Scan: US recession by end of 2026?

date	predicted	market price	confidence
2026-08-29	8%	8%	76%
2026-08-22	7%	8%	77%
2026-08-14	6%	8%	78%

1. Planner Fable 5.1 · Fermi decomposition + tool plan

Tools chosen: polymarket_direct kalshi_related polymarket_related fred article_search claude_news code_execution

Sub-questions (Fermi decomposition)

What are the most recent BEA real GDP quarterly growth prints for Q1 2025 through the latest available quarter, and has any quarter since Q2 2025 already printed negative (creating a 'one down' setup for the two-consecutive-quarters trigger)?
What is the current trajectory of key recession indicators — unemployment rate (Sahm rule), nonfarm payrolls, initial claims, yield curve (10y-2y / 10y-3m), ISM PMIs — and are any flashing recession-consistent readings?
What is the consensus probability of a US recession within 12 months from professional forecasters (WSJ survey, Bloomberg, Fed staff, NY Fed model, Goldman/JPM) and how does it compare to the Polymarket price?
Are there active macro shocks (tariff escalation, oil price spikes, fiscal contraction, credit events, AI capex reversal) or policy easing (Fed rate cuts, tax cuts) that materially raise or lower recession odds through end-2026?
How does Kalshi's recession market (and other Polymarket GDP/recession markets) price the same or similar questions, and is there a cross-venue disagreement?
What is the historical base rate of NBER-declared recessions beginning within any given 18–20 month window, and how often does NBER announce a recession within ~12 months of its start (relevant to the announcement-by-Q4-2026-advance-release deadline)?

Planner reasoning

This is a Polymarket macro question with two resolution paths: two consecutive negative real GDP quarters (Q2 2025–Q4 2026) or an NBER recession declaration. The key empirical drivers are the current GDP trajectory (including any negative quarters already printed), labor market/leading indicator trends, tariff/fiscal shocks, Fed policy path, and how the professional forecaster consensus and cross-market prices (Kalshi recession markets) compare to Polymarket. Base rates for a recession over an ~18-month horizon (~15-25%) plus current-conditions data are the main inputs.

2. Research Parallel tool calls

tool	status	time	items	summary
polymarket_direct	OK	0.5s	1	## This Market's Polymarket Data US recession by end of 2026? - Current price (probability): 7.50% - 7-day price change: +0.00% - 30-day price change: +0.00% - Total volume: $1,732,989 (USD notional) - Price range: 6.50% - 19.50% - Data points: 90 days
kalshi_related	OK	3.5s	1	1 related markets / summaries. series KXRECSSNBER: 0 markets (skipped 2 no-signal) \| series KXGDP: 0 markets (skipped 36 no-signal) \| series KXGDPNEG: 0 markets (skipped 0 no-signal) \| keyword 'recession': no matches \| keyword 'GDP negative': ok \| keyword 'NBER': no matches
polymarket_related	OK	0.1s	1	Scanned 100 active Polymarket markets, kept 1 matches. keyword 'recession': 0 markets \| keyword 'GDP': 0 markets \| keyword 'NBER': 1 markets \| keyword 'unemployment': 0 markets
fred	OK	5.7s	6	Fetched 6 FRED series (lookback=730d).
article_search	OK	4.0s	30	Fetched 30 unique articles across 4 queries (lookback=45d). 'US recession probability forecast 2026': 10 articles \| 'US GDP growth advance estimate BEA quarter': 10 articles \| 'economists recession odds survey': 10 articles \| 'tariffs slowing US economy recession risk': 10 articles
claude_news	OK	31.1s	24	Based on my research (data through early September 2026): GDP Growth (BEA): - Q1 2025: real GDP decreased 0.6% (final revision). In the first quarter, real GDP decreased 0.6 percent (revised). - Q2 2025: revised up substantially across estimates — final (third) estimate showed Real gross
code_execution	OK	48.9s	0	## Historical Base Rate (recession start within a ~20-month window, 1945–2024) - Gaps between the 12 recession starts ranged from 18 to 146 months (mean 77.7 mo, median 74 mo), reflecting the long 1982–1990 and 1991–2001 and 2009–2020 expansions vs. the short 1980–1981 back-to-back rece

