administration and its president declared the university would not "surrender its independence," the statement was principled. But it was also affordable.

A federal judge would ultimately rule in Harvard's favor, finding that the administration had used antisemitism as a pretext for what amounted to ideological retaliation. The ruling established binding precedent that the executive branch cannot weaponize funding freezes as retaliation for protected speech which is a precedent that shields every federally funded university, including the ones that kept their heads down. CMU benefits from Harvard's fight without having borne any of its cost.

But it is a victory that only an institution sitting on $57 billion can pursue. The moral stand doubles as a four-year investment: Harvard gets to be the university that stood up to Trump, and the reputational return on that narrative (e.g. donor enthusiasm, faculty recruitment, applicant prestige) will compound long after the administration leaves office. In other words, Harvard's defiance was real, but so is the calculation that defiance is the more profitable play.

Carnegie Mellon's endowment is $3.5 billion. It does not have the reserves to absorb a funding freeze, the legal budget to sustain a multi-year federal lawsuit, or the brand equity to convert political confrontation into a fundraising narrative. For Carnegie Mellon, defiance is not an investment. It is an existential bet, and the stakes are not abstract: they are the graduate students whose stipends depend on federal grants, the international researchers whose visa renewals depend on institutional stability, the labs whose DARPA contracts depend on a relationship with the defense establishment that does not survive a public fight with the president.

The students who wanted Jahanian to take a stand were asking him to make a bet he could not afford to lose with money and livelihoods that were not his to wager.

And, to me, this is where the accusation of selling out fails on its own terms. The students who painted "No Rapists on Our Campus" treated the summit as an aberration, as if Farnam had broken character by hosting a president he should have opposed. But hosting presidents is not new behavior for Farnam.

From 2011 to 2014, he led NSF's Computer and Information Science and Engineering Directorate under Obama. In 2022, he stood on the White House South Lawn as Biden signed the CHIPS and Science Act, a $280 billion package he had advocated for in congressional testimony, calling it "the result of extraordinary bipartisan cooperation." He joined Biden's President's Export Council in 2023. He served on Governor-elect Shapiro's transition team in 2022. CMU faculty testify before Congress almost monthly.

The summit was not Jahanian selling out to Trump. It was Jahanian doing what he has done across three administrations of both parties: positioning Carnegie Mellon at the center of the federal policy apparatus, because that is where the money, the influence, and the institutional protection come from. The Students saw a university president capitulating to an authoritarian. What they were actually watching was a university president doing his job-- the same job he did when Obama's NSF funded cybersecurity research, when Biden's CHIPS Act funded semiconductor R&D, and when McCormick's summit brought $90 billion in announced investment to Pittsburgh.

Jahanian absorbed the reputational damage personally to shield the institution from material harm. That is not cowardice. It is the unglamorous work of institutional stewardship, and the balance sheet vindicates it.

Consider what the alternative looked like in practice. When Columbia's president initially resisted the administration's demands over protest encampments, the result was not a principled standoff. It was $400 million in canceled grants, a forced settlement, and 180 researchers terminated. The stand lasted months but the consequences will last a decade. The president who took the stand is no longer president and the researchers who lost their positions had no say in the confrontation that cost them their jobs.

This is the arithmetic that rarely enters the leadership narrative. The visible act of refusal is legible as courage because it resembles what courage looks like in individual life. But institutional leadership is not individual life. When an individual takes a principled stand and suffers for it, the costs and the principles belong to the individual. When a university president takes a principled stand, the principles belong to the president and the costs belong to the postdoc on a J-1 visa whose lab just lost its funding. The courage framing obscures this gap. Jahanian's approach closes it, not by finding the right thing to say, but by accepting that the institution's standing mattered more than his own.

This does not make it right in any absolute sense. It makes it prudent. And prudence, in a political environment where the executive operates with an expanding concept of unitary authority, is what keeps institutions intact long enough to matter when the environment changes.

V. Moral clarity

There is a temptation, when thinking about the Fence, to take things to their logical conclusion. The students who painted "No Rapists on Our Campus" were doing exactly this: if Trump has been found liable for sexual abuse, and the university is hosting him, then the university is welcoming a rapist onto campus, and saying so is simply stating fact. The logic is airtight. The conclusion is correct. And the act of stating it was still strategically unwise ... not because it was wrong, but because being right is not the same thing as being effective.

But there is a harder version of the students' case that the strategic argument does not reach. The pragmatic objection, that Carnegie-Mellon cannot afford to insult the executive branch, would apply to any president. It would apply equally to a president the students admired. The students' objection was not, at bottom, about strategy. It was about whether a university should provide a platform to a person found liable for sexual abuse by a civil jury, regardless of his office. That is a moral question, and the language of institutional stewardship is not equipped to answer it.

