Kalshi Politics — Mispricing Audit

Report date 2026-05-28 · category politics · horizon 45 days (markets closing on/before 2026-07-12) · portfolio capital $1,000

Bottom line: 3 picks, all on the NO side, all betting that institutional/legislative processes resolve slower than the market implies into a tight deadline. ~$899 deployed, ~$101 reserve, blended EV ≈ +19%.

1 · How this was researched

Pipeline, in order:

Filter. Pulled every status='active' market with close_time between now and 2026-07-12 from the local Kalshi SQLite mirror (/tmp/bot.db, snapshots fresh to 2026-05-27 21:27 UTC). The mirror has no usable category column, so politics was isolated by ticker-series + title keywords (president, senate, congress, election, tariff, governor, impeach, pardon, nominee, confirm, reconciliation, …), then "what will X say" / "mention" / Truth-Social-post / EO-count noise series were dropped. That left 158 political contracts.
Liquidity cut. Dropped volume24h < 1,500, spread > 5¢, and anything already pinned (YES bid ≥ 95¢ or YES ask ≤ 5¢). Politics this week is dominated by settled markets — impeachment (0¢), Walz-out (3¢), Fed-chair-confirmed (100¢), "try to fire Powell" (1¢) — so the tradeable set is thin.
Diligence. For each survivor: resolution rules from the mirror, 14-day intraday price history from market_snapshots, live orderbook depth from the Kalshi public API (top-of-book derived as YES ask = 100 − best NO bid), and news diligence against primary sources (Roll Call, CBS, The Hill, Al Jazeera, RealClearPolling).
Reject. Killed anything where the rules carry a definitional trap, the edge needs private info, the price already moved on the thesis, or I couldn't size without ≥3¢ slippage.

Top-of-book and history are reconstructed from the mirror + a live orderbook pull on 2026-05-28; quoted live prices can drift intraday. All probabilities below are my subjective estimates, shown with reasoning so you can disagree with the inputs.

2 · Markets reviewed

Survivors of the liquidity screen and their disposition. "YES px" is live top-of-book (bid/ask, ¢).

Ticker	Question (short)	Vol 24h	YES px	Verdict
KXSENATEREC-26MAY-JUN06	Senate passes a reconciliation bill before Jun 6	1,039	30 / 31	PICK 1 · BUY NO
KXKNESSET-27-JUN30	Israeli Knesset dissolved before Jun 30	4,494	54 / 55	PICK 2 · BUY NO
KXLEAVESTARMER-26JUL01	Starmer leaves / announces exit as UK PM before Jul 1	6,737	15 / 16	PICK 3 · BUY NO
KXSENATEREC-26MAY-JUN14	Senate reconciliation before Jun 14	2,586	59 / 60	WATCHLIST
KXSENATEREC-26MAY-26JUL01	Senate reconciliation before Jul 1	200	75 / 76	WATCHLIST
KXVOTEHUBTRUMPUPDOWN-26MAY28	Trump VoteHub approval > 38.6% on May 28	12,511	82 / 83	PASS — efficient
KXAPRPOTUS-26MAY29-*	RCP approval bucket ladder (8 contracts)	~2,400	ladder	PASS — efficient
KXNEWTARIFFS-26MAY-JUN01	New/increased tariff exec action in May	3,658	5 / 6	PASS — settled
KXDNCAUTOPSY-26-JUN01	DNC releases 2024 post-election report by Jun 1	1,663	25 / 30	PASS — stale/wide
KXMAMDANIEO-26MAY30-T0	Mamdani signs > 0 EOs (weekly count)	3,832	14 / 15	PASS — noise
KXIMPEACH / KXLEAVEWALZ / KXFEDCHAIRCONFIRMED / KXALIENS …	High-OI but pinned (0¢ / 3¢ / 100¢ / 3¢)	10k–28k	pinned	SKIP — fully priced

3 · Top picks — thesis & detail

Ordered by conviction (defensibility = source quality + rule clarity + orderbook depth), highest first. Conviction is not EV-ranked: Pick 2 has the fattest edge but more two-sided risk than Pick 1.

KXSENATEREC-26MAY-JUN06 · BUY NO @ 0.70 · High conviction

Live: YES 30 / 31 → NO ask 0.70 True P(YES) ≈ 20% NO fair ≈ 0.80 Edge ≈ +10¢ EV ≈ +14% Closes 2026-06-06

Question. Resolves YES if "a reconciliation bill has passed the Senate after Issuance and before Jun 6, 2026." Clean, unambiguous rule — passage = a recorded Senate vote, full stop.

