_____________________________________________________________________________________
1. INTRODUCTION.
Economic data and econometric modeling. Causality and the ceteris paribus notion in econometric analysis. Observational vs. Experimental data. Consequences on econometric modeling.
2. THE SIMPLE LINEAR REGRESSION MODEL.
Assumptions. Coefficient interpretation. The OLS (Ordinary Least Squares) estimator and its properties. The variance of the OLS estimator.
3. THE MULTIPLE LINEAR REGRESSION MODEL.
Assumptions. Relation between the linear simple and multiple regression model: short vs. long regression. OLS estimation of coefficients and variance estimation. Goodness of fit. Coefficient interpretation. Most usual transformations. Calculation of elasticities. Inference. Tests about a single parameter. Tests about a linear combination of parameters. Tests about several linear constraints.
4. REGRESSION ANALYSIS WITH QUALITATIVE INFORMATION: BINARY VARIABLES.
Binary variables. A single independent binary variable. Sets of binary variables for multiple categories. Interactions with binary variables. Coefficient interpretation.
5. SPECIFICATION ERRORS.
Inclusion of irrelevant variables and omission of relevant variables. Consequences on estimation. Measurement errors. Consequences on estimation.
6. MODELS WITH ENDOGENOUS EXPLANATORY VARIABLES.
The concept of endogeneity and its sources. The method of instrumental variables (IV). Valid instruments. The two-stage least squares estimator (2SLS). Endogeneity tests (Hausman tests). Instrument validation: tests of overidentifying restrictions (Hansen-Sargan tests).
7. HETEROSKEDASTICITY.
The regression model with heteroskedasticity. Consequences on OLS estimation and 2SLS estimation. Heteroskedasticity—robust inference: White or Eicker-White estimator of the variances.
8. AUTOCORRELATION.
The regression model with time-series data. Consequences on OLS estimation Autocorrelation-robust inference: Newey-West estimator of the variances. Autocorrelation tests.