Clusters and co-location : the spatial geography of innovation in Amsterdam

(1)

Clusters and Co-location:

The Spatial Geography of Innovation in

Amsterdam

Natalia Fedorova

Student number: 11371900

Thesis submitted for the degree

of

Master of Urban and Regional Planning

Supervisor: Bas Hissink Muller

Second assessor: Anna Nikolaeva

(2)

(3)

(4)

Acknowledgements

I would like to thank all the people that gave me guidance and support during this thesis. Firstly, the thesis would not have taken shape without the help of my supervisor, Dr. Bas Hissink Muller, who was always available for discussions and, even when confronted with material outside of his expertise, managed to extend a helping hand. Secondly, I would like to thank Dr. Sjoerd de Vos for his assistance with ArcGis and for his reassurance about the statistical analyses employed in this thesis. Moreover, I am grateful to the Urban and Regional Planning Master class of 2017, without whose companionship and humor this thesis would be a much less enjoyable experience. Likewise, I would like to thank my parents for giving me the opportunity to be forever-a-student and supporting me in my academic adventures. Finally, special thanks goes to my partner, Tomas Langer, who has had to bear the brunt of the madness that comes with writing a Master thesis.

(5)

(6)

Abstract

Innovation is an integral aspect of economic growth and societal transitions. As such, creating conditions to foster it are of paramount interest to policy makers and academics alike. Innovation is extremely clustered, creating an uneven spatial distribution. Understanding the geography of

innovation is thus an important avenue through which to address the determinants of innovation. This thesis aims to add to the literature on the geography of innovation by exploring the within-city geography of innovation in Amsterdam, paying close attention to the innovation policy context. Moreover, as divergence between cities, also in terms of innovation, has been argued to be a result of population factor differences, three population factor theories are tested in relation to the spatial pattern of innovation: human capital theory, creative class theory, and social capital theory. A dataset of innovative enterprises is constructed and mapped using GIS. Regression analyses are utilized to statistically test population factors. The results reveal a heterogeneous and variegated geography of innovation at the within-city scale of Amsterdam. Innovation appears to cluster in the center of the city, but notable innovation hubs also exists outside the center. Out of the population factors tested, the creative class emerges as the strongest predictor, however, the results also suggest the relationship is not causal; innovation and the creative class tend to locate in the same areas. The thesis provides a detailed discussion of the results, concluding that the within-city scale sheds lights on the spatial and population factor determinants of the geography of innovation, hopefully providing a more nuanced perspective on innovation for innovation fostering policy.

(7)

Introduction

“The question remains: why then did particular places seize the torch of innovation?” Peter Hall, “Cities in Civilization”, 1998

Innovation is not homogenously spread across the geographic landscape, it clusters in specific areas, resulting in a dramatically unequal spatial pattern. The basic observation that places differ in terms of innovative activity has led to decades of research on the geography of innovation, aiming to elucidate why certain places are more innovative than others. It is to this body of research that this thesis aims to add by addressing the spatial geography of innovation at a city scale.

Innovation has always been important in human society, allowing us to achieve feats such as the moon landing, various vaccines, as well as democracy (Henrich, 2010). However, innovation has become integral in today’s world due to its importance in the global economy and its ability to

provide solutions to societal problems. Today’s economy is a knowledge economy, where knowledge, rather than manufacturing capability, is the primary source of competitive advantage (Meusburger, Meskioui, & Gluckler, 2013). As innovation is the economic application of new knowledge, it has become a key driver of economic growth (Salter & Alexy, 2013).

However, innovation can also be the source of solutions. Globally, countries face challenges brought on not only by climate change, but also by widening social inequality and polarization. The central role of innovation in socio-technical transitions, the rising of new social and technological paradigms, has attracted attention from academics and policy makers alike to harness innovation and innovate out of the crises we face (Bettencourt & West, 2010; Grimm, Fox, Baines, & Albertson, 2013; Rennings, 2000). In particular, innovation from alternative sources, such as civil society, have come to the fore (G Seyfang & Smith, 2007; Gill Seyfang & Longhurst, 2013).

Thus, innovation is today’s panacea: not only an engine of economic prosperity, but also a mechanism through which to adapt to climate change and changing social patterns. Naturally, there is thus a premium available to any nation, region, or city, that is able to foster innovation. To construct policy that will benefit local innovative activity, it is integral to understand the factors that affect the process of innovation. In turn, spatial geography is fundamental because the process of innovation is

dependent on the concentration and proximity of agents (Asheim & Gertler, 2009). That is, innovation is dependent on localized knowledge flows and the face-to-face interactions necessary to access them (Storper & Venables, 2004).

(11)

Innovation is not a solitary activity, it is dependent on a diverse milieu of agents. Spatial geography reflects this milieu, and thus understanding where innovative activity concentrates allows us to explore the features of local milieus that foster innovation. That is, geography is important insofar as it reflects the social environment in which the interaction necessary for innovation occurs; space is relational (Boschma, 2005; Coenen, Benneworth, & Truffer, 2012; Morgan, 2004). Although innovation may seem like an ephemeral process, it is conducted by people and thus occurs in the reality of space. Understanding the properties of such real space thus becomes fundamental to understanding the whole process, and subsequently supporting it.

Cities have been identified as the crucibles of innovation because they aid the flow and mix of ideas. That is, through the heterogeneity and spatial proximity they provide to economic agents, cities foster innovation. However, research on the determinants of localized innovation focus on the regional scale, comparing across cities but not addressing the spatial geography within them. This is

problematic as it treats cities as homogenous cauldrons, where economic activity and populations are equally dispersed, a fact any city-dweller can reject. That is, cities are the de facto geography of innovation, but as innovation stems from much smaller units, economic agents, a much finer understanding of geography is required. Moreover, cities are the forefront of global challenges, but also solutions, and thus understanding them scientifically is of paramount importance (Bettencourt, 2012). Thus, the primary aim of this research project is to address the spatial pattern of innovation within a city, taking the case of Amsterdam.

Cities are not equal when it comes to innovation and the divergence between them has been explained in terms of population factors. That is, innovation results from the positive externalities of the

clustering of the local population, thus, qualitative population differences in the properties necessary for innovation are likely to affect innovation rates. Three theories have gained ground in the literature as explanations for urban divergence in terms of economic growth, a growth caused by innovation. The value of these theories consists of their explanation for discrepancies in human terms, making them suitable for addressing the geography of innovation. However, these theories have likewise been developed at the regional level, with a lack of research addressing the city scale. This project thus aims to test these theories as predictors for the geography of innovation at the city scale.

The aim of this study is thus twofold: firstly, the spatial pattern of innovation at the city scale of Amsterdam is addressed, secondly, population factor theories are tested in terms of their strength in explaining this geographic pattern.

