Kalshi Economics — Mispricing Audit

Report date 2026-06-06 · Category Economics · Horizon 45 days · Capital $1,000

A deep screen of every active Kalshi Economics market closing before 21 Jul 2026. The headline finding is uncomfortable but honest: this slate is mostly efficient. The macro-data markets here track public releases (CPI, PPI, payrolls, Treasury yields, AAA gas) that sharp traders price tightly. After diligence I deploy capital on one core edge and one small asymmetric lottery, and hold a deliberately large cash reserve rather than force money into fairly-priced contracts.

Bottom line: BUY NO on KXCPIYOY-26MAY-T4.2 (May CPI >4.2%) — the Cleveland Fed nowcast (4.18% YoY) points to a 4.2% print, but this contract only pays YES on a 4.3%+ print, and it trades at 48¢. Plus a capped-downside BUY NO lottery on the $4.16 gas leg. ~$528 deployed, ~$472 held in reserve.

1 · How this was researched

  1. Universe. Queried the Kalshi DB mirror for all status='active' markets whose event category='Economics' closing within 45 days — 400+ markets, mostly micro-strike ladders (gas, CPI, Treasury, housing).
  2. Liquidity screen. Joined the latest market_snapshots row per market and cut everything with volume24h < 1500, bid-ask spread > 5¢, or already pinned (YES bid ≥95 / ask ≤5). That left 26 candidates.
  3. Rules first. Pulled the full rules text for each. Two definitional traps surfaced immediately: (a) CPI markets resolve on the one-decimal BLS print, so ">4.2%" actually needs a 4.3% print; (b) some Treasury legs say "before month-end" (a touch / high-water-mark) while others say "at month-end" (a level).
  4. Live state. Re-pulled top-of-book and depth from the Kalshi public API at audit time (snapshots were ≤30 min stale) and confirmed fills.
  5. News diligence. Grounded every prior in a primary source: AAA's national gas average, the Cleveland Fed inflation nowcast, the BLS April CPI/PPI releases, the Fed H.15 yield release, Bloomberg/CNBC on the SpaceX IPO, and CNBC on the Iran-war oil move. Sources are linked in §6.
  6. Reject hard. For each survivor: are the rules unambiguous? is my evidence public? has price already moved on the thesis? can I fill without ≥3¢ slippage? Most candidates failed one of these.

2 · Markets reviewed

26 liquidity-screened candidates. Prices are live Kalshi top-of-book at audit time (¢). "Verdict" links to the pick or the reject rationale.

TickerMarketYES bid/askVol24hClosesVerdict
KXCPIYOY-26MAY-T4.2May CPI YoY >4.2% (needs 4.3% print)46 / 505,691Jun 10PICK 1 · BUY NO
KXAAAGASW-26JUN08-4.160AAA US gas >$4.160 on Jun 892 / 9315,054Jun 8PICK 2 · BUY NO (lotto)
KXCPIYOY-26MAY-T4.1May CPI YoY >4.1% (needs 4.2% print)89 / 904,115Jun 10PASS — hot tape backs YES
KXCPI-26MAY-T0.5May CPI MoM >0.5%29 / 307,170Jun 10PASS — ~fair vs nowcast
KXAAAGASW-26JUN08-4.170AAA US gas >$4.170 on Jun 851 / 5314,776Jun 8PASS — priced on the landing
KXAAAGASW-26JUN08-4.180AAA US gas >$4.180 on Jun 823 / 2516,293Jun 8PASS — priced on the landing
KXMUSKNW-26JUN30-T790Musk net worth >$790B (Bloomberg)88 / 915,188Jun 30WATCH — SpaceX IPO ≈ fair
KXUSPPIYOY-26JUN11-T6.0May PPI YoY >6.0%49 / 511,925Jun 11PASS — genuine coin-flip
KXUSGASCPI-26JUN10-T378May CPI gasoline index >37818 / 195,316Jun 10PASS — no clean index map
KXUSGASCPI-26JUN10-T377May CPI gasoline index >37735 / 401,842Jun 10PASS — 5¢ spread
KXUST2-26JUN30-T4.202Y yield >4.20% before month-end78 / 792,528Jun 30PASS — touch market, fair
KXUSTM-26JUN30-T4.002Y yield >4.00% at month-end92 / 933,036Jun 30PASS — ~fair (spot 4.05)
KXUST10-26JUN30-T4.6510Y yield >4.65% before month-end43 / 461,539Jun 30PASS — touch market
KXUST5M-26JUN30-T4.355Y yield >4.35% at month-end28 / 291,640Jun 30PASS — ~fair (spot 4.18)
KXUST30M-26JUN30-T4.9530Y yield >4.95% at month-end69 / 702,446Jun 30PASS — ~fair (spot 4.97)
KXUST30M-26JUN30-T5.2530Y yield >5.25% at month-end9 / 101,610Jun 30PASS — ~fair tail
KXCBDECISIONBRAZIL-26JUN17-HOLDBrazil COPOM holds Selic Jun 1757 / 583,226Jun 17PASS — real coin-flip
KXBRAZILINF-26JUN-T5.00Brazil inflation >5.00% in Jun13 / 141,826Jul 10PASS — ~fair tail
KXHOUSINGSTART-26JUN16-T1.400May housing starts >1.400M86 / 871,861Jun 16PASS — no edge
KXHOUSINGSTART-26JUN16-T1.450May housing starts >1.450M6 / 72,317Jun 16PASS — ~fair tail
KXPAYROLLS-26JUN-T40000June payrolls >40k75 / 771,730Jul 2PASS — unobserved, forecast
KXECONSTATU3-26JUN-T4.3June unemployment exactly 4.3%26 / 272,213Jul 2PASS — ~fair bucket

