Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
A statistical analysis of the formation and location factors of high -tech centers in the United States, 1950--1997: An evaluation using quasi -experimental control group methods
(USC Thesis Other)
A statistical analysis of the formation and location factors of high -tech centers in the United States, 1950--1997: An evaluation using quasi -experimental control group methods
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
A STATISTICAL ANALYSIS OF THE FORMATION AND LOCATION FACTORS OF HIGH-TECH CENTERS IN THE UNITED STATES, 1950-1997: AN EVALUATION USING QUASI-EXPERIMENTAL CONTROL GROUP METHODS by Junghoon Ki A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Patial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (PLANNING) December 2002 Copyright 2002 Junghoon Ki Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3093777 Copyright 2003 by Ki, Jung-Hoon All rights reserved. ® UMI UMI Microform 3093777 Copyright 2003 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 90007 This dissertation, written by under the direction ofPh Jll Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School in partial fulfillment of re quirements for the degree of DOCTOR OF PHILOSOPHY —Dean of Graduate Studies Date .... DISSERTATION COMMITTEE Chairperson Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENTS The Lord is my shepherd. I shall not be in want. (Psalm 23:1) Most of all, I thank God, for He has faithfully helped me start, work, and finish this dissertation throughout my doctorate years. Actually, this dissertation is totally attributed to God the Lord. I would like to express my gratitude to my supervisor, Professor Peter Gordon, for his patient guidance and understandings. He has taken care of my poor research papers and encouraged me to apply several research fellowship, finally made me a NSF Dissertation Fellowship recipient. Especially, he has instructed me about a basic frame of economic approach and logical thinking in my doctoral years and dissertation. Special thanks go to my committee members, Professor Harry W. Richardson and Professor Jeffrey Nugent. Professor Richardson has inspired me to begin this research and always given me a novel way out of problems of research. Professor Nugent kindly read my dissertation and asked indispensable questions for my research. I also give thanks to Professor James E. Moore, for his nice academic and administrative support throughout my doctoral years. I give thanks to many spiritual leaders, who have prayed for me and given me substantial advice and support. Brother Jung-Min Ha, missionary Eun-Pyo Lee, and Pastor Stephen Kim are those whom I would like to dedicate my dissertation. ii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. My family members have also greatly helped me with their prayer and effort. My parent, parent-in-law, brothers (Jung-Woo and the late Jung-Jin), brother-in-laws (Kune-Jang and Kune-Young), and especially my wife, Kune-Sook has prayed for me and made every effort to help me finish this dissertation. Thanks to them all. This material is based on work supported by the National Science Foundation under Grant No. 0138314. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS ACKNOWLEDGEMENTS ii LIST OF TABLES vii LIST OF FIGURES x ABSTRACT xii CHAPTER 1 INTRODUCTION 1 1.1 Problems Statement and Research Questions 1 1.2 Research Objectives 3 1.3 Significance of the Study 4 1.4 Limitations of the Study 5 1.5 Organization of the Study 6 CHAPTER 2 LITERATURE REVIEW 7 2.1 Chapter Summary 7 2.2 Definition and Characteristics of High-Tech Industries 7 2.2.1 Definitions of High-Tech Industries 7 2.2.2 Characteristics of High-Tech Industries 12 2.3 Landscape and location Factors of High-Tech Centers in the U.S. 14 2.3.1 High-Tech Centers: Metropolitan Agglomeration 16 2.3.2 Significance of Educated Workers in High-Tech Centers 22 2.3.3 The Creative Class: The Reality of Educated Workers 24 2.3.4 Historical Aspects of High-Tech Centers 30 2.3.5 Location Factors of High-Tech Centers 33 2.4 High-Tech Center Indices 38 iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4.1 High-Tech Employment Density 39 2.4.2 Location Quotient 43 2.4.3 Growth Rate 49 2.4.4 Total Wage Income 55 2.4.5 Personal Wage Income 57 CHAPTER 3 THEORETICAL FRAMEWORK AND RESEARCH HYPOTHESES 60 3.1 Chapter Summary 60 3.2 Regional Growth Theories and High-Tech Regions 60 3.3 Quasi-Experimental Control Group Method 66 3.3.1 General Concept 66 3.3.2 Merits of the Method in the Research 68 3.3.3 Weaknesses of the Method in the Research 69 3.4 Research Hypotheses 69 CHAPTER 4 RESEARCH DESIGN AND METHODOLOGY 73 4.1 Chapter Summary 73 4.2 Basic Database Buildup 75 4.2.1 Defining High-Tech Industry 75 4.2.2 Defining High-Tech Center 76 4.2.3 Defining Twin Counties (Control Group) 79 4.3 Statistical Tests 82 4.3.1 T-test I 82 4.3.2 T-test II 84 4.3.3 Regression 85 CHAPTER 5 RESULTS 86 5.1 Chapter Summary 86 5.2 Treatment Year (Period) 86 V Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.2.1 High-Tech Employment 87 5.2.2 High-Tech Total Wage Income 97 5.1.3 High-Tech Personal Wage Income 107 5.3 Significant Location Factors 117 5.3.1 Categories of Location Factors in the Study 117 5.3.2 Descriptive Statistics and T-test Result 119 5.3.3 Regression Result 123 CHAPTER 6 CONCLUSION AND POLICY IMPLICATIONS 132 6.1 Conclusion 132 6.2 Policy Implications 137 BIBLIOGRAPHY 139 APPENDICES 152 vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES Table 2.1 High-Tech Industries’ salaries in Massachusetts, 1998 14 Table 2.2 Contemporary Major High-Tech Centers in the U.S. 15 Table 2.3 Spatial Patterns of High-Tech Industries’ Employment, 1997 19 Table 2.4 High-Tech Industries’ Dispersion Jobs in Massachusetts, 1998 21 Table 2.5 Top 15 High-Tech Counties, by Employment Size, 1997 21 Table 2.6 Top 15 High-Tech Metros, by Real Output Size, 1998 22 Table 2.7 Top 10 High-Tech Center and Creative Class Centers 27 Table 2.8 Specialized Technologies of the High-Tech Centers 30 Table 2.9 General Categories of High-Tech Location Factors 37 Table 2.10 Spatial Patterns of Employment Density in High-Tech Industries, 1997 41 Table 2.11 Top 15 Counties by High-Tech Employment Density, 1997 42 Table 2.12 Spatial Patterns of High-Tech Industries by Location Quotient, 1997 46 Table 2.13 Top 15 High-Tech Counties by Location Quotient, 1997 47 Table 2.14 Top 15 High-Tech Metros by Location Quotient of Output, 1998 47 Table 2.15 Top 10 Milken Institute Tech-Poles, Composite Index, 1998 48 Table 2.16 Growth Rate of Total High-Tech and Manufacturing Employment 49 Table 2.17 Spatial Patterns of High-Tech Employment Growth, 1987-1997 51 Table 2.18 Spatial Patterns of High-Tech Employment Growth, 1987-1992 52 Table 2.19 Spatial Patterns of High-Tech Employment Growth, 1992-1997 53 Table 2.20 Spatial Patterns of Annual Wage Income in High-Tech, 1997 56 Table 2.21 Top 15 High-Tech Counties, by High-Tech Total Wage Income, 1997 57 Table 2.22 Spatial Patterns of Annual Personal Wage Income in High-Tech, 1997 59 Table 2.23 Top 15 High-Tech Counties, by Personal Wage Income, 1997 59 Table 4.1 Research Steps of the Research 75 Table 4.2 High-Tech Industry by 20 SIC codes 76 Table 4.3 High-Tech Centers and the Values of Three Criteria 78 vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.4 High-Tech Centers and Their Twin Counties in the U.S. 81 Table 4.5 Spatial Patterns of High-Tech Centers and their Twins in Metropolitan and Non-metropolitan Areas in 1997 82 Table 4.6 How to Interpret T-test Result to Decide the Period of High-Tech Center83 Table 5.1 Descriptive Statistics for High-Tech Employment (1951-1997) 87 Table 5.2 Descriptive Statistics for High-Tech Manufacture Employment 88 Table 5.3 Descriptive Statistics for High-Tech Service Employment (1951-1997) 88 Table 5.4 T-test Results: High-Tech Employment 89 Table 5.5 T-test Results: High-Tech Manufacture Employment 90 Table 5.6 T-test Results: High-Tech Service Employment 91 Table 5.7 Total High-Tech Employment in High-Tech Group and Twin Group 92 Table 5.8 Average High-Tech Employment in High-Tech Group and Twin Group 92 Table 5.9 Actual High-Tech Employment Growth in High-Tech and Twin Group 95 Table 5.10 Descriptive Statistics for High-Tech Total Wage Income (1951-1997) 97 Table 5.11 Descriptive Statistics for High-Tech Manufacture Wage Income 97 Table 5.12 Descriptive Statistics for High-Tech Service Wage Income 98 Table 5.13 T-test Results: High-Tech Total Wage Income 99 Table 5.14 T-test Results: High-Tech Manufacture Wage Income 100 Table 5.15 T-test Results: High-Tech Service Wage Income 101 Table 5.16 Total High-Tech Wage Income, (Annual, $1000) 102 Table 5.17 Average High-Tech Wage Income, (Annual, $1000) 102 Table 5.18 Actual High-Tech Wage Income Growth 105 Table 5.19 Descriptive Statistics for High-Tech Personal Wage Income 107 Table 5.20 Descriptive Statistics for High-Tech Manufacture Personal Wage 107 Table 5.21 Descriptive Statistics for High-Tech Service Personal Wage Income 108 Table 5.22 T-test Results: High-Tech Personal Wage Income 109 Table 5.23 T-test Results: High-Tech Manufacture Personal Wage Income 110 Table 5.24 T-test Results: High-Tech Service Personal Wage Income 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.25 Total High-Tech Personal Wage Income, (Annual, $) 111 Table 5.26 Average High-Tech Personal Wage Income, (Annual, $) 112 Table 5.27 Actual High-Tech Personal Wage Income Growth 115 Table 5.28 Category of Location Factors 118 Table 5.29 T-test Result: High-Tech Location Factors in 1970 120 Table 5.30 T-test Result: High-Tech Location Factors in 1980 122 Table 5.31 Descriptive Statistics of Variables (1980) 124 Table 5.32 Result of Correlation Analysis 124 Table 5.3 3 Result of Regression Analysis 127 ix Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES Figure 2.1 High-Tech Employments by the U.S. Counties, 1997 17 Figure 2.2 Top 100 High-Tech Counties by High-Tech Employment in the U.S. 18 Figure 2.3 High-Tech Employment Density by the U.S. Counties, 1997 40 Figure 2.4 High-Tech Centers by the Location Quotient, 1997 45 Figure 2.5 High-Tech Centers by the Growth Rate, 1987-1997 54 Figure 2.6 High-Tech Annual Total Wage Income by the U.S. Counties, 1997 55 Figure 2.7 High-Tech Annual Personal Wage Income by the U.S. Counties, 1997 58 Figure 3.1 Quasi-Experimental Design Diagram 67 Figure 3.2 Standardized Distance for a Single Feature 68 Figure 3.3 Mahalonobis Distance for Variable Matrix 68 Figure 4.1 Research Design and Research Targets 74 Figure 5.1 Average High-Tech Employment Trend, 1951-1997 93 Figure 5.2 Average High-Tech Manufacture Employment Trend, 1951-1997 93 Figure 5.3 Average High-Tech Service Employment Trend, 1951-1997 94 Figure 5.4 Average High-Tech Employment Growth Rates 95 Figure 5.5 Average High-Tech Manufacture Employment Growth Rates 96 Figure 5.6 Average High-Tech Service Employment Growth Rates 96 Figure 5.7 Average High-Tech Wage Income Trend, 1951-1997 103 Figure 5.8 Average High-Tech Manufacture Wage Income Trend, 1951-1997 103 Figure 5.9 Average High-Tech Service Wage Income Trend, 1951-1997 104 Figure 5.10 Average High-Tech Wage Income Growth Rates 105 Figure 5.11 Average High-Tech Manufacture Wage Income Growth Rates 106 Figure 5.12 Average High-Tech Service Wage Income Growth Rates 106 Figure 5.13 Average High-Tech Personal Wage Income Trend, 1951-1997 112 Figure 5.14 Average High-Tech Manufacture Personal Wage Income Trend 113 Figure 5.15 Average High-Tech Service Personal Wage Trend, 1951-1997 114 Figure 5.16 Average High-Tech Personal Wage Income Growth Rates 115 Figure 5.17 Average High-Tech Manufacture Personal Wage Growth Rates 116 X Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.18 Average High-Tech Service Personal Wage Income Growth Rates Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT In this thesis, I examine when high-tech centers were formed in the United States and what location factors contributed significantly to the formation of high- tech centers during the period. I examine these questions by using quasi- experimental control group methods with the high-tech center group and the twin group that is composed of U.S. counties whose educational and economic conditions are most similar to those of the high-tech center group. There are major three findings from the research results. First, the high-tech centers were substantially formed during the 1970-1980 period in the United States. High-tech industries had their origin in the 1940s or 1950s, but their becoming high- tech centers came true in the 1970-1980 period. Second, high-tech services played a leading role in the formation of high-tech centers during the 1970-1980 period in the United States. Meanwhile, high-tech manufacturing played a smaller role because it followed the general trend of manufacturing decentralization into non-metropolitan areas. Third, most of the conventional high-tech location factors did not play a role in building up high-tech centers during the 1970-1980 period in the United States. This finding confirms that high-tech centers were established spontaneously without specific planning or policy during the period of formation. More careful approach and interpretation of research results, however, provide some insights that a local educated and affluent class can be a significant location factor of high-tech centers through “quality of life” improvements. Universities and university medical centers seem to have played a unique role in xii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. incubating and spinning off such educated and affluent class and workers who have a higher statistical relation with the formation of high-tech centers in the United States. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 1 INTRODUCTION 1.1 Problems Statement and Research Questions Many academic and business research efforts have been undertaken to determine the reasons for successful high-tech centers, mostly using qualitative and quantitative methods. These studies seek to find an inevitable determinant or an essential factor from the performance of successful high-tech centers such as Silicon Valley in California, Route 128 in Massachusetts, or Research Triangle Park in North Carolina. Various ideas are tested and suggested in discussions of high-tech centers of the United States. One group of researchers ascribes a basic reason for their success to military spending and contracts; meanwhile another group of studies emphasizes the relationship between educational institutions (such as local research universities) and high-tech firms in the successful high-tech centers. Still other researchers claim that high-tech industries were bom and grew in better business environments, quality of life, and combinations of factors, including agglomeration economies. Recent researchers praise the role of quality of life, which has played an important part in the formation of high-tech center by attracting educated high-tech workers. Though there have been many studies of high-tech centers, most previous research has a methodological weakness in search of location factors or determinants that affect the formation and growth of high-tech centers. A serious gap in the study Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of location factors is that most high-tech centers research focuses on the performance of the high-tech centers themselves. The plausibility of conclusions from such studies is limited because no control groups were used. First, we may lose rigor in determining periods of high-tech formation and growth without using a control group. If we choose any period as a time of high-tech formation and growth without a control group, this decision may have a significant possibility of statistical error. Percentage increases of high-tech employment or high-tech wage income may mean nothing without comparison to a control group. Second, such research cannot really identify the effects of high-tech location factors or determinants. “What would have happened without these factors?” This problem of determining what would have happened without a particular factor is, of course, not unique to regional science and economics. “Before a cure can be attributed to a particular medicine or an improvement in test scores to a particular course, it is necessary to determine what would have happened without the medicine or the course.” This is a frequent question in medical field research according to Isserman and Merrifield (1982), who make the important point that although a vast literature on experimental and quasi- experimental designs exists, the use of such methods is rare in regional science. Acknowledging these problems of current high-tech studies, I formulate two research questions. These questions come from an awareness of the methodological weaknesses in most high-tech research. The answers to these questions will be addressed in a statistical analysis of the selected U.S. high-tech centers (counties) and their “twin” counties. 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Research Question 1 What is the period that high-tech centers started in the United States? Do the quasi- experimental control group method give a better insight for identifying this period? Research Question 2 What are the fundamental determinants and location factors associated with high- tech centers in the United States? Do suggested quasi-experimental control group methods produce results that are different from what previous studies and literature claim? 1.2 Research Objectives There are two main objectives of this research, which are drawn from the two suggested research questions. First, this research is designed to identify a period when high-tech centers started in the United States. Quasi-experimental control group methods, with statistical tests, will be used to decide the starting period of high-tech growth. Second, this research will also identify substantial location factors that play an important role in the rise of high-tech centers. To reach these objectives, theoretical frameworks and research hypotheses are suggested based on previous studies that deal with characteristics, major determinants and location factors of U.S. high-tech centers. Comparisons between treatment groups (high-tech centers) and control groups (twin counties) will be made over sequential time intervals. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.3 Significance of the Study This research is probably the first in identifying suitable control groups and then systematically testing hypotheses on the explanation of high-tech centers formation and their substantial location factors. Isserman has produced several papers on policy analysis and economic impacts using quasi-experimental control group methods (Isserman and Merrifield, 1982; Isserman and Merrifield, 1987; Isserman and Beaumont, 1989; Rephann and Isserman, 1994; Isserman, 1999). Though his studies contribute substantially to policy analysis and economic impact investigations in the United States, his research using quasi-experimentation did not deal with U.S. high-tech policy and high-tech centers. “Technology in the Garden” by Luger and Goldstein (1991) is a unique high-tech related study that adopts the quasi-experimental control group tool as a research park success indicator (Luger and Goldstein, 1991). However, it deals with only planned science parks in the United States and compares their performances. Our understanding of how and why certain regions prosper and others decline is still fragmentary. Whereas there are some well-known opinions on the reasons for high-tech center success, these are still too vague to allow policy makers to try implementing them. Significant innovation is now routinely taking place in selected high-tech centers. A better understanding of how these centers grow and flourish will add considerably to discussions of innovation diffusion. If the factors underlying high-tech center growth are understood well enough so that new centers 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. can be strategically planned, then innovation diffusion can, to some extent, fall within the realm of regional planning. There has also been considerable discussion of "digital divides" and such. One possible approach to remedying problems of this sort is to identify and implement policies that foster high-tech centers in places where they are less likely to otherwise emerge. Findings from the research will also inform this type of policy discussion. 1.4 Limitations of the Study There are still some limitations of the suggested approach. First, this research using a quasi-experimental control group method does more of testing the outcomes of current high-tech studies, rather than showing anything new. The result of this research may support a group of studies, but it may undermine another group of research. Second, the approach of this research can not wholly control for endogeneity because this research can not consider all social and economic conditions of base year (1950) in selecting a control group (twin counties). I control some of the most important social and economic conditions in the base year (1950): education, employment, spatial size and land use. The control group selection is designed to discover twin counties whose education base, employment structure, and land size are similar to high-tech centers in 1950. The selected twin group is composed of counties with similar education, employment, and spatial structure with the high-tech group. Third, there is a constraint of data availability in location 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. factors. Thus I just use twenty-eight possible location factors in this study, mostly obtained from the County and City Data Book (U.S. Census Bureau). 1.5 Organization of the study The rest of the research is organized into five chapters. The second chapter reviews the general definitions and characteristics of high-tech workers, high-tech industries, the spatial patterns of high-tech centers in the U.S. This chapter also introduces metropolitan advantages, historical aspects, location factors and some indices of high-tech centers in the U.S. The third chapter builds research hypotheses by means of pivotal regional growth theories and the quasi-experimental control group method. Chapter four is devoted to research design and methodology in the framework of the quasi-experimental control group method. This chapter describes how high-tech centers and their twin counties were selected, how to decide the period of the beginning of high-tech center, and how to select the substantial location factors of high-tech formation. Some indices for defining a high-tech center are also introduced in this chapter. Chapter five displays the results of this research, its significance and limitations of the suggested method and research. Chapter six discusses the main conclusions from the research and their implications for policy, particularly for planners, policy-makers, and decision-makers, who would like to encourage high-tech activities and location within their regions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 2 LITERATURE REVIEW 2.1 Chapter Summary Chapter two is committed to a literature review of high-tech industries and high-tech centers. I review previous studies and research to properly define high-tech industries, examine their major characteristics, show the spatial patterns of high-tech centers, and describe their metropolitan rationales, historical aspects and location factors. At the end of the chapter, I propose five high-tech center indices; three of them will be used to define the high-tech centers of this research. 2.2 Definitions and Characteristics of High-tech Industries 2.2.1 Definitions of High-tech Industries The term high-tech is frequently suggested to describe not only industries, but also occupations and products. High-tech industries have been introduced as a remedy for economic decline in older industrial sectors. They are generally defined as the kind of companies associated with innovations, new products, and new processes. Premus (1982) explains that high-tech industries consist of heterogeneous collections of firms that share several attributes as follows: First, the firms are labor-intensive rather than capital- intensive in their production processes, employing a higher percentage of technician, engineers and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. scientists than other manufacturing companies. Second, the industries are science-based in that they thrive on the application of advances in science to the market place in the form of new products and production methods. Third, R&D inputs are much more important to the continued successful operation of high technology firms than is the case for other manufacturing industries. (Premus, 1982: 4) Then, what are the main features that distinguish high-tech industries from other types of business enterprise? It is true that there seems no clear standard for making a distinction between firms that are high-tech and those that are not (Cordes et al., 1986), but there is fairly wide agreement on the general characteristics of high- tech industries and the criteria for developing lists of such industries (Premus, 1982; Hecker, 1999). A generally accepted approach for identifying high-tech industries emphasizes whether developing or applying new technological knowledge plays a vital role in the competitive strategy of the firm. A firm or an industry would be considered high-tech if one of its primary assets is the possession of advanced technological knowledge used to develop new products or processes (Cordes et al., 1999). The Congressional Office of Technology Assessment (1982) notes that “high-technology firms are engaged in the design, development, and introduction of new products and innovative manufacturing processes, or both, through the systematic application of scientific and technical knowledge.” The problem with this approach is, however, that technological knowledge is an intangible asset that is not as easily measured as are tangible assets, and identifying new products or innovative 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. manufacturing processes is subjective. To overcome these weaknesses, several ways have been developed to quantify a firm’s commitment to the developing new products and processes. Charles River Associates (1976) suggest six criteria1 for distinguishing technology-based firm provide twenty four-digit SIC code industries2 that satisfy all six criteria (Charles River Associates, 1976). A Congressional Office of Technology Assessment (1982) explains that high-tech firms include (i) a high proportion of expenditures to R&D and employ (ii) a high proportion of scientific, technical, and engineering personnel (Office of Technology Assessment, 1982). A National Science Foundation report (1988) provides a similar idea on science and technology resources, saying that (i) the employment of scientists, engineers, and technicians and (ii) measures of R&D activities are two most important parameters of innovation, and are surrogates for measuring the broader concept of innovation (National Science Foundation, 1988). Cordes and colleagues (1986) claims that firms are considered high-tech on the basis of (i) the extent of technology embodied in products and production processes, (ii) the determination that certain types of firms produce disproportionately more innovative outputs than others, and (iii) relative expenditures on innovative inputs, such as scientific and technical workers, and especially R&D expenditures (Cordes et al., 1986). Webre (1985) provides eight 1 They are (i) the degree to which a product is proprietary, (ii) how recently the underlying technology was developed, (iii) the extent to which a new market is created or an existing market is substantially transformed, (iv) the extent to which a product was based on scientific research, (v) rapidity of technological obsolescence, and (vi) the size of R&D expenditures required to develop a product 2 They clustered in the broad industrial groups of electrical equipment, electronic components, chemical and allied products, pharmaceuticals, professional and scientific instruments, and aircraft and missiles (Cordes et al., 1999:6) 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. •2 three-digit SIC industries as being high-tech using two attributes, which define high-tech industries: (i) high R&D intensity (ratio of R&D to sales one-third higher than manufacturing average), and (ii) rapid growth (ten-year increase in employment above all manufacturing industries average) (Webre, 1985). The Organization for Economic Cooperation and Development (OECD) suggests a definition of high-tech industries, which applies to all major industrialized countries. The classification is based on R&D intensities as measured by R&D expenditures as a percentage of production, and yields six industrial sectors4 as high-tech (National Science Board, 1993). Another approach to identifying high-tech industries comes from the recognition of the importance of diffusion of new technologies, products, and processes as well as their production. Some companies are more likely to play an important role in the diffusion of new technologies as users of new products or processes, even though such firms do not have a usual association with the development of such new products or processes (Cordes et al, 1999). The list of high-tech industries in the 1992 report of the U.S. President on the state of small business is compatible with this definition of high-tech industries (Executive Office of the President, 1993; Cordes et al., 1999). This list includes many industries in the service and manufacturing sectors that are important users of new technologies, but 3 They are drugs, industrial organic chemicals, computer and office equipment, communications equipment, electronic components and accessories, aircraft and parts, guided missiles, space vehicles and parts, and instruments. 4 They are aircraft (aerospace), office and computing equipment, communications equipment, drugs and medicines, scientific instruments, and electrical machinery. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. which are relatively less R&D-intensive. Therefore, the list of high-tech industries includes users as well as producers of high-tech goods. A variety of indicators are suggested to define the high-tech industries, which implies that it is very difficult to define high-tech industries using only one factor. The most commonly used indexes are “R&D intensity, or the percentage of sales expended on R&D” and “technical workers (scientists, engineers, and technicians) as a percentage of the workforce” (Charles River Associates, 1976; Office of Technology Assessment, 1982; Premus, 1982; Webre, 1985; Cordes et al., 1986; National Science Foundation, 1988; Malecki, 1991; National Science Board, 1993). McArthur’s diffusion-based definition such as “newly emerging and widely diffused technologies” also provides a substantial and detailed definition of high-tech industries (McArthur, 1990). These parameters converge in certain industrial sectors such as office machines, information technology, software, semiconductors, biotechnology, and new material industries (e.g. ceramics, superconductor). Their employment and establishment are operationally determined using two-digit, three- digit or four-digits Standard Industrial Classification (SIC) code industries (Premus, 1982; Hecker, 1999). High-tech occupations are scientific, technical, and engineering occupations, the same group of jobs that is utilized to define high-tech industries (Hecker, 1999). They contain engineers; computer specialists; mathematical specialists; life and physical scientists, technicians; and computer managers. These workers are frequently called educated professional workers, or technical worker, because they 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. require special and detailed knowledge of science, engineering, and mathematics, which is usually attained from post-high school education, or higher education such as master or doctoral degrees. They usually work in the R&D sectors or departments to increase scientific knowledge; development of products and production processes; design of equipment; computer application and system development; and management and administration (Hecker, 1999). High-tech firms make efforts to attract, gain, and retain this kind of educated and technical workers. 2.2.2 Characteristics of High-tech Industries High-tech industries’ distinctive features include knowledge-intensity with high levels of R&D investment, high value-added products, system-integration, and higher investment risk. More appealing traits of high-tech industries are, however, rapid growth rates, higher employment rates, and higher pay. First, high-tech industries are identified by very rapid growth rates compared to other industries (Webre, 1985; Goss and Vozikis, 1994). The U.S. Bureau of Labor Statistics in its reports of employment and earnings shows the fact that from 1975 to 1990, high-tech manufacturing grew 17.8% while non-high-tech manufacturing increased only 3.3% (Bureau of Labor Statistics, 1975,1990). Second, a higher employment rate is another sign of high-tech industries. The employment rate of high-tech industries is more than twice that of traditional industries. However, cuts in defense spending over the 1986-1996 period allowed some high-tech manufacturing employment to decline, and caused a slow down in 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. other industries’ growth. This phenomenon took place because employment of some high-tech industries depends substantially on defense spending and military contracts (Markusen, 1991; Hecker, 1999). Third, high-tech jobs are well paid as previous studies show (National Science Foundation, 1996: 6-5 and 6-16; American Electronics Association, 1999: 7). A recent report of the Massachusetts Division of Employment and Training briefly mentions that, “High-tech means high pay” (Massachusetts Division of Employment and Training, 1999). Hecker (1999) provides evidence that median wages in every high-tech industry in 1997 are higher than the median for all industries (Hecker, 1999: 20). In ten high-tech industries, wages are more than 50 percent higher than the median for all industries. In a statewide study in 1998, Massachusetts Division of Employment and Training (1999) shows that salaries in high-tech industries averaged $62,726 a year, which is $24,939 or 66 percent more than the $37,787 paid on average to the Massachusetts worker (Massachusetts Division of Employment and Training, 1999). On a regional basis, the percent differential is much bigger. In the Central Region of Massachusetts, high-tech industries pay on average 74 percent more than the average for all industries (See Table 2.1). High-tech workers are able to enjoy a better quality of life with higher income from higher wages. High-tech industries, thus, invigorate other sectors of economy, providing higher personal incomes. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.1 High-tech Industries’ salaries in Massachusetts, 1998 Greater Boston Western Northeast Central Southeast Total Statewide Total Jobs in Region $44,432 $28,958 $35,350 $32,579 $30,178 $37,787 Total Jobs in High-tech Industries $67,590 $49,006 $58,973 $56,560 $48,989 $62,726 Source: Massachusetts Division of Employment and Training (1999) There is wide agreement that the high-tech industry has been a pivotal engine for economic growth in the U. S. for the last decade. Over the past 20 years, high- tech industries have almost doubled their share of industry output in the United States to near 11 percent. It is not surprising that during the 1990-91 economic recession, growth in the high-tech sector was four times as fast as in the aggregate economy. DeVol (1999) suggests that America's unprecedented economic growth is spurred by high-tech industries; therefore, regions that fail to attract these industries could be left behind in the 21st century. 2.3 Landscape and Location Factors of High-tech Centers in the United States High-tech centers play an important role in the production and diffusion of innovations. Innovation is a concept that includes both the development and application of a new product, process, or service. It supposes novelty in device, application, or both. “Innovation encompasses many activities: scientific, technical, and market research; product, process, or service development; and manufacturing and marketing in order to support diffusion and application of invention.” (Office of 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Technology Assessment, 1995). High-tech centers are a place where these ideas and activities take place by virtue of their abundant educated and technical workers. High-tech centers, benefiting from agglomeration economies, produce innovations and spread them spatially and functionally. High-tech centers are created and developed through concentration of high-tech industries into a specific region or area, where is usually a metropolitan area. I summarize the contemporary major high-tech centers in Table 2.2, which are frequently cited and studied by researchers and planners. They are all located within metropolitan areas and have developed large cities and business districts. Table 2.2 Contemporary Major High-Tech Centers in the U.S. Name Location Name “ ‘ j Location Bionic Valley Salt Lake City, Utah Silicon Bayou Lafayette, Louisiana Electronics Belt Orlando-Tampa, Florida Silicon Beach Fort Lauderdale, Florida Golden Triangle South-eastern New Hampshire (north of Boston) Silicon Desert Phoenix, Arizona *Los Angeles Los Angeles County, California Silicon Forest Portland, Oregon ♦Orange County Orange County, California Silicon Hollow Oak Ridge, Tennessee Research Triangle Raleigh-Durham-Chapel Hill, North Carolina Silicon Mountain Colorado Springs, Colorado Route 128 Boston, Massachusetts Silicon Prairie Dallas-Fort Worth, Texas Satellite Alley Montgomery County, Maryland (north-west of Washington, DC) Silicon Valley Santa Clara County, California ♦Seattle Seattle, Washington Silicon Valley East Albany-Schenectady- Troy, New York Silicon Barrio Miami, Florida Tech Island Long Island (East of New York City), New York * No specific nickname Sources: Roger and Larsen (1984); Malecki (1991); Rosegrant and Lampe (1992); Lorek et al. (2000) 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.3.1 High-tech Centers: Metropolitan Agglomeration of Educated Workers High-tech industries have a strong tendency to concentrate in metropolitan areas, mainly because of their demand for educated technical workers. Successful high- tech centers are located within a small number of metropolitan areas or counties (Preer, 1992; DeVol, 1999). Using 1997 high-tech employment data (County Business Patterns, US Census), I show high-tech industries’ spatial patterns in the U.S. (see Figure 2.1). This figure demonstrates that the strong high-tech centers are distributed along with metropolitan areas, and each of these centers possesses more than 10,000 high-tech employments. Figure 2.2 supports the trend by displaying the top 100 largest high-tech counties in 1997 (see also Appendix 1). All of these counties are within metropolitan areas. It is apparent that high-tech centers have a strong preference for metropolitan areas in their location patterns at least in 1990s. 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.1 High-Tech Employment by the U.S. Counties, 1997 High-tech em ployment, 1997 n - 1 0 - 9 9 100 - 999 Hi 1 00 0 -9 9 9 9 1 0000-99999 H i 100000- 737652 Source: County Business Patterns (1997); Author Calculation P Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.2 Top 100 High-tech Counties by High-tech Employment in the U.S., 1997 ¥ s* 200 0 200 4 0 0 Mries □ P rim ary M SAs & M etropolitan SA s H igh-tech C ounties Top 100 H igh-tech C en ter 1 S ta te B oundary Source: County Business Patterns (1997); Author Calculation 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Using three or four digit SIC (Standard Industry Classification) codes5 , I demonstrate the degree of high-tech employment’s metropolitan agglomeration in a spatial analysis of high-tech employment patterns in the United States (1997) Table 2.3 Spatial Patterns of All and High-tech Industries’ Employment, 1997 CODE Definition # of counties All Industries High-tech industries High-tech m anufacturing High-tech service Metropolitan Counties * 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 57291889 (54.47%) 7668988 (78.74%) 2035786 (83.01%) 5533202 (77.30%) 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 32031261 (30.45%) 1881172 (19.31%) 387355 (15.79%) 1493817 (20.50%) Non-Metropolitan Counties I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 1283665 (1.22%) 15187 (0.16%) 1976 (0.08%) 13211 (0.18%) 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 670526 (0.64%) 12291 (0.13%) 1360 (0.06%) 10931 (0.15%) II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 3330377 (3.17%) 55665 (0.57%) 13453 (0.55%) 42212 (0.58%) 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 3267059 (3.11%) 21799 (0.22%) 5272 (0.21%) 16527 (0.23%) III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 3735036 (3.55%) 64248 (0.66%) 3121 (0.13%) 61127 (0.84%) 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 2869223 (2.73%) 19370 (0.20%) 4168 (0.17%) 15202 (0.21%) 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 704821 (0.67%) 1295 (0.01%) 0 (0%) 1295 (0.02%) Total 3140 (100%) 105183857 (100%) 9740015 (100%) 2452491 (100%) 7287524 (100%) Source: County Business Patterns, United States Census Bureau (1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation. In the Table 2.3,1 find that high-tech employment agglomerates in metropolitan area more than other industries’ employment, because 84.92 percent of employment of all industries concentrate in metropolitan area, while that of high- 5 SIC 281,283,2860,2911,357,362,366, 367,372, 376,381,382 for high-tech manufacturing; and SIC 481, 737, 7391, 7397, 781, 871, 873, 8922 for high-tech service; total twenty SIC codes are used. 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. tech industries is 98.05 percent. High-tech industries, especially, locate within large metropolitan areas more than small metropolitan areas. In all industries, 54.47 percent of employment is in large metropolitan areas, but 78.74 percent of high-tech employment locates in large metropolitan areas. In both high-tech manufacturing and high-tech service, employment trends with respect to metropolitan location are similar (high-tech manufacturing: 98.80 percent, high-tech service: 97.80 percent). High-tech manufacturing, however, has a stronger employment pattern of large metropolitan location than high-tech services (high-tech manufacturing: 83.01 percent; high-tech service: 77.30 percent located within large metropolitan areas). Compared to all industries, with 15.08 percent of jobs located in non-metropolitan areas, high-tech industries are not likely to locate in non-metropolitan areas; the employment share in the non-metropolitan areas is only 1.95 percent. Research by the Massachusetts Division of Employment and Training (MDET) (1999) displays a more detailed spatial aspect of high-tech centers. From the dispersion of high-tech industries across the state for each of the five regions within Massachusetts, MDET (1999) suggests that high-tech jobs have a stronger tendency to locate within large metropolitan areas than other industries, which suggests that high-tech employment agglomerates within large metropolitan areas more than other industries (see Table 2.4). 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.4 High-tech Industries’ Dispersion Jobs in Massachusetts, 1998 Greater Boston Western Northeast Central Southeast Total Statewide Total Jobs in Region 46% 11% 13% 10% 19% 100% Total Jobs in High- tech Industries 59% 4% 20% 7% 9% 100% Source: Massachusetts Division of Employment and Training (1999) High-tech’s higher propensity for the large metropolitan areas reflects on high-tech real output. Though there are some differences in the criteria of high-tech selection between, Table 2.5 and Table 2.6 show that high-tech employment scale is closely associated with high-tech output scale. Most high-tech metropolitan areas contain high-tech counties that have higher number of high-tech employment. Table 2.5 Top 15 High-tech Counties, by Employment Size, Percentage of National High-tech employment, 1997 Ranking County Em ploym ent 1 California Los Angeles 737652 7.57% 2 California Santa Clara 465282 4.78% 3 M assachusetts Middlesex 295365 3.03% 4 Illinois Cook 251848 2.59% 5 California Orange 241698 2.48% 6 Texas Dallas 236298 2.43% 7 New York New York 209866 2.15% 8 Virginia Fairfax 201850 2.07% 9 California San Diego 191547 1.97% 10 Texas Harris 190813 1.96% 11 Arizona Maricopa 148354 1.52% 12 Washington King 138143 1.42% 13 Minnesota Hennepin 135053 1.39% 14 Michigan Oakland 114738 1.18% 15 California Alameda 111882 1.15% Source: County Business Patterns (1997); Author Calculation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.6 Top 15 High-tech Metros, by Real Output Size, Percentage of National High-tech Real Output, 1998 Ranking State Metro Percent 1 California San Jose 5.79% 2 California Los Angeles-Long Beach 5.11% 3 New York New York 4.23% 4 M assachusetts Boston 4.18% 5 Illinois Chicago 3.76% 6 Texas Dallas 3.67% 7 DC-MD-VA-WV Washington 3.50% 8 Georgia Atlanta 2.53% 9 Washington Seattle-Bellevue-Everett 2.52% 10 Pennsylvania Philadelphia 2.09% 11 California Orange County 1.85% 12 Texas Houston 1.84% 13 Arizona Phoenix-Mesa 1.78% 14 California Oakland 1.55% 15 New Jersey Middlesex-Somerset-Hunterdon 1.48% Source: Devol (1999) 2.3.2 Significance of Educated Workers in High-Tech Centers In traditional industrial areas, their raw materials are ports, coal, iron, railroads, and highways, which attracted firms to locate there. In contemporary high-tech centers, their raw materials are not traditional ones but concentrations of educated workers. High levels of educated workers characterize high-tech centers such as Austin (TX), Irvine (CA), and Raleigh (NC) (Kotkin, 2000: 9). Kotkin (2000) says this trend of high-tech firm location as follows: Companies prefer these locations for a host of reasons, including the relative lack of distractions, low crime, and often lower taxes, but again, the most critical reason, according to numerous studies, is with the availability of and attraction for needed employees. The movement is particularly marked in large corporate headquarters operations and high-technology industries. (Kotkin, 2000: 9, 37) 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kotkin and Siegel (2000) claim that a new paradigm in economic development has been shifting emphasis from the traditional urban center’s ports, railroads, and large pools of manual labor to those places where concentrations of educated workers can be lured and harnessed. Then, why does the change of paradigm take place? The progress of technology makes this shift of paradigm possible. Increasing electronic transactions, globally inclining business, and telecommunications allow work distributed anywhere, firm location more elastic (Kotkin, 2000:7). The growth of a region is increasingly dependent upon decisions of individual entrepreneurs or educated workers who locate there. These “placeless” or “footloose” individuals include engineers, system analysts, scientists, and creative workers. These workers are called “very sophisticated consumers of place” (Knight, 1989), who make various regions compete for their amenities and attention (Kotkin, 2000: 7). Kotkin and Siegel (2000) use the term “nerdistans” to refer communities built around universities and highly evolved clusters of science-based industries. These communities are new urban regions built by their ability to attract the rising technological elite. Though the nerdistans often lack the social diversity and cultural richness in relation with urban areas, they re-create the suburban dream, “but in a more conscious, less egalitarian manner” (Kotkin, 2000: 40). Joel Kotkin (2000) says, “Rather than provide a home and garden to the average worker, the nerdistans seeks primarily to lure only the better-educated, more affluent workers critical to the digital economy” (Kotkin, 2000: 40). Therefore, communities wanting to be 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. prosperous in this digital age must find ways to attract enough entrepreneurial, technical, and creative talent, who are mostly better-educated, more affluent workers. 2.3.3 The Creative Class: The Reality of Educated Workers in High-Tech Centers The scale of educated workers by itself can not denote a strongly innovative high- tech center. An innovative high-tech center is made by a specific kind of workers, who can be a source of creativity. Florida (2002) claims that this group of workers shapes a new social class, “the Creative Class”, which gains its identity from its members’ roles as “purveyors of creativity”. This class includes people in science and engineering, architecture and design, education, arts, music and entertainment; their role is to create new ideas, build up new technology, and spread new creative subject (Florida, 2002: 8). According to Florida, “about 38 million people in the Untied States belong to the Creative Class, which accounts for 30 percent of all employed people”. The Creative Class has distinctive work habits, lifestyles, and communities. They tend to set their own working hours, dressed in casual clothes, and worked in stimulating conditions. Creative-class workers could never be impelled to work, however, “they were never truly not at work”; meanwhile, the Creative Class is most likely to work the longest hours according to figures from the Bureau of Labor Statistics (Florida, 2002: 13, 146). They want intense experiences in the real world, and seek opportunities to practice their creativity. The creative-class workers do not 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. stay tied to companies anymore; therefore, they move from job to job to pursue interesting projects and activities though not all creative people are self-employed or job-hopping free agents (Florida, 2002: 104,135). To capture this emerging class, monetary incentives, orders or negative sanctions are not as effective as for blue- collar workers. To the Creative Class, “money is an important but insufficient motivator”, because they favor challenge, flexibility and stability over pay (Information Week, 2001, Florida, 2002: 89-92). In front of the surge of creative- class workers, companies are adjusting to this change by making efforts to provide a more amenable working environment to the Creative Class. The new lifestyle of the creative-class people is based upon individuality, self-support, acceptance of differences, and a desire for diverse experiences. Along with job mobility trends, creative-class workers are highly mobile, not fixed to any particular place, residence, and community. They choose cities that provide tolerant environments, diverse population, multi-cultural experiences, and active outdoor recreations as well as good jobs. Good wages are a necessary but not a sufficient condition to attract this class because the Creative Class is not simply following traditional economic theory, where workers settle in cities that offer them the highest-paying jobs in their fields (Eakin, 2002). Creative-class people want places that satisfy their particular lifestyles and activities. Florida (2002) reflects this trend in his Creativity Index: this index includes the Gay Index, as a sign of tolerant environment, the Bohemian Index, a percentage of artistically creative people, and the Melting Pot Index, the relative percentage of foreign bom people in a region. 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Creative-class people of all background are migrating to the same kinds of cities, not only by economic opportunities but a desire for a clearly different way of life. Florida (2002) says that the Creative Class is strongly oriented to large cities and regions mainly due to a variety of economic opportunities, a stimulating environment and amenities for every possible lifestyle. Florida’s quantitative analysis shows that the statistical correlation between Creative Class locations to rates of high-tech industry is positive and statistically significant. Creative-class location rationale can account for the metropolitan advantage of high-tech centers. Five points summarize Florida’s results from the following Table 2.7. • The San Francisco Bay Area is the nation’s undisputed leadership in creativity. • Other winners include Boston, New York and Washington D.C., as well as younger high-tech regions such as Austin, Seattle, San Diego and Raleigh-Durham. • Large places have an obvious advantage in spawning and capturing creativity because large regions can provide plentiful choices. • Still, large places do not have an exclusive hold on creativity. • Some cities that are parts of broader such as Ann Arbor and Boulder are significant Creative Centers in their own right. These cities are usually home to major research universities. (Florida, 2002: 245-246) 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.7 Top 10 High-Tech Centers and Creative Class Centers (1999) A. Large Regions with over 1 million in population B. All Regions for which complete data are available Ranking High-Tech Jflcfex j 'f Creatiwtytndex Ranking H l K i l i i S t e x Creativity Index* 1 San Francisco San Francisco 1 San Francisco San Francisco 2 Boston Austin 2 Boston Austin 3 Seattle Boston 3 Seattle Boston 4 Los Angeles San Diego 4 Los Angeles San Diego 5 Washiongton, D.C. Seattle 5 Washiongton, D.C. Seattle 6 Dallas Raleigh-Durham 6 Dallas Raleigh-Durham 7 Atlanta Houston 7 Atlanta Houston 8 Phoenix Washington, D.C. 8 Phoenix Albuquerque g Chicago New York 9 Albuquerque Washington, D.C. 10 Portland Minneapolis 10 Chicago New York Source: Florida (2002: 251) According to the indices that Florida (2002) has adopted and developed, I gain some insights about metropolitan location trend of high-tech centers. High-tech centers are positively associated with immigration statistically, even though places that are open to immigration do not inevitably count among the major Creative Class Centers. Immigrants and minority groups usually settle in large cities or metropolitan areas, which provide more job-opportunities than other smaller or rural areas. High-tech centers are also positively correlated with other types of diversity such as the Gay Index, which is a measure of coupled gay people in a region relative to the United States as a whole. A community that accepts gay neighborhoods welcomes all kinds of people, because homosexuality represents the last frontier of diversity in this society (Bishop 2000). For this rationale, openness to the gay community is a good clue of low entry barriers to those who are very important to inspiring creativity and generating high-tech growth (Florida, 2002: 256). 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Metropolitan areas are more open to various ideas and diverse values than any other regions; in addition large cities provides anonymous communities, where people can hide aspects of identity to the neighborhood. The third relationship is found between high-tech places and high levels of amenities. Big cities, which have been centers of culture and fashion, are also attracting talented people and generating new technology-intensive industries (The Economist, 2000). Florida, in particular, emphasizes “more active, informal, street-level irregular amenities” rather than big- ticket events and attractions (Florida, 2002: 259-260). Austin, Texas, is his example of this relationship, which is a highly ranked location for high-tech firms although this city does not have major league sports and few world-class cultural organizations. Lastly, Florida contends that there is a strong relationship between high-tech centers and amount of artistically creative people like writers, designers, musicians, composers, actors, directors, painters, sculptors, artist printmakers, photographers, dancers, artists, and performers. Artistically creative people, who tend to stay in metropolitan areas, lead a region’s cultural amenities that are attractive to creative high-tech workers. There are some critics of Florida’s creative class. Samuelson (2002) criticizes his definition asking a keen question; “Can creativity define a class?” He says that the phrase “creative class” is not interesting, its meaning is not obvious, and it may not be a class at all (Samuelson, 2002). Eakin (2002) raises questions about Florida’s research outcomes. She emphasizes the investment to higher education rather than attraction for the Creative Class. Her criticism, moreover, deals with 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. weaknesses of Florida’s theory that fail to explain the extraordinary success of some non-high-tech centers such as Las Vegas, which has more gamblers and tourists than computer scientists or artists. Eakin (2002) is skeptical about tolerance that is suggested by Florida as an indicator of creative class location, a precursor to urban growth and high-tech prosperity. Though there are weaknesses, Florida’s discussion of the Creative Class is a significant contribution. First, his research augments human capital theory that claims that concentrations of educated people spur regional economic growth. He names his new theory “the creative capital theory” in which regional economic growth is driven by creative people who prefer places that are diverse, tolerant and open to new ideas (Florida, 2002: 249). His new theory opens the possibility of a significant role of a certain class in originating high-tech centers in the United States. Second, his findings add social perspectives to the present high-tech location research, which usually deal with economic aspects of high-tech workers and high- tech firms. Though many of his ideas depend on interviews and surveys to catch basis for testing creative-class people’ work habits, lifestyles, and locales, he implicitly provides a high-tech workers’ social and personal attributes beyond economic motives. 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.3.4 The Evolutionary Landscape: Historical Aspects of High-tech Centers High-tech centers are not all created equally. Most high-tech centers are heavily concentrated in one industry, which has developed from the region’s intrinsic advantages and knowledge base (Iswalt, 2001). According to Is wait (2001), for example, New York City has a large amount of high-tech employment in Silicon Alley, but high-tech industries in New York City rely on prevailing financial industries. Energy-dependent Houston has a large tech base on due to NASA, its medical device industries, and Compaq. Detroit has a large amount of high-tech economy associated with the big automakers. With the exception of Silicon Valley, which contains almost all kinds of high-tech industries; “high-tech employments in most high-tech centers are highly specialized in just a few technologies” (Cortright and Mayer, 2000a) (see Table 2.8). 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.8 Specialized Technologies of the High-tech Centers State Count* Cil\ Specialized Tcchnoloi>\ California Santa Clara San Jose Microprocessor; Personal Computer; Semiconductor; Database Software California San Diego San Diego Wireless Communication; Biotechnology California Sacramento Sacramento Computer & Semiconductor Manufacturing Colorado Denver Denver Cable & Telecommunication; Computer Storage Firms Utah Salt Lake Salt Lake City Medical Devices; Software Washington King Seattle Aerospace; Avionics; Software; Biotechnology Oregon Multnomah Portland Semiconductor; Wafers; SME/EDA; Displays; Computers Texas Travis Austin Semiconductors; (Personal) Computers Arizona Maricopa Phoenix Semiconductors Georgia Fulton Atlanta Airline Transportation Massachusetts Suffolk Boston Computers; Medical Devices; Software Minnesota Hennepin/Ramsey Minneapolis/ St. Paul Computers; Data Processing; Medical Devices North Carolina Wake/ Durham/ Orange Raleigh- Durham- Chapel Hill Microelectronics; Biomedicine & Genetic Engineering Source: Cortright and Mayer (2000a) Silicon Valley has played a unique role in developing high-tech industries and shaping other high-tech centers in the U.S. As the single largest concentration of high-tech employees, it accounts for more patents than any other areas. Its role in the U.S. is still secure as in the 1970s and 1980s. Silicon Valley has a higher level of patenting per employee than any other region; it accounts for almost 40 percent of all of the reported venture capital investment made in the U.S. in 1999 (Cortright and Mayer, 2000a: 51). Though Silicon Valley recently began to depend more on foreign firm operations, it still maintains corporate leadership function (Cortright and Mayer, 2000a, Gray et al., 1996). Many of the largest high-tech firms in Silicon Valley locate headquarters locally. Cortright and Mayer (2002a) report that at least 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. one of the ten largest high-tech firms in every region they examined counted a corporate headquarters in the city of San Jose. History plays a pivotal role in high-tech development and the evolutionary landscape of high-tech centers. According to Cortright and Mayer (2002a), the role of history is that time refines technological specialization, and enables successive generation of firms. The creation history of Silicon Valley by Hewlett Packard, Fairchild, and Intel is well known. However, the role of history is not only technological refinement, and successive generation of firms. Its substantial role is in the commercialization of research breakthrough, government policy and supports to specific regions, and industrial relocation. As industries and technologies change, firms and knowledge in specific regions and metropolitan areas evolve. Cortright and Mayer (2002a) say that “(high-tech) development processes seem to be cumulative and path dependent: a region’s opportunities for development today are determined, patially, by the developmental patterns of the past”. Shockley’s semiconductor firm started in Santa Clara County, and Howard Vollum’s oscilloscope was in Portland: they are powerful examples of the commercialization of research breakthroughs. One instance of government policy and support for a specific region is the massive defense spending on electronics that held up Route 128 firms in Boston area. There are many cases of industrial relocation and expansion. Manufacturing plants in Silicon Valley moves to surrounding regions, other states, and foreign countries. IBM moves its tape drives from Silicon Valley to Boulder (Texas), and its PC to Austin (Texas). Throughout these periods and processes, 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. many location factors and determinants determine evolutionary landscape of high- tech centers. 