Competitive Advantage in the Tile Industry

Competitive Advantage and Upgrading in the Tile Industry:
Integrating a Cluster and a Value Chain Perspective

Jörg Meyer-Stamer
Claudio Maggi
Silene Seibel

Second Draft

February 2002

Content

1 Introduction 1

2 Conceptual framework: Clusters and value chains 2

3 General features of the tile industry 8

3.1 Features of the product and the production process 11

3.2 The role of suppliers of capital goods and glazing materials 14

3.3 The demand side 15

4 The competitive game in the tile industry: Italy vs. Spain 18

4.1 The Sassuolo cluster 19

4.1.1 Stage of industrial evolution: Consolidation 19

4.1.2 Structure of the value chain: Downstream activities 23

4.1.3 Structure of the value chain: Upstream activities 25

4.1.4 Structure of supporting institutions 30

4.1.5 Trends in upgrading 33

4.2 The Castellón cluster 35

4.2.1 Industrial evolution 36

4.2.2 Structure of the value chain: Downstream activities 38

4.2.3 Structure of the value chain: Upstream activities 41

4.2.4 Structure of supporting institutions 44

4.2.5 Trends in upgrading 46

4.3 Comparing governance patterns in Sassuolo and Castellón 47

5 Latecomer industrialisation and value chains in the tile industry: The case of Santa Catarina, Brazil 48

5.1 Location, value chain and competitive advantage 49

5.2 Santa Catarina in the global value chain 54

6 Conclusions 57

6.1 The quest for rents and the power structure in the global tiles value chain 57

6.2 The relevance of location 62

6.3 Conclusions for cluster research and cluster promotion 63

6.4 Conclusions for value chain research 65

7 Bibliography 66

1Introduction

This paper is about competitive advantage and upgrading in the tile industry.¹ Why did we choose the focus and the industry? The reason for choosing the focus is the interest in advancing the understanding of industrial change, in particularly with respect to latecomer industrialisation. This is an issue which is not only of academic interest but also politically relevant. What is the pattern of upgrading in industries which have to compete an increasingly globalised world, and what are the chances for industries in latecomer countries to enter competitive global markets? Apart from intra-firm efforts, what is involved in upgrading – what is the role of collective action and business associations, what is the role of government?

The starting point for the research was a territorial focus, related to an issue which was very prominent in the latecomer industrialisation discussion of the 1990s: to identify how local clusters connect to global markets. But the guiding question was not only addressing geographical but also governance issues: What are the power structures involved in the relationships between local producers and their global customers, and how do they influence local governance, in particular local collective upgrading efforts? This is an important question since the prevailing view, especially in the cluster policy community, is quite optimistic regarding the latitude for local policy making. Obstacles to local policy making seem to be determined mostly by local factors, such as weak business associations or lack of trust. How these obstacles are related to the insertion of local clusters into the global economy has only occasionally been addressed in any systematic way. We will show that in the case of the tile industry the mode of connection to global markets plays a decisive role in shaping, and also limiting, the options for collective action to upgrade at the local level.

We selected the ceramic tile industry because it is at the same time a highly globalised and a highly localised industry. It is dominated by companies from Italy and Spain, and in both countries production is concentrated in one cluster (Sassuolo / Italy, Castellón / Spain), both of which have a high export ratio (70 % in Italy, 50 % in Spain). For a long period, the cluster in Sassuolo was the leading one in the world, both in terms of production and technology. Castellón is a case of latecomer development where the transition from craft to industry began in the early 1970s. Apart from those two clusters we investigate the case of Santa Catarina, Brazil which is one of the most dynamic tile clusters in the developing world.

The structure of the paper is as follows. In Section 2 we present the conceptual framework for this paper. In Section 3 we introduce a number of key facts the reader should know about the tile industry. In Section 4 we investigate the clusters in Italy and Spain, as they shape the evolution of the industry. In order to understand how the interplay between Italy and Spain affects industrialisation efforts in latecomer countries, we present a case study of the Santa Catarina cluster in Section 5. In Section 6 we draw a number of conclusions.

