Better Neighbors: Toward a Renewal of Economic Integration in Latin America

Introduction and Summary of Results

In a clear break from its past, Latin America and the Caribbean (LAC), particularly South America, experienced a growth spurt with equity during the first decade of the 21st century.¹ In fact, LAC’s gross domestic product (GDP) growth rate over the past decade stood at about 4 percent, well above the region’s historic average of 2 percent. Moreover, the incomes of the poorest households in many countries in the region grew at a faster pace than those of high-income households. Unfortunately, the latest period of prosperity seems to have waned; and, with a few exceptions in Central America and the Caribbean, countries in the region confront once more a reality of low growth.

As the good times have faded, there is now a clear understanding that such impressive performance was the result of a demand boom fueled by an increase in the price of LAC’s exports relative to the price of the region’s imports (a terms-of-trade improvement) (de la Torre, Filippini, and Ize 2016). Moreover, the slowdown has brought back fears of economic instability. To be sure, there is no evidence to date suggesting that LAC is returning to the volatile days of the 1980s, partly because of improvements in its macrofinancial framework. Yet it is undeniable that such fears exist, especially in the context of expected increases in global interest rates. Against this backdrop, policy makers in LAC are now in search of sources of long-term growth and stability.

One policy area that has moved back to center stage is regional integration. Indeed, since at least the 1960s, LAC has experimented with various forms of regional integration with the hope that fostering regional economic ties can yield the type of economic success that the region has long sought. The current push toward regional integration has been influenced by the success of the East Asia and Pacific region (EAP), where intraregional trade, exports to the rest of the world, and incomes have risen together as the region continues to catch up to the income levels of the United States (de la Torre, Lederman, and Pienknagura 2015). Whether this coincidence of trade and growth outcomes is the result of regional commercial or other policies—or whether regional growth itself caused the rise of intraregional trade and global exports—remains an open question. Still, EAP continues to be a source of inspiration for Latin Americans.

Hence, underlying the push in favor of deeper integration at the regional level is the belief that part of LAC’s low growth problem is its low level of intraregional economic integration. In fact, LAC has levels of regional integration that pale in comparison to those of the European Union (EU), EAP, and Eastern Europe and Central Asia (ECA). Taken at face value, this suggests that pursuing formal policy arrangements with the potential of strengthening economic links within the region might boost growth in LAC.

The goal of leveraging formal trade arrangements to accelerate growth is evident in many of the trade agreements that are in place in the region. For example, an objective of the Pacific Alliance—the 2012 integration agreement between Chile, Colombia, Mexico, and Peru—is driving further growth, development, and competitiveness of the economies of its members.² Similarly, the Dominican Republic–Central America Free Trade Agreement (CAFTA-DR) lists the creation of new opportunities for economic and social development and new employment opportunities and improved working conditions and living standards in their respective territories as some of its resolutions.³ And, although not explicitly stated in the agreement, many view Mercosur—the customs union comprising Argentina, Brazil, Paraguay, Uruguay, and República Bolivariana de Venezuela—as a useful vehicle to achieve higher growth for the countries in the Southern cone (Fanelli 2007).

The objective of taking advantage of regional integration to boost growth is not new to LAC. In fact, the region has explored several models of integration to achieve this goal—from the old regionalism that prevailed until the late 1980s, which emphasized the role of regional integration and import substitution as integral parts of industrialization strategies, to the new regionalism that emerged amid the wave of reforms that the region implemented in the 1990s (see IDB 2002). Importantly, the latter form of regional integration views regionalism as a stepping-stone toward the goal of global integration, hence earning the label of open regionalism (hereafter OR).⁴

This report revisits the concept of OR and presents evidence supporting the idea that a revitalized OR strategy can contribute to growth with stability by exploiting the complementarities between regional and global economic integration. It proposes a five-pronged strategy, including (i) reducing external most-favored-nation (MFN) tariffs; (ii) deepening economic integration between South America and Central and North America; (iii) harmonizing rules and procedures governing the exchange of goods, services, and factors of production; (iv) stepping up efforts to reduce LAC’s high trade costs; and (v) integrating labor and capital markets in the Americas. This agenda is nothing short of a wholesale renewal of the notion of OR.