2.3.5 Location Factors of High-tech Centers A key area of high-tech centers research deals with the role of federal support and military spending. There are many reports that emphasize government roles in R&D, innovation, and high-tech center development. Lederman and Windus (1971) explore federally sponsored R&D in accordance with national goals and priorities (Lederman and Windus, 1971). They report that military spending and R&D expenditures maintained a higher portion of federal funding during the 1961-1971 period due to national security rationale. March (1970) focuses his research on federal budget priorities for R&D during 1950-1970 (March 1970). Murphy (1971) notes the geographical distributions of (1) federal R&D funds in general, (2) federal support of university R&D, (3) federal prime contract and subcontract awards to private firms, and (4) location of federal facilities (Murphy 1971). His data demonstrate that R&D funding starts in several regions in the west and east coasts, university support is increasing but with the absence of an overall federal policy, and policy issues relating to contracting for R&D are important to the national economy and regional growth. Nelson contributes books and papers on government support of technical progress and high-tech industry in America and other industrialized countries (Nelson, 1982,1984a, 1984b). He describes the nature of public policies that have influenced the pace and pattern of technical progress and tries to assess the 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. broad effects of these policies. Markusen’s studies specify the relationship between defense spending and the geography of high-tech industries (Markusen and Bloch 1985, Markusen 1984,1989,1991). Her research includes evidence that suggests defense spending has been a powerful force in supporting high-tech innovation and has been a major factor in the growth of economies in several American regions. Rosegrant and Lampe (1992) provide the general view that a major role of federal government and its policies affect the overall environment in which innovation takes place through university research contracts, federal laboratories, and procurement contracts in a diverse range. Another group of researchers examine the role of infrastructure such as highways, the Internet, and higher education in the formation and growth of high- tech centers and “technopolis”. Echeverry-Carroll and Brennan (1999) emphasize the role of airports by highlighting the demand for an infrastructure that permits frequent flights for skilled employees among the fast innovators (Echeverry-Carroll and Brennan, 1999). They conclude that policymakers who want to attract innovative high technology firms should focus on providing infrastructure that strengthens external networks, such as a good airports and airlines that offer frequent direct flight. Williams (1992) suggests an increasing role of information infrastructure within the technopolis - as well as between them. He emphasizes the growth of high-performance scientific networks such as computer-based networks and Internet connections between major universities and between science and research communities. Suarez-Villa notes that education infrastructure plays a 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. critical role in innovative capacity, in the analysis of age cycles for investment and patenting in the U.S. (1920-1990) (Suarez-Villa, 1997a; Suarez-Villa, 2000). It is a generally accepted idea that education improves economic growth and productivity via several paths (Dean 1984). High-tech centers research in this field usually deals with the relationships among research universities, engineering graduate students, faculty, university patents and the growth of high-tech centers. There are two competing arguments on the role of universities and their engineering graduates in the formation of high-tech centers. Some support a significant and positive role of universities in the high-tech centers because they are providing skilled laborers, R&D facilities, and new firms (Nelson, 1986; Lund, 1986; Sivitanidou and Sivitnides, 1995; Gottlieb, 2001). Others discuss a weak, inconsistent, or even no relationship between university research activities and high- tech location (Markusen, Hall, and Glasmeier, 1986; Florax and Folmer, 1992; Bania, Eberts, and Fogarty, 1993) Every high-tech location factor belongs to one of three categories: business climate, quality of life, or overlapping factor. In the first kind, local business environments are considered most significant for high-tech firm operations. These authors argue that local business amenities play a pivotal role in successful high-tech centers, which includes local patent, R&D subsidies/expenditures, available venture capital, local industry size, local market size, transportation linkages, office rents, zoning, and foreign investments (Malecki, 1985; Timmons and Fast, 1987; Florida and Kenney, 1988; Acs and Audretsch, 1988; Scott and Drayse, 1993; Goss and 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Vozikis, 1994; Haug, 1995; Sivitanidou and Sivitanides, 1995; Lyons, 1995; Auselin, Varga, and Acs, 1996; Audretsch and Stephan, 1996; Irwin and Klenow, 1996; Suarez-Villa, 1997b; Gray and et al., 1998; Zucker and et al., 1998; Sivitanidou, 1999; Varga, 1999; Alarcon, 1999; Cordes, Hertzfeld, and Vonortas, 1999). In the second kind, local or metropolitan quality of life is suggested as the most important factor explaining successful high-tech centers because quality of life attracts and holds “more footloose” technical and professional workers. Included factors are climate, cost of living, traffic congestion, crime, pollution, recreation, public education, public services, health care providers, and poverty (Malecki, 1985; Malecki and Bradbury, 1992; Herzog and et al., 1986; Haug, 1991; Sivitanidou and Sivitanides, 1995; Gottlieb, 1995). The third group highlights combinations of factors such as city size, population, and taxation, which influence both residential and business amenities (Goss and Vozikis, 1994; Suarez-Villa, 1997b; Sivitanidou, 1999). Table 2.9 shows the three categories of high-tech location factors mentioned in this chapter. 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.9 General Categories of High-Tech Location Factors General Category LocationFactor Business Climate Government policy and support (R&D etc.) University and higher education Transportation The Internet infrastructure Local financial condition local industry size Office rent, zoning Foreign investment quality of life climate Cost of living crime Pollution Recreation Public education public service (welfare) Health services Poverty (income etc.) housing Overlapping factors Population City size Taxation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4 High-Tech Center Indices Many regional indices have been developed to measure regional economic situations, dynamics, or quality of life. The National Real Estate Index (by CB Richard Ellis) is a perennial report for business and residential conditions around all metropolitan areas. A regional government association in Oregon publishes a “Qualify of Life Index” to examine the quality of life conditions for the Southern Oregon region (Rogue Valley Civic League and Southern Oregon Regional Services, 2000). Recently, indices have been developed for measuring “new economy” conditions or regional innovation and technology position as more people are interested in local level high-tech potential and adaptability to the “new economy” (Kansas Technology Enterprise Corporation, 1999; Massachusetts Technology Collaborative, 2000; Washington State, 2000; National Science Foundation, 2001; Office of Technology Policy, 2002; Atkinson, 2002). The indices suggested in these papers include innovation capacity, knowledge jobs, competitiveness, economic dynamism, the digital economy, human potential, and quality of life factors. These indices have a common weakness of a biased measure: many indices favor metropolitan areas because they are correlated with population, employment, and city size. Therefore, the Milken Institute Tech-Poles, a high-tech index recently developed by the Milken Institute is valuable, in which “the composite index is equivalent to the percent of national high-tech real output multiplied by the high-tech real output location quotient for each metro” (Devol, 1999: 67). This index is a less biased measure because the former number favors large metropolitan areas; while the 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. latter number favors smaller regions with large high-tech sectors (Devol, 1999; Florida, 2002). In this chapter, I introduce five high-tech indices: high-tech employment density, location quotient, high-tech growth rate, high-tech total wage income, and high-tech personal wage income. Four of them (except “total wage income”) are relative values and less biased compared to high-tech employment size or high-tech total wage income. Among the suggested indices, high-tech employment density, high-tech employment location quotient and high-tech personal wage income are to be used to define high-tech centers in this research. 2.4.1 High-tech Employment Density High-tech employment size has a weakness in representing a high-tech center because a large numbers of high-tech employment are likely to be found in metropolitan counties with large population sizes and urbanized areas. Thus, we have to look at high-tech employment density to identify more substantial high-tech centers. Figure 2.3 and Table 2.10 shows the spatial patterns of high-tech centers by high-tech employment density. 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.3 High-Tech Employment Density by the U.S. Counties, 1997 A 300 H ig h -te ch em p lo y m e n t d e n sity . 1997 (p e rs o n /s q m ile) 300 600 Miles 1 - 1 0 ■ ■ 10-100 100 - 1000 1000-10000 Source: County Business Patterns (1997); Author Calculation 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.10 Spatial Patterns of Employment Density in High-Tech Industries, 1997 (person/square mile) CODE Definition # of counties All Industries High-tech industries High-tech m anufacturing High-tech service Metmpoi&n Counties 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 268.85 35.99 9.55 26.43 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 66.92 3.93 0.81 3.12 '' Nor-M ^rnp^itnn ^ n r itie l ■ ■ I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 17.59 0.21 0.03 0.18 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 8.98 0.16 0.02 0.15 II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 20.12 0.34 0.08 0.25 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 6.14 0.04 0.01 0.03 III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 11.45 0.20 0.01 0.19 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 3.01 0.02 0.00 0.02 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 0.91 0.00 0.00 0.00 Total 3140 (100%) 29.33 2.72 0.68 2.03 Source: County Business Patterns, United States Census Bureau (1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation. From the Figure 2.3 and Table 2.10, we find that high-tech industries concentrate in large metropolitan areas whose populations are larger than 1 million. When I compare the ratios of small metropolitan to large metropolitan (density in large metro/density in small metro), they are 4.02 in all industries, 9.16 in high-tech industries, 11.79 in high-tech manufacturing, and 8.47 in high-tech services. Table 2.11 displays 15 counties that are top 15 high-tech employment density value. I can find that all of them are within metropolitan areas. 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.11 Top 15 Counties by High-Tech Employment Density, 1997 Ranking STATE COUNTY High-tech Employment Density (1997, person/ sq mile) 1 NY New York County 9209.254 2 VA Arlington County 1910.174 3 CA San Francisco County 1389.111 4 VA Alexandria city 1146.895 5 MA Suffolk County 843.473 6 DC District of Columbia 751.997 7 VA Fairfax city 650.407 8 VA Fairfax County 508.455 9 CO Denver County 482.466 10 PA Philadelphia County 448.812 11 MO St. Louis city 384.872 12 CA Santa Clara County 362.398 13 MA Middlesex County 347.381 14 VA Falls Church city 342.011 15 CA Orange County 303.452 Source: County Business Patterns (1997); Author Calculation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4.2 Location Quotient The location quotient (L.Q.) is a standard index for defining a high-tech center. “It is a measure of the relative significance of a phenomenon in a region compared with its significance in a larger region,” (Hayter, 1997: 434) and is a way of measuring the relative contribution of one sector of an economy to the whole economy. The location quotient may provide an answer to the question: “How important are high- tech industries in California compared to the rest of the United States?” For example, the location quotient for employment can be defined as follows: L Qi = Where e< = Regional employment in industry i in year t er = Total regional employment in year t - National employment in industry i in year t Et = Total national employment in year t Klosterman (1990) explains the location quotient as follows: Industries with location quotients equal to 1.0 have a local employment share exactly equal to their national share. Regional production in these sectors is supposed to be just sufficient to meet with local demand, and the industries are to have no basic employment. Industries with location quotients less than 1.0 have local employment shares smaller than Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. their national shares, and they are insufficient to satisfy local demand, which requires products to be imported. Industries with location quotients greater than 1.0 have local employment shares bigger than their national shares. Local production is specialized in these industries relative to the nation, and their production surpasses local demand, which allows the surplus to be exported. (Klosterman, 1990: 129) Thus, local economic activity of an industry, when its location quotient is bigger than 1, includes basic sector activity regardless of non-local economic environments. That is, industries export goods and services outside the local economy and in so doing pull money into the region. The greater the location quotient exceeds 1, the larger the importance of the industry as a basic industry. I describe 100 high-tech counties by the 100 location quotients (See Figure 2.4 and Appendix 2). Figure 2.4 shows that higher location quotients in high-tech employment are discovered within metropolitan areas, though twelve counties show higher location quotients located out of metropolitan areas. Most of these twelve counties, however, locate adjacent to or nearby metropolitan areas. The only exception is Logan County (NE), a totally rural area, 150 mile far from the nearest metropolitan area. This county is ranked first by the location quotient because it has 24 high-tech service workers among regional total 37 employment. Thus, I exclude this county from the high-tech center category. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.4 High-Tech Centers by the Location Quotient, 1997 I I Primary MSAs & Metropolitan SAs High-tech Counties The L.Q. is larger than 1.25 E g g State Boundary Source: County Business Patterns (1997); Author Calculation When I analyze the location quotient patterns with high-tech employment in metropolitan areas and non-metropolitan areas, I find that high-tech employment is concentrated and specialized in the metropolitan areas. Their concentration degree is higher within the large metropolitan areas than the small ones (See Table 2.12). The location quotient of high-tech employment in large metropolitan areas is 1.446; and that in small metropolitan areas is 0.634. Table 2.9 shows that high-tech manufacturing has a stronger concentration in large metros than high-tech services 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (1.524:1.394). The non-metropolitan high-tech concentration degree is so low that I may ignore their role in non-metro areas. Table 2.12 Spatial Patterns of High-Tech Industries by the Location Quotient, 1997 CODE Definition # of counties High-tech industries High-tech manufacturing High-tech service Metropolitan Counties 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 1.446 1.524 1.394 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 0.634 0.519 0.673 Non-Metropolitan Counties I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 0.128 0.066 0.149 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 0.198 0.087 0.235 II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 0.181 0.173 0.183 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 0.072 0.069 0.073 III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 0.189 0.036 0.236 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 0.073 0.062 0.076 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 0.020 0.000 0.027 Total 3140 (100%) Source: County Business Patterns, United States Census Bureau (1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation. I construct Table 2.13 from location quotients of high-tech employment size, and quote Table 2.14 from Devol’s paper (Devol, 1999). Devol (1999) provides location quotients of high-tech centers in the United States, using total high-tech real output. In Table 2.13 and Table 2.14, higher scores of the location quotients are different from those of high-tech employment size; however some counties and metros maintain higher scores in both standards (See Table 2.13 and Table 2.14). 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Santa Clara County in the San Jose Metro (CA) and Boulder County in Boulder- Longmont Metro (CO) have higher scores in both tables; Santa Clara County has the highest scores in high-tech employment size, high-tech real output size, and the location quotients of high-tech employment and output. Table 2.13 Top 15 High-Tech Counties, By Location Quotient of High-tech Employment, 1997 Ranking State County L.Q.(1997) 1 Nebraska Logan 7.00 2 Indiana Owen 6.76 3 California Santa Clara 5.63 4 Virginia Fairfax 5.19 5 Virginia King George 5.02 6 Virginia Arlington 4.86 7 Idaho Bonneville 4.26 8 M assachusetts Middlesex 4.12 9 Maryland St. Marys 3.91 10 Colorado Boulder 3.76 11 Kansas Sedgwick 3.17 12 Maryland Montgomery 3.15 13 Tennessee Anderson 3.13 14 Alabama Madison 3.05 15 Texas Travis 3.03 Source: County Business Patterns (1997); Author Calculation Table 2.14 Top 15 High-Tech Metros, By Location Quotient of High-tech Output, 1998 Ranking State Metro L.Q. (1998) 1 Minnesota Rochester 5.56 2 California San Jose 4.09 3 New Mexico Albuguergue 3.55 4 Texas Lubbock 3.08 5 Iowa Cedar Rapids 3.07 6 Colorado Boulder-Longmont 2.89 7 Idaho Boise City 2.68 8 Michigan Kalamazoo-Battle Creek 2.66 9 Washington Richland-Kennewick-Pasco 2.41 10 New Jersey Middlesex-Somerset-Hunterdon 2.30 11 Washington Seattle-Bellevue-Everett 2.06 12 Florida Melbourne-Titusville-Palm Bay 2.00 13 North Carolina Raleigh-Durham-Chapel Hill 2.00 14 Idaho Pocatello 1.99 15 Georgia Albany 1.97 Source: Devol (1999) 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I see that high-tech industries concentrate into several particular Metropolitan Areas such as San Jose (CA), Boston (MA), Dallas (TX), and Los Angeles-Long Beach (CA) from high-tech employment (or output) size and its location quotients. The dominance of Silicon Valley (San Jose, CA) is clearly displayed. Devol produces a composite measure of technology production centers, the Tech-Pole from the combination of the location quotient with the share of national high-tech output in a multiplicative fashion (Devol, 1999). He finds that the composite index of Silicon Valley (San Jose, CA; index: 23.7) is more than three times the size of the second ranked metro (Dallas, TX; index: 7.06), which is hardly larger than the third ranked metro (Los Angeles-Long Beach, CA; index: 6.91), and the fourth ranked metro (Boston, MA; index: 6.31) (See Table 2.15). Table 2.15 Top 10 Milken Institute Tech-Poles, Composite Index, 1998 Ranking Tech-Poles Composite Index* 1 San Jose, CA 23.69 2 Dallas, TX 7.06 3 Los Angeles-Long Beach, CA 6.91 4 Boston, MA 6.31 5 Seattle-Bellevue-Everett, WA 5.19 6 Washington, DC-MD-VA-WV 5.08 7 Albuquerque, NM 4.98 8 Chicago, IL 3.75 9 New York, NY 3.67 10 Atlanta, GA 3.46 * Composite Index is equivalent to the percent o f national high-tech real output multiplied by the high-tech real output location quotient for each metro Source: Devol (1999) 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4.3 Growth Rate Growth patterns of high-tech centers tell a different story from high-tech employment size or high-tech location quotients. Premus (1982) says that total high- tech employment growth rate has fluctuated 17.6% (1955-1965), 7.2% (1965-1975), and 19.5% (1975-1979), using four 2-digit SIC codes. These percentages of high- tech employment are bigger than those of other manufacturing industries and those of total manufacturing industries in the same periods (See Table 2.16). Table 2.16 Growth Rates of Total High-Tech and Other Manufacturing Employment 1955-1965 1965-1975 1975-1979 High-tech Employment 17.6% 7.2% 19.5% Other Manufacturing 2.6% -1.8% 11.3% Total Manufacturing 7.0% 1.5% 14.5% Source: Premus (1982), pp 5 Hecker (1999) says that high-tech employment increased 13 percent in comparison with 20 percent increase in total employment for the economy as a whole (Hecker, 1999). My calculation, using twenty 3-digit or 4-digit SIC codes (12 codes for high-tech manufacturing, 8 codes for high-tech services) shows a different result from Hecker’s (See Table 2.17). It is true that during the 1987-1997 period, the employment of high-tech manufacturing experiences an -8.82 percent decrease. However, strong increase in high-tech services compensates for this, and maintains 124.02 percent total high-tech employment growth. It seems that cuts in defense 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. spending have greater impact on high-tech manufacturing employment than on high- tech service. Another finding is that high-tech manufacturing employment decreases in the large metropolitan areas, but still increases in the small metropolitan areas. High-tech services employment undergoes a similar experience. Its growth rate in the small metropolitan areas is almost 1.7 times bigger than that of the large metropolitan areas. This trend of high-tech employment’s metropolitan growth pattern is, especially, strong during the 1987-1992 period, and becomes weak during the 1992-1997 period (See Table 2.18 and Table 2.19). The surprising growth of high-tech employment in the non-metropolitan areas is not significant because the absolute numbers are small. 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.17 Spatial Patterns of High-Tech Employment Growth, 1987-1997 CODE Definition # of counties All Industries High-tech industries High-tech m anufacturing High-tech service Metropolitan Counties 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 19.33% 100.96% -16.00% 296.50% 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 28.00% 286.94% 64.44% 496.10% ", N o n-M et Wpdlitltt: CdUfitJdf- I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 31.18% 1591.20% 3366.67% 1470.87% 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 33.07% 926.82% 457.38% 1047.01% II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 24.93% 228.10% 23.38% 596.34% 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 29.47% 1028.90% 569.04% 1345.93% III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 32.37% 1059.08% 131.70% 1356.79% 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 28.78% 1.91% -75.74% 733.90% 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 27.18% 1650.00% N/A 1650.00% Total 3140 (100%) 23.30% 124.02% -8.82% 338.69% Source: County Business Patterns, United States Census Bureau (1987,1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation. 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.18 Spatial Patterns of High-Tech Employment Growth, 1987-1992 CODE Definition # of counties All Industries High-tech industries High-tech m anufacturing High-tech service M etropoltari'C cisiitB tes". > ’ 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 6.17% 66.37% -13.11% 204.05% 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 11.50% 195.57% 58.60% 324.32% Non-Metroodlitanddontieis^" I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 13.55% 924.50% 2077.19% 846.37% 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 10.33% 429.32% -100.00% 564.85% II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 9.75% 131.57% -29.83% 421.89% 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 12.35% 461.94% 25.38% 762.90% III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 14.95% 601.61% 301.56% 697.93% 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 12.49% -35.90% -81.31% 392.10% 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 12.44% 955.41% N/A 955.41% Total 3140 8.64% 81.77% -7.13% 225.39% Source: County Business Patterns, United States Census Bureau (1987,1992); Urban Influence Code, United States Department of Agriculture (1997); 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.19 Spatial Patterns of High-Tech Employment Growth, 1992-1997 CODE Definition # of counties All industries High-tech industries High-tech m anufacturing High-tech service Metropolitan Counties 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 12.40% 20.79% -3.32% 30.41% 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 14.80% 30.91% 3.68% 40.48% Non-Metropolitan Counties I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 15.53% 65.08% 59.23% 65.99% 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 20.61% 93.99% N/A 72.52% II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 13.83% 41.68% 75.83% 33.43% 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 15.24% 100.89% 433.60% 67.57% III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 15.15% 65.20% -42.30% 82.57% 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 14.48% 58.99% 29.76% 69.46% 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 13.11% 65.81% N/A 65.81% Total 3140 13.50% 23.24% -1.82% 34.82% Source: County Business Patterns, United States Census Bureau (1992,1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation The growth rates of high-tech employment in each county are large during the 1987-1997 periods, because many counties had a small number of high-tech employment even in the 1987. Five hundred thirty five counties among total three thousand and one hundred forty U.S. counties are reported to increase more than 100 percent in high-tech employment during the periods. Four hundred forty eight counties experience more than 200 percent high-tech employment growth in the same periods. In one hundred sixty counties, the growth rate is more than 1000 percent from 1987 to 1997. However, there are only eighty two counties selected whose growth rate during the periods is more than 200 percent because I restrict the 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. high-tech employment of the Counties in 1987 is more than two thousand (See Figure 2.5 and Appendix 3). Figure 2.5 High-Tech Centers by the Growth Rate, 1987-1997 t: 200 400 Mi es I I Primary MSAs & Metropolitan SAs High-tech Counties m The Growth Rate is larger than 200% n State Boundary Source: County Business Patterns (1987,1997); Author Calculation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4.4 Total Wage Income High-tech wage income data are useful because they represent indirectly the trend of output. Figure 2.6 and Table 2.20 are provided to show spatial patterns of total wage income in high-tech industries. Figure 2.6 High-Tech Annual Total Wage Income by the U.S. Counties, 1997 (dollar) 300 0 300 600 Miles High-tech annual total wage income, 1997 ( $ ) r n 0-100000 L Z J 100000-1000000 1000000 - 10000000 10000000-33871307 Source: County Business Patterns (1997); Author Calculation 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.20 Spatial Patterns of Annual Total Wage Income in High-Tech Industries, 1997 ($ 1,000) CODE Definition # of counties All Industries High-tech industries High-tech m anufacturi ng High-tech service Metropolitan Counties 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 1866716790 (61.34%) 389658224 (81.79%) 101062258 (85.29%) 288595986 (80.64%) 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 832689321 (27.36%) 79612714 (16.71%) 16287548 (13.75%) 63325176 (17.69%) !'" : ■ ■ ■ ■ ■ ■ ■ I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 30211994 (0.99%) 492359 (0.10%) 66831 (0.06%) 425528 (0.12%) 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 1403855 (0.46%) 611256 (0.13%) 63996 (0.05%) 547257 (0.15%) II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 76774028 (2.52%) 2029339 (0.43%) 492983 (0.42%) 1536356 (0.43%) 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 68755701 (2.26%) 760907 (0.16%) 231491 (0.20%) 529416 (0.15%) III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 81762680 (2.69%) 2456267 (0.52%) 82575 (0.07%) 2373692 (0.66%) 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 58689736 (1.93%) 725459 (0.15%) 203491 (0.17%) 521968 (0.15%) 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 13535888 (0.44%) 39310 (0.01%) 0 (0.00%) 39310 (0.01%) Total 3140 (100%) 3043170993 (100%) 476385862 (100%) 118491173 (100%) 357894689 (100%) Source: County Business Patterns, United States Census Bureau (1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation. Table 2.20 and Figure 2.6 demonstrate that high-tech industries are producing output in large metropolitan areas (81.79 percent of total wage income) compared to total industrial sectors’ output (61.34 percent of total wage income). I display the top 15 total wage income counties in Table 2.21, which is very similar to Table 2.5 top 15 high-tech employment size counties. 