(Also screened and cut: the same-day AAA daily gas ladder closing Jun 6, already effectively resolved.)

3 · The picks

Pick 1 — KXCPIYOY-26MAY-T4.2 · BUY NO @ 0.55 · Medium Conviction

Market: Will May 2026 CPI inflation be above 4.2% YoY? NO entry: 54–55¢ Resolves: Jun 10, 2026 (BLS 8:30am ET)

The mispricing. The contract title says "above 4.2%," but the rule resolves on the one-decimal value reported by BLS — so YES requires a printed 4.3% or higher. The Cleveland Fed's inflation nowcast — the single best real-time predictor, because it already ingests the daily oil and weekly retail-gas data that drive this print — sits at 4.18% YoY, which rounds to a 4.2% headline. A 4.2% print resolves this market NO. The market nonetheless prices YES at 48¢, implying a ~48% chance of a 4.3%+ print, i.e. a center near 4.23–4.25%. That is hotter than the nowcast by 0.05–0.07pp.

Why NO is the value side. Centering reported May YoY on the 4.18% nowcast (modal print 4.2%), a 4.3% print needs the actual to land ≥4.25% — roughly a 0.07pp upside miss versus the best model. Even widening the error band and crediting the current hot-surprise regime, P(4.3%+) lands ~25–35%, not 48%. NO bought at 54¢ is therefore worth ~65–75¢.

Tail risk (the clean one): a hot energy surprise. April CPI beat consensus and April PPI beat hard (6.0% vs 4.8% expected) — in this oil-shock regime, inflation prints have been surprising up. A single 4.3% headline (actual ≥4.25%) loses the whole position. That upside skew — plus my inability to load the live nowcast page directly (cited value is from mirrors) — is exactly why this is sized as a medium-conviction core, not a max bet.

True P(YES, 4.3%+) ≈ 28% NO fair ≈ 72¢ NO price 54¢ Edge ≈ +18¢ EV ≈ +33%/contract

Liquidity / fills: NO buyable at 54¢ (≈290 contracts), 55¢ (≈1,000), 57¢ (≈500) — a few-hundred-dollar clip fills inside 1¢ of the 55¢ limit. Price history: last 48¢, traded a tight 46–54¢ band over the past 24h on ~5.7k contracts; OI ~34.9k. No 48h thesis-moving jump.

Pick 2 — KXAAAGASW-26JUN08-4.160 · BUY NO @ 0.10 · Low Conviction · Asymmetric Lottery

Market: Will AAA US gas be above $4.160 on Jun 8? NO entry: 8–10¢ Resolves: Jun 8, 2026

The setup. AAA's national average is $4.220 (Jun 5 report) and falling steadily: $4.241 (Jun 4), $4.391 a week ago, $4.483 a month ago — a remarkably linear ~2.0–2.4¢/day decline, driven by WTI down ~20% on ceasefire optimism. A simple extrapolation of the print that resolves "on Jun 8" (which reflects ~Jun 7 collection) lands at roughly $4.15–$4.17 — straddling the $4.160 strike. Yet the market prices YES (gas stays >$4.16) at 92%, leaving NO at just .

Why a small NO. This is not a claim that the market is dumb — these AAA ladders are traded by sharp, gas-watching desks. It is a pure asymmetry play: downside is capped at ~10¢, while my landing model puts a realistic ~25–35% chance the Jun 8 figure prints ≤$4.16 as the decline grinds into the strike. At a ~10¢ entry that is materially +EV even if I'm fading smart money, because an 8% NO probability is hard to defend when the underlying is sitting just 6¢ above the line and dropping 2¢/day.

