Commodity Chains and Economic Development:

One and a Half Proposals for Spatially-Oriented Research

Richard P. Appelbaum

working paper prepared for CSISS/IROWS Specialist Meeting,

Globalization in the World System: Mapping Change Over Time

Session 1: Commodity Chains and Labor in the Global Economy

University of California at Riverside

February 7-8, 2004

The purpose of this paper is propose two vastly different approaches to studying the role

of commodity chains in the global economy. Both use the commodity chains framework to

analyze the possibilities for industrial upgrading. The first proposes to develop an index of

industrial upgrading in individual countries, and then use the index as the dependent variable in

causal models incorporating various predictors of industrial upgrading. The second, somewhat

more adventurous strategy, proposes a commodity chains-based decision approach that would

attempt to model the complex interactions between the commodity chain and its regional

environment. The first approach is developed considerably more extensively than the second

(which is barely developed at all), both because it builds on former work I have done with others

(including David Smith, who is part of this workshop), and because it seems reasonably possible

to accomplish empirically. The second approach is developed more briefly and speculatively,

mainly because I really have no idea how to proceed further.

Before proceeding to the two proposed approaches, it is important to review the

underlying theoretical frameworks, along with some recent changes in global production systems

that are consequential for both approaches.

SOME THEORETICAL CONSIDERATIONS

In this section we briefly review the concept of global commodity chains, discuss the

importance of social networks in an increasingly globalized economy, and briefly review the

possible role of state policies in development.

Global Commodity Chains

This notion of an increasingly integrated global economy – where countries come to

occupy distinct export niches and where industrial upgrading is a key strategy – can be fruitfully

understood through the notion of global commodity chains, “network[s] of labor and production

processes whose end result is a finished commodity” (Hopkins and Wallerstein, 1986: 159).

Global commodity chains consist of a number of ‘nodes’ or operations that comprise pivotal

points in the production process: raw materials supply, production, export, and marketing, taking

us “across the entire spectrum of activities in the world-economy” (Gereffi, 1992: 94). The study

of global commodity chains, which originated with the work of sociologist Gary Gereffi and his

colleagues has spawned a major cottage industry in the sociology of development.

As originally conceived by Gereffi, global commodity chains have three main

dimensions: an input-output structure comprised of a set of products and services linked together

in a sequence of value-adding economic activities; a territoriality that identifies the geographical

dispersion or concentration of raw material, production, export, and marketing networks; and a

governance structure of power and authority relationships that determines how financial,

material, and human resources, as well as economic surplus, are allocated and flow within a

chain. While there is a large and growing body of empirical work on all three of these

dimensions, that work has consisted entirely of case studies of specific industries, most notably

low-wage industries such as apparel and electronic assembly.

Gereffi has also distinguished between two distinct types of global commodity chains –

those that are controlled by producers, and those that are controlled by buyers (see Gereffi, 1994,

for the original formulation). Producer-driven commodity chains refer to those industries “in

which large integrated industrial enterprises play the central role in controlling the production

system (including its forward and backward linkages)” (Appelbaum and Gereffi, 1994: 44). This

is most characteristic of capital- and technology- intensive industries dominated by transnational

corporations. Buyer-driven commodity chains, on the other hand, refer to those industries

in which large retailers, marketers and branded manufacturers play the pivotal roles in

setting up decentralized production networks in a variety of exporting countries, typically

located in developing countries. This pattern of trade-led industrialization has become

common in labor-intensive, consumer-goods industries such as garments, footwear, toys,

handicrafts and consumer electronics. Tiered networks of third-world contractors that

make finished goods for foreign buyers carry out production. Large retailers or marketers

that order the goods supply the specifications (Gereffi and Memedovic, 2003: 3)

This pattern of trade-led industrialization is common in labor-intensive, consumer goods

industries such as garments, footwear, toys, and consumer electronics. In the current phase of

globalization, abetted by revolutions in information technology and logistics, there has been a sea

change in global industrial organization: Producer-driven commodity chains, which dominated

during an era of Fordist production, are rapidly giving way to buyer-driven commodity chains in

which giant retail conglomerates call the shots. Wal-Mart, not General Motors, is the world’s

largest corporation.

In buyer-driven commodity chains, profits “derive not from scale, volume, and

technological advances as in producer-driven chains, but rather from unique combinations of

high-value research, design, sales, after-sales services, marketing, and financial services that

allow the buyers and branded merchandisers to act as strategic brokers in linking overseas

factories and traders with evolving product niches in their main consumer markets” (Gereffi,

1994: 99). In other words, the highest value-added activities are often more closely associated

with consumption than production. Because constant design changes for customized markets is

the primary source of competitive advantage, products have become increasingly aestheticized,

emphasizing elements of style, fad, and mystique, all of which increases the contribution of

design to the value of the product. Thus, design-intensive activities have increased their

proportion of value generated relative to manufacture and assembly activities. So one aspect of

the shift to buyer-driven commodity chains is the creation of competitive advantages through

product differentiation and customization for distinct market segments, rather than merely by

cutting labor costs: it is no longer possible to complete exclusively on the basis of low-cost labor.

The economic success of newly industrializing nations will largely depend on their firms’ ability

to “move up” into these higher value-added economic activities.

A handful of peripheral countries have engaged in industrial upgrading, shifting from

commodities like textiles, apparel and footwear to higher value-added, technologically

sophisticated production that requires a strong and well-integrated industrial base. This was the

pathway followed by the East Asian newly-industrializing economies (NIES) during the 1980s

and 1990s, when regional growth rates averaged 7-8 percent annually despite escalating wages,

labor shortages, and currency appreciation that threatened competitiveness in the very laborintensive

industries upon which they built their economic successes. Their pattern involved

continuous technological improvement of production processes, the production of new products

and the provision of new services, and otherwise engaging in higher value-added economic

activities. East Asian firms were able to move up from what Gereffi terms “captive networks” (in

which producers are limited to assembly of cut fabric following detailed instructions) into

“relational value chains” entailing “more complex forms of coordination, knowledge exchange,

and supplier autonomy,” permitting full-package production, the ability to go beyond simple

assembly and supply the client with a completely finished product by providing designing,

sourcing, cutting, sewing, assembling, labeling, packaging, and shipping (Gereffi, Humphrey,

and Sturgeon, 2003: 12).

