– DRAFT COPY –
PLEASE DO NOT CITE WITHOUT AUTHORS’ PERMISSION
April 20, 2012
Arts Districts, Universities, and the Rise of Media Arts
Shiri M. Breznitz and Douglas S. Noonan
Introduction
According to the new economic growth theory, people and ideas create technological change,
which results in economic growth. In his work, Romer refers to the use of existing resources in a
different way as a new recipe for growth. Similarly in the last decade, arts and culture have been
placed at the center of attention when discussing economic growth. In particular, studies on the
“creative class” have been using arts and culture as an important factor impacting local
economies. In addition, studies on local economic development have frequently viewed
universities as a major factor in economic growth. In the middle of this discussion is new
economic growth via creativity, via new recipes and new combinations of local capital, and via
innovation centers. Combining these disparate literatures brings to center stage both clusters of
arts and culture and concentrations of research and human capital development. Hence, the
focus of this paper is to analyze the dual impacts of universities and arts districts on innovation
and economic growth through employment in media arts.
There are several ways to identify clusters of arts and culture in an economy. The approach
taken here opts for officially designated arts districts as a formally recognized cluster that
1
generally manifests as part of some targeting public investment. For simplicity only, throughout
we refer to cultural districts and arts districts synonymously. Frost-Kumpf (1998) lists over 90
US cities with cultural districts, which gives us an excellent measure of arts districts observed
roughly in the middle of the available timespan of data. When coupled with the long-standing
existence of research intensive (i.e., R1) universities, these two major investments (universities,
arts districts) are seen as “treatments” that are hypothesized to affect long-term development of
arts-related activities. Simple statistical inferences in this context, especially for arts districts,
will be limited by obvious endogeneity concerns. We address these preliminarily by employing
a difference-in-difference estimator to see how trends in employment (and innovation in) in
media arts differ in cities with arts districts (or universities) and those without. By measuring
arts-related employment and innovation as shares of overall employment and innovation, our
approach resembles a sort of triple-difference estimator to identify how media arts employment
(and innovation) differs from overall employment (innovation) in the city, and how trends in that
difference differ by treatment.
The results indicate that cultural districts have a consistently positive effect on local media arts
economic activity – employment and innovation. The same cannot be said for research
universities. The results show that R1 universities in the broader urban area positively impact
media arts employment levels and trends. R1 universities in the same city have no impacts on
arts and media arts employment. But universities do not appear to differentially impact media
arts-related patenting. Cultural districts appear to foster arts and media arts related employment
between 1999 and 2006. However, the employment trend is transitory when the analysis period
2
is extended to 2011. The more stable effect of arts districts arises in the patenting analysis, where
these clusters appear to promote innovation.
Culture, Universities, and Economic Development: What do we know?
Economic development studies view industrial clusters, universities, and culture as important.
Each of these factors is significant but requires additional factors in the form of human capital,
knowledge, and absorptive capacity to make local economic contribution. In particular, studies
show that the strength of universities in economic impact is not limited to technology
commercialization. Other activities in the form of policy, real estate, and neighborhood
revitalization have shown as much vitality although less quantifiably. Similarly, culture in
general and art in particular have been increasingly viewed as economic-growth generating tools
through the creation of cultural districts and institutions as well as a tool for economic
development through the attraction of high quality labor and large corporations. The following
section examines each of these factors in more detail.
The role of the university in economic development
Starting in the 19th century, universities increasingly have become important factors for local
economic growth. Moreover, in the late 1980s and early 1990s,with the development of the
knowledge economy, where knowledge is a product, led to an increased reliance on universities’
contribution to financial economic development by way of focusing on outcomes of academic
research (Etzkowitz and Leydesdorff, 1997, Goddarrd and Chatterton, 1999). As a result,
universities have played a central role in contributing to regional and national economies.
3
Universities are an important source of new knowledge and technology, with the potential to be
commercialized (Scott, 1977). The pressure on modern universities to pay back to the
community has created what is known as the “third role” of universities, by which many
universities are now obliged to make a contribution to society through research and development
(R&D), collaborations, and technology transfer with industry (Minshall et al., 2004).
Universities use a variety of mechanisms to transfer technology to industry. Technology
commercialization is the common way by which universities’ economic development is
evaluated (Etzkowitz and Leydesdorff, 1997, Feldman and Breznitz, 2009, Lawton Smith, 2000,
Lerner, 2004, Markman et al., 2005, Nelsen, 2003, Rahm et al., 2000, Siegel et al., 2004).
Formal mechanisms include patents, licenses, and spinout firms (Di Gregorio and Shane, 2003,
Link and Siegel, 2005, Mowery et al., 1999, Owen-Smith and Powell, 2001, Sampat, 2006,
Siegel et al., 2003, Thursby et al., 2001). Though not the only economic contribution universities
make, patents, licenses and spinout firms are easy to quantify and use to measure university
productivity.
In addition, studies have found a direct link between the distance of firms and universities as a
proof for the knowledge spillover generated by university research (Feldman, 1994, Jaffe et al.,
1993). This points to a geographic cost to the transmission of knowledge and creativity out of
university centers, and it raises interesting questions about the mechanisms behind the spatial
clustering of this knowledge and creativity (e.g., geographic sorting by innovators, transportation
costs related to communication costs, the role of in-person experiences in creating or consuming
the knowledge). In the center of many successful technological regions we find an innovative
4
university with a research focus on the same technology. The mere existence of a university in a
region, however, is not a guarantee for economic success. While considerable research effort has
gone into understanding the role of universities in generating innovations, transferring
technologies and new ideas, and stimulating (local) economic development, nearly all of that
work has focused on various technologies with scientific emphases or with commercial and
industrial applications. Universities’ roles in spawning innovations and economic impacts in
arts-related fields have not received much attention to this point. While this oversight may be for
several fairly obvious reasons, we argue that arts-related impacts of universities are particularly
interesting not just to the arts sector but also as a way of understanding broader university
knowledge spillovers in areas not typically the focus of “tech transfer” or local economic
stimulus.
When we examine the overlapping relationships between culture and universities we find that
universities are found in most highly ranked cities in the “creative class index” (Florida, 2002).
Universities are an important part of local economic growth. Florida (2002) claims that the main
contribution of universities is in the 3Ts: technology, talent, and tolerance. Hence, the impact of
university R&D is important, but its main effects might arise through a different role: its ability
to draw talented students and faculty and grow a talented work pool, as well as its education
program and diversity that creates an environment of tolerance, to which, according to Florida,
many firms are drawn.
