- Email: [email protected]
Leisure choices of the creative class
Cities 41 (2014) 38–43
Contents lists available at ScienceDirect
Cities journal homepage: www.elsevier.com/locate/cities
Leisure choices of the creative class Eric Joseph Van Holm ⇑ Georgia State University & Georgia Institute of Technology, United States
a r t i c l e
i n f o
Article history: Received 16 December 2013 Received in revised form 15 May 2014 Accepted 17 May 2014
Keywords: Creative class Amenity choice Cultural economics City development
a b s t r a c t Richard Florida’s theory of the creative class has led cities to offer amenities as a way to attract the knowledge workers that he argues drive the economy. At present, scant evidence exists concerning the leisure activities different worker classes choose to engage in. Using logit analysis on data from the American Time Use Survey, I analyze the leisure preferences of the creative class and find limited support for Florida’s assertions about their amenity use. Cities and policymakers should consider the evidence here before trying to create unique amenities to attract the creative class, because they may not require distinctive activities to choose a location. Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
Literature review
Cities are constructing bike paths, financing public art, and encouraging diversity, all in an effort to attract knowledge workers. The Rise of the Creative Class by Richard Florida (2002) has become a touchstone for policymakers throughout the United States and Western Europe, and it has offered planners a new path towards prosperity. Traditionally, urban development theory stated that individuals follow jobs, making attracting employers the key to promoting economic growth. Florida (2002) and Clark (2011) flip that traditional narrative arguing that in the knowledge economy jobs follow people. Florida advises cities to build a ‘‘people climate’’, because talented individuals are the new primary resource for a strong economy, meaning educated, creative people will find jobs wherever they locate (Florida, 2002, p. 283). Florida argues that the creative class chooses locations based on high diversity and natural and cultural amenities. He speculated on the cultural amenities the creative class are most likely to enjoy, but was unable to test his hypotheses empirically lamenting that data on the subject was not available. Using data from the American Time Use Survey, released since the publication of the book, I test several of his assertions. First, I will test what leisure activities the creative class is more likely than working and service class employees to enjoy. Secondly, I will assess what other factors explain differences in leisure preferences.
Florida’s theory of the creative class centers upon their attraction to the 3 T’s of technology, talent, and tolerance. That is, Florida posits that regions with a high-acceptance of diversity and alternative life-styles and the right mix of cultural amenities attract individuals with greater talent. These talented individuals, a valuable resource in the modern economy, attract high-technology industries, spurring economic growth. Therefore, rather than the traditional narrative that jobs attract workers to a region, Florida argues that workers attract jobs, placing the emphasis for cities on building a people climate in order to attract high-skilled workers. Florida makes repeated claims about the types of lifestyle amenities that attract the creative class; he concludes from his interviews and focus groups that they enjoy participatory experiences that evolve organically and events that are not prepackaged. Because the creative class place such a high value on time, their entertainment must be continuously engaging; their ‘‘lifestyle comes down to a passionate quest for experience’’ (Florida, 2002, p. 166). Thus, two of Florida’s central claims are that the creative class drives economic development, and that they are a unique demographic. However, both claims have been weakened by later research. Studies have shown traditional measures are better predictors of economic health than Florida’s measures of technology, tolerance and talent (Donegan, Drucker, Goldstein, Lowe, & Malizia, 2008; Hoyman & Faricy, 2009; Sands & Reese, 2008). Furthermore, the creative class is not very distinctive in their residential choice (Andersen, Bugge, Hansen, Isaksen, & Raunio, 2010;
⇑ Corresponding author. Tel.: +1 916 396 7806. E-mail address: [email protected] http://dx.doi.org/10.1016/j.cities.2014.05.006 0264-2751/Ó 2014 Elsevier Ltd. All rights reserved.
E.J. Van Holm / Cities 41 (2014) 38–43
Asheim & Hansen, 2009; Frenkel, Bendit, & Kaplan, 2013; Lawton, Murphy, & Redmond, 2013), mobility (Hansen & Niedomysl, 2009; Martin-Brelot, Grossetti, Eckert, Gritsai, & Kovacs, 2010), or amenity use (Bille, 2010). Bille (2010) was the first to test Florida’s claims about amenity use by the creative class. He compared creative class and service class workers on the use of 35 amenities from a Danish survey. Bille finds that the creative class is more likely than the service class to attend contemporary concerts, to visit art museums and cultural landscapes, to engage regularly in fitness and play sports, and to use the internet than the service class. The creative core, a subset of the creative class, is significantly more likely than the service class to enjoy creative self-expression, to engage in noninstitutional fitness activities, to read a book, and to visit historic sites, museums, and classical art forms. The creative class is neither more nor less likely to watch TV, watch films at home, listen to recorded music, play computer games, read magazines, or go to the zoo, amusement park or aquarium. Bille’s findings agree with Florida in that the creative class uses the internet more, is more likely to exercise, and is less likely to attend spectator sports events. Contrary to Florida’s assertions, however, Bille’s results show that the creative class is more likely than the service class to attend a museum, or scheduled activities like evening class or lectures, and is as likely to watch TV or a movie. Data In 2002, comprehensive data was not available to test amenity use by the creative class or others. Since 2003, the Bureau of Labor Statistics (BLS) has been administering the American Time Use Survey (ATUS) to collect detailed data on how Americans spend their time. The BLS randomly selects participants from the Current Population Survey (CPS) to fill out a time diary for all household members for a specific 24 h period, which it then combines with demographic information from the CPS (Abraham, Flood, Sobek, & Thorn, 2011). I combine three years of the data, (2008–2010), providing a sample size of 18,386 respondents, allowing sufficient observations for even such rare events as attending performing arts. One limitation of the ATUS data for this study is the distinction between revealed and stated preferences. The data used here is of the revealed variety, in that it is activities actually done by respondents. Stated preferences are also important for cities, as individuals use cultural and natural amenities as signals of the type of city or community they would like to reside in, regardless of whether they actually partake in those activities. There has been work on Florida’s theory of the stated preference variety (Frenkel et al., 2013; Lawton et al., 2013) but because Florida discusses amenities as activities the creative class not only uses as signals but participates in, it is appropriate to use revealed preference data. Methodology I am able to test five specific claims made by Richard Florida concerning amenity use by the creative class. ATUS respondents report the number of minutes they spent on each activity in the specified day; with the exception of television watching, I recode each as a dummy variable with 1 for having done that activity because I am uninterested in the duration of any given amenity but rather its use. I divide the dependent variables into four categories of arts, time wasting, exercise, and late night dining. Using the four hypotheses and five dependent variables described below, I will test Florida’s claims that the creative class is less likely to attend performing arts, and watch television and more likely to practice outdoor sports, exercise, or eat out late at night. I use the same model for all dependent variables.
