US recession by end of 2026?

Nearly every verifiable quantitative anchor for a 2026 US recession clusters between roughly 15% and 25%, above Polymarket's 10.5% yes price. That includes the post 1948 historical base rate, the NY Fed yield curve model, the Philadelphia Fed's Survey of Professional Forecasters, the Wall Street Journal's July 2026 economist poll, and a comparably defined Kalshi market. The gap has narrowed sharply since spring, economist recession odds ran as high as 33% to 49% in March and April 2026 during an oil price and Iran war shock, and have come down as Q1 2026 GDP printed a positive 2.0%, unemployment held near 4.2% to 4.4%, and the yield curve stayed positively sloped (10y minus 3m about plus 0.77 points as of July 14, 2026). Even so, the most recent, most directly comparable readings found (WSJ survey 25%, Kalshi about 17% to 18%, NY Fed model about 15%) still sit meaningfully above the market price, suggesting 10.5% may be priced on the low side of consensus rather than in line with it. Part of the gap is likely structural, not mispricing: Polymarket's rule needs either two consecutive negative BEA GDP quarters or a formal NBER declaration by the time Q4 2026 GDP is published, a stricter, slower trigger than open ended "will there be a recession" survey questions. This is a comparison of price to available evidence, not a prediction, and not investment advice.

• Historical base rate: the US has had 11 recessions since 1948, about one every six years, implying a roughly 15% to 18% unconditional chance of a new recession starting in any given year, close to but above the market's 10.5%.
• The NY Fed yield curve model puts 12 month ahead recession probability near 15% as of mid July 2026, based on a 10 year minus 3 month Treasury spread of about plus 0.77 percentage points, well below the roughly 30% threshold that has preceded every recession since 1969.
• Philadelphia Fed's Q2 2026 Survey of Professional Forecasters put the probability of a GDP contraction in the current quarter at 17.9%, and forecasters see contraction risk rising to roughly 23% to 26% for each quarter from Q3 2026 through Q1 2027.
• The Wall Street Journal's July 2026 survey of 72 economists, taken July 2 to 7, put average 12 month recession odds at 25%, down from 33% in April but still well above Polymarket's implied price.
• Kalshi's comparably worded market (BEA two quarter negative GDP definition) fell from a 36.9% high in March, during the Iran war oil price spike, to roughly 17% to 18% recently, still notably above Polymarket's 10.5% for a similar event.
• This Polymarket market has itself been volatile: about 31% in December 2025, a spike above 30% in April 2026 on roughly 13 million dollars of volume during the tariff and Iran war oil shock, then a decline to 10.5% now. No quarter since mid 2025 has shown negative GDP growth, so the market's built in two consecutive negative quarter trigger has not yet been armed.

Polymarket's "US recession by end of 2026" market has priced YES down to about 10.5% from roughly 28-30% in April 2026, as Q1 GDP was revised up to 2.1% and the labor market avoided an outright break. The mechanical recession signals are mixed rather than uniformly reassuring: the yield curve has un-inverted (10s2s near +0.35 percentage points) but analysts attribute that mainly to a rising term premium, not growth optimism, and the Sahm rule, which called all 11 US recessions since 1950, does not look triggered given unemployment's narrow 4.2 to 4.3% range this year. The historical matchup that stands out is oil shocks: 10 of the last 11 postwar US recessions were preceded by a sharp oil price spike, and the 2026 Iran conflict's disruption of Strait of Hormuz traffic pushed Brent sharply higher and lifted CPI to 4.2% year over year in May, a combination of oil-driven inflation and a Fed holding or leaning toward hikes rather than cuts that has historically preceded recessions more often than not. Consensus forecasters and markets also have a poor recent record calling recessions correctly in advance, since about 70% of economists expected one within 12 months as of late 2022 and none arrived in 2023, and NBER itself typically takes 6 to 21 months to formally declare a recession once it has started, so today's roughly 10% price is a snapshot of sentiment, not a resolved fact, and could still be wrong given how long official confirmation lags real conditions. This is a record of how comparable setups have resolved historically, not a prediction or financial advice.