The number of leading global apparel exporting countries has increased sharply between

1980 and 2000, with many formerly lower-tier countries “moving up” the commodity chain into

higher value-added activities. Countries whose apparel exports exceeded US$1 billion in 1980

included only the East Asian NIEs (Hong Kong, Taiwan, and South Korea), along with China

and the U.S. A decade later, the list also included Indonesia, Thailand and Malaysia; India and

Pakistan; Turkey (which had emerged as the world’s fifth-largest apparel exporter); and Tunisia.

By 2000, the list included the Philippines and Viet Nam; Bangladesh and Sri Lanka; Morocco

and Mauritius; four East European countries; and of course Mexico, who apparel exports had

grown from virtually nothing in 1990 to $9.3 in 2000. In that year the top five apparel exporters

were China ($39.2 billion), Hong Kong ($24.7 billion), the United States $9.3 billion), Mexico

($9.3 billion), and Turkey ($7.0 billion). Yet there remains substantial variation in the degree to

which apparel remains a principal export item among the world’s 25 largest apparel exporters:

In Northeast and Southeast Asia, [apparel] has declined in importance, except in China

where it remains the top export item, and in Indonesia and Viet Nam where apparel has

climbed to third place. However, in South Asia, Africa, the Caribbean Basin and Central

and Eastern Europe, apparel is the leading export, and frequently has been for a decade or

more. (Gereffi and Memedovic, 2003: 26)

If one looks at changing geogrpahical patterns for U.S. apparel imports (see Figure 1)

during the past decade, it is clear that Northeast Asian countries are declining in importance,

South and Southeast Asia have stabilized, and China, Mexico and to some extent the Caribbean

Basin have increased; only China and Mexico are core suppliers, however. For most countries

there was little change between 1990 and 2000 (Mexico being the principal exception, thanks in

large part to NAFTA). The countries that have been most successful in exporting to the U.S. are

those that do not engage in simple assembly, but have developed, or are developing, full-package

production capabilities – Hong Kong, Taiwan, South Korea in the first instance, China and

Mexico in the latter.

Figure 1: Shifts in the regional structure of United States’ apparel imports, 1990-2000*

*Note: The 2000 position corresponds to the ring where the country’s name is located; the 1990

position, if different, is indicated by a small circle. The arrows represent the magnitude and

direction of change over time. Source: Gereffi and Memedovic, 2003: p. 18

Social Networks: Personal Ties and Spatial Proximity

Although labor costs often are a crucial component of the calculations of businessmen

and investors, other factors (such as market proximity, access to skilled labor, and trade barriers)

also figure in decision-making about industrial location (Dicken, 2003). One sent of important

factors has to do with social networks. Two different (although often overlapping) types of

social networks haved receive prominent attention in the development literature: those stemming

from personal ties and connections, and those stemming from spatial proximity.

Personal Ties: The ability of firms to create informal business networks in service of

global production has received extensive attention in the development literature, and is believed

by some scholars to be a key ingredient in East Asia’s economic success. Chinese businesses in

particular are said to prosper as a result of their reliance on informal personal networks and

connections – guanxi obligations of mutual obligation and reciprocity that are frequently

mediated through family or community ties. Integration tends to be horizontal and informal,

rather than vertical and contractual, with horizontal coordination based on short-term needs

rather than long-term obligations. Firms can therefore remain small and more responsive to

quickly changing market conditions, while at the same time gaining access to the large capital,

resource, and information pools of the business group. Such informal alliances between firms in

business groups allow the network as a whole, rather than individual firms, to organize and

manage a large portion of the commodity chain. Rather than using vertical integration to solve

problems of opportunism and information flow, these problems are managed through interfirm

trust and communication. Firms can therefore remain small and more responsive to quickly

changing market conditions, while at the same time gaining access to the large capital, resource,

and information pools of the business group (Orru, Biggart, and Hamilton, 1992; Hamilton and

Kao, 1990; Smart and Smart, 1991; Lui, 1998; Gerlach, 1992; Whitely, 1992, 1996; Appelbaum,

1998; Cheng, 1993; Chan, 1993; Walton, 1993; Birnbaum, 1993; Appelbaum, Felstiner, and

Gessner 2001).

Spatial Proximity: The agglomeration effects associated with spatially concentrated,

tightly integrated metropolitan regions (“industrial districts”) are believed to confer

competitiveness by permitting a quick and flexible response to rapidly-changing market

demands. Such flexibility, which results from the transactions-intensive production and supply

networks, results in a shift away from standardized assembly-line mass production to much more

flexible, segmented production. Industrial districts confer competitive advantage through

externalities resulting from the physical presence of numerous suppliers and producers,

concentrated in geographically interdependent networks of small firms, factories, and specialized

local labor markets. Information flow is facilitated by family connections, personal relationships,

professional and community-based ties, trade associations, tight lines of communication between

neighboring suppliers, and common culture. Such flows permit a highly flexible organization of

production, with quick response to shifts in market demand. Transaction costs are lowered

through proximity to markets, the ability to quickly acquire producer goods and services, lowered

transportation and communications costs, access to suppliers, and in general the rapid exchange

of information and knowledge (Scott, 1988; Storper and Walker, 1989). The presence of a strong

support infrastructure – for example, business associations, supplier clubs, and private or statesupported

research and development facilities – can also contribute to globally competitive firms.

There is also some evidence that small- and medium-sized enterprises may be better able to

respond flexibly to changing market conditions than large ones, particularly if informally

networked into strong business groupings (Doner and Hershberg, 1996).

The Role of State Policy

Firm and industry characteristics by themselves do not account for successful upgrading.

Both unique historical circumstances and state policy also contribute to economic growth. In East

Asia’s rapid development during the 1970s-1990s, for example, the Cold War funneled vast

amounts of foreign aid into the region, while the “long boom” in the core economies during the

1950s and 1960s provided markets for exports (Appelbaum and Henderson, 1992).

Developmentally-oriented state bureaucrats sought legitimacy by pursuing policies intended to

raise overall living standards. As Evans (1995) has demonstrated with regard to the South

Korean information technology industry, becoming a global competitor can benefit from the

interventions of an activist state (what Evans refers to as ‘entrepreneurial bureaucrats’) that is

strongly connected to social and political groups committed to development.

