Policy Research Working Paper 8871 Latin American Growth A Trade Perspective Augusto de la Torre Alain Ize Finance, Competitiveness and Innovation Global Practice June 2019 Policy Research Working Paper 8871 Abstract This paper reviews the determinants of Latin America’s dynamics: the first is centered on commodities and South uneven growth based on an accounting decomposition that America, the second on manufactures and Mexico, and the breaks down countries’ growth (relative to the world) into third on services and Central America. The evidence points three trade-related channels: (i) an export pull measuring toward the need for a trade-oriented growth agenda that the traction exerted by the country’s exports, (ii) an exter- puts a premium on raising exports and making countries nal leverage measuring the impact of the country’s use of more attractive to people, not just capital. The latter in external resources, and (iii) a domestic response measuring turn adds urgency to healing the region’s social fractures the impact of the country’s imports on its domestic income. and dealing with its institutional weaknesses. This decomposition brings to light three regional growth This paper is a product of the Finance, Competitiveness and Innovation Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at adelatorre@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Latin American Growth: A Trade Perspective Augusto de la Torre and Alain Ize JEL classification codes: 040, 054, F10. Keywords: growth, convergence, Latin America, export-led growth, import substituting industrialization, commodity dependence, natural resource curse, export diversification.  Augusto de la Torre (adelatorre@worldbank.org and apd2151@columbia.edu) Alain Ize (alain.ize@gmail.com) are adjunct professors at Columbia University’s School of International and Public Affairs. Augusto de la Torre is also a World Bank Consultant. and the Director of the Economics Research Center at Universidad de las Américas, Ecuador. 1. Introduction This paper reviews the determinants of Latin America’s (LatAm) uneven economic progress from an international trade and macro perspective that looks at both historical trends (the second half of last century) and recent events (the China-induced commodities boom).1 The analysis is backed by an accounting decomposition that links a country’s growth relative to the world (G) to its participation in world export markets through three trade-related channels, an export pull (EP) that measures the traction exerted by the country’s exports on its growth; an external leverage (EL) that captures the impact on the current account of the country’s real export growth relative to its real import growth, hence its use of external resources; and a domestic response (DR) that measures the impact on the country’s domestic income of the imports resulting from the combined export pull and external leverage. This growth decomposition brings to light three key regional dynamics. The first centers on the links between growth and commodities, as illustrated most closely by the South American countries; the second on the links between growth and manufactures, as illustrated in particular by the case of Mexico; and the third on the links between growth and services, as illustrated by Central American countries. As regards the growth-commodities interplay, growth in the more specialized South American commodity producers has historically tended to lag behind that of the world as a result of secularly declining export pulls (EPs). And the growth spurt experienced by these countries during the more recent commodity boom was mainly driven by rising export prices, rather than rising export volumes. This creates doubt about growth sustainability going forward, particularly for the more specialized commodity exporters, not only because of the uncertainty surrounding future trends in commodity prices, but also because of their proverbial volatility. Moreover, periods of terms of trade bonanzas have tended to elicit pro-cyclical macro-financial (particularly fiscal) policies, exacerbating spending binges during booms, and thereby making the adjustment pains and growth collapses more acute during the subsequent busts. The second dynamic concerns the links between growth and manufactures. It occupied center stage throughout LatAm in the times (roughly 1960-1980) of the grand experiment of import substitution industrialization (ISI). Paradoxically, ISI accentuated the dependence in the very commodities it was intended to shed, as the proceeds from commodity exports became even more indispensable to finance the voracious appetite of the local (highly protected) manufacturing sector for intermediate and capital goods imports. ISI did result in some important growth episodes although mainly in Brazil and Mexico, countries with large domestic markets. However, the exhaustion of ISI and the eventual unsustainability of the growth spurts it created had dire consequences for the region, which had to be painfully absorbed over two long decades of crises and adjustment (1980s and 1990s). After the 1980s, Mexico became the most visible case of a major reorientation of manufacturing production towards the outside. Boosted by NAFTA, 1 For the purposes of this paper, “LatAm” is composed of Mexico, the Spanish speaking countries in South and Central America, and the Dominican Republic. 2 Mexico’s success in expanding and diversifying its exports was disrupted by the collision with Chinese manufactures in the 2000s. Moreover, the trade liberalization required for the shift towards outward-looking manufacturing unleashed a surge in imports that undermined Mexico’s local production of importables, with lasting adverse consequences for Mexico’s capacity to transform export growth into per capita GDP growth. The third dynamic centers around the interplay between services and growth. This dynamic is arguably best illustrated by the Central American countries, which have over the past three decades become substantial producers and exporters of services. A substantial per capita growth differential has separated the high performers (Panama, Costa Rica and the Dominican Republic) from the low performers (Guatemala, Honduras and El Salvador). While all these countries run relatively large and persistent current account deficits, they differ markedly in the way these deficits are financed: by FDI in the case of the high performers, remittances in the case of the low performers. The other side of the coin of this contrasting feature is that the high performers keep their labor force at home and attract foreigners and foreign equity finance, reaping high growth dividends from both stronger exports and more vibrant domestic activities. In contrast, the low performers, unable to sustainably attract foreign capital, see their workers migrate abroad in droves. While the remittances sent back home help sustain consumption, they do not raise productivity, hence growth. The stark contrast between pulling in people and equity finance, on the one hand, and driving workers out, on the other, can be readily linked to radical differences between the two groups of countries as regards rule of law, as measured by indices of institutional quality and the incidence of crime and violence. Faster growth across the region will require a policy and reform agenda aimed at boosting the production of tradable goods and services. This agenda puts a premium on building knowledge, attracting talent and fostering all relevant forms of cross-border connectivity. This agenda will need to be supported, in the case of commodity exporters, by a continued effort at export diversification and stronger countercyclical macroeconomic management policies. In the case of manufacture exporters, the agenda will need to focus on the exports of goods and services that do not unduly displace labor in the destination markets, thereby limiting the scope for protectionist responses in those markets. A greater orientation toward innovation and in favor of exports whose world demand rises faster than world income—including of course tourism but also personal services such as health, wellness, and old age care—should ease the way. In all cases, but arguably more so for services exporters, the agenda will unavoidably have to focus on the rule of law and the related institutional strengthening. Reaching these goals would in turn be facilitated by making LatAm countries more attractive places to produce in as much as places to visit or reside in, which puts a premium on safe and predictable living environments, green development, good infrastructure, and an educated and friendly labor force. At the same time, for such growth strategies to be politically viable among LatAm’s deeply fractured societies, solid improvements will be needed in social protection systems (especially regarding pensions and health) and skill formation (including through high quality education for all). 3 The rest of the paper is organized as follows. Section 2 provides the conceptual and methodological background needed to anchor the subsequent empirical analysis of trade and growth in LatAm countries. Section 3 explores the roots of the region’s growth gap during the second half of the twentieth century. Section 4 focuses on LatAm’s growth response to China’s trade expansion during the recent commodities boom. Section 5 concludes with a brief discussion of possible policy responses to the growth challenges faced by the region. 