Higher educated residential preferences and the marketing of private housing in the Amsterdam Metropolitan Area and Hong Kong

(1)

Higher educated residential

preferences and the marketing of

private housing in the Amsterdam

Metropolitan Area and Hong Kong

Thesis submitted as part of the Research Master Urban Studies, University of Amsterdam Author: Michael Stuart-Fox

Student number: 10011455

Email address: michaelstuartfox@yahoo.com Supervisor: Prof. Sako Musterd

Second reader: Dr. Marco Bontje Date of submission: 19 June 2015

(2)

Acknowledgements

The completion of this thesis would not have been possible without the help of others. Here I would like to thank Ilona Vierveijzer and Margriet Doorten of Onderzoek, Informatie en Statistiek (OIS) Amsterdam for their kind help in adapting and hosting the Hong Kong version of the online questionnaire and converting the dataset into a file analysable in SPSS. I know the data collection department of OIS has been very busy in recent times so I sincerely appreciate the time Ilona Vierveijzer and Margriet Doorten have put into helping me. I also thank OIS more generally for covering the financial costs that have been associated with both the time investment of staff members and the hosting of the online questionnaire.

I thank all respondents in the MRA for completing the questionnaire as part of the Higher Educated Location Preferences (HELP) project on which this thesis builds. I also very much thank all respondents in Hong Kong for taking the time to complete the Hong Kong version of the questionnaire. Knowing that employees in Hong Kong companies and organisations are generally very busy increases my appreciation for the time investment they made. I also appreciate the help of fellow students at The University of Hong Kong (HKU) who helped check the suitability of the questionnaire to the Hong Kong context and assisted in finding respondents. I also thank dr. Roger Chan at HKU for sharing his insights into the creative industries of Hong Kong and for practical help where needed. Last but not least, final words of thanks go to my supervisor prof. Sako Musterd for his very timely feedback, even if I sent chapters a little later than would be ideal, and to my father for his conscientious review of the grammar and spelling.

Amsterdam, 19 June 2015, Michael Stuart-Fox

(5)

Remarks

Photograph on cover page: modern high-rise and older low-rise residential buildings in Aberdeen, southern Hong Kong Island, November 2014. Photograph by author.

References in the text include page numbers of books, articles or reports in the following cases: a direct quote or paraphrasing, specific statistical data, and if a particular finding, statement or conclusion that is referenced to is contained in a text that is primarily about a somewhat different topic or theme. This makes it easier for the reader to look up the exact section of the article that is referenced to. An example is the reference in Chapter 5 (page 51) to the statement that rent and price levels in Hong Kong rise for residential units the higher they are from the ground. This statement can be found in an article by La Grange (2011) which is a study about class differences in various forms of neighbourhood attachment and not about the relationship between floor height and rent and price levels.

When the text mentions ‘Amsterdam’ without any further qualification, it always refers to the municipality of Amsterdam and not any wider entity or region. The Amsterdam Metropolitan Area (MRA) is always referred to by either ‘MRA’ or, occasionally, ‘Amsterdam region’.

In the text (a conversion into) ‘number of square metres’ is always provided alongside ‘number of square feet’ to ease interpretation.

In the text amounts in Hong Kong dollar (HKD) are not converted into amounts in euro to increase the readability of the text. As of 18 June 2015, 1 HKD = 0.1135 euro. Amounts in HKD can thus be roughly divided by ten and then increased by 10% to get the equivalent amount in euro. The HKD/euro exchange rate fluctuates significantly as the HKD is pegged to the US dollar and thus fluctuates with the changing USD/euro exchange rate.

Total number of words in Chapters 1 to 8, excluding all tables, the summary and appendices: 29,636 words.

(6)

1. Introduction and research questions

Introduction

The dwelling is a central feature in the daily lives of people. A dwelling provides shelter and security and is a place of relaxation. As the most important spatial node of individuals it also structures the spatial pattern of daily life in terms of where people study, work, play, recreate, shop, dine and many more regularly recurring activities. It also often expresses one’s societal status and identity. The choice of dwelling is thus an important life choice, even more so because dwellings are relatively durable and people usually spend many years and sometimes decades in one particular home. One’s residential choice is thus often not made overnight. Relocating is done on the basis of residential preferences that are continuously subject to change. These residential preferences are the topic of this thesis. Researchers in the social sciences have long sought to study what preferences people have, and equally important, which factors determine those preferences. It is now increasingly accepted that the traditional explanatory factors of, for example, household composition, income, and educational

attainment are not adequate anymore to explain residential preferences (e.g., Clapham, 2005). Various other concepts have been introduced as explanatory factors, such as lifestyles and values, to various degrees of success. This thesis continues with this approach and seeks to investigate the importance and meaning that occupational group has for residential

preferences by comparing workers in the creative sector with those in the high-tech sector. These two sectors are both growing and increasingly important in developed economies around the world in the twenty-first century, such as the Amsterdam Metropolitan Area (MRA) and Hong Kong, which are the two urban regions that I study in this thesis. This thesis is an international comparative case-study research in which the residential preferences of those in the creative sector and high-tech sector in both regions are compared in order to show what the influence of occupational group, other individual-level factors, and the urban context is on those housing preferences. It builds on the Higher Education Location Preferences (HELP) project carried out recently at the UvA and VU.

Research questions

The two main research questions are as follows: how do the stated residential preferences of workers in the creative and high-tech sectors in the Amsterdam Metropolitan Area and Hong Kong differ and what is the influence of individual factors and the urban context in this regard? And to what extent do residential property developers in these two urban regions appeal to target groups’ identities and lifestyles in their promotion of private housing? The main research questions can be dived into four parts which correspond to the four following subquestions:

1) What are the stated residential preferences in terms of relative location,

neighbourhood and dwelling of workers in the creative sector and high-tech sector in the Amsterdam Metropolitan Area and Hong Kong and how do they differ between the two occupational groups?

This first subquestion encompasses the quantitative-descriptive part of the thesis where the stated residential preferences in terms of relative location, neighbourhood characteristics, and dwelling characteristics are described and interpreted for the two occupational groups in the two urban regions.

(7)

2) What is the influence of individual- and household-level factors on the stated residential preferences of workers in the creative sector and high-tech sector in the Amsterdam Metropolitan Area and Hong Kong?

This second subquestion encompasses the quantitative-explanatory part of the thesis where the effects of individual- and household-level factors, including occupational group, on residential preferences will be examined through regression analyses. Separate analyses will be made for the two urban regions (two samples).

