1 MEASURING REGIONAL AND LOCAL INNOVATIVE OPPORTUNITY By IRIS E. PATTEN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF P HILOSOPHY UNIVERSITY OF FLORIDA 2014
2 Â© 201 4 Iris E . Patten
3 ACKNOWLEDGEMENTS I cannot express enough thanks to my committee for their continued support , encouragement , and patience : Dr. Paul Zwick , my committee chair; Peggy Carr, my committee co c hair ; Dr. Joseli Macedo ; and Dr. Barbara McDade . I offer my sincere appreciation for the learning opportunities provided by my committee. My completion of this project could not have been accomplished without the support of my family. You were there ever y step of the way and especially when your encouragement and support thank everyone who, willingly or not, came along for the ride.
4 TABLE OF CONTENTS page ACKNOWLEDGEMENTS ................................ ................................ ............................... 3 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 2 LITERATURE REVIEW ................................ ................................ .......................... 15 Industrial Performance in the New Economy: Geography ................................ 17 Industrial Performance in the New Economy: Geography ................................ 26 The Urban System ................................ ................................ ................................ .. 30 Components ................................ ................................ ................................ ........... 37 Components: Inno vative Agents ................................ ................................ ...... 41 Components: Innovative Inputs ................................ ................................ ....... 43 Components: Innovative Outputs ................................ ................................ .... 47 Relationships ................................ ................................ ................................ .......... 48 Measurement ................................ ................................ ................................ .......... 54 Conclusion ................................ ................................ ................................ .............. 70 3 METHODOLOGY ................................ ................................ ................................ ... 78 Theoretical Study Area: Metro Washington, District of Columbia ........................... 80 Comparative Study Area: Pima County, Ariz ona ................................ .................... 81 Understanding the Innovative System: The Actors ................................ ................. 83 Understanding the Innovative System: Methods ................................ ..................... 85 Data ................................ ................................ ................................ ........................ 87 Innovative Agents: Temporal ................................ ................................ ............ 89 Innovative Agents: Spatial ................................ ................................ ................ 91 Innovative Inputs: Temporal ................................ ................................ ............. 91 Innovative Inputs: Spatial ................................ ................................ ................. 95 Innovative Outputs ................................ ................................ ........................... 96 Suitability Analysis and Economic Opportunity ................................ ....................... 97 4 RESULTS AND ANALYSIS ................................ ................................ .................. 105 Capitol Region Innovative Agents ................................ ................................ ...... 105
5 Capitol Region Innovative ................................ ................................ .................. 111 Capitol Region Innovative Outputs ................................ ................................ .... 113 Theoretical Regional Regression Equation ................................ ........................... 114 Capitol Region Innovative Networks ................................ ................................ .. 114 Theoretical Local (Spatial) Regression Equation ................................ ............ 115 Pima County Region Innovative Agents ................................ ...................... 116 4 DISCUSSION ................................ ................................ ................................ ....... 127 5 CONCLUSION ................................ ................................ ................................ ...... 130 APPENDIX A DATA LIST ................................ ................................ ................................ ............ 134 B OCCUPATIONAL PROF ILES ................................ ................................ ............... 138 C SECTOR INDUSTRY TRENDS ................................ ................................ ............ 139 D CAPITOL REGION SCATTERPLOTS BY AGE COHORT ................................ 142 E REGRESSION RESULTS BY AGE COHORT ................................ ................... 155 F SPATIAL STATISTICS METRO WASHINGTON, DC REGION ......................... 167 G CAPITOL REGION SCATTERPLOTS EDUCATIONAL ATTAINMENT FACTOR ................................ ................................ ................................ ............... 181 H REGRESSION RESULTS EDUCATIONAL ATTAINMENT FACTOR ................. 185 I REGRESSION R ESULTS PRINCIPAL COMPONENTS ANALYSIS, CAPITOL REGION ................................ ................................ ................................ ................ 188 REFERENCE LIST ................................ ................................ ................................ ...... 189 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 194
6 LIST OF TABLES Table page 2 1 The OECD/EC classification of knowledge intensive sectors ............................. 71 2 2 A reformulation of the components of local economic development ................... 72 2 3 Preference value descriptions ................................ ................................ ............ 73 2 4 Three example stakeholder preference value explanations. .............................. 73 3 1 A&E firms per capita ................................ ................................ ......................... 101 3 2 Degrees and disciplines considered for architectural and engineering in dustries ................................ ................................ ................................ .......... 102 3 3 Education factor for weight distribution ................................ ............................. 103 3 4 Traditional measures of concentration ................................ .............................. 103 4 1 Top 10 places of residence for metro Washington, DC region workers ............ 119 4 2 Capitol region R2 values, age ................................ ................................ ........... 121 4 3 R square regression statistic by age cohort ................................ ...................... 121 4 4 Capitol region R 2 values, educational attainment factor ................................ ... 121 4 5 R square regression statistic for educational attainment factor ........................ 121 4 6 Capitol region architecture and engineering firm study count ........................... 122 4 7 Location quotients for Capitol area study counties ................................ ........... 122 4 8 Horizontal clustering quotient for metro Washington, DC region ...................... 122 4 9 Pima County population, 1990 2008 ................................ .............................. 123 4 10 Places of residence for Pima County, Arizona workers ................................ .... 124 4 11 Pima County regression values ................................ ................................ ....... 124 4 12 Pima County architecture and engineering firm study count ............................ 124
7 LIST OF FIGURES Figure page 2 1. Skill composition of economic sectors in the European Union, 2000. (Recreated from Miles, I. (2008) Knowledge Policy: Challenges for the 21 st Century, 16) ................................ ................................ ................................ ........ 74 2 2. Illustrates the relationship between residential traditional business factors, residential amenities, population growth, and firm location and employment growth. (Source: Gottlieb, P.D. (1995)) ................................ ............................. 75 2 3. LUCIS hierarchical structure ................................ ................................ ................. 76 2 4. LUCIS preference ................................ ................................ ................................ . 77 4 1. Capito l region A&E firm density ................................ ................................ ........... 125 B 1. Occupational profiles architecture and engineering firms (NAICS 5413) in the Greater Washington area. ................................ ................................ ................ 138 D 1. Business patterns for 20 to 24 year olds in the Capitol Region, 1990 ................. 143 D 2. Business patterns for 25 to 34 year olds in the Capitol Region, 1990 ................. 144 D 3. Business patterns for 35 to 44 year olds in the Capitol Region, 1990 ................. 145 D 4. Business patterns for 45 to 54 year olds in the Capitol Region, 1990 ................. 146 D 5. Business patterns for 20 to 24 year olds in the Capitol Region, 2000 ................. 147 D 6. Business patterns for 25 to 34 year olds in the Capitol Region, 2000 ................. 1 48 D 7. Business patterns for 35 to 44 year olds in the Capitol Region, 2000 ................. 149 D 8. Business patterns for 45 to 54 year olds in the Capitol Region, 2000 ................. 150 D 9. Business patterns for 20 to 24 year olds in the Capitol Region, 2008 ................. 151 D 10. Business patterns for 25 to 34 year olds in the Capitol Region, 2008 ............... 152 D 11. Busi ness patterns for 35 to 44 year olds in the Capitol Region, 2008 ............... 153 D 12. Business patterns for 45 to 54 year olds in the Capitol Region, 2008 ............... 154 F 1. District of Columbia spatial statistics summary ................................ .................... 167 F ............................ 169
8 F 3. Fairfax County, Virginia spatial st atistics summary ................................ .............. 171 F 4. Montgomery County, Maryland spatial statistics summary ................................ .. 173 F 5. Fairfax City, Virginia spatial statistics summary ................................ ................... 175 F 6. Arlington County, Virginia spatial statistics summary ................................ ........... 177 F 7. Alexandria City, Virginia spatial statistics summary ................................ ............. 179 G 1. Business patterns and Educational Att ainment Factor in the Capitol Region, 1990 ................................ ................................ ................................ ................. 182 G 2. Business patterns and Educational Attainment Factor in the Capitol Region, 2000 ................................ ................................ ................................ ................. 183 G 3. Business patterns and Educational Attainment Factor in the Capitol Region, 2008 ................................ ................................ ................................ ................. 184
9 LIST OF ABBREVIATION S A&E Architecture and Engineering AHP Analytic Hierarchy Process CBD Central business district C PT Central Place Theory EAF Educational attained factor ECF Education conferred factor GDP Gross domestic product HC Horizontal clustering HHI Herfindahl Hirchman Index LBIO Literature based innovative output LIN Local innovative network LGC Locational Gin i Coefficient LQ Location quotient LUCIS Land Use Conflict Identification Strategy LUCIS plus Land Use Conflict Identification Strategy Planning Land Use Scenario ( Plus ) MSA Metropolitan Statistical Area NAICS North American Industrial Classification System OECD Organisation for Economic Co operation and Development PCA Principal component analysis RIF Regional innovation factor R&D Research and development SME Small to medium sized enterprises TREO Tucson Regional Economic Opportunities Inc. USPTO United St ates Patent and Trademark Office
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MEASURING REGIONAL AND LOCAL INNOVAT IVE OPPORTUNITY By Iris E. Patten August 2014 Chair: Paul Zwick Cochair: Margaret Carr Major: Design, Construction and Planning While some progress has been made in identifying cities that stimulate creativity to explain patterns of economic growth in k nowledge based industries, no existing research has used land use suitability to justify similar growth patterns. This study establishes the role and contribution of innovative agents, innovative inputs, and innovative outputs on the regional potential to attract architecture and engineering (A&E) firms as well as the ability to create innovative networks at the local scale using land use suitability. Using longitudinal interactions and the Land Use Conflict Identification Strategy to measure suitability and spatial inter actions two factors were created that identified the potential of Pima County, Arizona to create similar innovative networks found in Washington, DC. The results of this analysis indicates that by using longitudinal data and spatial data t o measure innovative potential, you can assess local and regional performance in achieving innovative networks.
11 CHAPTER 1 INTRODUCTION Cities that have the potential to attract companies with the greatest capacity to expand and attract employees whose s ole role is to innovate experience greater economic expansion than those cities whose economic base is composed of more manufacturing or agricultural related activities. Innovative industries derive their value from the networks created by the specific us e and arrangement of land uses that support their activities. In agriculture, this productive value is measured by the fertility of land. In manufacturing, productive value is measured by consumer demand for a specific good. In innovative or knowledge b ased industries, productive value is measured by the spatial arrangement of knowledge networks. Innovative labor influences which imported goods are substituted through local, more advanced activities. Jane Jacobs believed that innovative labor is a part of not add new kinds of goods and services, but continue only to repeat old work, do not of this phenomenon are exhibited in cities like Detroit or Cleveland, where local economies have faltered due to a continuation of business practices that encourage producing goods that are known to sell rather than producing new goods that encourage effi ciency and forward thinking. Furthermore, these cities also exhibit land use patterns that challenge the creation of innovative networks. As national and global markets become more unstable, cities supported primarily by non innovative or manufacturing re lated industries seek alternative market economies that serve as substitutes to mitigate negative economic impacts (DeVol et al . 2008).
12 Economic growth or development is a structural transformation of the economy in respect of both inter sectoral and int rasectoral employment and income. The process of growth is associated with the emergence of those sectors, institutions and basic attitudes as well as with such changes in structure and organization [sic] of the economy which would not only prevent the eco nomy from slipping back but would ensure continuous long run development. (Husain 1967, 40; 241) These additional sectors indicate a shift in the role of the city; in an industrial society, physical growth and land use are slaves to the instruments of pro duction whereas in a post industrial society where knowledge rules and economic activity relies heavily upon information, people care where they live and lifestyle quality has replaced many of the tradit ional determinants of location choice (Kotkin 1999; H earn 2008). Innovative industries create greater economic efficiency as a result of the et al. 2007, 214). Innovation is analogous to ecological succession. An urban economy, like a natural community, is constantly seeking to achieve an optimal range with low environmental stress. Change occurs unpredictably and several different patchworks of ecosyste ms essential contributions made within the conduit are created by diverse biological ts activities are environments that do not seek optimal efficiency. There are two general schools of thought regarding the criteria and development patterns of knowledge based industries and employees. First, unique resources such as density, access to ca pital, or sufficient labor pool places a premium on land in specific locations. These resources illustrate the first school of thought which is more focused on the actual production of new ideas and the importance of space in
13 urages spatially concentrated capital formation (buildings) and accentuates the need to produce at discrete points in space because of increasing returns to scale in production (Anas et al. 1998, 1427 28). Secondly, amenity based characteristics like open space, diverse populations, openness and acceptance, and culture influence how we value our communities. These amenities illustrate the second have, the more creative s (Florida 2008, 158). Unfortunately, local economic development strategies have used either school of tho ught to validate the other. Yet relying upon resources (natural resources availability, local, labor, capital investment, entrepreneurial climate, transport, communication, industrial composition, technology, size, export market, international economic si tuation, and national and state government spending) or capacity (economic, social, technological, and political activity) (Blakely and Bradshaw 2002) independently create weak economies that fail to attract businesses or yield an overabundance of low payi ng jobs. While some progress has been made in innovation geography (Ratanawaraha and Polenske 2007; Prosperi 2005) and proximity analysis (Gottlieb 1995; McHarg 1969; Carr and Zwick 2005; Collins et al. 2001) no existing research has used spatial land use conflict to justify growth patterns of knowledge based firms and land use opportunity that facilitates innovative networks. The available body of research that identifies amenities as a signifi cant influence on the location decisions of people and industr ies focuses on sociological factors as a means of determining regional advantages in the
14 creative economy (Florida 2005; Krupka 2004; Clark 2002). Other research identifies business factors that serve as an attractor of creative netw orks (Gottlieb 1995; H usain 1967). This paper points out methodological inefficiencies that could significantly influence spatial behavior. Explaining the relationship between economic and non economic factors of production in a knowledge based economy requires much greater e mphasis on the role of land use suitability in the development of networks that stimulate innovation and creativity. Accordingly, by customizing the Land Use Conflict Identification Strategy (LUCIS) (Carr & Zwick 2005), a geographic information system (GI S) model, to examine the causal relationship between business factors, amenity factors, and the intrinsic value of land a relative comparison of regions can be developed that reflects the attractiveness of cities for knowledge based firms based upon the cr eation of environments conducive to economic advantages for innovative industries.
15 CHAPTER 2 LITERATURE REVIEW In the history of economics, none is more unique than the creative economy; its industries and entire economy bridges factors of production wit h factors of consumption. Cunningham describes the difference between creative industries and the creative specific arguments to creative skill sets as occupational inputs into the broader economy, and creative outputs as intermediate 2008, 71). It is this intricate relationship that demands a sy stems based approach to understanding and determining economic opportunity. A systems based approach in regional economics is not new. The role of cities has traditionally been explained by the competitive positioning and wealth generation in subareas (Blakely and Bradshaw 2002). Innovation systems, commonly defined as et al. 2002, 233; Ratanawaraha and Polenske 2008), play a similar role to that of cities and are pa rticularly sensitive to spatial and locational attributes (Scott 2006). This review of literature will highlight research that addresses the structure of environments conducive to innovation creation and its importance in the economic structure of jurisdi ctions. Theoretical Perspectives of Traditional Economic Theory The Australian economist Colin Clark (1940) divided the pre industrial era economy into three sectors: primary (agriculture and mining), secondary (manufacturing), and tertiary (services). Es sentially, primary industries supplied raw
16 materials to be turned into finished goods by secondary industries, which are assisted in the tasks of distribution, finance and so on by tertiary industries (Hearn 2008). During nomy was measured by the degree of productivity (output per capita) in each sector. As productivity rose in each sector (i.e., fewer people were needed to achieve particular levels of output), labor could be transferred from one sector to another. Econom ic progress, thus, was defined as a function of the differential services wer ew of production and exchange was really innovation in action. Productivity gains during the pre industrial period was described by Schumpeter as methods, and product desi gns in favor of newer and more economically performative The post industrial economy can no longer be categorized into three sectors as Clark suggested, but additional sectors of the economy are needed and, depending upon the theo rist, the additional sectors differ. For example, Bell (1973) suggests adding tertiary (transportation and utilities), quaternary (trade and finance), and quinary (health and education) sectors that fuse theoretical knowledge with science and technology ( Bell 1973, xiv). Barry Jones (1982) proposed a quaternary sector covering employment in information processing and related activities (Hearn 2008, 176). This
17 difference in additional sectors illustrates the new economic paradigm. The new economy springs from the shift in natural resource based activities to human resource based activities. In a human resource based economy economic advantage is based upon the ability to transform an import into a new exporting enterprise through innovation; the core of performance in the new economy can be predicted by three key factors: geography, the entrepreneur/innovator, and rates of productivity. Through systems of innovation, the productivity of the new econom y is optimized. Industrial Performance in the New Economy: Geography Anas et al. insights into economy wide growth processes and sheds light on economic concepts of long standing i nterest: returns to scale, monopolistic competition, vertical integration, broad classes of theories help us understand the process of local economic development: loca tion theories, which focus on geographical factors, and economic base theories, which look at the flow of economic activity into and out of the local economy to identify and explain which firms and industries have the greatest capacity to and Bradshaw 2002, 57). Economic base theories prescribe a e of their locational attributes in relation to to optimize their economic base. Both are needed to optimize the economic potential of a region.
