1.1 A Decade of Debate
1.2 Objective and Structure of the Thesis
2 Regional Integration
2.1 Definition, Forms and Objectives
2.2 Welfare Implications
2.2.1 Static Welfare Effects
2.2.2 Dynamic Welfare Effects
2.3 North-South versus South-South Integration
2.4 Spatial Inequality in the Integration Process
3 NAFTA as an Example of Regional Integration
3.1 Mexico’s Way into NAFTA
3.2 Design and Coverage of the NAFTA Provisions
3.3 NAFTA in the Light of Integration Theory: Expected Effects
4 NAFTA’s Impact on Mexico after One Decade
4.1 Static Benefits
4.1.1 NAFTA’s Impact on Trade
4.1.2 Trade Creation, Diversion and Mexican Terms of Trade
4.2 Dynamic Benefits
4.2.1 NAFTA’s Impact on Foreign Direct Investment
4.2.2 Dynamic Spillovers, Productivity and Growth
4.2.3 Catching Up with the North?
4.3 Adjustment and Divergence
4.3.1 Growing Disparities and Intra-Mexican Divergence
4.3.2 Migration - Adjustment Mechanism for the NAFTA-Neglected?
5. Conclusions and Policy Implications
5.1 Summary and Discussion of Results
5.2 Policy Implications and Outlook
Figure 1: Trade Creation and Diversion in a Free Trade Area
Figure 2: Exemplary Effects of the Formation of an FTA on Countries’ Industry Share and Welfare
Figure 3: Mexican Trade Flows, 1980-2002
Figure 4: FDI Inflows to Mexico, 1980-2002
Figure 5: GDP Per Capita Growth of Mexico and the U.S., 1993-2002
Figure 6: Migration from Rural Mexico to the United States, 1980-2002
Table 1: FDI Inflows as a Percentage of Gross Fixed Capital Formation
Table 2: Poverty in Rural and Urban Areas, 1984-1994
Table 3: Income from Household Remittances, 1996-2002
illustration not visible in this excerpt
1.1 A Decade of Debate
In January 1994, after two and a half years of negotiation, the North American Free Trade Agreement (NAFTA) came into force. The treaty between Canada, Mexico and the United States has created the largest economic area in the world, slightly surpassing the European Union in market size. But NAFTA is also outstanding in a second aspect: it has constituted the first major regional integration arrangement between two highly developed countries, the United States and Canada, and a developing country, Mexico.
The North-South nature of North American integration has polarized the debate about NAFTA from the earliest stage on. On the one hand it was unclear how much the U.S. would gain from the agreement. Would it stabilize its southern neighbor and thus benefit the U.S. economically and politically? Or would it cause the “giant sucking sound” Ross Perot feared, drawing thousands of jobs from the U.S. over the border (Thorbecke/EigenZucchi 2002, p. 648)? Regarding these concerns, Canada was at most a side-player, possessing neither intense trade relations nor geographical proximity to Mexico.
Mexico’s gains from NAFTA, on the other hand, seemed even more unsure. The agreement’s effects on the southern member state, whether positive or negative, were expected to be unequally greater than on the U.S. On the one hand, it seemed, Mexico could gain immensely through improved access to the North American market, increasing trade, attracting foreign investment, and importing growth and stability. On the other hand, some trade economists, such as Arvind Panagaria (1996, pp. 512-513) warned that Mexico could only lose when opening its market to its powerful northern neighbors, while receiving little in return that it would not have obtained anyway. Furthermore, would Mexico’s move towards regional integration hamper any further step into the direction of multilateral opening, after promising reforms had been started in the mid-1980s?
Concerns also regarded the adverse effects of NAFTA within Mexico. These centered around large adjustment costs from sectoral restructuring and resource reallocation. This would occur if inefficient, partly subsidized Mexican industries declined after removing tariffs and non-tariff barriers, allowing the North American competition to enter the national market. In addition, would this hit mostly those Mexican regions that were poor anyway?
The Mexican government was convinced of the miracles of regional integration and acted as the main driver in the negotiation process between the three North American nations. Nevertheless, ten years after NAFTA’s start the debate about its benefits for the Mexican economy is more polarized than ever. On the one hand, NAFTA’s champions, such as the then Mexican minister of trade and industrial development, Jaime Serra, still argue forcefully in favor of NAFTA’s blessings regarding trade and investment (Serra/Espinosa 2002, pp. 60-62). On the other hand, many Mexicans appear disappointed by unattained social goals. In addition, activists from all over the world regard NAFTA as a prime example of the negative effects of free trade, denouncing working conditions in and the environmental impact of the maquiladora plants, essentially seeing Mexico on the ground after a race-to-the-bottom.
In the face of the controversy about NAFTA’s effects, it is important to establish an unbiased evaluation of the benefits and costs that regional integration has brought to Mexico after a decade. Not least because the outcome of the experiment of North-South integration is of considerable importance for other developing countries trying to liberalize, whether preferentially or multilaterally, and for which NAFTA may or may not serve as a model.