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.21 Top 15 High-tech Counties, by High-tech Total Wage Income, 1997 Ranking ' C T A T C / ( ' c tx $ rrY High-tech Wage I n c o r ^ •' 1 CA Los Angeles County 33871307 2 CA Santa Clara County 33646872 3 MA Middlesex County 16238883 4 TX Dallas County 12199317 5 IL Cook County 11890510 6 CA Orange County 11729011 7 VA Fairfax County 11726083 8 NY New York County 11484599 9 TX Harris County 9957224 10 CA San Diego County 9159817 11 WA King County 8036988 12 AZ Maricopa County 6637669 13 CA Alameda County 6637589 14 Ml Oakland County 6284281 15 MN Hennepin County 6216636 Source: County Business Patterns (1997); Author Calculation 2.4.5 Personal Wage Income Personal wage income data will provide a more substantial index for high-tech industries, and their output than total wage income because personal wage income offsets a bias of employment size. Figure 2.7, Table 2.22, and Table 2.23 show that large metropolitan areas provide the highest personal wage incomes to high-tech employees. Higher personal wage incomes over $ 50,000 are found in large metropolitan areas. 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.7 High-Tech Annual Personal Wage Income by the U.S. Counties, 1997 (dollar) I " h-tech annual personal wage income, 1997 ( $ ) 0 - 20000 20000 - 50000 50000 - 200000 Source: County Business Patterns (1997); Author Calculation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.22 Spatial Patterns of Annual Personal Wage Income in High-Tech Industries, 1997 ($) CODE Definition # of counties All Industries High-tech industries High-tech m anufacturing High-tech service Metropolitan Counties 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 32582.6 50809.2 49641.5 51231.3 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 25996.1 42322.3 42054.6 42391.7 Non-M etrorioiitah S iin W l'" ^:, I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 23535.7 32419.8 33821.4 32210.1 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 20903.1 49731.8 47055.9 50064.7 II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 23052.7 36416.4 36479.4 36396.2 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 21054.0 35973.3 43909.5 33338.5 III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 21890.7 37801.5 26457.9 38373.9 8 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 20451.4 37547.7 49499.1 34317.4 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 19204.9 32274.2 0.0 32274.2 A verage 3140 (100%) 28932.2 48910.1 48314.6 49110.5 Source: County Business Patterns, Census Bureau (1997); Urban Influence Code, United States Department of Agriculture (1997); Author Calculation. Table 2. 5 High-Tech Counties, by Personal Wage Income, 1997 ($) Ranking COUNTY “ Personal W age Income 1997: 1 GA Upson County 198357 2 NM Roosevelt County 159000 3 TN Dickson County 115000 4 TX Waller County 76103 5 MO McDonald County 72389 6 CA Santa Clara County 72315 7 KY Boyd County 72120 8 TX Rockwall County 71400 9 LA Iberville Parish 71191 10 CA San Mateo County 70217 11 LA Ascension Parish 68756 12 TX Parker County 68069 13 NJ Hunterdon County 67542 14 NH Grafton County 66958 15 KS Butler County 65516 16 CA Marin County 65168 17 VA Halifax County 65000 18 NJ Salem County 64018 19 NJ Somerset County 63903 20 GA Forsyth County 63591 Source: County Business Patterns (1997); Author Calculation 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 3 THEORETICAL FRAMEWORKS AND RESEARCH HYPOTHESES 3.1 Chapter Summary This chapter introduces the theoretical frameworks for this study and its research hypotheses. The theoretical frameworks consist of two parts: regional growth theories and quasi-experimentation. In the regional growth theory part, I explain the high-tech phenomenon for the perspective of growth pole theory, neoclassical economic location theory, “quality of life” theory, creative capital theory, and time sensitive theory of technology. In the quasi-experimentation part, I describe the general concept of quasi-experimental design, its merits and weaknesses when applied to this study. Four hypotheses will be submitted here based on theoretical frameworks and other literature. These hypotheses will be tested to produce results and policy implications. 3.2 Regional Growth Theories and High-tech Centers Some industrial and regional growth theories are applicable to shape this research and research hypotheses. First, growth pole theory can explain the basic characteristics of high-tech centers. Perroux introduced the concept of growth poles, and claimed that innovative firms that were clustered in propulsive industries 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. generate economic growth (Perroux, 1955). These poles spread growth to surrounding economic space. Perroux noted that, “Growth does not appear everywhere and all at once; it reveals itself in certain points or poles with different degrees of intensity” (Perroux, 1961). The growth pole concept was built on Schumpeter’s emphasis on innovation and entrepreneurship. Hirschman (1958) and Mydal (1957), with similar ideas, suggested “trickling down” and “backwash effects”, respectively. Using spread and backwash effects, Richardson (1979) suggested a logistic function to explain the technology diffusion process from core to periphery areas. High-tech centers are built up from stable propulsive industries and their spatial and functional diffusions. Second, the neoclassical economic approach provides a background for business-amenity rationales, of which “a common manifestation is the business climate idea, whereby a set of qualities attractive to business are defined to rank regions by how well they supply these attributes” (Preer, 1992). Traditional business conditions for business agglomeration include access to raw materials, labor, and markets along with low transportation costs. Whereas, an attractive business climate for high-tech firms includes local patent policies, R&D subsidies/expenditures, available venture capital, local industry size, local market size, airport/airline connection, computer networks, office rents, and foreign investments, these are somewhat different from traditional good climates cited for business operations and profits. One important contribution of neoclassical location theory is its early identification of agglomeration economies. Though the decline of traditional 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. manufacturing centers casts some doubt on the validity of agglomeration economies, the rise of high-tech industrial clusters provides a new perspective on agglomeration economies. High-tech centers display many traits of traditional industrial agglomerations in which the interactions among firms involve information flows as well as backward and forward linkages. Different from more traditional firms, high- tech firms gain the essential advantages from large pools of technical and professional workers that arise around high-tech centers (Premus, 1982). There is literature on the relationship between quality of life (quality of place) and high-tech firm location. Quality of life plays an important role in attracting and keeping the labor force demanded by high-tech firms, although quality of life factors themselves cannot create a high-tech center. As quality of life factors interact with other influences in developing high-tech centers, they frequently play a pivotal role in supporting innovation and high-tech entrepreneurship. Compared to other types of workers, workers in the high-tech occupations demonstrate higher levels of geographical mobility (Herzog, Schlottmann, and Johnson, 1986). Such highly skilled workers are likely to prefer locations with higher levels of amenities and move towards such locations (Gottlieb, 1995). Empirical findings in this literature support the idea that region-specific amenities are especially important for higher-skilled labor, using land rents, wages, and incomes data (Arora, Florida, Gates, and Kamlet, 2000). Roback (1982), using an equilibrium model, explains that areas with high amenities will have high land costs and relatively lower wage rates, so firms that demand high-skilled labor with limited land constrains (e.g., high tech 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. firms) will tend to locate in amenity-rich areas. Gottlieb (2001) suggests several high-tech development strategies from a close look at high-tech workers' inter-state mobility trends. The role of quality of life and research universities as factors in high-tech development attracts special attention in the most recent high-tech studies. Most regional science literature agrees that large urban regions such as metropolitan areas are the convergent points of good business climate and better quality of life. Recently, Florida (2002) appraised metropolitan areas as creativity receptacles in his creative capital theory. His findings demonstrate that metropolitan areas provide abundant options for creative-class people in both work and lifestyle (quality of life). The value of his research is finding a link between business climate theory and quality of life theory. Desrochers (2001) emphasizes the business aspects of metropolitan advantage in human creativity and technological innovation. He concludes that individual movements among different types of production or companies take place in urban areas, which finally yield to a new activity or a new product/process (Desrochers, 2001: 381-383). Metropolitan areas are where people from any background are accepted to turn their energy and ideas into innovation and wealth (Jacobs, 1969), and the location of a diverse set of firms and industries (Quigley, 1998). Jacobs (1968) broadly discussed the concept of cities as intellectual furnaces where new ideas are formed; whereas Lucas (1988) suggests that human capital spillover in cities are inherent to the creation of new ideas to support economic growth. Audretsch and Feldman (1997) say that intellectual innovations are strongly concentrated in metropolitan areas. Glaeser (1997) claims that the role 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of cities in generating new innovations may be “not in creating cutting edge technologies,” “but in creating learning opportunities for everyday people.” Dense urbanization economies in metropolitan areas provide a faster rate of contact between individuals, and each new contact offers an opportunity for learning (Glaeser, 1997: 2). Large cities are also the place of information institutions such as public research institutes, universities, and knowledge and technology transfer centers, where learning for innovation actively takes place. There are phase/time-sensitive theories concerning business and technology development, which can be applied to this study. First, the entrepreneurship/seedbed theory has been applied to emphasize the activities of scientists and engineers in high-tech development. This theory suggests that the main economic and social advantage of a high-tech center is achieved when the local environment supports new firm formation by innovative entrepreneurs. In particular, the theory emphasizes an entrepreneurial phase of high-tech development that takes place after an institutional phase. The entrepreneurship/seedbed theory highlights incubator space for start-up entrepreneurs, as opposed to permanent facilities for established firms or new branch plants (Luger and Goldstein, 1991). The role of research universities as a source of entrepreneurs is stressed in this approach. Second, the evolutionary theory of economic change also plays a role in the theoretical foundations of this study. Nelson and his colleagues maintain that there is a sharp inconsistency between neoclassical growth theory and micro studies of technological change (Nelson and Winter, 1974; Nelson, 1996). Evolutionary theory offers a 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. framework for a rigorous and rich analysis of the processes of technological change and dynamic competition, embracing several Schumpeterian notions. The main commitment of evolutionary theory is to a behavioral approach to the study of individual firms. Nelson's premise is that a firm at any time operates largely according to a set of decision rules that connect a range of business conditions to a range of responses on the part of firms. As technology evolves, firms deal with new conditions via changes in management, firm strategies, and firm organization (Nelson, 1996). This suggests that high-tech firms' contact points (or location factors) change according to changes in economic conditions. To be sure, studies of long waves of development may also reveal that government policy plays a pivotal role in developing high centers. Identifying the spatial shift and the emergence of new growth industries in the technology-induced phases of a long wave also seems plausible, though “the question of whether the upswing phases of long waves are really caused by endogenous technological innovation (technology push), or whether caused by factors such as wars or historical contingencies (demand pull), cannot be answered vaguely” (Sternberg, 1996). Though there are rare references to regulating long waves in the work of Kondratieff and Schumpeter, recent research finds the potential influence of government policy on the allocation of resources for R&D, and the possibility of alleviating the transition from one technology to the next by means of regional innovation policy. This theory implies a decreasing scale in the role of government policy after an initial period of high-tech development. It appears that 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. phase/time-sensitive theories suggest that the weight of each determinant and each location factor may change in accordance with time and business stages. 3.3 Quasi-Experimental Control Group Method 3.3.1 General Concept Applied social science researchers are frequently interested in evaluating effects or treatments. Randomized experimental and quasi-experimental designs are two major research designs that are used to evaluate the effects of treatments (Reichardt and Mark, 1997). In both designs, comparisons are made between what happens when a treatment is present and what happens when no treatment is present. The difference of these two designs is that in a randomized experimental design, a random process decides who receives the treatment, whilst in a quasi-experimental design, assignment to treatment conditions is determined deliberately. Thus, we can distinguish quasi-experimental designs form true experimental designs by the absence of randomized assignment of units to treatments (Campbell, 1988). Quasi- experimental design comes about because of difficulty of applying the classic natural science method to the (applied) social sciences, high costs of typical natural science methods, and development of new statistical tools that allowed for statistical control. In a quasi-experimental design, researchers substitute statistical controls for the absence of physical control of the experimental conditions. The most commonly used quasi-experimental design is the comparison group pre-test/post-test design. This design is the same as the typical controlled experimental design except that the 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. subjects cannot be randomly assigned to either the experimental or the control group, or the researcher is not able to control which group will have the treatment. Therefore, participatants do not always have the same opportunity of being in control or the experimental groups. The diagram of this design is as follows: Figure 3.1 Quasi-Experimental Design Diagram Source: St. Germain (2002) In a quasi-experimental design, how to select the best control group is the most important process after the researcher selects the treatment group. A distance measure is required to minimize the value between the variable matrix of treatment group and that of control group. Euclidean distance is not practicable in this method because of the problem of scale. When I use a different scale or unit, the Euclidean distance will be different. Thus I normalize the distance by using standardized distance, which can scale the data. Figure 3.2 is a basic function of standardized distance. Pre-test Experimental: O Comparison: O Treatment X Post-test O O 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.2 Standardized Distance for a Single Feature r2 = x - m = (x - m ) \ ( x - m) s x: treatment group, m: control group, s: standard deviation Source: Knapp (1998) Mahalanobis distance is introduced for utilizing standardized distance via a variable matrix to determine the best control group. Mahalonobis distance scales and weighs the variables by the variances and co-variances in the data. Figure 3.3 is the Mahalanobis distance function. Figure 3.3 Mahalonobis Distance for Variable Matrix r2 = ( x - m x)‘C~1(x -m x) x: treatment group, mx: control group, Cx: co-variance matrix Source: Knapp (1998) 3.3.2 Merits of the Method in this Research The quasi-experimental control group method used in this research has significant merits that other research methods can not claim. I minimize the common effects and endogenous growth elements in the group of high-tech centers. It is more reasonable to examine the suggested theories, and to support the research hypotheses, by using actual values of research variable the control group method provides than without screening common effects. 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This research shows the periods of high-tech center formation by the actual values of high-tech employment and income. Some high-tech location factors can be re-appraised though these have been highly praised so far, or their functional periods can be re-estimated. 3.3.3 Weaknesses of the Method in this Research The main weakness of the method in this research is its limitation in controlling for endogeneity. It can not wholly control for endogeneity mainly because data are not available. This research can not consider all social and economic conditions of the base year (1950) in selecting a control group (twin counties). Therefore, I control presumably the most important social and economic conditions for high-tech generation in base year (1950); education, employment, and land size. The control group selection is designed to discover twin counties whose education base, employment structure, and land size were similar to the eventual high-tech centers in 1950. The selected twin group consists of counties with similar education, employment, and land size with high-tech group. 3.4 Research Hypotheses From the theoretical framework and its practical applications, the following research hypotheses are established to test the formation of high-tech centers and their location factors. 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Research Hypothesis 1 Different from traditional firms, high-tech firms gain essential advantages from large pools of educated entrepreneurs that arise around high-tech centers (Premus, 1982). The role of research universities and educated graduates is to be put in the initial and incubating stage of high-tech centers (Luger and Goldstein, 1991; Nelson. 1996). This hypothesis is an application of entrepreneurship/seedbed theory and Nelson’s business evolution rationale. The former theory emphasizes role of institutions as incubators, and entrepreneurship as venture spirit for creating technological innovation. The latter idea denotes that firms change their management, strategies, and organization according to technological shift. The initial phase of high-tech industries, dependence to universities and role of engineering graduate are more important than any other period of high-tech centers. Most researchers agree that high-tech industries and high-tech centers were discovered from the 1950s to middle 1960s (Saxenian, 1981,1985; Castells, 1985; Hall, 1998). Thus, the role of universities and graduate students is expected to be the highest in this period; but after this period their role declines. Research Hypothesis 2 Business and residential amenity factors may start working after the incubating period. This research hypothesis is an application from Nelson’s evolutionary theory of firm behavior (Nelson, 1996). There are two reasons that these amenity factors apply later than the presence of universities and research universities’ graduates. 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. First, high-tech firms begin searching for better business conditions for spatial expansion and functional diffusion in case they have a good amount of educated and technical workers for firm operations. After high-tech firms capture enough graduates for their operation, or when graduate students start high-tech firms, these firms are likely to look for better business conditions, or quality of life factors for more qualified potential labor pool. Quality of life factors increase in importance because the educated and technical workers in high-tech firms have a wider range of choice to move to similar high-tech firms, which demand similar skills, and promise better wages and positions. Educated and technical workers will go to firms whose locations provide better residential and community amenities than others (Florida, 2002). This period may start just after the incubating period, that is the 1970s or the early 1980s. Research Hypothesis 3 It will be harder to sift the role of residential amenity factors than to examine business amenity factors or the role of universities and graduates in the formation of high-tech centers. It is because educated and technical workers in high-tech firms tend to live in wealthy communities (within metropolitan areas) already or after they obtain high-paying high-tech jobs (Herzog, Schlottmann, and Johnson, 1986; Gottlieb, 1995; Florida 2002). Higher wages provide high-tech workers with better living conditions and housing in wealthy communities. 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Research Hypothesis 4 Disaggregating high-tech employment into manufacturing and services, manufacturing decentralization into nonmetropolitan areas and its application to high-tech manufacturing decentralization (Norton and Rees, 1979; Barkley, 1988, 1993), experiences a different location trend and relation with location factors. High-tech services are expected to show a stronger location attraction to metropolitan areas than high-tech manufacturing, and they will play an important role in the formation of high-tech centers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 4 RESEARCH DESIGN AND METHODOLOGY 4.1 Chapter Summary This study is conducted using a high-tech county group and a twin county group; each group consists of fifty-five counties, respectively. The high-tech county group has significant high-tech employment size, high-tech employment density, and high- tech personal wage income in the terminal year (1997). The twin county group had similar conditions with the high-tech county group in education, employment, and geographical traits in the baseline year (1950), but has no meaningful high-tech employment size, high-tech employment density, and high-tech personal wage income in the terminal year (1997). The baseline year plays the role of “pre-test”, and the terminal year’s function is similar to “post-test” in a control group design. To select twin counties (control group), I apply the quasi-experimental control group method, and practically a combination of non-equivalent control group and interrupted time-series methods. Figure 4.1 demonstrates the basic design of this research. The purpose of the research is to investigate treatment year or period (Y) and actual significant location factors (X ) for the formation and growth of high- tech centers in the United States, and then examine the relationship between treatment year or period ( Y ) and actual significant location factors ( X ) using statistics. 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.1 Research Design and Research Targets Year (Pre-test) 7 (Treatment year or period) (Post-test) Experimental (High-tech g ro u p ): a a a a ~ x ~ (Location factors w orks) b b b B Control (Twin county g ro u p ): a a a a X (Location factors fails) a a a A * “a” and “b” can represent employment, total wage, or personal wage. This research focuses on U.S. counties’ experience for three reasons. First, there are about 3140 treatment and control group candidates among the U.S. Counties. It is a large enough pool to select high-tech and twin counties. Second, high-tech centers such as Silicon Valley, Route 128, and Research Triangle are well developed in the U.S., and their boundaries match counties. Third, county-level social, economic, and education data are available for public use. The research period is from 1950 (the baseline year) which was before the high-tech centers first emerged, until 1997 (the terminal year) when many metropolitan areas have one or more high-tech centers (Devol 1999) and the share of high-tech employment in the U.S. reached 10% (author calculation). This study includes two main research steps, and each step has three sub steps. Table 4.1 shows which investigations defined each step, the data used in each step with its source, and what methods and data ranges are applied to each research step. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.1 Research Steps of the Research Research Stops Data & Sources Methods & Ranges 1. Basic Database Buildup (1) Define High-tech Industry County Business Patterns, Census Bureau 3 or 4 digit SIC codes (2) Define high-tech (Counties) Centers County Business Patterns, Census Bureau 3 criteria, High-tech Industry = High-tech manufacturing+ High-tech services (3) Define Twin Counties US Historical Census Data Browser, Census Bureau Mahalanobis Distance, Quasi-experimental control group method 2. Statistical Tests (Basic statistics are included in each step) (4) T-test I: High-tech Center & Twin, To determine formation & growth period County Business Patterns, Census Bureau Employment, total wage income, personal wage income (1951,62,70, 80, 90, 97) (5) T-test II: Location Factors between High-tech Centers & Twin Counties County and City Data Book, Census Bureau Period of high-tech formation and growth, 28 location factors (6) Regression: High-tech Employment, Wage-Location Factors County Business Patterns, County and City Data Book, Census Bureau High-tech growth start point year, Significant location factors in step (5) 4.2 Basic Database Buildup 4.2.1 Defining High-tech Industry I use 3 or 4 digit SIC (Standard Industry Classification) codes in order to define high-tech industries. Many researchers agree that various SIC code represent the high-tech industry, and use 3 or 4 digit SIC codes in their research to define high- tech industries. Thus, I follow previous studies that define high-tech industries using SIC codes this step6. The high-tech industry has two sub categories: one is high-tech manufacturing and the other is high-tech services. Table 4.2 shows the selected 3 or 4 digit SIC codes in this research. In this research, the high-tech industry consists of twenty 3 or 4 digit SIC codes, which includes 12 SIC codes of high-tech manufacture and 8 SIC codes of high-tech service. 6 The following literature is used to define high-tech industries in my research: Barkley (1988), Devol (1999), Glasmeier et al (1983), Glasmeier (1985, 1986), Haug (1991), Lyons (1994), Malecki (1991), McArthur (1990), OECD (1982), Phillips (1991), Saxenian (1985), Scott (1992), Storper and Scott (1993), Vinson and Harrington (1979). 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.2 High-Tech Industry by 20 SIC codes Category ,S(Q' code Description High-tech Manufacturing 281 Industrial inorganic chemical 283 Drugs 2860 Industrial organic chemical 2911 Petroleum refining 357 Computer & office equipment 362 Electrical industrial apparatus 366 Communication equipment 367 Electronic components 372 Air crafts & parts 376 Guided missile & space vehicle 381 Search & navigation equipment 382 Measuring & control devices High-tech Service 481 Telephone communication 737 Computers & data processing services 7391 R&D laboratories 7397 Commercial testing laboratories 781 Motion picture production & services 871 Engineering & architectural services 873 Research & testing services 8922 Non-commercial education & science research organization Source: U.S. SIC Description, Census Bureau (1987) 4.2.2 Defining High-tech Centers There are indices for defining high-tech centers. Among them, I use three criteria to define high-tech counties (or high-tech centers). They are the high-tech employment location quotient, high-tech employment density quotient, and high-tech personal wage quotient in 1997 respectively. The high-tech employment location quotient represents the significance of the high-tech industry in the entire industrial structure. The high-tech employment density quotient stands for the degree of high-tech employment agglomeration in a region. The high-tech personal wage (income) quotient implicitly shows high-tech output. In my research, high-tech centers (counties) are defined to score more than 1.0 on all of three criteria. Three criteria 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are introduced here, and 60 counties are shown in Table 4.3, which satisfy these criteria. Three Criteria for High-tech Centers 1. High-tech Employment Density Location Quotient (1997) : Representing employment agglomeration (County _Hightech_ Employment / Nationwide _Hightech_Employment) jg larger than 1.00 County _ Area _Size / Nationwide _ Area _Size AND 2. High-tech Employment Location Quotient (1997) : Representing employment significance (County_Hightech_Employment / Nationwide_Hightech^Employment) is larger than 1.00 County_Total_Employment / Nationwide_Total_Employment AND 3. High-tech Personal Wage Income Quotient (1997) : Representing high-tech output ( County_Hightech_Wage_ Income / Nationwide_ Hightech_Wage_ Income ) is larger than County_Hightech_Employment / Nationwide_Hightech_Employment 1.00 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.3 High-tech Centers and the Values of Three Criteria 1 CA Alameda County 55.36 2.10 1.21 2 CA Contra Costa County 27.59 2.15 1.12 3 CA Marin County 12.40 2.02 1.33 4 CA San Francisco County 511.49 1.37 1.13 5 CA San Mateo County 62.24 2.53 1.44 6 CA Santa Barbara County 2.42 1.58 1.09 7 CA Santa Clara County 133.44 5.63 1.48 8 CA Santa Cruz County 9.06 1.66 1.11 9 CA Sonoma County 3.81 1.23 1.11 10 CO Arapahoe County 24.40 2.28 1.12 11 CO Boulder County 22.05 3.76 1.04 12 CO Denver County 177.65 1.46 1.07 13 CT Fairfield County 32.88 1.52 1.07 14 CT Hartford County 34.03 1.63 1.10 15 DC District of Columbia 276.90 1.35 1.11 16 GA DeKalb County 71.81 1.72 1.08 17 GA Fulton County 53.99 1.35 1.18 18 IL DuPage County 92.27 1.59 1.05 19 IN Kosciusko County 3.11 1.66 1.05 20 KS Johnson County 33.11 1.92 1.04 21 KS Sedgwick County 23.81 3.17 1.03 22 KY Marshall County 1.85 2.02 1.16 23 LA Iberville Parish 1.06 1.90 1.46 24 LA St. Jam es Parish 1.15 1.35 1.06 25 MD Howard County 49.47 3.01 1.09 26 MD Montgomery County 80.26 3.15 1.07 27 MA Essex County 25.13 1.57 1.01 28 MA Middlesex County 127.91 4.12 1.12 29 MA Norfolk County 36.17 1.40 1.03 30 MA Suffolk County 310.58 1.19 1.05 31 M l Oakland County 46.10 1.72 1.12 32 M l Washtenaw County 10.88 1.59 1.01 33 NJ Bergen County 85.29 1.40 1.04 34 NJ Hunterdon County 4.12 1.31 1.38 35 NJ Mercer County 27.30 1.15 1.06 36 NJ Middlesex County 87.04 2.25 1.24 37 NJ Monmouth County 17.52 1.23 1.07 38 NJ Morris County 47.37 2.49 1.25 39 NJ Som erset County 33.46 1.82 1.31 40 NY New York County 3390.97 1.18 1.12 41 NY W estchester County 28.90 1.12 1.11 42 NC Durham County 30.54 1.75 1.08 43 PA Chester County 16.57 2.11 1.16 44 PA Delaware County 39.76 1.08 1.04 45 PA Montgomery County 58.96 1.78 1.04 46 PA Philadelphia County 165.26 1.23 1.02 47 SC Greenville County 9.73 1.02 1.12 48 TX Brazoria County 1.67 1.24 1.14 49 TX Collin County 13.22 2.43 1.21 50 TX Dallas County 96.80 1.86 1.06 51 TX Fort Bend County 4.30 1.68 1.30 52 TX Harris County 40.35 1.36 1.07 53 TX Travis County 36.18 3.03 1.01 54 VA Arlington County 703.35 4.86 1.09 55 VA Fairfax County 187.22 5.19 1.19 56 VA Loudoun County 4.98 1.66 1.15 57 VA Alexandria city 422.30 2.95 1.01 58 VA Fairfax city 239.49 2.81 1.08 59 WA Benton County 1.43 1.60 1.10 60 WA King County 23.50 1.61 1.19 Source: Author Calculation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.2.3 Defining Twin Counties (Control Group) The method introduced here to select a control group is a combination of non equivalent control group and interrupted time-series methods (Campbell and Stanley, 1963; Isserman and Rephann, 1995). These are applied to multidimensional data in a way that allows the behavior of several variables to be examined simultaneously. This builds on the methods that Isserman utilized by selecting a control group from both theoretical and statistical perspectives, assessing the characteristics and suitability of the control group, and introducing statistical tests for making inferences about the effects of each high-tech location factor; the essence of this method to match treated counties (a successful high-tech center) with untreated ones that had similar economic and spatial characteristics in 1950, the baseline year for American high-tech formation. To obtain the best twins, it is necessary to use a measure of similarity to rank all potential twins. The choice of metric is the choice of the weight to assign to each variable. Mahalanobis distance, a metric usually utilized in statistical analysis, will be adopted to measure similarity. Mahalanobis distance is defined as follows: d \X „ ,X c) = (X „ -X cy S-‘( X „ - X c) where X is the vector of selection variables, H is the high-tech county, C is a possible control county, d is the distance between the two vectors, and ^ is the variance-covariance matrix of possible control counties. 