18 Prior to 1840, the shape o f cities was tied directly to the location of natural features such as waterways or rivers which provided a means to transport goods over long distances. This created substantial scale economies at these transport access points and the average cost of pro cessing freight fell sharply with the quantity produced at a particular port. A similar phenomenon occurred with the advent of the railroad for smaller quantity goods and the telegraph for transferring information between cities (Anas et al. 1998). For t he first half of the twentieth century technological improvements and advanced methods of transport were credited with increased complexity of urban structure. Technology favored decentralization within metropolitan areas, with the automobile contributing the most to this out migration. Land use patterns shifted with improvements in technology; retailers could follow their customers and manufacturers could follow the labor force. Spatial theory also evolved during this period. For example, Burgess devel oped his Concentric Zone Model in 1925, during the time when the segregation of work and residence was normal and considered acceptable urban form. In 1939 Hoyt determined that land use patterns were not randomly distributed, nor were they a series of nea t concentric circles. Instead, land use in North America could be defined as sectors dependent upon access to transportation and communication (Figure 1 2). Following this realization by Hoyt, Harris and Ullman of urban land uses. [They concluded] that many towns and nearly all large cities do not grow around one [central business district] (CBD), but are formed by the progressive integration of a number of separate nuclei in ome specialized and differentiated in the growth process and were not located in relation to any distance attribute, but were bound by
19 accessibility, land use compatibility, land use incompatibility, and location suitability. This was the first example of a polynuclear structure of urban activity. Yet this structure was not a reflection of technological and transportation improvements rather it was a reflection of a shift in the economic relationships between firms as well as a shift in the organization o f work. The role of manufacturing is central to the process of economic growth and its role changes significantly as economies develop. Chapman and Walker use the analogy of a motor car to illustrate the changing role of manufacturing in a growing econom y. the economic vehicle which it drives, but there is no sign of revolutionary change which nomic growth or development is a structural transformation of the economy in respect of both inter the essence of growth has been calculated from the relative decline of agr iculture in the economy both as a source of employment and income. More recently, economic growth has been calculated from the relative decline or reshuffling of manufacturing in the economy. Over the past forty years manufacturing has been in a state of transition. manufacturing sectors of the economy was more rapid than for manufacturing, giving rise to the notion that the United States was becoming a service oriented economy. This is all true, but until 2000 the phenomenon was more about growth in the service sectors, and not about the decline emergence of those sectors, institutions and basic attitudes as well as with such
20 ch anges in structure and organization [sic] of the economy which would not only prevent (Husain 1967, 241). The growth in economic sectors since Clark indicate a shift in the r ole of the city; in an industrial society, physical growth and land use are slaves to the instruments of production whereas in a post industrial society where knowledge rules and economic activity relies heavily upon information, people care where they liv e and lifestyle quality has replaced many of the traditional determinants of location choice (Kotkin 1999; Hearn 2008) contemporary urban planner that economic growth influence s spatial structure. This model is shaped by location al requirements of specific industries, often traced to eras of economic growth, and reflects the post industrial expansion of economic sectors beyond primary, secondary and tertiary. Parr believes th at the agglomeration of (Parr 2003, 11). The evolution of spatial structur e during the twentieth century indicated the point at which economic theory and the physical structure of cities caught up with each other. intangibles as reciprocity, exchange, from an industrial to a post industrial economy. Adam Smith believed that economic growth and these intangibles were r ealized by increases in productivity yet limited by
21 occasion to the division of labour, so the extent of the division of labour must always be limited by the extent of the m The departure from spatial theory in understanding the new economy is due to the relationship between markets and productivity. This relationship is not perfect, partly because location is affected not only by costs of product ion but also by demand and the Walker 1987, 8). This conflict is addressed by locat ion attempts to set orthodox microeconomic production theory within an explicitly spatial framework. These analyses then attempt to explain how changes in spatial economic (McCann 2002, 3). These models demonstrate that the optimum location of the firm and the optimum production relationships are interdependent and co determined (McCann 2002). The problem with most early location production models is that they are one dime nsional, just like early spatial theory. In these cases, modes of transportation and quantities shipped are fixed. Other assumptions include the fact that location does not 2, 12). Given a profit location model, market failure is inevitable. This is explained by neoclassical economic theory. Neoclassical theories revolve around utility and profit maximization. Two major concepts underlie neoclassical economic theory: equilibrium of economic systems and
22 become equal at an equilibrium that leads to an optimal allocation of resources (Buitelaar 2007, 4). Since markets are rarely fully competitive, the allocation of resources is sub optimal, since a competitive equilibrium cannot be achieved. The result of a non equilibrium system is market failure (Buitelaar 2007). Viewing the fir m as a rational maximizer is a psychologically unrealistic behavioral assumption (Buitelaar 2007; McCann 2002; Chapman and Walker 1987). minimized, with emphasis given to the trans port costs involved in assembling materials at the manufacturing site and in delivering the finished product to the market. This is probably the most significant limitation of the Weber model seeing that typically transport costs tend to be only a very sm all percentage of total costs for most firms influence of labour costs and the possibility that economies may be achieved as a result of the agglomeration of several plants in Walker 1987, 19). This problem can be overcome through substitution, which is not privately controlled infrastructure, if the exis ting infrastructure is poor, can diminish the impact of poor mobility. The problem with most substitutions is that they are often more expensive or less satisfactory. Many economists acknowledge that Weber was the most influential early contributor to nor mative location theory (Chapman and Walker 1987). Normative theory considers basic objectives for industrial location in hopes of determining the best or
23 models that have gr eatly contributed to urban spatial structure have been inductive. Christaller propose d the Central Place Theory that described the distribution of development in terms of a hierarchy of centers, each with a hexagonal market area (Anas et al. of different industries are merged to form a regional system of cities. The [CPT] is a model of market oriented firms, as firms that base their location decisions exclusively on 20). The model is not oriented to the location of inpu efficient spatial allocation of production and housing in the presence of economies of scale and transport costs. Economies of scale encourage the concentration of production at a rela tively few locations, each of which achieves an efficient level of output. Transport costs encourage the adjacent location of production and population (residential housing) to minimize shipping and commuting costs. Production (especially agricultural pro duction) and housing require contiguous space, so they are concentrated in any one location only up to the point at which space requirements increase transport Collectively, thi s model introduces the concept of urban systems in location theory. Shortly after Christaller described the CPT, the German economist August Losch proposed that forces more fundamental than local resource based differences were at work, which accounted for that the goal of the rational entrepreneur should be to select the location at which profits
24 are maximized and that, in reality, neither demand nor costs are spatial constants as assumed by the le ast (Chapman and Walker 1987, 19). This is seen in contemporary Chicago. Fyfe and Middle West; location at the s outhern end of Lake Michigan is a secondary factor. If there were no Lake Michigan, the urban population of the Middle West would in all Although fundamentally important, the CPT is fla wed in that it is customer driven, spatially limited, and restricts economic growth. Christaller and Losch both assumed a homogenous geographical landscape with a uniform geographical population distribution; assumptions that encourage spatial inefficienc ies. Consider this, from 2000 to 2006, emerging suburbs and exurbs grew nearly three times as fast as the U.S. population, as Americans moved further out in search of more affordable or bigger homes (NJ Report 2008). Developers saw cheap land and rising home prices and residents saw a slew of federal and private loan programs that would help them achieve the American Dream. To fulfill this dream, national investors invested in areas open to outsiders flooding the local market with capital. For investors , the costs and barriers associated with risky deals across regions in locations too far away to benefit from regional economic development and distributive equity increased spatial inefficiencies. For homeowners, costs associated with commuting were unde restimated. When the price of gasoline rose to record levels during the summer of 2008 the ability to reduce combined with a lack of adequate transportation infrastructure, wh at could be perceived
25 Romo 2008, 2 3). Therefore to create a truly polycentric structure that encourages economic growth and optimizes the intangibles created through eff icient geography, it is not simply a reorganization of firms that needs to occur, but an efficient reorganization of labor as that is what helps to support strong markets. 1999a, concern is with the flow of goods and services between the central place and market. lower levels, so that upward trade flows from urban [centers] are wholly absent from the spatial economics in a service sector environment or outside of manufacturing, the urban system is dynamic, requires backward and forward linkages, and is independent of fixed transportation routes to define its market. An additional flaw is that the CPT holds the view that efficient pricing of public facilities alone will make land use patterns more efficient thereby saving resource lands for resource uses and facilitating efficient and Duncan 1995, 113). Yet in reality, public facilities use an average cost pricing structure, where everyone is charged equally for the same service, regardless of the real cost to provide that service to a particular customer. This pricing structure e ncourages sprawled development patterns where higher density developments subsidize the cost of services for outlying low density developments. In a
26 spatial context, by measuring the conflict between land uses or development goals (i.e., what is to be acc omplished) all costs are considered, including transaction costs. spatial theory, especially with regard to the systems based approach for understanding the spatial relat ionship between employment sectors and markets. According to McCann, from this theory we can glean two distinct components of functional urban systems: 1) the use of centrality to determine the locational patterns of economic activity and 2) the impact of a diverse set of influences on specialized function activity (McCann 2002). Industrial Performance in the New Economy: Geography Jacobs, Blakely and Bradshaw, and others use economic base models to support the need for an economic class whose sole role i s to create a stable economy and quickly adapt to any disturbances. This economic class has a national or international market presence and takes advantage of economic opportunity zones that contribute to the expansion of economic activity or expands the markets for goods. Modern economic base theories concentrate on the transactions within an economic system as the engine behind wealth and regional growth. Innovation and the new economy rely upon reducing the failures and inadequacies of those transacti ons maximizing internal institutional linkages in the public and private sector (Blakely and Bradshaw 2002). Since the time of Adam Smith, productivity has been used to measure economic performance. One of the primary indicators used to measure economic health is the represents the total dollar value of all goods and services produced over a specific time mic Co operation
27 and Development (OECD) Science, Technology, and Industry Scoreboard 2007 contrasts with the late 1990s when investment in knowledge outpaced growth of GDP (OECD 2007). This trend is projected to continue, especially in sectors that fuse theoretical knowledge with science and technology ( Table 2 1 ). According to employment projections by the Bureau of Labor Statistics, almost all of the fastest growing in dustries between 2006 and 2016 are in high tech or knowledge intensive industries. Such industries require a high share of high advanced skills (Figure 2 1 ) . To support future industry growth in these sectors additional investment of knowledge is require d and firms often spend more on knowledge based workers as a percentage of their total wages expenditure than other sectors (Hearn 2008, 72). The OECD defines investment in education (public required to catalyze it, is the primary determinant of competitive advantage in the current economic landscape, and that this can involve the re definition of industry industrial society innovation and globalization are the new sources of job creation and have led to major developments in al l sectors of the economy. Unlike the traditional industrial sectors, the success of information based industries is not directly based on monetary returns and relies heavily upon largely undirected, spontaneous, and unmotivated collaboration (Hearn 2008, 1 79). As Hearn
28 in such an economy will be dominated by the sale or licensing of intellectual property, n 2008, 180). example, Silicon Valley is regarded as a region that has successfully executed the complex process of knowledge transfer. The economic base in this r egion is an innovative network of high value products and services based on its proximity to an excellent source of research, principally at Stanford University, and the clustering of high technology companies (Brown 2007, 1). To achieve a technological r evolution, which is the product of an innovative network and of which productive value is gained, the importance of space must be realized. Lundvall et al. (2007, 216) conclude that g run, suppress the importance of space, while radical innovation makes face to face The impact of location on a technological revolution is dista nce; on the other hand, the radical change taking place within the technology itself gives a privilege to agents interacting face to face and with opportunities for hands on et al. 2007, 216). In Silicon Valley, the focus from a tradi tional input output supply chain linkage to a wide range of collaborative relationships, and the capacity to maintain considerable knowledge flows and innovation between organizations represents a shift in industry spatial organization from the individual firm to a productive system (Fingleton et al. 2007, 62).
29 they require, and the competiti ve pressures that they face. Firms understand that innovation, and the knowledge required to catalyze it, is the primary determinant of competitive advantage in the current economic landscape, and that this can involve the re definition of industry bounda ries and positions of market dominance on, historically speaking, rapid timescales. The wealth of information made available at low cost by the Internet and through increased exposure to international trade should guarantee that those involved in enterpri se (who are able to effectively compete) accept and understand the importance of innovation as it relates to their own opportunities, risks, Innovation is a market based activity. As Hearn states (2008), innovat catay[z]ed by science and other non test of innovation is whether market selection has acted to generate a new source of economic value (usually understood as a new source of profitability). Anoth er way of saying this is that science [and information] produces knowledge that does not necessarily lead to innovation. It is only when demand for a solution based on new knowledge has been realized or created that innovation can be said to have occurred (Hearn 2008, 166). Furthermore, innovation and creativity are analogous terms. Given the definition of innovation stated above, creativity is defined by Robinson (2001) as uires Creative Task Force in Britain adds to this definition that creative industries possess eneration and et al. 2008, 60). For this paper,
30 innovation is the extension of knowledge in creating something new and of value. Creativity is the intersection of science and innovation. Creativity appli es that new knowledge in a manner that is understood and valued by a wider audience. The Urban System As described above, industrial location has become more complex as technology has improved. The transition from an industrial to a knowledge economy has been products and services, rather than deriving greater efficiency and economies of scale nt interests of a region are often not in the interests of a given locality within that region, and vice versa. Economic strategies are often localized and the spat ial distribution of benefits bear little resemblance to the spatial distribution of costs (Parr 1999a, 1262). This explains why although regions such as Silicon Valley in California and Research Triangle Park in North Carolina are centers of rapid economi c growth, not every helps to explain why some areas are increasingly advantaged whil e others are disproportionately disadvantaged. The cumulative causation theory regards economic change as an endogenous phenomenon in the market system (Fujita 2007). This logic follows positive feedback. forces, by their nature, pull capital, skill and expertise to certain areas. These areas accumulate a large scale competitive
31 Schumpeter (1934) states that economic deve lopment cannot solely be described by the narrow view provided in traditional economic theory. Economic development is not a phenomenon to be explained economically, but that the economy, in itself without development, is dragged along by the changes in the surrounding world, that the causes and hence the explanation of the development must be sought outside the group of facts which are described by economic theory. (Schumpeter 1934, 63) Helen Brown (2007) describes innovation in terms of three stages: r esearch, refer to the culmination of all three stages, from knowledge creation to diffusion of , 2). Collectively, these three stages are an innovation system. et al. 2002, 233). Innovation systems can be viewed in terms of the physical or geographical dimension and in terms of time. The spatial and traditional economic system necessary to support an innovative economy goes against the existing tradition of scientific practice; or normal science. Innovative economies support scientific r evolutions. Thomas Kuhn (1996) states that innovation is shattering complements to the tr adition The geographic or physical dimension of innovation systems is an important dimension. Sometimes the most important dimension in economic development is the industry which then determines the geogra phic boundary but the physical dimension of innovative economic opportunity gains value through finding a location where innovation
32 roduct development must take place in areas with greater wealth and capital to invest in the process of inventing and developing new products, supported by local markets that can pay higher prices for products that have Ratanawaraha and Polenske (2007) indicate that three primary issues affect innovation geography analysts. The first issue concerns the question of whether innovation, however defined, is spatially conce ntrated or dispersed in certain areas. Once analysts confirm that an innovation is spatially concentrated/dispersed, the second issue they investigate concerns the various mechanisms and factors that underlie such a distribution and concentration/dispersi on, such as regional resource endowments, knowledge spillovers, and industrial organization. The third issue is how the spatial concentration/dispersion affects other variables, such as regional and national economic growth (Ratanawaraha and Polenske 2007 , 44). Through December 2007 the United States continued to steadily produce jobs but geographic inequalities remained. Across the nation a spatial mismatch exists between the people seeking work, the location of work, and the resources needed to stimulat e job creation (Blakely and Bradshaw 2002, 10). Blakely and Bradshaw believe that global market demands, obsolete plants, uncompetitive industries, locational disadvantages, inferior skills, and racial discrimination further contribute to spatial inequiti es. Even in a free market system reducing spatial economic inequality requires intervention. Litvak and Daniels state that, Regardless of how it manifests itself, the existence of relatively depressed communities in substate regions means a certain segme nt of the population is cut off from the fruits of national economic development. People in these localities will not simply migrate to healthier areas. On the contrary, better educated people with more promising job prospects are likely to move from pla ce to place looking for employment. Moves by poor people tend to be within the same county or city. Clearly there is a need to try to bring jobs to people rather than counting on people to move to jobs (Blakely and Bradshaw 2002, 11).
33 Martin and Feldman argue that a unique jurisdictional advantage is available in cities that provide positive economic value through knowledge spillovers, clustering, and believe that true j Trans it and transportation corridors, recreation facilities, crime, and land values are all affected by aggregate county level decisions and events. Furthermore, counties facilitate polycentric spatial patterns that provide a more organized structure for idea exchange and knowledge absorption. An urban economy, like a natural community, is constantly seeking to achieve an optimal range with low environmental stress. Change occurs unpredictably and several different patchworks of ecosystems work together to fo rm a community. In Richard umulating changes in temporally, a snapshot of the system at a particular point in time may differ substantially from another snapshot of the same system at a different time. A comparison of two distinct time periods offers insight into system performance. One argument against innovative firms is that although the result yields a more efficient or productive good, the length of time associated with creating that new good can b e lengthy. But as Schumpeter points out, the length of production is irrelevant. I n
34 an industrial based economy consumption drives production and profit and changes in production is driven by incremental changes. In an innovation based economy consumpti on adapts itself to run at the rate of production (Schumpeter 1934, 36 37) and application, is seldom or never just an increment to what is already known. Its assimilation requires the reconstruction of prior theory and re evaluation of prior fact, an intrinsically revolutionary process that is seldom completed by a single man and never Carlsson et al. define the elements of the innovation system as follows: are the operating parts of a system. Relationships are the links between the components. Attributes are the properties of the components and the relationships between et al. 2002, 234). perspective focuses on emer concentrates on the object of activity. These perspectives are summarized in four stages. The first stage involves the formulation of the problem or idea. The second stage involves modeling a publicl y transmittable representation of this. In the third stage, the model is tested in simulation or real life. Finally, in the fourth stage, the model is consolidated and stabilized in the acquisition of new practices (Brown 2007, 8). From the perspectives summarized above, four general components of the innovative production process can be defined: innovative networks, innovative agents, innovative inputs, and innovative outputs (Ratanawaraha and Polenske 2007). These components are activities due to lab or. The first concept, innovative networks , is what
35 differentiates innovative activities from any other production activity (i.e. manufacturing or agriculture). Innovative networks describe the entire system and add value to not only the production proce ss but through the specific use [and arrangement] of land productive value is achieved. In agriculture this productive value is measured by fertility. In manufacturing productive value is measured by consumer demand for a specific good. In innovation, pr oductive value is measured by the spatial arrangement of knowledge networks. Innovative economies support a research agenda rooted in economies of agglomeration. The knowledge networks create a premium on land in produce at discrete points in space because of et al. 1998, 1427 1428). An innovative system is simply a reformulation of the four components of economic development: locality, business and economic base, em ployment resources, and community resources (Table 2 2) new growth theory, relies upon endogenous factors to measure economic growth through technological progress. Very few sectors of the economy continue to base land use theory upon the relationship of markets to proximity of fixed physical transportation many people and industries, digital connections seem to trump geography as a gauge In attraction theory, communities are products. As communit ies improve their quality of life they tend to attract new populations, which may improve their level of
36 for innovation (Blakely and Bradshaw 2002, 66). The linkages be tween locational advantage and the built environment derive from the planning theory of Jane Jacobs. She asserts that creativity originates from the built environment, which encourages the arrangement of unique economic dynamics. More specifically, Jacob s states that preserved and increased through a proce ss of human physical development that is history of economics, none is more unique than the creative economy; its industries and entire economy bridges factors of produ ction with factors of consumption. 69). It is this complexity that challenges the econo mic development professional with creating a plan that integrates the concepts of innovation and the needs of a community. publicly financed research is organized in a give schooling, training and financial institutions. Production of economically useful new technological knowledge results from collective actions of different actors of the system connected by various linkages ranging from informal to formalized network et al. 2002, 1070).
37 Components Husain states that The phenomenon of growth, though defined in measureable terms requires a set of tangible and intangible forces making for change. The former are usually the material forces of production and the latter consist of the environment causing and facilitating such a change. These two forces are, however, not independent of each other. They mutually act and react in such a manner that the process becomes cumul ative. But the distinction between material forces and the environment of production is not necessarily the same as between economic and non economic factors. It is not easy to classify the factors of growth into economic and non economic since the forme r include the material agents of production as well as a part of environment in which economic decisio ns are made. (Husain 1967, 67) Development is caused by economic factors of production as well as facts described outside of traditional economic theory (Schumpeter 1934, 63). The economic forces provide order in the larger environment. This order is formed by credit institutions, fiscal and monetary policy, and price levels. The non economic forces effects economic decisions indirectly. These forces are defined by the culture, values, customs, institutions, social structure, status system, etc. of the people (Husain 19641967, 72). When comparing two companies in the same field in two different cities from the outside the companies may seem similar or Economically, the two companies seek the same thing to survive in their respective economic environment, access to capital, purchasing power to support research and development activities, and access to a customer base. I n reality the differences between these two firms and their eventual success or failure can be traced to the causal relation that exists between the economic and non economic factor. Economic factors are based upon logic. Non economic factors explain the nuances generated by the uniqueness of isolated communities. For example, Schumpeter (1934) uses the
38 example of ground rent and explains that the variations across a given space are due to lar price movements to political regulations of commerce, then I have done what I can as an economic theorist, because political regulations of commerce do not aim immediately at the acquisition of goods through exchange or production, and hence do not fal l within our concept of In 2006 two rural Virginia cities, Lebanon and Rose Hill, in desperate need of an economic boost were given access to high speed Internet (Kang 2009). One town attracted two large compan ies, creating 700 jobs with average salaries of $50,000 for residents. The other town was unable to attract any large companies and Internet access resulted in the creation of only a few home based businesses. The difference between these two towns was th e additional investment one city undertook to educate its residents and prepare them for positions which utilize the Internet and technology to do business. The willingness of the local government in Lebanon to train its workers and voluntarily increase t he educational level of its workers is what attracted Northrop Grumman and software ma ker CGI to this rural community . According to the Washington Post, about 71 % compared with Rose Hill, where only 29 % do . In Rose Hill the education gap and lack of training of its residents have presented challenges in attracting new industry. At face value locating in either of these rural areas would present an opportunity to offer lower ly competitive firms recognize that the firm and the Bradshaw 2002, 69). Furthermore, regardless of the industry, a strong human resource
39 base is a major attractor and (Blakely and Bradshaw 2002, 69). As in the case of Lebanon and Rose Hill an initial investment by governm ent was needed to create a desirable environment for commerce thus stimulating the economy. about adjusting, influencing, and regulating the action of other owners of us er rights. Institutions are the rules of the game in a society or, more formally, the humanly devised Virginia Governor Mark Warner helped in getting $2.3 million in gran ts to bring fiber optic pipes to homes and business parks. In Rose Hill, the costs of laying the fiber lines and building cell towers was funded from a state tobacco settlement fund for broadband projects and a rural telecommunications program run through the Department of Agriculture (Kang 2009). Both of these funding efforts facilitate increased human interaction and increase the productive value of land. Although, the production costs in dge based or otherwise. A review of literature by Acs et al. equally distributed in space. Production of new scientific and technological knowledge growth pole strategy model proposes a shift in the relationship between firms and locations (usually as part of a deliberate effort to modify region al spatial structure), in an
40 attempt to encourage economic activity and thereby raise levels of welfare within a engine to development that subsequently attracts a c omplex set of industries needed for industries would cause growth to diffuse thr oughout a specific zone of influence (Parr 1999b, 1198) . The principles behind a natural growth pole strategy support the fundamental arguments for a creative economy: when the intrinsic value of land supports exploiting the potential advantages of a loca tion for innovative, highly competitive industries there will be a cumulative buildup of the various types of agglomeration economy that create beneficial inter sectoral and inter industry externalities (Parr 1999a; Parr 1999b; Perroux 1955). Although th e Christaller approach represented the first comprehensive outline of a hierarchical urban system (McCann 2002), the difference between historical perceptions of market area and the growth pole strategy is that Perroux represents an open urban system with spatially flexible supply areas. A market area represents the territory served by a particular urban center (McCann 2002); suggesting an equilibrium there commodities withou t complements. The sellers of all commodities appear again as buyers in sufficient measure to acquire those goods which will maintain their consumption and their productive equipment in the next economic period at the level so (Schumpeter 1934, 9). The household or firm will respond
41 tightly as possible to habitual economic methods and only submit to the pressure of circumstances as it becom es necessary. Thus the economic system will not change capriciously on its own initiative but will be at all times connected with the preceding state of (Schumpeter 1934, 8 9). Conversely, a supply area is a territory that serves a particular urb an center . Within a supply area ideas and new products are generated that improve efficiency by demand side economics and have the benefit of quickly adjusting to negative feedback or emergency adaptations thus reducing market instabilities created by realistic behaviors exhibited by firms and individuals. Components: Innovative Agents In McClellan entrepreneurial fash sain 1967, 70). Taken a step further, innovative agents are those who produce innovations that expand development. Economic progress is achieved because of individuals who understand where inefficiencies exist and have the capability to synthesize existi ng practice and expertise and relate them to the need for in spirit and are ready for experimentation and putting hitherto untried technique into practi 1 967, 236). These are innovative agents. of the individual entrepreneur. Innovation reflects cumulative processes of interaction