39
H1. The creative class is less likely to enjoy traditional performing arts. ‘‘In many cities recently, museums and the symphony opera and ballet have fallen on hard times . . . Meanwhile, the Creative Class is drawn to more organic and indigenous street-level culture’’ (Florida, 2002, p. 182). Florida argues at length that traditional forms of high-culture, such as the symphony, opera, and ballet are less attractive to members of the Creative Class than past generations. The ATUS asks how much time individuals spent at the performing arts, using examples such as the theatre, opera, or musicals. I code the variable as a 1 if respondents indicate they attended any performing arts; only .8% did so on the survey day. H2. The creative class is less likely to watch television. ‘‘But if you spend your workday in front of a computer screen or an artist’s canvas, you are probably not eager to spend your leisure time in front of a TV screen’’ (Florida, 2012, p. 139). While some may consider an hour of the news as less of a waste of time than an hour of reality television, Florida does not differentiate, rather asserting that television itself does not appeal to the creative class. However, because watching 1 min of television is different from watching all day, I code television watching as 1 if the individual watched more than 4 h of television. 22% of the sample watched more than 4 h of television within the 24-h period used for the survey. H3. The creative class is more likely to exercise ‘‘The force behind the Creative Class obsession with being in shape is more than a concern with health . . . I see it as a growing awareness of the body as an arena for creative expression’’ (Florida, 2002, p. 177). ‘‘Creative Class people in my studies are into a variety of active sports, from traditional ones like bicycling, jogging and kayaking, to newer more extreme ones like trail running and snowboarding’’ (Florida, 2002, p. 173). I code outdoor sports as 1 if the individual did any biking, climbing, hiking, or running; .25% of respondents reported doing these activities. I create the dummy exercise as 1 if they have done aerobics, walked, lifted weights, done yoga or used cardiovascular equipment. By this definition 10% of the sample exercised. H4. The creative class is more likely to eat out late at night. ‘‘They need to have options around the clock. . . the highestrated nightlife options were cultural attractions and late-night dining’’ (Florida, 2002, p. 225). Finally, late night dining is coded to encompass anyone that ate out at a bar or restaurant from 10 p.m. until 2 a.m.; 1% of the sample reported dining out between those hours. Independent variables The Current Population Survey data provides major occupational code for each respondent, which allows me to test my key independent variable creative class. I was able to nearly replicate Florida’s definition of the creative class, which he categorizes as containing the major occupational codes for: business and financial operations; computer and mathematical science; architecture and engineering; life, physical, and social science; community and social service; legal, education, training, and library arts; design,
40
E.J. Van Holm / Cities 41 (2014) 38–43
entertainment, sports and media; healthcare practitioner; and high-end sales and sales management. The only occupation I was unable to include were high-end sales and sales management, because the occupational codes do not differentiate between high-end and low-end sales. I removed all sales from the sample, because placing them in either the creative or service class would create bias. The percentage of workers in the creative class for my sample was 39.3%, higher than Florida’s estimate of 33%. Other than the issue with sales, I was able to replicate perfectly Florida’s definition of the service class: health care support; food preparation and food-service-related; building and grounds cleaning and maintenance; personal care and service; low-end sales; office and administrative support; and protective services. My definition of the working class had no differences from Florida’s, containing construction and extraction; installation, maintenance, and repair; production; and transportation and material moving. Agriculture is the final class of workers, but the final data sample contained few observations for agricultural workers, so I dropped it from the data. My first analysis will use logit analysis to regress the amenities on worker class. Initially, I will test Florida’s hypotheses by finding the relative propensity of the creative class to use each amenity compared to other worker classes. If the creative class is no more likely than other classes to enjoy an amenity, then the government offering it is not truly attracting the creative class, so much as attracting all workers. The bivariate regression is used to establish the importance of preceding and intervening variables in the later analysis. Secondly, I will conduct multivariate analysis to test whether other variables are able to explain the relative propensity of different worker classes to enjoy each activity. It may be that other factors are a bigger determinant of individual’s leisure activities than worker class. I expect a variety of demographic variables to influence amenity choice. A gender dummy is included as well as four dummy variables that define white, African–American, Latino, Asian, and other as mutually exclusive. I coded education, representing the highest grade of education reached for the respondent, divided into four levels of high school, some college, college, or graduate school. Age may have a non-linear relationship to many amenities, so I include a set of 5 dummies capturing varying points in an individual’s life to capture the different leisure choices of different age groups. Income is measured with 5 dummy variables, divided by intervals of $25,000 ending with $100,000 and above. Because a disability may keep an individual from enjoying the amenities tested for, I have included a dummy variable, which will be 1 if any impairment is present. As individuals marry and have children, amenity use may change. Therefore, I control for their marital status, with variables for whether they have never been married, are married, or are divorced or separated, as well as the number of children under 18 in the household. In order to account for regional and locational differences in amenity use, I create two sets of dummy variables. The first indicates what Standard Federal Region the respondent lives in and the second set of dummy variables designates whether the individual lives in the central city of a metropolitan statistical area (MSA), suburbs of an MSA, or not in an MSA. Finally, the day of the week the individual records in their time journal will affect the amenities they use. Seeing the ballet is associated with ‘date nights’ on Friday or Saturday evening, but an individual surveyed about their amenity use for a Tuesday is less likely to attend the performing arts. However, the ATUS does not report what day the 24-h period occurs on or whether it is a weekend or weekday. I include a variable to indicate whether the individual worked more than 4 h on that day to attempt to capture the influence of work on amenity choice.
Bivariate regression findings I display the first sets of regressions in Table 1 as odds-ratios from a logit regression, with creative class as the reference group. In order to provide context for the magnitude of each change in probability, I provide the percentage of the sample doing each activity at the top of Table 1. The row for Florida’s hypothesis indicates the expected direction of the relationship, moving from the creative class to the working or service class. The odds that a member of the service or working class attends the performing arts were only 71% and 30% as high as the odds for someone from the creative class respectively. Likewise, the creative class has greater odds of practicing outdoor sports or performing traditional forms of exercise than both the working and service class. The only dependent variable for which the working and service class had greater odds was watching more than 4 h of television, where they are 45% and 102% more likely than the creative class respectively. Neither class of worker shows a significant difference in regards to eating out late at night. Florida correctly hypothesizes about the creative class in regards to television, outdoor sports, and traditional forms of exercise. However, the coefficients of the other three dependent variables are insignificant or in an unexpected direction meaning that even without controlling for other socioeconomic factors Florida is only accurate on half of his suggestions. To test whether the relationship of worker class with amenity use are created by spurious or indirect relationships, the next set of regression results will identify what other factors may explain the differences in leisure choice.
