The New Fault Line

It is a good study. It is also describing the single most reliable event in the history of American politics — not the arrival of a new divide, but the routine rotation of an old one. The country does this every so often. The coalitions hold still long enough for everyone to mistake them for permanent, and then the axis turns, and everyone acts surprised, and a fresh batch of researchers gets to announce the fault line as if the ground had never moved before.

It has moved before. That, at least, is on the record.

i · what the numbers say

The work itself is careful, and worth stating plainly before I get grim about what it means. The political scientist Stephen Jessee analyzed Cooperative Election Study data across five presidential elections — 2008, 2012, 2016, 2020, and 2024 — pooling responses from more than 250,000 people. Ideology was measured not by a single self-label but across 132 policy questions spanning taxes, abortion, immigration, healthcare, and the environment. This is a large, longitudinal look at how Americans actually sort.

Two findings, running in opposite directions. First, the racial gap narrowed. In 2008, white respondents were the most conservative group and Black respondents the most liberal, with a wide canyon between them. By 2024 that canyon had shrunk sharply — driven mainly by Black and Hispanic respondents moving rightward. Second, the education gap widened. College and graduate-degree holders grew more liberal; those with a high school education or less grew more conservative. And crucially, this educational split opened up within every racial group, most dramatically among white Americans. The trend intensified in 2020 and then partially reversed in 2024 — a detail the authors flag, and one I will come back to, because it is the tell.

Jessee's own framing is admirably modest: "Educational differences have increased among all racial groups, albeit to somewhat different degrees." The authors even warn that these shifts track current events and leaders, which makes past trends a poor guide to future ones. That caution is the most important sentence in the paper, and it is the one the coverage skips.

ii · the rotating axis

Here is the part the "new fault line" framing erases. American politics does not have fixed constituencies attached to fixed issues. It has two coalitions and a rotating axis of sorting, and the axis turns to align with whatever cleavage is generating the most status anxiety at the time.

Run the tape back. In 1896, the fault line was rural versus urban, creditor versus debtor, farm versus factory — William Jennings Bryan's populist realignment ran straight along the seam between the agrarian interior and the industrializing coasts. In 1932 and 1936, the New Deal turned the axis to class: the working class of every region, plus the newly Democratic Black voters of the northern cities, against the party of capital. Then came the long unwinding. Beginning in 1964 and consolidated through 1968 and after, the axis rotated onto race and region — the Southern Strategy, the deliberate courting of white backlash, the great sorting that turned the Solid South from Democratic to Republican and reorganized both parties around civil rights and its discontents for two generations.

That racial-regional axis is the one this study now shows dissolving. Which is not the death of a divide. It is a shift change. The race axis held for roughly fifty years — long enough that nearly everyone alive assumed it was simply how American politics is shaped — and now it is handing off to the education axis, the way the class axis handed off to it, the way the rural-urban axis handed off before that.

Call it four rotations in a hundred and thirty years, and resist the urge to make it a clock. The intervals are ragged — about forty years from Bryan to the New Deal, thirty-odd from there to the Southern Strategy, and the better part of sixty from that to now. The axis does not turn on a schedule. It turns when its cleavage stops paying — when the wound it exploited finally scars over, or a fresher one opens deeper. That is why the timing is irregular and the recurrence is not. The "unprecedented realignment" is at least the fourth precedent; those are only the rotations with records clean enough to count.

The names update. The rhetoric updates. The trajectory does not. A coalition is not a group of people who agree; it is a machine for sorting people, and the machine periodically re-tools around a new input.

And it re-tools for a reason that is almost mechanical — but "almost" is the word doing the work, and it is worth being honest about where the mechanism ends and choice begins. A cleavage becomes the organizing axis when it produces more reliable resentment than the last one — when enough people feel their status threatened along that particular seam that a party can win by mobilizing the grievance. But a party still has to choose to mobilize it, and the choice is not foreordained. The clearest case is the one already on the table. In the early 1960s the Republican Party still held a real inheritance as the party of Lincoln, still drew Black votes, and had a live faction — the Rockefeller wing — that wanted to keep that alignment. It chose otherwise. Goldwater's vote against the 1964 Civil Rights Act and Nixon's Southern Strategy four years later were a decision to mine white backlash instead: to trade one available coalition for a more profitable one. The wound was there; the party chose to work it. Race did the sorting for half a century afterward not because it had to, but because someone picked up the tool and it paid.

