The read · Metro Council
No party labels, top two advance: the new Metro Council math.
Louisville just ran its first nonpartisan council primaries. Six districts, twelve November finalists, and a May electorate small enough to fit in a high-school gym. Here's what the numbers can tell you — and the bigger thing they can't.
LOUISVILLE — This is the first Metro Council cycle in Louisville's history where the ballot doesn't say Democrat or Republican next to anyone's name. Thirteen odd-numbered districts are up in 2026. In six of them, enough candidates filed to force a May primary, and under the new rules the top two vote-getters advance to November regardless of party. Those six are the ones the May returns can tell us something about — so let's read them, carefully, because the most important thing about these numbers is how little they promise.
A ballot without party labels doesn't make voters nonpartisan. It just makes them unguided — and a campaign's job this fall is to be the guide the ballot stopped being.
The six boards, May results first, district context second:
District 3 — Keisha Dorsey 2,101, Kumar Rashad 1,395 advance (Shaun Spencer, 861, out). 4,357 total votes cast in a district with 17,191 registered voters — a quarter of the file showed up.
District 5 — Ray Barker 1,704 and Donna Purvis 1,040 advance from a six-name field. 4,298 cast, 18,049 registered.
District 9 — Alison Brotzge-Elder 2,722, Andrea Parr 2,390 advance from four (Mark England 1,835, Charles Todd 1,257). The May margin between the finalists is 332 votes — in a district with 23,141 registered. Hold that thought.
District 17 — Markus Winkler 3,957, Nancy-Lynne Walters 1,448 advance (Merideth Booth 1,201 out). The widest May spread on the board: Winkler took 60% of a three-way field.
District 21 — Betsy Ruhe 1,991, Lonnie Joseph 1,045 advance from four. 4,049 cast, 16,734 registered.
District 23 — Jeff Hudson 2,980, Ainsley Jones 1,924 advance (David Steff 886 out). 5,790 cast, 21,578 registered.
~4,000–8,200 · 16,700–23,100 · 3–5x
May votes cast per district · registered voters per district · how much bigger the November electorate can run
Primary totals from the SOS certified 2026 results via kypolitics; registration from the current Jefferson voter file. The seven other odd districts drew two or fewer candidates and go straight to November.
Now the honest part — what May does and doesn't tell you. A May primary electorate is the most engaged sliver of the file: older, more habitual, more partisan in behavior even when the ballot hides the label. In these six districts it ranged from about 4,000 to 8,200 voters, against registration files of 16,700 to 23,100. The November electorate in a midterm with a U.S. Senate race on top of the ballot will run a multiple of May. So a May margin — Winkler's 2,500-vote cushion, District 9's 332-vote squeaker — is a photograph of the sliver, not a forecast of the fall. Every one of these races resets to something close to even the day the electorate triples.
Two structural reads for anyone working these races, either side, any district:
The precinct logic flips. In the old partisan world, a campaign skipped the precincts registered heavily for the other party. In a nonpartisan top-two general, those precincts stop being enemy territory and start being persuadable territory — there is no party line on the ballot to hold a voter back from you. The campaigns that redraw their turf maps around that will find votes their old walk lists told them didn't exist.
Somebody has to replace the label. For decades, the party label did the heaviest lift in downballot races — thousands of voters walked in knowing nothing but the letter, and voted the letter. That letter is gone. The work it did doesn't disappear; it transfers to whoever does the voter-ID and the doors. In a low-information race with no label, the campaign that has actually identified its voters and physically pulls them in November holds an advantage the old system used to hand out for free.
The sleeper pools — registered Democrats who vote Novembers but skip Mays — run 2,777 to 4,118 across these six districts, and their Republican counterparts exist too. That's the composition change coming, in both directions. The May winners who treat their lead as real are making the oldest mistake in the book with a brand-new ballot.
What I'm watching: which of these twelve campaigns re-cuts turf for the no-label electorate by September, and whether November's returns in the old R-heavy precincts show the crossover the new rules just made possible. The votes, the files, the map — all public. I read them straight.
Disclosure: I work in Kentucky Democratic politics and consult campaigns in this state, and I hold a party role in Jefferson County. I read these races the same way regardless — that's the job.