Do not quote without the permission of the authors. Local ... - CiteSeerX

7 downloads 16115 Views 222KB Size Report
the importance of international knowledge for developing economies is ... (collected through a survey) of the software firms concentrated in the cluster of .... public or a private good does not depend on the intrinsic characteristics of knowledge; .... differentiates custom-made products from registered packages (Bitzer, 1997).
3rd Globelics Annual Conference - Africa 2005

Working Paper: Do not quote without the permission of the authors.

Local Knowledge Spillovers and Innovation: The software cluster in Uruguay1 Effie Kesidou* and Henny Romijn†

Abstract Currently, a number of economic geographers and economists of innovation incorporate knowledge spillovers in their analysis of clusters (Jaffe, Trajtenberg and Hesderson, 1993; Audretsch and Feldman, 1996; Verspagen and Schoenmakers 2000, Caniels 1999). They argue that local technological externalities are the main reason for the clustering of innovative activity. Empirical studies have been undertaken in economically advanced countries, which confirm that knowledge spillovers are geographically bounded. This trend is especially strong in high-tech sectors because they involve relatively more tacit knowledge than the traditional sectors. Yet, an important gap in the literature remains, in that research primarily has been limited to high-tech clusters in the advanced economies. There is no indication that this argument holds in less developed countries too. The purpose of this research, thus, is to find out whether Local Knowledge Spillovers (LKS) are important drivers of innovation and learning in a developing country context. This study develops a novel methodological approach, in which the importance of LKS (in relation to other mechanisms of knowledge flow) is measured and statistically analysed on the basis of new firm level data collected through fieldwork (survey and interviews from the software cluster in Montevideo, Uruguay. Our analysis supports the positive relation, propounded in the theory, between LKS and the innovative performance of firms. In particular, the probability that a firm comes up with a product innovation is enhanced by local knowledge flows from universities, informal knowledge flows from competitors and labour turnover. However, the commercial success that firms subsequently obtain with these product innovations depends on information flows from their international customers. 1. Introduction Globalization places new challenges to small countries, especially in the developing world. Economies of limited size redefine their position and search for those advantages that will enable them to compete in the new era. Despite the drawback of a country being small, there are cases of developing economies that have managed to catch-up2 (e.g. Taiwan, South Korea). However, this pattern of development is not easily replicated any 1

Key words: Local knowledge spillovers, Innovation, Geographical clusters Catch-up refers to the process of rapid economic growth that developing countries may experience. This happens due to the fact that developing countries are characterised by low levels of labour productivity. Thus, their productivity may grow faster than the one of developed countries that already operate in a high level of labour productivity (Abramovitz, 1986; Van Dijk, 2005). 2

1

3rd Globelics Annual Conference - Africa 2005

longer in a world where government initiative shrivels and trade liberalisation is greater than ever before. Like many small developing countries, during the last decade Uruguay has been experiencing a profound crisis. Thus, the need for economic development is urgent. As Uruguayans characteristically say, Uruguay is a small country between two giants –Argentina and Brazil. While the rest of their economic sectors are currently shrinking, the software sector in Uruguay has been growing rapidly. Software firms are going through a considerable growth of sales, mainly directed to the export market. Uruguay is characterised by a natural (demographic) concentration of economic activity in the capital city of Montevideo; where also software firms are clustered. While underresearched, this case captured our attention and led us to pose the following question: Is the growth of software sector related to benefits deriving from the geographical concentration of the software firms in Montevideo? This agglomeration of economic activity may generate benefits related to economies of scale, labour market concentration and rapid diffusion of knowledge. The literature on territorial theories3 emphasises the reasons that lay behind the clustering of economic activity. One of the advantages of agglomeration is the easy and fast circulation of knowledge (especially in its tacit form) among the actors that are located in the cluster. Studies on local knowledge spillovers (LKS)4 claim that local technological externalities are the main reason for the clustering of innovative activity. In addition, modern economic theory emphasises that innovation and technological change boost economic growth, due to the fact that innovation creates conditions of increasing returns in production (Romer, 1986, 1990; Griliches, 1992). Such conditions accelerate economic growth over the long run. Local knowledge spillovers are one of the key mechanisms through which this occurs. Empirical studies have been undertaken in economically advanced countries, which confirm that knowledge spillovers are geographically bounded. This trend is especially

3

Regional Systems of Innovation and Learning Region (Cooke, 2001; Morgan, 1997), Innovative Milieux (Aydalot, 1986; Camagni, 1992), New Industrial Spaces (Porter, 1990; Storper & Scott, 1988; Saxenian, 1994), Industrial District (Piore & Sabel, 1984; Becattini, 1990; Schmitz, 1999). 4 Knowledge spillovers are positive technological externalities that derive from the inability of firm A to retain the economic returns of its innovative activity. As a consequence, firm B can take advantage of the new product or the new knowledge directly and without compensating firm A.

2

3rd Globelics Annual Conference - Africa 2005

strong in high-tech5 sectors because they involve relatively more tacit knowledge than the traditional sectors. Yet, an important gap in the literature remains, in that research primarily has been limited to high-tech clusters in the advanced economies. There is no indication that this argument holds in less developed countries too. Although clustering is a phenomenon that has been identified and researched in developing countries, mostly in traditional sectors, little is known about the nature and the function of knowledge spillovers in less developed countries. Frontier technology and knowledge are concentrated in developed countries. Cluster of firms in the South access this knowledge centres by importing technology products or by collaborating closely with MNCs6 already established in developing countries. While the importance of international knowledge for developing economies is evident, it is not yet clear how local knowledge spillovers can have an economic relevance in the context of developing countries. The merit of local knowledge lays on the translation of the international knowledge into local meanings. In other words, the adoption of an advanced technology into another context and the search for economic value of the adapted or improved technology is a challenging process in many developing economies. In particular, it constitutes a process of technological upgrading (or innovation) in firms in located in less developed countries, and in particular in small and medium enterprises, which is demonstrated by the ability to adapt, improve and develop technologies as well as by the ability of problem solving (Bell M. 1984). On this ground the easy and rapid circulation of knowledge is crucial for seeing innovation flourishing in many developing economies. The purpose of this research, then, is to elucidate whether LKS are important drivers of innovation and learning in a developing country context. Using new firm-level data (collected through a survey) of the software firms concentrated in the cluster of Montevideo, we test whether LKS affect the innovative performance of the firms within the cluster or not. The outcome of this study will contribute to a better understanding of the current mechanisms of LKS in developing countries, and will indicate whether LKS can be a potential path for development and growth. The article is structured as follows: 5

Knowledge-based or high-tech sectors are characterised by a high proportion of investments in R&D (Lall, 2000). 6 MNCs stands for Multinational Cooperations.

3

3rd Globelics Annual Conference - Africa 2005

In the introduction we stress the objective of this research and in section 2 we discuss the literature on local knowledge spillovers while reflecting it on development issues. In section 3 we introduce the methodology that has been used. Section 4 features the conceptual model and the hypothesis. Finally, in the section 5 we discuss the results of the empirical analysis and offer some conclusions.

