US Recession Probability: 12-Month Outlook
May 5, 2026
Prediction markets price a 25-26% probability of an NBER-defined recession by May 2027, anchored to resilient near-term growth but vulnerable to geopolitical shocks and late-cycle risks. The yield curve is steepening (10Y-2Y at +50bp), credit spreads remain tight, and labor data is stable, all supportive signals; however, Q1 actual GDP missed expectations and Goldman Sachs flags a 30% 12-month recession risk, signaling material downside tail risk.
Key takeaways
- Polymarket consensus 25% vs. Kalshi 32.4%: prediction markets diverge by ~7pp, with Polymarket's lower read reflecting recent soft-landing optimism and Kalshi's higher tail suggesting tail-risk hedgers are active 12.
- Yield curve steepening is recession-bearish: 10Y-2Y spread at +50bp (P66 vs. 5y history) 3, well above the inversion threshold that historically preceded downturns; short-end (Fed funds at 3.64% 4) and long-end (10Y at 4.45% 5) both stable, pricing zero rate cuts through September and modest hikes into December 4.
- Labor market firm but cooling: unemployment at 4.3% (March 2026, 65 days old) 3, initial jobless claims at 189k (April 25, lowest since 2020 on 5y percentile) 3, both inconsistent with imminent contraction.
- Q1 GDP revised down to 1.2% but Q2 nowcast holds at 3.5%: Atlanta Fed GDPNow as of May 1 estimates Q2 at 3.5% 6; the Q1 miss (from 2.3% to 1.2%) signals deceleration mid-cycle but not a recessionary trough; Goldman Sachs projects H2 2026 growth at 1.25%-1.75%, below trend 6.
- Credit spreads tight, default risk muted: high-yield OAS at 277bp (latest 2.77% 3), down 12% over 30 days and P11 over 5y, indicating low near-term credit stress; investment-grade OAS at 81bp (81bp 3), also near 5y lows.
- Geopolitical tail risk priced: Polymarket odds for U.S. invasion of Iran by end-2026 at 30.5% 1, which, if realized, could trigger supply shocks and financial stress; Strait of Hormuz traffic normalization by May 15 priced at 2.9%, consistent with high geopolitical friction 1.
- Inflation expectations stable at 2.5%: 10Y breakeven inflation at 2.5% (P87 over 5y) 3, suggesting Fed's 2% target is credible and long-end real rates are pricing sustained non-recessionary conditions.
Signal table
| Signal | Value | Horizon | Note |
|---|---|---|---|
| Polymarket "US recession by end-2026" 1 | 25% | by Dec 31, 2026 | resolves on NBER call or 2 consecutive negative-real-GDP quarters |
| Kalshi recession probability 2 | 32.4% | by end-2026 | higher tail-risk premium vs. Polymarket |
| Goldman Sachs 12-month recession odds 6 | 30% | next 12 months | up from 25%, citing energy/financial conditions tightness |
| 10Y-2Y spread 3 | 50bp | spot (2026-05-04) | P66 historical; no inversion signal |
| Fed funds target mid-point 4 | 3.625% | spot | ZQ futures +11bp cumulative by Dec FOMC; zero cuts priced |
| Unemployment rate 3 | 4.3% | March 2026 (65d lag) | P73 over 5y; steady, no recessionary spike |
| Initial jobless claims 3 | 189k | April 25 (10d lag) | P0 over 5y (lowest); labor market resilient |
| Atlanta Fed GDPNow Q2 2026 6 | 3.5% | for Q2 2026 | down from 3.7%, but above trend; reflects ISM miss |
| Q1 2026 actual GDP (revised) 6 | 1.2% | Q1 2026 | revised down from 2.3% early estimate; mid-cycle deceleration |
| High-yield OAS 3 | 277bp | May 1 | down 12% in 30d; low default stress |
| 10Y breakeven inflation 3 | 2.5% | spot (2026-05-04) | P87 over 5y; Fed credibility intact |
| US-Iran invasion probability 1 | 30.5% | by Dec 31, 2026 | tail geopolitical risk; could trigger oil shock |
Cross-check
Polymarket (25%) and Kalshi (32.4%) diverge by 7 percentage points, with prediction-market consensus roughly 3pp below Goldman Sachs' 30% estimate. The gap likely reflects (i) Kalshi's structural design penalizing tail-event hedgers, and (ii) GS's incorporation of lagged deterioration in early-2026 jobs/credit data that has since stabilized. The yield curve's steepening (+50bp) and credit spreads' tightness both argue against imminent recession (they would be inverted and widening if one were truly imminent), yet the Q1 GDP miss and late-cycle positioning (unemployment low, growth decelerating, Fed holding) create a classic soft-landing vs. hard-landing fork. The 25-26% market price reflects roughly 3-to-1 odds favoring the soft landing, with meaningful tail risk if geopolitical escalation (Iran) or credit stress emerges.
Caveats
- Q1 GDP surprise was large (revised from 2.3% to 1.2%), raising model uncertainty; Atlanta Fed GDPNow carries a 1.17pp root-mean-squared error, so Q2's 3.5% nowcast has a ±1.2pp band and is not fixed.
- Unemployment data lags by ~65 days (latest March 2026 print as of May 5); monthly jobs report for April 2026 has not yet resolved and is flagged as a near-term catalyst 2.
- High-yield credit spreads are at 5-year lows, not fully reflecting tail scenarios (Iran conflict, geopolitical supply shock); sudden oil-price spikes could widen spreads sharply and flip recession odds.
- Prediction-market liquidity on ultra-long-duration contracts (e.g., recession by May 2027) is structurally thinner than near-term events, so the quoted 25-26% carries wider confidence bands than near-term FOMC odds.
References
Model-derived probabilities anchored to current data; not investment advice. Past base rates and current market-implied probabilities do not guarantee future outcomes.