(12)

Following this introduction, Chapter 1 provides detail on the themes introduced here, firstly addressing why innovation is integral and how we can conceptualize it. The chapter then proceeds with an empirical discussion of the geography of innovation, and a review of the literature seeking to explain this geography. Three population factor theories are discussed: human capital theory, creative class theory, and social capital theory, in terms of the geography of innovation. Subsequently, the research limitations of these theories are addressed. Importantly, as innovation never occurs in a vacuum, the relationship between innovation and planning policy is discussed, to provide a contextual understanding of the geography of innovation. The research methodology and design of this study is detailed in chapter 2. The results are presented in chapter 3, commencing with the spatial pattern of innovation at a city-scale, followed by statistical analyses of the population factor theories discussed. The results are discussed in chapter 4, which likewise addresses the limitations of this study and suggests avenues for future research. Closing remarks are provided in the conclusion, which provides a summary of the findings of this study.

(13)

(14)

Chapter 1: Theoretical concepts and perspectives

A review of the why, what, where of innovation and the population factor theories that explain it

1.1 Innovation: economic driver and solution provider

The study of innovation arguably began with the theorizing of Schumpeter, an economist who elucidated the integral role of innovation in capitalism (Martin, 2012). He framed innovation as the driver, not only responsible for economic cycles (Archibugi, Filippetti, & Frenz, 2013a) but also the main source of competitive advantage (Fagerberg, 2009). That is, innovation is the process by which firms can gain a temporary competitive advantage over other firms and thus reap the benefits of market leadership (Fagerberg, 2009). It is the consequences of this competitive advantage that frame innovation as a force in economic growth (Fagerberg, 2009; Salter & Alexy, 2013).

Innovation contributes to a strong economy through job creation, subsequent income increases which drive economic growth, and innovative products which can increase quality of life (Hausman & Johnston, 2014). Though this has always been true of innovation, the changing structure of the global economy has made innovation even more important. That is, our economy is a “knowledge

economy”, where instead of physical resources, knowledge has become the most valuable resource (Malmberg & Maskell, 2002; Meusburger et al., 2013; Morgan, 1997). No longer is economic growth driven by mass-production sectors such as steel and petroleum, rather, today’s “new” economy is driven by sectors dependent on knowledge and learning: high-tech, financial and business services, and art/artisanal industries, where innovation is an integral feature to every lifecycle phase of the industry (Scott, 2006). Due to the restructuring of the economy towards the dominance of knowledge industries, there has been an impressive increase in economic interest related to innovation around questions such as where it happens, why, and how, in order to spur economic growth (Salter & Alexy, 2013).

However, innovation is not merely a source of competitive advantage. Innovation is the leading force behind technological and social change, driving long-term socio-technological transitions (Geels, 2004; Hall, 1998). To consider history, the process of innovation is integral to both world-changing technology, from the wheel to the computer, and social transitions, such as the enlightenment, or more recently, the rise of neoliberalism (Hall, 1998; Johnson, 2010). As such, it has attracted attention from academics seeking to understand radical changes in the status-quo. In particular, due to the social and environmental problems stemming from the pressures of climate change, many theorists have turned to innovation as a potential solution, allowing us to avoid crisis by innovating out of it (Grimm et al.,

(15)

2013; Rennings, 2000). That is, innovation is seen as a social process that may allow us to emerge from crisis, whether through new technology, new forms of organization, or new social norms (Geels, 2004; Sabadie, 2014).

Although the economic and transition perspective take different views on innovation, one more short term and zoomed in, one more long-term and zoomed out, both consider innovation as an important process. As such, both views encompass a need to understand the process of innovation. However, an understanding of spatial factors, and thus the social environment in which innovation occurs, has stemmed mainly from the consideration of innovation as an economic process (Coenen et al., 2012). That is, literature considering innovation as a form of socio-technological transition has rather focused on diffusion and upscaling processes, giving spatial factors a limited treatment (Coenen et al., 2012; Ravena, Schota, & Berkhoutb, 2012). Thus, our current understanding of the geography of innovation is limited due to a focus on innovation as an economic process.

To summarize, this section has highlighted the dual role of innovation in human society: as a market mechanism providing competitive advantage, and as the mechanism of socio-technological

transitions. However, in order to study innovation, it is integral to have a clear definition of what it is. This is the subject of the subsequent section.

1.2 Innovation: Defining concepts

Like all terms relating to abstract process, defining innovation is notoriously difficult and a source of continuing debate (Fagerberg, 2009; Salter & Alexy, 2013) . However, conceptual clarity is a fundamental of empirical research and thus this section aims to explain how innovation is conceptualized in this research project.

Innovation can be conceptualized as one of three interrelated, but independent, processes. Invention is the first of these processes and relates to the first appearance of a new idea (Fagerberg, 2009). As such, invention is concerned with the production of new knowledge. Entrepreneurship is the second of these processes and relates to creative economic activity utilized by individuals (and organizations) when starting new enterprises; turning knowledge into something with economic value (Acs & Audretsch, 2005; Fagerberg, 2009). Entrepreneurship is thus concerned with the creation of new organizations and enterprises. Innovation, as the third of these processes can be seen as the

combination of invention and entrepreneurship. Innovation is the process by which new knowledge is turned into an economically valuable resource; innovation is the application and commercialization of new knowledge (Fagerberg, 2009).

(16)

To illustrate this point, let us discuss the example of the steam engine. Though the invention of the steam engine is generally attributed to James Watt, he merely (though significantly) improved upon a less efficient, working steam engine produced by Thomas Newcomen more than 50 years prior (McNeil, 1990, p33). However, Thomas Newcomen likewise did not invent the steam engine, rather improving on a model by Thomas Savery. The crucial distinction between Savery’s and Newcomen’s engine was that Newcomen’s was efficient enough to be useful (McNeil, 1990). Thus, while Savery’s steam engine worked, it can be considered an invention – it has potential but isn’t yet commercialized. In contrast, Newcomen’s engine is an innovation – it is an invention with practical, commercial, application.

In terms of geography, invention is concerned with the location of knowledge-producing institutes, such as universities and research labs, while entrepreneurship is concerned with the location of new enterprise, and thus focuses on the start-up ecosystem of an area (Fagerberg, 2009; Glaeser,

Rosenthal, & Strange, 2010). Although the above example details individual innovators, innovation has been most thoroughly studied at the level of the firm. That is, the unit of analysis is not generally the individual innovator, but an innovating firm (Audretsch & Feldman, 2004). As such,

understanding the geography of innovation is based on observing where firms high in innovative activity are located; a geography which is discussed in the subsequent section.

