Why people move : a quantitative analysis into the moving intentions of the Dutch population

Author:

Sean-Paul Stotesbury

Student:

10352090

Supervisor:

Dr. Aslan Zorlu

Faculty:

Graduate School of Social Sciences

Programme:

M.Sc. Human Geography

Why people move

:

2 GSSS

Graduate School of Social Sciences

M.Sc. Human Geography; Urban Geography Track University of Amsterdam

Geo Track: Master Thesis Project Human Geography: Urban Geography Course code: 735420022W.AJ.S01.GEO Course year: 2016/2017

Coordinator: J. (Jeroen) van Pelt (studieadviseur-GPIO@uva.nl) Supervisor: Dr. A. (Aslan) Zorlu (A.Zorlu@uva.nl)

Assessor: Dr. W.P.C. (Wouter) van Gent (W.P.C.vanGent@uva.nl) Sub-group: Quantitative approaches, cross-sectional designs Student: S.C. (Sean-Paul) Stotesbury

Address: Kuipersstraat 7, 1074EE Amsterdam Tel: +31 6 555 33 410

E-mail: seanpaulstotesbury@hotmail.com Student number: 10352090

3

ABSTRACT (EN)

The key feature of this piece of research is the moving intention of citizens in Dutch households. The objective of the research will be to provide an overview of the determinants of people in Dutch households’ intention to move house, related to the life course, whereby a more specific goal will be to ascertain the key factors that play a role in young people’s moving intentions and the differences that exists among various age groups. A quantitative statistical analysis was applied, in order to ascertain to what extent the various factors are determinants of moving intentions. The WoON 2015 survey data on which the statistical analysis was performed, was supplied by Statistics Netherlands (CBS). It features over 60,000 respondents across the whole of the Netherlands, thus it was possible for the scope of this piece of research to comprise the entire Dutch population. Both descriptive and regression analyses were carried out. Following the statistical analyses it became apparent that a number of factors are strong determinants of moving intentions, the interaction of these factors with age proves to be key. In terms of future research, it would be of great interest to carry out an in-depth qualitative analysis into moving intentions by means of multiple case-studies in various urban and rural locales, as a limitation of this piece of research is the lack of qualitative insights on a micro-level.

Key words: moving intentions, life course analysis, WoON 2015, demographics, housing, neighbourhood effects, age, quantitative statistics, logistic regression.

4

ABSTRACT (NL)

Het doel van dit onderzoek is om inzicht te bieden in de factoren die de verhuisintenties van de Nederlandse bevolking bepalen, gerelateerd aan de life course en de mate waarin deze verschillen per leeftijdsgroep. Hiertoe is een kwantitatieve statistische analyse toegepast, omwille van het achterhalen in hoeverre bepaalde factoren invloed hebben op de verhuisintenties van de Nederlandse bevolking. Deze statistische analyse is toegepast op data van het WoON-onderzoek van het Centraal Bureau voor de Statistiek (CBS). De dataset omvat meer dan 60.000 respondenten verspreid over het gehele land. Dit faciliteert het grote onderzoekskader van deze scriptie, welke de gehele Nederlandse bevolking omvat. Er zijn zowel beschrijvende analyses als regressieanalyses uitgevoerd. Uit de statistische analyses is gebleken dat meerdere factoren van invloed zijn op de verhuisintentie, waarbij een aantal sterke determinanten naar voren komen, en de interactie met leeftijd van groot belang blijkt. Wat betreft de mogelijkheden voor toekomstig onderzoek, zou het zeer interessant zijn om een diepgaande kwalitatieve analyse uit te voren in meerdere stedelijke en niet-stedelijke omgevingen. Een beperking van dit onderzoek is het ontbreken van kwalitatieve inzichten op microschaal.

Sleutelwoorden: verhuisintentie, mobiliteit, life course analysis, WoON 2015, demografie, woonbeleid, leeftijd, kwantitatieve methoden, logistische regressie.

