Abstract

The aim of this article is to investigate research and development (R&D) networks created under the scope of the R&D Programme implemented by the Brazilian Electricity Regulatory Agency (2006-2008). The main contribution of this article is its empirical analysis, drawing upon social network analysis and georeferencing to support the visualisation of the networks, as well as identify the location of actors and interactions in the Brazilian territory. We identified the features of the interactions with regards to their geographical location and distribution in the territory, as well as the main actors and their roles in the creation and circulation of knowledge based on R&D activities, and whether geographical proximity was present in the interactions. The results show that the networks are hierarchical, and electricity companies show a leading role in the governance of knowledge creation and circulation over external research organizations and project coordinators. Such companies interacted mainly with researchers based in regions where the capacity of research infrastructure has been consolidated.

Keywords:
ANEEL R&D; Social Networks Analysis; Georeferencing; Geographical Proximity

Resumo

O objetivo deste estudo é investigar as redes de P&D formadas no escopo do Programa de Pesquisa e Desenvolvimento da Agência Nacional de Energia Elétrica (Aneel) para o triênio de 2006 a 2008. A principal contribuição do artigo é empírica. São utilizadas métricas de análise de redes sociais e georreferenciamento para visualização da rede e mapeamento das interações no território nacional. Apresenta-se o perfil das interações quanto à sua localidade geográfica e distribuição no território e identificam-se os principais atores na criação e circulação do conhecimento a partir das atividades de P&D e quando a proximidade geográfica estava presente na criação de tais interações. Os resultados mostram que as redes são hierarquizadas e as concessionárias de energia elétrica exercem liderança na governança da criação e circulação de conhecimento sobre organizações de pesquisa extramuro e sobre os coordenadores de projetos. Estas criaram interações com pesquisadores baseados nas regiões do país que oferecem infraestrutura de pesquisa consolidada.

Palavras-chave:
P&D Aneel; Análise de Redes Sociais, Georreferenciamento; Proximidade; Geografia da Inovação

1. Introduction

The Brazilian energy matrix presents a high percentage of renewable energy sources (41.2% in 2015) compared to other countries (e.g., 9.4% in OECD countries) (BRASIL, 2016). Considering renewable sources, sugar cane derivatives present the main share of internal energy supply (16.9%) followed by hydropower (11.3%). Oil and its derivatives and natural gas (non-renewable sources of energy) were the main sources of the Brazilian energy supply in 2015 (37.3% and 13.7%, respectively). However, domestic electricity supply relies mainly on hydropower (58.4% in 2015) (BRASIL, 2016). Brazil is considered a low-carbon emission country due to its energy matrix and renewable electric energy sources: 3.6% and ranked 6th place globally, compared to the USA (17.9%), China (14.9%), Russia (5.8%), Indonesia (5.3%) and Japan (4%) (IPCC, 2014).

In this context, the Brazilian Electricity Regulatory Agency (ANEEL)¹ 1 In Portuguese Agência Nacional de Energia Elétrica (ANEEL). fosters research and development (R&D) investment in the national electricity sector with the aim to solve new market and technological challenges, including carbon emission mitigation (Pompermayer, 2009POMPERMAYER, F. M. Desafios e perspectivas para a inovação tecnológica no setor de energia elétrica. Revista de Pesquisa e Desenvolvimento da ANEEL, n. 3, p. 11, 2009.). Among such challenges, one finds support for the increase of electricity production from renewable sources other than hydropower, such as wind and solar energy.² 2 For a discussion on the Brazilian wind energy trajectory, see Camillo (2013).

Considering the structure of the Brazilian electricity sector, its privatisation and economic deregulation in the 1990s (FURTADO, 2010_________. O sistema setorial de inovação do setor elétrico brasileiro e o CTEnerg. Campinas: Instituto de Pesquisa Econômica Aplicada (Ipea), 2010.), ANEEL implemented its Research and Development Programme in 2000 (hereon ANEEL R&D) with the aim to foster innovation activities in the sector (POMPERMAYER et al., 2011POMPERMAYER, F. M. et al. Rede de pesquisa formada pelo programa de P&D regulado pela ANEEL: abrangência e características. In: POMPERMAYER, F. M.; DE NEGRI, F.; CAVALCANTE, L. R. (org.). Inovação tecnológica no setor elétrico brasileiro: uma avaliação do programa de P&D regulado pela Aneel. Brasília: Ipea, 2011. p. 13-). The program was implemented in accordance with Law n. 9.991/2000³ 3 This law was amended by Laws n. 10.438/2002, 10.848/2004, 11.465/2007, 12.111/2009 and 12.212/2010. and requires firms that operate in generation (G) and transmission (T) to invest 1% of their annual net operational revenue in R&D activities (Aneel, 2012a). Electric power distribution (D) firms are required to invest 0.75% of their annual net revenue in R&D activities and 0.25% in energy efficiency (Aneel, 2012a: 9).

The main premise of ANEEL R&D Program is that R&D activities will foster interactions among key actors in the national electricity sector (ANEEL, 2012b), such as firms, research institutes (public and private), and local universities, and therefore improve systemic interactions within the sector (EDQUIST, 2004EDQUIST, C. Systems of innovation: perspectives and challenges. In: FAGERBERG, J.; MOWERY, D. C.; NELSON, R. R. (ed.). The Oxford handbook of innovation. New York: Oxford Universty Press, 2004. p. 181-208.). Such interactions potentially foster: knowledge and learning exchange among actors; technological development; and improvement of local capacities for technological, market and regulatory challenges at the national and global levels. The structure of these interactions is defined by this study as networks (Burt, 1992BURT, R. S. Structural holes: the social structure of competition. Cambridge, MA: Harvard University Press, 1992.).

