CMStatistics 2017: Start Registration
View Submission - CMStatistics
Title: Multivalued treatments and decomposition analysis: An application to the WIA program Authors:  Sebastian Calonico - Columbia University (United States) [presenting]
Ying-Ying Lee - University of California, Irvine (United States)
Wallice Ao - Institute for Defense Analyses (United States)
Abstract: Efficient estimators are analyzed for multi-valued treatment effects on the treated that can be used to conduct distributional decompositions in the outcome variable. In particular, we propose two-step semiparametric efficient estimators that can be used to decompose differences in the outcome distribution into (i) a wage structure effects, arising due to the conditional outcome distributions associated with different levels of participation; and (ii) a composition effect, arising due to differences in the distribution of observable characteristics. These counterfactual differences reveal causal relationships under a conditional independence assumption. Moreover, we calculate the semiparametric efficiency bound for the multivalued treatment effects and we provide uniform inference results for efficient multi-valued nonparametric propensity score weighting estimators.We employ these procedures to study evaluate the Workforce Investment Act (WIA), a large US job service program. Our estimation results show that heterogeneity in levels of participation is an important dimension to evaluate the WIA and other social programs in which participation varies. The results, both theoretically and empirically, provide rigorous assessment of intervention programs and relevant suggestions to improve the performance and cost-effectiveness of these programs.