August 2026 Unemployment Rate Confirmed at 4.1% as Question Resolves YES
A prediction question on whether the August 2026 US unemployment rate would land at 4.1% resolved YES on September 4, 2026, defying market prices and one AI estimate.
A closely watched question about the August 2026 unemployment rate has landed on the side few traders were betting on. On September 4, 2026, the question of whether that rate would come in at exactly 4.1% resolved YES, according to the stored resolution authority. For anyone whose job, wages or household budget rides on the health of the labor market, this is a concrete data point rather than a guess.
The outcome is notable because a precise 4.1% is a narrow target. Hitting one specific decimal reading, rather than a range, is the kind of call that markets and models tend to price cautiously. Yet the resolution came down firmly on YES, as of the September 4 resolution date.
Markets leaned toward no
Leading up to the decision, the prices on Polymarket suggested doubt that the rate would land exactly at 4.1%. The recorded market prices ran at 40.5%, 40.0%, 37.0%, 34.0%, 33.5% and 36.5%, all below the halfway mark. In plain terms, the crowd of traders was assigning less than even odds to the outcome that ultimately proved correct.
That gap between market pricing and the final result is a useful reminder of how prediction markets work. A price of, say, 37.0% is a live estimate of probability, not a settled fact. When a low-probability outcome resolves YES, it shows the market was expressing genuine uncertainty rather than a prediction that was simply wrong.
Where the AI estimate stood
The stored model reasoning went even further toward doubt than the market. An AI agent, deepseek/deepseek-v4-flash running via openrouter, put the probability at just 12%, with a confidence of 0.85. Its logic focused on the difficulty of hitting one exact figure.
The agent explained its thinking directly: "Current unemployment is around 4.0-4.2%. A precise 4.1% is a specific point estimate; BLS rounding and monthly volatility make exact hits unlikely. Historical distribution suggests ~12% chance." The review, run on September 4, 2026, carried a single judge and zero dispersion, meaning there was no disagreement among agents to average out. In the end, the outcome fell into that unlikely band the agent had described.
A wider run of resolved calls
This result arrived alongside another economics question that resolved the same way. On September 2, 2026, the question of whether the Bank of Canada would make no change to its target for the overnight rate at the September interest rate announcement also resolved YES, per the stored resolution authority.
Both items share a common thread: central-bank policy and labor-market readings are the signals that shape borrowing costs and hiring. Seeing them settle within days of each other gives readers a snapshot of how the economic questions of early September 2026 were closing out. Each remains a discrete outcome, resolved on its own terms rather than one driving the other.