1.2 Objective and Structure of the Thesis
The question around which this thesis centers is, therefore, where and to what extent Mexico has gained from regional integration in the form of NAFTA and where it has experienced adverse effects and costs from it. These gains and losses can cover a large variety of issues, the examination of all of which would exceed the scope of any single thesis by far. The economic effects of trade liberalization are one possible area of investigation, another one would be the impact of NAFTA on Mexican politics, policy reform and political stability. Political economy approaches can explain the role of and the effect on interest groups in regional integration arrangements. Lastly, NAFTA’s environmental impact could be an issue.
This thesis focuses on the economic implications of NAFTA, leaving aside the other issues above. The problem is further broken down into three specific questions: First, what have been the traditional static gains from integration for Mexico? Second, which dynamic effects has NAFTA brought about, especially on foreign investment, its impact on productivity and growth, and have these led to economic convergence? And third, what have been the costs of adjustment resulting from the static and dynamic effects?
This allows a theoretically well-founded analysis of benefits and costs of this specific case of regional integration in order to carry through a reasonable evaluation of NAFTA’s impact on Mexico. A thorough comparison between regional integration and multilateral trade opening as alternative approaches to economic liberalization, however, lies beyond the scope of the thesis.
The existing literature on NAFTA’s effects on Mexico can essentially be divided into two groups. The first group consists of ex-ante studies that have been published since the first announcement of a North American free trade area in the early 1990s. Some of the literature qualitatively analyzes NAFTA’s potential effects applying concepts of the theory of regional integration, such as Ramirez de la O (1993) and Ohr (1995). A vast amount of ex-ante studies, however, conducts simulations based on computable general equilibrium (CGE) models, and range from estimations of trade and investment effects over the impact on certain sectors such as agriculture and manufacturing to the effects on labor and wages.1
The second group of literature consists of ex-post studies that evaluate NAFTA’s impacts with hindsight. Individual studies focus on a multitude of different issues, such as the impact on trade, using a traditional Vinerian framework2 of analysis (e.g. Krueger 2000), or foreign direct investment and dynamic effects (López-Córdova 2002; Romalis 2003). Some other works have used a more comprehensive approach and combined different aspects (e.g. Ramirez 2003). One of the most thorough analyses has been conducted by a recent World Bank report (Lederman et al. 2003), which combines existing literature with more far-reaching econometric exploration of several important issues.
The purpose of this thesis is to use the benefit of hindsight to combine, structure, and critically evaluate the results of the multitude of existing (ex-post) studies and thus provide answers to the three research questions detailed above. In order to do this, it first develops a theoretical framework using insights from theories of regional integration and the new economic geography. This framework is then used to examine the North American Free Trade Agreement and formulate expectations about its probable effects. These are subsequently discussed within the analysis and compared to the actual developments, thus bringing together theory and empirical evidence. In contrast to several existing re view studies3, special attention is paid to separating NAFTA’s impact from those of other influential factors to provide an as accurate as possible picture of the former. Chapter 2 provides the theoretical framework of the study. After introducing the main forms and characteristics of regional integration, potential welfare effects are considered. These are split into static and dynamic effects, an analytical separation that is upheld throughout the course of the study. Derived from these, the welfare implications of North-South and South-South integration for developing countries are discussed. Finally, the issue of convergence, divergence and regional inequality, mainly represented by the new economic geography, is added as an essential element of analysis. Chapter 3 then deals with NAFTA as an example of regional integration. First, Mexico’s way into NAFTA as a gradual path to liberalization is depicted, being crucial to the understanding of post-NAFTA developments. After briefly addressing the consequent expectations of Mexico versus North American free trade, the main provisions and negotiation outcomes of NAFTA are discussed. Subsequently, expectations of NAFTA’s effects in the areas depicted above are formulated in order to test and discuss them in chapter 4. Chapter 4 analyzes NAFTA’s impacts on Mexico after the first ten years of the agreement. The first area to be examined is the static effects concerning trade creation and diversion, based on a Vinerian framework. Second, dynamic effects are discussed, focusing on foreign direct investment, its contribution to technological spillovers and growth, and subsequent intra-NAFTA convergence. Lastly, the attention turns to the adjustment effects resulting from trade liberalization, with a special focus on intra-Mexican divergence and adjustment through migration.
Chapter 5 concludes the thesis by discussing the findings from chapter 4, outlining limitations of the study, deriving policy implications and presenting an outlook.
2 Regional Integration
2.1 Definition, Forms and Objectives
Different definitions exist for the term ‘regional integration’. One of the most common is Balassa’s4, which will be adopted for the purpose of this thesis:
“[Economic integration can be defined] as a process and as a state of affairs. Regarded as a process, it encompasses measures designed to abolish discrimination between economic units belonging to different national states; viewed as a state of affairs, it can be represented by the absence of various forms of discrimination between national economies.” (Balassa 1962, p. 1)
The dual nature of integration as process and state demonstrates that integration can be seen as a continuum (Blank et al. 1998, p. 31), and thus must be typified accordingly.5 A classification that has found wide acceptance is the distinction between preference area, free trade area, customs union, common market, and economic union (Balassa 1962, p. 2; Ohr/Gruber 2001, p. 3), where a preference area constitutes the weakest form of integration and an economic union the strongest.6 Table A.17 (p. 62) summarizes these forms of integration and their main characteristics.