ANEEL R&D regulation is defined by the Research and Development Program Handbook (ANEEL, 2008), which provides information on the criteria for the eligibility of projects submitted by GTD firms. Projects should present goals and outcomes aligned to (mainly) applied research activities; project implementation and results are also assessed by ANEEL. Misconduct by project grantees (e.g, not performing activities defined as R&D) may imply in penalties and fines by the Agency (ANEEL, 2008). The interactions created by GTD firms to implement approved projects are characterized by competition between research consortia for financial resources.

Studies on ANEEL R&D (POMPERMAYER et al., 2011POMPERMAYER, F. M. et al. Rede de pesquisa formada pelo programa de P&D regulado pela ANEEL: abrangência e características. In: POMPERMAYER, F. M.; DE NEGRI, F.; CAVALCANTE, L. R. (org.). Inovação tecnológica no setor elétrico brasileiro: uma avaliação do programa de P&D regulado pela Aneel. Brasília: Ipea, 2011. p. 13-) show that science and technology organisations are the most frequently accessed actors within the electricity sectoral system by applicant firms, and that only about 10% of performing organisations are firms ‘mainly related to the electricity sector’. These studies suggest a potential weakness in the development of the Brazilian electric sector because the latter is supplier-dominated (JACQUIER-ROUX; BOURGEOIS, 2002JACQUIER-ROUX, V.; BOURGEOIS, B. New Networks of technological creation in energy industries: reassessment of the roles of equipment suppliers and operators. Technology Analysis & Strategic Management, v. 14, n. 4, p. 399-417, 2002.Available in: <http://dx.doi.org/10.1080/0953732022000028791>.
http://dx.doi.org/10.1080/09537320220000... , p. 416-417), and firm-firm interactions could potentially improve innovation performance at the country level. In addition, ANEEL R&D interactions are mostly concentrated in the Brazilian Southeast region, highlighting the importance of geographic proximity (BOSCHMA, 2005_________. Editorial: Role of proximity in interaction and performance: conceptual and empirical challenges. Regional Studies, v. 39, n. 1, p. 41-45, 2005.) for interactions between actors involved in R&D activities (discussed in section 2).⁴ 4 According to Boschma (2005), there are other types of proximities for the interactions. This study aims to investigate only the relevance of geographic proximity.

This study aims to investigate networks (structure of interactions) created within the ANEEL R&D Programme for the period from 2006 to 2008.⁵ 5 It is relevant to mention that this study applied to the Brazilian Access to Information Law (application number 48700001493201791) aiming to obtain access to complementary data about firms that perform research and research coordinators for each ANEEL R&D project for all years that the programme has been undergoing. However, ANEEL provided only data that were freely accessed on their website and already used in this study. According to ANEEL, data on firms and research actors by project are not compiled and therefore could not be disclosed. We consider the organisational (i.e., applicant firms and other performing organisations) and individual (i.e., project coordinators and applicant firms) perspectives. Studies on R&D and innovation networks in Latin America were scarce until the beginning of the 2010s (CASAS; LUNA, 2011CASAS, R.; LUNA, M. De redes y espacios de conocimiento. Significados conceptuales y de política. In: ARELLANO, A.; KREIMER, P. (ed.). Estudio social de la ciencia y la tecnología desde América Latina. Bogotá: Siglo del Hombre Editores, 2011.; PAMPLONA DA COSTA, 2012PAMPLONA DA COSTA, J. O. Technology Policy, Network Governance and Firm level Innovation in the Software Industry: a Comparison of Two Brazilian Software Networks, Science and Technology Policy Research (SPRU). University of Sussex, Brighton, p. 299, 2012.). Since then, there has been renewed interested in network studies in Latin America, contributing to the understanding of innovation systems and knowledge production in the region. This study aims to contribute to the understanding of the role of R&D networks in Brazil. Hence, we investigate networks created under the ANEEL R&D Programme, their structures, the role of participating actors and their geographical distribution. We assume that R&D activities are knowledge intensive and that R&D (and innovation) networks are a subsystem of the innovation system (CANTNER; GRAF, 2010CANTNER, U.; GRAF, H. Growth, development and structural change of innovator networks: the case of Jena. In: BOSCHMA, R.; MARTIN, R. (ed.). The handbook of evolutionary economic geography. Cheltenham: Edward Elgad, 2010. p. 370-387.). The study proposes the following research question: How are the actors of the innovation system of the Brazilian electric sector structured, in terms of their R&D activities and within the scope of government policies which aim at the creation of networks, such as the ANEEL R&D Programme?

Although some studies have contributed to the understanding of interactions among ANEEL R&D actors using aggregate data (POMPERMAYER et al., 2011POMPERMAYER, F. M. et al. Rede de pesquisa formada pelo programa de P&D regulado pela ANEEL: abrangência e características. In: POMPERMAYER, F. M.; DE NEGRI, F.; CAVALCANTE, L. R. (org.). Inovação tecnológica no setor elétrico brasileiro: uma avaliação do programa de P&D regulado pela Aneel. Brasília: Ipea, 2011. p. 13-), this study created a database of ANEEL R&D projects (see section 3) for each grantee firm in the period from 2006 to 2008. We employ social network analysis (SNA) to investigate network structure and characteristics (DE NOOY; MRVAR; BATAGELJ, 2005DE NOOY, W.; MRVAR, A.; BATAGELJ, V. Exploratory social network analysis with Pajek. Cambridge: Cambridge University Press, 2005.) and therefore delve into the results obtained from previous studies on ANEEL R&D Programme.

