Abstract
In the mid-nineteenth century, Karl Marx issued several critiques of political economy writings stressing the exclusive duality of states and the national economies. He argued that capitalism had characteristic features quite apart from those shaped by the idiosyncrasies of national economies. In the first part of this article, we critique the contemporary state-centered explanations for the industrialization of East Asia on same grounds. We claim that most political economists misinterpret or entirely ignore the significance of export-led industrialization, which is a characteristic feature of East Asian capitalism. In the second part of the article, we demonstrate the importance of the retail revolution in the US and Europe on Asian industrialization. In particular, we show that the development of the sequential ordering system that is an inherent feature of Western-based contract manufacturing differentially shaped the industrial organization of East Asian economies. The resulting path-dependent trajectories of development, in turn, encouraged government policy-makers to ‘buy into’ trends of global capitalism in different ways. The trajectories also led business people to privilege and adapt some social institutions and cultural patterns over other ones.
Keywords
Karl Marx and the critique of political economy
In the face of the voluminous and growing literature about the institutional foundations of economies and about the resulting varieties of capitalism, we find ourselves drawn, rather suddenly and with sympathy, to the late writings of Karl Marx. We have a renewed appreciation for the intellectual problem that Marx took on as he worked towards his final project, Das Kapital, a planned three-volume work, of which he only completed the first. Marx’s problem was how to convince his readers that capitalism has essential characteristics quite apart from the idiosyncrasies of national economies.
Starting in the 1840s, Marx read and took copious notes on the voluminous writings of the eighteenth- and early nineteenth-century political economists. Working from his notebooks, Marx, during a short six-month period in 1857, compiled the Grundrisse, a 900-page manuscript not published in his lifetime. Two years later, in 1859, he summarized the main themes from the Grundrisse in a short book entitled A Contribution to the Critique of Political Economy. After another period of intensive study, eight years this time, rereading many of the same works, Marx published, in 1867, the first volume of Das Kapital, which Marx subtitled, A Critique of Political Economy.
What Marx, along with a few other observers, recognized was that a new, distinctive form of economy was taking shape, a form not recognized by the leading political economists of his day. In The Communist Manifesto ([1848] 1998), he referred to this new economic form as the ‘bourgeois’ mode of production. He continued to use this term in his 1859 book, in which he succinctly outlined his materialist interpretation of history. However, by the time he wrote Das Kapital, he had changed his terminology. Now the ‘bourgeois’ mode of production became the ‘capitalist’ mode of production, 1 and England became the ‘classic ground’ to observe capitalism in its purest form. So sure was Marx about the ‘natural laws of capitalist production’ that he wrote: ‘If … the German reader shrugs his shoulders at the condition of the English industrial and agricultural laborers, or in optimistic fashion comforts himself with the thought that in Germany things are not nearly so bad: I must plainly tell him, “De te fabula narratur!”’ 2
Between the time that Marx wrote The Communist Manifesto (with Engels) and Das Kapital, capitalism became a term denoting an economic system that is distinct from political economy. Adam Smith did not use the word ‘capitalism’ in The Wealth of Nations ([1776] 2000). Nor did he use the words ‘industrialization’, ‘industrialism’, or even ‘industrialist’. The first entry for capitalism in the Oxford English Dictionary dates from 1854. The other words begin to appear around the same time. Google’s Ngram Viewer, a database that includes over five million books, does not record any significant usage for ‘capitalism’ until a small bounce occurred in the 1860s, around the time when Das Kapital appeared. 3 In fact, none of these words conceptualizing the economy as an organized system of activities became commonplace until after the turn of the twentieth century, a full one hundred years after economic historians date the beginnings of industrialization.
The reason for the lag between the beginnings of this capitalist transformation and its recognition is obvious enough. It is difficult to put a name to a trend before the trend becomes sufficiently fixed for it to be an object of analysis. Although many observers in the opening years of the nineteenth century knew that mechanized factories marked the beginnings of something new, it was not until Marx and a few other observers writing in mid-nineteenth-century Europe first saw these new beginnings as a type of economic configuration possessing certain systemic characteristics. The early awareness on the part of a few, however, did not become common knowledge until several decades later when nineteenth-century theorizing became an easily recognized twentieth-century reality.
Marx’s writings, therefore, are all the more remarkable when we realize that he was describing an emergent economic system that would, in succeeding generations, diffuse throughout the world, a system that is linked to, but independent of, the attempts by nations to control its course. In this article, unencumbered by Marx’s theoretical determinism, we will make a similar argument: It is impossible to interpret late twentieth-century national economies without recognizing the trajectories of and processes embedded in global capitalism. Moreover, late twentieth-century capitalism is substantially different from the pre-World War II capitalism that Marx and later writers described. In distinguishing between the two capitalist eras, we will refer to the former as ‘supply-driven capitalism’ and the latter as ‘demand-responsive capitalism’. This article focuses on the emergence of demand-responsive capitalism that has occurred since the end of World War II, and on the role of East Asian economies in that development.
Marx developed his theory of capitalist transformation with relatively little assistance from the scholars of his day. By contrast, we have lots of help. For over 20 years, many scholars from a number of different disciplines have been writing about the late twentieth-century transformations of the global economy. The resulting literature is quite diverse and very detailed. In the same period, the literature on institutionalism and the varieties of capitalism has also grown and is also diverse and detailed. The issue at hand, therefore, is not exactly an empirical one, for there is plenty of evidence documenting different aspects of this transformation. Instead, the issue is largely a theoretical one, a question of interpretation and emphasis. In this article we will argue that the theoretical emphasis prioritizing the importance of the state and national institutions in the development of capitalism (or what we call here ‘theories of indigenous capitalism’) in East Asia is misplaced. This emphasis on ‘indigenous capitalism’ makes the global economy into an aggregation of national economies instead of the dynamic, rapidly changing, often competing, and increasingly borderless economic configurations that many players, including state officials, try to control for their own interests.
