Polymarket
KalshiThe North Country's Open Seat Is Still a Republican Hold on the Board, but October Keeps Widening the Discount
The NY-21 House race in 2026, Republican Anthony Constantino against Democrat Blake Gendebien, is priced as a Republican hold at the northeast corner of New York State, and priced well short of what the district's own record would suggest, because the member who produced that record is not on the ballot. The Clerk of the House lists Elise M. Stefanik (Republican) as the sitting member for NY-21, but the State Board of Elections' amended certification of September 17, 2026 names Anthony Constantino on the Republican line and Blake Gendebien on the Democratic line for the seat, and Stefanik appears nowhere on it. What the price rests on, then, is a district rather than a person, and through October the market has decided that distinction is worth a good deal more than it had thought. Upstate voters settle it on November 3, 2026.
A twenty-point district with nobody defending it
The fundamentals here are not close. As Ballotpedia's race page stood on October 23, 2026, it gave NY-21 a Cook Partisan Voter Index of R+10 and recorded a 2024 presidential result inside the district of 60.1% for Donald Trump to 39.4% for Kamala Harris. The certified House results run the same way and have widened rather than narrowed: 58.8% to 41.1% on November 3, 2020, 59.1% to 40.8% on November 8, 2022, and 62.0% to 38.0% on November 5, 2024. Three cycles, three wins by more than seventeen points, each larger than the last.
Every one of those results was produced by the same member, and that is the whole complication. A seat held without interruption since January 2015 by one person is a seat whose record and whose incumbency cannot be told apart, and 2026 is the first year since then in which the market has to try. Ballotpedia counts NY-21 among the sixty U.S. House races in 2026 in which the incumbent is not running for re-election, and lists Stefanik among the withdrawn candidates in the June 23, 2026 Republican primary, which Constantino won over Robert Smullen by 59.2% to 40.3%.
What the price is saying, on October 23, 2026, is that a district which votes Republican by twenty points is now only about two in three to send a Republican to the next Congress. Polymarket had the Republican contract at 65% and Kalshi had the Democratic contract at 33%, the best price on each side. The path there runs one way: on October 5 Kalshi had the Republican side at 82% and Polymarket had the Democratic side at 18%; on October 9 the best prices were 71% and 29%; on October 12, 73% on Polymarket and 28% on Kalshi. The distance between the district's record and the current number is the open-seat discount, and it has grown by about seventeen points since October 5, 2026, which means the market now treats the vacancy as worth nearly as much as the district's whole partisan lean.
The poll that moved the board
The new evidence arrived in the form of the first independent poll in a month. The Siena Research Institute surveyed 496 likely voters in the district from September 30 to October 6, 2026 for the Times Union, WPTZ-NBC5 and WAMC, and found Gendebien at 46% against 42% for Constantino, with independent Christopher Schmidt at 5% and 8% undecided or refusing, inside a margin of error of 4.6 points. The same sample gave Bruce Blakeman 55% to Kathy Hochul's 40% for governor, which means the poll did not find a Democratic district: it found Republican voters at the top of the ticket who were not, at that point, carrying their House nominee.
Siena joins four earlier surveys Ballotpedia lists. SurveyUSA, for WNYT-TV, had Gendebien ahead 43 to 39 among 612 likely voters from August 4 to 8; J.L. Partners had Constantino ahead 53 to 40 from August 15 to 17; Impact Research, for Gendebien's campaign, had Constantino ahead 47 to 44 from August 30 to September 2; and Public Policy Polling had Constantino ahead 48 to 43 among 537 registered voters on September 9 and 10. Two of five now show the Democrat ahead, and none shows the Republican lead at the seventeen points or more the seat produced in each election from 2020 to 2024.
The forecasters have moved with it, and on Ballotpedia's October 20, 2026 update they split. In the July 21, 2026 update the Cook Political Report and Inside Elections carried NY-21 as Solid Republican, Sabato's Crystal Ball as Safe Republican and Decision Desk as Likely Republican. Sabato was at Likely Republican by September 22, Cook by September 29 and Inside Elections by October 6, so for two weeks all four rated the seat Likely Republican. On October 20 Cook and Inside Elections moved again, to Lean Republican, while Sabato and Decision Desk stayed at Likely. Lean means a small edge in a competitive race, and a contract at about two in three says roughly that in numbers, so the market and the two most recent rating moves now point the same way.
Heavy trading behind the move
The repricing came on real money, and the money has kept arriving. By October 23, 2026 about 158,500 contracts had traded on Kalshi's two NY-21 contracts, against about 141,800 on October 12, and Polymarket's event had drawn about $108.8K, against $93.7K on October 12. On Kalshi most of the activity is still on the Democratic side: the Democratic contract had changed hands about 104,600 times over its life against about 53,900 for the Republican one, and about 2,100 Democratic contracts traded in the 24 hours before the October 23 reading against about 430 Republican ones. On Polymarket the Republican line had drawn about $57.6K and the Democratic line about $51.2K.