5

PROLOGUE

The following thesis marks the completion of the Human Geography programme at the University of Amsterdam. I have thoroughly enjoyed the programme, gained a lot of knowledge and met many interesting people.

Many thanks towards my supervisor, Aslan Zorlu, for his input and comments during the thesis project.

A further word of gratitude towards my colleagues in the Valuations department at Colliers Amsterdam, for the use of the facilities and the welcome distraction during the writing process. I sincerely hope that you, the reader, are able to enjoy reading this thesis and gain new insights regarding the subject of residential mobility.

Best regards,

6

1 TABLE OF CONTENTS

1.1 List of tables and figures ... 7

2 Introduction ... 8

3 Theoretical framework ... 10

3.1 Residential decision making process ... 10

3.2 Life course ... 12 3.3 Contextual factors ... 17 4 Methodology ... 20 4.1 Research question ... 20 4.2 Research design ... 21 4.3 Dependent variable ... 23 4.4 Independent variables ... 24 4.5 Hypotheses ... 29 4.6 Conceptual model ... 33 5 Analysis ... 34 5.1 Descriptive analysis ... 34 5.2 Regression analysis ... 43 6 Conclusion ... 51

6.1 Limitations and recommendations ... 52

7

1.1 LIST OF TABLES AND FIGURES

Table 4.2: Classification and frequency of Moving intention. ... 23

Table 4.3: Classification and frequency of Educational attainment. ... 24

Table 4.4: Classification and frequency of Age category. ... 24

Table 4.5: Classification and frequency of Ethnicity. ... 25

Table 4.7: Classification and frequency of Disposable income. ... 25

Table 4.8: Classification and frequency of Household composition. ... 25

Table 4.9: Classification and frequency of Tenure. ... 26

Table 4.10: Classification and frequency of Type of dwelling. ... 26

Table 4.11: Classification and frequency of Crowding. ... 26

Table 4.12: Classification and frequency of Attachment to dwelling. ... 27

Table 4.13: Classification and frequency of Sense of community. ... 27

Table 4.15: Classification and frequency of Degree of urbanization. ... 28

Table 4.16: Categorisation and frequency of Grey pressure... 28

Table 5.1: Interpretation of Cramér's V (University of Toronto, undated). ... 34

Table 5.2: Two-way frequency table of Moving intention and Educational Attainment. ... 35

Table 5.3: Two-way frequency table of Moving intention and Age. ... 36

Table 5.4: Two-way frequency table of Moving intention and Ethnicity. ... 37

Table 5.6: Two-way frequency table of Moving intention and Household income. ... 37

Table 5.7:Two-way frequency table of Moving intention and Household composition... 38

Table 5.8: Two-way frequency table of Moving intention and Tenure. ... 38

Table 5.9: Two-way frequency table of Moving intention and Type of dwelling. ... 39

Table 5.10: Two-way frequency table of Moving intention and Crowding. ... 39

Table 5.11: Two-way frequency table of Moving intention and Attachment to dwelling. ... 40

Table 5.12: Two-way frequency table of Moving intention and Sense of community. ... 41

Table 5.13: Two-way frequency table of Moving intention and Grey pressure of municipality. ... 41

Table 5.14: Two-way frequency table of Moving intention and the degree of urbanization in municipality. ... 42

Table 5.15: Variable overview per regression model. ... 43

Table 5.16: Measures of association per regression model. ... 44

8

2 INTRODUCTION

Why do people move? Moving house is a significant occasion in every person’s life. It can have all manner of reasons, ranging from the birth of a child to the completion of one’s education. The reasons for moving house can therefore be influenced through various events that take place throughout the course of one’s life, as well as factors related to one’s household characteristics, neighbourhood characteristics or current dwelling.

In the current economic climate, it is becoming increasingly hard for young people to find suitable, affordable housing. Moving house is a key life event for young people, which in many cases takes place more than once during the youthful years. Currently, there is a lack of insight into the factors that determine moving intentions, and more specifically the moving intentions among young people in the Netherlands.

