The “Creative Class” in the Uk: an Initial Analysis the “Creative Class” in the Uk: an Initial Analysis

Home , Break (work), Creative Cities, Creative class, Poverty

THE “CREATIVE CLASS” IN THE UK: AN INITIAL ANALYSIS THE “CREATIVE CLASS” IN THE UK: AN INITIAL ANALYSIS

by Nick Clifton

CLIFTON, N. (2008): ‘The “creative class” in the UK: an initial In a knowledge-based economy, the ability to at- analysis’, Geogr. Ann. B 90 (1): 63–82. tract and retain highly skilled labour is therefore perceived as crucial to the current and future pros- ABSTRACT. Richard Florida argues that regional economic outcomes are tied to the underlying conditions that facilitate creativ- perity of regions as well as entire nations. For ex- ity and diversity. Thus the Creative Class thesis suggests that the ample, Florida (2000) has argued that in the know- ability to attract creativity and to be open to diverse groups of peo- ledge economy, regions develop advantage based ple of different ethnic, racial and lifestyle groups provides distinct on their ability to quickly mobilize the best people, advantages to regions in generating innovations, growing and attracting high-technology industries, and spurring economic resources and capabilities required to turn innova- growth. In this paper we investigate the extent to which there tions into new business ideas and commercial prod- might be similar processes concerning the relationship between ucts. In particular, the ability to attract creative peo- creativity, human capital, and high-technology industries at work ple in arts and culture fields and to be open to di- in the UK as in North America. The approach taken is broadly sympathetic to the Creative Class thesis; critical perspectives and verse groups of people of different ethnic, racial reservations from the literature are introduced as appropriate re- and lifestyle groups provides distinct advantages to search is focused around the three principal research questions: regions in generating innovations, growing and at- Where is the creative class located in the UK? What is the impact tracting high-technology industries, and spurring of quality of place upon this dispersion? What is the connection between the location of the creative class and inequalities in tech- economic growth (Gertler et al. 2002). nical and economic outcomes within the UK? To this end, the cre- This research demonstrates that quality of place ative class and its subgroups are defined and identified. We then must be understood in broader terms than we have construct quality of place indicators relating to tolerance, diver- traditionally been accustomed to: while the attrac- sity, creativity and cultural opportunity. To these are added measures of public provision and social cohesion. Data are analysed by tiveness and condition of the natural environment means of correlations and regression. In general we find that, al- and built form are certainly important, so too is the though the distribution of the creative class is uneven and com- presence of a rich cultural scene and a high concen- plex, our results are consistent with the findings of the North tration of people working in cultural occupations American research with the notable exception of technology- based employment growth. Finally, priorities for further research (most specifically the “bohemians”). According to are discussed. The need to further investigate causality, variations the results from Florida et al.’s research the under- within the creative class itself, and the potential role of qualitative lying hypothesis is that the presence and concentra- data in this are highlighted, as is the potential fate of “non-creation of bohemians in an area creates an environment tive” workers and places. or milieu that attracts other types of talented or high Key words: Creative Class, spatial analysis, economic dynamism, human capital individuals. The presence of such hu- implications – policy and further research man capital in turn attracts and generates innovative, technology-based industries (Florida 2002b). Introduction However, given the interest Florida’s writings Much of the recent interest in the development of have received from academics, policy-makers and creativity has drawn upon Richard Florida’s the media alike, it is no surprise that they have been (2002b) book The Rise of the Creative Class. the object of a high degree of critical examination. Whereas in the Industrial Age classical and neo- This critique has centred most notably around the classical economic theory told us that ‘people fol- apparent fuzziness of some of the concepts, defini- lowed jobs’, in the modern knowledge economy tions and causal logic Florida employs, the seem- Florida describes how ‘jobs follow talented peo- ingly convenient appeal of his ideas to the agendas ple’. That is, places that display “creative class” of a multitude of urban actors and policy-makers, characteristics, meaning a high presence of profes- and conversely the minimal attention paid to diffi- sionals, technologists and bohemians, performed cult issues such as the potential inequalities and best economically in recent years. negative externalities implied by a creative class

© The author 2008 Journal compilation © 2008 Swedish Society for Anthropology and Geography 63 NICK CLIFTON model of regional development. More fundamen- play the lead role in knowledge-intensive produc- tally, a number of authors question the very concept tion and innovation – who provide the ideas, know- of a “new economy” that can deliver prosperity in how, creativity and imagination so crucial to eco- tandem with greater levels of self-determination to nomic success. The idea that growth-based devel- an ever-expanding body of knowledge workers. opment agendas can be actively pursed at the city Brown and Lauder (2006), for example, envisage a level is however not a new one – see, for example, future scenario of diminishing returns to human the “urban entrepreneurialism” documented by capital investments, with all the various discontents Leitner (1990). If we accept that the value creation this entails. Moreover, we also need to be aware in many sectors of the economy rests increasingly that “Culture” is now positioned at the centre of on non-tangible assets, the locational constraints of many urban policies. It has become a delivery ve- earlier eras – for example, the access to good nat- hicle for all manner of outcomes including social ural harbours or proximity to raw materials and cohesion, sustainability, economic growth, civic cheap energy sources – no longer exert the same pride, mental and physical well-being, social inclu- pull they once did. Instead, what Florida and his as- sion, and an ever-increasing array of other social, sociates assert matters most now are those at- economic and environmental goals. This trend is tributes and characteristics of particular places that highlighted by a recent report by the Department of make them attractive to potentially mobile, much Culture, Media and Sport which states: ‘Culture sought-after talent. A key reason for believing that drives regeneration in many ways, from inspiring a significant shift has occurred taking us into a landmark buildings through to reviving the decay- knowledge economy is that data suggest this to be ing centres of market towns to bringing a commu- true. Thus the book value of intangible assets com- nity together around an arts event’ (DCMS 2004, p. pared to raw materials has shifted from 20:80 in the 6). In this climate we need to guard against the Cre- 1950s to 70:30 in the 1990s (Cooke and De Lau- ative Class and the Creative and Cultural Industries rentis 2002). Consequently, the distribution of tal- becoming the latest policy panacea. ent, or human capital, is an important factor in eco- This paper essentially seeks to apply Florida’s nomic geography, as talent is a key intermediate models systematically in the UK context,1 using variable in attracting high-technology industries comprehensive datasets, for the first time (to the and generating higher regional incomes. This best of the author’s knowledge). The approach tak- makes it an important research task to explore fac- en is broadly sympathetic to the Creative Class the- tors that attract talent and its effects on high-tech- sis, although we are of course mindful of the critical nology industry and regional incomes (Florida perspectives and reservations noted above. The 2002c). structure of the paper is therefore as follows: the The replacement of raw materials or natural har- theoretical basis of the Creative Class approach is bours with human capital and creativity as the cru- outlined in more detail below, with significant res- cial wellspring of economic growth means that in ervations and criticisms from the literature intro- order to be successful in the emerging creative age duced as appropriate. Data sources and methodol- of the knowledge economy, regions must develop, ogies are then described in some detail, with the attract and retain talented and creative people who consequent results presented. These are focused generate innovations, develop technology-inten- around the three principal research questions: sive industries and power economic growth. Such Where is the creative class located in the UK? What talented people are not spread equally across na- is the impact of quality of place upon this disper- tions or places, but tend to concentrate within par- sion? What is the connection between the location ticular city-regions. According to Florida, the most of the creative class and inequalities in technical successful city-regions are the ones that have a so- and economic outcomes within the UK? Finally, cial environment which is open to creativity and di- conclusions are drawn and implications for future versity of all sorts. The ability to attract creative research considered in the light of these questions. people in arts and culture fields and to be open to diverse groups of people of different ethnic, racial and lifestyle groups provides distinct advantages to The creative class: review of theoretical regions in generating innovations, growing and at- perspectives tracting high-technology industries, and spurring A distinct advantage of city-regions is their ability economic growth. to produce, attract and retain those workers who As Wojan et al. (2007) describe, theorizing on

