Jump to ContentJump to Main Navigation
Beating the OddsJump-Starting Developing Countries$

Justin Yifu Lin and Célestin Monga

Print publication date: 2019

Print ISBN-13: 9780691192338

Published to Princeton Scholarship Online: May 2020

DOI: 10.23943/princeton/9780691192338.001.0001

Show Summary Details
Page of

PRINTED FROM PRINCETON SCHOLARSHIP ONLINE (www.princeton.universitypressscholarship.com). (c) Copyright Princeton University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in PRSO for personal use. Subscriber: null; date: 26 September 2021

Reaping the Dividends of Globalization: A Winning Road Map

Reaping the Dividends of Globalization: A Winning Road Map

(p.254) 7 Reaping the Dividends of Globalization: A Winning Road Map
Beating the Odds

Justin Yifu Lin

Célestin Monga

Princeton University Press

Abstract and Keywords

This chapter analyzes the conditions needed to design and implement successful special economic zones and industrial parks. It discusses long-term trends and fundamental issues in global trade since trade is the main source of growth for low-income countries that have limited domestic demand. In recent years the story of global trade has often been presented by some economists and development experts as a cause for concern on the export-led growth model that made possible the so-called Asian Miracle that is no longer available for poor countries in Africa or South Asia. But statistics appear to show a turning tide: the value of world merchandise exports rose from $2.03 trillion in 1980 to $18.26 trillion in 2011, equivalent to 7.3 percent growth per year on average in current dollar terms according to WTO trade statistics. But from 2012 to 2014, world trade growth averaged only 2.2 percent, well below the average for the proceeding 20-year period. This has raised the question whether the same shaping factors that have given rise to today's global trade system are likely to continue in the medium and long term.

Keywords:   special economic zone, industrial parks, global trade system, domestic demand, Asian Miracle, Africa, South Asia, world merchandise exports, world trade growth

One day in 1946 in Argentina, shortly after a president had come to power in a military coup, a representative of the new government went to the prestigious Buenos Aires Library to inform one of its staffers that his tenure was over. The staffer in question had written some fictional short stories and had contributed to a literary journal and was thus perceived in some political circles as a nuisance. His dismissal was conceived with some humor: he was “promoted” to inspector of rabbits and poultry of public markets. When he asked why, the response was straightforward: his humiliating new assignment was justified because of his (perceived) political opinions. He immediately resigned from the inspector position and found himself with no other career choice than to search for his true calling.

He had previously been an unsuccessful poet and writer, and he was shy and not considered articulate—not even by his own mother. He was also rapidly becoming blind. Regardless, some of his friends helped him find a job as a literature professor, which eventually led to invitations to teach, write, and lecture in some of the most prestigious venues and academic institutions in the world. He had carefully reflected on what he knew best and on what he could do better than almost anyone else. He prepared his public appearances meticulously and delivered them to make the maximum impact. And he realized that his escape route from failure was to look for intellectual and professional opportunities well beyond his country’s boundaries. He had no choice but to try to sell his skills on the global labor market.

(p.255) To his surprise and to the amazement of his family, friends, and adversaries alike, his search for a new career and life strategy was a success. His literary, oratory, and teaching skills quickly received recognition beyond the Argentinian market. His talents and services were in high demand in North America, Europe, and Asia. He became so good at his new occupation that he rapidly established himself as one of the most sought-after international lecturers in the world. He mesmerized global audiences in sold-out lectures across the globe. He wrote many books that became instant classics, earned him many prestigious literary prizes, and cemented his place as the master of the modern short story and one of the twentieth century’s finest writers. Thanks to his dismissal from the library, Jorge Luis Borges—yes, it was him—went from a complete unknown to one of the most towering figures in literary history. Losing his obscure assistant job turned out to be a major career blessing and source of revenue and fame—and eventually, after the military dictator was overthrown in 1955, he was even appointed director of the Buenos Aires Library.

Borges’s success story is a metaphor for the potential benefits that a crisis can bring to one’s life—and to economic policy making—and for why developing countries should never shy away from global trade. When hit suddenly by a negative shock that forced him to change his life trajectory, Borges had to do some soul searching and reallocate his (intellectual) resources to activities that he knew he could perform best. He found the right entry point for his work and focused on services that could quickly generate the highest payoffs. Because of his country’s political circumstances and the absence of a national market for his services, he was also forced to “think global” and to connect to international markets and networks where his expertise was not only most needed but also most highly rewarded. Perhaps unwittingly, he traded his skills on the international market, found a gratifying niche for his talents, and rapidly gained from it. Also, it is very likely that Argentina eventually reaped some balance-of-payments gains from his economic success abroad.

What Borges did was to identify and tap into his latent comparative advantage and exploit the opportunities beyond his nation’s borders. The major crisis that could have ended his career forced him to see opportunities beyond his traditional horizon. His story also provides a metaphor for any low-income country that has been facing economic shocks. The 2008–09 financial and economic crisis should have been a similar moment (p.256) of truth. Not for soul searching—the general recipe for success is readily available—but for redefining a country’s economic strategy and focusing its resources in activities where it has comparative advantage.

This chapter outlines a road map for doing just that. It starts by discussing long-term trends and fundamental issues in global trade—because trade is the main source of growth for low-income countries that have limited domestic demand. In recent years the story of global trade has often been presented by some economists and development experts as a cause for concern—concerns that the export-led growth model that made possible the so-called Asian Miracle is no longer available for poor countries in Africa or South Asia. Statistics appear to show a turning tide: the value of world merchandise exports rose from $2.03 trillion in 1980 to $18.26 trillion in 2011, equivalent to 7.3 percent growth per year on average in current dollar terms (WTO trade statistics). But in 2012–2014, world trade growth averaged only 2.2 percent, well below the average for the proceeding 20-year period. This has raised the question whether the same shaping factors that have given rise to today’s global trade system are likely to continue in the medium and long term. Development economists’ fear is palpable. They wonder whether transport and communication costs will maintain their dramatic, linear decline from continued incremental technological improvement or the introduction of entirely new technologies. They also debate whether the marginal improvements observed in recent decades will diminish in the future, making declining transport and communications costs a less salient shaping factor for world trade—even leading to a slowing of trade growth.

Such questions are the reaction to the analysis of fundamental changes in global trade patterns, showing that developing countries can still gain from a much larger world economy, but they must confront the new barriers to trade and organize their economies to enter global value chains. As Jawaharlal Nehru, India’s first prime minister, famously said, “Every little thing counts in a crisis.” The current global economic crisis provides a unique opportunity for developing economies to also carefully review their strategies, policies, and instruments for reaping new benefits from an increasingly interlinked world economy. The second part of this chapter outlines the benefits of clusters, discusses what is wrong with the current institutions for fostering trade (most notably the special economic zones), and presents guiding principles for building and running them effectively. (p.257)

Global Trade Trends and Patterns: Nothing to Fear but Fear Itself

An important new narrative in the economic discourse on global growth tells the story of a deep decline in global trade with potentially catastrophic consequences for the world economy. In the words of Gavyn Davies, “world trade has lost its mojo,” and global trends support his observation. From 1990 to 2008 global real GDP expanded at a 3.2 percent annual rate, while world trade volume grew at 6.0 percent. Since 2008, however, world trade has grown slightly slower than GDP, so the share of exports in GDP has actually fallen after a twenty-five-year uptrend (Davies 2013). The first reason many researchers provide for this trend is that protectionism has been on the rise (Evenett 2013). This raises fears of a repeat of the protectionist disaster in the 1930s and has led to the logical conclusion that multilateral negotiations on trade barriers should be revived and made the centerpiece of the world’s economic agenda.

Another, even more disturbing suggestion is the idea that the export-led growth model, allowing many previously poor countries such as Korea and China to lift large parts of their populations out of poverty and become dominant global economies in the past few decades, may have run its course. In the words of Howard Pack (2010):

The earlier experience is likely to be difficult to replicate. First, the generation of huge industrial complexes benefitting from agglomeration economies, particularly in China, makes it difficult for new entrants to compete. Second, over the next decade the major international macroeconomic adjustment will consist of the reduction in excess demand by a few countries, notably the United States and a concomitant increase in domestic absorption of GDP in a number of surplus countries mainly in Asia. To achieve a share of (relatively) contracting U.S. imports will not be easy nor will it be easy for non-Asian emerging markets to penetrate the Chinese and other Asian markets as they redirect their own production toward domestic uses.

These are valid concerns because they reflect the recent trends and changing patterns of global trade. But the facts and figures on which they are based can be analyzed differently. It would be unfair to compare the (p.258) new trade skeptics’ views to the export pessimism of the early structuralists in the 1950s. But their thesis is somewhat reminiscent of Raul Prebisch (1950) and Hans Singer (1950), who interpreted the international trade slump in the Great Depression as a sign that developing countries would not be able to gain from opening their economies. They believed that the decline in the terms of trade against the export of primary commodities was a secular and almost irreversible phenomenon, and that any attempt to boost developing countries’ exports would simply result in the transfer of income from resource-intensive poor countries to capital-intensive rich countries. In Latin America, where such views were most influential, political leaders and social elites chose to adopt inward-looking economic policies not based on their economies’ comparative advantage, which eventually proved to be a misguided strategy.

The kind of trade skepticism expressed today by leading economists is not based on the principles of early structuralism. But they use analogous reasoning. Let us review their arguments. First, the initial proposition that global trade must always expand at a much higher pace than global output and that anything less would be detrimental to global economic prospects is not supported by historical evidence. While trade has grown faster than output since 1950, up through about 1970 that only represented a return to levels of trade relative to output that prevailed before World War I (Krugman 2013). One should not set the almost exponential rate of trade expansion after 1970 to be the norm. It reflects the profound changes in the global economic environment, the declining costs of trade among nations, new technological developments, and the understanding that the rules governing commercial and financial transactions among firms in nations across the world should be amended to benefit all. In addition, business cycles tend to produce large fluctuations in trade, much bigger in percentage terms than the changes in GDP.

Over the long term, the trade-GDP relationship is usually not a static one. Douglas Irwin (2002) examined the statistical relationship between world trade and world income (GDP) over three different epochs: the pre–World War I era (1870–1913), the interwar era (1920–1938), and the post–World War II era (1950–2000). He found that trade grew slightly more rapidly than income in the late nineteenth century, with little structural change in the trade-income relationship. His study also showed that in the interwar and postwar periods, the trade-income relationship can (p.259)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.1. U.S. trade-weighted tariffs on dutiable imports (percent) and historical periods, 1930–2008.

Source: USITC staff compilation from U.S. Department of Commerce statistics.

be divided into different periods owing to structural breaks, but since the mid-1980s trade has been more responsive to income than in any other period under consideration. It is therefore not too surprising or even a major concern that world trade volume has been virtually stagnant in recent years—or even on a declining trend. Irwin’s study also highlighted that the trade policy regime differed in each period, from the bilateral treaty network in the late nineteenth century to interwar protectionism to postwar liberalization under the General Agreement on Tariffs and Trade (GATT) and the World Trade Organization (WTO). The commodity composition of trade has also shifted from primary commodities to manufactured goods over the past century, but the results cannot directly determine the reasons for the increased sensitivity of trade to income.

Is protectionism the culprit for the recent decline in global trade? That contention too is not supported by the facts. As noted by Paul Krugman (2013), the rapid trade growth since World War II was driven by major waves of trade liberalization, which took place in advanced economies until about 1980 (figure 7.1) and in developing economies since then (figure 7.2), and by the almost linear decline of trade barriers over the course of the twentieth century.1

Now, skeptics could argue that there has been a reversal of the declining trend in average tariffs after the Great Recession of 2008, which in (p.260)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.2. Evolution of average tariffs in developing countries, 1980–2009 (in percent).

their view is the reason world trade grew by less than 3.0 percent in 2012 and 2013, compared with the precrisis average of 7.1 percent (1987–2007). C. Constantinescu, A. Mattoo, and M. Ruta (2015) have examined whether the slower growth in trade reflects simply the sluggishness of GDP or there is a deeper structural shift in the relationship between trade and GDP. They find that the long-term elasticity of trade with respect to income was 1.3 between 1970 and 1985, rose to 2.2 in the period 1986–2000, but reverted back to 1.3 in the 2000s. Looking specifically at the 2000s, they find that a 1 percent increase in income was associated with a 1.5 percent increase in trade from 2001 to 2007 and a 0.7 percent increase from 2008 to 2013. This suggests that that the change in the trade-income relationship cannot be entirely attributed to the financial crisis. But it cannot be inferred from these numbers that protectionism is the culprit. In fact, their empirical analysis shows that the changing structure of global trade and aggregate demand accounts for much more than any resurgence of protectionism.2

The main reasons for the declining trends in tariffs are the successful outcomes of multilateral negotiations pursued under the GATT after World War II—including the Uruguay Round, completed in 1994—and the granting of special preferences by industrialized countries to developing countries for their exports.3

(p.261) But tariffs are only part of the story of global trade flows and patterns. Many other important factors determine the flow of trade among countries. Firms involved in international trade are also and increasingly confronted with nontariff measures (NTMs) that can seriously restrict their business transactions through important changes in the quantities of goods and services traded, or their prices. NTMs represent a wide range of regulations and requirements other than customs tariffs, which all countries apply to exports and imports of goods and services. They include customs procedures, technical regulations, conformity assessments, and so on.4 They vary across products and countries and are often changed quickly and unilaterally by national governments and other sovereign entities. Their purpose is generally not to impose new forms of protectionism but to ensure public health and safety, environmental standards, compliance with various standards, or even some broad ethical values that authorities in each country may subjectively determine.