Strengthening Florida’s notion as to the extended role of universities in economic development,
studies find that there are many services and activities provided by universities, such as
5
community and volunteer work that cannot easily be quantified and is not recognized by the
public service provided by universities (Breznitz and Feldman, 2012, Forrant et al., 2001,
Maurrasse, 2001). In particular, Breznitz and Feldman have identified two additional roles for
universities: policy advice and research as well as becoming an active player in local economic
initiatives. This includes efforts like workforce development, partnership development,
community development, real estate development (Breznitz and Feldman, 2012).
Culture and art in economic development
The importance of culture as a factor in economic development is not new. Starting in the 1960s
this argument was a contributing factor to the creation of many culture and culture-related
institutions such as the National Endowment for the Arts is the US (Markusen and Gadwa,
2010). The role of culture and art in local economies can be divided to two parts: generating
direct income from local institutions and artists, and as a powerful tool for drawing large
corporations and highly skilled labor to a region.
Artists as an industry sector on their own have been shown to make a direct contribution via
exporting goods and services. Though many artists do not sell their products in the local
economy they contribute to the exporting base of their region (Markusen and Schrock, 2006,
Markusen and King, 2003). Moreover, the existence of art and cultural institutions such as
museums and theaters draws outside funding and visitors to a region. Though, some studies
claim that the cost-benefit analysis is very difficult to measure and to compare with places that
have not fared so well due to cultural investments (Markusen and Gadwa, 2010).
6
The rise of the new media arts sector, defined in more detail below, in recent years in some
regions has demonstrated several important properties that we investigate further here. First, the
emergence of the media arts sector shows how the venerable arts industry sector can have a
large, disruptive, and exciting impact on economies. The rapid growth and sheer size of the
video game industry, for instance, reflects both a potential and a reality for arts-related
innovation to occupy a substantial role in an economy. Second, the (spatial) clustering of media
arts production is commonly observed at least anecdotally. The serious efforts from states and
localities to attract more of that business to their region point to media arts as an avenue (viable
or not) for local economic development. Third, how new media arts combines truly artistic and
creative content with emergent and novel technologies – the staple of most innovation studies –
offers a fascinating intersection between arts and technology and an area with seemingly greater
potential for economic impact and innovation for universities.
Importantly, the indirect contribution of culture and art can be measured through its impact on
other firms and industries (Markusen and Schrock, 2006, Markusen and King, 2003). In
particular, arts and artists attract firms and high level human capital to a region contributing to
the productivity of other industries (Florida, 2002). Similarly to universities, culture and art are
contributing factors to the environment in which firms and individuals choose to live and work
(Eaton and Bailyn, 1999).Studies claim that some cities or regions have added jobs more rapidly
over time as a result of cultural industries (Pratt, 1997). These additional positions are attributed
to large concentration of cultural industries, which have been viewed important to diverse
economic base in highly specialize cities (Pratt, 1997). However, studies find conflicting results
7
regarding the possible causality between cultural activity and the attraction of other firms and
high income workers (Markusen, 2006, Gleaser, 2004).
In addition, physical investment in arts helps revitalize neighborhoods or districts (Bianchini et
al., 1988, Landry et al., 1996). As stated by Perryman: “Even in less obvious places the cultural
industry plays a measurable economic role” (Perryman, 2000). Hence, culture can be used as an
economic development tool. Its contribution, however, depends on a change in the political
economy of cities and in the organization and mission of cultural institutes, which create close
and mutual benefits between urban political, economic, and cultural entrepreneurs (Florida,
2002, Strom, 2002).
Though studies on culture view its impact positively they also caution us about overestimating its
impact. “Conventional economic impact studies could be strengthen by stricter criteria for
determining the extent to which a given arts or cultural investment induces external income,
increased scrutiny with regards to the substitution effect, and attention to efficiency and equity
concerns through cost-benefit analyses and participation studies” (Markusen and Gadwa, 2010).
Research Design and Method
To assess the importance of universities and cultural districts in the economic development
literature, this paper analyzes the impact of universities and arts districts on economic growth
through employment and innovation in media arts. In particular, the paper focuses on three
research questions:
 How do trends in employment and innovation in media arts differ in cities with arts
districts and universities?
8
 How do media arts employment and innovation differ from overall employment and
innovation in the city, and how do trends in that difference differ by treatment (i.e.,
whether the city has an arts district or university)?
 What is the effect of the combination of a major university and a cultural district in the
same area on media-arts employment and innovation?
We choose to focus on media arts as our industry of choice because of its strong relationships to
both arts and university research. Moreover, media arts is a booming industry including but not
exclusive to: video games, motion pictures, television, internet publishing, and online
entertainment (streaming content, music & video downloads, etc.). For the purpose of this paper,
we define media arts as “An artistic and creative content in new (especially digital) formats and
media, such as digital art, computerized animation, internet and interactive art.”
The paper is based on a quantitative analysis of roughly 140 cities, with measures of local arts
district, research universities, employment related to arts and media arts, and media arts-related
patenting over the years. We define economic development mainly through two dependent
variables: (1) employment in arts and media arts, and (2) innovation, measured by total patents
and by the share of those patents related to media art. Our independent variables are the local
presence of cultural districts and the number of “research one (R1)” universities.
The preliminary research design starts with a sample of the 100 largest cities in the United
States, ranging from New York, NY to Spokane, WA. The cities in Frost-Kumpf (1998)’s
inventory of 89 cities with cultural districts, which were not already among the top 100 cities by
9
population, are added to this sample. This leaves 148 cities in the sample, although employment
data are not available for some smaller towns. Patent data (at various points in time) are
measured based on whether the patent assignee’s address matches that city or not. 1 Employment
data taken from the Current Population Survey (CPS) at various years (always the month of May
for any given year) are matched to the city based on the respondents’ Metropolitan Statistical
Area (MSA) for 1999 and earlier and based on the respondents’ Core Based Statistical Area
(CBSA) for 2006 and later. This difference is entirely the result of changes in the geography
reported in the public-use CPS data files. As geographic areas, MSAs capture the bigger (50,000
people or more) urban areas, which means that some smaller cities in our sample may slip
through the cracks for earlier years. CBSAs are based around urban centers of at least 10,000
people and adjacent areas that are socioeconomically tied to the urban center by commuting.
Thus, the CBSA captures micropolitan areas and is more inclusive than the match by CBSAs.