3) What is the influence of urban contextual factors on the stated residential preferences of workers in the creative sector and high-tech sector in the Amsterdam Metropolitan Area and Hong Kong and on expected differences between the two regions?

This third subquestion encompasses the qualitative-exploratory part of the thesis where previous research findings, statistical data, and qualitative data/information on the

characteristics of the urban context (built environment, commercial and cultural facilities, transport network, housing market characteristics) of the Amsterdam Metropolitan Area and Hong Kong will be used to explain expected differences in residential preferences between similar occupational groups in the two different urban regions. No own statistical analyses will be done here and this part also relies on existing qualitative data/information whereas the second subquestion relies exclusively on the quantitative datasets. This third subquestion is exploratory in the sense that it does not attempt to come up with definitive conclusions regarding the influence of the urban context given the fact that contextual factors other than those included in the analysis might also influence residential preferences and differences between the two urban regions.

4) To what extent do residential property developers in the Amsterdam Metropolitan Area and Hong Kong appeal to target groups’ (presumed) identities and lifestyles in their promotion of private housing projects, what differences can be discerned between the two urban regions, and how are these related to the urban context ? This fourth subquestion encompasses an additional qualitative-exploratory part of the thesis in which I conduct a content analysis of online promotional material – dedicated websites and online information documents, brochures, photographs, renderings, videos – that property developers use to market residential developments with private rental or owner-occupied housing units. The aim is to examine whether, and to what extent, property developers in their promotion emphasise and appeal to the (presumed) identities and lifestyles of target groups and to explain expected differences in this regard between the two urban regions. The rationale of and theoretical expectations for this fourth subquestion are further described in Chapter 7.

Theoretical expectations that I have for the first three subquestions are discussed in Chapter 2.

Scientific relevance and aims

The main aim of this thesis is to contribute to the relatively scarce body of literature on the stated residential preferences of occupational groups within the creative class – as Florida

(8)

(2002) defines it – at the intra-metropolitan level. In relation to the existing literature, the aim is to make the following two main contributions. Firstly, I will compare the Amsterdam Metropolitan Area to Hong Kong, and these kinds of cross-country and cross-cultural comparisons are quite rare in residential preferences research. Secondly, I will compare two occupational groups that are both part of Florida’s creative class. As will be detailed in Chapter 2, there have been several valuable studies that have compared various of these occupational groups, but the overall literature on this topic is still fairly limited.

According to Clapham (2005) there is little focus in housing research on how constraints and opportunities of residential choice come about and only limited understanding of their interrelationships with preferences and attitudes. According to him there is little empirical work that investigates and attempts to understand how different households perceive and react to housing market contexts facing them (2005: 10). I intend to make a small contribution to address this gap since I am specifically interested in how the constraints and opportunities of urban contexts influence residential preferences. According to MacLennan (2012) and Van Ham (2012), researching the preferences and beliefs that underpin household housing choices is less common in housing research than investigating the choices themselves. In housing research, especially housing economics, most empirical studies explore how variation in constraints shift choices, they do not study how preferences differ between certain groups and how they change and are formed. Since this thesis is focused on stated preferences I intend to add to the literature on stated housing preferences which is less developed than the literature on revealed preferences. Lawson (2012) states that analyses at the national level predominate in international comparative housing research, and that more attention should be paid to local and regional policies and effects. Hong Kong and the MRA are both subnational regions. My thesis will thus also make a contribution to address this gap in the literature.

Societal relevance

The societal relevance in a direct, practical sense is limited since the results of this thesis are not expected to be used by a government or other stakeholder. In a more indirect sense, the societal relevance stems from the fact that there has not been much research on the residential preferences and behaviour of the creative class, and subgroups within that broad group, with regards to their intra-metropolitan preferences. This means that governments and real estate developers do not have clear enough insights into what type of housing in what kind of neighbourhoods and at what kind of locations are most popular with certain occupational groups that have increased in size in recent history and will probably continue to do so in the foreseeable future. An attractive, diverse and affordable metropolitan housing market is, after all, generally seen as a crucial condition within a wider set of conditions that are important for a metropolitan region to have to be able to attract and retain creative class occupational

groups (Scott, 2010: 125-26).

Reading guide

The following Table 1.1 gives an overview of the chapters and indicates which subquestion is answered in which chapter. Subquestions 1 and 3 logically fit together in one chapter because the analysis of the influence of urban contextual factors for subquestion 3 is based on

descriptive statistics showing the residential preferences of workers in the creative sector and high-tech sector in the two urban regions, which is the focus of subquestion 1.

(9)

Table 1.1: Reading guide Chapter

number Chapter title Which research questions are answered?

1 Introduction and research questions -

2 Theoretical framework, conceptualisation, and

operationalisations -

3 Research design, cases, and methodology -

4 Descriptions of samples in the MRA and Hong Kong - 5 Stated residential preferences, differences between

the MRA and Hong Kong, and the influence of urban contextual factors

Subquestions 1 and 3

6 Stated residential preferences and the influence of individual- and household-level factors on residential preferences in the MRA and Hong Kong

Subquestion 2

7 Online marketing of private housing in the MRA and

Hong Kong Subquestion 4

8 Conclusion and discussion Main research questions

(10)

2. Theoretical framework, conceptualisation, and

operationalisations

This chapter first outlines the theoretical framework of the thesis. On that basis it presents the main conceptualisation of the concept ‘residential preferences’ and subsequently its

operationalisations, followed by the conceptual model of the thesis. The chapter ends with the theoretical expectations I have regarding the empirical findings.

2.1 Lifestyles, values and residential preferences in modern

societies

Traditional housing research and housing as a cultural practice

For decades, residential preferences and residential choice have been explained by a relatively limited set of traditional socio-economic and socio-demographic factors. Socio-demographic variables include household composition, number of household members, and age of

household member(s) or the household head. Socio-economic variables include the level of educational attainment, level of household income, (un)employment status, and the number of hours of work per week (Boumeester, 2011). In the last two decades or so, an increasing number of scholars have argued that these traditional variables are not adequate anymore in explaining and predicting residential preferences and behaviour, primarily because of increasing wealth and greater differentiation in household types and consumption behaviour (Jansen, 2011 ; Clapham, 2005 ; Maclennan, 2012 ; Nio, 2012 ; Coulter et al., 2015). For the Netherlands specifically, the VROM-raad (2009: 30-1) also mentions ‘ontzuiling’, or the eroding of the political-denominational segregation of society, as an additional contributing factor. Indirect contributing factors are also that increasing car ownership and digital

technologies which facilitate working from home have made living near one’s workplace less important, giving households the opportunity to look at larger spatial segments of the housing market (VROM-raad, 2009: 35). All these developments taken together render less possible any common cultural (dis)position that can be ascribed to, for example, a particular age group (e.g., people in their thirties), ‘household composition’ group (e.g., two-person households) or income group (Clapham, 2005: 224). This has also led to, and is simultaneously the result of, a process of increasing differentiation in the design and planning of residential properties and environments. A concomitant development is the increasing use of ‘branding’ to market certain neighbourhoods or housing projects (Meier, 2013 ; Reinders, 2012).