In a preference area the participating countries grant each other preferential conditions for the trade of certain, specified product categories, usually in the form of a partial or total removal of tariffs for these products.8
A free trade area (FTA) is an agreement between two or more economies in which tariffs and non-tariff barriers, such as import quotas, are abolished for imports from the participating economies. In this, FTAs can be seen as a special case of a preference area, extended to all product categories. The elimination of tariffs generally only relates to products that have been produced entirely or to a large extent within the FTA. Imported products from external countries are still subject to national tariffs. Respectively, within an FTA there is no coordination of trade policies between the member countries. Hence, each country individually pursues national trade policies versus non-member states. This results in the necessity of ‘rules of origin’. These rules prevent externally imported goods from entering the FTA via the member with the lowest tariffs, circumventing the higher tariffs of other members (‘trade deflection’). In contrast to stronger forms of regional integration, an FTA allows its members to retain a large amount of independence in trade policy-making (Blank et al. 1998, pp. 57-58; Kaiser 2003, p. 27). Prominent examples of FTAs are NAFTA and the European Free Trade Association (EFTA).
In customs unions, which have been in the center of integration research until the 1990s (Krueger 1997a, p. 177), member countries introduce a common external tariff (CET)
vis-à-vis non-member countries, while internal import and export tariffs are entirely eliminated. This implies that the members’ trade policies are jointly coordinated - members give up a significant part of their policy-making independence. This loss of independence may be a reason why today customs unions are less abundant than FTAs (Kaiser 2003, p. 27). One example of a customs union, is the European Economic Community (EEC) of 1957.
While FTAs and customs unions limit integration to the product markets, the move to a common market, as happened in the European Community after 1992, implies the integration of product and factor markets, in which there is unlimited mobility of labor and capital. Respectively, a higher degree of harmonization of economic policies becomes necessary, regarding both competition and fiscal policies (Blank et al. 1998, p. 33). The highest level of regional integration of sovereign states is reached by forming an economic union. As Balassa puts it, this “[…] total economic integration presupposes the unification of monetary, fiscal, social, and countercyclical policies and requires the setting-up of a supra-national authority whose decisions are binding for the member states” (1962, p. 2). Economic union can go hand in hand with monetary union, as is the case with the European Economic and Monetary Union (EMU).
Following Ohr/Gruber one may define the goal of regional (market) integration as the “optimization of the allocation of resources and hence, an enhancement of economic efficiency and macroeconomic welfare in the integration area” (2001, p. 4). According to economic theory, however, efficiency and welfare are maximized when there is worldwide free trade, as suggested by the most-favored-nation (MFN) principle of the General Agreement on Tariffs and Trade (GATT) and the World Trade Organization (WTO). Taking into account the imperfectness of the world trading system, however, regional integration arrangements (RIAs) may be seen as a second-best solution (Blank et al. 1998, p. 21).
2.2 Welfare Implications
2.2.1 Static Welfare Effects
During the first half of the 20th century it was widely assumed that the conclusion of preferential trading agreements was necessarily trade liberalizing and therefore welfare increasing (Krueger 1997a, p. 175).9 Regional integration arrangements have thus been welcomed as a move towards free trade. Viner’s (1950) comparative static analysis10 of customs unions, however, showed by introducing the concepts of trade creation and trade diversion that the welfare effects of regional integration arrangements are indeed ambivalent.11 Krueger notes that “[…] a customs union (or other preferential agreements) might result in the attainment of a Paretosuperior situation for one trading partner (due to the predominance of trade creation) and of a Pareto-inferior situation for the other trading partner, with either a Pareto-superior or -inferior situation for the union (or FTA) members as a group.” (Krueger 1997a, p. 176)
Trade creation occurs when regional integration leads to the shift of demand from inefficient, uncompetitive domestic producers to efficient producers in a partner country, thus resulting in a more efficient allocation of resources and increased trade between the countries within an FTA or a customs union. Trade diversion, on the other hand, occurs when trade is shifted away from a non-member country to a member country due to preferential treatment of the member’s goods, although the non-member country’s producers are more efficient and can, with non-discriminating tariffs, offer more competitive prices (Viner 1950, pp. 44-45; Balassa 1962, p. 25).
The effects of trade creation and diversion can be shown using a simple model with two countries, A and B, the world market, and a tradable good.12 Figure 1 shows the market for good x in country A, where DA is country A’s demand for good x; SA is country A’s domestic supply, which is imperfectly elastic; SW is the perfectly elastic world market supply; and SB is country B’s supply, which is also assumed to be perfectly elastic. Initially, country A raises a non-discriminatory tariff t against all imports, so that the price of goods from country B is PB+t and the world market import price is PW+t. In this situation country A produces OX0 and imports X0M0 from at PW+t. Country B’s prices are not competitive. If country A enters into a free trade area with country B, B’s price will be PB, while imports from the world market remain at PW+t. Hence, country A’s production decreases to X1 and imports amount X1M1.
