اثرات جانبی صنعت داخل و بین صنعتی از سرمایه گذاری مستقیم خارجی در بخش ساخت مکزیک: مدارک و شواهد جدید از مناطق مکزیکی

This paper presents new empirical evidence on externalities from Foreign Direct Investment (FDI) in several Mexican regions in the early 1990s. The main findings are threefold. First, the presence of FDI creates negative externalities within industries and positive externalities between industries through backward linkages. Second, FDI-externalities are stimulated by large technological differences between FDI and Mexican firms and by geographic concentration of industries. Third, we identify a substantial level of regional heterogeneity of the externality impact of FDI, in line with the notion that FDI may have contributed to processes of changing regional prosperity under trade liberalization. The findings also imply that maquiladora firms in the border states are generating positive externalities.

Recent research on the impact from Foreign Direct Investment (FDI) is exploring avenues to improve the statistical identification of FDI spillovers. One development is to broaden the scope of the estimations, by controlling for the full industry dimension of these externalities. The recognition of the importance of this dimension translates into distinguishing between intra- and inter-industry FDI externalities (e.g., Blalock and Gertler, 2008, Driffield et al., 2004 and Smarzynska, 2004). A second development is that empirical studies are focusing on identifying structural factors that influence the occurrence and the nature of FDI externalities (Blomström and Kokko, 2003 and Crespo and Fontoura, 2007). Two factors appear to be commonly accepted as potential determinants. The first factor is the concept of absorptive capacity, which proposes that only those domestic firms or industries in a host economy that possess sufficient technological knowledge and skills are able to absorb new technologies from foreign-owned firms. The second factor is the extent to which FDI and domestic firms are located in geographic proximity. In line with evidence that proximity between economic agents may foster the creation and the transmission of knowledge spillovers (Audretsch & Feldman, 2004), the hypothesis that FDI spillovers are facilitated by such proximity between FDI and domestic firms is easily put forward. The purpose of this paper is to build and extend on these recent developments in empirical research on FDI externalities. Using unpublished and thus far unexplored regional data from the 1994 Mexican manufacturing census, we estimate FDI spillovers in detailed manufacturing industries for a number of regional economies in the Republic of Mexico. The main contribution of our analysis is threefold. First, building on previous empirical findings on FDI externalities among national Mexican manufacturing industries, we present improved evidence of FDI spillovers in regional manufacturing industries. Importantly, in our analysis we are able to distinguish between intra-industry externalities and externalities that arise through backward linkages between FDI and Mexican suppliers, a distinction that is missing in previous research on FDI spillovers in Mexico. Second, we present new evidence on the effects on FDI spillovers of the technology gap between FDI and Mexican firms and the level of agglomeration of industries. Although geographic concentration is increasingly seen as a potentially important determinant of FDI externalities (Barba Navaretti & Venables, 2005), corroborating empirical evidence remains limited. Furthermore, as a recent paper by Crescenzi, Rodriguez-Pose, and Storper (2007) clearly shows, the effect of proximity on externalities may be place and context specific. In their analysis of innovation dynamics, Crescenzi et al.’s findings indicate that whereas proximity is essential for the diffusion of knowledge in the United States, this is much less so in the European Union. Our analysis addresses this issue of proximity by providing new evidence on the relation between agglomeration of industries and FDI externalities. As for the effect of the technology gap, the issue is of a different nature. The notion that the level of absorptive capacity in a host economy is important for positive FDI spillovers is straightforward (Blomström & Kokko, 2003). The problem here lies with the use of technological differences between FDI and domestic firms as proxy for the level of absorptive capacity of domestic firms, as it only partly captures the underlying notion of technology catch up (Jordaan, 2005). The catch up thesis consists of two elements. The first element is the existence of technological differences between advanced and lagging countries, indicating the scope of potential externalities to arise (Abramovitz, 1986 and Gershenkron, 1962). The second element is the level of absorptive capacity of the lagging countries to absorb new technologies (Griffith et al., 2003 and Keller, 1996). Using the technology gap as an indicator of absorptive capacity therefore only relates to the second element of this catch up thesis. This imperfect relation between the technology gap and the catch up thesis generates several interpretational problems. Although a small technology gap may indicate that domestic firms possess sufficient absorptive capacity, it also indicates that the scope of potential externalities is limited. In contrast, large technological differences can stimulate externalities, as “…the greater the backlog of available opportunities to exploit, the greater the pressure for change…” (Findlay, 1978, p. 2). Furthermore, a limited scope of potential externalities does not offer incentives to domestic firms to make externality-facilitating investments, making it less likely that externalities will occur (Rodrik, 1992).1 In addition, one could argue that limited technological differences may lead to intensified competition between FDI and domestic firms, which may promote the occurrence of negative externalities (Aitken & Harrison, 1999). In sum, these arguments point at the possibility that instead of expecting a positive effect from a small technology gap, positive FDI spillovers may be stimulated in those cases where technological differences between FDI and domestic firms are large. Our analysis will present new evidence on this issue, thereby contributing to this important debate.2 Our third contribution is that we explicitly examine whether the externality impact from FDI differs between regions. Following the introduction of trade liberalization in Mexico from the mid 1980s onwards, the Mexican economy underwent significant structural changes. Trade liberalization led to drastic changes in relative regional prosperity: the border states experienced marked economic growth, in contrast to Mexico City which experienced its heydays in the 1970s under import substitution (Chiquiar, 2005, Faber, 2007, Hanson, 1998 and Rodriguez-Pose and Sanchez-Reaza, 2002). Interestingly, the role of FDI spillovers in these structural spatial changes has been largely ignored. Our analysis makes a modest contribution in this respect, by assessing whether and how regional heterogeneity of the externality impact of FDI may have contributed to these changes. The remainder of the paper is constructed as follows. In Section 2, we discuss structural spatial changes in the Mexican economy under trade liberalization, focusing on changes in regional employment shares and on the volume, type, and regional destination of inward FDI. One of the main findings of this section is that FDI has concentrated in those regions that have been most affected by the introduction of trade liberalization, suggesting that FDI may have influenced the diversified spatial outcome of trade liberalization. Section 3 briefly discusses the main findings and limitations of previous research on FDI spillovers in Mexico, describes the dataset that we use in our empirical analysis, explains the specification of the econometric model and defines the variables. Section 4 presents the main empirical findings, of which the following are particularly important. First, contrary to majority of the previous findings for Mexico, the present findings suggest that intra-industry foreign participation creates negative externalities. At the same time, the estimations also provide evidence of positive FDI spillovers through input–output linkages with Mexican suppliers. Second, the findings show that both the level of geographic concentration and the technology gap influence FDI spillovers. Generally speaking, large technological differences promote the occurrence of positive FDI externalities, suggesting that our alternative interpretation of the meaning of technological differences is the more appropriate one. As for geographic concentration, the findings indicate that this factor fosters negative inter-industry externalities, indicating that agglomeration stimulates fierce competition between local Mexican suppliers. Third, there is a considerable level of regional heterogeneity of the impact of FDI. Broadly speaking, the findings show a structural difference between the effects of FDI in the border states and in Mexico City: whereas the border states enjoy additional positive externalities from FDI, Mexico City is subject to additional negative externalities. This regional difference of the impact of FDI suggests that FDI has indeed contributed to the changes in relative regional prosperity in Mexico under trade liberalization. It also implies that the concentration of maquiladora firms in the border states generates positive externalities. Finally, Section 5 summarizes and concludes.