2. Accounting for growth This section presents an accounting decomposition that identifies three channels through which trade and macro policies affect growth. It then discusses interpretation and causality issues. It concludes with some worldwide econometric evidence describing the relative importance of each channel in explaining a country’s growth and exploring the links between a country’s gains in world export participation and its relative per capita income growth. a) A simple growth decomposition Consider the following growth accounting decomposition, inspired from Thirlwall (2011), but differing from it in two key aspects: (i) rather than assuming current account equilibrium, it adds a “leverage” residual to incorporate deviations from that equilibrium; and (ii) rather than assuming constant trade (export and import) elasticities and using them to predict growth, it takes growth as given and derives the changes in trade elasticities that are consistent with such growth and the actual evolution of the country’s exports and imports. Let be the rate of real output growth for any given country and ∗ the rate of output growth for the world as a whole. Similarly, let and be the real (constant dollars) rates of growth of the country’s exports and imports, and ∗ ∗ those of the world’s exports and imports. As in Thirlwall (2011), we use a simple trade elasticities formulation: ∗ (1) (2) but instead of assuming current account equilibrium we posit an additional elasticity-like term relating a country’s imports to its exports: (3) It follows that the ratio /∗ can be written as: ∗ ∗ ∗ ∗ (4) 4 Thus, defining the link between imports and exports (the changes in the current account) as a residual elasticity that can take any value (instead of being constrained to equal one) converts Thirlwall’s predictive growth equation into a pure accounting decomposition that must hold at all times. Given that world exports must equal world imports (∗ =∗ ), (4) may be rewritten so that all country growth terms are expressed relative to the world: ∗ ∗ ∗ ∗ ∗ / ∗ (5) Consider now the following definitions and notation: ∗ (6) ∗ (7) (8) ∗ / ∗ (9) Then, (5) can be rewritten as: G = EP + EL + DR (10) To complete the decomposition, valuation gains and losses deriving from changes in the terms of trade should also be taken into account. When sizeable, terms of trade fluctuations can significantly alter the purchasing power of a country’s output, thereby becoming important drivers of aggregate spending. Terms of trade-driven valuation gains and losses can be captured by the difference between the nominal and real expressions of the growth accounting decomposition. Using the prefix N to identify nominal aggregates: G = NEP + NEL + NDR (11) It is then easy to show that real and nominal values differ due to two valuation-related terms, TOT and WEV: EP = NEP - WEV (12) DR= NDR - TOT + WEV (13) EL = NEL + TOT (14) 5 where TOT, the changes in the terms of trade (i.e., changes in a country’s export prices relative to its import prices), and WEV, the changes in world export values (i.e., changes in a country’s export prices relative to the world’s), are defined as: ∗ (15) ∗ ∗ ∗ (16) TOT and WEV gains or losses affect real domestic demand by changing the purchasing power of a country’s output. As can be seen in equations 12-14, the TOT valuation effects are captured in the EL term, whereas the WEV valuation effects are captured in the DR term. Instead, the real external pull (EP) reflects only volume effects, as it is obtained by subtracting WEV from the nominal export pull (NEP). b) Interpretation and causality G measures the changes in a country’s growth relative to world growth. EP is the country’s real export pull, which can be interpreted as the traction exerted by the country’s exports on its growth. EP should grow when the country becomes more efficient in producing the exports that the world demands (a favorable local supply response to world demand patterns). EL is the country’s real external leverage, which captures the impact on the current account of the country’s real export growth relative to its real import growth, hence its use of external (debt and equity) financing. DR reflects the country’s real output growth relative to the growth of its imports. It can be interpreted as the real domestic response to a change in real imports brought about by changes in the country’s combined export pull and external leverage. DR rises when the country’s GDP grows faster than its imports. Since EL is a function of the current account disequilibrium of the balance of payments, it should reflect the impact on the external gap (the difference between imports and exports) of deviations between domestic aggregate demand and output. Under stable real exchange rates, the external gap and the output gap (the difference between potential and actual output) should be closely related. Hence, EL should be a good proxy for excess aggregate demand. Strong real exchange rate depreciations, however, could channel aggregate demand away from tradables toward non-tradables, in which case the (positive or negative) growth impact of domestic demand would be distributed in the EL and DR terms. But otherwise DR should mainly reflect supply- based (rather than demand-based) factors, such as changes in productivity that allow a country to produce more for a given level of imports, or changes in trade policies that restrain or promote imports for a given level of output.2 2 Indeed, in the case of advanced economies for which good output gap historical series are available, EL is generally highly correlated to the output gap, which is not the case for DR. 6 As with all accounting decompositions in economics, how one interprets the observed patterns and correlations between growth and its channels depends fundamentally on causality.3 Clearly, correlation does not necessarily imply causation. For example, countries may export because they grow (rather than the other way around), as faster productivity gains give faster- growing countries a natural edge in export markets. If so, the export pull term could be a consequence of growth, rather than its cause. In this case, there would be no reason for attaching more policy importance to boosting exports than to promoting productivity across the board. Likewise, an increase in a country’s domestic response could be the result, rather than the cause, of a rise in the country’s growth. Yet, both empirical and conceptual reasons support the view that a crucial (if not unique) direction of causality goes from EP (exports) and DR (imports) to growth, rather than the other way around and, hence, that policies to promote the tradable sector should be good for the economy as a whole. Starting with the empirical side, simple Granger tests clearly indicate that exports precede rather than follow growth (see Appendix Table 2). While such timing-based tests do not prove causality (any more than the fact that night follows day proves that day causes night), they are suggestive of causality and, at a minimum, imply that a country’s rising share in global exports can be a good predictor of growth. In addition, changes in EP resulting from exogenous changes in trade policies (Mexico joining NAFTA) or world events (China joining the WTO) are naturally leaning in the direction of trade influencing growth rather than the other way around. Similarly, causality should be easier to infer when changes in DR come from its denominator (imports), rather than from its numerator (output). This would be the case for example when the boost in a country’s imports derives from an exogenous event such as trade liberalization. On the conceptual side, beyond the obvious gains of scale from operating in larger (international) markets, there are at least two reasons to think that the expansion of the tradable sector (i.e., involving goods and services that can be traded across borders) can itself be a source of productivity growth. First, the production of tradables is likely to generate superior learning externalities and technological spillovers compared to non-tradable activities (goods and services that cannot be traded across borders).4 Second, more dynamic exports enhance balance of 3 Like other decomposition methods in economics, ours introduces a residual into a theoretically-based equation, turning it into an identity. Thus, in the Solow-inspired growth accounting model, growth is decomposed into factor accumulation and a productivity residual. Instead, in the Oaxaca-Blinder decomposition for labor economics, differences in the average wages of two groups are decomposed into differences in the observed characteristics of the groups and a residual that accounts for the unobserved differences. In our approach we start from Thirlwall’s model, which introduces a plausible conceptual link between trade and growth, and then introduce a current account residual. As noted by Aghion and Howitt (2007) and Fortin et al. (2010), such decompositions lack an identification theory (they are, in some sense, “theory free”), do not formally provide causal explanations or predictions, and do not explicitly recover deep underlying behavioral relationships. However, they can reveal interesting patterns, thereby enabling fruitful interpretations and inferences to the best explanations, all guided by sound economic theory. 