Several effects can now be observed. First of all, consumption increases from M0 to M1 due to the lower price. The result will be a consumer surplus gain of area d. Areas i and j would be the additional gain from free trade. Second, domestic production decreases from X0 to X1, which is the replacement of domestic products by more efficiently produced foreign imports. This leads to production cost savings of area b as a net gain. With free trade, areas f and g would additionally be realized. The consumption and production effects jointly represent the trade creation effect. At the same time, area a shifts from producer to consumer surplus.
Figure 1: Trade Creation and Diversion in a Free Trade Area
illustration not visible in this excerpt
Source: Own illustration based on Blank et al. (1998, p. 60)
But the FTA in the model also diverts all trade away from the world market to less efficient producers in country B. While without integration areas c and h represented tariff revenues, area c now adds to the consumer surplus. Area h, however, is lost and represents the welfare loss from trade diversion. The net welfare effect of the FTA is consequently +b+d-h and depends on the relative size of the consumer surplus gains, on the one hand, and the loss in tax revenue on the other.
The ambivalent effects of regional integration on national and world welfare reflect the theory of the second best:13 they demonstrate that an incomplete move towards free trade, encompassing only a few countries, is not necessarily beneficial. The illustration above suggests that FTAs are beneficial as long as trade creation outweighs trade diversion (Balassa 1962, p. 26; Ohr/Gruber 2001, p. 10).
The relative sizes of the opposing effects depend on several factors.14 One important determinant thereof is the level of the preexisting tariffs of a country. The higher the tariffs, the smaller will be the amount of trade between countries A and C prior to regional integration, resulting in fewer trade diversion. Furthermore, the higher the tariff, the larger the potential for trade creation with country B. A higher preexisting tariff therefore increases the potential benefits of a free trade area or a customs union (Kaiser 2003, pp. 7778).
The amount of trade diversion also depends on the price difference between imports from the third country (country C) and the one with which a regional integration arrangement is to be formed (country B). The lower the price of imports from C compared to B, the greater are the efficiency advantages of country C and the negative impact of regional integration on country A in terms in trade diversion (Kaiser 2003, p. 78). A third determinant is the level of preexisting trade among members of a RIA before integration. The higher the trade among members, and the lower the trade between members and non-members, the more likely it is that trade creation outweighs diversion (Krueger 1997a, p. 176; Stehn 1993, p. 4). The assumption that this is mainly the case for countries located in one geographical region gives rise to the ‘natural trading partner hypothesis’ (Krugman 1991b, p. 19). Here, the main argument is that the level of trade strongly depends on transport costs, which are lower the closer countries are located to each other.15
In an FTA, the rules of origin necessary to prevent trade deflection may be an additional cause of trade diversion. The reason is that, in the face of content requirements which remove tariffs from products only if they contain a certain percentage of components from within the FTA, producers may have an incentive to source from FTA-internal high-cost suppliers instead of from external lower-cost suppliers in order to fulfill the rules of origin (Krueger 1997a, p. 179).
In the case of a customs union, the members may experience an increase in their terms of trade. This is the case if the customs union is sufficiently large to influence the world market price of a good through a change in its common external tariff. In this case a rising CET lowers the world market price and raises the value of the customs union’s export compared to its imports, leading to a welfare increase. This welfare increase, however, is achieved at the expense of the rest of the world and thus can be described as a “beggar-thy-neighbor effect” (Krugman 1991b, pp. 13-16).16
In conclusion, the comparative static analysis of regional integration arrangements introduces the concepts of trade creation and diversion, the relative size of which determines the welfare effect of an FTA or a customs union, abstracting from terms-of-trade effects.17 These effects, however, have to be complemented by the dynamic effects of integration, without which an evaluation of a RIA would be incomplete.
2.2.2 Dynamic Welfare Effects
Over time, the impact of dynamic welfare effects is considered more important than the purely static effects described above (Schiff/Winters 1998, p. 178; Ohr/Gruber 2001, p. 13).18 Dynamic effects are understood as those that occur in an economy over time as it adjusts to observed disequilibrium states. Within the present context, specifically those dynamic effects are considered which affect a country’s economic growth over the long and medium term.19
While a large amount of literature exists on the general relationship between growth and economic openness, few studies specifically focus on RIAs (Schiff/Winters 1998, p. 183). Those that do mainly discuss two channels through which integration can incur growth effects: technology and investment (Vamvakidis 1999, p. 45). Under the first of these one may subsume economies of scale, technological spill-over effects and enhanced innovation. Economies of scale can be achieved through the market enlargement that results from economic integration. Firms can hereby increase their production and, with increasing marginal rates of return, produce at lower average costs than before (Corden 1972, pp. 465-474). Through these efficiency improvements domestic firms may become competitive at an international level, even if the RIA has originally been trade diverting. Additionally to firm-internal economies, external, sector-wide scale economies can occur as stronger firm specialization leads to sector-wide infrastructure improvements or enhanced cooperation in research and development (Ohr/Gruber 2001, p. 13). Evidently, this type of dynamic gain will be the greater the larger the share of sectors with economies of scale in total output in an integration area.