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Mahalanobis distance scales and weights the variables by the variances and co-variances in the data, and allows trade-offs among competing selection variables (Isserman and Rephann, 1995). In this way, dissimilarity in one variable is exchanged for a tighter fit with another variable. The twin county selection process follows after selecting the high-tech centers. Twin counties are selected on the basis that their education, employment, conditions are most similar to high-tech counties in 1950 when high-tech centers were not yet established in the United States. This is the way that I minimize endogenous elements for testing the formation of high-tech centers. The following variables are used to define twin counties. Rationales of these variables are provided in Appendix 4. These variables are from the United State historical census data browser (http://fisher.lib.Virginia.edu/census/). In this process, two high-tech counties (District of Columbia, and Fairfax City in Virginia) are eliminated because the historical census browser does not provide their data for 1950. Table 4.4 demonstrates high-tech centers (counties) and their twin counties by the quasi- experimental control group method. Seven Variables Used to Define Twin Counties (Control Group) 1. Number of people with high school completed (1950) 2. Number of people with college or university completed (1950) 3. Number of total employment (1950) 4. Number o f government workers (1950) 5. Number of professional and technical workers (1950) 6. County total area size (1950) 7. Non-farm land size (1950) 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.4 High-Tech Centers and Their Twin Counties in the United States High-toch Centers Twin Counties No. State County State County 1 CA Alameda County OH Franklin County 2 CA Contra Costa County UT Salt Lake County 3 CA Marin County NY Tompkins County 4 CA San Francisco County M N Hennepin County 5 CA San Mateo County Wl Dane County 6 CA Santa Barbara County NY St. Lawrence County 7 CA Santa Clara County TX Tarrant County 8 CA Santa Cruz County VA Henrico County 9 CA Sonoma County NY Steuben County 10 CO Arapahoe County Wl Fond du Lac County 11 CO Boulder County PA Butler County 12 CO Denver County OR Multnomah County 13 CT Fairfield County NJ Union County 14 CT Hartford County NY Monroe County 15 GA DeKalb County VA Southampton County 16 GA Fulton County Rl Providence County 17 IL DuPage County NY Schenectady County 18 IN Kosciusko County IN Montgomery County 19 KS Johnson County TN Anderson County 20 KS Sedgwick County TX Jefferson County 21 KY Marshall County WV Doddridge County 22 LA Iberville Parish WV Nicholas County 23 LA St. Jam es Parish LA St. John the Baptist P 24 MD Howard County GA Spalding County 25 MD Montgomery County OK Oklahoma County 26 MA Essex County KY Jefferson County 27 MA Middlesex County NJ Essex County 28 MA Norfolk County MO S t Louis County 29 MA Suffolk County MD Baltimore city 30 M l Oakland County OH Summit County 31 M l W ashtenaw County IA Polk County 32 NJ Bergen County OH Hamilton County 33 NJ Hunterdon County NY G enesee County 34 NJ Mercer County VA Richmond city 35 NJ Middlesex County NJ Camden County 36 NJ Monmouth County NY Richmond County 37 NJ Morris County IL La Salle County 38 NJ Som erset County NY Rockland County 39 NY New York County NY Kings County 40 NY W estchester County NY Nassau County 41 NC Durham County WV Cabell County 42 PA Chester County NC Guilford County 43 PA Delaware County CT New Haven County 44 PA Montgomery County MA W orcester County 45 PA Philadelphia County M l Wayne County 46 SC Greenville County FL Orange County 47 TX Brazoria County Wl Marathon County 48 TX Collin County Wl Dodge County 49 TX Dallas County IN Marion County 50 TX Fort Bend County TX DeWItt County 51 TX Harris County NY Erie County 52 TX Travis County M l Ingham County 53 VA Arlington County TX Bexar County 54 VA Fairfax County SC Richland County 55 VA Loudoun County MN Nicollet County 56 VA Alexandria city MD Anne Arundel County 57 WA Benton County FL Broward County 58 WA King County Wl Milwaukee County Source: Aut ior Calculation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.5 shows the spatial patterns of high-tech centers and their twins in metropolitan and non-metropolitan areas. Table 4.5 Spatial Patterns of High-Tech Centers and their Twins in Metropolitan and Non-metropolitan Areas in 1997 CODE Definition Number of counties High-tech C enters (Treatm ent Group) Twin C ounties (Control Group) (94*8%) 46 (79.3%) 1 Counties in large metropolitan, 1 million or more residents 311 (9.9%) 47 (81.0%) 29 (50.0%) 2 Counties in small metropolitan, Less than 1 million residents 525 (16.7%) 8 (13.8%) 17 (29.3%) Non-Metropolitan Counties 3 (5.2%) 12 (20.7%) I. Adjacent to a large metro area and 3 Contains all or part of its own city of 10,000 or more residents 63 (2.0%) 0 (0.0%) 3 (5.2%) 4 Does not contains all or part of its own city of 10,000 or more residents 123 (3.9%) 0 (0.0%) 1 (1.7%) II. Adjacent to a small metro area and 5 Contains all or part of its own city of 10,000 or more residents 188 (6.0%) 1 (1.7%) 2 (3.4%) 6 Does not contains all or part of its own city of 10,000 or more residents 627 (20.0%) 1 (1.7%) 1 (1.7%) III. Not adjacent to a metro area and 7 Contains all or part of its own city of 10,000 or more residents 233 (7.4%) 0 (0.0%) 3 (5.2%) 3 Contains all or part of its own town of 2,500-19,999 residents 555 (17.7%) 1 (1.7%) 1 (1.7%) 9 Totally rural, not contain any part of a town of 2,500 or more residents 515 (16.4%) 0 (0.0%) 1 (1.7%) Average 3140 (100.0%) 58 (100.0%) 58 (100.0%) Source: United State Historical Census Data Browser, Census Bureau (1998), Urban Influence Code, United States Department of Agriculture (1997), Author Calculation. 4.3 Statistical Tests 4.3.1 T-test I (standard difference of means test I) To Determine Treatment Year (or Period) The first t-test is conducted to determine the year or period when the treatment starts in high-tech center group, that is, when the high-tech center group begins to be 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. different from the twin county group in high-tech employment, high-tech total wage income, and high-tech personal wage income. The treatment period is decided based on two parameters; one is t value at the 1 percent level significance, and the other is the actual growth rate of high-tech centers (employment, total wage, and personal wage). In t-test I, a standard difference-of-means of high-tech center group and twin county group is compared. A period is a treatment period when the starting year has no significant difference, but the ending year has a significant difference at the 1 percent level of significance, among periods of 1951-1962,1962-1970,1970-1980, 1980-1990, and 1990-1997. Table 4.6 demonstrates how to interpret t-test result to identify the period of high-tech formation and growth. In the table, the period between Year A and Year B is when high-tech centers are formed and grow. The actual growth rates, which confirm t-test results are calculated from the growth rate of [high-tech center employment (or wage)-twin county employment (or wage)] during 1951-1962,1962-1970,1970-1980,1980-1990, and 1990-1997 periods. These two parameters will be used to decide the period of high-tech center formation and growth. Data are obtained from County Business Patterns (U.S. Census Bureau) in 1951,1962,1970,1980,1990, and 1997. The U.S. Census Bureau provides CD ROM for 1990 and 1997 data, and rest in hard copy. Table 4.6 How to Interpret T-test Result to Decide the Period of High-Tech Center T value in 1% level significance Year 1950 (Pre-test) Year A Y earB Y earC 1997 (Post-test) Experim ental (High-tech g ro u p ): Sam e Sam e Different Different Different Control (Twin g ro u p ): * Results: High-tech centers were formed between Year A and Year B 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.3.2 T-test II (standard difference of means test II) To Examine Significant Location Factors The second t-test was designed to examine significant location factors that work during the treatment year or period. The process of t-test II is similar to that of t-test I, except that t-test II compares location factors, not high-tech employment, total wage, and personal wage, and t-test II is given for a period that is obtained from t- test I. Location factors will be considered to work for high-tech formation and growth if the starting year has no significant differences, but the ending year has a significant difference at the 1 percent (or 5 percent) level significance in the given high-tech treatment period. The candidates of substantial high-tech location factors from t-test II will be finally determined after they pass the regression process with high-tech employment, total wage income, and personal wage income. Data are obtained from County and City Data Book (Census Bureau). Census Bureau provides electronic data in 1988 and 1994 data, and 1952,1962,1972,1977, and 1983 by hard copy. 4.3.3 Regressions Regressions is conducted to confirm the relationship between the selected location factors and high-tech employment, high-tech total wage income, high-tech personal wage income. The model for the simple regressions is Yt = /30 + fix X i + s Yt is high-tech employment, total wage income, or personal wage income X i is location factors Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In cases where there are more than two location factors, I will conduct multiple regression analysis to examine a more comprehensive explanation and relative degrees of contribution of the various location factors. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 5 RESULTS 5.1 Chapter Summary Chapter five examines a treatment year (period) for the initial formation of high-tech center in the United States and their substantial location factors using statistical analysis (t-test, actual growth rate, and regression). The 1970-1980 period shows the highest numbers in high-tech employment and high-tech total wage; while, the 1951- 1962 period shows the highest in high-tech personal wage followed by the 1970- 1980 period. The 1970-1980 period obtains the highest point in general. In analysis of location factors during the 1970-1980 period, “the number of physicians per 100,000 resident population” as a proxy for quality of life shows a significant probability in the initial formation of high-tech centers in the United States. This result opens a feasibility of a certain class’ substantial role in the formation of high- tech centers in the United States. 5.2 Treatment Year (or Period) A treatment period is examined for high-tech employment, high-tech total wage income, and high-tech personal wage income. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.2.1 High-Tech Employment Descriptive Statistics Table 5.1 shows the descriptive statistics for high-tech employment through the research period (1951-1997), including high-tech manufacturing and high-tech service in Table 5.2 and Table 5.3 each. Table 5.1 Descriptive Statistics for High-Tech Employment (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 4770.3 9403 0 50015 276677 1962 58 7407.1 15141 0 88073 429611 1970 58 9831.9 15919 0 58024 570253 1980 58 27569 36056 0 205287 1599028 1990 58 45508 66464 0 405806 2639483 1997 58 67464 81269 778 465282 3912932 Twin County Group 1951 58 2815.3 5706.6 0 28663 163287 1962 58 4335.9 8606.9 0 44837 251485 1970 58 5681.5 12329 0 73575 329526 1980 58 13416 14428 0 50246 778099 1990 58 12915 16276 0 82672 749092 1997 58 21338 24754 0 135053 1237628 Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.2 Descriptive Statistics for High-Tech Manufacturing Employment (1951- 1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 4206.2 8284.5 0 38645 243960 1962 58 5642 11390 0 48556 327235 1970 58 7342.1 13676 0 56256 425843 1980 58 15735 27700 0 178257 912612 1990 58 15869 42771 0 310426 920381 1997 58 18420 42105 0 297977 1068336 Twin County Group 1951 58 2561 5515 0 27949 148536 1962 58 3701.8 7958.2 0 42149 214703 1970 58 4782.3 11718 0 70453 277375 1980 58 8240.1 10358 0 40386 477926 1990 58 3806.2 6346.9 0 35553 220758 1997 58 4929.8 8771.9 0 52683 285927 Source: County Business Patterns, Census Bureau (1951-1997) Table 5.3 Descriptive Statistics for High-Tech Service Employment (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 564.09 1654.8 0 11370 32717 1962 58 1765.1 5281 0 39517 102376 1970 58 2489.8 6101.9 0 44072 144410 1980 58 11835 16855 0 105545 686416 1990 58 29640 36516 0 168469 1719102 1997 58 49045 52537 0 209234 2844596 Twin County Group 1951 58 254.33 619.63 0 4479 14751 1962 58 634.17 928.43 0 5031 36782 1970 58 899.16 1174.1 0 5659 52151 1980 58 5175.4 5347.4 0 23327 300173 1990 58 9109.2 11105 0 47119 528334 1997 58 16409 17907 0 82370 951701 Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. T-test Results Table 5.4 shows t-test results for high-tech employment through the research period (1951-1997), including high-tech manufacturing and high-tech services in Table 5.5 and Table 5.6 each. Table 5.4 t-test Results: High-Tech Employment * p<0.01 ** p<0.001 Year Index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1951 N 58 58 Mean 4770.3 2815.3 1955 1.35 0.1791 Std Dev 9403 5706.6 7777.6 1962 N 58 58 Mean 7407.1 4335.9 3071.1 1.34 0.1826 Std Dev 15141 8606.9 12315 1970 N 58 58 Mean 9831.9 5681.5 4150.5 1.57 0.1194 Std Dev 15919 12329 14237 1980 N 58 58 Mean 27569 13416 14154 2.78* 0.007 Std Dev 36056 14428 27461 1990 N 58 58 Mean 45508 12915 32593 3.63** 0.0006 Std Dev 66464 16276 48386 1997 N 58 58 Mean 67464 21338 46126 4.13** 0.0001 Std Dev 81269 24754 60072 High-tech group and Twin group are not different in 1951,1962,1970 High-tech group and Twin group are different in 1980 at 1% level significance High-tech group and Twin group are different in 1990,1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Table 5.4 demonstrates that high-tech employment has been significantly different between the high-tech center group and the twin group since 1980. Thus, certain location factors had a higher probability to build up high-tech centers via 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. high-tech employment growth in the 1970s. I find the role of high-tech manufacturing and high-tech services to total high-tech employment growth during this period in Table 5.5 and Table 5.6. The role of high-tech services employment is more than that of high-tech manufacturing employment in the growth of total high- tech employment during 1970-1980. High-tech manufacturing employment has no difference between high-tech group and twin group in each year at the 1 percent level significance. High-tech service employment begins to be different between the groups since 1980 at the 1 percent level significance. Table 5.5 t-test Results: High-Tech Manufacturing Employment * p<0.01 ** p<0.001 Year Index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1951 N 58 58 Mean 4206.2 2561 1645.2 1.26 0.211 Std Dev 8284.5 5515 7037.4 1962 N 58 58 Mean 5642 3701.8 1940.2 1.06 0.2901 Std Dev 11390 7958.2 9824.9 1970 N 58 58 Mean 7342.1 4782.3 2559.8 1.08 0.2814 Std Dev 13676 11718 12735 1980 N 58 58 Mean 15735 8240.1 7494.6 1.93 0.0575 Std Dev 27700 10358 20912 1990 N 58 58 Mean 15869 3806.2 12062 2.12 0.0378 Std Dev 42771 6346.9 30575 1997 N 58 58 Mean 18420 4929.8 13490 2.39 0.02 Std Dev 42105 8771.9 30412 High-tech group and Twin group are not different in 1951,1962,1970,1980,1990, 1997 Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.6 t-test Results: High-Tech Services Employment * p<0.01 ** p<0.001 Year index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1951 N 58 58 Mean 564.09 254.33 309.76 1.34 0.186 Std Dev 1654.8 619.63 1249.5 1962 N 58 58 Mean 1765.1 634.17 1130.9 1.61 0.1134 Std Dev 5281 928.43 3791.5 1970 N 58 58 Mean 2489.8 899.16 1590.7 1.95 0.0558 Std Dev 6101.9 1174.1 4393.9 1980 N 58 58 Mean 11835 5175.4 6659.4 2.87* 0.0055 Std Dev 16855 5347.4 12504 1990 N 58 58 Mean 29640 9109.2 20530 4.1** 0.0001 Std Dev 36516 11105 26988 1997 N 58 58 Mean 49045 16409 32636 4.48** <.0001 Std Dev 52537 17907 39248 High-tech group and Twin group are not different in 1951,1962,1970 High-tech group and Twin group are different in 1980 at 1% level significance High-tech group and Twin group are different in 1990,1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Actual Growth Rates Table 5.7, Table 5.8, and Figures 5.1, 5.2, 5.3 demonstrate the general trends of high- tech employment growth in the various periods. In Figure 5.1, 5.2, and 5.3,1 find changes in slopes of actual growth (A-B) between 1970 and 1980 in high-tech total employment, high-tech manufacturing employment and high-tech services employment. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.7 Total High-Tech Employment in High-Tech Group and Twin Group Year 1951 1962 1970 1980 1990 1997 High-tech (A) 276677 429611 570253 1599028 2639483 3912932 Twin (B) 163287 251485 329526 778099 749092 1237628 (A)-(B) = (C) 113390 178126 240727 820929 1890391 2675304 (C)/(A) 41.0% 41.5% 42.2% 51.3% 71.6% 68.4% Year 1951 1962 1970 1980 1990 1997 High-tech Manufacturing (A) 243960 327235 425843 912612 920381 1068336 Twin Manufacturing (B) 148536 214703 277375 477926 220758 285927 (A)-(B) = (C) 95424 112532 148468 434686 699623 782409 (C)/(A) 39.1% 34.4% 34.9% 47.6% 76.0% 73.2% Year 1951 1962 1970 1980 1990 1997 High-tech Services (A) 32717 102376 144410 686416 1719102 2844596 Twin Services (B) 14751 36782 52151 300173 528334 951701 (A)-(B) = (C) 17966 65594 92259 386243 1190768 1892895 (C)/(A) 54.9% 64.1% 63.9% 56.3% 69.3% 66.5% Source: County Business Patterns, Census Bureau (1951-1997) Table 5.8 Average High-Tech Employment in High-Tech Group and Twin Group Year 1951 1962 1970 1980 1990 1997 High-tech (A) 4770 7407 9832 27569 45508 67464 Twin (B) 2815 4336 5681 13416 12915 21338 (A)-(B) = (C) 1955 3071 4150 14154 32593 46126 (C)/(A) 41.0% 41.5% 42.2% 51.3% 71.6% 68.4% Year 1951 1962 1970 1980 1990 1997 High-tech Manufacturing (A) 4206 5642 7342 15735 15869 18420 Twin Manufacturing (B) 2561 3702 4782 8240 3806 4930 (A)-(B) = (C) 1645 1940 2560 7495 12062 13490 (0)1(A) 39.1% 34.4% 34.9% 47.6% 76.0% 73.2% Year 1951 1962 1970 1980 1990 1997 High-tech Services (A) 564 1765 2490 11835 29640 49045 Twin Sen/ices (B) 254 634 899 5175 9109 16409 (A)-(B) = (C) 310 1131 1591 6659 20530 32636 (C)/(A) 54.9% 64.1% 63.9% 56.3% 69.3% 66.5% Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.1 Average High-Tech Employment Trend, 1951-1997 80000 70000 - 60000 B 50000 - g- 40000 a. E U J 30000 20000 10000 1951 1962 1970 1980 1997 1990 -High-tech (A) -Twin (B) - (A)-(B) = (C) Year Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.2 Average High-Tech Manufacturing Employment Trend, 1951-1997 20000 18000 16000 - 14000 g 12000 0 ) I 10000 Q . E 8000 11 1 6000 4000 2000 1951 1962 1970 1980 1990 1997 Year —■ — High-tech Manufacturing (A) — Tw in Manufacturing (B) - * - ( A ) - ( B ) = (C) Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.3 Average High-Tech Service Employment Trend, 1951-1997 60000 50000 40000 30000 20000 10000 0 1951 1962 1970 1980 1990 1997 Y ear —■ — High-tech Services (A) — Tw in Services (B) —X— (A)-(B) = (C) Source: County Business Patterns, Census Bureau (1951-1997) Table 5.9 and Figures 5.4, 5.4, 5.6 reconfirm t-test and employment trend results. Actual growth of high-tech employment shows the highest values between 1970-1980 in total high-tech employment, high-tech manufacturing, and high-tech service employment. 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.9 Actual High-Tech Employment Growth in High-Tech Group and Twin Group Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 High-tech growth 55.3% 32.7% 180.4% 65.1% 48.2% Twin growth 54.0% 31.0% 136.1% -3.7% 65.2% Actual high-tech growth 57.1% 35.1% 241.0% 130.3% 41.5% Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 High-tech Manufacturing growth 34.1% 30.1% 114.3% 0.9% 16.1% Twin Manufacturing growth 44.5% 29.2% 72.3% -53.8% 29.5% Actual high-tech manufacturing growth 17.9% 31.9% 192.8% 60.9% 11.8% Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 High-tech Services growth 212.9% 41.1% 375.3% 150.4% 65.5% Twin Services growth 149.4% 41.8% 475.6% 76.0% 80.1% Actual high-tech services growth 265.1% 40.7% 318.7% 208.3% 59.0% Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.4 Average High-Tech Employment Growth Rates A verage High-tech growth A verage Twin growth Actual high-tech growth Period Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.5 Average High-Tech Manufacturing Employment Growth Rates 250.0% 200.0% 150.0% A verage High-tech Manufacturing grow th Average Tw in Manufacturing growth Actual high-tech manufacturing growth 100.0% 50.0% 0.0% 951-191% 198M 9TO '' 'tStlW&BO -50.0% - 100.0% Year Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.6 Average High-Tech Service Employment Growth Rates 500.0% 450.0% - 400.0% - 350.0% - S 300.0% ( 9 B C | 250.0% O 200.0% 150.0% 100.0% 50.0% -A verage High-tech Services grow th -A verage Twin Services growth -Actual high-tech services grow th 1951 1962 1962 1970 197 0 1980 19B 0-1990 1 990 1997 Year Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.2.2 High-tech Total Wage Income Descriptive Statistics Table 5.10 shows descriptive statistics for high-tech total wage income through the research period (1951-1997), including high-tech manufacturing and high-tech service in Table 5.11 and Table 5.12 each. Table 5.10 Descriptive Statistics for High-Tech Total Wage Income (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum 1951 58 18683 37333 0 184988 1083608 1962 58 48137 99661 0 589220 2791936 High-tech 1970 58 94473 155247 0 561980 5479428 Center Group 1980 58 414554 716045 0 3760000 24044129 1990 58 1780000 2740000 0 17500000 103036964 1997 58 3820000 5220000 40332 33600000 221827845 1951 58 10847 21719 0 95312 629124 1962 58 28254 59455 0 329272 1638716 Twin County 1970 58 43137 82714 0 397404 2501964 Group 1980 58 144028 198090 0 727977 8353646 1990 58 417861 552475 0 2760000 24235920 1997 58 932725 1110000 0 6220000 54098072 Source: County Business Patterns, Census Bureau ( 951-1997) Table 5.11 Descriptive Statistics for High-Tech Manufacturing Wage Income (1951- 1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 16195 32617 0 130048 939300 1962 58 35782 72438 0 307816 2075360 1970 58 68708 130893 0 545620 3985068 1980 58 222149 523687 0 3310000 12884627 1990 58 597232 1790000 0 13300000 34639443 1997 58 1020000 2830000 0 20900000 59309010 Twin County Group 1951 58 9767 20790 0 92992 566464 1962 58 24359 54997 0 311136 1412848 1970 58 34689 74896 0 390932 2011980 1980 58 89149 157054 0 671791 5170624 1990 58 121308 208741 0 1160000 7035871 1997 58 210405 382788 0 2270000 12203514 Source: County Business Patterns, Census Bureau ( 951-1997) 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.12 Descriptive Statistics for High-Tech Service Wage Income (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 2488 7948 0 54940 144308 1962 58 12355 38484 0 288908 716576 1970 58 25765 64373 0 465548 1494360 1980 58 192405 370893 0 2360000 11159502 1990 58 1180000 1450000 0 6420000 68397521 1997 58 2800000 3120000 0 12700000 162518835 Twin County Group 1951 58 1080 3238 0 24216 62660 1962 58 3894 6444 0 37480 225868 1970 58 8448 14678 0 97708 489984 1980 58 54880 71782 0 259694 3183022 1990 58 296553 379524 0 1610000 17200049 1997 58 722320 815278 0 3950000 41894558 Source: County Business Patterns, Census Bureau ( 951-1997) t-test Results Table 5.13 shows t-test results for high-tech wage income through the research period (1951-1997), including high-tech manufacturing and high-tech service in Table 5.14 and Table 5.15 each. 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.13 T-test Results: High-Tech Total Wage Income * p<0.01 ** p<0.001 Year Index High-tech (1) Twin (2) Diff (1-2) T Value P r > |t| 1951 N 58 58 Mean 18683 10847 7835.9 1.38 0.1704 Std Dev 37333 21719 30541 1962 N 58 58 Mean 48137 28254 19883 1.3 0.1952 Std Dev 99661 59455 82059 1970 N 58 58 Mean 94473 43137 51336 2.22 0.0288 Std Dev 155247 82714 124385 1980 N 58 58 Mean 414554 144028 270526 2.77* 0.0072 Std Dev 716045 198090 525338 1990 N 58 58 Mean 1.78E+06 417861 1.36E+06 3.7** 0.0005 Std Dev 2.74E+06 552475 1.98E+06 1997 N 58 58 Mean 3.82E+06 932725 2.89E+06 4.13** 0.0001 Std Dev 5.22E+06 1.11E+06 3.77E+06 High-tech group and Twin group are not different in 1951,1962,1970 High-tech group and Twin group are different in 1980 at 1% level significance High-tech group and Twin group are different in 1990,1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Table 5.13 demonstrates that high-tech total wage income has been different between the high-tech center group and the twin group since 1980. Thus, certain location factors have a higher probability of working to build up high-tech centers by high-tech wage income growth during 1970-1980. I find the role of high-tech manufacturing and high-tech services to total high-tech wage income growth during this period in Table 5.14 and Table 5.15. The role of high-tech services wage income plays a leading role in the growth of total high-tech wage income during 1970-1980. High-tech manufacturing wage income shows no difference between the high-tech group and the twin group in each year at the 1 percent level significance. High-tech service wage income begins to be different between the groups since 1980 at the 1 percent level significance. The t-test in high-tech wage income was done 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with an expectation that there would be an earlier significant difference in wage income because of higher paying employees. However, the t-test results of high-tech total wage income are very similar to that of high employment in the previous section. Table 5.14 t-test Results: High-Tech Manufacturing Wage Income * p<0.01 ** p<0.001 Year Index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1951 N 58 58 Mean 16195 9766.6 6428.2 1.27 0.2087 Std Dev 32617 20790 27350 1962 N 58 58 Mean 35782 24359 11423 0.96 0.341 Std Dev 72438 54997 64311 1970 N 58 58 Mean 68708 34689 34019 1.72 0.0892 Std Dev 130893 74896 106636 1980 N 58 58 Mean 222149 89149 133000 1.85 0.0683 Std Dev 523687 157054 386597 1990 N 58 58 Mean 597232 121308 475924 2.01 0.0494 Std Dev 1.79E+06 208741 1.28E+06 1997 N 58 58 Mean 1.02E+06 210405 812164 2.17 0.0344 Std Dev 2.83E+06 382788 2.02E+06 High-tech group and Twin group are not different in 1951,1962,1970,1980,1990,1997 Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.15 t-test Results: High-Tech Service Wage Income * p<0.01 _______________________________________________________________________________ ** p<O.OQ1 Year index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1951 N 58 58 Mean 2488.1 1080.3 1407.7 1.25 0.2155 Std Dev 7948.1 3237.6 6068.6 1962 N 58 58 Mean 12355 3894.3 8460.5 1.65 0.1039 Std Dev 38484 6444.1 27591 1970 N 58 58 Mean 25765 8448 17317 2 0.0501 Std Dev 64373 14678 46687 1980 N 58 58 Mean 192405 54880 137526 2.77* 0.0074 Std Dev 370893 71782 267128 1990 N 58 58 Mean 1.18E+06 296553 882715 4.49** <.0001 Std Dev 1.45E+06 379524 1.06E+06 1997 N 58 58 Mean 2.80E+06 722320 2.08E+06 4.92** <.0001 Std Dev 3.12E+06 815278 2.28E+06 High-tech group and Twin group are not different in 1951,1962,1970 High-tech group and Twin group are different in 1980 at 1% level significance High-tech group and Twin group are different in 1990,1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Actual Growth Rate Table 5.16, Table 5.17, and Figures 5.7, 5.8, and 5.9 demonstrate the general trends of high-tech wage income growth in the given periods. In Figure 5.7, 5.8, and 5.9, we find changes in slopes of actual growth (A-B) between 1970 and 1980 in high- tech total wage income, high-tech manufacturing wage income and high-tech service wage income. 