Network and Spatial Panel Data Analysis using Stata

Monday 13/07/2026 to Friday 17/07/2026

Program

This is a 5-day event. The course consists of both theoretical sessions (in the form of lectures) and practical sessions (in the form of Stata exercises). Each practical session is designed to complement the lectures with empirical illustrations.

Participants are strongly encouraged to bring their own laptop with an installation of Stata. If this is not feasible, the participant should inform the organizers and a temporary license and/or laptop will be provided. Participants are expected to have basic knowledge of Stata and panel data analysis.

Daily structure

• 1st Morning Lecture: 09:30–11:00
• Coffee break: 11:00–11:30
• 2nd Morning Lecture: 11:30–13:00
• Lunch break: 13:00–14:00
• Afternoon Session (Stata lab): 14:00–15:30

Detailed Programme

Day 1 (Monday 13/07/2026)

09:30–11:00 – Lecture 1: Why Spatial and Network Models?
Lecturer: Professor V. Sarafidis

• The importance of modelling spatial and network interactions in economic and social data.
• Consequences of neglecting spatial dependence: bias, inconsistency, and misleading policy inference.
• Applications from economics, finance, trade, and regional studies.
• Diagnostics for spatial dependence: Moran’s I statistic and related tests.

11:00–11:30 – Coffee break

11:30–13:00 – Lecture 2: Spatial Analysis for Panel Data Models
Lecturer: Professor V. Sarafidis

• Overview of available approaches.
• Identification conditions and underlying assumptions.
• IV estimation with two-way fixed effects and latent common factors.
• Introduction to the Stata command spxtivdfreg.
• Comparing models with and without spatial terms, and across different spatial weighting matrices.

13:00–14:00 – Lunch break

14:00–15:30 – Stata Lab Session 1
Introduction to spatial panel data commands in Stata.
Lecturer: Professor T. Panagiotidis

Day 2 (Tuesday 14/07/2026)

09:30–11:00 – Lecture 3: Direct and Indirect (Spillover Effects) in Spatial Panels
Lecturer: Professor V. Sarafidis

• Definition and interpretation of direct, indirect, and total effects.
• Estimation of average effects in spatial models.
• When are spillovers small vs substantial?
• Extensions: dynamic spatial panels and long-run effects.

11:00–11:30 – Coffee break

11:30–13:00 – Lecture 4: Heterogeneous Spatial Panels
Lecturer: Professor V. Sarafidis

• Motivation for heterogeneous coefficients.
• Mean Group estimators.
• Computation of heterogeneous direct and indirect effects.
• Applications and interpretation.

13:00–14:00 – Lunch break

14:00–15:30 – Stata Lab Session 2
Estimation of spatial panels (homogeneous and heterogeneous) in Stata.
Lecturer: Professor T. Panagiotidis

Day 3 (Wednesday 15/07/2026)

09:30–13:00 – Case Study: Independent Empirical Project

Participants will work with a full-scale dataset (provided in advance) to apply the material covered in the preceding sessions.

• Specifying an appropriate spatial or network model.
• Estimating spatial models in Stata using both known weighting structures.
• Computing and interpreting direct, indirect, and total effects.
• Presenting results in tables and visualising networks graphically.

The session will be conducted in small groups or individually. Feedback will be provided during the Stata Lab Session 4.

Day 4 (Thursday 16/07/2026)

09:30–11:00 – Lecture 5: Spatial Panels with Time-Varying Weighting Matrices
Lecturer: Professor V. Sarafidis

• Motivation and examples (evolving trade, financial networks).
• Modelling strategies and identification challenges.
• Implementation in Stata.

11:00–11:30 – Coffee break

11:30–13:00 – Lecture 6: Introduction to Network Models
Lecturer: Professor V. Sarafidis

• Moving beyond exogenous spatial weights: estimating the connectedness structure from the data.
• Identification conditions and underlying assumptions.
• Key network metrics: density, homophily, in-degree, out-degree.

13:00–14:00 – Lunch break

14:00–15:30 – Stata Lab Session 3
Estimation of time-varying networks using Stata.
Lecturer: Professor T. Panagiotidis

Day 5 (Friday 17/07/2026)

09:30–11:00 – Lecture 7: Estimating Network Models
Lecturer: Professor V. Sarafidis

• High-dimensional estimation methods (Lasso, Elastic Net, Boosting).
• Visualising network structures.
• Detection of homophily patterns.
• Implementation in Stata.

11:00–11:30 – Coffee break

11:30–13:00 – Lecture 8: Extensions
Lecturer: Professor V. Sarafidis

• Dominant units and latent common factors.
• Robustness checks and diagnostics.
• Policy implications: how shocks propagate across networks.

13:00–14:00 – Lunch break

14:00–15:30 – Stata Lab Session 4
Full workflow example – from geographic contiguity models to data-driven estimation of latent networks using Stata.
Lecturer: Professor T. Panagiotidis

Readings

A reading list will be provided to the participants upon registration to the course.

ECTS Credits

Participants who successfully complete the course can earn 4 ECTS credits (reported in the participation certificate). The grade is based on the evaluation of the instructors. Participants should check with their own University whether these ECTS credits are transferable.