We use data from a longitudinal survey to estimate the effects of time in drug abuse treatment on post-treatment weeks worked and earnings for 2,420 clients in three treatment modalities. The regression analysis shows that time in treatment had a positive and statistically significant impact on these labor market outcomes, but the effects were small for all modalities. Although residential clients experienced the largest relative changes in weeks worked and real earnings, a benefit-cost calculation suggests that additional residential treatment cannot be justified from earnings improvements alone. These results may indicate a need for more employment services while in treatment.