Mispricing thesis. The market still hands the Senate a ~30% chance of passing the $70B "Reconciliation 2.0" (ICE/CBP/White-House-security) package within the first five calendar days back from recess. That is too generous given the chamber isn't even in session until June 1 and the substantive blockers are unresolved.

Evidence

Senate punted the vote past Memorial Day recess on May 21; senators return June 1. Majority Leader Thune: "The Senate wants to measure twice and cut once, and we're not quite ready to cut yet." (Roll Call, 2026-05-21)
The blockers are substantive, not scheduling: a $1.8B DOJ "anti-weaponization" fund GOP senators "remained unconvinced" on after meeting Acting AG Blanche, plus a $1B White-House-ballroom Secret Service provision the parliamentarian already flagged. (CBS News; Roll Call)
Schumer has said Democrats will contest both provisions during the June vote-a-rama — i.e. a full amendment marathon still lies ahead before any final vote. (Roll Call)
House Speaker Johnson signaled the House takes it up "as soon as July 2" with a July-4 signing goal — leadership's own timeline assumes Senate passage in late June, not the first week. (Conference Board / Ballotpedia)

Numbers

Buy NO @ 0.70 · payout $1.00 · EV/contract = 0.80 − 0.70 = +$0.10 (+14.3%) · max gain if NO = +42.9% · $400 → 571 contracts.

Liquidity

NO is bought by hitting YES bids: $1,270 resting at 0.30, $250 at 0.29, $500 at 0.27. ~$1.5k available inside 71¢ — a $400 fill costs <1¢ slippage.

14-day price history

YES: 05-14 94 · 05-15 93 · 05-16 91 · 05-20 98 → 05-27 40 → live 30/31

Was near-certain (98¢) on May 20; collapsed ~58¢ after the May 21 punt and is still drifting down — confirming the move is the recess news, not a stale print. The decline is continuing, which supports (not contradicts) the NO thesis.

Tail risk (cleanest way to lose): Republicans cut a fast deal on the anti-weaponization fund and Thune forces a compressed motion-to-proceed + vote-a-rama in the first week back, clearing final passage by Fri June 5. Possible under Trump pressure, but contradicted by Thune's own "not ready" posture.

KXKNESSET-27-JUN30 · BUY NO @ 0.46 · Medium conviction

Live: YES 54 / 55 → NO ask 0.46 True P(YES) ≈ 40% NO fair ≈ 0.60 Edge ≈ +14¢ EV ≈ +30% Closes 2026-06-30

Question. Resolves YES "if the Israeli Knesset is dissolved before Jun 30, 2026." Legal dissolution occurs only when the dissolution bill clears its final (third) reading — a high, well-defined bar.

Mispricing thesis. The market reads the May 20 preliminary vote as near-coin-flip-to-done. But a preliminary reading is the first of four hurdles; completing committee + three readings inside ~4.5 weeks, while Netanyahu actively negotiates with the Haredi factions, is less than the implied 54%.

Evidence

On May 20 only the preliminary reading passed (110–10). The bill "will now pass to a committee before three more parliamentary readings… a process that could take weeks." (Al Jazeera; The National, 2026-05-20)
Al Jazeera's own read: completion before June 30 "appears unlikely given the remaining three readings and the several-month procedural window"; the statutory election deadline is Oct 27, with September floated. (Al Jazeera)
The trigger is a Haredi fight over the military-service exemption law; reporting indicates ultra-Orthodox parties "may still be coordinating with the coalition behind the scenes on timing" — i.e. dissolution is being used as leverage, which historically stalls these bills. (Times of Israel; Haaretz, 2026-05-27)

Numbers

Buy NO @ 0.46 · payout $1.00 · EV/contract = 0.60 − 0.46 = +$0.14 (+30.4%) · max gain if NO = +117% · $350 → 760 contracts.

Liquidity

NO bought against YES bids: $2,115 at 0.54, $791 at 0.53, $1,000 at 0.27. ~$2.9k inside 47¢ — a $350 fill is well within 2¢ slippage.

