Elsevier

Research Policy

Volume 30, Issue 3, 1 March 2001, Pages 485-508
Research Policy

Technological change and network dynamics: Lessons from the pharmaceutical industry

https://doi.org/10.1016/S0048-7333(00)00094-9Get rights and content

Abstract

In this paper, we investigate how underlying relevant technological conditions induce distinguishable patterns of change in industry structure and evolution. A mapping is detected between the specific nature of problem decompositions and research techniques at the micro level of knowledge bases, and patterns of structural evolution at the macro level of the industry network. The graph-theoretic techniques we introduce map major technological discontinuities on changes observed at the level of dominant organization forms. They might have applications in other domains, whenever the identification of structural breaks and homological relationships between technological and industrial spaces are important issues.

Introduction

Networks of collaborative relationships among firms and other institutions are widely recognized as an important organization form of innovative activities.

One can find in the literature widely different interpretations of the nature, motivations, structure and functions of networks, ranging from more sociologically oriented approaches to economic explanations based on (various mixes of) alternative theoretical backgrounds, e.g. transaction costs, contract theories, game theory, competence-based accounts of firms and organizations.

These interpretations generate widely different predictions about the evolution of collaborative relationships over time.

However, most of these approaches and explanations seem to agree in principle that, especially in high-growth, technology-intensive industries, networks of collaborative relationships should be considered and analyzed as organizational devices for the coordination of heterogeneous learning processes by agents characterized by different skills, competencies, access to information and assets (see Pisano, 1991, Barley et al., 1992, Arora and Gambardella, 1994, Powell et al., 1996, Walker et al., 1997).

Beyond a rather generic agreement, though, the existing literature on networks does not address in detail the nature and specific properties of relevant knowledge bases and search activities to be used as explanatory constructs (see Dosi, 1982, Dosi, 1988).

Against this background, this paper aims to establish a closer connection between the structure and evolution of scientific/technological knowledge and the structure and evolution of organization forms in innovative activities. More precisely, we deal with the relationships between some fundamental attributes of the evolution of relevant knowledge bases in pharmaceutical R&D and relevant properties of the structure and evolution of the industry network.

Our findings strongly suggest that a mapping is in place between the specific nature of problem decompositions and research techniques observed at the micro level of knowledge and technology dynamics and the patterns of structural evolution detected at the macro level of the industry network.

Our empirical analysis of network evolution relies on graph theoretical tools and measures to investigate an extensive data set that covers more than 5000 collaborative agreements among around 2000 firms/institutions from 1978 to 1997.

The mathematical language provided by the theory of directed graphs enables us to show how the nature and evolution of underlying relevant technological conditions induce distinguishable patterns of change at the level of industry structure and evolution. A set of indicators is developed, which turn out to be very useful to unravel the complex properties of empirical objects such diverse as technological and industrial structures.

The paper is organized as follows.

Section 2 briefly highlights the nature and goals of some fundamental research heuristics and techniques developed by firms and institutions in the last 20 years in their efforts to discover and develop new effective drugs. A fundamental distinction is captured, between co-specialized and transversal research technologies/strategies. That is, between research hypotheses and techniques that tend to be specific to particular domains and research techniques that are generic and, at the same time, complement co-specialized hypotheses and techniques in the course of research activities.

In Section 3, we highlight some implications of the nature of these heuristics and research strategies on the organization of innovative activities and on patterns of evolution of the network of R&D collaborative relationships.

In Section 4, we turn to the empirical analysis of the evolution of the network. Graph theory and numerical representations of networks are introduced, coming to show the existence of a striking homomorphic relationship with the structure and evolution of most recurrent research hypotheses and techniques used in problem solving activities. We refer to the notion of Canonical Decomposition of a graph in order to disentangle two major drivers/components of the structural evolution of the net, i.e., co-specialized and transversal actors that rely on co-specialized and transversal research techniques.

The presentation of the main findings and the discussion of some implications for the analysis of organization and industrial dynamics close the paper.

Section snippets

The growth of scientific and technological knowledge in pharmaceutical R&D

The last 25 years have witnessed a revolution in biological sciences, with significant basic advances in molecular biology, cell biology, biochemistry, protein and peptide chemistry, physiology, pharmacology, and other relevant scientific disciplines. The application of these new bodies of knowledge to pharmaceutical industry has had an enormous impact on the nature of R&D activities, on organizational capabilities required to introduce new drugs, and on patterns of industry evolution (see

From growth of knowledge to network dynamics

So far, we have identified some properties of the processes of scientific discovery underpinning research activities in the pharmaceutical industry. An extensive literature has documented some of the consequences that the advent of molecular biology has produced on the organization of innovative activities, both at the firm level and at the industry level Orsenigo, 1989, Henderson, 1994, Gambardella, 1995, Mc Kelvey, 1995, Galambos and Sturchio, 1996. In particular, it has been emphasized that

The evolution of the industry network

This section analyzes in detail the transformations occurred in the organization of innovative activities within the international pharmaceutical industry from 1978 to 1997.

Several graph theoretical measures are applied to investigate the evolution of the inter-organizational R&D activity that has characterized the pharmaceutical industry after the emergence of molecular biology.

The analysis is based on a unique data set built at the University of Siene by integrating several fonts. In

Concluding discussion

In this paper we have analyzed the structural evolution of the network of collaborative agreements in pharmaceutical R&D in the last 20 years. Our results reveal that some fundamental properties of the processes of growth of relevant knowledge bases are preserved in the structural evolution of the net.

Specifically, both the growth of knowledge and the structural evolution of the network have been characterized by fast expansion, proliferation of research trajectories and techniques, and

Acknowledgements

The authors wish to thank Giovanni Dosi, Alfonsa Sambandella, Stanley Metcalfe, Keith Pavitt, Pier Paolo Saviotti, and paricipants to seminars held at Universidad Pompeu Fabra, Barcelona; Universidad Complutense, Madrid; SPRU, University of Sussex; INRA, Grenoble; Columbia University, New York; BETA, Strasbourg, for useful comments. Support from the Merck Foundation (EPRIS project), the European Commission, DG XII, and from the Italian Ministry of University and Scientific Research (MURST) is

References (35)

  • G. Dosi

    Sources, procedures and microeconomic effects of innovation

    Journal of Economic Literature

    (1988)
  • A.L. Dulmage et al.

    Some generalization of the problem of distinct representatives

    Canadian Journal of Mathematics

    (1958)
  • A.L. Dulmage et al.

    A structure theory of bipartite graphs of finite exterior dimensions

    Transactions of the Royal Society Canada Series

    (1959)
  • B.S. Everitt

    The Analysis of Contingency Tables

    (1977)
  • O. Frank et al.

    Cluster inference by using transitivity indices in empirical graphs

    Journal of the American Statistical Association

    (1982)
  • L. Galambos et al.

    The pharmaceutical industry in the twentieth century: a reappraisal of the sources of innovation

    History and Technology

    (1996)
  • A. Gambardella

    Science and Innovation. The U.S. Pharmaceutical Industry During the 1980s

    (1995)
  • Cited by (0)

    View full text