← Newsroom

Donors beat dollars

Since 2016, forty-five sitting House Democrats have faced a credible challenger measurably to their left, on a measure built from federal spending filings. Nine of them lost. The challengers who beat them did not have more of the money in their race. What they had was three and a half times as many donors.

Money did not separate them. Donors did.

Challengers who won raised more than those who lost — but so did the incumbents they beat, and by more: four times what an incumbent who won their primary raised. Either way the challenger was outspent by roughly 2.5 to 1. Money is not what tells the nine apart from the thirty-six. Every figure is a median, comparing the 9 challengers who won against the 36 who lost, counting only money raised before the primary from FEC filings.

What does separate them is the number of people writing the checks.

Our study of general elections found the same shape: doubling a challenger’s donor count was worth a point of vote share while the dollars stayed flat. That was 17,284 candidates over four decades. This is the primary-election version, and it points the same way.

A midterm, and a seat the Democrat cannot lose in November

Winners did not just run in a different year, they ran in a different kind of seat: one the Democratic nominee wins in November no matter what, so the primary is the whole contest. These are separate rates, not two halves of one number. Read each row as: of the challenges that won, this share had the trait — and of the challenges that lost, this share had it.

Challengers keep trying. Most cycles, almost none of them win.

Each column is one election cycle. The full height is how many progressive challenges were mounted; the solid tiles at the bottom are the ones that beat the incumbent. Results are from OpenElections and published election boxes. One in five of these challenges won, against roughly one in twenty-five for primary challenges generally.

New York won four. Everywhere else that won, won once.

Nine wins came from six states. New York accounts for four; the other five states managed exactly one each. Twelve states mounted 23 challenges between them and won nothing.

All 45 challenges and all 9 wins.

Is the gap just the kind of seat they ran in?

Winners ran in far safer Democratic seats than losers — 78.0% Democratic at the last general against 63.9% — and that is where donors already live. So the gap might be describing the district rather than the campaign. Narrowing the comparison says otherwise.

Safe seats are the only cut that can be tested this way: among the 34 challenges with a last-general result, the competitive seats hold a single win between them.

Every challenge that won

Vote shares are the primary result, not the November one. Eight of the nine came in a midterm year; the exception is Jamaal Bowman’s in 2020.

Texas’s 28th in 2022 is counted here as a challenge that lost. Our result file for that race covers a fraction of the vote and shows Jessica Cisneros ahead; the certified count has Henry Cuellar taking the March primary 48.7% to 46.9% and the May runoff 50.3% to 49.7%.

How this was built

The election results

Nobody publishes American primary results in one place, so we built the file: 6,948 result files from OpenElections, a volunteer archive covering 47 states, plus published election boxes for the recent cycles the volunteers have not reached yet.

Deciding who counts as progressive

We did not label anyone by hand. Instead we used the fact that outside groups tell you, by fighting. When one committee pays to defeat a Democrat in a Democratic primary and a different committee pays to defend them, those two committees are on opposite sides — and so are the candidates. Feeding every independent expenditure filing since 2014 into that logic sorts the committees into two camps on its own. It puts Working Families and Justice Democrats at one end and AIPAC’s political arm and DMFI at the other. We never told it which was which; it worked that out from the spending.

Filling the gaps

That only sees candidates somebody spent money against, which leaves out Ocasio-Cortez in 2018 — almost no outside money touched that race. So a second step looks at who donates to whom, learns which donor patterns go with the scores we already trust, and uses that to place the rest. It is right about two-thirds of the time, which is useful but not certain, and every figure on this page also appears for the group scored the more reliable way.

Who was the incumbent

From FEC candidate filings, matched within a single seat in a single year, and refusing to guess whenever a name could belong to two people.

The money window

Only what a candidate raised between the start of the election cycle and the day of the primary. Anything after the primary is money the contest had already decided.

One seat left that fits the pattern

Most of the 2026 calendar is done. Massachusetts votes on 1 September and New Hampshire on 8 September; Rhode Island and Delaware have no contested Democratic primary against a sitting member. Eight seats still have an incumbent facing a challenger. Two of them are not really challenges at all: Seth Moulton and Chris Pappas have each raised nothing for their House seat because both are running for the Senate, leaving those races open.

That leaves six. Applying what the nine past winners had in common — a midterm year (2026 qualifies), a safe Democratic seat, and a challenger with several times the incumbent’s donors — only one clears the bar.

Donor counts are individual contributors since the cycle opened in January 2025, from FEC filings. The nine challengers who won had a median of 905 donors.

---