2. Review of Literature The first author to raise the claim that clusters facilitate the diffusion of knowledge through the concentration and the mobility of specialised labour was Alfred Marshall (1920). Inspired by the cotton mills of nineteenth-century Manchester, he noticed the existence of production systems that are geographically concentrated. One of the ingredients of Marshall’s Industrial District Theory can be interpreted to refer to knowledge spillovers ‘The mysteries of the trade become no mysteries, but are as it were in the air’7 (Marshall, 1920, p. 225). The concept of knowledge spillovers reappeared in the literature, decades later, in the work of Scitovsky (1954). The so-called real or technological externalities resemble what Marshall referred to be ‘something in the air’, albeit, Scitovsky did not consider explicitly the spatial attributes of knowledge spillovers. According to Scitovsky, real external economies are the result of the interdependence between the decisions and actions of various firms. In the presence of interdependence between firms, the production of a firm can be modified by the behaviour and outcomes of another enterprise directly and not through the market. A major contribution of Scitovsky was the idea that the concept of real externalities and generally of external economies should be studied in two distinct contexts: The static theory of equilibrium and the theory of industrialisation of underdeveloped countries. His analysis together with the theory of balanced growth (Rosenstein-Rodan, 1943) provided theoretical support to the so-called import substitution policy of industrialisation (ISI) that has been applied in many developing countries during the 1960s. However, the failure of several countries to catch up following the ISI model of industrialisation in contrast to many other countries that pursued the so-called exportoriented strategy of industrialisation led several academic scholars and policy makers in 7

This refers to the informal exchange of knowledge or cafeteria effects.

4

3rd Globelics Annual Conference - Africa 2005

the 1980s to believe that trade liberalisation is the policy that will enable developing countries to grow fast. Twenty years later the difficulty of various countries to respond to the expectations of the trade liberalisation model of development made many of its initial supporters to reconsider this policy. In the last decade, Westphal (1990) claimed that the export-led model of growth entails a degree of intervention, and that it has been mistakenly argued that trade liberalisation was the only policy that reflected its content. In a globalised world it is most unlikely for small developing countries to follow the paradigms of Taiwan or South Korea. The pressure and difficulties are expanding under the new era of loose State control and increased trade liberalization. Under these conditions of uncertainty and hesitation to follow any of the two dominant models of development, much emphasis has been given to clusters of enterprises due to the success of these clusters in competing in international markets (Becattini, 1990; Schmitz, 1999). The success of clusters is attributed to the presence of external economies. Real external economies in the context of industrial district or cluster occupy a prominent role in the discussions of growth during the last decade. Simultaneously, this theoretical consideration yields a challenge for many developing countries: Do local knowledge spillovers foster innovation in clusters in developing countries? It is worth knowing whether they do it or not, because if high-tech clusters drive innovation and learning in developing countries, this fact could attract the attention of government policies and stimulate economic development by identifying and encouraging high-tech clusters. At this point it is important to stress that this study analyses the impact of LKS upon the innovation of firms within clusters. Competition based on innovation differs from the traditional notion of competitiveness based on squeezing prices (static competitiveness). In particular, we conceptualise competitiveness as a dynamic process of technological upgrading which usually in developing countries takes the form of incremental innovations. The tacit nature of knowledge, which is an important ingredient of innovation, constitutes the foundation of the current conceptualisation of local knowledge spillovers. Knowledge is an important component for any type of innovation. Various aspects of knowledge (information) can be codified and thus diffused, imitated and sold. Yet, the tacit part of knowledge cannot be traded but can only be learned. This learning process

5

3rd Globelics Annual Conference - Africa 2005

has two major characteristics: it is cumulative (Dosi, 1988) and it is achieved through interaction (Lundvall, 1988; 1992). Learning is not a rapid leap to wisdom but rather a gradual process whereby new knowledge is build upon previous understandings and incremental changes on the known parts unravel the unknown aspects of the matter (Nelson and Winter, 1982). This is one of the reasons why knowledge spillovers and innovation have a local dimension; regions that have accumulated knowledge can produce new knowledge easier than other areas that are in the beginning of the learning process. Moreover, it is acknowledged now (Lundvall, 1992) that innovation is the result of the interaction between firms, between functions within the firm, between producers and users, and between firms and the research institutes as well as the wider institutional infrastructure. In particular, face-to-face interaction is necessary for the exchange of knowledge that is tacit and complex. Face-to-face interaction facilitates the sharing of knowledge. During this process tacit knowledge becomes explicit and it is converted into another new knowledge (Nonaka, 1994). Nowadays a number of geographers, economic geographers and economists of innovation have incorporated knowledge spillovers in their analysis of clusters. Jaffe, Trajtenberg and Hesderson (1993) examined the geographic distribution of patents and the citations of these patents, and concluded that patent citations are highly localised, indicating that knowledge spillovers are spatially bounded. In addition, Audretsch and Feldman (1996) examined the spatial distribution of innovation using as a proxy new products introduced to the U.S. market. Their findings support the idea that innovation is spatially concentrated, due to the tacit nature of technological knowledge, which indicates that in-person interaction is necessary for the knowledge to spill over (Verspagen and Schoenmakers, 2000; Caniels, 1999). Another line of research, which has implicitly incorporated LKS, is qualitative case studies of clusters in different countries. One of the most influential works has been carried out in the high-tech cluster of Silicon Valley (Saxenian, 1994). This study supported empirically the idea that the particular cluster draws its advantages from the strong interdependence of firms, which allows the exchange of ideas and knowledge, facilitating in turn the learning process and, consequently, increasing the innovative

6

3rd Globelics Annual Conference - Africa 2005

activity of these firms. However, an important gap in the existing literature is that most of research works have been restricted to high-tech clusters in advanced economies. Not much is known about whether this argument holds in less developed countries as well. The most influential work on clusters in developing countries has been undertaken by Schmitz (1999). He has analysed the impact of external economies and of joint action upon the competitiveness of clustered firms. Schmitz concluded that firms which act jointly are more dynamic and competitive than those that receive passively the advantage of external economies within a cluster. In addition, he added the notion of “collective efficiency” that is defined as ‘...the competitive advantage derived from local external economies and joint action’. According to Schmitz the existence of external economies the so-called ‘passive collective efficiency’- is not sufficient to sustain the competitiveness of the clustered firms. It is the cooperation of firms that gives rise to the ‘active collective efficiency’, which is crucial for the long-term competitiveness of firms. However, at this point it is important to understand what Schmitz means by 'joint action' that is so significant for the competitiveness of the firms. Joint action or active CE is nothing but the utilisation or internalisation of external economies. Firms take into account external economies, to later pursue investments in a coordinated way. Nevertheless, when referring to external economies Schmitz does not separate technological from pecuniary externalities; rather, external economies are conceived as one thing. Therefore, by treating external economies as a ‘black box’ many aspects of the issue remain in the shadows. This is a crucial issue, because technological externalities are different from pecuniary externalities and subsequently they lead to distinct policies. In addition, Schmitz conceptualises competitiveness in a rather traditional way; that of increasing sales and exports, while overlooking the advances in innovation that firms might have achieved. In sum, the review of the literature indicates that LKS are important mechanisms that enhance the innovativeness of firms within clusters in advanced economies. However, the literature on developing countries is puzzling because LKS have not been distinctively examined. In order to assess the importance of LKS for firms’ innovativeness it is first crucial to see if LKS do exist. Thus, the mechanisms of knowledge flow will be examined and analysed in this study.