Since the 1990s, OR in LAC focused primarily on preferential trade agreements and their relationship with trade policies affecting trade with extraregional partners (Bergsten 1997). The ultimate goal of the renewal of the OR strategy is to enhance the region’s competitiveness with respect to the rest of the world, which depends on smart (yet complex) policies that enhance intraregional economic integration while also lowering barriers to international trade with the rest of the world. Because the magnitude of bilateral trade and migration flows is dependent on the geographic distance between economic partners, a key analytical challenge is to assess the potential of region-wide efficiency gains that can be attained through regional integration efforts (beyond the pull of geography), combined with domestic structural reforms and further liberalization of trade with the rest of the world. The preponderance of the evidence compiled for this study suggests that how the Americas become integrated can affect the region’s long-term growth prospects and stability, precisely because the forces of geography imply that pro-growth global integration cannot be achieved without strengthening our own neighborhood. A key implication is that the renewal of OR embraces domestic structural reforms that can raise the economic efficiency of the Americas as whole.

To be clear, the analysis presented in the report does not quantify the impact that OR has had on LAC economies in the past. Nor does it quantify the potential gains of the proposed renewal of the OR strategy. Rather, it relies on the economics literature to identify accepted channels through which different forms of international economic integration can stimulate growth and stability, which in turn can be quantified as an indirect way of assessing the priorities for the renewal of OR in the Americas. The report draws upon two prominent strands of economic theory. The first is the idea that the gains from trade depend on differences between countries. In these neoclassical models, these differences are usually modeled as arising either from factor supplies (for example, being labor abundant or capital abundant) or from technology. The second is the idea that trade facilitates learning, either through the experience of exporting or from the exposure to new products and ideas that are embodied in imports. Although these are not the only theories that explain trade and the gains from trade, these are two that have perhaps the longest and most established history in international economics.⁵ The intuition from these models can, in certain ways, also apply to factor market integration.

To keep the discussion focused, the report leaves aside two important aspects of regional integration. First, it does not discuss the effects of economic integration on inequality and poverty, a subject that has been widely discussed in the existing literature.⁶ The potential effects of international integration on inequality and poverty are important to consider but go beyond the scope of this report. For now, it suffices to say that there is evidence that global integration in LAC has probably helped reduce inequality. Following the extensive trade liberalization of the 1990s, wage inequality eventually fell throughout Latin America (Lopez Calva and Lustig 2010; Silva and Messina, forthcoming). Falling inequality could be linked to trade, as predicted by neoclassical trade theory (Robertson 2004). Perhaps more important, concerns about poverty and inequality are generally considered to be less effectively addressed through trade policies than through alternatives, such as expanding the coverage of public education, improving the quality of education in poor neighborhoods, or conditional cash transfers, among other policies that would not hamper growth and economic efficiency. The second limitation of this study is that it does not discuss noneconomic objectives and consequences of regional integration.⁷

The rest of this introduction briefly discusses some of the key findings of the report. It first discusses the importance of geography in shaping both economic performance and integration patterns around the world. Having discussed the role of geography, the overview analyzes observed regional integration patterns. Then it assesses the benefits of integration through two separate theories—one that argues that potential efficiency gains depend on how much countries can complement each other, and another one that argues that benefits depend on how much countries can learn from each other. With this evidence in hand, the introduction lays out the five-pronged strategy for renewing OR in the Americas. In discussing each area of the strategy, the introduction presents the current state of policies in the region as well as the challenges that lie ahead.

Even in the age of globalization, geography matters for trade, factor flows, and economic performance

In recent decades the world has experienced significant technological and economic changes that have transformed international economic relations. These changes have led many to claim that distance is dead or that the world is flat.⁸ In short, one expects that in a flat world a country’s economic performance should not be affected by its geographic location. The Internet and improvements in transportation have certainly affected trade patterns and facilitated new trade relationships

As significant as these changes are, however, the effect of distance does not seem to have disappeared. As discussed in chapters 1 and 2, geographic forces are also important drivers of economic integration. That is, even in the absence of policies favoring regional integration, proximity facilitates economic integration. This bias is the by-product of costs associated with the movement of goods, people, and, to a lesser extent, capital across borders. The literature has found that such costs increase with distance.