© The author 2008 64 Journal compilation © 2008 Swedish Society for Anthropology and Geography THE “CREATIVE CLASS” IN THE UK: AN INITIAL ANALYSIS how local environments influence economic out- (2002) cities on average are twice as advantaged by comes has a long and rich history, the two dominant their knowledge intensity over towns and rural ar- views of which may be traced back to Marshall eas compared to their already existing advantages (1920) – agglomerations, industry/firm-focused – from agglomeration economies. Thus if a city and Jacobs (1961, 1969) focusing on variety and scores 50 per cent above the mean in GDP per cap- people. Traditional theories of economic growth ita it is likely to score 100 per cent above it in terms and development tended to emphasize the role of of its knowledge-based industry. Thus there is more natural resources and physical assets. Such theories chance of knowledge economy employment in the were used to inform strategies typically based on city than in the country, a major contributory factor various incentives to try to alter the location deci- in the renewed migration of young people from ru- sion of firms. In recent years, several more related ral to urban areas in many European countries, theories have emerged. The first, associated with making the knowledge economy unevenly distri- the work of Porter (2000) and others, emphasizes buted and knowledge poverty a new kind of loca- the role of clusters of related and supporting indus- tional disadvantage (Cooke and De Laurentis tries. According to this work, clusters operate as 2002). geographically concentrated collections of interre- Thus, it is not enough to attract firms: the “right” lated firms in which local sophisticated and de- people also need to be attracted. Richard Florida manding customers and strong competition with calls for complementing policies for attracting other firms in the same industry drive the innova- firms with policies for attracting people, which tion process. A second view associated with Lucas means addressing issues of “people’s climate” as (1988) and Glaeser (1998) focuses on the role of well as of business climate (Florida 2002c). Indeed, human capital – that is, primarily highly educated the former is seen as basic to the latter, in that the people. It argues that places with higher levels of presence of human capital and talent is essential for human capital are more innovative and grow more attracting and developing high-tech industries and rapidly and robustly over time. A third approach, consequently for the economic growth of cities. associated with Richard Florida, emphasizes the This suggests that the attention of politicians and role of creative capital, arguing that certain under- planners should be directed towards people, not lying conditions of places, such as their ability to companies; that is, away from business attraction to attract creative people and be open to diversity, in- talent attraction and quality of place (Florida form innovation and growth.2 Further independent 2002c). Critics of this viewpoint have however research by Robert Cushing (2001) provides sup- pointed to what they consider the “easy-sell” of this port for the creative capital view. Peck (2005) re- message to urban policy-makers; Peck (2005, p. marks upon how the Creative Class thesis taps into 752) in particular has dissected in some detail what many of the same ‘cultural circuits of capitalism’ he regards as the eminently ‘deliverable supply- (Thrift 2001) as much of the immediately preced- side policy prescriptions’ that flow implicitly from ing work on the knowledge (or new) economy. It acceptance of the Creative Class thesis, and which should also be noted that there is work preceding thus (either by design or otherwise3) find a ready the first accounts of the creative class which makes market among these newly “legitimized” urban ac- explicit reference to quality of place and locational tors. choice factors, including Wong (2001) in the UK, The knowledge-based economy means then that which in turn may be linked back to Hall et al. the ability to attract and retain highly skilled labour (1987). is seen as crucial in terms of both the current and fu- Such constellations of talents and creative peo- ture prosperity of city-regions as well as entire na- ple are – as already mentioned – most commonly tions. Recent research on this question indicates found in large city-regions where the diversity of that talent is attracted to and retained by cities, but urbanization economies is most abundant. This, to- not just any cities. In their analysis of American gether with other factors such as labour markets metropolitan areas, Richard Florida and Gary characterized by high demand for qualified person- Gates have shed new light on those characteristics nel, cultural diversity and tolerance, low entry bar- of urban regions that seem to be most important in riers and high levels of urban service, largely de- this process (Florida and Gates 2001; Florida termine the economic geography of talent and of 2002a, 2002b, 2002c). The central finding of this creativity, both of which display concentration to work is that the social character of city-regions has large cities. According to Cooke and De Laurentis a very large influence over their economic success

© The author 2008 Journal compilation © 2008 Swedish Society for Anthropology and Geography 65 NICK CLIFTON and competitiveness. In particular, Florida and his (2000) actually merged the two words themselves colleagues have found that those places which offer in describing the emergence of the bobos – a new a high quality of life and best accommodate diver- group of people in which bourgeois and bohemian sity enjoy the greatest success in talent attraction/ values are blended into the creative, unconvention- retention and in the growth of their technology-in- al but also entrepreneurial (Brooks’ depiction is tensive economic activities. though an inherently less sympathetic one than Diversity as a key aspect of successful places Florida’s description of his creative class). This concerns entry barriers facing newcomers: cities idea has some intrinsic appeal when the nature of with great diversity are understood as places where “cool” jobs in technology, new media and so on is people from different backgrounds can easily fit in considered: as Heath and Potter (2006, p. 206) put (Florida 2002c). There are several quantitative in- it: ‘What people yearn for these days is no longer dicators used by Florida and colleagues to capture an old-fashioned “status” job, like being a Doctor. this. One of the most influential variables was The “cool job” has become the holy grail of the found to be a city’s “gay index”, measuring the modern economy.’ Indeed, with play-zones and prevalence of gay males in the local population. chill-out areas having become the stuff of hi-tech This index has been shown to imply a high degree start-up cliché the question remains as to whether of openness to newcomers of diverse backgrounds such idiosyncrasies are actually intrinsic to this with respect to nationality, race, ethnicity and sex- kind of creativity or merely ostentations, to be ual orientation, reflecting an environment that is dropped hastily in an economic downturn such as open to diversity, high in urban oriented amenities, the fallout from the dot.com shakeout. A full an- and characterized by low entry barriers (Florida swer to this question is however beyond the scope 2002c). Another associated indicator of diversity is of this paper. the “melting-pot index”, reflecting the proportion The use of the word class comes of course with of a city-region’s population that is foreign-born; a certain baggage in that it implies some kind of this is the indicator we employ in the present study. self-identity and consistent value system within a The findings from the Canadian study by Ger- socio-political hierarchy. Whether the creative tler et al. (2002) strongly indicate that the relation- class really is a class in some kind of Marxist sense ships first captured for US city-regions in the work is something of a moot point; the broad attitudes of Florida and colleagues are also evident in Ca- held and approaches to life that Florida (and others) nadian city-regions. If anything, the relationships describe does suggest that at least the term is not in Canada are stronger than those found in the wholly erroneous in this context, but this argument United States. In particular, the findings showed is not an entirely convincing one. These traits in- that a vibrant local creative class and an openness clude personal attire and style, beliefs and values, to diversity attracted knowledge workers in On- attitudes to work – to old-style demarcations of tario and Canada more generally. It was also found blue collar and white collar is added the “no-collar” that, in general, Ontario city-regions (of which To- workplace; viewed through the eyes of someone ronto is the largest) have a solid foundation in used to the traditional demarcations of the work- these areas to compete against US city-regions. In place these are people ‘who seem to be always summary, there appears to be a strong set of link- working, and yet never working when they are sup- ages between creativity, diversity, talent and tech- posed to’ (Florida 2002b, p. 5). nology-intensive activity that are driving the The points that Markusen (2006) makes regard- economies of Ontario’s – and Canada’s – city-re- ing the fuzziness of Florida’s definitions and caus- gions (Gertler et al. 2002). al logic are thus warranted in certain respects, albeit that the comment ‘it is rather amusing to think of the vast bulk of artists as making common urban Creativity, class and the economic mainstream or economic cause with bankers, real estate devel- The creative class moniker can be linked more gen- opers’ (Markusen 2006, p. 1937) does seem rather erally to a reassessment of the idea that “bourgeois” close to something of a straw man. Artists (i.e. the values of business, profit and so on (inherently bohemians) are typically treated as a distinct linked to a wider conservative value system) are by group even within Florida’s creative class. Indeed definition mutually exclusive to “bohemian” values the bohemians are essentially posited as a quality of creativity, the embracing of new ideas and valu- of place factor in their own right; that is, one which ing of diversity (in its various forms). Brooks is attractive to the much wider creative class in

