Growth and Human Capital: A Network Approach


  • We are grateful to Aqib Aslam, Paul Beaudry, Yann Bramoullé, Vasco Carvalho, Francesco Caselli, Edouard Challe, Dean Corbae, Tom Crossley, Tatiana Damjianovic, Partha Dasgupta, Steven Durlauf, Alessandra Fogli, Ben Golub, Sanjeev Goyal, Jonathan Heathcote, Migel Leon-Ledesma, Marco van der Leij, Omar Licandro, Yan Liu, Hamish Low, Juan-Pablo Nicolini, Kjetil Storesletten, José Tavares, Richard Todd, Frederik Toscani, Flavio Toxvaerd and Klaus Waelde for interesting conversations and insightful comments on this project. We also thank participants at the 2nd Cambridge-Oxford-London Macro Meeting, the 2009 SAET meeting, 18th CEPR ESSIM, the 10th CRETE conference, the 16th CEF international conference, and seminar participants at University of Manchester, University of Copenhagen, University of Illinois in Urbana-Champaign, UC Davis, SUNY in Stony Brook, University of Oxford, CREI Barcelona, Universitat Autonoma de Barcelona, Federal Reserve Bank of Minneapolis, University of Glasgow, University of Bristol, Copenhagen Business School, City University London, PIMES/UFPE, CREST/École Polytechnique, GREQAM and Universidad Torcuato di Tella. A part of this article was completed during Chryssi Giannitsarou's visit to the Federal Reserve Bank of Minneapolis – she is grateful for their hospitality.


We study the interactions of human capital, growth and inequality by embedding networks into an endogenous growth model with overlapping generations. Human capital depends on investment in education and the average human capital of a household's neighbourhood. High network cohesion leads to long-run equality, while for low network cohesion inequality is high and persists more often. During transition, high overall growth is achieved when the network has high degree centralisation, and high individual growth is achieved when the household has low human capital relative to its neighbourhood and is located in a neighbourhood with high average human capital.