To account for the larger geography in measuring employment, the R1 university counts are
measured alternatively based on location inside the same municipality, within that municipality’s
MSA, and within that municipality’s CBSA. The Frost-Kumpf inventory is used to measure the
presence of cultural districts in each municipality (circa 1998).
Summary data sources
1. A list of cities with cultural districts based on Frost-Kumpf (1998).
2. A list of the number of research extensive university in each city. This definition is drawn
using the Carnegie foundation Research I – Doctoral/Research Universities-Extensive
1
For the cities in the sample, a dictionary of alternate and misspelled city names were used in the matching.
Inaccuracies are common in the NBER patent database, and our approach to manually create city synonyms for our
sample of cities addresses this. Cities like New York has 19 separate spellings (e.g., “N.Y.”, “New York City”,
“New Yotk”) while Milwaukee had 13 (e.g., “Mulwaukee” and “Milwaukie”).
10
definition (Carnegie foundation for the advancement of teaching, 2012). Hence, these
institutions are referred to in the paper as R1 universities. These institutions typically
offer a wide range of baccalaureate programs, and they are committed to graduate
education through the doctorate. During the period studied, they awarded 50 or more
doctoral degrees per year across at least 15 disciplines.
3. Employment data from the Current Population Survey, matched using the metro area of
the respondents, up to 2011. Employment data is drawn from four categories: Arts-related
Industries, and Media Arts-related Industries, Arts-related Occupations, and Media Artsrelated Occupations. When available in the CPS, for recent years, we count a respondent
as employed in one of these categories if their primary or secondary job is coded as such.
Since our list of arts districts reflects the establishment of cultural districts in or before
1998, the analysis typically uses 1999 as the starting year.2 Our end year depends on data
availability. For patent data, the NBER files only reach 2006. The employment data
from the CPS cover as recently as 2011. We also use the 2006 CPS employment data for
comparison to the patent data.
4. Patents and media art patents are available from the NBER data base 1989 - 2006.
Employment definitions
The broader field of “the arts” and newer fields like media arts do not map particularly well onto
the industry and occupation categorizations used in the CPS. Further, classifications of
employment responses changed in the CPS over the timespan of the data. For 1989, 1990, and
2
For whatever reason, the May 1999 CPS data include far fewer MSAs that match to the sample of cities here than
in the other years.
11
1999, the analysis uses eleven categories for arts-related industries. 3 It uses eleven categories for
media arts-related industries, some of which overlap with arts industries.4 For 2006 and 2011,
the arts industry codes5 and media arts industry codes6 changed in the CPS. Similarly, for
occupation codes, several categories were combined for arts occupations 7 and for media arts
occupations8, some of which overlap, for earlier years of the CPS. Recent years required recategorizations for arts occupation codes9 and for media arts occupation codes10. (The research
3
These are: glass and glass products; cement, concrete, gypsum, and plaster products; structural clay products;
pottery and related products; fabricated structural metal products; photographic equipment and supplies; radio and
television broadcasting and cable; theatre and motion pictures; miscellaneous entertainment; museum, galleries, and
zoos; engineering, architecture, and surveying services.
4
These are: newspapers publishing and printing; printing publishing and allied industries; paints, varnishes, and
related products; photographic equipment and supplies; radio and television broadcasting and cable; paper and paper
products; advertising; computer and data processing services; theatre and motion pictures; miscellaneous
entertainment; museum, galleries, and zoos.
5
These are: printing and related support activities; pottery, ceramics, and related product manufacturing; structural
clay product manufacturing; glass and glass product manufacturing; cement, concrete, lime, and gypsum products
manufacturing; misc. nonmetallic product manufacturing; aluminum production and processing; cutlery and hand
tool manufacturing; coating, engraving, heat treated and allied activities; motion pictures and video industries; sound
recording industries; radio and TV broadcasting and cable; internet publishing and broadcasting; architectural,
engineering and related services; specialized design services; computer system design and related services;
advertising and related services; independent artists, performing art, spectator sports, and related industries;
museums, art galleries, and similar institutions.
6
These are: newspaper publishing; publishing except newspapers; software publishing; motion pictures and video
industries; sound recording industries; radio and TV broadcasting and cable; internet publishing and broadcasting;
specialized design services; computer system design and related services; advertising and related services;
independent artists, performing art, spectator sports, and related industries; museums, art galleries, and similar
institutions.
7
These are: buys, wholesale and retail trade; purchasing agents and buyers; architects; arts, drama, and music
teachers; authors; technical writers; designers; musicians and composers; actors and director; painters, sculptors,
craft-artists, and artists paintmakers; photographers; dancer; artists, performers, and related workers; editors and
reports; tool and die makers; tool and die apprentices; precision assemblers; patternmakers and model makers,
wood; hand molders; patternmakers jay out workers cutters; hand cutting and trimming; hand molding, casting, and
forming; hand painting, coating; hand engraving and printing; miscellaneous hand working; production helpers.
8
These are: designers; musicians and composers; actors and director; painters, sculptors, craft-artists, and artists
paintmakers; photographers; artists, performers, and related workers; editors and reports; computer operators;
classified ad clerks; production helpers.
9
computer and information systems managers; agents and business managers of artists, performers, and athletes;
purchasing agents, except wholesale, retails, and farm products; architects; artists and related workers; designers;
producers and directors; dancers and choreographers; musicians, singers, and related workers; entertainers and
performers; announcers; editors; technical writers; writers and authors; misc media; broadcast and sound;
photographers; television, video, and motion picture camera operators; media and communication equipment;
motion picture projectionists; misc personal appearance workers; extruding and drawing machine setters; forging
machine setters; rolling machine setters; tool and die makers; heat treating equipment; layout workers, metal and
plastic; printing machine operators; model makers and pattern makers wood; photographic process workers.
10
computer and information systems managers; agents and business managers of artists, performers, and athletes;
purchasing agents, except wholesale, retails, and farm products; computer programmers; computer software
12
design here, by measuring arts-related employment as a percent of total employment, should
avoid most of the problems associated with the re-categorization, because the re-categorization
affects all cities or observations equally and simultaneously.)
Patent definitions
Like employment, the field of new media arts does not map well onto existing technology
classifications. Media arts-related patents are classified here under media arts according to their
description. 11 When considering which patents relate to media arts, this paper used the widest
and most inclusive definition possible. 12
Descriptives
Table 1 depicts some descriptive statistics from our sample. While the arts and media arts
appear occupy small shares of overall employment and patenting in the sample, they are not so
small as to be undetectable. The R1 variables are significantly but imperfectly correlated with
each other: R1inCity and R1inCBSA have a ρ=0.46, R1inCity and R1inMSA have a ρ=0.66.