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.16 Total High-Tech Wage Income, (Annual, $1000) Year 1951 1962 1970 1980 1990 1997 High-tech (A) 1083608 2791936 5479428 24044129 103036964 221827845 Twin (B) 629124 1638716 2501964 8353646 24235920 54098072 (A)-(B) = (C) 454484 1153220 2977464 15690483 78801044 167729773 (C)/(A) 41.9% 41.3% 54.3% 65.3% 76.5% 75.6% Year 1951 1962 1970 1980 1990 1997 High-tech Manufacturing (A) 939300 2075360 3985068 12884627 34639443 59309010 Twin Manufacturing (B) 566464 1412848 2011980 5170624 7035871 12203514 (A)-(B) = (C) 372836 662512 1973088 7714003 27603572 47105496 (C)/(A) 39.7% 31.9% 49.5% 59.9% 79.7% 79.4% Year 1951 1962 1970 1980 1990 1997 High-tech Services (A) 144308 716576 1494360 11159502 68397521 162518835 Twin Services (B) 62660 225868 489984 3183022 17200049 41894558 (A)-(B) = (C) 81648 490708 1004376 7976480 51197472 120624277 (C)/(A) 56.6% 68.5% 67.2% 71.5% 74.9% 74.2% Source: County Business Patterns, Census Bureau (1951-1997) Table 5.17 Average High-Tech Wage Income, (Annual, $1000) Year 1951 1962 1970 1980 1990 1997 High-tech (A) 18682.9 48136.8 94472.9 414553.9 1776499.4 3824618.0 Twin (B) 10847.0 28253.7 43137.3 144028.4 417860.7 932725.4 (A)-(B) = (C) 7835.9 19883.1 51335.6 270525.6 1358638.7 2891892.6 (C)/(A) 41.9% 41.3% 54.3% 65.3% 76.5% 75.6% Year 1951 1962 1970 1980 1990 1997 High-tech Manufacturing (A) 16194.8 35782.1 68708.1 222148.7 597231.8 1022569.1 Twin Manufacturing (B) 9766.6 24359.4 34689.3 89148.7 121308.1 210405.4 (A)-(B) = (C) 6428.2 11422.6 34018.8 133000.1 475923.7 812163.7 (C)/(A) 39.7% 31.9% 49.5% 59.9% 79.7% 79.4% Year 1951 1962 1970 1980 1990 1997 High-tech Services (A) 2488.1 12354.8 25764.8 192405.2 1179267.6 2802048.9 Twin Services (B) 1080.3 3894.3 8448.0 54879.7 296552.6 722320.0 (A)-(B) = (C) 1407.7 8460.5 17316.8 137525.5 882715.0 2079728.9 (C)/(A) 56.6% 68.5% 67.2% 71.5% 74.9% 74.2% Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.7 Average High-Tech Wage Income Trend, 1951-1997 o v < / > E o u c 0 > O ) 3 c c < 4000000.0 3500000.0 3000000.0 2500000.0 2000000.0 1500000.0 1000000.0 500000.0 0.0 / f / / / a " / ■ ■ ’ j j t - , • f j f * , i‘, f ■ ! ) ■ ■ ■ j j v j i i * ? * * * 4 * r v j j x r s s s s r i i ........ - . * . h i * * .V** M «• ** i" ' j x * ' > " --------------- — J S M ^ g Z g T . ■ -■ • - ■ - >_,tv *" ' ' ' £& # - High-tech (A) - Tw in (B) ■ (A)-(B) = (C) 1951 1962 1970 1980 1990 1997 Y ear Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.8 Average High-Tech Manufacturing Wage Income Trend, 1951-1997 1200000.0 1000000.0 800000.0 600000.0 400000.0 200000.0 0.0 1951 1962 1970 1980 1990 1997 Year ■High-tech M anufacturing (A) ■Tw in M anufacturing (B) ■ (A )-(B ) = (C) Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.9 Average High-Tech Service Wage Income Trend, 1951-1997 3000000.0 2500000.0 5 2000000.0 E 0 1 1500000.0 0 ) U l £ I 1000000.0 c c 500000.0 - I 0.0 ■ H igh-tech S e rv ic e s (A) -T w in S e rv ic e s (B) ■ (A )-(B ) = (C) 1951 1962 1970 1980 Year 1990 1997 Source: County Business Patterns, Census Bureau (1951-1997) Table 5.18 and Figure 5.10, 5.11, 5.12 reconfirm t-test and wage income trends. Actual growth of high-tech wage income shows the highest values between 1970-1980 in total high-tech wage income, high-tech manufacturing, and high-tech service wage income. 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.18 Actual High-Tech Wage Income Growth Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 Average High-tech growth 157.7% 96.3% 338.8% 328.5% 115.3% Average Twin growth 160.5% 52.7% 233.9% 190.1% 123.2% Actual high-tech growth 153.7% 158.2% 427.0% 402.2% 112.9% Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 Average High-tech Manufacturing growth 120.9% 92.0% 223.3% 168.8% 71.2% Average Twin Manufacturing growth 149.4% 42.4% 157.0% 36.1% 73.4% Actual high-tech manufacturing growth 77.7% 197.8% 291.0% 257.8% 70.7% Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 Average High-tech Services growth 396.6% 108.5% 646.8% 512.9% 137.6% Average Twin Services growth 260.5% 116.9% 549.6% 440.4% 143.6% Actual high-tech services growth 501.0% 104.7% 694.2% 541.9% 135.6% Source: County Business Patterns, Census Bureau ( 951-1997) Figure 5.10 Average High-Tech Wage Income Growth Rates 450.0% 400.0% 350.0% 300.0% Average High-tech grow th Average Tw in grow th Actual high-tech grow th 250.0% 200 .0% 150.0% o > 100.0% 50.0% 0 .0 % 1951- 1962- 1970- 1980- 1990- 1962 1970 1980 1990 1997 P eriod Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.11 Average High-Tech Manufacturing Wage Income Growth Rates 350.0% 300.0% 250.0% 2 0 0 .0 % 150.0% 50.0% 0 .0 % 1951- 1970- 1980- 1990- 1962- 1962 1970 1980 Pe rlod 1990 1997 -A v e ra g e H igh-tech M anufacturing grow th -A v e ra g e Tw in M anufacturing grow th -A ctu al high-tech m anu factu rin g grow th Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.12 Average High-Tech Service Wage Income Growth Rates a 2 J Z I 0 d > E o o c i 800.0% ► * * * s * * % , < i 700.0% 600.0% 500.0% 400.0% 300.0% ■Igllilll 200.0% 100.0% 0 .0 % 1951- 1962- 1970- 1980- 1990- 1962 1970 1980 1990 1997 Period —• — Average High-tech Services grow th — ■— Average Tw in Services grow th — A— Actual high-tech services grow th Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.2.3 High-tech Personal Wage Income Descriptive Statistics Table 5.19 shows descriptive statistics for high-tech personal wage income through the research period (1951-1997), including high-tech manufacturing and high-tech service in Table 5.20 and Table 5.21 each. Table 5.19 Descriptive Statistics for High-Tech Personal Wage Income (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 2492 1750 0 4757 144509 1962 58 4694 2779 0 7809 272259 1970 58 7572 3922 0 13283 439163 1980 58 17900 5655 0 25729 1038226 1990 58 34922 9358 0 47758 2025453 1997 58 55306 5528 49240 72315 3207735 Twin County Group 1951 58 2180 1778 0 5027 126428 1962 58 3789 2809 0 8257 219768 1970 58 5907 3923 0 12513 342626 1980 58 13448 8185 0 25482 780011 1990 58 27140 11123 0 45462 1574149 1997 58 38621 11739 0 55251 2239993 Source: County Business Patterns, Census Bureau (1951-1997) Table 5.20 Descriptive Statistics for High-Tech Manufacturing Personal Wage Income (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 1803 1927 0 4966 104593 1962 58 3301 3169 0 7595 191460 1970 58 5709 4535 0 14579 331139 1980 58 14253 7277 0 26326 826660 1990 58 35840 10095 0 54288 1514761 1997 58 45184 15158 0 71519 2620692 Twin County Group 1951 58 1366 1814 0 4675 79206 1962 58 2405 2985 0 8257 139466 1970 58 3254 4100 0 10429 188746 1980 58 9460 8991 0 24918 548653 1990 58 27529 11324 0 45457 1021354 1997 58 28130 21191 0 73533 1631539 Source: County Business Patterns, Census Bureau (1951-1997) 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.21 Descriptive Statistics for High-Tech Service Personal Wage Income (1951-1997) Category Year Observation Mean Standard Deviation Minimum Maximum Sum High-tech Center Group 1951 58 2443 1701 0 5127 141686 1962 58 4842 2885 0 8429 280832 1970 58 7456 4104 0 12968 432431 1980 58 18437 6399 0 30297 1069343 1990 58 35840 10095 0 54288 2078738 1997 58 53431 12477 0 76112 3099018 Twin County Group 1951 58 3433 9883 0 76235 199117 1962 58 3689 2806 0 7450 213976 1970 58 6190 4144 0 17266 358993 1980 58 13649 8306 0 25482 791668 1990 58 27529 11324 0 45457 1596676 1997 58 38644 11623 0 52727 2241354 Source: County Business Patterns, Census Bureau (1951-1997) t-test Results Table 5.22 shows T-test results for high-tech personal wage income through the research period (1951-1997), including high-tech manufacturing and high-tech services in Table 5.23 and Table 5.24 each. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.22 T-test Results: High-Tech Personal Wage Income * p<0.01 ** p<0.001 Year index High-tech (1) Twin (2) Diff (1-2) t Value Pr > |t| 1951 N 58 58 Mean 2491.5 2179.8 311.75 0.95 0.3432 Std Dev 1749.6 1777.7 1763.7 1962 N 58 58 Mean 4694.1 3789.1 905.03 1.74 0.0838 Std Dev 2779.1 2808.7 2793.9 1970 N 58 58 Mean 7571.8 5907.3 1664.4 2.29 0.0242 Std Dev 3921.9 3922.7 3922.3 1980 N 58 58 Mean 17900 13448 4452 3.41** 0.0009 Std Dev 5655.1 8185.4 7034.9 1990 N 58 58 Mean 34922 27140 7781.1 4.08** <.0001 Std Dev 9357.9 11123 10279 1997 N 58 58 Mean 55306 38621 16685 9.79** <.0001 Std Dev 5528.2 11739 9174.9 High-tech group and Twin group are not different in 1951,1962,1970 High-tech group and Twin group are different in 1980,1990,1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Table 5.22 demonstrates that high-tech personal wage income has been different between the high-tech center group and the twin group since 1980. Thus, certain location factors have a higher probability of working to build up high-tech centers by high-tech personal wage income growth during 1970 to 1980. In Table 5.23 and Table 5.24, high-tech manufacturing personal wage income has been different between the high-tech group and the twin group since 1970 at the 1 percent level significance (at the 0.1 percent level significance in 1997). High-tech service personal income begins to be different between the groups since 1980 at the 0.1 percent level significance. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.23 t-test Results: High-Tech Manufacturing Personal Wage Income * p<0.01 _______________________________________________________________________________ ** p<0.001 Year index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1951 N 58 58 Mean 1803.3 1365.6 437.69 1.26 0.2104 Std Dev 1927.1 1813.9 1871.4 1962 N 58 58 Mean 3301 2404.6 896.44 1.57 0.1196 Std Dev 3168.7 2984.6 3078 1970 N 58 58 Mean 5709.3 3254.2 2455 3.06* 0.0028 Std Dev 4535.1 4100.1 4323.1 1980 N 58 58 Mean 14253 9459.5 4793.2 3.16* 0.0021 Std Dev 7276.5 8991.4 8179 1990 N 58 58 Mean 26117 17610 8507 3.11* 0.0024 Std Dev 13883 15515 14721 1997 N 58 58 Mean 45184 28130 17054 4.98** <.0001 Std Dev 15158 21191 18423 High-tech group and Twin group are not different in 1951,1962 High-tech group and Twin group are different in 1970,1980,1990 at 1% level significance High-tech group and Twin group are different in 1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Table 5.24 t-test Results: High-Tech Service Personal Wage Income * p<0.01 _______________________________________________________________________________ ** p<0.001 Year index High-tech (1) Twin (2) Diff (1-2) t Value Pr > |t| 1951 N 58 58 Mean 2442.9 3433.1 -990.2 -0.75 0.455 Std Dev 1701 9883 7091.1 1962 N 58 58 Mean 4841.9 3689.2 1152.7 2.18 0.0312 Std Dev 2884.8 2806.4 2845.9 1970 N 58 58 Mean 7455.7 6189.5 1266.2 1.65 0.101 Std Dev 4104.3 4144.1 4124.2 1980 N 58 58 Mean 18437 13649 4787.5 3.48** 0.0007 Std Dev 6399.1 8306 7414.1 1990 N 58 58 Mean 35840 27529 8311.4 4.17** <.0001 Std Dev 10095 11324 10727 1997 N 58 58 Mean 53431 38644 14787 6.6** <.0001 Std Dev 12477 11623 12058 High-tech group and Twin group are not different in 1951,1962,1970 High-tech group and Twin group are different in 1980,1990,1997 at 0.1% level significance Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Actual Growth Rates Table 5.25, Table 5.26, and Figures 5.13, 5.14, and 5.15 demonstrate general trends of high-tech personal income growth in the given periods. In Figures 5.13, 5.14, and 5.15, we find changes in slopes of actual growth (A-B) between 1951-1962, and 1970-1980 in high-tech personal wage income. In high-tech manufacturing personal income, a change in slope begins in 1962-1970 period. A change is slope in high- tech service personal wage income is during 1970-1980. Table 5.25 Total High-Tech Personal Wage Income, (Annual, $) Year 1951 1962 1970 1980 1990 1997 High-tech (A) 144509.4 272259.4 439163.0 1038225.9 2025453.0 3207734.9 Twin (B) 126427.8 219767.7 342625.7 780011.1 1574148.9 2239993.1 (A)-(B) = (C) 18081.6 52491.7 96537.3 258214.8 451304.1 967741.8 (C)/(A) 12.5% 19.3% 22.0% 24.9% 22.3% 30.2% Year 1951 1962 1970 1980 1990 1997 High-tech Manufacturing (A) 104592.7 191459.5 331139.0 826660.0 1514761.1 2620691.7 Twin Manufacturing (B) 79206.4 139466.3 188746.3 548652.6 1021353.7 1631539.4 (A)-(B) = (C) 25386.3 51993.2 142392.7 278007.5 493407.4 989152.3 (C)/(A) 24.3% 27.2% 43.0% 33.6% 32.6% 37.7% Year 1951 1962 1970 1980 1990 1997 High-tech Services (A) 141686.1 280832.0 432430.6 1069343.3 2078738.3 3099018.0 Twin Sen/ices (B) 199116.9 213975.7 358992.6 791668.2 1596676.1 2241353.8 (A)-(B) = (C) -57430.7 66856.3 73438.1 277675.1 482062.2 857664.2 (C)/(A) -40.5% 23.8% 17.0% 26.0% 23.2% 27.7% Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.26 Average High-Tech Personal Wage Income, (Annual, $) Year 1951 1962 1970 1980 1990 1997 High-tech (A) 2491.5 4694.1 7571.8 17900.4 34921.6 55305.8 Twin (B) 2179.8 3789.1 5907.3 13448.5 27140.5 38620.6 (A)-(B) = (C) 311.8 905.0 1664.4 4452.0 7781.1 16685.2 (C)/(A) 12.5% 19.3% 22.0% 24.9% 22.3% 30.2% Year 1951 1962 1970 1980 1990 1997 High-tech Manufacturing (A) 1803.3 3301.0 5709.3 14252.8 26116.6 45184.3 Twin Manufacturing (B) 1365.6 2404.6 3254.2 9459.5 17609.5 28130.0 (A)-(B) = (C) 437.7 896.4 2455.0 4793.2 8507.0 17054.3 (C)/(A) 24.3% 27.2% 43.0% 33.6% 32.6% 37.7% Year 1951 1962 1970 1980 1990 1997 High-tech Services (A) 2442.9 4841.9 7455.7 18437.0 35840.3 53431.3 Twin Services (B) 3433.0 3689.2 6189.5 13649.5 27528.9 38644.0 (A)-(B) = (C) -990.2 1152.7 1266.2 4787.5 8311.4 14787.3 (C)/(A) -40.5% 23.8% 17.0% 26.0% 23.2% 27.7% Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.13 Average High-Tech Personal Wage Income Trend, 1951-1997 60000.0 50000.0 40000.0 30000.0 £ 20000.0 2 10000.0 1951 1962 1970 1980 1990 1997 Year Source: County Business Patterns, Census Bureau (1951-1997) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.14 Average High-Tech Manufacturing Personal Wage Income Trend, 1951- 1997 60000.0 50000.0 High-tech Manufacturing 40000.0 □ ) Twin Manufacturing (B) 30000.0 (A)-(B) = (C) 20000.0 10000.0 1951 1962 1970 1980 1990 1997 Year Source: County Business Patterns, Census Bureau (1951-1997) 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.15 Average High-Tech Service Personal Wage Income Trend, 1951-1997 60000.0 50000.0 40000.0 30000.0 20000.0 0.0 1951 1962 1970 I960 1990 1997 Year - High-tech Services (A) •Twin Services (B) • (A)-(B) = (C) Source: County Business Patterns, Census Bureau (1951-1997) Table 5.27 and Figures 5.16, 5.17, 5.18 match t-test and personal wage income trends. Actual growth of high-tech personal wage incomes show the highest values between 1951-1962, but the second highest value is 1970-1980 in high-tech personal wage income. In high-tech service personal wage income, the highest value is between 1970-1980, but high-tech manufacturing personal wage income has its highest value in the 1962-1970 period. 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.27 Actual High-Tech Personal Wage Income Growth Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 Average High-tech growth 88.4% 61.3% 136.4% 95.1% 58.4% Average Twin growth 73.8% 55.9% 127.7% 101.8% 42.3% Actual high-tech growth 190.3% 83.9% 167.5% 74.8% 114.4% Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 Average High-tech Manufacturing growth 83.1% 73.0% 149.6% 83.2% 73.0% Average Twin Manufacturing growth 76.1% 35.3% 190.7% 86.2% 59.7% Actual high-tech manufacturing growth 104.8% 173.9% 95.2% 77.5% 100.5% Year 1951-1962 1962-1970 1970-1980 1980-1990 1990-1997 Average High-tech Sen/ices growth 98.2% 54.0% 147.3% 94.4% 49.1% Average Twin Services growth 7.5% 67.8% 120.5% 101.7% 40.4% Actual high-tech sen/ices growth 216.4% 9.8% 278.1% 73.6% 77.9% Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.16 Average High-Tech Personal Wage Income Growth Rates 200.0% 180.0% 160.0% 140.0% Average High-tech growth A verage Twin growth 120.0% 100.0% 80.0% Actual high-tech growth C l 60.0% 40.0% 20.0% 0.0% 1951- 1962- 1970- 1980- 1990- 1962 1970 1980 1990 1997 Period Source: County Business Patterns, Census Bureau (1951-1997) 1 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.17 Average High-Tech Manufacturing Personal Wage Growth Rates ♦ A v e r a g e H igh-tech M an u fa c tu rin g g ro w th —• — A v e r a g e Tw in M an u fa c tu rin g g ro w th —A — A c tu a l h ig h -tech m a n u fa c tu rin g g ro w th 1 9 5 1 - 1 9 6 2 - 1 9 7 0 - 1 9 8 0 - 1 9 9 0 - 1 9 6 2 1 9 7 0 1 9 8 0 1 9 9 0 1997 Period Source: County Business Patterns, Census Bureau (1951-1997) Figure 5.18 Average High-Tech Service Personal Wage Income Growth Rates — Average High-tech Services growth ■ Average Tw in Services growth — 6 — Actual high-tech services growth 1951- 1962- 1970- 1980- 1990- 1962 1970 1980 1990 1997 Period Source: County Business Patterns, Census Bureau (1951-1997) 2 0 0 .0 % 1 5 0 .0 % 1 00 .0 % 5 0 . 0 % Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.3 Significant Location Factors Location factors are tested by two statistical tests (a difference of means test and regression tests) among selected 28 location factors culled from the County and City Data Book (Census Bureau). T-tests will screen candidate location factors that have the same trend with high-tech employment, wage income, and personal wage income in the 1970-1980 period when the high-tech centers were shaped in the U.S. Location factors that have no difference in 1970, but have difference in 1980 in t-test will be the candidate location factors. Regression analysis will corroborate a decision on the significant location factors and their contributions to high-tech centers in the United States. 5.3.1 Categories of Location Factors in the Study As many as twenty-eight location factors are sorted into several categories. Population and population density imply the urbanization economies; college enrollment and rate of bachelor or higher degree explain the local educated labor pool and universities effects; health services includes physicians (rate), hospitals (rate) and local government expenditure rate to health and hospital; income effect includes median family income, per capita money income; housing effects and affordability includes recently built house rate, vacant house rate, median house value, and median rent; and bank deposit is used as proxy o f local financial conditions. Local government’s roles are also reflected on local governments’ revenue and expenditure; total expenditure, per capita expenditure, and that on education, highway, and public welfare respectively. Tax effects are considered, and 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the intergovernmental revenue rate is used as an index of federal or state support. Table 5.28 demonstrates these categories with their ultimate impacts on business, quality of life, or both. Table 5. 28 Category of Location Factors General Category Specific Category Location Factor university and higher education college enrollm ent bachelors o r m ore rate B usiness Climate local bank financial condition bank deposits T ransportation highway expenditure rate federal and state support intergovernm ental revenue rate physicians, physicians rate, Health services hospitals, hospitals rate, health and hospitals expenditure rate crim es, Crime crim es rate, police officers, incom e (poverty) m edian family incom e Quality of life per capita m oney incom e recently built h o u ses rate housing affordability (cost of living) V acant h o u ses rate Median house value Median rent Public education education expenditure rate social service (welfare) public welfare expenditure rate local governm ent revenue local governm ent's service level local expenditures per capita local expenditure urbanization econom ies Population Overlapping factors Population density Taxation total tax percent per capita property tax Source: County and City Data Book (Census Bureau), 1970 and 1980 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.3.2 Descriptive Statistics and T-test Result Tables 5.29 and 5.30 include descriptive statistics for high-tech location factors through the high-tech formation period (1970-1980). Location factors that follow the same trend with that of high-tech employment (wage, or personal wage) have a higher probability of effective location factors. Candidate location factors should not be different between the high-tech group and the twin group in 1970, but should be significantly different between them in 1980. Twenty-eight variables are compared in 1970, but only twenty of them are compared in 1980 because eight variables are eliminated in 1970. The eliminated eight variables are the proportion of respondents with a bachelor degree or higher, median family income, per capita money income, recently built houses rate, median house value, median rent, total tax percent, and per capita property tax. They are excluded in the 1980 descriptive statistics because they are significantly different between the two groups of counties at the 1 percent level significance t-test in 1970. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.29 t-test Results: High-Tech Location Factors in 1970 * p<0.05 ** p<0.01 *** p<0.001 No Location Factors index High-tech (1) Twin (2) Diff (1-2) T Value P r> |t| 1 Population N 58 58 Mean 503240 494406 8834.5 0.1 0.9241 Std Dev 456571 536624 498208 2 Population Density N 58 58 Mean 2985.9 2014.8 971.16 0.7 0.4843 Std Dev 9188 5146 7446.5 3 Physicians, 1975 N 58 58 Mean 1452.1 1064 388.09 1.3 0.1964 Std Dev 1923.7 1208.3 1606.3 4 Physicians rate, 1975 N 58 58 Mean 235.15 182.86 52.291 1.89 0.0609 Std Dev 168.49 125.68 148.64 5 Hospitals, 1975 N 58 58 Mean 24.569 11.534 13.034 1.16 0.2498 Std Dev 84.679 11.107 60.39 6 Hospitals rate, 1975 N 58 58 Mean 623.61 886.79 -263.2 -1.56 0.1244 Std Dev 408.15 1222.2 911.16 7 Crimes, 1975 N 58 56 Mean 40252 29779 10473 0.92 0.3615 Std Dev 78519 36632 61617 8 Crime rate, 1975 N 58 56 Mean 5945 5269.1 675.91 0.89 0.3773 Std Dev 5206.3 2508.5 4109.1 9 Police officers N 38 34 Mean 1799.4 834.26 965.13 1.23 0.2251 Std Dev 4714.9 990.96 3494.7 10 College enrollment N 58 58 Mean 24353 17916 6437 1.36 0.1765 Std Dev 31106 18109 25451 11 Bachelors or more rate N 58 58 Mean 16.566 11.134 5.431 5.28*** <.0001 Std Dev 6.3575 4.5834 5.5419 12 Median family income N 58 58 Mean 11316 9925.8 1390.1 3.89*** 0.0002 Std Dev 2092.3 1741.7 1925 13 Per capita money income N 58 58 Mean 3676.5 3121.1 555.43 4.35*** <.0001 Std Dev 780.07 579.57 687.18 14 Recently built houses rate N 58 58 Mean 30.94 23.024 7.9155 4.02*** 0.0001 Std Dev 11.384 9.742 10.595 15 Vacant houses rate N 58 58 Mean 1.219 1.0552 0.1638 1.55 0.1243 Std Dev 0.6039 0.5332 0.5696 Continue in the next page Source: County and City Data Book, Census Bureau (1970) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.29 t-test Results: High-Tech Location Factors in 1970, continued * p<0.05 ** p<0.01 *** p<0.001 No Location Factors Index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 16 Median houses value N 58 58 Mean 22967 16947 6020.5 4.58*** <.0001 Std Dev 8264.9 5662.5 7084.2 17 Median rent N 58 58 Mean 125.93 106.34 19.586 3.8*** 0.0002 Std Dev 29.02 26.39 27.736 18 Bank deposits N 58 58 Mean 2659.4 1304.3 1355.1 1.08 0.2863 Std Dev 9401.1 1923.9 6785.4 19 Local government revenue N 58 56 Mean 229.95 138.34 91.613 1.06 0.2947 Std Dev 636 175.07 470.01 20 Intergovernmental Revenue rate N 58 56 Mean 30.202 36.809 -6.607 -2.88** 0.0049 Std Dev 9.9583 14.137 12.191 21 Total tax rate N 58 56 Mean 54.857 48.288 6.5694 2.93** 0.0041 Std Dev 11.936 11.968 11.951 22 Per capita property tax rate N 58 56 Mean 154.47 124.45 30.019 2.9** 0.0046 Std Dev 58.908 51.645 55.46 23 Local expenditure N 58 56 Mean 225.05 143.74 81.305 1.04 0.303 Std Dev 568.47 177.87 424.26 24 Per capital local expenditure N 58 56 Mean 257.14 236.63 20.513 1.42 0.1598 Std Dev 82.166 72.425 77.535 25 Education expenditure rate N 58 56 Mean 52.638 52.307 0.3308 0.15 0.8811 Std Dev 12.255 11.29 11.791 26 Highway expenditure rate N 58 56 Mean 6.4672 8.4482 -1.981 -2.41* 0.0183 Std Dev 2.5962 5.593 4.335 27 Public welfare expenditure rate N 58 56 Mean 5.6612 5.1741 0.4871 0.51 0.608 Std Dev 5.4842 4.6 5.0693 28 Health and hospitals expenditure rate N 58 56 Mean 3.894 4.4143 -0.52 -0.74 0.4639 Std Dev 3.1703 4.2825 3.7578 High-tech group and Twin group are different in No. 26 at 5% level significance High-tech group and Twin group are different in No. 20. 21, and 22 at 1% level significance High-tech group and Twin group are different in No. 11-14, 16, and 17 at 0.1% level significance Source: County and City Data Book, Census Bureau (1970) In Table 5.29, eight variables (No. 11,12,13,14,16,17,21, and 22) are excluded from candidates of substantial location factors because they are already different in 1970 when high-tech employment and wage data are not different in high-tech group and twin group. Intergovernmental revenue rate (No.20) and highway expenditure rate (No.26) can be included because their T values are negative signs that imply that twin group is larger in these values in 1970. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.30 t-test Results: High-Tech Location Factors in 1980 * p<0.05 ** p<0.01 * ** p<0.001 No. Location Factors Index High-tech (1) Twin (2) Diff (1-2) t Value P r > |t| 1 Population N 58 58 Mean 546053 490429 55624 0.62 0.5378 Std Dev 481196 488104 484662 2 Population density N 58 58 Mean 2866.4 1382.5 1484 1.26 0.2106 Std Dev 8729.1 1909.3 6318.3 3 Physicians N 58 57 Mean 1603.9 1136 467.98 1.59 0.116 Std Dev 1913.2 1167.1 1587.9 4 Physicians rate N 58 57 Mean 254.23 188.31 65.924 2.38* 0.0196 Std Dev 183.45 103.91 149.42 5 Hospitals N 58 55 Mean 13.345 11.364 1.9812 0.93 0.3554 Std Dev 12.802 9.757 11.422 6 Hospitals rate N 58 55 Mean 669.04 766.03 -96.99 -0.94 0.351 Std Dev 583.47 516.77 552.03 7 Crimes N 58 56 Mean 47072 32426 14645 1.07 0.2893 Std Dev 97519 36902 74220 8 Crimes rate N 58 56 Mean 6004.4 5705.4 298.97 0.6 0.5509 Std Dev 2582.2 2746.9 2664.3 9 Vacant house rate N 58 58 Mean 5.7948 5.8741 -0.079 -0.17 0.8633 Std Dev 2.5206 2.43 2.4757 10 College enrollment N 58 58 Mean 56987 54126 2860.6 0.18 0.8579 Std Dev 70015 99150 85828 11 Local government revenue N 58 56 Mean 722823 414214 308609 1.17 0.2449 Std Dev 1940000 498567 1430000 12 Intergovernmental revenue rate N 58 56 Mean 76.507 78.995 -2.488 -1.18 0.2417 Std Dev 11.629 10.935 11.294 13 Local expenditure N 58 56 Mean 647232 406189 241044 1.11 0.2689 Std Dev 1570000 476075 1170000 14 Per capita local expenditure N 58 56 Mean 830.48 783.41 47.07 0.97 0.3342 Std Dev 261.68 256.49 259.14 15 Education expenditure rate N 58 56 Mean 46.888 46.938 -0.05 -0.02 0.9806 Std Dev 10.311 11.306 10.811 16 Highway expenditure rate N 58 56 Mean 4.7069 6.0563 -1.349 -2.42* 0.0175 Std Dev 2.0268 3.6613 2.9451 17 Public welfare expenditure rate N 58 54 Mean 4.481 6.2 -1.719 -1.63 0.1066 Std Dev 5.4268 5.7324 5.5761 18 Health and hospitals expenditure rate N 58 56 Mean 5.4836 6.0732 -0.59 -0.56 0.5738 Std Dev 4.8683 6.1846 5.5538 19 Police expenditure rate N 58 56 Mean 4.9914 4.8018 0.1896 0.61 0.5455 Std Dev 1.5225 1.7986 1.6638 20 Bank deposits N 58 58 Mean 5047.3 2436.2 2611.2 1.54 0.1273 Std Dev 12528 2952.6 9101.5 High-tech group and Twin group are different in physicians rate (+), and local government highway expenditure rate (-) at 5% level significance_________________________________ Source: County and City Data Book, Census Bureau (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Physician rate (No.4) and highway expenditure rate (No. 16) show a difference between the high-tech group and the twin group at the 5 percent level significance in 1980. The physician rate is not different in the 1975 data, so it can be selected a substantial location factor for high-tech centers in the United States. However, highway expenditure rate shows negative t-value in both 1970 and 1980 at the 5 percent level significance. Thus it does not satisfy the conditions of this analysis. Among twenty-eight location factors, only the physician rate satisfies the requirements for the substantial location factor for the formation and growth of high- tech center in the United States. The “physician rate” is used as a proxy for quality of life, including local health services, but it can be considered indicating educated and wealthy class without significant changes of other health service variables such as number of physicians, number of hospitals, rate of hospitals, and local governments’ expenditure rate for health and hospitals7. 5.3.3 Regression Results Correlations and simple regression are estimated between the physician rate, and high-tech employment (total wage income, and personal income) with high-tech manufacturing and high-tech services data in 1980 to examine the fitness of the location factor to high-tech growth. Table 5.31, Table 5.32, and Table 5.33 show the 7 It can be argued that the presence of medical centers jointly explains high-tech growth as well as the presence of large numbers of physicians. To control for this effect, a sub-sample of counties, that without a major medical center, and their twin counties were studied. The subset consisted of 41 counties. No substantial differences between the subset results and the results reported in this chapter were found. 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. descriptive statistics of the variables, the result of correlation analysis, and the result of regression models, each. Table 5.31 Descriptive Statistics of Variables (1980) No. Variable Unit N Mean Std Dev Sum Minimum Maximum x1 High-tech manufacturing personal wage income Dollars 58 14253 7276 826660 0 26326 x2 High-tech service personal wage income Dollars 58 18437 6399 1069343 0 30297 x3 High-tech personal wage income Dollars 58 17900 5655 1038226 0 25729 x4 High-tech manufacturing total wage income $1,000 58 222149 523687 12884627 0 3307364 x5 High-tech service total wage income $1,000 58 192405 370893 11159502 0 2356100 x6 High-tech total wage income $1,000 58 414554 716045 24044129 0 3763218 x7 High-tech manufacturing employment People 58 15735 27700 912612 0 178257 x8 High-tech service employment People 58 11835 16855 686416 0 105545 x9 High-tech employment People 58 27569 36056 1599028 0 205287 X10 Physicians rate Per 100,000 58 254.2 183.5 14746 40 871.8 Source: County Business Patterns, County anc City Data Boo k (1980) Table 5. 32 Result of Correlation Analysis * p<0.05 ** p<0.01 ***p<0.001 Variable X1 x2 x3 x4 x5 x6 x7 x8 x9 x10 High-tech manufacturing personal wage income 1.000 High-tech service personal wage income 0.449 1.000 High-tech personal wage income 0.536 0.861 1.000 High-tech manufacturing total wage income 0.371 0.210 0.216 1.000 High-tech service total wage income 0.256 0.277 0.320 0.260 1.000 High-tech total wage income 0.404 0.297 0.324 0.866 0.708 1.000 High-tech manufacturing employment 0.375 0.224 0.213 0.968 0.219 0.821 1.000 High-tech service employment 0.325 0.308 0.343 0.296 0.962 0.714 0.266 1.000 High-tech employment 0.440 0.316 0.324 0.882 0.617 0.965 0.893 0.672 1.000 Physicians rate 0.035 0.318* 0.303* -0.009 0.475*** 0.240 -0.041 0 504*” 0.204 1.000 xl-xlO follows the number of variable in the descriptive statistics in the past page Source: County Business Patterns, County and City Data Book (1980) 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In the correlation analysis, physicians rate has positive (+) relationship with high-tech service personal wage income (0.318, p<0.05), high-tech personal wage income (0.303, p<0.05), high-tech service total wage income (0.475, p<0.001), and high-tech service employment (0.504, p<0.001) in 1980. Thus, physician rate demonstrates higher relationship with high-tech service activity. In the regression analysis, the model is = A + P\X, + s in this model, ^ is the high-tech employment, total wage income, and personal wage income in 1980, which follows the number of variable in the descriptive statistics in the previous page y (from xl through xlO). And model numbers follow this rule; < in model 1 is high- y tech manufacturing personal income in 1980, > in model 9 is high-tech employment in 1980. is the physician rate in 1980, A is Y-intercepts, A is slopes, and s is error. Table 5.33 summarizes model 1 through model 9. According to the hypothesis for goodness of fit in chapter 4, physicians rate is not zero in model 2 (p<0.05), model 3 (p<0.05), model 5 (p<0.001), and model 8 (pO.001). In other models, A (physicians rate) can be considered 0. R2 values of those models are not too low. R values of model 2 are 0.1013 (10.13%), model 3 is 0.092 (9.2%), model 5 is 0.226 (22.6%), and model 8 is 0.2537 (25.37%). High-tech service total wage income (Model 5) and high-tech service employment (Model 8) are relatively well predicted by the rate of physician than high-tech services personal wage income (Model 2) and high-tech personal wage income (model 3). Statistical prediction functions for these two models are as follows. 