Related to these developments, it is also increasingly recognised by scholars that housing choice in developed economies – where most households have at least some sort of

meaningful choice – has increasingly become a cultural practice by which people signify their societal status and class and personal taste (Meier, 2013 ; Maclennan, 2012 ; Butler and Hamnett, 2012 ; Lawson, 2012 ; Lawrence, 2012 ; Nio, 2012 ; Van der Land, 2003 ; Bridge, 2006). This conceptualisation of choosing and living in housing as a cultural practice that, among other things, serves the goal of distinction, is one of the most significant developments in housing studies in the last few decades (Butler and Hamnett, 2012: 160 ; VROM-raad, 2009: 27). Preferences for specific locations or architectural styles can serve as a means of displaying distinction in Bourdieu’s sense, which is especially evident in themed residential neighbourhoods such as the Mediterranean-styled Le Medi in Rotterdam and the Dutch Golden Age-styled Brandevoort in Helmond (Meier, 2013 ; VROM-raad, 2009: 37). In general, the higher the cultural capital – here defined as ‘institutionalised cultural capital’ or

(11)

the (educational) qualifications one has – the more importance people attach to the symbolic and representational qualities of products rather than its instrumental characteristics (Meier, 2013: 95). In general over the last few decades, the symbolic, representational, and

experiential meaning of products and objects used in daily life has become more important, and this is also true for housing (Clapham, 2005: 135 ; Giddens, 1991: 198 ; Nio, 2012 ; VROM-raad, 2009: 36). The aesthetic, experiential, and identity aspects of dwellings and their surrounding residential environments have become more important, in addition to the ‘classic’ functional aspects of housing (VROM-raad, 2009: 3).

Reflexivity, lifestyle and the role of housing

Residential preferences and choice in contexts such as the MRA and Hong Kong take place in societies that are characterised by both an increased capacity of individuals to choose and exercise their own lifestyles and a greater need felt by individuals to differentiate from others’ lifestyles (Clapham, 2005: 221). Giddens (1991: 6) calls this the process of ‘life planning’ or ‘life politics’. This ‘planning’ takes place in modern social life which Giddens sees as being characterised by two fundamental characteristics: an institutionalised reflexivity and an expansion of disembedding mechanisms. Reflexivity means that individuals are constantly reflecting on their past, present and future lives and assessing ways they might change their lives to better suit their values and goals. Giddens terms this the ‘reflexive project of the self’ (1991: 20). The notion of lifestyles takes on a particular importance in this regard: the more tradition loses hold, the more individuals are forced to reflexively choose lifestyles among a generally increasing diversity of options. Lifestyles are routinised practices, the “routines incorporated into habits of dress, eating, modes of acting, and favoured milieu for

encountering others” (idem, italics mine). Significantly, Giddens observes that “lifestyles are characteristically attached to, and expressive of, specific milieu of action” and that “lifestyle options are thus often decisions to become immersed in those milieu at the expensive of possible alternatives” (1991: 83). A key ‘milieu of action’ is one’s residence. “Where a person lives, after young adulthood at least, is a matter of choice organised primarily in terms of the person’s life-planning” (1991: 147). This perspective on residential choice can also be found in more practical-oriented research. The second report in the study series ‘Monitor Creative Kennisstad’ by the Dienst Ruimtelijke Ordening (DRO), for example, concluded that creative knowledge workers in Amsterdam continually evaluate whether their place of residence still fits their lifestyle (DRO, 2008).

The other fundamental characteristic of late modernity, according to Giddens, is the process of disembedding which involves the “lifting out of social relationships from local contexts and their recombination across indefinite time/space distances” (1991: 242). These

disembedding mechanisms entail, among other things, transport and communication

technologies that make social contact with others less dependent on spatially fixed locations. This means that those locations of social contact and activity that are by definition fixed in space, acquire a higher degree of importance for expressing one’s identity and status. The most prominent of those locations is the residence, characterised by durability and fixedness in space and by having an important structuring and ordering influence on daily life activities (Coulter et al., 2015: 8), which set it apart from other locations where social interaction occurs.

Savage, Bagnall and Longhurst (2005: 155) also conclude that those social practices that continue to rely on spatially fixed spaces and places are most likely to be tied up with claims to social distinction compared to those practices that do not rely on spatial fixity. Savage et al. (2005: 179-80) contrast certain forms of consumptive behaviour such as watching television

(12)

and reading (not spatially fixed) with choosing and living at one’s residence (spatially fixed). According to Savage et al., housing, because of its spatial fixity, is among the ‘fields’

(defined as ‘domains of life’ such as work, leisure, residence etc.) in modern society which are most tied up with claims to distinction (2005: 12). Savage et al. (2005: 207) even write that “one’s residence is a crucial, possibly the crucial identifier of who you are”, and that the spatial sorting process by which people choose to live in certain places and others leave is at the heart of contemporary battles over social distinction.

Bridging the gap: new concepts in residential preference and choice research To bridge the gap between research based on the classic demographic and socio-economic variables and the increasing diversity in residential preferences and options, and to do justice to the conceptualisation of housing as a cultural practice, researchers have come up with ‘new’ explanatory factors that add to the explanation of residential preferences. An example is ‘place experience’ in a particular type of residential environment. Feijten et al. (2008) and Blaauboer (2011) show that the residential environment during childhood is strongly associated with the current residential environment and that individuals show a strong similarity to their parents and siblings in their residential environment, also after accounting for residential inertia (not moving) and return migration.