Depth has come back since the thin week after the poll. Polymarket carried about $149.8K of resting liquidity across the event on October 23, 2026, up from $70.3K on October 12 and above the $131.9K of October 9, so the further fall in the Republican price happened on a fuller book, not an empty one. That makes the move harder to dismiss as a few orders walking a thin market.
The two venues moved together
The two books have moved down together all month, though they no longer sit on top of each other. On October 12, 2026 Kalshi had the Republican contract at 74% and Polymarket at 73%, and both had the Democratic contract at 28%. On October 23, 2026 Kalshi had the Republican side at 68% and Polymarket at 65%, while Kalshi had the Democratic side at 33% and Polymarket at 35%. Two venues falling by similar distances across the same three weeks is a change of view, not one participant leaning on one book; the three-point gap that has opened between them is a difference over how far the move runs, not its direction.
The case against the read
The strongest case against the current read is that one poll inside its own margin of error has been allowed to override three cycles of twenty-point results. Siena's four-point Democratic lead sits inside a 4.6-point margin, the earlier independent surveys from J.L. Partners and Public Policy Polling had the Republican ahead by five to thirteen, and the same Siena sample put the Republican candidate for governor fifteen points up in the district. A reader who holds that view would say a district that gives Blakeman a fifteen-point lead is unlikely to elect a Democrat to the House, and that the week's move priced a headline rather than the electorate.
There is money behind that view too, and more of it than a month ago. Federal Election Commission summaries through October 14, 2026 credit Constantino's committee with about $17.1M in receipts, of which $17.0M is loans from the candidate himself and none came from individual contributors, and about $2.3M in cash on hand. Gendebien's committee reported about $7.2M through September 30, 2026, about $6.9M of it from individuals, and about $470K in cash. A nominee who has now lent his own campaign seventeen million dollars, and who holds roughly five times his opponent's cash at the last reports, is the configuration in which an open seat gives away the least, and on that reading a price in the mid-sixties is paying for one survey and two rating changes and discounting everything else.
What changes the read
October 24 to November 1, 2026: early voting, which the New York City Board of Elections lists as running from Saturday, October 24 to Sunday, November 1. Early turnout matters more where the names on the ballot are new to the voters reading them.
The other two ratings: Sabato's Crystal Ball and Decision Desk still rated the seat Likely Republican on Ballotpedia's October 20, 2026 update. If either follows Cook and Inside Elections to Lean, or any of the four moves to Toss-up, the read that the vacancy is doing most of the work gets stronger; a move back toward Likely or Solid would say October went too far.
A second October poll: Siena's survey closed on October 6, 2026, and Ballotpedia listed no later poll as of October 23. Another independent poll showing the Democrat ahead would confirm the move; one showing the Republican back at a high single-digit or double-digit lead would leave the repricing resting on a single survey.
Late money: Constantino's latest filing runs to October 14, 2026 and Gendebien's to September 30. Ballotpedia counted about $325K in outside spending in the race as of October 23, 2026; a large late outlay on either side would be the first sign that party committees see the race the way the market does.
November 3, 2026: the general election, with every New York congressional seat on the ballot. A price in the mid-sixties is better prepared for a close count than one in the eighties, and the read is wrong if Constantino wins by anything like the seventeen points the seat produced in each election from 2020 to 2024.
How the NY-21 contract resolves, and when
The two venues do not ask quite the same question. Kalshi's KXHOUSERACE-NY21-26 pair, open since January 7, 2026, resolves on the party of the House member sworn in for the term that begins in 2027, with accelerated determination allowed once a consensus of media organizations projects a winner; the contracts carry a listed expiration of November 3, 2027 and expect to settle on January 4, 2027. Polymarket resolves on the party of whoever wins the November 3, 2026 election, by ballot-listed affiliation at the point the 2026 House races are conclusively called, falling back to official Federal Election Commission results where the reporting is ambiguous, and ends its event on November 3, 2026. Polymarket also lists a contract on another candidate taking the seat, which carried no volume on October 23, 2026. The candidate names each venue attaches to its contracts are labels; settlement reads party alone.
Related New York markets
The closest comparison on the board is NY-01, a Republican seat on eastern Long Island with a weaker record than this one, an incumbent defending it and a Republican price sixteen points above NY-21's on Polymarket on October 23, 2026, which is the clearest measure of what the exchanges think incumbency is worth. In the Hudson Valley, NY-17 is the other New York seat whose price moved hard in October, in the opposite direction. The statewide race on the same ballot is the governor's contest; the politics board and the elections hub carry the rest.