4 This argument has been advanced by several authors. Rodrik (2008) uses it to highlight the importance for growth of a competitive real exchange rate. Hausmann and Rodrik emphasize the role of product complexity (particularly the complexity of export baskets) in boosting growth (see Hausmann et al., 2014; and Hausmann, Hwang and Rodrik, 2005). This focus on exports is complementary to the focus on productivity that has characterized recent studies on Latin American growth, such as Araujo et al. (2014) and Pages (2010). 7 payments viability and resiliency, thereby helping avert the negative growth impacts (externalities) of macro-financial instability and crises. More generally, moreover, observed changes in the components of growth (EP, DR and EL) should help flesh out a meaningful growth narrative whatever the direction of causality. Indeed, the growth accounting decomposition should be viewed as a convenient tool (an approximating lens) to view and interpret the data. The focus on macro fluctuations and international trade (rather than on factors of production and full employment) gives tradable goods and services a more central role in the growth process than in the conventional literature and places openness and international integration at the heart of the exercise. At the same time, the fact that all variables are expressed in growth terms and relative to the rest of the world promotes standardization, thereby enhancing the visual usefulness of country-specific dynamic analyses and facilitating their comparability across countries.5 c) Growth patterns and drivers How important on a worldwide basis has been the role played by the export pull compared to the other two terms of the growth accounting identity? To help answer this question, Table 1 presents a variance decomposition of the growth accounting identity; three features stand out:  For the sample as a whole, EP accounts for the bulk of the variance decomposition, with DR a distant second.  Yet, DR has a higher weight in the slower growing countries and, at least for the more recent period, in the higher income countries.  The negative signs of EL suggest that, rather than to boost growth, countries have tended to use external finance to dampen declines in growth associated with dips in exports or to invest abroad excess national savings associated with dynamic exports. Further insights into the drivers of growth and convergence can be gleaned from Figure 1, which plots the average yearly rates of change of countries’ shares of world exports (vertical axis) against the yearly rates of change in per capita income relative to that of the US, defined as convergence rate. The close correlation between both variables (very high for both periods, although with greater dispersion in 2000-2015) clearly suggests that convergence is unlikely to materialize in the absence of a vigorous export performance. As evidenced by the bar columns in each of the four quadrants of Figures 1a and 1b (the bar columns show the percentage of all the countries in the sample in each quadrant, in blue, and their median per capita income, in orange), there are important contrasts between the pre and post 5 By construction (i.e., relative to world growth) the cross-country distributions of growth and each of its three components should be distributed around zero. Likewise, growth and its components should fluctuate around zero for the average world country. 8 China periods. Relatively few countries converged towards the US per capita income in the 1960- 2000 period and, except for the fast-growing South East Asian “tigers”, most of those that converged were relatively high-income countries (chiefly in Europe). Most countries in the world (nearly 70 percent) diverged in that period. By contrast, nearly 80 percent of all countries, including LatAm countries, converged during the China-pull period of 2000-2015 and most of them were low-to-middle-income (the higher-income countries lost ground during this period). Table 2 in the Appendix provides econometric evidence that confirms the key messages of Figure 1. The table reports the results of regressing the rate of per capita relative income growth against the rate of change of world export shares, controlling for country size, rate of population growth and initial GDP per capita. The results highlight the overwhelming role of exports in growth. They also suggest the existence during 1960-2000 of an apparent commodity curse, as growth in commodity exporting countries tended to lag behind that in other countries. However, this effect was no longer present during 2000-2015. The results also show that population growth tends to hinder per capita income growth, an oft-overlooked feature that contradicts the theoretical steady-state predictions of plain-vanilla neoclassical growth models where population growth is assumed always to boost (rather than drag down) per capita income growth. The reality is that countries with fast population growth are often unable to fully integrate their labor force into productive, formal employment. The growth accounting decomposition method that was discussed in this section provides the conceptual and empirical background to the analysis of LatAm’s post-WWII growth experience undertaken in the next two sections. Taking into account the structural break in the data introduced by the momentous rise of China in the global economic landscape, this analysis will focus separately on the pre- and post-2000 subperiods. 3. Latin American growth during the second half of the twentieth century LatAm’s growth performance after WWII and until the rise of China in the early 2000s was rather dismal when compared to that of Southern Europe and South East Asia (Figure 2).6 LatAm’s per capita income fell steadily from being 50 percent above that of the world in 1950 to around 10 percent below by 2000. Over the same period, Southern Europe’s per capita income, which was similar to that of LatAm in the early 1950s, rose to two-and-a-half times that of the world by 2000. As strikingly, South East Asia’s per capita income, which was barely 40 percent of LatAm’s in the 1950s, rose rapidly starting in the 1970s to a level 60 percent higher than that of the region by 2000. The 1950-2000 period was marked by the rise of the US, Western Europe, and Japan as the three major engines of world growth, against the background of a powerful post-WWII economic reconstruction momentum. From a trade perspective, therefore, LatAm’s poor growth performance reflects an inability to take advantage of its links to the US economic locomotive, in sharp contrast 6 While cross-country data on GDP per capita start in 1950, comparable export data become available no earlier than 1960. 9 to Southern Europe and South East Asia, which harnessed successfully their trade connections to, respectively, Western Europe and Japan. LatAm’s failure to capitalize on trade-related sources of growth is not independent of two distinct features (veritable trademarks) of its economic history: (i) a long-lived dependence on commodity exports, common to virtually all countries in the region until the 1980s, as discussed in Section 4; and (ii) an enthusiastic and widespread embrace of an inward-looking growth strategy that promoted (and protected) import substituting manufacturing activities until at least the early 1980s. The rest of this section explores the growth implications of these two salient features. a) Commodity curse? To better grasp the links between growth and commodity dependence we focus on the major Latin American commodity exporting countries and classify them according to their gross exports’ composition (Figure 3). Countries are considered to be “commodity exporters” if in 1990 commodities accounted for at least 40 percent of their gross exports, “specialized commodity exporters” if their exports in any given class of commodities (agricultural, mineral, hydrocarbons) exceeded 45 percent of total exports, and “diversified commodity producers” if neither of the three classes of commodities exceeded the 45 percent threshold. With these definitions, Brazil, Colombia and Mexico are classified as “diversified exporters”, the same as Australia and Canada. Instead, Uruguay and Argentina are classified as “agricultural exporters”, the same as New Zealand; Chile and Peru as “mineral exporters”, the same as South Africa; and the República Bolivariana de Venezuela as an “oil exporter”, the same as many other Middle Eastern or African countries. How did these groups of countries perform in terms of growth during the 1960-2000 period, relative to each other and to their non-LatAm comparators? It is clear from Figure 4a that the diversified commodity exporters systematically outperformed the specialized commodity exporters. Their diverging paths were, however, remarkably similar to those followed by their respective non-LatAm comparator countries (Figure 4b). This suggests that commodity dependence and export composition mattered a lot for growth during this period, with the more specialized commodity exporters having experienced commodity curse-like symptoms.7 This view is strengthened by the striking similarity of the declining growth and export paths followed by the specialized agricultural commodity exporters—Argentina, Uruguay and New Zealand (Figure 5). Likewise, specialized oil exporting República Bolivariana de Venezuela’s growth and export paths looks eerily similar to those followed by other oil producers (Figure 6). And so are the growth and export paths of the mineral exporters Peru and South Africa (Figure 7). That the commodities curse is not destiny, however, is illustrated by Chile, the most salient counter case in LatAm during the 1960-2000 period (Figure 4a). After initially following the same diverging path as other specialized mineral exporters, Chile made a complete turnaround in the 7 The “natural resource curse” refers to the likelihood that countries with abundant natural resources grow slower (Sachs and Warner, 2001). 