This post analyzes why U.S. nonfarm payroll (NFP) figures have repeatedly been revised sharply downward months after optimistic initial releases. The author traces the root cause to the BLS Current Employment Statistics (CES) establishment survey, whose response rate has collapsed from about 72% in 2004 to below 40% in 2024-2026. To fill the gaps, the agency's net birth/death model extrapolates job creation from historical seasonal trends—an assumption that works in expansions but systematically overstates hiring during high-rate downturns, adding an estimated 50,000-120,000 phantom jobs per month. The Quarterly Census of Employment and Wages (QCEW), an administrative tax-record dataset covering over 95% of jobs with a 5-9 month lag, later exposes the gap: a benchmark revision erased 818,000 reported jobs in a single year, the largest since 2009. The author argues this is not deliberate manipulation but an institutional structure with asymmetric incentives—markets price the initial print instantly while corrections arrive after attention has faded. The post closes with a quantitative cross-section (SPY, QQQ, IWM, VIX, 10-year yields) and a two-week tactical outlook. Tags: nonfarm-payrolls, BLS, QCEW, data-revision, labor-market, macro, quant
818,000 Ghost Jobs: A Mathematical Dissection of Nonfarm Payrolls' 'Rally First, Slash Later' Revisions
Summary
This post analyzes why U.S. nonfarm payroll (NFP) figures have repeatedly been revised sharply downward months after optimistic initial releases. The author traces the root cause to the BLS Current Employment Statistics (CES) establishment survey, whose response rate has collapsed from about 72% in 2004 to below 40% in 2024-2026. To fill the gaps, the agency's net birth/death model extrapolates job creation from historical seasonal trends—an assumption that works in expansions but systematically overstates hiring during high-rate downturns, adding an estimated 50,000-120,000 phantom jobs per month. The Quarterly Census of Employment and Wages (QCEW), an administrative tax-record dataset covering over 95% of jobs with a 5-9 month lag, later exposes the gap: a benchmark revision erased 818,000 reported jobs in a single year, the largest since 2009. The author argues this is not deliberate manipulation but an institutional structure with asymmetric incentives—markets price the initial print instantly while corrections arrive after attention has faded. The post closes with a quantitative cross-section (SPY, QQQ, IWM, VIX, 10-year yields) and a two-week tactical outlook. Tags: nonfarm-payrolls, BLS, QCEW, data-revision, labor-market, macro, quant
This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634498