We draw upon the geography of innovation approach (FELDMAN, 1994FELDMAN, M. P. The geography of innovation. Boston: Kluwer Academic Publishers, 1994.; BOSCHMA; MARTIN, 2010BOSCHMA, R.; MARTIN, R. Introduction - The new paradigm of evolutionary economic geography. In: BOSCHMA, R.; MARTIN, R. (ed.). The handbook of evolutionary economic geography. 1. ed.. Cheltenham: Edward Elgar, 2010. p. 3-39.) and geoprocessing tools to provide the geospatial analysis of the networks. The cartographic representation of the networks allows the visualisation of actors’ geographical localisation, interactions driven by geographical proximity and at the regional level. The use of point mapping and network mapping indicates that partnerships among the actors outweigh geographical barriers and motivate the process of specialization and competition between regions in attracting investments.

Considering previous studies, we expected ANEEL R&D networks to show low density, potentially limiting knowledge circulation among participating actors. Our results confirm such characteristics using SNA. Hence, our main contribution to the literature on the Brazilian electric sector is empirical, we reveal the structure of the network and the role of geographical proximity. The study brings evidence on the sectoral system of innovation (SSI) of the Brazilian electric sector (MALERBA, 2002MALERBA, F. Sectoral Systems of innovation and production. Research Policy, v. 31, n. 2, p. 247-264, 2002. Available in: <http://www.sciencedirect.com/science/article/B6V77-459H02Y-5/2/805c2d6a3af53b43a33f2a221520a7f0>.
http://www.sciencedirect.com/science/art... ), enabling the visualization of interactions among actors which may potentially foster shared learning.

The article is organised as follows, in addition to this Introduction. Section 2 presents the electric sector and the ANEEL R&D Programme under the analytical framework of innovation systems; section 3 presents the methodology employed in the study. Section 4 presents and discusses the data and Section 5 concludes the article pointing to the main contribution, limitations and policy implications of the study.

2. The Brazilian electric sector and the ANEEL R&D Programme

2.1. The sectoral approach and geographic proximity

The electric sector employs mature technologies related to coal, oil, natural gas and nuclear fusion. Recently, emerging technologies such as wind, wave and tidal energy (Foxon et al., 2005FOXON, T. J. et al. UK innovation systems for new and renewable energy technologies: drivers, barriers and systems failures. Energy Policy, v. 33, n. 16, p. 2123-2137, 2005. Available in: <http://www.sciencedirect.com/science/article/pii/S030142150400120X>.
http://www.sciencedirect.com/science/art... ) renewed the sector.⁶ 6 Hydroelectricity is a renewable source of energy generation, but hydropower technology is considered mature when related to large hydroelectric power plants. The global supply of primary energy (86.6%) and electricity generation (80.5%) in 2009 were based on non-renewable sources (IEA, 2011, p. 6, 24). Brazil shows a different path but faces technological challenges. Figure 1 presents features and barriers to the development of the electric sector in developing countries, as discussed below.

Multinational companies (MNCs) are the main actors of the SSI in electric energy generation and mainly based in developed economies (JACQUIER-ROUX; BOURGEOIS, 2002JACQUIER-ROUX, V.; BOURGEOIS, B. New Networks of technological creation in energy industries: reassessment of the roles of equipment suppliers and operators. Technology Analysis & Strategic Management, v. 14, n. 4, p. 399-417, 2002.Available in: <http://dx.doi.org/10.1080/0953732022000028791>.
http://dx.doi.org/10.1080/09537320220000... ; FURTADO, 2008FURTADO, A. T. O arranjo institucional da inovação no setor elétrico brasileiro. Campinas: GEOPI/DPCT. Campinas, março 2008; ROGGE; HOFFMANN, 2010ROGGE, K. S.; HOFFMANN, V. H. The impact of the EU ETS on the sectoral innovation system for power generation technologies - Findings for Germany. Energy Policy, v. 38, n. 12, p. 7639-7652, 2010. Available in: <http://www.sciencedirect.com/science/article/pii/S030142151000577X>.
http://www.sciencedirect.com/science/art... ). Power generation equipment suppliers develop most of the new technologies in the sector and are responsible for interactions with organizations that create new knowledge (JACQUIER-ROUX; BOURGEOIS, 2002). This feature of the sector is a potential barrier for R&D collaboration between equipment suppliers and firms based in developing countries, the geographical distance may restrict user-supplier interactions. For instance, GTD firms based in developing countries may have lower benefits related to knowledge spillovers from suppliers’ technological capacities.⁷ 7 Local conditions in developing countries should be considered for local technological development, such as technological training and industrial production capacity (BELL; PAVITT, 1993). To overcome user-supplier limitations for local technological development in Brazil, the ANEEL R&D Programme aimed to foster interactions among local actors for R&D activities.

The main components and actors of the Brazilian electric SSI are: governmental organizations, firms, universities, vocational education organizations, public and private research organisations, non-governmental organisations and associations; and intermediary actors. Figure 2 presents the SSI of the electric sector and the focus of this study.