2Conceptual framework: Clusters and value chains

For several years now, the academic as well as the policy discussion on latecomer industrialisation has been passing through a phase of profound redefinition. The 1980s and 1990s have witnessed the demise of heavy-handed, government-driven, top-down, centralist approaches to the promotion of industrial development in latecomer countries. As part of the “Washington Consensus”, it was suggested that government intervention did more damage than benefit, even in the successful newly industrialising countries of East Asia (World Bank 1993). Even though this argument has repeatedly been challenged,¹ it is probably fair to argue that in the course of the 1990s even those who saw industrial development in East Asia as a success story of developmental government started to have increasing doubts regarding the applicability of this approach in less-advanced countries.

The increasing scepticism regarding the effectiveness of centralised government intervention and traditional industrial policy was one of the reasons why industrial clusters and cluster promotion received increasing attention, not only among donor agencies and governments in developing countries but also among enterprise promotion agencies in industrialised countries. Clusters, commonly defined as concentrations of firms of a given sub-branch of industry plus supporting industrial and service firms within a delimited region, started to become prominent in the discussion on industrial promotion in the 1980s, initially based on the striking growth and export performance of Italian industrial districts, mostly populated by small and medium-sized enterprises / SME (Piore and Sabel 1984, Pyke, Becattini and Sengenberger 1990). As clusters are frequent in developing countries, and the private sector there mostly consists of SME, promoting clusters appeared as a promising new approach to stimulate latecomer industrialisation (Schmitz 1989). Clusters promised to reduce all sorts of barriers – barriers to intra-firm competence building, as firms could specialise more; barriers to exports, as local firms could work jointly in export consortia; barriers to upgrading, since an agglomeration of many firms of the same branch created strong demand for business development services.

Firms in clusters which operated individually did already enjoy all sorts of advantages (e.g. the availability of specialised, experienced workers), but firms in clusters which managed to act collectively had a chance to create a competitive advantage beyond this. Building on the concept of “collective efficiency” introduced by Schmitz (1995), Nadvi (1999, 1608) has elaborated on this:

"Collective efficiency is defined as having two aspects to it: external economies that clustered agents accrue by virtue of their location, and joint action benefits that arise from deliberate cooperation between local agents. I view external economies as the 'passive' dimension of collective efficiency. The term passive describes the nature of ties required between local agents in order to obtain externality gains. In contrast, joint action is the 'active' dimension of collective efficiency requiring deliberate and active cooperation. These two aspects can also be clearly linked; joint action by some agents can generate cluster-specific externality gains for others. This process of upgrading by facilitating the flow of technical information on standards and by assisting in managerial training. Local institutions can also play a potentially key function in defining and regulating local product standards, and thus in creating a reputational basis for the cluster's products. This provides a powerful example of what I refer to as 'externalities of joint action'".

Collective efficiency is not just a microeconomic issue but also involves a governance dimension. Passive advantages (easy access to suppliers, buyers and specialised workers) are based on market governance. It is the invisible hand of the market which is creating these advantages without deliberate collective action or government intervention. Active advantages may be based on hierarchical governance, if it is government which supports businesses’ efforts. More often, they will be based on network governance, if there is close interaction between government and non-governmental actors and joint upgrading efforts are based on negotiation and shared commitments. Enright and Ffowcs-Williams (2000, 4) summarise the rationale of cluster promotion as follows:

“Membership of clusters and networks can enhance the productivity, rate of innovation and competitive performance of firms. Clusters and networks can allow small firms to combine advantages of small scale with various of the benefits of large scale. Public policy on clusters and networks can help SMEs realise the opportunities and meet the challenges associated with globalisation. Essentially, a policy on clusters provides a framework for dialogue and co-operation between firms, the public sector (particularly at local and regional levels of government) and non-governmental organisations. This dialogue can lead to efficiency-enhancing collaboration amongst firms, such as in joint marketing initiatives, the creation of mutual credit guarantee associations, joint design and sponsorship of training, a more efficient division of labour among enterprises, etc. Such a dialogue can also lead to an improved quality of policy and government action (such as in training, the provision of information, and infrastructure supply).”