Capitalism in Asia: a new critique of political economy and institutionalism
Chalmers Johnson was the first to develop the concept of the ‘developmental state’. However, the idea of a strong state creating the conditions for rapid industrialization has a long history in the theories of capitalism. Marx’s contemporary, Herbert Spencer, noted the connection between Germany’s military state and economic development as early as the 1860s. Writing during World War I, Thorstein Veblen ([1915] 1954) presented a reasoned analysis linking Germany’s unusually rapid and successful late industrialization to its centralized state and strong bureaucracy. Much later, Alexander Gerschenkron (1962) and Barrington Moore (1966) made similar arguments, applying the thesis to Japan as well as to Germany. However, none of these earlier works was as influential as Johnson’s clear-minded conceptualization, first made in his (1982) book, MITI and the Japanese Miracle: The Growth of Industrial Policy, 1925–1975.
What made Johnson’s reconstruction of the strong-state theory of capitalist development so influential was not because the Japanese case fit the theory so well. As Johnson tells the story (1999: 32–60), the developmental state theory was an afterthought, a theory added to the first and last chapters of the book at the request of his editor, who wanted a ‘socko’ finish to the book (1999: 42). He had written a book about the organization and operation of the Ministry of International Trade and Industry (MITI), but the editor wanted a broad conclusion applicable to other societies as well. Johnson added the ‘take-home message’, which turned out to be the Japanese ‘model’ of development. He made this addition even though he remained dubious about the model, feeling that ‘it is hard to abstract a “model” from historical reality’. In the end, however, Johnson came to the conclusion that ‘“the developmental state” actually exists in time and space in East Asia and also exists as an abstract generalization about the essence of the East Asian examples. It is both particular and generalizable’ (1999: 42–3).
The take-home message – the developmental state model of East Asian industrialization – was not, however, distilled out of the case history of MITI. To come up with the model, Johnson read widely, and in his readings picked up and elaborated the reconstruction of Marxian theory that was then in progress. Signs of this reconstruction show up in the first and last chapters of the book. In the Introduction, Johnson (1982: 24) refuted Marx’s statement about the modern state being a committee representing the interests of the bourgeoisie, and in so doing acknowledged the link between nationalism and late industrialization. Then, in the concluding chapter, he contrasts the Japanese political system with the ‘bureaucratic-authoritarian regimes of Argentina, Brazil, Chile, and Uruguay’, regimes in which ‘the ruling elites seek to promote industrialization by excluding from power the previously mobilized economic groups and by developing collaborative relationships with multinational corporations’ (1982: 316–17). Instead, the Japanese state is democratic, and the governing coalition is not ‘hospitable to collaboration with foreign capital. The Korean developmental state, by contrast’, continues Johnson, ‘seems to share some of the bureaucratic-authoritarian characteristics and should to that extent be distinguished from the postwar Japanese case.’
In the late 1960s and throughout the 1970s, this same thesis in nearly the same words, though without explicit references to Asia, was part of an ongoing debate about how to reconstruct the Marxian theory of the state. In the waning years of the Vietnam War, the ‘New Left intellectuals’ in the US turned toward European Marxism, in particular the writings of Poulantzas, whose ‘work came as a revelation’ (Block 1987: 6). Poulantzas’s reconstruction of Marxian theory (Poulantzas 1969) led the American New Left intellectuals to ask ‘why the state in capitalist society functions as it does’ and allowed the focus to shift away from the economy ‘toward the state itself’ (Block 1987: 6).
At the center of this discussion were the early writings of several New Left intellectuals, in particular, Theda Skocpol (1973, 1979) and Peter Evans (1979), among many others. In 1973, Skocpol wrote a detailed and widely read critique of Moore’s Social Origins of Dictatorship and Democracy, which was then one of the most influential Marxian interpretations of how different types of political structures arise from different paths of capitalist development. She argued that Moore repeats the ‘fatal shortcoming of all Marxist theorizing about the role of the state … [the] propensity to explain political struggles and structures as functions of class structures and struggles’ (1973: 18). Instead, scholars needed to recognize ‘the independent roles of state organization and state elites in determining agrarian societies’ and landed upper classes’ responses to challenges posed by modernization’ (1973: 30). ‘Economic prowess’, she concludes, in the case of late industrializing societies, could come from ‘structural changes administered “from above” by state elites free from class controls’. She followed this article with her well-known (1979) book, States and Social Revolution: A Comparative Analysis of France, Russia, and China.
Even more importantly, Peter Evans carried the New Left thesis forward in a series of writings culminating in his very influential book, Dependent Development: The Alliance of Multinational, State, and Local Capital in Brazil (1979). In this book, Evans accomplished two important tasks that were crucial for the later theories of the Asian developmental state. First, he provided introductions to the writings of Fernando Cardoso (Cardoso and Faletto 1979), and Guillermo O’Donnell (1973), writings in which they developed comparative analyses of Latin American paths of capitalist development. O’Donnell’s advance of the concept of bureaucratic-authoritarian state regimes and the impact of these regimes on development was soon picked up in Johnson’s work 4 and figured prominently in Bruce Cumings’s (1984) very important article, ‘The origins and development of the Northeast Asian political economy: industrial sectors, product cycles, and political consequences’. In this article, Cumings (1984: 26–8) wrote: ‘Readers who know Latin America and especially the work of Guillermo O’Donnell will have noticed that Taiwan and Korea went through industrialization in phases that resemble the sequence in Brazil, Argentina, and other states, even though the import-substituting phase was much shorter in East Asia.’ Cumings concluded that state regimes in South Korea and Taiwan could be confidently characterized as ‘bureaucratic-authoritarian industrializing regimes (BAIR)’.