The incidence of geography on trade and factor flows (capital and labor) has important implications for the patterns of regional integration observed across regions. Countries that are geographically closer to their regional partners are expected to have higher levels of regional integration compared to those that are more distant.

Chapter 1 shows that economic performance around the globe is also geographically clustered. In particular, a country’s economic performance in both the long run and the short run is highly correlated with that of its neighbors. The likelihood that two countries will simultaneously experience prolonged episodes of either high growth or low growth falls with geographic distance. Similarly, the probability of two countries going through the same phase of a business cycle decreases with geographic distance. Moreover, the geographic forces that shape economic performance haven’t diminished over time. On the contrary, they have increased. Regional forces affecting a country’s GDP growth have gained prominence in the recent past, to the point that, for the average country in the world, they have surpassed country-specific and global factors as key determinants of macroeconomic fluctuations.

Regional forces have become increasingly important over time for Latin American countries as well. By the 1995–2011 period they were as important as country-specific factors. To be sure, the lion’s share of this increase is due to the rising prominence of forces that similarly affect many countries in the region but are linked to developments in other corners of the world. This finding is likely explained by China’s rise in the global economy and the impact it has had across LAC (see de la Torre et al. 2015).

There are at least two hypotheses for the geographic clustering of economic performance. One is that it is a consequence of trade and the integration of capital and labor markets, which themselves are geographically clustered (see Calderón, Chong, and Stein 2007). Another is that endowments, institutions, and other determinants of economic performance are geographically clustered.

Regardless of the explanation, the geographic clustering of economic performance and the way it has evolved over time affect the gains from integration predicted by neoclassical models of trade based on endowment or technological differences and by models of learning through trade. Similarly, the forces of geography are expected to affect the observed levels of regional integration. The rest of this chapter analyzes in more detail these two points. The analysis will then guide the policy discussion that lays the ground for the proposed renewal of OR.

Regional trade in LAC and in the Americas: International comparison and determinants

Regional trade flows are an integral part of international trade flows. As chapter 2 shows, approximately half of total trade flows occur between regional partners. There are, however, significant differences in the incidence of intraregional trade flows in total flows across regions. At one extreme stand EU15+ (European Union 15 extended) and EAP, regions where intraregional exports accounted for 60 and 50 percent of total trade in 2014, respectively. At the other extreme stand regions such as South Asia (SAR), Sub-Saharan Africa (SSA), and the Middle East and North Africa (MENA), where intraregional exports accounted for a meager 10 to 15 percent of total trade in 2014.

As mentioned above, the remarkable performance of EAP in terms of regional trade integration has caught the attention of other regions, including LAC. However, replicating EAP’s experience has proven a difficult challenge for LAC. The region has pursued regional integration efforts through formal trade integration agreements since the 1960s, efforts that have only intensified since the mid-1990s. Indeed, prior to the year 2000 the average country in LAC held a preferential trade agreement with about 4 regional partners; by 2013 this number rose to close to 10. Despite these efforts, the incidence of intraregional exports in LAC’s total exports has remained stable at about 20 percent.

The discussion above raises a question: Why is the region not more integrated? Or, more precisely, what are the constraints to boosting regional trade that face policy makers in LAC? To answer these questions, the rest of this section explores a potential explanation that follows the insights of the international trade literature pointing to economic size and trade frictions associated with geographic distance as gravitational forces shaping trade flows.

Size and geographic distance matter for trade flows

Understanding the determinants of international trade patterns is a research goal that dates back to the early 1800s. One empirical model that appears to fit the trade data particularly well is the so-called gravity model of trade.⁹ Its central tenet is that trade flows should be proportional to the GDP of trading partners and inversely proportional to their geographic distance. The positive relationship between bilateral trade flows and the GDP of trading partners captures the idea that large, wealthy countries demand and supply more goods from and to the rest of the world relative to smaller countries, yielding high levels of trade between them.¹⁰ The inverse relationship between trade and distance captures the idea that trade implies moving goods, and that the cost of moving goods is expected to increase with distance.¹¹ Hence, the prices charged by more distant producers are expected to be higher than those of producers nearby, resulting in lower demand for exports (varieties) from more distant countries. The effects of distance, therefore, may prevent countries from realizing the benefits of trade predicted by neoclassical models.