Innovation presents a geographic dimension related to the locality of SSI actors (THOMAS, 1985THOMAS, M. Regional economic development and the role of innovation and technological change In: OAKLEY., A. T. T. A. R. P. (ed.). The regional economic impact of technological change. New York: St. Martins Press, 1985.). This study highlights the importance of a geographic analysis regarding interactions among actors of the Brazilian electric SSI and we consider the regional scale of analysis (i.e., regions of the country). From the innovation perspective, proximity (potentially) favours personal interaction in R&D projects and enables cognitive-based relationships and trust building among actors (STORPER; VENABLES, 2001STORPER, M.; VENABLES, A. J. Buzz: the economic force of the city. In: INTERNATIONAL SEMINAR ON ECONOMY AND SPACE. Proceedings... Ouro Preto, MG: Faculdade de Economia, Cedeplar, Universidade Federal de Minas Gerais, 2001.). It can also promote learning interactions between R&D organisations, suppliers and technology users. Both the organisational and individual dimensions of interactions are developed in network arrangements.

With regards to R&D and innovation (R&D&I) activities and the role of geographic location (BOSCHMA; MARTIN, 2010BOSCHMA, R.; MARTIN, R. Introduction - The new paradigm of evolutionary economic geography. In: BOSCHMA, R.; MARTIN, R. (ed.). The handbook of evolutionary economic geography. 1. ed.. Cheltenham: Edward Elgar, 2010. p. 3-39.), geographic proximity is assumed to have a central role in the choices of actors for the creation of partnerships (for both organisational and individual relations). According to Feldman (1994FELDMAN, M. P. The geography of innovation. Boston: Kluwer Academic Publishers, 1994.), individual proximity may enable information exchange and reduce the risks and costs of innovation activities. At the organisational level, product innovation relates to the technological infrastructure available, and the strategic location of knowledge sources strengthens knowledge inputs that can be crucial for the competitive advantages of a given location or region. Geographic proximity enables tacit knowledge exchange and the creation of trusted interactions among actors (BOSCHMA, 2005).

Networks materialize in the “geographical space” (SANTOS, 1996SANTOS, M. A natureza do espaço. São Paulo: Hucitec, 1996.) and cause economic, political and social changes. Mapping the proximity of relations is central to the analysis of the geography of innovation (FELDMAN, 1994FELDMAN, M. P. The geography of innovation. Boston: Kluwer Academic Publishers, 1994.) to explain phenomena occurred in the geographic space and their relation with technological development, where innovation sources and knowledge flows materialize and create (un)expected results (SAXENIAN, 1994SAXENIAN, A. Regional advantage: culture and competition in Silicon Valley and Route 128. Fourth. Cambridge, MA: Harvard University Press, 1994.; STORPER, 2013STORPER, M. Keys to the city: how economics, institutions, social interaction, and politics shape development. Princeton, NJ: Princeton University Press, 2013.). However, network actors must also show some level of cognitive proximity when accessing and creating links with regional external actors (BOSCHMA, 2004BOSCHMA, R. Competitiveness of regions from an evolutionary perspective. Regional Studies, v. 39, n. 9, p. 1001-1014, 2004., p. 1006). In addition, they must have minimum absorptive capacity to internalize knowledge (COHEN; LEVINTHAL, 1990COHEN, W. M.; LEVINTHAL, D. A. Absorptive capacity: a new perspective on learning and innovation. Administrative Science Quarterly, v. 35, n. 1, p. 128-152, 1990. Available in: <http://www.jstor.org/stable/2393553>.
http://www.jstor.org/stable/2393553... ). A minimum level of institutional proximity is required when actors share values and expectations. This study presents empirical data regarding this in Section 4.

Regions are central to foster interactions that increase and improve local learning capacities (COOKE, 1992COOKE, P. Regional innovation systems: competitive regulation in the new Europe. Geoforum, v. 23, p. 365-382, 1992.; DOLOREUX, 2002DOLOREUX, D. What we should know about regional systems of innovation? Technology in Society, v. 24, n. 3, p. 243-263, 2002.; DOLOREUX; PARTO, 2005). They leverage local vocations over time and become a strategic asset for local competitiveness and the formation of regional innovation systems. In turn, Regional innovation systems aim (among other things) to attract new investments related to specific sectors to a region. Brazilian regions that have undergone industrial concentration over time have managed to attract scientific infrastructure investments, while regions with lower industrial concentration have become peripheric in the country (DE NEGRI; SQUEFF, 2016DE NEGRI, F.; SQUEFF, F. V. D. H. S. Sistemas setoriais de inovação e infraestrutura de pesquisa no Brasil. Brasília: Ipea, 2016.).

The geographical dimension of the Brazilian territory and its regional diversity create a geographical distance between network actors in their interactions, challenging the advantages of geographical proximity. The actors who participate in the ANEEL R&D are not geographically close in all interactions. Network creation supersedes the regional scale, and interactions are located throughout the national territory. Regional inequality and heterogeneity are likely to increase as a result of differences in access to capital, availability of trained human resources and scientific and technological infrastructure (DINIZ; GONÇALVES, 2005DINIZ, C. C.; GONÇALVES, E. Economia do conhecimento e desenvolvimento regional no Brasil. In: DINIZ, C. C.; LEMOS, M. B. (ed.). Economia e território. Belo Horizonte: Editora UFMG, 2005. v. 1, p.131-170.). The latter is unequal among Brazilian regions and this asymmetry creates inequalities in the potential for regional innovation, which is crucial for sectorial competitiveness (DE NEGRI; SQUEFF, 2016DE NEGRI, F.; SQUEFF, F. V. D. H. S. Sistemas setoriais de inovação e infraestrutura de pesquisa no Brasil. Brasília: Ipea, 2016.).

