An econometric model of link formation with degree heterogeneity

The first draft of this paper dates to March 2014 and was titled “An empirical model of network formation: detecting homophily when agents are heterogenous”. A Python 2.7.11 module implementing the methods described in this paper is available for download from my web page. I thank Jinyong Hahn for sharing an unpublished appendix of Hahn and Newey (2004), Joachim De Weerdt for generously making his Nyakatoke dataset available, Eric Auerbach, Peter Bickel, Stephane Bonhomme, Jesus Carro, Gary Chamberlain, Xiaohong Chen, Nicholas Christakis, Ivan Fernandez-Val, Jim Heckman, Yuichi Kitamura, Elena Manresa, Costas Meghir, Roger Moon, Jong-Myun, Matt Shum, Martin Weidner and Bin Yu for numerous helpful suggestions. This paper has also benefited from the reactions of audiences at the 2014 Seattle-Vancouver Econometrics Conference, the 2014 All California Econometrics Conference, the 2015 Econometrics Journal Conference on Networks, the 2016 BeNet Conference at UCL/CORE, as well as from participants in seminars and workshops at New York University, UC Berkeley, the Toulouse School of Economics, Oxford, Cambridge, St. Gallen, CEMFI, the Yale Institute of Network Sciences, Stanford, USC, UCSD, Chicago, Michigan, Brown and Harvard-MIT. This revision incorporates many useful suggestions made by the co-editor, Elie Tamer, and four anonymous referees. Finally, I am especially grateful for helpful conversations and suggestions from my three immediate colleagues in the Economics Department at Berkeley: Michael Jansson, Demian Pouzo and Jim Powell. Financial support from NSF grant SES #1357499 is gratefully acknowledged. All the usual disclaimers apply. The views expressed herein are those of the author and do not necessarily reflect the views of the National Bureau of Economic Research.

• online appendix

• July 24, 2014
• August 4, 2015