The key feature of this piece of research is the moving intention of citizens in Dutch households. The objective of the research will be to provide an overview of the determinants of people in Dutch households’ intention to move house, whereby a more specific goal will be to ascertain the key factors that play a role in young people’s moving intentions and the differences that exists among various age groups.

By utilising a quantitative statistical analysis, the key determining factors of people’s moving intentions will become apparent, as well the relative importance of each factor. An analysis will be carried out in order to find out whether these determinants differ for young people in the Netherlands, whereby the differences between various characteristics will be analysed in relation to each other, such as educational attainment, level of income and the household composition.

In order to carry out what is stated above, the following research question has been formulated: - To what extent are the moving intentions of the Dutch population influenced by the life

course, and how do these differ for the young population?

My interest in this topic arose as a result of the combined experiences in my professional career in the real estate sector and academic career in urban geography. In modern society, issues related to housing are highly relevant in both academia and the real estate markets. In the real estate sector, I have noticed a lack of appreciation of the underlying factors affecting the housing market. The focus of the sector is very much on ‘hard’ factors such as house price increases and investment yields, as opposed to personal and contextual factors, which are also likely to play a key role in relation to these ‘hard’ factors.

Furthermore, being a young resident of the Netherlands myself, I am very interested in the moving intentions and differing personal and life characteristics of my contemporaries. I have therefore chosen to specifically look into the moving intentions of young people in the Netherlands, following a broad view of the entire population.

9 I believe that insight into people’s moving intentions could offer valuable information for the real estate sector, to provide a context for understanding the fluctuations in the real estate market and offer insight into the characteristics of arguably the most important actors in the housing market, which are the residents. For example, when planning new developments, it could be highly beneficial in the planning phase to know what type of housing to build, or how to price the rents in order market the property in an optimal manner. As an urban geographer, I believe I am well placed to provide these insights.

In addition to providing useful insights for the private sector, this piece of research can provide a significant contribution to society on both a municipal scale and a national scale. Information regarding the influence of neighbourhood characteristics on people’s mobility intentions could offer policy makers valuable insights when planning housing developments, urban renewal schemes or the allocation of social housing. Especially in relation to the preferences of young people, as this is the group that is most likely to benefit from improvements to such policy developments, due to the propensity for young people to move house.

With regards to academia, I believe that there is a lack of in-depth research carried out based on the Dutch WoON survey data. As this dataset provides, in relative terms, an large cross-section of the Dutch population, I believe it provides high potential for an in-depth statistical analysis and a subsequent opportunity for a relevant contribution to the available research based on the WoON 2015 data, because similar research based on this dataset has not yet been carried out.

10

3 THEORETICAL FRAMEWORK

The following section of this thesis will provide an overview of the literature and theories that are deemed to be relevant in relation to moving intentions. Various authors’ work has been reviewed in order to form a sound theoretical foundation to base this piece of research on. The various theories have been laid out below and structured according to the overarching theme of the literature. A mix of seminal literature from the field of residential mobility and moving intentions, and more recent literature will be provided.

The reason for providing an overview including historic literature, is that it provides a strong foundation for understanding the processes and context of residential mobility and moving intentions. The recent literature provides an overview of the current state of the available knowledge and theories regarding residential mobility and moving intentions. Following the theoretical framework, the research methods and operationalisation will be discussed, after which expectations will be formed based on the theoretical framework, regarding the outcome of the statistical analysis.

3.1 RESIDENTIAL DECISION MAKING PROCESS

With regards to mobility intentions, Kley (2011) proposes a framework that makes use of psychological theories about decision-making. The framework that Kley (2011) proposes entails the linking of a sequence of stages in the migration process to certain phases of the decision-making process. The division of stages that Kley (2011) proposes is: the pre-decisional phase (in which one considers migration), a post-decisional phase (in which one plans migration) and the moving phase (in which the mobility intention becomes actual behaviour).