So answer it on its own terms. Universities host people who have done terrible things. They hosted Kissinger. They host heads of state with records far worse than a civil liability finding. The question has never been whether a university should only platform the morally unblemished: that standard would empty every lecture hall and cancel every diplomatic visit in the history of higher education. A sitting president convened by a sitting senator for an energy policy summit is not a speaker invitation. It is a state visit in all but name. And universities do not decline state visits on moral grounds because they cannot afford to, not financially, though that is also true, but structurally. The moment a university arrogates to itself the authority to decide which elected officials are morally fit to appear on its campus, it has made itself a political actor, and political actors get treated politically. That is precisely what happened to Harvard and Columbia.

The students' moral clarity is grounded. But moral clarity that ignores institutional position is not courage. It is the luxury of people who do not have to sign the grant applications. It is the luxury of people whose stipends do not depend on the relationship they are asking the university to torch. And it is, in a specific and uncomfortable way, the luxury of transience: the students who painted the Fence will graduate. The international researchers whose visas depend on institutional stability will not. The labs running on DARPA contracts that require a functional relationship with the defense establishment will not. Jahanian does not get to be morally clear. He gets to be responsible, and responsibility means weighing the protest of students who will leave against the livelihoods of people who will stay.

That said, this is not an argument that students should be silent. Instead, it is an argument that the cost of speech is not equally distributed, and that the people who bear the least cost are often the most confident that the cost is worth paying.

This is the trap of rationalism applied to politics. In a system governed purely by logical consistency, you would always say the true thing, and the true thing would always produce the just outcome. But institutions do not operate on logical consistency. They operate on relationships, leverage, timing, and the slow accumulation of trust.

A university that loudly opposes the executive branch on moral grounds and subsequently loses its federal funding has not advanced the cause it fought for. The students who built a second fence demonstrated this in practice. They did not repeat the confrontation on the administration's terms. Instead, they changed the medium, reframed the narrative, and created a symbol that was harder to censor precisely because it was more creative.

What makes the Fence worth examining beyond its immediate politics is what it reveals about the relationship between authority and the social contracts that legitimize it.

A university's social contract is specific. Students pay tuition and produce academic work. Faculty conduct research and teach. The administration manages the institution and secures the resources that allow the whole enterprise to function. This contract is not the same as the social contract of a democracy, or a corporation, or a religious community, and the error that both sides of the Fence debate committed was treating it as though it were.

The students who painted the Fence were operating under the social contract of a democracy: citizens have the right to protest their government, and institutions that suppress protest are illegitimate. This is a reasonable framework, but it is not the framework that governs a private university. The administration that censored the Fence was operating under the social contract of a corporation: leadership must protect the organization's material interests, and expressions that threaten those interests can be curtailed.

The social contract of a university is neither of these. It is the contract of a community organized around inquiry, in which authority derives its legitimacy not from democratic consent or shareholder value but from its stewardship of the conditions that make inquiry possible. Under this contract, the administration's obligation is not to maximize institutional safety at any cost, nor to permit any expression without consequence, but to maintain the environment in which students and faculty can do the work that justifies the institution's existence in the first place.

A grave error in institutional policy (one I see repeated across every domain where authority meets dissent) is the assumption that policy can be designed agnostic to the social contract it serves. Federalism assumes the ability to provision rights from a central authority. Communism assumes negligible operating costs at the governmental level. Market capitalism assumes that entrepreneurship, left alone, generates growth. Each of these frameworks produces coherent policy within its own assumptions, and incoherent policy the moment it is transplanted into a system that operates on different principles. The Fence is a small example of this larger pattern.

Jahanian defaulted to corporate reasoning because that is what boards of trustees incentivize. The students defaulted to democratic reasoning because that is what they have been taught citizenship looks like. But a university is not a democracy, and it is not a corporation, and the Fence broke because neither side reckoned with that.

Variation 1

In my mind, the inquiry contract has its own hierarchy, and once you specify it, the conclusion is not comforting. Inquiry requires an institution. The institution requires survival. Survival, at a federally funded research university, requires a functional relationship with the state. This ordering means that when expression and survival conflict, expression is subordinate. Not because it does not matter, but because it cannot exist without the thing it is being sacrificed to protect.

Carnegie Mellon did sell out. The students were correct that a man found liable for sexual abuse was welcomed onto their campus and their speech was censored to make the visit comfortable. That is what happened, and no amount of institutional reasoning changes what it was. But more voices in the room would not have changed the outcome, because the outcome was not a failure of process. It was the framework working. A research university that depends on federal funding will, when forced to choose, protect the funding and sacrifice the expression. The students' moral clarity was not wrong. It was subordinate. And that distinction is the one no one wants to make.

Variation 2

Carnegie Mellon did sell out. I think it was probably the right call. But the fact that one person could make that call, unilaterally, for fifteen thousand students: that is the problem worth solving.