Kalshi maintains a ‘perfect forecast record’ in predicting Fed rate decisions, beating professional forecasters, study finds

An NBER working paper finds Kalshi’s prediction-market forecasts for FOMC rate decisions (modal and median) were as accurate as Wall Street and outperformed fed funds futures from 2022 through June, with a perfect modal track record in that period; the platform notably anticipated the September 2024 50-basis-point move that other forecasters missed. Kalshi’s inflation and unemployment forecasts are statistically similar to Bloomberg consensus, and a 2025 Kalshi study reported a 40.1% lower mean absolute error on inflation shocks versus consensus; prediction markets’ annualized volume reportedly rose from $300m to roughly $40–50bn since August 2025. The article notes operational advantages (real-time updates) and growing liquidity but flags model risks after an alleged $400k insider-bet on Polymarket; on Kalshi, traders were 99% certain the Fed would hold in January, though longer-term rate paths remain split.

Analysis

Market structure: Real‑time prediction venues like Kalshi become information monopolists for short‑dated Fed expectations—beneficiaries include exchange/data vendors, HFT/prop desks and sell‑side quant teams that ingest minute‑by‑minute probabilities. Losers are slower signal providers (six‑week surveys, consensus polling) and any desk that pays a premium for outdated fed‑funds futures signals; expect a re‑pricing of front‑end liquidity and narrower risk premia in 0–18 month rate products. Cross‑asset impact will concentrate on short rates, front‑end swap spreads and USD; a clearer Fed path should compress short‑dated rate vols and can transiently steepen the curve if markets price later cuts with more conviction.

Risk assessment: Key tail risks are regulatory clampdowns (CFTC/SEC restricting event classes), high‑profile insider trades or manipulation that trigger litigation, and liquidity blackouts from concentrated players — any of which could invert the signal’s value. Near term (days–weeks) the reliability is high for next meeting calls; medium (3–12 months) depends on liquidity growth and composition (retail vs institutional); long term (1–3 years) is regulatory and market structure dependent. Hidden dependencies: predictive accuracy degrades if participation concentrates (whales) or if markets become self‑fulfilling and amplify Fed reaction functions.

Trade implications: Use Kalshi probabilities as a trigger overlay for front‑end rate trades and vol trades rather than a lone signal. Tactical plays: size small directional 2y positions ahead of shifts in implied cut probability and sell short‑dated rate vol when Kalshi consensus >90% for “hold”; buy protection when sudden skew toward cuts arises. Broader: data/exchange providers (CME, ICE) should outperform pure retail platforms if commercialisation continues; consider exposure with regulatory hedges.

Contrarian angles: Consensus overweights Kalshi’s accuracy and underweights manipulation/regulation risk — the market may be underpricing a regulatory shock (probability 10–25% over 12 months). Historical parallel: when new price discovery venues (e.g., electronic FX in 2000s) scaled, volatility fell until concentrated liquidity events reversed it; expect intermittent violent repricings. Unintended consequence: herd trading from a single venue can make Fed reaction function more sensitive to that venue, increasing systemic fragility.

Key Decisions for Investors

• Establish a 1.0–2.0% NAV tactical long in 2‑year Treasury futures (CME ZT) if Kalshi’s implied probability of a 25bp cut before the next FOMC rises above 20% within a rolling 14‑day window; target a 20–30bp yield rally, hard stop at +10bp adverse move.
• When Kalshi shows >90% probability of a ‘hold’ for the next meeting, sell 30‑day ATM front‑end rate straddles (options on 2y note futures or 1‑month SOFR caps) sized to 0.5–1.0% NAV and hedge with 10–15bp OTM wings to cap tail risk; unwind if Kalshi cut probability moves above 25%.
• Initiate a 1.0% NAV long in exchange/data plays (CME: CME, ICE: ICE) over 6–12 months to capture incremental trading/data licensing revenue from prediction markets; take profits at +10–15% or if regulatory comment letters indicate material restrictions.
• Avoid or trim (‑50%) direct equity exposure to retail crypto/prediction platforms (e.g., COIN) until CFTC/SEC provide clarity; monitor for CFTC rulemaking and DOJ insider‑trading actions over the next 30–60 days and re‑rate positions if enforcement probability rises above 20%.