Some case studies on clusters in developing countries indicated that local governance, especially network governance, can contribute significantly to solving competitive challenges (Schmitz 1995, Nadvi 1996, Meyer-Stamer 1998). This was not necessarily surprising as one might have expected that local actors would respond more swiftly, and in a more targeted way, to new challenges than a distant, less informed central government ever could. Moreover, a stronger local economic development effort fits into the larger picture of decentralisation.

Over time, however, increasingly evidence surfaced which lead to a questioning of the prevailing view of clusters as favourable places for SMEs to become internationally competitive, and in particular as favourable places for collective efforts to create competitive advantage. Most relevant are two types of evidence: on structural change in clusters, and with respect to the interaction between clusters and the economy at large.

Regarding structural change inside clusters, it was first and foremost ongoing observation of industrial districts in Italy which lead to the discovery of different trajectories in the evolution of clusters. In the 1990s, it has been found that inside many of them concentration processes occurred, and that some others began to de-verticalize, i.e. to relocate certain activities to other locations (Brusco et al. 1996, 28 f., Ottati 1996, 45 ff., Crestanello 1996, 72 ff.). In Emilia-Romagna, a case study of four industrial districts found that 55.2 % of firms and 89.9 % percent of workers belonged to economic groups, a finding which indicates that the prevalence of SME is a feature of the past (Brioschi, Brioschi and Cainelli 2001). Some industrial districts managed to adjust to increasing competitive pressure, whereas others started to decline (Belussi 1999). Government efforts to promote the competitiveness of clusters have become less effective, as the demands of firms are increasingly differentiated, specialised and sophisticated, whereas public support institutions are increasingly bureaucratised (Whitford 2001).

Regarding interaction between clusters and the economy at large, some authors have early on criticised the analysts of industrial districts for their excessive focus at the local and the neglect of global economic issues (e.g. Amin and Thrift 1994). In the course of increasingly deep investigation of clusters, notably in developing countries, it became obvious how important the interaction with the global economy actually is – sometimes in terms of stimulating massive local upgrading efforts, such as in the case of the surgical instruments cluster in Sialkot, Pakistan (Nadvi 1999), sometimes in terms of creating obstacles to local collective upgrading efforts, such as in the case of the footwear cluster in Sinos Valley, Brazil (Schmitz 1998).

These kinds of observations led some cluster researchers to have a closer look at one specific feature of international trade: the existence of global value chains. Economics textbooks suggest that international trade involves transactions between anonymous sellers and buyers. In practice, however, quite the opposite is the case. Rauch (1999) finds that what he calls “differentiated products”, i.e. products which are not standard commodities and usually involve some degree of direct interaction between producer and buyer to determine product quality and specifications, accounted for 67.1 % of world trade in 1990, up from 56.5 % in 1970. Direct interaction may take different forms; typical examples are international business networks established by ethnic minorities (e.g overseas Chinese), international networks managed by trading companies (Rauch 2001), or intra-firm trade. Cluster researchers, however, have chosen to investigate yet another variation of direct interaction, namely the one which has been addressed under the heading of “global commodity chains” (GCC; Gereffi 1996a).

The work of Gereffi is mostly based on the investigation of production networks in the garment industry and trade, and even though he introduces a distinction between buyer-driven commodity chains (e.g. in garments) and supplier-driven commodity chains (e.g. in cars), his own work is mostly on the buyer-driven apparel chain. Other authors have investigated buyer-driven chains in other industries, such as footwear (Schmitz and Knorringa 1999), horticulture (Dolan, Humphrey and Harris-Pascal 1999), coffee (Fitter and Kaplinsky 2001), and furniture (Kaplinsky and Morris, undated). The common feature of these industries is the dominating position of buyers vis-à-vis producers in developing countries, which is due to a high degree of concentration among buyers in industrialised countries and relatively low barriers to entry for producers in developing countries. What these studies mostly do not investigate is the competitive pattern among the buyers in the industrialised countries, and the extent to which producers in developing countries might be able benefit from oligopolistic competition between buyers.