For instance, in many industries, compliance with international standards has become a critical criterion for firm performance in global markets. Goods exported from developing countries are often rejected at their destination country’s border on quality or sanitary grounds. In the case of perishable foods, entire consignments are destroyed at the point of entry—after the producers have already incurred substantial production and transportation costs. In other instances, these undesirable consignments must be shipped back to their place of origin. Such rejections not only cause major financial losses for the producers, but they can also damage their home country’s reputation and discourage business initiatives in other potentially competitive and profitable sectors. The rejection problem illustrates the limitations of the Doing Business prescriptions in developing countries: no amount of procedural changes or streamlining of documentation can overcome such situations.

Empirical studies based on various methodologies have estimated the negative impact of NTMs on trade, focusing on “direct” and “indirect” approaches.5 José de Sousa, Thierry Mayer, and Soledad Zignago (2012) have provided evidence that the difference between de jure protection (tariffs) and de facto protection (the tariff equivalent of crossing borders) is almost always much greater for developing countries, leading to the conclusion that most trade costs—other than those associated with geographical distance—are associated with NTMs. In most countries, few individual (p.262) firms (especially small and medium-sized enterprises) involved in international trade can comply with NTMs.

One obvious solution is to ensure that firms in developing countries can produce better-quality goods and services, which have been inspected, tested, and certified—through an internationally recognized accredited body—before they are shipped overseas. Although such a requirement may appear to burden production costs and slow down business processes, the savings will ultimately far outweigh the initial costs. Conformity assessment certificates will facilitate and expedite the transit process, and goods and services that go through such a rigorous process will pass through borders more quickly because the risk of rejection will be minimized.6 Another solution, not exclusive of the first one, is to develop special zones where such problems are most efficiently handled. This is discussed below.

Despite the NTMs, the general, long-term trend of global trade is still a very positive one for developing countries. Moreover, the declining general trend in average tariffs around the world is unlikely to be rolled back given the structural changes declining tariffs have induced in the global production system and the enormous win-win opportunities they have created for advanced and developing economies. The best indicator of that evolution is that many goods are manufactured now in several countries at the same time. Global trade is therefore no longer a series of transactions between countries producing individual goods and services within their national boundaries and exchanging them in international markets. It is often about collaboration and partnerships, even in an intensively more competitive world. Manufacturing is increasingly a network of global supply chains in which the various production stages take place in the most cost-efficient locations—regardless of where they are in the world (Baldwin 2011).

Perhaps nothing better illustrates the changes in recent years in world trade than a disaggregated picture of an aircraft, revealing the distribution of productive knowledge among nations and the need for cooperation. As shown in figure 7.3, the Boeing 787 aircraft, one of America’s more visible industrial flagship products and a major symbol of modern times, is actually the work of a very large number of firms from many different countries. Years ago the management team at Boeing adopted a business model aimed at cutting costs and reducing its employee count by outsourcing the design and manufacture of 787 parts all over the planet. The rationale was to avoid “betting the company” on 787 development and instead to convince (p.263)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.3. Partners across the globe are bringing the 787 together.

suppliers from various countries—including firms that were competitors to Boeing—to spend hundreds of millions of their own dollars to design and produce the parts. Even within the United States, the company managed to obtain substantial fiscal advantages from states and local governments that were interested in its potential for employment creation. Just to secure the final assembly jobs, the state of Washington granted Boeing a tax break so massive that the WTO deemed it an illegal subsidy (Groves 2013).

The Boeing production strategy has led to some controversial financial results.7 But it was successful in bringing together an unlikely coalition of firms in a carefully choreographed industrial ballet—one that illustrates well some of the major trends in today’s global economy. One of the engines of the 787 aircraft is produced in California, the other in the United Kingdom; the wing tips are from Korea; the movable trailing edge from (p.264)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.4. 100 years of changes and opportunities: product shares in world merchandise exports, 1900–2011 (percent).

Source: WTO (2014, 54). NES (“not elsewhere specified”) is used (a) for low-value trade and (b) if the trading party designation was unknown to the country or if an error was made in the partner assignment.

Australia; the horizontal stabilizer of the tail from Italy; the cargo access doors from Sweden; the forward fuselage from Japan; the passenger entry door from France; and so on. That example is just one illustration of the importance and dominance of global value chains in the world economy today and in the decades ahead.

The profound changes that have occurred in the composition of global trade in the past decades confirm the emergence of new patterns. Up to the end of World War II, commercial exchanges among nations were largely dominated by the exchange of raw materials and agricultural products for manufactured goods. Since 1945 manufactured goods or the components of manufactured goods have become the main features of trade, with their value increasing from 40 percent of world trade in 1900 to 75 percent in 2000, while agriculture’s relative share of world trade has steadily declined (figure 7.4). That change also reflects the shift from a world where trade took place between national economies selling nationally produced goods to an era of international exchange of goods often coproduced by competing national economies. Such an evolution leaves less room for high-impact protectionist policies. (p.265)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.5. Volume of world merchandise exports and GDP (percentage change), 1950–2013.

Source: WTO (2014).

If protectionism is not really a threat for world trade, what about the problem of global imbalances and the need for macroeconomic adjustment in advanced economies, said to limit the volume of exports that developing countries can expect in the future? That line of argument deserves scrutiny, too. Pack (2010), one of its main proponents, writes: “This is not to say that world exports will be absolutely stagnant, simply that rates of growth that prevailed from the 1960s to 2007 are extremely unlikely to continue. Realistically domestic markets will prove critical over the next decade.”

In fact, the actual size of the recent decline in global trade should be put in perspective: international trade has grown dramatically over the past three decades. According to WTO statistics, the value of world merchandise exports rose from $2.03 trillion in 1980 to $18.26 trillion in 2011, which is equivalent to 7.3 per cent growth per year on average in current dollar terms (figure 7.5). “Commercial services trade recorded even faster (p.266) growth over the same period, advancing from US$ 367 billion in 1980 to US$ 4.17 trillion in 2011, or 8.2 per cent per year. When considered in volume terms (i.e., accounting for changes in prices and exchange rates), world merchandise trade recorded a more than four-fold increase between 1980 and 2011” (WTO 2014, 55). The arithmetic logic mitigates the negative implications of a small decline in global trade. In other words, a reduction in a pie that had grown much bigger in size over decades sheds light on the fundamental observation that things may not be as bad as they appear.

Moreover, the world trading system has become more balanced, a positive development for global economic stability. Between 1980 and 2011 developing economies raised their share in world exports from 34 percent to 47 percent and their share in world imports from 29 percent to 42 percent (UNCTAD 2013).

It should also be noted that the world has entered a new era, with emerging economies becoming new growth poles. In the 1980s and 1990s, among the top five contributors to global growth all except China were G7 industrialized countries. But in 2000–2009 all except the United States were emerging economies—with China having become the top contributor. The trend has been reinforced in the aftermath of the 2007–2009 global crisis. The recovery was characterized by a two-speed pattern, with developing countries as a group growing more than twice as fast as high-income countries. That shift in economic weight is likely to produce major benefits for the world economy, with positive effects for both high-income and developing countries. For high-income countries, the growth of emerging economies will expand markets for their exports of capital goods and intermediate goods. For many developing countries that are still major producers of agricultural and natural resource commodities, greater consumption and production in the new growth poles will continue to support adequate prices for their commodity exports. And, thanks to the enormous benefits reaped from several decades of successful performance in trade, firms and governments in emerging economies are well positioned to provide funds for infrastructure and natural resource investment in developing countries.

The broader argument about global imbalances made by trade skeptics also deserves critical assessment. First, many believe that global imbalances, driven by an undervalued renminbi, were a major cause of the 2007–2009 global crisis (Bernanke 2007a, b). But it has become clear that the cause was the loose monetary policy introduced in 2001 by the U.S. Federal Reserve (p.267) Bank in response to the bursting of the dotcom bubble, magnified by financial deregulation and innovations in financial instruments, resulting in a boom in the U.S. housing market (Lin 2012a, b). The imbalance was caused by the large current account deficits in the United States and western Europe, except Germany, on the one hand, and the current account surpluses accumulated by oil-exporting countries, China, and Japan, on the other.8

Second, although size is always an important issue to consider, one should not forget that the real issue with large current account deficits is their sustainability, that is, whether they will be met by sufficient, timely, and affordable foreign capital inflows. Célestin Monga (2012) suggests that these deficits mainly reflect the general equilibrium interaction between many macroeconomic variables (national rates of saving and investment, fiscal, monetary and exchange rate policies, patterns of growth, and international trade, and so on). Moreover these variables themselves reflect the prevailing deeper macropolitical and sociocultural choices, which must be taken into consideration in the analysis of current account deficits. Philip Lane and Gian Maria Milesi-Ferretti (2014) have documented the substantial narrowing of current account global imbalances following the financial crisis of 2008, with projections suggesting a further compression in current account imbalances in the medium term.

Finally, Pack’s pessimism about the ability of small, poor economies of Africa and South Asia to outcompete large manufacturing groups may not fully take into account the importance of new developments in global trade, and the increasing dominance of global value chains (GVCs), which have become powerful vehicles to link small and large firms from around the world. In fact, economic history shows that the small size of an economy has really never been a binding constraint to growth and success. Small economies such as Luxemburg, Switzerland, and Singapore have been able to position themselves well on the global scene by developing competitive industries and sometimes exploiting linkages with some of their bigger and more successful neighbors. Following the flying-geese model, others such as Taiwan-China or Mauritius have created domestic firms that link up with foreign firms and are able to access supplier networks, technology, and knowledge. As discussed in chapter 5, rising wages in large, middle-income countries such as China provide even greater opportunities for today’s lower-income economies where labor-intensive activities can be developed in many industries.

(p.268) Perhaps a more credible and more immediate threat to industrial development in low-income countries that have to rely primarily on low-skilled labor is that of technology—and its corollary, the risk that many of the jobs being performed in light manufacturing industries will soon be carried out more efficiently and cheaply by machines and robots (Brynjolfsson and McAfee 2012). It is true that throughout history, breakthrough technologies with major economic impact have been accompanied by stagnant or even declining wages for some workers along with rising inequality. That was certainly the case during the Industrial Revolution. The threat is often perceived to be even more daunting today given the wide scope and rapid pace of changes that information technology brings to industrial production and managerial processes (many white-collar jobs are being automated, just like blue-collar ones).

But studies devoted to the impact of technology on low-wage workers show that fears are often exaggerated. Looking at the impact of adopting transformative new equipment in the textile industry in the United States in the nineteenth century, James Bessen (2015a) notes that most gains from this new technology took a long time to materialize and required the skills and knowledge of many people—including factory workers with little education who were considered unskilled.

Although the early mill workers had little formal schooling, they learned skills on the job, skills that were critical to keeping the strange, new, expensive machines running efficiently. Their skills were narrow compared to those of traditional craftsmen but valuable nonetheless. These skills eventually allowed factory weavers to earn far more than earlier artisan weavers; steel workers with narrow skills earned more than craft ironworkers with broad skills; typographers on the new Linotype machines earned more than the hand compositors they replaced. Moreover, employers paid these workers well at a time when unions had little power. Technical skills learned through experience allowed blue-collar workers with little education to enter the middle class.

Similar slow patterns of change were observed with the advent of steam engines, factory electrification, and petroleum refining (Bessen 2015b). The obvious policy lesson is for firms and governments to adopt the appropriate (p.269) modernization strategies to mitigate the short-term negative impacts of automation and new technologies on low-skilled workers, by providing strong incentives for learning on the job.

But while the short-run effects of technology may apply to workers with few skills and in industries that require routine and arduous efforts, it should be stressed that the relationship between job growth and productivity is not unidimensional. Beyond the first-order effects of automation where the machine replaces workers in some industries, there are also important second-order effects. Firms using new technology may save money, which flows back into the economy either through lower prices for their goods and services, higher wages for the remaining workers, or higher profits. Thanks to these three channels, the money saved in the use of new technologies eventually gets spent, creating new demand that leads other companies to hire more workers. In the longer run, the net effect of technological change is always positive for wages and for aggregate employment (Atkinson and Ezell 2012). This is also the conclusion of a comprehensive report by the Organization for Economic Cooperation and Development (1994, 33) reviewing several empirical studies on the subject: “Historically, the income-generating effects of new technologies have proved more powerful than the labor-displacing effects: technological progress has been accompanied not only by higher output and productivity, but also by higher overall employment.”