1.3 The geography of innovation

The fundamental observation sparking interest in the geography of innovation is that innovative activity is not equally spread across the geographic landscape (Asheim & Gertler, 2009; Audretsch & Feldman, 2004). Instead, it clusters in specific areas, leading to a highly polarized pattern whereby some areas have high levels of innovative activity while others remain completely devoid of innovation (Florida, 2008; Johnson, 2010) (see figure 1).

Cities have long been framed as the hubs of innovation by both researchers and policy makers, as empirical observation suggests innovative activity tends to be located in urban rather than rural environments (Audretsch & Feldman, 2004; Glaeser & Gottlieb, 2009). In fact, work in complexity studies suggests a scaling law – the bigger the city, the more innovation produced (Bettencourt, Lobo, Helbing, Kühnert, & West, 2007). At a national level, it can thus be observed that cities tend to be the sources of innovative activity (i.e. firms in them are more productive in terms of innovation) while rural areas do not produce significant innovations (Florida, 2008). This observation has been termed an urbanization economy, whereby a firm gains by being in an urban agglomeration (Black, 2004;

(17)

Malmberg & Maskell, 2002). Thus, the argument in terms of innovation tends to be causal; cities aid the innovation process (Asheim & Gertler, 2009).

Figure 1: World map illustrating peaks of innovation in the form of patents granted1

Historically, certain cities have held innovative prominence, as world leaders in innovation at

particular times (Hall, 1998). In the industrial revolution, these were cities such as Manchester, while during the 15th century, Venice was the economic and thus also innovation hub (Hall, 1998).

Innovation, as arising from cities, can also be identified at the within-city scale. Specifically, the coffee houses of Paris and Vienna were seen as absolutely integral to the rise of local political thought and culture (Laurier, 2008), while areas such as Greenwich village, NY, were seen as centers of avant-garde movements (Johnson, 2010). As such, cities not only had a positive effect on the innovative activity of firms, they also inspired innovation in the cultural realm.

Moreover, cities are also influenced by the innovation that occurs in them. That is, innovation is an important source of city growth, which is related not only to innovations role in growing the

economy, but also in the pull of innovation-rich cities in terms of attracting population growth (Acs,

1_{Reproduced from Florida’s “Who’s your city? ” (Florida, 2008). Source:}

(18)

2002; Jacobs, 1969). In fact, this observation was key in Jane Jacob’s reasoning that cities exist to foster innovation, as will be discussed in the subsequent section (1.4).

The alternative to an urbanization economy is termed a localization economy. Localization economies are based on the clustering of similar enterprises, whereby such clustering increases the productivity and innovative activity of such firms due to the externalities present in the cluster (Malmberg & Maskell, 2002). The most prominent example of such an industrial cluster is Silicon Valley, which started as a hub of semi-conductor firms and is now the most developed high-tech cluster in the world (Salter & Alexy, 2013). However, Silicon Valley is not a unique example as industrial clusters are present all over the world: the film industry in Los Angeles, financial services in London, heavy industry in the Baden-Württemberg region of Germany and biotech in Boston (Asheim & Gertler, 2009; Salter & Alexy, 2013).

However, evidence for the positive externalities of industrial clusters is lacking, that is, it is unclear whether industrial clustering benefits innovation (Gordon & Mccann, 2000; Gordon & McCann, 2005; Malmberg & Maskell, 2002). Notably, the assumption of interfirm learning and communication that is fundamental to why industrial clusters exists has not been confirmed by empirical study (Gordon & McCann, 2005).

The difference between cities and industrial clusters can be conceptualized in terms of heterogeneity vs. specialization – cities provide a diverse milieu of various firms and inhabitants, while industrial clusters provide a specific milieu – a dense agglomeration of one type of industry (Glaeser &

Gottlieb, 2009). Research comparing the heterogeneity and specialization suggests that heterogeneity is integral to innovation: a diverse mix is better than a homogenous milieu when innovation is concerned (Acs, 2002). Thus, this study focuses on cities – as engines of innovation both in the economy and society in general. The subsequent sections aim to address the relevant theoretical developments aiming to explain why innovative activity clusters in cities.

1.4 Explaining the geography of innovation: A story of human externalities

The role of knowledge

Understanding why innovation arises from cities requires an understanding of what innovation as a process requires. Innovation involves the formulation and recombination of knowledge towards novelty which can be applied, commercialized and up-scaled through entrepreneurship (Salter & Alexy, 2013). Innovation is dependent on new knowledge that arises through invention, and

(19)

entrepreneurship – the economic creativity to apply that knowledge and commercialize it, thereby turning it into an economic good (Fagerberg, 2009).

Knowledge, as a resource, is quite unlike any other. Knowledge can be reproduced infinitely and does not deplete with use – it is non-rival, like a recipe, it can be consulted again and again, spreading further and further, without losing its potency (Audretsch & Feldman, 2004; Meusburger et al., 2013). The reusable nature of knowledge allows in to “spill over”. That is, knowledge can spread from agent (firm, organization, or even individual) to agent in geographic space. Cities foster knowledge

spillovers because they facilitate communication between diverse agents and the flow of agents between firms. That is, the clustering of people and a richer and more dynamic labor market present in cities aids knowledge spillovers (Meusburger et al., 2013). In fact, it can be observed that the more knowledge-dependent an industry, the more it will cluster (Audretsch & Feldman, 1996).

Despite a long academic tradition in studying inventors and innovators that would suggest that these individuals are almost super-human, possessing high intelligence and able to come up with

unpresented ideas (Akcigit, Grigsby, & Nicholas, 2017), a soberer evaluation of the history of innovation suggests innovators do not act alone. Innovations arise from a complex interaction between agents (Johnson, 2010; McNeil, 1990; Scott, 2006; Storper & Venables, 2004). A complex interaction that in turn leads to the new knowledge production and use that is integral for the innovation process.

Jane Jacobs not only highlighted the knowledge flow of cities in some of her seminal works, she saw knowledge flows as the primary reason for why cities exist in the first place (Jacobs, 1961a, 1969). Termed the “information city” thesis, she proposed that cities exist, and by extension benefit

innovation because they aid the flow of ideas (Audretsch & Feldman, 2004; Glaeser & Gottlieb, 2009; Glaeser & Saiz, 2003; Jacobs, 1969). Cities are clusters of people, diverse in their backgrounds, motivations, and activities, coming together to mutually create and extract from, a pool of local knowledge (Jacobs, 1961b). It is this pool of local knowledge, appropriate proximity for chance encounters and subsequent mixing of ideas that Jacobs saw as fundamental to city function and the reason for why innovation arises in cities.

