چه کسی از یک دنیای جهانی شده می ترسد؟ سرمایه گذاری مستقیم خارجی، دانش محلی و تخصیص استعداد

We study the distributional effects of globalization within a model of heterogeneous agents where both managerial talent and knowledge of the local economic environment are required in order to become a successful entrepreneur. Agents willing to set up a firm abroad incur a learning cost that depends on how different the foreign and domestic entrepreneurial environments are. In this context, we show that globalization fosters FDI and raises wages, output and productivity. However, not everybody wins. The steady state relationship between globalization and income is U-shaped: high- and low-income agents are better off in a globalized world, while middle-income agents (domestic entrepreneurs) are worse off. Thus, consistently with recent empirical evidence, the model predicts globalization to increase inequality at the top of the income distribution while decreasing it at the bottom.

Who opposes globalization? Who favors it? It is well-known that in a Hecksher–Ohlin context the process of globalization produces winners and losers as a consequence of the changes in the relative abundance of factors. Despite its obvious relevance, this issue has been so far hardly analyzed in the context of intraindustry trade models à laMelitz (2003), where gains from trade do not arise from international differences in factor endowments, but from consumers' love for variety and from the ability of the entrepreneurs to overcome the barriers that distance generates. So far this literature has focused on models with “heterogeneous firms” and “homogeneous agents”. This paper is an attempt to analyze the distributional effects of globalization within a Melitz-type model with heterogeneous agents. Our main finding is that the effect of globalization on the individuals' well-being is non-monotonic. A higher degree of inter-connectedness among countries has a U-shaped effect on the income distribution, improving the position of both those at the top and the bottom of the distribution and harming those in the middle. This prediction is consistent with recent empirical evidence showing that since the 1990s both in the U.K. and the U.S. inequality went up in the upper tail of the distribution and decreased in the lower tail ( Autor et al., 2005, Autor et al., 2006 and Machin and Van Reenen, 2007). We obtain this result in a model of Foreign Direct Investments (FDI), one of the most prominent (and debated) features of globalization. FDI grew dramatically over the last decades far outpacing the growth of trade and income.2 Another salient feature of FDI is that they take place mostly between developed countries, i.e. between countries that are similar in terms of natural endowments and relative supply of inputs.3 We provide additional empirical evidence in line with this fact, documenting that bilateral FDI are also higher between countries that have more similar entrepreneurial environments.4 Consistently, we propose a model in which both managerial talent and knowledge of the local entrepreneurial environment are required in order to set up a firm and earn positive profits. The main trade-off that arises in the model depends on how individuals with different abilities are allocated to the different types of jobs available in the economy. To be more specific, a first key feature of the model is that agents with different levels of managerial ability are allowed to select their occupation and choose whether to become entrepreneurs or workers. Those who become entrepreneurs may engage in FDI and set up a firm abroad. However, in order to become a successful entrepreneur in a given country, managerial ability is not sufficient: some knowledge of the local economic environment is also required. A second key feature of the model is that agents are assumed to know more about the domestic economic environment (e.g. domestic consumers' tastes) than about the foreign environment. Domestic agents have to learn how the foreign economic environment works in order to profitably set up a firm abroad. Thus, both managerial ability and nationality affect career choices. The idea is that a certain level of managerial talent, though allowing agents to profitably produce within the domestic economic environment, may be of little help when setting up a firm abroad: the more different the foreign and the domestic economic environments, the more difficult it is to succeed in the foreign market. This distance between entrepreneurial environments is the only explicit barrier to capital movements that matters in the model. It may be overcome only at the cost of learning how the foreign environment works. Of course, in equilibrium, only the most talented entrepreneurs, who run – in line with the empirical evidence – the largest and most productive firms, have incentives to pay the learning cost and produce abroad. The model endogenously determines the allocation of talents between (domestic and international) entrepreneurial activity and salaried work. It follows that FDI, Total Factor Productivity (TFP), GDP and wages depend on how efficiently talents are allocated. Talent allocation, in turn, depends on how hard it is to learn about the foreign entrepreneurial environment. A lower distance between entrepreneurial environments reduces the learning cost and raises the inflow of foreign-owned firms into the domestic market. This increases the domestic wage and makes the entrepreneurial activity less profitable, driving a fraction of low-ability domestic entrepreneurs out of the market. This general equilibrium effect improves the allocation of talents and increases both TFP and GDP.5 Conversely, a larger distance between entrepreneurial environments protects low-ability entrepreneurs from foreign competitors and reduces output, wages and TFP. Thus, globalization fosters aggregate efficiency. Still, not everybody wins when the degree of globalization increases. The welfare of the individuals with the lowest and highest levels of entrepreneurial talent (who choose to become workers and multinational entrepreneurs, respectively) increases as learning costs go down and GDP, TFP and wages rise. Differently, the welfare of the individuals with an “intermediate” level of talent is decreasing in the degree of globalization. The reason is that, in a globalized world, domestic entrepreneurs pay the cost of tougher competition without enjoying the benefits of accessing to wider markets. In a non-globalized world they enjoy higher entrepreneurial profits as they are sheltered from foreign competition. Even in the absence of any pro-competitive effects of FDI working through lower prices,6 the general equilibrium effect through wages is sufficient to expel mediocre entrepreneurs from the market when the difference between economic environments becomes smaller. As a result, globalization increases inequality in the upper tail of the distribution and decreases it in the lower tail of the distribution. As the model is based on the idea that globalization reduces the distance between economic environments and therefore leads to higher FDI, we test this relationship against the data. We proxy the distance between economic environments exploiting measures of Product Market Regulation and interpret the difference between languages as an additional qualitative proxy of the distance between economic environments. Our results indicate that, controlling for the levels of regulation, GDPs and populations in both countries, host and source countries fixed effects, time effects, and a set of geographical variables, a higher distance between economic environments negatively affects the size of bilateral FDI stocks. This paper is obviously related to the recent trade literature that, since Melitz (2003), develops dynamic industry models with heterogeneous firms, in which only the most efficient firms engage in cross-border activities and where more openness forces the least productive firms out of the market. The key difference with Melitz (2003) is that in this paper firms' heterogeneity stems from the heterogeneity (in managerial talent) of the agents who are allowed to make career choices. This feature of the model allows us to emphasize the (endogenous) mechanism by which exposure to foreign competition improves the allocation of talents and, most importantly, to discuss the distributional implications of globalization. Another strand of literature related to this paper is the one that analyzes the distributional effects of decreasing trading costs. While this issue has been widely studied in the context of models à la Hecksher–Ohlin, the literature on the distributional effects of globalization in the context of intraindustry trade models is much thinner. Closest to us is Helpman et al. (2010) that study the distributional consequences of international trade in a model with heterogeneous firms and workers in which labor markets are imperfect.7 In their context, the distributional effects of globalization in developed economies are akin to those derivable in Hecksher–Ohlin models: the most efficient workers benefit from globalization because their firms (the most efficient ones) do, while the least skilled workers suffer because their firms (the least efficient ones) also suffer. One key difference between our approach and their model is that we allow for endogenous career choices and learning of the foreign environment. Thus, in our context the welfare effects of globalization are U-shaped. The individuals at the low-end of the income distribution improve their position because the demand for their labor services is larger when foreign firms have access to the local market. We finally relate to the extensive (theoretical and empirical) literature that studies the driving factors of FDI.8 The rest of the paper is organized as follows. Section 2 presents new empirical evidence on the effects of the distance between economic environments on FDI. Section 3 describes the model. Section 4 solves for the closed economy while Section 5 discusses the open economy framework. Section 6 analyzes the distributional effects of globalization and Section 7 concludes.