7

3rd Globelics Annual Conference - Africa 2005

3. Methodological Issues In contrast to other types of knowledge flow, LKS do not involve any market transactions and no compensation is given for the acquired knowledge (i.e. they are direct and free). Although much has been said and written about the concept, not much is actually known about LKS themselves, because this characteristic makes them hard to capture. So far, research works in advanced countries have used indirect proxies of LKS, such as patent citations or R&D-based proxies to justify their importance for the localized nature of knowledge and innovation. However, such indirect data have apparent shortcomings. For instance, patents do not cover all the outcomes of innovative activity. While this is true even for advanced countries, it would be a more severe problem for developing countries, where only a fraction of innovation is ever patented. Thus, using patents as proxies in this research would pose a great risk of misrepresenting innovative activity. Besides, in developing countries adequate patent and R&D data of this type are hardly available. Much of firms' innovation is informal, and does not feature in any statistical database (Bell, 1984). Even if we could apply similar methodologies as the ones that have been used in developed economies, inherited problems of these approaches have raised strong criticisms. In particular, Breschi and Lissoni (2001) argued that the association of patents’ distribution with LKS constitutes only an indirect evidence of the presence of local knowledge spillovers. The fact that patents and patents’ citations are locally distributed advocates that knowledge flows more frequently among local firms than among firms situated in long distance. There is not any indication that knowledge circulates freely and without compensation between the firms. Zucker et al (1998) provided empirical evidence showing that the knowledge, which is exchanged between local firms and Universities, is pecuniary. Additionally, the notion of tacit knowledge and the way in which has been often used to justify the presence of LKS is now challenged. According to economics of knowledge the strict division between tacit and codified knowledge is an oversimplification of a more complex problem that derives not only from the features of knowledge, but also from the characteristics of the recipients of this knowledge (Cowan, David and Foray,

8

3rd Globelics Annual Conference - Africa 2005

2000). In particular, technical knowledge is not tacit only because it cannot be articulated. Rather, technical knowledge is highly contextual and specific and, thus, not everybody is aware of the way that this knowledge is translated into simple meanings. For example, in order to imitate or change a software program it is essential to understand how it is built. For that it is necessary to know the particular programming language that the code is written in. For an economist the produced code might be tacit knowledge. However, for a software engineer who is aware of the particular vocabulary, each code represents a welldefined meaning that he can manipulate, reproduce or improve. On the same ground, Cowan et al (2000) argued that knowledge is highly contextual and that specific institutional incentives determine its public or private character. Whether knowledge is a public or a private good does not depend on the intrinsic characteristics of knowledge; rather, ‘...the economically relevant characteristics of a good or service derive from the structure of incentives provided for its production and/or consumption’ (Cornes and Sandler, 1996; sited in Breschi and Lissoni, 2001, p. 13). We addressed these important methodological problems by developing a novel approach, in which the importance of LKS (in relation to other mechanisms of knowledge flow) has been measured and statistically analysed on the basis of new firm level data collected in the field. Detailed information on firms’ innovative activities and capabilities, and their sources of knowledge was collected directly from the firms themselves. This research will thus contribute to the debate on the localised nature of innovation/knowledge through a direct study of the different mechanisms of knowledge flow at the micro-level, and in particular of the impact of LKS upon firms' innovative performance. Since Audretsch and Feldman (2003) claimed that LKS are very important for the innovative performance of clustered firms compared to market-based knowledge flows, we seek to assess the significance of local knowledge spillovers (LKS) versus other mechanisms of knowledge flow for firms' innovative performance. According to NACE Rev. 1 72 (Tether et al. 2001, p101) computer services include hardware and software consultancy services, analysis, design and programming of ready to use software, data processing and database activities, as well as the maintenance and repair of office machinery. The computer service sector is a hybrid of technology providing, product oriented activities (such as the production of packaged software) that

9

3rd Globelics Annual Conference - Africa 2005

are very similar to manufacturing, as well as service oriented and technology using activities (such as data processing and database analysis). The bulk of the firms in Montevideo began their operations during the 1990s. This period is characterised by excess demand for technical products in Latin America. However, demand is not a sufficient factor that can explain the development of a software sector in Uruguay and not in the neighbouring Paraguay or Bolivia. The main rational behind the development of the software cluster in Montevideo was the presence of a group of well qualified professionals. Education has been constantly a priority of the Uruguayan State, and it has succeeded in achieving one of the lowest illiteracy levels8 in Latin America (World Development Indicators, 2002). These groups of professionals retained a hybrid type of knowledge; they hold technological knowledge and knowledge of a specific market (e.g. financial, health, construction etc.). Thus, they managed to respond to the increasing demand in Latin American markets. The Information Technology Industry in Uruguay is comprised by the following sections: Software Developers, Consultancy and Services, Internet and Data Transmission, Hardware and Sales. In total there are 2216 companies registered in the Uruguayan Chamber of Information Technologies (CUTI). The later is an institution that assists firms to develop their business capabilities and reinforces common action for the promotion of the Uruguayan software products in foreign markets. Audretsch and Feldman (1996) emphasised that LKS are present in knowledge intensive sectors. Therefore, to select the sample we have relied upon the criteria that firms have to be ‘knowledge intensive’. Among these four sections, the sub-sector of Software developers is the most knowledge intensive. Consequently, the initial decision was to include mainly firms of this section. Once on the field, at the suggestion of local researchers and software firms, we took the decision to include consultancy services as well. In total these firms mount to around 149 (Stolovich, 2003)9. We obtained a list of these firms from CUTI. However, after the first contact it became apparent that some of 8

Adult illiteracy rate is the percentage of people ages 15 and above who cannot read and write. In 1990 3% of the Uruguayan population could not read and write. This was the lowest rate among the Latin American Countries, followed by Argentina 4%, Cuba 5% and Chile 6%. A decade later Uruguay still holds the lowest illiteracy levels in Latin America; 2% is the illiteracy level in Uruguay, 3% in Argentina and Cuba and 4% in Chile (World Development Indicators, 2002). 9 These are the firms without considering one-person companies (1600), Hardware and sales (371) and Internet and data transmission firms (96).

10

3rd Globelics Annual Conference - Africa 2005

the firms that were placed in the aforementioned list were not carrying out any kind of software development. On the other hand, many other firms were included in the research after locating their contact information from the local telephone guide, and also from the initial interviews with firms. In total, we calculate that the number of firms that develop software and provide consultancy is approximately 150. For the sample to be representative of this heterogeneous population, the second criterion was to obtain a sufficient representation of all firms into the sample. The section of Software Developers and Consultancies in Montevideo provide a range of products and services, which are usually software applications in the form of a standard or customised product (Stolovich, 2003; Mejía and Rieiro, 2002; Failache et al., 2004). Figure 2 demonstrates a classification of the software products produced in Uruguay. The degree of standardisation of the product is the main factor that differentiates custom-made products from registered packages (Bitzer, 1997). The sale of a customised product is in the form of services; implementation or adaptation of products, ad hoc solutions provided at one point in time in the form of consultancy, maintenance and training. The sale of a registered package is the form of a product similar to manufacturing products. In particular software firms in Montevideo develop products to cover the needs of the financial market (banking, credit cards), the vertical market such as health, education, transport, and the horizontal market such as the management solutions for SMEs. Additionally, software firms develop tools that are used by other firms in the sector for their applications. Primary data were collected through a field study in Uruguay (October-December 2004). The research unit is the technologically dynamic cluster of software firms in Montevideo. The unit of analysis consists of the individual firms within the cluster. The survey is based to a certain degree on Community Innovation Survey and has also been adjusted to reflect the peculiarities of the software sector in a Developing country. A structured questionnaire was administered by means of face-to-face interviews with the director or/and the chief engineer of the R&D department of the companies. The whole population of 150 firms has been approached and asked to take part to the survey. Finally, 97 firms were willing to participate to the survey (this represents a 64 per cent response rate).