10 late 1980s and started converging rapidly towards the per capita income of the advanced economies, eventually becoming a high-income country. This suggests that even specialized commodity exporters can sustain high growth as long as they manage to continuously raise their shares in global exports. How easily can this be achieved over the long haul is an open question, however. Indeed, as discussed in the next section, Chile is now facing export and growth challenges that are at least in part related to the commodity intensive composition of its exports. Given that the diversified commodity producers were also relatively larger countries (Brazil, Colombia, and Mexico), it could be argued that their better growth performance during the 1960-2000 period, compared to that of the specialized commodity exporters (Figure 4), is attributable more to their size than to their more diversified export structure. Larger domestic markets would have, under this view, helped enable import substitution industrialization (ISI) to yield better fruits. But if size was the decisive factor, one would expect the growth difference between diversified and undiversified countries to have domestic roots (a strong DR), rather than external roots, whether via exports (EP) or external leverage (EL). However, Figure 8 shows that this was not the case: during the fragment of the import substitution period for which data permit a growth decomposition (1965-1982), the diversified commodity exporters clearly dominated the undiversified as regards their EP (or the sum of EP+EL), but not their DR. This suggests that external drivers (export diversification) mattered more for growth than internal drivers (import substitution). We will revisit this important point in the next section when discussing growth in LatAm during the import substitution era. b) Import substitution industrialization and its aftermath Many LatAm countries aimed at freeing themselves from their commodity dependence through import substitution industrialization. ISI used (and abused) the infant industry argument to develop manufacturing products to be sold initially and mainly in regional markets protected by common import barriers. However, it paradoxically relied on the very commodity exports it wanted to shed: those exports were the source of the foreign exchange needed to finance the imports of inputs and equipment required by the inward-looking (highly-protected and import- intensive) manufacturing sector. South East Asia, by contrast, had no commodity exports to rely on and, hence, had no choice but to pursue an outwardly-oriented industrialization strategy. To be sure, ISI did accelerate growth during the 1960s and until the late-1970s in the larger LatAm countries, where scale effects or broader export diversification could play a larger role. The heyday of ISI in fact featured “growth miracles” for Brazil and Mexico, which significantly outperformed the regional average (Figure 4a). The conventional view is that this resulted from large productivity gains associated to the migration of labor from the low productivity agricultural sector to the more productive industrial sector (Pages et al., 2010).8 However, as already noted above in Figure 8, the growth decomposition comparing specialized versus diversified commodity 8 Argentina was also decidedly committed to ISI, yet it performed poorly comparatively speaking (Figure 6) because, arguably, the favorable impact of ISI was overwhelmed by the downward trend driven by commodity specialization. 11 exporters suggests that this might also (and perhaps even more importantly) have reflected a broader export diversification capacity. While ISI gave LatAm a head start over South-East Asia as regard the participation of manufacturing in GDP, this was dramatically reversed during the trade liberalizations that started in the 1980s. Manufacturing shrunk fast relative to GDP in LatAm while gaining importance in South-East Asia (Figure 9a). At the same time, the share of manufacturing exports in LatAm’s total exports lagged considerably that of South-East Asia (Figure 9b). Indeed, under ISI, LatAm ended up promoting inefficient manufacturing sectors with little exporting potential. ISI reached exhaustion in the late 1970s as commodity prices weakened and commodity exports could no longer generate sufficient foreign exchange to finance the import-intensive manufacturing sector. A widespread borrowing binge ensued—a last-ditch and ultimately unsuccessful effort to prolong the life of ISI. That led to profound debt and macro crises, which are at the heart of LatAm’s dismal “lost decade” of the 1980s and the painful “adjustment decade” of the 1990s. LatAm’s per capita income contracted in the 1980s, diverging sharply away from that of the world. While its per capita income growth rose in the 1990s, roughly matching that of the world, it continued to lag that of the US. Mexico is a paradigmatic example of the high economic costs involved in exiting ISI, a saga that can be neatly recounted with the help of the accounting decomposition method presented in Section 2 and depicted for the case of Mexico in Figure 10. A key development that forced Mexico out of ISI was arguably the collapse in growth during the 1980s, which resulted from the combination of a real export dive (a decline in EP; Figure 10a) and a drop in the external leverage (Figure 10b). The latter in turn resulted from the debt crisis of the early 1980s—which abruptly closed Mexico’s access to international financial markets—and the sharp deterioration of the terms of trade caused by the collapse of oil prices.9 This massive external adjustment led to imports falling faster than GDP, hence to a rising domestic response (DR). The gravity of the situation left Mexico with little choice but to abandon a long period of heavy protectionism and shift to an aggressive process of trade liberalization starting in the second half of the 1980s. However, this led to a surge of imports and a collapse of the local production of importable goods which, together with a delayed response in exports (EP continued to fall until 1993-94, when Mexico joined NAFTA), prolonged the economic stagnation. In this phase, as imports expanded faster than GDP, DR collapsed, pulling along Mexico’s growth downward. Indeed, it has taken over 20 years for DR to return to zero, let alone to positive levels. To be sure, boosted by NAFTA, trade liberalization did trigger a surge of exports by the mid-1990s. But that 9 Figure 10b shows relatively moderate fluctuations in EL, which do not seem to do justice to the depth and importance of the indebtedness that led to the debt crisis. This is because the EL term measures net rather than gross flows. In Mexico, the huge public borrowing of the late seventies and early eighties was in large part offset by nearly equally huge private capital flight. 12 was insufficient to offset the impact on growth of the concomitant rise in imports and persistently low DR.10 4. Latin America’s growth response to the rise of China This section analyzes LatAm’s economic growth since 2000, a period strongly marked by the rise of China. Some of the graphs, however, present data from 1990 onwards to better display the China-driven structural break. By exploring contrasts between converging and non-converging countries in the region, the analysis provides further insights into the “natural resource curse” debate as well as into the role of countercyclical policies, international competition and the rule of law in fostering the expansion of tradable sectors. a) The overall picture China’s “big bang” changed the global economic landscape profoundly. It boosted relative growth among the lower and middle-income countries (Figure 11).11 While LatAm also benefitted, the rise in its relative growth was more subdued and went into a reversal after 2012, as the commodity price cycle entered its downswing phase. The comparatively modest impact of China on LatAm’s relative growth was partly the result of strong cross-country heterogeneity exacerbated by a sharp bifurcation of trade patterns from the 1980s onwards (Figure 12). While South American (henceforth, LatAm South) exports remained commodity centered, LatAm North diversified away towards manufactures (Mexico) or services (the rest of LatAm North, namely, Central America and the Dominican Republic). As a result, the terms of trade of LatAm North and LatAm South bifurcated, with LatAm South’s moving with commodity prices and LatAm North’s tending to move in the opposite direction. This bifurcation in the trajectories of the terms of trade deepened after Mexico joined NAFTA in 1993-1994 and China joined the WTO in 2000. Interestingly, not all LatAm commodity exporters converged towards the standards of living of the advanced economies, nor all commodity importers diverged. To drive this point home more clearly, we classify LatAm countries in terms of their convergence performances over the past 25 years or so, distinguishing between net commodity exporters and net commodity importers. To broaden the perspective, we choose two measuring subperiods, the first (1990-2017) starting a decade ahead of the rise of China and the second (2000-2017) coinciding with it. The result of this exercise is summarized in Figure 13 and Table 2, where commodity exporters (CE) and commodity importers (CI) in the region are classified as “strongly converging” (SC) if their per capita income converged towards the US per capita income at a minimum rate of 1 percent per year during both subperiods; “moderately converging” (MC) if they converged by at least 1 percent per year during one of the two subperiods; “weakly converging” (WC) if they 10 Indeed, the underestimation of the transitional costs of market liberalization was arguably one of the salient weaknesses of the “Washington Consensus” agenda. 