Breaking Down the U.S. Employment Multiplier Using Micro-Level Data

We leverage contract-level data on federal procurement to analyze the employment effects of federal spending. Using establishment-level data from the Quarterly Census of Employment and Wages (QCEW) linked with the Federal Procurement Data System (FPDS), we decompose the employment impact of Nakamura-Steinsson defense spending shocks, revealing that at least 45% of MSA-level employment impacts stem from indirect effects on non-contractor firms. We identify a subset of procurement contracts that are plausibly unanticipated by recipient firms and find that these contracts have significantly greater employment impacts per dollar than ongoing contracts, which dominate defense procurement. Our analysis yields a cost-per-job estimate of $38,000 for these unanticipated contracts—substantially lower than the $436,000 typically associated with overall defense procurement. This discrepancy is attributed to the unexpected nature of these contracts, which more effectively stimulate local economic activity. Additionally, we leverage restricted QCEW data to provide the first estimates of the direct employment effects of government procurement on small non-public U.S. firms (the extensive margin), demonstrating that small firms are significantly more responsive to additional spending. These findings highlight the potential for significant heterogeneity in the impacts of government spending, suggesting opportunities for designing more effective stabilization policies.