3. Evidence Brief Sonnet 5 · 7237 chars

# Current state No quarter within the resolution window (Q2 2025–Q4 2026) has yet printed negative real GDP growth; growth has decelerated but remains positive, and NBER has made no recession announcement. The market therefore hinges on either an abrupt back-to-back negative GDP surprise in the two remaining measured quarters (Q3/Q4 2026) or an NBER declaration reaching the public before the Q4 2026 advance estimate (~late Jan 2027). # Timeline of key events - 2025-01 to 2025-04 (Q1 2025): Real GDP **-0.6%** annualized (confirmed, BEA/FRED) — outside resolution window (window starts Q2 2025). - 2025-07 (Q2 2025): Real GDP **+3.8%** (confirmed, BEA/FRED). - 2025-10 (Q3 2025): Real GDP **+4.4%** per FRED (confirmed); note claude_news cites a lower ~3.2-3.4% GDP/GDI average figure — reconciled to FRED's official real-GDP series as primary. - 2025-12: Bloomberg survey ~30% recession odds; Moody's Zandi ~42% (reported). - 2026-01 (Q4 2025... wait, actually Q4 2025 = 2025-10-01 date; Q1 2026 = 2026-01-01): Real GDP **+2.1%** (confirmed, FRED). Note: FRED's 2025-10-01 obs (+0.5%) corresponds to Q4 2025 — slower but still positive. - 2026-03: Iran war triggers oil shock (Brent ~$71→~$101-110), recession-odds forecasts spike across Wall Street (Goldman 30%, Zandi 49%) (reported). - 2026-03: NY Fed DSGE model recession probability 35.8% (confirmed, Liberty Street Economics). - 2026-04 (Q2 2026): Real GDP **+1.5%** (second estimate, confirmed, FRED). - 2026-07: WSJ Economic Forecast Survey — 25% avg 12-month recession probability, down from 33% in April (confirmed). - 2026-08: Unemployment stable 4.1-4.4%, initial claims 189k-227k range (confirmed, FRED); NY Fed yield-curve model recession odds eased to ~18.8% from a ~32% November 2025 peak (reported). - 2026-09-04: August payrolls +162k (vs. 53k consensus), unemployment steady 4.1% (confirmed, CNBC/BLS); Fed officials signaling more concern about inflation than labor market, markets pricing possible hike (reported). # Event Will the US enter a recession (per NBER declaration or two consecutive negative real-GDP quarters between Q2 2025–Q4 2026) by the time BEA releases the Q4 2026 advance GDP estimate (~late Jan 2027)? # Outcomes to forecast Yes / No # Kalshi market anchor No kalshi_direct price was returned; the provided ticker is a hex-format identifier matching Polymarket's schema, not a Kalshi ticker. Best available direct-market anchor: **Polymarket "US recession by end of 2026?" = 7.5% YES**, flat over 7d and 30d, range 6.5%-19.5% over 90 days (high volume, $1.73M total). Kalshi-related searches (KXRECSSNBER, KXGDP, KXGDPNEG) returned no matching live market for direct comparison. # Sub-question answers 1. **GDP prints since Q2 2025 / one-down setup?** No negative quarter in-window yet: Q2'25 +3.8%, Q3'25 +4.4%, Q4'25 +0.5-0.7%, Q1'26 +2.1%, Q2'26 +1.5% (FRED, confirmed). No "one down" trigger exists; both Q3 and Q4 2026 would need to print negative. 2. **Recession indicators?** Unemployment stable 4.1-4.5% (no Sahm-rule trigger), payrolls positive (Aug 2026 +162k), initial claims low/stable (189k-227k), 10y-2y spread positive (~0.40-0.50, not inverted). No indicator currently flashing recession (FRED, confirmed). 3. **Forecaster consensus vs. Polymarket?** WSJ survey 25% (Jul 2026, down from 33%), Bloomberg 30% (Dec 2025), JPMorgan 35% (Jan 2026), Moody's/Zandi 42-49% (peaked during Iran shock), NY Fed model 18.8% (eased from 32% peak). These 12-month "any recession" odds (~20-40%) run notably higher than Polymarket's 7.5%, partly because Polymarket prices the narrower GDP/NBER trigger with a now-shortened remaining window. 