11 Interestingly, China’s impact on world growth was much less dramatic than on world convergence, reflecting in large part the growth-impairing sequels for the advanced economies of the 2009 global financial crisis (Figure 13b). 13 exhibited a positive (but less than 1 percent per year) rate of convergence during either subperiod; and “non-converging” (NC) if they failed to converge during either period. Perhaps unexpectedly, the best performers include three commodity exporters (Chile, Peru, and Uruguay, the SCCEs) and three commodity importers (Costa Rica, Dominican Republic, and Panama, the SCCIs). As expected, the República Bolivariana de Venezuela is by far the worst performer (the only NCCE), its per capita GDP diverging pronouncedly away from that of the US. Puzzlingly, Mexico is the only other non-converging country in the region (the only NCCI), despite its success in export expansion and diversification. Why were some commodity exporting countries in LatAm South unable to latch more vigorously onto the convergence train pulled by the China locomotive? Instead, why were some commodity importing countries in the region unable to converge despite a shift away from (or at least a significant attenuation of) commodity dependence? The rest of this section sheds light into these questions and aims at clarifying at least in part the reasons behind the great heterogeneity in the recent growth performances across the region. The focus is first on the South American commodity exporters; then on Mexico, the main (if not the only) manufacture exporter in the region; finally, on the Central American services exporters. b) Commodity exporters: The challenges of steady growth One message of Table 2 is that export dynamism per se may be more fundamental than export diversification or complexity in avoiding the natural resource curse.12 Even where export baskets remain concentrated in commodities, countries might converge towards the standard of living of advanced countries if they can continue to raise their shares in global exports (a strong EP) and translate such an export pull into a vigorous domestic response (a strong DR). Neither of these two conditions is easy to ensure, however. Consider first the possible unsustainability of commodity export-based growth with the help of our accounting decomposition tool. The boost in exports during the China-led commodities boom was largely driven by rising prices (terms of trade gains) rather than rising volumes. Accordingly, ELs have risen while EPs have remained relatively flat (Figure 14a) or even declined sharply, as in the case of Chile (Figure 14b).13 In effect, with the recent decline in copper prices, Chile has been losing ground in global export markets and hence losing its growth momentum. This raises therefore the broader question of whether LatAm commodity producers can continue to increase their shares in world exports and hence accelerate their growth relative to the rest of the world. To be sure, the strong growth rates that countries like Chile and Peru registered in the upswing phase of the recent commodities boom can in part be attributed to the progress they made in upgrading value added within commodities.14 12 This echoes the message in Lederman and Maloney (2012). 13 During the 1990s, by contrast, the rise in Chilean exports was driven mainly by increases in volumes. 14 Mandel (2011) provides evidence of significant upgrading towards higher-quality, higher-value-added varieties within minerals in Chile and Peru. He also shows that, contrary to popular perception, international trade in metals is characterized by a high degree of intra-industry trade and the room to upgrade within metal goods compares well within other manufacturing exports. 14 Yet, it is doubtful that solely relying on commodity upgrading will do the trick over the long haul. The scope for gaining ground through continuously rising commodity prices in the future, while possible in principle, is very much an open question. At the same time, these countries’ ability to gain ground in global markets through higher export volumes—relative to world exports—may also be limited as it hinges not only on constant improvements in production efficiency but also on whether world demand for commodities will rise faster than world income. Consider next the domestic response problem. Commodity dependence exposes countries to pronounced terms-of-trade cycles, which, in the absence of strong countercyclical policies, may induce major macroeconomic excesses during the upswing, followed by painful adjustments (collapsing DRs, as output growth falls behind import growth) during the downswing. Figure 14a illustrates this problem for the LatAm South countries. Their growth was clearly lifted by the strong positive terms of trade gains resulting from the China-induced boost in commodity prices. The latter translated into a parallel surge in EL, as the windfall gains allowed countries to boost their spending, thereby allowing real imports to jump beyond real exports without undermining nominal current account deficits. Yet, by spilling over into non tradables, the underlying boost in demand triggered overheating pressures in the non-tradable sector. The resulting (temporary) appreciation of the real exchange rate likely caused durable damage to the tradable sector. The large real exchange rate depreciations that took place in the downswing phase of the commodities boom, however, depressed the supply of non-tradables, and hysteresis prevented an elastic supply response in the tradable sector. While the strongly converging commodity exporters (the SCCEs—Chile, Peru, and Uruguay) were able to sustain their ground in terms of GDP growth during the downswing phase of the commodities cycle (starting in 2012), the weakly converging commodity exporters (the WCCEs—Argentina and Brazil) did not. As a result, a strong end-of-cycle growth differential (WCCE growth minus SCCE growth) emerged between the two groups (Figures 15a and 15b). The key difference originated from the much more aggressive domestic demand response of the WCCE countries, as captured by their much larger EL demand surge (Figure 14b). This surge reflected more procyclical policies—particularly fiscal (Figure 16a), but also monetary— that led to stronger real exchange rate appreciations during the boom (Figure 16b). Thus, the reversal of these higher real exchange rate appreciations during the downswing led to higher inflation (particularly for Argentina). At the same time, the bias in favor of non-tradable production during the boom weakened the regeneration of international market niches for the tradable sector, thereby prolonging the recessionary consequences of the commodity price reversal. As a result, the WCCEs ended up with much weaker DRs (Figure 15b), resulting in much lower GDP growth over the entire cycle, boom and bust. c) Mexico: A shift towards manufacturing exports that did not boost growth That Mexico is the second worst performer (after the República Bolivariana de Venezuela) in the region in terms of per capita income convergence over the past 25 years or so is quite a puzzle. Even more so considering that Mexico excels in other relevant respects, including the quality of its macro-financial policies, a high degree of trade openness, the tight links to NAFTA, and its successful shift towards rather diversified and complex manufacturing exports. Thanks to 15 this shift, Mexico quickly moved up the ladder of “economic complexity” and now ranks 1st in the region and 25th in the world (ahead of Canada, Hong Kong, and Spain, for instance) in this regard, according to the MIT’s Observatory of Economic Complexity. That alone should have led to higher growth, yet it did not.15 This section sheds light into Mexico’s nonconvergence puzzle, again with the help of the growth decomposition method discussed in Section 2.16 As Figure 9a clearly illustrates, Mexico’s real exports (the EP term) have been the main driver of growth over the last 35 years. Mexico’s entry into NAFTA in the early 1990s caused its exports to skyrocket, lifting growth along. Yet, Mexico’s exports collapsed 5 years later, once China entered the WTO, dragging growth down again. Chinese exports to the US clearly displaced Mexican exports (Figure 17), thereby dragging down Mexico’s EP (Figure 9a). This suggests that the timing of Mexico joining NAFTA—and not NAFTA per se—was an important factor behind its poor economic performance. Had Mexico joined NAFTA 10 years earlier, the rise in its export share could have endured enough to more likely boost the country’s growth. But the timing of NAFTA relative to the rise of China can hardly be the only explanation of Mexico’s depressed growth. Another reason is that