B1666
Title: The economics of monotonicity conditions: Exploring choice incentives in IV models
Authors: Rodrigo Pinto - University of Californai at Los Angeles (United States) [presenting]
Moshe Buchinsky - Science Po (France)
Abstract: The purpose is to examine how to use economic incentives to aid the identification of treatment effects in multi-valued choice models with categorical instrumental variables. A general yet simple framework is employed that utilizes revealed preference analysis to translate choice incentives into identification conditions. It is demonstrated that popular identification assumptions that rely on the monotonicity and separability of the choice equation can be traced back to specific properties of choice incentives. It is also shown that novel identification assumptions emerge when individuals face non-standard choice incentives. Finally, the usefulness of the approach is illustrated by revisiting prominent policy evaluation studies in the economic literature.