Specifically, Jacob’s focus is on what she calls “diversity”, but may be better understood as

heterogeneity; the presence of various types of industry and thus knowledge (Jacobs, 1961a). In terms of knowledge spillovers, her thesis suggests that it is the presence of multiple such knowledge flows, each overlapping and interacting with the others, that leads to the “buzz” that is city life which benefits innovation (Bettencourt, 2012; Storper & Venables, 2004). Jacob’s idea contrasts with specialization, which we encountered at the end of the previous section. It is the presence of different

(20)

forms of knowledge, rather than clustering of one particular type of knowledge, that makes cities centers of innovative activity (Acs, 2002; Glaeser & Gottlieb, 2009; Meusburger et al., 2013)

However, the reliance of innovation on knowledge poses a conundrum. As previously stated knowledge is a non-rival resource and so it is unclear as to why it should respond to city boundaries (Audretsch & Feldman, 2004). In fact, why should any institutional or geographic boundaries influence knowledge spillovers and the ephemeral flow of ideas required for innovation? The answer lies with one integral distinction: not all knowledge is the same.

Fundamentally, knowledge can be distinguished into codified or tacit knowledge (Asheim & Gertler, 2009). Codified knowledge, or information, is easily transferred, as it can be explicitly formulated - a car production manual, or cake recipe (Meusburger et al., 2013, p5-6). Tacit knowledge on the other hand is difficult to verbalize and can only be learned by practice, the concept of “you just had to be there”- the specific way to tighten the screw in a car hood, or the particular way your grandma whisks the eggs (Asheim & Gertler, 2009; Meusburger et al., 2013).

Going back to the question of why cities are centers of innovative activity, it is now clear that tacit knowledge is the main culprit. Codified knowledge can flow around the globe through email or skype, but “know-how”: specific local customs, implicit ways of doing things, these are all extremely

geographically bound, or “sticky”, and passed on namely through face to face interaction (Asheim & Gertler, 2009; Audretsch & Feldman, 2004; Storper & Venables, 2004).

The implications of this point can be most clearly illustrated by the effect rising information and communication technologies (ICTs) have had on the clustering of innovative activity in cities. During the advent of ICT’s, it was predicted that improved communication technologies would make

geography obsolete (Asheim & Gertler, 2009; Morgan, 2004). Everyone everywhere would have access to knowledge and thus industries and people could innovate wherever, dispersing far more widely that was previously possible. Instead, the empirical reality suggests the rise of ICTs have exaggerated the clustering process, placing a premium on “being there”. The failure of ICTs to

ameliorate the spatial heterogeneity of innovation is directly related to the type of knowledge involved in the clustering of innovative activity (Morgan, 2004). Since everyone has improved access to codified knowledge, the source of competitive advantage, is tacit knowledge – knowledge that can only be gained through face-to-face interaction.

Yet, not all cities are equally innovative. Although size plays an important part, and larger cities are more innovative, there are cities that outperform, or do not meet, expectations – cities which are more or less innovative then their size would suggest (Bettencourt, Lobo, Strumsky, & West, 2010). There

(21)

must thus be a qualitative difference between the cities in terms of factors that are important for innovation.

Human externalities

Drawing on Jacob’s work, the economist Robert Lucas was perhaps the first to frame an explanation for city divergence in economic success and rates of innovation in human terms. He postulated that the tendency of people to cluster in cities generated externalities from human capital which could account for the economic growth and innovative activity observed in these cities (Axtell & Florida, 2006; Lucas, 1988). Human capital is a labor pool’s stock of knowledge, practices, norms, and characteristics that allow it to produce economic value, of which innovation is a significant part (Lucas, 1988). On an individual level, a person’s human capital is their education, training, skills and practices that allow him/her to produce economic value.

As such, Lucas was arguing cities differed because they differed in terms of the skill level embodied by the local population, and thus not only the ability of that population to innovate, but also the ability of firms to make use of the local knowledge (Gennaioli & Porta, 2013). Lucas’s model has been developed into endogenous growth theory which frames economic growth as a result of investments in human capital and innovation, rather than, for example, infrastructure (Gennaioli & Porta, 2013; Romer, 1994).

Research conducted by Glaeser and Saiz (2003) suggests skilled cities, cities that have high

percentages of skilled workers, tend to have higher amounts of innovative activity (Glaeser & Saiz, 2003). This suggest that externalities from the human capital stock of a city are positively associated with the amount of innovative activity observed in that city, providing support for a population-centered view on innovation. Moreover, Bettencourt et al (2007) find that some cities are more innovative because they have more innovators, not because the innovators that reside in them are more productive (Bettencourt et al., 2007). Likewise, Lobo and Strumsky (2008) suggest that it is the agglomeration of innovators in cities, not the interconnections between them, that drives innovation in cities (Lobo & Strumsky, 2008). These finding suggests the population composition of a city

influences innovative activity, and cities that happen to have or attract innovators will have higher amounts of innovation. That is, the innovative capacity of a city rests not with increases in

productivity, but in increases of a certain type of population that is innovative. The question thus remains what kind of population that is.

(22)

Of course, population factors are not the only determinants of innovation; the institutional and

infrastructural makeup of specific areas also plays a part (Acs, 2002). However, focusing on the social aspect of cities – their populations – allows us to consider innovation inclusively. That is, addressing populations allows us to get at the common denominator of all forms of innovation – the people that conduct it. Thus, three theories which most lend to an explanation for the geography of innovation are discussed: human capital, creative class, and social capital, which are discussed in the following section.

1.5 Population factor theories

Although the aim of population factor theories is to explain discrepancies in local economic growth; innovation’s role in the current knowledge economy makes it an integral mechanism for explaining this growth. As such, population factor theories can be adapted to provide a theoretical basis for the geography of innovation. That is, population factor theories are subsequently discussed because they are candidates for a population-level explanation for why innovation is where it is. To reiterate the argument presented in previous sections: a population level explanation for the geography of innovation is valuable as it provides a more inclusive perspective on innovation, addressing the common denominator of all innovation processes – people.

In this section, three theories are discussed. It is outside of the scope of this study to provide a thorough review of each, and as such, emphasis is rather directed at elements of the theory that explicitly address innovation. That is, the point of this section is to draw the spatial link between population factors and innovation. Limitations of the research are addressed after all three of the theories are discussed as in many instances the separate literatures share similar drawbacks in relation to the question at hand.

Human capital

Human capital is the skills and knowledge embodied within an individual that contribute to that individual’s ability to create economic value through his/her labor. Human capital theory, as developed by Edward Glaeser in the urban context, argues that educated people are the engine of urban economic growth (Gennaioli & Porta, 2013; Glaeser, 1994). Though several hypotheses about the mechanism through which human capital affects economic growth have been explored in the literature, the one which has gained the most ground is related to Jacob’s information city hypothesis. That is, human capital fosters economic growth because the clustering of skilled workers leads to higher rates of innovation. That is, we should “expect cities to be increasingly oriented around the

(23)

skilled because the skilled specialize in ideas” (Glaeser & Saiz, 2003), and flows of ideas lead to innovation, which results in economic growth. Thus, the clustering of skilled workers (i.e. areas with high human capital) generates positive externalities that result in innovation and subsequently economic growth.

