Panel data and non-linear models

Lecturer: Patrick Gagliardini.

Week 1 (16 - 20 August, 2021)

The workshop is in jointly held with the Swiss School of Public Health (SSPH+) Summer School.

For these reason, SSPH Students have a reduced workshop fee of 300 CHF.

 

Workshop contents and objectives
This course introduces participants to methodological tools for conducting empirical analysis with panel data observations. Panel data consist of repeated observations over time for a group of individuals. This observational structure characterizes many empirical studies in social sciences, e.g. when data on individual workers in the labor market are collected across time, or health surveys are conducted in a population, and in many other fields where statistical analysis with individual-level data is used. In addition to continuous data, many applications involve observations in the form of binary or polytomous data (discrete choices), counts and duration data. The course will therefore cover the necessary tools for analyzing those data types in a panel framework. At the end of the workshop the participants would know how to (1) specify a suitable model for investigating the problem under study; (2) compare the theoretical model with empirical observations; and (3) draw proper conclusions based on the results. The workshop will address the key issues of unobserved individuals’ heterogeneity (with either fixed, or random effects), endogeneity, dynamic models, sample selection, and nonlinear modeling, e.g. when the outcome variable is dichotomous and represents a discrete choice of the individuals. Among the workshop objectives is the discussion of a series of empirical studies conducted with panel data, notably in the field of labor and health economics. Finally, part of the workshop is devoted to offering participants the possibility to discuss their research work related to the topics of this course, and econometric analysis in general, with the instructor and the class.

Prerequisites
Workshop’s participants are expected to be familiar with basic econometric analysis (linear model, least square estimation, hypothesis testing on the regression coefficients). Some prior knowledge of statistical software (such as R or Stata) is required; participants without such experience are invited to take some advance tutorials.

Bibliography

Cameron, A. and Trivedi, P., (2005) Microeconometrics, Cambridge University Press.

Hsiao, C. (2003) Analysis of panel data, Cambridge University Press; 2nd edition.

Kerkhofs, M., and M., Lindeboom (1997): “Age Related Health Dynamics and Changes in Labour Market Status”, Health Economics, 6, 407-423.