More controversially, researchers have also increasingly investigated whether lifestyles might help explain residential preferences (VROM-raad, 2009: 33). Jansen (2011), Pinkster and Van Kempen (2002) and Heijs et al. (2009) have explored the usefulness of the concept of

lifestyles for research into residential preferences and they all identify a host of problems with its usage. First of all, there is no consensus about the definition of lifestyle although most include consumption, activity and value patterns. Lifestyle typologies are not always explained clearly and there can still be substantial heterogeneity in residential preferences within a particular lifestyle group. Moreover, individuals belonging to different lifestyle groups can still share similar residential preferences. It is also not always clear how certain lifestyle characteristics can be ‘translated’ into residential preferences (Pinkster and Van Kempen, 2002 ; VROM-raad, 2009). Despite this, typologies of lifestyle and value are still used by municipalities, housing associations, and commercial developers (Nio, 2012). Empirical quantitative research by Pinkster and Van Kempen (2002) shows that in the

Netherlands the traditional socio-demographic and socio-economic variables are still the most important determinants of housing preferences while lifestyle variables – in terms of activities in leisure or free time – only marginally improve the explanation in statistical analyses. Empirical research by Heijs et al. (2011) finds that lifestyles – in terms of contacts with neighbours and participation in the neighbourhood – of tenants of an Eindhoven-based

housing association are not related to the large majority of measured housing preferences. In a somewhat different approach, Meesters (2009) investigated, on the basis of a nationwide survey of mostly home-owning households with above-average gross household income in the Netherlands, the meanings that the dwelling has for residents in a city centre, suburban and rural type of residential environment. She found that those living in city centres were three times more likely to mention ‘going out’ (to the cinema, theatre, concert or restaurants and cafes) and ‘fun shopping’ as everyday activities than those living in suburban or rural environments, after controlling for the conventional individual- and household-level

variables. Meesters concluded that the ‘meanings’ people attach to activities carried out in the residential environment, influence the likelihood that people conduct those activities, after controlling for socio-demographic variables and dwelling and residential environment features. In all three types of residential environments ‘going out’ was connected to the

(13)

meanings ‘relaxation’ (value type ‘hedonism’) and ‘social contacts’ (value type ‘benevolence’). ‘Going out’ among city centre dwellers was also associated with the meanings ‘getting away from things’ and ‘taking a break from work’ (corresponding to the value type ‘stimulation’), which was not the case for suburban and rural respondents. This shows that the desire for (places of) stimulation in the residential environment is greater among city centre residents than the other groups, which helps explain their preference for amenity-rich environments (2009: 160). More generally, Meesters shows how the different meanings that features of the dwelling, neighbourhood and location have for different groups of people can help explain their residential preferences.

Other researchers have investigated the importance of latent lifestyles – in terms of the values and attitudes one holds – for the explanation of residential preferences and choice. Research by Aero (2006) in Denmark shows that the meaning and function of the home in family life, which differs between people, influences the choice for residential environments regarding the density and height of buildings. Jansen (2011) found, controlling for the influence of traditional socio-demographic variables, that people who attach importance to the value type ‘hedonism’ (gratification, pleasure) had a relatively strong preference for living in a lively neighbourhood. Further research by Jansen (2014) compared two ‘value groups’: those in the ‘security’ group who find the safety, harmony and stability of relationships and of oneself of great importance and those in the ‘self-direction’ group who find freedom, independency, self-respect and creativity important values. She found that those in the ‘security’ group have a stronger preference for living on the city edge or in a small municipality, a newly-built dwelling, a more homogenous residential composition, and a neighbourhood with only housing compared to those in the ‘self-direction’ group after controlling for the conventional individual-level factors. The overall importance or added value of ‘value types’ for predicting preferences was relatively limited, however (Jansen, 2014).

Nijënstein et al. (2015) conclude on the basis of a survey among students living in student housing in Breda and Tilburg, and controlling for socio-demographic variables, that those who indicate that openness to change is an important value in their lives show a more than average concern for living near or in the city centre, likely because the many facilities and activities in an urban centre help satisfy the value types ‘stimulation’ and ‘hedonism’ that they find particularly important as guiding values in life. Finally, other support for the use of meanings as explanatory factors comes from Coolen (2008). When ‘meaning of tenure’ was added as independent variable to his analysis of intended tenure choice, the explained variance of tenure preference was substantially increased. ‘Meaning of tenure’ became the second-most important explanatory factor, after ‘age’ but ahead of ‘current tenure’ (2008: 106). Those who associated (one of) the meanings ‘one’s own home’, ‘independence’, ‘freedom’ and ‘financial security later in life’ to tenure were more inclined towards owning, whereas those who associated the meaning ‘(lack of) maintenance’ to tenure were more inclined towards renting (2008: 108).

In conclusion, there are mixed results as to the value of lifestyles – either conceptualised as behaviour or as underlying values and attitudes – for the explanation of residential

preferences and choice. Conventional socio-demographic and socio-economic variables are certainly of enduring importance.

(14)

2.2 Occupational groups, orientation of capital, and residential

preferences

Orientation of economic and cultural capital

Besides lifestyles, scholars have also researched whether and to what extent ‘occupational group’ is related to residential choice in any meaningful way. The use of occupational group has two major advantages over the use of lifestyles: someone’s occupational group is

relatively stable over time and is easier to measure.

De Wijs-Mulkens (1999) looked at the residential location of five occupational groups within the Dutch ‘cultural and economic elite’: writers and visual artists, professors, judges and public prosecutors, members of the boards of directors of major companies, and chartered (registered) accountants. She finds that among her respondents four groups could be distinguished on the basis of a) the orientation of the capital that is used to signify status – either with an emphasis on material assets (income and property as part of economic capital) or with an emphasis on cultural assets (knowledge and cultural knowhow as part of cultural capital) – and b) the level of ‘seniority’ in the upper class, with individuals being either newcomers into the upper layer of society or having already ‘arrived’ in the upper layer by virtue of their father’s position. These four groups are all overrepresented in a specific type of residential environment. The individuals whose capital is culturally oriented and have been born in the upper class are overrepresented in a type of residential environment that has a relatively high density – is relatively urban in that regard – and is exclusive in the sense that the environment is more homogenous with respect to the population in socio-economic terms. Those whose capital is also culturally oriented but are newcomers to the upper class are overrepresented in a type of residential environment that has a relatively high density but is less exclusive in the sense that the population is more mixed socio-economically. This pattern is mirrored among the two groups whose capital is materially oriented: those who are born in the upper class are overrepresented in a type of residential environment that has a lower density – is less urban – and is homogenous and exclusive whereas those who are newcomers to the upper layer of society are overrepresented in a type of residential environment that also has a relatively low density but is less homogenous and exclusive. The fifth residential environment that is distinguished, villages in rural areas, are popular with both groups whose capital is materially oriented.

