Identification and inference in nonparametric models

The econometric evaluation of the impact of an economic policy on an outcome variable of interest is critical to society. Since social and economic systems are extremely complex and economic theory does not fully specify the relationships between variables of interest, modelling of these variables involves the use of nonparametric models. Moreover, relevant variables are usually very difficult to fully control and even to observe making the analysis fairly complex. This project aims at advancing the methodology for identification, estimation and inference in nonparametric models used for policy evaluation. The results will have applications to labour market policies, health care policies and education policies among others.