42 where different organizations and individu als combine efforts in creating, diffusing, and et al. 2007, 214). Madric believes that with each major period of growth the same sorts of factors contribute to economic expansion. Recently, in addition to a reduced cost of com puter power and an expansion in how products are used, the suppression of wages and new attitudes toward business and entrepreneurialism contributed to the expansion during the 1990s (Madric 2002). Human capital, or people as economic agents, affects the geography and population composition of the system. People with marketable skills highlight the or a cluster cannot function without labor and history has shown that the a vailability of a educated young people, in particular, are more likely to have mastered newer technologies and ideas. Many view these younger workers, who typically have fewer family responsibilities, to be more entrepreneurial and risk The perce ntage of college graduates reflect general human capital. Such broad measures accept that innovation is not restricted to laboratories and research centers; a workforce with general skills is better prepared to provide the flexibility necessary to meet op 110). More precise as a percentage of the population. Due to the mobility of highly educated human
43 capital, states that are unable to retain or recruit such skilled workers never reap the full benefit of their economic value. Regions with significant levels of human capital reinforce the importance of a supply area in determining the spatial geography of an inn ovative system. Husain (1967) and Hearn (2008) indicate that investment in human capital creates more progressive and affluent consumers in stimulating innovation on the part of the producers and services providers and facilitating the adoption of inventi development requires a change in the composition of skills and attitudes of the people. There arises the need for a transformation of illiterate, backward and ill fed people to illiterate, skilled, healthy, efficient and trained workers; fro m the attitude of asceticism and empty spirituality to that of achievement and success; and from resistance to when the rules run out or when there are no rules in the fir a concentration of intelligent people is important, but in terms of our economic future creativity is about more. Components: Innovative Inputs Innovative inputs are the process or efforts used in creating knowledge. Ther e is a lack of understanding in the study of innovation, which involves innovation diffusion and diffusion, analysts must not consider something after it has been innovated, but where, i.e. geographical place the innovation occurs, and under what conditions. If all innovators and people using the innovation were clustered in the same location, those who write about tacit versus codified knowledge would blend with the tran smission of
44 between declining sectors and companies and expanding ones. Abo ve all perhaps, standardised knowledge and 56). The central role of tacit knowledge is the process of learning through interacting, and this method overextends the tools whose use neoclassical economists regard as the most fundamental criterion for remaining scientific (Lundvall et al. 2007). The meeting and interaction of people in the workplace through which idea s can be negotiated, transferred or developed is crucial for development. More specifically, knowledge and innovation is developed through interaction in the workplace itself (Power and Lundmark 2003). This interaction is evident in the cluster. Feser et al. manifestation of the mutually reinforcing influences of first mover effects, conventional business agglomeration economies, localized technology spillovers, and geographical path d marshal diverse resources and programs behind groups of related industries that have et al. 2001, 1). With respect [that] generate sustained competitive advantage for the firms and institutions within them by: increasing productivity through innovation by making information about new
45 opportunities more widely available; and promoting new business formation in related based clustering can be explained using five indicators: 1. The importance of specific forms of labor [sic] input, a nd the quality of each specialized labor [sic] and associated forms of tacit knowledge; 2. The organization of production in dense networks of small to medium sized enterprises (SMEs) that are strongly dependent upon each other for the provision of special ized inputs and services; 3. The employment relation in creative industries, which is frequently characterized by intermittent, project based work, meaning that recurring job search costs can be minimized through co location in particular areas; 4. The ind irect, synergistic benefits that result from the interaction of individual creativity with collective learning, tacit knowledge and historical memory, through the co existence of people and enterprises engaged in inter related activities; and 5. The devel opment of associated services and institutional infrastructure, and the priority that the relevant industry sectors have in the thinking of local and regional governments. (Hearn 2008, 65) from the firm to productive systems and an understanding of the phenomena of competitiveness as a et al. 2007, 62). In a knowledge economy, less tangible variables, such as th e quality of life overshadow advantages in market area or natural resources (Blakely and Bradshaw 2002). Although the exportable commodity plays an important role in supporting an es in the level of (Arora et al. 2000; Hall 2007) indicate a clear association between knowledge workers
46 (individuals with high levels of human capital) and location pref erences; particularly with those places that have a higher than average quality of place. Financial resources are indisputably an essential component in the process of creating knowledge. Particular geographical areas demonstrate higher than average perfo rmance in research and development due to access to primary sources of government investment. Financial capacity for innovative activities is typically measured by research and development (R&D) investment. In the science and engineering fields this info rmation is available from several sources, yet for other industries this information may be more difficult to measure. Therefore, Hall mentions that financial capacity of private knowledge can be measured by industry R&D spending and public knowledge sour ces are often measured in terms of university R&D expenditures (Hall 2007). Unlike science, which generates ideas but does not guarantee the creation of new economic value (Hearn 2008, 164), economic value is applied to information through innovation. In ). Innovation exceeds convention; the inadequacy of tradition d rives the development of new practices (Brown 2007). Yet even the best intentions surrounding plans to maximize economic conditions through technological the technolog
47 the economic syst em; technology only develops productive methods for goods market institutions; they both contribute to the growth of knowledge and act as mechanisms that further innovation. Research and de velopment activities possess potential economic value. As mentioned previously, human capital is an integral part of innovation and in many cases the innovator is an individual who can use a process or service in a more economically efficient manner, but that does not make the innovator an entrepreneur. Often the entrepreneur leads production into new channels (Schumpeter 1934, 89). to pursue R&D goals. Aside from traditional government investment, another means of funding is through venture capital sources. Venture capital investment is often used to capital does not represent capacity to innovate but, rather, capacity to commercialize. Economic growth follows commercialization, which supports the distinction between Components: Innovative Outputs Innovative output is evidence of what is produced by knowledge based activity. innovation p rocess: the commercialization of technical ideas as tangible products (Polenske 2007, Acs et al. 2002). Previous research by Ratanawaraha and Polenske (2007) indicates that patent data is the most widely used data in innovation geography measurement (20 07, 36). This is due to the ease of access to such data, the ability of
48 researchers to conduct both longitudinal and cross sectional analyses with this data, and because patent data extends beyond just the actual count but it identifies locational special ization. According to Acs et al. (2002), indicators such as patents may misrepresent true output because firm size may influence the need to announce new products. But Acs et al. fairly reasona ble measure of innovative activity at the industry level, and some evidence that patents and innovations behave similarly at the state level however, this has not et al. 2002, 1071). Other methods of determining inno vation counts are through literature based journals for new this method over patents include that they docume nt the commercialization of every innovation but these types of counts are very expensive to produce and typically only represent select years across select geographies. Relationships Initiative, creativity, problem solving, and openness to change are inc reasingly important skills in a knowledge economy (Houghton and Sheehan 2000, 9). The market and physical environment contribute to creating an atmosphere where innovation can thrive. This is why cities like New York, San Francisco or Chicago have advant ages in (Florida 2008, 135). Knowledge workers seek high returns socially and financially for y and capital flow to where the returns are greatest, and people move where opportunity lies. Both
49 capital and talent concentrate where opportunities for productivity and returns are The criteria and development patterns of kn owledge based industries and employees are based upon two principles. First, unique resources such as density, access to capital, or a sufficient labor pool places a premium on land in certain places. These resources illustrate the first principle, which is more focused on the actual production of new ideas and the importance of space in economics. Secondly, amenity based characteristics like open space, diverse populations, openness and acceptance, and culture influence how we value our communities. Th ese amenities illustrate the (Florida 2008, 158). Unfortunately, local economic development strategies have used either principle to validate the other. Yet relying upon resources (natural resource availability, location, labor, capital investment, entrepreneurial climate, transport, communication, industrial composition, technology, size, export market, international economic situation, and national and state government spending) or capacity (economic, social, technological, and political capacity) (Blakely and Bradshaw 2002, 55) independently create weak economies that fail to attract businesses or yield an overabundance of low paying jobs. et al. 2002, 234). A review of lit rank[ed] both labour supply and quality of life as top location factors, raising the
50 1414). Amenities contribute to the attractiveness of a commuter shed for a firm, where employees live (Gottlieb 1995; Gottlieb 2004; Florida 2005). Although Gottlieb was among the first to recognize the relationship between amenities and business locat progressive from previous scholars, particularly with general audiences, who chose to ignore the relationship of residential amenities in the attraction of the firm. Employees of high technology firms demand more of residential amenities when considering where to locate. Therefore high technology firms within metropolitan areas are more likely to consider residential amenities in their firm location (Gottlieb 1995). Traditional economic factors of pro duction represent endogenous causes of grow and stay competitive and therefore create high value jobs need: 1. cost effective access to reliable transport, energy, and telecommunications infrastructure; 2. to operate in a competitive environment in terms of input markets; 3. access to finance on terms and conditions appropriate to their needs; 4. a stable macroeconomic environment, appropriate company, commercial an d intellectual property law, and minimum (by the standards of global best practice) business regulation, and red tape; and 5. Access to an appropriately skilled and experienced labor [sic] force (or human capital base). (Morrison and Potts 2007, 169) Ite ms 1, 2, 3, and 4 above relate to the inputs necessary to create an appropriate environment and affects the acceptance of outputs created by innovative firms and agents outside of the immediate market area. Item 5 is necessary to incite the flow of
51 knowle dge. Furthermore, to strengthen metropolitan economies Bruce Katz of the Brookings Institution suggests that the above factors with quality places can have significant implications on the form and functions of a region by creating a strong and diverse mid dle class with the potential to grow in more environmentally sustainable ways (Katz 2008). Companies that innovate provide the functional environment in which new ideas can flow but without the people to manipulate these inputs, ideas are simply free advic e. Individuals that possess the education or acumen to fully develop a new product from an idea are a scarce resource, and as such must make tradeoffs when choosing a community that satisfies their intellectual needs. Innovative companies realize this, as do the communities in which they locate. Charles Tiebout outlined a powerful framework for identifying the tradeoffs involved in choosing a place, rooted in the local value of public goods. Therefore if innovative agents, who place a high value on ameni ties, are presented with several employment choices Tiebout argued that their final decision will reflect a more localized fitness advantage: aesthetic appreciation (Jacobs 2000; Florida 2008). based goods or services that enter the utility amenity maximizing agent, thus cities and firms use amenities as an economic development strategy for three reasons: 1. Executives consistent ly rank both labor supply and quality of life as top location factors, raising the possibility that amenities are viewed as a separate factor, 2. Firms may locate in high amenity areas, not only to tap an existing labor for ce, but also to recruit a new one. The elite firm may act as an amenity maximizing
52 agent, blurring the distinction between residential and non residential location behavior. 3. Amenities presumably affect firm location through compensating wage differentials . If the price rather than the quantity of skilled labor is key, then a focus on migration or labor supply could be misleading (Gottlieb 1995, 1414). Gottlieb developed a method for spatially weighting amenity variables for high tech employment concentra tions. His research concluded that at the municipal scale corporations are agglomerative and there was evidence of amenity optimization over hypothetical commuter sheds, supporting the hypothesis of the firm as an amenity maximizing agent. In his researc h only the most distressing disamenities matter at the place of work, while surrounding residential locations must pass a higher amenity standard. Gottlieb developed the rel at ional illustration in Figure 2 2 for the location of an amenity oriented firm. Typical location theory, illustrated by the solid lines, considers the existing labor force as the primary attractor. Business factors contribute to firm location and amenities draw population to a region. New residents appeal to firms looking to reloca te and with the addition of employment opportunities and the expansion of business sectors employees possessing specific skills are drawn to the area. This model also indicates that the relationship between the firm and quality of life is ignored, the das hed arrows. After developing this model, Gottlieb realized that amenities had to be considered in firm location since amenities are a priority to the quality and type of employee high tech innovative firms seek. Gottlieb studied a region in northern New Jersey connected by mass transit, highways, and a diverse economy. The area had an abundance of skilled professionals with the highest per capita income and housing costs (Gottlieb 1995, 1416). Gottlieb
53 acknowledged the polycentricity of the region and t use regulation and the provision of services. He used six business variables that highlighted the locational concerns of the high tech or professional service sector and several amenity variables. The Gottlieb study excluded tax rates, utility rates, and land costs because he believed that the high tech sector was less sensitive to these variables than manufacturing firms. These variables were excluded because he believed that these costs were influenced by the attributes (i. e. amenities) of that municipality and since these attributes have been included as individual factors, to include tax rates and land infrastructure factors such as the number of college graduates, municipal employment density, and access to transportation. Generally, Gottlieb used proximity analysis, reflected in the positive or negative coefficients on the independent variables represented in his firm location function. Spatiall y, Gottlieb included variables like distance away from minority residential neighborhoods (to measure where employees are less likely to live), distance away from congested roads, proximity to recreational opportunities, and proximity to local public serv ices and public education as values that locate. To measure the response of high tech firms to amenities and agglomeration in their location decisions, Gottlieb used a c ombination of a gravity formula to determine the weighting of business variables to explain consumption behavior and the negative exponential formula in conjunction with questionnaires to determine the weight of amenities and density gradient of high tech workers around the facility where they
54 worked (Gottlieb 1995). To measure the spatial effects of amenities Gottlieb uses distance decay parameters to determine the coefficients of his empirical model. The more significant problems with this model include imprecision in determining the distance decay properties of consumption or physical geography without collecting massive amounts of data or conducting a large detailed study. Measurement Polenske (2007) identified three main issues in current literature a bout innovation. First, there seems to be a bit of difficulty in defining innovation. Secondly, there is a lack of consensus on a framework both to define the theory of innovation and the ways to measure it, and by the vast number of empirical studies th at are done, but using relatively simplistic measures. Lastly, as mentioned previously, there is a lack of understanding in the study and measure of innovation, The experimental design of innovative geography projects involve determining two primary fac tors: geographic unit and measuring relational innovative activity. Apiwat Ratanawaraha and Karen R. Polenske completed an extensive literature review of the spatial measurement of innovation. In their research, they concluded that the three primary obs tacles in the study of innovation geography are limited data availability, inadequate theoretical models, and conceptual precision (Ratanawaraha and Polenske 2007, 30). Together these hindrances prevent the development of ices, test (Ratanawaraha and Polenske 2007, 30). These obstacles are created due to the lack of existing innovation geography research that details data and underlying mechanisms for attracting innovation. Additionally, Ratan awaraha and Polenske note that previous research is ineffective in identifying and measuring the indicators that explain the spatial
55 distribution of innovation. Although traditional location theory is used to measure the agglomeration/dispersion of innova tion it is insufficient. Geographers, planners, and researchers should cons ider the trade offs in measurement that result from separating the product that is created from the people and process needed to develop a new idea. It is relatively easy to understa nd the theoretical framework behind innovation and the four general components of the innovative production process (innovative agents, innovative inputs, innovative outputs, and innovative networks). What is more difficult is designing an equivalent fram ework for temporal and spatial measurement. Ratanawaraha and Polenske developed a comprehensive survey of data and sources for each step of the production process (Appendix A). As mentioned previously, the value of land is determined by unique resources in certain places and amenity based characteristics dictate how we value our communities. From Table 3 we see that the study of innovation geography to this point has occurred at a very gross scale, the census tract or larger. Yet within an innovative sys tem microclimates exist that may transcend standard political or geographical boundaries. The dilemma is attempting to measure value gained from inputs, outputs or agents beyond a predetermined scale. A greater difficulty lies in obtaining comparable empi rical data that illustrates a discrete measurement of economic or social value. Land use suitability analysis is a method that merges the temporal and spatial aspect of measurement and provides a value reflective of localized and global actors. Anas et al . provided a review of various techniques used to measure urban structure at finer resolutions. One approach uses fractals, or geometric figures which display ever
56 fractals to exami ne the irregularity of the line marking the outer edge of urban development in a particular urban region. The fractal approach highlights the inadequacy of a deterministic view of development in accounting for the irregularities in et al. 1998, 1433). A more intuitive approach used by urban employment density exceeding some minimum D [density], and together containing total employment exceeding some et al. 1998, 1434). This method does not consider irregularity and scale dependency of employment patterns and agglomeration clustering. the sources of their economic growth lie within themselves, in the processes and growth systems that go on within them. Cities are not ordained; they are wholly existent ial. To are the result of a gestalt approach; there is an intrinsic suitabi lity of land uses that dictate urban form. Growth responds to natural processes, which are clearly visible in the pattern and distribution of development in its density (McHarg, 1969, p. 160). Using scenario modeling and suitability surfaces, Ian McHarg developed methods that relied upon the land to dictate its best sites for development (McHarg, 1969, 197). The core of characteristics lend to maximizing future utility and optim al use. Conflicts between the
57 values of multiple uses was resolved either through a decision made by an actor or co existence of multiple uses. His suitability model also deviated from more absolute models of land use assessments and value analysis. It is easy to understand the establishes a method to understand the relative importance and value of systems and processes. The original McHargian process was rational and expl icit meaning that the process and results were reproducible. Land use suitability analysis is an analytical process that combines inventory information to determine the fitness of a given unit of land. The result is either tabular data, a single map o r a series of composite maps that display the relative suitability [or appropriateness] for a specific use (in siting studies) or a number of uses (in comprehensive planning) (Randolph 2004, 591). As landscape architects in the late 1800s, Charles Eliot a nd Warren Manning used suitability analysis in their environmental planning pursuits to measure the relative degree lands in Boston were fit for integration into the Boston Metropolitan Park System. Central to this process was developing a systematic appr oach to inventory site resources and, through the use of overlay mapping, analyze the natural fitness of the land . The various approaches to between human and natural p rocesses. Some are innovative and sensitive to the future, while others are the repackaging of the same approaches under different names An extension of traditional suitability analysis is lan d use conflict. An example of a land use conflict method is the Land Use Conflict Identification Strategy (LUCIS),
58 developed by Professors Paul Zwick and Peggy Carr at the University of Florida. use suitabilities are in conflict or where one land Zwick 2007a, 26) LUCIS stops short of representing alternative futures, but instead focuses on the comparison of the results of three suitability analysis purposefully designed to capture biases inherent in the motivations of three stakeholder groups: conservationists, developers, and farmers and ranchers dedicated to an agricultural future (Carr and Zwick 2005). The evolution of suitability analys accurate, legally defensible, technically and ecologically sound, and open to public even more important because determin ing the optimal use of the landscape for embracing an understanding of changing economic circumstances: the supply and demand of land, varying human needs and values, pol itical realities, and new (Ndubisi 2002, 142). Planners, economists, and decision makers can be so deeply involved in maximizing their economic base that trade offs of that performance goal are not considered, even when these trade offs are highly relevant to social well being, as is the increasing possibility of gentrification. A broader perspective supported by appropriate data and decision support tools is needed in order to have livability given a serious consideration in planning and to have it viewed as a legitimate part of the set of goals to be served by creative employment decision making (National Research
59 Council 2002, p11). Suitability analysis is a spatial measurement of the intrinsic suitability of land. It provides a method of allocation that recognizes the social, economic, and political values people place on specific land uses (e.g. creative employment or commercial/service employment). According to t he National Research Council this effort can be hampered by several factors: 1. Addressing the complex issue of liveability requires access to a wider variety of information than is traditionally used by the various planning organizations. 2. Communities n eed to be able to measure whether their actions are improving liveability, but they often lack necessary data and face challenges in developing sound methodologies. 3. Organizations and stakeholders often do not have consistent or comparable data, making t he analysis of options and decisions more difficult. 4. The information needed to make good decisions may not be available in useable forms. Land use suitability analysis provides a quantitative and qualitative medium to measure innovative economic potenti al and accounts for structural, residential, and non residential location behavior (i.e. quality of life). LUCIS specifically, and suitability analysis in general, resolves the complex association in measuring the relationship between markets and producti vity. Geographers, planners, and researchers should consider the trade offstradeoffs in measurement that result from separating the product that is created from the people and process needed to develop a new idea. Measurement: The LUCIS Approach : The LUCI S Strategy is a six step suitability analysis process. The first step is the development of a hierarchical set of goals and objectives for each stakeholder group that become suitability criteria or ind icators of measurement (Figure 2 3 ). The stakeholder groups are identified as land use types or categories integral to the development process. Generally, the grossest
60 scale for measurement and group categorization is agriculture, conservation, urban. The hierarchical procedural tree, developed by Alexander and Mannheim in 1962, was used as a method for combining factors. This is an important aspect of suitability analysis as it helps to organize the components of the environment that are important. Generally, a systematic assessment of the environment in volves a three tiered framework: I MPACT V ARIABLES . Identifying components of the environment that are important (e.g., water quality) I MPACT I NDICATORS . Measures that indicate change in an impact variable (e.g. dissolved oxygen) I MPACT T HRESHOLDS OR S TAND ARDS . Values of impact indicators above or below which there is a problem; used to evaluate the impact (e.g., 6 ppm minimum of dissolved oxygen) (Randolph 2004, 613) The first tier, Impact Variables, is analogous with the LUCIS hierarchical goals, object ives, and sub objectives. Husain (1967) understood the interrelated nature of economic and non economic factors of production. LUCIS goals lend to this interaction. By identifying the multiple purposes within each stakeholder, the complexity of producti on can be measured. For example, amenity based characteristics influence how we value our communities. Within the LUCIS structure, the values of amenities are listed as parallel goals with productive value. The intent of each goal is dichotomous, econom ic and non economic or physical. These are represented by objectives. The variables that contribute to achieving the intent are identified as sub objectives. LUCIS goals, objectives, and sub objectives allows for a clear method to organize indicators of innovation opportunities, reduce uncertainty, and explain the inter dependence of explanatory factors. The hierarchical structure separates considerations
61 of cost, in terms of real estate prices and costs of production, from value, in terms of desirable l ocal attributes or amenities given cost. The hierarchy provides a relational structure for measuring the transaction costs in determining creative economic production. Tr ansaction costs are the costs that are made to increase the information 31). In suitability analysis this is accomplished through the measurement of suitability for each component factor. As uncertainty is reduced a framework is created that illustrates the comparative advantage of knowledge based firms integrating the influence of amenities, as a measure of quality of life, and resources needed to encourage the production of new ideas. The s uitability of objectives and sub objectives contribute to the overall goal productive value. The hierarchical structure lends to concluding that the explanatory factors of economic and non economic factors of production, as illustrated in terms of distinct goals, are inter dependent. Conversely, the complexity of many aspects of knowledge based activity creates difficulties in identifying components of the environment that are important. The second step of the LUCIS strategy is to develop an inventory of available data that best demonstrate the suitability of the feature(s) identified in each objective or subjective. The second step of LUCIS is analogous to the second tier, Impact Indicators, of systematic assessment. Economic performance is typically m easured by discrete numeric indices, as in the case of GDP. According to some, the material well being measured by GDP is not
62 monetary values to the various fact ors and intangibles that comprise a wider measure of socio economic well create non monetary indices of social and economic well and arbitrariness in the factor s that are chosen to assess quality of life and, even more suitability analysis the importance of the factors chosen that contribute to innovative systems can be measured statistically and through existing and historical land use wages in order to attract labour independent of local amenities. Similarly, with land of a given accessibility i n less than perfectly elastic supply, high wage levels boost real Leven 1992, 738). Using suitability analysis you can spatially measure the inter dependence of land v alue and the suitability of land for creative employment. Using a historical dataset you compare the location of new creative employment and change in future land value as it relates to the influence of explanatory factors by performing a sensitivity anal ysis on the appropriate suitability surface. Returning to the example of GDP, critics suggest that using GDP illustrates a clear and substantive meaning to quality of life and that prices are objective weights for the goods and services that make it up ( taxes, tariffs, or speculation l ocal economies can randomly contract. Since suitability is
63 a measure of the intrinsic value of land the data used to measure a feature neglects any artificial influences. Determining creative employment suitability introduces various difficulties in spati al measurement. As Husain stated above, economic forces are influenced by tangible and intangible economic factors. Examples of such economic forces include taxation and government spending. In spatial suitability, impact indicators are represented by a geographic relational database. The first issue that arises is related to data availability. Determining spatial suitability is not only dependent upon understanding what factors contribute to measuring explanatory factors of fitness but also having ac cess to spatial data to complete the measurement. For example, in the context of fiscal and monetary policy the data used in measurement would include descriptive information about the rate of taxes and associated descriptive and topological information i dentifying spatial geography. This introduces issues related to scale and data availability. If spatial data is available the scale of fiscal and monetary policy is gross; its spatial extent is larger than a municipality or county. Also, the impact of a fiscal policy over a given area is typically uniform and local sensitivities between the economic (human) or natural processes are difficult or impossible to measure. Additionally, to perform suitability analysis on a factor with such a gross scale would not provide any useable or direct suitability results. Factors unavailable in a geographic or spatial format would be used to facilitate the synthesis and analysis of relationships typical of the gestalt method. The gestalt method is based upon direct o bservation and views the site as a whole. In contemporary suitability analysis, including LUCIS, this is associated with community values or weighting.