Kley (2011) elaborates further on the various stages and hypothesizes certain causes of the phases. The pre-decisional phase of considering a move is mainly a result of the perception of opportunities and the differences in these opportunities between the current place of residence and alternative places of residence. The perception of these differing opportunities, and the accompanying mobility intentions are most likely to coincide with changes in the life-course. Examples of such changes are the commitment to a partner, the birth of a child, or the abandonment of a job. The post-decisional phase, in which the intent to move has been established and the planning of the move takes place, is mainly caused by the anticipation of such life-course events. The actual realisation of the move can be viewed as a result of planning the move. The opportunity differentials and life-course events mentioned above are only of an indirect influence on this stage.

Kley (2011) also states that there are a number of socio-economic variables that likely have a stronger influence on mobility intentions than life-course characteristics. For example, a higher level of education could enhance the possibility of mobility intentions due to the availability of adequate jobs being more spatially concentrated, and therefore necessitating residential moves in order to gain suitable or desirable employment.

11 Coulter et al. (2011) also show that education level is an indicator of moving intentions. In their research, they found that education level was one of the strongest indicators linked to desiring and expecting a move, whereby the highly educated are the most likely to hold these feelings. A key reason that Coulter et al. (2011) mention, similarly to Kley (2011), is that career progression in highly skilled professions often requires spatial flexibility. Another possible explanation put forward by Coulter et al. (2011) is that highly educated people can be more spatially aware than those with low levels of education, and therefore may show an increased likelihood to move due the spatial knowledge they have.

In further research, Coulter & Van Ham (2013) emphasize the relationship between education (and knowledge) and residential mobility. Similar to the premise of spatial awareness mentioned above, the authors state that household moving decisions are influenced by information asymmetries with regards to housing information. To put this simply, households with a low level of education and a lack of knowledge are less likely to be able to seek out the relevant information that is necessary in order to move house. An example of such information is financial information regarding mortgages and housing costs, whereby knowledgeable and highly educated households are more likely to be able to seek out relevant information to keep costs to a minimum. These households are also not likely to feel daunted by the process of information gathering, whereas households with a low level of education may be reluctant to engage in the information gathering process.

Coulter & Scott (2015) follow up on the previous research (Coulter et al., 2011; Coulter & Van Ham, 2013) by examining why people desire to move and how these desires affect their subsequent moving behaviour. They find that concerns about the local area strongly influence moving desires, but are not prominent determinants of actual moving behaviour. Additionally, they state that highly targeted reasons for desiring to move, such as employment opportunities, are far more likely to lead to an actual move than diverse area preferences. They suggest that factors that ‘push’ people towards forming moving desires are not the same as those ‘pulling’ them to choose particular dwellings and locations.

This assertion is subsequently linked to the life course concept, whereby Coulter & Scott (2015) emphasize the importance of the biographical timing of mobility decision-making. They show that employment reasons and the desire for independent living are important and highly targeted motives for young people, but for middle aged and older people it is more diffuse reasons relating to housing that are more important.

3.1.2 Intentions

Dommermuth & Klüsener (2017) state in their research piece that stated moving intentions are highly predictive of subsequent moving behaviour. They hereby also recognise that moving intentions are not equal to moving desires, an observation that has previously been put forward by Coulter et al. (2011) and De Groot et al. (2011). A moving desire differs from a moving intention due to a moving desire reflecting a wish to move house whereby an actual move may not be possible due to various constraints. A moving intention on the other hand, can predict moving behaviour more accurately because with an intention to move the element of desire is bypassed, as an intention implies that a person or household expects to make a move. Whether

12 the person or household actually wishes to do so is less relevant (Dommermuth & Klüsener, 2017).

De Groot et al. (2011) also make this distinction between moving intentions and moving desires. They state that an intention to move is an indicator of the willingness to change residence. Regarding positive attitudes towards moving, they further identify a difference between constrained and unconstrained attitudes, whereby unconstrained attitudes are wishes and desires, and constrained attitudes are expectations. With constrained attitudes, people take into account the constraining factors that may limit their desire to move. A desire to move may therefore not transform into an intention to move, if people that they cannot overcome the moving constraints that they have identified.