The second task that Evans accomplished was his inversion of Moore’s thesis of capitalist development by shifting the explanatory focus from typologies based on class struggle to ones based on state/business relationships, a focus that he later called ‘comparative political economy’, which is a precursor to the ‘varieties of capitalism’ theme that appeared later in the political science literature (Evans and Stephens 1988).
The most immediate reason for the success, within the academic community, of Johnson’s developmental state theory was that it was an instantly creditable theory; it just ‘rang true’. It fit an emerging explanatory paradigm. The obviousness of the theory’s validity, therefore, arose not from its application to Japan, for that part of Johnson’s thesis soon came under attack, 5 but rather from the fact that the theory seemed to prove what other political economists were trying to argue: The state has an independent role in the creation and maintenance of capitalism. At about the same time that Johnson’s book was published, many other works appeared extolling the importance of the state, the most important of which was a book edited by Evans, Rueschemeyer, and Skocpol entitled, Bringing the State Back In (1985). 6
The statist turn in the social sciences magnified the importance of the developmental state theory of East Asian capitalism. Johnson’s developmental state theory, as applied to key East Asian cases, provided the proof that the overall perspective was correct: If properly constituted and appropriately managed, the state could not only provide support for industrialization, but it could also create industrialization itself. Quickly, a new literature proliferated, a literature arguing that the best examples of the East Asian model of development were the South Korean and Taiwanese cases, with Japan, Hong Kong, and Singapore providing lesser but still relevant examples. This thesis, with some variation for particular locations, is found in Cumings (1984), Gold (1986), Deyo (1987), Winckler and Greenhalgh (1988), Amsden (1989), Wade (1990), Woo (1991), Evans (1995), Kim (1997), and Woo-Cumings (1999), as well as in scores of other articles and books. One of the more important recent books to take up this theme is Atul Kohli’s State-Directed Development, Political Power and Industrialization in the Global Periphery (2004).
A number of sociologists and other like-minded scholars (Dore 1983; Hamilton and Biggart 1988; Redding 1990; Whitley 1992) countered the state-centered interpretations of East Asian industrialization with their own accounts. These accounts stressed the importance of culture, social institutions, and history, and in this sense they are more a variation on a theme of indigenous capitalism than a genuine alternative. Nonetheless, these more sociological accounts differ from the developmental state interpretations in two important ways. First, whereas the developmental state interpretations lumped all the developing states together, making each a greater or lesser variation of the strong-state theme, the sociological versions uniformly distinguished among the developing economies, making each a separate and unique case of development. Second, the developmental state interpretation made the state the creator of industrialization, the central and most active agent that ‘governed the market’ (Wade 1990). By contrast, the sociological versions did not so much explain the origins of industrialization as they did the trajectory of development. Each case, based on a distinct set of culturally disposed actors in institutionally distinct settings, led to a differently organized form of capitalism, thus yielding, over numerous cases, a variety of capitalisms (e.g., Whitley 1992).
Although the sociological interpretations downplayed the singular importance of the state, they still shared several features with political economy interpretations. First, the narrative structure of both sets of interpretations is similar. The professed target of these interpretations was the economists’ neo-classical interpretations of Asian capitalism, and the test of their accuracy was how much better they explained the patterns of industrialization than the neo-classical alternative. Their arguments differed, however. Whereas the developmental state theorists contrasted what Johnson (1982: 17–34) called market-rational systems, such as are found in the US, with plan-rational systems, such as are found in East Asia, the sociologically inclined theorists argued that a ‘market explanation’ (Hamilton and Biggart 1988), or what Whitley (1992: 2) called ‘market determinism’, could not explain the distinctly different trajectories of industrialization.
In all of these swipes at the economists’ disciplinary paradigms, the characterizations of the neo-classical ideal of markets remained unfailingly abstract. Most writers provided general comments about how neo-classical markets are supposed to work in theory, but few writers followed those observations with a detailed analysis about how these markets actually worked in fact. The developmental state theorists are particularly negligent in this regard, for their main focus involved the state. By contrast, the sociological explanations were more focused on firms and on organizational patterns in the economy. But, even with this emphasis, there was surprisingly little actual analysis of what Marx saw as the defining feature of capitalism, ‘the conditions of production and exchange’, or of the defining feature of East Asian capitalism, export-led industrialization.
This omission appears to be analytically very curious. While analysts spent tremendous interpretative energy defining what they attributed to be the prime causal variables (i.e., bureaucratic officials, the state’s economic policies, and state/business relationships, social institutions, firm structure, patterns of authority), they more or less ignored the dependent variable, the outcome, the prevailing features of Asia’s capitalist expansion, the dynamic and rapidly changing processes and organization of production and exchange that center on the organizational links between Asian manufacturers and global retailers and brand name merchandisers.
It is easy to criticize these theorists in retrospect for their failure. In fact, in the 1970s and 1980s, it was difficult, if not impossible, for anyone to define the form of capitalism that was then emerging in Asia. Many authors described the facts of trade, the rates of growth, the organization of firms, the role of multinationals; all the evidence was there before them, but the trees obscured the forest. The general patterns could not be recognized for two reasons. First, there were a number of alternative theories (e.g., Wallerstein’s (1974) ‘world systems theory’, Frank’s (1967) ‘dependent development theory’, or Braudel’s (1981–82) theory of capitalism’s ‘longue durée’) available to explain the big picture, but none of those theories seemed to work. None of these theories squared with what seemed to be happening in Asia in the 1970s and 1980s, but no one offered an equally general theory to counter those theories. Second, and much more importantly, none of the players themselves knew exactly what was happening either. Even business people were unsure what the trends were and where those trends were leading them. 7 Similar to the development of the first age of capitalism in the nineteenth century, the rise of post-World War II capitalism in Asia was truly an emergent phenomenon and very difficult to distinguish from earlier periods of capitalist development in Europe, Japan, and the US. Quite logically, most writers just examined the process of industrialization country by country, concentrating on the internal causes and organizational patterns of economic expansion, and for that limited goal, the developmental state theory or the sociological alternative seemed to many to be the most viable theories – ‘the state willed it so’ or ‘the institutional arrangements made it happen that way’.