The relationship predicted by the gravity model has important implications for understanding the regional integration patterns discussed above. First of all, the negative relationship between trade flows and distance predicted by the gravity model and observed in the data implies that, all other things equal, trade flows between nearby partners are expected to be higher than between faraway partners. In other words, even if trade policy around the world were nondiscriminatory, the gravity model predicts that trade should be largely regional because of trade costs that vary systematically with geographic distance.

Another important implication of the gravity model is that differences in the size and distance between countries within regions can play an important role in explaining differences in the incidence of regional trade across regions. In particular, regions comprising countries with large GDP values and with short distances between them are expected to exhibit higher regional trade flows as a share of total trade than others, all else equal.

The insights of the gravity model suggest that, in order to carefully assess LAC’s standing in terms of regional integration and to understand the factors underpinning it, one should take into account the impact of geography and size on trade flows. The analysis presented in chapter 2 follows this approach and compares LAC’s standing relative to other regions in terms of intraregional trade, after stripping away the impact of these variables.¹²

The results in chapter 2 show that the average pair of countries in LAC, which originally ranked poorly in terms of the incidence of regional trade, has intraregional trade flows that are in line with or exceed what is predicted by gravity variables. In contrast, EAP, a region that ranked second in the original comparison of intraregional trade flows, presents levels of intraregional trade that are statistically lower than those predicted by gravity variables.

Importantly, chapter 2 also highlights that the conclusions of the gravity benchmarking are sensitive to the definition of region because the inclusion or exclusion of countries can change the size and distance of the average pair of countries in the region. This will, in turn, affect intraregional trade patterns because of the role of geography and size in shaping trade flows. For instance, an assessment of integration in the Americas (as opposed to LAC alone) provides substantially different conclusions—intra-Americas trade is statistically larger compared to what gravity variables would predict, suggesting that the inclusion of the United States and Canada boosts trade in the Americas beyond what would be predicted by their economic size and distance to LAC countries.

The results from chapter 2 show that, if one of their objectives is to increase intra-LAC or intra-Americas trade flows, Latin American policy makers have two options. Countries in the region could grow at a rate higher than that of the average country in the world, or they could reduce trade frictions associated with policies and distance. But clearly growth is a policy goal in its own right, and arguably a more important one than regional integration per se. Thus, instead of focusing on policy actions that have the sole objective of boosting regional integration, the rest of this introduction discusses integration strategies that can help LAC achieve high and stable growth.

The conceptual arguments for a renewal of open regionalism

This study assesses the benefits of different integration strategies through the lens of two prominent strands of economic theory. The first is inspired in neoclassical models, which suggest that the gains from trade and economic integration more broadly crucially depend on how different economies are in terms of their technologies and their endowments. The gains are expected to be larger when partners are more different. Likewise, the gains in terms of stability are also expected to be larger when trade occurs between dissimilar countries because they are exposed to different types of shocks.

The second strand of theory highlights the role of economic integration as a conduit for technological diffusion and learning. According to these theories, countries could, for example, learn from the technological content embodied in the goods they import. This knowledge content depends on the innovation efforts of a country’s partners and those of their partner’s partners (Coe and Helpman 1995; Lumenga-Neso, Olarreaga, and Schiff 2005). Similarly, economic integration may allow firms in one country to learn about the goods, production processes, and business relationships in third markets of the firms with which they interact in another country. This, in turn, may facilitate productivity improvements and entry and survival in third markets (Morales, Sheu, and Zhaler 2014; Chaney 2011). Importantly, according to these theories the characteristics of a country’s partners matter for the benefits stemming from learning. The gains are expected to be larger when a country’s partners are knowledge hubs (invest in research and development [R&D]) or when they are open (have more business connections and trade with knowledge hubs)

What do these two strands of the theory imply for the attractiveness of different economic integration strategies for LAC countries? From the point of view of neoclassical models, countries in the region would benefit the most by seeking trading partners that are not near them. In particular, chapter 2 shows that, in all regions, integration with the rest of the world appears to provide larger potential efficiency gains compared to regional integration. In LAC, however, the average pair of countries appears to be much more similar compared to the average pair of countries in developing regions, such as EAP or ECA.

