Abstract This study analyzes the relationship between the economic cycle and work accidents in the United States. The empirical strategy is based on vector autoregression models (VAR) for time series and panel-data settings on a sample of 40 US States during 2003-2018. The results confirm a bidirectional causal relationship in the short-run between economic activity-i.e., Gross Domestic Product (GDP) per worker-and work accidents in 28 States. Additionally, the empirical evidence suggests that this relationship is heterogeneous. In line with increased awareness on how the economic cycle affects the temporal trajectory of work accidents across territories, policy implications and future research avenues are discussed.