Technology and products may change very fast. But the nature of production activity will not change at the same pace unless workers have the skills to adapt. Furthermore, it is quite conceivable that even in industries that are more prone to automation, some segments of the value chain may remain labor intensive or require the kind of human touch that gives special value to the goods and services. This is true not only for textiles, garments, and footwear but also for capital-intensive industries such as the automobile industry. The new Rolls-Royce Dawn unveiled in September 2015 as the most technologically advanced luxury cabriolet ever produced displays under the company logo: “Hand Built in Goodwood, England.”9 It has been known for decades that machines and robots are excellent at assembling parts into final products, including in the automobile industry. But increasingly firms in all industries advertise the value that handicraft and direct, tedious human involvement bring to their goods and services. (p.270)

Entering Global Value Chains to Gain Credibility, Learn, and Grow

For developing countries, international trade used to generate important benefits through various preference schemes. Under the U.S. Generalized System of Preferences (GSP),10 they had access to the markets of rich economies where they could export their goods (typically commodities) with fewer constraints, including reduced tariff rates and less restrictive quotas. Although the GSP also includes rules of origin as the criteria needed to determine the national source of traded products, which sometimes reduced the effectiveness of preferences, the system generally has worked well and has often provided preferential duty-free entry for various products from designated beneficiary countries and territories. In addition to the GSP, other preference schemes, such as the European Union’s Lomé and Cotonou Agreements, the Caribbean Basin Initiative, the Andean Trade Promotion Act, and the United States’ African Growth and Opportunity Act, were adopted to facilitate participation by developing economies in the international trading system.

The benefits of trade preferences typically accrue through several channels: first, there is the transfer of rents to developing countries (tariff revenue or quota rents that would normally be received by the developed importer country are instead gained by recipient, developing countries). The preference margin is therefore transferred to producers in exporting countries. Second, there is the potential for a substantial export supply response, which can generate foreign exchange revenue and create employment in developing countries. Although agricultural and natural resource–based exports can generate rents from trade preferences, they eventually face limitations due to land availability. By contrast, manufacturing exports generally provide much greater potential: they do not face the constraints of market size or domestic endowments and can therefore be expanded virtually without running into diminishing returns to scale. Therefore trade preferences can stimulate production and exports if they are designed to allow import of complementary inputs and to facilitate connecting domestic firms operating in competitive industries to international production and trade networks.

Thanks to the considerable changes that have occurred in the organization and structure of global trade in recent decades, trade preferences (p.271) can indeed contribute to boost the economic performance of developing countries—provided that they are used to promote manufacturing and develop industries and sectors that are economically viable. As noted by Paul Collier and Anthony Venables 2007, 4),

modern sector production is [no longer] simply a matter of transforming primary factors into final output. It requires primary factors and many other complementary inputs, ranging from specialist skills and knowledge to component parts. These are frequently supplied by many different countries, with design, engineering, marketing, and component production occurring in different places—a process known as fragmentation of production. Furthermore, productivity in these different activities is not exogenously fixed. They are shaped by learning and by complementarities with other activities. These processes often give rise to increasing returns to scale, and imply that clusters are more productive than is dispersed activity.

The emergence of intrafirm trade as the primary vehicle for international trade raises new analytical questions for economists, but it mainly offers new opportunities for policy makers across the planet. Each country actually imports intermediate products—whether goods or services—to which they add value before reexporting them or using them domestically, whether for consumption or for incorporation in new productive processes. Pascal Lamy (2013), the former director-general of the World Trade Organization, often stressed the common benefits of the new trade landscape:

We may still think in terms of Ricardo’s world of trade between nations, but in reality most trade now takes place within globe-spanning multinational companies and their suppliers. The results of this “trade in tasks” are all around us. With value chains, it is no longer necessary to be competitive in producing a particular product or service; it is enough to be competitive in delivering a particular task. The growing weight of services in the business portfolios of countries and the increase in the reach of technology and transportation are fast narrowing the distances between and to markets and creating new opportunities for all countries (developed or developing) to grow through trade.

(p.272) He also observed that only two decades ago, 60 percent of world trade was between developed countries (North–North), 30 percent was between developed and developing countries (North–South), and 10 percent was South–South. By 2020 it is expected that world trade will be split equally three ways, so the relative weight of North–North trade will have been halved in just thirty years or so.

International production fragmentation, taking place in GVCs, has indeed generated new and stronger incentives for economies at different development levels to cooperate.11 By unbundling the value chains, it also generates win-win opportunities for all parties involved. Because the different stages involved in producing a particular final good are performed in different countries, specific tasks can be outsourced and undertaken separately, simultaneously, or consecutively in different countries. Whether these new production processes occur within a single multinational firm or through several production networks of supplier firms does not really matter. Advanced-economy firms can take advantage of the factor price and productivity differences and focus on the high-skilled/capital-intensive aspects of production while firms in developing countries can concentrate on the labor-intensive activities in which they have comparative advantage.

In fact, GVCs are typically organized around a dominant or lead firm located in an advanced economy and rely on a dense network of suppliers from across the globe. Participating firms may simply supply intermediate goods that are put together in the lead firm’s home country. They may also put together the final good with the role of the lead firm centered in activities such as research and innovation, product design, advertising, and distribution. GVCs are therefore potentially powerful vehicles for structural transformation in developing countries. First, these countries no longer need to build entire industries from scratch to industrialize and assert themselves as credible competitors in world markets—they can simply specialize in activities that they do best and at the lowest cost (Baldwin 2012).

Second, GVCs allow even the poorest countries with the worst business environments to be associated with high-quality manufactured goods and to learn through cooperation with partner firms in advanced economies. As noted by Neil Foster-McGregor and colleagues (2015):

(p.273) Through participation in GVCs, and the exposure to international markets and foreign competitors, the potential for technology transfer and spillover effects arises. Such effects can take a number of forms, for example by providing access to best practice management and business methods, through the use of high-quality and high-tech intermediates, through developed country intellectual property and trademarks, through lead firm knowledge and technology sharing, through skills demand and upgrading, and through learning from customers. Such effects can impact upon local firms not engaged in GVCs as well as those that are involved in GVCs, with the development of a part of a GVC in a country potentially also leading to spin-off firms and industries.

It is therefore important for developing-country policy makers to recognize that their domestic firms that are connected to GVCs can learn and access supplier and financial networks. Governments that are able to build the right institutions to foster firm participation in GVCs (mainly special economic zones and industrial parks) and adopt the appropriate policies to make them work properly can actually shift resources out of traditional agriculture and other low-productivity primary activities into industry to boost productivity.12 In doing so, they could increase the part of the working population engaged in manufacturing, an essential feature of sustainable and inclusive growth.

The advent and dominance of GVCs also has substantial macroeconomic implications. In a world economy with integrated production networks and businesses collaborating across boundaries—including firms from “competing” countries—domestic firms must purchase goods and services abroad and use them as intermediate inputs in their own production, which is then sold on world markets. Likewise, their export products are used as contributing inputs to producing final goods and services in other countries, which are also traded in international markets. Such patterns of industrial interdependency change some of the fundamentals of macroeconomic policy. The quantity, quality, and cost of imports become as important for developing and developed countries as those of exports.

Policy makers should be aware that imports have become direct contributors to economic growth and employment generation. Lamy (2013) stresses that point:

(p.274) One lesson is that to be able to export, you must know how to import. When an industry’s competitiveness relies on the cost effectiveness of the components and intermediate goods and services making up the production chain, strong performance in all segments of the value chain is essential. Indeed, there is a positive correlation between the buoyancy of a country’s exports and its integration in value chains through imports of intermediate goods. Importing competitive components where necessary enables developed-country firms to generate margins for investing in those segments where their real comparative advantages lie. Far from killing jobs, this enables Europe, the United States, and Japan to maintain industrial activities linked in particular to research and development, industrial engineering, and high value-added services. These are the activities that can and will generate the best-paid jobs.

It follows that in some instances exchange rate policies such as competitive devaluation, typically used to reduce imports and boost exports, may have adverse effects on domestic firms and burden their production costs. This would penalize their competitiveness, with potential consequences on foreign exchange, export and fiscal revenues, and economic growth.

The increasing dominance of GVCs and the opportunities they provide for developing countries are not just a matter of conjecture. There is now ample empirical evidence to document their role as new anchors of global trade and potential vehicles for rapid growth. Statisticians have to redesign national accounting systems in each country in a value chain to capture the industrial interactions between the different countries and world regions. Such tedious work necessitates constructing a large international input-output matrix, which contains all interindustry trade that precedes production and consumption of a final good or service. It also requires properly harmonizing each of the trading partners’ national accounts and detailed analyses of how the traded goods and services are used, that is, either for consumption or investment purposes or for further use in a new production process.

The empirical analyses that must be carried out to measure the various trade flows into and out of GVCs are challenging because they require detailed data across countries and industries. Any country involved in GVCs produces goods and services using both imported inputs as intermediates (imported value added) and its own domestic value added. A (p.275) straightforward approach often used by trade economists starts with the observation that Country A’s exports (or final demand) can be divided into two components: one capturing domestically produced value added and the other capturing the imported value added from other countries that is incorporated into Country A’s exports (final demand). A complete analysis of GVCs should also account for the fact that a country’s exports need not constitute final goods only and can be used as inputs into other countries’ production (and exports).

To measure Country A’s participation in GVCs, one should therefore think of at least two main components:

  • ➔ The share of foreign value added (FVA) used in its own exports and calculated as a percentage of Country A’s total exports (this is the indicator of how much Country A is involved in downstream production in GVCs because it measures its use of foreign inputs in the processing of export products)

  • ➔ The share of Country A’s domestic value added (DVX) that is absorbed as intermediate inputs into the value added exported by all other countries in the world, calculated as a percentage of Country A’s total exports (this is the indicator of how much Country A is involved in upstream production in GVCs because it measures its inputs—value added—into other countries’ production and export of goods and services)

Thus GVC participation combines the FVA and DVX measures by summing up the foreign value added used in a country’s own exports and the value added supplied to other countries’ exports and taking the sum as a ratio to gross exports (figure 7.6).13

Using that calculation framework, Neil Foster-McGregor and colleagues (2015) shed light on participation in GVCs by world region for 1995, 2000, 2005, and 2010.14 Their empirical analyses highlight several interesting facts. First, they confirm that GVC participation has been increasing over time in most regions of the world, with the share of exports that are part of a multistage process increasing from around 41 percent in 1995 to just almost 50 percent in 2010. The only exceptions are Central America and the Caribbean, where GVC participation has tended to decline somewhat (figure 7.7). (p.276)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.6. Measuring GVC participation as FVA + DVX.

Source: Authors.

Advanced countries—the EU member states in particular—are heavily integrated in GVCs, which is not surprising. The intense GVC participation of East and Southeast Asia and to a lesser extent West Asia is also in line with expectations. Perhaps the most encouraging evidence that even low-income countries can perform well in global trade is African economies’ (p.277)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.7. Participation in global value chains by region, 1995–2010.

Source: Foster-McGregor, Kaulich, and Stehrer (2015).

rising participation in GVCs. In fact, Africa has some of the highest rates of GVC participation, matching the levels found in Asia. In 1995 Africa’s GVC participation was about the average for all regions (about 41 percent), though its involvement is mainly in upstream production (that is, providing inputs to other countries) rather than downstream production (processing for exports). The growth rate of GVC participation between 1995 and 2010 in Africa has also been similar to that for all countries, with GVC participation increasing by 19.9 percent for Africa and 20.5 percent for all countries over the period 1995–2010.

Perhaps the most encouraging news for developing countries is the story derived from the empirical analyses of GVCs by sectoral category (figure 7.8). High-tech sectors represent, on average, 58 percent of GVCs’ participation in all countries, with the numbers being larger for developed countries (62 percent) as well as Central America (69 percent). Asia’s contribution of high-tech sectors to GVCs is around the average, though contributions in South and West Asia lag behind those for East and Southeast Asia. Most striking is Africa’s sectoral contribution, which seems quite balanced, with the primary sector’s share at 26 percent, two services categories (high-and low-tech) at around 20 percent each, and the two manufacturing (p.278)

Reaping the Dividends of Globalization: A Winning Road Map

Figure 7.8. GVC participation by region and by sector category, 2010.

Source: Foster-McGregor, Kaulich, and Stehrer (2015).

categories (high-and low-tech) at around 17 percent each. But Africa’s contribution of manufacturing goods and services to GVCs (33 percent) lags behind that of all other regions, with the average for all countries being 66 percent—West Asia and the Caribbean are the only other regions with shares below 40 percent.

Skeptics of this overall positive story could still observe that even involving its firms in GVCs does not guarantee that a poor economy will eventually grow at sustained high rates. They would point to the experiences of countries such as Bangladesh that have done quite well in developing labor-intensive industries (textiles and garment) but seem trapped in low value-added segments of the GVC, where there is little possibility for innovation or technology transfer. It is true that the world economy has witnessed many growth-acceleration episodes that did not lead to convergence between low-and high-income countries. Ricardo Hausmann, Lant Pritchett, and Dani Rodrik (2005) identify eighty such growth-acceleration episodes since the 1950s, which they find to be “highly unpredictable.” They conclude pessimistically that the vast majority of these accelerations are unrelated to standard determinants in the growth literature and that most instances of economic reform do not produce growth acceleration.