4. **Active macro shocks/easing?** Iran-war oil shock (Q1 2026, Brent to ~$110) temporarily raised recession odds sharply; by late summer 2026 this faded, and Fed is now leaning toward inflation-fighting (possible hike) rather than recession-fighting cuts (CNBC, reported). 5. **Cross-venue signals?** No Kalshi-direct equivalent found; Kalshi's long-dated GDP-growth-bucket markets (2035/2036) show modest ~9-35% probabilities on unrelated horizons, not directly comparable. Polymarket-related search found no other matching recession/GDP markets. 6. **Historical base rate?** ~23-25% empirical/Poisson base rate for a recession starting in any ~20-month window; NBER announcement lag averages 7.3 months (range 4-12), meaning a recession starting after mid-2026 has a sharply declining (<50%, often <15%) chance of NBER confirmation before the Jan 2027 deadline (code_execution analysis). # Key facts (high-confidence, factual) 1. [FRED] Real GDP growth in-window: Q2'25 +3.8%, Q3'25 +4.4%, Q4'25 +0.5%, Q1'26 +2.1%, Q2'26 +1.5% — no negative quarter yet. 2. [FRED] Unemployment 4.1-4.5%, initial claims 189k-227k — no acute deterioration. 3. [FRED] 10y-2y spread positive ~0.40-0.50, not inverted as of Sep 2026. 4. [Polymarket] Current YES price 7.5%, flat 7d/30d, 90-day range 6.5-19.5%. 5. [code_execution] NBER announcement lag mean 7.3 months; recessions starting after ~mid-2026 unlikely to be confirmed before resolution deadline. # Cross-market signals - Kalshi related: no matching recession/GDP-negative market found; unrelated long-horizon GDP-bucket markets only. - Polymarket: 7.5% YES, high liquidity ($1.73M volume), stable recently — reflects narrowed remaining resolution window with no negative print yet. - Sportsbook implied: n/a. # Analyst opinions and speculation - Forecaster 12-month recession odds cluster 18-35% (WSJ, JPM, NY Fed, Bloomberg), higher than Polymarket's 7.5%, likely because these surveys measure broader/rolling 12-month risk vs. this market's specific narrowing GDP/NBER trigger. - Moody's Zandi remains the outlier bear (42-49%, tied to oil-shock scenario), a minority view not shared by consensus. - Fed commentary (Aug-Sep 2026) leans toward inflation concern over recession risk, implying no imminent easing-driven recession signal. # Directional lean per outcome - **Yes**: supported by cooling growth trend (Q4'25 low of 0.5%), historical base rate (~23-25%), forecaster surveys (18-35%), and residual tail risk from oil/tariff shocks; opposed by two consecutive positive-growth quarters most recently (Q1/Q2 2026 +2.1%/+1.5%), stable labor market, and narrowing time window with NBER lag risk. - **No**: supported by Polymarket's own 7.5% pricing, all in-window GDP quarters positive to date, stable unemployment/claims, non-inverted yield curve, and mechanical NBER-lag issue reducing effective "Yes" probability even if a late-2026 downturn began. # Gaps / unknowns - No true Kalshi-direct price was retrieved for this specific ticker; relying on Polymarket as best proxy anchor. - Q3 2026 GDP (not yet released) is the critical unknown data point. - Discrepancy between FRED (Q3'25 +4.4%) and claude_news narrative (~3.2-3.4% GDP/GDI average) — resolved in favor of FRED's official real-GDP series. # Calibration anchors - Polymarket current YES price: 7.5% (anchor, flat trend). - Historical base rate for a recession starting within ~20-month window: ~23-25%; blended indicator-adjusted estimate 30-38% before NBER-lag discount. - Forecaster consensus 12-month recession odds: ~20-30% (WSJ 25%, NY Fed 18.8%, Bloomberg/JPM 30-35%).