© The author 2008 66 Journal compilation © 2008 Swedish Society for Anthropology and Geography THE “CREATIVE CLASS” IN THE UK: AN INITIAL ANALYSIS general, which does not necessarily imply a com- plete commonality of values and beliefs.4 Further- Research questions more, the ingenious study by Wojan et al. (2007) Although North America and Europe share many did find evidence for a positive “artistic milieu” common values and institutions, there are aspects effect upon regional growth outcomes; these au- of their respective societal development that show thors attempt to avoid the potential bohemians- strong divergence; for example, with regard to po- lead-to-growth, growth-leads-to-bohemians cir- litical priorities, functioning of labour markets, cularity by controlling for both supply and de- economic growth processes and social outcomes mand-side factors for artists, and in turn using the (see e.g. Hall and Soskice 2001). Given that virtu- residuals from this model (i.e. unexplained varia- ally all of the systematic research of the phenome- tions in the presence of bohemians) to predict eco- non has taken place in North America, the Creative nomic growth. The positive result does not of Class thesis warrants investigation in a European course necessarily mean that artists themselves context.5 Therefore, the research reported in this are the conscious hard-nosed economic drivers of paper represents an analysis of quality of place and the localities in which they find themselves con- the dispersion of the Creative Class in the UK, centrated. Either way, a conceptual route in to building upon the work described above that has these issues of shared versus divergent preferenc- been undertaken in North American cities in order es is offered by an analysis of the knowledge bases to understand whether similar processes concern- which differentiate occupations within an aggre- ing the relationship between creativity, human cap- gate creative class (Hansen et al. 2005). We are not ital and high-technology industries are at work in able to emulate that here, but we can at least spec- Europe as clamed within North America. The ques- ify separate models for the Creative Core, Crea- tions we seek to address are then essentially: tive Professionals and combined Creative Class, and indeed omit the bohemians from the analysis 1 Where is the Creative Class located in the UK? when they are also involved as an independent 2 What is the impact of quality of place upon this quality of place variable. dispersion? Does the Creative Class thesis ap- However, even when the creative class is defined pear valid within the UK context? in its widest sense, this still implies that around 60 3 What is the connection between the location of per cent of the workforce is engaged in “non-crea- the Creative Class and inequalities in technical tive” activities. A number of authors make the point and economic outcomes within the UK? that the role of this non-creative class is neglected. Peck (2005) goes so far as to suggest that Florida Our research is taking place within the context of a advocates a form of what he (Peck) terms ‘creative wider project entitled ‘Technology, talent and tol- trickle-down’; that is, there is an acknowledgement erance in European cities: a comparative analysis’, of the potential inequalities and negative external- involving matched datasets and research partners in ities associated with high-growth locations, but Denmark, Sweden, Norway, Finland, the Nether- very little in the way of concrete policies for how lands and Germany.6 To get at these issues, the role these might be addressed, and generally nothing of human capital, creative capital and diversity in within the remit of central government. There has technology-based economic development in the been some attempt to deal with these concerns out- UK is investigated. The research uses the two side the US context (Cannon et al. (2003) in the measures developed by the North American stud- UK; Bradford (2004) in Canada) within some of ies: the Bohemian index to reflect creative capital, the creative-cities literature, but typically these and the Diversity (mosaic) index to reflect open- amount to little more than somewhat ad hoc social ness and diversity. This suggests that there will be objectives bolted on to an underlying Creative a relationship between openness to creativity and Class orthodoxy. Whatever the subtleties of the de- diversity and the ability to support high-tech indus- bate however, the fundamental point remains that a tries and economic development based on talented dichotomous split between the bohemian and the workers. New indices are developed in order to bourgeois, or the business like and the creative, is grasp the fundamental differences in certain as- no longer adequate in describing how a significant pects of life between North American and Europe- group of people live and work, and that this creative an societies (see the following section). As noted, class is now very much part of the economic main- this kind of analysis has not yet been performed for stream. European cities, and has the potential to shed im-

© The author 2008 Journal compilation © 2008 Swedish Society for Anthropology and Geography 67 NICK CLIFTON portant new light on the role of quality of place in and between Europe and the USA and Canada. Var- shaping the competitiveness of city-regions in the iation in availability across the partner nations has UK. inevitably imposed constraints upon the data used by individual partners, and in some cases involved compromises around lowest common denominator Data and methodology levels of detail and time frames. The data for the quantitative statistical analyses are derived primarily from the 2001 Census of Popu- Creative class: Clearly of central importance is the lation, supplemented by the Labour Force Survey ability to actually quantify the size of the creative and the Annual Business Inquiry. Various indices class present in any given location. These are es- are constructed from these data as described below, sentially people who as a key constituent of their and explored through means of plots, regressions work are involved in the creation of new know- and correlations. In addition, this is being supple- ledge, or use of existing knowledge in new ways, mented by the collection of qualitative empirical combinations and so on. In the absence of a primary material in the form of interviews with key actors, dataset relating to the actual engagement in such as well as with theoretical representative groups of activities, this is proxied by the use of occupational “talents” employed in urban governance and plan- categories. We subdivide the Creative Class into the ning in the case study cities, high-technology in- Creative Core (scientists and engineers, architects dustry and service, and higher education and re- and designers, academics and teaching profession- search institutions. These interviews are being als),7 and the Creative Professionals (associated complemented with additional interviews with rep- professional and technical occupations of the Cre- resentatives from creative/artistic occupations as ative Core, managers, financial and legal profes- well as from different ethnic groups. The qualita- sionals). Data for this index (and the others using tive analyses are aimed at obtaining a subjective occupational data) were derived from the 2001 evaluation and assessment of the relative impor- Census of Population, the only source of sufficient- tance of the various indices used in the quantitative ly similar size to allow four digit Standard Occu- analyses in order to obtain a more comprehensive pational Classification 2000 (SOC2000) break- picture of which preferences talented people actu- downs at the levels of regional disaggregation really have and why they behave the way they do. quired. These interviews are ongoing at the time of writing; thus most of the material presented here derives Bohemian index: The Bohemian index is defined from the quantitative analyses. using employment in artistic and creative occupations. It is a locational quotient that compares the region’s share of the nation’s bohemians to the re- Key variables for quantitative analyses gion’s share of the nation’s population (local prom- The key variables for quantitative analyses are the inence of employment in artistic/creative occupa- Bohemian index, the Talent index, the Diversity in- tions compared to national prominence of employ- dex, and the Tech-Pole index. These mirror varia- ment in the same occupations). Data for the UK are bles were employed in previous research by Florida delineated using the SOC2000 system; within this (2002a, 2002b, 2002c) and Gertler et al. (2002) on it is convenient to use sub-major group 34 (Culture, the geography of talent and the rise of the creative Media and Sports) to define the bohemian occupa- class. In addition, indicators for cultural and recre- tions.8 ational amenities, which were also used in Flori- da’s studies, will be considered. A pair of new in- Cultural Opportunity index: The opportunity to en- dicators, developed to reflect characteristics of Eu- joy cultural and recreational activities can be cal- ropean cities and their national political econo- culated more directly than other indicators on qual- mies, is also introduced. These two new indicators ity of place. According to Florida, such opportuni- are social cohesion, and a public provision index ties play an important role in cities’ ability to attract measuring the supply of public sector goods such the creative class. The proportion of employees in as education, healthcare, social security and so on. the cultural and recreational industries within an In general, the variables used were designed to area is used as an indicator of this cultural oppor- jointly maximize consistency between the different tunity. This involves using a number of three digit European countries involved in the wider project, Standard Industry Classification (SIC) groups from

Annual Business Inquiry to account for employ- some kind of gini coefficient with respect to in- ment within restaurants and bars, libraries, muse- come distribution, but these kinds of data are not ums and other cultural and entertainment activities. available at the spatial level we are using (this is also true for the other European project partners). Talent index: Talent is defined as the proportion of As such, the apparent level of exclusion from main- the population over 18 years of age with a Bach- stream economic activity was used, as represented elor’s degree or higher (local proportion compared by the ILO definition of unemployment.9 Using a to national). These data were obtained from the labour market variable as a proxy for social cohe- 2001 Census of Population. sion introduces a new net of complicating factors – as with other issues relating to the operationaliza- Diversity (Openness) index: The Diversity index is tion of Creative Class concepts, these are revisited the counterpart of Florida’s melting pot index. It is in later sections. The data were taken from the La- calculated as the proportion of the total population bour Force Survey with respect to the twelve- that is foreign-born. A second index, which meas- month period March 2001 to February 2002. ured the proportion of total population that is foreign-born of non-Western origin, was also con- Public Provision index: Having acknowledged the structed; the logic behind this was that a more vis- potential implications for social cohesion implied ible manifestation of diversity may be a more valid within the Creative Class model, there is a need to tool, or at least provide greater insight into the con- further take account of the European context; given cept of tolerance (see later discussion of this issue). the high levels of private employment-based provi- These data were provided by the 2001 Census of sion in the USA, particularly for healthcare, to what Population. extent might high-quality social welfare provision be a quality of place factor for the European crea- Tech-Pole index: Following the method of the tive class? In this index, levels of employment in Milken Institute (e.g. DeVol et al. 2007), this index education and healthcare (SIC2003, two-digit shows local employment in technology-intensive codes 80 and 85) for a given locality are expressed manufacturing and service sectors (specialization as a proportion of the resident population, using and size). The index compares a region’s share of data from the Annual Business Inquiry. national employment in high-technology industries to the region’s overall share of national em- Levels of geography: As described above, data were ployment; this is then adjusted for city size by mul- collected such that the triple goals of country-spe- tiplying by a region’s overall share of national high- cific, inter-country and Europe–North America technology employment. Therefore, it reflects both comparisons could best be achieved. This was also the region’s degree of specialization in technology- true with respect to the levels of geography (i.e. spa- intensive activity as well as its sheer scale of em- tial units employed in the analysis). With the partner ployment in these sectors. The index includes tech- countries accounting for large variations in size, nology-intensive sectors in both manufacturing governance structure, patterns of population disper- and services. Data for constructing the index were sion and so on it was impractical to impose a single extracted from the Annual Business Inquiry (ABI) standard definition; in practice such definitions ac- employee analysis for 2002. tually have different meanings dependent upon the context in which they are applied; i.e. the same def- Social Cohesion index: Florida himself has ex- inition will actually define a spatial unit that has a pressed concern regarding the polarization in terms different meaning from one country to another. of prosperity that high levels of creative class em- As we were seeking definitions that encapsulat- ployment has been associated with in the USA, to ed something approaching functional labour mar- the point where this is becoming a negative quality kets (analogous at least in part to the municipal of place factor for these creative workers them- city-regions employed in the North American selves. As such it was felt that some attempt to fac- research10) it was decided that the most meaningful tor this into the research design should be made; functional unit in each national context would be that is, not cohesion in the sense of a social capital used, subject to this also being a level at which the indicator (which may indeed be incompatible with necessary statistical data were available from the the concept of openness), but rather one based on relevant national agencies. As Parr (2005) notes, perceived inequality. Ideally this should involve standard UK administrative geographies do not

