Planting the seed to grow local creative industries: The impacts of cultural districts and arts schools on economic development

Abstract

The impact of arts and culture on local economies has been studied extensively. However, a review of the literature finds conflicting and critical results regarding the impact of cultural on economic outcomes. In this paper, we shift attention to examine different intermediaries and concentrations of cultural agents that can influence growth and innovation in the “creative economy.” Thus, we build on previous work and expand on it by refining the scale of analysis (zip-code level). The paper focuses on education in the arts and digital media in all arts-related programs at universities as well as accredited art schools across the United States. Further, employing more observations for larger cities allows a richer depiction of the rather urban nature of the arts and digital media industries. We find that, by going to the zip-code level, we can say that both districts and arts programs (especially at schools that specialize in arts education) have a positive relationship with the share of jobs in the arts and digital media. Moreover, when we evaluate the impact of schools versus districts, we find that schools have a greater role.

Keywords

Cultural districts digital media economic impact employment intermediaries

Introduction

The impact of arts and culture on local economies has been studied extensively (Markusen and Gadwa, 2010b; Rantisi, 2004; Rushton, 2013; Scott, 2001, 2006). However, a review of the literature finds conflicting and critical results regarding the impact of cultural on economic outcomes. In this paper, we shift attention to examine different intermediaries and concentrations of cultural agents that can influence growth and innovation in the “creative economy” (Jakob and van Heur, 2015). We focus on the role of universities and arts districts¹ in the arts and digital media² as intermediaries that contribute to the employment and industrial base of occupations in the arts and digital media.

This paper makes two scholarly contributions: First, we expand the work on education intermediaries in cultural geography, and, second, we add the unique component of art districts.¹ Continuing the work of Breznitz and Noonan (2014) as well as Rantisi and Leslie (2015), we focus on universities with arts and digital media programs (UADMs) and schools specializing in the arts and digital media (SSADMs) as an important driver of firms’ location. We use a new database of US cultural districts to control for concentrations of arts and digital media jobs outside university campuses and to observe how specialized arts and digital media education that is spatially concentrated fosters job growth independently and through those cultural districts.

Why focus on intermediaries? The study of intermediary organizations is not new. It has been conducted in sociology, innovation, management, and cultural economics (Howells, 2006). The literature on cultural intermediaries focuses on actors working in connected production or consumption (Jakob and van Heur, 2015; Negus, 2002). Specifically, the literature recognizes the importance of these institutions in bridging the creator and the market. Like intermediaries in other fields, intermediaries in culture play roles as both gatekeepers and mediators.

These discourses work through a range of institutional apparatuses to assemble people and things within specific relations and technologies framed and mutually understood as “economic.” By implication this process of “assemblage” requires intermediaries who work to connect these institutional discourses with the interests and contexts of the specific actors whose behaviours are to be modified. (O’Connor, 2015)

Hence, they are very active agents who pick “winners” as well as provide support from the initiation of the idea to the creation of the final product (Howells, 2006; Nixson and Du Gay, 2002; Power and Scott, 2004).

The paper attempts to explain the role of educational institutions in employment in the arts and digital media, and hence focuses on the question: in what way do arts-related intermediaries (art schools and cultural districts) contribute to regional economic development? We start by examining the different employment trends associated with the presence of cultural districts and the number of SSADMs and UADMs. We examine this relationship for a variety of arts-related jobs: arts, cultural, and digital media jobs. In addition, we evaluate the connection to high-tech jobs in order to capture the more technological aspect of digital media. To identify the role of districts and schools, both separately and together, we follow by evaluating the geographic proximity of employment to schools and districts. We evaluate the relationship at the 5-mile as well as the 50-mile radius. The 5-mile radius analysis is used to evaluate the clustering effects of districts and schools. The 50-mile radius digs deeper to explore commutability (e.g., living downtown and working in the suburbs and vice versa). When educational institutions or cultural districts play intermediary roles as gatekeepers and mediators, we expect them to influence where digital-media–related jobs locate.

In order to better understand the role of arts-related educational institutions and cultural districts as important intermediaries in the local economy, we build on previous work and expand on it by refining the scale of analysis (zip-code level). The paper focuses on education in the arts and digital media in all arts-related programs at universities as well as accredited art schools across the United States. Prior studies at the scale of a city or county overlook the critical within-metro variation and local hotspots that are detectable at a finer resolution (e.g., Grodach et al., 2014). Further, employing more observations for larger cities allows a richer depiction of the rather urban nature of the arts and digital media industries. We find that, by going to the zip-code level, we can say that both districts and arts programs (especially at schools that specialize in arts education) have a positive relationship with the share of jobs in the arts and digital media. Moreover, when we evaluate the impact of schools versus districts, we find that schools have a greater role.

The analysis overlays three sorts of national data for the United States to map a spatially detailed landscape of employment, higher education resources, and formal clusters of cultural activity. First, establishment and employment data from the US Census Bureau’s Zip Code Business Patterns (ZBP) resource provide highly disaggregated measures of relevant jobs. The ZBP enables the construction of annual job counts by narrow industrial categories at the zip-code level for the entire country. Second, a compilation of institutions of higher education with accredited arts-related programs or with digital media degree programs provides geographic locations for the concentration of institutions of higher education. Third, a novel inventory of cultural districts around the country is leveraged for spatially detailed information on other local concentrations of cultural economic activity. The particular notion of a cultural district is defined and discussed at length below.

The paper continues with a review of the literature about the importance of art and culture industries in economic development with a particular focus on intermediaries. We discuss art and culture in the field of economic geography, as well as the role of institutions of higher education as intermediaries and as important contributors to local economic development. We follow with sections on research design and method, model results, discussion, and conclusion.

Economic geography, culture, and intermediaries

Culture in general and the arts in particular have been viewed increasingly as tools for generating economic growth through the creation of cultural districts and institutions as well for economic development through the attraction of high-quality labor and large corporations (Rushton, 2013).

The importance of arts and cultural industries to economic development in a city region can be viewed in different studies. For example, in her study of Silicon Valley versus Route 128 outside Boston, Saxenian (1994) sees the communication culture within the region as a unique factor in the ability of these regions to prosper. Scott (2006) shows that the success of specific regions is based on their ability to extrapolate arts-related goods and services and that the development of cultural products requires constant development of new ideas. In addition, studies show that the arts and artists attract firms and high-level human capital to a region, which contributes to the productivity of other industries (Florida, 2002). Culture and art are contributing factors to the environment in which firms and individuals choose to live and work (Eaton and Bailyn, 1999). Physical investment in the arts helps revitalize neighborhoods or districts (Ladry et al., 1996). Studies claim that some cities or regions have added jobs more rapidly over time as a result of cultural industries (Tomusk, 2011). These additional positions are attributed to a large concentration of cultural industries, which have been viewed as important to having a diverse economic base in highly specialized cities (Tomusk, 2011).