64 The third step of the LUCIS strategy makes use of methods employed by Carr and Zwick, including proxi mity and statistical analysis, to measure the suitability of a specific unit of land within the region with respect to the values and bias of each stakeholder. The third step of LUCIS is analogous to the third tier, Impact Thresholds or Standards, of sys tematic assessment. Depending upon the intent of the objective and/or sub objective, GIS models are developed, which are a workflow of spatial data and geoprocessing tools that measure suitability in terms of utility value. LUCIS employs a standardized v alue range of 1 to 9, with 1 representing low suitability and 9 representing high suitability. Once the suitability of each objective and/or sub objective has been determined, individual layers within a goal are combined using utility values (i.e., weight s) that equal 1.0 (100%). The result is a single GIS raster layer that illustrates the final suitability for a specific stakeholder group: agriculture, conservation, or urban. These suitabilities are combined to create land use preference. The Impact T hresholds or Standards tier of general assessment identifies variables given two general indicators of change (or influence). The first is based upon proximity. This idea is based upon the principle of diminishing or increasing return; either how far awa y from or closer to you are to a particular impact indicator. The second indicator of change is based upon the idea of with between mutually different levels of suitability. For each factor, criteria are established to identify the minimum acceptable ratings which serve as cut off points. All sites with a rating below the minimum acceptable rating fail or are unsuitable, while all sites with a
65 1991, 15). Most variation in the suitability surfaces can be explained either statistically or due to the indicator of change (i.e. proximity or with without). A significant part of determining creative employment opportunity over a given area employs the inclusion of non economic factors that are based upon qu ality of life standards, which are often subjective. These standards are reflective of personal and community values and within suitability analysis are best used for deriving weights for various quality of life objectives. The weights are determined usin g a pair wise comparison method. Carr and Zwick calculate community preference using the analytic hierarchy process (AHP). AHP is a non general form of pairwise comparison that assists in setting priorities and makes the best decisions when both qualitat ive and quantitative aspects of a decision need to be considered (Expert Choice 2008). Combining spatial data is best completed using map overlay. Warren Manning introduced the map overlay concept in 1912 and since then it has been instrumental to the pr technique is often used in conjunction with a weighting scheme. The weight values are 2001 , 61). As described in the description of third step of the LUCIS Strategy, individual objective suitabilities within a goal are combined using weights representative of community values and goals are combined using weights reflective of the values from t he AHP process. Surveys have also been used to gauge community values. For
66 determining the importance of the future location of growth exercises such as visioning sessions, which link questions about personal preference for future growth to a solid con nection with the landscape through either maps or other visual aids, encourage participants to think realistically about their goals and how to achieve them. Suitabilities related to quality of life create difficulty in determining community values. On the one hand you could conclude that repeated future patterns of development indicate an environment that is amenable to residents. On the other hand you feel in an enviro satisfaction surveys as a determinant of quality of life; thus can be applied as weights to ask people the simple question of how satisfied they are with their lives in general. The simple measure of life satisfaction [as opposed to surveys of the related concept of happiness] has been found to correlate highly with more sophisticated tests, ratings by Life satisfaction is seen as a judgment that depends on social and culturally specific frames of reference, which is exactly how community values are formed. Unfortunate ly, community values, if existing policy is not used as a guide, can often introduce bias into the larger suitability process. Depending upon w ho defines the end of t he gradient is an approach that looks to citizen stakeholder groups to define internally consistent narrative assumptions about how future land [use] will unfold. The citizen driven approach produces alternative futures that typically have the advantages
67 of integral citizen involvement, greater political plausibility and increased likelihood of et al. 2004, 326). The disadvantage of this approach is that it is difficult to statistically aggregate their preferences into a s maller number of driven approach, with experts in the bio physical and social sciences or planning professions defining a set of decision or transition rules, often with input from other groups, that determine future land [use] conditions. The decision rules are generally constructed to optimize for particular endpoints or illustrate focal policy options (e.g., improved water quality, better wildlife habitat, lower infrastructure costs, less highway c ongestion, etc.). Alternative futures produced using this approach typically have the advantages of quantifiable statistical likelihood and the disadvantages of unclear political plausibility, which may be due to the encoded decision or transition rules l ying outside the political processes actually et al. 2004, 326). The fourth step of the LUCIS strategy combines goal suitabilities to represent stakeholder preference. Carr and Zwick (2007b: 14) define prefer suitability raster developed in step three for each stakeholder has values that range between 1 and 9, but may not include the value 9. For a value o f 9 to result, at least one cell in the study area would have to be optimally suited for every measure of suitability included in the goals, objectives, and sub objectives for that land use category. The probability of this occurring is very low. Carr an d Zwick (2007b) recommend normalizing final suitability values before comparing preferences. Normalization only occurs on areas with development potential.
68 for each ras ter cell. For these multi attribute problems, the individual cell size of the raster representation of a given geographic region that results in larger raster databases is not a limiting factor for the procedure because raster cells can be processed indep argue that scale is an issue in multi objective land use allocation because two or more non en tirely true. Take the case of determining individual suitabilities within a goal. The objectives are aggregated to the goal level and the resultant suitability surface comprises a common extent with unique cell values based upon mathematical calculations . But in land use conflict, the non complementary activities (i.e. individual stakeholders) are linked together by individual preference value. The fifth step of the LUCIS strategy reclassifies the preference of each stakeholder into three classes that correspond to high, medium, and low preference. There are several methods available that produce an even distribution of preference values. The distributed values are characterized using a designation of 1, 2, or 3, which describe the level of preference (Table 2 3 identifies relationships among the three stakeholders. The sixth and final step to develop a conflict surface compares areas of preference to determine the quantity and spatial distribution of potential land use conflict. As stated above, the collapsed preference surface is characterized using values of 1, 2, or 3. To compare the preferences of each stakeholder the collapsed preference surface is combined into one GIS raster layer. This G IS raster layer is known as the final
69 conflict surface. The values in the final conflict surface ranges in value from 111 to 333. The first, second and third digit in each number sequence is representative of agricultural, conservation, and urban prefere nce, respectively (Figure 2 4 ) . The LUCIS conflict surface illustrates which lands a single stakeholder has the highest preference value, which lands two stakeholders have the same preference value (i.e., moderate conflict), and which land all three stake holders have the same preference value (i.e., severe/major conflict) (Table 2 4 ). The value in LUCIS is the land use conflict feature which identifies locations for specific land use opportunity. This feature sets LUCIS apart from other suitability models in that other models gain their value in the assumptions used or the role of lay and expert stakeholders in the scenario modeling process. With the LUCIS Conflict table, differences between stakeholder preferences are identified before synthesis products are developed. Land use conflict is a synthesis of the outcomes of the conflict provides spatial context insight into resource competition. The relative conflict values illustrate opportunity costs associated with choosing specific geographic units of land relative to other land uses. Although this research does not allocate actual creative employment centers, it is important to note that allocating development based upon land use conflict introduces the concept of differential rent. For example, if more land is needed for urban use than is available in urban preference and the ass umption is made that agriculturally preferred
70 land can be used to satisfy urban need then the rent of agricultural lands with a higher land use suitability increases. Conclusion Creating an environment or system conducive to innovation is about more than d eveloping new products. Innovation and the activities it spurns are a source of growth economic and physical. The structure of the innovation system is inextricably linked to markets, education, asset value, and land use patterns. Madrick sums it up ni cely, capital investment easier and less expensive to undertake, just as growing savings helps facilitate growth. A growing economy motivates people to educate themselves and make the public investments in transportation and communications often necessary for further growth, just as more investment in education raises the potential rate of ne arly as multifaceted as the attributes associated with measuring the performance of the system.
71 Table 2 1. The OECD/EC classification of knowledge intensive sectors Styles Apply to High Technology Aerospace (17 2011); Pharmaceuticals (29 1051); Computers , office machinery; Electronics communications; Scientific instruments Medium high technology Electrical machinery; Motor vehicles; Chemicals excluding pharmaceuticals; Other transport equipment; Non electrical machinery Medium low technology Coke, refi ned petroleum products and nuclear fuel; Rubber and plastic products; Non metallic mineral products; Shipbuilding; Basic metals; fabricated metal products Low technology Other manufacturing and recycling; Wood pulp, paper products, printing and publishing ; Food, beverages, and tobacco; Textile and clothing Knowledge intensive high tech services Post and Telecommunications; Computer and related activities; Research and development Knowledge intensive market services (excl. financial intermediation and hig h tech services) Water transport; Air transport; Real estate activities; Renting of machinery and equipment without operator, and of personal and household goods; Other business activities Knowledge intensive financial services Financial intermediation, e xcept insurance and pension funding, except compulsory social security; Activities auxiliary to financial intermediation Other knowledge intensive services Education; Health and social work; Recreational, cultural and sporting activities Less knowledge i ntensive market services Sale, maintenance and repair of motor vehicles and motorcycles/retail sale of automotive fuel; Wholesale trade and commission trade, except of motor vehicles and motorcycles; Retail trade, except of motor vehicles and motorcycles/r epair of personal and household goods; Hotels and restaurants; Land transport/transport via pipelines; Supporting and auxiliary transport activities/activities of travel agencies
72 Table 2 1. Continued Styles Apply to Other less knowledge intensive servic es Public administration and defense/compulsory social security; Sewage and refuse disposal, sanitation and similar activities; Activities of membership organizations n.e.c.; Other service activities; Private households with employed persons; Extra territo rial organizations and bodies (Source: Miles, I. (2008) Knowledge Services, Knowledge Policy: Challenges for the 21st Century, 14 15) Table 2 2 . A reformulation of the components of local economic development Component Old Concept New Concept Locality (Innovative Inputs) Physical location (near natural resources, transportation, markets) enhances economic options. A quality environment and strong community capacity multiply natural advantages for economic growth. Business and economic base (Innovative Inputs) Export base industries and firms create jobs and stimulate increased local business. Clusters of competitive industries linked in a regional network of all types of firms create new growth and income. Employment resources (Innovative Agents) Mor e firms create more jobs, even in many are minimum wage. Comprehensive skill development and technological innovation lead to quality jobs and higher wages. Community resources Single purpose organizations can enhance economic opportunities in the communi ty. Collaborative partnerships of many community groups are needed to establish a broad foundation for competitive industries. Source: Blakely and Bradshaw 2002, 67, Table 3.2
73 Table 2 3 . Preference v alue d escriptions Cells with a value of Indicate 1 Low preference 2 Moderate preference 3 High preference Table 2 4 . Three example stakeholder prefer ence value explanations. Conflict Value Preference Value Stakeholder Preference 333 Major Conflict Agriculture, Conservation, and Urban have the same preference value 113 No Conflict Urban has the highest preference 122 Minor Conflict Conservation and U rban have the same preference value
74 Figure 2 1. Skill composition of economic sectors in the European Union, 2000. (Recreated from Miles, I. (2008) Knowledge Policy: Challenges for the 21 st Century, 16)
75 Figure 2 2 . Illustrates the relationship b etween residential traditional business factors, residential amenities, population growth, and firm location and employment growth. (Source: Gottlieb, P.D. (1995) )
76 Figure 2 3 . LUCIS hierarchical structure
77 Conflict Value: 2 1 3 Figur e 2 4 . LUCIS preference Agriculture Preference Conservation Preference Urban Preference
78 CHAPTER 3 METHODOLOGY The ability of a single new idea to seamlessly integrate itself into multiple industrial sectors is the product of a knowledge economy. Houghton and Sheehan characterize the emergence of the knowledge econ as simple as opening up a few engineering firms or research facilities and suddenly an explosion of economic growth occurs . Again, the purpose of this research is to identify physical and economic factors of urban economics that leverage land use suitability in the location of in novative industries. The outcome of this research is to develop two relative measures one local and one regional that reflect the opportunity available for innovative agents, in terms of people or industries, to locate to a particular region. While so me progress has been made in identifying cities that stimulate creativity to explain patterns of economic growth in knowledge based industries, no existing research has used land use suitability to justify similar growth patterns. The available body of re search that identifies amenities as a significant influence on the location decisions of people and industries focuses on sociological factors as a means of determining regional advantages in the creative economy (Florida 2005, 76). These factors include diversity in local and regional populations, parks and open space, and economic growth is associated with tight knit communities where people and firms form and sh traditional factors of production that result in an increase in economies of scale. Florida
79 a fundamental way from more traditional factors of production like land or raw materials; they are not fixed stocks, but transient flows. Technology and talent are highly mobile plaining the relationship between amenities and neo classical definitions of economic factors of production in a knowledge based economy requires much greater emphasis on the role land use suitability has played in the development patterns of knowledge bas ed industries. Earlier research on the determinants of the motivation of innovative firm location and the preferences of their employees provided the framework for subsequent analysis (Donegan et al. 2008, Florida 2008, Florida 2002, Gottlieb 1995, Marku sen 2006). Nevertheless, attention was limited to discussing the relationship between traditional economic and spatial theory and the changing spatial relationships needed for firms to thrive in a knowledge economy. As spatial data that more accurately i llustrates physical features becomes more readily available, using spatial analysis methods that account for the intrinsic properties of land enable more detailed assessments of land use suitability. Nonetheless, little research has utilized land use suit ability analysis to evaluate the determinants of innovative firm location. Traditionally, the focus of economic growth has been at the national level, however units def ined as clusters of industrial activities or alternately as regions. The literature suggests that economic growth is a local process and that cities [and counties] are an important, if not most important economic unit, in generating new development,
80 compe development is a slave to the activities that occur within the counties and cities that operate within them. 2 2 ) he believed that firms on ly responded to a pre existing labor force and quality of life was ignored in the location of an amenity oriented firm. This study proposes that the pre existing regional labor force, amenity value, and demographic trends create local environments conduci ve to the cre ation of new knowledge (Figure 3 1 ). Furthermore, understanding the suitability or appropriate opportunities for specific land use types aids in evaluating the tradeoffs associated with the location of employment centers at the local scale an d influences more local location decisions. Theoretical Study Area: Metro Washington, District of Columbia Defining the scope of a region through urban structure is important in evaluating social and economic performance and provides a framework for policy intervention (Parr 2003, 14). The study area consists of two distinct geographic areas: metro Washington, DC and Pima County, Arizona. The metro Washington, DC region in this study consists of seven counties in Virginia and Maryland (Virginia: Alexandr ia City, Arlington County, Fairfax County, Fairfax City, Falls Church City; Maryland: Montgomery County, and the most Inc. 500 fastest growing private businesses in the U.S. and has been for the years the Capitol region has experienced a significant shift from primarily federal employment to high tech industries, occupations, and products . According to the National Science Foundation, the metro Washington region leads the country in
81 government research and development spending per capita, beating out Beijing, S Washington Initiative 2010). Maryland ranks second among U.S. states with regard to federal R&D obligations, ranks sixth in terms of federal R&D expenditures to academ ic institutions, and ranked fourth in the number of Small Business Innovation Research (SBIR) awards granted between 1995 and 2000 (Feldman 2007, 242). Comparative Study Area: Pima County, Arizona Unlike the metro Washington DC megaregion, Pima County, Ari zona has yet to be recognized as a major player in the larger national or western economy. The county seat of Pima County is Tucson and, although unofficially, is part of the Sun Corridor megapolitan region. Recently, more studies have suggested the grow ing benefit for regional cooperation between Tucson and Phoenix. Even though the Sun Corridor would provide the largest concentration of economic power in the eight states of the Intermountain West (Sonoran Institute 2010), this study considers only Pima County. In addition to providing a basis for comparison for the regional comparison measure being developed, the outcomes of this study will support or reject further discussion concerning the need or benefit of regional economic development. Also unlike the Washington, DC region the Pima County region is self contained within one county. Pima County contains four municipalities and an unincorporated area. The municipalities are: Marana, Oro Valley, Sahuarita, South Tucson, and Tucson. The county has ex perienced a steady rate of population growth for the past 20 years with significant growth in the municipalities of Sahuarita and Marana, primaril y due to incorporation . The Town of Sahuarita was incorporated in 1994 and originally
82 covered approximately 9 square miles. Today, the town boundaries encompass more than 29 square miles. The Town of Marana was incorporated in 1977 but after starting out as a town of about 10 square miles and a population of about 1,500 townspeople an aggressive annexation polic y and access to water from the Colorado River in 1992 has supported additional growth and burgeoning populations. Today Marana covers almost 120 square miles and has a population of over 34,000 people. Researchers from the Sonoran Institute (2010) believe larger regional economy is to provide an environment conducive to small and medium sized high waged industries. Tucson offers a lower cost of living that cities such as Phoenix or those within the metro Washington, DC region an d by catering to smaller enterprises local planners believe that fewer infrastructure improvements will be needed. Compared to the Capitol region who ranked 4th, in 2008 Tucson ranked 55th in population with a little more than 1 million residents (Sonoran Institute 2010). Among all the attributes that contribute to an innovative system the current status is not very promising in Tucson. Between December 2006 and December 2008 Tucson experienced a higher job loss rate, 2.8%, than the national average, 1.5 %. The average annual pay lags behind the national average and other cities in the West and the percentage of Tucson workers with a college degree continues to be low by comparison to other western cities (Sonoran Institute). In 2007, the Tucson Regional Economic Opportunities Inc. (TREO) released a blueprint identifying five areas needed to develop a more competitive economy: high skilled/high wage jobs, educational excellence, livable communities, urban renaissance, and collaborative governance and stew ardship.
83 All of these areas are included in this research study as key aspects of developing an innovative system. As stated previously, the metro Washington, DC region is the baseline to which Pima County will be compared. Questions may arise concerning the scale of comparison, given that the innovative propensity of eight counties will be compared to the lone county of Pima. First, although Pima County could benefit from joining the Sun Corridor and leveraging the stronger Phoenix economy, there is stro ng political will Therefore, there will need to be an overwhelming motivation for these two mini regions e larger metropolitan city of Tucson and the suburbs of Oro Valley, Marana, South Tucson, Sahuarita, and Green Valley. Although these geographies are much smaller than a county scale, the relative measure of innovation and methods outlined below really il lustrate the interaction between places. Lastly, the following two chapters, Analysis and Discussion, will outline the regional and local effects of innovation in the metro Washington, DC region and discuss whether similar patterns are visible in Pima Coun ty and its suburbs. Understanding the Innovative System: The Actors Since growth pressure is directly related to pop ulation and employment, given an employment that facilitates new ideas, this study asserts that firms will follow the agents that create the se ideas thus enhancing the regional economic climate and subsequently attracting new firms and people to the area. This study analyzes this relationship specifically for architecture and engineering firms. Ar chitecture and engineering are two sectors of i ndustry as old as the world itself. Creating a seamless connection between the practical functions, the purpose and
84 relationships of its various parts, and structural principles is the nature of building and design. Architecture and engineering are among a short list of industries in which an artistic expression has an infinite benefit across disciplines and applications. Kimball fusion of expressions in a single build ing involves the sacrifice of many others, and is a These fields associate practitioners and researchers with specialists of other knowledge and expand the application of traditional measurement (Pinson 2004). For data collection and standardization purposes, firms included in the architectural and engineering sector are those included in the NAICS category 541300. This level of sectoral data is available at the state, county, and metropolitan statistical area from the U.S. Department of Labor Bureau of Labor Statistics. Nationwide, the average annual salary for these occupations in 2007 is $ 66,280, which is significantly above the average American income in 2007 of $ 50 ,233. According to the Greater Washington Initiative, which includes a larger study area than that discussed in this research, there are over 71,000 employees within architecture and engineering occupations ( Appendix B ). Among these occupations, which ran ge from landscape architects and cartographers and photogrammetrists to materials and electrical engineers, the average is $77,930. Nationally, the value added (in current dollars) of A&E has grown almost 100% since 1998 (Bureau of Economic Analysis 2010). Analysis 2010). Over the next ten years the wage and salary growth of arc hitecture and
85 engineering industries will continue to rapidly grow, while industries associated with manufacturing will continue to decline (Appendix C). A sample of A&E firms in both metro Washington and Arizona were obtained from three different sources: Brian Spivak at Arizona State University, Site to do Business online (www.stdb.com), and Dun and Bradstreet data. Data from the first two sources are based upon information gathered from InfoUSA. For the sake of standardization, unless otherwise noted, total A&E population estimates will be based upon this data. To illustrate the saturation of A&E firms in the metro Washington region, Table 3 1 illustrates the number of A&E firms per capita compared to that of Pima County. Understanding the Innovative S ystem: Methods This research project integrates qualitative, quantitative, and spatial methods to accomplish methodological triangulation. This type of mixed method research involves d Gaber 2008, 136). More specifically, this project is a between method investigation, which 2008, 136). The research methods described below use descriptive, predictive, an d evaluative analytics that reduce the complex system of an innovation economy into elementary units. These elements facilitate the development of a detailed analysis to understand the types of interaction between spatial and aspatial determinants. Additi onally, this research examines the innovative network of the metro Washington, DC region and through spatial analysis and descriptive statistics determines the temporal and spatial significance of possible explanatory factors for innovative agents, inputs, and outputs of A&E firms. Once explanatory factors for innovation have been stated, the innovative capacity for Pima County will be
86 determined. Given the factors that encourage innovation in the Capitol region, this treatment will be applied to Pima Co unty to determine whether its impacts the innovation potential of A&E firms across the region and locally. lin 2008, 63). Gottlieb cautions the transferability of the results in a model such as the one described in this study. It is easy to find a region with similar characteristics as the New Jersey region Gottlieb studied and transfer his conclusions regard ing the relationship of amenities, business factors and firm location to another area but he warns the political climate significantly influences development (Gottlieb 1995, 1416). Economic expansion through innovation can only be achieved by creating act ivity systems that facilitate the cluster or agglomeration of clusters of particular industries by investing in attributes that make it particularly attractive to firms and talent in those clusters. Not all cities and regions will succeed with similar str ategies. The key lies in diversity, not copy cat strategies to attract the most current vogue industries (Feldman and Martin 2004). Additionally, the usefulness of data is also determined by the spatial extent of information available. For example, the s tudy area for my research is the Capital County (Maryland), Montgomery County (Maryland), and Fairfax County (Virginia). For data available (e.g. High Tech Manufacturing Employment) available at only the Metropolitan Statistical Area (MSA) scale, the data would not be useful for two reasons: (1) the Washington MSA includes the counties in my study area but is also comprised of
87 three additional counties in Maryland and nine additional counties in Virginia; and (2) the data at such a gross scale would not allow me to understand the sensitivities that exist within my study area due to the influence of the High Tech Manufacturing Employment data. Data Independent variables were chosen from readings of the survey literature on amenities based location, econometric literature on intra metropolitan firm location, lit erature for high technology firm location, and the literature on the place based location preferences of scientists a nd engineers. The goal was to select business and amenity variables studied previously and variables that matter in business geography decisions. Research results, implications, and conclusions are most directly affected by spatial units. The spatial comp onent in this research project relies heavily upon geo spatial data at a parcel or smaller scale to capture the complex relationships between dependent and independent variables. Therefore, to take advantage of the micro and macro scales of data that meas ure innovation, during analysis that integrate both economic and amenity based factors data will be summarized up to the smallest common spatial unit. Analysis in this research project will be assessed using a raster based GIS. In raster GIS, decision pr oblems involving multi objective or multi attribute determining the opportun ities for creative employment. analysis but is significant in drawing conclusions about the effects of your data on whatever you are testing. For example, land value is a factor that will be represented by
88 a continuous raster surface. Each raster cell in that surface will have a unique land value. Additionally, the geographic location of cultural and entertainment facilities is currently available as a static vector format. In my analysis, land value and cultural and entertainment facilities will be evaluated together creating a common scale is imperative in analysis. This common scale is achieved by using the same raster cell size for all data within my analysis and ensuring tha t all data represents the same spatial extent. For discrete locations that may not have continuous values over the entire surface, cells outside of those with a value representing a specific feature is given a dummy value so that in the end all datasets us ed are on the same scale; each dataset has a representative cell over the exact same extent with a unique value. Conversely, scale becomes a significant issue when drawing conclusions over the urban landscape. The smaller the unit of input data, the more likely sensitivities in the pattern of development can be identified. and how one averages these local irregularities determines the look of the r esulting et al. 1998, 1440). The number, size, and arrangement of employment densities (sub centers) are spatial scale and definition. The temporal and physical dimension will be examined within each innovative component (i.e., innovative agent, innovative input, and innovativ e output). The temporal dimension examines the relationship and trends of the independent variables over time. More importantly, the temporal dimension is used as a way to analyze the regional impact of each variable. Theoretically, it is an application of putting the horse before the cart; before innovation can blossom from smaller microclimates the larger (regional) environment must be conducive to growth and expansion.
