US employers added 165,000 non-farm jobs in October, the Bureau of Labor Statistics reported last week, more than double economists' median forecast of 70,000. The surprise extended a run of large forecasting misses in 2026, highlighting the difficulty of interpreting monthly changes in the world's largest economy.
A Bloomberg analysis published by Mint linked the widening gaps partly to reduced resources at US statistical agencies, alongside modest underlying job growth and a change in how the government estimates employment at new and closing businesses. Originally reported by mint.
The divergence matters because unexpected payroll figures can increase financial-market volatility. In Bloomberg's survey of more than 70 economists, even the highest October forecast, at 131,000, fell short of the reported gain.
Forecast misses grow beyond October
October's 95,000 gap was not the largest this year. Forecasts for February, March, July and August each missed the reported result by more than 100,000 jobs.
The average absolute forecasting error has reached about 88,000 in 2026, compared with roughly 50,000 during 2010-2019. Those figures measure the size of the miss regardless of whether economists overestimated or underestimated hiring.
Budget pressure weakens survey collection
Inflation-adjusted BLS funding has dropped by well over 20% since 2010, according to the analysis. Lawmakers resisted deeper reductions sought by the second Trump administration, but the agency has faced shrinking samples and difficulty securing survey responses.
Former BLS commissioner Erika McEntarfer told the Financial Times in July that people and businesses had become harder to contact just as budget cuts were squeezing field resources. Reaching respondents now costs more, adding pressure to data collection.
The resulting uncertainty makes small monthly movements harder to interpret. The analysis put the 90% confidence interval at around 122,000, meaning gains below that level cannot be statistically distinguished from zero; roughly half of 2026's reports have fallen into that category.
A revised model adds another complication
The BLS surveys employers but must also estimate jobs associated with businesses opening and shutting. Its birth-death model changed in January 2026, combining historical extrapolation with more current information rather than relying on the smoother older approach.
That makes the estimate more variable and harder to forecast. Against that statistical uncertainty, October's stronger-than-expected result is less informative on its own than the longer-term employment trend.
Originally reported by mint.