Moreover, studies in economic geography link work on culture with industrial clusters. Accordingly, firms concentrate in specific locales to take advantage of specialized resources, including labor (Rantisi, 2004; Scott, 2000; Storper, 1997). Studies on the geography of innovation show the localization of patenting and the importance of knowledge sharing and knowledge creation between and within different actors in certain regions (Feldman, 1994; Jaffe et al., 1993). In both Jaffe and Feldman’s work, the role of universities as a source and anchor of knowledge is highlighted. In addition, Sedita (2008) sees the arts and cultural district as the ideal locus for developing relationships and networks. These studies demonstrate that proximity provides access to trade networks and information (Grodach, 2011; Markusen and Schrock, 2006).

It is important to note that studies have conflicting results regarding the possible causality between cultural activity and the attraction of other firms and high-income workers (Markusen and Gadwa, 2010a). Some studies show inverse relationships between cultural clusters and economic growth (Brooks and Kushner, 2001) and neighborhood dynamics (Noonan, 2013). Taylor (2015) discusses the fuzziness of the creative economy concept, claiming that its ability to explain the phenomenon is weak. The author holds that, while an attempt to explain the notion of the creative economy is made in a phrase such as the “creative industries,” no consensus has been reached on a definition of what the creative industries include, leaving much room for speculation.

Studies observing the importance of culture on economic development highlight the existence of regional intermediaries (Jakob and van Heur, 2015; Negus, 2002; Taylor, 2015). In particular, studies view intermediaries as important contributors to social capital, network development, and knowledge transfer (Vinodrai, 2015). Scott explains the importance of intermediaries by saying that knowledge is not evenly distributed, and it tends to concentrate in specific places among individuals and within firms and specialized institutions, such as universities:

Such sites constitute the atoms and neurons of the creative field, so to speak, but their power to generate new knowledge is magnified many times over when they come into definite interrelationship with one another. (Scott, 2006: 8)

Existing studies show that employees and arts-related industries locate in proximity to specific institutions—that is, intermediaries (Currid and Williams, 2010; Lloyd, 2010). Negus (2002) highlights the role of workers in the creative industries and their ability to change production and its symbolism. Intermediaries are able to create “atmospheres and environments conducive to new rounds of creative production and consumption” (Taylor, 2015: 367). Hence, intermediaries in cultural industries create certain “scenes” that allow the different players in these industries to meet (O’Connor, 2015). This is not much different from what Marshall (1890) identifies about industrial districts as a specific culture that is “in the air.” O’Connor (2015) adds that the “new cultural intermediaries” refuse to see art as separate from life. They provide a more modern interpretation of art in real-life products. In a way, one can see this in the ability of someone such as Steve Jobs to use the calligraphy he loved in Apple’s Macintosh computer. Chapple et al. (2010) find that “strong leadership and action within informal networks catalyzed action” (p. 233) in developing arts districts. These networks, and the role of planners and city leaders in coordinating and supporting them, are an important theme in many arts districts, such as Baltimore’s Arts and Entertainment District (Ponzini and Rossi, 2010).

In this paper, we follow the call of Jakob and van Heur (2015) to investigate “sector-specific dynamics,” with a specific focus on two types of cultural agents: cultural districts and universities with arts education programs. Prior research on cultural intermediaries does not typically examine institutions of higher education or arts districts. Especially in the creative economy, intermediaries are often limited to individuals or organizations, such as consultants, music industry managers, or designers (Taylor, 2015). Although empirical work links the density of cultural activity to economic development (Chapple et al., 2010; Silver, 2012), the institutional context and connections in that sector are often neglected. Many economic development studies examining the cultural sector overlook the role of postsecondary educational institutions, even though universities can both anchor the local cultural sector and drive stronger economic growth. In their paper, Rantisi and Leslie (2015) explain the importance of intermediaries in cultural industries and the move from specific production and consumption intermediaries to more day-to-day intermediaries. The authors discuss the lack of study of institutions of higher education in general and institutes of applied arts and culture as important intermediaries in cultural economics. For the knowledge economy in particular, higher education has been viewed as a new source of knowledge, basic research, and entrepreneurship for today’s globalized economy (Asheim and Gertler, 2006). However, higher education’s role in spawning innovations and economic impacts in arts-related fields has not received much attention until now. Breznitz and Noonan (2014) analyze the impact of arts districts and research-intensive universities on innovation and employment. In their work, the authors focus on research-intensive universities and find that arts districts have a stronger effect than research-intensive universities on employment related to the arts and digital media and innovation in major US cities. Research-intensive universities in urban areas have few effects on patenting and employment related to the arts and digital media. The positive impact of research-intensive universities is evident only in employment growth in the arts, if at all, and not in digital media.

However, arts and culture are not taught only at universities. A multitude of institutions of arts and culture offers educational programs in these fields. For example, Rantisi and Leslie’s (2015) study of the National Circus School (NCS, Ecole Nationale de Cirque) in Montréal and the circus industry in the city is a new and important contribution to this wave of studies. In their study, the authors analyze the importance of the NCS to the development of a new contemporary art form, the relationship of artists with local industry (the network), and policy. The study emphasizes the importance of higher education institutions in general and for cultural industries in particular, especially those with the ability to support the transfer of tacit knowledge. Currid (2007) finds similarly important roles for cultural institutions and universities in providing skills and networks crucial to artists. Asheim and Gertler (2006) identify the interactions among firms, research organizations (including universities), and public agencies as the base on which innovation develops. In their work, they specifically identify the importance of work experience, which is achieved through “learning by doing” and interacting. This highlights the importance of professional schools.

In addition, a special issue of Regional Studies on intermediaries and the organization of the creative economy (Jakob and van Heur, 2015) calls for the study of spaces in which cultural producers can be celebrated and on a range of different levels meet their publics. We believe that such spaces can be identified within arts and cultural districts. As explained in the following section, cultural districts are policy-driven areas that allow for certain arts and cultural industries to receive public support. Many claims have been made about the benefits of arts development in general and cultural districts in particular. Frost-Kumpf (1998) lists several of these benefits, such as increases in employment, population, property values, and education levels. Markusen and Gadwa (2010b) cite the positive impacts of creative places on jobs and income. Mommaas (2004) argues that public policies (including those related to the creation of formal districts) can create conditions favorable for the development of arts and cultural industry agglomerations:

We see the rise of a common place-based cultural development strategy, linking cultural activities and amenities to economic, spatial, and social policy goals. (Mommaas, 2004: 514)

In particular, the author identifies five justifications for policy to cluster creative activity, of which, three are related to increased economic activity. In their, 2014 work, Breznitz and Noonan find that cultural districts have a consistently positive effect on local digital media economic activity—employment and innovation. Hence, similar to art-related educational programs, cultural districts are an important intermediary to study.