14-day price history

YES: 05-14 62 · 05-15 48 · 05-16 44 · 05-20 46 → 05-27 54 → live 54/55

Drifted up ~8¢ into the late-May procedural votes. Liquidity was thin before May 16; treat early prints with caution.

Tail risk: the coalition genuinely "completes its days" (chairman Ofir Katz's phrase) and rams all three readings through in June — Israeli dissolutions can move fast once leadership commits. This is real two-sided, headline-driven risk, which is why it's sized below Pick 1 despite the bigger edge.

KXLEAVESTARMER-26JUL01 · BUY NO @ 0.85 · Low conviction

Live: YES 15 / 16 → NO ask 0.85 True P(YES) ≈ 11% NO fair ≈ 0.89 Edge ≈ +4¢ EV ≈ +5% Closes 2026-07-01

Question. Resolves YES if Starmer "has either officially announced their intention to leave… or has actually left as Prime Minister" before Jul 1. Clean rule.

Mispricing thesis. A high-probability anchor, not a fat edge. The acute post-local-election crisis has peaked and Starmer has dug in; 15-16¢ slightly overstates the chance he announces or exits within five weeks.

Evidence

After Labour's wipeout (~1,500 council seats lost; Reform UK +1,454), Starmer told ministers he "had no intention of quitting" and would not "plunge the country into chaos." (CBS News; CNN live blog, 2026-05-12)
Wes Streeting resigned and is challenging — but a challenge needs 81 of 403 Labour MPs behind one name, with "no consensus… not likely any time soon." (CNN; Wikipedia: 2026 UK government crisis)
He survived the make-or-break cabinet meeting; the resignation rumor that spiked this contract to 66¢ on May 16 has fully unwound. (Townhall, 2026-05-18)

Numbers

Buy NO @ 0.85 · payout $1.00 · EV/contract = 0.89 − 0.85 = +$0.04 (+4.7%) · max gain if NO = +17.6% · $150 → 176 contracts.

Liquidity

Deep: best YES bid 0.15 backs ~$11,470 of NO at 0.85 — effectively no slippage at this size.

14-day price history

YES: 05-14 31 · 05-15 26 · 05-16 67(spike) · 05-20 28 → 05-27 22 → live 15/16

Caveat (rejection-criterion flag): the thesis has already been mostly priced (66→15 over ten days, still falling). The easy money is gone; this is a small, high-win-rate carry position, hence Low conviction and the smallest allocation.

Tail risk: a fresh shock (more cabinet resignations, a confidence vote, or Starmer choosing to jump before being pushed) flips it fast — UK leadership crises are non-linear. At 85¢ cost the downside is the full stake.

4 · Recommended $1,000 portfolio

Deploy ~$899 across the three NO picks; hold ~$101 (≈10%) as a cash reserve for opportunistic adds (see triggers below). Use limit orders only — never cross the spread on these.

#	Ticker	Action	Limit	Contracts	Cost	Max payout	EV ¢/ct	EV %	Conv.
1	KXSENATEREC-…-JUN06	BUY NO	0.70	571	$399.70	$571	+10.0	+14.3%	High
2	KXKNESSET-27-JUN30	BUY NO	0.46	760	$349.60	$760	+14.0	+30.4%	Medium
3	KXLEAVESTARMER-26JUL01	BUY NO	0.85	176	$149.60	$176	+4.0	+4.7%	Low
Total deployed					$898.90	$1,507	—	+19.0%	—
Cash reserve					$101.10	—	—	—	—

Blended EV ≈ +$170 on $899 at risk (+19.0%). Figures use my subjective true-probabilities; they are only as good as those inputs.

Risk profile

Worst case (all three resolve against me): −$899 (full deployed stake). Requires the Senate to rush passage by June 5, the Knesset to complete three readings by June 30, and Starmer to quit — three independent surprises across three countries.
Best case (all three NO): payout $1,507 on $899 → +$608 (+68%).
Most-likely outcome: Picks 1 and 3 resolve NO (78% / 89%), Pick 2 is a genuine coin-flip leaning NO (60%). Expected ≈ +$170.
Concentration / shared thesis: all three are "status quo persists past a tight deadline / process moves slower than implied" bets. The underlyings are uncorrelated (US Senate vs Israeli coalition vs UK Labour), but they share a behavioral premise — if you think markets systematically underprice institutional speed, treat this as one thesis, not three, and trim.