As satisfying as their approaches appeared at the time, most theorists, in retrospect, made three methodological errors. The first error is that they assumed as true from the outset the thesis they were trying to prove; most of them never seriously considered a viable alternative. Instead, they offered ‘straw men’ neo-classical alternatives that were easily knocked over. As Veblen (1898) noted long ago and as many economists (Nelson and Winter 1982; Kaldor 1985, 1996; Hodgson 1996; Setterfield 1998, 2013) have repeated since then, neo-classical economics, with its inherent equilibrium bias and its static assumptions about human nature, cannot readily explain transformative changes. More importantly, the neo-classical economic concept of ‘market’ is a formal term, which is abstractly and quantitatively conceived so that it needs no exact empirical definition. The conventional economist’s idea of a market thus exists on an analytic plane that is quite different from the analytic plane needed for an explanation of East Asian industrialization. In both respects, this neo-classical conception was never a viable alternative to explain Asia’s rapid industrialization.
The second error is an almost exclusive focus on the independent variables, the imputed causes of development – either the political economy or the institutional environment – and a superficial and underspecified analysis of the dependent variables – the assumed outcomes, the processes and products of industrialization itself, that is, capitalism by another name. This concentration on the independent variable leads to convoluted explanations about what causes differences in the process of industrialization from place to place. We might call this methodology a ‘Goldilocks’ approach. Defining the outcomes as differential rates of growth, Evans (1995) and Kohli (2004) line up selected case studies by the reputed strength of ‘state-directed development’. Some states are too predatory (Evans 1995) or neo-patrimonial (Kohli 2004); other states are too fragmented-multiclass (Kohli 2004) or lack state capacity (Evans 1995); and yet others are just right, are cohesive-capitalist states (Kohli 2004) or have the right amount of ‘embedded autonomy’ (Evans 1995). The necessary and sufficient causes (variation in state structures) are so classified based on the presumed but under- or unspecified outcomes. Accordingly, in cases where supposedly little or no development occurs, such as Zaire (Evans 1995) or Nigeria (Kohli 2004), the cause is obviously due to the state’s predatory and neo-patrimonial characteristics. In cases of rapid, successful development, such as South Korea (Evans 1995; Kohli 2004), the cause is just as obviously due to the state having the right amount of cohesiveness and embedded autonomy. Between these two extremes are the ‘intermediate’ states (Evans 1995) that lack some characteristic that the more successful cases possess. This ‘pseudo-sampling’ rests on the assumption that the imputed cause is correct from the outset and does not need to be proven. The entire attempt to present a series of cases is simply an illustrious demonstration of an obvious thesis, or what some social scientists (Bates et al. 1998) call an ‘analytic narrative’, which is an account in which the presumed proof of a thesis is embedded in the provided story.
Those advocating the sociological alternatives fared somewhat better than the political economists, largely because their independent variables are more closely involved with the organization of businesses than were the variables favored by political economists. For instance, Hamilton and Biggart (1988) concentrated on the differences among business networks and Whitley (1992) on the differences among business systems in East Asia, and both explained the observed economic configurations in terms of a variety of contextual institutional factors. Implicitly, in these and other similar studies, the organization of business is connected to the rapidly industrializing economy, but the link between the two is unspecified and largely assumed without showing how a more or less static set of institutional variables could result in such a dynamic pattern of economic growth. Instead their reasoning led to the conclusion that different institutional environments resulted in different types of economies, and from here it was a short leap to the varieties of capitalism perspective, an emerging viewpoint that was largely in place before the Hall and Soskice volume appeared in 2001.
The third error that both sets of analysts made, in retrospect, was to give local causes for what is a general occurrence. 8 By the 1990s, it was becoming clear that industrialization in East Asia was not just an East Asian phenomenon, but rather was part of a general global phenomenon, whose causes could not be isolated in specific developing locales. Starting soon after World War II, some locations in Asia, and usually multiple locations, experienced extremely rapid economic growth. The economic expansion started in Japan and Hong Kong about the same time, and subsequently diffused throughout the region, with the less developed countries, starting from a lower base, expanding at a faster rate than the more developed ones. Recognizing this pattern of growth, we should see that Asian industrialization is definitely not a zero-sum game. Almost every country in East Asia (except for North Korea) and most countries in South and Southeast Asia have industrialized to a significant extent, and this industrialization continues to deepen. The success of one economy has not meant the failure of another, though in the case of Japan this statement needs qualification. 9 The rise of capitalism in Asia is a regional and global phenomenon, rather than a country-by-country phenomenon; it is one case of capitalism, instead of a variety of individual cases.
Instead of recognizing a general pattern of diffusion, both sets of analysts concentrated on specific variables in each location to explain the patterns of development, some arriving at an ‘Asian model’ or later a ‘Chinese model’ of development. As significant as these variables might be, they were still insufficient to explain the rapid pan-Asian changes that occurred. General patterns of change need equally general sets of causes that drive those changes, and if we are dealing with one case of capitalism, we do not need a strictly Asian model of development. What we need, instead, is a global history, a developmental and inclusive account of the diffusion of capitalism to Asia, an account that stresses ‘cumulative causation’ and ‘increasing returns’, as well as the key features of earlier social science explanations. In the next section, using diverse sets of studies, we will outline the key elements of such a developmental history.