89 The spatial dimension serves two purposes. First, to validate local responses to r egional variables (i.e., the relationship between A&E firms to education, demographics, income, etc.). Once it has been determined that the regional climate can support innovative endeavors the spatial aspect identifies microclimates or more localized net works which offer the greatest opportunity for innovative firms. Innovative Agents : Temporal As stated above, the metro Washington, DC region has evolved into a metropolitan area well known and well respected for its growth in research and high skill emplo yment. Using census data for 1990, 2000, and 2008 a correlation between A&E business patterns and demographics is determined. Specifically, residents within the 20 54 age range are compared. Although innovative agents continue to create and explore new ideas past the age of 54, the rate of innovation decreases with age. Younger agents tend to be less focused and may develop many dissimilar processes that further an exponential number of new products, but as the agent gets older time and experience begin to narrow the focus of product creation. The coefficient of determination is then calculated for each age group to determine the percentage of variance in the business pattern variable that is accounted for by the variance in each age range variable. Alt hough not all agents need formal training to create, a connection with a research institution enables the agent to explore new ideas with financial support. The provision of research funding for post secondary education is related to level of degree. Th erefore, the lower the level of degrees the less funding available for research. Using data provided by the University of Arizona Career Services, the study considered disciplines most associated with architecture and engineering and determined an
90 educatio nal conferred factor (ECF) that considers the influence of degrees awarded in the following fields: agriculture, architecture, engineering technology, engineering, and natural resources and conservation. The degrees included within each discipline vary ac cording to the institution. Table 3 2 summarizes the range of degrees awarded in each discipline for the 26 schools within the metro Washington, DC region and Pima County. Using data from the National Education Administration, a count of conferred degrees at the certificate, associates, bachelors, masters, and doctorate level was recorded. A weight was applied to the total number of degree s awarded at each level (Table 3 3 ). An average ECF for all schools within each county was calculated. Alexandria City and Fairfax City were aggregated to Arlington County and Falls Church City was aggregated to Fairfax County, since they are consolidated metropolitan areas and not true counties. Although some of the disciplines listed in Table 8 seem outside the scope of NAICS 5413, when sub sector products and line of work is examined for existing A&E firms within the study area these fields are deemed appropriate. Example line of business descriptions include Mechanical Engineering Consulting, Building Construction Cons ultant, Architectural and Interior Design Service, and Engineering Services: Civil. There are two measures of education used in this study. The first, described above, determines the relationship between degrees conferred and presence of A&E firms using b usiness pattern data from the Economic Census. The second measure is the educational attainment factor (EAF) . Educational attainment measures population characteristics whereas degrees conferred measures characteristics of the educational institutions in the region . The EAF and ECF will be compared using basic descri ptive
91 statistics to note differences/similarities in the educational level of the general population vs. those of a specific demographic group. Using a regression, the relationship between ed ucational attainment and business patterns was measured temporally for the years 2008, 2000, and 1996. Innovative Agents: Spatial Ideally, this project would include a spatial analysis of the variables examined temporally for innovative agents . Using spati al autocorrelation, two demographic spatial relationships were measured to determine the degree of clustering or dispersion . These measures were 1) the location patterns of county residents between the ages of 20 and 54, and 2) income. Spatial autocorrelat 2009, 107) given a continuous set of data. In areas indicating clustering spatial descriptive statistics are determined aga inst existing A&E firm locations . Innovative Inputs : Temporal Innovative inputs introduces the concept of standardizing the dependent variable, in this study architecture and engineering firms, to control for the size and economic base of each municipality (Gottlieb 1995, 1418). The dependent variable will be standardized in two ways. First, the employment in the architecture and engineering sectors will be measured as a function of density. As in each innovative production phase, temporal and spatial me asurements are employed to determine employment density. The second method to standardize the dependent variable calculates professional service employment in architecture and engineering sectors relative to other firms measured using the location quotien t. Gottlieb notes the advantage a
92 municipality is heavily influenced by a number of factors, including agglomeration economies, competition with residential locators and zoning. The [location quotient] effectively controls for fixed area effects that influence business location in general, Previous research indicates numerous methods to identify employmen t centers. McDonald (1987) introduced the first formal procedure for identifying sub centers (McMillen 2001, 17). By estimating a simple employment density function, McDonald creates a linear relationship between the number of employees per acre and dista nce from the CBD. Sub centers are then identified as clusters of positive residuals in the sub center as a set of contiguous tracts that have a minimum employment density of 10 2001, 18; Giuliano et al. 2007, 2939). Both of these approaches suffer from several pretation and the results are sensitive to the unit of analysis. ultimately affects the number of residents that are classified as a sub center. In itrary cutoffs for employment density and jobs, leading to the identification of fewer centers. If a sub center is an area of large employment then the extent of the sub center may include several contiguous tracts but because of their density are disprop ortionate in size to center tends to flatten the estimated employment density function, which reduces the probability of identifying sub r both models McMillen
93 Craig and Ng use a nonparametric estimation procedure to obtain smooth employment density estimates, eliminating many of the problems with earlier methods percentile of the employment density distribution. The quantile regression approach is attractive in this context because a subcenter is defined using the extremes of the as those mentioned previously and is readily reproducible by other researchers. This method also requires little knowledge of the local area. McMillen states that the arbitrariness caused by the McDonald and Guiliano because the local rise that defines a subcenter is subject to tests of statistical 2001, 18). For this research project, I will identify employment centers as Craig and Ng describe. Developing a methodology for measuring the spatial patterns of innovation is often difficult. Traditional methods involve using case studies and in the 19 80s quantitative methods were introduced using indicators of geographic concentration. A fundamental argument supporting innovation geography is that spillover and the transfer of knowledge is the catalyst to creativity and economic growth. From the argu ments presented previously in the review of literature, spillover is most successfully achieved when relational industries are located in close proximity to one another. From previous innovative geography research measures of concentration can be divided into two categories: concentration indicators that do and do not measure spillover .
94 From the review of Ratanawaraha and Polenske (2007), they concluded that there is no single indicator or measure that is ideal to measure innovation (Table 3 4 ) ( Ratanawara complex as innovation, analysts need to consider the multi dimensional economic, Polenske 2007, 54). For this s tudy, a comparison of location quotients was used to measure spillover in the study region. average proportion of ng area has a higher or lower share of a particular industry than the national share, but does not include any information related to the absolute size of the industry (Fingleton 2007, 69). This problem can be resolved using a method developed by Fingleton, Igliori, and Moore which takes into account the relative local importance of an industry and the size of the agglomeration in terms of number of jobs. is defined as the number of jobs in the local industry that exceeds its expected number. The expected number is then defined by the number of jobs in the industry that would correspond to the area having the national share of the indu stry and therefore produce known as horizontal clustering (HC), by dividing the employment in industry i in area j by the area, A, in each jurisdiction. This is repres ented by . Fingleton admits concentration with regards to the national average it does capture the mass effects,
95 which could be responsible for the gener ation of positive externalities advocated in Innovative Inputs: Spatial The capacity to innovate is based upon how knowledge is dispersed and the organization of firms in space. As mentioned previously, the depend ent variable will be standardized using density. This density is representative of employment centers. Empirically, employment centers are not an appropriate measure or description of the distribution and concentration of employment. Giuliano et al. (200 7, 2938) identifies extensive research that larger US cities have become polycentric using definitions of job density and total jobs to define employment concentration. Additionally, Gottlieb believes the use of polycentric regions facilitate the firm loc ation (Gottlieb 1995, 1416). Understanding the distribution and concentration of i nnovative firms will result in a greater understanding of whether the mechanisms that facilitate clustering are significant. As Giuliano et al. advancements in communication technology and lower tran sportation costs are indeed reducing the need for clustering then we should see an overall reduction in employment et al. 2007, 2938). edict the consequences of discontinuous changes in the traditional way of doing things; it can neither explain the (Schumpeter 1934, 62).
96 Due to a lack of temporal spatial data, a comparison of the distribution over time was not available. As mentioned previously, firm locations were provided from three different sources: Brian Spivak at Arizona State University, Site to do Business online (www.stdb.com), and Dun and Brads treet data. Firm location data provided by Site to Do Business included a business inception date but did not indicate when or if the business failed. Therefore, this information was excluded from the analysis to reduce error . This study measures the i mpacts and transaction costs resulting from the concentration and location of A&E firms in the goals and objectives within the land use conflict model. Innovative Outputs In the review of literature two research projects stood out which measured innovation capacity, a similar goal of this research. The first completed by Hall (2007), groups individual variables into new uncorrelated variables referred to as common 2007, 112). In this work, patents are removed from determining capacity since it follows rather than precedes innovation. A similar process is followed in the work of Acs et al. (2002) where patents were used to explain the relationship of regression var iables. This study uses a combination of the two methods. First, this study employs factor analysis to reduce the variables identified as significant for innovative inputs and innovative agents. The principal component analysis (PCA) factor method is use d because it gives the original factors in terms of differences and similarities between the factors. Theoretically, the first two production phases of innovation, inputs and agents, should lead to innovative outputs. Therefore, a regression equation tha t includes
97 innovative inputs and innovative agents will be used to predict the innovative outputs, or patent counts. The differences in the predicted score and the actual number of patent counts for each county and regionally, also known as the error of e stimate, will be used to validate how appropriate the use of patent counts is in predicting the outcomes of innovation. The calculation of innovation outputs in this study is used more as a validation tool than as a true measure of innovation capacity for one primary reason. The patent data that was collected was based upon patents granted during the following year increments: 1989 1990, 1999 2000, and 2007 2008. All data was collected from the United States Patent and Trademark Office (USPTO) through the online search tool ( http://www.uspto.gov/patents/process/search/index.jsp ). Filings through the USPTO are only published when a patent is granted. Furthermore, patents may not be grante d within eighteen months, the typical time period for application processing and new methods could be in use before formal approval is given by the USPTO. Also, patent information is not available by county. Patent information was retrieved according to major cities included in the sample of A&E firms used in this study. Therefore, it is possible the patent counts represent a sample of the total patents available for a county then the error of estimate would be based upon a sample patent count and not an actual count. Suitability Analysis and Economic Opportunity The variables introduced to this point describe the relationship of variables that reduce barriers to entry. The variables and statistical values calculated do not indicate where the specific ne tworks or microclimates that offer the greatest opportunity for information exchange and may ultimately yield the greatest productive value. By using
98 key elements of the LUCIS method as factors of innovation, opportunities for innovative networks are iden tified. The LUCIS method as applied to economic models uses the first tier of the LUCIS method with a modified second tier. The first tier consists of the process to determine land use conflict. The process includes 1) determining land use suitability bas ed upon the pre determined goals and objectives; 2) determining land use preference; and 3) identifying conflict. In the standard LUCIS process the second tier consists of the allocation of employment and population in the development of alternative future s. The second tier in this study employs the Combine Grid for allocation but consists of different variables that facilitate the allocation of innovative employment in areas that present the greatest opportunities for innovative networks. The standard LUC IS plus Combine Grid uses of the following layers in scenario building: Transit access layer TAZ layer Political Areas Developments of Regional Impacts (DRI) (if applicable) Clustering These layers describe typical considerations of residential and employm ent location. For example, priorities may be given to allocating a specific population into DRIs based upon phasing plans approved by the regional planning agency. In the location of innovative employment the rationale of location is a bit different. Al though it is easy to allocate employment into particular areas because of directed capital investments or regulatory incentives, these mechanisms ignore the dynamic system created by the natural organization and presence of amenities and innovative system.
99 The modified Innovation LUCIS plus Combine Grid uses the following layers in scenario building: Transit access layer . The traditional LUCIS plus model uses the following methodology to dete rmine transit access (Arafat, Zw ick, and Patten 2010): Calculating the network distance from the upstream nearest stop to job related stops downstream Calculating the distance from the downstream stop to the job location Calculating the job opportunity associated with downstream stops Applying the gravity model to calcula te an accessibility index for each stop Summarizing the transit scores within walking distance from parcels using an opportunity/distance accessibility model For simplicity, this model utilized the determined transit access by calculating Manhattan Distanc e TAZ layer . Used for aggregation of final employment Political areas . The traditional LUCIS plus model uses the Political Areas field to apply a geographic weighting in the prioritization of employment and residential allocation Firm Density (IP_Density) . Describes the A&E firm density by census tract (unit of measurement is # of firms/square mile) Average values for transit access and service suitability were taken in census tracts with A&E firm densities within one standard deviation of the mean. The PC A is then re run twice: once using the average values for transit access as the predicted value, by county; and secondly re run using the average transit access value as the predicted value. Final Multiple Regression Equation . The first PCA analysis examin es the
100 of contribution to the innovative outputs using longitudinal data. The second and third PCA analysis, which integrates the spatial component, measures the abilit y for actual on the ground innovative networks to be created and describe the microclimates in which innovative firms prefer. To determine innovative economic opportunity a two step process is employed: 1. A regression equation is formed using the variables included in the first PCA analysis to understand the regional context for innovation. Coefficients of the regression equation are based upon the PCA values in the reference region, the metro Washington, DC region. 2. A regression equation is formed using th e transit access, firm density, and average service suitability values, again, using the values from the PCA as the coefficients for the regression equation. The following chapter, Results & Analysis, will discuss in detail the outcomes of each step descr ibed in this Methodology for the metro Washington, DC region and Pima County.
101 Table 3 1. A&E firms p er c apita County A&E Firm Population 2008 Population A&E Firm (per capita, in thousands) Alexandria City 224 140,657 1.6 Arlington County 308 204,889 1 .5 District of Columbia 733 588,373 1.2 Fairfax County 1,403 1,005,980 1.4 Fairfax City 116 23,281 5.0 Falls Church City 37 11,169 3.3 Montgomery County 1,306 942,747 1.4 County 118 825,924 1 TOTAL 4,245 3,743,020 1.1 Pima County 53 9 1,018,012 0.5
102 Table 3 2 . Degrees and d isciplines c onsidered for a rchitectural and e ngineering i ndustries Discipline Degrees Included Agriculture, Agricultural Operations, and Related Services Landscaping and groundskeeping Architecture and related services Architecture, Landscape Architecture Engineering Technologies/Technicians Engineering/Industrial Management; Computer Engineering Technology/Technician; Architectural Drafting and Architectural CAD/CADD; Computer Technology/Computer Systems Techn ology; Aeronautical/Aerospace Technology/Technician; Civil Engineering Technology/Technician; Construction Engineering Technology/Technician; Electrical/Electronic Engineering Technologies/Technicians, Other; Electromechanical Instrumentation/Maintenance T echs, Other; Water Quality & Wastewater Treatment Mgmt & Recycling Tech Engineering Civil Engineering, Other; Computer Engineering, General; Computer Software Engineering; Electrical, Electronics and Communications Engineering; Engineering, Other; Systems Engineering; Aerospace, Aeronautical & Astronautical Engineering; Agricultural/Biological Engineering & Bioengineering; Chemical Engineering; Engineering, General; Engineering, Other; Materials Engineering; Mechanical Engineering; Architectural Engineerin g; Natural Resources and Conservation Environmental Science; Environmental Studies; Natural Resources Management and Policy; Natural Resources/Conservation, General
103 Table 3 3 . Education f actor for w eight d istribution Degree Education Factor Weight Ce rtificate 0.067 Associates 0.134 Bachelors 0.201 Masters 0.268 Doctorate 0.335 Table 3 4 . Traditional m easures of c oncentration Concentration indicators that DO measure spillover Concentration indicators that DO NOT measure spillover Location Quot ient (LQ) Ellison Glaeser Geographic Concentration Index (EGGCI) Horizontal Clustering Location Quotient (HC) Geographic Coincidence (Concentration) Index (GCI) Locational Gini Coefficient (LGC) Gene Related Herfindahl Hirchman Index (HHI)
104 Figure 3 1. Proposed theoretical model
105 CHAPTER 4 RESULTS AND ANALYSIS Capitol Region Innovative Agents As stated in the literature review, existing human capital is fundamental to on with an ample supply of young and growing population provides greater meaning to the dashed lines connecting overlap between business factors and amenities to firm location and is no relationship between innovative agents and the location of A&E firms. The Capitol region has experienced a steady rate of population growth for the past 20 years. Table 6 illustrates that between 1990 and 2000 the Capitol region experienced a cumul ative growth of about 11 % and through 2008 growth was much slower at 5 % . Demographically, there was a decline in population of 20 to 24 year olds in all counties within the region from 1990 to 2000. During this same time period there was a decline in popu lation among 25 to 34 year olds except in Alexandria City and Arlington County, among 34 to 44 year olds there was an increase in population in all counties except in the District of Columbia and Falls Church City, and all counties during this decade exper ienced an increase in population among 45 to 54 year olds. The following decade offered more promise for all age cohorts. Among 20 24 year olds between 2000 and 2006/ 2008, all counties experienced an increase in population except Alexandria City, Arling ton County, and Fairfax City. Among 25 to 34 year olds all counties declined in population except the District of Columbia and among 35 to 44 year
106 County all experienced population declines. Among the last age cohort, 45 to 54 year olds, all counties saw an increase in population. Regionally, from 1990 to 2000 there was a cumulative loss in population only among the 20 to 24 year olds and among the 25 to 34 year olds. D uring this decade 45 to 54 year olds experienced the largest gain in population and 25 to 34 year olds experienced the largest decline in population with 150,932 people and 69,843, respectively. During 2000 to 2006/08, there was a loss in population amon g 25 to 34 year olds and 35 to 44 year olds. The greatest population gain was among 45 to 54 year olds and the greatest decline was found among 25 to 34 year olds with 58,913 and 87,278, respectively. Although the decline in residents during 2000 2006/ 08 is noteworthy, remember that general Census demographic data reflects the residential location of individuals. Housing opportunities in the Capitol region beyond the counties included in this study were better than those available within the included c ounties and residents were willing to live further and commute into the metro Washington, DC region. This is evident in data liste d in Table 4 1 , which illustrates the county of residence for firms located within each study area county. ificant about innovative agents and firm location is the availability of young, educated workers. The relationship between the number of firms and the age composition for each county was calculated using a regression to determine if in fact a relationship existed in the Capitol region. Firm location information was based upon business pattern data provided in the Census Business Statistics for 2008, 2000, and 1998. To measure this statistical relationship the hypothesis is restated and a research
107 hypothes is is given: the regression equation does not account for a significant portion of the variance in the location of A&E firms. Unfortunately, a direct comparison to parallel time periods was unavailable for 1998 so this study used 1990 demographic data in this comparison. First, a scatterplot of business patterns against population quickly illustrated the correlation between both variables. The log of the raw business patterns and population data was taken to normalize the data. For each age cohort the fig ures used were the cohort population as a percentage of the total county population. Appendix D illustrates the scatterplots for 1990, 2000, and 20 08 for each age group. Table 4 2 summarizes the correlation between these two variables and the direction o f the relationship. Given the three time periods this study examines, no strong relationship was evident between any age cohort and business patterns. Although it is important to note that the strength of the relationship increased as time progressed. Th e greatest increase in strength was not ed in the 20 to 24 year old and 35 to 44 year old age cohorts, which illustrated a direct and indirect relationship, respectively. The foundation of this research is to predict innovation based upon various variables described in the Methodology section. Essentially this study contends that given independent variables that describe innovative agents, innovative inputs, and innovative outputs business patterns in our study region can be predicted. For innovative agent s a regression is used to predict the business patterns given the population within each age cohort. The following equation is used in this prediction: Log (# of businesses) = Log (total population) + % of each age range of total population
108 This equation describes the contribution of each age range to business firm location, by county, while adjusting for total population. The resultant regression statistics for the age cohorts ( Appendix E ) during the three time periods indicates that a high percentage of variation in business patterns is explained by the relationship between the total county population and population of that specific cohort. The adjusted R square more accurately reflects the bivariate relationship of the degree of fit in the population. T he adju sted R square values in Table 4 3 indicate that in 1990 and 2000 age described business patterns to a greater degree than in 2008. A 95% confidence interval was used in the regression analysis. The F ratio critical ratio for 2008 with n=7 has df =2 , 4 and a critical value of 6.94. For all 2008 age groups, the F statistic is less than the critical value therefore we fail to reject the null hypothesis .. The F ratio critical ratio for 2000 and 1990 with n=8 has df=2, 5 and a critical value of 5.79. T he F ratio critical value in 2000 for the 20 to 24 year old and 35 to 44 year old cohorts in addition to all cohorts in 1990 were greater than the critical value therefore rejecting the null hypothesis. Therefore, as time has progressed the age cohorts ex amined in this study have explained less of the variation in A&E business patterns. In space, spatial autocorrelation was used to determine the degree things near each other were more alike than things far apart. Spatial autocorrelation indicates the distribution of values is dependent on the spatial distribution of the are more like each other than they are like more distant features. Negative
109 autocorrelation i s when neighboring features are unlike each other. No autocorrelation results when there is a random pattern among features (Mitchell 2008). The advantage to using this statistic is that it can handle continuous data like age although spatial autocorrela is for high values or low values. Using Census 2000 data the age cohorts in the Washington, DC, Virginia and the County, and Alexandr ArcGIS also produces a z score when spatial autocorrelation is calculated to indicate confidence of the spatial statistic. The analysis of the age cohorts will demonstrate whether the data is randomly distributed. Assuming a 95% confidence limit, the z score for all age cohorts except Fairfax City and Arlington County are outside of the 1.96 to 1.96 range, thus indicating that the null hypothesis cannot be rejected. An additional argument used to support attracting creative industries to cities is that they provide higher wages than non knowledge based industries. An example is in the case of A&E firms in which the average salary is higher than the national average for all industries. Onc e again, Washington, DC and the two Maryland counties indicated a clustered distribution for median income, whereas the three Virginia counties illustrated a dispersed pattern with z scores within the 95% confidence limit indicating that null hypothesis ca nnot be rejected. An educated population is not only beneficial for municipalities in attracting new employment but the educational level attained by a population can influence available services and amenities. As described in the Methodology, the Educati onal Attainment
110 population. Greater weight was given to more advanced degrees, thus the higher the factor the more residents with advanced degrees. Similarly, the EC F was calculated based on the actual number of degrees awarded by universities in each county. Unlike the age cohort data, degrees conferred were available for each respective year represented in the longitudinal study. Therefore, for 1998 business patte rns degrees conferred were used from the 1997 1998 academic year. Table 16 shows the EAF, ECF, and A&E business patterns count for each county in which there are universities with disciplines related to A&E. During each of the three time periods the educ ational attainment of the county is greater than that of the average number of degrees awarded by universities in that county. A scatterplot of business patterns against the educational attainment factor quickly illustrates the correlation between both var iables. Appendix E illustrates the scatterplots for 2008, 2000, and 1990 f or each time interval. Table 4 4 summarizes the correlation between these two variables and the direction of the relationship, which is weak and indirect. For the educational attai nment factor a regression is used to predict the business patterns for each year. The following equation is used in this prediction: Log (# of businesses) = Log (total population) + educational attainment factor This equation describes the contribution of the level of education to business firm location, by county, while adjusting for total population. The resultant regression statistics for the three time periods (Appendix F) indicates that as time progresses the percentage of variation in business patter ns explained by
111 the county educationa l attainment decreases . The adju sted R square values in Table 4 5 indicate that in 1990 and 2000 the educational attainment of the county in which A&E firms were located accounted for more than half of the variation in firm location. A 95% confidence interval was used in the regression analysis. The F ratio critical ratio for all three time periods with n=8 has df=2, 5 and a critical value of 5.79. The F ratio critical value in 1990 is greater than the critical value which supports a rejection of the null hypothesis. The F ratio critical value in 2000 and 2008 is less than the critical value, resulting in a failure to reject the null hypothesis. Therefore, as time progressed the linear relationship between educationa l attainment and business patterns have decreased. Capitol Region Innovative A rand om sample of A&E firms (Table 4 6 ), the dependent variable, was used in this study to reflect innovative inputs. Spatially, the dependent variable was standardized as a f unction of densit y in each census tract (Figure 4 1 ). The magnitude of the z scores indicate that census tracts with the number of A&E firms furthest away from the mean are located in the District of Columbia or within one mile of the DC boundary. For ea ch county, the average nearest mean distance for a hypothetical neighbor index considers the relationship between fe atures and indicates whether features are randomly distributed (i.e., the null hypothesis). In every county examined in this study ( Appendix G ), the nearest neighbor index indicates clustering and the z score for each county indicates that the probability is less than 0.05, which translates into a rejection of the null hypothesis and the features are not randomly distributed.