Research design and method

The empirical analysis investigates local patterns of job growth in certain industrial sectors across the United States. The structure of the empirical models builds on the Breznitz and Noonan (2014) approach. Local zip-code-level jobs are explained by their proximity to UADMs, arts districts, and their interaction. To control for the propensity of different sectors of jobs to locate closer to urban cores, the full national sample models include indicators for zip code locations in metropolitan statistical areas (MSAs), suburbs, or core cities. The models also control for the starting point for employment in case of a tendency toward convergence in zip-code-level employment or high-density employment clusters to gain even more jobs. The model seeks to explain trends in employment, rather than just levels of employment (which might be driven more by other local, permanent unobserved factors), as in Grodach et al. (2014). The basic model takes the linear form:

\begin{array}{l} Δ J o b S h a r e = β_{0} + β_{1} J o b S h a r e_{98} + β_{2} D I S T R I C T \\ + β_{3} S C H O O L S + β_{4} D I S T * S C H + β_{5} U R B A N I C I T Y + ε \end{array}

JobShare refers to the share of all zip-code jobs in the jobs group (e.g., arts, digital media) being examined. Thus, ΔJobShare = JobShare₁₃ − JobShare₉₈ reflects growth in the share of jobs, in percentage points, over the period, 1998–2013.³ The DISTRICT vector contains two variables, the number of cultural districts in the zip code (District Count), and the number of districts within 5 miles of the zip code (District Count5). Similarly, the SCHOOLS vector shows the number of arts schools in the zip code (ArtSchoolCount) and within 5 miles (ArtSchoolCount5). The DIST*SCH interaction term is a dummy variable that has a value of 1 if the zip code contains both a district and an arts school (i.e., District Count*ArtSchoolCount > 0). The URBANICITY vector contains dummy variables for zip codes located in an MSA, in a suburb, or in a core city of a metro area. For the “city” subsample analyses (see Appendix 2), the URBANICITY vector is omitted.

Although this basic empirical model builds directly on Breznitz and Noonan (2014), this analysis offers many important improvements over that earlier effort. First, the present analysis employs data that are far more spatially disaggregated (zip codes vs. cities). Second, it examines the locations of work, rather than the locations of residence, thus better measuring spatial clusters in creative production, rather than the locations of employees’ residence. Third, unlike the previous study, which had only about 100 US cities, this analysis has all US cities and expands the scope to include the entire country, rather than just major cities. Fourth, many more postsecondary schools—and an indicator of whether they are schools that specialize in the arts—are part of this analysis than the previous study, which used only research-intensive universities. Fifth, this analysis makes use of an expanded set of arts and cultural districts—143 in all, compared with just 41 in Breznitz and Noonan (2014). All these differences add up to a substantial enrichment of the analysis: far more disaggregated data, with better measures, an expanded sample, and greater coverage of schools and districts. At the same time, this analysis omits the parallel inquiry into patenting patterns in favor of this much richer analysis of employment growth patterns. Any differences in results here compared to those in Breznitz and Noonan (2014) result from these significant improvements in the data. The shift in spatial scale, for instance, allows the detection of the kind of reversal in results that Grodach et al. (2014) observe when switching from a between-city analysis to a within-city analysis.

The US Census maintains the ZBP data set, which measures annual counts of establishments by six-digit North American Industry Classification System (NAICS) codes for each zip code from 1998 to 2013. Establishment counts are further broken down by their size in terms of the number of employees.⁴ Although the ZBP includes data earlier than 1998, those years classify jobs using Standard Industrial Classification (SIC) codes and thus present problems for converting between industrial classification systems. Accordingly, 1998 is selected as the starting date for this analysis in order to maximize the time span available while maintaining the same industrial classification codes. As mentioned earlier, the zip codes being measured here represent the location of the establishment and employment, not the location of residence. Likewise, these codes indicate the establishment’s industry, rather than the sort of work done by the employees. The analysis below estimates employment trends across select groups of industries by aggregating six-digit NAICS codes into appropriate categories. The main industrial groups of interest are the arts, culture, digital media, and high tech (see Appendix 1 for details). Arts and digital media job groupings stem from the classification schemes used by Breznitz and Noonan (2014). The cultural grouping includes other aesthetic, creative, and cultural industries exclusive of the arts grouping. The high-tech grouping, selected to include industries largely separate from the arts and cultural sector, allows for a sort of falsification test in which we can confirm that the empirical models do not show the same relationships across any employment grouping.

To capture other efforts at spatially concentrating and developing the local cultural sector, the analysis overlaps the notion of formal arts districts on the ZBP data. Arts districts represent a discretely identifiable geography with strategically targeted arts and cultural activity or investment (Frost-Kumpf, 2001). We distinguish arts districts here from other forms of cultural clusters (see, e.g., Santagata’s, 2002 typology). These arts districts are distinct from other cultural clusters or areas with high density that arise informally or organically. Those clusters are what the ZBP data can already identify (and appear as JobShare₉₈ in the model), whereas arts districts reflect a formally designated zone for arts or cultural activity (regardless of whether the district actually has the highest concentration of activity in the region). Districts also differ from “cultural campuses” (where land uses are exclusively cultural) and “historic districts” (where the cultural content is merely preservation and not also ongoing cultural economic activity). These districts include prominent places such as the New Orleans arts district, New York’s Broadway theater district, and the Miami design district, as well as less renowned districts such as the River District in Rockford, IL, and the theater district in Buffalo, NY. Many well-known and long-standing clusters (e.g., New York’s garment district, Music Row in Nashville) do not qualify as arts districts under this definition.⁵ A total of 176 districts are identified and mapped for use in this study. This inventory extends the list of 99 districts used in Noonan (2013), which was based on the inventory of Frost-Kumpf (2001), by continuing to search for newer and less prominent districts.⁶ The resulting list of 169 districts established before, 2013 is thus the most complete set of US arts districts that we know of. Most of these districts involve largely coordinated marketing and branding efforts with some fairly passive local government support, although some do have more active government involvement (e.g., infrastructure investment, grants, tax benefits). In particular, these districts occupy the intersection of many interests and audiences: arts management, urban planning, economic development, public policy, and others. That cultural districts have captured the attention of audiences as disparate as scholars, planners, business groups, nonprofits, and artists suggests that the phenomenon has many dimensions and can serve many important roles.

Several districts are part of business improvement districts, which offer tax advantages and additional local control over how that revenue is spent. A few districts (e.g., Chicago’s theater district, Haynie’s Corner in Evansville, IL) receive funding via tax increment financing mechanisms that pay for current infrastructure improvements by earmarking the future stream of increased tax revenues in the area. At times, the arts community or arts agencies themselves seek the establishment of a cultural district to develop local arts and culture. Downtown business coalitions often form to promote the establishment of a cultural district for commercial purposes. City planners and regional economic development interests have also initiated cultural districts as a strategy to encourage citywide economic growth and prosperity. The variation and idiosyncrasy are substantial given the particular historic and evolutionary path by which cultural producers settled in the area (Noonan, 2013).