But this is only one of the limitations of the available literature on commodity chains. To start with, the term chosen by Gereffi is a contradiction in terms. As Humphrey and Schmitz (2000, 10) point out, he investigates commodity chains precisely because they do involve differentiated products, i.e. they are not about standardised commodities. This is one of the reasons why they and other authors prefer the term “value chain”.² The main deficiency, however, is the mostly descriptive nature of Gereffi’s work. It involves dense descriptions based on extensive field research, but little theory. The purpose of his work is to identify power structures within the international political economy, but he does not provide a clear theoretical framework in doing so. Similarly, other authors (e.g. Humphrey and Schmitz 2001) emphasise that global value chains involve governance, but the do not conceptualise this governance from a theoretical perspective. Kaplinsky (2000, 9 ff) presents “three key elements of value chain analysis”, name barriers to entry and rent, governance, and systemic efficiency. The latter is another descriptive category, not being embedded in a theoretical framework. In terms of governance, he distinguishes between legislative, judicial and executive governance, which is helpful in his description but also not exactly state-of-the-art in terms of theory. What is most useful, however, is his reference to barriers to entry and rent, which are, from an economic perspective, two key conceptual elements.

Another deficiency of the GCC concept is that it lumps together entirely different arrangements. What Gereffi calls producer-driven commodity chains are arrangements which are co-ordinated by transnational industrial corporations. To some extent they organise their supplies on a market basis, to some extent as relational contracting. What Gereffi calls buyer-driven commodity chains is a different arrangement altogether, and in fact are two different arrangements. First, there are companies such as Nike, which have organising production as their main rationale. Second, there are companies such as Ikea, which have selling products to final individual customers as their main rationale.

Another weakness of the literature on value chains is its relatively limited focus, and this in different respects. It suffers from the same weakness Bell and Albu (1999) identified in the cluster discussion, namely neglecting a systematic discussion of technology issues. Related to this is the fact that many case studies also suffer from a very narrow focus. Gereffi’s research on the garment value chain is a case in point, as he looks exclusively at apparel producers and retailers (Gereffi 1996b, 1999; Bair and Gereffi 1999). However, it is no secret that this value chain starts on cotton fields and in chemical plants, and it involves not only spinning and weaving, as well as dyeing and finishing, but also the chemical and other inputs used for these processes, as well as the machinery. Technical progress in this value chain is to a large extent driven by capital goods manufacturers and chemical firms (Breitenacher, Vieweg & Vogler-Ludwig 1995). This suggests that including them in the analysis might render some relevant insights on the dynamics of the apparel value chain. A different school of value chain research, namely the one which looks at ecological issues, has been pursuing this much broader view and delivered a more differentiated perspective at the structure of international value chains (Aarts 1996, Claus & Völkle 1996).

This is not to say that value chain analysis is not useful. The opposite is the case, and this case study will illustrate how useful it actually is. The problem with value chain analysis so far is its too narrow focus, its bias for buyer-driven chains and its shallow theoretical foundation; even clumsy attempts to formulate typologies (e.g. Sturgeon 2001) can make their way past referees. But what are adequate theoretical points of reference in analysing value chains, including their interaction with local clusters?