One might furthermore expect the removal of trade barriers to trigger intensified competition. Monopolies and oligopolies existing prior to integration are, in the face of competition, pressurized to increase their productivity, leading to the reduction of xinefficiencies20. Additionally, firms will have incentives to pursue process and product innovations (Stehn 1993, p. 5).21.
But especially developing countries can also benefit from foreign research and development (R&D). Coe et al. have found that trade is an important vehicle of technological spillovers (1997, p. 147). Specifically, this can happen through the import of larger varieties of intermediate products and capital equipment, enhancing the productivity of a country’s own resources, as well as through cross-border communication and learning, and the copying and imitation of foreign technologies (Coe et al. 1997, p. 136).
The second channel through which growth may be fostered is investment. For developing countries the role of foreign direct investment (FDI) is of particular importance. FDI serves as a harbinger of confidence and, secondly, as a means of modernizing the economy (Schiff/Winters 1998, p. 180). Here, the potential of FDI to transfer production know-how, technology and managerial skills is well recognized. Additionally, spillover effects to other firms can be expected. In developing countries technological progress usually accounts for a relatively small proportion of growth, because the conduct of research and development (R&D), and thus the generation of new knowledge, is constrained by a low volume of human capital and adequate skills. FDI, as well as trade, can bridge this gap by importing the lacking skills.22 This finally enhances the marginal productivity of the host countries’ capital stock and thus promotes growth (Balasubramanyam et al. 1996, pp. 6-7).
Unfortunately, theory does not convey a very clear picture of how FDI flows will be affected by regional integration. Traditionally, trade and capital movement in form of foreign investment have been regarded as substitutes, making FDI a device for “tariff jumping”. One may hence reason that regional integration leads to decreasing FDI from bloc partners for whom the tariff reduction makes the trade option more attractive (Blomström/Kokko 1997, p. 5). This argument is, for instance, valid for horizontally organized multinational enterprises (MNEs)23. However, investment in developing countries is often based on the use of differential factor endowments, such as the abundance of unskilled labor, presuming the predominance of vertical MNEs (Waldkirch 2002, p. 2). Further, a trade-creating RIA could entail shifts in production structures which may result in additional intra-regional FDI (Blomström/Kokko 1997, p. 5).24 FDI from third countries could be expected to increase, because the possibility of accessing the larger market at low (or zero) tariffs from the internal country presents an additional incentive to invest. Conversely, a reduction of external FDI is thinkable if companies possess horizontally organized subsidiary networks. In the case of integration, such a structure might become sub-optimal and result in a rationalization of the network, leading to disinvestment somewhere in the integration area (Blomström/Kokko 1997, p. 6). Hence, from the trade-barrier motive, it is difficult to draw conclusions, although especially for FDI from third countries the pro-FDI argument seems stronger. While important, trade-barrier related motives are not the only determinants of FDI. For once it must be recognized that specific investment provisions included in preferential trade agreements may have a large influence on the development of internal FDI.25 Their practical relevance, though, depends strongly on the level of preexisting investment barriers abolished by the provisions and on the extent of host government discrimination against foreign investors (Blomström/Kokko 1997, p. 9).
Secondly, investor confidence is a prerequisite for access to international capital. Confidence can be provided through the credibility of government reforms.26 Especially reforms in developing countries often lack credibility due to time-inconsistent policies (Schiff/Winters 1998, p. 181).27 It is therefore worthwhile to explore the relationship between credibility and regional integration.28 Fernández/Portes suggest that a RIA can be more efficient than multilateral liberalization under GATT since in the former there are higher incentives to punish members deviating from the rules, and because the intrinsic incentive to stay within a RIA is stronger than for the latter (1998, p. 205).29 Two conditions for a RIA to enhance credibility in trade policy are, however, that a country’s policies are time-inconsistent prior to the RIA and that the costs of exiting the RIA are higher than the gains of returning to time-inconsistent policies (1998, p. 204). Others have suggested that RIA’s can also serve as a commitment device for credibility in other policy areas, such as micro- and macroeconomic reforms (Whalley 1996, p. 16).
In conclusion, regional integration arrangements may incur dynamic effects that potentially outweigh static gains from trade, since they can influence a country’s economic growth past the short run and lift a country on a higher growth path in the long run.
2.3 North-South versus South-South Integration
After having discussed static and dynamic effects of regional integration, this section takes on the perspective of developing countries. Which types and which properties of economic integration are more likely than others to have a beneficial impact on them? Developing countries generally possess different factor endowments and economic structures from more advanced economies, the former being rather labor- and natural resource-intensive, and the latter predominantly capital- and technology-intensive. Considering that developing countries wishing to join a RIA generally have the options of South-South integration (i.e. integration between developing countries) and North-South integration (i.e. integration between developing and industrialized countries), the question of South-South versus North-South can at least partially be discussed on the basis of integration of rival versus complementary economic structures.
