The Death of the Loyal Opposition

What The Liberal Patriot’s closure tells us about the epistemology of political coalitions

Last week Ruy Teixeira and John Halpin shut down The Liberal Patriot. Five years of consistent, data-driven, often uncomfortable analysis of the Democratic Party’s structural problems, and the thing that killed it wasn’t being wrong. It was being right in the wrong rooms.

I have relied on this publication for years. I’m a political scientist by trade, and The Liberal Patriot was one of a vanishing number of outlets that treated empirical evidence as more important than team loyalty. Teixeira’s work on working-class realignment, the Hispanic voter shift, the cultural disconnect between the party’s professional-class base and the median voter it needed to win: this was careful, methodical, sometimes wrong on the particulars but usually vindicated on the structural trends by the next election cycle. His “Fox News Fallacy“ concept alone was worth the subscription. (The idea is simple: if Fox News criticizes Democrats for X, then there must be absolutely nothing to X. It’s a reflexive epistemological move that substitutes source-sorting for evidence evaluation. Democrats dismissed concerns about crime, immigration, and school curricula not because the data didn’t support the concerns but because the wrong people were raising them.)

The Fox News Fallacy isn’t a political strategy. It’s an epistemological commitment. And the world it created is the one in which The Liberal Patriot couldn’t survive.

The Standard Explanations

The commentary on the closure has been sympathetic but somewhat shallow. Chris Cillizza called it “a warning to everyone who cares about heterodox political coverage” and noted the structural problem: there’s a built-in funding mechanism for partisan sites, but no stable revenue model for the middle. Yascha Mounk wrote about donor pressure and foundation reluctance to fund publications that challenge orthodoxy. Patrick Ruffini called it “a sad commentary on the state of the Democratic coalition, where heterodox ideas are no longer tolerated.”

All true. All insufficient.

The deeper question isn’t why donors pulled funding. It’s why the ecosystem is structured such that a publication doing exactly what a healthy political party should want (honest internal criticism grounded in evidence) registers as a threat rather than an asset. And not just on the left, by the way — we’ll get there. The donor story is the mechanism. The question is what produced the mechanism.

The Donor Story

Teixeira has been open about what happened. A donor approached through an advisor offering multi-year support. The Liberal Patriot expanded, hired staff, developed a policy arm. Then Teixeira and his colleagues started writing about climate. They argued the catastrophist narrative was wrong on the politics, wrong on the policy, and in some respects wrong on the science. They published work questioning the 1.5-degree Celsius target as arbitrary and unscientific.

The donor’s contact in the environmental movement said something remarkable. Teixeira paraphrased it: this may be technically correct, but it’s harmful and pernicious and undermines the case for the climate movement.

Sit with that for a second. Technically correct, but harmful. The objection isn’t that the analysis was wrong. The objection is that correct analysis, deployed outside the boundaries of acceptable discourse, threatens the coalition. Truth is evaluated not by its correspondence to reality but by its utility to the project.

That’s not politics. That’s an epistemology. And it’s one I’ve been writing about.

Phase 3 Eats Its Own

I’ve spent the last several weeks tracing what I call the mutation arc in my “Reality Bats Last” series. The short version: over fifty years, a set of originally reasonable intellectual cautions (be careful about objectivity, power shapes knowledge, perspectives matter) hardened into philosophical dogma (objectivity is a myth, all knowledge is social construction) and then settled into what I call Phase 3: institutional common sense.

Phase 3 isn’t enforced by committee though. It’s enforced the way all orthodoxies are enforced: through the slow accumulation of what sounds smart and what sounds naive in professional conversation. You know which words to think twice about in a job talk. You know which framings get nods and which get raised eyebrows. You absorb what counts as sophisticated without anyone having to tell you explicitly.

You could call this groupthink and leave it there. But groupthink is about social pressure: people conforming because they want to belong. Phase 3 is about something deeper. It’s about what happens when a community’s evaluative criteria shift from is this true? to is this useful to the project? A different theory of knowledge operating under the surface, often without anyone noticing the swap. The donor who said “technically correct but harmful” wasn’t caving to social pressure. He was applying a coherent evaluative standard. It just wasn’t an empirical one.

TLP ran headfirst into that standard. Not in the academy (though it operates there too) but in the Democratic Party’s donor and institutional ecosystem, the thing John Judis and Teixeira called the “Shadow Party”: the NGOs, foundations, mainstream media, and academic institutions that surround the party and provide cultural infrastructure. Phase 3 is the operating system of the Shadow Party. It determines which questions are serious and which are harmful, which analyses are sophisticated and which are naive, which dissent is constructive and which is pernicious.

