Open AccessBounds on treatment effects in regression discontinuity designs with a manipulated running variable
Author(s) -
Gerard François,
Rokkanen Miikka,
Rothe Christoph
Publication year - 2020
Publication title -
quantitative economics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 4.062
H-Index - 27
eISSN - 1759-7331
pISSN - 1759-7323
DOI - 10.3982/qe1079
Subject(s) - regression discontinuity design , causal inference , variable (mathematics) , econometrics , range (aeronautics) , regression , inference , computer science , discontinuity (linguistics) , cutoff , point (geometry) , empirical research , economics , statistics , mathematics , artificial intelligence , engineering , mathematical analysis , physics , geometry , quantum mechanics , aerospace engineering
The key assumption in regression discontinuity analysis is that the distribution of potential outcomes varies smoothly with the running variable around the cutoff. In many empirical contexts, however, this assumption is not credible; and the running variable is said to be manipulated in this case. In this paper, we show that while causal effects are not point identified under manipulation, one can derive sharp bounds under a general model that covers a wide range of empirical patterns. The extent of manipulation, which determines the width of the bounds, is inferred from the data in our setup. Our approach therefore does not require making a binary decision regarding whether manipulation occurs or not, and can be used to deliver manipulation‐robust inference in settings where manipulation is conceivable, but not obvious from the data. We use our methods to study the disincentive effect of unemployment insurance on (formal) reemployment in Brazil, and show that our bounds remain informative, despite the fact that manipulation has a sizable effect on our estimates of causal parameters.