Both constellations receive some theoretical support. According to the Heckscher-Ohlin theorem30, a wealth increase is a function of the difference in the countries’ factor endowments that determine the pattern of trade, where gains are the larger the greater these libria, if the promise not to intervene is not credible. E.g. country may use tariffs to compensate wage differentials across sectors, resulting from a terms-of-trade shock that leads to a productivity loss. The result would be that, in anticipation of the intervention, more workers than otherwise remain in the injured sector (Fernández/Portes 1998, p. 203). differences are (Tichy 1992, p. 109).31 The result would be a high amount of interindustry trade and, where factor movements are liberalized, high streams of factor migration to where their marginal productivity is highest. On the other hand, Viner suggests that, with respect to ex-ante protected industries, rival structures result in higher welfare than complementary ones, where high trade diversion may outweigh the positive effects (1950, p. 51).32 Furthermore, efficiency gains from specialization are most likely where intra-industry trade is strong, which predominantly is the case for similar countries (Tichy 1992, p. 109).
Which kind of integration, then, will be more beneficial to low income countries? Weighing up different factors, Schweickert, for example, is convinced that South-South integration must be excluded as a viable option for a catching-up strategy for developing countries (1994, p. 12). Moreover, Anne Krueger states that “[…] the developing countries’ comparative advantage and opportunities for gains from trade lie predominantly in the wide divergence between their factor endowments […] and those of their developed countries” (1997a, p. 177).33
Regarding the static effects of integration, however, Spilimbergo and Stein present a convincing argument for case distinction. They argue that in a situation where the tariffs of all countries are equal, poor countries would choose North-South over South-South integration based on comparative advantage arguments, thus supporting Krueger’s statement (rich countries would in this case choose North-North integration). For a situation with differential tariffs between the integration countries, however, the situation is different. If the tariffs of a rich country are sufficiently low, then the result of integration may be deteriorating terms of trade of the poor country.34 Therefore, the poor country in this situation prefers South-South integration (Spilimbergo/Stein 1996, pp. 28-32). Additionally, in this situation poor countries may suffer disproportionate losses in tariff revenue.
The Spilimbergo-Stein result is valid ceteris paribus, but it has to be qualified in so far as other factors influencing static gains, such as size of the FTA or amount of common trade, may outweigh disadvantages from tariff differentials, and the question remains of what is “sufficiently low” in practice.
The major gains for a poor country in favor of North-South integration, however, are dynamic ones. Here, some of the arguments are the access to larger markets, technology spill-overs and increased investment. These opportunities lie certainly rather with large, advanced countries than with poor ones. Additionally, the credibility gains for developing countries are larger if they bind their commitment to wealthy, stable economies rather than with volatile countries which have themselves problems of time-inconsistency.
Hence, while large differences between the economic structures of integration partners are generally rather disadvantageous, for less advanced countries seeking North-South arrangements seems to be the only viable option, assuming the countries somehow resemble natural trading partner. An important consideration, however, is how these structures influence the way in which integration affects the geographic distribution of benefits, dealt with in the next section.
2.4 Spatial Inequality in the Integration Process
Over time, how will the benefits of integration be distributed over the integration area? Will regional disparities between regions in terms of productivity and income converge or further diverge? Is industrial activity likely to spread evenly across the integration area or will it concentrate in certain loci, resulting in industrial agglomerations? These questions may concern the distribution across different countries, as well as the development of the regions of a particular member country.
Neoclassical theory, following the factor-price-equalization theorem, predicts equalization of factor income and productivity across countries, where factors move into the direction of their highest marginal productivity, assuming perfect competition, factor mobility and constant returns to scale (Samuelson 1948, pp. 169-172).35 The result is that in a regional arrangement capital flows into the lower income country, in which the marginal productivity of capital is higher. This raises the productivity of labor and, as a consequence, per-capita-income increases, while in a high-income, capital-intensive country the opposite happens (Ohr/Gruber 2001, pp. 26-27).
According to this model industrial activity spreads out evenly across space even with the assumption of positive transport costs: product and factor market competition would be stronger where more firms are geographically concentrated, driving the firms apart. Differences in production structures are explained by underlying differences between the regions, such as natural resources, geography, or technology, that lead regions to specialize according to their comparative advantages (Ottaviano/Puga 1998, p. 709).
The more recent literature on the ‘new economic geography’, on the other hand, provides endogenous mechanisms that explain the uneven distribution of industrial activity across regions.36 The core of new economic geography-models consists of firms’ location decisions, driven by several centripetal and centrifugal forces (Krieger-Boden 2000, pp. 6-7): The centripetal forces, which drive firms towards agglomerations, consist of firm-internal economies of scale that induce firms to produce in few, centralized places; economies of localization that are internal to an industry sector37 ; and economies of urbanization, which are not industry-specific38. Centrifugal forces, on the other hand, which cause dispersion, consist mainly of the scarcity of immobile factors and their respective price increase after agglomeration, and congestion costs such as pollution or high traffic.