That the orientation of one’s capital helps explain residential preferences has also

demonstrated more recently by research by Boterman (2012) who investigated the residential behaviour of middle-class couples expecting and welcoming their first child and who were living in the inner-city boroughs (within the A10 ring road) in Amsterdam. Controlling for a set of individual- and neighbourhood-level variables, he finds that couples who possess relatively high cultural capital – the attainment of a university degree by the spouses and their parents; an education in the humanities, arts or social sciences; having a knowledge or

creative job; and visiting classical concerts, museums, galleries and theatres – have a lower propensity to leave the inner-city boroughs than couples who possess high economic capital (gross annual household income) and low cultural capital. The latter group leaves the inner-city boroughs more often to live in a suburban environment. High cultural capital leads to “attachment to the city in both symbolic and social ways” according to Boterman (2012: 2410).

Metaal (2011) focused on spatial identities, i.e. the degree to which people identify with urban, suburban and village environments. He distinguished between the lower class, the

(15)

cultural middle class and the economic middle class. The cultural middle class consists of people who have completed higher education in the social or cultural sciences and work in those fields. The economic middle class, by contrast, consist of people who have completed higher education in law, economics and finance, and medical or natural sciences and work in those fields. His research in the case-study areas of Oosterparkbuurt in Amsterdam, the neighbourhood Ridderveld in Alphen aan den Rijn, and the village of Kockengen (Utrecht) shows that those belonging to the cultural middle class self-identified less with the suburb and more with the inner city as types of residential environments than did the economic middle class and lower class. The cultural middle-class self-identified more strongly than the other classes with being a ‘new urbanite’.

The capital of workers in the creative sector in the MRA and Hong Kong is more culturally oriented because of the content of their education and, especially, profession. On the basis of the findings by De Wijs-Mulkens, Boterman and Metaal, they can be expected to have a stronger urban residential preference than those in the high-tech sector.

2.3 Studies on the intra-metropolitan residential preferences of

creative class occupational groups

Foreign studies

To the best of my knowledge only a few studies have been carried out regarding the intra-metropolitan residential preferences of occupational groups within the creative class (in Florida’s (2002) definition). Lawton et al. (2013) have, on the basis of a survey, studied the residential preferences of workers in the creative industries (game and software development, publishing, and radio, TV and film production) and knowledge-intensive industries (law, accounting, finance, R&D and higher education) in Dublin. They concluded that both groups primarily take dwelling size and cost, distance and commuting time to the location of

employment, and the accessibility of public transport and roads into account when making a decision to live in a particular neighbourhood in the Dublin area. For these workers in the creative and knowledge-intensive industries ‘hard’ location factors were thus found to be more important than ‘soft’ location factors such as local ‘amenities’ (restaurants, cafes, cultural facilities), social atmosphere, demographic diversity, perceived tolerance, or the perceived authenticity of neighbourhoods. Frenkel et al. (2013) conducted a similar study of workers employed at various high-tech companies and firms in the financial services industry in Tel Aviv and concluded that for them the affordability of dwellings and the distance and commuting time were the primary factors that they took into account when making a decision for a particular neighbourhood within the Tel Aviv metropolitan area. Cultural amenities were of secondary importance. The research also showed the importance of lifestyle: knowledge workers with a ‘culture-oriented lifestyle’ (who visit cultural activities relatively often) had a stronger preference for living in the centre of the Tel Aviv area than knowledge workers with a ‘home-oriented lifestyle’.

Woldoff et al. (2012) studied the residential preferences (upon graduation) of undergraduate students in creative majors (art, music, theater, and graphic design) versus those in other majors (accounting, statistics, math, and business) studying at a “large public (state) university located in a rural state in the Mid-Atlantic region of the U.S.” (not further specified). Creative undergraduates showed a substantially stronger preference than

noncreative undergraduates for living near the ‘bohemian cultural amenities’ of coffee shops, used bookstores, tattoo parlors, live music venues, antique stores, vintage clothing stores,

(16)

organic food markets, and libraries. They also showed a considerably stronger preference for living in an urban residential environment after graduating than those in the other majors.

Dutch studies

In the Dutch context, Atzema en Snel (2007) studied the residential preferences of individuals belonging to the creative class (in Florida’s (2002) definition) in the ‘Noordvleugel Utrecht’ (cities of Utrecht, Hilversum, Amersfoort and neighboring municipalities). They concluded that the presence of entertainment and cultural facilities was of secondary importance for the choice of a particular neighbourhood in comparison with the most important factors of ‘green spaces’, primary facilities (supermarket, school), public safety, and nearby employment. This was also the case among the reference group of individuals not in ‘creative class occupations’. Only in the city centre of Utrecht were entertainment and cultural facilities among the five most important reasons to live there for the creative class.

The third report (2011) in the aforementioned study series ‘Monitor Creative Kennisstad’ by DRO investigated the aspects of Amsterdam as a place of residence that ‘creative knowledge workers’ and other higher educated workers (all resident in the municipality Amsterdam) attach importance to (DRO, 2011). DRO used the definition of ‘creative industries’

formulated by TNO on the basis of the Standaard Bedrijfsindeling (SBI) to define ‘creative knowledge worker’. Creative knowledge workers were those working in the arts (performing arts, museums, art galleries, public libraries), media and entertainment (publishing, radio and television, film production and cinemas, and games) and creative business services (public relations, advertisement, architecture, industrial design, fairs and conference organisation) (CBS, 2010: 4 ; OIS, 2013: 10). In terms of revealed preferences, the study (2011: 8) showed that a larger proportion of creative knowledge workers live in districts within the A10 ring road (84%) than higher educated workers not in creative occupations (hbo or higher), which for example includes workers in the high-tech sector as defined and used in this thesis (72%), and lower educated workers (lower than hbo) (62%). The 2009 Regionale Enquête

Beroepsbevolking of OIS (using the same definitions of occupational groups) shows a similar pattern: Stadsdeel Centrum is more popular with creative knowledge workers (23%) than non-creative higher educated workers (18%) and lower educated workers (10%) (DRO, 2011: 8). The study also looked into stated preferences. Of the respondents that expressed a wish to move home, creative knowledge workers had a somewhat stronger preference for staying within the A10 ring road (56%) than non-creative higher educated workers (49%) and lower educated workers (46%). However, for all three groups, a location within the A10 was the most favoured location. For a move outside of Amsterdam, creatives had a stronger

preference for Haarlem than the other categories. In terms of preferred tenure there were no meaningful differences between the groups. Regarding the type of dwelling, creatives had a somewhat lower preference for single-family dwellings and apartments on the ground floor and a somewhat higher preference for multi-floor apartments. In terms of neighbourhood facilities, creative workers attached more importance to public places such as parks, libraries and cafes than non-creative higher educated workers. Creatives also attached somewhat more importance to street markets and somewhat less importance to sports facilities and (car) parking places than the non-creatives. For other neighbourhood facilities, the differences between the groups were smaller (2011: 26). Creatives also had a somewhat stronger preference than non-creatives for neighbourhoods with a functional mix (living, working, recreation). In terms of facilities and amenities on a city-wide level, creative knowledge workers found theatres and music venues, cinemas, museums and art galleries, restaurants and