4. Forecasters Fable 5.1 + GPT-5.6 Sol · in parallel

claude-fable-5-1

68% conf

Yes	6%
No	94%

Resolution now requires either both Q3 and Q4 2026 real GDP to print negative or an NBER declaration before ~late Jan 2027; every in-window quarter so far has been positive (Q1'26 +2.1%, Q2'26 +1.5%) and no 'one-down' setup exists. Coincident indicators are not flashing recession: August payrolls +162k, unemployment steady at 4.1%, claims 189k-227k, 10y-2y positive, and the Fed is leaning toward inflation concern rather than easing. The NBER lag (mean ~7.3 months) makes a declaration before the deadline nearly impossible for any downturn that began in H2 2026. Polymarket's 7.5% YES (flat, liquid) is the best anchor; higher 12-month survey odds (18-35%) measure a broader, rolling risk and are not directly comparable. I shade slightly below the market given the strong labor data and mechanical constraints.

gpt-5.6-sol

78% conf

Yes	7%
No	93%

The GDP evidence points toward No because every in-window quarter through Q2 2026 was positive, so both Q3 and Q4 would now need to contract for the mechanical trigger. Labor-market data, low claims, and a positive yield-curve spread also point toward No because none signals an imminent recession, while the earlier oil shock has faded. Although 12-month forecaster estimates of 18-35% and the historical base rate preserve meaningful tail risk, they cover broader horizons than this narrowing resolution window, and typical NBER announcement lags make a late-2026 recession unlikely to qualify in time. I therefore stay close to the best direct-market anchor, Polymarket's 7.5% YES price, with a slight downward adjustment to 7%.

5. Devil's Advocate Sonnet 5 · red-team critique

Both forecasts anchor almost entirely on Polymarket's 7.5% and land within 1-1.5pp of it (6.0% and 7.0%), effectively deferring to the market rather than independently reasoning from the sub-question evidence—neither explains why they don't weight the base rate (23-25%) or survey consensus (18-35%) more heavily given these come from more directly relevant, higher-n sources than a single prediction market with unclear depth/composition.

Neither forecast grapples with the discrepancy noted in the brief itself: "these surveys measure broader/rolling 12-month risk vs. this market's specific narrowing GDP/NBER trigger" is asserted but not quantified—how much of the 18-35% consensus should mechanically collapse once you condition on zero negative quarters through Q2'26? Without that math, both forecasts risk anchoring bias rather than genuine reconciliation.

Neither forecast discusses the Q3 2026 GDP release explicitly as a near-term binary catalyst still pending at the time of forecast (the brief flags this as "the critical unknown data point")—if Q3 comes in negative (even modestly), the "two consecutive negative quarters" path suddenly becomes live for Q4, which could reprice this market sharply; both forecasts treat the mechanical trigger as effectively dead without addressing this live risk.

Both forecasts nearly ignore the Fed's hawkish pivot (leaning toward a hike over cuts per Sep 2026 commentary) as a source of forward recession risk—tightening into decelerating growth (Q4'25 dropped to +0.5%) is a classic recession-inducing pattern, yet neither treats this as raising rather than lowering forward-looking risk.

Neither forecast addresses possible Kalshi/Polymarket divergence or thin-market risk explicitly: the brief notes no genuine Kalshi-direct price was found and the Polymarket range (6.5%-19.5% over 90 days) shows meaningful historical volatility, which should widen uncertainty bands rather than support two forecasts converging so tightly (6-7%) with high stated confidence (0.68-0.78).

Both forecasts underweight tail/shock scenarios—the brief explicitly flags "residual tail risk from oil/tariff shocks" (the Iran war already caused one spike to 49% Zandi odds within the window), yet neither forecast meaningfully reserves probability mass for a second unforeseen shock in the remaining ~4 months before the Q3/Q4 GDP prints.

6. Aggregator Fable 5.1 · final calibrated forecast

6% (73% confidence)

Yes: 6% No: 94%

With every in-window quarter through Q2 2026 positive, resolution now requires both Q3 and Q4 2026 to print negative (or an NBER declaration before late Jan 2027). Coincident data for Q3 (Aug payrolls +162k, claims 189k-227k, unemployment 4.1%, non-inverted curve) make a negative Q3 print unlikely (~8-10%), and even then Q4 would also need to contract, giving a mechanical-trigger path of roughly 3-4%; the NBER path adds only ~1% given the 7+ month announcement lag. The critique's points about base rates and 12-month surveys are valid in spirit but those measure broader rolling risk not conditioned on the now-shortened window with no negative print, so I don't move materially toward them. I stay near the liquid Polymarket anchor (7.5%) with a slight shade lower, consistent with both forecasters.

Pipeline Timing

Total pipeline time: 171.5s

Per-tool research timings shown in the Research section above.