Do Overtime Regulations Reduce Overtime Work in Japan?

This paper estimates the causal effect of Japan’s overtime-cap reform on the incidence of extreme overtime and its heterogeneous impacts. Exploiting the statutory coverage threshold by firm size, we implement a regression-discontinuity design to identify local average treatment effects on the probability that workers exceed legally salient overtime cutoffs. In the full sample, the reform significantly reduces the incidence of monthly overtime exceeding eighty hours, while we detect no statistically significant effects at the forty-five-hour cutoff. We then study two sources of heterogeneity. First, conditioning on workers’ pre-policy propensity to work long hours, the effect at the eighty-hour exceedance is substantially larger for high-propensity workers; within the top quartile of this propensity, the reform also lowers the probability of exceeding the statutory forty-five-hour limit, indicating progress on the policy’s primary objective among those most at risk. Second, using industry-level unionization as a proxy for bargaining power, stronger effects arise only at the eighty-hour exceedance in more unionized industries, with no significant effects at forty-five hours—consistent with Japan’s historically limited union involvement in policing moderate overtime. Overall, the evidence indicates that enforcement primarily trims the extreme upper tail and is most effective where worker predisposition or institutional bargaining amplifies compliance at high-risk margins.