In relation to the relationship between human capital and innovation, there is some evidence suggesting that educated individuals tend to be more innovative. A study by Akcigit et al (2016) found that historically, American innovators tend to be highly educated (Akcigit et al., 2017). Glaeser and Saiz (2003) likewise find some support for skilled workers being more innovative, as increases in educated individuals predict increases in innovative activity (Glaeser & Saiz, 2003). Taking a

different measure of innovation – new firm formation – Barreneche Garcia (2014) finds that human capital at a city level correlates with new firm formation (Barreneche García, 2014). As such, there is strong empirical support for the finding that human capital is positively related to innovation. In terms of geography, it is thus clear that areas – namely, cities, with high human capital likewise are highly innovative (Glaeser & Saiz, 2003; Hoyman & Faricy, 2009).

As human capital is a measure of the knowledge embodied in the workforce, we would expect it to foster innovation and economic growth through knowledge creation. That is, we would expect human capital to support the invention component of innovation. However, Glaeser and Saiz (2003) suggest human capital actually impacts economic growth through entrepreneurship – as they find no

relationship between patents and urban growth (i.e. innovation and economic growth) when controlling for human capital (Glaeser & Saiz, 2003). As such, they suggest human capital may impact economic growth by making people more creative (i.e. entrepreneurial) (Glaeser & Saiz, 2003). Gennaioli et al (2013) make a similar argument, suggesting that entrepreneurs utilizing their human capital may be the real factor in explaining economic growth (Gennaioli & Porta, 2013). That is, it is the presence of skilled entrepreneurs, not just skilled workers, that is a determinant for innovative activity and thus economic growth

Regardless of the specific mechanism through which human capital fosters innovation and economic growth, human capital theory explains divergence in city economies through the finiteness of human capital. There are only so many skilled workers, and thus the city in which they cluster will benefit the most, possessing higher amounts of innovation and economic growth.

(24)

Creative class

Despite the strong empirical support enjoyed by human capital theory, it is the “creative class” thesis that has risen to infamy (Peck, 2005). The creative class stresses that human capital, as measured by education, is not a measure of economically valuable knowledge; it doesn’t matter how much education an individual has if he/she is not applying it in an economically valuable way (Boschma & Fritsch, 2009; Florida, Mellander, & Stolarick, 2008). As such, proponents of the creative class thesis argue occupation is a better measure of economically valuable skill than education. Furthermore, due to the nature of the economy and the premium put on knowledge and innovation, the occupations suggested as the most integral for economic growth are the “creative” occupations, where people earn with their heads (Florida, 2002, 2008). These are occupations such as research, ICT development, consulting, and the creative industries such as media, advertising, and the arts (Florida, 2002).

The creative class thesis is a reformulation of human economic value based not on knowledge possession (as education or training) but economic creativity. It is thus a continuation of individual-level characteristics, which, when the people possessing them cluster, lead to positive externalities in the form of innovation and subsequent economic growth (Storper & Scott, 2009). It is important to stress that the creative class thesis is founded in the same academic origins as human capital –

inspired by Jacobs and stemming from the concepts of endogenous growth theory (Hoyman & Faricy, 2009).

Although the creative class thesis frames economic growth in terms of an interaction between technology (high-tech industry), talent (the creative class), and tolerance (a tolerant culture with low entry barriers) – attractively dubbed the three t’s, the thesis nonetheless argues that, to foster

innovation in a city, cities should aim to attract the creative class (Florida, 2002). As such, both human capital theory and the creative class thesis argue that labor is extremely mobile and thus the personal decisions of labor have a direct effect on local economic growth (Storper & Scott, 2009). As such, both theories suggest attracting economically desirable (either high in human capital or part of the creative class) individuals is necessary to foster innovation and economic growth (Storper & Venables, 2004). Thus, both theories hold that the presence of certain individuals and the subsequent interaction between them leads to innovation. The question is thus which indicator for human economic value is a better predictor of the geography of innovative activity.

There is in fact some support for the idea that the creative class is a better measure of human

economic value. As discussed in the human capital section, Gennaioli et al (2013) and Glaeser & Saiz (2003) both suggest that human capital affects local economic growth by having a positive effect on

(25)

economic creativity, or entrepreneurship – providing better input through higher knowledge content (Gennaioli & Porta, 2013; Glaeser & Saiz, 2003). As such, occupational measures may in fact be aiming to measure the combination of human capital and entrepreneurship.

Moreover, empirical studies also provide support for the association between creativity and

innovation. A study exploring firm innovation in different cities found firms in creative cities, cities with high numbers of creative industries (i.e. arts and media), were more innovative than those in less creative cities (N. Lee & Rodríguez-Pose, 2014). Additionally, a study comparing innovation (as new firm formation) and various measures of local social diversity found that the presence of creatives was positively associated with higher innovation rates (S. Y. Lee, Florida, & Acs, 2004). Although both these studies take a view of the creative class that is limited to creative industries, they nonetheless provide some support for the claim that creative individuals and industries are associated with innovative activity.

Florida (2008) argues that human capital and creative class both affect economic growth, but through different channels (Florida et al., 2008). The creative class has a positive impact on wages and labor productivity, while human capital has a positive impact on regional income and wealth (Florida et al., 2008). This again suggests that the creative capital converts the human capital into something that is economically useful, suggesting the creative class thesis alludes to the entrepreneurship component of innovation. These results stress the conceptual similarities and frame occupational and educational measures as alternative indicators for human economic value– either in the form of knowledge, or in the form of creativity.

A key feature of the creative class thesis is how dramatically it has been applied. In fact, it has led to the “creative city” boom, experienced in the early 2000s whereby cities put forward urban planning agendas that aimed to attract the creative class, in an attempt to foster economic growth (Evans, 2009; Geenhuizen & Nijkamp, 2012; Peck, 2005). Thus, the creative class thesis forms the backbone of many policy agendas trying to foster innovation in cities as will be discussed in section 2.7. In terms of the geography of innovation, the creative class thesis suggests innovation should occur in areas with high numbers of the creative class, a claim which has received mixed results in the literature at the regional level (i.e. across cities) (Florida, 2002, 2008; Hoyman & Faricy, 2009).

Social capital

Both human capital theory and the creative class thesis focus on individual level characteristics that, when agglomerated, result in emergent innovation and subsequent economic growth (Florida, 2002;

(26)

Glaeser & Gottlieb, 2009; Storper & Scott, 2009). However, if we consider that the interaction between individuals necessary for innovation is not random, then certain properties of such a network make innovation more or less likely. That is, there can be population level, rather than individual, factors that influence innovation (Dakhli & De Clercq, 2004). Social capital, comprised of networks and norms of trust and cooperation, speaks to the social environment of innovation (Patulny & Svendsen, 2007). As a feature of networks, social capital can be addressed at a variety of spatial scales – from the neighborhood or firm, to the national level (Lochner, Kawachi, & Kennedy, 1999).

