Partial Identification in Regression Discontinuity Designs with Manipulated Running Variables

In a regression discontinuity design, units are assigned to receive a treatment if their value of a running variable lies above a known cutoff. A key assumption to ensure point identification of treatment effects in this context is that units cannot manipulate the value of their running variable to guarantee or avoid assignment to the treatment. Standard identification arguments break down if this condition is violated. Christoph Rothe presents evidence that treatment effects remain partially identified even if some units are able to manipulate their running variable.