Networks of venture capital firms in silicon valley

International Journal of Technology Management

Volumn 25, Issue 1-2, 2003, Pages 113-135

  Castilla, Emilio J ^a  

^a STANFORD UNIVERSITY   (United States)

Author keywords

Regional development; Route 128; Silicon Valley; Social networks; Venture capital firms (VCs)

Indexed keywords

ELECTRONICS INDUSTRY; INDUSTRIAL ECONOMICS; MARKETING; SOCIAL ASPECTS; SOCIETIES AND INSTITUTIONS; TECHNOLOGY;

REGIONAL DEVELOPMENT; SILICON VALLEY; SOCIAL NETWORKS; VENTURE CAPITAL FIRMS;

INDUSTRIAL MANAGEMENT;

EID: 0037210978     PISSN: 02675730     EISSN: None     Source Type: Journal    
DOI: 10.1504/IJTM.2003.003093     Document Type: Article

References (44)

1. For comprehensive reviews of the wide-ranging literatures on the role of social networks in economic life and network analysis.
2. Wasserman, S. and Faust, K. (1994) Social Network Analysis: Methods and Applications, Cambridge University Press, New York.
3. Powell, W.W. and Smith-Doerr, L. (1994) 'Networks of economic life', The Handbook of Economic Sociology, N.J. Smelser and R. Swedberg (Eds.) Princeton University Press, New Jersey, pp.368-402.
4. Saxenian, A. (1990) 'Regional networks and the resurgence of Silicon Valley', California Management Review, Vol. 33, pp.89-112.
5. Saxenian, A. (1994) Regional Advantage: Culture and Competitions in Silicon Valley and Route 128, Harvard University Press, Massachusetts.
6. Florida, R.L. and Kenney, M. (1987) 'Venture capital and high technology entrepreneurship', Journal of Business Venturing, Vol. 3, pp.301-319.
7. Nohria, N. (1992) 'Information and search in the creation of new business ventures: The case of the 128 Venture Group', Networks and Organizations: Structure, Form and Action, N. Nohria and R.G. Eccles (Eds.) Harvard Business School Press, Massachusetts, pp.240-261.
8. Harmon, S. (1999) Zero Gravity: Riding Venture Capital from High-Tech Start-Up to Breakout IPO, Bloomberg Press, New Jersey.
9. Mintz, B. and Schwartz, M. (1985) The Power Structure of American Business, University of Chicago Press, Illinois.
10. Granovetter, M. (1985) 'Economic action and social structure: The problem of embeddedness', American Journal of Sociology, Vol. 91, pp. 481-510.
11. Granovetter, M. (1974) Getting a Job: A Study of Contacts and Careers, Harvard University Press, Massachusetts,
12. Granovetter, M. (1995) 'Coase revisited: business groups in the modern economy', Industrial and Corporate Change, Vol. 4, pp.93-131.
13. More detailed results, including an expanded electronic version of this report, can be found in the Money Tree web site at www.pwcmoneytree.com.
14. Joint Venture (2000) 2000 Index of Silicon Valley: Measuring Progress Toward Silicon Valley 2010
15. Joint Venture (1999) Silicon Valley 2010: A Regional Framework for Growing Together. For more information, visit: http://www.jointventure.org/resources.html.
16. Wilson, J.W. (1985) The New Venturers: Inside the High-Stakes World of Venture Capital, Addison-Wesley Publishing Company, California.
17. Asset Management Company's 'West Coast Venture Capital - 25 years' Genealogy Chart 1983. There are some omissions from the chart such as Tom Flowers, Bud Moose and Ray Lyon, who started early SBICs and were leaders of the old Western Association of Small Business Investment Companies (WASBIC). These people were not included since they were primarily real estate brokers. WASBIC was the predecessor of today's Western Association of Venture Capital (WAVC). The information is current from 1958 up to December 1983, to the best of Asset Management Company's knowledge.
18. For a complete analysis of the historical development of networks of VC firms in Silicon Valley see [16].
19. Castilla, E.J.; Hokyu, H.; Granovetter, E. and Granovetter, M. (2000) 'Social networks in Silicon Valley', The Silicon Valley Edge: A Habitat for Innovation and Entrepreneurship, C-M. Lee, W.F. Miller, M.G. Hancock and H.S. Rower (Eds.) Stanford University Press, California, pp.218-247.
20. A general note that includes references to venture capital literature: see [18-22].
21. Bartlett, J.W. (1998) Venture Capital: Law, Business Strategies and Investment Planning, John Wiley & Sons, New York.
22. Bygrave, W.D. and Timmons, J.A. (1992) Venture Capital at the Crossroads, Harvard Business School Press, Massachusetts.
23. Eccles, R. and Crane, D. (1998) Doing Deals: Investment Banks at Work, Harvard Business School Press, Massachusetts.
24. Halloran, M.J., Benton, L.F., Lovejoy, J.R., Kearney, K.L. and Gunderson, R.V. (1991) Venture Capital and Public Offering Negotiation, Prentice-Hall Law & Business, New Jersey.
25. Timmons, J.A. and Sapienza, H.J. (1991) 'Venture capital: More than money?', S.M. Pratt and J. Morris. (Eds.) Pratt's Guide to Venture Capital Sources, Venture Economics, Inc., Massachusetts, 15th Edition, pp.35-41.
26. One of the data sets analysed in this paper was collected from the ACI Genealogy chart as part of the project 'Social Networks in Silicon Valley' funded by the Bechtel Initiative, Stanford University, with Mark Granovetter, Emilio J. Castilla and Hokyu Hwang. I especially want to thank Michel Ferrary (ESSEC, France) for his help in accessing the database of venture capital funds and start-up companies organised by PricewaterhouseCoopers. The PricewaterhouseCoopers Money Tree Survey is a quarterly study of equity investments made by the venture capital community in the USA. Survey questionnaires are mailed. For more information about the survey, go to: www.pwcmoneytree.com
27. Richardson, D.C. and Richardson, J. S. (1992) 'The Kinmage: A tool for scientific communication', Protein Science, Vol. 1, pp.3-9.
28. MAGE was developed as a device to be used in molecular modelling. It produces three-dimensional scientific illustrations that are presented as interactive computer displays. For learning and downloading the latest version of the MAGE program, visit: http://www.faseb.org/protein/kinemages/kinpage.html. See [26].