Table 1. Creative Class as a share of the labour force (%).

England and Wales Highest locality Lowest locality

Creative Core 9.7 13.7 5.0 Creative Professionals 25.5 44.3 17.7 Bohemians 2.1 8.8 0.8 Creative Class in total 37.3 64.9 24.1

Source: Census of Population (2001). typically relate in a systematic way to any theoret- able variation around the England and Wales per- ical construct of the city-region; for the UK this centages. It is worth noting here that the two meaningful unit was primarily the NUTS3 defini- NUTS3 areas of Inner London West and Inner Lon- tion (105 spatial units in England and Wales). Due don East between them account for the highest per- to the sometimes complex and non-hierarchical na- centages across all four creative class categories ture of these standard geographies, this is supple- shown. Other than Inner London West (64.9%), mented by analysis using the Unitary/County Au- only Inner London East (at just under 52%) pos- thority level (171 units). sesses a labour force of which the Creative Class A significant difference between our geography comprises more than half. For the bohemians an and that of the research undertaken in the USA and even greater concentration is observed; after the Canada is that it is totally inclusive (i.e. no locali- two Inner London areas (both above 7%) the high- ties are excluded), whereas the North American est percentage is found in Brighton and Hove at 4.4 analysis has tended to focus exclusively on large per cent of the labour force. urban centres. This was essentially done for rea- At the opposite end of the spectrum, the same lo- sons of consistency; the European countries are calities also tend to crop up across the board, albeit generally more densely populated than North with some variation; Stoke on Trent possesses the America, the demarcation of city-regions less ob- lowest proportion of total Creative Class and also vious and so partial selection is potentially more ar- Creative Core, Gwent Valleys the lowest share of bitrary. Once information has been included in a bohemian occupations, while the lowest share of database it can always be manipulated further, re- Creative Professionals is found in Hull. Despite scaled or partially excluded, while the reverse is this apparent variation, the same localities are typ- much more problematic. ically found within a few places of each other at both the top and bottom of the rankings. This ef- fectively demonstrates that although there are var- Results and analysis iations in the overall make-up of the Creative Class Mapping the creative class in England and Wales in any given place, there is little correlation be- As shown in Table 1, the creative class in England tween this variation and the actual size of the Cre- and Wales accounts for some 37.3 per cent of the ative Class in that place. workforce, substantially greater than the ‘more Table 2 shows the top fifteen and bottom fifteen than 30%’ figure that Florida (2002b) himself NUTS3 areas in England and Wales ranked in quotes with regard to the USA. Problems in obtain- terms of their population size. The actual popula- ing consistent occupational time series data mean tions vary from just under 1.8m down to around that it is difficult to draw many conclusions with re- 69 000, with a median size of just over 360 000. The gard to how the size of the creative class may be corresponding ranks in terms of Creative Class Lo- changing over time. However, if the major group cation Quotient (LQ) are also shown. The LQ itself Professional Occupations is taken as a proxy for the is a measure of spatial concentration, expressed as Creative Core, then an increase is observed from a proportion such that the average for England and below 9 per cent of the workforce in the 1991 Cen- Wales is 1. sus to over 11 per cent in 2001, suggesting signif- These data highlight the sometimes arbitrary na- icant growth in these occupations. ture of NUTS3 geography in the UK, with counties, The total figure for the Creative Class is split be- towns, cities and urban metropolitan areas all being tween the Creative Core (9.7% of the workforce), represented. Many of the smaller areas are partic- the Creative Professionals (25.5%) and the Bohe- ular accidents of geography such as islands and rel- mians (2.1%). As Table 1 shows there is consider- atively isolated smaller towns and cities.

Table 2. Creative Class location and population rank.

Population size Creative class LQ rank Locality (NUTS3) rank

1 London (inner, west) 1 2 London (outer, west and N west) 4 3 London (outer, east and N east) 29 4 Kent 36 5 Greater Manchester (south) 35 6 Essex 25 7 Hampshire 17 8 Greater Manchester (north) 75 9 London (outer, south) 6 10 Lancashire 50 11 Surrey 3 12 Hertfordshire 9 13 Birmingham 54 14 London (inner, east) 2 15 Calder, Kirlees and Wakefield 66

91 York 32 92 Swindon 47 93 Herefordshire 51 94 Southend 26 95 Telford and Wrekin 73 96 Peterborough 55 97 Thurrock 97 98 Blackpool 86 99 Blackburn and Darwen 87 100 Isle of Wight 65 101 Torbay 68 102 Powys 80 103 Gwynedd 70 104 Darlington 56 105 Anglesey 76

Source: Census of Population (2001).

One thing which Table 2 does highlight are the As the UK’s “second city”, Birmingham is an in- differences that exist within London, particularly teresting case; with statistics relating to the central the relatively low ranking in terms of Creative area (population of about 1m) it ranks proportion- Class LQs of the more peripheral areas in contrast ally much lower for the Creative Class (54th) than to the central areas. Although London looms large it does in terms of population (13th); variations over the UK picture, there are other regional cen- around this urban core are however apparent in Fig. tres of high creativity (and high concentrations of 1, discussed below. Despite these discrepancies bohemians). The situation tends to differ in the and data issues, there is some evidence of an asso- smaller countries, where the capital may be the ciation between size of agglomeration and creative only realistic locational choice for many special- class concentration, a rank correlation between the ized professional workers. Moreover, the status of two producing a coefficient of 0.41. In order to ex- London as a genuine global city is significant, as it plore this agglomeration effect further, the associ- is therefore competing for creativity (to use the ation between population density and Creative fashionable rhetoric) on the world stage, with the Class concentration was investigated; somewhat different tensions that this brings. One of these may surprisingly, virtually no link was found, with a be a latent conflict between national capital versus correlation coefficient of only 0.1.11 Clearly then it international roles, perhaps generating competing is not just size or density of population, the higher pressures or funding priorities; for example, re- levels of which are typically found in metropolitan sources which are ostensibly earmarked for local areas and urban centres, that is associated with the regeneration might be spent with one eye on the po- location of the creative class. This could reflect dif- tential national or international payoffs. ferent preferences within the agglomerated crea-

Table 3. Creative Core location by unitary authority/county.