11

3rd Globelics Annual Conference - Africa 2005

4. Conceptual Model and Research Questions LKS is one of the mechanisms by which knowledge is diffused within a cluster. Knowledge may also be transferred through partners that are involved in a formal cooperative relation. In particular, this study focuses upon four mechanisms of knowledge flow: Intra-Cluster: 1. Local Knowledge Spillovers (LKS): Informal spontaneous and unpaid-for knowledge flow (Audretsch and Feldman, 1996; Allen, 1983; Lundvall, 1992). 2. Local Pecuniary Flow of knowledge (LPF): Agreed sharing of knowledge (through collaborations, etc) between actors within the cluster. Like LKS, these represent local knowledge flows, but they occur as a result of deliberate coordinated action and are not free (Zucker et al, 1998). Extra-Cluster: 3. International Knowledge Spillovers (IKS): Informal (spontaneous and unpaid-for) knowledge flows. Firms imitate each other through reverse engineering, attending trade fairs, following scientific or technical journals and, of course, through patent disclosures. 4. International Pecuniary Flow of knowledge (IPF): Agreed sharing of knowledge (through collaborations, etc) between firms in the cluster and actors outside of the cluster. Like non-local KS, they represent non-local knowledge flows, but they occur as a result of deliberate co-ordinated action and are not free. A first prerequisite towards operationalising this approach has been the appropriate comprehension of the firms’ mechanisms of learning in less developed countries. In other words, we tried to understand how firms acquire knowledge from external sources (interfirm learning) and how they process this knowledge internally (intra-firm learning). LKS is a type of knowledge, which flows when a firm attempts to learn from external sources (see figure 1). However, LKS are not the only mechanisms of knowledge flow identified in the existing literature. It is worth knowing which are the mechanisms of knowledge flow within a cluster for two main reasons. Firstly, the empirical identification and

12

3rd Globelics Annual Conference - Africa 2005

analysis of the different types of knowledge flow within a cluster will help us to understand the functioning of the mechanisms of LKS. As Audretsch et al (2003, p.13) admit ‘...there is no understanding of the way in which spillovers occur and are realised at the geographic level’. A micro study may be useful in distinguishing LKS from other mechanisms of knowledge flow. Secondly, thanks to this analysis, we will be in a favourable position to verify the presence of LKS, and also to assess their strengths in comparison to other types of knowledge flow. We address these issues by answering the following questions, throughout this article: Do local knowledge spillovers foster innovation? If so, to what extent do LKS occur? How important are they in comparison to other knowledge flows? And is the role played by LKS in developing country clusters in any way different from the function of LKS in high tech clusters in economically advanced countries?

5. Empirical Analysis and Results In order to examine to what extent LKS affect the innovative performance of the firms within the cluster we formulated two empirical models. First we discuss all variables and subsequently we present the two empirical models. The dependent variable measures the innovative performance of the firm. Dependent variables The indicators that have been used to denote the innovative performance of the firm consider product and service innovation while overlooking process innovation. Pavitt (1984) classified software firms as specialised suppliers. He argued that this type of firms are characterised by a high rate of product innovations. In addition, we have tested the survey with one European (based in the Netherlands) and one American (based in the Silicon Valley) software firms, which confirmed the tendency of software firms to undertake predominantly product innovations. Moreover, after the initial interviews in Uruguay and discussions with local experts, we verified the use of product and service innovation as the appropriate indicators for the innovative performance of the software firms in Uruguay. We have constructed two models that enable us to assess the research questions (see section 4. Table 1 displays the description of the variables that we use for the two models

13

3rd Globelics Annual Conference - Africa 2005

while table 2 presents the descriptive statistics. In the first model we use the variable New Market as an indicator of the innovative performance of the firm. It is a yes/no answer to the question: Did your firm develop product and/or service innovations that were technologically new in the market during the period 1999-2004? A logit model has been used to analyse the dependent variable New Market due to its dichotomous response outcomes10. In the second model we use the variable Innovation Output as an indicator of the innovative performance of the firm. This variable refers to the percentage of sales that derive from product and/or service innovations on 2003. This is a censored variable (i.e. lower limit equal to zero) and thus we use a Tobit model for its estimation.

Independent variables A. External Learning Firms use different mechanisms to acquire external knowledge. According to the Oslo Manual (OECD/EC/Eurostat, 1996) these learning mechanisms are the following: ¾ Formal and Informal Linkages between firms: These could be networks of small firms, user-producers relationships, relationships between competitive firms, relationships of firms with Universities or Research Institutes. These interactions can produce information flows, which can stimulate innovation (explicitly or implicitly). ¾ International links: Networks of international experts (epistemic communities, a-spatial, ‘invisible colleges, conferences) is an important channel through which frontier knowledge is transferred to less developed countries. ¾ Reverse engineering: Knowledge embodied in machinery involves knowledge flow. ¾ The degree of mobility of experts: An important part of knowledge is embodied in people. Labour mobility is used to measure knowledge flows, as the person who moves to another employer brings along the

10

Limited-dependent variable models i.e. Logit, Probit, Tobit are widely used in the field of econometrics the last 20 years (Madalla, 1983; Anemiya, 1981; Hosmer and Lemeshow, 2002).

14

3rd Globelics Annual Conference - Africa 2005

cumulated knowledge (human capital) he has acquired during years. High labour turnover may bring new knowledge/information into the firm. ¾ Spin-off company formation: Involves the transfer of knowledge from a University or a MNC to the newly emerged firm. This is a way by which new developments are commercialised. ¾ Codified knowledge in patents, the specialised press and scientific journals. ¾ Access to public R&D capabilities. ¾ The presence of expert technological ‘gatekeepers’ or receptors: These are individuals that continue to be aware of new developments and maintain personal networks, which facilitate flows of information. The advocates of Economic Geography (Audretsch and Feldman, 1996; Verspagen and Schoenmakers 2000) claim that a particular type of knowledge flow plays the predominant role in increasing firms’ innovative capability; namely local knowledge spillovers. In general, LKS arise when a firm takes advantage of the novelty of a close located firm, without compensation. We apply this definition to the various categories of knowledge flows that are identified in the literature. A knowledge flow, thus, is a LKS only when satisfies two conditions: ¾ It is knowledge that flows locally, and ¾ It is knowledge for which no compensation (pecuniary) is given. Taking into account this definition, only some of the aforementioned mechanisms of knowledge flow can be considered as LKS. ¾ Formal and Informal Linkages between firms: By definition a formal relationship such as contract agreement, R&D cooperation, licensing does not constitute LKS. This is because a formal relationship presupposes a type of pecuniary compensation for the acquisition of the knowledge.11 On the other hand, an informal relationship between actors (on the basis of reciprocity, trust,

11

Formal relationship may result to informal contact. This is the case of Rent spillovers. However, this is out of the scope of this research, which adopts the criterion of no compensation (pecuniary) to define Spillovers.

