Aller à : contenu haut bas recherche
 
 
EN     FR
Vous êtes ici:   UNIL > HEC Inst. > HEC App. > SYLLABUS
 
 

Microeconometrics

  • Enseignant(s):   M.Huber  
  • Titre en français: Microéconométrie
  • Cours donné en: anglais
  • Crédits ECTS: 6 crédits
  • Horaire: Semestre de printemps 2019-2020, 4.0h. de cours (moyenne hebdomadaire)
  •  séances
  • Formation concernée: Maîtrise universitaire ès Sciences en économie politique

 

Objectifs

This course discusses microeconometric methods for causal inference, i.e. to assess the causal impact of a specific explanatory variable (also referred to as "treatment" or "policy intervention") on an outcome (or "dependent variable") of interest. This may, for instance, concern the effectiveness of public policies (e.g. training programs for unemployed, income support for poor families, public childcare,…), corporate policies (marketing campaigns, educational programs for employees,…), health interventions (new medical treatments…), among many other examples. This course reviews and extends methods for causal inference/policy evaluation partly covered in the course "Econometrics", including matching, inverse probability weighting, doubly robust estimation, instrumental variables, difference in differences, changes in changes, and regression discontinuity and kink designs. It also provides an introduction to machine learning and its application to causal inference in big data contexts. Finally, it discusses the evaluation of mechanisms through which a treatment may affect an outcome (e.g. training may affect mental health via finding employment), known as causal mediation analysis.

The methods are first introduced and discussed in the lecture and then applied to actual data in 4 PC sessions using the statistical software "R". For this reason, an introduction to R is provided in the second week of the semester. Prior to each PC session, students are asked to do a little home assignment (3-5 exercises) with a data set in groups of up to 4 students, which is then discussed in class (exercises can be solved in R or stata). By the end of the course, the students should be able to: (a) formulate empirical research questions involving microeconomic data; (b) understand the requirements, advantages, and disadvantages of various methods for assessing causal effects; (c) select appropriate methods for the research question at hand; (d) be able to implement such methods using appropriate software; (e) interpret the results of the analysis.

Contenus

The main topics covered in the course are the following:

- The definition of causal effects (review of the “potential outcome” notation)
- Methods for policy and impact evaluation under exogeneity: estimation based on matching, inverse probability weighting and tilting, and doubly robust estimation
- Methods for policy and impact evaluation under endogeneity: estimation based on instruments, regression discontinuities and kinks, differences in differences, and changes in changes
- Machine learning (e.g. lasso regression, random forests) for the analysis of big data and its application to causal inference
- Methods for assessing causal mechanisms underlying a particular causal effect (causal mediation analysis)

The lecture is accompanied by 1 tutorial providing an introduction to the software package "R" as well as 4 PC sessions, in which the methods are applied to empirical data.

Références

Basic reading - briefly covers most topics of the course:

Huber (2019): An introduction to flexible methods for policy evaluation

Further reading - these papers cover the topics of the course in more detail (and even exceed the content of the course):

Imbens and Wooldridge (2009): Recent developments in the Econometrics of Program Evaluation

Lee and Lemieux (2010): Regression discontinuity designs in economics

Lechner (2011): The Estimation of Causal Effects by Difference-in-Difference Methods

Card, Lee, Pei, and Weber (2016): Regression kink design - Theory and practice

Chernozhukov et al. (2018): Double/debiased machine learning for treatment and structural parameters

Huber (2019): A review of causal mediation analysis for assessing direct and indirect effects

Huber and Wüthrich (2019): Local average and quantile treatment effects under endogeneity - a review

Further reading for an introduction to machine learning (exceeds the content of the course):

James, Witten, Hastie, and Tibshirani (2013): An Introduction to Statistical Learning with Applications in R, Springer, New York

Pré-requis

Connaissances solides en économétrie de base.

Evaluation

1ère tentative

Examen:
Ecrit 1h30 heures
Documentation:
Non autorisée
Calculatrice:
Non autorisée
Evaluation:

La note finale est une moyenne pondérée des points obtenus dans l'examen (80%) et pour les travaux de groupes (20%).

Rattrapage

Examen:
Ecrit 1h30 heures
Documentation:
Non autorisée
Calculatrice:
Non autorisée
Evaluation:

La note finale est une moyenne pondérée des points obtenus dans l'examen de rattrapage (80%) et pour les travaux de groupes (20%).



[» page précédente]           [» liste des cours]
 
Recherche


Internef - CH-1015 Lausanne - Suisse  -   Tél. +41 21 692 33 00  -   Fax +41 21 692 33 05
Swiss University