The data on UADMs and SSADMs are drawn from several sources to capture a long and varied list of institutions for applied higher education. Their base sample starts with all US schools with programs accredited by the National Association of Schools of Art and Design or the National Architectural Accrediting Board. Because many programs related to digital media may not be in schools with accredited arts, architecture, or design programs, we then augmented this list by going through all the institutions for applied higher education in the top 100 cities (by population). All schools in these cities with (unaccredited) arts programs or with academic programs related to digital media are added to the list. Finally, to oversample the cities that have arts districts, we followed the same procedure for schools in 48 more cities with arts districts. This is same list of cities as in Breznitz and Noonan (2014). Ultimately, 1242 schools and universities are identified and mapped. We further code all schools as either schools specializing in the arts (SSADMs; e.g., Art Institute of New York City) or universities that have broader programs (UADMs; e.g., New York University). SSADMs make up 18% of the final sample.

The unit of analysis in this study is the zip code. This is the smallest geographic area for which jobs data are available in the ZBP. A total of 31,839 observations are available for the full sample; 7367 observations are available for a subsample of zip codes that are in urban sections of MSAs. Observations for rural and small towns are omitted for analyses of the “city subsample,” reported in Appendix 2, in order to examine whether the observed relationships are peculiar to city life or are more general phenomena around the country.

Map 1 offers some perspective on the scales involved in this analysis. Using the Minneapolis–St. Paul metro area as an example, Map 1 shows the relative size of zip-code areas, the density of UADMs, the density of SSADMs, and examples of cultural districts in that metro.⁷ This metro area has three downtown districts, one of which (Minneapolis North) has an SSADM (the Institute of Production and Recording). The institutions for applied higher education and districts in this metro area are clearly concentrated near the downtown, often clustering, but also exhibiting some dispersion. The analysis here uses 5-mile buffers around schools and districts. Map 1 depicts buffers for a sample school (Academy College, which offered a program in digital arts and design) and for a sample district (Lowertown). The MSA spans roughly 100 miles from north to south and from east to west, which puts a 5-mile buffer into perspective.

Map 1.

Schools and districts in the Minneapolis–St. Paul Metro area.

Results

Our data analysis focuses on the impact of the number of arts districts or arts and digital media schools on the share of job growth in the arts and digital media by zip codes and by focusing on urban areas. Our findings support existing studies on the role of intermediaries, especially the role of educational programs and universities in economic development and innovation. The results suggest that both districts and the number of arts schools in a zip code are positively related to the share of arts and culture jobs. The share of digital media jobs grows faster in proximity to schools. These results are even stronger in urban areas. Moreover, particularly with regard to the role of educational programs, we find that when we break down the results and evaluate the relationships with schools versus districts, we find that the existence of a school within 50 miles has a larger, positive coefficient for the growth of jobs in arts and cultural sectors. The positive association with geographically remote districts disappears without an institution for applied higher education.

The models in Table 2 explain arts job trends using both SSADMs and UADMs (column 1)⁸ and using only SSADMs (column 2). The small effect sizes in Table 2 (and subsequent tables) reflect the rarity of jobs in these sectors, as reflected in the small means for JobShare98 variables in Table 1. The results confirm that the number of arts districts in a zip code is positively and significantly associated with the growth of jobs in the arts. This is not the case if the arts district is within a 5-mile radius of the zip-code boundary. From previous studies, we know that arts districts alone may not be enough to promote economic development in general or job growth in particular. Nonetheless, a designated arts district requires targeting or concentrating investment, which appears to translate into jobs created or relocated. This added value disappears quickly beyond the district boundaries, however, with a negative effect in the 5-mile buffer consistent with the districts possibly displacing jobs from surrounding areas and concentrating them in or nearer to districts.

Table 1.

Descriptive statistics (full sample, N = 31,839).

Variable	M	SD	Min	Max
ΔJobShare[Arts]	−0.0005	0.0565	−3.0508	1.6916
ΔJobShare[Culture]	−0.0018	0.0572	−1.5625	1.8571
ΔJobShare[Digital media]	0.0026	0.0794	−1.9167	1.8571
ΔJobShare[Hi-tech]	0.0022	0.1163	−5.3686	2.0682
JobShare98[Arts]	0.0078	0.0499	0	3.0508
JobShare98[Culture]	0.0150	0.0506	0	1.5859
JobShare98[Digital media]	0.0208	0.0668	0	2.2222
JobShare98[Hi-tech]	0.0334	0.1101	0	5.3686
District Count	0.0116	0.1162	0	3
District Count—5mi	0.1164	0.5052	0	7
School Count	0.0382	0.2504	0	9
School Count—5mi	0.7378	2.9547	0	44
Both District & School	0.0033	0.0571	0	1
ArtSchool Count	0.0055	0.0912	0	5
ArtSchool Count—5mi	0.1672	1.2012	0	20
School in 50mi	0.7494	0.4333	0	1
Core	0.1043	0.3057	0	1
Suburb	0.1539	0.3608	0	1
Metroarea	0.5551	0.4970	0	1

Table 2.

Regression results for changes in arts-related jobs.

	Share of jobs in the arts		Share of cultural jobs
	All	Specialty	All	Specialty
Variable	Coef.(SE)	Coef.(SE)	Coef.(SE)	Coef.(SE)
JobShare98	−0.8177***	−0.8176***	−0.7760***	−0.7775***
	0.0367	0.0367	0.0283	0.0283
District Count	0.0105***	0.0104***	0.0178***	0.0181***
	0.0023	0.0023	0.0032	0.0032
District Count—5mi	−0.0015***	−0.0017***	0.0001	0.0005
	0.0006	0.0005	0.0009	0.0009
School Count	0.0065***	0.0121**	0.0039***	0.0132***
	0.0017	0.0049	0.0013	0.0047
School Count—5mi	0.0005***	0.0014***	0.0013***	0.0029***
	0.0001	0.0002	0.0003	0.0006
Both District & School	−0.0031	0.0021	0.0105	0.0111
	0.0048	0.0044	0.0105	0.0104
Core	0.0036***	0.0051***	0.0071***	0.0095***
	0.0011	0.0011	0.0011	0.0010
Suburb	0.0016**	0.0020**	0.0022***	0.0029***
	0.0008	0.0008	0.0007	0.0007
Metroarea	0.0021***	0.0022***	0.0018***	0.0020***
	0.0005	0.0005	0.0006	0.0007
Constant	0.0036***	0.0036***	0.0064***	0.0064***
	0.0003	0.0003	0.0004	0.0004
N	31839	31839	31839	31839
Adjusted R²	0.5175	0.5175	0.4624	0.4630

*, **, and ***Indicate significance at the 0.1, 0.05, and 0.01 levels, respectively.

