In this webinar, I will present a review of recent advances in applied econometrics, which is useful for researchers working on empirical questions related to program/policy impact evaluation. I will consider three major topics offering recommendations for applied work in each of them. I will start with the discussion of causal inferences with observation data; natural experiment design-based identification strategies including difference-in-difference estimator, synthetic control methods, instrumental variables method, and regression discontinuity design. These will be followed by a brief look into external validity, and the causal interpretation of regression methods. Secondly, I will present different supplementary analyses of these estimators with a focus on placebo and robustness tests to establish credibility of the identification strategies. Finally and more importantly, I will introduce recent advances in causal machine learning (ML) with a focus on (1) l^1-Least Absolute Shrinkage and Selection Operator (LASSO) for estimating treatment effects and model selection and (2) Causal Forest for identifying and estimating heterogenous treatment effects.
PRESENTER: Dr Dambala Gelo Kutela, Senior Lecturer, School of Economics and Finance, University of the Witwatersrand, South Africa; Member of AFAERE; Senior Research Associate, Environment for Development (EfD)
ZOOM MEETING LINK: https://gu-se.zoom.us/j/69977090559
TIME: 8 October 2020, 14h00 GMT