112 The second method of dependent variable standardization uses the location quotient to determine the influence of A&E firms to determ ine their influence on the larger economy. Location quotient (LQ) values were collected from 2001 , 2005, 2008, and 2009 (Table 4 7 ). All counties, except Fairfax City and Falls Church City, experienced a relative decline then increase over the course of the eight year time period. No county rebounded to its highest LQ value except Washington, DC. Additionally, in all study area counties there is an above average proportion of employment in A&E. The highest proportion of employment for each of the four time periods remains within three counties: Arlington County, Fairfax City, and Fairfax County. In addition to the LQ, Ratanawaraha and Polenske (2007) identify various other methods of concentration that also measure knowledge spillover. This study calcu lated the horizontal clustering (HC) location quotient for each metro DC county. Horizontal clustering also indicates cluster intensity. Given the total number of A&E firms in a county, Table 4 8 details the HC quotient for each Capitol Region county. T he HC quotient measures the cluster intensity of employment per unit area. Although it is easy to translate this value as a measure of concentration/dispersion of innovation this value does possess two weaknesses. First, the accuracy of the results is ba sed upon the Polenske 2007, 47). This study analyzed A&E firms at the 4 digit sector level, which according to research completed by Ratanawaraha and Polenske, may reveal one or two subsectors of innovation. The second potential weakness is the inability to translate the results of the HC quotient to explain the degree of dispersion over the region.
113 Capitol Region Innovative Outputs For each aspect of innovation discussed t hus far, innovative agents and innovative inputs, variables were identified that could explain the relationship with firm location. To reduce the variables examined in this study, factor analysis was used to determine which were important. The hypotheses for innovative agents are: H 0 : There is no relationship between innovative agents, amenity factors, and the intrinsic value of land in the creation of innovative networks. H 1 : There is a relationship between innovative agents in the creation of innovative networks. At the county level, firm location was examined given the following temporal variables: Age cohorts Education: degrees conferred and educational attainment Horizontal clustering quotient The research hypothesis, H 1 , is not specific since this stu dy only focuses on three possible treatment conditions. The principal components analysis (PCA) results provide summary statistics and a correlation matrix ( Appendix I ). A&E firms have a moderately strong relationship to the 20 to 24 year old age cohort. For all other age ranges there is a weak or negatively correlated relationship. This means that businesses may be moving to the region for reasons aside from the demographics of the general population. The eigenvalues resulting from the PCA exclude the 25 to 34 year old and 45 to 54 year old cohorts in 2008 because these two variables are negatively correlated. In 2008 the first three eigenvalues, business patterns, total population, and the 20 to 24 year old cohort correspond to more than 87% of the in itial variability of the data. The
114 correlation circle of axes F1 and F2 shows a projection of the initial variables in the factors space. Patent data has been used in the past as a measure of innovative output. The factor scores from PCA for each county was used as the predictor for innovative agents in the regression equation for patent counts, the nearest neighbor ratio reflected the degree of clustering for innovative inputs, and the actual number of patents granted was the predicted value. The adjust ed R square value of 0.737 indicates that the variability in patent counts is due in large part to cluster dynamics and innovative agents. Theoretical Regional Regression Equation The regression equation that describes the regional context of innovation (R CI) is: RCI = 0.662 (log total population) + 0.628 (20 to 24 year old %) + 0.059 (25 to 34 year old %) 0.560 (35 to 44 year old %) 0.267 (45 to 54 year old %) 0.122 (educational attainment factor) + 0.071 (educational conferred factor) When the regr ession equation is applied to the test site a graph of the expected (predicted) outputs, as a function of patents, and actual outputs is available. The resultant graph demonstrates whether the region underperforms (below the regression line) or over perfor ms (above the regression line) the regional innovative performance of the Capitol region. When applied to the metro DC region, the RCI index is: RCI = 0.662 (6.572) + 0.628 (0.070) + 0.059 (0.130) 0.560 (0.31) 0.267 (0.30) 0.122 (0.228) + 0.071 (0. 067) The resulting regional innovative value is 4.350 + 0.044 + 0.008 0.174 0.028, which equals 4.200. Capitol Region Innovative Networks Using the LUCIS hierarchical framework as a guide, suitability was determined for the entire Capitol region. Al though this study does not include future employment
115 allocation, a conflict surface was created and a Combine Grid was assembled using the following fields: Census Tract TAZGrid Conflict TransAcces Jurisdiction SvcSuit FirmDens PCA was run on the 668 cens us tracts with A&E firms in them using the TransAcces, SvcSuit, and FirmDens fields. None of the variables in the resulting correlation matrix are negatively correlated, so all variables affected the result. The firm density (IP_DENSITY) and transit acces s are different from zero with a significance level alpha=0.05. From the eigenvalue matrix it is evident that the first eigenvalue equals 1.379 and represents almost 46% of the total variability. Although the value is relatively low (i.e., less than 50%) , the first two factors represent almost 80% of the variability. Theoretical Local (Spatial) Regression Equation The regression equation that describes local innovative networks (LIN) is: LIN = 0.826 (IP_DENSITY) + 0.817 (TransAcces) + 0.171 (SvcSuit) When applied to the average density for the Washington, DC region the local innovative network value results in the following: LIN = 0.826 (38.96) + 0.817 (7) + 0.171 (7) The resulting local innovative value is LIN = 32.18 + 5.719 + 1.197, which equals 39.096.
116 Pima County Region Innovative Agents Table 4 9 illustrates that between 1990 and 2000 the Pima County region experienced a cumulative residential growth of about 27 % and through 2009 growth was slightly slower at 21 % . The municipalities within the regi on that experienced the largest amount of growth was due primarily to a significant influx of retirees and those and early millennium. County wide there was a cumulati ve increase in population within all age groups analyzed. Trends between jurisdictions indicate a general increase in populations among all age cohorts with the exception of Tucson. The 35 to 44 and 45 to 54 year old demographics illustrate a decline bet ween 2000 and the 2006 08 estimates. Tucson is home to the University of Arizona, the major university for southern Arizona, and the decline is most likely due to students graduating and moving out of the area or seeking housing after graduation in a Tucs on suburb due to the cost of housing within the City of Tucson. Although there are residents who work in Pima County and live in other adjacent counties (Table 4 10) , the proportion of these residents is much lower than those found in the Capitol region . indicated clustering only in the 45 to 54 cohort. The 35 to 45 cohort indicated randomness and both the 20 to 24 and 25 to 34 cohorts demonstrated dispersed distribution of age. Arc GIS also produces a z score when spatial autocorrelation is calculated to indicate confidence of the spatial statistic. Assuming a 95% confidence limit, the z score for the 35 to 44 cohort indicates that the pattern does not appear to be significantly dif ferent than random. For the 45 to 54 cohort there is less than 5% likelihood that the clustered pattern could be the result of random chance. This
1 17 indicates for both these age cohorts that t he null hypothesis cannot be re jected and for the cohorts betwee n ages 20 and 34 the null hypothesis can be rejected. An additional argument used to support attracting creative industries to cities is that they provide higher wages than non knowledge based industries. An example is in the case of A&E firms in which th e average salary is higher than the national average for all industries. Pima County demonstrates a clustered distribution with a z score within the 95% confidence limit indicating that there is le s s than 1 % likelihood that this clustered pattern could be the result of random chance. Therefore the null hypothesis cannot be rejected. From the reference Capitol region, the following variables were included in the final regional regression equation: Total population, 2008 20 to 24 year old % 25 to 34 year old % 35 to 44 year old % Educational attainment factor Educational conferred fa ctor Table 4 11 lists the values f or Pima County. Using the regional regression equation, regional innovation can be measured by: RI = 0.662 (6.0078) + 0.628 (4.8616) + 0.059 (5 .1221) 0.560 (5.1060) 0.267 (5.1338) 0.1222 (0.2324) + 0.071 (0.1117) The resulting regional innovation measure is 3.977 + 3.053 + 0.3022 + 2.859 1.37 0.0283 + 0.0079, which equals 8.8008. A rand om sample of A&E firms (Table 4 12 ), the dependent variable, was used in this study to reflect innovative inputs.
118 Spatially, the dependent variable was standardized as a function of density in each census tract ( Figure 4 2 ). The magnitude of the z scores indicates that census tracts with the number of A&E firms furthest away from the mean are located in the City of how similar the mean distance is to the expected mean distance for a hypothetical 2008, 88). The nearest neighbor index indicates clustering and the z score for Pima County reflects that there is less than 1% likelihood that the clustered pattern could be the result of random chance. To predict the innovative network potential, the lo cal innovative network equation was applied: IN = 0.826 (IP_DENSITY) + 0.817 (Trans_Acces) + 0.171 (SvcSuit) For each of the variables, the averages were applied for all 126 census tracts that contain A&E firms in Pima County. IN = 0.826 (3.323) + 0.817 ( 5) + 0.171 (5) The resulting local innovative network value is 2.745 + 4.085 + 0.855, which equals 7.685.
119 Table 4 1. Top 10 p laces of r esidence for metro Washington, DC region w orkers County of Employment 1990 Census County of Residence 2000 Census Co unty of Residence Arlington County, Virginia Fairfax County, VA (51,841) Arlington County, VA (34,382) (17,055) District of Columbia (13,393) Fairfax County, VA (48,670) Arlington County, VA (34,379) MD ( 15,912) District of Columbia District of Columbia (236,734) (141,590) Montgomery County, MD (103,320) Fairfax County, VA (94,502) Arlington County, VA (43,842) Alexandria City, VA (23,557) Prince William County, VA (13,547) Anne Arundel County, MD (11,964) Charles County, MD (9,976) Howard County, MD (7,917) District of Columbia (190,566) MD (126,138) Montgomery County, MD (99,672) Fairfax County, VA (88,908) Arlington County, VA (42,263) Alexandria City, VA (23,292) Anne Arundel County, MD (15,891) Prince William County, VA (15,368) Charles County, MD (10,785) Howard County, MD (8,461) Fairfax County, Virginia Fairfax County, VA (238,650) Prince William County, VA (32,934) Loundoun County, VA (18,055) Mon tgomery County, MD (16,177) Arlington County, VA (15,575) (15,362) Fairfax County, VA (278,064) Prince William County, VA (44,322) Loundoun County, VA (35,933) Montgomery County, MD (22,148) Arlington County, VA (20,476) Prince G MD (18,258) Alexandria City, VA (14,643)
120 Table 4 1. Continued County of Employment 1990 Census County of Residence 2000 Census County of Residence Montgomery County, Maryland Montgomery County , MD(251,949) D (40,560) District of Columbia (20,487) Frederick County, MD (18,887) Fairfax County, VA (15,001) Montgomery County, MD (267,128) MD (40,240) Frederick County, MD (22,867) District of Columbia (19,509) Fairfax County, VA (16,943) Maryland (167,418) Montgomery County, MD (26,879) Anne Arundel County, MD (23,758) Howard County, MD (13,202) District of Columbia (12,979) MD (155,671) Montgomery County, MD (26,82 5) Anne Arundel County, MD (26,271) Howard County, MD (14,538) Charles County, MD (13,834) Source: US Census Bureau Journey to Work and Place of Work http://www.census.gov/population/www/socdemo/journey.html
121 Table 4 2 . Capitol r egion R 2 values, age Age C ohort 1990 R 2 2000 R 2 2008 R 2 20 24 year old 0.0198 (Positive) 0.1949 (Positive) 0.3947 (Positive) 25 34 year old 0.0004 (Positive) 0.0031 (Positive) 0.0035 (Positive) 35 44 year old 0.0761 (Negative) 0.2160 (Negative) 0.3140 (Negative) 45 54 y ear old 0.0213 (Negative) 0.2501 (Negative) 0.0711 (Positive) Table 4 3 . R square regression statistic by age cohort Age Cohort 2008 R 2 2008 R 2 Adj 2000 R 2 2000 R 2 Adj 1990 R 2 1990 R 2 Adj 20 24 year old 0.6604 0.4906 0.7266 0.6172 0.7884 0.7038 25 34 year old 0.5012 0.2517 0.6737 0.5432 0.7858 0.7002 35 44 year old 0.5404 0.3105 0.8430 0.7802 0.8029 0.7240 45 54 year old 0.5576 0.3364 0.6923 0.5692 0.7969 0.7156 Table 4 4 . Capitol r egion R 2 values, educational attainment f actor 2008 R 2 2000 R 2 1990 R 2 Regional EAF 0.0621 (Negative) 0.0464 (Negative) 0.0312 (Negative) Table 4 5 . R square regression statistic for educational attainment factor 2008 R 2 2008 R 2 Adj 2000 R 2 2000 R 2 Adj 1990 R 2 1990 R 2 Adj Region 0.6390 0.4947 0.6799 0. 5503 0.8034 0.7248
122 Table 4 6 . Capitol r egion a rchitecture and e ngineering f irm s tudy c ount County Population Study Sample Alexandria County 224 146 Arlington County 308 140 Washington, DC 733 463 Fairfax County 1,403 86 Fairfax City 116 60 Falls Ch urch City 37 18 Montgomery County 1,306 602 118 240 Table 4 7 . Location quotients for C apitol a rea s tudy c ounties 2001 2005 2008 2009 Arlington County (5.82) Arlington County (5.71) Arlington County (4.79) Arlington County (5. 07) Fairfax City (3.42) Fairfax City (3.92) Fairfax City (4.17) Fairfax City (4.47) Fairfax County (2.59) Fairfax County (3.21) Fairfax County (2.67) Fairfax County (2.62) Table 4 8 . Horizontal c lustering q uotient for m etro Washington, DC r egion Coun ty Land Area, 2000 (sq miles) Persons per sq mile, 2000 Horizontal Clustering (HC) Quotient 2001 2005 2008 2009 Arlington County 25.87 7286.7 309.6 288.9 273.6 281.8 District of Columbia 61.4 9378.0 100.9 110.1 134.4 129.7 Fairfax County 395.04 2455 .1 36.4 47.7 43.7 40.3 Fairfax City 6.27 N/A 100.5 163.8 159.5 161.6 Falls Church City 2 N/A 57.5 87.0 105.5 104.0 Montgomery County 495.52 1760.8 19.1 16.4 7.6 13.4 County 485.43 1652.6 8.3 8.5 14.0 7.7
123 Table 4 9 . Pima County popu lation, 1990 2008 1990 Population 2000 Population 90 00 Population Change 90 00 % Population Change 2008 Population 00 08 Population Change 00 08 % Population Change Pima County 666,880 843,746 176,866 27% 994,244 150,498 18%
124 Table 4 1 0 . Places of r e sidence for Pima County, Arizona w orkers County of Employment 1990 Census County of Residence 2000 Census County of Residence Pima County, Arizona Pima County, AZ (282,789) Maricopa County, AZ (1,111) Pinal County, AZ (1,000) Cochise County, AZ (805) Santa Cruz County, AZ (741)) Pima County, AZ (359,296) Pinal County, AZ (2,601) Cochise County, AZ (1,711) Maricopa County, AZ (1,214) Santa Cruz County, AZ (978) Source: US Census Bureau Journey to Work and Place of Work http://www.census.gov/population/www/socdemo/journey.html Table 4 1 1. Pima County regression values Regression Variable Pima County Value Log total population 2008 6.0078 20 to 24 year old percentage 4.8616 2 5 to 34 year old percentage 5.1221 35 to 44 year old percentage 5.1060 45 to 54 year old percentage 5.1338 Educational attainment factor 0.2324 Educational conferred factor 0.1117 Table 4 12 . Pima County a rchitecture and e ngineering f irm s tudy c oun t County Population Study Sample Pima County 523 484
125 Figure 4 1. Capitol region A&E firm density
126 Figure 4 2 . Pima County A&E firm density
127 CHAPTER 5 DISCUSSION Existing research described in the Literature Review identified several elements tha t contributed to an innovative environment. Some of these elements included: a young population to serve as a knowledge base for employers (innovative agents), co location of existing A&E companies and business factors that provide financial benefits for location (innovative inputs), evidence of the productive value of a knowledge network through patents (innovative outputs), and the spatial relationship of existing firms (innovative networks). The Methods section outlined the elements to be examined in t his study that would reflect the four elements of innovation: innovative agents, innovative inputs, innovative outputs and innovative networks. From the reference Washington, DC metro region it was determined that the size of the population, the 20 to 24 y ear old cohort, the 25 to 34 year old cohort, and the education conferred factor all contributed to the locations of A&E firms. Interestingly, in Pima County the two age cohorts have declined in population since 1990. As mentioned previously, the Pima Co unty region is comprised of five jurisdictions and it would be easy to conclude that the decline is due to the burgeoning senior populations in the large retirement communities of Sahuarita and unincorporated Pima County. Unfortunately, this is not the ca se. In Tucson, where the greatest densities of A&E firms are located (Figure 4 2 ), the population is getting older; the higher age cohorts possess a larger share of the total population . The State of Arizona has three universities and the University of Ar izona in Tucson is a land grant university with an enrollment of approximately 37,000 students and is the second largest in the state behind the Arizona State University in Tempe, which has
128 approximately 52,000 students. Although the Capitol region is lar ger and made of multiple counties, the number of degrees in the disciplines considered in this study are, by number, greater than those of the University of Arizona alone. The other advantage to the Capitol region is that in addition to the major universi ties in the region (i.e., American University, the University of Maryland, etc.) the region has smaller colleges and community colleges that confer a large number of degrees that contribute to jobs in architecture and engineering fields. In Pima County, t here are smaller educational facilities that provide degrees in A&E related fields, but the per capita numbers are much less. Regionally, when the regional innovation regression equation is applied to the Capitol region the RI value equals 4.200. When appl ied to test sites larger values indicate an over performance of innovation and values below indicate an underperformance when compared to the Washington, DC region. Locally, the innovative network measure is calculated to be 41.799 and the same rules appl y in the determination of over and underperformance as stated above. When the same tests were applied to Pima County the regional innovation value equals 8.8008 and the local innovative network value equals 10.3875. For Pima County, on a regional scale i t over performs the Capitol region but underperforms on a local level. This can be translated to mean that although when both regions are compared at face value and then considering the methods described in this study it seems as if the Washington, DC reg ion is thriving in creating quality places that A&E firms want to locate even though regionally there may be other industries that may have
129 offers much in terms of attract ing A&E firms, but the potential exists to leverage regional resources. Yet on a local level (i.e., within census tracts), Pima County is unable to leverage and translate regional demographic and educational benefits in to sound land use policy that creat e the networks needed to attract A&E firms. Furthermore, the location quotient for Pima County is 0.78, which indicates that there is underemployment for A&E industries.