Multivariate regression findings Table 2 contains results from the multivariate regression as odds-ratios. I tested for multicollinearity for the full model and the variation inflation factors (VIF) for all variables were below five, indicating there was not a problem. Being a member of the creative class has a diminished effect on amenity use when other factors are included in the regression. While none of the signs change direction in the new analysis, four of the eight significant worker class coefficients from the bivariate regression lose statistical significance. Of the four that retain significance, only two are in the predicted direction. The following sections will discuss how the worker class variables reacted to the inclusion of other factors, and what those other variables indicate about the relationship of being a member of the creative class and leisure use. Table 3 displays the relationship of the three worker classes to each of the other independent variables, which will help to identify relationships that affected the significance of worker class coefficients. All variables discussed are significant at .05 or greater, unless otherwise specified.
Performing arts The odds of the working class attending the performing arts are 50% lower than the odds for the creative class, while the difference with the service class is not significant. That this result remains significant despite the inclusion of other factors indicates the creative class has a predilection towards traditional arts that cannot be explained with other socioeconomic factors; however, this result is also in the opposite direction predicted by Florida. The results here generally concur with Bille (2010), who found the creative class more likely to attend classical arts than the service class. In addition, my results connect well with Frenkel et al. (2013) who showed that the creative class prefer to live near
41
E.J. Van Holm / Cities 41 (2014) 38–43 Table 1 Bivariate regression results. Variable
Performing arts
Television
Outdoor sports
Exercise
Late night eating
Creative class (reference group) (%) Service class Working class Florida’s hypothesis
1.02 0.71 0.30 Positive
16.98 1.45 2.04 Positive
3.4 0.56 0.50 Negative
12.47 0.68 0.51 Negative
1.07 0.94 0.73 Negative
cultural attractions including the theater and museums; here I show they are also more likely to partake in similar activities. The loss of significance for the service class is likely connected to ethnicity, income and education. Variables representing Latinos and Asians are also significant and below one, meaning neither ethnic categories is more likely than whites to enjoy the symphony, opera, or ballet. Conversely, an income of $75 and $100 thousand and having a graduate degree, both of which the creative class is more likely to have, increase the odds of attending the performing arts. Lastly, having worked 4 h on the
day the time journal is kept lowers the chances of attending the performing arts. Television Individuals in the creative class are less likely than both the service and working class to watch more than 4 h of television. The service and working class have odds 13% and 19% higher respectively, holding race, age, sex, income, family status, education, and location constant. Bille (2010) found that the creative
Table 2 Multivariate regression results.
*
Performing arts
Television
Outdoor sports
Exercise
Late night eating
Creative class (reference group) (%)
1.02
16.98
3.4
12.47
1.07
Service class Working class
0.93 0.50**
1.13** 1.19***
1.02 0.85
1.01 0.84**
1.13 1.02
Latino Black Asian Other race
0.49* 0.88 0.44 0.55
1.04 1.49*** 0.84 0.94
0.70** 0.44*** 0.71 1.53
1.02 0.85** 1.17 0.71
0.98 0.51** 1.41 0.99
Age 30–50 Age 50–65 Age above 65 Female
1.15 1.18 1.2 1.24
1.38*** 1.61*** 1.67*** 0.55***
0.9 0.62*** 0.57** 0.66***
1.08 1.05 1.35** 1
0.59*** 0.45*** 0.11*** 0.94
25–50k income 50–75k income 75–100k income Income above 100k
1.07 1.57 1.72* 1.35
1.08 0.93 0.88* 0.81***
0.94 1.63*** 1.42* 2.00***
1.12 1.23** 1.31*** 1.36***
1.3 1.21 1.35 1.78**
Married Divorced or separated
1.03 1.24
0.79*** 0.89*
0.79* 0.77
0.99 1.05
0.43*** 0.8
One child Two children Three or more children
0.89 0.75 0.64
0.74*** 0.64*** 0.64***
0.89 0.89 0.78
0.71*** 0.69*** 0.76**
0.99 0.84 0.48*
Some college College Graduate degree
1.42 1.47 1.88**
0.79*** 0.67*** 0.51***
1.2 1.89*** 2.55***
1.31*** 1.75*** 2.39***
1.13 1.59* 1.47
Any handicap
0.95
1.01
0.74
0.71**
0.79
Central city Suburbs
1.33 1.06
1.09 1.09*
1.14 0.94
1.22*** 1.13*
2.09*** 1.44*
Worked
0.39***
0.20***
0.47***
1.14***
0.92
New England New York/New Jersey Mid-Atlantic Southeast North South Plains Mountain West Southwest
1.07 1.65 1.3 0.81 0.98 0.74 1.11 0.51 1.07
0.98 1.07 1.1 1.24** 1.07 1.15 1.12 0.9 1.06
0.57 0.9 0.37*** 0.77 0.84 0.76 0.75 1 1.33
1.08 0.93 0.86 0.72*** 0.85 0.84 0.78* 1.17 1.06
0.86 1.02 0.69 0.81 0.78 0.86 0.67 0.74 0.63
2009 sample 2010 sample
1.31 0.83
0.96 0.92
1.09 1.17
1.17* 1.14
0.75 0.93
McFadden’s R2
0.05
0.12
0.07
0.03
0.05
p < 0.1. p < 0.05. *** p < 0.01. **
**
42
E.J. Van Holm / Cities 41 (2014) 38–43
Table 3 Summary statistics (in %). Variable
Creative class
Service class
Working class
Count Percent
8057 43.8
6370 34.6
3959 21.5
White Latino Black Asian Other race
76.9 7.1 9.6 5.3 1.1
60.7 18 17.7 2.4 1.5
62.6 21.1 13.7 1.8 1.3
18–30 30–50 50–65 Over 65
10.9 56.7 27 5.2
19.7 45.2 26.1 6.4
16.1 50.4 28.1 4.7
Male Female
45.2 54.8
30.5 69.5
82.8 17.2
Under $25,000 $25,000–50,000 $50,000–75,000 $75,000–100,000 Over $100,000
6.1 16.1 19.4 17 33.9
21.9 30.2 18.6 10.7 10.1
19.4 32.4 21.6 9.9 8.4
Never married Married Divorced or separated
19.6 62.8 15
30.6 45.7 19.1
23.4 55.6 18.5
No kids One child Two children Three or more children
45.9 21.9 23.1 25.3
48.9 22.3 18.3 26.5
49.9 19.9 18.9 26.9
High school Some college College Graduate degree
10.5 22.6 38.2 28.7
43.5 36.3 15.3 4.9
60.9 29.4 7.9 1.8
Any handicap
3.2
5.4
5
Worked
27.5
27.9
21.8
Central city Suburbs Rural
45.6 12.3 42.2
39.5 17 39.6
39.4 22.1 41
New England New York/New Jersey Mid-Atlantic Southeast North South Plains Mountain West Southwest Northwest
6.1 8.4 11.7 13.9 17.9 11.2 5.6 4.5 13.1 7.7
4.6 8.3 10.1 15.7 19.5 11.4 5.7 4.1 13.1 7.5
3.5 6.1 10 15.9 21.1 12.3 6.7 4.4 12.2 7.8
