Modélisation et exploration des données multivariées - Applied Econometrics

Enseignants

Draye Xavier; Gaspart Frédéric; Govaerts Bernadette;

Langue
d'enseignement

Français

Thèmes abordés

For the various types of non-experimental data (cross-sections, time series, panel data), the course outlines the main problems that are expected in linear regression models, some typical ways of identifying them with the help of theory or of statistical tests, and their most usual solutions. The main focus is, respectively, endogeneity for cross-sectional data, nonstationarity for time series and correlated random effects for panel data.

Acquis
d'apprentissage

A la fin de cette unité d’enseignement, l’étudiant est capable de :
1	a. Contribution de l’activité au référentiel AA (AA du programme) 1.1-1.5, 2.1-2.5 multivariate statistics, statistical inference, econometrics, 3.2-3.3 matching nonexperimental data sets and empirical questions with archetypal statistical models and their most frequent problems 3.4 estimation techniques 3.6-3.8 interpreting the results of empirical estimations 4.1-4.2 identifying typical problems in complex empirical situations 4.4-4.7 drawing lessons from abstract properties and statistical theory for complex, real situations linked with empirical estimation 6.2 & 6.6-6.7 homeworks 7.1 & 7.5 econometrics b. Formulation spécifique pour cette activité des AA du programme At the end of the course, students will be able : - to understand in depth the limits to empirical questions that can be answered with each type of data (experimental vs non-experimental, cross-sections, time series, panel data). - to identify a typical problem (amongst the most frequent ones) in a given linear model and a usual solution to this problem. - to formulate empirical strategies connecting the type of data, the expected statistical problems and the appropriate estimation techniques for linear models. - to estimate a linear model with the appropriate estimation technique. - to interpret the results of estimated linear models (including formulated predictions) with the help of statistical tests. - to validate or question a given empirical strategy in the light of its most frequent problems.

Contenu

1. Introduction to the different data types (cross-sections, time series, panel data) and to the small-sample and large-sample justifications of the OLS estimators
2. Cross-sections : typical exceptions to the Gauss-Markov assumptions, sources of endogeneity, IV estimators
3. Time series : the problem of non-stationarity, unit root tests, a few typical econometric specification for time series (Koyck, ECM,...)
4. Panel data : fixed effect model vs random effect model, the unifying Mundlak approach

Méthodes d'enseignement

En raison de la crise du COVID-19, les informations de cette rubrique sont particulièrement susceptibles d’être modifiées.

Classes and homeworks

Modes d'évaluation
des acquis des étudiants

En raison de la crise du COVID-19, les informations de cette rubrique sont particulièrement susceptibles d’être modifiées.

Homeworks in groups

Autres infos

Langue d'enseignement : anglais

Ressources
en ligne

Teams, Moodle

Faculté ou entité
en charge

AGRO