3.1.3 Areas of origin and destination

Lee (1966) provides a theoretical framework for the study of mobility decision-making processes. Factors deemed important to the decision are factors associated with the area of origin, factors associated with the area of destination, intervening obstacles, and personal factors. Factors associated with the areas of origin and of destination can affect most people in the same way, such as a good climate, or can be diverse in their effects on people, such as good schools, which are only relevant for households with children.

Lee (1966) notes important differences between the factors associated with the areas of origin and of destination. People who live in an area have a direct and often long-standing relationship with the area and are therefore able to make carefully considered judgements regarding these areas. Contrastingly, the knowledge one has of the area of destination is often significantly less precise and therefore decision-making can be hampered. A second important difference between the two types of factors is related to stages of the life cycle. There are many stages and events in the life cycle that can affect the mobility decision-making process, such as relationship formation and dissolution, job attainment or loss, childbirth or adult children moving out.

3.2 LIFE COURSE

As stated by Rossi (1955) the decision-making process is instigated by a trigger that sets off a positive attitude towards moving. A positive attitude can be triggered by a variety of stimuli. For example, dissatisfaction with one’s residential location or home and the desire to change this situation.

The study by Rossi (1955) was the first piece of academic research that differed from the general ideas that social scientists held at the time regarding household mobility decision making processes. Rossi (1955) demonstrated that household mobility decisions were underpinned by changes in the household composition, most often due to life-cycle changes and life-course events taking place within households. Furthermore, he stated that renters had a significantly higher inclination to move than homeowners. Renters were more likely to move than homeowners because their moving costs were significantly lower than those of homeowners. This implied that neighbourhoods of homeowners were more stable (people moved less) than

13 neighbourhoods of renters, because homeowners were more able to control their residential environment and it was more costly for them to move. In short, according to Rossi (1955) area mobility depended on the household composition characteristics and the housing characteristics of an area.

The premise of the life-cycle was also a central element in the categorization of families by Lansing & Morgan (1955). Their theory was based on the premise that households pass through a number of ‘’stages’’ in the life-cycle. They defined a series of categories whereby a linear progression through the stages in the life-cycle is assumed. Each stage is characterized by a certain size and composition of the family. The move from one stage of the life-cycle to the next provides the impetus for mobility and housing change. In this model, a change in the life-cycle is associated with a change in housing needs. The life-cycle approach is based on a traditional notion of household composition and structures, and the stages are organized to adhere to this traditional view. The life-cycle approach is therefore fairly inflexible.

3.2.2 Life course paradigm

A similar, but more flexible approach is that of the life course paradigm (Clark & Dieleman, 1996). The concept of life course incorporates parts of the life-cycle analysis but is more flexible and responsive to change in family and household structures. The life course approach emphasizes that an individual’s process through life can be seen as a sequence of events. These events include a wide range of possibilities that range from occurrences such as family formation, education and career changes, and housing decisions. The aim of life course analysis is to view individual life events and life patterns in the context of the social processes that shape, and are shaped by, these events. Individuals’ housing preferences and mobility intentions are an example of social processes that are strongly influenced by life course events.

Pinkster et al. (2014) state in their qualitative piece of research that a key common factor for households that determines their moving intentions is (a change in) the stage of the life course. When peoples’ work or family situation changes, their expectations and wishes regarding their neighbourhood also change. The characteristics of the neighbourhood are weighed differently according to the stage in the life that one has reached and is about to enter. A change in the life cycle therefore creates an imbalance in the housing situation, relating both to the housing itself and the surrounding neighbourhood. This imbalance subsequently creates an intention to move house.