• Polymarket YES on US recession by end of 2026 fell from about 28-30% in April 2026 to roughly 10.5% by mid-July 2026, tracking stronger growth data and a Fed that has signaled no 2026 rate cuts.
• Q1 2026 real GDP was revised up to 2.1% annualized, but the Atlanta Fed's GDPNow estimate for Q2 2026 had slipped to 1.3% as of July 8, a downward revision from earlier in the quarter.
• June 2026 unemployment fell to 4.2% from 4.3%, but payrolls grew just 57,000 and the drop was driven largely by labor force participation falling to 61.5%, the lowest since March 2021, a weaker underlying signal than the headline suggests.
• The 10-year/2-year Treasury spread has un-inverted to about +0.34 to +0.36 percentage points as of mid-July 2026, but market analysts describe this as a 'bear steepener' driven by rising term premium rather than genuine growth optimism, making it a weaker green light than a typical steepening.
• The Sahm rule has triggered in all 11 US recessions since 1950 with essentially one ambiguous false positive in 1959 (which preceded a recession by six months); it does not appear triggered now given unemployment's narrow band this year, though the next official update was due August 7, 2026.
• Track record check: 10 of the last 11 postwar US recessions were preceded by a sharp oil price spike (Hamilton research), and the 2026 Iran conflict's Strait of Hormuz disruption pushed oil and CPI (4.2% YoY in May) higher while the Fed held rates at 3.50-3.75% with more officials leaning toward a hike than a cut, a stagflationary pattern that historically preceded recession more often than not; separately, about 70% of economists wrongly expected a recession within 12 months as of late 2022, and NBER typically takes 6 to 21 months to formally declare a recession after it starts.

The 10.5% YES price on "US recession by end of 2026" is being tested by a stagflation-shaped setup: Q2 GDPNow growth cut from 2.5% to 1.3% in two weeks, June payrolls came in at just 57K, and core PCE inflation is running at 3.4% (a three-year high), which leaves new Fed Chair Kevin Warsh boxed in between hiking and cutting. Several hard dates cluster in the next 90 days, the July 24 expiration of Section 122 tariffs, the July 28-29 and September 15-16 FOMC meetings, the July 30 Q2 GDP advance estimate, and a September 30 government funding deadline, any of which could reprice the contract sharply in either direction. Resolution itself carries real mechanical risk: the market can resolve YES purely on two consecutive negative BEA advance-estimate GDP quarters without any NBER recession call, or it can miss a genuine recession entirely if NBER's typically 6-21 month-lagged determination comes after the market closes. Wall Street's own house views (Goldman 20-30%, JPMorgan 35%) sit well above the market price, and a Bitcoin ETF outflow streak plus AI-capex-valuation stretch are plausible risk-off triggers, though none of this is a forecast or advice.

• Section 122 global tariffs (10%, raised to 15%) expire July 24, 2026 under their 150-day statutory cap; Congress has not extended them and the Court of International Trade already ruled them invalid in May 2026 with an appeal pending, so what replaces or doesn't replace them in the following weeks is a live catalyst for growth and inflation expectations.
• Growth is decelerating fast: the Atlanta Fed's GDPNow model for Q2 2026 fell from 2.5% on June 25 to 1.3% on July 8, and June nonfarm payrolls added only 57,000 versus a 115,000 consensus, with April and May revised down a combined 74,000; the drop in unemployment to 4.2% was driven by labor-force participation falling to 61.5% (lowest since March 2021), not hiring strength.
• Core PCE inflation hit 3.4% in May 2026, the highest since October 2023 and well above the Fed's 2% target, while growth slows, a stagflation combination that constrained the June 16-17 FOMC (new Chair Kevin Warsh) to hold at 3.50-3.75% while dropping its easing bias and floating a possible hike; the next decisions land July 28-29 (no economic projections) and September 15-16 (with projections).
• Resolution mechanics cut both ways: the market resolves YES if BEA advance-estimate real GDP is negative for two consecutive quarters between Q2 2025 and Q4 2026, or if NBER declares a recession by the Q4 2026 advance-estimate release; a weak Q3 2026 print (tariff whiplash, an AI-capex pause) could trigger YES mechanically even without NBER ever confirming a downturn, while NBER's historical 6-21 month lag means a genuine H2 2026 recession might not be declared before the market closes.
• Labor-market and data-integrity risk: continuing jobless claims rose to 1.814 million, a three-month high (watch level is 1.85 million for a structural signal), even as initial claims stayed low near 215,000, a pattern consistent with labor hoarding; separately, BLS staffing is down about 20% with a third of leadership roles vacant after the 2025 commissioner firing and a 911,000-job downward benchmark revision, raising the odds of another large, market-moving revision.
• A September 30, 2026 government funding deadline, with only 2 of 12 appropriations bills passed the House and none through the Senate as of mid-July, risks a third 2026 shutdown (two already occurred, one lasting a record 75 days) that could delay the GDP and jobs data this market depends on; meanwhile Goldman Sachs (20-30%) and JPMorgan (35%) both carry recession odds well above the 10.5% market price, and record June 2026 Bitcoin ETF outflows plus AI-capex valuation stretch are candidate risk-off triggers, though this is not a prediction.