The seminal paper of the new economic geography, Krugman (1991a), uses a model of two regions, two sectors, manufacturing and agriculture, and two production factors, known as the core-periphery (CP) model.39 Agglomeration is mainly driven the firms’ desire to locate where most customers live to buy their products. Concerning the relative strengths of dispersion and agglomeration forces, Krugman finds that the former dominate when costs of trade are high. Hence, with high trade barriers, both regions will have fairly equal shares of manufacturing. But as trade becomes freer, agglomeration forces start to dominate, with the result that all industry will, in a self-reinforcing process, locate in one region when trade costs fall below a certain point, creating a core and a periphery region (Baldwin et al. 2003, p. 11).40
The assumption of mobile labor, which puts limits to the model’s applicability to FTAs41, is dropped in later models (Venables 1996; Krugman/Venables 1995). Here the driving mechanism is backward and forward linkages to suppliers and buyers, to which firms locate geographically near. According to these models industry starts to agglomerate in one country as trade costs fall (and integration proceeds), leading to a rise in real wages in this, and falling real wages in the other country.42 At an even higher degree of integration, though, the countries again converge, thus producing a U-shaped relationship between the degree of integration and industry share (Krugman/Venables 1995, pp. 860862).43
Models that specifically examine FTAs yield similar results (Puga/Venables 1997; Baldwin et al. 2003, pp. 330-361). An FTA-internal agglomeration effect occurs, resulting in an unequal development of industrial activity within the integration area. The implication is a relocation of industry towards the largest member country. Large asymmetry between the economies within the FTA can amplify the agglomeration forces, as can higher trade barriers versus the rest of the world (Baldwin et al. 2003, p. 339). Puga/Venables find that, as trade is further liberalized, the smaller economy may be able to re-attract some industry due to developing centrifugal forces, shown exemplarily in Figure 2 (1997, p. 364).44
As a consequence, the relocation of industry into the larger country is likely to cause welfare to fall in the smaller country as real wages drop due to fewer demand for labor and higher costs of imported goods.45 As the smaller country re-attracts industry in the further liberalization process, welfare is likely to rise again (Puga/Venables 1997, p. 361).
1 Comprehensive and readable surveys of the studies using CGE models are Brown (1992), Brown et al. (1992) and Kehoe/Kehoe (1994). Models of labor implications are reviewed by Hinojosa Ojeda/Robinson (1992). Generally these models predict income gains for Mexico, which are higher in the models accounting for dynamic growth effects.
2 See chapter 2.
3 A notable exception being Lederman et al. (2003).
4 Balassa’s definition in fact relates to ‘economic integration’. Both terms will be used interchangeably in this study. A similar definition is given by El-Agraa (1989, p. 1).
5 However, while different forms of integration to be presented are continuously measured by the degree of integration, they each represent a single, independent state the attainment of which does not necessar ily require to undergo previous states or proceed to higher degrees of integration. See, e.g., Stange (1994, pp. 9-10).
6 Interestingly, the preference area does not form part of Balassa’s original classification. See also Blank et al. (1998), pp. 32-33. A slight variation of this typology is, e.g., presented by Jovanovic (1992, p. 9).
7 Tables and figures that are labeled with an “A” can be found in the Appendix.
8 One example of a preference area is the British Commonwealth Preference Scheme, that has been founded in 1932 by Great Britain and former Commonwealth countries (Blank et al. 1998, p. 32).
9 The scope of this thesis does not allow for a detailed discussion of why trade liberalization in general is considered welfare increasing. For a short overview over the subject see, e.g., Krugman/Obstfeld (1994, pp. 1-8).
10 As commonly understood, the comparative static analysis comprises changes in the equilibrium states of a model as assumptions are altered. Here, static effects are primarily understood as the changes in the structure of trade as an economy participates in a regional integration scheme.
11 Most of the fundamental analytical work from Viner on is concerned with customs unions, while free trade areas and other forms of regional integration have only recently received more attention (Krueger 1997a, p. 177). Most of their treatment is based on the adaptation of customs union theory, so that a theory of free trade areas has not existed independently. Rather, relative advantages and dis advantages of free trade areas versus customs unions have mostly been analyzed (Pelkmans 1980, pp. 340-341; Stange 1994, p. 37). Here, however, the example of a free trade area is chosen due to its relevance for the thesis’s subject.
12 This example is based on Blank et al. (1998, pp. 60-61).
13 The theory of the second best says that, “[…] if is impossible to satisfy all the optimum conditions […], then a change which brings about the satisfaction of some of the optimum conditions may make things better or worse” (Lipsey 1960, p. 498, emphasis in original).
14 For a more complete overview of these factors see Viner (1950, pp. 51-52).
15 The notion of natural trading partners has, however, received substantial amounts of criticism. See, e.g., Bhagwati et al. (1998, pp. 1132-1134) and Panagariya (1996, pp. 488-492).
16 For a graphical analysis of the terms-of-trade effects in customs unions see Blank et al. (1998), pp. 98 102. See also Viner (1950), pp. 55-58.
17 Some argue that trade creation or diversion misrepresent static welfare effects. For example, if in addi tion to the ‘inter-country substitution’ effect analyzed by Viner the effect of consumer’s changes in consumption due to the relative prices of commodities (‘inter-commodity substitution’) are taken into account, the latter can partly compensate welfare losses from trade diversion (Lipsey 1960, p. 504).
18 One reason why still the majority of studies focuses on static effects of integration (Kaiser 2003, p. 94) might be seen in the fact that dynamic effects are considerably harder to measure than static effects.
19 For such an understanding of dynamic integration effects see also Schiff/Winters (1998, p. 179).
20 X-inefficiencies are inefficiencies that result from slack within the firm if available resources are not used efficiently.
