Objective: There is moderate evidence that workers in experience-rated firms sustain less injuries when compared with workers in firms that are not experience rated. This study aims to provide more insight on this issue. Methods: Panel data from the Bureau of Labor Statistics and National Academy of Social Insurance between 1999 and 2006 were used. A theoretical framework was developed, and a fixed effects vector decomposition model was estimated. Results: Self-insuring was positively associated with relatively low worker injury and illness incidence rates when compared with insuring (including experience rating and manually rating). After controlling for workforce characteristics, industrial composition, firm size, and state-specific laws, states with an above the median percentage of self-insured firms had incidence rates that were lower than rates in states with a below the median percentage of self-insured firms. Conclusion: A higher degree of' experience rating seems to better align the economic incentive to invest in prevention and the intended outcome of reducing worker injury and illness.