UNIVERSITY RESEARCH AND REGIONAL INNOVATION: A Spatial Econometric Analysis of Academic Technology Transfers

UNIVERSITY RESEARCH AND REGIONAL INNOVATION: A Spatial Econometric Analysis of Academic Technology Transfers

Economics of Science, Technology and Innovation VOLUME 13 Series Editors Cristiano Antonelli, University o/torino, Italy Bo Carlsson, Case Western Reserve University, USA. Editorial Board: Steven Klepper, Carnegie Mellon University, USA. Richard Langlois, University o/connecticut, USA. J.S. Metcalfe, University o/manchester, UK. David Mowery, University o/california, Berkeley, USA. Pascal Petit, CEPREMAP, France Luc Soete, Maastricht University, The Netherlands The titles published in this series are listed at the end of this volume.

UNIVERSITY RESEARCH AND REGIONAL INNOV ATION: A Spatial Econometric Analysis of Academic Technology Transfers by ATTILA VARGA Institute for Urban and Regional Research Austrian Academy of Sciences Vienna, Austria ~. " SPRINGER SCIENCE+BUSINESS MEDIA, LLC

Library of Congress Cataloging-in-Publication Data Varga, Attila, 1958- University research and regional innovation : a spatial econometric analysis of academic technology transfers / by Attila Varga. p. cm. - (Economics of science, technology, and innovation ; v. 13) Includes bibliographical references and index. ISBN 978-1-4613-7556-2 ISBN 978-1-4615-5587-2 (ebook) DOI 10.1007/978-1-4615-5587-2 1. Technology transfer-united States. 2. Research, Industrial-United States. 3. Industry and education-united States. 4. High technology industries-united States. 1. Title. II. Series. T174.3.V37 1998 338.97307-dc21 98-29954 CIP Copyright ~ 1998 by Springer Science+Business Media New York Originally published by Kluwer Academic Publishers in 1998 Softcover reprint ofthe hardcover Ist edition 1998 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, mechanical, photo-copying, recording, or otherwise, without the prior written permission of the publisher, Springer Science+Business Media, LLC. Printed on acid-free paper

10 Ani and Zsombi

TABLE OF CONTENTS LIST OF TABLES... IX LIST OF FIGURES... XI FOREWORD BY ZOLTAN J. ACS... XIII PREFACE... XV 1 INTRODUCTION... 1 2 UNIVERSITIES AND THE LOCAL HIGH TECHNOLOGY INDUSTRY: WHAT DO WE ALREADY KNOW?... 7 2.1. INTRODUCTION... 7 2.2. UNIVERSITY KNOWLEDGE EFFECTS ON THE REGIONAL ECONOMy... 8 2.2.1. Technology transfer... 8 2.2.2. University knowledge effect on the regional economy: the experience of the well known high technology concentrations...... 10 2.3. UNIVERSITY EFFECT ON THE LOCATION CHOICE OF HIGH TECHNOLOGY COMPANIES... 12 2.3.1. Factors affecting high technology location... 12 2.3.2. Universities and the choice of high technology location: case studies, surveys, and descriptive works on existing high technology centers... 13 2.4. UNIVERSITIES AND THE SPATIAL DISTRIBUTION OF HIGH TECHNOLOGY PRODUCTION: ECONOMETRIC STUDIES... 14 2.5. UNIVERSITY RESEARCH AND THE SPATIAL DISTRIBUTION OF INDUSTRIAL RESEARCH AND DEVELOPMENT... 20 2.6. MODELS OF TECHNOLOGY TRANSFER: PATENTS, INNOVATIONS, AND KNOWLEDGE PRODUCTION... 22 2.7. SUMMARy... 24 3 SPATIAL DATA ANALySIS... 27 3.1. INTRODUCTION... 27 3.2. SPACE AND DATA ANALYSIS... 27 3.2.1. The nature of spatia I data... 28 3.2.2. Modeling space... 29 3.2.3. Spatial stochastic processes... 30 3.2.4. Unsolved methodological problems..... 31 3.3. EXPLORATORY SPATIAL DATA ANALYSIS... 31 3.3.1. Global measures of spatial association... 32 3.3.2. Local measures of spatial association... 33 3.4. ESTIMATION AND HYPOTHESIS TESTING... 35 3.4.1. The spatial lag model..... 35 3.4.2. The spatial error model... 36 3.4.3. Specification diagnostic and spatial effects... 37