(p.279) It is indeed possible that an economic development strategy can initially succeed—even foster the emergence of competitive firms, generate employment, and contribute to a country’s growth—but eventually fail because it was not designed to sustain continuous industrial and technological upgrading. This raises the broader problem of the validity of the economic development strategy in question, which must evolve to reflect the economy’s changing endowment structure and constantly adjust to its institutional and factor requirements. Countries that have succeeded in boosting production and in developing trade as a steady and reliable source of economic growth have often built clusters of firms for their most competitive industries. They have generally done so by ensuring that the government allows domestic and foreign investors to establish strong partnerships and nurture firms that can succeed in international markets.

How to Boost Production and Trade? Beyond a Random Theory of Clusters

The wisdom inherent in the well-known saying “one hand cannot tie a bundle” is generally offered as justification for unity and collaboration within social groups. Surprisingly, it appears to be valid also in the realm of business, even when capitalist principles and war-sounding mottos such as “survival of the fittest” are prevalent. For firms competing not only to gain bigger market shares but also to push one another out of business, the benefits of collaborating with the enemy often outweigh the costs. This was made obvious by Paul Krugman’s (1991) seminal work on the geography of trade, which suggested that clustering—the phenomenon of firms in the same industry (or in neighboring industries) gathering in the same geographic location—was a very common pattern of economic activity.15 With the growth of global trade and the changes that have occurred in recent decades in the ways in which goods and services are produced and exchanged among economies, the importance of clusters is now widely accepted: they offer opportunities for firms to benefit from increasing returns to scale and external economies of scale.16

Marshall (1890) offered a first compelling theoretical rationale explaining and justifying that phenomenon. His insights have sparked a rich economic literature looking at clusters mainly through the prism of increasing returns to scale, or economies of scale, which convert increased levels of (p.280) output into downward-sloping average costs curves. These insights were derived from Marshall’s industrial district analysis, which showed that economies of scale may even be “external,” emerging from outside the firm because of asset sharing, such as the provision of specific goods and services by specialized suppliers or the emergence of a localized labor pool due to the concentration of production. Also, the very proximity of firms working on similar products or competing closely against one another eventually yields collective benefits in new research, managerial, and organizational practices (Griliches 1979). Such learning dynamics and spillovers increase the stock of knowledge available for individual firms.

By building factories or offices next to those of their competitors, firms can benefit from some critical aspects of their production, including tapping into the neighborhood’s pool of expertise and skilled workers, accessing a larger network of component suppliers, and learning from one another (even simply through the informal channels of gossiping or closely monitoring best business practices). Moreover, as clusters attract more firms, a network of specialist input suppliers develops and markets for intermediate goods expand, opening up new activities, and transport and infrastructure support tends to improve along with the local labor market. The combined spillover effects of firms setting up next to one another generally outweigh the cost of competition among them. It has been shown empirically that productivity is higher in areas of dense economic activity (Duranton and Puga 2005), and analyses focusing on cities have noted that, over a wide range of city sizes, each doubling of size raises productivity by 3 to 8 percent (Rosenthal and Strange 2005). Most of these insights have been known intuitively for centuries. In eighteenth-century England industrial districts (clusters) were common. Adam Smith described them in his works. The town of Staffordshire was the location of many potteries, and that region is still known today as “The Potteries.” The town of Nottingham was home to many lace-makers, and so on. Since French economist François Perroux suggested the notion of growth poles in 1949, countries as different as Austria, Belgium, Bolivia, France, Great Britain, Italy, Peru, Spain, the United States, and Venezuela have adopted and attempted various interpretations of it (Perroux 1955; Darwent 1969; Christofakis and Papadaskalopoulos 2011).

Today’s best-known clusters include Silicon Valley in the San Francisco Bay Area (California), where many firms have gathered to produce some of the most innovative, life-changing technologies in use around the world, (p.281) and Hollywood (California) and Bollywood (Mumbai) where the film industry cohabits, competes, and collaborates. It may seem counterintuitive that highly competitive firms would be willing and even eager to locate their headquarters and operational units close to one another and thus expose their business strategies and trade secrets to spying risks. It generally turns out that the potential costs of such risks are outweighed by their benefits. By working close to one another, firms in almost all industries realize that they have unparalleled access to financiers, the best experts in the business, and a culture of innovation and risk taking. It has been observed that “new information technology and Internet firms continue to gather there in spite of the high prices of local property and the danger of earthquakes. Ironically, they find that much of the most valuable information that they obtain comes not electronically but from face-to-face meetings” (Hindle 2008, 36).

The success of clusters has led to a broad consensus among economists on the importance of policies that facilitate their development. The implications of industrial policy, the welfare gains from trade, and all the Marshallian externalities for the patterns of international trade have been studied extensively, most notably by Krugman (1991, 1995, 2008), Paul and Siegel (1999), Rodriguez-Clare (2005), Aghion (2009), and Harrison and Rodriguez-Clare (2010). The topic is particularly important in an increasingly globalized world economy. Recent work based on quantitative analysis that looks at whether Marshallian externalities lead to additional gains from trade indicates that they do lead to gains, and that the externalities increase overall gains from trade by around 50 percent (Lyn and Rodriguez-Clare 2011).

In recent decades analyses of the Marshallian externalities have been enriched by the experiences of developing economies that found clusters to be the optimal tool for circumventing some of the structural issues they face. The infrastructure and human capital deficits, the rigid labor laws and regulations, and the weak governance that afflicts low-income economies and reflects their low capacity and limited resources can be effectively addressed within the circumscribed and more easily managed cluster (Ayele et al. 2010).

But there are still unanswered questions. Is the clustering of firms in a particular location a phenomenon that occurs randomly as one of the many enigmas of economic development? Or are there government policies that can be implemented to establish clusters and then foster dynamics that (p.282) create viable, competitive firms in which workers (unskilled or educated) are given incentives and opportunities to acquire the skills that help them prepare for the constantly changing demands of the global economy? The answers to these questions concerning the causes and optimal conditions for the emergence of clusters have remained a mystery, even for economic theorists who have focused their efforts on the topic.

Theories of agglomeration and clusters proposed by various authors since Adam Smith and Alfred Marshall generally have suggested that successful clusters emerge randomly, according to some spontaneous generation process. As a result, such theories have had a limited role in government policies to create clusters from scratch. Michael Porter (1998) observed that in today’s world of global competition, rapid transport, and high-speed telecommunications, location should no longer be a source of competitive advantage. Yet he pointed to puzzling cases such as the flower-growing industry in the Netherlands, which could not have been an obvious first choice for anyone starting a flower-growing business today except that the business is already there. The cluster granted new entrants with strong competitive advantage that consisted of the sophisticated Dutch flower auctions, the flower-growers’ associations, and the country’s advanced research centers. Porter did not explain the genesis of clusters, but his analysis was consistent with the random theory that they emerge in unpredictable ways.

For several decades the emergence of clusters was explained in the economic literature simply as illustrating “the economics of QWERTY.” It was based primarily on the work of Paul David (1985), who chronicled the QWERTY keyboard’s rise to dominance. Although it was not the most efficient layout in terms of finger movement, it forced typists to work slowly and mitigated the risk of mistakes caused by the jamming keys of the early machines. With innovation and technical progress, the jamming problem was subsequently corrected, but a path had already been set, and manufacturers and typists were accustomed to the bizarre keyboard layout. In sum, a historical accident had set the stage for a long-lasting technical standard and the development of keyboards.

The theoretical lesson derived from this story was straightforward:

A path-dependent sequence of economic changes is one in which important influences upon the eventual outcome can be exerted by temporally remote events, including happenings dominated by (p.283) chance elements rather than systematic forces. Stochastic processes like that do not converge automatically to a fixed-point distribution of outcomes, and are called non-ergodic. In such circumstances “historical accidents” can neither be ignored, nor neatly quarantined for the purpose of economic analysis; the dynamic process itself takes on an essentially historical character.

(David 1985, 332)

David tried not to draw definitive conclusions about economic phenomena from his investigation of the origins of the QWERTY keyboards. He wrote: “Standing alone, my story will be simply illustrative and does not establish how much of the world works this way. That is an open empirical issue and I would be presumptuous to claim to have settled it, or to instruct you in what to do about it” (332). However, his main conclusion that “it is sometimes not possible to uncover the logic (or illogic) of the world around us except by understanding how it got that way” led some economic theorists to assert that clusters are always random phenomena.

Arguing that the QWERTY keyboard was “not just a cute piece of trivia” but “a symbol for a new view about how the economy works” and “a parable that opens our eyes to a whole different way of thinking about economics,” Krugman (1994, 223) hailed it as evidence that neither the market nor the government can manufacture good economic outcomes. He wrote: “That different way of thinking rejects the idea that markets invariably lead the economy to a unique best solution; instead, it asserts that the outcome of market competition often depends crucially on historical accident. … And this conclusion is fraught with political implications, because a sophisticated government may try to make sure that the accidents of history run the way it wants.” He then went on to compare the randomness of QWERTY emergence and dominance to that of the film industry in Hollywood or the concentration of banking and financial institutions in New York.

Krugman’s use of the QWERTY story may just have been a rhetorical device to make the broader and theoretically valid points about the importance of clusters once they have been created and are functioning well. But his skepticism about the effectiveness of government policy agendas designed to facilitate the emergence of clusters implied that he adhered to the random theories, which have dominated the literature since Smith and Marshall. While it is true that historic clusters such as Silicon Valley or (p.284) Hollywood may have been random phenomena, the notion that governments should refrain from getting involved in the emergence of clusters has been proved to be inaccurate. As shown by Xiaobo Zhang, Jin Yang, and Thomas Reardon (2015), several developing countries have actually been successful in using smart government policies to deliberately create successful clusters in places where nothing existed before.

Alfred Marshall observed that many of Britain’s successful industries in the late nineteenth century were concentrated in specific industrial districts: cotton around Manchester, ironworking in Birmingham, cutlery in Sheffield, and so on. Subsequent theories of agglomeration and clustering have explained well the underlying reasons for success and highlighted their perceived unpredictability. But the more recent successes of some developing countries that defied geographic randomness and that engineered clusters in specific locations provide useful lessons for policy making.

“Look down at the shirt you’re wearing. Chances are the buttons came from Qiaotou. The small Chinese town, with about 200 factories and 20,000 migrant workers, produces 60 percent of the world’s supply” (Lim 2006).17 It all started in the early 1980s when Qiaotou in the remote Zhejiang Province was a poor small town with unproductive paddy fields and dusty roads, no infrastructure, and no capital. Two village brothers started buying buttons from Hubei Province and selling them to local garment manufacturers. They quickly became successful and inspired a large number of other merchants. Everything changed in the 1990s when entrepreneurs from Qiaotou visited Italy to convince button producers that they could import equipment and start low-end manufacturing in China using Italian designs and material. “Faced with rising costs, exhaustion of raw materials, and competition from emerging economies, Italian button producers saw outsourcing of the low end activities to Chinese manufacturers as the right solution to remain in business” (Rasiah, Kong, and Vinanchiarachi 2012, 32). They saw win-win opportunities in relocating button production to a place where production costs would make the industry more competitive.

In just two decades Qiaotou converted itself into a manufacturing powerhouse, with hundreds of family-run new firms so competitive that they put out of business some of the established European firms that had been producing buttons for centuries. The low-investment, labor-intensive industry was perfectly suited for Qiaotou and consistent with the country’s comparative advantage at that time. The Chinese government worked (p.285) closely with local communities and business associations to complement the development of the button industry. In recent years, despite the normal challenges associated with rapid industrial development, Qiaotou has also become the world capital for other industries, such as zipper manufacturing. In other industries in China, local governments were even more proactive with creating clusters; they provided targeted infrastructures and specific incentives for the private sector and facilitated the creation of cooperatives so that small local entrepreneurs could have a voice and have their interests preserved (Zhang et al. 2015; Dinh et al. 2013).

More recently, the magic of government-engineered or government-supported clustering has been observed well beyond the traditional Marshallian externalities. In trying to determine why agricultural production in China has increased steadily despite small landholdings, a high degree of land fragmentation, and rising labor costs, Zhang and colleagues (2015) find that the main explanation may lie in new clustering strategies. Confronted with unsustainable labor costs, farmers organized themselves not just to act as single small-producer households but to cooperate (while still competing against one another) to share the main element of their cost structure. They outsourced some power-intensive stages of production, such as harvesting, to specialized mechanization service providers, which are often clustered in a few counties and travel throughout the country to harvest crops at very competitive prices. Through such an arrangement, smallholder farmers could stay viable in agricultural production. Again, local and central governments intervened proactively to facilitate the development of such activities with agricultural clusters.18

Stories of successful government-engineered clusters have been recorded from well beyond China. The Penang electrical-electronic cluster, one of the largest of several major clusters of this type in East Asia, also resulted not from random theory but from smart, activist government policy. It is another successful case of “artificial agglomeration.” In 1972 Malaysian authorities decided to create an export-processing zone (EPZ) where the country could develop its electric and electronic industry. Although it was a case of learning by doing, the project took shape through several industrial master plans, each including a series of policies to attract foreign direct investment. The second of such plans (1996) was specifically designed to strengthen linkages and complementarities between foreign investors and local firms.