15

3rd Globelics Annual Conference - Africa 2005

belonging to the same epistemic community etc.) located within a close geographical distance represents a potential LKS. ¾ International links: By definition an international link does not constitute a LKS. ¾ Reverse engineering: Reverse engineering is not a mechanism of LKS because a product is not restricted in one location. It can be bought and consequently imitated in any R&D laboratory. ¾ The degree of mobility of experts: As far as the skilled employees move within the cluster we can consider them as channels of LKS. ¾ Spin-off company formation: As far as the spin-off occurs at the same area as the parent University/company this can be consider a channel of LKS. ¾ Codified knowledge in patents, the specialised press and scientific journals: Knowledge within patents does not constitute LKS since this information can travel in any location and is not restricted to space. ¾ Access to public R&D capabilities: If the cooperation of the firm with the public research institute is formal this does not constitute a LKS. On the contrary, if the cooperation is informal it could be a potential channel of LKS. ¾ The presence of expert technological ‘gatekeepers’ or receptors: Networking (informally) with these key agents constitutes a LKS. Consequently, there are three main mechanisms through which LKS arise: ¾ Spin-of of new firms: When a firm is a spin-off of a local actor (University, MNC, large firm) this implies, that crucial knowledge (know-how and problem solving skills) is circulated within the cluster. A person learns this knowledge within a University/MNC and then he creates his/her own firm. The identification of a firm that was a (local) spin-off indicates that knowledge has (at one point in time) spilled over from a University/MNC to a firm. ¾ Labour mobility:

16

3rd Globelics Annual Conference - Africa 2005

When a firm is characterised by high labour turnover, this implies that employees represent a channel for the acquisition of knowledge. At the same time, when the new employees originate from the cluster, it means that knowledge spills over locally through the mobility of labour. ¾ Interaction of local actors: When the most important source of knowledge for a firm is a local actor (University, Supplier/User, Competitor) this constitutes a significant channel for LKS. A prerequisite, for this would be that the interaction between the two actors is informal. Subsequently the independent variables, pertaining to the mechanisms of knowledge flow for the acquisition of external (to the firm) knowledge, are the following: 1. Intra-cluster: Local Knowledge Spillovers (LKS): LKS_S: This is a dummy variable that takes the value of 1 if a firm is a spin off of a university/MNC that is located within the cluster and the value of 0 in other case. LKS_L: This variable denotes the percentage of employees (inflow) of a firm that come from within the cluster during the last five years (1999-2004). LKS_I: Importance of intra-cluster information sources that is free. 2. Intra-cluster: Local Pecuniary Flow (LPF): LPF: Importance of intra-cluster information sources that is pecuniary. 3. Extra-cluster: International Knowledge Spillovers (IKS): IKS: Importance of extra-cluster information sources that is free. 4. Extra-cluster: International Pecuniary Flow (IPF): IPF: Importance of extra-cluster information sources that is pecuniary.

B. Internal Learning We use R&D as a proxy of the internal capability (or absorptive capacity) of the software firm. The proxy reflects the cumulative R&D effort of the firm during 1999-2004. In particular, this indicator measures in man-years the time that firms have spent on the research and development of their innovative products and/or services. This variable is denoted as R&D.

17

3rd Globelics Annual Conference - Africa 2005

C. Control Variables In both models we control for the age of the software firms. This is a continued variable – denoted as Age- with reference year 2004. In addition we control for the size of the firm. The variable Size is measured by number of employees on 2004. Firms were asked in the survey to assess the importance of various sources of knowledge for innovation in a Likert scale (1=unimportant, 2=less important, 3=important, 4=very important, 5=crucial). We provided them with thirteen different potential sources of knowledge (Group, New Personnel, Customers, Suppliers, Competitors, Alliances, Consultants, Research Institutes, Universities, Innovation Centres, Sector Institutes, Exhibitions, Electronic Information). Secondly, firms were requested to report where the sources of knowledge that they use were geographically located (Local, International). Thirdly, firms were asked to clarify the type of relationship between their firm and each source of knowledge that they use (Formal or Informal). Using the three attributes (Importance, Location and Type) we have constructed the different types of knowledge flow. For instance, International Pecuniary Flows of knowledge (IPF) were constructed in the following way: for every case (firm) we added up the level of importance of the various sources of knowledge that are acquired internationally in a pecuniary way. All the relations between firms and sources such as Group, New Personnel, Customers and Suppliers were classified as formal. Even though user-producer interaction is not a strictly pecuniary relation, still the knowledge flow between a firm and its supplier or customer is the result of a formal market transaction and thus it is treated as a pecuniary knowledge flow. In contrast all the relations between firms with Competitors are informal and thus considered to be knowledge spillovers. Likewise acquisition of Electronic Information is for free and thus considered as a spill over of knowledge. Finally, the relation of firms with sources of knowledge such as Alliances, Consultants, Research Institutes, Universities, Innovation Centres, Sector Institutes, Exhibitions, is ambiguous. For example, some firms form alliances in a formal way (e.g. sharing of R&D outcomes) while others keep it informal (e.g. sharing information regarding problem solving activities). A knowledge that flows between these ambiguous sources of knowledge and the firms can be pecuniary or free. Therefore, the type of knowledge flow between these

18

3rd Globelics Annual Conference - Africa 2005

sources of knowledge and the firm varies for each case and we considering it differently for each firm. Table 3 depicts a correlation matrix of all the variables.

Model 1 – Dependent variable; New Market The results12 of the first model13, displayed at Table 4, indicate that the variable local knowledge spillovers by interaction (LKS_I) has a significant and positive impact upon the innovative performance of the firm. Likewise, the variable local knowledge spillovers through labour mobility (LKS_L) has a significant and positive impact upon the innovative performance of the firm. In addition, the other type of knowledge flow – international pecuniary flow of knowledge (IPF) - has a significant and positive impact upon the innovative performance of the firm. Finally, the variable R&D intensity of the firm has a significant although negative impact upon the innovation of the firm. Consequently, to address the first question, these findings show that local knowledge spillovers foster the innovation of firms within clusters. Local knowledge spillovers occur very often and take the form of two main mechanisms: they are transferred through interactions and through the mobility of labour. In comparison to other knowledge flow LKS play a more important role in enhancing the innovation of firms with clusters.14 The magnitude of the effect of the variable that represents international pecuniary flow of knowledge –IPF- is obviously smaller than the impact of every one of the variables

12 13

The software program that has been used for the statistical analysis is SAS 9.1 for Windows. Univariate analysis for each variable shows that the variables LPF, R&D and Age have a p-value>0.25

and therefore, they are not included into the multivariable model. We proceed by forming the full Logit model. We examine the Wald statistic for each variable and we eliminate the variables Spin-off, IKS and Size for not providing significant reults (not presented here). We compare the new Logit model with the full Logit model using the likelihood ration test (G statistic). We test the null hypothesis that the excluded coefficients are equal to zero. The value of the log likelihood, of the first model, is -2LogL = 112,038. The log likelihood for the new model yields the log likelihood -2Log L = 114,522. Thus the value of the likelihood ratio test is G = 2,484. The p-value for the test, with 4 degrees-of-freedom, is P[X2(4)>2,484]>0,50. Thus, we accept the null hypothesis; the new model (presented at Table 4) is better than the full model. 14

For the comparison to be meaningful all the variables are standardised.