Top 10 localities (LQ) Bottom 10 localities (LQ)

1. Wokingham 1.46 1. Barnsley 0.63 2. Reading 1.42 2. Tameside 0.62 3. Cardiff 1.39 3. NE Lincolnshire 0.61 4. Oxfordshire 1.34 4. Knowsley 0.60 5. London* 1.33 5. Kingston upon Hull 0.58 6. Newcastle 1.32 6. Sandwell 0.57 7. Cambridgeshire 1.31 7. Thurrock 0.56 8. Brighton and Hove 1.31 8. Blackpool 0.52 9. Windsor and Maidenhead 1.31 9. Blaenau Gwent 0.51 10. Trafford 1.27 10. Stoke on Trent 0.49

Source: Census of Population (2001). Note: * Combined NUTS1 region. tive class, as suggested by some of the literature re- chester in the northwest (Trafford lying just to the viewed earlier. It also suggests that different spatial west of the city centre with Manchester itself levels of analysis might be appropriate – in order, ranked only four places below at 14), and Newcas- for example, to zoom in on micro-level neighbour- tle in the northeast. Finally, our rankings conﬁrm hood effects not currently captured in the data. the perception of Brighton and Hove as a creative More generally, these issues highlight the impor- centre with its unique bohemian image and relative tance of investigating additional quality of place proximity to London.13 factors (quantitative and if possible qualitative) in Turning attention to the bottom ten UAs, a seeking to explain the distribution of the creative number of these are places suffering the protracted class; see the section on the creative class and qual- after-effects of the loss of heavy industry, either as ity of place below. distinct localities (Blaenau Gwent, Stoke on Trent, Table 3 provides a slightly different view of Cre- Barnsley) or the deindustrialized areas of large cit- ative Class distribution, in this case showing the ten ies, for example, Tameside (Manchester), Knows- highest and lowest unitary/county authorities, ley (Liverpool) and Sandwell (Birmingham). In ad- ranked with respect to their Creative Core LQs. dition, a couple of places associated with old-style There are in total 171 of these localities in England seaside holiday resorts also make an appearance and Wales, and this level of geography allows a lit- (northeast Lincolnshire, Blackpool). Although lo- tle more detail to become apparent than is the case cations in the left-hand column are generally asso- with NUTS3. Further to this end, the thirty-three ciated with higher levels of growth and prosperity, unitary authorities that comprise London have been the patterns revealed here do not necessarily imply combined into the single standard NUTS1 UK re- an economic consequence; for example, Blackpool gion; this provides a view of how London as a although the ninth least Creative (Core) locality in whole is positioned, but equally important is the England and Wales has had some success in recent fact that eighteen of the top twenty Creative Core years, reinventing itself from a traditional “bucket UAs in England and Wales are located in the capi- and spade” destination into one that seeks to attract tal,12 and so collapsing these into a single ﬁgure al- a younger party-oriented clientele. lows detail elsewhere to emerge. Figure 1 shows how the Creative Core are dis- As might be expected (see e.g. Hall et al. 1987), tributed within England and Wales, with London localities in the west-of-London M4 corridor area excluded from this representation for the reasons (Wokingham, Reading, Oxfordshire, Windsor and outlined above. A number of locations are high- Maidenhead) feature heavily in the top ten Creative lighted on the map; this is not necessarily a com- Core LQs. In addition to London, ranked at number prehensive listing of the highest or lowest ranked seven Cambridgeshire completes the third facet of places, rather they are intended to serve as illustra- the “Golden Triangle” of the UK’s knowledge tive examples. The concentration Creative Core in economy. What is perhaps more interesting is that the southeast of England generally and the M4 cor- in addition to those areas which might be expected ridor area in particular is apparent; within this area to feature, a number of less obvious regional cen- the Unitary Authority of Reading is highlighted, tres of creativity emerge – Cardiff in the west, Man- immediately to the southeast is Wokingham, and

Fig. 1. Location of the Creative Core in England and Wales (except London).

Windsor and Maidenhead lie to the east again (not this are apparent in Birmingham, with the district of shown). Sandwell immediately to the west placed ﬁfth low- The three cities of the East Midlands (Derby, est (see Table 3) while directly to the east lies Soli- Leicester and Nottingham) serve to illustrate the hull (twenty-second from the 138 Unitary Authori- way in which urban centres do not necessarily pos- ties and Counties outside London). A similar situa- sess a consistent relationship with their surround- tion is observed with respect to Manchester: Tame- ing hinterland in terms of where the Creative Core side to the east, Trafford to the (south) west, placed congregate – one of these cities (Leicester) has a in the bottom ten and top ten respectively of the Cre- signiﬁcantly higher concentration than the corre- ative Core ranking. The northeast of England pro- sponding county area, while in the other two cases vides another example of contrasting creative class it is approximately equal. There is a wider point to patterns existing in close proximity; this is however note on this in that as alluded to earlier, explana- different in one key aspect in that although adjacent tions of creative class distribution go beyond a sim- to one another, Newcastle and Sunderland are two plistic urban vs. rural issue, although this may play distinct cities with their own identities, not to men- some part in the process. It is also worth acknowl- tion rivalries. Newcastle’s much higher levels of edging that such idiosyncrasies do tend to get lost Creative Core may in part be assigned to its role as in a NUTS3 level or user-constructed city-region- a regional capital and administrative centre, but in type analysis. recent years it has seen extensive cultural develop- This type of pattern is further revealed when the ment and has acquired something of a boho image major cities are examined, in that localities which that is more complex to explain. make up the wider city area are often very different Focusing on those areas of England and Wales with respect to levels of the Creative Core, despite that are less well placed in terms of creativity, as quite small distances being involved. Examples of discussed above a number of larger city areas fall

Table 4. Quality of place indicators: overview (%).

England and Wales Highest locality Lowest locality

Openness (diversity) 8.9 36.2 1.6 Bohemians 2.1 8.8 0.8 Cultural opportunity 2.9 15.2 1.3 Public provision 9.0 17.6 5.4 Unemployment 5.1 10.5 2.1

Source: Census of Population (2001); Annual Business Inquiry, Employee Analysis (2002). into this category, along with distinct ‘post-indus- We do not suggest that all aspects of a concept trial’ regions such as the South Wales valleys, and as nebulous as quality of place can be perfectly cap- the former coalfield areas in the north of England. tured by these relatively simplistic statistics; it is It is notable that the former exist in relative prox- possible though to construct some general indica- imity to Cardiff, one of the highly placed cities for tors, underpinned by Creative Class theory – in this Creative Core in England and Wales, and it is sig- section we analyse the association between these nificant that of these Blaenau Gwent is one of the indicators and the location of the creative class in two valley areas that does not share a border with England and Wales. An analysis is conducted of the either Cardiff or Newport to the east (the other be- bivariate correlations between the individual qual- ing Merthyr Tydfil). ity of place indicators and the location of the creative class; we then combine these indicators into a single multiple regression model, which allows es- The creative class and quality of place timation of the overall explanatory value of these As outlined earlier, Florida suggests that the crea- variables on the distribution of the creative class. tive class is highly mobile, with strong preferences Table 4 shows summary values of the quality of for certain aspects of quality of place. In conjunc- place indicators for England and Wales. The distri- tion with Canadian researchers (see e.g. Gertler et bution of the bohemians has been discussed above al. 2002) he shows that in North America cities and so will not be commented upon here. Levels of with high levels of creative class tend to be open, diversity, defined as the percentage of residents that tolerant and diverse places, with high levels of rec- are foreign-born, shows quite considerable varia- reational and cultural opportunity. According to tion; the three highest values are accounted for by Florida (2002b) an open environment is one in London NUTS3 areas (of which there are five) with which people are accepted and allowed in, on the Inner London East being the highest. The highest basis of their skills rather than their similarity to the placed non-London locality is Leicester, a city with existing gatekeepers. As such, another indicator of a long tradition of immigration from the Indian tolerance we can employ is the proportion of bohe- subcontinent. This result does raise concerns over mians in any given location. This group often this particular indicator, and these are subsequently stands out compared to mainstream culture, life- discussed. Conversely, the Gwent Valleys is the style and values, and thus in order to thrive requires least diverse locality in England and Wales by this high levels of tolerance (Florida 2002b, pp. 260– measure, closely followed by the other Valleys area 261). The idea therefore is that bohemians seek the (central). opportunity to experience a diversity of impres- Cultural opportunity also sees wide variations, sions, and are thereby themselves indicators of a however, Inner London west is a massive outlier, tolerant and open environment (it may be recalled unsurprisingly, given the concentration of high pro- from previous sections that this view has been cri- file museums and galleries therein. Perhaps more tiqued as revealing a somewhat simplistic under- surprising is that the next highest placed locality is standing of the locational choices of arts- and cul- Blackpool at 4.1 per cent; again this highlights how ture-based employment; e.g. Markusen 2006; Wo- certain quality of place indicators can be influenced jan et al. 2007). Finally, to these indicators have by underlying factors which are not necessarily been added those of Public Provision and Unem- consistent – the Blackpool figure is almost certain- ployment (Social Cohesion) in order to reflect the ly largely derived from a high concentration of bars European context of our research. and amusements rather than the “high” culture