Social capital is split into two forms in terms of how inward or outward looking the network is, these are termed bonding and bridging social capital (Patulny & Svendsen, 2007). Bonding capital consists of inward looking networks, whereby close and long lasting ties between individuals lead to group homogeneity and exclusiveness. In contrast, bridging social capital is present in outward looking networks whereby ephemeral and short term ties between individuals lead to groups that lack cohesion but have low entry barriers (Patulny & Svendsen, 2007). Robert Putman originally put forward the concept of social capital in a sort of nostalgic perspective on declining bonding social capital (Florida, 2002). However, it has been argued that bonding capital is actually detrimental to innovation as it stifles creativity and makes it difficult for new ideas to take hold (Florida, 2002). In contrast, it has been argued that bridging social capital, in the form of local tolerance and spurious interactions, may aid innovation as it is more supportive to new ideas and low entry barriers into the network (Florida, 2002; Hoyman & Faricy, 2009).

Alternatively, a study by Kaasa (2009) questions this binary distinction by looking at the effect of several dimensions of social capital on innovation (Kaasa, 2009). Out of general trust and networks, institutional trust, norms of helping, norms of active social participation, norms of orderliness, and civic participation, civic participation has the strongest positive relationship with innovation. General trust was the second strongest and several factors, such as norms of orderliness, actually had negative relationships with innovation (Kaasa, 2009). It is thus clear that social capital and innovation have a complex relationship, and it would be interesting to elucidate how this complexity is reflected in geographic space.

1.6 Research limitations of population factor theories

The main contribution presented by human capital, creative class, and social capital, is their identification of the human perspective. They have identified the importance of population composition and characteristics in innovation and local economic growth, thereby constructing a powerful lens through which to view the question of spatial heterogeneity. However, there are several

(27)

limitations to the theories, based on their central assumptions and key measures, that require more explicit research. It is the aim of this research project to address these gaps and thus these limitations are discussed in this section.

Perhaps the biggest limitation of the theories in question is their indirect treatment of the process of innovation. Because innovation is considered merely as an intermediate factor between population factors and economic growth, it lacks conceptual clarity. Firstly, innovation is framed only as a market mechanism, not as a feature of society. Secondly, the related but independent processes of invention, entrepreneurship, and innovation are largely conflated.

The theories consider innovation only as a market mechanism; a process stemming exclusively from firms. However, innovation can also arise from civil society, in the form of grassroots innovations (Pansera, 2013; G Seyfang & Smith, 2007). Currently, these are most often observed in civil action towards sustainability, the creation of urban gardens, neighborhood energy collectives or community currency (Gill Seyfang & Longhurst, 2013). That is, grassroots innovation may be more need oriented, labelled as “social innovation”, which aims to fill the gaps that the private sector does not (Moulaert, Martinelli, & Swyngedouw, 2005). Additionally, non-profit organization such as charitable foundations, which form a middle ground between grassroots and private companies in terms of profit vs. need orientation, also haven’t been geographically considered. Thus, the literature concerning the geography of innovation has not adequately dealt with innovation sources outside of the firm, limiting our understanding of innovation and its geography (Moulaert et al., 2005). Future research thus needs to elucidate whether alternative forms of innovation have the same geography as commercial innovation and whether this geography can be explained with the same theories.

In relation to invention, entrepreneurship, and innovation, there is a distinct lack in differentiating between these processes and establishing their unique roles. Blurring these concepts becomes problematic when measures of innovation are chosen. In fact, determining a valid measure for innovation has been the main limitation of innovation research (Acs, Anselin, & Varga, 2002).

Patent counts are the most prevalent measure used in the literature discussed above. However, a granted patent does not give any indication of whether the idea in question has been commercialized (Acs et al., 2002). As such, metropolitan patent counts may, at best, be a representation of invention rather than innovation. On the other hand, several studies use new firm formation as a proxy for innovation. However, new firm formation is a measure of entrepreneurship; there is no distinction between new firms which utilize new knowledge and those that are merely new firms. In order to measure innovation, it is integral to measure both whether the idea in question is novel and whether it has been commercialized, at least to some extent.

(28)

Furthermore, spatial scale is also a limitation of the research on population factor theories. Most of the research considers the regional scale, engaging in comparative work on cities. However, if innovation is framed as a social process dependent on face to face interaction, and cities are the engines of such innovation – why don’t we look below the hood? Cities are presented as homogenous entities when it comes to innovation, however, the few studies that have looked at the within-city scale suggest human capital and creativity is not equally distributed throughout the studied cities (Brennan-Horley & Gibson, 2009; Hansen & Winther, 2010). Of course, this is obvious to anyone that lives in a city. However, if human capital is not spread equally in cities, what does this mean for the geography of innovation?

Furthermore, understanding the city-scale of innovation could in fact provide support to some of the claims presented in the theories. Specifically, if the idea is that face to face interaction between skilled individuals leads to innovation, we certainly would not expect innovation to arise in areas that have no skilled population living or working there. That is, the finer spatial scale provided by a within-city perspective can give us a better idea of the causality involved in population factor theories – do certain populations cause (i.e. conduct) innovation?

Moreover, innovation policy tends to have direct effects at the city level, even if formed at national or regional levels, with strategies such as the “creative city” assuming dominance (Evans, 2009; Peck, 2005; Tödtling & Trippl, 2005). In the absence of city-level perspectives, this policy is based on research that either compares cities or is conducted at a regional level (e.g. Florida, 2002). A within-city spatial scale could provide a better test for the hypotheses of population factor theories and a more relevant research base for policy formulation.

To summarize, in terms of explaining the geography of innovation, the three population-factors are limited due to the broad spatial scale addressed, un-inclusive formulation of innovation, and poorly conceptualized measures of innovation. However, despite these limitations, and most importantly, despite a lack of urban perspective on the effect of these population factors on innovation, the theories have been applied with gusto to urban planning policies aiming to foster innovation and economic growth (Hoyman & Faricy, 2009). The reasons for this speedy application and an explicit

consideration of the urban context is explored in the subsequent section.