And here’s the thing about Phase 3: it has a motte and a bailey. The motte is the defensible position you retreat to when challenged. Of course we believe in honest self-criticism. Of course we want evidence-based policy. We’re the party of science. The bailey is the territory you actually occupy when nobody’s pushing. And in the bailey, “technically correct but harmful” is a perfectly coherent reason to defund someone.

Teixeira’s climate analysis threatened the bailey. His work on working-class realignment threatened the bailey. His willingness to say plainly that the party’s cultural positions were alienating voters, that the progressive moment was over, that the left’s 21st-century project had failed: all of it was Phase 3 heresy. Not because it was wrong (most of it has been vindicated repeatedly) but because it violated the institutional common sense about what serious reform-minded Democrats say and don’t say.

But Phase 3 isn’t just about policing ideas. It’s about protecting identities. I made this argument in Part 2 about humanities departments: scholars who’ve built careers on constructivist commitments don’t experience AI as an intellectual challenge. They experience it as an existential threat. The resistance is paradigm defense, not Luddism. It’s rational, given what’s at stake for them personally.

The same dynamic operates in the Shadow Party. There’s a charitable reading here: maybe the donors were doing rational coalition management. Climate activism mobilizes a huge chunk of the Democratic base. Questioning the 1.5-degree target, even if empirically defensible, might demoralize those activists, fracture fundraising networks, and weaken the coalition’s ability to mobilize on other fronts. That’s a legitimate strategic calculation, and I don’t want to dismiss it.

But I think something else was also operating, layered underneath the strategy. Reform-minded Democratic donors aren’t just people with policy preferences. They’re people whose social identities, professional networks, and moral self-understanding are organized around being the kind of person who cares about climate, who stands on the right side of history, who supports the arc of progress. The Liberal Patriot didn’t just challenge their strategy. It challenged who they understood themselves to be. Telling a donor that climate catastrophism is empirically wrong isn’t a policy disagreement. It’s an identity threat. And people respond to identity threats the way they always respond: by expelling the source of the threat, not by updating the belief.

This is why even the reform-minded donors couldn’t tolerate it. They could absorb “we need better messaging” because that leaves the identity intact (we’re right, we just need to communicate it differently). They couldn’t absorb “our core positions are alienating the voters we claim to represent” because that strikes at the root. If your climate commitments are wrong, you’re not who you thought you were. That’s not a funding decision. That’s self-preservation.

The Ratchet

The really brutal part is the ratchet effect. After 2024, you’d think The Liberal Patriot’s analysis would have been vindicated beyond any reasonable doubt. Trump won. Working-class voters of all races continued their migration away from Democrats. Hispanic support was halved. Young men swung 29 points. The “Shattering of the Democratic Coalition“ that Teixeira had been warning about for years happened on schedule.

And the response from the reform-minded Democratic donor class was: no money.

Not because Teixeira was wrong. Because the party’s favorable results in 2025 special elections and a few gubernatorial races made it easy to tell a different story. Democrats are winning again. The resistance is working. Affordability is the new message. The system is self-correcting. No fundamental reckoning needed.

Teixeira saw this coming. His final piece was titled “No Learning Please, We’re Democrats!“ and the core argument was bleak: Democrats mistake thermostatic reactions against an unpopular incumbent for genuine ideological support. Working-class voters who’ve soured on Trump haven’t warmed to Democrats. They’ve disengaged. The party’s favorability remains low. Its brand remains broken. But short-term electoral wins create enough noise to drown out the structural signal.

Phase 3 doesn’t just suppress dissent. It generates its own confirmation. Every piece of ambiguous evidence gets sorted through the existing framework, and the framework always survives. This is the same unfalsifiability problem I identified in the constructivist response to AI. The theory explains everything, which means it explains nothing, which means the theory survives regardless of what happens.

Which means I owe you a falsification condition for my own argument, or I’m running the same play I’m criticizing. So here’s what would change my mind: if Democrats nominate someone who explicitly repudiates the party’s cultural orthodoxies (not dog-whistle moderation but actual substantive breaks on, say, climate or immigration or gender) and the donor class funds that campaign enthusiastically, Phase 3 isn’t as entrenched as I’m arguing. If working-class voters return to Democrats without the party changing its positions, the structural problem is overstated. I don’t think either is likely. But you have to be willing to say so, or what are we even doing here.