Fig. 2. Creative Core and Bohemi- ans (Boho LQ).

found in central London. Once again, the lowest integration, or for highly educated foreign-born value of the indicator (1.3%) is accounted for by the workers might, capture the openness of a commu- Gwent Valleys region. nity better. These are however restricted by data Finally, with regard to the Public Provision in- problems, an issue that forthcoming qualitative dex (PPI) it is quite hard to discern any pattern work will seek to address. However, from the above within the results, with two of the top three being we can tentatively conclude that the Creative Class cities of the East Midlands (Nottingham and and tolerance (measured as diversity, and the pres- Leicester respectively), separated only by Inner ence of bohemians) do correlate in the same way London west. Again, different forces are likely to that Florida and his associates found in the North be at work here. American analysis. Figures 2 and 3 show graphical As shown in Table 5, the indicators for both di- representations of the relationship between the versity and the bohemians are positively correlated Creative Class and respectively the bohemians and with the localization of the Creative Class and both openness. Figure 2 demonstrates a strong and pos- of its subgroups. This means that the creative class itive relationship as predicted, with the four in England and Wales tends to live in places that NUTS3 areas of London, along with Brighton and also have high levels of bohemians and diversity. Hove as signiﬁcant outliers (i.e. they have higher Both relationships are quite strong, particularly so concentrations of bohemians for their observed between the creative class and the location of the levels of Creative Core employment than is typi- bohemians. The Openness index is a fairly simplis- cal). This makes intrinsic sense given the locations tic measure of tolerance and as such might not be in question, but without additional data we can only sufﬁcient. A more focused measure on the effect of speculate on the relative importance of supply ver-

Table 5. Quality of place – bivariate correlations.

Correlation with Correlation with Correlation with Creative Class Creative Core Creative Professionals

Openness (diversity) 0.52** 0.44** 0.50** Bohemians 0.72** 0.58** 0.71** Cultural opportunity 0.52** 0.29** 0.56** Public provision 0.02 0.21* –0.07 Unemployment –0.31** –0.21* –0.33**

Source: Census of Population (2001), Annual Business Inquiry, Employee Analysis (2002). Notes: * Significant at the 95% level. ** Significant at the 99% level.

Fig. 3. Creative Core and Open- ness (diversity).

sus demand factors within this pattern. Figure 3 il- levels of employment in these sectors. Conversely, lustrates the point made earlier that despite a sig- unemployment (i.e. exclusion from the labour mar- nificant trend line, localities can possess identical ket) is one of the main manifestations of social ex- levels of openness (as we measure it), but very dif- clusion. For public provision, a significant and pos- ferent levels of the Creative Core. Different local itive (but weak) association is observed for the Cre- and regional trajectories (and thus potentially dif- ative Core only. For unemployment, the relation- ferent relationships) are likely to underpin this ob- ship is consistently negative, although moderate servation. and weak with respect to the Creative Core, mean- With regard to cultural opportunity, a positive ing that municipalities with high levels of unem- and statistically significant relationship exists be- ployment tend to have a low concentration of the tween the presence of the creative class in a loca- creative class. tion, and that location’s cultural and recreational Overall, it would appear that the creative class in offering. This finding is again compatible with England and Wales shows a similar pattern of dis- those of the North American research. This rela- tribution with respect to quality of place, as is ob- tionship is quite strong for the Creative Class as a served in North American cities. High concentra- whole and for the Creative Professionals, but some- tions are typically found in places which are toler- what puzzlingly weaker for the Creative Core. ant, diverse, bohemian, socially cohesive and The concepts of public provision and social co- which offer higher levels of cultural opportunity, hesion are actually quite abstract, and are therefore but the overall pattern is quite complex. somewhat difficult to operationalize in practice. In order to test a unified Creative Class model, The relative provision of public goods and services the multiple regression method is used. The ad- in healthcare and education is measured here by vantage of using this method is that all indicators

Table 6. Results of the combined model: summary.

Total Creative Class Creative Core Creative Professionals

Openness (diversity) + n/s + Bohemians + + + Cultural opportunity n/s – + Public provision n/s + – Unemployment – – – Adjusted R2 .664 .484 .698

Table 7. Results of the Combined Model: detail.

Unstandardized Standardized coefficients coefficients Collinearity statitics Model B Std. error Beta t Sig. Tolerance VIF

1 (Constant) 27.882 1.097 25.406 .000 Openness .143 .060 .218 2.360 .020 .345 2.897 Bohemians 3.235 .906 .414 3.572 .001 .221 4.530 Public provision –.423 .126 –.223 –3.356 .001 .670 1.492 Cultural opportunity .877 .265 .302 3.313 .001 .356 2.807 Unemployment –.903 .145 –.385 –6.231 .000 .776 1.289

Note: Dependent variable: Creative Professionals

Model summary

Model R R square adjusted R square std. error of the estimate

1 .844(a) .713 .698 2.19272

are tested in one model, and therefore it is possible fessionals. This means that, as we would expect to control for any multicollinearity between the from the theory of the Creative Class, wherever independent variables. A total of six models were these quality of place indicators are high, levels of produced; for each of the Creative Class and the Creative Professionals will also tend to be higher. two subgroups (the Creative Core and the Creative Moreover, from the standardized coefficients Professionals) two models were calculated for shown in Table 7 we can infer that of these varia- each dependent variable, including either all for- bles it is the presence of the bohemians that has the eign-born citizens or non-Western foreign-born greatest influence. With regard to the other two in- citizens as the Openness variable. These results dependent variables, as expected unemployment are summarized in Table 6: this reveals some var- is negatively associated with the location of the iation across the regressions: the model for the Creative Professionals; the nature of causality be- Creative Core had the lowest explanatory power hind this relationship is open to debate. On the one (.484), while that of the Creative Professionals hand it may represent an association with higher had the highest (.698). There was also some vari- levels of social cohesion, while on the other it ation across the quality of place variables: for ex- could be seen as purely labour market related in ample the Cultural Opportunity was negatively as- that creativity is a growing area of employment sociated with the Creative Core (the opposite di- and as such would be expected to coincide with rection predicted), and not significant within the lower unemployment. Finally, the public provi- Total Creative Class model. The Public Provision sion index is negatively linked to the distribution index was also not significant for the Total model, of the Creative Professionals, which is the oppo- but positively associated with the Creative Core site of what might be expected. This could be due (the direction predicted). to the nature of public sector employment in the The model shown in full in Table 7 – Creative UK, which tends to be proportionally higher in Professionals as the dependent variable with all less prosperous areas, reflecting a lack of private foreign-born citizens as a dependent variable – had sector jobs in combination with a conscious policy the greatest explanatory power. In total, this model of employment redistribution. On the one hand, explains around 70 per cent of the distribution of Tables 6 and 7 would appear to be prima facie ev- the Creative Professionals. All of the independent idence of systematic variation between sections of variables are significant at the 99 per cent level, the Creative Class; however, we need to be wary with the exception of Openness (95%). of data or other specification issues; this is a clear The location of the Bohemians, Openness and area for further research and one in which quali- the Cultural Opportunity index is positively cor- tative (casual) evidence will play a useful part in related with the localization of the Creative Pro- the near future.

Table 8. Indicators of prosperity – bivariate correlations.

Correlation with Correlation with Correlation with Creative Class Creative Core Creative Professionals

Population change 0.51** 0.29** 0.56** Employment change 0.27** 0.23** 0.25** Milken Tech-pole index 0.64** 0.49** 0.63** Employment change (hi-tech) 0.04 0.01 0.05 New firm formation 0.76** 0.46** 0.82**

Source: Office of National Statistics, various datasets14 Notes: * Significant at the 95% level. ** Significant at the 99% level.