This study introduced the use of geotechnologies to map whether there is a connection between geography and innovation. Georeferenced information includes remote sensing, digital cartography, geostatistics and geographical information systems (GIS). These allow the analysis, management and representation of geographical location as well as of the phenomena developed in the territory (MATIAS, 2002MATIAS, L. F. Sistemas para In[form]ação. Espaço & Geografia, v. 5, n. 1, p. 101-118, 2002.; SOUZA FILHO; CROSTA, 2003SOUZA FILHO, C. R.; CROSTA, A. P. Geotecnologias aplicadas à geologia. Revista Brasileira de Geociências, v. 33, n. 2, 2003.; ROSA, 2011ROSA, R. Análise espacial em geografia. Revista da ANPEGE, v. 7, n. 1, p. 275-289, 2011.). This study proposes to map the networks created under the ANEEL R&D Programme; according to Garcia (2012GARCIA, R. Resenha: The handbook of evolutionary economic geography. Revista Brasileira de Inovação, Campinas, v. 11, n. 1, p. 233-240, 2012.), it is necessary to bring more empirical evidence for Brazil on the topics discussed by the evolutionary economic geography. Mapping networks through georeferencing contributes to this debate.

2.2 The ANEEL R&D Programme: a short review

The ANEEL R&D Programme was implemented in a highly regulated market.⁸ 8 This study considers the 2006-2008 period. ANEEL has changed some practices of its R&D Programme over time, a different period of analysis must consider the practices that regulate the programme for the relevant period. The Brazilian State has been historically present in the electric sector through state-owned companies, which have had a central role in the creation and expansion of the energy sector. The local electric sector went through a deregulation process in the 1990s which settled the institutionalization of a regulatory agency (ANEEL) and the increase of private firms’ share in the sector (MASSAGUER, 2013MASSAGUER, P. X. Perspectivas para a Pesquisa, Desenvolvimento & Inovação no setor elétrico brasileiro. 2013. Dissertação (Mestrado em Política Científica e Tecnológica) - Departamento de Política Científica e Tecnológica, Instituto de Geociências, Universidade Estadual de Campinas (Unicamp), Campinas, 2013.). However, the presence of state-owned companies was still relevant in the period covered by this study.

The Programme was part of the Brazilian federal government’s technology policy, which aimed at fostering local sectoral R&D activities. The policy framework deemed systemic interactions a condition to the formation of a dynamic innovation system; the objective was to create knowledge exchange between its actors (ANEEL, 2011). The Programme was implemented in conformity with the Law n. 9.991/2000 and it demands that electrical energy operators located in Brazil apply 1% of their annual net revenue in R&D activities. According to the R&D ANEEL Handbook, every project approved must comply in terms of: theme and sub-theme; stage in the innovation chain; and type of product. For each theme there is a classification of sub-themes, which indicates the wide scope of R&D activities.⁹ 9 For more information, refer to http://www.aneel.gov.br/area.cfm?idArea=641&idPerfil=6; accessed in March 17 2014.

The projects are framed into specific phases of the innovation chain, as well as according to specific types of products.¹⁰ 10 For further details on the specificities of the Programme, see the ANEEL Handbook (2012a), this discussion is beyond the purpose of this paper. ANEEL evaluates submissions to classify whether or not they comply with the Programme¹¹ 11 The criteria include: “originality; applicability; relevance, cost reasonableness. For each criteria a pointing scheme is attributed determining the classification of the project, which will determine its partial or complete approval or even its rejection” (Manual da ANEEL, 2012a, p. 18, our translation). and follows their implementation (when approval is granted). If project outcomes are not approved, participant must pay a penalty and return the value of the grant to ANEEL (ANEEL, 2012a).

The purpose of this study is to investigate the networks resulting from the ANEEL R&D Programme to elucidate the interactions created through geographical proximity in the period from 2006 to 2008. In this way, the paper will not assess the efficiency of the Programme in relation to the outcomes of its projects. The Programme is an object of study that will enable to examine whether or not geographical proximity relates to the electric sectoral innovation systems in Brazil. In order to achieve this purpose, we applied the methodology of SNA and geoprocessing, as will be explained in the next section.

3. Research method

Data on approved projects are made available by ANEEL R&D including the following variables: project title; name of applicant firm; firm’s acronyms and ANEEL code; project theme; stage in the electric production chain; type of products; cost; preliminary starting and ending date, and final report; and ANEEL’s project assessment report.¹² 12 http://www.aneel.gov.br/area.cfm?idArea=75&idPerfil=6, last accessed in 04 April 2014. However, available data on projects do not include other relevant information for this study, such as: performing organisation; and the coordinator of each project. These data were accessed only through manual search for each individual ANEEL R&D project in ANEEL’s website.¹³ 13 https://ped.aneel.gov.br/consultaPublica/resultadoConsultarProjetos.asp. Completing the database with information on performing organisations and coordinators was the first step of the data collection. Final data refers to the 2006-2008 period and to 278 R&D projects (i.e., 52.2% of the submitted projects for the analysed period).¹⁴ 14 As mentioned above, data analysed in this study includes information for the 2006-2008 period shared by ANEEL.

In addition, the following data for projects are not available: geographical location of applicant firms and performing organisations; andinstitutional affiliation of project coordinators and their geographical location. We completed the database with information for each ANEEL R&D project on: coordinators’ institutional affiliation through searches in the Lattes Platform; and the geographical location of each coordinator and applicant firms using ArcGiS (ÒESRI). These data are crucial for explaining ANEEL R&D networks with regards to the type of performing organisation that is involved (organisational perspective), and project coordinators (individual perspective). Interactions between actors located in the same state or Brazilian region are a proxy for geographical proximity (GARCIA et al., 2015GARCIA, R. et al. Looking at both sides: how specific characteristics of academic research groups and firms affect the geographical distance of university-industry linkages. Regional Studies, Regional Science, v. 2, n. 1, p. 517-533, 2015.), this highlights the relevance of the regional innovation system, e.g., sharing the same scientific infrastructure (POWELL; GRODAL, 2005POWELL, W. W.; GRODAL, S. Networks of innovators. In: FAGERBERG, J.; MOWERY, D. C.; NELSON, R. R. (ed.). The Oxford handbook of innovation. New York: Oxford University Press, 2005. p.56-85.).

