Social Experiments and Instrumental Variables with Duration Outcomes

Abstract

This paper examines the empirical analysis of treatment effects on duration outcomes from data that contain instrumental variation. We focus on social experiments in which an intention to treat is randomized and compliance may be imperfect. We distinguish between cases where the treatment starts at the moment of randomization and cases where it starts at a later point in time. We derive exclusion restrictions under various informational and behavioral assumptions and we analyze identifiability under these restrictions. It turns out that randomization (and by implication, instrumental variation) by itself is often insufficient for inference on interesting effects, and needs to be augmented by a semi-parametric structure. We develop corresponding non-and semi-parametric tests and estimation methods.

Get Training or Wait? Long-Run Employment Effects of Training Programs for the Unemployed in West Germany

“…This is a direct consequence of modeling heterogeneity with respect to the starting time of the treatment relative to the length of elapsed unemployment. Both the k-group and the l-group at the start of the treatment are affected by the dynamic sorting effects taking place before, see Abbring and van den Berg (2004) for a recent dis-cussion of this problem in the context of estimating duration models. Thus, the estimated treatment parameter depends dynamically on treatment decisions and outcomes in the past when taking the timing of events seriously (Abbring and van den Berg, 2003;Johansson 2003, 2004;Sianesi 2003Sianesi , 2004.…”

Employment effects of the provision of specific professional skills and techniques in Germany

“…This is a direct consequence of modelling heterogeneity with respect to the starting time of the treatment relative to the length of elapsed unemployment. Both the treatment group and the group of nonparticipants at the start of the treatment are affected by the dynamic sorting effects taking place before, see Abbring and van den Berg (2004) for a recent discussion of this problem in the context of estimating duration models. Thus, the estimated treatment parameter depends dynamically on treatment decision and outcomes in the past when taking the timing of events seriously (Abbring and van den Berg, 2003;Fredriksson and Johanson, 2003;Sianesi, 2004).…”

Identification of a competing risks model with unknown transformations of latent failure times

“…For example, Bond and Shaw (2003) obtain bounds for covariate effects under the assumption that the copula associated with the joint distribution of latent failure times is invariant to the value of covariates. Abbring and Van den Berg (2005) apply the result of Bond and Shaw (2003) to bound the treatment effects on duration outcomes. Honoré and Lleras-Muney (2004) derive bounds in an accelerated failure time competing risks model with discrete covariates.…”