viii 3.5. SUMMARy... 40 4 UNIVERSITY RESEARCH AND THE SPATIAL DISTRIBUTION OF HIGH TECHNOLOGY INNOVATIONS AND PRIVATE RESEARCH... 45 4.1. INTRODUCTION... 45 4.2. THE INNOVATION, PRIVATE RESEARCH AND UNIVERSITY RESEARCH DATA... 46 4.3. THE SPATIAL DISTRIBUTION OF HIGH TECHNOLOGY INNOVATIONS... 50 4.4. THE SPATIAL DISTRIBUTION OF HIGH TECHNOLOGY R&D ACTIVITIES... 55 4.5. THE SPATIAL DISTRIBUTION OF HIGH TECHNOLOGY UNIVERSITY RESEARCH AND ITS RELATION TO INNOVATIONS AND INDUSTRIAL RESEARCH... 58 4.6. SUMMARy... 63 5 LOCAL KNOWLEDGE TRANSFERS: STATE LEVEL ANALYSIS... 67 5.1. INTRODUCTION... 67 5.2. STATE ANALYSIS AND LOCAL KNOWLEDGE TRANSFERS: SUMMARY OF EARLIER FINDINGS... 68 5.3. ALTERNATIVE INDICATORS OF LOCAL UNIVERSITY KNOWLEDGE TRANSFERS 72 5.4. EMPIRICAL RESULT... 74 5.5. SUMMARy... 78 6 THE SPATIAL EXTENT OF UNIVERSITY EFFECTS: MSA LEVEL ANALySIS... 81 6.1. INTRODUCTION... 81 6.2. THE MODEL... 82 6.2. ESTIMATION ISSUES... 89 6.3. EMPIRICAL RESULTS... 90 6.4. SUMMARy... 97 7 FACTORS GOVERNING UNIVERSITY EFFECTS... 101 7.1. INTRODUCTION... 101 7.2. THE EMPIRICAL MODEL... 102 7.3. REGRESSION RESULTS... 105 7.4. SPATIAL VARIATION IN THE INTENSITY OF UNIVERSITY KNOWLEDGE TRANSFERS... 107 7.5. THE "CRITICAL MASS" OF AGGLOMERATION... 112 7.6. SUMMARy... 119 8 SUMMARY AND CONCLUSIONS... 121 APPENDIX A: DEFINING HIGH TECHNOLOGY FOR THE EMPIRICAL STUDY... 127 APPENDIX B: VARIABLE DEFINITIONS AND SOURCES... 131 REFERENCES... 135 INDEX... 149

List of Tables Table 2.1. The literature on high technology plants and universities 15 Table 2.2. The literature on high technology labor and universities 17 Table 2.3. The literature on industry R&D and universities 21 Table 3.1. Tests for spatial dependence in regression models 39 Table 4.1. Linking industries to university departments 48 Table 4.2. Characteristics of the innovation, R&D employment and university research data 49 Table 4.3. The distribution of innovations and US counties by innovation value ranges 51 Table 4.4. The distribution of R&D employment and US counties by R&D value ranges 56 Table 4.5. The distribution of university research expenditures and US counties by university research ranges 59 Table 4.6. The distribution of innovations and innovating counties by university research ranges 62 Table 4.7. The distribution of R&D employment and R&D counties by university research ranges 63 Table 5.1. The literature on knowledge transfer from universities 71 Table 5.2. Innovations, R&D employment, and university research expenditures for 43 US states - 1982 75 Table 5.3. Regression results for Log (Innovations) at the state level (N = 43, 1982) 76 Table 6.1. Innovations, R&D lab employment, and university research expenditures for 125 US MSAs - 1982 86 Table 6.2. OLS regression results for Log (Innovations) at the MSA level (N = 125,1982) 91 Table 6.3. Regression results for Log (Private R&D) at the MSA level (N = 125, 1982) 94 Table 6.4. Regression results for Log (University Research) at the MSA level (N=125,1982) 96 Table 7.1. Parameter expansion regression results for Log (Innovations) at the MSA level (N = 125, 1982) 106 Table 7.2. Innovation elasticities and some local area characteristics for 125 MSAs -1982 109 Table 7.3. Innovations and the values of certain indicators of agglomeration by innovation elasticity categories 112 Table 7.4. Predicted and observed innovations for 125 MSAs - 1982 114