Execution notes

Limit orders at the quoted prices (NO 0.70 / 0.46 / 0.85). All three fill inside ≤2¢ slippage at these sizes per the live books; do not chase if they move >3¢ against you.
Watchlist adds (deploy the $101 reserve):
- KXSENATEREC-…-JUN14 (YES 60): if the Senate is still tangled in the anti-weaponization fund past ~June 8 with no scheduled final vote, NO at 0.40 becomes attractive — a second, later expression of the same Pick-1 thesis.
- KXKNESSET-27-JUN15 (YES ~20): cheap NO optionality if you want to fade the earlier dissolution date with less premium.
What invalidates each thesis (exit triggers):
- Pick 1 — Thune schedules a final passage vote for the week of June 1, or a deal on the DOJ fund is announced. Exit NO.
- Pick 2 — the dissolution bill clears its committee and a first reading is fast-tracked with a fixed second/third-reading date in June. Exit NO.
- Pick 3 — a second cabinet resignation, a formal no-confidence motion, or 81+ MPs coalescing behind a challenger. Exit NO immediately (this one is non-linear).
Hedges: Picks are small and independent; no cross-hedge needed. The natural hedge on Pick 1 is simply not also buying the JUN14/JUL01 NO — keep the reconciliation exposure to one node unless a watchlist trigger fires.

5 · What I rejected and why

KXVOTEHUBTRUMPUPDOWN-26MAY28 (YES 82/83) & the KXAPRPOTUS RCP bucket ladder — efficient. RCP averaged 39.8% (May 1–25) and the buckets sit exactly on that value; VoteHub is the live resolution source and the market tracks it better than a day-old web figure can. Both close tomorrow with definitional/timing finickiness (which aggregator snapshot, rounding). No defensible edge.
KXNEWTARIFFS-26MAY-JUN01 (YES 5/6) — settled state. "New/increased tariff exec action in May." With ~3 days left and the contract at 5¢, the market already encodes that it hasn't happened this month. I can't beat a realized state without inside knowledge of an imminent action — exactly the private-info edge this exercise forbids.
Reconciliation JUN14 (60¢) / JUL01 (76¢) — roughly fair. Both are the same underlying as Pick 1 but at dates where the July-4 forcing function makes passage plausible. JUL01 at ~76% is close to my own estimate; no separate edge → watchlist, not capital.
KXDNCAUTOPSY-26-JUN01 (25/30) — stale & wide. Snapshot is a week old (May 20), the spread is the full 5¢, and depth is thin. Can't size cleanly; fails the slippage test.
KXMAMDANIEO / Spencer-Pratt & Newsom endorsement ladders / pardon-count ladders — noise. Weekly EO/post counts and novelty-endorsement markets are low-information and easily gamed by single events; not a probability edge.
Pinned high-OI markets — no edge by construction. Impeachment by Jun 1 (0¢), Walz leaves governorship (3¢), Fed-chair-confirmed (100¢), "try to fire Powell" (1¢), aliens confirmed (3¢). All correctly priced at the rail; included only to show the screen worked.
Starmer JUN01 (1/2¢) — settled. He won't exit in the next four days; nothing to trade.

6 · Sources

Primary and aggregator sources used for the diligence above.

Reconciliation: Roll Call — "Reconciliation bill punted until after Memorial Day recess" (2026-05-21)
Reconciliation: CBS News — Senate punts ICE funding over DOJ anti-weaponization fund
Reconciliation: NLIHC — Members leave for Memorial Day recess without voting on "Reconciliation 2.0"
Reconciliation: Conference Board — House passes reconciliation bill; Senate next steps
Knesset: Al Jazeera — Israeli lawmakers advance bill to dissolve parliament (2026-05-20)
Knesset: The National — Israeli MPs vote to dissolve Knesset
Knesset: Haaretz — Coalition parties schedule first vote on dissolution bill (2026-05-27)
Starmer: CNN — Starmer faces critical cabinet meeting amid calls to resign (2026-05-12)
Starmer: CBS News — Starmer faces resignation calls after local-election losses
Starmer: Wikipedia — 2026 United Kingdom government crisis
Approval: RealClearPolling — Trump job-approval average
Approval: The Hill — Trump disapproval average hits 58.3% (RCP)
Approval: VoteHub — Trump approval polling average (market resolution source)
Market data: Kalshi public API — api.elections.kalshi.com/trade-api/v2 (orderbooks); local SQLite mirror (rules, snapshots).