Reported Operator Injuries and Wage Expenditure in U.S. Bus Agencies: A Fiscal-Year Panel, 2019–2024

This working paper examines whether prior reported operator injuries are associated with later real wage expenditure per worked operating hour in U.S. bus agencies. National Transit Database records for directly operated conventional bus service are matched to agency fiscal years for 2019–2024. The screened panel contains 1,670 agency-years from 304 agencies; the principal one-year-lag models use 1,345 observations from 284 agencies.

The originally planned all-staff coefficient is −0.000149 (95% confidence interval −0.008255 to 0.007957). The exploratory operator-specific coefficient is −0.000917 (−0.009713 to 0.007879). These are conditional within-agency associations with an accounting expenditure ratio, not identified wage-setting responses or causal hazard premiums.

This deposit preserves manuscript v6 dated October 14, 2026. A subsequent October 16 exploratory audit reports a negative operator association when weighting by prior vehicle revenue hours (coefficient −0.019266, p = 0.0217). That result changes the estimand and is not a confirmatory or causal finding. The manuscript predates that audit; the research progress note and supporting code and outputs are available at https://github.com/KiwiHazards/transit-injuries-wage-expenditure/blob/main/PROGRESS.md.

The released processed panels reproduce the headline estimates and weighted diagnostic. Injury reporting scope, occupational allocation, sparse counts, payroll composition, and bargaining timing remain limitations. AI-assisted internal checks are not independent human peer review. Author: Kruthik Ankam, University of Washington Seattle. Whitepaper license: CC BY 4.0.