(17)

cafes, and monuments and historical architecture somewhat more important than

non-creatives. The observed differences were not substantial, but they were consistent (2011: 28). Earlier research among different categories of knowledge workers whose location of

employment is in the municipality of Amsterdam, reported by Musterd (2004 ; 2006) and Musterd and Deurloo (2006), also showed that the urban orientation varies between the different categories of workers. Of all studied occupational groups, architects working in Amsterdam lived there most often (71%), followed by employees of the municipal

government (60%). People working in the advertising sector (52%), in higher education social sciences (52%) or humanities (50%) or law (47%), and in the media sector (48%) also lived in Amsterdam relatively often. Those working in higher education medical sciences (37%) or the natural sciences (35%) or economics (33%) in Amsterdam lived in the city less often. Accountants 19%), workers at banks (19%) and workers in ICT firms (17%) lived in the city least often. Within Amsterdam, the aforementioned occupations in the creative sector

(advertising, architecture) were overrepresented in the central and ‘most urban’ areas of the municipality, revealing a preference for inner-city living. The workers in the business and financial services and ICT sectors had a relatively stronger preference for suburban locations either within the Amsterdam municipality or within neighbouring (suburban) municipalities in the Amsterdam region. Workers in medical sciences, natural sciences and economics had an orientation in between the other two groups (Musterd, 2004 ; 2006 ; Musterd and Deurloo, 2006).

2.4 Conceptualisation and type of preferences: stated and

intra-metropolitan

This thesis continues with the dominant approach in housing studies of the last few decades that goes beyond the traditional socio-demographic and socio-economic variables in

explaining residential preferences and choice. More specifically, it is part of the approach used by scholars mentioned in Sections 2.2 and 2.3 who have researched the relationship between occupational group, or a similar concept such as ‘orientation of capital’, and

residential preferences and behaviour. On the basis of, among others, Giddens (1991), Savage et al. (2005), and Clapham (2005) I conceptualise residential preferences as expressions of the wishes of individuals regarding their place of residence that take into account not only

‘functional’ concerns such as housing size, affordability, and access to employment but also ‘lifestyle and identity concerns’ in such a way that one’s residence, the surrounding

neighbourhood and its relative location are congruent with one’s preferred activity pattern and with the identity one would like to signify. The operationalisations of ‘residential preferences’ fit this conceptualisation (see Section 2.5).

Most research on the residential preferences and behaviour of the creative class or subgroups has focused on the preferences at the inter-metropolitan level, i.e. which metropolitan regions attract the creative class and which factors are the most important pull factors. By

comparison, research into the residential preferences of the creative class at the intra-metropolitan level has been more scarce. This thesis studies those intra-intra-metropolitan

preferences. Researchers can study either the stated preferences or the revealed preferences. Stated preferences are stated intended or hypothetical preferences expressed in response to survey questions. This is the type of preferences studied in this thesis. Research into revealed residential preferences, by contrast, tries to identify preferences on the basis of where people actually live. An assumption of this approach is that people’s actual observed housing choices

(18)

reflect their preferences. Both approaches to the study of residential preferences have their own merits and demerits (Jansen et al., 2011 ; Lennartz, 2013: 150). Which approach is most suitable depends on the specific research question at hand and on the type of data which is available. A difficulty of revealed preferences research is that one usually does not know which characteristics of a particular neighbourhood or type of residential environment attracted households to live there. Additionally, the assumption that people’s actual housing choices reflect their preferences is in some cases not valid (Boumeester, 2011). An advantage of stated preference research is that one can ask questions that are difficult to answer with revealed preference research – i.e., how important is it to live near facility X – and that one can ask respondents to rank various residential environments, neighbourhoods or types of housing. A difficulty with stated preferences is that it can be difficult to know for sure whether the preferences respondents report take into account household financial constraints and housing market constraints or not. Households might adjust or completely abandon their ‘real’ preferences to what is possible within their realistic choice set, which is a form of cognitive dissonance reduction (Van Ham, 2012: 48-9 ; Coulter et al., 2011: 2757). In housing markets such as Hong Kong, where some types and locations of housing are very expensive, this is a potential problem when interpreting the results. This problem is partly overcome in this thesis by the fact that the samples in both regions consist of respondents that have relatively high household incomes (see Chapter 4) and thus have more realistic choices on the housing market (Meesters, 2009: 252).

A related ‘problem’ with stated preferences research is that one does not know how meaningful or valuable the stated preferences are in terms of predicting future behaviour (Lennartz, 2013: 150). The studies by (among others) Lu (1999) and De Groot (2011) on moving intentions, residential preferences, and actual moving behaviour in the United States and the Netherlands respectively showed that most people do not actually move house when they state that they intend to move within a particular time period and vice versa. Moreover, many people do not realize their stated preferred options, for example in terms of tenure or dwelling size. One might expect this issue to be particularly pronounced in a tight and expensive housing market with substantial supply-side constraints, such as in Hong Kong. Moreover, these discrepancies between intention and behaviour are not the same for all population subgroups, rather they are found to be selective on the basis of, for example, current tenure, age and income (Lu, 1999 ; Coulter, 2013) and type of motive(s) for moving. People are more likely to actually move if they have ‘targeted’ reasons like employment opportunities or independent living than if they want to relocate for more ‘diffuse’ reasons relating to dwelling and (especially) neighbourhood characteristics (Coulter and Scott, 2014). Another issue that complicates residential preferences research is the level at which the decision-making process of residential mobility occurs. The online questionnaire used to measure residential preferences is completed by individuals, not, as is relevant in multi-person households, collectively by the decision-makers who sometimes do not hold the same

preferences or views on the desirability of moving in the first place (Coulter et al., 2012). The issues described in this paragraph make clear that one should not ascribe predictive qualities to the results in this thesis.

Research on residential preferences can also be distinguished on the basis of their measurement approach: a compositional approach versus a conjoint approach. In the compositional approach each dwelling attribute (garden, number of rooms, size etc.) is measured separately whereas in the conjoint approach housing profiles are measured, which are bundles of attributes (house A versus house B with the two houses having different features) (Coolen, 2008). My approach is the compositional approach: the online

(19)

questionnaire consists of questions which separately measure the preferences of respondents regarding various separate features of the residence.

