RT Generic T1 Networking entrepreneurs A1 Vega-Redondo, Fernando A1 Pin, Paolo A1 Ubfal, Diego A1 Benedetti, Priscilla A1 Domínguez, Magdalena A1 Rubera, Gaia A1 Hovy, Dirk A1 Fornaciari, Tommaso A2 Universidad Carlos III. Departamento de Economía, AB Can peer interaction foster entrepreneurship in large-scale environments? This paper addresses the question empirically and theoretically. Empirically, we tested the effects of peer interaction on the number and quality of business plans submitted in a Pan-African RCT including 5,000 entrepreneurs. Thetreatment provided the possibility of interaction in different interaction settings (face-to-face or virtual, peers being of the same or diverse nationalities). We find that, while estimated network effects are almost always strong, the treatment effect is not so and displays a non-monotone trade-off between diversity and interaction "bandwidth." We develop a model that sheds light on this behavior by differentiating between constructive and disruptive interaction. It is also qualitatively supported by our experimental evidence SN 2340-5031 YR 2024 FD 2024-06-06 LK https://hdl.handle.net/10016/43954 UL https://hdl.handle.net/10016/43954 LA eng NO Preliminary versions of this paper have been presented at the Paris School of Economics, Queen Mary University, University of Alicante, the Barcelona GSE Summer Forum, Boston University, NYU at Abu Dhabi, Central European University, University of Wisconsin at Madison, University of California at Berkeley, and Stanford University. We thank the useful feedback provided in all of them. In particular, Yann Bramoulle and Marcel Fafchamps were specially helpful. An extensive account of the empirical results discussed in the first part of the paper was first circulated in the report by Vega-Redondo et al. (2019). They were the result of a field experiment that was funded by the PEDL program, received ethical approval from the ethics committee of Bocconi University, and were registered in the American Economic Associations registry (ID AEARCTR-0004017). We want to thank Chris Woodru , the Scientific Coordinator of the PEDL program, who provided both early encouragement and later valuable advice in the realization of the experiment. We also want to acknowledge the support of Village Exchange Ghana { and in particular its director Christiane Milev, as well as Kofi Nyalimba and Richard Kpobi { for the small pilot experiment in Ho (Ghana) that started our experimental work. In that first pilot, the research team included Fernando Vega-Redondo, Cristiana Benedetti-Fasil, Emmanuele Colonnelli, Paolo Pin, and Nicholas Yannelis. At a later stage, Chris Roth, Adam Telek, and Alessio Muscillo also provided useful feedback and help. And throughout the whole experimental work, we have been fortunate to enjoy superb assistance by an RA team working on it. It included (at di erent phases) the following members: Silvia Barbareschi, Chiara Dall'Aglio, Domenico Fabrizi, Gianluca Gaggioti, Francesca Garbin, Long Hong, Ruggero Jappelli, Giulia Mariani, Nelson Mesker, Laura Montenovo, The Lin Bao Nguyen, Giulia Olivero, Massimo Pulejo, Andrea Pugnana, Tiziano Rotesi, Ferran Vega-Carol, and Krzysztof Zaremba. Naturally, all those mentioned here are innocent of any mistakes or misunderstandings that might still remain. F. Vega-Redondo acknowledges financial support by grant CEX2021-001181-M financed by MICIU/AEI /10.13039/501100011033, María Zambrano Program 2021, and Comunidad de Madrid, grant EPUC3M11 (V PRICIT). DS e-Archivo RD 16 jul. 2024