The rise of demand-responsive economies: a developmental history
We will start this history with the point on which all analysts agree: In the early period of rapid growth, roughly from 1965 to 1985, the economies of South Korea, Taiwan, Hong Kong, and Singapore were truly export-led. As exports of finished goods increased, so too did imports, but imports at the time were mostly intermediate inputs for goods that would later be exported or were capital goods, such as machine tools, which would be used to make finished goods. Much of the domestic capital went into building infrastructure, constructing factories, and establishing subsidiary businesses. In this sense, export production shaped the entire economy.
Exports and the retail revolution
The first question we should address, therefore, is what were these exports? If we take exports to the US as representative of all exports from these countries, 10 then this question can be answered in detail and with certainty. Feenstra and Hamilton (2006: 238–52, 301–41) have examined these exports to the US from Taiwan and South Korea. What they found for the entire period from 1972 until 2000 (which are the first and last years of the US Customs database they used) was, first, that the goods were overwhelmingly manufactured goods; second, that even in the first years of the record (1972–75) both countries exported thousands of distinct products (as classified in the seven-digit TSUSA system of the US Customs Service); third, that most of the values of these goods, year by year, were concentrated in the top ten, and overwhelmingly in the top one hundred, exported goods (as defined at the seven-digit level); and, fourth, the composition of the top products changed, sometimes drastically, year by year. Based on the export data, as well as on additional primary and secondary sources, their conclusion concerning these exports was, first, that the products had been manufactured in locally owned factories and not in factories owned by multinational Western or Japanese firms, and, second, that Asian manufacturers produced these goods on contract to be sold under the buyers’ brand names. The conclusion applies to exports from Taiwan for the entire period (1972–2000) and from South Korea for the 1970–88 period. After 1985 or so, when South Korean chaebol began in earnest to build their own brand names, South Korean factories still produced goods on order, but less frequently as contract manufacturers.
Based on these realizations, the question that Feenstra and Hamilton (2006; also see Hamilton 2006) ask is, what, on the demand side, caused this sudden growth of contract manufacturing? They show, first, that, before 1965, US imports provided a very small proportion (always less than 10 percent, and usually less than 5 percent) of the market share for consumer goods. In other words, what little contract manufacturing occurred in the US happened in factories inside that country. Moreover, because of Fair Trade laws enforced in the US before 1965, 11 one can conclude that contract manufacturing contributed relatively little to the US economy.
Second, Feenstra and Hamilton (2006) trace the growth of contract manufacturing to what they call the ‘retail revolution’ that began in the late 1950s in the US and that quickly expanded into a worldwide phenomenon in the decades after 1970. 12 In 1955, there were only around 500 shopping centers throughout the US, but by 1964 there were over 7,600. Today, fifty years later, there are over 50,000 shopping centers in the US alone and tens of thousands more scattered around the world, a few of which are larger than anything in the US. 13
With reference to Asian manufacturing, the most important aspect about these shopping centers is that they fostered the development of chain stores. Consider the fact that, before 1965, the retail trade sectors of all national economies were decentralized, fragmented, and mostly locally owned. At the same time, as Chandler and colleagues (1977, 1990; Chandler and Daems 1980) show, manufacturing industries had centralized their production facilities, vertically integrated a portion of their inputs, and distributed their products nationally and often internationally. In the case of major consumer goods, these corporate manufacturers developed distribution channels (often via wholesalers) that they controlled directly or indirectly. Simply put, high levels of concentration are a lot easier to accomplish in manufacturing sectors than in retailing. The reason for this difficulty is because retailers would need to deal with hundreds, even thousands of suppliers for the goods they sell, and if they tried to establish branches, they would also have to deal with different sets of inventories for different sets of customers in different locations – all of which would be a logistics nightmare. Therefore, retailing remained at low levels of concentration until such time as transportation and communication technologies had reached a point that retail entrepreneurs could harness those technologies to reorganize the retail sectors of national economies (Abernathy and Volpe 2011).
That point occurred in the 1960 and 1970s, when a convergence of factors accelerated qualitative changes in the organization of global capitalism, promoting a shift from supply-driven to demand-responsive industrialization. 14 The shopping center boom, the end of Fair Trade Laws, the beginning of discount merchandising, the completion of an interstate highway system, the development and standardization of container shipping, the switch from rail to trucking as the main mode of distribution within the country, the establishment of Uniform Product Codes, the development of barcodes and scanning devices, the beginnings of computerized point-of-sales inventory systems, and the start of widespread use of credit cards – all of these technological developments occurred in the US during these two decades. These shifts in technology occurred along with a sea change in consumer behavior, including women joining the workforce in large numbers and the emergence of a youth culture. This convergence of technology and new consumer life-styles encouraged the development of chain stores – much needed to line the aisles in the new shopping malls – that offered differentiated products targeting special consumer niches: working women, pre-teens, adolescents and young adults, sports enthusiasts, and so forth. All these foundations of the retail revolution were in place in the US before 1980, but in the decades after 1980s these innovations, merging with improved forms of high technology in communications and transportation, spread throughout the world. Today, the retail sectors in most developed economies are, by contrast to the earlier period, much more highly concentrated and, with the introduction of internet retailing, still rapidly evolving. 15
The third point that Feenstra and Hamilton (2006: 169–298) make is that the industrialization of East Asia is directly connected to, and an integral part of, the retail revolution. At first, Japanese trading companies played a central role by obtaining contracts from Western retailers and developing suppliers that could supply the ordered goods (Kojima and Ozawa 1984; Hamilton and Kao, forthcoming). In many cases, these suppliers were first-time factory owners whom the Japanese firms had to finance, train, and supply. This all-inclusive arrangement did not last long, because by the early 1970s US retailers had begun to establish buying offices in East Asia, where they soon began to purchase a large percentage of their total inventories (Gereffi and Pan 1994). These buying offices, as had the Japanese trading companies before them, worked with local trading companies in both Taiwan and South Korea to source the products they needed. 16 The expansion and success of an increasing number of newly established general retailers (such as Wal-Mart, Target, K-Mart), specialty retailers (such as ToysRUs, Gap, The Limited, Home Depot, and Best Buy), and factory-less brand name merchandisers (such as Nike, Ralph Lauren, Anne Klein, Dell Computers) led to the expansion and increasing success of Asian manufacturers. This reciprocal growth is captured in the import statistics, as well as in the rapid rate of growth experienced by these firms in the US and (for some) internationally.