There is considerable variation in the combinations of R1 universities and arts districts across
cities in the sample. This is evident in the cross-tabulation in Table 2.
engineers; computer support specialists; artists and related workers; producers and directors; dancers and
choreographers; musicians, singers, and related workers; entertainers and performers; announcers; public relations
specialists; editors; technical writers; writers and authors; misc media; broadcast and sound; photographers;
television, video, and motion picture camera operators; media and communication equipment; motion picture
projectionists; hairdressers, hairstylists; misc personal appearance workers; printing machine operators;
photographic process workers; helpers- production workers; production workers - all others.
11
This includes: A63J, B41F, B41G, B41M, B42D, B44B, B44C, B44D, F21D, G03B, G03C, G04F, G06F, G06K,
G06Q, G06T, G10B, G10C, G10D, G10F, G10G, G10H, G10K, G10L, H04B, H04H, H05K, H04N, H04R, H04S.
12
For patenting classifications, industry codes, and occupation codes, alternate (more restrictive and more
expansive) coding was experimented with, and the results do not differ substantially.
13
Table 1: Descriptive Statistics
Variable
Definition
top100
1 if metro-area is in the top 100 city pop. List;
0 otherwise
Frost
1 if city listed in Frost-Kumpf's catalog of
cultural districts; 0 otherwise
R1inCity
count of R1 universities in the city
R1inMSA
count of R1 universities in the city’s MSA
R1inCBSA
count of R1 universities in the city’s CBSA
patsMA99
# MA patents in '99
patpMA06
% of patents in MA in '06
patsMA06
# MA patents in '06
patpMA9906 % of patents in MA in '06 - % in '99
patsMA9906 # of patents in MA in '06 - # in '99
ArtsInd99
% employed in Arts industry in '99
ArtsOcc99
% employed in Arts occupation in '99
MAInd99
% employed in MA industry in '99
MAOcc99
% employed in MA occupation in '99
ArtsInd06
% employed in Arts industry in '06
ArtsOcc06
% employed in Arts occupation in '06
MAInd06
% employed in MA industry in '06
MAOcc06
% employed in MA occupation in '06
ArtsInd11
% employed in Arts industry in '11
ArtsOcc11
% employed in Arts occupation in '11
MAInd11
% employed in MA industry in '11
MAOcc11
% employed in MA occupation in '11
ArtsInd9911 % employed in Arts industry in '11 - % in '99
ArtsOcc9911 % employed in Arts occupation in '11 - % in
'99
MAInd9911
% employed in MA industry in '11 - % in '99
MAOcc9911 % employed in MA occupation in '11 - % in
'99
Obs
Mean
148
Std.
Min
Max
Dev.
0.676
0.470
0
1
148
0.601
0.491
0
1
148
148
148
148
139
148
133
148
79
79
79
79
126
126
126
126
126
126
126
126
121
121
1.047
1.459
0.554
3.027
0.021
5.365
0.006
4.047
0.019
0.010
0.033
0.010
0.029
0.015
0.024
0.024
0.030
0.015
0.024
0.020
0.010
0.004
1.486
2.094
0.950
9.000
0.034
17.540
0.040
16.345
0.011
0.006
0.016
0.006
0.012
0.008
0.011
0.013
0.014
0.008
0.012
0.011
0.016
0.008
0
0
0
0
0
0
-0.13
-7
0
0
0
0
0
0
0
0
0
0
0
0
-0.04
-0.02
8
9
6
72
0.2
154
0.2
144
0.08
0.03
0.09
0.03
0.06
0.05
0.06
0.09
0.06
0.04
0.05
0.06
0.028
0.013
121
121
-0.010
0.010
0.015
0.012
-0.06
-0.03
0.042
0.020
Note: “MA” represents “media arts-related” in this table.
14
Table 2: Cross-Tabs of “Treatment” Variables
R1inCity
0
1
2
3
4
5
6
total
No
District
40
17
0
1
1
0
0
59
District
51
27
8
0
1
1
1
89
Total
91
44
8
1
2
1
1
148
Identification approach
The analysis here seeks to measure the impact of cultural districts and R1 universities – analogs
to “treatment variables” in this context. The simplest version of the model looks like:
(1)
Y = α + βDISTRICTS + δUNIVERSITIES + ε
where Y is a measure of (media) arts innovation or (media) arts employment. The analysis that
follows uses four separate measures for arts employment: two for arts (employment in an artsrelated industry and employment in an arts-related occupation) and two for media arts
(employment in a media arts-related industry and employment in a media arts-related
occupation). Just one, the share of all patents related to media-arts, is used for patenting.
One problem with estimating equation (1) and drawing inferences about the impact of arts
districts or universities arises if there is some other force or factor that affects everything in the
city, not just media arts employment or media arts innovation. Similarly, a problem arises if it is
just a matter of scale, and cities of certain sizes tend to have more or less arts-related innovation
or employment. To address this issue, the Y variables are measured as “shares of overall
innovation” or “shares over total employment.” In this way, we are interpreting our impacts of
15
districts and universities as impacts on media arts relative to the local economy as a whole. This
approach implicitly controls for factors like the Great Recession, for example, with are large
external shocks that affect the entire local economy. Only if that shock affected media arts
innovation or employment disproportionately and that shock was correlated with cities with
districts or universities would the results be biased. Similarly, measuring everything in terms of
“shares” or percentage of overall employment or patenting controls for scale effects (e.g.,
innovators disproportionately congregate in bigger cities) because everything is measured
relative to non-arts related activity in that city.
