A0719
Title: Tests of exogeneity in duration models with censored data
Authors: Gilles Crommen - KU Leuven (Belgium) [presenting]
Ingrid Van Keilegom - KU Leuven (Belgium)
Jean-Pierre Florens - Toulouse School of Economics (France)
Abstract: Consider a duration time of interest $T$, a possibly endogenous treatment variable $Z$, and a vector of exogenous covariates $X$ such that $T=\varphi(Z,X,U)$ is increasing in $U$ with $U \sim U[0,1]$. Moreover, let $T$ be right-censored by a censoring time $C$ such that only their minimum, denoted by the follow-up time $Y=\min\{T,C\}$, is observed. Test statistics are constructed for the hypothesis that $Z$ is exogenous w.r.t. $T$, that is, $Z$ is independent of $U$. The tests makes use of an instrumental variable $W$ that is independent of $U$ given $X$, since it can be shown that $Z$ is exogenous w.r.t. $T$ if and only if $V_T = F_{T \mid Z,X}(T \mid Z,X)$ is independent of $(W,X)$ jointly. The asymptotic properties of the proposed test are proved for the case where $Z,X$, and $W$ are categorical, and possible bootstrap approximations for the critical value of the tests are shown to have a good finite sample performance via simulations. Lastly, an empirical example is provided using data from the National Job Training Partnership Act (JTPA) Study.