130 CHAPTER 6 CONCLUSION The purpose of this study was to determine the role and contrib ution of innovative agents, innovative inputs, and innovative outputs on the regional potential to attract A&E firms as well as the ability to create innovative networks at the local scale. This study has achieved this by examining various measures of inn ovation at each innovation level and determining a measure or predictor of innovative potential based upon the contribution of various factors. Theoretically, this study resembles the Gottlieb residential amenities and firm location study. The difference is that this study provides a way to measure local and regional firm advantage and provides a way to explain the advantage at various scales. This study also more clearly measures land use suitability using the LUCIS methodology. One aspect of the spati al component not explored in this study but could serve as a basis for future research is to determine if in locations where there is a regional but not local advantage for A&E firms whether by allocating A&E, or more generally service, employment in areas illustrated with a high opportunity or conflict value would change the innovative network potential. From a practical point of view, this study could aid economic development offices in understanding how to better leverage traditional factors of business location with land use policy to attract firms of particular types. The LUCIS model, and models which employ similar McHargian methods, is important in that in addition to the ability to create a land use conflict value, it provides a common scale to mea sure land use potential within regional microclimates using values that describe achieving the greatest use of the land.
131 This paper implicitly re introduces the region vs. non region debate in that the county and successfully ach ieves regional jurisdictional advantage. The test region, Pima County, is a region defined using municipalities to define its regional context. Again, future research may use the methods employed in this study to determine whether a greater share of innov ative networks can be created if the Sun Corridor were used as the region of study. Furthermore, it would be interesting to undertake a comparative study for Maricopa County to determine how it fared against the Capitol region in measure of RI and IN. Du e to study funding constraints, data was obtained from various sources and during quality control some discrepancies were documented. Secondly, the land use suitability model built for Pima County was more detailed than that of the Capitol region due to l imitations on available data. The District of Columbia has a very thorough and complete GIS database available for free, whereas free spatial data for the other municipalities was limited and the costs associated with many datasets was beyond the budget f or this project. The result of these differences could potentially surface in determining the suitability value for service related industries. My interest in my research topic came from two unique perspectives. After reading a book by Richard Florida on the location characteristics of the creative class I was left thinking that the success of cities and the attraction of specific employees and employment types was more complex than the framework and methodologies that seemed to encourage socially enginee ring cities with an appropriate mix of beatniks, parks, and removing any evidence of professional sports stadiums. Florida mentioned throughout his book that his results were supported by science but I was curious how
132 the advancement of predictive spatial and statistical models contributed to understanding the dynamics behind thriving cities and communities and how these methods would encourage more proactive planning. The linkage between location advantage and the built environment derive from the plann ing theory of Jane Jacobs. She asserts that creativity originates from the built environment, which encourages the arrangement of unique economic dynamics. More creativ outcomes of my research would be considered scientific in the planning community because I am not criticizing the conceptual weaknesses in economic theory, but using economic a nd non economic factors to explain the casual relationship between creative community acquires with a paradigm is a criterion for choosing problems that, while the paradig m is taken for granted, can be assumed to have solutions. To a great extent these are only the problems that the community will admit as scientific or encourage its resea rch, diverse contributions specifically within economics have been implicit to urban planning (e.g. neoclassical economics, land rent, agglomeration economy, etc.). In the shadow of these contributions, specific knowledge for urban planning has developed through history (Pinson 2004, 506). contributions meet the standards of a science (which differ from a non science), but to see how a specific contribution adds something to t he current knowledge about human
133 Science is called upon to provide an understanding of the complexity of objects and e unity but the coherence of marries science with planning knowledge. This study provides a scientific measure for understanding regional and local potential for architect ure and engineering industries.
134 APPENDIX A DATA LIST Table A 1. Data used for spatial concentration/dispersion of innovation measurement General data Specific data Examples of empirical work Geographical scale and location Time frame Industries Data sour ces Innovative inputs R&D Expenditures Funding for research projects Feldman and Lichtenb erg (1998) Country (EU) 1962 1996 All CORDIS RTD database Laboratory R&D budget Adams (2002) Distance from laboratory location (200 miles) (US) 1991 1996 All Surve y of Industrial Laboratory Technologies R&D personnel Number of scientists and engineers engaged in R&D Porter and Stern (1999) Country (OECD) Various years All OECD Science and Technology Indicators Number of scientists and collaborators Zucke r, Darby, and Brewer (1994) Zip code, County, and Functional Economic Area (US) 1975 1989 Biotechnology GenBank database Employment Employment Fingleton, Igliori, and Moore (2003) UALAD (UK) 2000 Annual Busin ess Enquiry of the Office for National Statistics
135 Table A 1. Continued General data Specific data Examples of empirical work Geographical scale and location Time frame Industries Data sources High tech manufacturing employment Malecki (1985) Metropol itan Area (US) 1983, 1993 4 high tech Dun & Bradstreet Corporate Market Identifiers File Alecke et al. , (2003) County, Labor Market Area, and Planning Region (Germany) 1996 9 high tech NA Maggioni (2002) State, Census divisions (US), County, Region (UK), Departement, Region (France), Provincia, Regione (Italy) 1991 1995 5 high tech County Business Patterns (US), ONS (UK), Institut National de la Statistique et des Etudes Economiques (INSEE) (France), ISTAT (Italy)
136 Table A 1. Continued General data Specific data Examples of empirical work Geographical scale and location Time frame Industries Data sources Innovative outputs Patent counts Number of applied patents Guerrero and Sero (1997) Province (Spain) 1989 1992 16 productive sectors RIP data Number of granted patents Thompson (1962) Standard Metropolitan Area (US) 1947 16 patent classes Official Gazette of the Patent Office Patent families Criscuolo (2005) Countries (EU, Japan, US) 1989 2000 All EPO, USPTO, and Japanese Patent Office (JPO) Innovation counts Number of new products introduced to market Feldman (1994) State (US) 1982 All SBA database Innovative outputs (direct innovation surveys) Hinloopen (2003) Country (EU) 1992, 1996 All Eurostat CIS I and II databases Publications i n R&D Feldman and Lichtenberg (1998) Country (EU) 1962 1996 All CORDIS RTD database Potential innovations Awards in the SBIR Wallsten (2001) State and Metropolitan Statistical Area (US) 1993 1996 7 technology areas SBIR database
137 Table A 1. Continued Ge neral data Specific data Examples of empirical work Geographical scale and location Time frame Industries Data sources Innovative agents Innovation related establishme nts Universities, R&D institutes, and new firms, Venture capital firms Zucker, Darby, a nd Brewer (1994) Zip code, County, and Functional Economi Area (US) 1976 1989 Bio technology GenBank database Formation of new firms Number of single location firms Malecki (1985) Metropolitan Area, State, Region (US) 1986 1993 4 high tech Dun & Bradstr eet Corporate Market Identifiers File Innovative networks Innovation networks Research and technology Vonortas (2002) Country 1980 1998 All MERIT CATI database Knowledge spillovers Patent citations Jaffe, Trajtenberg, and Henderson (1993) State, Standar d Metropolitan Statistical Area (US) 1975, 1980 All industries USPTO
138 APPENDIX B OCCUPATIONAL PROFILE S Figure B 1. Occupational profiles architecture and engineering firms (NAICS 5413) in the Greater Washington area.
139 APPENDIX C SECTOR INDUSTRY TREN DS Table C 1. Industries with the fastest growing and most rapidly declining wage and salary employment, 2008 and projected 2018 Industry description Sector 2007 NAICS Thousands of jobs Change Average annual rate of change 2008 2018 2008 18 2008 18 F astest growing Management, scientific, and technical consulting services Professional and business services 5416 1,008.9 1,844.1 835.2 6.2 Other educational services Educational services 6114 7 578.9 894.9 316.0 4.5 Individual and family services Health care and social asistance 6241 1,108.6 1,638.8 530.2 4.0 Home health care services Health care and social asistance 6216 958.0 1,399.4 441.4 3.9 Specialized design services Professional and business services 5414 143.1 208.7 65.6 3.8 Data processing, h osting, related services, and other information services Information 518, 519 395.2 574.1 178.9 3.8 Computer systems design and related services Professional and business services 5415 1,450.3 2,106.7 656.4 3.8 Lessors of nonfinancial intangible assets ( except copyrighted works) Financial activities 533 28.2 37.9 9.7 3.0 Offices of health practitioners Health care and social asistance 6211, 6212, 6213 3,713.3 4,978.6 1,265.3 3.0 Personal care services Other services 8121 621.6 819.1 197.5 2.8 Outpatien t, laboratory, and other ambulatory care services Health care and social asistance 6214, 6215, 6219 989.5 1,297.9 308.4 2.8 Facilities support services Professional and business services 5612 132.7 173.6 40.9 2.7
140 Table C 1. Continued Industry descript ion Sector 2007 NAICS Thousands of jobs Change Average annual rate of change 2008 2018 2008 18 2008 18 Fastest growing Software publishers Information 5112 263.7 342.8 79.1 2.7 Independent artists, writers, and performers Leisure and hospitality 711 5 50.4 64.8 14.4 2.5 Local government passenger transit State and local government NA 268.6 342.6 74.0 2.5 Elementary and secondary schools Educational services 6111 854.9 1,089.7 234.8 2.5 Scientific research and development services Professional and b usiness services 5417 621.7 778.9 157.2 2.3 Waste management and remediation services Professional and business services 562 360.2 451.0 90.8 2.3 Other miscellaneous manufacturing Manufacturing 3399 321.0 399.4 78.4 2.2 Community and vocational rehabili tation services Health care and social asistance 6242, 6243 540.9 672.0 131.1 2.2 Most rapidly declining Cut and sew apparel manufacturing Manufacturing 3152 155.2 66.7 88.5 8.1 Apparel knitting mills Manufacturing 3151 26.2 12.5 13.7 7.1 Textile a nd fabric finishing and fabric coating mills Manufacturing 3133 48.3 23.5 24.8 7.0 Fabric mills Manufacturing 3132 65.4 35.0 30.4 6.1 Audio and video equipment manufacturing Manufacturing 3343 27.0 14.6 12.4 6.0 Apparel accessories and other appar el manufacturing Manufacturing 3159 17.0 9.2 7.8 6.0 Fiber, yarn, and thread mills Manufacturing 3131 37.4 20.7 16.7 5.7 Textile furnishings mills Manufacturing 3141 75.4 41.9 33.5 5.7 Railroad rolling stock manufacturing Manufacturing 3365 28.4 1 7.5 10.9 4.7 Footwear manufacturing Manufacturing 3162 15.8 10.0 5.8 4.5 Pulp, paper, and paperboard mills Manufacturing 3221 126.1 81.9 44.2 4.2 Basic chemical manufacturing Manufacturing 3251 152.1 99.9 52.2 4.1
141 Table C 1. Continued Industry description Sector 2007 NAICS Thousands of jobs Change Average annual rate of change 2008 2018 2008 18 2008 18 Most rapidly declining Semiconductor and other electronic component manufacturing Manufacturing 3344 432.4 286.8 145.6 4.0 Computer and peripheral equipment manufacturing Manufacturing 3341 182.8 124.7 58.1 3.8 Other textile product mills Manufacturing 3149 72.2 49.4 22.8 3.7 Federal enterprises except the Postal Service and electric utilities Federal government NA 63.5 44.9 18.6 3.4 Leather and hide tanning and finishing, and other leather and allied product manufacturing Manufacturing 3161, 3169 17.8 13.0 4.8 3.1 Cutlery and handtool manufacturing Manufacturing 3322 49.1 35.9 13.2 3.1 Manufacturing and reproducing magnetic and optical media Manufacturing 3346 34.9 26.0 8.9 2.9 Ventilation, heating, air conditioning, and commercial refrigeration equipment manufacturing Manufacturing 3334 149.5 112.8 36.7 2.8 Source: Employment Projections Program, U.S. Department of La bor, U.S. Bureau of Labor Statistics (2010)
142 APPENDIX D CAPITOL REGION SCATTERPLOTS BY AGE COHORT
143 Figure D 1. Business patterns for 20 to 24 year olds in the Capitol Region, 1990
144 Figure D 2. Business patterns for 25 to 3 4 year olds in the Capi tol Region, 1990
145 Figure D 3. Business patterns for 35 to 44 year olds in the Capitol Region, 1990
146 Figure D 4. Business patterns for 45 to 54 year olds in the Capitol Region, 1990
147 Figure D 5. Business patterns for 20 to 24 year olds in the Capit ol Region, 2000
148 Figure D 6. Business patterns for 25 to 34 year olds in the Capitol Region, 2000
149 Figure D 7. Business patterns for 35 to 44 year olds in the Capitol Region, 2000
150 Figure D 8. Business patterns for 45 to 54 year olds in the Capito l Region, 2000
151 Figure D 9. Business patterns for 20 to 24 year olds in the Capitol Region, 2008
152 Figure D 10. Business patterns for 25 to 34 year olds in the Capitol Region, 2008
153 Figure D 11. Business patterns for 35 to 44 year olds in the Capit ol Region, 2008
154 Figure D 12. Business patterns for 45 to 54 year olds in the Capitol Region, 2008
155 APPENDIX E REGRESSION RESULTS BY AGE COHORT Table E 1. Regression summary output 1990 , 20 to 24 year old Regression Statistics Multiple R 0.8879449 98 R Square 0.788446319 Adjusted R Square 0.703824847 Standard Error 0.339765802 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.151200431 1.075600216 9.317331616 0.020585028 Residual 5 0.577204 0.1154408 Total 7 2.728404431 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 1.627557887 1.052570514 1.546269695 0.182706808 4.33327653 1.078160756 4.33327653 1.078160756 log total pop 1990 0.72 0940328 0.169144801 4.262267148 0.007997541 0.286139775 1.155740881 0.286139775 1.155740881 1990 20 24 Y.O. % 2.377452531 7.280057476 0.326570572 0.757219655 16.33653098 21.09143604 16.33653098 21.09143604
156 Table E 2. Regression summary output 19 90, 25 to 34 year old Regression Statistics Multiple R 0.886476598 R Square 0.785840758 Adjusted R Square 0.700177062 Standard Error 0.341851729 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.144 091407 1.072045704 9.17355647 0.021224724 Residual 5 0.584313024 0.116862605 Total 7 2.728404431 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 1.283672176 1.181279234 1.086679711 0.3267667 95 4.320247119 1.752902766 4.320247119 1.752902766 log total pop 1990 0.728108461 0.170028131 4.282282331 0.007846448 0.291037236 1.165179687 0.291037236 1.165179687 1990 25 34 Y.O. % 0.823584981 3.903356738 0.210994033 0.841221706 10.85748291 9.2 10312948 10.85748291 9.210312948 Observation Predicted log 1998 business patterns Residuals Standard Residuals 1 2.167685419 0.017006012 0.058861146 2 2.313679539 0.001925678 0.006665148 3 2.76248848 0.544793567 1.885637504 4 2.862108262 0.0187053 31 0.064742822 5 1.676496782 0.09435523 0.326581977 6 1.470239358 0.108511522 0.375579685 7 2.834916104 0.051012525 0.176564364 8 2.81444041 0.513410414 1.777014251
157 Table E 3. Regression Summary Output 1990, 35 to 44 year old Regression Stati stics Multiple R 0.896028544 R Square 0.802867151 Adjusted R Square 0.724014012 Standard Error 0.327981139 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.190546293 1.095273147 10.18180325 0.01725 4306 Residual 5 0.537858138 0.107571628 Total 7 2.728404431 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 0.437296077 1.691050127 0.25859439 0.806255936 4.784278816 3.909686663 4.7842788 16 3.909686663 log total pop 1990 0.707495303 0.164781434 4.293537733 0.007762934 0.283911143 1.131079462 0.283911143 1.131079462 1990 35 44 Y.O. % 5.181371716 7.476997133 0.69297495 0.519188848 24.40160473 14.0388613 24.40160473 14.0388613 RESI DUAL OUTPUT Observation Predicted log 1998 business patterns Residuals Standard Residuals 1 2.183092874 0.001598557 0.005766911 2 2.325668746 0.013914885 0.050198966 3 2.841502091 0.465779956 1.680335271 4 2.745326316 0.135487276 0.488780262 5 1 .809470118 0.038618106 0.139317644 6 1.341864504 0.019863332 0.071658423 7 2.806412168 0.022508588 0.08120138 8 2.848717537 0.547687541 1.975822877
158 Table E 4. Regression summary output 1990, 45 to 54 year old Regression Statistics Multiple R 0.892684033 R Square 0.796884783 Adjusted R Square 0.715638697 Standard Error 0.332920548 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.174223974 1.087111987 9.808285133 0.018593286 Residual 5 0. 554180457 0.110836091 Total 7 2.728404431 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 0.580321293 1.76449572 0.328887901 0.755568985 5.11610194 3.955459354 5.11610194 3.955459354 log t otal pop 1990 0.722132921 0.165271512 4.369373235 0.007226317 0.297288974 1.146976868 0.297288974 1.146976868 1990 45 54 Y.O. % 7.44330785 13.18266885 0.564628296 0.596715724 41.33043693 26.44382123 41.33043693 26.44382123 RESIDUAL OUTPUT Obse rvation Predicted log 1998 business patterns Residuals Standard Residuals 1 2.275696304 0.091004874 0.323435784 2 2.400789966 0.089036105 0.316438683 3 2.835080537 0.47220151 1.678227325 4 2.71030847 0.170505123 0.605983568 5 1.67331396 0.09753805 1 0.346655018 6 1.389337147 0.027609311 0.098124846 7 2.791530581 0.007627002 0.027106739 8 2.825997388 0.524967392 1.865759859
159 Table E 5. Regression summary output 2000, 20 to 2 4 year old Regression Statistics Multiple R 0.852398722 R Square 0.726583582 Adjusted R Square 0.617217015 Standard Error 0.439092351 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.561785357 1.280892679 6.643562113 0.039089622 Residual 5 0.964010464 0.1928020 93 Total 7 3.525795821 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 1.961657445 1.200901351 1.633487583 0.163294596 5.048672644 1.125357755 5.048672644 1.125357755 log total pop 2000 0. 702621429 0.225330507 3.118181554 0.026306243 0.123390922 1.281851936 0.123390922 1.281851936 2000 20 24 Y.O. % 9.211217101 9.176285062 1.003806773 0.361548107 14.3771746 32.7996088 14.3771746 32.7996088 RESIDUAL OUTPUT Observation Predicted log 2000 business patterns Residuals Standard Residuals 1 2.339674316 0.152153595 0.410006144 2 2.55036096 0.251507883 0.677734741 3 2.918090831 0.70629441 1.903241571 4 2.742659598 0.104913061 0.282707744 5 1.7 38363635 0.022360292 0.060253962 6 1.204531605 0.117687689 0.317131354 7 2.67348527 0.085426623 0.230197914 8 2.860141619 0.588300013 1.585283735
160 Table E 6. Regression summary output 2000, 25 to 3 4 year old Regression Statist ics Multiple R 0.820784457 R Square 0.673687126 Adjusted R Square 0.543161976 Standard Error 0.479690018 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.375283252 1.187641626 5.161358764 0.0608254 77 Residual 5 1.150512569 0.230102514 Total 7 3.525795821 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 1.778427148 1.44102792 1.234137884 0.271991002 5.482707344 1.925853047 5.482707344 1.925853047 log total pop 2000 0.761486591 0.237557481 3.205483524 0.023848663 0.150825646 1.372147536 0.150825646 1.372147536 2000 25 34 Y.O. % 0.712151714 3.875283076 0.183767663 0.861415881 9.249580569 10.673884 9.249580569 10.673884 R ESIDUAL OUTPUT Observation Predicted log 2000 business patterns Residuals Standard Residuals 1 2.292190153 0.104669433 0.258180362 2 2.419529261 0.120676184 0.297663035 3 2.732469543 0.891915698 2.20002261 4 2.890677454 0.043104795 0.106323415 5 1.639427211 0.076576132 0.188884692 6 1.374670073 0.052450779 0.129376464 7 2.848911689 0.089999796 0.221995853 8 2.829432451 0.557590844 1.375368173
161 Table E 7. Regression summary output 2 000, 35 to 4 4 year old Regression Statistics Multiple R 0.918131524 R Square 0.842965495 Adjusted R Square 0.780151694 Standard Error 0.332767667 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.97 2124221 1.48606211 13.42006805 0.009772096 Residual 5 0.5536716 0.11073432 Total 7 3.525795821 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 3.824828878 2.50376725 1.527629566 0.187139887 2 .611309735 10.26096749 2.611309735 10.26096749 log total pop 2000 0.737700352 0.165114056 4.467822856 0.006592529 0.31326116 1.162139545 0.31326116 1.162139545 2000 35 44 Y.O. % 30.78807144 13.17605347 2.336668677 0.066655805 64.65819516 3.082052 283 64.65819516 3.082052283 RESIDUAL OUTPUT Observation Predicted log 2000 business patterns Residuals Standard Residuals 1 2.028459657 0.159061063 0.565570538 2 2.409805839 0.110952763 0.394512727 3 3.353340923 0.2 71044319 0.963747368 4 2.562304504 0.285268155 1.014322805 5 1.786381381 0.070378038 0.250241912 6 1.301590347 0.020628948 0.073349975 7 2.718435873 0.04047602 0.143919849 8 2.866989311 0.595147704 2.116155896