1.7 Planning to innovate - Why is innovation an urban planning issue?

So far, this thesis has remained relatively abstract, treating innovation as place-independent, reliant on factors that can be anywhere and everywhere. However, this is not the reality; the process of

(29)

innovation is always deeply embedded in place, and actively sought to be influenced by policy (Geenhuizen & Nijkamp, 2012; Meusburger et al., 2013). Although evaluating innovation policy seldom has a clear spatial element, it is clear that policy can have an effect on the geography of innovation through the creation of industrial clusters, science parks, or creative districts (Evans, 2009; Geenhuizen & Nijkamp, 2012; Meusburger et al., 2013). Thus, understanding the geography of innovation cannot take place in a policy vacuum. This section aims to address the wider neoliberal turn in urban policy that encapsulates why cities have sought to foster innovation to provide a more localized context for understanding the forces acting upon the geography of innovation.

To recapitulate, innovation plays an important role in localized economic growth. In fact, innovation holds a premium position in the knowledge economy (Meusburger et al., 2013). As such, it is too valuable to merely leave to develop organically, and governments ranging from the national to the municipal have in the past few decades developed policies to try to foster innovation (Tödtling & Trippl, 2005)2_{, notwithstanding the theoretical limitations and, in many cases, lacking empirical}

evidence (Gordon & McCann, 2005; Meusburger et al., 2013). Since cities are seen, both historically and theoretically, as the crucibles of innovation, much of the policy focuses on the urban level (Tödtling & Trippl, 2005). That is, though policy aiming to foster innovation exists at various spatial scales, its focus tends to be on within-city elements, with different governance levels organizing different aspects: for example, the regional level may deal with clusters, while knowledge production through universities is organized at the national level (Tödtling & Trippl, 2005).

It is integral to clarify that cities do not develop policies solely focused on innovation. As reflected by the theories discussed above, innovation is mainly seen as an intermediate to the real goal of

economic growth (Brenner & Theodore, 2002; Harvey, 1989; Tödtling & Trippl, 2005). As such, interventions specifically aimed at innovation are absorbed into a wider policy agenda aiming to improve the economic competitiveness of a metropolitan area (Geenhuizen & Nijkamp, 2012).

The advent of the knowledge economy, and the subsequent increasing importance of urban areas to the economic wellbeing of regions and countries, has combined with the process of neoliberalisation to create an elixir of entrepreneurialism – an entrepreneurial way of governing urban development (Harvey, 1989). Neoliberalism, an aggressively free-market ideology whereby the private sector has more power and is free from government regulation, has led to urban austerity – city budgets have

2_{In fact, in Europe, innovation fostering policy exists at the level of the European Union as part of the Europe}

2020 program which aims to support the economic competitiveness of the EU region: http://ec.europa.eu/research/innovation-union/index_en.cfm

(30)

become increasingly constrained (Brenner & Theodore, 2002; Peck, 2012a). As such, cities have been pushed to operate more like firms, strategically planning economic growth in order to flourish

(Harvey, 1989).

Neoliberal urban policies are explicitly market oriented; their main agenda is to foster economic growth and competitiveness (Brenner & Theodore, 2002; Harvey, 1989). Here the link with

innovation becomes clear, innovation is the intermediate to economic growth and thus naturally fits into the neoliberal agenda. In fact, it is interesting to note that innovation as a process is affected by neoliberalisation: while innovation in more traditional coordinated markets focus more on long-term, coordinated innovation, liberal markets are characterized by high rates of change, which give agents the opportunity to adjust institutional structures dynamically (Asheim & Gertler, 2009). As such, neoliberalisation prioritizes and accelerates innovation.

In relation to the theories discussed in this research project, one in particular has achieved dominance in the policy sphere. The creative class thesis has proved to be extremely attractive to policy makers (Evans, 2009; Hoyman & Faricy, 2009; Peck, 2005). One postulation as to why is that, in contrast to human and social capital, the creative class thesis has always been presented in a neatly packaged array of recommendations (Peck, 2005). It was only too easy for policy makers to take these policy recommendations and indiscriminately apply them to a variety of cities: Barcelona, Berlin, Sheffield, the list goes on (Evans, 2009). In what critics call the “creativity craze”, the “creative city” agenda become a popular neoliberal urban development model from the turn of the millennium onwards (Peck, 2005).

To recapitulate the creative thesis recommendations: the theory rests on the idea that the creative class is the engine of economic growth – responsible for innovation and productivity growth. As such, cities should attract the creative class to foster economic growth. The assumption is that the creative class is hyper-mobile and does not migrate only based on employment opportunities, but rather makes more nuanced migration decisions based on quality of place – local culture, customs, and most importantly, amenities (Florida, 2002). Amenities in this case encapsulate a vibrant local consumer culture – hip café’s, a bustling music scene, a lively shopping district, authentic street culture, you name it (Peck, 2005). As such, localized policies mainly took the form of funding for cultural activity – whether grassroots or otherwise, and rejuvenating certain neighborhoods that were seen as

potentially attractive for the creative class (Evans, 2009).

More than a decade has passed since the emergence of the creative class thesis and the craze has all but ended. Firstly, the gentrifying and neoliberal nature of the creative city agenda became associated with urban segregation and polarization. That is, it has become clear that cities with high numbers of

(31)

the creative class likewise experience widespread inequality, with creative city policies furthering the gap (Donegan & Lowe, 2008; Peck, 2005). Moreover, the creative class thesis was widely criticized for its glorification of the creative class, many of who actually work precarious jobs without economic security (Pratt, 2011). Perhaps more abstractly, the thesis and corresponding development agenda were criticized for overt commercialization of local culture – turning local authenticity into part of the city brand, and thus into an economic good (Evans, 2009; Pratt, 2011).

Most importantly however, cities suffered disillusionment because the policies were not the quick fixes they anticipated (Pratt, 2011). Not every “creative city” became the next San Francisco.

Recently, cities have sought to establish more nuanced creative city agendas that take into account the problematic nature of gentrification; a creative city with social consciousness (Pratt, 2011).

Alternatively, the creative city has been replaced by other, similarly neoliberal, agendas – the smart city, for example (Söderström, Paasche, & Klauser, 2014). However, these are variations of a similar theme in which cities act in order to foster economic growth, the original “entrepreneurial city” described by Harvey (1989) in his seminal work outlining the new, market oriented structure of urban development (Harvey, 1989).

Regardless of the disillusionment of the creative city agenda, innovation remains an important goal for cities to pursue.. As innovation is not only is it integral for the current economy, it is also capable of providing solutions and societal transitions it is paramount to deliver policy that does foster all sorts of innovative activity. In order to deliver policy which can take an inclusive look at innovation, understanding the geography of innovation provides an important starting point. However, to the best of my knowledge, policy evaluations are scant, and thus it is hard to establish causality between urban policy and the subsequent geography of innovative activity (Evans, 2009). Nonetheless, the policy context must be accounted for when addressing the geography of innovation.

Chapter 1: Conclusion

To conclude, this chapter has discussed the theoretical underpinnings of this research project.