Even that might not be enough. A second concern is that the share of arts or media-arts activity
(e.g., employment, patenting) is correlated with something else unobserved about the city that is
also correlated with the location of arts districts or universities there. Not addressing this could
bias the results. For instance, a city might have a long history of an active artist community, or it
might have long had strong arts departments at local universities. In those cases, high levels of
media arts employment or patenting might appear to be due to the presence of an arts district or
university when, in fact, it is really due to the latent artistic strength of the town. Estimating
equation (1) in this situation would lead to upwardly biased estimates of the impact of districts
and universities on media arts activity. To mitigate this problem, a first-differenced approach is
also taken. Because our “treatment” variables are time-invariant – the presence of districts and
universities is constant throughout time in our data – then as a practical matter we can only
estimate the impact of districts and universities on the trends in Y. To see this algebraically,
consider a more flexible model that incorporates a temporal dimension with subscript t:
(2)
Yt = αt + βtDISTRICTSt + δtUNIVERSITIESt + εt
16
(3)
Yt+1 = αt+1 + βt+1DISTRICTSt+1 + δt+1UNIVERSITIESt+1 + εt+1
Comparing equation (2) to equation (1) shows how there is now a time-varying effect (β, δ)
effect of universities and districts. Note that DISTRICTSt = DISTRICTSt+1 because it is timeinvariant, as is UNIVERSITIES. Take equation (2) and subtract from it equation (3), use ∆ to
represent the difference between values in period t+1 and in period t (i.e., the trend), and write
this as equation (4):
(4)
∆Y = ∆α+βt+1DISTRICTSt+1–βtDISTRICTSt+δt+1UNIVERSITIESt+1–δtUNIVERSITIESt +∆ε
This can be rewritten as equation (5), dropping the subscripts t for the time-invariant variables:
(5)
∆Y = αt + ∆βDISTRICTS + ∆δUNIVERSITIES + ∆ε
Equation (5) closely resembles equation (2) in that the regressors are the same; the difference is
in the interpretation. When estimating the model for the trends in Y, employment and patenting,
the effects (coefficients) of districts and universities are interpreted as the changing impact of
those treatments on media arts activity. The trends models show how having an arts district or
R1 universities alters the trajectory of media arts-related employment and innovation. In a sense,
this controls for any (time-invariant) latent, underlying differences across the cities that might
affect the level of media arts employment or innovation. Instead, the trends models show how
the impact of these local economic development engines has changed over time. If cities with
cultural districts, for instance, have seen their media arts employment (as a share of total
employment) rise faster than it has in cities lacking arts districts, this approach will identify that
effect.
Impacts on media arts employment or patenting that also affect all other employment or
patenting will not be detected here; it’s only differential effects. And impacts on media arts
17
employment or patenting that might be due to some intrinsic characteristics of the cities with arts
districts (or more universities) will not be detected in the trends analysis – only how the
performance of those cities has changed differentially over time relative to cities without arts
districts (or universities). Of course, problems remain if (a) there is some other force that
happened to affect media-arts related employment or innovation more than other sectors and (b)
that force’s impact changed over time and (c) tended to affect cities with cultural districts (or
more universities).
Research Findings
Employment
This section reviews the effect of arts districts and R1 universities on art and media art
employment. We start with the share of total employment for arts and media arts jobs, followed
by employment trends, and lastly examining the interaction effects of R1 universities and
cultural districts on arts and media arts employment. The basic models (essentially, equations
(1) and (5) above) are all estimated using OLS with Huber-White robust standard errors. As the
R2 statistics indicate, most of the models perform modestly overall and a few of them offer
extremely limited explanatory power. Overall, arts districts and research universities clearly
explain little of the variation across cities in arts and media arts economic activities.
The first finding from the analysis shows that arts districts are associated with more arts and
media arts employment in 2006, across all four employment of measures. These results shown in
Table 3 are robust to a variety of alternative specifications (and also for levels of employment in
2011). These results support existing studies that claim culture and arts in particular to have a
18
direct impact on local economies. In this case we see that the establishment of arts districts by
the late 1990s had a positive impact on employment in arts and media arts in their respective
cities.13
There does not appear to be an impact on arts and media arts-related employment in 2006 of the
number of R1 universities in the city. That said, R1s in the broader urban area (i.e., CBSA) have
a consistently positive effect on both arts and media-arts employment in 2006. Although these
results on their face seems to contradict studies that view universities as an important factor in
economic growth, this may not be the case. First, research-intensive universities do not
necessarily a focus on arts or media arts. Hence, these results should not be that surprising.
Moreover, expanding the geographical level of analysis from the city to the CBSA changes the
results and hence, implies that artists do not concentrate in the same city as R1 universities but
rather in the surrounding urban area.
Table 3 – Employment shares, 2006
dependent
variable
variable
District
R1inCity
R1inCBSA
constant
N
R2
Arts
Industry
coef. (t)
0.0041 **
2.01
-0.0008
-0.87
0.0021 ***
4.25
0.0242 ***
14.66
126
0.1326
Arts
Occupations
coef. (t)
0.002395
1.62
-0.00044
-0.61
0.001091 ***
3.33
0.01237 ***
12.07
126
0.0761
Media Arts
Industry
coef. (t)
0.004567 **
2.36
-0.00023
-0.21
0.002141 ***
4.63
0.018783 ***
12.42
126
0.1648
Media Arts
Occupations
coef. (t)
0.005087 **
2.26
-0.00106
-1.08
0.001269 ***
2.95
0.019739 ***
13.42
126
0.0692
13
Interestingly, the same model estimated for 1990 data indicates no significant relationship at all between arts
districts and arts-related employment. This is also true for 1989 data. This suggests that arts districts did not just
develop in cities with greater arts employment.
19
To make clear the impact of districts and universities on employment trends and relative growth
of the arts or media arts sector, the analysis turns to a differenced model. Importantly, the results
in Table 4 show that there is no effect of R1 universities or arts districts in the city on arts or
media-arts employment trends in the urban area between 1999-2011. That said, R1 universities
in the broader urban area (CBSA) have a fairly consistently positive effect on arts and media-arts
employment trends. These result appear consistent with arts districts providing a temporary
boost to arts- and media arts-related employment during the mid-2000s and arts-related
employment not clustering in the same cities as the R1 universities that do seem to modest
promote their growth. Given that the sample average of city’s growth in arts-related industry
employment shares is only +1%, the small effect of +0.3% in Table 4 is relatively important.
Table 4 – Employment Trends, 1999 – 2011
dependent
variable
variable
District
R1inCity
R1inCBSA
constant
N
R2
Arts
Industry
coef. (t)
0.0029
0.75
0.0015
1.24
0.0030 ***
4.65
0.0030
0.99
72
0.1568
Arts
Occupations
coef. (t)
-0.0005
-0.35
-0.0005
-0.79
0.0008 *
1.72
0.0033 ***
3.13
72
0.0544
Media Arts
Industry
coef. (t)
0.0005
0.15
0.0002
0.15
0.0007
0.45
-0.0115 ***
-3.58
72
0.0093
Media Arts
Occupations
coef. (t)
-0.0011
-0.45
-0.0010
-1.25
0.0014 ***
2.93
0.0088 ***
6.2
72
0.0594
Table 5 shows the results of the analysis examining the interaction effect between arts districts
and R1 universities when predicting employment shares in 2006. In general, we find that there is
20
minimal significant interaction effects. Arts districts do not have stronger employment effects if
they are in cities with more R1 universities. There is some evidence, however, that the positive
effect of R1 universities in the broader urban area (CBSA) on employment shares in arts- and
media arts-related industries shares is significantly reduced in cities that have arts districts.