x Table 7.5. Average marginal university research expenditures and average total university research expenditures by innovation elasticity categories - 1982 I 18 Table A.I. High technology industries based on the ratio of R&D expenditures to sales 128 Table A.2. High technology industries based on the percentage of engineers, engineering technicians, scientists, and mathematicians of total industry occupations 129 Table A.3. High technology industries according to the R&D, employment structure, and innovation criteria 130

List of Figures Figure 4.1. The spatial distribution of innovations Figure 4.2. Innovation clusters Figure 4.3. Moran scatterplot : high technology innovations Figure 4.4. Moran scatterplot map: high technology innovations Figure 4.5. The spatial distribution of industrial R&D employment Figure 4.6. Industrial R&D employment clusters Figure 4.7. Moran scatterplot: industrial R&D employment Figure 4.8. Moran scatterplot map: industrial R&D employment Figure 4.9. The spatial distribution of university R&D expenditures Figure 4.10 University R&D clusters Figure 4.11. Moran scatterplot: university R&D expenditures Figure 4.12. Moran scatterplot map: university R&D expenditures Figure 7.1. MSAs by predicted innovation elasticity ranges Figure 7.2. Expected innovations by innovation elasticity ranges 51 52 54 54 55 56 57 58 59 60 61 61 113 117

Foreword By Zoltan J. Acs There is little question that economic activity of all types is moving in the direction of globalization. Globalization refers to the web of linkages and interconnections between states, societies and organization that make up the present world economic system. Globalization creates new structures and new relationships, with the result that business decisions and actions in one part of the world have significant consequences in other places. Global changes in the nature of economic activity are provoking a reexamination of cornerstone factors of production, investment and trade. Because of our growing understanding of constructed endowments, such as knowledge, the literature dealing with the economics of innovation, growth, and regional development has allowed us to move beyond the neo-ricardian notion of landlocked 'comparative advantage' and beyond Porter's variation of market-locked 'comparative advantage' to the point where we can now speak of 'constructed advantages'. These 'constructed advantages' are conditioned by bounded rationalizes, technological and organizational complexities, and badly behaved dynamics. It allows us to begin to analyze the knowledge-laden content of regional development. As we approach the 21st century a worldwide system of production and distribution is evolving in much the same way as national markets evolved from local and regional networks during the 19 th century. In fact a new paradoxical logic appears to drive the world wide production systems towards an increasing degree of globalization and an increasing degree of sub-national regionalization. The paradoxical consequences of globalization have been not only to generate balkanization but also to create the need for new forms of regional organizations. In the new global knowledge intensive economy, despite continued predictions of the end of geography, regions are becoming more important centers of economic and technological activity. Although there have been numerous excellent studies of the dynamics of individual regions, the role of regions in the new global knowledge-based economy remains rather poorly understood. And, while several outstanding studies have chronicled the rise of the knowledge-based economy, outlined the contours of learning regions, virtually no one has developed a comparable theory of what such changes portend for regions and regional organizations. A central aspect of studies of knowledge-based innovation at the regional scale is the role of the university research in regional innovation systems. Universities playa central role in this process as producers of basic research. The importance of basic (university research) is derived from the public good nature of the research, and the resulting positive externalities to the private sector in the form of knowledge spillovers. However, a fundamental issue, which remains unresolved

