This paper evaluates the effects of the Hungarian disability employment quota, which requires firms over a certain size to employ people with disabilities or pay a noncompliance tax. We employ a regression discontinuity design on firm-level data to estimate the effect on the quota on the employment of persons with disabilities, when the tax increased from very low levels to 170 percent of the lowest wage cost required to meet the quota. We estimate a lower bound of the effect that takes into account the bias resulting from bunching of firms below the threshold. Firms hire 0.24–0.29 additional disabled workers on average when the tax increased, with a lower bound of 0.16. When the threshold is raised from 20 to 25 employees, bunching of firms and the estimated effect disappears around the old threshold. The policy effect is weaker in regions with few disabled individuals, implying that the policy outcomes are hampered by low labor supply, materializing in high fixed costs of hiring, as predicted by our model.