Examples of state policies that promoted development include maintaining low wages

through the labor repression in South Korea, Taiwan, and Singapore; large-scale underwriting of

a social wage in the form of extensive public housing schemes in Singapore and Hong Kong;

investment in education and training throughout the NIEs; and various forms of industrial policy

during the latter phases of export-led growth and secondary import substitution in South Korea,

Taiwan, and Singapore. Examples of industrial policy included credit control and price-rigging as

a means to prod companies into higher value-added, higher wage and more technology-intensive

forms of production; enforced savings, as exemplified by Singapore’s Central Provident Fund;

public investment in the creation and refinement of new technologies, such as government R&D

centers whose results were made available to private companies; state creation of industrial

sectors that did not previously exist either through state companies or through the supply of

credit and financial guarantees to private companies; and state discouragement of speculative

domestic or overseas investment, thus indirectly ensuring its flow into manufacturing.

Occasionally developmental policies even called for direct state ownership of key industries – for

example, banks in South Korea, or airlines, armaments, ship-repairing in Singapore (see the

writings in Appelbaum and Henderson, 1992; Henderson and Appelbaum, 1992; Henderson,

1993; Evans, 1987, 1995; Amsden, 1989; Wade, 1990).

RECENT CHANGES IN GLOBAL PRODUCTION

There are two relatively recent developments in global production that have must be

taken into account in any effort to model the possibilities for economic development, because

both modify the prospects for industrial upgrading through movement up the commodity chain.

The first is the growing power of large retail multinationals; the second the emergence of a

stratum of giant multinational factories that are increasingly playing the role of intermediaries

between manufacturers and retailers on the one hand, and labor on the other.

The Growing Importance of Large Retailers

One of the principal changes in global apparel commodity production has been the

growing economic power of giant retailers, who exert growing control over prices and sourcing

locations, both through price pressures they exert on the independent labels they carry, and

through their growing volume of private label production (now estimated to encompass as much

as a third of all U.S. retail apparel sales). As Hamilton and Kotha (2003: 2-3) describe it,

the event of crucial historical importance was the “retail revolution” of 1965-1980 which

created mass merchandising giants such as Wal-Mart, K-Mart, and Target; and, later,

specialty retailers such as Home Depot, Best Buy, Circuit City, and Office Depot, which

today, together with the earlier established Sears, Penney's, and major grocery chains,

procure a substantial amount of products sold to final consumers. The success of these

discount general merchandisers and “category killers” also provided a context for the

success of specialized distributors, marketers, and assemblers such as Nike, The Limited,

Dell, and Gateway; as well as for an increasingly intermediary position of major

manufacturers and technology innovators such as AT&T, GE, Compaq, and HP. Internetbased

retailing, which took off in the last five or so years, most likely represents another

“revolution” in distribution with profound effects on the consumer-oriented industries.

Giant retailers have grown in size to surpass the largest manufacturers in terms of

revenues. Among retailers, the U.S. dominates the world, and Wal-Mart dominates the U.S. The

four largest U.S. retailers account for about a tenth of total U.S. retail sales. The world’s 40

largest retailers accounted for nearly $1.3 trillion in revenues in 2001, nearly 4 percent of the

world GDP (derived from Fortune, 2002). Among the top forty, twelve are based in the U.S.

accounting for nearly half (43%) of total sales; almost all the rest are from the EU (accounting for

46%). The only Asian firms in the top forty are five Japanese retailers (accounting for 11%). Wal-

Mart accounts for nearly a fifth of the combined sales of the top 40, more than three times those

of its nearest competitor, France’s Carrefour. In fact, Wal-Mart’s 2002 revenues of $246 billion

made it the world’s 18th largest economy, roughly tied with Switzerland. In the last few years the

giant retailer has surpassed Exxon, General Motors, British Petroleum, and Ford Motors in

revenues, signaling the rising power of retailers in the world economy. This suggests an

important emerging dynamic in the global economy: the US and EU overwhelmingly control the

retail end, at a time when retailers in general are exerting increasing control over the global

economy (Appelbaum, forthcoming 2004).

In terms of labor, the dominance of giant retail transnationals poses a significant

challenge to working class organization, since their buyer-driven commodity chains are

characterized by extreme post-Fordist production involving networks of global outsourcing and

high levels of capital mobility. In the classical global buyer-driven commodity chain formulation,

retailers have disproportionate control over the manufacturers who design the goods they sell and

the factories where those goods are made (Appelbaum and Gereffi, 1994; Gereffi, 1994, 2001).

The Gap, to take one example, sources from 4,000 factories in 55 countries; Disney, to take

another, from 30,000 factories. Because these giant firms can place their orders anywhere on the

planet they choose, their contractors are seen as relatively powerless price-takers, rather than

partners and deal-makers. The effects on labor of this arrangement are mixed: one outcome is

the “race to the bottom,” where retailers and manufacturers play off competing contractors to

force prices (and wages) down and thwart unionization drives. Another outcome, however, is that

if large retailers and manufacturers can be made to pressure their suppliers by consumer pressure,

gains for labor can also be achieved – as occurred in Mexico’s Kukdong (Mexmode) factory and

the Dominican Republic’s BJ&B cap company.

Large retailers characteristically have large volume requirements, leading them to only

consider large producers (1000+ workers) as potential suppliers. In the words of one African

supplier, success requires “never deviating from a chosen product type, not trying to be versatile,

seeking efficiency on single styles and going for longer and longer runs” (Gibbon, 2003: 33).