Recent research on FDI externalities has broadened its research scope, by controlling for the effect of intra-industry and inter-industry foreign participation and by obtaining evidence on structural factors that may influence the occurrence of such externalities. In this paper, we have extended on these issues, in the context of estimating FDI externalities among detailed manufacturing industries for several regions in the Republic of Mexico for the early 1990s. The empirical findings are noteworthy for several reasons. First, contrary to previous research on Mexico, we find that—in line with recent research on other host economies—the presence of foreign-owned firms generates two types of externalities: negative externalities within industries and positive externalities through backward linkages. The occurrence of negative externalities can be explained by the occurrence of a negative competition effect, whereas positive externalities through backward linkages occur when FDI offers various types of support to local suppliers. Second, we find that large technological differences promote the occurrence of FDI externalities. This finding cannot be explained by interpreting the technology gap as an inverse indicator of the level of absorptive capacity of Mexican firms. Instead, as we have argued, the positive effect of large technological differences on FDI spillovers indicates the importance of a large scope of potential externalities, which provides incentives to domestic firms to make externality- facilitating investments. Also, it is less likely that industries with large technological differences are subject to negative competition effects. As for the effect of geographic concentration on FDI spillovers, the findings are more diverse. The estimations indicate that the declining region Mexico City is subject to additional negative externalities in industries that are geographically concentrated. We see two possible explanations for the occurrence of this effect in this particular region. One is that the decrease in demand by FDI for local inputs leads to a decrease in the production volume among local suppliers, resulting in an efficiency loss. Second, the geographic concentration of Mexican suppliers may lead to an intensified level of competition, which, possibly in combination with a decreased level of support offered by FDI, may cause domestic firms to offer too many concessions in terms of, for example, price and quality. If so, this would result in negative externalities. In contrast to this, Baja California enjoys additional FDI spillovers in agglomerated industries. This finding is in line with the notion that geographic proximity facilitates the occurrence of positive spillovers. Third, we have assessed our findings against the background of structural regional changes in Mexico following the introduction of trade liberalization. In particular, the border states benefited from the liberalization of the Mexican market, in contrast to Mexico City which enjoyed its heydays under import substitution. Our findings suggest that FDI has contributed to these spatial changes, as they show that FDI has generated additional positive externalities in the border states, whereas Mexico City has been subject to additional negative externalities. Having said so, the role of FDI in the process of regional change requires further investigation. Importantly, recent findings by Faber (2007) shows clear differences between the spatial development of export oriented and import competing industries following the introduction of trade liberalization. In particular, the feature that import competing industries have fared relatively well in regions such as Mexico City, characterized by relative poor international market access, contrasts with our finding that this region has been subject to additional negative externalities. Further research is necessary to improve our understanding of the role that FDI has played in the structural spatial changes following the introduction of trade liberalization. Finally, the states which have been enjoying additional positive externalities from the presence of FDI are also the states that incorporate the vast majority of maquiladora FDI. The usual assessment of these foreign firms is that they are modestly integrated into the local economy, which implies that their presence is not likely to generate externalities among Mexican firms. Our findings contradict this common assessment, as they suggest that maquiladora-intensive states do enjoy additional positive externalities, including positive spillovers materializing through backward linkages between FDI and local Mexican suppliers. Future empirical research should therefore focus on acquiring more evidence on the full local sourcing behavior of maquiladora firms and on obtaining more direct evidence from local Mexican suppliers on the level and nature of maquiladora-induced externalities in the border states.