(p.286) As a result, the electrical-electronic industry has been a major source of manufacturing value added, employment, and exports in Malaysia for several decades.

In Penang, excellent basic infrastructure—good transport services, power supply, water supply, and telecommunications—was combined with superior supply of social services, such as public health facilities and schools to make the region attractive to skilled workers and managers. Institutional reforms were also introduced to improve the performance of the security and customs within the EPZ. Drawn by these investments and financial incentives, Japanese, European, and giant North American firms such as Hitachi, Sony, Siemens, Advanced Micro Devices, Hewlett Packard, Intel, National Semiconductor, and Seagate moved to Penang. The agglomeration of these flagship firms helped stimulate the development of local supplier firms.”

(UNIDO 2009, 34)

Similar successes in creating new clusters were observed in countries as different as Mauritius (Lall and Wignaraja 1998), Costa Rica (Ciravegna 2011), and the United Arab Emirates (Ketels 2009).

The main lesson from these experiments in “artificial agglomeration” and from the fragmentation of global value chains is that the old theories of clusters as random phenomena may be obsolete. It is actually possible for a developing economy to create a successful strategy that facilitates its engagement with the international trading system. But few policy makers in these developing countries have managed to take full advantage of the new opportunities offered by changes in trade patterns and the evolving global economy. One of the main reasons for this lack of initiative is that the traditional policy advice given to developing countries by most mainstream economists and development institutions has been to stick to minimalist government intervention and adopt “neutral,” “horizontal” economic strategies. This includes implementing prudent macroeconomic policies and improving the business environment through broad microeconomic and institutional reforms without giving special consideration to particular industries. The value of such advice is questionable. Evaluation studies often show that such generic prescriptions have rarely yielded sustained and inclusive growth. In fact, many of the success stories in economic (p.287) development (most notably China, Brazil, and Vietnam in recent years) are countries where policy makers did not follow that advice, often simply because it would have required the abrupt removal of all the distortions that stifled the economy in the first place, with the high risk of creating social and political disruptions.

Economic development requires uninterrupted and coordinated upgrading of physical and human capital and institutions. For poor economies with limited financial resources and administrative capacity, it is therefore essential that economic policies be geared toward the changing patterns of industrial structure and technology diffusion and choosing production bundles and modernization and innovation strategies that are consistent with their comparative advantage and development level (Lin 2012a, b). At lower incomes, the main economic policy challenge is therefore to break into global industrial markets and find their niches or to organize their economies to take advantage of the opportunities being vacated by middle-income countries that are forced out of these niches because of rising wages, rising productivity, and the need for industrial upgrading. Despite the recent declines in global trade, the opportunities to attract foreign capital in potentially competitive industries, to create employment, and to develop production and exports have never been so numerous for developing countries. Clusters in well-targeted industries are viable if not essential policy tools to circumvent the well-known constraints on growth in developing economies (weak governance, infrastructure bottlenecks, and insufficient human capital).

The real questions then become: Why have only a few developing country’s governments been able to successfully facilitate the formation of such clusters? Why have the main instruments used to build clusters (SEZs) often failed to yield positive economic results? What are the guiding principles for success?

Why Special Economic Zones Often Fail to Generate Viable Clusters

Certain ideas are unlucky. Despite their potentially great value, they are too often associated with past failed experiments and doubt. Or they are casually discussed in public policy debates mainly with skepticism. Even when they are put to good use and lead to success, these cases are quickly (p.288) dismissed as anomalies and exceptions that confirm the rule. Consider SEZs (broadly defined as administratively separated areas where investors may run their business activity based on specific preferential conditions such as tax and tariff incentives, streamlined customs procedures, and less regulation), often seen as the best modern institutions to generate industrial clusters.

The well-known rationale for SEZs is to provide special policy incentives and infrastructure in a circumscribed geographic location to firms that can attract foreign direct investment, create employment, develop and diversify exports (even when economy-wide business environment problems and protective barriers are not yet resolved) and foreign exchange earnings, and serve as “experimental laboratories” for new pricing, labor, financial, or labor policies. The ultimate expectation is that the knowledge spillovers of these experiments eventually translate into private-sector development, sustained growth, productivity increases, and other financial and economic benefits for the entire economy.

The most popular form of SEZs are export-processing zones, which typically operate under a few basic principles: allowing investors to import and export free of duties and exchange controls; streamlining customs and administrative controls and procedures; facilitating licensing and other regulatory processes; and freeing firms from obligations to pay corporate taxes, value-added taxes, or other local taxes (Farole 2011). To ensure effective monitoring of their activities, export-processing zones are often fenced-in estates with customs controls at the entrances, and sales are typically restricted to export markets. Export-oriented SEZs are generally intended to “convey ‘free trade status’ to export manufacturers, enabling them to compete in global markets and counterbalance the anti-export bias of trade policies” (FIAS 2008, 12).

SEZs and export-processing zones have been around for a while. It is believed that they were initiated in ancient Phoenicia as basic free-trade zones. In recent times their track record is stellar: they were a central pillar of China’s growth strategy, which allowed the most populous nation on earth to lift out of poverty about 700 million people in a rather short period of time (1979–2015). An SEZ was established in 1937 in New York. Another one was created in 1942 in Puerto Rico. Iceland and Taiwan-China also established SEZs in 1960. In 1959 one was set up at Shannon Airport in Ireland. SEZs were used effectively by some latecomers such as Korea (p.289) and Taiwan-China to build clusters, emulate the economic development strategies of leader countries, and even catch up with them in the race to economic prosperity. Some developing countries in other parts of the world have been quite successful in establishing various types of well-functioning SEZs to boost their export and growth strategies, which allowed them to generate much-needed employment and also spark their industrial upgrading process. Well-known cases are those of Costa Rica, Honduras, El Salvador, Bangladesh, Vietnam, and Mauritius.

The best SEZ success story—at least in recent memory—occurred when Deng Xiaoping’s China embraced them in the 1980s, using them, for instance, to convert a poor, sleepy fishing village such as Shenzhen with no natural resources and no infrastructure into one of the world’s premier export powerhouses, with a 10.7 million population and a $24,000 per capita GDP in 2014.19 In fact, various types of SEZs and industrial clusters initiated by the government have been the main laboratories and engines of China’s remarkable economic development.20

Such success stories cannot go unnoticed. Developing countries in various parts of the world have used SEZs as the main instrument to attract FDI and transform their economies. Thomas Farole (2011) notes that SEZs have allowed the Dominican Republic to create more than 100,000 manufacturing jobs and shift dramatically away from reliance on agriculture. Qatar has launched SEZs to diversify its economy from a hydrocarbon-led economy—which made it one of the richest in the world but is subject to volatile global oil prices—to a knowledge-based economy. Ethiopia has followed the same path, mainly to address its logistics woes (infrastructure) and to experiment with new models of governance that give more flexibility to investors and allow them to recruit workers for light manufacturing industries in which the country has comparative advantage. Advanced industrialized economies are also using various forms of SEZs at the federal, state, and local levels to attract investors, stimulate economic activity, and create employment.

Yet there has been widespread skepticism about the potential economic value of SEZs in developing economies, and understandably so. Although the general recipe of such zones has long been understood by economists and policy makers, few countries have actually managed to design and make good use of policy frameworks and instruments to achieve their development goals. In fact, most countries that have tried to replicate this strategy (p.290) have not gained the expected benefits. Globally there are now about 4,500 such special zones, with few actually delivering the expected results. In most countries, the benefit-cost ratio for setting up and running SEZs has been disappointing. Personal income tax on employment, permit fees and services charges, sale and rental fees on public land to developers, import duties and taxes on products from the zones sold to the domestic customs territories, concession fees for facilities such as ports or power plants, and corporate income tax (when assessed) totaled only negligible amounts. In the meantime, import duties and charges lost from the smuggling opportunities created by SEZs, tax revenue forgone from firms relocating from the domestic customs territory into the zones, and public investment for (often untargeted) infrastructure and recurrent expenditures (mainly the wage bill of public-sector workers needed to run and regulate the zones) often represented substantial costs to governments. Even in China, some of these initiatives failed to attract competitive industries and generate employment, requiring the authorities to reengineer them (Chenggang 2011, Zhang 2012). Over the decades, many SEZs have indeed failed and were simply abandoned by policy makers who did not know how to make them function properly. As a result, SEZs are often dismissed in mainstream economic circles as mere illusions, or worse, a “Chinese” affair—with the usual and not-so-subliminal tone of political and ideological disdain.

Nothing could be further from the truth. First, as the historical record clearly shows that China did not invent SEZs. In fact, when Deng Xiaoping became China’s leader in 1978, he actively sought new ideas and useful policies wherever he could. Brushing aside the ideological battles of the past, he famously commented, “Who cares if a cat is black or white, as long as it catches the mice.” In 1980 a younger reformist, engineer Jiang Zemin (president of China, 1993–2002), and a group of other Chinese officials were sent around the world to study how SEZs worked. When they reached Ireland, they experienced a revelation:

The delegation was introduced to the world’s first duty free shop, an airport-focused infrastructure, and a special zone of low tax and free trade. They welcomed the pragmatic approach of Shannon Development. Unlike all the other countries [Jiang] Zemin visited—most of which had a stake in the cold war—Ireland was not touting an ideological agenda. The officials in Shannon were focused on jobs (p.291) creation, practical education and transfer of skills, and facilitation of each new company’s needs. This was a cat that could catch mice. After touring various economic zones around the world, the Shannon model and that of Singapore were the only two that the Chinese decided to follow.

(Quigley 2012).

More important, SEZs have become the most effective channel for attracting FDI and for building the kinds of clusters or industrial districts that allow large or small economies to take advantage of the new patterns of global trade. If fragmentation (as highlighted above) is indeed the “new normal” in global exchanges for the foreseeable future, then even countries with poor infrastructure, limited human capital, or weak governance can find pragmatic solutions to position their economies as credible business environments for global supply chains. Successful SEZs are not only zones of excellence where such problems can be addressed but also essential places where business linkages can be built between domestic firms (small and large) in various sectors and international firms.

Economists’ and policy makers’ reasons for why most SEZs fail range widely—each makes sense individually but they are overwhelming and confusing when lumped together. They include generic opposition to the very idea of SEZs, that is, of “government bureaucrats in business”—a view primarily based on ideological grounds (World Bank 1995): the suggestion that they are “political priorities” that often lead to creating so-called white elephants (Economist 2015); problems of poor governance; lack of institutional framework and political commitment; weak implementation capacity; and improper monitoring and evaluation mechanisms (Zeng 2012).

These explanations for the high failure rates of SEZs tend to merge the causes of the problem with the symptoms or consequences. Although many types of mistakes can explain the inability of policymakers in developing countries to replicate the success of SEZs recorded in China and in a small number of other countries, the fundamental problem has often been one of strategic selection. The first-order condition for building a viable SEZ is to choose the “right” industries—not necessarily the most “modern” or the most attractive ones. The industries with economic viability potential are those that reflect the economy’s comparative advantage, not the political ambitions of the leaders, so that they can develop into industries with competitive advantages quickly. As noted by Farole (2011): “Increasingly (p.292) it is not the existence of a special-economic-zone regime, a compelling master plan, or even a fully built-out infrastructure that will make the difference in attracting investment, creating jobs, and generating spillovers to the local economy. Rather, it is the relevance of the special-economic-zone programmes in the specific context in which they are introduced, and the effectiveness with which they are designed, implemented, and managed on an ongoing basis that will determine success or failure.”

In retrospect, the main reason for the weak performance of SEZs appears to be the belief that they should emerge randomly (just like QWERTY keyboards). Most SEZs do not live up to the expectations of policy makers because of poor targeting of appropriate industries or their generic nature, which makes them too broad to attract enough competitive firms. Many developing-country policy makers who have attempted to create SEZs did not properly reflect on the conditions needed for success in an increasingly globalized world and on how best to exploit the comparative advantage of their economies. Poor countries typically face two broad types of constraints that impede private-sector development: high factor costs (for skilled labor and capital) and high transaction costs (mainly due to poor infrastructure, unfriendly business environment, and weak administrative capacity), which are often compounded by political capture and rent seeking.

Thanks to globalization and free movement of labor and capital, high factor costs have generally declined in recent decades, even in remote places. In 2015 a country like Bolivia that needs highly trained engineers could attract large numbers of them from Mexico or Spain at costs lower than in 1970. Likewise, Sri Lanka can attract private capital from Qatar or Malaysia (countries with well-endowed sovereign wealth funds) much more easily today than three decades ago. But the best way for most low-income countries to circumvent the high factor cost constraint and launch a development strategy with the highest likelihood of success is to formulate economic development strategies that aim at using their only competitive factor costs (low-skilled labor and land). Factor costs can indeed be lowered if economic development strategies are consistent with a country’s comparative advantage and the abundant factors (unskilled labor, land, or natural resources) are used extensively. The first constraint is therefore removed when the industries selected for and attracted into SEZs are primarily those making good use of low-skill labor.