Emphasis has been placed on the city-scale, the value of population factor explanations and the wider policy context in which innovation occurs. Namely, this chapter extends the argument that a city-scale perspective on the geography of innovation is necessary both theoretically, as a better test and

extension of existing theories and practically, to provide a better theoretical basis for innovation-fostering policy. The subsequent chapter discusses the research methodology employed in this study.

(32)

Chapter 2: Research design and methodology

Research design considerations, methodological procedures relating to data collection and processing as well as data analyses

2.1 Research Design

Problem statement and aim

Although theories focusing on population factors have proved instrumental in explaining regional and cross-city differences in innovation and economic growth, they have not directly addressed the geography of innovation in relation to these factors. Without an understanding of where innovation happens at a city scale, it is impossible to know how it is related to human capital, the creative class, or social capital, as cities are not homogenous bodies in which people and economic activity are evenly distributed. Moreover, as policy is applied at the city-scale, the theoretical base needs to adopt a scale that is relevant for policy formulation.

Likewise, the market oriented conceptualization of innovation present in both the theory and in the policy arena means that innovation outside of the market, originating from grassroots movements for example, is not considered or understood in terms of geography (Coenen et al., 2012). As such innovation may be important for solutions to current problems – such as a transition to sustainability, it requires consideration and an integration into theories of the geography of innovation. The point of this research project is thus to address some of the theoretical gaps in the literature by providing a within-city perspective on the geography of innovation in relation to the geography of human capital, creative class, and social capital. Moreover, this research project aims to provide a more inclusive view of innovation considering innovation both from private firms, grassroots organizations, and foundations.

Research question

To address the issues discussed in the previous section and problem statement, it is necessary to address the spatial geography of innovation within a city. As such, the research question of this study is:

(33)

A quantitative research design is necessary to address this research question because it allows for a rigorous test of human factor theories which can subsequently be generalized beyond the test population. In relation to the specifics of the research question: the selection of Amsterdam as a case is reviewed below. The extent of human capital, creative class, and social capital, is addressed as these variables will never be the whole story. Innovation is affected not just by human factors, but also by institutional and infrastructural factors (Acs, 2002), which are not addressed in this research project. Thus, this project addresses the strength of population factors as predictors of innovation. Moreover, as three theories are compared, it is important to assess their individual strengths as predictors of innovation.

To address the research question, it is first necessary to describe and understand the spatial pattern of innovation in Amsterdam. Secondly, to establish the role of population factors in predicting

innovation, these theories must be tested in relation to the observed spatial pattern. Thus, the research question can be separated into two sub-questions:

Question 1: What is the spatial geography of innovation in Amsterdam?

Question 2: To what extent can human capital, creative class, and social capital predict the geography of innovation in Amsterdam?

Case under study

The Dutch Capital, Amsterdam, has been selected as the case of this research project. Amsterdam has been selected due to its long history as an innovative city and due to the emphasis placed on

innovation in the city in the past two decades. In particular, Amsterdam’s pursuit of the creative city agenda illustrates its economic growth agenda and thus implicit support for innovation. Though municipal support for innovative activity is not a precondition for this study, it makes it easier to construct the large database of innovative activity necessary for this study. Moreover, the creative city pursuit has led to a rich literature and research on Amsterdam which provides a detailed context in which to understand the results of this study.

To look at a within-city scale, it is important to study a city which has enough innovation present to provide a large enough sample. In this sense, Amsterdam is an appropriate case. Moreover, as this research project aims to address not only commercial innovation, a city with an active civil society is required. Amsterdam won the 2016 iCapital – European capital of innovation – award for its high

(34)

levels of “social innovation”3_{. Furthermore, Amsterdam is known as a hotbed of civil activity and}

since the 2008 financial crisis, Amsterdam has seen a boom in bottom-up initiatives in diverse fields (Nijman, 1999; Savini & Dembski, 2016). As such, Amsterdam provides an interesting environment in which to observe innovation stemming from different sources.

Finally, Amsterdam has high-quality and reliable statistical data on population composition and characteristics available at neighborhood level, allowing for a spatially detailed analysis of the

geography of innovation in relation to these factors. However, as innovation is by no means merely an organic activity occurring in cities, but rather actively encouraged, the urban planning policies

pursued by Amsterdam require discussion. As such, a review of secondary and primary material on urban policies relating to innovation is conducted to provide the policy context for the geography of innovation addressed in this study. The results of this review are presented in section 3.1, prior to the empirical results of this study.

Research phases

Addressing the research question of this project requires a four-phase approach. Phase one consists of data collection, whereby data on the dependent variable (innovation) and the relevant statistical data are obtained. Moreover, desktop research is conducted on primary and secondary sources relating to policies employed by Amsterdam in relation to innovation. This research aims to provide an

understanding of the policy context present in Amsterdam.

Phase two consists of mapping innovation locations and conducting a cluster analysis to describe the geography of innovation in Amsterdam, identifying whether innovation clusters in particular areas of the city. Phase two utilizes a quantitative descriptive research methodology. This research is inductive and does not depend on hypothesis testing.

Phase three involves statistically testing the relationship between the location of innovation and human capital, creative class, and social capital. That is, phase 3 aims to address whether high

instances of innovation are related to high levels of human capital, creative class, and/or social capital, and if so, to what extent. To this end, a quantitative comparative research design is utilized, consisting namely of multiple linear regression analyses. This research design is deductive and explicitly tests hypotheses derived from the literature.

3

(35)

Finally, phase four takes the form of exploratory post hoc research. As part of this phase, a longitudinal perspective on the geography of innovation is explored by way of a longitudinal

visualization of the data. Additionally, an interview with two municipality employees is conducted to provide an Amsterdam-specific understanding of the results.

Hypotheses

Phase one, two, and four of this research project consist of inductive research and thus do not consist of hypothesis testing. Phase three of this research project involves statistically testing the relationship between innovation and population factors and thus depends on the formulation of testable

hypotheses.

Drawing on the literature discussed in the previous chapter, there is evidence that both human capital, and creative class, are important determinants of innovative activity. As such, this research

hypothesized that high innovative activity would be observed in areas with high amounts of human capital and creative class.

Hypothesis 1: Neighbourhoods with higher amounts of human capital will have higher amounts of innovative activity

Hypothesis 2: Neighbourhood with higher amounts of creative class will have higher amounts of innovative activity

In terms of social capital, the theory suggests that bridging social capital supports innovation while bonding social capital is detrimental to innovation. As such, this research hypothesized that high innovative activity would be observed in areas with high bridging capital but low bonding capital

Hypothesis 3: Neighbourhood with higher amounts of bridging social capital will have higher amounts of innovative activity

Hypothesis 4: Neighbourhood with higher amounts of bonding social capital will have lower amounts of innovative activity

Clusters and co-location : the spatial geography of innovation in Amsterdam