19

3rd Globelics Annual Conference - Africa 2005

representing local knowledge spillovers –local knowledge spillovers by interaction (LKS_I) and local knowledge spillovers through labour mobility (LKS_L). Finally, the variable which conveys the R&D intensity of the firm influences innovation in a negative way. This outcome contradicts theoretical views arguing that the internal capability of the firm and, thus, the effort spent in learning has a positive impact upon its innovative performance.15 However, taking into account the life cycle of high technology products created by small firms may elucidate this controversial finding. Oakey (1991) argues that during the innovation cycle of high technology products, R&D investments take place well in advance to their production. The subsequent stages of production and marketing are considered to be parts of the innovation cycle of such products. The R&D intensity variable represents the recent R&D activity of the firm. In particular, the R&D intensity variable is constructed by dividing the number of labour working in the R&D department with the total amount of labour in 2003. Thus, this variable takes into account the recent effort and human resources that the firm devoted to the building of its own capabilities. On the other hand, the innovative performance of the firm, measured by the variable New to the Market, considers the innovative outcome of the firm of the previous five years. The innovative performance of the previous five years is the result of prior building up and accumulation of capabilities. Thus, it is possible that the inconsistency with the capability theory that this result conveys, could be attributed to the fact that R&D intensity variable considers the effort of the firm at a point in time (2003) while the innovative performance (1999-2004) is the result of learning in prior years. On this ground, a negative relationship between R&D intensity and innovation could imply that the firm has already created its core competencies in the past and that it is currently focusing on other innovative activities (e.g. marketing). Model 2 – Dependent variable; Innovation Output Table 5 exhibits the results of the second model. The variables local knowledge spillovers through interaction (LKS_I) as well as local knowledge spillovers through labour mobility (LKS_L) have a significant and positive impact upon the innovative performance of the firm. Likewise, the variable R&D has a significant and positive impact upon the innovation of the firm. 15

This refers to capability theory (Lall 1987; 1992; Romijn, 1998; Cohen and Levinthal; 1990).

20

3rd Globelics Annual Conference - Africa 2005

This model yields similar results as the first model; the two variables that represent local knowledge spillovers - LKS_I and LKS_L - affect positively the innovative performance of the firm. Finally, the variable which conveys the R&D of the firm influences innovation in a positive way. This outcome supports the theory of capabilities. This happens because the variable R&D is measured in such a way as to capture the cumulativeness of firm’s capabilities for a period of five years. In particular, the R&D variable is constructed by adding the man years that firms have spent in R&D for the previous five years. Thus, this variable considers the recent effort and as well as the past human resources that the firm devoted to the building of its capabilities.

5. Conclusions Throughout this paper we have examined the importance of a particular mechanism of knowledge diffusion -namely local knowledge spillovers. Using new firm-level data of software firms concentrated in the cluster of Montevideo we tested the hypothesis of whether LKS affect the innovative performance of the firms within the cluster or not. The contribution of this paper is twofold; it examines a neglected though important cluster in Uruguay by means of a survey; and it sheds light upon local knowledge spillovers by examining the whole range of knowledge flow mechanisms. The results of the empirical analysis support the presence of local knowledge spillovers and their positive influence upon the innovation of firms within the cluster. In both models, LKS have a significant and positive impact upon innovation. Model 1 reveals that IPF affect innovation but its effect is smaller that the one of LKS. The other mechanisms of knowledge flow, namely LPF and IKS do not have a significant impact upon innovation. Furthermore, we examined three mechanisms by which knowledge spillovers occur: spin off, labour mobility, and LKS through interaction. The first mechanism –spin offdoes not produce significant results therefore, we cannot discern its impact upon innovation. The other two mechanisms – LKS through interaction and LKS through labour mobility- play the most important role for the innovation of firms with the cluster (model 1 and model 2).

21

3rd Globelics Annual Conference - Africa 2005

Our results demonstrate that LKS is an important mechanism through which knowledge circulates among actors within the cluster of Montevideo. Therefore, we conclude that the argument -found in theoretical and empirical works- that LKS foster innovation within high-tech clusters holds for our case in Uruguay. LKS occur in the context of software firms in Montevideo and are more important for innovation than the other types of knowledge flow. To the extent that it is possible to generalise based on a case-study we may conclude that the role played by LKS in the Latin America context is similar to that in high tech clusters in advanced countries. Policies supporting LKS encompass mainly government subsides towards Universities and firms that conduct substantial R&D. However, these policies have been strongly criticised because of the extensive role played by the governments and also because of the potential failure of the State to accommodate resources efficiently. Despite these criticisms, in the past developmental states have played an important role in enhancing the human capital and the capabilities of their countries, negotiating with international capitalists, attracting investment, and forwarding investment into potential dynamic sectors (Amsden, 2001; Kesidou, 2004). In the case of the software cluster in Uruguay, labour mobility and the non-pecuniary interaction of agents within the cluster seem to be the two most important mechanisms for the transfer of knowledge. Thus, it would be essential for the Uruguayan State to continue investing in education (formation of high skilled employees), to endow more subsidies for R&D, and to facilitate labour mobility by relaxing labour law, especially for SMEs. Awareness of the importance of LKS for innovation is crucial for drawing policies that enhance accumulation and circulation of knowledge. LKS can be a potential path of learning, innovation and thus of economic development for developing countries and in particular for small countries with many potentials.

22

3rd Globelics Annual Conference - Africa 2005

Appendix: Figure 1. Conceptual Framework: Mechanisms of Knowledge Flow

CLUSTER Internal Learning Absorptive capacity Intra-firm Learning Indicator: R&D man-years R&D intensity

NATIONAL or INTERNATIONAL Context

FIRM

External learning

Innovative Performance Indicators: New Product Innovation Output

Intra-cluster Knowledge Flow Local Knowledge Spillover Local Pecuniary Flow Extra-cluster Knowledge Flow International Knowledge Spillover International Pecuniary Flow

Indicator: IKS

Indicator: IPF

23

Indicator: LKS_S LKS_L LKS_I

Indicator: LPF

LKS

3rd Globelics Annual Conference - Africa 2005

Figure 2: Classification of Software products in Uruguay

DEGREE of Standardisation

Low High Customised Software

Standardised Systems, Registered packages

TYPE of Software in Uruguay – APPLICATION SOFTWARE

Services: • Implementation and adaptation of a product (their own or third party). • Maintenance • Training • Consultancy



• • •

24

Horizontal Market: Management solutions for SMEs (accounting, human resources) Vertical Market: (Education, Health, Transport) Development Tools Financial market

3rd Globelics Annual Conference - Africa 2005

Table 1: Summary Variables Variable Name Dependent Variables New Market

Definition Dummy variable, which takes the value of =1 when the firm undertook innovation which is new to the market, and =0 when the firm carried out innovation which is new to the firm. It measures the innovation of firms during the period 1990-2004.