class did not differ from the rest of the population with respect to many commonplace amenities, a result that differs from mine although television is the only such activities I test. The creative class is the most likely to be highly educated or have high incomes, which are both significant and negative for television watching. That is, even when controlling for the fact that the creative class is more likely to have finished a postsecondary education and more likely to earn above $75,000, they are still significantly less likely to watch 4 h of television. Exercise Both variables for class of worker lose their significance for outdoor sports as does exercise for the service class. Similar to television watching above, education and income differences are able to explain part of the difference in the results. Income and education have consistently large effects on exercise and outdoor sports, with increasing positive changes the greater the difference from the reference group. In addition, race may be a factor. Working and service class members are more likely to be minorities, the variables for which are mostly significant and negative. Therefore,
in addition to income and education explaining why the creative class is more likely to exercise, race may explain why the service and working class are less likely to do so. Lastly, an individual living in the central city has odds 22% higher for exercise, probably due to the nature of individual’s employment and the availability of gyms. Despite being more likely to live in the city, the creative class is still more probable than the working class to exercise. The results for exercise differ strongly with past analyses, which have found a strong attraction for the creative class to such activities. McGranahan, Wojan, and Lambert (2011) found that the creative class was more likely to live near outdoor sport opportunities and Bille’s (2010) results showed the creative class was more likely to do traditional exercise and walk or cycle than the service class. Late night dining In both regressions there were no significant differences between worker classes in terms of late night dining. The variable that had the most consistent effect on eating around midnight was age, an area that worker classes did not differ significantly in the sample. Central city is also significant, likely due to the availability of restaurants open at any hour. Conclusion I find narrow support for Florida’s claims about how the creative class differs from the service and working class in the amenities it enjoys, as most of Florida’s assertions are not supported by the evidence presented. Even in the bivariate analysis, where the creative class diverged from the service and working class on four of the activities tested, differences on one of those activities were in the opposite direction of what Florida hypothesized. In the bivariate test the creative class was, more likely to go to the performing arts than both the service and working class. Furthermore, when I control for other factors, the majority of differences between worker classes disappear. In the multivariate analysis, the creative class differs from both classes in television watching, and from the working class in the performing arts and exercise. My results are much more pessimistic than Bille’s (2010) study, which found general agreement with the broad claims of Florida even when controlling for other socioeconomic factors. Cities and policymakers should consider the evidence presented before trying to create unique amenities to attract the creative class. The fact that occupation did not make for consistently large changes in amenity preferences for either the bivariate or multivariate analysis indicates that the creative class may not require unique lifestyle options to enjoy a location. Demographics do play a part in determining leisure choices, but Florida’s creative class is the wrong way to understand these differences. The creative class may still enjoy bicycling and kayaking, but no more so than other workers with similar socioeconomic status; if a city endeavors to create a people climate it will be doing so to improve its livability of all residents, not a specific class of workers. The demographics of the roughly average member of the creative class are also worth discussion. A reader could come away from The Rise of the Creative Class with an image of the creative class as young, tattooed, and untraditional. However, nearly 55% of the creative class is between the ages of 35 and 55, the vast majority is married, most are female, and more than half have at least one child. If a city really wants to attract the broadest segment of the creative class, it appears that sound areas to invest in would be safe streets, good schools, and numerous parks in order to create family-friendly amenities.
E.J. Van Holm / Cities 41 (2014) 38–43
References Abraham, K. G., Flood, S. M., Sobek, M., & Thorn, B. (2011). American Time Use Survey data extract system: Version 2.4. Andersen, K. V., Bugge, M. M., Hansen, H. K., Isaksen, A., & Raunio, M. (2010). One size fits all? Applying the creative class thesis onto a Nordic context. European Planning Studies, 18(10), 1591–1609. Asheim, B., & Hansen, H. K. (2009). Knowledge bases, talents, and contexts: On the usefulness of the creative class approach in Sweden. Economic Geography, 85(4), 425–442. Bille, T. (2010). Cool, funky and creative? The creative class and preferences for leisure and culture. International Journal of Cultural Policy, 16(4), 466–496. Clark, T. N. (2011). The city as an entertainment machine. Plymouth, United Kingdom: Lexington Books. Donegan, M., Drucker, J., Goldstein, H., Lowe, N., & Malizia, E. (2008). Which indicators explain metropolitan economic performance best? Traditional or creative class. Journal of the American Planning Association, 74(2), 180–195. Florida, R. (2002). The rise of the creative class and how it’s transforming work, leisure, community & everyday life. New York: Basic Books.
43
Florida, R. L. (2012). The rise of the creative class: Revisited. Basic Books. Frenkel, A., Bendit, E., & Kaplan, S. (2013). Residential location choice of knowledgeworkers: The role of amenities, workplace and lifestyle. Cities, 35, 33–41. Hansen, H. K., & Niedomysl, T. (2009). Migration of the creative class: Evidence from Sweden. Journal of Economic Geography, 9(2), 191–206. Hoyman, M., & Faricy, C. (2009). It takes a village: A test of the creative class, social capital, and human capital theories. Urban Affairs Review, 44(3), 311–333. Lawton, P., Murphy, E., & Redmond, D. (2013). Residential preferences of the ‘creative class’? Cities, 31, 47–56. Martin-Brelot, H., Grossetti, M., Eckert, D., Gritsai, O., & Kovacs, Z. (2010). The spatial mobility of the ‘creative class’: A European perspective. International Journal of Urban and Regional Research, 34(4), 854–870. McGranahan, D. A., Wojan, T. R., & Lambert, D. M. (2011). The rural growth trifecta: Outdoor amenities, creative class and entrepreneurial context. Journal of Economic Geography, 11(3), 529–557. Sands, G., & Reese, L. A. (2008). Cultivating the creative class: And what about Nanaimo? Economic Development Quarterly, 22(1), 8–23.