Once the database was completed, we used SNA measures and representation (DE NOOY; MRVAR; BATAGELJ, 2005DE NOOY, W.; MRVAR, A.; BATAGELJ, V. Exploratory social network analysis with Pajek. Cambridge: Cambridge University Press, 2005.) to create the following ANEEL R&D networks: applicant firms and performing organisations; and applicant firms and project coordinators. Network analysis centrality measures (Table 1) were combined with geospatial analysis.

The final step was the use of georeferencing (Geographic Information System - GIS) to map interactions between ANEEL R&D network actors. Cartography data was added to the project database for each project: geographic coordinate system (degrees, minutes, seconds) for applicant firms and performing organisations were converted into decimal degrees, allowing the use of ArcGis 10.1 (ÒESRI). Each latitude (X) and longitude (Y) provides the geographical location of SSI actors based on the ANEEL R&D Programme, providing a point mapping. In geospatial analysis points are isolated and positioned geometric location, which in group create a spatial arrangement. The pattern of this arrangement emerges from the distance between points. The dispersion is the degree of spatial positions related to a frame in which points are present. Density is a characteristic of dispersion, which is related to the measure of an area but independent of its shape or dispersion of points (FERREIRA, 2014FERREIRA, M. C. Iniciação à análise geoespacial. São Paulo: Editora Unesp, 2014.).

Networks are a spatial arrangement of links, nodes and regional nodes. Networks are integrated structures that represent the possibilities of movement and spatial interaction among SSI actors of the electric sectoral system. The ANEEL R&D network is represented by a graphic model (graph), points connected through lines.¹⁵ 15 This model does not consider the length and direction of lines, neither if lines are curved or straight. Graphs provide the construction of a connective matrix between nodes. In each square matrix (i, j), there is value for the interaction between the location i and the location j (nodes) of the ANEEL R&D network. This study uses the binary connectivity matrix that registers the occurrence (or absence) of connections between locations i e j. The matrix allows the calculation of number of access by the applicant firm to the performing organisation and to other actors of the network. The result was a network mapping to identify the quantity of interaction flows between applicant firms and performing organisations (FERREIRA, 2014FERREIRA, M. C. Iniciação à análise geoespacial. São Paulo: Editora Unesp, 2014.). Geoprocessing allowed the visualisation of the cartographic representation of the ANEEL R&D network (Section 4).

4. ANEEL R&D mapped networks: results and discussion

This study analysed two networks created under the ANEEL R&D Programme: the organisational network of applicant firms and performing organisations, i.e., research institutions, universities and firms; and ii) the individual network of applicant firms and project coordinators.¹⁶ 16 This study aimed to analyse networks based on project themes and performing organisations, but there was no complete consistency on data collected from ANEEL. Hence, the analysis about the network of themes and areas of expertise was initiated but not completed. It was not possible to obverse systematically whether R&D projects related to technologies are closer or more distant to the technological frontier. Such study is a future topic of research.

4.1 Applicant firms and performing organisations: the organisational network

The organisational network (Figure 1) shows low density, high fragmentation and therefore low level of interactions among actors. These features suggest that knowledge circulation among actors engaged in ANEEL R&D is lower than its potential (see Table 1 for a summary of centrality measures). Degree and closeness centralities show that a few actors are relevant as intermediaries in knowledge exchange and information diffusion within the network, among those we highlight the following performing organisations: LACTEC,¹⁷ 17 Previously known as Technological Institute for Development (in Portuguese, Instituto de Tecnologia para o Desenvolvimento) and now LACTEC Institutes (in Portuguese, Institutos LACTEC), based in Curitiba (capital of the state of Paraná). It is a Civil Society Organisation for the Public Interest (in Portuguese, Organização da Sociedade Civil de Interesse Público-OSCIP) since 2000. which accumulated a variety of research capabilities in the electric sector, and the Universidade Federal Fluminense (UFF) and its Euclides da Cunha Foundation, both based in the state of Rio de Janeiro.¹⁸ 18 The identification of UFF as a relevant actor does not allow conclusions about its capabilities, once the data refer to the university level, not to a specific research centre. The same applies to its Foundation, which manages research projects and funds at the university level.

In addition, a few actors are related to knowledge production through basic research activities. Applicant firms are the main intermediaries in the network, suggesting that access to and circulation of frontier knowledge offered by actors that conduct research, such as universities and public and private research institutes, may be lower than its potential.

