Detecting Grouped Local Average Treatment Effects and Selecting True Instruments

In the context of an endogenous binary treatment with heterogeneous effects and multiple instruments, we propose a two-step procedure to identify complier groups with identical local average treatment effects (LATE), despite relying on distinct instruments and even if several instruments violate the identifying assumptions. Our procedure is based on the fact that the LATE is homogeneous for any two or multiple instruments which (i) satisfy the LATE assumptions (instrument validity and treatment monotonicity in the instrument) and (ii) generate identical complier groups in terms of treatment propensities given the respective instruments. Under the (plurality) assumption that for each set of instruments with identical treatment propensities, those instruments satisfying the LATE assumptions constitute the relative majority, our procedure permits identifying these true instruments in a data driven way. We also provide a simulation study investigating the finite sample properties of our approach and an empirical application investigating the effect of incarceration on recidivism in the US with judge assignments serving as instruments.