Martin Weidner - Oxford University

Abstract

This paper reviews identification strategies for binary choice logit models with fixed effects in panel and network data settings. We consider both static and dynamic models, allowing for general fixed effect structures, including individual effects, time trends, and two-way or dyadic effects. A central challenge in these models is the incidental parameter problem, which arises when the number of fixed effects grows with the sample size. {\color{red} We discuss the two main approaches to eliminating the nuisance parameters completely}: conditional likelihood methods, which eliminate fixed effects by conditioning on sufficient statistics, and moment-based methods, which construct fixed-effect-free moment conditions. We summarize key results from the existing literature and illustrate how these methods apply across a range of models. 

Additional information:

• Speaker: Martin Weidner
• Time: Thursday, 03.07.2025, 16:00 - 17:00
• Location: Faculty Lounge, Room 0.036
• Further links:
  • Martin Weidner's webpage
• Organizer: Statistics Group
• Contact:
  • Almut Lunkenheimer
  • +49 228 73-9228
  • ifs@uni-bonn.de