COURSE ID: D-EF13 LANGUAGE: MICROECONOMETRICS USING STATA

Microeconometrics using Stata offers participants with a comprehensive applied and theoretical overview of the principle methodologies implemented in the analysis of microeconomic data. More specifically, the course focuses on instrumental variable analysis, non-linear least squares estimation, binary variable models, multi-nominal models, Tobit models and count data models, panel data models, IV estimators and GMM estimators. Although the course is entitled “Microeconometrics using Stata”, as the examples discussed relate to economic data, the techniques developed through the courses can of course are extensively implemented in other social sciences.

In common with TStat’s course philosophy, each individual session is composed of both a theoretical component (in which the techniques and underlying principles behind them are explained), and an applied (hands-on) segment, during which participants have the opportunity to implement the techniques using real data under the watchful eye of the course tutor. Throughout the course, theoretical sessions are reinforced by case study examples, in which the course tutor discusses and highlights potential pitfalls and the advantages of individual techniques.

Researchers and professionals working in biostatistics, economics, epidemiology, finance, psychology, social and political sciences needing to acquire the necessary statistical requisites required to independently conduct empirical analysis using micro data.

It is assumed that course participants have at some point followed a basic course in econometrics or statistics. Previous exposure to Stata or other statistical software packages would also be an advantage.

SESSION I: PRELIMINARY TOPICS

Stata 15 – a quick review
Linear and non-linear models in Econometrics
Estimators and tests for linear models with endogenous variables: Instrumental Variables and Generalized Method of Moments (ivregress, ivreg2, gmm, treatreg)
Estimators and tests for non-linear models
Estimating marginal effects with margins

SESSION II: COUNT MODELS

The Poisson model

Estimators: Non-Linear Least Squares (nl), GMM (gmm), Maximum likelihood (poisson)
Endogenous regressors (gmm and ivpoisson)
Overdispersion: the Negative Binomial Model (nbreg)

SESSION III: Discrete dependent variable models

Univariate models

Linear Probability Model, Probit and Logit (regress, probit, logit)
Ordered models (oprobit, ologit)

Multivariate models

Bivariate and multivariate Probit models (biprobit, mvprobit, cmp)
Multinomial models

(Conditionally) independent latent heterogeneity in probit models

Estimation of average partial effects

Endogenous regressors in probit models

The control function approach (CFA) to continuous endogenous regressors: test and estimation
Bootstrap standard errors and covariance matrix in the CFA
Maximum likelihood estimation with continuous endogenous regressors (ivprobit)
A multivariate probit solution to binary endogenous regressors (biprobit, mvprobit, cmp)

SESSION IV: PROBIT AND LOGIT PANEL-DATA MODELS

The ancillary parameter problem in non-linear models with correlated latent heterogeneity (LH)

Logit and probit panel data models with LH

Models with independent LH: Random effect models (xtlogit, xtprobit)
Models with correlated LH: Fixed effect models
The Chamberlain-Mundlak approach for probit models
The Fixed effect logit model (xtlogit)

SESSION V: ModelS with Censoring and sample selection (TIME Permitting)

Censoring
Tobit models: ML and Two-step Least Squares (tobit)
The CFA to continuous endogenous regressors: test and estimation
Maximum likelihood estimation with continuous endogenous regressors (ivtobit)
Panel data tobit models with LH
Sample selection

Tests and corrections a la Heckman (heckman) for linear models
Tests and corrections for linear panel-data models
Attrition in panel-data models: Inverse Probability weighting (IPW)
Bootstrap standard errors with IPW

to be confirmed.

NAME

EMAIL

OBJECT