The Income Ladder Flipped

In 2008, the income ladder still worked the way every postwar textbook said it should. Voters reporting household incomes under $30,000 gave Barack Obama 62.4% of their two-party presidential vote; voters above $150,000 gave him 53.3%. Reading down the ladder from poor to rich, Democratic support fell — a 9.1-point pro-Democratic tilt at the bottom.

Sixteen years later the ladder is not flat. It is inverted. In 2024, the single most Democratic income bracket in the Cooperative Election Study was the richest one: voters above $150,000 gave Kamala Harris 57.2% of their two-party vote. The poorest bracket had slid to 51.5%, and every middle bracket — the $30,000-to-$150,000 majority of the electorate — sat between 46.9% and 49.3%. Democrats now do best at the very top of the income distribution, cling to a fading edge at the very bottom, and lose the middle outright.

This is not a one-cycle fluke, and the slope chart understates how steady the rotation has been. The rich-minus-poor gap — Democratic share among $150k+ voters minus the share among sub-$30k voters — ran −9.1 points in 2008, narrowed to −3.4 in 2012, crossed zero to +3.4 in 2016, held at +2.6 in 2020, and widened to +5.7 in 2024. The crossing happened the year Donald Trump first appeared on a ballot, but both ends of the ladder were already in motion before he arrived, and both kept moving after.

Two ladders creep, but only one flips

One objection should be dealt with immediately: the brackets are nominal dollars, and $150,000 bought a lot more social position in 2008 than in 2024. The weighted share of voters in the top bracket roughly doubled, from 6.7% to 13.3%, so today’s “rich” bracket reaches further down into the professional class than it used to. But bracket creep cuts against the finding, not for it. Diluting the top bracket with merely comfortable households should pull its politics toward the middle of the electorate — toward those 46.9-to-49.3% middle brackets — and yet the top bracket moved the other way. If anything, the inversion among the genuinely affluent is understated here.

A second objection — that online survey respondents lean left — has less bite than usual in this file. The CES’s weighted two-party vote tracks the official national count closely: 49.0% Democratic in the survey against 49.2% in the certified returns for 2024, and within about a point in every other cycle. The level may wobble; the sixteen-year rotation of the income gradient is far too large to be a house effect.

The diploma did it

Why would the rich move left while the poor move right? Slice the affluent by education and the mechanism is hard to miss. Among high-income voters, college graduates went from 55.0% Democratic in 2008 to a peak of 62.1% in 2020. High-income voters without a degree went the opposite direction in every cycle but one: 47.0% in 2008, a tick up to 48.0% in 2012, then 44.9%, 42.3%, and 41.1% in 2024. The diploma gap inside the top income group exploded from 8.0 points in 2008 to 19.8 by 2020.

Then 2024 broke the pattern — in one direction only. Affluent college graduates snapped back toward the Republicans, falling from 62.1% to 56.8% Democratic, a 5.4-point retreat that erased most of their Trump-era drift. Affluent non-graduates did not reciprocate; they kept sliding, to 41.1%. The within-affluent diploma gap narrowed to 15.6 points, but it narrowed because the Democrats’ best new constituency wobbled, not because their lost one returned. The party-identification series says the wobble is real rather than an artifact of the vote question: among the same high-income college group, the Democratic share of two-party party ID fell from 62.1% to 58.2% over the same four years.

Among white voters the same split is starker and the floor is lower. High-income college whites rose from 51.7% Democratic in 2008 to 59.0% in 2020 before falling back to 53.7%; high-income non-college whites dropped from 41.0% to 35.8%. By 2024, white non-college voters had converged to nearly the same number at every income level — from 35.8% to 38.1% across the low, middle, and high income groups. For white voters without a degree, income no longer carries detectable political information. The diploma is the class line.

The compositional consequence is the quiet headline of the whole series. High-income college graduates were 14.8% of Democratic presidential voters in 2008; in 2024 they were 29.5% — the single largest of the six income-by-education cells in the party’s coalition. Low-income voters without degrees traveled the reverse arc, from 30.5% of Democratic voters down to 20.8%. Part of this is electorate-wide drift — the whole country got richer on paper and more credentialed, and the Republican coalition’s high-income college share also grew, from 14.4% to 22.0%. But the normalized version survives the adjustment: relative to their share of all voters, high-income college graduates went from essentially proportional representation among Democrats (+0.2 points) in 2008 to overrepresentation of +3.8 points in 2024.

The union card lost to the diploma

If education has displaced income as the class divide, it should also outrank the older badges of working-class politics. It does. The CES asks whether anyone in the household currently belongs to a union, a series available from 2012 onward. Among voters without a college degree, current-union households gave Democrats 59.7% of the two-party vote in 2012. By 2024 that number was 46.9% — a 12.8-point collapse that carried the union household without a diploma below half. Never-union non-college voters fell too, from 50.4% to 42.3%, so the “union premium” among non-graduates shrank from 9.3 points to 4.5.

The crossover came in 2016. That year, for the first time in the series, the college graduate who had never been near a union (58.3% Democratic) out-voted the union household without a degree (55.4%). By 2024 the diploma-without-union beat the union-without-diploma by 7.2 points, 54.0% to 46.9%. Union membership still matters where it stacks with education — college-educated union households held solidly Democratic, 63.6% in 2012 and 66.0% in 2024 — but as a standalone class marker, the card now tells you less than the degree.

Housing tenure, the other classic marker, tells the same story in miniature. The renter-minus-owner Democratic gap narrowed from 18.7 points in 2012 (renters 66.0%, owners 47.3%) to 12.6 in 2024 (58.5% vs 45.9%) — and almost all of the narrowing came in one cycle, with renters swinging from 66.2% Democratic in 2020 to 58.5% in 2024, a 7.7-point lurch toward Trump among the younger and poorer half of the housing divide.

What this shows, and what it doesn’t

What it shows: in the largest continuous academic survey of American voters, the income gradient in presidential voting inverted between 2008 and 2024, the inversion is concentrated where income meets education, and the older markers of class politics — the union card, the rent check — have lost most of their independent signal. All figures are weighted, two-party shares, with cell sizes in the thousands.

What it doesn’t show: that the rich as a fixed social stratum became Democrats. Brackets are nominal; the 2024 top bracket is a broader, less elite slice than the 2008 one. Vote choice is self-reported and validated against nothing finer than national totals. Income goes unreported by 6.3% to 11.2% of two-party voters depending on the cycle (8.2% in 2024), and those non-reporters are simply absent from the ladder. The 2024 college snapback is a single observation — it could mark the start of a re-sorting or just an inflation-year incumbency penalty paid by the party in power. One more cycle will tell. What sixteen years of data already tell is that “working-class party” and “party of the poor” have stopped being descriptions of either side, and the ladder Americans climb no longer points the way it used to vote.