Based on their analysis, Feenstra and Hamilton argue that East Asian industrialization was and continues to be demand-led, but not by final demand created by consumers, but rather by intermediate demand created by retailers and merchandisers in anticipation of final demand. In a felicitous phrase, Gary Gereffi (1994) calls these intermediate agents ‘big buyers’. What the retail revolution accomplished is to close the gap between retailer and customer by using point-of-sales information to determine what specific products big buyers will next order from their suppliers. This information gave retailers and factory-less merchandisers considerable leverage over would-be suppliers, and it further encouraged most big buyers to order goods on contract and to closely supervise the manufacturing process. The economic power of these big buyers made manufacturers (i.e., their suppliers) into an organizational extension of retailing. A number of analysts have identified the resulting transformative configurations, variously, as ‘global value chains’ (e.g., Gereffi et al. 2005; Cattaneo et al. 2010) or ‘global production networks’ (e.g., Coe, Hess, and Dicken 2008; Yeung and Coe 2014).
Now we come to two critical issues: first, what are the effects of intermediate demand on East Asian economies, and, second, what roles do the state and social institutions play in the process of industrialization? We will take each of these questions in turn.
The effects of demand on industrial organization
Political economists and conventional neo-classical economists are biased toward supply-side explanations of export growth, and, therefore, most of them fail to theorize the effects of demand on economic growth. 17 Instead, they analyze factors relating to production, which for political economists would include the actions of the state. However, the factors relating to demand are just as important to consider. Both sets of activities are organized and intertwined with each other. Moreover, in examining the industrialization of East Asia, we would argue that the organization of demand substantially shapes the organization of production.
To illustrate the effects of demand on production, we will narrow our focus to Taiwan and South Korea in the first decades of industrialization. The export manufacturing sectors of these economies mostly consist of firms producing goods that big buyers had ordered. Big buyers specify the terms of their purchase relative to price, quantity, and quality. From the big buyers’ point of view, Asian manufacturers constituted a supplier market for specific goods. 18 In the early 1980s, for instance, there were many firms located in both countries where big buyers could order shoes that the buyers designed and specified. In a very real sense, there was an Asian (and perhaps even a global) price-setting supplier market for shoes, for which the big buyers acted as market-makers. Big buyers set the terms for the exchange in full recognition of the competition for selling shoes that they faced in the consumer markets in their home territory. In entering into this exchange, big buyers recognized that different firms in different locations had different capabilities. Some big buyers, such Nike, needed a range of different types of shoes, some designed for the mass market and sold in Wal-Mart, and others designed for special activities (like soccer or long-distant running) and sold by specialty retailers. Faced with ordering such differentiated products, some big buyers, such as Nike, split their orders, with a portion going to factories that specialized in large runs of inexpensive shoes and another portion going to factories producing smaller quantities of high quality shoes. 19 Other big buyers of shoes would specialize in only one side of this consumer market. For instance, a big buyer specializing in women’s dress shoes, which are fashion products, would always favor factories able to deliver small runs of distinctly designed shoes very quickly.
Given such diversified orders for products within the same three- or four-digit SIC classification, it is not surprising that the profile of exports from the East Asian countries looks similar. However, a closer examination at the seven-digit level reveals that different countries produced different types of goods, even goods within the same general classification (Feenstra and Hamilton 2006: 240–52; Hamilton 2006: 146–83). Examined over time, this trend becomes clearer. The economies diverged not only in what they produced, but also how they produced those goods. The process of ordering goods served as a driver of this divergence.
Now examining this situation from the suppliers’ points of view, we should recognize that manufacturers faced real constraints in terms of which products they could reasonably manufacture within the buyers’ specifications. In part, the constraints were path-dependent, were in place at the outset of export production, and became more clearly defined over time. In the 1960s, the East Asian economies differed in terms of which groups in society could best take advantage of new economic opportunities. For a variety of reasons, the South Korean economy was centered in suburban areas around Seoul and was concentrated in business groups (i.e., chaebol) from the very first. As industrialization moved forward, the largest of these chaebol were in a position to grasp most of the new opportunities. By contrast, and also for a variety of reasons, the rural sector of the Taiwanese economy had the resources and the connections to take advantage of the new opportunities. (We will discuss some of these factors below. Also see Feenstra and Hamilton 2006: 169–211.) In both cases, initially, Japanese trading companies, which heavily invested in both South Korea and Taiwan, played a decisive role in developing which firms in which locations were able to obtain which orders. Because Korean factories were larger and more ‘Fordist’ in orientation than those in Taiwan, they received more orders requiring techniques of mass production. Taiwanese factories, however, were mostly located in the countryside, were small and organized through satellite assembly systems that allowed different small firms to produce different parts of the final product. These inter-firm networks could work flexibly, but were not well suited to economy-of-scale mass production; these networks landed many of those orders requiring small batches of distinctive products.