Table 5 – Employment Shares, 2006 – Interaction Effects
dependent
variable
variable
District
R1inCity
R1inCBSA
District×
R1inCity
District×
R1inCBSA
constant
N
R2
Arts
Industry
coef. (t)
0.0062
2.19
-0.0016
-1.01
0.0033
3.84
0.0014
0.69
-0.0020
-2.02
0.0229
11.62
126
0.1537
Arts
Occupations
coef. (t)
**
***
**
***
0.0028
1.42
-0.0012
-1.21
0.0015 ***
2.73
0.0012
0.9
-0.0007
-1.05
0.0121 ***
11.16
126
0.0824
Media Arts
Industry
coef. (t)
0.0072
2.71
0.0003
0.16
0.0032
3.93
-0.0005
-0.23
-0.0017
-1.84
0.0172
9.74
126
0.1861
Media Arts
Occupations
coef. (t)
***
***
*
***
0.0064 **
2.17
-0.0012
-0.9
0.0019 ***
2.99
0.0003
0.17
-0.0011
-1.33
0.0189 ***
11.36
126
0.0748
The results of the interaction models for employment trends from 1999-2011, evident in Table 6,
indicate no significant interaction effects overall. Trends in arts- and media arts-related
employment shares in this sample of cities appear unrelated to the joint presence of arts districts
and R1 universities. Any synergy that is there does not manifest in trends in employment shares
either within the city or in the broader metropolitan area.
21
Table 6 – Employment Trends, 1999–2011 – Interaction Effects
dependent
variable
variable
District
Arts
Industry
coef. (t)
R1inCity
R1inCBSA
District×
R1inCity
District×
R1inCBSA
constant
N
R2
0.0004
0.07
-0.0003
-0.15
0.0026 ***
3.14
0.0033
1.21
0.0007
0.55
0.0042
1.21
72
0.1683
Arts
Occupations
coef. (t)
-0.0001
-0.05
0.0002
0.23
0.0007
1.57
-0.0012
-1.04
0.0001
0.17
0.0031 ***
2.87
72
0.0612
Media Arts
Industry
coef. (t)
-0.0024
-0.49
0.0004
0.28
-0.0007
-0.76
0.0002
0.07
0.0022
0.88
-0.0099 ***
-2.84
72
0.027
Media Arts
Occupations
coef. (t)
-0.0013
-0.37
0.0005
0.61
0.0007
1.5
-0.0024
-1.61
0.0010
1.26
0.0089 ***
5.85
72
0.0709
Robustness Tests
The employment section results involve an analysis of the trends from 1999 – 2011, instead of
1990 – 2011. The 1999 – 2011 analysis has the advantage that it is the best date range to detect
an effect of arts districts, because our list/inventory was done in 1998. It has the disadvantage of
the CPS sample drawn from May 1999 that, for whatever reason, only has about two thirds of the
sample of cities in it, so the sample is inconveniently small. An analysis of 1990 – 2011 trends,
which has a larger sample, suffers from the problem that the arts district “treatment” might not
have occurred until well after 1990, thus likely biasing the resulting effects toward zero (i.e.,
attenuation bias).
Hence, when extending the timeframe of the analysis to start at 1990, the results changes
substantially. For media arts-industry employment or for arts occupations there is no significant
effect of arts districts or of R1 universities. For media arts occupations, there is a negative effect
22
of R1s in the city (and no significant effect of R1s in the CBSA or arts districts). Only for arts
industry employment trends are the results for 1990-2011 similar to those in Table 4.
Another robustness check performed here considers the possibility that the District variable is
endogenous in equation (1) or (5). To assess this possibility, District is instrumented for using
the 1989 levels of arts- and media arts-related employment, the share of patents in 1989 for
media arts-related classifications, and the share of all patents granted in 1989 that were assigned
to universities. The premise here is that the 1989 arts-related employment shares and the patent
share variables do not belong in either equation (1) or (5) (i.e., they do not predict arts
employment shares in 2006 condition on the presence of arts districts and the count of R1
universities), and that they also predict the likelihood of an arts district being in that city.
Estimating the various models in Tables 3-6 via instrumental variables reveals two consistent
results. First, the exclusion restrictions appear valid – the proposed instruments indeed do not
belong in the equation of interest. Second, the instruments perform very poorly as the 1989
levels of arts-related employment and innovation are only weakly correlated with the presence of
an arts district. This second result is rather surprising, but, if anything, seems to suggest that the
arts district “treatment variable” may be sufficiently idiosyncratic as to resemble an exogenous
treatment.
Innovation and Patents
This section analyzes the impact of R1 universities and arts districts, as well as their interaction,
on media arts patents. We start with an analysis of their impact on the share of media arts
23
patents, followed by the impact on total numbers, impact on the share of patents over time, and
the trend in total number of patents.
Examining the share of media arts patents in 1999 we find no correlations with the existence of
an arts district or a research one university in the city. Examining the patenting share of media
arts in 2006, however, we find that towns with arts districts have higher shares of media arts
patents. Table 7 reports the results for 2006 media arts patenting. The analysis did not find any
university impact on the share of patenting. The differences between the 1999 and the 2006
results could be due to the possibility that arts districts’ impacts on innovation were not felt until
years after 1998 (the date of the Frost-Kumpf inventory).
Table 7 – Patent Shares and Counts, 2006
dependent
variable
variable
District
R1inCity
R1inCBSA
District×
R1inCity
District×
R1inCBSA
constant
N
R2
Media Arts
patent share
coef. (t)
0.0182 ***
3.42
0.0005
0.17
0.0012
0.63
0.0077 ***
3.10
139
0.0803
Media Arts
patent share
coef. (t)
0.0186 ***
2.89
0.0016
0.62
0.0011
0.83
-0.0014
-0.28
0.0002
0.07
0.0074 ***
3.21
139
0.0806
Media Arts
patent count
coef. (t)
4.8533 **
2.34
5.3325 ***
3.18
Media Arts
patent count
coef. (t)
2.9130
1.65
2.2066 ***
3.74
4.1520 *
1.74
-0.5081
-0.59
148
0.1118
0.7634
1.39
148
0.121
An analysis of the total number of media arts patents granted in 2006 indicates that cities with
arts districts and R1 universities are associated with more media arts patents. (See the third
column of results in Table 7.) The effect of arts districts is +4.9 more media arts patents in that
24
city in 2006 and the effect of additional R1 universities in the same city is +5.3 more media arts
patents in 2006. When you allow for districts and universities to have an interaction effect,
however, then we find arts districts and universities have weaker individual effects while their
interaction becomes the more impacting variable. (See the fourth column of results in Table 7.)