The creative class and indicators of prosperity, tors, the relationship between the Creative Class growth and technology and share of “high-tech” employees is positive and Having examined where the creative class is locat- significant; this relationship is somewhat stronger ed in England and Wales, and how this distribution for the Creative Class in general and the Creative is associated with various indicators of quality of Professionals in particular than it is for the Creative place, attention is now turned to the relationship be- Core. This means that a high concentration of res- tween the creative class and basic indicators of ident creative class tends to be associated with the prosperity. Constraints of space prevent our report- presence of relatively high levels of employment in ing a full analysis of the geographical variations technology-based businesses. This preponderance within each of these indicators here; thus Table 8 for the creative class to be collocated with high- shows overall bivariate correlations between the technology activities is consistent with what is ob- three creative class groupings and the five indica- served in the North American research (Florida tors of prosperity, growth and technology. 2002a, 2002b, 2002c; Gertler et al. 2002). There is The association between the two general indica- however no evidence to suggest that the presence of tors of prosperity (i.e. population growth and em- the creative class (in any of its guises) is associated ployment growth) and the localization of the Cre- with any growth in employment in these activities, ative Class is statistically significant and positive. these correlations not being significantly different The relationship between Creative Class location from zero. This is on the face of it a major differ- and population growth is quite strong for both the ence between the UK and the US/Canada creative Creative Class as a whole and the Creative Profes- class results, and one that requires further investi- sionals, but weak for the Creative Core. For em- gation given the central tolerance and diversity be- ployment growth, the relationship is however quite gets creativity begets technology-based employ- weak for both the Creative Class and its two sub- ment growth theme of the Creative Class model. groups. Overall therefore, a high concentration of Finally, the correlation between the presence of the creative class tends to be found in places that the creative class (in all its forms) and the rate at have growing populations and rising employment. which new firms are created is positive and signif- This may be interpreted as evidence in support of icant. The relationship is strong with respect to the the ‘jobs follow people’ aspect of the Florida thesis; Creative Class as a whole, and for the Creative Pro- that is, that the creative class creates prosperity in fessionals but less so for the Creative Core. This general by its very presence. Bivariate correlations means that localities in which the creative class in can however only indicate association (i.e. covari- England and Wales is concentrated typically exhib- ance) and do not themselves imply any causal re- it higher levels of new firm formation; this is con- lationships between the variables involved. Put sistent with the Florida thesis but again open to in- simply, it may be that the creative class also follows terpretation with regard to causality. prosperity, rather than creating it. Conclusions regarding causality will need to be informed by more qualitative data, currently being gathered (at the Conclusions and issues for further research time of writing). Our first research question was concerned with With regard to employment in the Tech-pole sec- mapping the size and distribution of the creative

© The author 2008 78 Journal compilation © 2008 Swedish Society for Anthropology and Geography THE “CREATIVE CLASS” IN THE UK: AN INITIAL ANALYSIS class in England and Wales. In its widest defini- ators of symbolic, analytical, and synthetic knowl- tion it currently accounts for just over 37 per cent edge. A valid alternative (or indeed complement) to of all employment, yet the dispersion of the cre- this approach might be to conduct a series of parative class throughout England and Wales is allel focused studies on the distribution of narrowly somewhat uneven. High concentrations are typi- defined occupational groups; that is, to minimize cally an urban and metropolitan phenomenon, but the conflation of identities, processes and identi- not exclusively so, a number of inner city areas ties: the bohemians might represent the most ready- having some of the lowest rankings observed, in made starting point for such an exercise – averag- common with a number of localities and regions ing around 2 per cent of the labour force. Another afflicted by the long employment decline within way to achieve greater understanding could be to old traditional industries. It should be noted that explore life-stage effects of location choice, for ex- the level of geography employed does have an im- ample, as per the Markusen and Johnson (2006) pact here, and striking the balance between high- study of artists in Minneapolis. lighting distinct localities and functional labour More generally it may also be insightful to bring markets on the one hand and being able to observe together research on the hard and soft infrastruc- interesting variations and idiosyncrasies on the tures of a given place. For example, what types of other is not always easy. Particularly within the spaces, natural or built environments might aid (or larger cities, the findings of Markusen (2006) and hinder) milieu effects? At present these are essen- Wojan et al. (2007) suggest the need to investigate tially assumed to happen inside a spatial black box. the subtleties of bohemian and creative class dis- Indicators of prosperity, technology and growth tribution at the sub-metropolitan (or even neigh- are all significant and positively associated with the borhood) level. presence of the creative class, and again variations Attention was focused upon how variations in in the significance and strength of relationships quality of place are related to the location of the cre- were observed by creative class subgroups. The no- ative class. This analysis of this second research table exception to this pattern was technology- question showed statistically significant relation- based employment growth, which showed no asso- ships between the localization of the creative class ciation with the location of the creative class. This and these indicators. Tested individually, all the in- suggests the representation of both technology en- dicators were significantly correlated in the direc- trepreneurs and bohemians in the same places in tions hypothesized with the exception of the Public North America while in the UK, by contrast, bohe- Provision index, which was partially so. The mians are concentrated in the London area and in strength and significance level of these correlations a few other creative centres (Brighton, for exam- showed some variation by creative class type (i.e. ple); technology entrepreneurship is less common Creative Class as a whole, Creative Core and Cre- but may also be found in smaller university cities. ative Professionals). The combined quality of place Overall, we generally conclude that the creative regression model produced high levels of explana- class in England and Wales displays similar prop- tory power with independent variables largely sig- erties to those ascribed to it in the USA (Florida nificant in the direction suggested by Florida’s the- 2002a, 2002b, 2002c) and Canada (Gertler et al. ory: there are however some variations depending 2002), albeit with certain caveats as noted. It should upon which creative class subgroup is employed as also be noted here that although our quantitative re- the dependent variable. sults provide evidence of consistency with the Cre- Do these variations reflect “real” differences in, ative Class theory through numerous correlations for example, preferences for quality of place? Al- and associations, it is much more difficult to draw ternatively are they more likely to derive from data/ inferences relating to actual causality. This will be model specification issues? These observed differ- an issue that the qualitative part of the research ad- ences between the Creative Class groupings under- dresses, with qualitative interviews and case stud- line the potential value in seeking to unpack the ies designed to uncover the motivations and sometimes amorphous set of occupations included; thought processes behind locational choices (i.e. the knowledge base approach outlined by Hansen designed to probe the causations that underpin the et al. (2005) may represent a fruitful method of do- observed correlations). A good example of this is ing just that. Essentially this requires the re-coding the association of creative class and openness: does of the Creative Class dataset (or gathering of a new this correlation between the creative class and di- one) into (in their methodology) the users and cre- versity arise merely from a more or less coinciden-

© The author 2008 Journal compilation © 2008 Swedish Society for Anthropology and Geography 79 NICK CLIFTON tal location in certain metropolitan and urban areas creativity much in the same way that cities and re- with little or no interaction, is it primarily a kind of gions used to enter into zero-sum games chasing consumption relationship in which the creative old-style inward investment. Perhaps it would be class are able to access, for example, exotic ethnic an interesting exercise for further research to re- shops and a wide variety of interesting restaurants, visit the sites of creativity attraction programmes, or is it a reflection of the creative class’ genuine to systematically evaluate their impacts, success- preference to live and work in an open, tolerant and ful or otherwise. diverse community as Florida describes? Each of these outcomes would potentially be consistent with our findings so far. Acknowledgements On the issue of diversity, Fig. 2 suggests a Research was funded by the UK Economic and So- number of places that have arrived at similar levels cial Research Council as a project entitled ‘Tech- (as we are able to measure it) but almost certainly nology, Talent and Tolerance in European Cities: A for very different reasons, and with different out- Comparative Analysis’ (Grant No. RES-000-23- comes likely. Clearly the whole issue of diversity 0467), and we extend our gratitude to all the Euro- and tolerance needs further exploration; one possi- pean partners for their various efforts in ensuring bility is using the census data to examine variations the project was interesting and enjoyable for all within the overall foreign-born or non-Western fig- concerned. The author is grateful to the sponsors ures; at present it is not possible to distinguish be- for making this research possible. The author is tween places that have equal-sized foreign-born also thankful for the insightful comments provided communities originating from one or a hundred dif- by the Editor and anonymous referees. The usual ferent parts of the world: intuition alone would sug- disclaimer applies. gest that two such places are likely to be very different. In addition, on the wider point of measuring tolerance, one of the key insights of Florida and Notes Gates (2001) is the linking of gay and hi-tech indi- 1. At present data from Northern Ireland and Scotland are ab- ces, something not possible in the UK using sec- sent from our analysis. This omission will be addressed during subsequent iterations of the analysis. ondary data. Brighton is the classic example from 2. Markusen (2006) proposes that Florida’s results in the USA the UK of a locality whose position in the tolerance actually derive at least in part from human capital effects, rankings would be significantly affected if this al- rather than from creativity effects per se. However, short of ternative measure were used. To widen this point assessing all work-related creativity on an individual basis, proxy by occupation is the only way large-scale studies of out, Florida himself suggests that the USA is now this nature can be undertaken realistically. starting to lose its dominance in what he refers to 3. Peck (2005) also notes the use of disarming terminology as ‘the new global competition for talent’ (Florida that invites positive association, and asks who indeed would 2004); if this is indeed the case, what might be the seek to be rigid, narrow-minded and conservative when they could be flexible, open and creative? potential impact on both policy and potential eco- 4. Indeed, as Markusen (2006) herself notes, visual artists may nomic outcomes for Europe in general, and the UK (for example) have some different locational preferences to, more specifically? say, performing artists. Perhaps the most pressing research priority 5. There has been some work on Europe (Florida and Tinagli 2004) but this been at the national level only. Within this, though is the fate of the non-creative class, which with regard to creativity, the traditional large economies (in- after all comprises more than 60 per cent of the la- cluding the UK) are seen as losing ground, Southern Europe bour force. Will they be limited to those meagre continues to lag behind, while the Netherlands and the Nor- trickle-down benefits of creativity, or can every- dics forge ahead. 6. This involves the construction of a dataset that allows three one become more creative; and who will be re- distinct fields of analysis: examining how the Creative Class sponsible for making this happen? The urban model applies within each partner country, the performing leaders driving forward the Creative Class agen- of inter-country comparisons across partner countries, and da? National governments? Or will it be left to the finally a comparison between the “TTT group” as a whole and existing US and Canadian studies. Results relating to creative class themselves to “grow up” and take these latter two project aims will be reported elsewhere in responsibility as Richard Florida himself propos- due course. es? Analogously, there are similar questions to be 7. As we define it, the Creative Core maps very closely to the resolved regarding non-creative places. Peck major occupational group Professional Occupations, albeit with a small number of occupations therein placed in our (2005) warns against the creation of “cargo Creative Professionals category. cults”, vainly attempting to flag down footloose 8. Mainland European countries use the International Stand-