The answer is relatively simple: Both value chains and clusters are about patterns of transaction which are neither markets nor hierarchies, but networks. Powell (1990) argues that it is useful to introduce networks as a third pattern of transaction. He suggests that networks are not an intermediate pattern to be located on a continuum somewhere between the extreme points of market and hierarchy, but rather a distinctive pattern of organising economic transactions. Storper and Harrison (1991) address networks in terms of the link between the spatial organisation of production and governance. They provide a typology which straddles the gap between cluster- and value-chain-research, as clusters are usually integrated into value chains and value chains often connect different clusters. They also point at the fact that clusters are not necessarily agglomerations of SMEs, thus anticipating the findings of Markusen (1996). Hollingsworth and Streeck (1994) emphasise that not only networks but also hierarchies and markets are profoundly shaped by different patterns and traditions of governance in different countries. Hollingsworth (1998) reminds us that economic co-ordination happening at different, interrelated levels, from the local to the global, is not exactly a new phenomenon, though the different levels have become more interdependent. In other words, looking at the interaction between local clusters and international chains is nothing radically new. Neither is the introduction of the governance issue (in the sense of political governance, not corporate governance). What we can do is develop a better understanding of the interaction between clusters and chains.

What about the theoretical perspective at networks? From an economic perspective, the rationale of networks is based on the existence of transaction costs and principal-agent problems (Richter & Furubotn 1996). Market-based transactions can involve high transaction costs if products are not standardised, contract enforcement is unpredictable, costly and/or time intensive, or for other reasons. Hierarchies may involve high costs of supervision due to the agency problem. Companies prefer network-based relationships in those cases where they appear as more efficient than market-based transactions or hierarchies. Typical examples are long-term relationships with suppliers of differentiated products which are more efficient than market-based transactions (because the reliability and competence of the supplier is known) and hierarchy (because the supplier is specialised in what he does and thus more efficient). Chains emerge in situations where buyers find it inefficient to draw on markets, in particular due to the necessity to have some control over the product characteristics (due to customer preferences, or legal requirements such as phyto-sanitary standards, or issues such as labour, environmental or other standards which are difficult to supervise and enforce), and where buyers also find it inefficient to organise the whole operation as a hierarchy, i.e. a vertically integrated firm. Chains are the more globalised, the bigger the difference is between chain elements in terms of cost composition, economics of scale and technological sophistication.

Complementary to the perspective of institutional economics, innovation economics emphasise the importance of networks as places for learning-by-interacting (OECD 1992). This refers both to local and international networks. Their emergence reflects the fact that technical change is much more a process, involving permanent interaction between different actors, than the outcome of isolated events (such as the sale of a license or blueprint). An important factor in this respect is that technical know-how is never fully codifyable, and that the transfer of tacit know-how is usually based on ongoing interaction.

From a governance perspective, networks are an alternative to hierarchy in those cases where resources are distributed among different actors so that decision-making is based on negotiation rather than command, even though there are usually power asymmetries. This has been thoroughly investigated with respect to governmental policies which rely on the know-how and other resources, as well as the voluntary co-operation, of non-governmental actors (Mayntz & Scharpf 1995, Messner 1997). The categories which have been formulated in the conceptual work on policy networks can as well be applied to business networks which are based on negotiation (Messner & Meyer-Stamer 2000).

What is the use of the value chain concept for this study? First and foremost, it is the emphasis on factors which shape competitiveness and upgrading which are located outside local clusters. As the subsequent sections will show, in the case of the ceramic tile industry it is tempting to choose the local cluster as the focus of research. But we will see that limiting the focus in this way would lead to an entirely inadequate understanding of competitive patterns and upgrading. It might lead to entirely unrealistic suggestions for collective action and government intervention to strengthen the competitiveness of the clusters.

In this paper we will show that local factors played the crucial role in shaping the structure of the tile industry. Until recently, a cluster-centred perspective rendered most of the insights necessary to understand the evolution of the industry and the dynamics of competitive advantage. This, however, is no longer the case. Local factors continue to play a very important role. But in order to understand the competitive dynamics in the tile industry it is essential not only to analyse the interaction between the two leading clusters, in particular with reference to the upstream activities in the tile value chain. It is also essential to look at their interaction with downstream activities in the value chain, i.e. commercialisation.

In the following sections, we will look at each of the three clusters in turn, addressing their competitiveness both from a local cluster and a global value chain perspective. We will take a wide perspective at the value chain, including activities both upstream and downstream from tile manufacturing. We will investigate the value chain looking at technological interaction, but also at the power structures.