This result is consistent with studies that view both culture and universities as contributing but
not sufficient factors in local economic growth. Hence, their combined effect is larger than their
direct contribution. That universities positively affect the count of media arts-related patents but
not the share of patenting for the media arts is consistent with the role of universities as
promoting more patenting overall. As total patent counts rise, so does the media arts-related
patent subset, but the growth rate for media-arts patents is not differentially larger or smaller than
other technology classes.
Examining the share in media arts patenting over time, we find that arts districts lead to a steeper
trajectory in the share of media arts patents. (See Table 8.) The share of media arts patents rose
from 1999 to 2006, on average, by two percentage points more than in cities lacking arts
districts. Contrast that effect with the sample average media arts patent growth rate of 0.6%.
Arts districts are associated with higher shares of media arts patenting in 2006 and also with
faster growth rates in the share of patenting that goes to media arts. We find that the count of R1
universities does not lead to a rise in share of patents for media arts. The second column of
results in Table 8 shows that there is no significant interaction effect between districts and
universities when it comes to patenting trends. Overall, the results in Tables 7 and 8 support
existing studies that claim that arts districts contribute to the productivity of industry and firms in
the region. It appears that arts districts promote innovation in media arts in particular.
25
Table 8 – Trends in media arts patenting, shares and counts, 1999-2006
dependent
variable
variable
District
R1inCity
R1inCBSA
District×
R1inCity
District×
R1inCBSA
constant
N
R2
Media Arts
patent share
coef. (t)
0.0192 ***
2.87
0.0015
0.4
-0.0010
-0.47
-0.0042
-1.2
133
0.06
Media Arts
patent share
coef. (t)
0.0152 *
1.88
0.0045
0.85
-0.0043
-1.11
-0.0054
-0.77
0.0051
1.11
-0.0014
-0.42
133
0.0709
Media Arts
patent count
coef. (t)
3.8634 **
2.46
0.8028
0.31
-0.3081
-0.83
0.0195
0.02
148
0.0222
Media Arts
patent count
coef. (t)
4.2651 **
2.35
0.5291
0.56
0.0063
0.04
0.4211
0.12
-0.4442
-0.78
-0.2742
-0.48
148
0.023
Examining the trends in the total number of media arts patents between 1999 and 2006 , we find
that cities with more R1 universities do not produce more media arts-related patents. As for
patenting shares, however, the presence of an arts districts does have a positive effect on media
arts patenting trends. Having both more R1 universities and an arts district does not have a
differential impact on media arts patenting. The consistent result is that arts districts have a much
stronger role on media arts-related patenting, although universities do impact patenting in
overall. Perhaps because patenting universities do not focus on media arts limits their impact in
this sector.
Discussion and Conclusions
Overall, in this simple model, R1 universities in the urban area (rather than in the same city
itself) seem to have the strongest effects on arts- and media arts-related employment levels.
26
Cities with cultural districts had higher arts employment shares in 2006, but that effect is not
reflected in longer-term trends in employment shares. There, only R1s in the CBSA seem to
drive arts- and media arts-related employment growth. With respect to patenting, the story is
even more clear: cultural districts promote more media arts patenting and faster growth in
innovation in that sector, R1 universities typically do not differentially favor media arts
patenting. Further, there do not appear to be consistently significant interaction effects between
districts and universities. Perhaps unintuitively, universities seem to promote more arts-related
employment (rather than media arts patenting and innovation) whereas cultural districts seem to
promote more media arts patenting and innovation (rather than arts-related employment). A
plausible story, worth further and future investigation, is that universities play a major role in
developing arts and cultural industries by attracting and training artists and arts-related labor,
while cultural districts attract innovative entrepreneurs in the related field of media arts.
These results strengthen studies regarding the role of culture and universities in economic
growth. We find that arts districts and universities impact employment levels. The results
indicate that the impact of arts district, however, is confined to a short period of time. Moreover,
the impact of R1 universities implies that artists do not live in the same city as the R1 university
and hence suggest a need to evaluate the relationships between the geographical location of the
R1 university and employed artists.
These results also conform to cultural studies literature and indicate that culture has a direct
positive connection to economic development. Hence, art and culture impacts economic growth
both directly through employment and also indirectly through innovation. Universities’ impact
27
specifically in media arts, on the other hand, may be more limited to promoting innovation
(patenting) more generally and attracting or training labor for arts and media arts jobs.
Future Work
Future work includes several extensions of this preliminary investigation. First, we are
collecting information specific arts- (and media arts-) related degree programs at universities.
This will allow us to test whether more specific or targeted university efforts impact media arts
employment and innovation. In parallel, we are also exploring richer measures of cultural
clusters, beyond the inventory of cultural districts. This includes controlling for the age and size
of the cultural districts as well as using Strom’s (2002) inventory of 71 major cultural facilities
being built or renovated between 1985-2005. Second, we will extend the analysis to include
other years of CPS (employment) and patent data, overcoming some of the current limitations.
Third, we will carry out a qualitative investigation of the top- and bottom-performing cities in
terms of media arts employment and innovation. This will allow a richer exploration of the
relationship between universities, cultural districts, and the arts and media arts sectors. Fourth,
using more geospatially refined data, we intend to investigate more localized impacts of cultural
districts on employment patterns.
28
References
BIANCHINI, F., FISHER, M., MONTGOMERY, J. & WORPOLE, K. 1988. City Centres, City
Cultures: The Role of the Arts in the revitalisation of Towns and Cities, Manchester, UK,
Center for Local Economic Development Strategies.
BREZNITZ, S. M. & FELDMAN, M. P. 2012. The Engaged University. The Journal of
Technology Transfer, 37, 139-157.
CARNEGIE FOUNDATION FOR THE ADVANCEMENT OF TEACHING. 2012. 1994
Edition Data File [Online]. Available:
http://classifications.carnegiefoundation.org/resources/ [Accessed 2012].
