Geography, Technology, and Optimal Trade Policy …

1 geography , technology , and Optimal TradePolicy with heterogeneous FirmsAntonella Nocco, Universit del Salento Gianmarco Ottaviano, LSE, University of Bologna, CEP and CEPRM atteo Salto, European CommissionThis version: December 2015 PRELIMINARY AND INCOMPLETE DRAFTDO NOT CITE WITHOUT AUTHORS PERMISSIONA bstractHow should multilateral Trade Policy be designed in a world in whichcountries di er in terms of technology and geography , and rms withmarket power di er in terms of productivity? To answer this question wedevelop a normative analysis of the model by Melitz and Ottaviano (2008)in the case of an arbitrary number of asymmetric countries. We rst takethe viewpoint of a global planner maximizing global welfare constrainedonly by technology and endowments. We characterize the corresponding rst best allocation in terms of cross-country specialization, Trade pat-terns, rm productivity distribution and product variety.

2 We show howthe rst best allocation departs from the free market equilibrium alongall four dimensions, pinning down the Policy tools needed for its decen-tralization. As these tools are hardly enshrined in any real multilateraltrade agreement, we then study alternative scenarios, in which the set ofavailable Policy tools is increasingly :international Trade , Trade Policy , monopolistic competition,heterogeneity, selection, Classi cation:D4, D6, F1, L0, L1. Nocco: University of Salento, Department of Management, Economics, Mathematics andStatistics, Ecotekne, via Monteroni, 73100, Lecce, ITALY, Ot-taviano: London School of Economics, Department for Economics, Houghton Street, WC2A2AE, UK, Salto: European Commission, Rue de la Loi 200, B- 1049,BELGIUM, We are grateful to participants at the seminar of theCenter of th orie conomique, mod lisation et applications (THEMA) of the Universit Cergy-Pontoise, the XX DEGIT Conference (Graduate Institute Geneva, Universit de Gen ve, 3-4 September 2015) and the ETSG Conference 2015 (Universit Paris 1 Panth on-Sorbonne, 10-12 September 2015), for useful comments and suggestions.

3 The views expressed here are thoseof the authors and do not represent in any manner the EU IntroductionHow should multilateral Trade Policy be designed in a world in which countriesdi er in terms of technology and geography , and rms with market power di erin terms of productivity? Should Policy tools di er in kind or implementationbetween more and less developed countries? Should smaller less productive rms be protected against larger more productive (foreign) rivals? Should na-tional product diversity be defended against competition from cheaper importedproducts? What is the Optimal degree of product diversity on a global scale?To answer these and other related questions we propose a normative analysisof the monopolistic competition model by Melitz and Ottaviano (2008). Thismodel exhibits several features useful for our purposes.

4 As even with heteroge-nous rms it is analytically solvable with all sorts of asymmetries in countrysize, technology and accessibility, the model allows for transparent comparativestatics. As it features variable markups, it allows for rm heterogeneity to be-come a key source of income e ects are neutralized due toquasi-linear preferences, constant marginal utility of income allows for a con-sistent global welfare analysis based on a straightforward de nition of globalwelfare for an economy with heterogeneous countries as the sum of all individ-uals indirect utilities. While the absence of income e ects gives our results apartial equilibrium avor, this approach shares the focus on social surplus withmainstream Policy and Ottaviano (2008) model an economy in which countries di er interms of market size, barriers to international Trade and state of are active in two sectors with labor as the only production factor.

5 A traditional perfectly competitive sector supplies a freely traded homogeneousgood. A modern monopolistically competitive sector supplies varieties of ahorizontally di erentiated good. In each country the productivity of entrantsin the modern sector is dispersed around a country-speci c mean dictated bythe national state of technology , which thus de nes the country s comparativeadvantage in the modern sector with respect to the other countries. More-over, in the modern sector countries face di erent physical transport costs fortheir international and domestic Trade as dictated by geography . In equilibriumtechnology and geography endogenously determine a country s intersectoral spe-cialization as well as its patterns of inter- and intra-industry Trade . They alsodetermine the productivity distribution of a country s modern producers as wellas the number and the prices of varieties the country s consumers have accessto and the dimension of the di erentiated good sector in each and Ottaviano (2008) do not provide any normative analysis of theirmodel.