Examining the role of arts schools, we again find a positive relationship. Having more schools in a zip code is associated with larger increases in arts-related jobs. Unlike the case of arts districts, this finding still holds when we examine the number of schools within 5 miles of the zip code. This is true whether we examine the broad set of schools with arts and digital media programs or limit the analysis to include only SSADMs. The effect sizes from the rarer SSADMs are several times larger. These findings support previous studies regarding the role of universities in general and arts programs in particular as a positive impact on jobs in the arts and economic development (Breznitz and Noonan, 2014; Rantisi and Leslie, 2015; Rushton, 2013). The results for districts and schools have the same strong results when we analyze them only in urban areas.

Table 2 also shows results for (non-arts) cultural-sector jobs. These results look generally quite similar to those for jobs in the arts. The main difference between jobs in the arts and cultural jobs can be seen in the 5-mile buffer. Growth in both kinds of jobs occurs close to the districts, but districts may actually relocate jobs in the arts by drawing them in from surrounding areas while cultural jobs in surrounding areas appear largely unaffected. The positive effect of hosting institutions for applied higher education is smaller in magnitude for cultural jobs than for jobs in the arts, but the effect fades more slowly in the 5-mile buffer. Apparently, cultural jobs do not need to be in such close proximity to the educational programs. The interaction between schools and districts appears insignificant in Table 2 (and in all the tables that follow). In zip codes with districts or institutions for applied higher education, the share of arts and cultural jobs grew faster, but the advantage is not greater when both institutions were present.

Examining the relationships between arts districts and schools and trends in digital media jobs (Table 3) tells us a different story. The arts districts’ and art schools’ significant associations with arts or cultural jobs disappear here. The only consistently significant factors are the number of schools within 5 miles of the zip code. These results hold even when we reduce our sample and focus only on urban zip codes. These findings correspond with what we know about the digital media industry. Many of the occupations in digital media, such as software engineers, graphic designers, and computer scientists, do not require them to live in close proximity to the school where they studied. Unlike firms in the sciences, which need connections to sources of knowledge, such as university labs, most digital media firms can locate almost anywhere. However, because these firms are very “artsy” in nature and need to be close to the business community and other resources, they tend to grow more rapidly within a 5-mile radius of schools. Hence, we can say that by being both high tech and arts based, digital media jobs show patterns that are similar to those in both. In one way, digital media jobs show patterns similar to those in high-technology jobs (or even biotechnology jobs), which need to remain close to the tacit knowledge at the UADMs (Breznitz, 2007; Breznitz and Anderson, 2006; Rantisi and Leslie, 2015). However, when only SSADMs are included in the analysis, being within a 5-mile buffer of arts districts has a positive (but imprecisely estimated) effect on growth in the share of digital media jobs. This result is consistent with the different location patterns for digital media firms and arts-related establishments (comparing District Count—5mi in column 2 in Tables 2 and 3), and these digital media firms’ interest in being somewhat near sources of talent.

Table 3.

Digital media and high-tech jobs.

	Share of digital media jobs		Share of high-tech jobs
	All	Specialty	All	Specialty
Variable	Coef.(SE)	Coef.(SE)	Coef.(SE)	Coef.(SE)
JobShare98	−0.5871***	−0.5872***	−0.6430***	−0.6429***
	0.0392	0.0392	0.0362	0.0362
District Count	0.0042	0.0042	0.0001	−0.0014
	0.0050	0.0051	0.0057	0.0056
District Count—5mi	0.0027	0.0035*	0.0017	0.0030
	0.0019	0.0019	0.0024	0.0023
School Count	0.0003	0.0020	0.0044*	0.0038
	0.0018	0.0045	0.0023	0.0038
School Count—5mi	0.0009**	0.0017**	0.0003	0.0001
	0.0004	0.0008	0.0003	0.0007
Both District & School	0.0118	0.0117	−0.0008	0.0061
	0.0114	0.0113	0.0087	0.0080
Core	0.0203***	0.0217***	0.0315***	0.0329***
	0.0025	0.0023	0.0029	0.0028
Suburb	0.0137***	0.0142***	0.0260***	0.0262***
	0.0015	0.0015	0.0020	0.0020
Metroarea	0.0066***	0.0067***	0.0076***	0.0076***
	0.0008	0.0008	0.0012	0.0012
Constant	0.0058***	0.0059***	0.0116***	0.0116***
	0.0006	0.0006	0.0008	0.0008
N	31,839	31,839	31,839	31,839
Adjusted R²	0.2396	0.2395	0.3678	0.3676

*, **, and ***Indicate significance at the 0.1, 0.05, and 0.01 levels, respectively.

When we examine the distribution of job growth within the urban core, the suburbs, and the metro areas, we find that concentrations of digital media jobs are growing especially rapidly in the core of urban areas and less rapidly in the suburbs but still faster than elsewhere in metro areas or in rural areas. This simply reflects the tendency of digital media jobs to be drawn to urban cores and to employment density, rather than more remote “greenfields.” Although the model does not provide explicit answers to this phenomenon, we believe that the increased access to amenities and knowledge spillovers in higher density areas have probably attracted digital media jobs over the period under study here.

Because digital media jobs are closely related to many high-tech jobs, we wanted to verify that the growth in the share of digital media jobs is not based on general trends among high-technology jobs. Observing patterns in the digital media sector similar to those among high-tech firms would cast doubt on the role of institutions for applied higher education—with programs related to the arts and digital media—as intermediaries. Our model results (Table 3) confirm our emphasis on institutional intermediaries in this sector. High-technology jobs react as expected by concentrating in the urban core and less so in the suburbs—a pattern similar to that in digital media jobs. That strong geographic sorting for high-technology jobs, however, is largely unrelated to these concentrations of arts and cultural activity. The results indicate that arts districts have no significant association with high-tech job growth—even when the sample is limited to urban core areas. Only in zip codes with schools—yet not SSADMs—might high-tech job growth be greater. Except for generally weak clustering around universities, these models indicate little substantial role for arts districts or general proximity to arts programs beyond the general effects of urban density. Schools with arts programs tend to witness stronger growth of arts, cultural, and digital media jobs in close proximity that we simply do not see for high-tech jobs. And proximity to cultural districts is associated only with stronger growth of arts and cultural jobs, with basically no effect on the growth of the share of high-tech jobs and weaker effects on the growth of digital media jobs.