Contents lists available at ScienceDirect
Cities journal homepage: www.elsevier.com/locate/cities
Leisure choices of the creative class Eric Joseph Van Holm ⇑ Georgia State University & Georgia Institute of Technology, United States
a r t i c l e
i n f o
Article history: Received 16 December 2013 Received in revised form 15 May 2014 Accepted 17 May 2014
Keywords: Creative class Amenity choice Cultural economics City development
a b s t r a c t Richard Florida’s theory of the creative class has led cities to offer amenities as a way to attract the knowledge workers that he argues drive the economy. At present, scant evidence exists concerning the leisure activities different worker classes choose to engage in. Using logit analysis on data from the American Time Use Survey, I analyze the leisure preferences of the creative class and find limited support for Florida’s assertions about their amenity use. Cities and policymakers should consider the evidence here before trying to create unique amenities to attract the creative class, because they may not require distinctive activities to choose a location. Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
Literature review
Cities are constructing bike paths, financing public art, and encouraging diversity, all in an effort to attract knowledge workers. The Rise of the Creative Class by Richard Florida (2002) has become a touchstone for policymakers throughout the United States and Western Europe, and it has offered planners a new path towards prosperity. Traditionally, urban development theory stated that individuals follow jobs, making attracting employers the key to promoting economic growth. Florida (2002) and Clark (2011) flip that traditional narrative arguing that in the knowledge economy jobs follow people. Florida advises cities to build a ‘‘people climate’’, because talented individuals are the new primary resource for a strong economy, meaning educated, creative people will find jobs wherever they locate (Florida, 2002, p. 283). Florida argues that the creative class chooses locations based on high diversity and natural and cultural amenities. He speculated on the cultural amenities the creative class are most likely to enjoy, but was unable to test his hypotheses empirically lamenting that data on the subject was not available. Using data from the American Time Use Survey, released since the publication of the book, I test several of his assertions. First, I will test what leisure activities the creative class is more likely than working and service class employees to enjoy. Secondly, I will assess what other factors explain differences in leisure preferences.
Florida’s theory of the creative class centers upon their attraction to the 3 T’s of technology, talent, and tolerance. That is, Florida posits that regions with a high-acceptance of diversity and alternative life-styles and the right mix of cultural amenities attract individuals with greater talent. These talented individuals, a valuable resource in the modern economy, attract high-technology industries, spurring economic growth. Therefore, rather than the traditional narrative that jobs attract workers to a region, Florida argues that workers attract jobs, placing the emphasis for cities on building a people climate in order to attract high-skilled workers. Florida makes repeated claims about the types of lifestyle amenities that attract the creative class; he concludes from his interviews and focus groups that they enjoy participatory experiences that evolve organically and events that are not prepackaged. Because the creative class place such a high value on time, their entertainment must be continuously engaging; their ‘‘lifestyle comes down to a passionate quest for experience’’ (Florida, 2002, p. 166). Thus, two of Florida’s central claims are that the creative class drives economic development, and that they are a unique demographic. However, both claims have been weakened by later research. Studies have shown traditional measures are better predictors of economic health than Florida’s measures of technology, tolerance and talent (Donegan, Drucker, Goldstein, Lowe, & Malizia, 2008; Hoyman & Faricy, 2009; Sands & Reese, 2008). Furthermore, the creative class is not very distinctive in their residential choice (Andersen, Bugge, Hansen, Isaksen, & Raunio, 2010;
⇑ Corresponding author. Tel.: +1 916 396 7806. E-mail address: [email protected] http://dx.doi.org/10.1016/j.cities.2014.05.006 0264-2751/Ó 2014 Elsevier Ltd. All rights reserved.
E.J. Van Holm / Cities 41 (2014) 38–43
Asheim & Hansen, 2009; Frenkel, Bendit, & Kaplan, 2013; Lawton, Murphy, & Redmond, 2013), mobility (Hansen & Niedomysl, 2009; Martin-Brelot, Grossetti, Eckert, Gritsai, & Kovacs, 2010), or amenity use (Bille, 2010). Bille (2010) was the first to test Florida’s claims about amenity use by the creative class. He compared creative class and service class workers on the use of 35 amenities from a Danish survey. Bille finds that the creative class is more likely than the service class to attend contemporary concerts, to visit art museums and cultural landscapes, to engage regularly in fitness and play sports, and to use the internet than the service class. The creative core, a subset of the creative class, is significantly more likely than the service class to enjoy creative self-expression, to engage in noninstitutional fitness activities, to read a book, and to visit historic sites, museums, and classical art forms. The creative class is neither more nor less likely to watch TV, watch films at home, listen to recorded music, play computer games, read magazines, or go to the zoo, amusement park or aquarium. Bille’s findings agree with Florida in that the creative class uses the internet more, is more likely to exercise, and is less likely to attend spectator sports events. Contrary to Florida’s assertions, however, Bille’s results show that the creative class is more likely than the service class to attend a museum, or scheduled activities like evening class or lectures, and is as likely to watch TV or a movie. Data In 2002, comprehensive data was not available to test amenity use by the creative class or others. Since 2003, the Bureau of Labor Statistics (BLS) has been administering the American Time Use Survey (ATUS) to collect detailed data on how Americans spend their time. The BLS randomly selects participants from the Current Population Survey (CPS) to fill out a time diary for all household members for a specific 24 h period, which it then combines with demographic information from the CPS (Abraham, Flood, Sobek, & Thorn, 2011). I combine three years of the data, (2008–2010), providing a sample size of 18,386 respondents, allowing sufficient observations for even such rare events as attending performing arts. One limitation of the ATUS data for this study is the distinction between revealed and stated preferences. The data used here is of the revealed variety, in that it is activities actually done by respondents. Stated preferences are also important for cities, as individuals use cultural and natural amenities as signals of the type of city or community they would like to reside in, regardless of whether they actually partake in those activities. There has been work on Florida’s theory of the stated preference variety (Frenkel et al., 2013; Lawton et al., 2013) but because Florida discusses amenities as activities the creative class not only uses as signals but participates in, it is appropriate to use revealed preference data. Methodology I am able to test five specific claims made by Richard Florida concerning amenity use by the creative class. ATUS respondents report the number of minutes they spent on each activity in the specified day; with the exception of television watching, I recode each as a dummy variable with 1 for having done that activity because I am uninterested in the duration of any given amenity but rather its use. I divide the dependent variables into four categories of arts, time wasting, exercise, and late night dining. Using the four hypotheses and five dependent variables described below, I will test Florida’s claims that the creative class is less likely to attend performing arts, and watch television and more likely to practice outdoor sports, exercise, or eat out late at night. I use the same model for all dependent variables.