Related to these trends, since the mid-1980s, there has been a move toward “lean

retailing,” particularly in the U.S. but also in Europe and Japan. Traditionally, apparel producing

firms and retailers were relatively independent of one another. Led by Wal-Mart and other large

U.S. retailers, and enabled by technological changes that permitted a high degree of data sharing

and other electronic interchanges, retailers increasingly brought their suppliers under much more

direct control, requiring them to “implement information technologies for exchanging sales data,

adopt standards for product labeling, and use modern methods of material handling that assured

customers a variety of products at low prices” (Abernathy et al, 1999: 3). Such changes in

retailing favor Hong Kong, Taiwanese, and South Korean garment firms (Gereffi, 2003), who are

well positioned to manage triangle manufacturing (so-called because a foreign buyer places an

order with an East Asian firm which manages the production, completing the triangle by shipping

the goods to the foreign buyer; see Gereffi and Pan, 1994: 127). As Thun (2001: 15) notes in his

study of Taiwanese firms,

small, local firms in Southeast Asia or mainland China may be able to undercut a

Taiwanese firm on labor costs, but they are unlikely to be able to make the investments in

electronic data interchange that make rapid response possible. In short, being able to

handle electronic orders from buyers, effectively forecast, plan, track production, and

manufacture apparel quickly and flexibly, are skills that provide a far more enduring form

of comparative advantage for Taiwanese firms than constantly scouring the globe for the

lowest cost labor.

One study of European retailing (focusing on Britain, France, and Scandinavia) found that

Scandinavian retailers tended to concentrate their purchases among a relatively small number of

foreign suppliers, while French retail sourcing was more dispersed (British retailers were in

between). The study identified three different models of supply base management (Palpacuer,

Gibbon, and Thomsen, 2003):

_ a rules-based UK model emphasizing rationalization of the supply chain through formal

supply chain management (SCM) doctrines, with specialized functions centralized at

corporate headquarters

_ a market-based Scandinavian model emphasizing concentrated sourcing networks,

achieved by establishing strong personal relations with overseas manufacturers

_ a socially-embedded French model emphasizing more open, informal, and dispersed

sourcing networks

The growing size and dominance of larger EU and U.S. retailers suggests an important

dynamic in the world economy: the experience of Hong Kong, Singapore, Taiwan, and South

Korea – newly-industrializing economies that relied on apparel and textile production as integral

parts of successful development strategies – may prove difficult to replicate in a world where the

retail end is much more tightly controlled today than it was 20-30 years ago.1 Only countries with

sizeable internal markets, such as China and India, may prove capable of moving up the apparel

chain into higher value-added activities, insofar as they are able to capitalize on their internal

markets in developing indigenous retail capabilities.

1 There are other factors which make it less likely that other countries will be able to replicate the

original East Asian experience. For a more complete discussion, see Henderson and Appelbaum (1992).

The Growing Importance of Major Producers

This system of retail dominance is being challenged somewhat by the rise of global

contractors, typically from South Korea or Taiwan, many of whom began as small local

producers in their home countries, using their know-how to go multinational. A handful of these

have grown to giant size, where they often have as much power as all but the largest retailers,

constituting still another layer of price-making and profit-taking. Consider, for example, the

following examples of giant global contractors:

_ Nien Hsing Corporation, a Taiwanese multinational that employs more than 20,000

workers in five Central American factories, as well as thousands of workers in a Mexican

factory and two in Lesotho. Founded in 1986, Nien Hsing is currently the world's largest

jeans maker, with an output of 40 million pairs in 2000, making jeans for Wal-Mart, JCPenny,

K-Mart, the Gap, Sears and Target. It is also the sixth-largest denim maker,

producing 60 million yards per year.

_ Yupoong, Inc., a South Korean multinational, which is the world’s second largest cap

manufacturer, exporting their “flexfit” hats (motto: “worn by the world”) to some 60

countries. Yupoong (2003) operates the BJ&B hat factory in DR, the scene of the second

recently successful labor struggle that we will consider, as well as Dhakarea Ltd. in

Bangladesh.

_ Boolim, a South Korean multinational that was founded in 1994 by Y.S. Lim, who had

headed up Macy’s in South Korea for 14 years. Boolim makes athletic, casual wear, and

knitwear in some countries, including China, Indonesia, Sri Lanka, Bangladesh, Saipan,

Thailand, Philippines, Malaysia, Myanmar, Guatemala, Mexico, Dominican Republic,

Nicaragua, Honduras, El Salvador and Vietnam Its clients include Nike, Polo Ralph

Lauren, Kenneth Cole, Calvin Klein, and NBA Properties.

_ Pou Chen, a Taiwanese multinational, is 50% owner of Tue Yen Industrial, a Hong

Kong-listed shoe manufacturer that is the world’s largest, employing 150,000-170,000

workers worldwide. Yue Yen, which makes shoes for Nike (about half of its total

production), as well as adidas-Saloman, Reebok, New Balance, Asics Tiger, Converse,

Puma, Keds, Timberland, and Rockport, controls 17% of the world market. Most of its

shoes are made in low-cost factories throughout southern China; its Yue Yen II factory

complex in Dongguan, China, employs more than 40,000 workers. The company is

Nike's biggest supplier, providing 15% of Nike’s shoes, with one Indonesian factory

turning out a million shoes a month for Nike. The company’s Huyen Binh Chanh megafactory

in Vietnam will be the largest footwear factory on the planet, employing 65,000

workers (Bailey, 2003; Boje, 2000).

One study of changing patterns of imports to Britain, France and Scandinavia concluded

that as recently as the late 1980s, southern Europe (mainly Portugal and Italy) was by far the

leading source of imports to the three countries combined. Today the picture is far different:

…by 2000, this picture changed so that Asian and ‘greater European’ producers were of

roughly equal significance, ahead of their Southern European counterparts…. Importing

countries’ increasing dependence on a combination of ‘low price’ and ‘medium

price/short lead time’ producing countries lends support to the idea that there are now

commonly acknowledged ‘global production centres’… Factors to do with history,

language and proximity play a role in determining the weight that specific supplying

countries and regions enjoy in specific end-markets, even within this framework

(Palpacuer, Gibbon, and Thomsen, 2003: 7-8).

Finally, it should be noted that the growing importance of giant producers may

paradoxically be facilitating worker organizing, since the large factories are vulnerable to

pressure from the large retailers and manufacturers that use them. A number of successful

unionization drives have occurred in such factories in recent years, including the Kukdong (now

Mexmode) apparel factory in Mexico, the BJ&B hat factory in the Dominican Republic (owned

by Yupoon); and Hien Hsing factories in Mexico (Chentex) and Lesotho. In these examples,

pressure on the factories and their clients (which included Nike, Reebok, the Gap, and other

major U.S. companies) by local independent labor unions, supported by U.S. and EU unions and

NGOs, have caused the parent companies to allow the formation of independent unions.2

ESTIMATING THE DETERMINANTS OF INDUSTRIAL UPGRADING

One approach would empirically estimate the circumstances under which labor-intensive

industrialization – which played a key role in the early development of the growing economies of

East Asia – contributes to economic development. It builds on my earlier work with David

Smith, Brad Christerson, and Herbert Wong (see, for example, Appelbaum, Smith, and

Christerson, 1993; Appelbaum, Smith, and Wong, 1998).