2.5 Operationalisations

Residential preferences in this thesis are measures in various different ways: there is no single operationalisation. I have used the operationalisations that had already been formulated as part of the Higher Educated Location Preferences (HELP) project at the UvA and VU and that had been incorporated in the questionnaires for the Amsterdam and Eindhoven cases. These operationalisations are well suited to internationally comparative research because they are not dependent on or relevant only to a particular urban context. The operationalisations can be divided in five different types, explained below. The complete Hong Kong version of the questionnaire, which includes the exact wording of all questions, statements and answer categories, can be found in Appendix A.

Type 1: Absolute importance of aspects of and amenities in the urban region and residential neighbourhood

In this type, respondents are asked to indicate the level of (absolute) importance they attach to thirteen aspects and amenities of the urban region and nine aspects and amenities of the neighbourhood, which are described in Chapter 5.

Type 2: Relative ranking of importance of features of the dwelling, neighbourhood, and relative location

In this type, respondents are asked to rank seven features of the dwelling, rank seven features of the neighbourhood and rank seven features of the relative location, which are described in Chapter 5. Respondents also indicated which three of the twenty-one characteristics would be most decisive in the case of moving. The advantage of relative rankings is that one avoids the problem of respondents who find almost everything important: respondents are forced to think about which dimensions for them take on greater importance than others. For both type 1 and type 2 operationalisations, all dimensions are relevant for both the MRA and Hong Kong. Type 3: Assessment of chance of moving and reasons for (not) being likely to move In this type, respondents are asked to assess the chance of moving home within the next two years, on a 0-100% scale. Respondents with an assessment of 51% or higher are asked to indicate the reasons, out of a list provided, for being likely to move. Those with an assessment of 50% or less are asked to indicate the reasons for not being likely to move.

Type 4: Dwelling size, type, and tenure

In this type, respondents are asked to indicate the size – in terms of square feet or square metres – of the dwelling and the type and tenure they would look for in the case of moving home.

Type 5: Geographical location and ranking of images of residential environments and names of towns and neighbourhoods

In this type, respondents are asked to indicate which general areas of the urban region they would primarily consider in the case of moving home. They are also asked to rank the names of ten neighbourhoods and towns in the region and to rank ten images of different residential environments. The ten neighbourhoods and towns vary primarily in terms of their urban or suburban location within the region as a whole and the types of housing typically found in

(20)

those areas. Respondents, of course, could have other associations – such as population composition or available facilities – with these locations as well and choose largely on that basis. The ten images of residential environments vary primarily in terms of the types of housing and the wider built environment found there. The locations and images are further described and depicted in Chapter 5 and Chapter 6.

2.6 Conceptual model

The (first) main research question is focused on the effects of individual- and household-level factors and urban contextual factors on residential preferences. These are the two sets of independent variables of which I investigate their influence on housing preferences . The individual- and household-level factors are: age, gender, household size and composition, country of birth, level of educational attainment, level of household income, and occupational group (creative versus high-tech sector). The effects of these factors are uncovered through multivariate regression analyses in Chapter 6. The urban contextual factors are geography and compactness of the urban regions; relative attractivity of various transport modalities; supply, spatial location and usage of cultural facilities; historical inner-city and diversity in the built environment; and housing market features such as types of housing available, affordability, and government policies for home-buyers and tenants. The effects of these factors, which are more elaborately described in Chapter 5, are explored through comparing the MRA and Hong Kong on these dimensions in Chapter 5. Figure 2.1 contains the conceptual model of this thesis by schematically showing the relationships between independent factors and dependent variables. The two sets of independent factors are shown on opposite sides of the figure because I do not investigate the effects of these two sets in the same type of analysis: the regression analyses per urban region are explanatory whereas the comparison between the MRA and Hong Kong on the contextual dimensions is more exploratory in the sense that it does not statistically test the effect of these contextual factors or attempt to come up with definitive conclusions regarding the influence of the urban context. It should be noted that the conceptual model of Figure 2.1 is not relevant to subquestion 4. To avoid confusion, no separate conceptual model for that subquestion is made.

Figure 2.1: Conceptual model of this thesis

(21)

2.7 Definitions and importance of the creative and high-tech

sectors

The ‘creative sector’ and the related concepts of ‘creative industries’ and ‘cultural industries’ have been defined in various ways by different researchers and governments (Foord, 2008), leading to different conclusions as to the importance of these sectors to urban and regional economies (Markusen et al., 2008). An abstract definition of the ‘creative sector’ is: a sector where producers are “concerned with producing and marketing goods and services that are permeated in one way or another with broadly aesthetic or semiotic attributes” (Scott, 2000a: 2). Almost always definitions of cultural or creative industries include at least the following subsectors: architecture, advertising, design, publishing, fashion, and the production of music, (news) television and film (OIS, 2013 ; CBS, 2010 ; BOP Consulting, 2010: Kloosterman, 2004 ; Power, 2002 ; Scott, 1997 ; 2000b ; Denis-Jacob 2012). This is also the definition of ‘creative sector’ that will be used in the thesis although the exact delimitation of the ‘creative sector’ is not of great importance since the sampled workers in the MRA and Hong Kong do not ‘cover’ all the aforementioned subsectors. Detailed information on the sample is given in Chapter 4 and Appendix B.

Compared to the creative sector, the high-tech sector has been studied less in urban and economic geography so there are fewer studies on its composition and definitions. The high-tech sector can also be defined in various ways but almost always includes at least the following subsectors: the electronics, semiconductor, nanotechnology, robotics and ICT (software development) industries. Again, the precise delimitation is not of great importance since the sampled respondents in the high-tech sector do not ‘cover’ all subsectors within the high-tech sector (see Chapter 4). The creative and high-tech sectors as defined above are both leading sectors of what Scott terms ‘cognitive-cultural capitalism’ (Scott, 2008 ; 2014). This notion refers above all, according to Scott (2008: 64), to the fact that labour processes in general have come to depend more and more on intellectual and affective human assets. Metropolitan areas such as the MRA and Hong Kong are relatively important locations for the creative and high-tech sectors and are therefore appropriate and relevant cases for a study into the residential preferences of some of the workers in those sectors. In the MRA, ‘creative industries’ as defined by TNO (see earlier paragraph) make up 5% of total value added and 6.2% of total employment in the region, whereas nationally those shares stand at 2.6% and 3.3% (Gemeente Amsterdam Economie, 2015: 169). In the municipality of Amsterdam these shares are even higher than in the MRA, with 6% of total added value and 10% of total employment (OIS, 2013: 5) The advertising sector, in which all MRA questionnaire respondents are employed, is the most important subsector within the ‘creative business services’ segment of ‘creative industries’, with 58% of all jobs in that segment (OIS, 2013: 49). Seen from the Netherlands as a whole, creative business services are most strongly concentrated in the COROP region ‘Groot-Amsterdam’ (Raspe et al., 2012: 54-55). There has been less research on the development and spatial distribution of the high-tech sector in the MRA so it is more difficult to establish what kind of level of specialisation (concentration) the MRA has in this sector, but it is almost certainly less strong than with the creative sector. The sector ‘cultural and creative industries’ (CCI’s) in Hong Kong comprises advertising, architecture, arts, antiques and crafts, design, film, video and music, performing arts, publishing, and television and radio (CSD, 2014a: 5 ; 2015a). This sector has grown from accounting for 3.8% of GDP in 2005 to 5.1% in 2013 and from accounting for 5.1% of employment in 2005 to 5.6% in 2013 (idem). All questionnaire respondents in the creative