3.2.3 Parallel careers

Mulder (1993) explains the life course paradigm by connecting it to the concept of parallel careers. As stated by Willekens (1987) migration is not and end in itself, but rather a means of attaining certain goals. Mulder (1993) applies this to the life course paradigm by stating that in an individual’s life course, the migration career and mobility pattern is of secondary importance to other, parallel careers. Migration, thus, is an adjustment mechanism for meeting needs arising from these parallel careers: the household and family, education, occupation and housing. Mulder (1993) defines the life course concept as the way in which an individual progresses through certain stages or statuses in various parallel careers in life, without the normative connotations associated with the life cycle concept. Mulder emphasizes that the

14 definition reflects a focus on individual life courses as opposed to household life courses. The reason is that households are volatile structures that can be formed, dissolved, split or combined. Household characteristics are viewed as influencing factors on individual migration behaviour. An example of a household characteristic which has a strong effect on an individual’s mobility intentions is the household composition. Mulder (1993) shows that households with children are less likely to develop mobility intentions, as parents are very reluctant towards having to move their children to a different school. Contrastingly, individuals in a single-person household are more likely to develop mobility intentions, as they do not have the constraints of partners or children. Furthermore, singles are more likely to move for reasons of relationship formation and co-habitation (Helderman et al., 2004).

When individuals form housing preferences, and the mobility intentions that result from these preferences, they are influenced by their personal values and expectations. This influence takes place based on the importance that individuals attach to certain goals and the expectancy they have of attaining these goals at their current location or an alternative location. De Jong and Fawcett (1981) state that an individual’s choices stem from the values that they hold regarding certain key themes. Among the seven types identified, a few are highly relevant in regard to residential preferences and mobility intentions. Example are wealth, status, comfort and affiliation. Wealth and status are important in their relation to housing preferences and consumption. Pursuing the goals of wealth and status may increase an individual’s intention to move. The same applies to comfort and affiliation. If an individual feels a strong affiliation to a neighbourhood, for example due to family ties or having a strong sense of community, the intention to move may decrease as a result.

Mulder (1993) states that the De Jong & Fawcett’s (1981) migration related goals can be linked to the parallel career theory mentioned previously. She provides the examples of the household and family career generating affiliation goals, and the educational and occupational careers being connected with wealth and status goals. However, the career with the clearest connection with the migration career is that of the housing career. Steps in an individual’s housing career are most often taken through migration. The parallel career which produces the goal that migration seeks to achieve is the triggering career. Willekens (1991) distinguishes two types of dependence in the intertwining of the migration and triggering careers: event dependence and state dependence. Event dependence is defined as the effect that the occurrence of an event in a parallel career has on the migration career. Examples of such occurrences are marriage, divorce, entering higher education, or changing jobs. State dependence is defined as the (long-term) effect of occupying a certain state in a career.

De Groot et al. (2011) echo the theories regarding triggering careers. They state that both anticipated and unanticipated life events play a key role in people’s development of moving intentions. This is because life events can trigger moving intentions by creating the need for independent housing or changing current housing needs. De Groot et al. (2011) mention a number of key life events that play a role in the formation of intentions to move. The formation and dissolution of co-habiting relationships is a clear trigger affecting moving intentions. It also illustrates the difference between anticipated and unanticipated moves. Co-habiting after relationship formation is an anticipated event, in contrast to divorce or break-up, which in many

15 cases is an unforeseen event. Another element of co-habitation that may positively influence residential mobility is the fact that couples can pool their financial resources and therefore financial constraints become less problematic.

Another event mentioned by De Groot et al. (2011) is the occurrence of childbirth. In most cases couples will have an idea if they desire to have children in the future. When childbirth occurs, be it the first child or more, the residential preferences of the household are subject to change. Childbirth can create the need for a larger dwelling or the desire to live in a child-friendly area. In case of the current housing situation not matching the household’s changed preferences, childbirth can form the trigger for moving intentions. De Groot et al. (2011) emphasize the simultaneous consideration of childbirth in combination with the suitability of the current dwelling. Their analysis proves that childbirth in a crowded home results in moving significantly more often than childbirth in a spacious home.