21 One example demonstrating such positive effects is the European Common Market, founded in 1958, where strong diversionary effects seemed likely - however, a large increase in intra-industry trade in manufactures and a rationalization of production resulted (Krugman 1991b, p. 13).
22 However, imported skills and technologies are still property of the foreign entity. Hence the benefit of FDI also depends, among other things, on the ‘absorptive capacity’ of firms in the host country (Kinoshita 2000, p. 1).
23 An MNE is horizontally organized when it produces similar products or services in multiple countries. It is vertically organized when it fragments the production process and distributes various stages of pro duction over different countries (Aizenman/Marion 2001, p. 2).
24 In this sense, FDI can also be a direct consequence of the static effects of integration. Analytically, how ever, FDI will be treated as a dynamic phenomenon in accordance with the understanding of dynamics presented (see p. 10).
25 Yet another motive may be the internalization of intangible firm-specific assets (Blomström/Kokko 1997, p. 7). But since these do not help to resolve the relevant debate about the nature of integration induced FDI, they will not be looked at further.
26 Also, provisions of integration agreements, such as dispute settlement mechanisms, can help to increase the confidence of intra-regional investors (Blomström/Kokko 1997, p. 10).
27 In international trade, problems of time inconsistency can occur if governments are tempted to use sur prise trade policy actions in the absence of other first-best instruments, resulting in suboptimal equi-
28 This link is, among others, explored by Whalley (1996), Francois (1997), and Baldwin et al. (1996).
29 This, however, presumes that the benefits, e.g. from foreign investment, for a country will indeed be higher than in a multilateral liberalization.
30 For a detailed discussion of the Heckscher-Ohlin model see Siebert (1991), pp. 53-78.
31 In the early literature on customs unions, as Viner points out, “ […] it is almost invariably taken for granted that rivalry is a disadvantage and complementarity is an advantage in the formation of cus toms unions” (Viner 1950, p. 51).
32 If, for instance, two countries A and B with complementary structures integrate, the chances that certain imports are diverted away from third countries are relatively high. If the structures are rival, than the goods that have been imported from the third country are less likely to be diverted, since B will unlikely be able to replace imports from A to C.
33 The advantageousness of North-South integration also seems to be supported by the empirical literature relating integration and growth (see, e.g., Vamvakidis 1998, p. 165).
34 In the extreme case of a poor country A with positive, and a rich country B with zero tariffs the forma tion of an FTA between these countries will have the following consequences: country A will deviate trade away from a third country C in favor of B and will also shift demand from itself to B. Since in B the tariff structure does not change, it neither creates nor diverts trade. Overall, demand for goods in A will fall, while demand in B rises. A’s terms of trade deteriorate, B’s improve (see Spilimbergo/Stein 1996, pp. 31-32).
35 According to the Heckscher-Ohlin-Samuelson theorem of factor price equalization, free trade is a perfect substitute for factor mobility “unless initial factor endowments are too unequal” (Samuelson 1948, p. 169, italics removed).
36 The endogenous determination of differences also contrasts with the ‘new trade theory’ (e.g. Krug man/Venables (1990)). The new economic geography takes into consideration a decentralized market process, scale economies, non-homogeneous products, non-competitive markets, transportation costs, and factor mobility (Krieger-Boden 2000, p. 6). For surveys of the new-economic-geography litera ture see Ottaviano/Puga (1998), Schmutzler (1999), and Neary (2001).
37 E.g. forward and backward linkages, proximity to important suppliers or knowledge bearers.
38 Such as general knowledge and information spillovers.
39 Initially, both regions are assumed to be similar. One of the production factors, industrial workers, is mobile, while the other, farmers, is immobile. Transport costs are modeled as iceberg costs (i.e. as a percentage of the goods “melting away” during transportation). The agglomeration forces consist of a market-access effect (firms locate of big markets and export to small ones) and an cost-of-living effect (goods are cheaper in the big market since consumers have to pay less transport costs for imported products), while the dispersion force consists of a market-crowding effect, leading firms to locate where there is less competition. For a detailed analysis of the core-periphery model see Baldwin et al. (2003, pp. 9-67).
40 In fact, the model predicts an “catastrophic” agglomeration, in the sense that once openness surpasses a certain point, the only stable outcome is complete agglomeration (Baldwin 1993, p. 35).
41 This is because FTAs do not usually comprise complete factor market liberalization.
42 The effect on real wages is due to the rising demand for labor in the country where the manufacturing sector concentrates and falling demand for labor where manufacturing declines (Krugman/Venables 1995, p. 861).
43 For a similar result see Puga (1999). In Krugman/Venables (1990) the pattern of divergence and conver gence for increasing integration is similar, however no circular causation results, since agglomeration forces are not endogenous.
44 In Figure 2 overall welfare of the trade bloc rises due to a second effect which causes production to shift from non-member countries into the FTA due to its enlarged internal market (Baldwin et al. 2003, pp. 330, 332-337).
45 It has to be noted that the model of Baldwin et al. results no welfare decrease for any member country (2003, p. 340).