Econometrics of Micro, Macro, and Multilevel Panel Data 2026

Econometrics of Micro, Macro, and Multilevel Panel Data

Bertinoro, 14-20 June 2026

Coordinator

Maria Elena Bontempi  
University of Bologna, Department of Economics, Piazza Scaravilli 2, 40126 Bologna  
e-mail: mariaelena.bontempi@unibo.it  
website: https://sites.google.com/site/mariaelenabontempi/home 

Lecturers

Maria Elena Bontempi, University of Bologna
Roberto Golinelli, University of Bologna
Irene Mammi, University of Venezia Ca' Foscari

The course will be delivered in presence and online. The maximum number of allowed participants in presence is 50.

Requirements

Basic knowledge of econometrics to a level comparable with Introductory Econometrics SIDE course.

Examples of reference are represented by:

** Verbeek M. (2017), A guide to Modern Econometrics, 5th ed, Wiley, Ch. 1, 2, 3, 4, 5, 6 or
** Wooldridge J. M. (2020), Introductory Econometrics, a Modern Approach, 7th ed., Cengage, Ch. 1-6, 8-9, 15

Reference textbooks for the course:

**Wooldridge J. M (2010), Econometric Analysis of Cross-Section and Panel Data, 2nd ed, Cambridge Mass.: MIT Press, Ch. 10- and 11 (panel), Ch. 7 (SUR)
**Handouts, readings and further material will be provided during the lectures.

Course outline, objectives and learning outcomes
Panel and multilevel datasets – covering repeated observations over time and structures with higher- and lower-levels – are now fundamental to empirical research in economics, political economy, and social sciences. These data structures raise econometric challenges: unobserved heterogeneity, dependence across units, endogeneity, dynamic adjustment. The aim of this course is to provide a coherent theoretical and applied overview of econometric models for panel and multilevel data, understood as datasets with at least two dimensions (e.g. individual over time, units nested in higher-level groups). To easily introduce the two main parts of the course, let N denote the cross-sectional dimensions and T the time dimension. The first part focuses on micro panels and multilevel data (N > T), where the key issue is how to model/control for unobserved heterogeneity across units and levels. Starting from static models, we introduce fixed, random, and correlated effects, emphasising their interpretation from a multilevel perspective and their implications for causal inference. We then address endogeneity arising from individual heterogeneity, simultaneity, measurement errors, and dynamics, with instrumental variable estimators and the Generalized Method of Moments (GMM) as the main methodological tools. The second part of the course covers macro and long panels (T > N), where non-stationarity, cointegration, heterogeneous dynamics, and cross-sectional dependence become central. We study these issues in relation to short- and long-run relationships and discuss estimation strategies that allow for parameter heterogeneity and common factors. At each stage, theoretical material is complemented by hands-on empirical applications, allowing participants to implement the methods and interpret results using real data. 

Learning outcomes: At the end of the course, participants will be able to:

• recognize when the data structure requires standard panels or multilevel methods;
• specify/estimate static and dynamic panel models based on different heterogeneity assumptions;
• distinguish between exogenous and endogenous regressors and select appropriate identification strategies;
• apply instrumental variable and GMM estimators in micro panel contexts;
• model heterogeneous dynamics and cross-sectional dependence in macro panels;
• diagnose non-stationarity and cointegration and interpret long- and short-run effects;
• critically evaluate empirical studies based on panel or multilevel data and design your own empirical analyses.

Schedule of the course:

1. Crash course: OLS, GLS, IV; heteroscedasticity and clustering; introduction to dynamics and  unit roots in empirical models.
2. Static panels and multilevel structure. Understanding the clustered data structure. Decomposition of variance between levels, unobserved heterogeneity and heterogeneity bias. Methods: fixed, random, and correlated effects (Mundlak-Chamberlain); random slopes. Practical  session.
3. Dynamic panels. Sources of endogeneity in panels (Nickell’s bias). Instrumental variables and identification. Weak instruments and instrument proliferation. Methods: first difference, IV, GMM-DIF, GMM-LEV, GMM-SYS estimators; reduction of instrument sets (PCA). Practical session.
4. Non-stationary panels. Integration, cointegration, and common factors. Methods: first- and second-generation panel unit roots tests; Pedroni and Westerlund cointegration tests; PANIC and PANICCA. Practical session.
5. Heterogeneous panels and cross-sectional dependence. Short- and long-run dynamics in heterogeneous panels. Poolability and model selection. Methods: ARDL models, Mean-Group (MG) and Pooled Mean Group estimators; demeaning and common correlated effects. Practical sesssion.

Practical Sessions

The theoretical lessons are supplemented by practical sessions during which you will receive the necessary guidance on how to use econometric software and perform your empirical analyses. Stata will be used, and assistance will be provided during the lessons to new Stata users. The data sets and programming files for applied econometrics will be provided during the lessons in Bertinoro. 

Remark on the software

Please note that the course is NOT about Stata, BUT about the econometrics of panel data. Stata is just a tool, like any other econometric package capable of handling panel data. Furthermore, we do not have and cannot provide Stata software installation files and licences in Bertinoro. If you wish to participate in the lectures and practical sessions actively, we strongly recommend that you have a software pre-installed on your laptop before coming to Bertinoro

Assessment

The certificate of attendance will be based on one final “take-home in Bertinoro” exam. The assignment will consist of an empirical exercise covering the topics addressed during the lectures, to be carried out using econometric software and the theoretical knowledge acquired. Working in groups is strongly encouraged to improve exchanges of ideas, but each of you must be identifiable in your individual contribution.

 Venue and timetables

Each Module requires a full-time attendance and participation is not compatible with other jobs at the same time (e.g. preparation of other exams). The Module will last one week and will be held in the University Residential Centre, Via Frangipane 6, 47032 Bertinoro (FC). Participants will be hosted in the Centre guest quarters, (as an exception, in case of reduced availability of rooms in the Centre, they will be accommodated in local hotels).

Lectures will be in English, with the following schedule:

** Monday  to Friday: lectures from 9am to 1pm and 3pm to 5pm

** Daily individual practice sessions: 5pm to 7pm

** Saturday: lectures from 9am to 1pm

Contacts

• For more information: Roberta Partisani phone: +39 0543446554 ; e-mail: info@side-iea.it
• For administrative issues :  Fabrizio Zanotti (admin@side-iea.it),  phone:0543446561;
• For travel and accommodation: Roberta Partisani (rpartisani@ceub.it), phone: +39 0543446554