Measuring Industrial Upgrading

Appelbaum, Smith, and Wong (1998) suggested developing an index of industrial

upgrading in individual countries, estimating causal models using the index as the dependent

variable. We proposed analyzing exports from all non-core developing countries to the United

States for 35 period 1965-2000, at the broad (two-digit) SITC level, in order to discern different

paths of industrial transformation, as well as conducting a more nuanced analysis of highly

specific trade flows for two commodities, apparel and consumer electronics.

‘Moving up the value chain’ is typically taken to mean that producers adopt more

capital-intensive processes and techniques, while at the same time switching to the production of

more sophisticated and expensive ‘high-end’ goods. Measuring this type of change would

capture an important component of industrial upgrading. Fortuitously, international trade data

are available on a yearly basis from the United Nations that provide standardized comparable

information across a range of countries. Data are coded using the hierarchically ordered Standard

International Trade Classification (SITC), which allows us to examine a level of detail ranging

from either very broad (one- or two- digit categories) or extremely specific (seven- to nine-digit

categories). These data also include information on the unit volume and dollar value of the

international commodity flows.3 Smith and Nemeth (1988) attempted to empirically sort

commodities into ‘bundles’ of exports which flow together in the circuits of world trade. By

factor analyzing all bilateral trade for every country with a population greater than one million

which provides complete import and export data, they identified five major groups or “bundles”

2 For more detailed discussion see Espenshade, 2003, forthcoming.

3 For a general discussion of the data see Nemeth and Smith, 1985; Smith and White, 1992; for specific

examples see Appelbaum, Smith, and Christerson 1993.

of two-digit commodities (from food products and low wage/light manufacture to hi tech/heavy

manufacture; see Smith and Nemeth, 1988: Tables 2 and 3).

The Smith/Nemeth strategy could be replicated, but using international commodity trade

data for all countries in the most recent year available (the Smith/Nemeth analysis relied on 1980

data). This would provide one measure of the level of upgrading that characterizes a country’s

exports. It is important to note that this operationalization of upgrading is partial. One of the key

insights of the commodity chain approach is the importance of considering non-production

aspects such as design, distribution and marketing of final products. Data classifying

manufacturing output, even if it is by very specific product types, does not offer direct evidence

about the extent to which there is a move to local design or brand name marketing.

Measuring Changing Export Profiles

The analysis of commodity trade from non-core nations to the US between 1965 and 2000

would yield a detailed image of how each country’s export profile has changed over the last 35

years, revealing differences in the path of industrial transformation between countries. This in

turn would provide a gauge of changing commodity export mixes that reflect the ebbs and flows

of technologically-driven and fashion-related product cycles. There are a number of possible

measures that tap into dimensions of the production side of industrial upgrading, which can be

arrayed from the simplest to the most complex:

a. Changing average unit value of trade in all products.

b. Changing average unit value amount for major product groups. A simple analytic

strategy would be to compare the changing production levels of different commodities

(at either grouped, generic, or very specific-levels of classification) by calculating

autocorrelation models of changes in either volume or value over the 35 year period

(or any shorter periods). The coefficient of the time variable estimates the annual

growth rate for that type of export (cf. O’Hearn 1994).

c. Changing index of dissimilarity, calculated from the largest fifteen two-digit SITC

categories in each country. This measure gauges export diversification: countries

undergoing industrial upgrading should have a higher degree of dissimilarity over

time. Both weighted and unweighted measures could be constructed in a range

between 0 and 100.

d. Changing concentration measures, also calculated from the largest fifteen two-digit

SITC codes for each country. This measure gauges export specialization: countries

undergoing industrial upgrading are likely to have a lower degree of concentration

over time. This measure also ranges from 0 to 100.

e. Changing index of industrial transformation, calculated using recalibrated Smith-

Nemeth “bundles.” This measure is defined as the total value of export in hitech/

heavy manufacture to low wage/light manufacture. For countries undergoing

industrial upgrading the index should increase over time.

There should be major differences between countries on these indices. In particular, the

established East Asian NIEs are likely to stand out with a steady pattern of upgrading over almost

the entire period. Has the upward arc slowed or stagnated in light of the East Asian slowdown of

1997-8? One would expect the second-tier East and Southeast Asian NIEs to begin this process

later and to score more modestly, with latecomers like China and Vietnam starting their

upgrading even later (but, perhaps, to have particularly steep recent increases). It will be of great

interest to determine whether the various latecomers simply follow a trajectory that replicates the

initial group of NIEs, whether their upgrading is more rapid and skips stages. Finally, it should

be possible to determine whether there is a distinctive “Asian model” that is distinguishable from

less-developed countries in other regions, like Latin America or Africa.

Analysis of Upgrading in Apparel and Consumer Electronics

A more fine-grained analysis of upgrading is possible using seven- and nine-digit SITC

categories, focusing in particular on apparel and consumer electronic assembly. Data could be

analyzed for the period 1965-2000 for all non-core countries, in order to facilitate a comparison

with the East Asian NIEs, since both of these industries served as critically important motors of

export-led industrialization in that region.4 In apparel manufacture, Hong Kong and Taiwan

moved from sewing, to sourcing offshore production for U.S. and European designers; they are

now moving up into designing and marketing branded labels themselves. A similar process has

occurred in South Korea and Singapore’s consumer electronics industries, where the movement

has been from component assembly to engineering and design. It seems reasonable to assume

that these two industries are playing the same role throughout East and Southeast Asia, and may

potentially play this role in other countries.

Yet apparel and consumer electronic assembly differ in significant ways as well: unlike

electronic assembly, apparel production remains greatly resistant to technological upgrading

(Taplin, 1989, 1994; Bonacich and Appelbaum, 2000; Waldinger, 1986; Dicken, 2003). The

principal technological changes have been in automated fabric cutting, specialized operations

such as embroidering and button-holing, and electronic point-of-sales (EPOS) inventory systems.