39
H1. The creative class is less likely to enjoy traditional performing arts. ‘‘In many cities recently, museums and the symphony opera and ballet have fallen on hard times . . . Meanwhile, the Creative Class is drawn to more organic and indigenous street-level culture’’ (Florida, 2002, p. 182). Florida argues at length that traditional forms of high-culture, such as the symphony, opera, and ballet are less attractive to members of the Creative Class than past generations. The ATUS asks how much time individuals spent at the performing arts, using examples such as the theatre, opera, or musicals. I code the variable as a 1 if respondents indicate they attended any performing arts; only .8% did so on the survey day. H2. The creative class is less likely to watch television. ‘‘But if you spend your workday in front of a computer screen or an artist’s canvas, you are probably not eager to spend your leisure time in front of a TV screen’’ (Florida, 2012, p. 139). While some may consider an hour of the news as less of a waste of time than an hour of reality television, Florida does not differentiate, rather asserting that television itself does not appeal to the creative class. However, because watching 1 min of television is different from watching all day, I code television watching as 1 if the individual watched more than 4 h of television. 22% of the sample watched more than 4 h of television within the 24-h period used for the survey. H3. The creative class is more likely to exercise ‘‘The force behind the Creative Class obsession with being in shape is more than a concern with health . . . I see it as a growing awareness of the body as an arena for creative expression’’ (Florida, 2002, p. 177). ‘‘Creative Class people in my studies are into a variety of active sports, from traditional ones like bicycling, jogging and kayaking, to newer more extreme ones like trail running and snowboarding’’ (Florida, 2002, p. 173). I code outdoor sports as 1 if the individual did any biking, climbing, hiking, or running; .25% of respondents reported doing these activities. I create the dummy exercise as 1 if they have done aerobics, walked, lifted weights, done yoga or used cardiovascular equipment. By this definition 10% of the sample exercised. H4. The creative class is more likely to eat out late at night. ‘‘They need to have options around the clock. . . the highestrated nightlife options were cultural attractions and late-night dining’’ (Florida, 2002, p. 225). Finally, late night dining is coded to encompass anyone that ate out at a bar or restaurant from 10 p.m. until 2 a.m.; 1% of the sample reported dining out between those hours. Independent variables The Current Population Survey data provides major occupational code for each respondent, which allows me to test my key independent variable creative class. I was able to nearly replicate Florida’s definition of the creative class, which he categorizes as containing the major occupational codes for: business and financial operations; computer and mathematical science; architecture and engineering; life, physical, and social science; community and social service; legal, education, training, and library arts; design,
40
E.J. Van Holm / Cities 41 (2014) 38–43
entertainment, sports and media; healthcare practitioner; and high-end sales and sales management. The only occupation I was unable to include were high-end sales and sales management, because the occupational codes do not differentiate between high-end and low-end sales. I removed all sales from the sample, because placing them in either the creative or service class would create bias. The percentage of workers in the creative class for my sample was 39.3%, higher than Florida’s estimate of 33%. Other than the issue with sales, I was able to replicate perfectly Florida’s definition of the service class: health care support; food preparation and food-service-related; building and grounds cleaning and maintenance; personal care and service; low-end sales; office and administrative support; and protective services. My definition of the working class had no differences from Florida’s, containing construction and extraction; installation, maintenance, and repair; production; and transportation and material moving. Agriculture is the final class of workers, but the final data sample contained few observations for agricultural workers, so I dropped it from the data. My first analysis will use logit analysis to regress the amenities on worker class. Initially, I will test Florida’s hypotheses by finding the relative propensity of the creative class to use each amenity compared to other worker classes. If the creative class is no more likely than other classes to enjoy an amenity, then the government offering it is not truly attracting the creative class, so much as attracting all workers. The bivariate regression is used to establish the importance of preceding and intervening variables in the later analysis. Secondly, I will conduct multivariate analysis to test whether other variables are able to explain the relative propensity of different worker classes to enjoy each activity. It may be that other factors are a bigger determinant of individual’s leisure activities than worker class. I expect a variety of demographic variables to influence amenity choice. A gender dummy is included as well as four dummy variables that define white, African–American, Latino, Asian, and other as mutually exclusive. I coded education, representing the highest grade of education reached for the respondent, divided into four levels of high school, some college, college, or graduate school. Age may have a non-linear relationship to many amenities, so I include a set of 5 dummies capturing varying points in an individual’s life to capture the different leisure choices of different age groups. Income is measured with 5 dummy variables, divided by intervals of $25,000 ending with $100,000 and above. Because a disability may keep an individual from enjoying the amenities tested for, I have included a dummy variable, which will be 1 if any impairment is present. As individuals marry and have children, amenity use may change. Therefore, I control for their marital status, with variables for whether they have never been married, are married, or are divorced or separated, as well as the number of children under 18 in the household. In order to account for regional and locational differences in amenity use, I create two sets of dummy variables. The first indicates what Standard Federal Region the respondent lives in and the second set of dummy variables designates whether the individual lives in the central city of a metropolitan statistical area (MSA), suburbs of an MSA, or not in an MSA. Finally, the day of the week the individual records in their time journal will affect the amenities they use. Seeing the ballet is associated with ‘date nights’ on Friday or Saturday evening, but an individual surveyed about their amenity use for a Tuesday is less likely to attend the performing arts. However, the ATUS does not report what day the 24-h period occurs on or whether it is a weekend or weekday. I include a variable to indicate whether the individual worked more than 4 h on that day to attempt to capture the influence of work on amenity choice.
Bivariate regression findings I display the first sets of regressions in Table 1 as odds-ratios from a logit regression, with creative class as the reference group. In order to provide context for the magnitude of each change in probability, I provide the percentage of the sample doing each activity at the top of Table 1. The row for Florida’s hypothesis indicates the expected direction of the relationship, moving from the creative class to the working or service class. The odds that a member of the service or working class attends the performing arts were only 71% and 30% as high as the odds for someone from the creative class respectively. Likewise, the creative class has greater odds of practicing outdoor sports or performing traditional forms of exercise than both the working and service class. The only dependent variable for which the working and service class had greater odds was watching more than 4 h of television, where they are 45% and 102% more likely than the creative class respectively. Neither class of worker shows a significant difference in regards to eating out late at night. Florida correctly hypothesizes about the creative class in regards to television, outdoor sports, and traditional forms of exercise. However, the coefficients of the other three dependent variables are insignificant or in an unexpected direction meaning that even without controlling for other socioeconomic factors Florida is only accurate on half of his suggestions. To test whether the relationship of worker class with amenity use are created by spurious or indirect relationships, the next set of regression results will identify what other factors may explain the differences in leisure choice.