Mexico’s export pull (EP), while rising, has lacked buoyancy because Mexico’s export penetration into the US market (Mexico’s dominant destination market) was dampened by the combination of a relatively slow growing US economy and the concentration of Mexican exports on relatively income-inelastic products—that is, products whose demand in the US does not grow faster than US income. Indeed, as illustrated by Figure 18, Mexico’s growth has closely followed that of the US from the time Mexico joined NAFTA (1995) until 2010.17 While the US is a country with a much higher per capita income than Mexico, hence with an inherently slower equilibrium growth rate, Mexico’s population still grows much faster than that of the US. Thus, coupling Mexico’s GDP growth with that of the US has resulted in a continued divergence of per capita incomes between the two countries. Moreover, compared to the South East Asian countries, Mexico continues to display a rather weak domestic response (DR) to its export pull (EP). To be sure, Mexico’s DR recovered somewhat during the China boom period from its earlier collapse. Yet it has continued to drag down growth to the present (Figure 10b). As documented in Ize (2019a), this was partly the result of limited factor mobility inside Mexico, which allowed for a very slow growing South to co-exist with faster growing and more outward-oriented Center and North regions, as well as a sharp worsening of the rule of law associated with a burst of drug-related criminality.18 15 According to Hausmann et al (2014), “countries whose economic complexity is greater than what we would expect, given their level of income, tend to grow faster than those that are ‘too rich’ for their current level of economic complexity. In this sense, economic complexity is not just a symptom or an expression of prosperity: it is a driver.” 16 A deeper analysis of Mexico’s growth challenges, following a similar growth decomposition methodology as in this paper, can be found in Ize (2019a and 2019b). 17 While Mexico’s recently rising exports suggests a welcome ability to adapt and reconstruct export niches while taking advantage of China’s rising wage costs, Mexico’s domestic response has again declined in recent years, suggesting that the rise in exports was partly offset by a concomitant rise in imports. 18 While Mexico’s GDP expanded as fast as that of the Eastern European countries (which underwent a similar process of vigorous trade integration as Mexico), Mexico’s per capita income lagged significantly that of Eastern Europe. This again indicates that, despite its successful shift to manufacturing exports, the Mexican economy has a deep-seated inability to absorb productively its relatively fast-growing labor force, much of which ends up in the informal sector. 16 d) Central American services exporters: A matter of rule of law The rest of LatAm North listed in Table 2 encompasses Central American countries and the Dominican Republic. These countries have two features in common. First, they are services exporters—their shares of services in total exports lie in the 30 to 40 percent range, compared to around 15 percent in the case of the South American commodity exporters (Figure 19). Second, they run relatively large deficits in the trade of goods and non-factor services account, averaging around 5 percent of GDP in the past 15 years. Yet they have had very different growth performances, with Costa Rica, Dominican Republic, and Panama classified as “strongly converging commodity importers” (SCCI); and El Salvador, Guatemala, and Honduras as “weakly converging commodity importers” (WCCI). The SCCIs’ faster growth was driven by a higher DR at the beginning of the period and a more dynamic EP towards the end (Figure 20). This reflected a significant initial expansion in industry (likely driven by a construction boom), followed by a substantial growth of services (Figure 21a), much of it exported (Figure 19). In the WCCI countries, services were also the most dynamic sector of the economy (Figure 21b), but arguably much of it took place in the informal sector and, thus, did not add much to overall growth. Notably, the SCCIs and the WCCIs differed markedly in the way in which they financed their current account deficits: with remittances in the case of the WCCIs, FDI in the case of the SCCIs (Figure 22). While the labor force in the SCCIs has stayed at home to work with incoming FDI, much of the labor force in the WCCIs has emigrated to work abroad. The preponderance of FDI inflows (which facilitate learning and technology transfer) is consistent with the superior growth performance of the SCCIs. Instead, the preponderance of remittances inflows (which help to support consumption, hence alleviate poverty) seems to have systematically undercut growth and convergence in the WCCIs.19 Hence, the better performance of the SCCI group can be traced to a stronger services- related export pull as well as an easier availability and higher quality (FDI-based) of external finance. While a full explanation of the contrast between FDI-reliance versus remittances-reliance cannot be reduced to a single factor, the marked differences in the rule of law between the two groups of countries stand as a first, rather obvious underlying determinant (Figure 23). Services seeking customers and FDI inflows are unlikely to flock in from abroad to a country where the rule of law suffers from major weaknesses.20 19 Shapiro and Mandelman (2014) find adverse productivity effects of remittances, a result of negative work incentives and weaker firm dynamics. Higher remittances are also associated with lower saving rates, another factor behind slower growth. 20 More generally, converging countries, whether in LatAm North or LatAm South, feature better institutional quality and a lower incidence of violence. To be sure, causality between institutional quality and growth runs in both directions. Weak rules of law discourage investment and limit countries’ export potential. But the lack of growth can further weaken institutions, partly by dampening the demand for and supply of quality governance. 17 5. Looking ahead To achieve faster growth, LatAm needs to boost its production of tradable goods and services, both in terms of exportables (to raise their EPs) and importables (to raise their DRs). The road ahead is complicated by a host of potential threats associated with the global environment that domestic policies cannot influence. These include the risk of rising protectionism in advanced economies, a possible secular decline in the world demand for commodities (particularly in fossil fuels and minerals), and the unpredictable impacts on world trade (particularly, on labor-based manufacturing) of the unrelenting progress toward digitalization and automation. In view of these uncertain (if not downright adverse) external prospects, a policy option that may appeal to many a politician in the region but that must be vigorously avoided is to shift again toward an inward-oriented growth strategy supported by a boost in domestic demand and heavily protectionist policies. LatAm’s history and the worldwide empirical evidence provide a clear warning that while this course of action might produce short-term gains, it would be disruptive of macroeconomic stability and fatal to long-term growth and convergence. Instead, the emphasis needs to be placed on stimulating the supply of tradable goods. Where possible, this might imply an expansion in the local production of importables. In all cases, policies should facilitate moving up the value-added ladder in the production of exportables while diversifying exports and export destinations as much as possible. In part, export diversification may be attained via deeper LatAm-LatAm or South-South integration. The South has probably by now acquired the critical mass necessary for such course of action to be viable. In fact, the South is likely to grow faster than the North, even with limited trade interaction within the South. And notwithstanding the meager returns thus far from regional integration, further efforts should be made to turn Latin America into an entirely free trade region, thereby better leveraging larger markets and the associated economies of scale. The pursuit of South-South integration has limits, however, and is not a good substitute for expanding export niches into the richer economies. Hence, LatAm should put a premium on building up and diversifying its export base with an eye towards global markets, with South-South integration pursued so as to boost global integration. Moreover, for much of the region (particularly Mexico), sitting next to the largest consumer market on earth remains a critical asset. Given anti-globalization sentiments in the US and other advanced economies, export promotion efforts should emphasize win-wins over zero-sums. Gaining exporting ground in richer destination markets will be politically easier to achieve if, instead of displacing firms and dislodging jobs in the importing countries, LatAm exports are mainly focused on goods and services that are naturally buoyant and do not enter into direct conflict with the local production in the advanced countries. This implies a boost in creativity (not just productivity), and a repositioning toward tradable services of all kinds (tourism, health and wellness services, old age, or education, to cite just a few), which are on average more demand elastic (i.e., they account for a rising share of global