1、 Kaplinsky&Readman, 2001; Porter, 1990）. Capitalizing on one of the most productive areas of the recent literature on SMEs, we restrict our field of research to small enterprises located in clusters. There is now a wealth ofempirical evidence （Humphrey, 1995; Nadvi &Schmitz, 1999; Rabellotti, 1997）
2、showing that small firms in clusters, both in developed and developing countries, are able to over come some of the major constraints they usually face: lack of specialized skills, difficult access to technology, inputs, market, information, credit, and external services. Nevertheless, the literatur
3、e on clusters, mainly focused on the local sources of competitiveness coming from intracluster vertical and horizontal relationships generating collective efficiency （Schmitz, 1995）, has often neglected the increasing importance of external link ages. Due to recent changes in production systems, dis
4、tribution channels, and financial markets, and to the spread of information technologies, enterprises and clusters are increasingly integrated in value chains that often operate across many different countries. The literature on global value chains （GVCs） （Gereffi, 1999; Gereffi& Kaplinsky, 2001） ca
5、lls attention to the opportunities for local producers to learn from the global leaders of the chains that may be buyers or 1 producers. The internal governance of the value chain has an important effect on the scope of local firms upgrading （Humphrey& Schmitz, 2000）. Indeed, extensive evidence on L
6、atin America reveals that both the local and the global dimensions matter, and firms often participate in clusters as well as in value chains （Pietrobelli& Rabellotti, 2004）. Both forms of organization offer opportunities to foster competitiveness via learning and upgrading. However, they also have
7、remarkable drawbacks, as, for instance, upgrading may be limited in some forms of value chains, and clusters with little developed external economies and joint actions may have no influence on competitiveness. Moreover, both strands of literature were conceived and developed to overcome the sectoral
8、 dimension in the analysis of industrial organization and dynamism. On the one hand, studies on clusters, focusing on agglomerations of firms specializing in different stages of the filiere, moved beyond the traditional units of analysis of industrial economics: the firm and the sector. On the other
9、 hand, according to the value chain literature, firms from different sectors may all participate in the same value chain （Gereffi, 1994）. Nevertheless, SMEs located in clusters and involved in value chains, may undertake a process of upgrading in order to increase and improve their participation in
10、the global economy, especially as the industrial sector plays a role and affects the upgrading prospects of SMEs. The contribution this paper makes is by taking into account all of these dimensions together. Thus, within this general theoretical background, this study aims to investigate the hypothe
11、sis that enterprise upgrading is simultaneously affected by firm-specific efforts and actions, and by the environment in which firms operate. The latter is crucially shaped by three characteristics: （i） the collective efficiency of the cluster in which SMEs operate, （ii） the pattern of governance of
12、 the value chain in which SMEs participate, and （iii） the peculiar features that characterize learning and innovation patterns in specific sectors. The structure of the paper is the following: in Section 2, we briefly review the concepts of clustering and value chains, and focus on their overlaps an
13、d complementarities. Section 3 first discusses the notion of SMEs upgrading and then 2 introduces a categorization of groups of sectors, based on the notions underlying the Pavitt taxonomy, and applied to the present economic reality of Latin America. Section 4reports the original empirical evidence
14、 on a large sample of Latin American clusters, and shows that the sectoral dimension matters to explain why clustering and participating in global value chains offer different opportunities for upgrading in different groups of sectors. Section5 summarizes and concludes. 2. CLUSTERS AND VALUE CHAINS
15、During the last two decades, the successful performance of industrial districts in the developed world, particularly in Italy, has stimulated new attention to the potential offered by this form of industrial organization for firms of developing countries. The capability of clustered firms to be econ
16、omically viable and grow has attracted a great deal of interest in development studies. 1 In developing countries, the sectoral and geographical concentration of SMEs is rather common, and a wide range of cases has since been reported. 2 Obviously, the existence of acritical mass of specialized and
17、agglomerated activities, in a number of cases with historically strong roots, does not necessarily imply that these clusters share all the stylized facts which identify the Marshall type of district, as firstly defined by Becattini （1987）. 3 Nonetheless, clustering may be considered as a major facil
18、itating factor for a number of subsequent developments （which may or may not occur）: division and specialization of labor, the emergence of a wide network of suppliers, the appearance of agents who sell to distant national and international markets, the emergence of specialized producer services, th
19、e materialization of a pool of specialized and skilled workers, and the formation of business associations. To capture the positive impacts of these factors on the competitiveness of firms located in clusters, Schmitz （1995） introduced the concept of collective efficiency （CE） defined as the competi
20、tive advantage derived from local external economies and joint action. The concept of external economies 4 was first introduced by Marshall in his Principles of Economics（1920）. According to Schmitz （1999a）, incidental external economies （EE） are of importance in explaining the competitiveness of in
21、dustrial clusters, but there is also a deliberate force at work: consciously pursued joint action 3 （JA）.Such joint action can be within vertical or horizontal linkages. 5 The combination of both incidental external economies and the effects of active cooperation defines the degree of collective eff
22、iciency of a cluster and, dynamically, its potential for fostering SMEs upgrading. Both dimensions are crucial: Only incidental, passive external economies may not suffice without joint actions, and the latter hardly develop in the absence of external economies. Thus, our focus is on the role of int
23、racluster vertical and horizontal relationships generating collective efficiency. However, recent changes in production systems, distribution channels and financial markets, accelerated by the globalization of product markets and the spread of information technologies, suggest that more attention ne
24、eds to be paid to external linkages. 6 Gereffis global value chain approach （Gereffi, 1999） helps us to take into account activities taking place outside the cluster and, in particular, to understand the strategic role of the relationships with key external actors. From an analytical point of view,
25、the value chain perspective is useful because （Kaplinsky,2001; Wood, 2001） the focus moves from manufacturing only to the other activities involved in the supply of goods and services, including distribution and marketing. All these activities contribute to add value. Moreover, the ability to identi
26、fy the activities providing higher returns along the value chain is key to understanding the global appropriation of the returns to production. Value chain research focuses on the nature of the relationships among the various actors involved in the chain, and on their implications for development （H
27、umphrey & Schmitz, 2002b）. To study these relationships, the concept of governance is central to the analysis. At any point in the chain, some degree of governance or coordination is required in order to take decisions not only on what should be, or how something should be, produced but sometimes al
28、so when, how much, and even at what price. Coordination may occur through arms-length market relations or non market relationships. In the latter case, following Humphrey and Schmitz （2000）, we distinguish three possible types of governance:（a） network implying cooperation 4 between firms of more or
29、 less equal power which share their competencies within the chain; （b） quasi-hierarchy involving relationships between legally independent firms in which one is subordinated to the other, with a leader in the chain defining the rules to which the rest of the actors have to comply; and （c） hierarchy when a firm is owned by an external firm. Also stressed is the role played by GVC leaders, particularly by the buyers, in transferring knowledge along the chains. For small firms in less developed countries （LDCs）, participation in value chains is a way to obtain i
