Cluster-Based Industrial Development

How to develop industries was the main issue in development economics in the 1950s and 1960s, when this new branch of economics was being established. For example, criticizing Rosenstein-Rodan’s (1943) influential “Big Push” or “Balanced Growth” theory, Hirschman (1958) proposed the strategy of focused or “unbalanced” industrialization based on his pioneering analysis of forward and backward linkages among industries. Lewis (1954) and later Fei and Ranis (1964) formulated a model of industrialization which is based upon the transfer of labor force from agriculture to industries. As a comprehensive review of the early literature on industrialization by Sutcliffe (1971) clearly attests to, the issue of industrialization was heatedly and widely debated among a large number of economists and other social scientists. The excitement about this issue, however, quickly faded away in the 1970s, with a major exception being the issue of the appropriate technology or the choice of optimum technology by developing countries from the “shelf” of available technologies in developed countries (Stewart, 1978; White, 1978). This issue, too, lost momentum shortly thereafter. As a result, the discussion of industrialization occupies only a small space in modern textbooks of development economics.¹

As discussed in Chapters 1 and 2, it has been increasingly recognized that industrial clusters play a critical role in the development of industries. The case of Japan is particularly noteworthy, as the fine division of labor among a large number of small- and medium-size assemblers and part-suppliers located in small industrial districts is regarded as a major characteristic of the post-war industrialization (Asanuma, 1989; Kawasaki and Macmillan, 1987; Whittaker, 1997). Yet, according to the statistical analyses of changes in industrial locations in Japan (Fujita and Tabuchi, et al., 1997; Mano and Otsuka, 2000), manufacturing industries tend to disperse geographically over time, rather than concentrate in certain areas. The same tendency is found in the US (Glaeser et al., 1992; Henderson 1995; Kim, 1995). Does this imply that industrial clusters are dissolved rather than strengthened in the process of industrial development?

We undertake a comparative study of the development of the garment clusters in China and Japan in this chapter. The cluster in China produces children’s clothes and is located in Jili (or Zhili) in the city of Huzhou in Zhejiang province (see Figure 3.3). The Japanese cluster produces working clothes and is located in Shin-ichi town in the Bingo area, which is the traditional regional name of the whole neighborhood of Fukuyama city in Hiroshima prefecture (see Figure 4.1). The case of Jili is interesting not only because it has grown rapidly but also because its development pattern is so similar to that of the garment cluster in Wenzhou, which is known to be born in the “dirt floor of poor farmer’s house.”¹ As will be discussed in Chapter 6, Wenzhou is well-known for its successful cluster-based development led by private enterprises, dubbed the “Wenzhou model” of industrial development. While the garment cluster in Wenzhou had already developed fast in the 1980s, Jili’s cluster developed throughout the 1990s, thus making it easier to collect quantitative information on the relatively early phase of the development process from enterprises in Jili. On the other hand, we chose by chance the working clothes cluster in Bingo among several large garment clusters in Japan.

In the previous chapter we examined the hypothesis that the development of market transactions facilitated the entry of new enterprises, thereby expanding the quantity of production, which, in turn, induced the quality improvement by reducing the profitability of producing low-quality products. Since marketing is the key to success in the garment industry, former merchants played the role of entrepreneurs in the transition from the quantity expansion to the quality improvement phases. In the motorcycle industry, as well as the machinery industry to be analyzed in the next chapter, it is often engineers who initiate a new business and undertake innovations, as production is technically more complicated than garment production. In this chapter, we would like to explore first, whether clusters are indeed important for the development of the motorcycle industry and second, whether the evolutionary processes of industrial development, consisting of the initiation, quantity expansion, and quality improvement, take place in this engineer-led industry.

As was discussed in Chapter 1, industrial clusters have received attention from economic geographers. Since their main interests lie in the spatial dimensions of industrialization, the dynamic process of industrial development that takes place inside the clusters over time is not the main theme of their studies. While there are a group of economists who are interested in the role of industrial clusters in developing countries, their analyses do not generally cover the long-term development process of industrial clusters. One of the major themes of this study is to demonstrate that there are economic forces, which are endogenously generated in the process of cluster-based industrial development, that expand the opportunity to innovate, thereby accelerating the growth of the industry.

Regarding what we call information (or knowledge) spillovers today, Marshall (1920, p. 271) pointed out nearly a century ago that “the mysteries of the trade become no mystery” in industrial clusters, which means that important business knowledge or know-how is easily diffused in an industrial cluster. Consistently, we have found in Chapters 5 and 6 that early followers are almost all former employees who learned the mysteries of the trade by working at the pioneering enterprises. It is only after these spin-off enterprises have formed an industrial cluster that other types of entrepreneurs with different occupational backgrounds enter the industry.

We firmly believe that the compilation of careful empirical studies is imperative to answer the questions of how industries develop and how industries can be promoted in developing countries. Motivated by this belief, we visited a number of enterprises and conducted personal interviews with entrepreneurs in many industrial clusters in Taiwan, China and Japan. While doing so, we became aware that the different industries in the different countries have followed surprisingly similar development processes. Similarities are striking especially in the manner in which the quality improvement phase arises out of the quantity expansion phase. There are, of course, dissimilarities among industries, such as merchant-led and engineer-led development patterns, but even the dissimilarities display regularity (see Table 2.2).