Multivariate regression findings Table 2 contains results from the multivariate regression as odds-ratios. I tested for multicollinearity for the full model and the variation inflation factors (VIF) for all variables were below five, indicating there was not a problem. Being a member of the creative class has a diminished effect on amenity use when other factors are included in the regression. While none of the signs change direction in the new analysis, four of the eight significant worker class coefficients from the bivariate regression lose statistical significance. Of the four that retain significance, only two are in the predicted direction. The following sections will discuss how the worker class variables reacted to the inclusion of other factors, and what those other variables indicate about the relationship of being a member of the creative class and leisure use. Table 3 displays the relationship of the three worker classes to each of the other independent variables, which will help to identify relationships that affected the significance of worker class coefficients. All variables discussed are significant at .05 or greater, unless otherwise specified.
Performing arts The odds of the working class attending the performing arts are 50% lower than the odds for the creative class, while the difference with the service class is not significant. That this result remains significant despite the inclusion of other factors indicates the creative class has a predilection towards traditional arts that cannot be explained with other socioeconomic factors; however, this result is also in the opposite direction predicted by Florida. The results here generally concur with Bille (2010), who found the creative class more likely to attend classical arts than the service class. In addition, my results connect well with Frenkel et al. (2013) who showed that the creative class prefer to live near
41
E.J. Van Holm / Cities 41 (2014) 38–43 Table 1 Bivariate regression results. Variable
Performing arts
Television
Outdoor sports
Exercise
Late night eating
Creative class (reference group) (%) Service class Working class Florida’s hypothesis
1.02 0.71 0.30 Positive
16.98 1.45 2.04 Positive
3.4 0.56 0.50 Negative
12.47 0.68 0.51 Negative
1.07 0.94 0.73 Negative
cultural attractions including the theater and museums; here I show they are also more likely to partake in similar activities. The loss of significance for the service class is likely connected to ethnicity, income and education. Variables representing Latinos and Asians are also significant and below one, meaning neither ethnic categories is more likely than whites to enjoy the symphony, opera, or ballet. Conversely, an income of $75 and $100 thousand and having a graduate degree, both of which the creative class is more likely to have, increase the odds of attending the performing arts. Lastly, having worked 4 h on the
day the time journal is kept lowers the chances of attending the performing arts. Television Individuals in the creative class are less likely than both the service and working class to watch more than 4 h of television. The service and working class have odds 13% and 19% higher respectively, holding race, age, sex, income, family status, education, and location constant. Bille (2010) found that the creative
Table 2 Multivariate regression results.
*
Performing arts
Television
Outdoor sports
Exercise
Late night eating
Creative class (reference group) (%)
1.02
16.98
3.4
12.47
1.07
Service class Working class
0.93 0.50**
1.13** 1.19***
1.02 0.85
1.01 0.84**
1.13 1.02
Latino Black Asian Other race
0.49* 0.88 0.44 0.55
1.04 1.49*** 0.84 0.94
0.70** 0.44*** 0.71 1.53
1.02 0.85** 1.17 0.71
0.98 0.51** 1.41 0.99
Age 30–50 Age 50–65 Age above 65 Female
1.15 1.18 1.2 1.24
1.38*** 1.61*** 1.67*** 0.55***
0.9 0.62*** 0.57** 0.66***
1.08 1.05 1.35** 1
0.59*** 0.45*** 0.11*** 0.94
25–50k income 50–75k income 75–100k income Income above 100k
1.07 1.57 1.72* 1.35
1.08 0.93 0.88* 0.81***
0.94 1.63*** 1.42* 2.00***
1.12 1.23** 1.31*** 1.36***
1.3 1.21 1.35 1.78**
Married Divorced or separated
1.03 1.24
0.79*** 0.89*
0.79* 0.77
0.99 1.05
0.43*** 0.8
One child Two children Three or more children
0.89 0.75 0.64
0.74*** 0.64*** 0.64***
0.89 0.89 0.78
0.71*** 0.69*** 0.76**
0.99 0.84 0.48*
Some college College Graduate degree
1.42 1.47 1.88**
0.79*** 0.67*** 0.51***
1.2 1.89*** 2.55***
1.31*** 1.75*** 2.39***
1.13 1.59* 1.47
Any handicap
0.95
1.01
0.74
0.71**
0.79
Central city Suburbs
1.33 1.06
1.09 1.09*
1.14 0.94
1.22*** 1.13*
2.09*** 1.44*
Worked
0.39***
0.20***
0.47***
1.14***
0.92
New England New York/New Jersey Mid-Atlantic Southeast North South Plains Mountain West Southwest
1.07 1.65 1.3 0.81 0.98 0.74 1.11 0.51 1.07
0.98 1.07 1.1 1.24** 1.07 1.15 1.12 0.9 1.06
0.57 0.9 0.37*** 0.77 0.84 0.76 0.75 1 1.33
1.08 0.93 0.86 0.72*** 0.85 0.84 0.78* 1.17 1.06
0.86 1.02 0.69 0.81 0.78 0.86 0.67 0.74 0.63
2009 sample 2010 sample
1.31 0.83
0.96 0.92
1.09 1.17
1.17* 1.14
0.75 0.93
McFadden’s R2
0.05
0.12
0.07
0.03
0.05
p < 0.1. p < 0.05. *** p < 0.01. **
**
42
E.J. Van Holm / Cities 41 (2014) 38–43
Table 3 Summary statistics (in %). Variable
Creative class
Service class
Working class
Count Percent
8057 43.8
6370 34.6
3959 21.5
White Latino Black Asian Other race
76.9 7.1 9.6 5.3 1.1
60.7 18 17.7 2.4 1.5
62.6 21.1 13.7 1.8 1.3
18–30 30–50 50–65 Over 65
10.9 56.7 27 5.2
19.7 45.2 26.1 6.4
16.1 50.4 28.1 4.7
Male Female
45.2 54.8
30.5 69.5
82.8 17.2
Under $25,000 $25,000–50,000 $50,000–75,000 $75,000–100,000 Over $100,000
6.1 16.1 19.4 17 33.9
21.9 30.2 18.6 10.7 10.1
19.4 32.4 21.6 9.9 8.4
Never married Married Divorced or separated
19.6 62.8 15
30.6 45.7 19.1
23.4 55.6 18.5
No kids One child Two children Three or more children
45.9 21.9 23.1 25.3
48.9 22.3 18.3 26.5
49.9 19.9 18.9 26.9
High school Some college College Graduate degree
10.5 22.6 38.2 28.7
43.5 36.3 15.3 4.9
60.9 29.4 7.9 1.8
Any handicap
3.2
5.4
5
Worked
27.5
27.9
21.8
Central city Suburbs Rural
45.6 12.3 42.2
39.5 17 39.6
39.4 22.1 41
New England New York/New Jersey Mid-Atlantic Southeast North South Plains Mountain West Southwest Northwest
6.1 8.4 11.7 13.9 17.9 11.2 5.6 4.5 13.1 7.7
4.6 8.3 10.1 15.7 19.5 11.4 5.7 4.1 13.1 7.5
3.5 6.1 10 15.9 21.1 12.3 6.7 4.4 12.2 7.8
