Background Workers' compensation systems throughout Australia implement "step-downs," which cut the amount paid to injured workers after they have received benefits for several months. Though initially introduced to control rising insurance premiums, step-downs are currently justified as an incentive return to work. Whether they have this effect has never been formally tested. Methods Using administrative claims data, we applied a regression discontinuity study design to test whether step-downs

more »

... affected weekly scheme exit rates, a proxy for return to work, within eight state, territory, and Commonwealth workers' compensation systems. We also examined effects by injury type (fractures, musculoskeletal, mental health, and other trauma). To derive generalised effects, we combined results using meta-analyses and conducted meta-regressions to determine whether timing or magnitude of step-downs significantly moderated effects. Results Step-downs reduced scheme exit by 0.86 percentage points (95% CI: -1.45, -0.27). Neither timing nor magnitude of step-downs was a significant effect moderator. There were significant effects in fractures (-0.84, 95% CI: -1.61, -0.07) and sensitivity analysis suggested possible effects within mental health and musculoskeletal conditions. Conclusions The negative effects suggest some workers' compensation recipients anticipate step-downs and return to work early to avoid a reduction in income. However, the effects were small and suggest step-downs have marginal practical significance. We conclude that step-downs are generally ineffective as a return to work policy initiative. Key messages Reducing the amount of compensation paid to injured workers has a minor incentivising effect on their return to work rates.