Reviewed by
Tommaso Ciarli
Max Planck Institute of Economics, Jena
As the editors readily suggest at the beginning of their introductory
chapter, in the last decade we have witnessed a "network hype". This is
particularly true for Economics. As often happens, economists are eager
to "import" and (some times) adapt methods and results from other
disciplines, with a reverential attachment to the rigor of physics, as a
scientific method. Surprisingly, it took a great deal of time for
economists to be inspired by a much closer discipline, sociology, and in
particular by scholars that have explicitly referred to social networks
since no less than a century. It is probably not by chance that the
results on the "small word" structure of social relations put forward by
Milgram (a social psychologist) have attracted the interest of
physicists first (thirty years later) and that of economists only
afterwards. This is not to say that economists had not realized that social - as well as economic - relations could be represented by a network in which the nodes are the economic agents and the links between them the relations (e.g., see Jackson 2007 for a review of the analysis of networks in economics). But it took some time to take on board a new paradigm (as defined by the editors in the introduction to the book) within the realm of accepted methods in the study of economics. Here physics and economics have followed quite different paths. While economists were mainly interested in the result of given networks' topologies (e.g., in terms of efficiency, social and individual welfare, diffusion of knowledge), and in their formation as an outcome of pairwise coalitions (incentive driven), physicists have devoted most of their research to derive general laws of network formation and topologies, with particular attention to physical as compared to social networks.
This book is a worthwhile contribution aimed at narrowing down the divergence between the socioeconomic research and the physicists' research on networks. Provided networks can map any interaction between objects or agents, the editors rightly focus on one specific topic that may benefit from this cross-fertilization: innovation. Both in terms of creation of novelty - defined differently in the different contributed papers - via collaborations of networked agents (most chapters) and in terms of diffusion of novelty through the networked system (particularly chap. 9 and partially chap. 6). The declared aims of the Editors are (i) to provide the reader with different concepts and formalizations of networked systems (within the "new paradigm" of complex networks), and (ii) to create an exchange between the different approaches of economists and physicists ("create a negotiation platform for notions and research intentions of the different groups analyzing networks, and innovation networks in particular").
The potential of the book for achieving these aims is limited by its genesis, a collection of essays presented in a workshop. Many of the papers are actually summaries or re-elaborations of some of the authors' previous research, or reprints from previous publications. This aspect thus constrains the authors' search for convergence but spurs the diversity of contributions from which the reader may choose. In fact, the book provides a valuable resource for those interested in how network structures affect innovation and their outcome (e.g., in terms of industrial structure (chap. 2), international scientific collaborations (chap. 3), and performance of an industry (chap. 5)), and how different types of knowledge exchange affect the formation of networks (e.g., in terms of subgroups (chap. 7), or the dual relation between agents' and the network's evolution (chap. 8)). A notable exception to the lack of convergence between the different chapters is the contribution in chapter 2, which discusses the relation between network theory and other complexity theories. In sum, the book privileges knowledge variety, leaving to the reader the choice of the contributions that best serve her/his research. The following may serve as a guide through the book.
The book is divided into two parts. The first part binds together the contributions from economists, while the second focuses on the contributions of physicists, in particular on statistical physics. Broadly speaking, the first part is more appealing to those interested in the effect of networks on innovation and knowledge diffusion, with both theoretical and empirical contributions, while those more interested in the descriptions of how networks may emerge (or die) will be more attracted by the second part. Although the second part contains no empirical contributions as such, some of the chapters put their results in the perspective of existing empirical accounts. As a caveat, although all papers in the second part reflect an appreciable effort to introduce the language of the economics of innovation, the hiatus that remains between the description of a physical system and the description of a system that results from human choices and actions is also perceivable. However, this caveat should not be interpreted as a critique of these chapters, which provide well presented methods and results for the social scientists - where convenient and useful - willing to advance the understanding of network economics.
Chapter 2 starts with a discussion of the connections between network theories and other complexity theories. The author then provides a valuable representation of knowledge as a web connecting different domains. This interpretation is then used to discuss existing empirical results on the knowledge networks in firms and between firms. The third chapter focuses on the production of knowledge, representing the scientists on a network that covers different geographic regions. The authors then test the death-of-distance hypothesis, showing that geographic distance reduces collaborations between scientists. Chapter 5 uses a model to show the evolution of a network in which firms actually exchange knowledge in order to innovate, creating networks of collaborations. The results show the cyclical and persistence patterns of knowledge collaborations. Chapter 6 focuses more on the relation between the architecture of the network and the firm level and system level efficiency attached to it. Architectures may differ in terms of stars, degree, clustering, and abundance of knowledge. Firms are modelled as givers or traders of knowledge, which is used as production input. The results show the conditions under which both strong social capital and structural holes are associated with high or low performance, due to low knowledge diffusion. The model in chapter 4 is an attempt to analyze the effect of regional innovation networks on the innovation performance of firms. Firms have the option to innovate individually, or in cooperation with other firms, when they can exchange knowledge. The willingness to cooperate depends on, among other things, transaction costs and the related trust, which are shaped by local institutions. These, in turn, have an effect in shaping the regional innovation network, in terms of innovation dynamics.
Chapter 7 provides an informative introduction to the use of network theories in physics, with special focus on the homophily features. The main contribution describes the use of spin glasses to analyze the emergence of sub-communities within networks. The method is then applied to the identification of communities of research. It is instructive to compare these results with those in chapter 3. Another parallel can be drawn between chapters 6 and 8. The latter aims at showing, by way of a physical model, the relation between network evolution and the creation of knowledge by the nodes. Knowledge can also be a gift, or part of an exchange. The network changes because of the exogenous replacement of a link or because of a node choice. The different combinations of exchange and network evolution provide very different results on the equilibrium reached at the end of the process and its efficiency. Chapter 9 is closer to chapter 4 only in its application at a regional level: the Silicon Valley and Route 128. The model analyzes the evolution of networks and knowledge in an avalanche model framework, where initial adopters can be imitated or not by neighbours, depending on their diversity and homophily. Chapter 10 develops another approach to the design and analysis of networks, starting from the common framework of knowledge networks. Here knowledge is seen embedded in nodes present in the system, becoming part of the system only when the node and its link are activated. The authors then study how the network evolves via a stochastic process.
In conclusion, the book presents a rich set of models and empirical evidence of networks that mainly represent the share of knowledge between nodes (firms). Readers looking for methods and modelling techniques across innovation studies and statistical physics will find this book of valuable use.
References
Return to Contents
of this issue
© Copyright Journal of Artificial Societies and Social Simulation, 2010