Nakizito & Darkoh/lNCIPIENT INDUSTRIALIZATION Developing countries and incipient industrialization: a case study of enterprise clustering and inter- firm relations in small and medium-scale manufacturing industries in Lobatse, Botswana J. NAKIZITO AND M.B.K. DARKOH1 ABSTRACT Botswana s small and large towns offer good examples of incipient industrialization and enterprise clustering in a developing economy. Using data from Lobatse, a small industrial centre in Botswana, this brief paper shows that clustering in developing countries does not induce high inter-jirm relationships as is the case in necessarily industrialized countries. Relations are limited among firms within clusters. Introduction THE SMALL OR MEDIUM-SCALE SUBSECTOR is now widely recognized by governments, businesses and the academic community as playing a vital role in the development of an economy. There isrecent international literature suggesting that when these firms operate in clusters, it may help them overcome many growth constraints by promoting collective efficiency, facilitating the growth of firms in small steps and making it easier to respond to opportunities and crises (McCormick 1999). Much of the literature is on developed countries and, more recently, also on Latin America and Asia. There is very little material on Africa. I Both in the Department of Environmental Science, University of Botswana, Gaborone, Botswana. VOL 17 NO 1 JANUARY 2002 JOURNAL OF SOCIAL DEVELOPMENT IN AFRICA 19 The term "cluster" has been used in different ways in the literature relating to industrial development. Porter (1990), for example, uses the term as a means of identifying a group of firms that are engaged in related or similar activities within a national economy. To Porter, the relationships within an industry cluster benefit from firms being located near one another, but he does not emphasize geographical proximity as a defining characteristic of clusters. Schmitz (1992), however, makes an explicitly geographic use of the term, defining clusters as geographic and sectoral agglomeration of enterprises. McCormick (1999) notes that the latter approach is rooted in Marshall's (1890) observation on textile and metal industries in the nineteenth century that economic gains could be realized when small firms within a specific industrial activity clustered in close proximity to each other. In such a setting, individual firms could specialize in a particular stage of the production process. It is this specialized pool available to all firms within the cultural and social value that constitutes what Marshall termed the local "Industrial Atmosphere" (Marshall 1919). Clustering brings gains that are not likely to be attained by individual firms if they were located in isolation. It is the concept of collective efficiency, defined as the competitive advantage derived from local external economies and joint action (Schmitz 1995), which helps the firms to capture these gains, though such gains may not necessarily result from clustering. The concept can be examined from two angles, that is, the vertical and horizontal inter-firm relations. If these relation- ships are highly developed collective efficiency is likely. The existence of clusters could be a result of a number of factors that include spontaneity, urban zoning regulations, entrepreneurship or forced agglomeration. The absence of a location policy in Botswana has greatly affected small enterprises in such a way that some firms may be forced to agglomerate in spaces within urban centres, some of which may not even meet the necessary conditions for firm location. Whatever is the initial cause of clustering may not be important as the processes and forms of interaction existing in the clusters. Such processes could be firm relations, geographical proximity, entrepreneurs' mutual support 20 JOURNAL OF SOCIAL DEVELOPMENT IN AFRICA VOL 17 NO 1 JANUARY 2002 Nakizito & Darkoh/lNCIPIENT INDUSTRIALIZATION and the emergence of auxiliary services to meet the needs of the businesses in the agglomeration of firms. The major objective of this paper is to find out whether there are any industrial clusters in Lobatse and, ifthere are, to determine the existing inter-firm relations within the clusters. For the purposes of this investigation, small- and medium-scale industries are regarded as any units engaging in processing and manufacturing, found in the informal and the formal sector employing less than one hundred persons. This definition is basically the same as that proposed by the Small, Medium and Micro Enterprises (SMMEs) Task Force of Botswana, (Republic of Botswana 1999). The study's general hypothesis is that there are industrial clusters in Lobatse that result in high inter-firmrelations. For operational the study adopts Schmitz's definition of that is, the spatial agglomeration of small and medium-scale clustering, manufacturing enterprises. As McCormick (1999: 1532) explains, Schmitz's approach to clustering is more applicable inregionsof incipient like Africa where poor infrastructure, weak inform- industrialization ation systems and cultures that place high value on face-to-face communication are the norm. purposes, to study clustering of small enterprise development Review ofthe literature in Attempts developing countries are a relatively recent development that does not go beyond the late 1980s. The few studies that have been carried out in this field cover a number of developing