DI GREGORIO, D. & SHANE, S. 2003. Why do some universities generate more start-ups than
others? Research Policy, 32, 209-227.
EATON, S. C. & BAILYN, L. 1999. Work and life strategies of professionals in biotechnology
firms. Annalas. American Academy of Political and Social Science, 562:, 159-173.
ETZKOWITZ, H. & LEYDESDORFF, L. 1997. Universities and the Global Knowledge
Economy: A Triple Helix of University-Industry-Government Relations, London, UK,
Pinter.
FELDMAN, M. P. 1994. The Geography of innovation, Kluwer Academic Publishers.
FELDMAN, M. P. & BREZNITZ, S. M. 2009. The American Experience in University
Technology Transfer. In: MCKELVEY, M. & HOLMÉN, M. (eds.) European
Universities Learning to Compete: From Social Institutions to Knowledge Business.
Edward Elgar.
FLORIDA, R. 2002. The Rise of the Creative Class, New York, NY, Basic Books.
FORRANT, R., PYLE, J., LAZONICK, W. & LEVENSTEIN, C. (eds.) 2001. Approaches to
Sustainable Regional Development: The Public University in the Regional Economy:
University of Massachusetts.
FROST-KUMPF, H. A. 1998. Cultural Districts: The Arts As a Strategy for Revitalizing Our
Cities Washington, DC: Americans for the Arts.
GLEASER, E. 2004. Review of The Rise of the Creative Class by Richard Florida [Online].
Available: http://www.economics.harvard.edu/faculty/glaeser/files/Review_Florida.pdf
[Accessed].
GODDARRD, J. & CHATTERTON, P. 1999. Regional development agencies and the
knowledge economy: Harnessing the potential of universities. Environment and Planning
C: Government and Policy, 17:, 685-699.
JAFFE, A. B., TRAJTENBERG, M. & HENDERSON, R. 1993. Geographic localization of
knowledge spillovers as evidenced by patent citations. The Quarterly Journal of
Economics, 108, 577-598.
LANDRY, C., BIANCHINI, F., EBERT, R., GNAD, F. & KUNZMAN, K. 1996. The Creative
City in Britain and Germany.
LAWTON SMITH, H. 2000. Technology Transfer and Industrial Change in Europe, Macmillan
LERNER, J. 2004. The University and the Start-Up:Lessons from the Past Two Decades. The
Journal of Technology Transfer, 30, 49-56.
LINK, A. & SIEGEL, D. 2005. Generating science-based growth: an econometric analysis of the
impact of organizational incentives on university-industry technology transfer. European
Journal of Finance, 11, 169-181.
29
MARKMAN, G. D., GIANIODIS, P. T., PHAN, P. H. & BALKIN, D. B. 2005. Innovation
speed: Transferring university technology to market. Research Policy, 34, 1058-1075.
MARKUSEN, A. 2006. Urban Development and the Politics of a Creative Class: Evidence from
the Study of Artists. Environment and Planning A, 38, 1921-1940.
MARKUSEN, A. & GADWA, A. 2010. Arts and Culture in Urban or Regional Planning: A
review and Research Agenda. Journal of Planning Education and Research, 29, 379-391.
MARKUSEN, A. & KING, D. 2003. The Artistic Dividend: the Hidden Contributions of the
Arts to the Regional Economy. Arts Research Monitor.
MARKUSEN, A. & SCHROCK, G. 2006. The Artistic Dividend: Urban Artistic Specialisation
and Economic Development Implication. Urban Studies, 43, 1661-1686.
MAURRASSE, D. J. 2001. Beyond the campus : how colleges and universities form
partnerships with their communities, New York, Routledge.
MINSHALL, T., DRUILHE, C. & PROBERT, D. 2004. The evolution of "Third Mission"
activities at the university of Cambridge: Balancing strategic and operational
considerations. 12th High Tech Small Firms Conference. University of Twente, The
Netherlands: University of Twente.
MOWERY, D., ROSENBERG, R. R., SAMPAT, B. N. & ZIEDONIS, A. A. 1999. The Effects
of the Bayh-Dole Act on U.S. University Research and Technology Transfer. In:
BRANSCOMB, L. M., KODAMA, F. & FLORIDA, R. L. (eds.) Industrializing
knowledge : university-industry linkages in Japan and the United States
Cambridge, Mass.: MIT Press.
NELSEN, L. 2003. The Role of University technology Transfer Operations in Assuring Access
to Medicines and Vaccines in developing Countries. Yale Journal of Health, policy, Law,
and Ethics, III, 301-308.
OWEN-SMITH, J. & POWELL, W. 2001. To Patent or Not: Faculty Decisions and Institutional
Success in Academic Patenting. Journal of Technology Transfer 26, 99-114.
PERRYMAN, R. M. 2000. The Catalyst for Creativity and the Incubator for Progress: The Arts,
Culture, and the Texas Economy The Perryman Group for the Texas Cultural Trust.
PRATT, A. 1997. The Cultural Industries Production System: A case Study of Emplyment
Change in Britain 1984-1991. Environment and Planning A, 29, 1953-1974.
RAHM, D., KIRKLAND, J. & BOZEMAN, B. 2000. University-industry R & D Collaboration
in the United States, the United Kingdom, and Japan, Dordrecht, The Netherlands,
Kluwer Academic Publishers.
SAMPAT, B. N. 2006. Patenting and US academic research in the 20th century: The world
before and after Bayh-Dole. Research Policy, 35, 772-789.
SCOTT, P. 1977. What Future for Higher Education, London, Fabian Tracts.
SIEGEL, D. S., WALDMAN, D. & LINK, A. 2003. Improving the effectiveness of commercial
knowledge transfers from universities to firms. Journal of High Technology Management
Research, 14, 111-133.
SIEGEL, D. S., WALDMAN, D. A., ATWATER, L. E. & LINK, A. N. 2004. Toward a model
of the effective transfer of scientific knowledge from academicians to practitioners:
qualitative evidence from the commercialization of university technologies. Journal of
Engineering and Technology Management, 21, 115-142.
STROM, E. 2002. Converting Pork into Porcelain: Cultural Institutions and Downtown
Development. Urban Affairs Review, 38.
30
THURSBY, J. G., JENSEN, R. & THURSBY, M. C. 2001. Objectives, Characteristics and
Outcomes of University Licensing: A Survey of Major U.S. Universities. The Journal of
Technology Transfer, 26, 59-72.
31