(p.293) To circumvent the second constraint, developing countries can build SEZs as zones of excellence, where the pervasive infrastructure problems, unfriendly business environment, and poor governance and administrative capacity can be addressed effectively. It is indeed much easier even for a government with very limited financial resources and administrative capacity to deliver first-rate infrastructure and to combat red tape and corruption within the geographic boundaries of an SEZ than across the entire country.

But for this second set of constraints to be removed and for SEZs to function effectively, the firms operating in these zones of excellence must be economically viable in the first place, and the conditions and the policy incentives must be appropriate for them to cooperate and compete. Therefore SEZs are best conceived initially as industrial parks (“specialized” SEZs) in which the government only has to provide the same kind of infrastructure to all firms. Generic SEZs that attract firms from different industries need different kinds of infrastructure, requiring extensive government financing and making the sustainability of the SEZs much more challenging.

Because the industries attracted to the SEZs have often defied the country’s comparative advantage, they are not viable without a strong set of protection policies. In most instances, policy makers either have identified the industries they favor for personal and political reasons or have not actively attempted to identify the industries that may be most suited to their country’s endowment structure (e.g., labor-intensive industries). They have assumed that foreign firms willing to join an SEZ or EPZ would create employment, which would be better than nothing. One consequence of the absence of identification strategies has been the random emergence of small single firms from very different types of industries. But given the limitations of state budgets and weaknesses of public investment programs, few governments can provide them the industry-specific infrastructure support they need.

The choice of the location for SEZs has not always been optimal. Although some zones are built in port cities that are already growth poles or near transport hubs, others are created as isolated geographic zones or in remote areas, not on the basis of an economic rationale but as a way of appeasing political constituencies. This has resulted in increased production and transaction costs for the few firms willing to build factories there. Reducing transaction costs has not been part of the strategic focus. Because of the randomness in industry selection and the limited government financial (p.294) resources, even basic utilities and services are sometimes not made available in many of these zones. Governments have not proactively played their indispensable facilitating role. They have not provided basic industry-specific infrastructure and often wait (in vain) for qualifying firms to finance investment in electricity, water, or telecommunication within the zone. They have not coordinated the design and implementation of the investment needed and used collectively by firms in their industries (storage facilities, for example).

Another major problem has been the limited volume of business transactions generated by SEZs. The likelihood that a small number of firms specializing in many different industries within the same SEZ can generate the critical mass of business transactions necessary for Marshallian externalities to materialize and make firms in that location credible partners in global markets is generally low. In today’s global economy, the firms that are able to negotiate the lowest prices for their inputs and intermediate goods and services from their suppliers are typically those that are part of a network—by buying in large quantities, they are better positioned to obtain the best unit costs. SEZs that host many different industries are therefore at a disadvantage when placing single-firm orders for inputs and intermediary goods and services. In other words, even an SEZ that hosts industries that are economically viable (and consistent with the economy’s comparative advantage) can still fail if it is trying to develop too many different activities and sectors at the same time and therefore not attracting enough market power for its firms—such SEZs are in fact spread too thin. In sum, the removal of the second constraint—high transaction costs—necessitates the development of SEZs that create clusters of large numbers of firms in industries where economies of scale, intraindustry knowledge spillovers, and other agglomeration effects can be realized. Governments in many developing countries have simply created generic SEZs with broad fiscal incentives across industries and firms and have failed to facilitate the process of industrial agglomeration, which requires the provision of industry-specific infrastructures and managing the coordination and externalities issues that always arise in collective-action situations (Lin 2012a, b).

Another first-order, strategic mistake that often leads to the failure of many SEZs across the developing world is the inability of their promoters (governments or private-sector actors) to establish effective linkages between these zones of excellence and the rest of the domestic economy. They (p.295) do not consider the political economy issues that are always at the heart of the development process. By its very nature, a gradual economic development strategy based on strategic selectivity and geographical targeting of the most promising growth potential necessarily creates groups of winners and losers, at least in the short term. It is essential to deliver quick wins (time and results) and to establish backward and forward linkages that mitigate the risks of social tensions.21 In some countries the local business community perceives SEZs suspiciously as geographic enclaves and closed “special clubs” where opaque business practices take place, often involving foreign firms and a few rich business leaders who are well connected to the ruling political class. In India, for instance, a country where federal and state laws and regulations often overlap, it has been reported that subsidies and tax incentives have been abused in SEZs and that land acquisition for zones has often led to protests. According to Arpita Mukherjee:

Land taken under the SEZ policy is sometimes misused. States have their own SEZ act, which sometimes differs from the central act. Multi-layered government and multiple policies create scope for corruption. Too many SEZs have been awarded and government continues to award SEZs in spite of the fact that existing ones are not performing. Corruption can be addressed … through more stringent evaluation of feasibility of project proposals and online application and approval systems. Also, project evaluations should be independent of the government.

(Mukherjee, quoted by Gray 2013)

It should be noted that in almost all poor countries, corruption is likely to be a pervasive problem, not only in SEZs but also across the entire economic spectrum. It is a stylized fact that has been observed across time in all regions of the world (Lin and Monga 2012). The reason is the high costs of running a well-staffed, well-equipped, and well-functioning national judicial system are often beyond what the public sector in a low-income country can afford. The problem is compounded in some countries where corruption is embedded in societal, economic, and power relations and virtually all state institutions, including the judicial system, are caught in the low-equilibrium dynamics of what Richard Joseph (2014) called “prebendal politics.” If that is the case, and if virtually all governments in the world—including those in high-income and democratic countries—must (p.296) constantly combat corruption, then the issue at hand has little to do with the existence of SEZs but rather involves the policies adopted to mitigate the risks. The challenge then is to understand which policy circumstances provide the best incentives for successful anticorruption mechanisms and for good governance in general.

Countries that successfully limit the prevalence of corruption in SEZs are generally those that lift trade restrictions, price controls, and multiple exchange rates, which A. O. Krueger (1974) has identified as some of the main causes of the problem. However, because these well-functioning SEZs have been designed to support well-targeted industries with competitive potential, there is no need for governments to provide protection or subsidies. Thus the risks of state capture and rent seeking are minimal. In that context, the sustainability of SEZs is determined by comparative advantage and the economic viability of the firms they host. Government intervention and public policies are indeed in place to support these SEZ firms, but they are carefully targeted incentives (of limited amount and time) and are allocated in a transparent manner to compensate for the externality generated by pioneer firms. The only beneficiaries are firms that will be viable in open, competitive markets. The investment and survival of such SEZ firms do not depend on protection, large budgetary subsidies, or direct resource allocations through measures such as monopoly rent, high tariffs, quota restrictions, or subsidized credits. In the absence of large rents embedded in generic SEZs that try to attract all kinds of firms from all industries, there will not be distortions that easily become the targets of political capture. In sum, the likelihood of the pervasive governance problems observed in many SEZs would be much reduced by government facilitating the development of new industries that are consistent with the country’s changing comparative advantage determined by the change in its endowment structure.

SEZs also fail because of poor institutional organization and ineffective management. But while such reasons are certainly valid, they are less important than the first-order problems discussed earlier. And often these second-stage issues are consequences of the strategic mistakes committed in launching SEZs for industries that defy comparative advantage. In some instances, government policies to support the newly created SEZs were either insufficient or inappropriate. Other failed experiments involve SEZs that were exclusively developed, regulated, and operated by governments or public entities. Beyond the obvious issues of expertise and capacity, their (p.297) institutional arrangements often led to conflict-of-interest situations, with regulatory agencies also engaged in zone development activity, especially when public zones compete with private firms outside the zone.

Investment climate surveys also indicate that SEZ managers in many countries did not realize that successful integration into the world economy increasingly requires behind-border measures that fall under the heading of trade facilitation. Too often they did not alleviate the burden of red tape, nor did they provide services such as customs and port efficiency. In some countries, it often took more than a year for a foreign firm to obtain necessary permits to operate. They also had to deal with heavy and complex bureaucratic rules and procedures, a very high cost of infrastructure (communications, energy, water), and constraining labor regulations. In addition, they had to commit their companies to unrealistic employment creation goals and high requirements for initial investment. In other places, qualifying firms that managed to join SEZs still had serious difficulties accessing foreign exchange and other financial services.

In sum, the belief in allowing clusters to emerge randomly generally led to disappointing results. Because of their poor design, ineffective management, and misguided policies, most SEZs did not attract enough firms in competitive industries. Moreover, their firms did not generate enough backward linkages and subcontracting business relationships with local enterprises. Too often local firms either had no interest in supplying cluster-based firms in the zones or failed to meet world market standards for quality, price, and delivery times. SEZ-based firms themselves tended to use domestic factors and inputs only to a limited extent and condemned themselves to remaining in small enclaves in poor economies. Given the often inappropriate strategic focus of these zones (where a few firms often benefited from lucrative special deals with influential politicians and could afford to produce the wrong goods in otherwise uncompetitive factories), their status as enclaves of foreign corporations with limited interactions with the local private sector exacerbated the economy-wide distortions. The disconnect with the domestic private sector worsened local business people’s perception of them. In some cases, the poor logistics and weak supply chain (both a reflection of limited clustering) led these firms to rely heavily on imports (with industries such as electronics or even apparel often showing imports ratios well over 60 percent). In such situations, currency devaluations compounded the distortion of net exports. Eventually they faced high transaction costs. Despite the (p.298) benefits of distortive protection by governments, they failed to yield enough business volume to be credible entities.

Basketball Hall of Famer Michael Jordan, whom many consider the greatest player ever, is often quoted as saying, “I’ve failed over and over and over again in my life and that is why I succeed.” Perseverance and drawing the right lessons from failed experiences are certainly keys to success. However, promoters of SEZs need not experience several rounds of painful and costly failures before getting it right. The world has recorded enough success stories in different country contexts to allow development economists to derive broad, useful guidelines for success.

Building Successful SEZs and Industrial Parks: A Few Guiding Principles

To maximize the chances of success of SEZs (making them the main vehicles for the journey toward industrial and technological upgrading, sustainable growth, and employment), developing countries should be more selective in the choice of industries developed, ensure that they are equipped with the infrastructure needed by potentially competitive firms, remove the distortions and inefficiencies that have characterized many of them in the past, and adopt a policy and institutional framework that provides the right incentives for growth and for linkages to the domestic economy. Instead of creating generic, broad-purpose SEZs, they should consider building zones with specialized facilities that are configured to the needs of specific industries and sectors—cluster-based industrial parks (CBIPs). How they are developed will depend on the industries to be promoted, all of which should be consistent with the country’s revealed or latent comparative advantage.22 With their specialized facilities customized to the unique needs of target industries, they may be airport-based zones to support air-based activities (fruits and vegetables or cut-flower exports for instance), agriprocessing zones, or even simply financial services zones aimed at promoting off-shore activities.

Good General Principles

The industries in CBIPs should be carefully selected and consistent with each country’s revealed or latent comparative advantage to ensure that they make the best possible use of the abundant factor (typically low-skilled (p.299) labor) and can become competitive in international markets without excessive forms of government protection. At least in their initial phase, they should host labor-intensive, assembly-oriented activities such as textiles, apparel, and footwear, and electrical and electronic goods. Within such industries, the scope of activities should be expanded to include not only manufacturing and processing but also commercial and professional services such as warehousing or transshipment.

All investors (foreign and local) should be treated equally. Appropriate legislation, rules, and regulation should therefore be in force to reduce the probability of distortions in incentives. Moreover, there should be a unique set of fiscal incentives for all promoted industries, regardless of their location (within the zone or outside). Never before have political leaders around the world been confronted with the difficult sociopolitical challenges posed by increasingly large, demanding, and often educated crowds. In fact, it has become very costly to remain in power without delivering tangible results, especially on the employment front. With the emergence of a new, more pragmatic leadership in developing countries, policy makers are much more likely to respond to electoral politics and be more accountable for their economic policy choices.

Deliberate efforts should be made to integrate CBIPs into national economies, not just because exporting firms operating in well-protected SEZs are better accepted politically and socially when they establish business linkages with the local private sector but also because they are more effective at building the foundations of sustained and inclusive growth when they proactively give a stake to domestic firms (especially small and medium-sized enterprises) in their activities. There is nothing worse for a potentially successful SEZ that has attracted private firms with strong competitive potential than to be (mis-)perceived as a domestic enclave for shady business ventures between corrupt politicians and foreign industrialists. Building economic and social connections with a local network of small firms and other actors (such as academic institutions) can help reduce the suspicion and resentment often faced by foreign investors arriving in developing countries.

To preempt the inevitable domestic criticism, social fears, and other political economy issues, CBIPs should strive to generate quick wins soon after they have been established.23 This is done most effectively when the newly created zones rapidly create manufacturing jobs and absorb sizable (p.300) segments of the low-skill labor force. Their promoters should encourage linkages between CBIPs-based firms and local firms so that the zones can serve as examples for success and as catalysts to broader reforms and work with local authorities and business associations to comply with ILO labor standards. It is indeed important to communicate the message that for most people in the labor force in poor countries, the alternative to employment in such CBIPs would be low-productivity, low-income informal activities, underemployment in urban areas, unprofitable and highly risky agricultural work in rural areas, unemployment, and the perpetual trap of poverty. Even with minimal formal education, many unskilled workers could still be employed in CBIPs that specialize in basic assembly operations. In the medium and long run, the strategic focus of CBIPs should be to improve the economy’s endowment structure by moving toward higher-value activities but at a realistic pace. This can be achieved only by promoting skills development for the workforce and setting industrial and technological upgrading as the ultimate goal of the firms hosted in CBIPs.