consumption) and generate more employment than manufactures. By 18 pulling labor out of the informal sector into more productive formal employment, the expansion of tradable services can also deliver a needed productivity boost. The shift toward more creative exports puts the accent on knowledge and on attracting and retaining the talent that is germane to the development of export niches. Given globalization, many non-commodity goods and services can increasingly be produced anywhere in the world; hence, to raise growth, LatAm needs to become an attractive place to produce in. Moreover, a shift toward the export of tradable services would require making the region unquestionably attractive to foreigners as a place to visit or live in. This highlights the need for a more sophisticated outwardly- oriented strategy. Making the region an enticing hub to visit, live in, or work in puts a premium on stronger institutions (particularly the rule of law), more efficient and integrated infrastructure (including digital infrastructure), safer living environments, friendly and educated citizens, clearer air and greener development, and well preserved natural and cultural capital. Finally, growth agendas in LatAm will need to be supported by stronger countercyclical macro-financial policies, which are essential to avoid the growth-impairing consequences of amplified boom-bust phenomena, particularly in commodity exporting countries. Of course, implementing countercyclical policies when external factors are favorable (yet unlikely to last long) may be an uphill battle politically speaking. The technical challenges are indeed easier to handle than the political ones. Pressures induced by rising expectations in deeply unequal societies make it difficult for policy makers to postpone spending today (when external factors are auspicious) in anticipation of large adjustment costs tomorrow. Social pressures and distributional concerns intensify when the fiscal profligacy in the boom requires belt-tightening in the bust. The above clearly suggests that tackling social inequities will be as crucial to unleashing export and GDP growth as other growth-oriented structural policies (for instance, regarding education, skill formation, infrastructure, and contract rights). Given that social and regional fractures and inequities will otherwise undermine the domestic response to an export pull, the region will need to significantly improve its social protection systems (pensions, health) and provide quality education for all. Rather than an afterthought, good social policy should be thought as a core component of a sound growth-oriented reform program. 19 References Aghion, P. and P. Howitt (2007), “Capital, Innovation and Growth Accounting”, Oxford Review of Economic Policy Vol. 23, No 1, pp. 79-93. Araujo, J.T., M. Brueckner, M. Clavijo, E. Vostroknutova and K. Wacker (2014), Benchmarking the Determinants of Economic Growth in Latin America and the Caribbean, World Bank. De la Torre, A. and A. Ize (2019). “Accounting for Growth: An Open Macro Perspective.” Mimeo. Fortin, N., T. Lemieux and S. Firpo (2010), “Decomposition Methods in Economics”, NBER WP16045. Hausmann, R. et al. (2014). The Atlas of Economic Complexity: Mapping Paths to Prosperity, MIT Press Books. Hausmann, R. J. Hwang and D. Rodrik (2005). “What you Export Matters.” NBER WP11905. Ize, A. (2019a). “A Note on Mexico’s Growth”, Estudios Economicos, vol. 34(1), pp. 123-155. Ize, A. (2019b). “El Modesto Crecimiento de México: Una Descomposición Macroeconómica Contable”, Revista de Economía Mexicana, UNAM (forthcoming). Lederman, D. and W. Maloney (2012). Does What You Export Matter? In Search of Empirical Guidance for Industrial Policies. World Bank. Mandel, B. (2011). “The Dynamics and Differentiation of Latin American Exports”. Federal Reserve Bank of New York Staff Report 508. Pages, C., ed. (2010). The Age of Productivity. Inter-American Development Bank, Washington, D.C. Prados de la Escosura, L. (2007). “When did Latin America Fall Behind,” in The Decline of Latin American Economies: Growth, Institutions and Crises, edited by S. Edwards, G. Esquivel, G. Márquez, The University of Chicago Press (by NBER). Sachs, J. and A. Warner (2001). "The Curse of Natural Resources". European Economic Review, vol. 45, issue 4-6, pp. 827-838. Shapiro, A. and F. Mandelman (2014). “Remittances, Entrepreneurship, and Employment Dynamics over the Business Cycle.” FRB Atlanta Working Paper, No. 2014-19. Rodrik, D. (2008). “The Real Exchange Rate and Economic Growth”, Brookings Papers on Economic Activity, 39:2 (Fall), pp. 365-439. 20 Tables Table 1. Growth Variance Decomposition 1970‐2000  2000‐2017  EP  DR  EL  EP  DR  EL  TOTAL  0.95  0.23  ‐0.20  0.77  0.27  ‐0.06  HIGH INCOME  1.00  0.06  ‐0.11  0.69  0.40  ‐0.12  LOW INCOME  1.02  0.23  ‐0.28  1.21  ‐0.13  ‐0.09  HIGH GROWTH  1.60  ‐0.56  ‐0.07  1.14  0.05  ‐0.20  LOW GROWTH  0.92  0.45  ‐0.40  0.49  0.42  0.06  Source: WDI Table 2. LatAm Countries’ Convergence Patterns During 1990-2017 Commodity Exporters Commodity Importers SCCIs Strongly SCCEs Panama, Costa Rica, converging Chile, Peru, Uruguay Dominican Republic MCCEs Moderately MCCI Bolivia, Colombia, Ecuador, converging Nicaragua Paraguay WCCIs Weakly WCCEs El Salvador, Guatemala, converging Argentina, Brazil Honduras Non NCCE NCCI converging Venezuela, RB Mexico Figures Figure 1. Per Capita Income Relative Growth and Export Shares (a) 1960-2000 (b) 2000-2015 Note: Per capita GDP is in constant US dollars; export shares are in current US dollars. Each dot represents a country and the value it takes along the vertical and horizontal axes is the yearly geometric average of the respective rates of change over each period. The orange and blue bars give the median income per capita and the share of countries in each of the four quadrants. Source: WDI. Figure 2. GDP per Capita Relative to the World’s, 1950-2000 3.00 2.50 2.00 1.50 1.00 0.50 0.00 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 LatAm South East Asia Southern Europe Notes: LatAm includes Argentina, Brazil, Chile, Colombia, Costa Rica, Ecuador, Mexico, Panama, Peru, Uruguay and the República Bolivariana de Venezuela. South-East Asia includes China, Indonesia, Malaysia, Singapore, Thailand and the Philippines. Southern Europe includes Greece, Italy, Portugal and Spain. Source: Maddison Project. Figure 3. Gross Export Composition for Selected Countries, 1990 DIVERSIFIED Note: A country is classified as a “commodity exporter” if in 1990 at least 40 per cent of its exports originated from the primary sector (raw or processed commodities); and as a “specialized” commodity exporter when its exports in any given class of commodities (agricultural, mineral, hydrocarbons) exceeded 45 per cent of total exports; otherwise, it is classified as a “diversified commodity producer”. Source: Comtrade. Figure 4. GDP Per Capita for Commodity Exporters (1950=1) (a) LatAm Countries (b) LatAm vs Comparator Countries 1.6 1.30 1.20 1.4 1.10 1.2 1.00 0.90 1 0.80 0.8 0.70 0.60 0.6 0.50 0.4 0.40 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 1 95 0 1 95 1 1 95 2 1 95 3 1 95 4 1 95 5 1 95 6 1 95 7 1 95 8 1 96 9 1 96 0 1 96 1 1 96 2 1 96 3 1 96 4 1 96 5 1 96 6 1 96 7 1 96 8 1 97 9 1 97 0 1 97 1 1 97 2 1 97 3 1 97 4 1 97 5 1 97 6 1 97 7 1 97 8 1 98 9 1 98 0 1 98 1 1 98 2 1 98 3 1 98 4 1 98 5 1 98 6 1 98 7 1 98 8 1 99 9 1 99 0 1 99 1 1 99 2 1 99 3 1 99 4 1 99 5 1 99 6 1 99 7 1 99 8 2 00 9 0 1 95 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 Argentina Brazil Chile Colombia Mexico Peru Uruguay Venezuela LatAm specialized LatAm diversified LatAm special ized LatAm diversified World di versified World specialized Notes: LatAm specialized is the simple average of Argentina, Chile, Peru, Uruguay and the República Bolivariana de Venezuela. LatAm diversified is the simple average of Brazil, Colombia and Mexico. World specialized commodity exporters include New Zealand, South Africa and an average of eight oil exporters (Algeria, Bahrain, Cameroon, Gabon, Kuwait, Nigeria, Qatar and United Arab Emirates). World diversified commodity exporters include Australia and Canada. Countries’ per capita GDPs are divided by the GDP of the US. Source: Maddison Project. Figure 5. GDP Per Capita and Real Exports for Agricultural Exporters (a) GDP Per Capita (1950=1) (b) Export Volume (1960=1) 1.3 1.10 1.2 1.00 1.1 0.90 1 0.9 0.80 0.8 0.70 0.7 0.6 0.60 0.5 0.50 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 00 98 0.4 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 19 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 Argentina Uruguay New Zealand Argentina Uruguay New Zealand Note: Countries’ GDP per capita and real exports are divided by the world’s. Source: Maddison Project and WDI. Figure 6. GDP Per Capita and Real Exports for Oil Exporters (a) GDP per Capita (1950=1) Export Volume (1980=1) 1.2 1.40 1.1 1.20 1 1.00 0.9 0.80 0.8 0.60 0.7 0.40 0.6 0.20 0.5 0.00 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 0.4 80 81 82 83 84 85 86 87 88 9 0 1 2 3 4 5 6 7 8 9 0 98 99 99 99 9 9 9 9 9 99 99 00 19 19 19 19 19 19 19 19 19 19 19 19 19 19 1 1 1 1 1 1 2 Venezuela Other oil producers Venezuela Other oil producers Note: Countries’ GDP per capita and real exports are divided by the world’s. The comparator group of oil exporters includes Algeria, Bahrain, Cameroon, Gabon, Kuwait, Nigeria, Qatar, and the United Arab Emirates. Source: Maddison Project and WDI. Figure 7. GDP Per Capita and Real Exports for Mineral Exporters (a) GDP per Capita (1950=1) (b) Export Volume (1960=1) 1.20 1.6 1.10 1.4 1.00 1.2 0.90 1 0.80 0.8 0.70 0.6 0.60 0.4 0.50 0.2 0.40 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 0 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 Chile Peru South Afri ca Chile Peru South Africa Note: Countries’ GDP per capita and real exports are divided by the world’s. Source: Maddison Project and WDI. Figure 8. Growth Decompositions for LatAm Commodity Exporters (a) EP + EL (b) DR 0.80 0.50 0.40 0.60 0.30 0.40 0.20 0.10 0.20 0.00 0.00 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 96 96 96 97 97 97 97 97 98 98 98 98 98 99 99 99 99 99 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ‐0. 