29. Freeman, L.C., Webster, C.M. and Kirke, D.M. (1998) 'Exploring social structure using dynamic three-dimensional color images', Social Networks, Vol. 20, pp.108-118.
30. The diagram is therefore the visualisation of a valued graph or network: given that in the case of a link between two VC firms, I also know the number of instances in which both VCs co-invested in the same company. The diagram displays the VC firms in a three-dimensional x by y by z space. Location of the points is produced by the multidimensional scaling technique (MDS). So two VC firms with four companies in common, for example, will be closer to one another than two companies which have just three companies in common; how much closer they are in the three-dimensional space is determined by the MDS procedure. The MDS technique is designed to represent proximities among a set of entities in a low-dimensional space so entities that are more proximate to each other in the input data are closer in the space and entities that are less proximate to each other are further apart in the space. The usual input to a multidimensional scaling is a one-mode symmetric matrix consisting of measures of proximity between the pairs of entities. In this case, the input matrix was the adjacency matrix of VC firms in which the cells measure the number of investee companies in common during the 1995-1998 period. The output of multidimensional scaling is a set of estimated distances among pairs of firms, which can be expressed as coordinates in a three dimensional space. Results are also displayed as a diagram in which the coordinates are used to locate the entities in the resulting three-dimensional space. Using multidimensional scaling to study a network shows which subsets of firms are relatively close to each other in a graph theoretic sense and which set of firms are isolates. Because the adjacent matrix of firms by firms to analyse was valued (i.e., non-binary matrix where each of the cells shows the number of investee companies in common for a given pair of VC firms), I use a non-metric method of multidimensional scaling (starting with a metric solution). I use UCINET to run MDS and then MAGE to represent the plot in a three-dimensional space. Multidimensional scaling is a very general data analysis technique. There are numerous texts and articles describing multidimensional scaling. For more information, see [28-30].
31. Scott, J. (1991) Social Network Analysis, Sage Publications, California.
32. Krustal, J.B. and Wish, M. (1978) Multidimensional Scaling, Sage, California.
33. Coxon, A.P.M. (1982) The User's Guide to Multidimensional Scaling, Heinemann, New Hampshire.
34. This does not mean that both VC firms necessarily invested their capital the same quarter or year, or at the same stage of development of the investee company. Rather they could both have participated in the financing of the investee company at any stage in the 1995-1998 period of time here analysed.
35. Freeman, L.C. (1979) 'Centrality in social networks: conceptual clarification', Social Networks, Vol. 1, pp.215-239.
36. The standardisation of such crude degree measure is achieved through dividing degree centrality by the number of firms in the network minus one. Thus, the standardised centrality degree thus is the proportion of other VC firms who are connected to a given VC firm.
37. th firm. So, this index, which counts how 'between' each of the firms is, is a sum of probabilities. It has the minimum of zero, attained when the firm falls on no geodesics. Its maximum happens when the firm falls on all geodesics [2, Ch. 5]. In the example of the A, B and C case, the standardised betweenness centrality is 1 for B (or 100%, if betweenness is measured as a percentage) and 0 for A and C.
38. In a network of N organisations, an organisation obtains the highest possible betweenness score when all N-1 other organisations are tied only to that organisation. In this case the focal organisation would lie on all the geodesics in the network and would be called a 'star'. The relative betweenness of a point is a ratio that measures the extent to which a point in a network approaches the betweenness score of a star [32]. An organisation's relative betweenness can vary from a minimum of 0, when it lies on no geodesics, to a maximum of 1, when the organisation is a star. See [32].
39. Then non-standardised centrality measures are dependent on the group size and hence are difficult to compare across different networks. In general, the standardised centrality indexes are independent of the group size and can be compared across networks of different sizes [2]. The minimum value of any centrality index is 0 but the maximum value is 100 (when the standardised centrality measures are expressed as percentrages).
40. UCINET 5 for Windows, a computer software, was used to perform all network analyses of this paper. Borgatti, S., Everett, M. and Freeman, L. (1999) UCINET 5.0 Version 1.00, Analytic Technologies, Massachusetts.
41. Within a component, all points are connected through paths, but no paths run to points outside the component [28, p.104]. If there is only one component, the graph is connected. If there is more than one component, the network is disconnected [28].
42. Clique is a very strict definition of a cohesive group [2, p.256]. The absence of a single tie or 'cooperation' in the funding of a similar company will prevent a subgroup of firms from being a clique. This definition is, still, very useful for the purposes of this paper since one of the points I want to make is that in a sparse network of VC firms such as Route 128 there are very few cliques. By contrast, in Silicon Valley, I find a larger number of cliques, which are quite large and which overlap one another.
43. Kraemer, K., Dedrick, J., Hwang, C-Y., Tu, T.-C. and Yap, C-S. (1996) 'Entrepreneurship, flexibility and policy coordination: Taiwan's computer industry', The Information Society, Vol. 12, pp.215-249.
44. Mathews, J.A. 'A Silicon Valley of the East: Creating Taiwan's semiconductor industry', California Management Review, Vol. 39, No. 4.