(22)

sector in Hong Kong – who are employed in the subsectors advertising, architecture, design, and television – are part of this CCI definition. Advertising accounted for 9.1% of the total employment within CCI’s, architecture for 7.3%, design for 7.1% and television and radio for 2.9% (CSD, 2014a: 8-10). For the high-tech sector the situation is more complex. Some Hong Kong respondents in this sector are part of the so-called ‘innovation and technology’ cluster, defined by the Census and Statistics Department as covering “research and development (R&D) activities as well as activities relating to the commercialisation of R&D outputs. They drive the development and introduction of technologically new or significantly improved products or processes in relevant organisations for commercial purposes”( CSD, 2014a: 8-10). This cluster in 2013 accounted for 0.7% of GDP and 0.9% of total employment and has been growing in the last few years (idem). Other high-tech respondents in Hong Kong are

employed in the ICT sector, more specifically computer application development. In 2014 a total of 31,400 employees were employed in IT/software development, which constituted 0.81% of total employment in Hong Kong, up 8.7% compared to 2012 (HKTDC, 2015). A final group of high-tech respondents in Hong Kong is employed in higher education and academic research (The University of Hong Kong) but since this sector is so broad in terms of employees it is not meaningful to mention those statistics here.

2.8 Theoretical expectations

On the basis of the studies cited above, I expect that preferences for characteristics of inner-city or urban living are more strongly present for creative workers than for high-tech workers. I expect that in Hong Kong the differences in residential preferences between workers in the creative sector and high-tech sector are not as pronounced and clear as in the MRA. I have two main reasons for this expectation. The first is that suburban living in Hong Kong seems to be relatively more attractive because of Hong Kong’s general compactness, the

well-developed, time-efficient and extensive public transportation network, and the lack of a historic core (attractive monumental architecture) in the city centre. The second reason is that I expect workers in Hong Kong to be more concerned with ‘hard’ location factors and

‘functional’ concerns and considerations, primarily housing affordability since that constitutes more of a problem there than in the MRA. These expectations relate to the first three

subquestions (Chapter 1).

The theoretical expectations that I have regarding the emphases in online marketing of private housing in the MRA and Hong Kong are formulated in Chapter 7 where they are most useful to the reader.

(23)

3. Research design, cases, and methodology

This chapter first outlines the various features of the research design: the case selection process and the type of cases will both be discussed. Thereafter, the research design will also be discussed with reference to the dimensions of cross-case versus within-case analysis and causes-of-effects versus effects-of-causes research mentioned by Goertz and Mahoney (2012). Subsequently, the methodology of the thesis will be outlined by explaining the methods of data collection and data analysis.

3.1 Case selection and type of case study

The two cases are the Amsterdam Metropolitan Area (MRA) and Hong Kong (Hong Kong Special Administrative Region or SAR). Neither of these entities are physically contiguous urban areas, but rather metropolitan areas that consist of one core with several subcentres. Table 3.1 below summarizes the main characteristics of the two regions. Maps of the MRA and Hong Kong are provided in Chapter 5 where they are most useful to the reader. Table 3.1: Basic characteristics of the MRA and Hong Kong, year 2014 unless otherwise stated

Feature Amsterdam Metropolitan Area Hong Kong

Total area 2580 sq. kilometres 1104 sq. kilometres Total land area 1604 sq. kilometres 1045 sq. kilometres Total population 2,367,809 people 7,264,100 people

Population density 1476 inhabitants per square km. 6571 inhabitants per square km. Number of households 1,130,289 households 2,449,400 households (1/2015 –

3/2015) Average household size 2.15 persons 2.9 persons Labour force (employed +

unemployed) 1,161,000 persons 3,919,800 persons (2/2015 – 4/2015) Employed labour force 1,056,000 persons or approx. 91% of

labour force 3,792,300 persons or approx. 97% of labour force (2/2015 – 4/2015)

Source: MRA: OIS, 2014. Hong Kong: CSD, 2014b.

These two cases were originally chosen for pragmatic reasons related to the Research Master Urban Studies programme. I found the courses at the University of Hong Kong interesting, I had a degree of familiarity with Hong Kong as a city, and Amsterdam as home base was a logical case to be included in the comparison. In addition, the questionnaire data for the MRA case were already available since they had previously been collected for the UvA-VU Higher Educated Location Preferences (HELP) project. But from a theoretical perspective it can be argued that Amsterdam and Hong Kong are relevant cases to compare. Amsterdam and Hong Kong would be, in the terminology offered by Gerring (2007: 89-90), diverse cases because these two cases exhibit variation on several contextual factors which I expect to influence residential preferences (subquestion 3) and the marketing of private housing (subquestion 4). Gerring (2007: 97-8) explains that diverse cases are cases that exhibit variation (high and low values) on “relevant dimensions”. These dimensions can be either explanatory factors (X-variables), outcomes (Y-variables) or X/Y-relationships. In the (first) main research question formulated in Chapter 1, the influence of contextual factors (a group X-variables) on

residential preferences constitutes an important element. The ‘dimensions of difference’ on which the diverse-cases analysis in this thesis rests are threefold. Firstly, the urban

contextual factors have different values in the two regions. These differences are relevant to both subquestion 3 and subquestion 4. Secondly, I expect the outcome (i.e., residential preferences of similar occupational groups) to be different in the two regions. And, thirdly, I

Higher educated residential preferences and the marketing of private housing in the Amsterdam Metropolitan Area and Hong Kong