In this paper, we propose a new perspective to analyze the impact of institutions, environmental standards, and globalization on relocations of polluting rms in countries with lax environmental regulation (called pollution havens). Via a simple theoretical extension from the Economic Geography literature, we characterize the main features of pollution havens: a good market access to high-income countries and corruption opportunities. Using structural and reduced-form estimations, we analyse these determinants by exploiting a unique database on the number of European aliates located abroad. A 1% increase in access to the European market from a pollution haven fosters relocation there by 0.1%. We also nd that corruption in these countries lowers environmental standards, which strongly attract polluting rms: a 1% increase in this indirect eect of corruption fuels relocation by 0.28%. We test the economic signicance of these empirical ndings via simulations. The protection of the European market (e.g., a carbon tax on imports) to stop relocations to pollution havens must be high (a decrease of the European market for Morocco and Tunisia equivalent to 13%) not to say prohibitive (31% for China).Keywords: Multinational rms; Environmental Regulation; Europe; Corruption; Market Access, Trade.JEL: F12;Q5;Q53 * Corresponding author: fabien.candau@univ-pau.fr † We thank Simone Bertoli, Antoine Bouët, Pierre-Philippe Combes, Matthieu Crozet, José de Sousa, CarlGaigné, Michaël Goujon, Jacques Le Cacheux, Serge Rey, Frédéric Robert-Nicoud, François-Charles Wol for their attentive reading and comments. We also thank seminar and conference participants at CERDI, ETSG, U. Geneva, U. Toulon.

The contribution of this paper is twofold. Firstly, we explore the e¤ects of trade liberalization and commuting costs on the location of entrepreneurs. The model reveals a dispersion-agglomeration-dispersion con…g-uration when trade gets freer. Furthermore we prove that when both commuting costs and trade integration are high, then dispersion Pareto dominates agglomeration. Secondly, we use this framework to investigate the e¤ect of trade on corruption at di¤erent levels of democracy and instability. We show that corruption is bell-shaped with respect to trade liberalization in stable and democratic regimes but also in unstable dictatorships.JEL classi…cation: D73; H25; R12; F12

By introducing heterogeneity between high-skilled mobile and low-skill immobile workers in a model with land rent and commuting costs we develop a new model in the …eld of the new economic geography. This model, which can reveal a dispersionagglomeration-dispersion con…guration when trade gets freer, is used to compare via the Pareto criterion the two possible market outcomes, i.e. agglomeration and dispersion. When these equilibria can be ranked it is shown that dispersion can be a Pareto-e¢ cient outcome. JEL classi…cation: F12; R13

1 Simonovska (2015) proposes a model with non-homothetic preferences from a hierarchic choice of consumption (Jackson, 1984). In this model, where the marginal utility is bounded (consumers may have zero demand for some varieties, see also Saur e, 2012), the relative price of a variety is higher in relatively richer markets which contradicts Hummels and Lugovskyy's (2009) results.

This study is a theoretical and empirical analysis of the effects of regional trade integration on the spatial distribution of skills. We first develop a theoretical model in the economic geography field to integrate heterogeneous workers, housing, local entrepreneurs and skill upgrading by unskilled workers. We then analyse how the domestic integration of each state in the U.S., approximated by truck registrations, influenced the location choice of skilled and unskilled workers in 1940-1960. By using inter-and intrastate trade flow from the U.S. Commodity Flow Survey, we also analyse the impact of regional trade costs for the contemporary period (1997, 2002, 2007). The theoretical model shows that the bell-shaped curve of spatial development displays a sorting of individuals and firms. Only high-skilled workers increasingly choose the core region during the process of regional integration while intermediate-skilled workers move to the periphery due to the increase in the price of housing. By impacting differently on the opportunity cost to invest in skill acquisition in the core and the periphery, this sorting influences the regional creation of human capital. First a regional divergence in education investment occurs, and then a convergence, but only for high level regional integration. The empirical analysis confirms that regional trade integration has been a determinant of the spatial distribution of skills in the United States.