We present a new identification condition for regression discontinuity designs. We replace the local randomization of Lee (2008) with two restrictions on its threat, namely, the manipulation of the running variable. Furthermore, we provide the first auxiliary assumption of McCrary's (2008) diagnostic test to detect manipulation. Based on our auxiliary assumption, we derive a novel expression of moments that immediately implies the worst-case bounds of Gerard, Rokkanen, and Rothe (2020) and an enhanced interpretation of their target parameters. We highlight two issues: an overlooked source of identification failure, and a missing auxiliary assumption to detect manipulation. In the case studies, we illustrate our solution to these issues using institutional details and economic theories.
翻译:暂无翻译