xiv in the economics of technology, is the identification and measurement of regional R&D spillovers. Jaffe was the first to identify the extent to which university research spills over into the generation of commercial inventions. His statistical results provided evidence that corporate patent activity responds positively to commercial spillovers from university research. Building on Jaffe's work Feldman expanded the knowledge production function to innovative activity and incorporate aspects of the regional knowledge infrastructure. She found that innovative activity is conditioned by the knowledge infrastructure, and responds favorably to spillovers from university research strengthening Jaffe's findings. Attila Varga builds on this solid foundation by directly examining the role of universities in regional innovation. His main concern is whether universitygenerated economic growth observed in certain regions and for select industries can be achieved by other regions. He extends the Jaffe-Feldman approach by focusing on a more precise measure of local geographic spillovers. Varga approaches the issue of knowledge spillovers from an explicit spatial econometric perspective and implements the classic knowledge production function for 125 Metropolitan Statistical Areas, yielding more precise insights into the range of spatial externalities between innovation and R&D in a regional context. The Jaffe-Feldman-Varga research into R&D spillovers takes us a long way toward understanding the role of R&D spillovers in knowledge-based economic development.

Preface A sizable literature has documented the important role of universities in the development of the world's largest high technology concentrations: Stanford University in Silicon Valley, MIT in Boston and Cambridge University in Cambridge (U.K.). The basic research question behind this book is to assess the extent to which local university knowledge impacts are unique, non-repeatable phenomena or whether they can be experienced in other regions as well. This book represents the first study in the literature that provides a systematic, US-wide analysis of university-high technology connections at the lowest possible level of spatial aggregation. Its contributions are twofold: (I) it focuses on the regional aspects of the interaction between high technology innovations, university and private R&D at the proper spatial scale, at the level of counties and metropolitan areas; (2) it uses the specialized methodology of spatial econometrics to explicitly deal with potential spatial effects in cross-sectional data. The study is based on a unique data set of high technology innovations and industrial R&D employment in the US in 1982. From an exploratory spatial data analysis implemented at the county level, it is shown that the concentration of high technology product innovations, private R&D and university research follow a similar spatial pattern across the USA. A formal regression analysis is carried out for 43 US states and 125 Metropolitan Statistical Areas which implements the Griliches-laffe knowledge production function framework. This provides strong evidence that university research expenditures have a positive and very significant effect on aggregate high technology innovation. Moreover, this university impact follows a distance decay pattern. Additionally, results for US MSAs indicate that research employment in high technology R&D laboratories is significantly determined by the level of local university research expenditures. However, the intensity of knowledge transfers between university research and regional innovation is not constant over space. It is demonstrated in this book that the same amount of university research expenditure yields substantially different levels of local innovation activity depending on the concentration of economic activities in the metropolitan area. The findings in this book strongly indicate that a stimulation of research activities in universities located in existing agglomerations of high technology production and research has significant positive effects on the regional economy. This book is aimed to serve the interests of both academic researchers in the fields of regional science and economics of technological change and economic developers concerned with practical problems of innovative technology regions. The text may be used in graduate level courses of regional economics, economics of technological change, economics of education, and applied spatial econometrics.

xvi This work originated in my dissertation research at West Virginia University and benefited greatly from the support of the members of my doctoral committee. First, I would like to express my gratitude to my major advisor Luc Anselin, who directed me towards the topic and gave me the first stimulus to study spatial econometrics. His insightful research suggestions and highly valuable comments are greatly appreciated. I am indebted to Zoltan Acs for providing me the innovation data for the research, for the numerous fruitful conversations with him and for his constant encouragement. I am also grateful to Raymond Florax for his helpful discussions with me and his written comments on the manuscript and to Andrew Isserman and Stratford Douglas for their intuitive suggestions for further research. The Academic Press granted permission to use part of the material published in the paper entitled "Local Geographic Spillovers between University Research and High Technology Innovations," which appeared in Volume 42, 1997, of the Journal of Urban Economics. I would like to thank Jean Dailey for carefully editing the manuscript. This book is dedicated to my wife Ani and my son Zsombi. Thank you for your constant emotional support and understanding without which this work would have never been accomplished.