class did not differ from the rest of the population with respect to many commonplace amenities, a result that differs from mine although television is the only such activities I test. The creative class is the most likely to be highly educated or have high incomes, which are both significant and negative for television watching. That is, even when controlling for the fact that the creative class is more likely to have finished a postsecondary education and more likely to earn above $75,000, they are still significantly less likely to watch 4 h of television. Exercise Both variables for class of worker lose their significance for outdoor sports as does exercise for the service class. Similar to television watching above, education and income differences are able to explain part of the difference in the results. Income and education have consistently large effects on exercise and outdoor sports, with increasing positive changes the greater the difference from the reference group. In addition, race may be a factor. Working and service class members are more likely to be minorities, the variables for which are mostly significant and negative. Therefore,
in addition to income and education explaining why the creative class is more likely to exercise, race may explain why the service and working class are less likely to do so. Lastly, an individual living in the central city has odds 22% higher for exercise, probably due to the nature of individual’s employment and the availability of gyms. Despite being more likely to live in the city, the creative class is still more probable than the working class to exercise. The results for exercise differ strongly with past analyses, which have found a strong attraction for the creative class to such activities. McGranahan, Wojan, and Lambert (2011) found that the creative class was more likely to live near outdoor sport opportunities and Bille’s (2010) results showed the creative class was more likely to do traditional exercise and walk or cycle than the service class. Late night dining In both regressions there were no significant differences between worker classes in terms of late night dining. The variable that had the most consistent effect on eating around midnight was age, an area that worker classes did not differ significantly in the sample. Central city is also significant, likely due to the availability of restaurants open at any hour. Conclusion I find narrow support for Florida’s claims about how the creative class differs from the service and working class in the amenities it enjoys, as most of Florida’s assertions are not supported by the evidence presented. Even in the bivariate analysis, where the creative class diverged from the service and working class on four of the activities tested, differences on one of those activities were in the opposite direction of what Florida hypothesized. In the bivariate test the creative class was, more likely to go to the performing arts than both the service and working class. Furthermore, when I control for other factors, the majority of differences between worker classes disappear. In the multivariate analysis, the creative class differs from both classes in television watching, and from the working class in the performing arts and exercise. My results are much more pessimistic than Bille’s (2010) study, which found general agreement with the broad claims of Florida even when controlling for other socioeconomic factors. Cities and policymakers should consider the evidence presented before trying to create unique amenities to attract the creative class. The fact that occupation did not make for consistently large changes in amenity preferences for either the bivariate or multivariate analysis indicates that the creative class may not require unique lifestyle options to enjoy a location. Demographics do play a part in determining leisure choices, but Florida’s creative class is the wrong way to understand these differences. The creative class may still enjoy bicycling and kayaking, but no more so than other workers with similar socioeconomic status; if a city endeavors to create a people climate it will be doing so to improve its livability of all residents, not a specific class of workers. The demographics of the roughly average member of the creative class are also worth discussion. A reader could come away from The Rise of the Creative Class with an image of the creative class as young, tattooed, and untraditional. However, nearly 55% of the creative class is between the ages of 35 and 55, the vast majority is married, most are female, and more than half have at least one child. If a city really wants to attract the broadest segment of the creative class, it appears that sound areas to invest in would be safe streets, good schools, and numerous parks in order to create family-friendly amenities.
E.J. Van Holm / Cities 41 (2014) 38–43
References Abraham, K. G., Flood, S. M., Sobek, M., & Thorn, B. (2011). American Time Use Survey data extract system: Version 2.4. Andersen, K. V., Bugge, M. M., Hansen, H. K., Isaksen, A., & Raunio, M. (2010). One size fits all? Applying the creative class thesis onto a Nordic context. European Planning Studies, 18(10), 1591–1609. Asheim, B., & Hansen, H. K. (2009). Knowledge bases, talents, and contexts: On the usefulness of the creative class approach in Sweden. Economic Geography, 85(4), 425–442. Bille, T. (2010). Cool, funky and creative? The creative class and preferences for leisure and culture. International Journal of Cultural Policy, 16(4), 466–496. Clark, T. N. (2011). The city as an entertainment machine. Plymouth, United Kingdom: Lexington Books. Donegan, M., Drucker, J., Goldstein, H., Lowe, N., & Malizia, E. (2008). Which indicators explain metropolitan economic performance best? Traditional or creative class. Journal of the American Planning Association, 74(2), 180–195. Florida, R. (2002). The rise of the creative class and how it’s transforming work, leisure, community & everyday life. New York: Basic Books.
43
Florida, R. L. (2012). The rise of the creative class: Revisited. Basic Books. Frenkel, A., Bendit, E., & Kaplan, S. (2013). Residential location choice of knowledgeworkers: The role of amenities, workplace and lifestyle. Cities, 35, 33–41. Hansen, H. K., & Niedomysl, T. (2009). Migration of the creative class: Evidence from Sweden. Journal of Economic Geography, 9(2), 191–206. Hoyman, M., & Faricy, C. (2009). It takes a village: A test of the creative class, social capital, and human capital theories. Urban Affairs Review, 44(3), 311–333. Lawton, P., Murphy, E., & Redmond, D. (2013). Residential preferences of the ‘creative class’? Cities, 31, 47–56. Martin-Brelot, H., Grossetti, M., Eckert, D., Gritsai, O., & Kovacs, Z. (2010). The spatial mobility of the ‘creative class’: A European perspective. International Journal of Urban and Regional Research, 34(4), 854–870. McGranahan, D. A., Wojan, T. R., & Lambert, D. M. (2011). The rural growth trifecta: Outdoor amenities, creative class and entrepreneurial context. Journal of Economic Geography, 11(3), 529–557. Sands, G., & Reese, L. A. (2008). Cultivating the creative class: And what about Nanaimo? Economic Development Quarterly, 22(1), 8–23.