economies like Brazil (Schmitz 1993), Ghana (Dawson 1992), India(Kashyap 1992),Indonesia(Weijland 1992), Kenya (McCormick 1993, 1997, 1998;Kinyanjui 1998),Mexico (Rabellotti 1993), Peru (Villaran 1993),Tanzania(Hansohm 1992),South Korea (Levy 1991), Sudan (Hansohm 1992),and Zimbabwe (Sverrisson 1992). Some of the studies are still under way and as yet report only initial findings. Few adequately examinethe inter-fmnrelationshipswithin clusters. At present there are no studies that have been done in relation to industrial clustering in Botswana; the literature reviewed, therefore, is drawn from other developing countries. The empirical framework of VOL 17 NO 1 JANUARY 2002 JOURNAL OF SOCIAL DEVELOPMENT IN AFRICA 21 this review rests on the objective of the study, which, as outlined above, is the determination of inter-firm relations within clusters. Generally, firms are expected to be rivals because of their competition for resources, power, technology, market, space and even labour. As a result, one would expect firms to work in isolation and independently so as to guard their market and other factors of production. Studies on industrial clusters, however, paint a different picture and there is increasing evidence from the literature that firm relationships exist. Clusteringfacilitatesverticalproduct relations between fums; relations that range from orchestration by large firms to arrangements amongst small firms and from casual exchange of information and tools to close inter-firm collaboration. Vertical subcontracting and the specialized division of labour have been observed in many the small-firm clusters of developing countries. In some cases process specialization is at an early stage and mainly constitutes the provision of specialized services by a few individual units to other firms in the cluster. Aeroe (1992) notes that in a woodworking cluster in three small towns in Tanzania, only one unit (in a sample of73) is process specialized. There are indications that, despite the absence of inter-firm division of labour, clustering does encourage firms to co-operate. Sverrisson's study on small woodworking units in intermediate towns of Kenya and Zimbabwe testifies that the main form of technical linkage is the borrowing back and forth of sophisticated and rare tools, which proceeded very much on informal terms (Sverrisson 1992). Similarly, Aeroe (1992) has observed that workshops in Tanzania's carpentry clusters exchanged tools, ideas and designs. In cases where manufacturing processes are technically more sophisticated, inter-firm relation based on a vertical division of labour among small firms are often an important aspect of production organization. Levy reports that Taiwan's overwhelming small firm footwear industry is distinguished by a proliferation of subcontracting relations and finds that it is rare for a Taiwanese footwear firm to perform in-house more than at most two of the various sub- processes of footwear production (Levy, 1991). 22 JOURNAL OF SOCIAL DEVELOPMENT IN AFRICA VOL 17 NO 1 JANUARY 2002 Nakizilo & DarkohllNCIPIENT INDUSTRIALIZATION Vertical production relations are also experienced in footwear clusters in Brazil (Sinos Valley) and India (Agra), which are characterized by extensive specialization and vertical finn linkages. The Sinos Valley is striking for the exhaustive range of inputs, components and intermediate products that are required for the manufacturing of shoes. Despite some large firms in the Sinos Valleybecoming vertically integrated,most small and medium sized shoe finns within the cluster produce only parts of shoes. Process specialization is also an important feature in the success of the Indian cotton knitwear industry. Tiruppur's knitwear enterprises are involved in interactive networks with local specialist job-firms that undertake cloth fabrication, calendaring, bleaching and fmishing, among others (Cawthorne 1993). Cawthorne also draws attention to the process of the division ofIabour within finns, where, to keep labour supervision costs and capital investments low, finns often split production activities across semi-independent units specialized for urban and production organization functions. Kumasi's Suame district also displays signs of extensive inter-firm division of urbanization. Dawson (1992) reports that a number of enterprises develop specialized expertise, produce specific parts or acquire exclusive technological capabilities. A significant degree of co- operation that has developed enables small workshops with only limited equipment and staff to take on substantial pieces work, parts of which are shared or subcontracted out to neighbouring enterprises (Dawson 1992). Although the predominant practice observed is that of restricting inter-firm subcontractual rules solely to small finns within the cluster, it cannot be ruled out that large-scale finns fall within the production arrangements. Dore's (1983) remarks on relational subcontracting in Japan, as well as Schmitz's (1992) observations on the Stuttgart region of South Gennany indicate that not only can large firms be a part ~fthe cluster, but that they often play the leadrole by weavingwebs of vertically delineated production subcontractual arrangement with small firms located in close proximity to the large units. VOL 17 NO 1 JANUARY 2002 JOURNAL OF SOCIAL DEVELOPMENT IN AFRICA 23 Fig. 1: Location of Lobatse In the Southeast District GHANZI DISTRICT .Ci¥Town + lEGEND N Moj