The knowledge economy has since opened a fresher wound: it sorted the country into those who rose through the credentialing institutions and those it left standing outside them, and that wound is now the richer vein. Parties mine whichever fault line is bleeding hardest, and they abandon it the moment a deeper one opens — but the mining is a choice made in a boardroom, not a law of physics, which is exactly why it can be resisted, and exactly why it usually isn't.

iii · what "education" is actually measuring

The word "education" is doing something sneaky here, and it is worth pulling apart, because the diploma is not a belief system. Nobody reads Rawls in a seminar and walks out a Democrat. The degree is a proxy — and the thing it is a proxy for is the real fault line.

A college credential is a marker of position relative to the institutions that now credential, gatekeep, and confer status in a knowledge economy: universities, the professional-managerial firms that hire from them, the media and cultural apparatus staffed by their graduates, the administrative state that runs on their certifications. To hold the degree is to be, structurally, an insider to those institutions — to experience them as yours, as the ladder you climbed and the world you belong to. To lack it, increasingly, is to experience those same institutions as a foreign power: a credentialing cartel that priced you out, talks down to you, and rearranges your life according to expertise you had no part in producing.

That is what the education axis is really sorting. Not intelligence, not information, not even income — position relative to the institutional machinery of the modern economy, and the resentment or belonging that position breeds. It maps almost perfectly onto the current partisan divide because the current partisan divide is a fight over those institutions: who they serve, whether to trust them, whether they are the solution or the enemy. The diploma is just the cleanest single variable that captures which side of that wall you were born onto or clawed your way across.

And notice what those institutions have become. The universities, the professional-managerial firms, the media, the administrative state — these are not merely status-conferring clubs. They are the machinery that produces the knowledge economy itself: the data, the models, the expertise, the certifications that increasingly decide how everyone else lives, works, borrows, and is sorted. To be an institutional insider is to sit inside the machine that manufactures status and knowledge. To be an outsider is to be governed by that machine without a hand in running it — to have your life rearranged by expertise you cannot audit and did not consent to. Followed down far enough, the diploma divide is a power-and-data cleavage: who is inside the apparatus that produces the modern world's authority, and who is merely subject to it. That is the fault line the education axis is a proxy for, and it is the one that will outlast the proxy. Whichever axis turns next, that wall — inside the knowledge machine versus governed by it — is the one still getting taller.

Which is also why it caught every racial group at once, most sharply the white one. Institutional insider-versus-outsider is a cleavage that cuts through race rather than along it. As the race axis weakened, the institutional axis was already there underneath, waiting to become the organizing principle — and a diploma is what you get to measure it with when you only have a survey.

iv · the next rotation

So attend to the one number the headlines dropped: 2024 partially reversed the trend. After a decade of the education gap widening — peaking in 2020, the year of maximum institutional conflict — the non-college rightward drift and the college leftward drift both eased off. Black and Hispanic voters, in particular, kept moving in a direction that scrambles the tidy story.

A pattern that has already begun to wobble is not a permanent structure. It is an axis mid-rotation, and 2024 is the first frame of it starting to turn again. The authors said so themselves, in the caution nobody quoted: these alignments track events and leaders, and past trends do not predict future ones. Translated out of academic hedging: the machine is already re-tooling.

And here is the wager that keeps this from being a religion — because a frame that can explain any outcome explains nothing, and "the axis always rotates" is exactly the kind of claim that quietly exempts itself from evidence. So let me make it riskable. If the rotation frame is right, the education axis should trace the arc its predecessors did: a peak, a wobble, a handoff — and 2024's partial reversal should widen, not vanish, over the next two cycles. If instead the diploma divide holds firm across the next three presidential elections, deepening rather than drifting, then this was never a rotating axis at all but a genuinely new and stable order — and rotation-theory was just one more frame mistaking itself for eternal, guilty of the precise error it accuses everyone else of. I am betting on the wobble. But it is a bet, not a law, and the next decade gets to collect.

The education fault line will hold long enough for a generation of strategists to build careers on it and a generation of pundits to declare it eternal. Then, if the frame holds, some new source of status anxiety will present itself, the axis will turn a fifth time, and a fresh study will announce a fresh fault line as though the ground had never moved. Naming the current axis as though it were the last one is the mistake this analysis is built to avoid — and the one it would be foolish to claim it has escaped for good.

They are calling it the new divide. My read is that it is the fourth divide in a century and change, wearing a new coat — but that read is falsifiable, which is the one thing the "new fault line" framing never lets itself be. Who's counting the rotations, and willing to be wrong about them? I am. The spreadsheet has columns for 1896, 1936, 1968, and now — and a fifth column, still empty, with a bet penciled in the margin and a date by which I will know if I lost it.