Mulder (1993) also states that the triggers for moving indicate the goals that people try to achieve with the change of residence; changing residence is never a goal in itself. Parallel careers do not only provide the triggers for moving, they are also key determinants of resources and constraints. An example is the course of an individual’s occupational career, which determines an individual’s financial resources. An individual’s household career is a key factor in the creation of coupling constraints , which restrict an individual’s freedom in terms of personal migration. Another possible constraining factor that individuals can face is investment in one’s housing career. The parallel careers which provide resources and constraint are named conditioning careers (Mulder, 1993).

3.2.4 Age

Willekens (1987) notes that changes in parallel careers are not evenly distributed over the life course and therefore neither is migration. In the life course of an individual, transitional periods can be observed which host many decisions and in which many changes occur. These transitional periods typically take place around adolescence and young adulthood. The young adult years are period in which individuals shape their occupational, household and housing careers and are likely to mover several times before settling down. It can therefore be said that migration behaviour is highly age-specific.

Bernard et al. (2014) state that migration is an age selective process, and that moving intentions and behaviour are strongly linked to age and the corresponding stage in the life course that age represents. A key peak for moving intentions and migration takes place in the young adult years, which then declines as age increases. It rises again in the mid-adult years, due to the occurrence of childbirth and family formation among this age group. Bernard et al. (2014) further state that the propensity to migrate may also occur around retirement age.

The key assertion that is put forward by Bernard et al. (2014) is that the relationship between age and moving intentions is closely linked to life course events. This is illustrated by their premise that migration ages differ between countries, as a result of variations in the structure of the life course among different countries and cultures. They provide the example of marriage and relationship formation, whereby in certain countries it is deemed normal for individuals to marry at an early age, but others whereby this is not the case. The age profile of migration is

16 therefore dependent on the prevalent norms with regards to life course events (Bernard et al., 2014).

Kley & Mulder (2010) state that moving intentions and migration mainly take place in early adulthood. This is due to the issue of life course events being framed by institutions, such as education, and therefore being closely connected to certain phases in the life course and specific ages. In their research piece, Kley and Mulder (2010), show that migration decision making is mainly driven by life-course events and perceived opportunities in various life domains. Key periods for migration decision-making are the transition from school to higher education and from higher education to the labour market.

In a longitudinal analysis by Coulter et al (2016) it becomes apparent that moving desires and expectations decline as people grow older. The authors demonstrate that a connection exists between moving desire and expectation, and actual mobility. The rate of actual mobility in young people is significantly higher than the stated desire and expectation. The rates of actual mobility and stated desire and expectation move closer together as age increases, whereby the rates convene around from the age of 40 onwards. This is an interesting process, which can be interpreted as life course events related to age having a stronger effect on actual mobility behaviour than on moving intentions. This ties into the theory put forward by Kley and Mulder (2010), which states that life course events are largely framed by institutions and therefore closely connected to certain stages in the life course.

Clark (2013) continues with researching the role of age in relation to residential mobility processes, and presents further findings concerning the relationship between age and life course events. He states that age serves a proxy for a variety of life course events, as age captures many of these events due to the occurrence of these events at relatively early points in the life cycle. Clark (2013) therefore states that it is not age per se, that creates moving intentions, but rather the events that occur within the aging process.

Stockdale & Catney (2014) also contribute to the literature regarding age and the life course, more specifically of the relationship that they hold with urban - rural migration. Besides the likelihood of moving being linked to varying ages or life course stages, they also state that the motivation for moving and subsequent direction of the intended move are also linked to age and the stage in the life course. With regards to direction of the intended move, they distinguish a difference in urban to rural moves and vice versa. Reasons put forward for these differences are the declining importance of the employment considerations with age, and the increasing importance of housing characteristics, amenities and comfort with age. It becomes apparent from their analysis that young adults are likely to move to urban settings, and the older population are more likely to move to rural settings.