ard Classification of Occupations (ISCO) system; due to FLORIDA, R. (2002a): ‘Bohemia and economic geography’, the nature of the wider research being undertaken it was Journal of Economic Geography 2 (1): 55–71. with reference to this system that our occupational group- FLORIDA, R. (2002b): The Rise of the Creative Class. Basic ings (bohemian, creative class) were defined. In many cas- Books, New York. es ISCO maps directly to SOC2000, but in other cases it FLORIDA, R. (2002c): ‘The economic geography of talent’, An- does not. nals of the Association of American Geographers 92 (4): 743– 9. I.e. those people who have not worked more than one hour 755. during the short reference period regardless of whether or FLORIDA, R. (2004): The Flight of the Creative Class: The New not they are in receipt of unemployment related benefits. Global Competition for Talent. HarperCollins, New York. 10. Not in terms of absolute size – the USA and Canadian re- FLORIDA, R. and GATES, G. (2001): Technology and toler- gions typically being much larger – but rather having a sim- ance: the importance of diversity to high-technology growth. ilar role within the national context in question. Survey Series 1, Center on Urban and Metropolitan Policy, 11. This figure rises slightly to 0.28 for the bohemians only. The Brookings Institution, Washington, DC. 12. This statistic should not obscure the fact that massive vari- FLORIDA, R. and TINAGLI, I. (2004): Europe in the Creative ation exists within London, with some very low Creative Age. Carnegie Mellon Software Industry Center and Demos, Core LQs found therein (for example, Barking and Dagen- Pittsburgh, London. ham, at 0.51 the third lowest in England and Wales). GERTLER, M.S., FLORIDA, R., GATES, G. and VINODRAI, 13. It is worth noting here that although not a directly compa- T. (2002): Competing on creativity: placing Ontario’s cities rable analysis, there are generally some consistencies with in North American context. A report prepared for the Ontario the indices produced by the Demos think tank (Demos Ministry of Enterprise, Opportunity and Innovation and the 2003). Institute for Competitiveness and Prosperity. Munk Centre for International Studies, University of Toronto, Toronto, Nick Clifton ON. GLAESER, E.L. (1998): ‘Are cities dying?’, Journal of Econom- Centre for Advanced Studies ic Perspectives 12 (2): 139–160. Cardiff University HALL, P. and SOSKICE, D. (2001): ‘An introduction to varie- 44–45 Park Place ties of capitalism’, in HALL, P. and SOSKICE, D. (eds): Va- Cardiff CF10 3BB rieties of Capitalism: The Institutional Foundations of Com- parative Advantage. Oxford University Press, Oxford, pp. Wales 1–68. United Kingdom HALL, P., BREHENY, M., MCQUAID, R. and HART, D. E-mail: [email protected] (1987): Western Sunrise: The Genesis and Growth of Brit- ain’s Major High Tech Corridor. Allen & Unwin, Hemel Hempstead. HANSEN, H.K., VANG, J. and ASHEIM, B.T. (2005): ‘The cre- References ative class and regional growth: towards a knowledge based BRADFORD, N. (2004): Creative cities: structured policy dia- approach’, paper presented at the Regional Growth Agendas logue report. Background paper F45, Family Network, Cana- Conference, Regional Studies Association, Aalborg, 28–31 dian Policy Research Networks, Ottawa, ON. May. BROOKS, D. (2000): Bobos in Paradise: The New Upper Class HEATH, J. and POTTER A. (2006): The Rebel Sell: How the and How They Got There. Simon & Schuster, New York. Counter Culture Became the Consumer Culture. Capstone, BROWN, P. and LAUDER, H. (2006): ‘Globalisation, knowl- Chichester. edge, and the myth of the magnet economy’, Globalisation, JACOBS, J. (1961): The Death and Life of Great American Cities. Societies and Education 4 (1): 25–57. Random House, New York. CANNON, T., NATHAN, M. and WESTWOOD, A. (2003): JACOBS, J. (1969): The Economy of Cities. Random House, New Welcome to the Ideopolis. The Work Foundation, London. York. COOKE, P. and DE LAURENTIS, C. (2002): The index of knowl- LEITNER, H. (1990): ‘Cities in pursuit of economic growth: the edge economies in the European Union: performance rank- local state as entrepreneur’, Political Geography Quarterly 9 ings of cities and regions. Regional Industrial Research Re- (2): 146–170. port 41, Centre for Advanced Studies, Cardiff University, LUCAS, R.E. (1988): ‘On the mechanics of economic develop- Cardiff. ment’, Journal of Monetary Economics 22 (1): 3–42. CUSHING, R. (2001): Creative capital, diversity and urban MARKUSEN, A. (2006): ‘Urban development and the politics of growth. Unpublished paper, University of Texas at Austin, a creative class: evidence from a study of artists’, Environ- Austin, TX. ment and Planning A 38 (10): 1921–1940. DCMS (2004): Culture at the Heart of Regeneration. Department MARKUSEN, A. and JOHNSON, A. (2006): Artists’ Centers: for Culture Media and Sport, London. Evolution and Impact on Careers, Neighborhoods and Eco- DEMOS (2003): ‘Manchester is favourite with “new bohemi- nomies. Hubert H. Humphrey Institute of Public Affairs, Uni- ans”’ [WWW document]. URL http://www.demos.co.uk/ versity of Minnesota, Minneapolis, MN. media/pressreleases/bohobritain [accessed in November MARSHALL, A. (1920): Industry and Trade. Macmillan, Lon- 2007]. don. DEVOL, R., BEDROUSSIAN, A. and KIM, S. (2007): Best Per- PARR, J.B. (2005): ‘Perspectives on the city-region’, Regional forming Cities 2007: Where America’s Jobs are Created and Studies 39 (5): 555–566. Sustained. Milken Institute, Santa Monica, CA. PECK, J. (2005): ‘Struggling with the creative class’, Internation- FLORIDA, R. (2000): Competing in the Age of Talent: Quality of al Journal of Urban and Regional Research 29 (4): 740–770. Place and the New Economy. R.K. Mellon Foundation, Pitts- PORTER, M.E. (2000): ‘Location, clusters, and company strate- burgh, PA. gy’, in CLARK, G.L., FELDMAN, M.P. and Gertler, M.S.

(eds): The Oxford Handbook of Economic Geography. Ox- (2007): ‘Emoting with their feet: Bohemian attraction to cre- ford: Oxford University Press, pp. 253–274. ative milieu’, Journal of Economic Geography 7 (6): 711– THRIFT, N.J. (2001): ‘“It’s the romance, not the finance, that 736. makes the business worth pursuing”: disclosing a new market WONG, C. (2001): ‘The relationship between quality of life and culture’, Economy and Society 30 (4): 412–432. local economic development: an empirical study of local au- WOJAN, T.R., LAMBERT, D.M. and MCGRANAHAN, D.A. thority areas in England’, Cities 18 (1): 25–32.