Moment Inequality Estimators in Competition Analysis: A Practitioner’s Guide with Synthetic Data Applications
3. Uluslararası Veri Analitiği ve Makine Öğrenmesi Konferansı (DATAMACLEA’26), Ankara, Türkiye, 11 - 13 Mayıs 2026, ss.116-133, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Basıldığı Şehir: Ankara
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.116-133
- Ankara Yıldırım Beyazıt Üniversitesi Adresli: Evet
Özet
Moment inequality estimators have become a central tool in empirical industrial organization yet remain largely inaccessible to competition policy practitioners. This paper provides a self-contained tutorial on moment inequality methods for antitrust economists, competition authorities, and economic consultancies. I explain the conceptual shift from point identification to set identification, present the mathematical foundations in competition-relevant terms, and provide an implementation template adapted from Canay, Illanes, and Velez (2023). Two synthetic data applications illustrate the methods in core antitrust settings: (i) a twoplayer entry game with multiple equilibria, following Ciliberto and Tamer (2009), where the identified set for competitive-effect parameters is constructed via proper probability bounds over the multiplicity region; and (ii) a stylized logit demand example with an unknown outside-good share, where valid unconditional moments bound the price coefficient and the resulting confidence set is projected into post-merger price effects. Both applications are implemented in Python with valid critical-value computations. I discuss extensions to collusion screening, market definition, and dynamic entry, and address when moment inequality methods outperform standard techniques and how to communicate set-identified results in enforcement proceedings.