Effective Institutional Arrangements

An important question is the distribution of roles between the public and private sectors in the design, ownership, and management of CBIPs. CBIPs that are privately owned, managed, and operated should be encouraged. But they could start as public-private partnerships, with public provision of off-site infrastructure such as roads and public-private funding of on-site facilities. Governments can provide direct financial support or guarantees to build infrastructure and facilities in the zone. Private-sector participation can take many different forms: basic partnership with shared risks and rewards with governments; concession agreements; or “build-own-operate,” “build-operate-transfer,” or “build-own-operate-transfer” arrangements (FIAS 2008). Successful models of CBIPs include a variety of contract types, often with public-private partnerships that evolve over time. A model that has been popular recently involves “equity-shifting” arrangements, which allow a private contract manager of a government zone to exercise a purchase option once predefined performance levels have been reached.

Even well-designed CBIPs can succeed only if they are backed by strong political commitment from the highest government levels to improve the business environment and quickly remove all obstacles to implementation. A good institutional framework for preparation could be an interministerial (p.301) committee headed by a political “champion” who has the credibility and power to make things happen. That champion should also be the main interface between CBIPs developers and firms and all government entities. He or she should be able to respond quickly and effectively to the requests from the business community. But the champion should be insulated from political pressures to please any domestic political constituency.

Land, Facilities, and Services

Building clusters is less challenging when governments are willing to find land parcels and secure titles for lease to private zone developers. In many poor countries, the legal framework allows for an enduring influence of state bureaucracy on land distribution and land rights. Governments are reluctant to hand over the power of land distribution, and state control is legitimized as historically and socially fair. Such control opens the potential for rent seeking and bureaucratic arbitrariness. State ownership, and especially the power to redistribute land plots, makes citizens and business people vulnerable to arbitrary actions of local bureaucrats who decide to whom access to land is granted. CBIPs represent a good opportunity for implementing land reforms gradually, in a way that can generate quick wins for all stakeholders and improve their collective welfare. Even countries such as Ethiopia or Tanzania, with a long history of strong resistance to land privatization and property rights for individual plot holders, are considering changes in their land tenure policy—a sign of progress and the recognition that it may be the most viable alternative.

To expand the range of facilities and amenities available within CBIPs, public and private partners should consider not only industry-specific factories and infrastructure but also a wide array of services such as high-speed telecommunications and Internet services, common bonded warehouse facilities, training facilities, maintenance and repair centers, product exhibition areas, on-site customs clearance and trade logistics facilities, on-site housing, and on-site banking, medical clinics, shopping centers, and childcare facilities. Developing a cluster zone as an integrated industrial, commercial, residential, and recreational entity—not as a stand-alone—allows developers to diversify their potential sources of revenue and offset the potential low profitability of certain activities with higher margins in others. In many well-managed private zones in East Asia, as much as half of total annual revenue is derived from business support services and other sources of income.

(p.302) Solving the Infrastructure Problem

Poor infrastructure (inadequate and poor-quality energy supplies, high utility prices, lack of railway transport) negatively affects business. To alleviate the problem, policy makers in low-income countries are typically recommended simply to build infrastructure projects, preferably across the country and across borders. Yet large regional infrastructure projects that are supposed to yield cross-border economies of scale and link national markets often end up being costly, risky, and ineffective. The needs are enormous, and no country may ever be able to safely assert that the quantity and quality of its infrastructure is appropriate to sustain high rates of economic growth for a long period of time. It is and will remain a constant problem to be addressed in both rich and poor economies.24

Infrastructure projects are perhaps the most profitable investments any society can make. When they are productive they contribute to and sustain a country’s economic growth and therefore provide the financial resources to do everything else. But many policy makers do not find the right strategy to tackle the problem. Either they try to do too much at the same time and end up not actually doing much, or they give priority to the wrong industries and sectors and devote their limited financial, administrative, and human resources to activities that are not competitive enough and cannot generate enough payoff to sustain the development process.

The problem of infrastructure finance is therefore one of market failures, government failures, and donor failures:

  • Market failures stem from the fact that infrastructures are generally public goods. Once they are built, the marginal costs of additional consumers tend to decline toward zero. But the challenge is obviously to find the financing to invest in the heavy part. Private investors need to recover their fixed and sunk costs and make profits. If their perspective is that there is no market or credible and stable stream of revenue in a strong legal and regulatory framework, investors will obviously not get involved—even when they are granted monopolistic rights. From the perspective of the private sector, where the focus is normally on profitability and money, there is also a deficit of wisdom and innovation. Some private investors are concerned that infrastructure projects tend to offer lower-than-average (p.303) return. Others are hesitant to invest in them because the existing asset classes do not provide the structure needed for these projects to compete with traditional equity or debt.

  • Government failures are the consequences of selecting priority industries and sectors, even on the basis of traditional rate-of-return analyses, that often lead to disappointing results. The infrastructure projects that are selected do not yield expected results because of their inappropriate location, poor design, and so on. Moreover, infrastructure decisions are often political in nature and involve parliamentary processes that do not necessarily follow technical analyses and transparent processes.

  • Donor failures are due to the ineffective policy advice given by external development institutions that too often conform to dominant intellectual paradigms without seriously accounting for both market and government failures and the constraints they pose in countries with limited financial resources and administrative capacity.

These problems can be addressed effectively in countries where governments have the credibility to commit to long-term institutional and regulatory arrangements and most effectively with the involvement of international partners and development finance institutions. In an ideal world, advanced economies would be willing and even eager to transfer trillions of dollars to developing countries for infrastructure finance. This would create a global win-win, because these rich economies would benefit even more from such bold ventures. But that first-best solution would not be politically easy to carry out—even though during the Great Recession of 2008 all advanced economies could mobilize a large amount of fiscal stimulus to stimulate their domestic demand, often with limited results.

No developing nation will ever have the financial resources and administrative capacity to build roads, highways, railways, seaports, and airports across its entire territory or to foster regional interconnection with other (mainly poor) neighboring countries. The only sensible solution left is to set the correct priorities and to identify the geographic locations where high-quality infrastructure is necessary to support export-oriented light manufacturing industries for the global market. That second-best but still optimal solution is a carefully designed strategy that locates infrastructure (p.304) investments primarily in and around industrial parks that connect domestic firms to foreign firms and domestic economies to GVCs.

The provision of industry-specific, on-site infrastructure is indeed an important determinant of transaction costs and competitiveness in CBIPs. It helps attract firms and facilitate the clustering and the development of subcontracting relationships among them. Policy makers should work closely with private-sector operators to fully equip and service CBIPs with purpose-built facilities, which can then be put up for sale or lease. Private zone developers should be allowed to supply utilities services (water, power, sewerage, and telecommunications) to cluster-based firms.

Over several decades the ownership and financing of many infrastructure assets have been transferred to the private sector. A new model of privately owned and privately financed infrastructure utilities has emerged, with the government mainly playing the role of regulator. But the privatization model cannot cover all infrastructure categories, especially those that are circumscribed to small geographic areas covered by CBIPs. There is a need for project-specific institutional arrangements such as public-private partnerships and private finance initiatives. Developing countries can profit from the low interest rates across the world’s capital markets and find ways of bridging the gap between the relatively low cost of debt for existing regulated assets and the high cost of new projects.

In view of capital and regulatory constraints on banks in the wake of the financial crisis, capital markets are an increasingly important source of finance for infrastructure projects. Justin Yifu Lin, Kevin Lu, and Cledan Mandri-Perrott (2015) have proposed new financial instruments that could help channel FDI into infrastructure, including in CBIPs located in developing economies. Specifically, they recommend creating an asset class called “buy-and-hold equity,” which is between traditional equity and debt and held for fifteen years or longer. It would offer returns close to those of equity investments but with some of the risk offset by its long-term nature. The private sector would bring in infrastructure investment expertise, while sovereign funds and international financial institutions would provide the capital and stability. The platform would focus on projects with defined cash flows and contractual terms (which could include associated risk-mitigation strategies) guaranteed for twenty to thirty years.

Another viable solution to credit financing for infrastructure and economic development is strengthening public investment banks—providing (p.305) long-term financing while maintaining sustainable fiscal balance and avoiding undue strains on the borrowing capacity of central government (Aryeetey 2015; Monga 2012). Such development finance institutions can borrow on the capital markets to finance economically viable projects in competitive industries and sectors. They could offer partial or full guarantee of repayment on bonds issued by project initiators by bearing the risk and therefore substantially reducing the cost of funding. The Korea Development Bank and Development Bank of Japan offer good models of institutional and governance setups that allow them to fund major infrastructure projects while consistently avoiding losses and maintaining a tight hold on credit risk.

Mitigating the Risks of Corruption and Rent Seeking

Political economy concerns often identified in the theoretical literature on clusters are legitimate only for the traditional type of SEZs and EPZs that host firms in industries that defy comparative advantage. Firms in these industries are not viable in an open, competitive market. Their existence and continuous operation often depend on large subsidies and protection, which create opportunities for rent seeking and corruption and make it difficult for the government to abandon interventions and exit from distortions. CBIPs should promote a completely different development model: the industries that are consistent with an economy’s latent comparative advantage. Firms are viable once the constraints to their entry and operation are removed. Government incentives provided to the first movers must be transparent, targeted, temporary, and small solely to compensate for their information externality. In that context, the issues of pervasive rent seeking and the persistence of government intervention beyond its initial timetable can be mitigated. Selecting labor-intensive industries with economies of scale (so that there are incentives for foreign investors to localize in lower-wage countries) and potential for upgrading (to open up future possibilities for domestic value-added creation) would generate the kind of quick wins that policy makers need to build their own domestic political capital and to pursue reforms.

Not all developing countries are confronted with the kind of poor incentive systems and extreme rigidity in labor market rules that either stimulates rent seeking or impedes the effective development of special zones such as CBIPs. In some of them minimum wage and other labor (p.306) law rigidities are actually much less binding in practice than they appear on the books. In such countries (especially those where basic transportation, energy, and telecommunication infrastructure could be improved quickly), CBIPs could be much bolder in their design and implementation and become “freeports.” Instead of being mainly export drivers, they could be large platforms for private investment and catalysts for knowledge spillovers throughout the entire national economy and beyond and could even serve as a basis for regional hubs in specific industries. CBIPs that are selected on the basis of their economic rationale and not for political considerations can then cover much larger areas. These would therefore allow greater flexibility to firms in their choice of plant location and opportunities for interfirm linkages. They would allow full access to the domestic markets on a duty-paid basis—that is, lift the traditional requirement of exporting 80 percent or more of the production and allow instead unrestricted sale to domestic consumers as long as all applicable import taxes and other duties are paid. They would also enable firms to engage in any legal economic activity they deem profitable, including manufacturing, warehousing, and transshipment. Registered firms or individuals could be offered duty-free privileges to introduce all types of merchandise, which can then be sold at the retail or wholesale level or even consumed within the zone area.

Developing country policy makers should also learn from best practices in Ireland, Taiwan-China, and Korea and allow duty-free access to inputs for local firms, as is the case for CBIP-based firms. Domestic producers, especially small and medium-sized enterprises, could then benefit from tax credits and rebates on duties paid on imported goods and services used in products sold to CBIP-based firms. Local suppliers could import intermediary products and components using letters of credit initiated by CBIP-based firms. The latter could also provide domestic firms with technical assistance or financing arrangements as part of subcontracting arrangements. Such policy measures aimed at fostering backward linkages would eventually help diffuse political opposition to CBIPs.

Governments should also work closely with firms in competitive industries to support training and apprenticeship for workers, promote study tours and personnel exchanges, and implement programs tailored for purchasing and technical managers of export-oriented firms based in CBIPs to help their local suppliers achieve high-quality standards and meet the (p.307) required delivery times. By bringing local business leaders into the picture and creating the conditions for them to fully share the success of CBIPs, governments will foster job generation and weaken domestic sociopolitical resistance to the new policy (including from trade unions).

Finally, governments should make clear their political commitment to potential foreign investors to convince them that all constraints on businesses in CBIPs will be removed quickly. Personal engagement by presidents, prime ministers, and other high-level government officials will be needed to convey the message that once the policy is adopted, there will be no reversal. Well-organized and well-targeted (to specific industries) visits to countries where potential investors are located (China, Thailand, India, Brazil, Qatar, and so on) would help overcome skepticism and give credibility to the new policy.

The Great Recession, which plunged the world economy into turmoil in 2008–09, and its lasting negative impact on global trade and employment have also provided new opportunities for developing countries to address some of the difficult challenges they face—especially in the domain of infrastructure. Just as Jorge Luis Borges was forced to explore his true self and find a new calling, eventually making him not only Argentina’s most famous writer but also one of the greatest of the modern era, developing countries can seize this moment to reassess their economic strategies and reposition themselves to take full advantage of the infinite possibilities offered by the new world.