Even so, chapter 2 shows that there are still important differences among LAC countries that could lead to neoclassical-style gains from trade. It shows that there is a positive relationship between the similarity of the revealed comparative advantages (RCAs) of a given pair of countries and that pair’s similarity in terms of economic size. Likewise, there are marked differences in terms of patterns of RCA between countries in South America and those in Central and North America. In fact, the average efficiency gains that LAC countries could obtain from trade with regional partners outside their subregion are comparable to those that could be attained from trade with partners elsewhere in the world. These findings suggest that deeper integration between small and large countries in LAC and between South America and Central and North America could yield efficiency gains if the neoclassical theories are valid.

In addition to limiting the efficiency gains predicted by neoclassical trade models, the similar trade structures observed between LAC countries, especially those that are nearby, also limit the prospects for regional integration to deliver stability. LAC economies are typically exposed to similar shocks (for example, terms-of-trade shocks), thus limiting the scope for regional integration to diversify country-specific risks. This point is supported by chapter 2, which shows that the volatility of LAC’s exports would rise if the weight of regional partners on a country’s export basket increases. This is due to the volatile and correlated import demands observed in LAC countries.

Learning models do not provide much more support to regional integration. Chapter 2 shows that countries in LAC do not have as many trade connections as do other countries in the world and invest too little in R&D, thus limiting the benefits predicted by these models. To be sure, the desirability of integrating with specific partners depends on how transferable knowledge is between countries. If knowledge is fully transferable, the characteristics of a country’s partners become irrelevant because countries can build upon the stock of knowledge of the world. In this case, the stock of knowledge of a country can be appropriated by that country’s trading partners, the partners of its partners, and so on. In contrast, if knowledge transfers and learning are not easily diffused across space, the characteristics of a country’s trading partners become more important and countries will differ in their stock of knowledge. The assumption that there are frictions to knowledge diffusion and learning is supported by evidence, suggesting that the identity of partners matters (Keller 2002). Moreover, frictions appear to increase with distance, both geographic distance and distance in levels of development, which means that the potential gains from trade from the point of view of learning models will depend on the characteristics of a country’s nearby partners. A country is expected to have more scope for learning when its nearby partners are knowledge hubs or have strong commercial ties with knowledge hubs.

There seems to be tension, therefore, between geographic forces and the policies that facilitate regional integration and the predictions of economic models that drive countries in LAC to look for efficiency gains beyond their immediate neighbors. Indeed, there is a tension between preferential trade arrangements that provide incentives for intraregional trade perhaps at the expense of trade with the rest of the world and the realization that geography naturally favors intraregional trade. Why would LAC pursue an integration strategy that combines global and regional integration? The short answer is that there are important complementarities between regional integration and global integration that make LAC’s international competitiveness and its ability to reach extraregional markets dependent on regional integration. Thus, a comprehensive renewal of OR can make LAC more competitive in global markets.

There are several reasons why a balance between regional and global integration efforts can boost LAC’s competitiveness. First, the impact of geography is unlikely to disappear any time soon. This implies that trade links with nearby countries will affect the global competitiveness of countries in the region. The link between regional trade and global competitiveness is most clearly illustrated in the case of regionally traded goods (see chapter 2). These are goods and services for which the costs associated with distance are so high that they are typically exchanged only by neighboring countries and for which the policy-related barriers to trade are not import tariffs per se, but rather differences in regulatory schemes. For these goods and services, regional integration efforts are equivalent to global integration. Notable examples of these goods and services are electricity and land transportation. Hence, regional efforts to assure the quality and the efficient provision of these types of goods and services will be crucial for the growth and stability prospects of LAC and for the ability of the region to gain international competitiveness in sectors that use these regionally traded goods intensively.

Similar arguments can be made in the case of