Figures 4 and 5 are maps which present the geographical location of applicant firms and performing organisations. The map in Figure 4 suggests that interactions created for R&D projects follow the pattern of the Brazilian land occupation (GARCIA et al., 2015GARCIA, R. et al. Looking at both sides: how specific characteristics of academic research groups and firms affect the geographical distance of university-industry linkages. Regional Studies, Regional Science, v. 2, n. 1, p. 517-533, 2015.; FERNANDES, 2016FERNANDES, A. C. Sistema territorial de inovação ou uma dimensão de análise na geografia contemporânea. In: SPOSITO, E. et al. (ed.). A diversidade da geografia brasileira: escalas e dimensões de análise e da ação. Rio de Janeiro: Consequência Editora, 2016. p.113-143.). The majority of applicant firms are located close to coastal areas. Geographical proximity is also an asset for project performing organisations located in the South and Southeast regions of the country, which are responsible for the majority of projects - as shown by other studies (POMPERMAYER et al., 2011POMPERMAYER, F. M. et al. Rede de pesquisa formada pelo programa de P&D regulado pela ANEEL: abrangência e características. In: POMPERMAYER, F. M.; DE NEGRI, F.; CAVALCANTE, L. R. (org.). Inovação tecnológica no setor elétrico brasileiro: uma avaliação do programa de P&D regulado pela Aneel. Brasília: Ipea, 2011. p. 13-). Figure 5 suggests that geographical proximity is not a relevant issue for applicant firms based in the Northeast region, such as Energipe, CEMAR (both established the highest number of links in the network) and Coelce. Georeferencing showed that 35.5% of interactions occurred inside the same Brazilian state, whereas 64.5% between two Brazilian states, as shown in Table 3.

It is noteworthy that cognitive proximity seems to be a determinant issue for interactions in R&D activities for Northeastern actors; knowledge and trust are essential features for the creation of this type of link (GARCIA et al., 2015GARCIA, R. et al. Looking at both sides: how specific characteristics of academic research groups and firms affect the geographical distance of university-industry linkages. Regional Studies, Regional Science, v. 2, n. 1, p. 517-533, 2015.). In addition, project performing organisations specialised in knowledge production in the sector may not be present in locations close to applicant firms (based in the Northeast region), or there may not be the necessary knowledge capacities that foster the creation of collaborations.¹⁹ 19 Note that this study does not include the motivation for the creation of interactions from the perspective of applicant firms. Explaining such motivations would demand an in-situ fieldwork data collection or a survey research. Neither of them are within the scope of this study. See more on this topic in the concluding section. It is important to highlight that the creation of ANEEL R&D networks are limited to ANEEL calls for project submissions that follow the scope of the programme, which specifies themes and areas within the sector.²⁰ 20 We thank the comment of one of the peer reviewers, who brought our attention to this issue as a relevant fact in the creation of ANEEL R&D networks.

The study shows that the majority of projects are performed by network actors that are located in the most developed regions of the country (both in number of projects and in financial resources, see Figure 6 below). These results are in line with the fact that Brazil presents regional inequality with concentration of capacity of knowledge production and science and technology infrastructure in the South and Southeast regions (DE NEGRI; SQUEFF, 2016DE NEGRI, F.; SQUEFF, F. V. D. H. S. Sistemas setoriais de inovação e infraestrutura de pesquisa no Brasil. Brasília: Ipea, 2016.), which present vocational assets related to the themes of R&D projects.

Network betweenness centrality elucidates the relevance of a particular actor within the network for the creation of connections, the network hierarchy, and which actors are more central in interactions. Our analysis of the ANEEL R&D organizational network shows that a few actors are central for interactions (ten out of 198) and, therefore, that the network is very hierarchical. The main actors of the network are: Enerpeixes SA; INVESTCO; Termo Norte Energia Ltda; El Paso Amazonas; LACTEC; Bandeirante; Ampla; Energipe; Universidade Federal Fluminense; and Fundação Leonel Franca. Most of those actors are applicant firms, suggesting that project performing organisations have a subordinate position in knowledge generation and circulation in the Brazilian electric sector.

4.2 Applicant firms and project coordinators: the individual network

The network of applicant firms and project coordinators (the individual network) identifies which actors develop projects with grantee firms and to which organisations such actors are affiliated (Figure 7). The network shows low density; the interaction among actors are low when compared to its potential, revealing that a few coordinators can intermediate knowledge production and circulation in the electric sector. Degree and closeness centralities (Table 4) show that 28 nodes above the upper quartile (above 95%) are in majority (21 project coordinators). However, Figure 7 shows that the main actors of the network are applicant firms, such as Bandeirante and Ampla (followed by Energipe and CELPE).

The analysis of eight coordinators (higher values in the upper quartile) showed that most of them are located in the Southeast region (see Figure 8), but are accessed by actors present in other Brazilian regions, to which there is geographical distance. This result suggests that project coordinators might have specialised knowledge about the sector. It also suggests that geographically distant partners access project coordinators because they produce knowledge that is strategically relevant. Cognitive proximity may be a relevant issue in the creation of links when the necessary knowledge for the development of the sector is strategic (R&D activities show this feature). Another possible interpretation is that the knowledge accessed is geographically distant and not locally available. Even though geographical distance has not been a barrier for the creation of links, it reveals that there might be a cognitive proximity among actors.

The individual network betweenness centrality shows high hierarchy among actors in which, instead of project coordinators, applicant firms play a major role. However, two project coordinators (out of ten most important nodes) stand out: RRomano - at the time based at the Networks Laboratory and Infrastructure at the Centre for Research and Development in Telecommunications (CPqD), located in Campinas (state of São Paulo), which develops R&D activities in high technology sectors, including the electric sector (mainly related to the application of information technology in the electric network); and HHenriques - then based at the Technological Centre of the Engineering School of the Universidade Federal Fluminense (state of Rio de Janeiro).