Organizationally, a few factories have replaced the bundling system with unit production, thereby

reducing the time spent on handling. Second, both industries are characterized by flexible

production systems, which are themselves viewed by many theorists as an important key to

global competitiveness (Storper and Walker, 1989; Scott, 1988; Malecki, 1991). Insofar as

flexibility calls for simultaneously minimizing production costs while rapidly responding to

frequent demand, it has strong appeal in industries with tight coordination between design,

production, and marketing (Dicken, 2003). In both industries, the need for flexibility translates

into layers of subcontracting in which manufacturer-designers contract to numerous factories,

resulting in an uncoupling of the various components of manufacturing. This disintegrated form

of flexible accumulation greatly increases the importance of personal networks, which is another

feature of economic development we wish to investigate.

One approach would therefore be to construct 35 year sequences of export profiles to the

United States for all countries, with special attention given to those in East and Southeast Asia.

One would expect varying degrees of upward movement across different countries, as well as

across specific commodities. Previous research suggests that export-oriented manufacturing

4 The simplicity of this equation and the wide availability of worldwide cross-national on trade and GDP

make this feasible. The ten East and Southeast Asian countries include the Four Tigers (Hong Kong,

Singapore, South Korea and Taiwan) and six latecomers (China, Indonesia, Malaysia, the Philippines,

Thailand, and Vietnam).

economies, particularly as they move beyond the most labor intensive, low value-added ,

manufacturing, are likely to move toward more specialized export niche production to bolster

international competitiveness. This sort of commodity-specific pattern, likely to be obscured by

the aggregation of products into broad export categories, should manifest itself in this finergrained

analysis. It is also likely that the rate of upgrading in either of these specific sectors will

vary over time within each country. A careful examination of such patterns would make it

possible to discern the developmental sequences that each country has followed. These

sequenced paths of upgrading, graphed across the years, could be used to make some interesting

comparisons between countries. For instance, a retrospective look at patterns of apparel or

electronics upgrading in South Korea or Taiwan from the 1970s could be compared to more

recent changes in China or Vietnam.

The use of time-series data permits thus makes it possible to quantitatively assess the

determinants of upgrading. One strategy would involve pooled panel regression in order to

estimate models that control for the initial values of the dependent variable while assessing the

impact of the independent variables over time.5 Based on the preceding discussion, the principal

independent variables for this analysis might include:

1) Firm competitiveness, as indexed by average measures of labor cost and productivity,

quality, reliability, etc (some of these ratings may have to be subjectively based on the

perceptions of experts familiar with the industries of different countries)

2) Time-to-market (this would be one principal spatial component of the model –

estimating the relative importance of spatial propinquity in commodity flows,

looking, for example, at changing regional patterns of import-export relations)

3) The degree to which highly networked, spatially concentrated industrial districts exist

that reduce transaction costs and enable firms to engage in all aspects of production

(measuring this and estimating effects would provide another spatial component of

the model)

4) The social organization of a country’s firms into mutually supportive networks of

producers and suppliers, in particular the presence of informal (e.g., Chinese) business

networks (operationalizing this could be difficult; at worst, dummy or simple ordinal

variables could be developed as subjective measures based on existing research)

5) The role of retailers relative to manufacturers as the principal customer for exports

from each countries (suggestions for estimating this would be welcome; I can find no

systematic source of data on this, although information could possibly be gleaned –

laboriously – from the annual reports of publicly-traded retailers and manufacturers)

6) The relative importance of transnational producers in each country’s factory sector

(this would require a country-by-country survey of knowledgeable experts)

7) Changing trade barriers, including preferential trade agreements such as the North

American Free Trade Agreement (NAFTA), the African Growth and Opportunity Act

(AGOA), the Caribbean Basin Trade Partnership Act (CBTPA), and the Andean

5 The dependent variable, Yt, is regressed on itself at the earlier point in time, Yt-1, as well as on the

independent variables, Xi(t-1).

Trade Preferences Act (ATPA). These could be incorporated by means of dummy

variables, or perhaps ordinal measures reflecting experts’ perception of their impact

on trade

8) State role, as indexed by the proportion of government spending in business,

infrastructure, and education; the extent of national industrial policy (most likely

these would be dummy variables intended to capture the degree of marketization vs.

central planning, based on the characterization of these economies in the literature)

9) The role of labor, in particular the presence of an independent labor movement,

strikes and work stoppages, etc.

Secondary variables of substantive interest could include:

10) Flexibility/adaptability, as indexed by the average manufacturing firm size; market

concentration; and percent of GDP generated by SMEs

11) Human capital development, as indexed by the percentage of the adult population

with secondary education or more; the percentage of the population with tertiary

education; and the percentage of the population with technical/engineering education

12) Foreign penetration: ratio of FDI/DBI

13) Domestic economic conditions, as indexed by the absolute size of the economy (an

index of the size of the domestic market), domestic savings rates; the unemployment

rate; the ratio of public to private investment ratio; the percent GDP that is generated

by exports

14) Entrepreneurship, as indexed by the percent of the working population that is selfemployed;

and the new business start-up rate

15) Demographic characteristics, such as the age structure of the population which could

effect workforce participation

Contrasts between Hong Kong, Taiwan, Singapore, and South Korea, China and

Vietnam, and the other East Asian countries are of obvious interest. These sorts of comparisons

(and separate country-by-country analysis) can be carried out either by using dummy coding or

comparing parameter estimates from separate equations of sub-samples.

A COMMODITY CHAINS-BASED DECISION APPROACH TO MODELING

INDUSTRIAL UPGRADING

The global commodity chains framework lends itself to a decision model approach to

understanding the determinants of a firm’s locational decisions (i.e., to move production to – or

out of – a particular location), the regional impacts of those locational decisions, and the impact

of any resulting regional changes on subsequent decisions. To my knowledge this approach has

never been attempted (which probably says something about its feasibility, if not its merit). In

this final section I schematically outline such an approach, in hopes that someone will discern a

plausible modeling and research strategy. The basic logic would be as follows:

1. Construct a hypothetical global commodity chains for a product, modeled on the actual

structure of any existing firm - for example, the U.S.’s largest apparel retailer, The Gap

(with 2003 sales of $14.5 billion). One would begin by mapping out all of the networks

on the commodity chain. In the apparel commodity chain, for example, on Gereffi and

Memedovic (2003) have identified five categories of networks: those having to do with

raw material inputs (both natural and synthetic fibers), components (yarn, fabrics,

petrochemicals, synthetic fibers), production (divided into different geographic regions),

export (branded companies, overseas buying offices, trading companies), and marketing

(department stores, specialty stores, mass merchandise chains, discount chains, and offprice

outlets).