10 65 7 69 71 73 75 77 79 81 83 85 87 89 1 93 95 97 99 6 99 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 1 ‐0. 20 ‐0.20 ‐0. 30 ‐0.40 ‐0. 40 ‐0.60 ‐0. 50 ‐0. 60 ‐0.80 Diversified Undiversified Diversified Undiversified Note: The variables EP, EL, and DR are calculated according to the growth decomposition method explained in Section 2a. “Diversified” include Brazil, Colombia and Mexico. “Undiversified” include Argentina, Chile, Peru and Uruguay. Source: WDI. Figure 9. Manufacturing: ISI LatAm Countries and South-East Asia (a) As Share of GDP (b) As share of Exports 30 80 28 70 26 60 24 50 22 40 20 30 18 16 20 14 10 12 0 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 LatAm ISI South East Asia LatAm ISI South East Asia Note: Countries in South-East Asia includes the Republic of Korea, Malaysia and Thailand; countries in LatAm ISI includes Brazil, Colombia and Mexico. Source: WDI. Figure 10. Mexico: Growth Decomposition, 1965-2017 (a) G and EP (b) EP, DR, and TOT Mexico into NAFTA China into WTO Trade liberalization Note: The variables G, EP, EL and DR are calculated according the growth decomposition method explained in Section 2a. All variables are computed taking the logs of the average yearly growth rates over a ten-year moving window. Source: WDI. Figure 11. The Impact of China on LatAm’s GDP Per Capita Growth 0.21 China's entry  0.19 into the WTO 0.17 0.15 0.13 0.11 0.09 0.07 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 LatAm Upper Middle Income South East Asia Source: WDI Figure 12. Terms of Trade Bifurcation: LatAm North and LatAm South 1.2 1.4 1.1 1.3 Regional bifurcation 1 1.2 0.9 1.1 0.8 1 Region-wide 0.7 0.9 commodity dependence 0.6 0.8 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 LatAm South LatAm North (right axis) Note: LatAm South includes Argentina, Brazil, Chile, Colombia, Bolivia, Peru, Ecuador, Uruguay and the República Bolivariana de Venezuela; LatAm North includes Mexico, Costa Rica, Panama, El Salvador, Dominican Republic, Panama, Guatemala, Honduras, and Nicaragua. Source: WDI. Figure 13. LatAm and the China Pull: Convergence Patterns Source: WDI. Figure 14. Growth Decomposition (a) LatAm South (b) Chile 0.50 0.60 0.40 0.50 0.30 0.40 0.30 0.20 0.20 0.10 0.10 0.00 0.00 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 ‐0.10 ‐0.10 ‐0.20 ‐0.20 ‐0.30 ‐0.30 ‐0.40 G EP EL DR TOT G EP EL DR TOT Note: The variables G, EP, EL and DR are calculated according the growth decomposition method explained in Section 2a. All variables are computed taking the logs of the average yearly growth rates over a ten-year moving window. Source: WDI. Figure 15. Growth Decomposition (a) SCCEs (b) WCCEs 0.80 0.80 0.60 0.60 0.40 0.40 0.20 0.20 0.00 0.00 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 ‐0.20 ‐0.20 ‐0.40 ‐0.40 ‐0.60 ‐0.60 G EP EL DR G EP EL DR Note: The variables G, EP, EL and DR are calculated according the growth decomposition method explained in Section 2a. All variables are computed taking the logs of the average yearly growth rates over a ten-year moving window. Source: WDI. Figure 16. WCCEs vs. SCCEs: Selected Macroeconomic Indicators (a) Fiscal Balances (b) Real Exchange Rates 6 0.10 4 0.08 0.06 2 0.04 0 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 00 00 00 00 00 00 00 00 00 00 01 01 01 01 01 01 01 01 01 0.02 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 ‐2 0.00 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 ‐4 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 ‐0. 02 ‐6 ‐0. 04 ‐8 ‐0. 06 ‐10 ‐0. 08 WCCEs SCCEs SCCE e WCCE e Note: Real exchange rates are computed taking the logs of the average yearly growth rates over ten-year moving windows. Fiscal balances are year to year data. Source: WDI and WEO (IMF). Figure 17. China’s Impact on Mexican Trade: US Import Shares Mexico in NAFTA China in WTO China’s net impact Source: US Census Bureau and WDI. Figure 18. Mexico and US Growth 0.5 0.4 0.3 0.2 0.1 0 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 ‐0.1 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 ‐0.2 ‐0.3 ‐0.4 ‐0.5 G Mex G US A Note: G is the country’s GDP growth rate relative to the world’s, taking the logs of the average yearly growth rates over a ten-year moving window. Source: WDI. Figure 19. Services Exports as Percent of Total Exports 40% 35% 30% 25% 20% 15% 10% 5% 0% WCCEs SCCEs WCCIs SCCIs 1990 2015 Note: Country groupings follow Table 2. Source: WDI. Figure 20. Growth Decomposition in Central American Countries (a) SCCIs (b) WCCIs 0.60 0.60 0.50 0.50 0.40 0.40 0.30 0.30 0.20 0.20 0.10 0.10 0.00 0.00 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 ‐0.10 ‐0.10 ‐0.20 ‐0.20 ‐0.30 ‐0.30 ‐0.40 ‐0.40 G EP EL DR G EP EL DR Note: All variables are derived using the methodology presented in Section 2.b, taking the logs of the average yearly growth rates over a ten-year moving window. Grouping follows Table 2. Source: WDI. Figure 21. Growth by Sectors in Central American Countries (a) SCCIs (b) WCCIs 0.07 0.07 0.06 0.06 0.05 0.05 0.04 0.04 0.03 0.03 0.02 0.02 0.01 0.01 0.00 0.00 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Services Agriculture, forestry, and fishing Manufacturing Industry Services Agriculture, forestry, and fishing Manufacturing Industry Note: The growth rates are measured over a ten-year moving window, as in the case of all terms of the growth accounting decompositions. Country groupings follow Table 2. Source: WDI. 42 Figure 22. Commodity Importers: External Financing Items (a) Foreign Direct Investment as % of GDP (b) Remittances as % of GDP 10 12% 9 8 10% 7 8% 6 5 6% 4 3 4% 2 2% 1 0 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 WCCEs SCCEs WCCIs SCCIs SCCIs WCCIs 1990 2016 Note: Country groupings follow Table 2. Source: WDI. Figure 23. Rule of Law in Latin America (a) Crime Statistics (b) Institutional Quality 600 3.0 0.7 500 2.5 0.6 0.6 400 2.0 0.5 300 1.5 0.5 200 1.0 0.4 100 0.5 0.4 0 0.0 0.3 SCCEs MCCEs & SCCIs MCCIs & SCCEs MCCEs & SCCIs MCCIs & WCCIs WCCEs WCCIs WCCEs Homicides Kidnappings (right axis) Absence of corruption Police and Justice Quality Note: Country groupings follow Table 2. Homicides are based on reported cases; kidnappings are probability estimates based on surveys; data on corruption and institutional quality are also survey-based (a higher number reflects better quality). Source: World Justice Project, United Nations Office on Drugs and Crime (UNDOC) and NYA International. Appendix Tables Table 1. Regressions of Per Capita Income Convergence Equation (1) (2) (3) (4) Dependent Variable Income Exports _____________________ Coeff. t Stat. Coeff. t Stat. Coeff. t Stat. Coeff. t Stat. Constant ‐0.001 ‐1.56 ‐0.001 ‐1.78 ‐3.573E‐05 ‐1.05 0.000 ‐1.27 Income 0.010 3.55*** 0.010 3.68*** L(income) 0.100 2.25*** 0.131 2.96*** 0.003 1.175 LL(income) ‐0.037 ‐0.849 ‐0.021 ‐0.487 0.003 1.26 Exports 2.561 3.55*** 1.654 2.38*** L(exports) 2.824 3.76*** ‐0.351 ‐7.80*** ‐0.342 ‐7.64*** LL(exports) 2.055 2.90*** ‐0.151 ‐3.39*** ‐0.137 ‐3.13*** ______________________________________________________________________________________________________ R Square 0.06 0.03 0.13 0.12 Adjusted R Square 0.05 0.02 0.12 0.12 Standard Error 0.01 0.01 0.00 0.00 Observations 482 482 482 482 SSE 0.065 0.067 0.000 0.000 MSE 0.000 0.000 0.000 0.000 F value: 8.66 1.67 P statistic 0.0002 0.1884 Table 2. Granger Causality of Per Capita Income Convergence (1) (2) (3) (4) (5) (6) (7) 1960‐2000 2000‐15 Full Sample Full Sample Full Sample Full Sample Full Sample Small Countries Large Countries 0.50*** 0.45*** 0.26*** 0.29*** 0.24*** Rate of change of Export Shares (0.04) (0.04) (0.024) (0.046) (0.03) 2.12E‐11*** 2.27E‐11 2.21E‐11*** 2.73E‐11*** 3.55E‐11*** 2.02E‐11*** 2.18E‐11*** Population (1.04E‐11) (1.35E‐11) (0.90E‐11) (7.88E‐12) (10.19E‐12) (7.3E‐10) (7.60E‐12) ‐0.54*** ‐0.34 ‐0.48*** ‐0.38*** ‐0.31*** ‐0.41*** ‐0.45*** Population Growth (0.16) (0.21) (0.16) (0.08) (0.11) (0.11) (0.14) ‐0.53E‐07 ‐1.11E‐07 ‐2.6E‐08 ‐2.18E‐08*** ‐4.25E‐07*** ‐1.07E‐07 ‐4.29E‐07*** GDP x Capita (1.94E‐07) (2.44E‐07) (1.88E‐07) (6.8E‐08) (9.09E‐08) (9.25E‐08) (1.1E‐07) ‐0.017*** ‐0.00 ‐0.0059 Commodity Exporters (0.0030) (0.044) (0.0034) Multiple R 0.89 0.70 0.90 0.77 0.50 0.73 0.83 R Square 0.80 0.49 0.81 0.60 0.25 0.53 0.69 Adjusted R Square 0.79 0.46 0.79 0.59 0.23 0.50 0.67 Standard Error 0.01 0.01 0.01 0.01 0.02 0.01 0.01 Observations 73 69 66 143 143 71 73 Note:The table shows the coefficients (and standard deviations) of least squares regressions where the dependent variable is the average yearly rate of growth of per capita country income; GDP per capita is taken at the beginning of each period; countries are classified as commodity exporters when the share of manufacturing in merchandise exports is less than 50% and the ratio to GDP of trade in services is less than 40%. Source: WDI. 46