As orders rapidly accelerated into both economies, two important aspects of industrialization occurred. First, the factories in both economies began to specialize in those products that they could best produce and to rationalize the actual system of production to gain greater efficiencies. These measures, in turn, rationalized the ordering system used by big buyers, so that they could more easily select the right location, if not always the right firm, to place their order. This cycle of rationalization for both suppliers and buyers resulted in what economists call increasing returns to each location for that specific system of production (Helpman and Krugman 1985; Arthur 1994). In the 20 years between 1966 and 1986, a period during which both economies boomed, ‘the average firm size [in South Korea] jumped by 300 percent and its firms grew in number by only 10 percent’. In the same period in Taiwan, ‘the number of reported firms increased by 315 percent and the average firm size expanded 15 percent’ (Biggs 1988: 3–4). Accelerating orders drove the process of industrial organization.
Second, during the same 20 years, both economies began to ‘fill up’ with new firms that supplied the needed inputs and services that enhanced, and made money out of, the export sector. In these 20 years, a ‘self-organizing’ demand-responsive economy emerged. 20 In South Korea, the small numbers of new firms were mostly established inside the chaebol. These firms were large and vertically integrated into the top chaebols’ expanding system of export production. In keeping with the patterns of demand-responsive industrialization, the chaebol established the final assembly firms for their main exports first and the firms producing the essential components and the services for those consumer products later. Earlier, in the initial phase of production, the chaebol typically imported the essential components and outsourced the needed services, but later they filled in their ‘one-set’ production systems with firms producing intermediate inputs and services.
By contrast, a process opposite to vertical integration occurred in Taiwan. As more orders flowed in, inter-firm networks proliferated, as did the number of firms associated with those networks. Everyone wanted to start their own firms and be their own laoban (boss), and everyone had their own connections to family and friends who helped them do so. 21 By 1986, around 700,000 firms had registered with the government, which works out to be one firm for every 15 people in Taiwan (Chang 1988: 10). Over 40,000 of these firms were registered as trading companies. As the output of these networks of small and medium-sized firms rose, a considerable number of larger firms were established to supply the upstream inputs and services that the smaller firms needed, but these firms did not vertically integrate either upstream or downstream. Despite the differences in industrial organization between South Korea and Taiwan, the ultimate outcome was similar: an increasingly organized demand-responsive economy emerged.
More than a few political economists have argued that state officials were instrumental in creating the industrial structure in South Korea (Amsden 1989; Woo [Woo-Cumings] 1991, 1999; Evans 1995) and in Taiwan (Amsden 1985; Gold 1986; Wade 1990) that led to industrialization. Feenstra and Hamilton (2006) put forward the counter-argument that the demand structure and not the state played the decisive role in shaping the emerging economic organization of both economies. They tested their alternative thesis by developing an economic model based on the market power of firms in determining whether it is more ‘rational’ to buy inputs from another firm or supply those inputs internally to the firms. This is a question of whether to vertically integrate or not. The mathematical solution to the model reveals a multiple equilibrium: In those ‘high concentration’ economies where there are only a few big groups at the outset of industrialization (such as South Korea), the ‘rational’ course of action is not to buy from one’s competitors, because they will raise the cost of inputs in order to give themselves an economic advantage. It makes sense to vertically integrate by making all essential components yourself. By contrast, in ‘low concentration’ economies, where there are many players (such as Taiwan), it makes sense to sell your intermediate products to all possible buyers. It is not an ‘efficient’ outcome to vertically integrate because more money can be made by selling key components than is possible by making a final product with those components, simply because the competition downstream is so intense. The gold rush analogy applies here: When miners rush to the gold fields, the player who makes the most money is the one who sells shovels to all the miners.
The mathematical solution to this highly stylized model led to some testable hypotheses. A ‘high concentration’ economy is only stable at equilibrium when there are very few groups at the top. As more groups are added to this economy, the high concentration solution loses its stability. By contrast, the ‘low concentration’ economy is a stable solution as more firms are added to the mix. Using a database of internal transactions for all the top business groups in both economies, Feenstra and Hamilton (2006: 78, 120) demonstrate that the actual measures of vertical integration in both economies are consistent with the model, and, moreover, that the pattern of bankruptcies during the 1997–98 financial crisis in South Korea is also consistent with the predictions of the model. Therefore, they conclude that it is ‘objectively possible’ (Ringer 1997) that the organization of demand alone, without any actions on the part the state, could have produced the observed outcome in terms of industrial organization.
In summary, primary data from the US Customs and from the internal transactions of business groups in both economies, as well as a considerable amount of secondary material, support the thesis that industrialization in East Asia is demand-led and not supply-driven. In substantive terms, demand-responsive industrialization arises from the mutually causative relationship between global buyers and Asian manufacturers.
The effects of demand-responsive industrialization on the state and on social institutions
Given that the above thesis is correct, what then is the relationship between East Asian governments and local social institutions, on the one hand, and the processes of development and the organization of the economy, on the other? Most analysts cast the state and social institutions as the independent causes of industrial outcomes, all this in a framework that is remarkably static, even though the economies themselves are rapidly changing. By contrast, our analysis suggests that the changing organizations of demand (e.g., accelerating orders, the development of global logistics and communication technologies, the global diffusion of shopping malls and chain stores, and the spread of consumer-oriented societies) are also more accurately seen as the driving force behind state actions, the state’s economic policy, and even emerging social institutions. Of course, we acknowledge reciprocal causation between the global economy and local political and social institutions, but the previous accounts of industrialization have been so one-sided that it makes sense to stress the other side, the effects of demand-responsive industrialization on state actions and social institutions. Once we conceptualize the reciprocal effects of each on the other, we also enter a framework of ‘circular cumulative causation’, in which continual reciprocal effects of different economic configurations produce a developmental sequence. 22
At the outset of industrialization, East Asian societies were socially and politically different from each other, but for many analysts the differences did not appear to be that great. Accordingly, most stressed the similarities: strong states, Confucian cultures, high levels of literacy, and hard-working, family-oriented populations. All these were institutions that supposedly had effects on industrialization. Although we would also agree that these institutions did play a role, we believe, in addition, that these institutions were selectively invigorated by, and thus were partly an outcome of, rapid economic growth stemming from increasing demand, as well as the changing organization that went with that demand.