2. Conceptualize each of these networks as a set of decisional nodes. For example, if the

hypothetical firm is engaged in making a part cotton / part synthetic blouse, it needs to

make decisions about where to source the cotton; where to source the synthetics; where to

acquire bolts of fabric; where to assemble the blouse; etc.

3. Each hypothetical decision-maker then conducts an “environmental scan,” looking at

different locational options for the activity in question. Should garment assembly be done

in a contract in Los Angeles? Mexico? Bangladesh? China? Model the determinants of

this decision, based on what we know about such decisions. The model would incorporate

such factors as labor costs and productivity, labor militancy, the presence (or absence) of

labor unions, production quality, transportation costs, time-to-market (including

reliability), preferential trade treatment, the presence (or absence) of large producer

transnationals, the presence of supportive social networks and viable industrial districts,

state policies, etc. In other words, all of the predictors that are considered in the previous

approach to estimating the impact of industrial upgrading.

4. Assume a set of locational decisions, based on the foregoing considerations, for each

decisional node. Begin by focusing on assembly, which is the most crucial node from an

economic development perspective. Model the impact of each decision on the location

that is chosen. One set of impacts would have to do with industrial upgrading – for

example, prospects for the development of indigenous full-package production

capabilities, the development of local backward and forward linkages in the commodity

chain, including developing textile suppliers at one end, and original brand manufacturing

(OBM) capabilities on the other. The other set of impacts would have to do with labor –

for example, the effect on wages, inequality, and labor militancy.

5. Taking these impacts on the chosen location into account, what is the likely feedback on

the decision-maker in that particular node of the commodity chain? At what point does

the retailer or manufacturer decide to move production elsewhere? What are the

determinants of such a decision (for example, local labor shortages, that result in rising

wages)? What could be done locally to discourage such a decision?

6. Calibrate the accuracy of the model by comparing its results with actual results in

comparable production systems over the past decade.

7. Repeat with other commodity chains – for example, in a capital intensive industry (e.g.,

automobile manufacture), or a labor-intensive industry that is more capital-intensive than

apparel manufacture (e.g., footwear). Aggregated across firms, what does this approach

tell us about the prospects for industrial upgrading?

 

References Cited

Abernathy, Frederick H., John T. Dunlop, Janice H. Hammond, and David Weil (1999) A Stitch

in Time: Lean retailing and the transformation of manufacturing – Lessons from the

apparel and textile industry. Oxford: Oxford University Press

Amsden, Alice H. 1989. Asia's Next Giant: South Korea and Late Industrialization. Oxford

University Press.

Appelbaum, Richard P. (1998) “The Future of Law in a Global Economy,” Social & Legal

Studies 7: 2 (June): 171-192

Appelbaum, Richard P. (forthcoming) “Fighting Sweatshops: Problems of Enforcing Glbal Labor

Standards,” in Richard P. Appelbaum and William I. Robinson, Critical Globallization

Studies. New York: Routledge

Appelbaum, Richard P., William L.F. Felstiner and Volkmar Gessner, eds. (2001) Rules and

Networks: The Legal Culture of Global Business Transactions. Oxford, England: Hart

Appelbaum, Richard P., David Smith, and Herbert Wong (1998) “Moving Up: Industrial

Upgrading, Social Networks, and Flexible Production in East and Southeast Asia,”

proposal to the National Science Foundation

Appelbaum, Richard P. and Gary Gereffi. 1994. "Power and Profits in the Apparel Commodity

Chain," in Edna Bonacich, Lucie Cheng, Norma Chinchilla, Norma Hamilton, and Paul

Ong (eds.), Global Production: The Apparel Industry in the Pacific Rim. Philadelphia,

PA: Temple University

Appelbaum, Richard P. and Jeffrey Henderson. 1992. States and Development in the Asian

Pacific Rim. Newbury Park, CA: Sage.

Appelbaum, Richard, David A. Smith and Bradley Christerson. 1993. "Commodity Chains in the

Pacific Rim: Garment Trade and Manufacturing." Pp. 187-204 in Gary Gereffi and

Miguel Korzeniewicz (eds.), Commodity Chains and Global Capitalism. Westport, Conn:

Greenwood.

Birnbaum, David G. 1993. President, Third Horizon Ltd., Hong Kong (Brad Christerson

interview, July 1)

Bonacich, Edna and Richard P. Appelbaum (2000). Behind the Label: Inequality in the Los

Angeles Apparel Industry. Berkeley, CA: University of California Press

Chan, Joseph. 1993. General Manager, Innova Ltd., Hong Kong (Appelbaum interview, April 22

and May 2)

Cheng, Francis Man-Piu. 1993. Assistant Managing Director, Wing-Tai Exporters, Ltd., Hong

Kong (Christerson interview, July 19)

Dicken, Peter. 2003. Global Shift: Reshaping the Global Economic Map in the 21st Century, 4th

edition. New York: Guilford Publications

Doner, Richard and Eric Hershberg. 1996. “Flexible Production and Political Decentralization:

Elective Affinities in the Pursuit of Competitiveness?” paper presented at the annual

meeting of the American Political Science Association Meeting, San Francisco (August)

Esbenshade, Jill (2003) “Leveraging Neo-Liberal ‘Reforms:’ How Garment Workers Capitalize

on Monitoring,” paper given at the annual meetings of the American Sociological

Association, Atlanta (August)

Esbenshade, Jill (forthcoming) Monitoring Sweatshops: Workers, Consumers and the Global

Apparel Industry. Philadelphia: Temple University Press

Evans, Peter B 1987. "Class, State, and Dependence in East Asia: Lessons for Latin

Americanists." Pp. 203-26 in The Political Economy of the New Asian Industrialism,

edited by Frederic C. Deyo. Cornell University Press.