This paper first presents empirical evidence on the positive effect of cross-country proximity in “entrepreneurial environments” on bilateral FDI. By exploiting the OECD International Direct Investment Statistics and data on nationwide regulation levels from the OECD and the World Bank, we find evidence that larger similarities in the economic environment tend to be associated with larger bilateral FDI, after controlling for the levels of regulation in both countries, for countries fixed effects and for time effects. Motivated by this evidence, we build a general equilibrium model that allows us to study the distributional effects of globalization. In the model, agents are heterogeneous and differ both in their ability to be entrepreneurs and their nationality. Entrepreneurs may set up a firm abroad, i.e. engage in FDI. If they do so, they incur in the additional cost of learning how the foreign environment works. In this framework, globalization fosters FDI and improves the allocation of talents in the economy boosting wages, output, and productivity. However, not everybody wins. Low-ability agents always gain, as the demand for their labor services increases. Individuals in the middle of the distribution are always worse-off: they lose a valuable asset (exclusive knowledge of the local economy) without gaining anything, as they are not able to compete with high-ability agents in a globalized world. High-ability individuals typically win, and their relative position vis-à-vis the middle-income individuals always improves. Even though they pay higher wages, they reap the benefits of accessing to larger markets: they lose on one asset (knowledge of the local economy) but gain on a second one (cheaper access to the foreign economy). Thus, consistently with recent empirical evidence, the model predicts globalization to have a U-shaped effect on the distribution of income, worsening the situation of the middle class vis-à-vis the top and the bottom of the distribution. We leave a number of interesting questions for future research. First, the model suggests that the dynamics of globalization are interesting per se. Fig. 3 implies that the opposition to a marginal increase in globalization is smaller the larger the level of globalization, suggesting that the number of opponents should decrease over time, as more and more individuals benefit from globalization by either becoming workers or starting to invest abroad. This question is best addressed looking at the transitional dynamics of a full-blown dynamic model. Second, the model may be extended to endogenize the proximity between economic environments in order to analyze the process of nation-building that derives from the fragmentation of a larger unit. We intend to explore these issues both at the empirical and the theoretical level.