6 The aim of this paper is to show that lling this gap yields new insightson Optimal multilateral Trade Policy . In particular, we rst take the viewpointof a global planner maximizing global welfare constrained only by technologyand endowments. We characterize the corresponding rst best allocation interms of cross-country specialization and rm productivity distribution, tradepatterns, size of the di erentiated good sector and product variety. We show howthe rst best allocation departs from the free market equilibrium along all thesedimensions, pinning down the Policy tools needed for its decentralization. Theseinclude country-speci c lump-sum instruments for both rms and consumers as1 See the discussion in Nocco, Ottaviano and Salto (2014).2well as production and consumption subsidies and taxes di erentiated across rms both between and within the wide array of Policy tools needed to decentralize the rst best is hardlyenshrined in any real multilateral Trade agreement, we then study alternativemultilateral n-th best scenarios, in which the set of available Policy tools isincreasingly restricted.

7 In the second best , lump-sum instruments are stillavailable but subsidies and taxes can not be di erentiated across rms. In the third best , also lump-sum instruments for rms are ruled out. Finally, whenalso lump-sum tools for consumers are removed, the economy can not be drawnaway from the free market analysis speaks to two main literatures. First, there is the literature onoptimal Trade Policy both under perfect competition (see, , the discussion inCostinot, Donaldson, Vogel and Werning, 2014) and imperfect competition (see, , Grossman, 1992). This literature does not feature an arbitrary number ofcountries and rms, and rm heterogeneity has been introduced only recentlyin models that analyse the impact of non wasteful import tari s (that is, tradebarriers that generate revenues) in monopolistically competitive case, tari s are either set unilaterally by a country or in a non-cooperativeframework in a setup that considers the case of a small open economy or that oftwo large open economies.

8 This literature basically explores the e ects producedin a framework with heterogeneous rms by Policy intervention on the terms oftrade ( a tari produces an increase in the domestic wage that creates an ar-gument in favour of Policy intervention) and/or on the monopolistic competitiondistortion ( a tari on imports, or a subsidy on domestic sales, increases thenumber of varieties that are o ered in the market, where variety is ine cientlylow and, thus, corrects the entry distortion and, eventually, also the mark-updistortion).3 Using the Melitz and Ottaviano (2008) model, Bagwell and Lee(2015) analyze Trade policies nding that in the case of two symmetric countriesthere is an incentive for a country to introduce a small import tari and theyalso identify the conditions under which two symmetric countries have unilateralincentives to introduce beggar-thy-neighbor export subsidies.

9 Considering thecase of symmetric Trade policies, they nd that global free Trade is generally note cient. Within the same framework, Demidova (2015) nds that the intro-duction of pro-competitive gains from Trade due to variable markups does notchange the relationship between the impact of unilateral Trade liberalization onthe welfare level of the home country and the presence of an outside good: witha wasteful import tari , a unilateral reduction in Trade tari increases welfarein the absence of an outside good both in the case of two large economies and inthe case of a small home economy; with non wasteful import tari , on the con-trary, in the absence of an outside good unilateral Trade liberalization reduceswelfare at home in both cases. Our perspective is, however, di erent from that2 The results with homogeneous rms and monopolistic competition can be recalled makinguse of the words by Felbermayr, Jung and Larch (2013, p.)

10 13) who write that: "[f]irst, Flamand Helpman (1987) show that tari s correct for a mark-up distortion due to monopolypricing. Second, Venables (1987) argues that a tari can induce welfare-enhancing additionalentry. Third, as in other Trade models, there is a terms-of- Trade rationale for import tari s;see Gros (1987).3 The small-country analysis with one sector built on the Melitz (2013) model started byDemidova and Rodriguez-Claire (2009) has been extended to consider two large countries byFelbermayr, Jung and Larch (2013) or to introduce an outside good sector by Haaland andVenables (2015).3taken in all these works, as we analyze the "global" welfare properties of theMelitz and Ottaviano (2008) model identifying the ine ciencies of the marketoutcome and showing how the di erent n-th best allocations can be attainedwith di erent set of Policy , there is the literature on Optimal product variety without rmsheterogeneity (Spence, 1976, and Dixit and Stiglitz, 1977) and with rm het-erogeneity (Dhingra and Morrow, 2012; Melitz and Redding, 2012).