The Real Effects of Government-Owned Banks: Evidence from an Emerging Market



    Search for more papers by this author
    • Carvalho is at Marshall School of Business, University of Southern California. I am indebted to my advisors Jeremy Stein, Andrei Shleifer, Michael Kremer, and Gary Chamberlain for numerous discussions. I would like also to acknowledge helpful comments from Alberto Alesina, Effi Benmelech, Shawn Cole, Harry DeAngelo, Mihir Desai, Serdar Dinc, Mark Garmaise, Oliver Hart, Campbell Harvey (the Editor), Larry Katz, Asim Khwaja, John Matsusaka, Paul Niehaus, Kevin J. Murphy, Walter Novaes, Oguzhan Ozbas, and David Scharfstein, an Associate Editor, and an anonymous referee. This work benefited greatly from seminar participants at Brown, Dartmouth College Tuck, Harvard, Stockholm School of Economics, University of Illinois at Urbana-Champaign, USC Marshall, Washington University at St. Louis Olin, Yale SOM, the USC-UCLA Finance Day, and the Western Finance Association conference. Many thanks to Bruno Araujo, Joao De Negri, and IPEA for granting me access to data. All errors are mine.


Using plant-level data for Brazilian manufacturing firms, this paper provides evidence that government control over banks leads to significant political influence over the real decisions of firms. I find that firms eligible for government bank lending expand employment in politically attractive regions near elections. These expansions are associated with additional (favorable) borrowing from government banks. Further, these persistent expansions take place just before competitive elections, and are associated with lower future employment growth by firms in other regions. The analysis suggests that politicians in Brazil use bank lending to shift employment towards politically attractive regions and away from unattractive regions.