This Isn’t Just a Left Problem

Now, I can already hear a version of this piece that turns into a 3,000-word dunk on progressives. That’s not what I’m doing. Partly because Teixeira already wrote those pieces better than I could (seriously, go read “The Hills the Left Will Die On“ and “The Progressive Moment Is Over“ before the archive disappears). But mostly because Phase 3 isn’t a left-wing phenomenon. It’s a coalition phenomenon.

You don’t have to hypothesize about this. When conservative economists and media figures started questioning Trump’s tariff regime with real trade data, the right-wing media ecosystem fractured along exactly the lines you’d predict: the empirically grounded dissenters were treated not as loyal critics strengthening the movement but as traitors undermining it. “Technically correct but harmful to the movement” operates with equal force on the other side. Try writing for a right-populist audience that the institutional expertise they despise actually got some things right. Try suggesting that the administrative state, whatever its pathologies, also does things people rely on and can’t easily replace. The immune response is the same.

Teixeira himself acknowledged this. His report with Yuval Levin at AEI, “Politics Without Winners,” argued that neither party is capable of or perhaps even interested in expanding its coalition. They toggle back and forth, each over-pursuing its agenda when in power, each failing to build a durable majority. Phase 3 is the epistemological mechanism underneath that toggling. Both coalitions have developed institutional immune systems that treat internal dissent as infection rather than information.

Rob Henderson’s concept of “luxury beliefs“ is useful here too. The positions that defunded TLP (climate catastrophism, unrestricted immigration, the full suite of progressive cultural commitments) are luxury beliefs in Henderson’s sense: they confer status on the professional class that holds them while imposing costs on the working-class voters the party claims to represent. But this isn’t a one-sided diagnosis. The right has its own luxury beliefs (that tariffs won’t raise consumer prices, that slashing government agencies has no consequences for the people who depend on them). Every coalition has positions it holds for signaling reasons that it can’t afford to examine empirically. Phase 3 is what happens when the signaling calcifies into common sense.

What Was Actually Lost

The Liberal Patriot wasn’t a moderate publication in the sense that word usually implies (cautious, both-sidesy, splitting every difference). It had positions. Strong ones. Teixeira argued Democrats needed to embrace equality of opportunity over equality of outcome, support border enforcement, acknowledge biological sex in relevant policy domains, abandon climate catastrophism, and prioritize economic growth over cultural signaling. These weren’t centrist hedge-trims. They were direct challenges to the party’s institutional consensus.

What made the publication valuable wasn’t the specific policy positions (smart people can disagree about all of them). It was the method. Start with the data. Follow where it leads. If it leads somewhere uncomfortable for your coalition, say so anyway. Evaluate claims by their correspondence to evidence, not their utility to the project.

That method is Phase 1 thinking. It’s what the original intellectual cautions were supposed to protect: careful, empirically grounded, reflexive but not relativist. The Liberal Patriot practiced what the academy preaches and the political ecosystem punishes.

The Question

I keep coming back to something I wrote at the end of Part 2: Can institutions acknowledge that they’ve overcorrected without destroying the real insights they were trying to protect?

The Liberal Patriot’s closure suggests the answer, at least for political institutions, is: not yet. The donor who said “technically correct but harmful” wasn’t being cynical. He was being Phase 3-rational. Within the logic of coalition maintenance, empirically uncomfortable truths really are harmful if they weaken the coalition’s ability to win elections and implement policy. The problem is that suppressing those truths doesn’t make them less true. It just means the coalition keeps running into the same wall, over and over, and keeps blaming the wall.

Teixeira will keep writing at the Free Press, Persuasion, the Washington Post, and other venues. The analysis won’t disappear. But it will be dispersed, occasional, shaped by each outlet’s audience and editorial needs rather than by the method itself. A dedicated publication creates something a guest column can’t: accountability to a framework over time. The Thursday column built a cumulative record. Predictions were testable. When the 2024 data came in, readers could check TLP’s 2021 forecasts against what actually happened. That kind of longitudinal accountability is what Phase 1 thinking looks like in practice, and it’s exactly what dispersed publication dissolves.

The loss isn’t the ideas. The loss is the institution, however modest, that was organized around the principle that your own side’s data matters more than your own side’s comfort.

That’s a hard thing to fund in a Phase 3 world. It turns out the market for honest self-examination is thin on all sides.

Reality still bats last. It just doesn’t have a Substack anymore.