Tail risk (the clean one): resolution timing. If "on Jun 8" is read off an earlier daily print (≈$4.18–4.20) rather than the freshest one, $4.16 is cleared comfortably and NO loses — and that ambiguity is the most likely reason the market sits at 92%. A renewed oil spike (ceasefire collapse) would also halt the decline. Hence: small size, capped loss, lottery label.

True P(NO, ≤$4.16) ≈ 30% NO fair ≈ 30¢ NO price ≈ 9–10¢ Edge ≈ +20¢ EV ≈ +200% (high variance)

Liquidity / fills: NO thin at the top — ~21 @8¢, ~130 @9¢, ~504 @10¢; a ~$60 clip fills near a 10¢ average. Price history: YES firmed from ~88¢ to 92–93¢ over the evening on heavy volume (15k/24h) — the market is leaning into "gas holds above $4.16," which is the signal I am deliberately fading in small size.

4 · Recommended $1,000 portfolio

One steady core, one capped lottery, and a large deliberate cash reserve. I am not going to manufacture diversification by deploying into the dozen fairly-priced contracts above — that would be paying the spread for negative edge.

PickSideLimitContractsCostMax payoutEV ¢/ctEV $Conv.
KXCPIYOY-26MAY-T4.2BUY NO 0.55850$467$850+17+$145Med
KXAAAGASW-26JUN08-4.160BUY NO 0.10600$60$600+20+$120Low
Total deployed$527$1,450+$265
Cash reserve$473

Blended EV on deployed capital ≈ +50%, but that figure is flattered by the high-variance gas lottery; the steadier read is the CPI core at ~+31% EV. Dollar edge ≈ +$265 expected on $527 at risk — with a wide outcome distribution.

Risk profile

Execution notes

5 · What I rejected, and why

The rejects are the point — they show the screen worked. Five instructive misses:

Treasury "mispricings" that were definitional traps

At a glance, KXUST2-26JUN30-T4.20 ("2Y >4.20%") at 78% looks absurd: the 2Y closed at 4.05% on the Jun 4 H.15. But the rule says "before month-end," making it a touch / high-water-mark market — it pays YES if the 2Y is ever above 4.20% during June. From 4.05% with normal volatility, a +15bp touch over 24 days is genuinely ~70–80% likely. The "at/for month-end" legs (KXUSTM 2Y>4.00 @92, KXUST5M 5Y>4.35 @28, KXUST30M 30Y>4.95 @69) are all consistent with the live curve (2Y 4.05 / 5Y 4.18 / 10Y 4.47 / 30Y 4.97). No edge — just two different contract mechanics that look contradictory until you read them. Fed H.15

The rest of the gas ladder — priced right on the landing

With AAA at $4.220 and falling ~2¢/day, the Jun 8 print lands near $4.17. The ladder reflects exactly that: >$4.17 at a coin-flip (51¢) and >$4.18 at 23¢. Those are fair-to-the-landing, and the resolution is knife-edge sensitive to which day's figure is used — a reason to avoid, not bet. Only the far-from-landing $4.16 leg offered the capped-downside asymmetry worth a small stake.

Musk >$790B — a real catalyst, fairly priced

KXMUSKNW-26JUN30-T790 trades 88/91. The catalyst is concrete: SpaceX is set to debut on Nasdaq ~Jun 12 at $135/share (~$1.77T), which by Bloomberg's own math marks Musk to ~$988B — ~$200B above the strike. Build it up: P(IPO debuts by Jun 30) ~90% × P(stays >$790B | listed) ~97% ≈ ~88%. That is the market. Buying YES at the 91¢ ask is negative edge; an 88¢ limit is roughly zero. Watchlist, not a deploy. CNBC, Bloomberg

May CPI >4.1% — fighting a hot tape

The sister leg KXCPIYOY-26MAY-T4.1 (YES needs a 4.2% print) trades 89%. With the nowcast at 4.18% → modal 4.2%, that's broadly right, and the upside-surprise regime (April CPI and PPI both beat) backs it further. Betting NO here at 11¢ means betting on a soft 4.1% print in a month where energy ran hot — fighting the tape. Pass.

Brazil COPOM, PPI, payrolls, housing, unemployment — honest coin-flips

KXCBDECISIONBRAZIL-26JUN17-HOLD (57%) is a genuine hold-vs-cut toss-up after the BCB began easing — and the public rate path I could assemble was too muddled to claim an edge. KXUSPPIYOY-26JUN11-T6.0 (49/51) is a true coin-flip on whether May PPI holds April's 6.0%. June payrolls, housing starts, and the "exactly 4.3%" unemployment bucket are unobserved forecasts with no public edge. Forcing a view on any of these would be storytelling, not trading.

6 · Sources