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (1) High-tech service personal wage income = (11.1 x physician rate) +15614, with 10.23 percent R value at the 5 percent level significance in goodness of fit (1980). (2) High-tech personal wage income = (9.3 x physician rate) + 15524, with 9.2 percent R value at the 5 percent level significance in goodness of fit (1980) (3) High-tech service regional wage income = (961.1 x physician rate) - 51994, with 22.6 percent R2 value at the 0.1 percent level significance in goodness of fit (1980). (4) High-tech service regional employment = (46.3 x physician rate) + 68.3, with 25.37 percent R2 value at the 0.1 percent level significance in goodness of fit (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.33 Result of Regression Analysis, continued Model 1 1 I I Dependent Variable: High-tech manufacturing personal wage income Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 3699200 3699200 0.07 0.7942 Error 56 3014299265 53826773 Corrected Total 57 3017998464 Root MSE 7336.67313 R-Square 0.0012 Dependent Mean 14253 Adj R-Sq -0.0166 CoeffVar 51.47545 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 13900 1655.79464 8.39 <.0001 physician rate 1 1.38866 5.29712 0.26 0.7942 Model 2 | | | Dependent Variable: High-tech service personal wage income Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 236486945 236486945 6.31 0.0149 Error 56 2097569697 37456602 Corrected Total 57 2334056642 Root MSE 6120.17988 R-Square 0.1013 Dependent Mean 18437 Adj R-Sq 0.0853 Coeff Var 33.19518 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 15614 1381.24745 11.3 <.0001 Physician rate 1 11.10311 4.41881 2.51 0.0149 Source: County Business Patterns, County and City Data Book (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.33 Result of Regression Analysis, continued Model 3 | | Dependent Variable: High-tech personal wage income Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 167631538 167631538 5.67 0.0207 Error 56 1655234738 29557763 Corrected Total 57 1822866276 Root MSE 5436.70518 R-Square 0.092 Dependent Mean 17900 Adj R-Sq 0.0757 CoeffVar 30.37189 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value Pr > |t| Intercept 1 15524 1226.99583 12.65 <.0001 Physician rate 1 9.348 3.92534 2.38 0.0207 Model 4 | | | Dependent Variable: High-tech manufacturing total wage income Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 1266165960 1266165960 0 0.9465 Error 56 1.56309E+13 2.79123E+11 Corrected Total 57 1.56321E+13 Root MSE 528320 R-Square 0.0001 Dependent Mean 222149 Adi R-Sq -0.0178 CoeffVar 237.82283 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 228680 119235 1.92 0.0602 Physician rate 1 -25.69132 381.45073 -0.07 0.9465 Source: County Business Patterns, County and City Data Book (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.33 Result of Regression Analysis, continued Model 5 | | Dependent Variable: High-tech service total wage income Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 1.77205E+12 1.77205E+12 16.35 0.0002 Error 56 6.06896E+12 1.08374E+11 Corrected Total 57 7.84101E+12 Root MSE 329203 R-Square 0.226 Dependent Mean 192405 Adj R-Sq 0.2122 CoeffVar 171.09854 Parameter Estimates Param eter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 -51944 74297 -0.7 0.4874 Physician rate 1 961.12388 237.68631 4.04 0.0002 Model 6 | | Dependent Variable: High-tech total wage income Analysis of Variance Source DF Sum of Squares Mean Square F Value P r> F Model 1 1.67858E+12 1.67858E+12 3.41 0.07 Error 56 2.75465E+13 4.91902E+11 Corrected Total 57 2.92251 E+13 Root MSE 701357 R-Square 0.0574 Dependent Mean 414554 Adj R-Sq 0.0406 CoeffVar 169.18359 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 176736 158287 1.12 0.269 Physician rate 1 935.43256 506.38442 1.85 0.07 Source: County Business Patterns, County and City Data Book (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.33 Result of Regression Analysis, continued Model 7 | | Dependent Variable: High-tech manufacturing employment Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 72388631 72388631 0.09 0.7617 Error 56 43664424890 779721873 Corrected Total 57 43736813520 Root MSE 27924 R-Square 0.0017 Dependent Mean 15735 Adj R-Sq -0.0162 CoeffVar 177.46458 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 17296 6301.98205 2.74 0.0081 Physician rate 1 -6.14294 20.16095 -0.3 0.7617 Model 8 | | Dependent Variable: High-tech service employment Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 4109099547 4109099547 19.04 <.0001 Error 56 12084835931 215800642 Corrected Total 57 16193935479 Root MSE 14690 R-Square 0.2537 Dependent Mean 11835 Adj R-Sq 0.2404 CoeffVar 124.1272 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 68.29696 3315.38271 0.02 0.9836 Physician rate 1 46.28224 10.60639 4.36 <.0001 Source: County Business Patterns, County and City Data Book (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.33 Result of Regression Analysis, continued Model 9 | | Dependent Variable: High-tech employment Analysis of Variance Source DF Sum of Squares Mean Square F Value Pr > F Model 1 3090704808 3090704808 2.44 0.1241 Error 56 71011293725 1268058817 Corrected Total 57 74101998532 Root MSE 35610 R-Square 0.0417 Dependent Mean 27569 Adj R-Sq 0.0246 Coeff Var 129.16404 Parameter Estimates Parameter Standard Variable DF Estimate Error t Value P r> |t| Intercept 1 17365 8036.68603 2.16 0.035 Physician rate 1 40.1393 25.71051 1.56 0.1241 Source: County Business Patterns, County and City Data Book (1980) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 6 CONCLUSIONS AND POLICY IMPLICATIONS 6.1 Conclusions My research shows three major findings on the formation and location factors of high-tech centers in the United States. The explanations of these three findings can provide answers to the research hypotheses in Chapter Three. First, the high-tech centers in the United States are substantially established during the 1970-1980 period. Many researchers claim that high-tech industries began from the middle 1940s, or early years of the 1950s (Saxenian 1981,1985, Castells 1985, Scott and Drayse 1993, Hall 1998). It seems to be true that high-tech industries had their origin from these periods (1940s or 1950s); however, the industrial growth of high-tech industries into high-tech centers took place during the 1970-1980 period. Supporters for the rise of high-tech centers (industries) in the 1940s-1950s base their ideas on the role of government spending and military contracts released during this period. However, this research shows that the rise of high-tech centers in the U.S. started about 1970-1980 period accompanied by local accumulated educated and wealthy workforce. Historical trends on high-tech employment and high-tech total wage income support this argument. Historical trends on high-tech personal wage income, however, show the second largest number during the 1970-1980 period with slight difference from the top period. Aggregating these three trends, I select the 1970-1980 period as the substantial formation period 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. for high-tech centers in the United States. An interesting connection is that this result matches the historical trend of American science parks. Luger and Goldstein say that American local governments began their active development of “science parks” from the early 1970s to the middle 1980s (Luger and Goldstein 1991). They find that only 12 science parks were established during 1950-1970 period in the United States (0.6 science parks/year); but, there are 33 science parks founded during 1973-1984 period (3 science parks/year) (Luger and Goldstein 1991). Further research is worthwhile concerning the relationship between high-tech centers and science parks. These findings do not mean that there is a widely implemented high-tech policy in the United States during the 1970-1980 period. It is likely that most high- tech centers (counties) thrived without specific government policies designed for high-tech development. High-tech centers do have some spontaneous formation. Most location factors are not found to be effective for the formation of high-tech centers during the period. These location factors seem to work together to build up high-tech centers, rather than one location factor or one amenity factor leads all high- tech center development. The various location factors work in time or phase sensitive patterns, in which each factor has its own functional period. Second, high-tech services play a leading role in the formation of high-tech center in the United States. Compared to high-tech services’ contribution to high- tech center formation, high-tech manufacturing’ role is not important because high- tech manufacturing is shrinking and decentralizing to nonmetropolitan areas. This finding supports and applies the previous studies that claim the general trend of 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. manufacturing decentralization into nonmetropolitan areas (Norton and Rees 1979, Barkley 1988, 1993), onto high-tech manufacturing sectors. High-tech services activity plays more critical roles in total high-tech trends in employment, wage income, and personal wage income. In t-tests, high-tech manufacture has no difference of employment between the high-tech group and the twin group in the whole period at the 1 percent level significance, but high-tech service employment begins to be different in 1980 at the 1 percent level significance. High-tech wage income has a similar trend with high-tech employment in manufacturing and services in t-tests. High-tech personal wage income has a little bit different result. High-tech manufacturing personal wage income begins to be different in both groups in 1960 at the 1 percent level significance. High-tech service personal wage income begins to be different between the groups in 1980 at the 0.1 percent level significance. In the actual growth rates, high-tech manufacturing is different from high-tech service activity. High-tech manufacturing shows a mounting trend that keeps increasing until 1970-1980 period (employment, wage income) or until 1960-1970 period (personal wage income) and then declining trend. While, high-tech service shows a zigzag style in actual growth rate, usually their peaks are in 1970-1980 period. Third, I find that roles of many high-tech location factors are not significantly important in building up high-tech centers in the United States. This trend of insufficient high-tech location factor implies that high-tech industries and high-tech centers have a tendency to be established spontaneously without a specific planning or policy during the period of formation. In addition, it is because the quasi- experimental control group method can consider the effect of control group, and 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reflect it to the analysis. Among twenty-eight selected location factors, only one factor (physician rate) satisfies requirements for substantial high-tech location factor in the formation and growth of high-tech centers in the United States (1970-1980 period). In a simple regression analysis, this factor shows about 9.2-25.37 percent R2 values. And this factor (physician rate as a proxy for quality of life) has relationships with (1) high-tech personal wage income, (2) high-tech service personal wage income, (3) high-tech service wage income, and (4) high-tech service employment (See Tables 5.31 and 5.32). Its four fitting dependent variables, three are high-tech service activities, including employment, wage income, personal wage income. The location factor (physician rate) works for the inspiration of high-tech service activity in metropolitan areas, which leads to substantial growth of high-tech centers. I try to be careful in approaching the location factor (physician rate), because variables such as the numbers of physicians, hospitals, and hospital bed rate during the period 1970-1980 could explain the presence of high-tech centers as well as the large number of physicians. Yet, controlling for this variable did not substantially change my findings. A significant change in physician rate without any meaningful increases of other medical services may not mean only medical service improvements, but its purpose may have a significant increase of a certain class in the region and its rate increases. Physicians are a good representation o f a relatively highly waged and educated population. They have enough assets to build up or move into higher quality residential communities such as gated community, and suburban housings. Those residential communities play a role in attracting high-tech 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. workers and professionals into the region. From t-tests of location factors in 1970,1 also find that the high-tech group is already different from the twin group in wealth and property factors such as per capita money income, median family income, recently built house rate, median house value, and per capita property tax rate (Table 5.30). The highly waged and educated class in this paper includes the highest income group such as medical doctors, lawyers, and CEO. As Florida (2002) mentions, this class may represent “creative professionals”, a part of the Creative Class, which includes “a wide range of knowledge-intensive industries such as high- tech sectors, financial services, the legal and health care professions, and business management” (Florida 2002: 69). However, there are counter-reasons of the role of educated and affluent Creative Class for building up high-tech centers in the United States. First, educated and affluent people tend to be more “footloose” than less educated and poor classes. They do not stay twenty years in one place only to build up a remarkable high-tech center; they are more likely to look for an established high-tech center that is well equipped with amenities and campus airs. Second, mobility of their money is faster than their migration to other regions. There might be a mismatch between high-tech centers and educated wealthy class because such people can invest for high-tech businesses and high-tech centers although they live far from where high-tech businesses and high-tech centers locate. Beyond these three findings and results, I find the role o f universities as a unique foundation for high-tech formation. This finding favors some of the previous literature (Nelson 1986,1996, Lund 1986, Sivitanidou and Sivitnides 1995, Gottlieb 2001) and entrepreneur ship/seedbed theory (Luger and Goldstein 1991); but rejects 136 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the opposite studies (Markusen, Hall, and Glasmeier 1986, Florax and Folmer 1992, Bania, Eberts, and Fogarty 1993). In the pre-test, I select twin counties of which conditions of education, professional employment, and geography are most similar to high-tech county group members in 1950. Thus, educated people are very similar in both groups in 1950. The 1970 data, however, shows different values in both groups. In the t-test, “the percentage of bachelors or higher degrees” is different between the high-tech group and the twin group in 1970 at the 0.1 percent level significance. Educated human capital starts working for the formation of high-tech center throughout more than 20 year’s accumulation. This research implicitly emphasizes the role of research universities and educated human capital as infrastructure for high-tech prosperity. Through Chapter six, I find the role of university medical centers as a hub for making educated and affluent class who can build up high-tech centers by higher quality of life. However, it needs further inspections to have a rationale. 6.2 Policy Implications High-tech centers are not easily established by specific planning, policy, or government support. Such efforts can make a high-tech center in a region or area, but a high-tech center by only exogenous inputs may have difficulty in maintaining long prosperity. High-tech centers are spontaneously established where there are abundant educated labor pools accumulated from 1950s to 1970s. Long time- committed support to education, especially research universities and higher education are important in this beginning stage of high-tech formation. Commitment 137 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to higher education is an infrastructure provision, and will build a strong foundation, which yield a high-tech center in the beginning. However, it has been argued that without proper subsequent actions, those educated workers will depart to other states or other regions. My research provides guidelines for the after-actions for high-tech flowering. Creating and attracting a highly waged and educated labor pool, or the Creative Class into the region is the basis for stable high-tech prosperity. According to the research results and interpretation, this class has a higher probability to attract high-tech workers and high-tech firms via residential improvement. Local governments and communities that want high-tech prosperity should seek a long term strategy for providing the creative-class workers with open and tolerant environments (Florida 2002), supplying them with various dimensions (or options) of urban life of quality (Herzog, Schlottmann, and Johnson 1986, Costa and Kahn 1999), or buying the class with assistance of local research universities (Gottlieb 2001). At the same time, however, the local governments and communities should prepare actions for aggravating social segregation, economic inequality, degrading amenities and rising cost of living due to advent of both educated affluent class and high-tech chore workers (Saxenian, 1981; Kotkin, 2000; Kotkin and Siegel, 2000). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. B i b l i o g r a p h y Acs, Z. J. and D. Audretsch. 1988. Innovation in Large and Small Firms: An Empirical Analysis, American Economic Review, v.78: 678-690. Alarcon, R. (1999), Recruitment Processes Among Foreign-Born Engineers and Scientists in Silicon Valley, American Behavioral Scientist, v.42, n.9, June/July 1999: 1381-1397. Anselin, L., A. Varga, and Z. Acs (1997), Local Geographical Spillovers between University Research and High Technology Innovations, Journal o f Urban Economics, v.42: 422-448. Arora, A., Florida, R., Gates, G. J., and Kamlet, M. (2000), Human Capital, Quality o f Place, and Location, Research Paper, Richard King Mellon Foundation. Association of American Medical Colleges (2002), M edical Schools o f the United States -G eographical Listings-, http://www.aamc.org/meded/medschls/start.htm, Association of American Medical Colleges. Atkinson, R.D. (2002), The 2002 State New Economy Index -Benchmarking Economic Transformation in the States, Technology and New Economy Project, Progressive Policy Institute. Audretsch, D.B. and Feldman, M (1996), R&D Spillovers and the Geography of Innovation and Production, American Economic Review, v.86: 630-640. Audretsch, D.B and Stephan, P.E. (1996), Company-Scientist Location Links: The Case of Biotechnology, The American Economic Review, v.86, n.3: 641-651. Bania, N. and Eberts, R. and Fogarty, M.S. (1993), Universities and the startup of new companies: Can we generalize form Route 128 and Silicon Valley?, The Review o f Economics and Statistics, v.75: 761-766. Barkley, D. L. (1988), The Decentralization of High-Technology Manufacturing to Nonmetropolitan Areas, Growth and Change, v.29, n.l, Winter 1988. Barkley, D.L. (1993), Manufacturing Decentralization: Has the Filtering-Down Process Fizzled Out?, in Barkley, D.L. (ed.), Economic Adaptation - Alternatives fo r Nonmetropolitan Areas-, Westview Press, Inc. 139 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Beardsell, M. and Henderson, J.V. (1999), Spatial Evolution of the Computer Industry in the USA, European Economic Review, v.43, n.2 (February 1999): 431-56. Bishop, B. (2000), Technology and Tolerance: Austin Hallmarks, Austin American-States-man, June 25. Bureau of the Census (1952), County and City Data Book 1952, U.S. Department of Commerce. Bureau of the Census (1962), County and City Data Book 1962, U.S. Department of Commerce. Bureau of the Census (1972), County and City Data Book 1972, U.S. Department of Commerce. Bureau of the Census (1977), County and City Data Book 1977, U.S. Department of Commerce. Bureau of the Census (1983), County and City Data Book 1983, U.S. Department of Commerce. Bureau of the Census (1988), County and City Data Book 1988: http://fisher.lib.virginia.edu/ccdb/, U.S. Department of Commerce. Bureau of the Census (1994), County and City Data Book 1994: http://fisher.lib.virginia.edu/ccdb/, U.S. Department of Commerce. Bureau of the Census (1951), County Business Patterns 1951, U.S. Department of Commerce. Bureau of the Census (1962), County Business Patterns 1962, U.S. Department of Commerce. Bureau of the Census (1970), County Business Patterns 1970, U.S. Department of Commerce. Bureau of the Census (1980), County Business Patterns 1980, U.S. Department of Commerce. Bureau of the Census (1987), County Business Patterns 1987: CD ROM, U.S. Department of Commerce. Bureau of the Census (1990), County Business Patterns 1990: CD ROM, U.S. Department of Commerce. 140 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bureau of the Census (1992), County Business Patterns 1992: CD ROM, U.S. Department of Commerce. Bureau of the Census (1997), County Business Patterns 1997: CD ROM, U.S. Department of Commerce. Bureau of the Census (1998), United States Historical Census Data Browser: http://fisher.lib.virginia.edu/census/, U.S. Department of Commerce. Campbell, D.T. (1988), Methodology and Epistemology fo r Social Science, The University of Chicago Press: Chicago and London. Campbell, D. and Stanley, J. (1963), Experimental and Quasi-Experimental Designs fo r Research, Chicago, Rand McNally College Publishing Company. Castells, M. (ed.) (1985), High Technology, Space, And Society, Sage Publication. CB Richard Ellis (1999), National Real Estate Index, CB Richard Ellis. Cook, T.D. and Campbell, D.T. (1979), Quasi-Experimentation, Design & Analysis Issues fo r Field Settings, Rand McNally Publishing Company. Cordes, J.J. and Hertzfeld, H.R. and Vonortas, N.S. (1999), A Survey O f High Technology Firms, United States Small Business Administration. Cortright, J. and Mayer, H. (2000a), A Comparison o f High Technology Centers, Regional Connections Project, Institute for Portland Metropolitan Studies, Portland State University. Cortright, J. and Mayer, H. (2000b), The Ecology o f the Silicon Forest, Regional Connections Project, Institute for Portland Metropolitan Studies, Portland State University. Costa, D.L. and Kahn, M.E. (1999), Power Couples: Changes in the Location Choice o f the College Educated, 1940-1990, National Bureau of Economic Research. Dean (1984), Education and Economic Productivity, Ballinger Publishing Company, Cambridge, Massachusetts. Desrochers, P (2001), Local Diversity, Human Creativity, and Technological Innovation, Growth and Change, v.32, n.3, Summer 2001. 141 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DeVol, R.C. (1999), America's High-Tech Economy, Santa Monica, CA, Milken Institute. DeVol, R.C. and Koepp, R. and Fogelbach, F. (2002), State Technology and Science Index -Com paring and Contrasting California-, Milken Institute. Drucker, P (1998), Management’s New Paradigm, Forbes, 1, October 5: 152- 177. Eakin, E. (2002), The Cities And Their New Elite, Arts & Ideas/Cultural Desk, New York Times, Final, Section B, Page 7, Column 4, June 1, Saturday. Echeverri-Carroll, E.L. and Brennan, W. (1999), Are Innovation Networks Bounded by Proximity?, in Fischer, M.F., Suarez-Villa, L. and Steiner, M. (eds.) (1999), Innovation, Networks and Localities, Advances in Spatial Science series. Heidelberg and New York: Springer. Florax, R and Folmer, H. (1992), Knowledge Impacts of universities on industries: An aggregate simultaneous investment model, Journal o f Regional Science, v.32: 437-466. Florida, R (2001), Worklife: The Great IT Worker Awakening, Information Week, April. Florida, R (2002), The Rise o f the Creative Class -And How I t’ s Transforming Work, Leisure, Community and Everyday Life-, Basic Books Florida, R. and Kennedy, M. (1988), Venture capital, high technology and regional development, Regional Studies v.22: 33-48. Glaeser, E. L. (1997), Learning in Cities, Working paper series 6271, National Bureau of Economic Research, Inc. Glasmeier, A., Markusen, A., and Hall, P. (1983), Defining High-Technology Industries, Working paper 407, Berkeley: University of California, Institute of Urban and Regional Development. Glasmeier, A. (1985), Innovative Manufacturing Industries: Spatial Incidence in the United States, in Castells, M. (ed.) (1985), High Technology, Space, And Society, Sage Publication. Glasmeier, A. (1986), High-Tech Industries and the Regional Division o f Labor, Industrial Relations, Institute of Industrial Relations, University of California. Berkeley 25, n.2: 197-211. Spring. 142 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Goss, E. and Vozikis, G.S. (1994), High Tech Manufacturing: Firm Size, Industry and Population Desity, Small Business Economics, v.6: 291-297. Gottlieb, P. D. (1995), Residential Amenities, Firm Location and Economic Development, Urban Studies, v32, n9: 1413-36 Gottlieb, P. D. (2001), The Problem o f Brain Drain in Ohio and Northeastern Ohio: What is it? How severe is it? What should we do about it?, Research Paper, Center for Regional Economic Issues, Weatherhead School of Management, Case Western Reserve University, Cleveland, Ohio. Gray, M., E. Golob, and A. Markusen (1996), Big Arms, Wide Shoulders: The Hub and Spoke Industrial District in the Seattle Region, Regional Studies, v.30: 651-666. Gray, M., E. Golob, Markusen, A., Park, S.O. (1998), ‘New Industrial Cities? The Four Faces of Silicon Valley’, Reviews o f Radical Political Economics, 30(4): 1-28. Hall, Sir Peter Goeffrey (1998), Cities In Civilization, New York: Pantheon Books. Hamilton, D.P. and Gepert, K. (2002), Desperately Seeking to Lure Biotechnology Businesses -States and Cities Craving Jobs Look to Gene- Based Science, But Critics Predict Disappointment-, Wall Street Journal, June 11. Haug, P. (1991), Regional Formation of High-Technology Service Industries: The Software Industry in Washington State, Environment and Planning A, v.23: 869-884. Haug, P. (1995), Formation of Biotechnology Firms in the Greater Seattle Region: An Empirical Investigation of Entrepreneurial, Financial, and Educational Perspectives, Environment and Planning A, v.21 n.2, Feb. 1995: 249-67. Hayter, Roger. (1997), The Dynamics o f Industrial Location : The Factory, the Firm and the Production System. John Wiley & Son Ltd. Hecker, Daniel. (1999), High-technology employment: a broader view, Monthly Labor Review 112(6): 18-28. Herzog, H.W., Schlottmann, A.M., and Johnson, D.L. (1986), High-technology Jobs and Worker Mobility, Journal o f Regional Science, v.26, n.3: 445-459. 143 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Hirschmann, A.O. (1958), The Strategy o f Economic Development, New Haven: Yale University Press. Information Week (2001), National IT Salary Survey, Information Week. Irwin, D.A., and Klenow, P.J. (1996), High-tech R&D subsidies Estimating the effects of Sematech, Journal o f International Economics, v.40: 323-344. Isserman, A. (1999), Making Regional Planning More Effective Through Use o f Performance Measures and Control Groups, Paper presented at the Annual North American Meeting Regional Science Association International, Montreal, Canada, November 1999. Isserman, A. and Rephann, T. (1995), The Economic Effects of the Appalachian Regional Commision, An Empirical Assessment of 26 Years of Regional Development Planning, Journal o f the American Planning Association, v. 61, n.3, Summer 1995: 345-364. Isserman, A.M. and Beaumont, P.M. (1989), New Directions in Quasi- Experimental Control Group Methods for Project Evaluation, Socio- Economic Planning Sciences, v.23, n.1/2: 39-53. Isserman, A.M. and Merrifield, J.D. (1982), The Use of Control Group in Evaluating Regional Economic Policy, Regional Science and Urban Economics, v.12: 43-58. Isserman, A.M. and Merrifield, J.D. (1987), Quasi-Experimental Control Group Methods for Regional Analysis: An Application to an Energy Boomtown and Growth Pole Theory, Economic Geography, v.63, n.l: 3-19. Iswalt, Mike (2001), High-tech Centers Are Not A ll Created Equal, http://www. dismal, com/thoughts/article. a sp ? a id = ll 78 Jacobs, J. (1969), The Economy o f Cities, New York, NY, Vintage. Kansas Technology Enterprise Corporation (1999), Kansas Innovative Index, Kansas Technology Enterprise Corporation. Klosterman, Richard E. (1990), Community Analysis and Planning Techniques, Rowman & Littlefield Publishing. Knapp, R.B. (1998), Class Note, http://www. engr. sjsu. edu/~knapp/HCIRODPR/PR_Mahal/cov_mat. htm 144 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Knight, R.V. (1989), City Development and Urbanization: Building the Knowledge Based City, in Cities in a Global Society, Knight, R.V. and Gappert, G. (eds.), Sage Publications. Kotkin, J. (2000), The New Geography -H ow the Digital Revolution Is Reshaping the American Landscape-, Random House Kotkin, J. and DeVol, R. (2001), Knowledge-Value Cities in the Digital Age, Milken Institute. Kotkin, J. and Siegel, F. (2000), Digital Geography -The remaking o f city and countryside in the new economy-, Hudson Institute. Lederman, L. and Windus, M. (1971), Federal Funding and National Priorities: An Analysis o f Programs, Expenditures, and Research and Development, Praeger Publishers. Luger, M.I. and Goldstein, H.A. (1991), Technology in the Garden: Research Parks and Regional Economic Development, The University of North Carolina Press, Chapel Hill & London. Lucas, R. (1988), On the Mechanics of Economic Development, Journal o f Monetary Economics, 22:1: 3-42. Lyons, D.L (1995), Agglomeration Economies among High Technology Firms in Advanced Production Areas: The Case of Denver/Boulder, Regional Studies, v.29, n.3 (June 1995): 265-78. Malecki, E. J. (1985), Industrial Location and Corporate Organization in High- Tech Industries, Economic Geography, v.61: 345-369. Malecki, E. J., (1991), Technology and economic development: the dynamics o f local, regional and national change, Longman Science & Technical. Malecki, E. J., and S. L. Bradbury (1992), R&D Facilities and Professional Labour: Labor Force Dynamics in High Technology, Regional Studies, v.26: 123-136. March, M.S. (1970), Federal budget priorities fo r research and development, University of Chicago, Center for Policy Study. Markusen, A., P. Hall, and A. Glasmeier (1986), High Tech America: The What, How and Why o f Sunrise Industries, Boston, MA: Allen and Irwin. 