NBER WORKING PAPER SERIESFROM GROUNDNUTS TO GLOBALIZATION: A STRUCTURAL ESTIMATE OFTRADE AND GROWTHChristian BrodaJoshua GreenfieldDavid WeinsteinWorking Paper 12512http://www.nber.org/papers/w12512NATIONAL BUREAU OF ECONOMIC RESEARCH1050 Massachusetts AvenueCambridge, MA 02138September 2006Christian Broda is at the University of Chicago, Graduate School of Business; Joshua Greenfield isat Columbia University, and David Weinstein is at Columbia University and the NBER. Broda andWeinstein are grateful to the National Science Foundation for its support of this research (NSF grant#0214378). We wish to thank Daron Acemoglu, Charles Jones, Andres Rodriguez-Clare, Robert Feenstra,Elhanan Helpman and Paul Romer for excellent comments. The views expressed herein are those ofthe author(s) and do not necessarily reflect the views of the National Bureau of Economic Research.© 2006 by Christian Broda, Joshua Greenfield, and David Weinstein. All rights reserved. Short sectionsof text, not to exceed two paragraphs, may be quoted without explicit permission provided that fullcredit, including © notice, is given to the source.From Groundnuts to Globalization: A Structural Estimate of Trade and GrowthChristian Broda, Joshua Greenfield, and David WeinsteinNBER Working Paper No. 12512September 2006JEL No. E00,F43,O4ABSTRACTStarting with Romer [1987] and Rivera-Batiz-Romer [1991] economists have been able to model howtrade enhances growth through the creation and import of new varieties. In this framework, internationaltrade increases economic output through two channels. First, trade raises productivity levels becauseproducers gain access to new imported varieties. Second, increases in the number of varieties drivesdown the cost of innovation and results in ever more variety creation. Using highly disaggregate tradedata, e.g. Gabon's imports of Gambian groundnuts, we structurally estimate the impact that new importshave had in approximately 4000 markets per country. We then move from groundnuts to globalizationby building an exact TFP index that aggregates these micro gains to obtain an estimate of trade onproductivity growth for each country. We find that in the typical country in the world, new importedvarieties account for 15 percent of its productivity growth. These effects are larger in developing countrieswhere the median impact of new imported varieties equals a quarter of national productivity growth.Christian BrodaUniversity of ChicagoGraduate School of Business5807 South Woodlawn AvenueChicago, IL 60637and [email protected] GreenfieldColumbia [email protected] WeinsteinColumbia University, Department of Economics420 W. 118th StreetMC 3308New York, NY 10027and [email protected]. Introduction Economists have long postulated that trade may raise growth. However, it was not until the work of Romer [1987] and Rivera-Batiz and Romer [1991] that we had a general equilibrium model that would let us understand how trade might bring this about. While this seminal work has spawned the development of the vast endogenous growth literature, it has fallen short of one of its main objectives. As Rivera-Batiz and Romer state in their opening paragraph, they did not write their article because no one had thought that new traded varieties could foster growth – after all, the idea was already widely believed at the time – but rather because “it would be difficult for any of us to offer a rigorous model that has been (or even could be) calibrated to data.” Seen in this light, the failure of anyone to calibrate or structurally estimate their model (or even related models, e.g. Jones [1995]) means that while most economists continue to believe that globalization raises growth, we still know precious little about the magnitudes or mechanisms (c.f. Easterly and Levine [2001]). A major reason for this failure stems from the difficulty of bridging the gap between micro and macro evidence on the effects of trade on growth. Trade and development economists have estimated rich models of what happens to particular firms and individuals in particular cases of liberalization, but it is very hard to generalize from these detailed econometric case studies. Empirical macroeconomic studies, by contrast, have tended to use “one regression fits all” specifications to examine how globalization affects growth around the world, but their grand assumptions leave us with little to say about the precise mechanisms underlying the results. It is as if the only way that economists can view the world is through microscopes and telescopes. As a result, these two strands of literature try to describe the same processes, but they have little influence on each other, and a deep skepticism has developed regarding the robustness of the links between trade and growth. For example, in the excellent survey by Hallak and Levinsohn [2004], the authors identify three main classes of “basic methodological shortcomings” in the cross-country evidence. First, trade policy or openness is typically summarized by a one-dimensional index that has little theoretical foundation. Second, there are severe omitted variables biases, which lead to results that are not robust (c.f. Sala-i-Martin [1997], Rodriguez and Rodrik [2001], Noguer and Siscart [2005, and especially 2006]). Finally,2 there is so much heterogeneity in economic conditions across countries that it is doubtful that there is a unique mapping of trade into growth. The problems in the cross-country growth regressions have lead many researchers to focus their attention on specific cases of liberalizations. These “micro-econometric studies” of particular sectors such as “groundnuts”, constitute a second approach to understanding globalization. The enormous advantage of these studies is that they can provide rich and compelling econometric case studies of particular liberalizations. In the best examples, one can often be extremely precise about what exactly is being estimated and how the pieces fit together. The disadvantage is that it is very hard to extrapolate from groundnuts to globalization. This paper can be thought of as a hybrid approach to understanding how trade affects growth. By breaking world trade down into 6-digit bilateral import flows and estimating hundreds of structural parameters per country, we are able to build estimates that preserve the cross-country and cross-industry richness of the global economy. The