For instance, the inter-firm networks in Taiwan certainly build on top of the social etiquette prescribed in reciprocal relationships, often identified as guanxi relationships. And the patrilocal, patrilineal patterns of partible inheritance in Taiwan certainly had effects on dispersed patterns of family ownership and the divided succession of firms. Equally, the patterns of patriarchy and partial primogeniture had effects on the ability of South Korean entrepreneurs to create large vertically integrated chaebol that could be passed intact from one generation to another. But once the two economies began to industrialize rapidly, the very complex but very different and diverging economic organizations of the two economies can only be explained by the demand-driven economic processes outlined above. These very processes, in turn, fed back on the business people who could see the importance of these patterns in the conduct of their businesses and reinforced them. In this sense, these organizational and institutional patterns should be conceptualized as reflexive in nature. In reference to East Asian economies, the organization of these economies did not precede the activities that set them in motion. Things did not have to turn out as they did. The organization of the economy in North Korea is not like that of South Korea; Taiwan’s economy is not organized like that of China. People may draw on a cultural repertoire (Swidler 1986), but the institutionalization of a specific pattern of action comes out of the activity that gives that pattern its social significance to those involved.
Much the same reasoning can be applied to the state. We can show, in both cases, that the state did not purposefully set the parameters for economic growth. At some point along the way, political leaders, such as Park Chung Hee, forcefully encouraged the growth of an export-led economy, but the group of business leaders in Seoul was already small when he arrested them and made them promise to support the state. 23 Moreover, most of those making the promise went bankrupt or left their business or never grew their firms beyond a modestly sized business group – so much for the state picking champions (Feenstra and Hamilton 2006: 198–9).
Rather than viewing the state as driving the economy, it makes more sense to see rapid economic growth as a force driving the state to make decisions about how to channel and make money out of industrialization. In other words, the South Korean and Taiwanese states did not lead economic development but rather followed the lead of the newly emerging patterns of global capitalism. Take one famous example that political economists universally cite as evidence for a developmental state in South Korea, the Big Push, which was President Park’s plan, in 1973, to make heavy and chemical industries (HCI) the foundation for Korean industrialization. The small group of officials at the top of Park’s centralized regime worked out the Big Push with the top chaebol owners ‘without much consultation with the technocrats in the Economic Planning Board’ (Kim 1997: 148–9). What this plan produced was a way to establish upstream firms inside the leading chaebol to produce inputs that they needed for their export products. What is touted as the best evidence for the developmental state in South Korea is, in fact, an integral part of demand-responsive industrialization. This reversal in interpretation is an important point. The better way to describe the developmental states in East Asia is not argue that the ‘state governs the market’ (Wade 1990), but rather that state officials try to ‘buy into’ a rapidly changing global economy in such a way that they can further their own political and economic interests.
Conclusion
Marx’s analysis of capitalism centered on factory production and on the relations between owners of production and the laborers hired to do the work, and for the century after Das Kapital appeared, this bias toward production seemed to be a reasonable starting point for a theoretical description of capitalism. Indeed, it was during these hundred years, roughly from 1870 to 1970, that large businesses grew into oligopolies controlling most of the market shares across all the major sectors in the world’s most advanced national economies. Because this production bias privileged the geographical and institutional locations of these huge corporations, political economy interpretations of national economies often became the leading theories of capitalism as well.
This article argues, however, that the older form of production- and nation-based capitalism is a thing of the past. New and more global forms of production, distribution, and exchange have emerged that have substantially changed the tone and tenor of capitalism today. In this article, we focused on East Asia, but today all economically engaged countries around the world are increasingly demand-responsive. Demand-responsive capitalism is not merely a global trend, but is, more accurately, a lived-in reality. In the past 25 years, the retail sectors of nearly every developing and developed society have been transformed (Shin and Hamilton 2014). Retailers now rank among the world’s largest firms (Deloitte); shopping malls are ubiquitous; internet shopping is creating global markets for all kinds of products; and the shelves of most stores, regardless of the products being sold, are stocked with goods that are globally sourced. Once global markets become routinized, then everyone can participate. People around the world have become consumers, buyers of global products that retail stores of every shape and size have bought for them.
We can no longer think of capitalism as Marx did, as a troubled system of production built on the backs of the proletariat whose surplus labor is squeezed out of them by profit-seeking capitalists. Equally, we can no longer think of national economies, as Adam Smith did, as being autonomous expressions of a market economy shaped by a political umbrella. Our analysis needs to reflect the world we live in, a rapidly changing world that has been, and continues to be, transformed by what Weber called ‘the most fateful force of our modern life, capitalism’, a force characterized today, as it was in Weber’s time, by continual advances in calculation and control.
Footnotes
Acknowledgements
The first draft of this article was presented at the workshop for ‘Changing Asian Business Systems: Globalization, Socio-Political Change, and Economic Organization’, Manchester Business School. We want to thank Richard Whitley and the other participants for their comments. The mistakes are ours alone.