145 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Markusen, A.R. (1984), Defense spending and the geography o f high tech industries, Institute of Urban and Regional Development, University of California, Berkeley. Markusen, A.R. (1989), The Economic, Industrial, and Regional Consequences o f Defense-Led Innovation, Knowledge and industrial organization: 251-69. Markusen, A.R. (1991), Government as Market: Industrial Location in the U.S. Defense Industry, Industry location and public policy: 137-68. Markusen, A.R. and Bloch, R. (1985), ‘Defensive Cities: Military Spending, High Technology, and Human Settlements, in Castells, M. (ed.) (1985), High Technology, Space, And Society, Sage Publication. Markusen, Ann. R. 1991. “Government as Market: Industrial Location in the U.S. Defense Industry.” pp. 137-68 in Industry Location and Public Policy, edited by Henry W. Herzog Jr. and Alan M. Scholottmann. Knoxville: The University of Tennessee Press. Massachusetts Technology Collaborative (2000), Index o f the Massachusetts Innovation Economy, Massachusetts Technology Collaborative. McArthur, R. (1990), Replacing the concept of high technology: toward a diffusion-based approach, Environment and Planning A, v.22: 811-828. Moore, B. and Rhodes, J. (1973), Evaluating the effect of British regional economic policy, Economic Journal, v.83: 87-110. Murphy, T.P. (1971), Science, Geopolitics, and Federal Spending, Health Lexington Books. Mydal, G. (1957), Economic Theory in Underdeveloped Regions, London, Duckworth. National Science Foundation (2001), Experimental Program to Stimulate Competitive Research -Statistics-, www. ehr. nsf.gov/epscor/statistics, National Science Foundation. National Science Foundation (2002), Doctoral Record Files: Fiscal Year 1966- 2000, httv://casvar. nsf. eov. National Science Foundation. National Science Foundation (2002), Federal Obligations fo r Science and Engineering: Fiscal Year 1971-1999, http://caspar. nsf. eov. National Science Foundation. 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. National Science Foundation (1967), Federal Support to Universities & Colleges; Fiscal Years 1963-66, National Science Foundation. National Science Foundation (2002), R&D Expenditures: Fiscal Year 1966- 2000, httv://casvar. n sf gov. National Science Foundation. Nelson, R.R. (1982), Government and Technical Progress: A Cross-Industry Analysis, Pergamon Press. Nelson, R.R. (1984a), High-technology policies: a five-nation comparison, Washington, D.C.: American Enterprise Institute for Public Policy Research. Nelson, R.R. (1984b), Policies in Support o f High Technology Industries, Working Paper, Institution for Social and Policy Studies. Nelson, R.R. (1986), Incentives for Entrepreneurship and Supporting Institutions, in Balassa, B. and Giersch, H. (eds.) Economic Incentives, St Martin’s Press: 173-187. Nelson, R.R. (1996), Why Do Firms Differ, and How Does It Matter?, The sources o f economic growth, Cambridge and London: Harvard University Press: 147-58. Nelson, R.R. and Winter S.G. (1974), Neoclassical vs. Evolutionary Theories of Economic Growth: Critique and Prospectus, The Economic Journal, v.84, Issues 336: 886-905 Norton, R.D. and Rees, J. (1979), The Product Cycle and the Spatial Decentralization of American Manufacturing, Regional Studies, v.13: 141- 151. Office of Technology Assessment (1995), Innovation and Commercialization o f Emerging Technologies, Office of Technology Assessment. Office of Technology Policy (2002), The Dynamics o f Technology-based Economic Development, State Science and Technology Indicators, 2nd Edition, U.S. Department of Commerce. Organisation for Economic Co-operation and Development (1982), Innovation in small and medium firm s, Committee for Scientific and Technological Policy, Organization for Economic Co-operation and Development 147 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Perroux, F. (1955), Note sur la notion de pole de croissance, Translated in Livingstone, I. (ed.) (1971), Economic policy fo r development, Baltimore: Penguin Books: 278-89. Perroux, F. (1961), The economy o f the twentieth century, Paris: PUF. Phillips, B.D. (1991), The Increasing Role of Small Firms in the High- Technology Sector: Evidence from the 1980s, Business Economics, Jan. 1991:40-47. Preer, R.W. (1992), The emergence o f technopolis: knowledge-intensive technologies and regional development, New York, Praeger. Premus, R. (1982), Location o f High Technology Firms and Regional Economic Development, Washington D.C.:U.S. Government Printing Office. Quigley, J.M. (1998), Urban Diversity and Economic Growth, Journal o f Economic Perspectives, 12(2), Spring: 127-138. Reichardt, C.S. and Mark, M.M. (1997), Quasi-experimentation, In Bickman, L. and Rog, D.J. (1997), Handbook o f Applied Social Research Methods, SAGA Publications. Rephann, T. and Isserman, A. (1994), New highways as economic development tools: An evaluation using quasi-experimental matching methods, Regional Science and Urban Economics, v.24: 723-751. Richardson, H.W. (1979), Regional Economics, Illinois University Press (Champaign). Roback, J. (1982), Wages, Rents, and Quality of Life, Journal o f Political Economy, v.90: 1257-1278. Rogue Valley Civic League and Southern Oregon Regional Services (2000), Southern Oregon Quality o f Life Index, Rogue Valley Civic League and Southern Oregon Regional Services Rosegrant, S., and Lampe, D. (1992), Route 128 -Lessons from Boston’ s High- Tech Community, Basic Books. Saint-Germain, M.A. (2002), Class Note: PPA 696 Research Methods, http://www. csulb. edu/~msaintg/ppa696/696quasi. htm. Samuelson, R.J. (2002), But Does It Own the Means of Production?, Book Review, Los Angeles Times, Sunday, July 14. 148 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sanders, E. and Huffstutter, P J. (2000), Dreams of High-Tech Glory Passing Orange County By, Los Angeles Times, Sunday, July 9. Saxenian, A. (1981), Silicon Chips and Spatial Structure: The Industrial Basis o f Urbanization in Santa Clara County, California, Working Paper 345, Institute of Urban and Regional Development, University of California, Berkeley. Saxenian, A. (1985), Silicon Valley and Route 128: Regional Prototypes or Historic Exceptions?, in Castells, M. (ed.) (1985), High Technology, Space, And Society, Sage Publication. Scott, A. J. (1992), The Role of Large Producers in Industrial Districts: A Case Study of High Technology Systems Houses in Southern California, Regional Studies, v.26: 265-275. Scott, A. J., and M. H. Drayse (1993), Electronics Industry in Southern California: Growth and Spatial Development 1945-1989, The Review o f Regional Studies, v.20: 1-14. Sivitanidou, R. (1999), The Location of Knowledge-Based Activities: The Case of Computer Software in Fischer, M.M., L. Suarez-Villa, and M. Steiner (eds.), Innovation, Networks, and Localities, Advances in Spatial Science, Springer. Sivitanidou, R., and P. Sivitanides (1995), The Intrametropolitan Distribution of R&D Activities: Theory and Empirical Evidence, Journal o f Regional Science, v.35: 391-415. Sommers, P., and Carlson, D. (2000), Ten Steps to a High Tech Future: The New Economy in Metropolitan Seattle, Working Paper, Center on Urban and Metropolitan Policy, The Brookings Institute. Steinberg, R. (1996), Regional Growth Theories and High-Tech Regions, International Journal o f Urban and Regional Research, v.20, n.3: 518-38. Storper, M. and Scott, A.J. (1993), The Wealth o f Regions: Market Forces and Policy Imperatives in Local and Global Context, Lewis Center for Regional Policy Studies, Working Paper no. 7, University of California: Los Angeles. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Suarez-Villa, L. (1997a), ‘Innovative Capacity, Infrastructure and Regional Inversion: Is There a Long-Term Dynamic?’ in Bertuglia, Cristoforo S., Lombardo, Silvana; Nijkamp, Peter (eds.) (1997b), Innovative behaviour in space and time, Advances in Spatial Science series. Heidelberg and New York: Springer. Suarez-Villa, L. (2000), Invention and the Rise o f Technocapitalism, Rowman & Littlefield Publishers, Inc. Suarez-Villa, L., and W. Walrod (1997b), Operational Strategy, R&D and Intrametropolitan Clustering in a Poly centric Structure: The Advanced Electronics Industries in the Los Angeles Basin, Urban Studies, v.34: 1343- 1380. The Economist (2000), The Geography o f Cool, April 15:9. Timmons, J.A., and Fast, N.O. (1987), The Flow of Venture Capital to Highly Innovative Technological Ventures, in Churchill N.C. et al (eds.) Frontiers o f Entrepreneurial Research, Wellesley, MA, Babson College, Center for Entrepreneurial Studies: 109-123. Trochim, W.M.K. (eds.) (1986), Advances in Quasi-Experimental Design and Analysis, Jossey-Bass Inc., Publishers. Varga, A. (1999), Time-Space Patterns of US Innovation: Stability or Change?, in Fischer, M.M., L. Suarez-Villa, and M. Steiner (eds.), Innovation, Networks, and Localities, Advances in Spatial Science, Springer. Vinson, R and Harrington, P. (1979), Defining “high-tech industries” in Massachusetts, Commonwealth of Massachusetts, Department of Manpower Development. Washington State (2000), Index o f Innovation and Technology, Working Paper, Washington State. Westhead, P., and Batstone, S. (1998), Independent Technology-based Firms: The Perceived Benefits of a Science Park Location, Urban Studies, v.35, n. 12: 2197-2219. Williams, F. (1992), The Information Infrastructure in Technopolis: The Intelligent Network, Gibson, D., Kozmetsky, G. and Smilor, R.W. (eds.), The Technopolis Phenomenon: Smart Cities, Fast Systems, Global Networks, Rowman & Littlefield Publishers, Inc. 150 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Zucker, L.G., Darby, M.R., and Brewer, M.B. (1998), Intellectual Human Capital and the Birth of U.S. Biotechnology Enterprise, American Economic Review (March 1998): 290-306 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A p p e n d i x 1 100 Largest high-tech employment counties in 1997 Rank State County Employment Rank State County Employment 1 California Los Angeles 737652 51 Oregon Washington 48796 2 California Santa Clara 465282 52 New York Nassau 48701 3 M assachusetts Middlesex 295365 53 Florida Broward 47597 4 Illinois Cook 251848 54 Missouri Jackson 46356 5 California Orange 241698 55 Florida Orange 45360 6 Texas Dallas 236298 56 Colorado Boulder 45034 7 New York New York 209866 57 California Ventura 44983 8 Virginia Fairfax 201850 58 Florida Pinellas 44849 9 California San Diego 191547 59 Michigan Wayne 44815 10 Texas Harris 190813 60 Kansas Johnson 42504 11 Arizona Maricopa 148354 61 Maryland Prince Georges 40959 12 Washington King 138143 62 Massachusetts Norfolk 40445 13 Minnesota Hennepin 135053 63 Indiana Marion 39681 14 Michigan Oakland 114738 64 New Mexico Bernalillo 39065 15 California Alameda 111882 65 Connecticut New Haven 38742 16 Maryland Montgomery 108820 66 Georgia Gwinnett 37975 17 Texas Travis 103302 67 M assachusetts Essex 37912 18 New York Suffolk 86442 68 New York W estchester 37454 19 Pennsylvania Allegheny 84757 69 Wisconsin Milwaukee 37084 20 Illinois Du Page 83725 70 New Jersey Essex 36108 21 Georgia Fulton 80885 71 New York Monroe 36053 22 Ohio Cuyahoga 78619 72 Georgia Cobb 34982 23 Pennsylvania Montgomery 77254 73 Oklahoma Tulsa 34870 24 California San Mateo 74941 74 Colorado El Paso 34467 25 New Jersey Middlesex 73533 75 Pennsylvania Chester 34217 26 Connecticut Hartford 68655 76 Nebraska Douglas 34114 27 Pennsylvania Philadelphia 65835 77 Alabama Madison 33669 28 Kansas Sedgwick 65787 78 Oregon Multnomah 32904 29 California San Francisco 65633 79 Maryland Howard 32742 30 New Jersey Morris 61962 80 Minnesota Ramsey 32571 31 Missouri St. Louis 60374 81 Texas Collin 31539 32 Ohio Hamilton 60164 82 New York Erie 31368 33 Connecticut Fairfield 59051 83 Florida Brevard 29932 34 M assachusetts Suffolk 57972 84 Florida Duval 29850 35 Florida Hillsborough 57633 85 Florida Palm Beach 29162 36 New Jersey Bergen 57446 86 Pennsylvania Bucks 28389 37 Texas Tarrant 57184 87 M assachusetts Worcester 28234 38 Utah Salt Lake 56880 88 New Jersey Somerset 27893 39 Texas Bexar 56685 89 Missouri St. Louis City 27702 40 California Contra Costa 55861 90 Maryland Baltimore 27110 41 California Sacramento 55298 91 New York Albany 26949 42 Florida Dade 55052 92 New Hampshire Hillsborough 26800 43 Ohio Franklin 54159 93 North Carolina Durham 24675 44 Colorado Arapahoe 52959 94 Kentucky Jefferson 24222 45 Georgia De Kalb 50188 95 Wisconsin Waukesha 24034 46 North Carolina Mecklenburg 50181 96 Nevada Clark 23704 47 Colorado Denver 49994 97 Colorado Jefferson 23546 48 D.C. Washington 49679 98 T ennessee Shelby 23470 49 Virginia Arlington 49433 99 T ennessee Davidson 23213 50 North Carolina Wake 48861 100 Arizona Pima 22736 Source: County Business Pattern, 1997; Author Calculation 152 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 2 High-tech Counties whose L.Q. is larger than 1.25 in 1997 Rank State County L.Q.(1997) Rank S tate County L.Q.(1997) 1 Nebraska Logan 7.00485232 51 North Carolina Durham 1.75009169 2 Indiana Owen 6.76422898 52 Michigan Oakland 1.72467571 3 California Santa Clara 5.62963791 53 Georgia De Kalb 1.72235621 4 Virginia Fairfax 5.18585775 54 Georgia Gwinnett 1.71778710 5 Virginia King George 5.01882507 55 New Mexico Bernalillo 1.69220617 6 Virginia Arlington 4.85909041 56 Texas Fort Bend 1.68493673 7 Idaho Bonneville 4.26317975 57 New Hampshire Hillsborough 1.68278503 8 M assachusetts Middlesex 4.11775888 58 Indiana Kosciusko 1.65969361 9 Maryland St. Marys 3.91022845 59 California Santa Cruz 1.65954376 10 Colorado Boulder 3.76380523 60 Virginia Loudoun 1.65677295 11 Kansas Sedgwick 3.17332279 61 Connecticut Hartford 1.63405269 12 Maryland Montgomery 3.15283692 62 Washington King 1.60589148 13 T ennessee Anderson 3.13374197 63 Washington Benton 1.60216628 14 Alabama Madison 3.04514574 64 California Sacramento 1.59391882 15 Texas Travis 3.03397259 65 Illinois Du Page 1.59199951 16 Maryland Howard 3.01098237 66 Michigan W ashtenaw 1.59181618 17 Virginia Alexandria 2.95058670 67 New Jersey Burlington 1.58867708 18 Virginia Fairfax City 2.81481223 68 Colorado Jefferson 1.57906429 19 Oregon Washington 2.71179083 69 California Santa Barbara 1.57849621 20 Ohio Greene 2.54247156 70 M assachusetts Essex 1.56603849 21 California San Mateo 2.53163817 71 Connecticut Fairfield 1.51676347 22 New Jersey Morris 2.48955185 72 Colorado Denver 1.46490600 23 Texas Collin 2.43210422 73 Georgia Cobb 1.45825027 24 California Ventura 2.29583510 74 Missouri Jackson 1.44865734 25 Colorado Arapahoe 2.28257173 75 South Carolina Aiken 1.43120704 26 California San Diego 2.28077746 76 New Jersey Bergen 1.40121702 27 New Jersey Middlesex 2.25373909 77 M assachusetts Norfolk 1.39885507 28 California Los Angeles 2.21966780 78 New Mexico Los Alamos 1.38054253 29 Florida Brevard 2.15313957 79 Rhode Island Newport 1.38002033 30 California Orange 2.15235094 80 California San Francisco 1.37143671 31 California Contra Costa 2.15213954 81 Utah Salt Lake 1.36587424 32 Pennsylvania Chester 2.11345537 82 Pennsylvania Allegheny 1.36385264 33 California Alameda 2.09529377 83 Texas Harris 1.36292803 34 California Marin 2.01689340 84 D.C. Washington 1.35365364 35 Kentucky Marshall 2.01598989 85 Georgia Fulton 1.35299845 36 Colorado El Paso 2.01293709 86 Pennsylvania Bucks 1.34986370 37 Minnesota Winona 1.95471579 87 Louisiana St. Jam es 1.34643215 38 New York Suffolk 1.94969411 88 Arizona Maricopa 1.34384582 39 Kansas Johnson 1.92300939 89 Virginia M anassas City 1.34148224 40 North Dakota Walsh 1.91560438 90 Colorado Larimer 1.32964100 41 Louisiana Iberville 1.90230820 91 Florida Hillsborough 1.32846233 42 Pennsylvania Elk 1.87721039 92 New Jersey Hunterdon 1.31317892 43 Maryland Prince Georges 1.86602377 93 Nebraska Douglas 1.31006515 44 Texas Dallas 1.86327251 94 Connecticut New Haven 1.28797872 45 North Carolina Wake 1.82998244 95 Wisconsin W aukesha 1.28693614 46 New Jersey Somerset 1.82003116 96 Arizona Cochise 1.28136797 47 Minnesota Hennepin 1.80435811 97 Oklahoma Tulsa 1.27891927 48 New York Albany 1.78881704 98 Utah Utah 1.27230499 49 Kansas Sumner 1.77899861 99 Florida Pinellas 1.26401064 50 Pennsylvania Montgomery 1.77658147 100 New Jersey Passaic 1.25620743 Source: County Business Pattern, 1997; Aut nor Calculation 153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 3 High-tech Counties whose growth rate is larger than 200% during the 1987-1997 periods Ranking State County Growth Rate (%) Ranking State County Growth Rate (%) 1 Kansas Sedgwick 1479.00 42 New York Kings 331.00 2 Maryland Howard 1376.00 43 California San Mateo 331.00 3 New York Albany 904.00 44 Alaska Anchorage 325.00 4 Texas Collin 892.00 45 Michigan Macomb 322.00 5 North Carolina Mecklenburg 853.00 46 Nevada W ashoe 314.00 6 New Mexico Bernalillo 720.00 47 Indiana Marion 314.00 7 Arkansas Pulaski 719.00 48 Tennessee Knox 313.00 8 Tennessee Davidson 711.00 49 Kansas Johnson 302.00 9 Kentucky Jefferson 683.00 50 Georgia Fulton 298.00 10 Pennsylvania Philadelphia 652.00 51 Utah Utah 297.00 11 Texas Travis 648.00 52 California San Francisco 297.00 12 Virginia Alexandria 636.00 53 Washington Snohomish 294.00 13 New Jersey Somerset 603.00 54 California Marin 292.00 14 Hawaii Honolulu 584.00 55 New Jersey Burlington 289.00 15 Missouri St. Louis City 545.00 56 North Carolina Wake 280.00 16 Oklahoma Oklahoma 539.00 57 Colorado Arapahoe 268.00 17 Louisiana East Baton Rouge 527.00 58 Oregon Washington 262.00 18 California Riverside 525.00 59 Pennsylvania Chester 250.00 19 Maryland Anne Arundel 518.00 60 Texas Tarrant 250.00 20 Texas Bexar 514.00 61 Florida Hillsborough 250.00 21 Colorado Denver 510.00 62 North Carolina Guilford 249.00 22 North Carolina Durham 508.00 63 Minnesota Ramsev 245.00 23 California Sacramento 463.00 64 Texas Harris 245.00 24 California San Bernardino 460.00 65 Iowa Polk 238.00 25 Missouri Jackson 453.00 66 Virginia Fairfax City 238.00 26 istrict of Columbi Washington 440.00 67 New Jersey Morris 236.00 27 Oregon Multnomah 428.00 68 New Hampshire Rockingham 231.00 28 Ohio Lucas 422.00 69 Wisconsin Waukesha 231.00 29 Pennsylvania Allegheny 422.00 70 Virginia Virginia Beach 227.00 30 California Contra Costa 408.00 71 Alabama Mobile 227.00 31 Florida Seminole 405.00 72 Virginia Fairfax 226.00 32 Georgia Cobb 389.00 73 Pennsylvania Cumberland 222.00 33 Tennessee Shelby 383.00 74 Massachusetts Suffolk 221.00 34 Colorado Jefferson 377.00 75 Alabama Jefferson 219.00 35 Ohio Franklin 368.00 76 New York New York 211.00 36 Florida Dade 365.00 77 Washington Clark 208.00 37 Pennsylvania Delaware 342.00 78 Massachusetts Norfolk 208.00 38 Wisconsin Dane 340.00 79 Michigan Washtenaw 204.00 39 Florida Palm Beach 338.00 80 Nebraska Douglas 201.00 40 Virginia Arlington 334.00 81 Florida Duval 201.00 41 Washington King 333.00 82 Florida Orange 201.00 Source: County Business Pattern, 1987, 1997; Author Calculation Restriction: High-tech Employment in 1987 is larger than 2,000 154 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 4 Rationales for Seven Variables Used to Define Twin Counties (Control Group) : There are many variables that are found to have a direct impact on high-tech industries and high-tech centers (Kansas Technology Enterprise Corporation, 1999; Massachusetts Technology Collaborative, 2000; Washington State, 2000; National Science Foundation, 2001; Office of Technology Policy, 2002; Atkinson, 2002). However, all these variables are not available in county level, especially in 1950 data. The suggested seven variables are most relevant and available in county level and in 1950 data to measure high-tech centers formation. These seven variables are included in three categories (education, employment, and geography). The units of these variables are number or size, rather than density, number per population or number per employment because variables for scaling such as number of employment and area size are already included among the seven variables. <Education Variables> 1. Number of people with high school completed (1950) : This number is important in two major reasons. First, as a region continues to shift to a high-technology economy, high school dropouts will have fewer job opportunities. Second, in addition to the employment barriers, young people who are not in school are much more likely to bring out some kind of trouble than those who complete high school (The Legislative Assembly Oregon Progress Board, 1999:24). Drug and alcohol use, teen pregnancy, teen parenting and crime behavior are all more prevalent among teenage dropouts. Higher this number promises better quality of life and better labor pool for high-tech firm location. 155 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2. Number of people with college or university completed ( 1950) : An educated workforce is important to increasing productivity and fostering innovation in high-tech firms. This number is important because it gives an indication of both the level of educational attainment and the type of skills that are demanded by high-tech firms. The greater this number, the higher the perceived value of an education to the youth culture. The greater this number means the higher the amounts of wages gained by the workforce. The more educated the population, the more attractive the region is for high- tech firms where knowledge and skill assets are relatively more important than tangible assets (DeVol, Koepp, and Fogelbach, 2002: 81). <Employment Variables> 3. Number of total employment (1950) : This number is important in selecting a twin county that has a similar economic condition with its corresponding high-tech center. The other two employment variables can be a number not a density or ratio because local employment size is already considered. 4. Number of government workers (1950) : In 1950, a large portion of educated workers is allocated to government workers. Like professional and technical workers, government workers are a representation of educated workers who can contribute technological development in the region. In addition, government workers play an important role in government funding and support programs for high-tech industrial development and other science and technology policy including public R&D investments. 5. Number of professional and technical workers (19501 : In high-tech industries, most of a product’s value is added by professionals or technical workers before and after manufacturing (The Legislative 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Assembly Oregon Progress Board, 1999:12). Therefore, these professional and technical workers are increasingly important to regional economic competitiveness of high-tech businesses. These workers include managers, engineers and scientists, health professionals, lawyers, teachers, accountants, bankers, consultants, and engineering technicians (Washington State, 2002). These workers consist of an important part of the Creative Class (Florida, 2002) and contribute communities’ quality of life, which attracts high-tech firms’ location. In addition, advances in technical workers lead to increased R&D funding and opportunities from both within and outside the region. <Geography Variables> 6. County total area size (1950) : This size is important in selecting a twin county that has a similar physical and geographical condition with its corresponding high-tech center. The other variables can be a number not a density or ratio because county area size is already considered. 7. Non-farm land size (1950) : Even though two counties have a similar area size, they will have different potentials and developmental patterns if their land uses are very different. Thus, non-farm land size is important because this variable can promise a similar land use pattern in future progress including high-tech center development. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 5-1 Geographical Listing of Medical Schools of the United States (1) J S te Med cai Sch I clH lliH l 1 CA University of Southern California Los Angeles O 2 CA UCLA Los Angeles O 3 CA UC Irvine Orange O 4 CA Loma Linda San Bemandino X 5 CA UC San Diego San Diego O 6 CA UC San Francisco San Francisco O 7 CA Stanford Santa Clara O 8 CA UC Davis Yolo X 9 CO University of Colorado Denver O 10 CT University of Connecticut Hartford O 11 CT Yale University New Haven X 12 GA Mercer University Bibb X 13 GA Emory University De Kalb O 14 GA Morehouse University Fulton O 15 GA Medical College of Georgia School of Medicine Richmond X 16 I L Loyola University Cook X 17 I L Northwestern University Cook X 18 I L Rush University Cook X 19 I L University of Chicago Cook X 20 I L University of Illinois College of Medicine Cook X 21 IL Finch University of Health Science Lake X 22 I L Southern Illinois University Sangamon X 23 IN Indiana University M arion X 24 KS University of Kansas Wyandotte X 25 K Y University of Kentucky Fayette X 26 K Y University of Louisville Jefferson X 27 L A Louisiana State University Caddo X 28 L A Louisiana State University Orleans X 29 L A Tulane University Orleans X 30 M A Boston University Suffolk 0 31 M A Harvard Medical School Suffolk 0 32 M A Tufts University Suffolk 0 33 M A University of Massachusetts Worcester X 34 M D Johns Hopkins University Baltimore City X 35 M D University of Maryland Baltimore City X 36 M D Uniformed Service University Montgomery 0 37 M l Michigan State University College Ingham X 38 M l University of Michiga Washtenaw 0 Source: Association of American Medical Colleges (2002), continued Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 5-2 Geographical Listing of Medical Schools of the United States (2) No s ate Medical School 39 M l Wayne State University Wayne X 40 NC Duke University Durham O 41 NC Wake Forest University Forsyth X 42 NC University of North Carolina Orange X 43 NC East Calolina University Pitt X 44 NJ UMDNJ-New Jersey Medical School Essex X 45 NJ UMDNJ-Robert Wood Johnson Medical School Middlesex O 46 NY Albany Medical College Albany X 47 NY Yeshiva University Bronx X 48 NY University at Buffalo State University of New York Erie X 49 NY State University of New York Kings X 50 NY University of Rochester Monroe X 51 NY Columbia University New York 0 52 NY Cornell University New York 0 53 NY New York University New York 0 54 NY New York University School of Medicine New York 0 55 NY State University of New York Onondaga X 56 NY Stoney Brook Unversity Suffolk X 57 NY New York Medical College Westchester 0 58 PA University of Pittsburgh Allegheny X 59 PA Pennsylvania State University Dauphin X 60 PA Thomas Jefferson University Philadelphia 0 61 PA Drexel University Philadelphia 0 62 PA Temple University Philadelphia 0 63 PA University of Pennsylvania Philadelphia 0 64 SC Medical University of So. Carolina Charleston X 65 SC University of South Carolina Richland X 66 TX University of Texas at San Antonio Bexar X 67 TX The Texas A & M University Brazos X 68 TX University of Texas at Dallas Dallas 0 69 TX University of Texas at Galveston Galveston X 70 TX Baylor College of Medicine Harris 0 71 TX University of Texas at Houston Harris 0 72 TX Texas Tech University Lubbock X 73 VA University of Virginia Charlottesville X 74 VA Eastern Virginia Medical School of the Medical College Norfolk X 75 VA Virginia Commonwealth University Richmond City X 76 WA University of Washington K ing 0 Source: Association of American Medical Colleges (2002) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
An essay on Korean reunification: Spatial effects of regional integration and its implications for regional and transportation policies
PDF
Infrastructure investment and regional productivity of the manufacturing industries in Mexico
PDF
Economic analysis of ground lease-based land use system
PDF
Decentralizing of public finance: Centralizing forces in developing countries
PDF
Evaluation of policy alternatives to alleviate transportation-related air pollution
PDF
Intra-metropolitan variation in residential mortgage prepayment risks: Unobserved factors and estimation bias
PDF
Institutional analysis of evolution of housing cooperatives in India
PDF
Accuracy, richness of outcomes, and internal-consistency of integrated metropolitan input-output models: An empirical test of their relationships
PDF
Essays on commercial real estate and commercial mortgages
PDF
Income inequality and economic growth: A theoretical and empirical analysis
PDF
Differences in the development of citizenship: Residents' motivations and capacities for participating in neighborhood associations in Los Angeles
PDF
Dependence on foreign labor, quality of education and unemployment in the GCC countries: In search of solutions
PDF
CoCoA, a novel coactivator which acts through the interaction with the bHLH -PAS domains of p160 coactivators, AHR, and ARNT
PDF
Immigrants and housing attainment: Different forms of residential assimilation
PDF
Essays on agrarian institutions
PDF
Farewell to the empire? National identity, domestic structures, and foreign economic policies of the post -Soviet states
PDF
Jump -starting Main Street: A case study of the Los Angeles Neighborhood Initiative (LANI)
PDF
A study on the impact of congestion pricing: Benefits and revenue redistribution for the case of Seoul and Los Angeles
PDF
Globalization and the decline of the welfare state in less developed countries
PDF
An enquiry into the effectiveness of export assistance organizations serving southern California SMEs
Asset Metadata
Creator
Ki, Jung-Hoon
(author)
Core Title
A statistical analysis of the formation and location factors of high -tech centers in the United States, 1950--1997: An evaluation using quasi -experimental control group methods
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Planning
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
economics, labor,OAI-PMH Harvest,Statistics,Urban and Regional Planning
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Gordon, Peter (
committee chair
), Nugent, Jeffrey B. (
committee member
), Richardson, Harry W. (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-259386
Unique identifier
UC11339230
Identifier
3093777.pdf (filename),usctheses-c16-259386 (legacy record id)
Legacy Identifier
3093777.pdf
Dmrecord
259386
Document Type
Dissertation
Rights
Ki, Jung-Hoon
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
economics, labor