Our measurement of arts districts and arts-related programs counted how many districts or schools are in a zip code or within 5 miles of a zip code. But what about less geographic proximity to a school? In knowledge-based industries, proximity matters. Hence, what about jobs in the arts and digital media? Maybe people went to a school and then commuted to a job in a suburb. Or the opposite: perhaps they graduated from an arts program in the suburbs and then worked downtown. The previous models explore only very close proximity, within a zip code or within 5 miles of it, but miss intermediate (i.e., within-MSA) distances in which schools, districts, and jobs may be interconnected. These metro-scale connections may be very important and, as Grodach et al. (2014) point out, different from those observed at shorter distances. To address these questions, we established the dummy variable School in 50mi to control for whether a school or university with an arts program is located within 50 miles of the zip code. This term is also interacted with District Count to see whether the role of hosting arts districts differs if that district is within commuting distance of SSADMs or UADMs.

Like Tables 2 and 3, Tables 4 and 5 each have four columns,. Columns 1 and 3 refer to models including the full set of UADMs, whereas columns 2 and 4 use only SSADMs. (Appendix Tables 8 and 9 are analogous tables with the city-only subsample of zip codes.) Table 4 shows considerable change from Table 2, thus indicating that the inclusion of more flexible controls for the joint locations of SSADMs, UADMs, and arts districts matters. The effects identified in Table 2—in which zip codes with districts and those with schools experienced more robust growth in the share of arts and cultural jobs—differ if we consider whether an SSADM is within commuting distance of the zip code. The main effect of a zip code with an arts district becomes insignificant, revealing that the arts district effect observed in Table 2 is driven by zip codes that have art disticts and are also within 50 miles of an SSADM. For cultural districts to be associated with stronger growth in arts and cultural jobs, a nearby UADM is apparently a prerequisite.

Table 4.

Regression results for trends in the share of arts and cultural jobs and intermediate proximity.

	Share of jobs in the arts		Share of cultural jobs
	All	Specialty	All	Specialty
Variable	Coef.(SE)	Coef.(SE)	Coef.(SE)	Coef.(SE)
JobShare98	−0.8178***	−0.8177***	−0.7760***	−0.7776***
	0.0367	0.0367	0.0283	0.0283
District Count	0.0035	0.0034	−0.0070*	0.0058
	0.0021	0.0022	0.0040	0.0039
District Count—5mi	−0.0015***	−0.0017***	0.0001	0.0005
	0.0006	0.0005	0.0009	0.0009
School Count	0.0065***	0.0121**	0.0039***	0.0132***
	0.0017	0.0049	0.0013	0.0047
School Count—5mi	0.0005***	0.0014***	0.0013***	0.0029***
	0.0001	0.0002	0.0003	0.0006
Both District & School	−0.0033	0.0018	0.0101	0.0106
	0.0049	0.0045	0.0105	0.0105
School in 50mi	0.0005	0.0006	0.00004	0.0003
	0.0006	0.0006	0.0006	0.0006
District Count * School in 50mi	0.0072**	0.0072**	0.0111**	0.0127***
	0.0033	0.0033	0.0050	0.0048
Core	0.0036***	0.0050***	0.0071***	0.0094***
	0.0011	0.0011	0.0011	0.0010
Suburb	0.0015*	0.0020**	0.0021***	0.0029***
	0.0008	0.0008	0.0007	0.0007
Metroarea	0.0019***	0.0020***	0.0018***	0.0019***
	0.0005	0.0005	0.0006	0.0006
Constant	0.0033***	0.0032***	0.0034***	0.0063***
	0.0004	0.0004	0.0005	0.0005
N	31839	31839	31839	31839
R² adjusted	0.5175	0.5175	0.4624	0.4630

*, **, and ***Indicate significance at the 0.1, 0.05, and 0.01 levels, respectively.

Table 5.

Regression results for trends in the share of digital media and high-tech jobs and intermediate proximity.

	Share of digital media jobs		Share of high-tech jobs
	All	Specialty	All	Specialty
Variable	Coef.(SE)	Coef.(SE)	Coef.(SE)	Coef.(SE)
JobShare98	−0.5873***	−0.5875***	−0.6432***	−0.6431***
	0.0392	0.0392	0.0362	0.0362
District Count	0.0411	0.0399	0.0799	0.0782
	0.0409	0.0409	0.0675	0.0675
District Count—5mi	0.0027	0.0035*	0.0017	0.0029
	0.0019	0.0019	0.0024	0.0023
School Count	0.0002	0.0019	0.0042*	0.0036
	0.0018	0.0045	0.0022	0.0038
School Count—5mi	0.0009**	0.0017**	0.0003	0.0001
	0.0004	0.0008	0.0003	0.0007
Both District & School	0.0134	0.0130	0.0026	0.0091
	0.0113	0.0112	0.0083	0.0076
School in 50mi	0.0026***	0.0027***	0.0034***	0.0035***
	0.0008	0.0008	0.0012	0.0012
District Count * School in 50mi	−0.0383	−0.0370	−0.0827	−0.0824
	0.0412	0.0411	0.0677	0.0677
Core	0.0202***	0.0215***	0.0314***	0.0327***
	0.0025	0.0023	0.0029	0.0028
Suburb	0.0135***	0.0139***	0.0257***	0.0259***
	0.0015	0.0015	0.0020	0.0020
Metroarea	0.0058***	0.0059***	0.0066***	0.0066***
	0.0008	0.0008	0.0012	0.0012
Constant	0.0044***	0.0044***	0.0097***	0.0096***
	0.0006	0.0006	0.0010	0.0010
N	31839	31839	31839	31839
Adjusted R²	0.2398	0.2398	0.3680	0.3679

*, **, and ***Indicate significance at the 0.1, 0.05, and 0.01 levels, respectively.

Remarkably, however, Table 5 shows relationships in explaining the growth of digital media and high-tech jobs very similar to those in Table 3. When we examine the share of growth of digital media jobs in all the zip codes, even after we add the School in 50mi variable, we see no difference in the role of hosting arts districts or schools with arts and digital media programs. Being within 50 miles of UADMs is associated with faster growth of digital media jobs in the full sample, not in the city-only subsample (Appendix Table 9). This significant coefficient, as in the share of high-tech jobs models in Table 5, likely reflects simply a richer characterization of broader geographic factors (i.e., the URBANICITY vector) than it does a direct effect of those schools. Nonetheless, the District Count*School in 50mi interaction term is insignificant across Table 5, suggesting that arts districts’ relationships with technology-based job growth does not depend on whether an SSADM or UADM is within commuting distance.

Discussion and conclusion

Regardless of whether the contribution is locational or indirect, studies have shown that art and culture play an important role in today’s knowledge economy. Education, by contrast, has been an important factor in studies of economic development and innovation. However, education has not had as prominent a role in the literature related to cultural industries. Recent studies in economic geography have added a new approach by focusing on universities and applied arts programs as important factors in employment and innovation growth in industries related to the arts and digital media. In view of institutional theory literature, these studies analyze institutions of applied higher education as intermediaries that promote economic growth.