Foreign direct investment has been the main engine of economic growth for centuries, stimulating industrial, technological, and institutional upgrading and fostering knowledge transfer and learning opportunities. It is therefore the main ingredient for igniting and sustaining the dynamics of change that allow societies to combat poverty and achieve shared prosperity. FDI is at its best when it focuses initially on infrastructure financing in select industries and geographic locations with big potential for positive spillovers. Paradoxically, the still sluggish state of the world economy actually offers new avenues for growth to developing countries. In recent years, the cost of building infrastructure and launching economically viable new projects and programs has been much cheaper, thanks to excess capacity in (p.308) advanced economies (especially in the construction sector) and to record low interest rates.

Development experts and policy makers should find innovative ways of channeling the surplus savings from rich countries where there is excess capacity and fewer investment opportunities to developing countries where there is an urgent need for profitable ventures. This could be achieved through a new global pact between advanced and developing countries, the development of new instruments for infrastructure finance, and the rediscovery of the almost magical virtues of well-conceived, well-designed, well-managed industrial clusters.

To maximize the impact of infrastructure investment, industrial parks should be considered the most effective way of developing clusters. They can yield the clear benefits that are exemplified in economic theory dating back to Alfred Marshall. Marshall showed that concentrating production in a particular geographic area brings major external benefits for firms in that location through knowledge spillovers, labor pooling, and close proximity of specialized suppliers. Beyond his theoretical arguments are crucial and practical ones. Industrial parks provide great opportunities for building islands of excellence, even in economies afflicted with many other problems. A country with a poor business environment can still develop industrial parks and clusters with high-quality infrastructure and excellent governance and chart its path toward economic prosperity and social peace. (p.309)

Appendix 7.1: GVC Participation: Sector Groupings

Table A7.1. Eora sectors

Sector Number

Short Name









Mining and quarrying



Food and beverages

Low-tech manufacturing


Textiles and apparel

Low-tech manufacturing


Wood and paper

Low-tech manufacturing


Petroleum and chemicals

High-tech manufacturing


Metal products

Low-tech manufacturing


Electrical and machinery

High-tech manufacturing


Transport equipment

High-tech manufacturing


Other manufacturing

Low-tech manufacturing



Low-tech manufacturing


Electricity, gas, and water

Low-tech services



Low-tech services


Maintenance and repairs

Low-tech services


Wholesale trade

Low-tech services


Retail trade

Low-tech services


Hotels and restaurants

Low-tech services



Low-tech services


Post and telecommunications

High-tech services


Financial intermediation

High-tech services


Public administration

High-tech services


Education, health, and other services

High-tech services


Private households

Low-tech services



Low-tech services


(1.) UNCTAD (2013, 55) defines trade barriers as including all costs of getting a good to the final consumer other than the cost of producing the good itself: transportation costs (both freight costs and time costs), policy barriers (tariffs and nontariff barriers), and internal trade and transaction costs (including domestic information costs, contract enforcement costs, legal and regulatory costs, local distribution, customs clearance procedures, administrative red tape, and so on).

(2.) According to Constantinescu, Mattoo, and Ruta (2015), the explanation for the lower responsiveness of trade to income lies primarily in changes in international vertical specialization, most notably in the United States and China. For example, they note that Chinese exporters are now using more domestically produced inputs than imported inputs; the share of Chinese imports of parts and components in total exports has decreased from 60 percent in the mid-1990s to 35 percent today.

(3.) In 2015 it was estimated that manufacturing exports from developing economies faced a mean theoretical ad valorem tariff of only 0.63 percent in developed countries, while those from least developed countries (LDCs) faced an average tariff of 0.15 percent. Agricultural exports from developing countries and LDCs were subject to average tariff rates of 7.42 percent and 2.21 percent, respectively (ITC 2015, 1).

(4.) The UNCTAD classification of NTMs encompasses sixteen chapters (A to P), and each chapter is divided into groups with a depth of up to three levels.

(5.) Direct approaches attempt to collect information on measures of the NTMs (for example, customs procedures). This information is then introduced into gravity equations, which explain bilateral trade using a series of country characteristics, various trade cost factors, and the available information on NTMs. An assessment of the trade impact of these measures on quantities, price, or price-cost margins is then carried out. Indirect approaches use benchmarks such as traded quantities or prices (p.336) and estimate NTM ad valorem equivalents from the deviation between observed trade and the benchmark. A typical benchmark is trade within the border of the national economy (intranational trade), normally not subjected to or impeded by borders, foreign exchange issues, or the types of communications problems encountered when goods and services are exported. Trade among regions and cities of the same country is also normally subject to the same regulations, which provide an adequate basis for comparison with external trade. See Fontagné and Mitaritonna (2013), Cadot and Gourdon (2012), or Disdier, Fontagné, and Mimouni (2008).

(6.) International public institutions such as the United Nations Industrial Development Organization, the International Trade Centre, and the World Bank have launched effective programs of trade capacity building and trade facilitation to help firms in developing countries address NTMs.

(7.) It was subsequently reported that the outsourcing experiment has been too costly to Boeing. According to a “conservative” estimate by the Seattle Times, the cost of developing the 787 and rebuilding already-assembled but unusable jets may have amounted to $32 billion. The Wall Street Journal has also estimated that Boeing will have to deliver 1,100 of the 787s before the program will return a profit—at that time it had delivered fewer than 100. The more than three-year delay in the 787’s rollout also weakened its competitive position versus Airbus because it substantially delayed Boeing’s entire jet development schedule, including its next major aircraft program, the 777X.

(8.) While the focus has been on the United States and China, the global macroeconomic imbalances involved many other emerging economies, including oil-exporting Gulf countries. See Lane and Milesi-Ferretti (2014).

(9.) The press release for the car reads: “In common with the entire Rolls-Royce family of fine motor cars, the new Dawn is at the very vanguard of automotive design and technology. Dawn presents drivers with a suite of discreet technologies that ensure their leisure time in the car is a super-luxurious effortless experience.”

(10.) In July 2015 the U.S. Congress adopted an act to extend the GSP program, which had expired July 31, 2013, until December 31, 2017. The U.S. Congress first enacted the GSP in the Trade Act of 1974, which authorized the granting of duty-free treatment to specified goods from designated lesser-developed countries called “Beneficiary Developing Countries” (BDCs). There are currently 122 designated beneficiary countries and territories. Eligible articles from BDCs may be entered duty-free provided that 35 percent of the value of the article originates from one of the BDCs. The program promotes economic development by eliminating duties on up to 5,000 types of products when imported from one of the BDCs. Since its inception, GSP legislation has always included a sunset component, and the issue of lapse has been a major source of uncertainty for developing country exporters and U.S. importers.

(11.) The term “fragmentation” was originally proposed by Jones and Kierzkowski (1990). A GVC “describes a full range of activities undertaken to bring a product or service from its conception to its end use and how these activities are distributed over geographic space and across international borders” (DFAIT 2011, 86). Amador and Di Mauro (2015) (p.337) note that the economic literature has used a wide range of terminologies to describe the same phenomenon, including “vertical specialization,” “outsourcing,” “offshoring,” “multistage production,” “intraproduct specialization,” and so on.

(12.) Policies to build successful SEZs and industrial parks are discussed in the final section of this chapter.

(13.) In the example described in figure 7.6, Country A’s total exports amount to $100 million, of which its own value added (created by its own firms within its national borders) is $80 million and the imported intermediate input (value added) is $20 million. Hence FVA in Country A’s export is 20 percent. Calculating its domestic value added in global exports by other countries follows a similar logic: Country A’s total exports of $100 million can also be broken down from the perspective of the value of its intermediate inputs used in exports by other countries of $40 million, and the valued added in its own exports of final products for consumption in foreign markets is $60 million. DVX is therefore 40 percent.

(14.) The data used to construct the indicators of GVC participation come from UNC-TAD’s Eora database. It provides multiregion input-output tables at the world level, with international input-output tables reported for 187 countries over the period 1970–2011. The database reports information on between 25 and 500 industries, depending on the country. Foster-McGregor, Kaulich, and Stehrer (2015) use the twenty-five-sector database over the period 1995–2010, with data on basic prices, which reports consistent data at this level of aggregation for all 187 countries. See appendix 7.1 for the description of the twenty-five Eora sector groupings.

(15.) Krugman (1991) used a modified Gini coefficient to show that several traditional industries display a higher spatial concentration than even high-tech industries, which were thought to be the real engines of clustering. Porter (1988) defines a cluster as “a geographic concentration of interconnected companies and institutions in a particular field.” Swann, Prevezer and Stout (1998) offer a similar definition: “a large group of firms in related industries at a particular location,” which is echoed by Bresnahan, Gambardella, and Saxenian (2001): “a spatial and sectoral concentrations of firms.”

(16.) Increasing returns to scale typically refer to gains due to declining costs, which tend to fall with longer production runs. They are therefore internal to the firm. External returns to scale are typically the location gains from the presence of other firms in the same industry or related activities. For instance, technological or managerial externalities may occur as firms learn from one another by observing and borrowing best-practice techniques or organizational processes.

(17.) One reporter has commented that Qiaotou is “located slap-bang in the middle of nowhere” and “is the sort of place you might drive through without noticing. It is too small to be marked on most Western maps of China, too insignificant to merit a mention in newspapers, and so little-known that few outside the local county have heard its name” (Watts 2005).

(18.) For instance, in Peixian County, Jiangsu Province, one of the first and largest cross-regional mechanization service clusters, the government built infrastructure to (p.338) connect the townships to the national transportation network. “Peixian Bureau of Agricultural Mechanization (PBAM) selected directors from 18 agricultural mechanization service stations dispersed in different townships in the county and organized a study tour to Weifang of Shandong province to learn about their mechanization experience. … After returning home from the tour, PBAM organized free demonstration and training sessions for farmers and technicians at the township agricultural mechanization service stations. After completing training, PBAM issued a certificate allowing the trainees to drive trucks and combines to provide harvesting services. In addition, PBAM gathered harvest information nationwide, printed a pocketsize harvest calendar covering major cropping areas, and distributed them to potential machinery operators for free” (Zhang, Yang, and Reardon 2015, 15).

(19.) Bulletin of Shenzhen’s National Economy and Social Development in 2015, http://www.sztj.gov.cn/xxgk/tjsj/tjgb/201504/t20150424_2862885.htm.

(20.) Zeng (2010) estimated that as of 2007, SEZs (including all types of industrial parks and zones) accounted for about 22 percent of national GDP, about 46 percent of FDI, and about 60 percent of exports and generated in excess of thirty million jobs. In addition, fifty-four high-tech industrial development zones (HIDZ) hosted about half of China’s national high-tech firms and science and technology incubators. They registered some 50,000 invention patents in total, more than 70 percent of which were registered by domestic firms. They also hosted 1.2 million R&D personnel (18.5 percent of HIDZ employees) and accounted for 33 percent of the country’s high-tech output. In just their first fifteen years of existence, HIDZs accounted for half of China’s high-tech gross industrial output and one-third of China’s high-tech exports. In addition, similar institutions called Economic and Technological Development Zones (ETDZs) were also responsible for another one-third of China’s high-tech industrial output and exports.

(21.) Backward linkages can be defined as the various channels through which money, goods, services, and information flow between a firm and its suppliers and create a network of interdependence and mutually beneficial business opportunities. Such linkages exist when the development of an industry stimulates the growth of the industries that supply it. For example, growth of the leather industry may encourage the development of the livestock industry, which will lead to higher incomes for farmers and will create a greater demand for goods and services in the rural areas. Some authors distinguish between direct backward linkages and indirect linkages; a typical example is that of the auto industry, which has a direct backward linkage to the steel industry and an indirect backward linkage to the coal and iron industries (since coal and iron are inputs to steel production). Forward linkages are similar connections between a firm and its customers. They emerge when the growth of an industry leads to the development of the industries that use its output as input, or when the output of an industry helps stimulate activities in another industry. For example, through a forward linkage the development of the mining industry in a remote location can help build a network of rural roads needed to transport mining products.

(p.339) (22.) CBIPs should not try to promote static comparative advantage. They should support industrial diversification and upgrading. But their goals should not be too ambitious, as is often the case in countries where policy makers advocate the promotion of dynamic comparative advantage. The nuance here is important. Theories of dynamic comparative advantage typically attempt to help firms enter industries that are a country’s future comparative advantage. Because of endowment constraints in the African context, firms in those industries would not yet be viable in a competitive market even if the government helped them with the coordination and externality compensation. By contrast, CBIPs should aim at helping firms enter industries with latent comparative advantage. Under that scenario, firms would be immediately viable and require no subsidies or protection once the government provides coordination and externality compensation.

(23.) It is estimated that SEZs in sub-Saharan Africa generally contribute nearly 50 percent of exports. It can be inferred from their impact on diversifying the region’s export base that they also contribute to skill upgrading. (p.340)

(24.) According to the Africa Infrastructure Country Diagnostic (AICD), part of the African Development Bank group, the infrastructure needs of sub-Saharan Africa exceed $93 billion annually over the next ten years. To date less than half that amount is being provided, leaving a financing gap of more than $50 billion.