Other coordinators identified in Figure 8 show the highest degree centrality; most of them are located in the Brazilian Southeast region (with the exception of MFilho, located in the state of Paraná, in the South region) and are based at universities (exceptions are JMak, SFerreira and GTorres). These project coordinators were accessed by actors present in other Brazilian regions, suggesting that there is some level of cognitive proximity among them. Applicant firms such as CAIUA-D, EATE, Enerpeixes SA, CELESCDIS, Itapebi, Termo Norte Energia Ltda, Bandeirante, Ampla, and Coelba also played a central position in the network. Such feature of the network was also present in the analysis of the organisational network (Figure 3); the implications of these features for the evolution of the electric sector are present in the concluding section below.

5. Conclusion

This study analysed the Brazilian electric sector investigating two networks of the ANEEL R&D Programme in the 2006-2008 period and proposed the following research question: “How are the actors of the innovation system of the Brazilian electric sector structured, in terms of their R&D activities and within the scope of government policies which aim at the creation of networks, such as the ANEEL R&D Programme?”

Data analysis drawing upon SNA and georeferencing enabled inferences about the main interactions within the SSI of the electric sector in Brazil and supported the clarification of: description of the main actors in the networks in relation to their proximity (closeness centrality) and intermediation (betweenness centrality), number of projects, financial resources, geographical location (by georeferencing networks); and possible geographical specialisation of the knowledge generated within the scope of the programme.

The data showed that, in the analysed period, the majority of actors accessed by applicant firms for the creation of links were research institutes and universities, in line with previous literature about the topic (POMPERMAYER et al., 2011POMPERMAYER, F. M. et al. Rede de pesquisa formada pelo programa de P&D regulado pela ANEEL: abrangência e características. In: POMPERMAYER, F. M.; DE NEGRI, F.; CAVALCANTE, L. R. (org.). Inovação tecnológica no setor elétrico brasileiro: uma avaliação do programa de P&D regulado pela Aneel. Brasília: Ipea, 2011. p. 13-). This study sought to contribute to the understanding of the ANEEL R&D networks investigating the structure of the organizational and individual networks. The actors of these two networks are components of the electric SSI, as shown in Figure 2 (focus of the study). The study highlighted the geographical location of network actors.

The governance of the R&D networks created under the ANEEL R&D Programme was mainly structured by applicant firms, with a secondary participation of actors that conduct external research to intermediate knowledge created about the sector. Furthermore, we found that suppliers of technology and energy technology were scantly accessed by applicant firms (that is, at a lower level than external actors) and also had a peripherical participation in a hierarchical network dominated by applicant firms. Considering that the electric sector is supplier-dominated, and that external research actors are in a subordinate position with regard to knowledge creation through R&D activities, the absence of suppliers shows a possible lack of internal dynamics in the local creation and circulation of the knowledge needed for the electric sector. In addition, the result shows frailty of the local sector in the creation of frontier knowledge.

The analysis about the locality of interactions shows that the sector mirrors the knowledge production structure located in the main regions (South and Southeast) and near the Brazilian coast, following the national pattern of availability of research infrastructure (GARCIA et al., 2015GARCIA, R. et al. Looking at both sides: how specific characteristics of academic research groups and firms affect the geographical distance of university-industry linkages. Regional Studies, Regional Science, v. 2, n. 1, p. 517-533, 2015.; DE NEGRI; SQUEFF, 2016DE NEGRI, F.; SQUEFF, F. V. D. H. S. Sistemas setoriais de inovação e infraestrutura de pesquisa no Brasil. Brasília: Ipea, 2016.). The interactions created between actors in distant localities suggest that they present minimal capacities required to access actors located in consolidated research organisations. However, the search for geographically distant actors may result from the absence of local actors with sufficient minimum capacities to produce knowledge. Given that the predominant feature of the network is the large number of interactions with actors from universities, it is possible that the creation of new higher education institutions in Brazil since 2004 could potentially promote the growth of interactions with actors based in distant areas. However, this phenomenon is expected to occur in the medium term, since it depends on the consolidation of the local capacity to produce knowledge.

This study investigated the structure of ANEEL R&D networks based on SNA and georeferencing. These methods do not allow the understanding of the type of knowledge produced (whether they are closer or further away from R&D and demonstration activities, and the technological frontier). It neither enables to explain the motivation for the creation of network links, that is, whether they were created by geographical proximity reasons, other types of proximities, organizational affinities, previous well-succeeded experiences or even for personal reasons. Another issue which was not analysed were the implications of the interactions over time, neither from the personal perspective of the actors involved in projects. The understanding of these elements for the creation of networks demands in-situ research (fieldwork interviews or surveys) with all the actors in the networks. We suggest this approach as an area for future research, as well as a study of the investigated networks for a longer period of time and the observation of the financial figures involved in the projects as a proxy of the intensity of interactions. These elements are limitations of this article. Empirical evidence of this type would support the formulation of effective public policy to amend the dysfunctionalities of the SSI of the Brazilian electric sector identified by this study and the literature on the topic.

Acknowledgements

We are thankful to PIBIC-UNICAMP/CNPq for the financial support to the project “Redes de P&D no setor elétrico brasileiro: a contribuição metodológica da análise de redes sociais para o estudo do Programa P&D ANEEL” that resulted in the database created and used in this study, we acknowledge support from CAPES Foundation Process Number 88882.180279/2018-01. To Dr. Anapatrícia de Oliveira Morales Vilha for suggesting that the initial version of this study presented at the VI Simpósio Nacional de Ciência Tecnologia e Sociedade (ESOCITE Br, 2015) was worth an article, and to Dr. André Luiz Sica de Campos for his comments and suggestions in preliminary versions. To the two anonymous referees for their valuable comments and suggestions, which we aimed to include in the final version of this article. The usual disclaimers apply.

References

Publication Dates

History