Based on a previous study by Breznitz and Noonan (2014), this paper continues the focus on the role of educational programs in the arts and digital media as a factor in the growth of jobs in the arts and digital media. The analysis here evaluates the presence of (1) arts districts and concentrated activity related to their establishment and (2) universities with arts-related educational programs. Education here is expanded from the analysis of research-intensive universities, as Breznitz and Noonan (2014) use, to include both programs at all levels of universities as well as all certified arts programs. The analysis has several layers. Although the paper examines the spatial relationships with education, it focuses on several sectors within the cultural industries: cultural industries in general, arts-related employment, and digital media. It also examines the roles of these spatially concentrated, arts-related activities at a smaller spatial scale than previous work, allowing for the impacts to diffuse at different spatial scales. By layering our analysis, we can evaluate the difference in sensitivities of the different sectors.

The results of this study are consistent with those of studies on the economic impact of education and indicate the consistent importance of educational programs. In particular, in this paper, we find that proximity to arts and digital media education programs are more important than proximity to art districts. Moreover, we find that job growth depends on the employment sector. Growth in jobs related to the arts and culture can be found in closer proximity to educational programs than is the case for digital media jobs. In particular, we find that digital media jobs were growing mostly in urban core areas, with a weaker effect for being within a few miles of arts education programs.

Our examination of the role of arts districts shows that they have a positive association with arts and cultural jobs but only in the zip code where the district is located and only when an SSADM is also within commuting distance. This indicates a very specialized reliance of the industry on arts programs in urban cores. The results for the growth of digital-media jobs—in which the share of jobs grows more rapidly within 5 miles of arts districts and within 5 miles of arts schools—highlight an important role of proximity (but not necessarily colocation) to these concentrations of arts-related activity. The lack of significant results for other high-technology jobs casts doubt on the importance of concentrating arts activity in attracting these workers.

The importance of proximity and density, as urban cores clearly exhibit different concentrations of arts- or technology-related jobs, recommends further investigation of these relationships. A comparison of large (rural) zip codes and small (urban core) zip codes indicates that their 5-mile buffers differ in size.⁹ Thus, the analysis here uses more observations from smaller, high-density zip codes in order to characterize the patterns in urban cores more richly, but proximity and buffer distance is held fixed. Future research could explore alternative units of analysis and buffer sizes as densities vary.

Overall, we have several major results. First, the results differ markedly from those in Breznitz and Noonan (2014). These differences stem largely from the great improvement in data quality employed here. As noted above, this analysis uses a much higher spatial resolution, wider sampling of the United States, and better spatial location of jobs (rather than residence). Rather than contradicting those previous results, we note that Grodach et al. (2014) observe a similar inconsistency in results when analyzing the across-city distribution of jobs versus the within-city distribution. The Breznitz and Noonan (2014) article centers on across-city differences among highly aggregated data, whereas the present analyses explain considerable within-city variation.

The major change here is that we reaffirm the importance of industry-specific training. It is not about research universities or arts schools. The job growth is related to the existence of education programs in the specific occupations.

Second, we find advantages for arts schools over universities. Yes, both show positive relationships. However, the results show strong positive arts schools’ coefficients for job growth in arts-related sectors, culture-related sectors, and digital media (distinct from high-tech jobs more generally) sectors. These results are robust across urban or all zip codes. They include positive associations with zip codes within 5 miles of arts schools, and—for arts and cultural jobs—even stronger trends for zip codes where SSADMs are located. These effects are strongest for SSADMs, but exist for all schools that have arts or digital media programs (both SSADMs and UADMs).

Third, the findings for schools specializing in the arts (SSADMs) are not matched by arts districts. Those districts’ roles are limited largely to sectors related to the arts and culture, and even then the geographic spillover of those districts depends on the sector. Zip codes within 5 miles of the districts do not enjoy additional arts-related job growth (as jobs may be relocated from “nearby” districts to within the district). Stronger cultural-job trends diffuse more gradually around districts. This more nuanced role of districts is further conditioned by the intriguing result that cultural districts’ relationships appear driven by those districts that are also within commuting distance of an arts school. Having an arts district can help a locality with arts-related job growth, but only if it also enjoys proximity to a source of supply for artistic talent. Moreover, previous arguments (e.g., Florida, 2002) that arts and cultural activity attract creative-class workers find little or no support here, beyond the sorting of arts jobs around arts districts. Rather, the districts do not attract the (prized) digital media or high-technology jobs, although at least the former are attracted to or generated by the intermediaries focusing on developing human capital (schools). Industry-specific capital (human, social, physical, etc.) remains important in explaining where jobs are locating, perhaps more than just concentrating activity (e.g., arts and cultural districts that often function like entertainment districts). Some caution in interpreting these district effects causally is warranted, especially given the nonrandom creation of arts districts and the possibility that job growth leads to district formation, rather than the other way around. Although we do not detect a substantive difference in district effects between the pre-1998 districts and newly added districts, future research should explore how districts’ roles evolve with age.

Fourth, one finding that persists across the relevant job categories—and is consistent with previous findings (Breznitz and Noonan, 2014; Rantisi and Leslie, 2015)—is the positive role of proximity to arts schools. Proximity to these intermediaries appears to effectively promote job growth in these creative cities. Further, they enable efforts such as arts districting (and perhaps creative placemaking more broadly) to catalyze job growth in arts and cultural sectors. These findings are also consistent with general studies on intermediaries as well as studies on institutions for applied higher education in regional economic development. The strength of the relationships of different actors in regions help promote knowledge transfer and “know-how,” which positively affect innovation and job creation.

The findings here hold important implications for policymakers. This emerging area of research shows that patterns that hold between cities may vary within cities. Because we cannot merely scale down intercity competition and sorting to intracity dynamics, this suggests that policymakers ought to take different approaches and use different tools depending on the scale. Further, arts specialty schools appear to be effective in shaping where arts and cultural job growth occurs, giving local policymakers another tool in their toolbox for creative placemaking. That capital-intensive arts schools in the metro area appear to be a prerequisite for arts districts to influence where these jobs grow offers some evidence on the limits and interdependencies for districts. Simply designating a district on paper without industry-specific capital nearby may accomplish little. Consistent with previous research on similar intermediaries, the strong effects of arts schools in promoting job growth and innovation are an important signal for policymakers who want to foster creative cities.

Further research is needed to evaluate the networks and varieties of relationships within the arts and digital media industries as a variable in job creation. This can be especially useful with different units of analysis, to avoid some aggregation problems of zip codes and possibly leverage more qualitative evidence. Connecting this information to policy and funding of the arts and the creation of arts districts is another aspect worth investigating. Going beyond the arts and digital media sectors may also identify additional intermediaries that are critical in shaping regional economic activity.