Innovation Output

Indicates the percentage of sales that derives from new product innovation. It measures the innovation of firms in the year 2003.

Independent Variables LKS_L

This is measured by the Inflow Rate: R(in)t = Σ imt-1 /Nt.. Where im = 1 when person has changed status from preceding years, 0 when not, N = number of persons i.e. stock in year t (STEP, 2003). It is a continuous variable which measures the labour turnover of the firm during the period 1999-2004.

LKS_S

Dummy variable takes the value =1 when a firm is a spin-off, and =0 in other case. One or more partners of the current firm acquired their experience by working in another similar firm.

LKS_I

Local Knowledge Spillovers through interaction. This is a constructed variable that indicates the importance of knowledge that flows locally in a non-pecuniary way, from various sources. To construct it, for every case (firm) we added the level of importance of the various sources of knowledge which are acquired locally in a free manner.

LPF

Local Pecuniary Flows of knowledge. This is a constructed variable that indicates the importance of knowledge that flows locally in pecuniary way, from various sources. To construct it, for every case we added the level of importance of the various sources of knowledge which are acquired locally in a pecuniary way.

IKS

International Knowledge Spillovers. This is a constructed variable that indicates the importance of knowledge that flows to the cluster from abroad, in a nonpecuniary way. To construct it, for every case (firm) we added the level of importance of the various sources of knowledge which are acquired internationally in a free manner.

IPF

International Pecuniary Flows of knowledge. This is a constructed variable that indicates the importance of knowledge that flows to the cluster from abroad, in a non-pecuniary way. To construct it, for every case we added the level of importance of the various sources of knowledge which are acquired internationally in a pecuniary way.

R&D

R&D effort measured in man-years. It measures the cumulative R&D effort of the firm for innovative products during the period 1999-2004.

R&D intensity Control Variables Age

The percentage of total labour that conducted R&D in 2003.

Size

Size of the firm measured by number of employees at year 2004.

Firm’s age (reference year 2004). Continuous variable.

25

3rd Globelics Annual Conference - Africa 2005

Table 2: Descriptive Statistics Variables New Market Innovation Output LKS_L LKS_S LKS_I LPF IKS IPF R&D R&D intensity Age Size

N 97 97 97 97 97 97 97 97 97 97 97 97

Minimum 0 0 0 0 0 0 0 0 0 0 1.00 1.00

Maximum 1.00 1.00 3.00 1 16.00 22.00 16.00 18.00 59.00 2.50 138.00 260.00

26

Mean 0.51 0.44 0.38 0.48 6.09 9.10 5.93 5.40 10.41 0.36 13.88 23.98

Std. Deviation 0.50 0.36 0.42 0.50 3.90 4.51 3.79 4.36 10.24 0.36 15.81 40.29

3rd Globelics Annual Conference - Africa 2005

Table 3: Correlation Matrix of the main variables

New Market Innovation Output LKS_L LKS_S LKS_I LPF IKS IPF R&D R&D intensity Age Size

New Market 1.00

Innovation Output 0.028

KLS_L

LKS_S

KLS_I

LPF

IKS

IPF

R&D

0.299***+

0.146

0.229**+

0.031

0.216**+

0.187*+

1.00

0.218**

0.090

0.106

0.059

0.043

1.00

0.029 1.00

-0.098 0.208**+ 1.00

-0.080 0.204**+ 0.198* 1.00

0.058 0.020 0.055 -0.189* 1.00

Note: *** Statistically significant at the 1% level. ** Statistically significant at the 5% level. * Statistically significant at the 10% level. + Spearman Correlation Coefficient.

27

Age

Size

0.130

R&D intensity -0.143

-0.148

0.275***+

0.108

0.269***+

0.077

0.199**+

0.211**+

0.213** -0.092 -0.153 0.056 0.229** 1.00

0.120 0.041 -0.209** 0.170* -0.064 0.397*** 1.00

0.320*** -0.003 -0.004 0.015 -0.163 0.002 0.060 1.00

-0.094 -0.357***+ 0.078 -0.002 0.178* 0.053 0.259**+ 0.006

0.327***+ 0.025 -0.018 -0.067 0.142 0.360*** 0.433***+ -0.243**

1.00

0.345***+ 1.00

3rd Globelics Annual Conference - Africa 2005

Table 4: Binary Logit Model -dependent variable New Market (n=97) Variable

Coeff.

Std.Err.

LKS_I IPF LKS_L R&D(intensity) Constant

2.821 1.751 4.597 -4.550 -1.457

1.0243 0.9938 1.8971 1.8489 0.6147

Wald Chi - Square 7.58 3.10 5.87 6.05 5.62

Pr > ChiSq 0.005 0.078 0.015 0.013 0.017

Table 5: Tobit Model -dependent variable Innovation Output (n=97) Variable LKS_I LKS_L R&D Constant

Coeff. 0.371 0.619 0.626 0.082

Std.Err. 0.1715 0.2872 0.2402 0.1035

28

t value 2.16 2.16 2.61 0.79

Pr > | t | 0.030 0.030 0.009 0.427

3rd Globelics Annual Conference - Africa 2005

REFERENCES • • • • •

Abramovitz M. (1986) Catching Up, Forging Ahead, and Falling Behind, Journal of Economic History, 46 (2), pp. 385-406. Allen R. C. (1983) Collective Invention, Journal of Economic Behavior and Organization, Vol. 4, pp. 1-24. Amsden A H (2001) The Rise of “the Rest”: Challenges to the West from LateIndustrializing Economies, Oxford: Oxford University Press. Anemiya T. (1981) 'Qualitative Response Models: A Survey' Journal of Economics Literature, 19, pp. 481-536. Audretsch D. and Feldman M. (1996) R&D Spillovers and the Geography of Innovation and Production, The American Economic Review, Vol. 86, No. 3.



Audretsch D. and Feldman M. (2003) Knowledge Spillovers and the Geography of Innovation, Prepared for the Handbook of Urban and Regional Economics, Volume 4.