Evans, Peter. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton, NJ:

Princeton University Press

Gereffi, Gary. 1992. "New Realities of Industrial Development in East Asia and Latin America:

Global, Regional, and National Trends." Pp. 85-112 in Richard Appelbaum and Jeffrey

Henderson (eds.), States and Development in the Asian Pacific Rim. Newbury Park, CA:

Sage.

Gereffi, Gary. 1994. "The Organization of Buyer-Driven Global Commodity Chains: How U.S.

Retailers Shape Overseas Production Networks." in G. Gereffi and M. Korzeniewicz

(Eds.), Commodity Chains and Global Capitalism (Westport, CT: Greenwood Press.

Gereffi, Gary (2003) personal communication (email December)

Gereffi, Gary, John Humphrey and Timothy Sturgeon (2003) “The Governance of Global Value

Chains,” Global Value Chains Initiative publications webpage

(http://www.globalvaluechains.org) (November 4) (forthcoming in Review of

International Political Economy)

Gereffi, Gary and Olga Memedovic (2003) The Global Apparel Value Chain: What Prospects for

Upgrading by Developing Countries. Vienna: United Nations Industrial Development

Organization (sectoral studies series) (http://www.unido.org/filestorage/

download/?file%5fid=11900)

Gerlach, Michael L. 1992. Alliance Capitalism: The Social Organization of Japanese Business.

Berkeley: University of California Press.

Gibbon, Peter (2003) “'The Africa Growth and Opportunity Act and the Global

Commodity Chain for Clothing,” World Development 31:11

Hamilton, Gary G. and Kao, Cheng-shu 1990. "The Instututional Foundations of Chinese

Business: the Family Firm in Taiwan," Comparative Social Research, Vol 12. pp. 135-

151.

Hamilton, Gary G. and Suresh Kotha (2003) “Marketing, Merchandising, and Retailing: The

Role of Intermediaries in Global Value Chains – Workshop Concept Paper,” Seattle, WA:

University of Washington, Industry Center Workshop Series

Henderson, Jeffrey (1993) “Against the Economic Orthodoxy: On the Making of the East Asian

Miracle,” Economy and Society 22:2: 200-217

Henderson, Jeffrey and Richard P. Appelbaum. 1992. "Situating the State in the East Asian

Development Process," in Richard P. Appelbaum and Jeffrey Henderson (eds.), States

and Development in the East Asian Pacific Rim. Newbury Park: Sage

Hopkins, Terence K. and Immanuel Wallerstein. 1986. “Commodity Chains in the World-

Economy Prior to 1800,” Review 10(1): 157-170

Lui, Tai-lok (1998) “Trust and Chinese Business Behavior,” Hong Kong: Department of

Sociology, Chinese University of Hong Kong (unpublished paper)

Malecki, Edward. 1991. Technology and Economic Development: The Dynamics of Local,

Regional, and National Change. New York: Longman

Nemeth, Roger and David A. Smith. 1985. "International Trade and World System Structure: A

Multiple Network Analysis." Review (Fernand Braudel Center) 8(4): 517-560.

O'Hearn, Denis. 1994. "Innovation and World-System Hierarchy: British Subjugation of the Irish

Cotton Industry, 1780-1830." American Journal of Sociology 100(3): 587-621.

Orru, Marco, Nicole Woolsey Biggart, and Gary G. Hamilton. 1992. "Organizational

Isomorphism in East Asia: Broadening the New Institutionalism." In The New

Institutionalism in Organizational Analysis, edited by Walter W. Powell and Paul J.

DiMaggio. Chicago: University of Chicago Press.

Palpacuer, Florence, Peter Gibbon, and Lotte Thomsen (2003) New Challenges for Developing

Country Suppliers in Global Clothing Chains: A comparative European Perspective,”

Institute of Development Studies, Global Value Chain Initiative

(http://www.ids.ac.uk/globalvaluechains/publications/clothingchains.pdf)

Scott, Allen. 1988. "Flexible Production Systems and Regional Development," International

Journal of Urban and Regional Research Vol. 12: 171-186

Smart, J. and A. Smart. 1991. "Personal Relations and Divergent Economies: A Case Study of

Hong Kong Investment in South China," International Journal of Urban and Regional

Research 15:2: 216-233

Smith, David A. and Roger Nemeth. 1988. "An Empirical Analysis of Commodity Exchange in

the International Economy, 1965-1980." International Studies Quarterly 32: 227-240.

Smith, David A., and Douglas White. 1992. "Structure and Dynamics of the Global Economy:

Network Analysis of International Trade, 1965-1980." Social Forces 70(4): 857-893.

Storper, Michael and Richard Walker. 1989. The Capitalist Imperative: Territory, Technology

and Industrial Growth. New York: Basil Blackwell

Taplin, Ian. 1989. "Segmentation and the Organization of Work in the Italian Apparel Industry"

Social-Science Quarterly, Vol. 70, No. 2: 408-424.

Taplin, Ian. 1994. "Strategic Reorientations of U.S. Apparel Firms" in G. Gereffi and M.

Korzeniewicz (Eds.), Commodity Chains and Global Capitalism. Westport, CT:

Greenwood Press.

Thun, Eric (2001) “Growing Up and Moving Out: Globalization of ‘Traditional’ Industries in

Taiwan,” Cambridge: MIT Industrial Performance Center Special Working Paper 00-004

(June) (http://ipc-lis.mit.edu/globalization/globalization%2000-004.pdf)

Wade, Robert. 1990. Governing the Market: Economic Theory and the Role of Government in

East Asian Industrialization. Princeton University Press.

Waldinger, Roger David. 1986. Through the Eye of the Needle: Immigrants and Enterprise in

New York's Garment Trades. New York: New York University Press.

Walton, Steven R. 1993. Managing Director, Innova Ltd., Hong Kong (Appelbaum interview,

May 2)

Whitely, Richard (1992) Business Systems in East Asia: Firms, Markets, and Societies. Newbury

Park, California: Sage.

Whitely, Richard, “Business Systems and Global Commodity Chains: Competing or

Complementary Forms of Economic Organization,” Competition and Change: The

Journal of Global Business and Political Economy I: 4 (fall 1996): 411-425; p. 412.