Footnotes

Acknowledgements

The authors would like to thank Paige A. Clayton, and two anonymous reviewers for the input and comments.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received the following financial support for the research, authorship, and/or publication of this article: Funding for the project was available through the National Endowment for the Arts.

Notes

Appendix 1. Main industry groups by NAICS codes

Appendix 2. Results for city-only subsamples

References

Asheim

Gertler

(2006) The geography of innovation: Regional innovation systems. In: Fagerberg

Mowery

(eds) The Oxford Handbook of Innovation. Oxford and New York: Oxford University Press, pp. 291–317.

Breznitz

(2007) From ivory tower to industrial promotion: The development of the biotechnology cluster in New Haven, Connecticut. Revue d’Economie Industrielle 120(4th trimester): 115–134.

Breznitz

Anderson

(2006) Boston metropolitan area biotechnology cluster. Canadian Journal of Regional Science 28(2): 249–264.

Breznitz

Noonan

(2014) Arts districts, universities, and the rise of media arts. The Journal of Technology Transfer 39(4): 594–615.

Brooks

Kushner

(2001) Cultural districts and urban development. International Journal of Arts Management 3(2): 4–15.

Chapple

Jackson

Martin

(2010) Concentrating creativity: The planning of formal and informal arts districts. City, Culture and Society 1(4): 225–234.

Currid

(2007) How art and culture happen in New York: Implications for urban economic development. Journal of the American Planning Association 73(4): 454–467.

Currid

Williams

(2010) The geography of buzz: Art, culture and the social milieu in Los Angeles and New York. Journal of Economic Geography 10(3): 423–451.

Eaton

Bailyn

(1999) Work and life strategies of professionals in biotechnology firms. Annals of the American Academy of Political and Social Science 562(1): 159–173.

10.

Feldman

(1994) The Geography of Innovation. Dordrecht: Kluwer.

11.

Florida

(2002) The Rise of the Creative Class. New York: Basic Books.

12.

Frost-Kumpf

(1998) Cultural Districts: The Art as Strategy for Revitalizing Our Cities. Washington, DC: Americans for the Arts.

13.

Frost-Kumpf

(2001) Cultural districts: Arts management and urban redevelopment. PhD dissertation, Pennsylvania State University, State College, PA1, USA.

14.

Grodach

(2011) Art spaces in community and economic development: Connections to neighborhoods, artists, and cultural economy. Journal of Planning Education and Research 31(1): 74–85.

15.

Grodach

Currid-Halkett

Foster

et al . (2014) The location patterns of artistic clusters: A metro- and neighborhood-level analysis. Urban Studies 51(13): 2822–2843.

16.

Howells

(2006) Intermediation and the role of intermediaries in innovation. Research Policy 35(5): 715–728.

17.

Jaffe

Trajtenberg

Henderson

(1993) Geographic localization of knowledge spillovers as evidenced by patent citations. Quarterly Journal of Economics 108(3): 577–598.

18.

Jakob

van Heur

(2015) Editorial: Taking matters into third hands: Intermediaries and the organization of the creative economy. Regional Studies 49(3): 357–361.

19.

Ladry

Bianchini

Ebert

et al . (1996) The Creative City in Britain and Germany. London: Anglo-German Foundation for the Study of Industrial Society.

20.

Lloyd

(2010) Neo-Bohemia: Art and Commerce in the Postindustrial City. 2nd ed. New York: Routledge.

21.

Markusen

Gadwa

(2010a) Arts and culture in urban or regional planning: A review and research agenda. Journal of Planning Education and Research 29(3): 379–391.

22.

Markusen

Gadwa

(2010b) Creative Placemaking. Washington, DC: Mayors’ Institute on City Design and the National Endowment for the Arts.

23.

Markusen

Schrock

(2006) The artistic dividend: Urban artistic specialization and economic development implication. Urban Studies 43(10): 1661–1686.

24.

Marshall (1890). A Principles of Economics. London: Macmillan.

25.

Mommaas

(2004) Cultural clusters and the post-industrial city: Toward the remapping of urban cultural policy. Urban Studies 41(3): 507–532.

26.

Negus

(2002) The work of cultural intermediaries and enduring distance between production and consumption. Cultural Studies 16(4): 501–515.

27.

Nixson

Du Gay

(2002) Who needs cultural intermediaries?. Cultural Studies 16(4): 495–500.

28.

Noonan

(2013) How US cultural districts reshape neighborhoods. Cultural Trends 22(3–4): 203–212.

29.

O’Connor

(2015) Intermediaries and imaginaries in the cultural and creative industries. Regional Studies 49(3): 374–387.

30.

Ponzini

Rossi

(2010) Becoming a creative city: The entrepreneurial mayor, network politics and the promise of an urban renaissance. Urban Studies 47(5): 1037–1057.

31.

Power

Scott

(eds) (2004) Cultural Industries and the Production of Culture. London: Routledge.

32.

Rantisi

(2004) The ascendance of New York fashion. International Journal of Urban and Regional Research 28(1): 86–106.

33.

Rantisi

Leslie

(2015) Significance of higher education institutions as cultural intermediaries: The case of the Ecole National de Cirque in Montreal, Canada. Regional Studies 49(3): 404–417.

34.

Rushton

(ed.) (2013) Creative Communities: Art Works in Economic Development. Washington, DC: Brookings Institution Press.

35.

Santagata

(2002) Cultural districts, property rights and sustainable economic growth. International Journal of Urban and Regional Research 26(1): 9–23.

36.

Saxenian

(1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Cambridge: Harvard University Press.

37.

Scott

(2000) The cultural economy of Paris. International Journal of Urban and Regional Research 24(3): 554–566.

38.

Scott

(2001) Globalization and the rise of city-regions. European Planning Studies 9(7): 813–826.

39.

Scott

(2006) Entrepreneurship, innovation and industrial development: Geography and the creative field revisited. Small Business Economics 26(1): 1–24.

40.

Sedita

(2008) Interpersonal and inter-organizational networks in the performing arts: The case of project-based organizations in the live music industry. Industry & Innovation 15(5): 493–511.

41.

Silver

(2012) The American scenescape: Amenities, scenes and the qualities of local life. Cambridge Journal of Regions, Economy and Society 5(1): 97–114.

42.

Storper

(1997) The Regional World: Territorial Development in a Global Economy. New York: Guilford Press.

43.

Taylor

(2015) Between culture, policy and industry: Modalities of intermediation in the creative economy. Regional Studies 49(3): 362–373.

44.

Tomusk

(2011) The micropolitics of knowledge in England and Europe. In: Rhoten

Calhoun

(eds) Knowledge Matters: The Public Mission of the Research University. New York, Columbia University Press, pp. 377–396.

45.

Vinodrai

(2015) Constructing the creative economy: Design, intermediaries and institutions in Toronto and Copenhagen. Regional Studies 49(3): 418–432.