Aydalot P. (1986) Presentation in Aydalot P. (ed) Milieux innovateurs in Europe Paris, Gremi (pp. 9-14). Becattini G. (1990) The Marshallian industrial district as a socio-economic notion in Pyke F., Becattini G., Sengenberger W. (eds.), Industrial Districts and Inter-Firm Co-operation in Italy, Geneva. ILO (pp.37-51). Bell M. (1984), Learning and the Accumulation of Industrial Technological Capacity in Developing Countries, in Fransman M. and King K. (eds.) Technological Capability in the Third World, Macmillan Press, Hong Kong. Bell and Pavitt (1993), Technological Accumulation and Industrial Growth: Contrasts between Developed and Developing Countries, Industrial and Corporate Change, Vol. 2. Bitzer J. (1997) The Computer Software Industry in East and West: Do Eastern European Countries Need a Specific Science and Technology Policy? Deutsches Institut fur Wirtschaftsforschung (DIW). Discussion Paper N 149. Berlin, May 1997. Breschi S. and Lissoni F. (2001) Knowledge Spillovers and Local Innovation Systems: A Critical Survey, Liuc Papers, N. 84, Serie Economia e Impresa. Camagni R. (1991) Innovation Networks: Spatial Perspectives (Ed), Belhaven Press, London/New York. Caniels M. (1999) Regional Growth Differentials: The Impact of Locally Bounded Knowledge Spillovers, PhD dissertation, University of Maastricht. Cohen W.M. and Levinthal D.A. (1990), Absorptive capacity: a new perspective on learning and innovation, Administrative Science Quarterly, 35. Cooke P. (2001) Regional Innovation Systems, Clusters, and the Knowledge Economy, Industrial and Corporate Change, vol. 10, N. 4, pp. 945-974. Cowan R., David P.A. and Foray D. (2000) The Explicit Economics of Knowledge Codification and Tacitness, Industrial and Corporate Change, Vol. 9, N. 2. Dosi G. (1988) Sources, procedures and microeconomic effects of innovation, Journal of Economic Literature, 26, 1120-1171. Failache C., Muinelo L., Hounie A. (2004) Estudio de Competitividad Sectorial: Technologías de la Información en Uruguay, [Sectoral Competitiveness Study: Information Technology in Uruguay], Banco Interamericano de Desarrollo. Griliches Z., (1992), The Search for R&D Spillovers, Scandinavian Journal of Economics, Blackwell Publishing, vol. 94(0), pages S29-47. Hosmer W. D. and Lemeshow S. (2002) Applied Logistic Regression, Second Ed. Jon Wilely & Sons, New York.

• • • • • • • • • • • • • •

29

3rd Globelics Annual Conference - Africa 2005 • •

• • • • • • •

• • • • • • • • • • • • • •

Jaffe A. B., Trajtenberg M. and Henderson R. (1993), Geographic Localization of Knowledge Spillovers as Evidence by Patent Citations, Quarterly Journal of Economics, August 1993, 63 (3), pp. 577-98. Kesidou E. (2004) 'The Political Economy of Technology Policy: The Automotive Sector in Brazil (1950-2000)' in: S. Mani and H. Romijn (Eds), Innovation, Learning and Technological Dynamism of Developing Countries, Paris: United Nations University Press, 2004. Lall S. (1987) Learning to Industrialize: The Acquisition of Technological Capability by India, London: Macmillan and New York: St. Martin's Press, 1987. Lall S. (1992) Technological Capabilities and Industrialization World Development, Vol. 20, No. 2, pp. 165-186. Lall, S. (2000), “The Technological Structure and Performance of Developing Country Manufactured Exports, 1995-1998, Oxford Development Studies, Vol. 28, No. 3, pp 337-369. Lundvall Bengt-Ake (1988) ‘Innovation as an interactive process: from user-producer interaction to the national system of innovation’, cited in Dosi, Freeman Nelson et al (1988), ‘Technical Change and Economic Theory’, Pinter Publishers, London and New York. Lundvall B. A. (1992) National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning. Pinter, London. Maddala G. S. (1983) Limited-Depedent and Qualitative Variables in Econometrics, Cambridge University Press. Mejía T. and Rieiro M. (2002) Aprendizaje, Innovación y Cometitividad de la Industria del Software en Uruguay, [Learning, Innovation and the Competitiveness of the Software Industry in Uruguay], Trabajo de investigacion monografica. Universidad de la republica (UDELAR). Marshall A. (1920) Principles of Economics, MacMillan. Morgan Kevin (1997) The Learning Region: Institutions, Innovation and Regional Renewal Regional Studies Vol. 31, No. 5, pp. 491-503. Nelson, R. and S. Winter (1982) An Evolutionary Theory of Economic Change, Harvard University Press, Cambridge, MA. Nonaka I. 1994 A Dynamic Theory of Organizational Knowledge Creation Organizational Science, Vo. 5, No. 1. Oakey R. (1991), Government policy towards high technology: Small firms beyond the year 2000, in Curran & Blackburn (eds.), Paths of Enterprise, Routledge; 1991. OECD/EC/Eurostat (1996), Oslo Manual, second edition. Pavitt (1984), Sectoral patterns of technical change: Towards a taxonomy and a theory, Research Policy 13, 343-373. Pietrobelli C. and Rabellotti R. (2004) Upgrading in Clusters and Value Chains in Latin America: The Role of Policies, Inter-American Development Bank, Washington, D.C. Piore M. and Sabel C. (1984) The Second Industrial Divide: Possibilities for Prosperity, New York, Basic Books. Porter M. (1990) The competitive Advantage of Nations. Macmillan, London. Romer Paul, Increasing Returns and Long Run Growth, Journal of Political Economy 94, October 1986, 1002-37. Romer M., (1990), Endogenous Technological Change, Journal of Political Economy, University of Chicago Press, vol. 98(5), pages 71-102. Romijn H. (1998) Acquisition of Technological Capabilities in Small Firms in Developing Countries, Macmillan, Basingstoke. Rosenstein-Rodan P. (1943), Problems of Industrialization of Eastern and Southeastern Europe, Economic Journal 43, (June-September 202-211).

30

3rd Globelics Annual Conference - Africa 2005 • • • • • • • • • •

Saxenian A. (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Harvard University Press, Cambridge, MA. Schmitz Hubert (1999) Collective Efficiency and Increasing Returns Cambridge Journal of Economics, Vol. 23, pp. 465-483. Scitovsky T. (1954) Two concepts of external economies, in Arrow K. and Scitovsky T. (Eds) Readings in Welfare Economics (1969) Allen and Unwin LTD, London. Stolovich, L. (2003), What does the data on Uruguay’s Information Technology Industry indicate, CUTI/PASS, mimeo, Montevideo. Storper M. and Scott A. J. (1998) The Geographical Foundations and Social Regulation of Flexible Production Complexes, in Wolch J. and Dear M. (Eds) The Power of Geography Allen and Unwin, London. Tether B. S., Miles I., Blind K., Hipp C., de Liso N. and Cainelli G. (2001) Innovation in the Service Sector: Analysis of data collected under the Community Innovation Survey (CIS-2). Van Dijk M. (2005) Industry Evolution and Catch Up: The Case of the Indonesian Pulp and Paper Industry, PhD Thesis, Eindhoven University Press. Verspagen B. and Schoenmakers W. (2000) The spatial dimension of knowledge spillovers in Europe:Evidence from patenting data, Paper presented at the AEA Conference on Intellectual Property Econometrics, Alicante, 19-20 April. Westphal Larry E. (1990) Industrial Policy in an Export-Propelled Economy: Lessons from South Korea’s Experience, Journal of Economic Perspectives, Vol 4, No. 3, 41-59. Zucker L. G., Darby M. R. and Brewer M. B. (1998), Intellectual human capital and the birth of US biotechnology enterprises, American Economic Review, 8 (1), pp. 290-306.

*

Eindhoven Centre for Innovation Studies (Ecis), Faculty of Technology Management, Eindhoven University of Technology (TU/e), P.O.Box 513, 5600 MB Eindhoven, the Netherlands. Tel: +31 40 2475095; Fax: +31 40 2474646. E-mail: [email protected]

Eindhoven Centre for Innovation Studies (Ecis), Faculty of Technology Management, Eindhoven University of Technology (TU/e), P.O.Box 513, 5600 MB Eindhoven, the Netherlands. Tel: +31 40 2474026; Fax: +31 40 2474646. E-mail: [email protected]

31