162 Table E 8. Regression summary output 2000, 45 to 5 4 year old Regression Statistics Multiple R 0.832037026 R Square 0.692285613 Adjusted R Square 0.569199858 Standard Error 0.4658193 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.440857721 1.220428861 5.62441701 0.052525416 Residual 5 1.0849381 0.21698762 Total 7 3.525795821 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 0.412894725 2.4 62415971 0.167678707 0.873407143 6.742736493 5.916947043 6.742736493 5.916947043 log total pop 2000 0.693733883 0.258804614 2.68053136 0.043792249 0.028455444 1.359012321 0.028455444 1.359012321 2000 45 54 Y.O. % 5.91294299 10.17033935 0.581390924 0.586200046 32.05663258 20.2307466 32.05663258 20.2307466 RESIDUAL OUTPUT Observation Predicted log 2000 business patterns Residuals Standard Residuals 1 2.317689384 0.130168663 0.330638166 2 2.441773596 0.14292052 0. 363028839 3 2.802814308 0.821570934 2.086851788 4 2.780828554 0.066744105 0.169535031 5 1.754020769 0.038017425 0.096567112 6 1.250090184 0.072129111 0.183213352 7 2.809111344 0.050199451 0.127510371 8 2.870979697 0 .59913809 1.521855683
163 Table E 9. Regression summary output 2008, 20 to 2 4 year old Regression Statistics Multiple R 0.812657215 R Square 0.660411748 Adjusted R Square 0.490617623 Standard Error 0.443892199 Observations 7 ANOVA df SS MS F Significance F Regression 2 1.532770563 0.766385282 3.889485253 0.115320181 Residual 4 0.788161139 0.197040285 Total 6 2.320931702 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0 % Upper 95.0% Intercept 1.967116657 1.723422523 1.14140127 0.317394024 6.752104685 2.817871371 6.752104685 2.817871371 log total pop 2008 0.552721331 0.312441522 1.769039302 0.151609372 0.314755402 1.420198065 0.314755402 1.420198065 2006/08 20 24 Y.O. % 22.0572137 13.63655341 1.617506493 0.181077599 15.80392828 59.91835567 15.80392828 59.91835567 RESIDUAL OUTPUT Observation Predicted log 2008 business patterns Residuals Standard Residuals 1 1.911012751 0.228866335 0.6 31466087 2 2.282703281 0.092371583 0.25486283 3 3.212607133 0.450811079 1.243834781 4 2.685744408 0.145485285 0.401409075 5 1.887665592 0.109514341 0.302161489 6 2.649160721 0.073473202 0.202720227 7 2.947169979 0.6 96749977 1.922405851
164 Table E 10. Regression summary output 2008, 25 to 3 4 year old Regression Statistics Multiple R 0.707928936 R Square 0.501163378 Adjusted R Square 0.251745067 Standard Error 0.537997614 Observations 7 ANOVA df SS MS F Significance F Regression 2 1.163165972 0.581582986 2.00932873 0.248837975 Residual 4 1.15776573 0.289441432 Total 6 2.320931702 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 2.36540616 2.553293789 0.926413627 0.406666309 9.454486203 4.723673883 9.454486203 4.723673883 log total pop 2008 0.759211039 0.380060283 1.997606889 0.116434283 0.296005473 1.81442755 0.296005473 1.81442755 2006/08 25 34 Y .O. % 4.725537461 6.655461915 0.710023965 0.516890551 13.7529872 23.20406212 13.7529872 23.20406212 RESIDUAL OUTPUT Observation Predicted log 2008 business patterns Residuals Standard Residuals 1 2.37195245 0.232073364 0.52831 2124 2 2.575390369 0.385058671 0.876581271 3 2.873066924 0.790351288 1.79922487 4 2.698248268 0.132981426 0.30273056 5 1.579708324 0.198442926 0.451752851 6 2.704228828 0.018405095 0.041898969 7 2.773468703 0.5230487 1.190713856
165 Table E 11. Regression summary output 2008, 35 to 4 4 year old Regression Statistics Multiple R 0.73508728 R Square 0.54035331 Adjusted R Square 0.310529965 Standard Error 0.516432129 Observations 7 ANOVA df SS MS F Significance F Regression 2 1.254123127 0.627061564 2.351168067 0.21127508 Residual 4 1.066808575 0.266702144 Total 6 2.320931702 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0 % Intercept 1.028329319 3.104164641 0.331274091 0.757065912 7.590213403 9.646872041 7.590213403 9.646872041 log total pop 2008 0.538546806 0.383734517 1.403435922 0.233151635 0.526871016 1.603964628 0.526871016 1.603964628 2006/08 35 44 Y.O. % 8.689702316 9.220597401 0.942422919 0.399337002 34.29018484 16.91078021 34.29018484 16.91078021 RESIDUAL OUTPUT Observation Predicted log 2008 business patterns Residuals Standard Residuals 1 2.008882462 0.130996625 0.310665111 2 2.103847136 0.086484562 0.205102507 3 2.877442227 0.785975985 1.863981737 4 2.875366915 0.044137221 0.104673648 5 1.960683067 0.182531816 0.43288342 6 2.901929008 0.179295085 0.425207348 7 2.847913052 0.5974930 5 1.416984939
166 Table E 12. Regression summary output 2008, 45 to 5 4 year old Regression Statistics Multiple R 0.746724713 R Square 0.557597797 Adjusted R Square 0.336396695 Standard Error 0.506652075 Observations 7 ANOVA df SS MS F Significance F Regression 2 1.294146403 0.647073202 2.520773145 0.19571971 Residual 4 1.026785299 0.256696325 Total 6 2.320931702 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 0.851766106 2.781676377 0.306206039 0.774723784 6.871405656 8.574937867 6.871405656 8.574937867 log total pop 2008 0.72663181 0.346474377 2.097216584 0.103979539 0.235335279 1.688598898 0.235335279 1.688598898 2006/08 45 54 Y.O. % 15.40626259 14.83362997 1.038603674 0.35764563 56.59102191 25.77849673 56.59102191 25.77849673 RESIDUAL OUTPUT Observation Predicted log 2008 business patterns Residuals Standard Residuals 1 2.406898371 0.267019285 0.6454735 24 2 2.481504917 0.291173219 0.703861534 3 3.028626148 0.634792064 1.534501415 4 2.575484698 0.255744996 0.618219855 5 1.575821982 0.202329269 0.489096456 6 2.659469317 0.063164605 0.152689647 7 2.848258432 0.59783843 1.445172314
167 APPENDIX F SPATIAL STATISTICS METRO WASHINGTON, DC REGION Figure F 1. District of Columbia spatial statistics summary
168 Figure F 1. Continued
169 Figure F
170 Figure F 2. Continued
171 Figure F 3. Fairfax County, Virginia spatial statistics summary
172 Figure F 3. Continued
173 Figure F 4. Montgomery County, Maryland spatial statistics summary
174 Figure F 4. Continued
175 Figure F 5. Fairfax City, Virginia spatial statist ics summary
176 Figure F 5. Continued
177 Figure F 6. Arlington County, Virginia spatial statistics summary
178 Figure F 6. Continued
179 Figure F 7. Alexandria City, Virginia spatial statistics summary
180 Figure F 7. Continued
181 APPENDIX G CAPTIOL REGION S CATTERPLOTS EDUCATIONAL ATTAINME NT FACTOR
182 Figure G 1. Business patterns and Educational Attainment Factor in the Capitol Region, 1990
183 Figure G 2. Business patterns and Educational Attainment Factor in the Capitol Region, 2000
184 Figure G 3. Busines s patterns and Educational Attainment Factor in the Capitol Region, 2008
185 APPENDIX H REGRESSION RESULTS EDUCATIONAL ATTAINME NT FACTOR Table H 1. Regression summary output Educational Attainment Factor 1990 Regression Statistics Multiple R 0.89634864 9 R Square 0.803440901 Adjusted R Square 0.724817262 Standard Error 0.327503501 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.192111715 1.096055857 10.21882103 0.017129034 Residual 5 0.536292716 0.1 07258543 Total 7 2.728404431 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 2.613651036 1.868309055 1.39893934 0.220702625 7.416292355 2.188990284 7.416292355 2.188990284 log total pop 19 90 0.76794189 0.173265563 4.432166885 0.006814324 0.322548582 1.213335199 0.322548582 1.213335199 1990 Educ Attainment Factor 4.587321384 6.512174632 0.704422354 0.512616885 12.15275644 21.3273992 12.15275644 21.3273992
186 Table H 2. Regression summary o utput Educational Attainment Factor 2000 Regression Statistics Multiple R 0.823864554 R Square 0.678752803 Adjusted R Square 0.550253925 Standard Error 0.475952103 Observations 8 ANOVA df SS MS F Significance F Regre ssion 2 2.393143798 1.196571899 5.28216908 0.058492255 Residual 5 1.132652023 0.226530405 Total 7 3.525795821 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 2.388471868 2.523568471 0.946466 044 0.387371171 8.875511141 4.098567405 8.875511141 4.098567405 log total pop 2000 0.792544546 0.25262851 3.137193609 0.025747674 0.143142288 1.441946804 0.143142288 1.441946804 2000 Educ Attainment Factor 2.625400181 7.805019521 0.336373301 0.75024715 2 17.43804123 22.68884159 17.43804123 22.68884159 RESIDUAL OUTPUT Observation Predicted log 2000 business patterns Residuals Standard Residuals 1 2.265210576 0.077689855 0.193136821 2 2.432165897 0.13331282 0.331415399 3 2.710114462 0.914270779 2.272875289 4 2.950765214 0.103192555 0.256536482 5 1.593083354 0.12291999 0.305578843 6 1.426175614 0.103956319 0.258435197 7 2.918793232 0.15988134 0.397464684 8 2.730999485 0.45915787 9 1.141465549
187 Table H 3. Regression summary output Educational Attainment Factor 2008 Regression Statistics Multiple R 0.799426593 R Square 0.639082878 Adjusted R Square 0.494716029 Standard Error 0.516823289 Observations 8 ANOVA df SS MS F Significance F Regression 2 2.364851374 1.182425687 4.426797998 0.078256193 Residual 5 1.33553156 0.267106312 Total 7 3.700382934 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Uppe r 95.0% Intercept 1.600391267 3.130452381 0.511233225 0.63095546 9.647475295 6.446692762 9.647475295 6.446692762 log total pop 2008 0.76430785 0.270330893 2.827304863 0.036791111 0.069400165 1.459215534 0.069400165 1.459215534 2008 Educ Attainment F actor 0.480750769 10.43031173 0.046091697 0.965021428 27.29272064 26.3312191 27.29272064 26.3312191 RESIDUAL OUTPUT Observation Predicted log 2008 business patterns Residuals Standard Residuals 1 2.220268729 0.080389643 0.18 4044159 2 2.335689737 0.145358039 0.332782896 3 2.706050694 0.957367518 2.191798528 4 2.875151221 0.043921528 0.100554006 5 1.631978618 0.146172633 0.334647828 6 1.37785814 0.122585635 0.280647723 7 2.854556473 0. 13192255 0.30202367 8 2.829782759 0.579362757 1.326393902
188 APPENDIX I REGRESSION RESULTS PRINCIPAL COMPONENTS ANALYSIS, CAPITOL REGION Table I 1. Regression summary output for the PCA of the Washington, DC region Regression Statistics Multiple R 0.890398147 R Square 0.79280886 Adjusted R Square 0.772347993 Standard Error 3.647723296 Observations 61 ANOVA df SS MS F Significance F Regression 2 3003.952771 1501.976385 112 .8806 1.00873E 20 Residual 59 785.0472295 13.30588525 Total 61 3789 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 0 #N/A #N/A #N/A #N/A #N/A #N/A #N/A IP_DENSITY 0.0293870 73 0.023845189 1.232410967 0.222684 0.018327041 0.077101186 0.018327041 0.077101186 TranAcces 1.012318841 0.176253759 5.743530517 3.43E 07 0.659635885 1.365001798 0.659635885 1.365001798
189 REFERENCE LIST Anas, A., R. Arnott, et al. (1998). "Urban Spatia l Structure." Journal of Economic Literature 36 (3): 1426. Bell, D. (1973). The coming of post industrial society; a venture in social forecasting . New York,, Basic Books. Blakely, E. J. and T. K. Bradshaw (2002). Plan ning local economic development : the ory and practice . Thousand Oaks, Calif, Sage. Blakely, E. J. and N. G. Leigh (2010). Plan ning local economic development : theory and practice . Los Angeles, Sage. Brown, H. (2007). Knowledge and innovation : a comparative study of the USA, the UK and Ja pan . Milton Park, Abingdon, England; New York, Routledge. Buitelaar, E. (2007). : applying transaction cost economics to planning & development . ; Malden, MA, Blackwell Publishing. Carr, M. and P. Zwick (2005). "Using GIS suitability analysis to identify potential future land use conflicts in North Central Florida." Journal of Conservation Planning 1 : 58 73. Carr, M. H. and P. D. Zwick (2007). Smart land use analysis: the LUCIS model land use conflict identification s trategy . Redlands, Calif., ESRI Press. Chapman, K. and D. F. Walker (1987). : principles and policies . Oxford, Ox., UK ; New York, NY, USA, Blackwell. Clark, T. N. (2002). Urban Amenities: Lakes, Opera, and Juice Bars Do They Drive De velopment? The City as an Entertainment Machine . New York, JAI Press/Elsevier. 9 . Cromley, R. G. and D. M. Hanink (2003). "Scale independent land use allocation modeling in raster GIS." Cartography and Geographic Information Science 30 (4): 343 350. DeVol, R., A. Bedroussian, et al. (2008). Best Performing Cities 2008: Where America's Jobs Are Created and Sustained, Miliken Institute. Donegan, M., J. Drucker, et al. (2008). "Which Indicators Explain Metropolitan Economic Performance Best? Traditional or Crea tive Class." Journal of the American Planning Association 74 (2): 180 195. Expert Choice (2008). "Collaboration and decision support software for groups and organizations." Retrieved December 1, 2008, from http://expe rtchoice.com/ .
190 Feldman, M. and R. Martin (2004). Jurisdictional Advantage. National Bureau of Economic Research (NBER) Innovation and the Economy Conference . Washington, DC, National Bureau of Economic Research. Feser, E. J., K. Koo, et al. (2001). "Incor porating Spatial Analysis in Applied Industry Cluster Studies." Economic Development Quarterly : 42. Fingleton, B., D. C. Igliori, et al. (2007). Employment growth and clusters dynamics of creative industries in Great Britain. The Economic Geography of Inno vation . K. R. Polenske. New York, Cambridge : 60 84. Fisher, M. (2009). After 1,250 Columns, It's Time to Shift Gears. The Washington Post . Washington, DC, The Washington Post Company : 1. Florida, R. (2005). Cities and the Creative Class , Routledge. Flori da, R. (2008). Who's Your City? New York, Basic Books. Florida, R. L. (2004). The rise of the creative class : and how it's transforming work, leisure, community and everyday life . New York, NY, Basic Books. Fujita, N. (2007). "Myrdal's Theory of Cumul ative Causation." Evolutionary and Institutional Economics Review 3 (2): 275 283. Fyfe, N. R. and J. T. Kenny (2005). The urban geography reader . London ; New York, Routledge. Gottlieb, P. D. (1995). "Residential Amenities, Firm Location and Economic Deve lopment." Urban Studies 32 (9): 1413 1436. Gottlieb, P. D. (2004). Labor Supply Pressures and the "Brain Drain": Signs from Census 2000. Living Cities Census Series . Washington, DC, The Brookings Institution : 24. Hearn, G. and D. Rooney (2008). Knowledge p olicy : challenges for the 21st century . Cheltenham, UK Northampton, MA, Edward Elgar. Hospers, G. J. and R. van Dalm (2005). How to create a creative city? The viewpoints of Richard Florida and Jane Jacobs. foresight , Emerald Group Publishing Limited. 7: 8 12. Houghton, J. and P. Sheehan (2000). A Primer on the Knowledge Economy. Melbourne City MC, Victoria Australia, Victoria University : 25. Hulse, D. W., A. Branscomb, et al. (2004). "Envisioning alternatives: Using citizen guidance to map future land an d water use." Ecological Application 14 (2): 325 341.
191 Husain, I. Z. (1967). Economic factors in economic growth . Bombay, New York,, Allied Publishers. Investopedia (2009). "Gross Domestic Product GDP." Retrieved December 11, 2009, 2009, from http://www.investopedia.com/terms/g/gdp.asp . Jacobs, J. (1970). ities . New York, Vintage Books. Jacobs, J. (2000). The nature of economies . New York, Modern Library. Kang, C. (2009). Rura l Riddle: Do Jobs Follow Broadband Access? The Washington Post . Washington, DC, The Washington Post Company. Kotkin, J. (1999). The future of the center: The core city in the new economy. Los Angeles, Reason Public Policy Institute and Pepperdine Universit y Institute for Public Policy. Krupka, D. (2004). Location specific human capital, migration and amenities, Georgia State Univerity Andrew Young School of Policy Studies : 45. Kuhn, T. S. (1996). The structure of scientific revolutions . Chicago, IL, Uni versity of Chicago Press. Levy, F. (2006). Education and Inequality in the Creative Age. Cato Unbound . Washington, DC, Cato Institute. 2008 . Lundvall, B. Ã…., B. o. Johnson, et al. (2007). National systems of production, innovation, and competence building. The economic geography of innovation . K. R. Polenske. New York, Cambridge University Press : 213 240. Madrick, J. G. (2002). Why economies grow : the forces that shape prosperity and how we can get them working again . New York, NY, Basic Books. McCann, P . (2002). Industrial location economics . Cheltenham, UK ; Northampton, MA, Edward Elgar. McHarg, I. L. and American Museum of Natural History. (1969). Design with nature . Garden City, N.Y. , Published for the American Museum of Natural History by the Na tural History Press. Miles, I. (2008). Knowledge Services. Knowledge policy: challenges for the 21st century . G. Hearn and D. Rooney. Northampton, Edward Elgar Publishing : 11 26. National Research Council (2002). Community and Quality of Life: Data needs f or informed decision making, Board on Earth Sciences and Resources.
192 Ndubisi, F. and Center for American Places. (2002). Ecological planning : a historical and comparative synthesis . Baltimore, Johns Hopkins University Press. North, D. C. (1955). "Locatio n Theory and Regional Economic Growth." The Journal of Political Economy 63 (3): 243 258. Parr, J. B. (1999). "Growth pole Strategies in Regional Economic Planning: A Retrospective View. Part 1. Origins and Advocacy." Urban Studies 36 (7): 1195. Parr, J. B. (1999). "Growth pole Strategies in Regional Economic Planning: A Retrospective View. Part 2. Implementation and Outcome." Urban Studies 36 (8): 1247. Parr, J. B. (2003). Reinventing Regions? The case of polycentric urban regions. Reinventing regions in a gl obal economy . Pisa, Italy, Regional Studies Association. Peters, D. M. (2006). "In defense of American manufacturing." Retrieved December 15, 2009, 2009, from www.profiles inc.com/newsletter.html . Polenske, K. R. (2007). Introduction. The economic geography of innovation . K. R. Polenske. New York, Cambridge Press : 3 12. Power, D. and M. Lundmark (2003). Working through knowledge pools: Labour market dynamics, the transference of knowledge and ide as, and industrial clusters. Clusters in Urban and Regional Development . A. Cumbers and D. MacKinnon. New York, Routledge : 55 74. Prosperi, D. (2005). Creative economy integration across space: Evidence for two groups. AESOP . Vienna : 8. Randolph, J. (2004) . Environmental land use planning and management . Washington, Island Press. Ratanawaraha, A. and K. R. Polenske (2007). Measuring the geography of innovation: a literature review. The economic geography of innovation . K. R. Polenske. New York, Cambridg e University Press : 30 59. Rios, V. and J. L. Romo (2008). Closeness to markets not institutional quality determines firm size, Harvard University. Schumpeter, J. A. and R. Opie (1934). The theory of economic development; an inquiry into profits, capi tal, credit, interest, and the business cycle . Cambridge, Mass.,, Harvard University Press. Stover, M. E. and C. L. Leven (1992). "Methodological issues in the determination of the quality of life in urban areas." Urban Studies 29 (5): 737 754.
193 The Economi st (2008) Surviving the crash. The Economist (2005). The Economist Intelligence Unit's quality of life index : 1 4. Xiang, W. N. (2001). "Weighting by choosing: a weight elicitation method for map overlays." Landscape and Urban Planning 56 : 61 73.
194 BIOG RAPHICAL SKETCH Iris Patten has a Doctor of Philosophy and Master of Arts degree in Urban and Regional Planning from the University of Florida. In 2004 she received a Bachelor of Science Degree in Environmental Science and Policy from the University of Mar yland, College Park. Currently, Iris is an Assistant Professor of Practice at the University of Arizona in the School of Geography and Regional Development . Her research focuses on using GIS as a tool to solve growth management issues, GIS as a tool for decision making, and using GIS to identify opportunities for renewable energy developments. Iris also works with organizations and governments to develop strategic plans. Previous international research projects include developing a water capture and fil tration system for a small village in the West African country of Burkina Faso, sponsored by the U.S. Environmental Protection Agency, and developing a feasibility study for a new housing development in rural South Africa.
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MEASURING REGIONAL AND LOCAL INNOVAT IVE OPPORTUNITY By Iris E. Patten August 2014 Chair: Paul Zwick Cochair: Margaret Carr Major: Design, Construction and Planning While some progress has been made in identifying cities that stimulate creativity to explain patterns of economic growth in k nowledge based industries, no existing research has used land use suitability to justify similar growth patterns. This study establishes the role and contribution of innovative agents, innovative inputs, and innovative outputs on the regional potential to attract architecture and engineering (A&E) firms as well as the ability to create innovative networks at the local scale using land use suitability. Using longitudinal interactions and the Land Use Conflict Identification Strategy to measure suitability and spatial inter actions two factors were created that identified the potential of Pima County, Arizona to create similar innovative networks found in Washington, DC. The results of this analysis indicates that by using longitudinal data and spatial data t o measure innovative potential, you can assess local and regional performance in achieving innovative networks.
1 MEASURING REGIONAL AND LOCAL INNOVATIVE OPPORTUNITY By IRIS E. PATTEN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF P HILOSOPHY UNIVERSITY OF FLORIDA 2014