3.2.5 Ethnicity

Another factor that is considered to be important with regards to mobility intentions is ethnicity. As ethnicity is a key determinant of segregation, it is highly relevant factor in the development of mobility intentions. Ethnic groups tend to form residential clusters as a result of two key processes. Firstly, as a result of the emergence of ethnic networks within neighbourhoods these neighbourhoods become attractive for individuals of the same ethnicity to live. The supply of

17 ethnic goods and services, and the sense of community within these neighbourhoods contribute to this process. Secondly, these neighbourhoods become increasingly unattractive to individuals from the dominant group, who avoid these neighbourhoods and therefore further the process of segregation (Zorlu, 2009). Ethnicity and segregation can therefore be viewed as important factors with regards to mobility intentions.

3.3 CONTEXTUAL FACTORS

Tenure can also be viewed as a key factor in the determination of residential mobility. The main reasons for tenure being an important factor are the long-term commitments, both financial and non-financial, that are necessary to own a home. To become a homeowner it is necessary for most people to enter into a long-term financial commitment in the form of a mortgage. This commitment is instrumental in tying people down and decreasing their intention to move (Helderman et al., 2004). Furthermore, a key deterrent for moving between owner occupied homes are the financial costs involved, i.e. transfer tax and brokerage costs. The cost of moving is significantly lower for renters. In addition to these financial reasons, Helderman et al. (2004) provide non-financial reasons as well. Homeowners are more likely to customize their home to their own personal preferences, and form emotional attachments to their home. Contrastingly, renters rarely make significant personal changes to their home and therefore are less likely to become emotionally attached to the dwelling.

Speare et al. (1975) state that homeowners are less likely to develop mobility intentions than renters due to owner occupied dwellings generally being of superior quality to rental dwellings. The superior quality mentioned by Speare et al. (1975) not only refers to the physical characteristics of the dwellings, but also regarding the quality of the neighbourhood. As owner occupied homes are often single-family dwellings as opposed to apartments, they are generally larger and offer more space for family expansion. The higher quality of both the dwelling and the neighbourhood lowers the likelihood of residential dissatisfaction, and therefore decreases the development of mobility intentions.

3.3.2 Urbanization

Bootsma (1998) states that the degree of urbanization plays a role in the development of mobility intentions. Due to the nature of urban living, which is largely in apartments and dwellings smaller than those in suburban and rural contexts, individuals are likely to develop mobility intentions in order to upgrade their residential situation. This is related to the individual’s household characteristics. For example, when anticipating family life and the arrival of children, individuals may intend to move away from an urban environment to a single-family dwelling more suitable for children. Generally, it is easier for people to find a dwelling to suit their taste in suburban areas than in urban areas.

3.3.3 Satisfaction and attachment

A key issue in relation to moving intentions is the neighbourhood satisfaction and neighbourhood attachment that people hold in relation to their place of residence. Various

18 theories provide insight into the processes that influence neighbourhood attachment and its relationship with the decision making process regarding one’s residential situation. Grinstein-Weiss et al. (2011) state that theoretical explanations of neighbourhood satisfaction consist of the main assertion that it is determined by a comparison between one’s actual situation and one’s preferred situation. In general terms, there are three influential factors that have an effect on this comparison. Individual sociodemographic attributes, subjective evaluations of the neighbourhood and objective neighbourhood conditions.

Morris and Winter (1975) proposed a theory based on the role of normative influences, whereby housing conditions are assessed by households and compared against relevant family and cultural norms. Having weighed these factors, households determine a ‘household deficit’ that influences neighbourhood satisfaction and the subsequent decision to move. Kahana et al. (2003) applied a person-environment theory whereby their research into satisfaction and well-being in older adults showed that important predictors of satisfaction were neighbourhood aesthetics, amenities, accessibility and social characteristics of the community. Sociodemographic characteristics are also relevant in predicting satisfaction. Spain (1988) shows that unmarried households are more likely to hold lower levels of neighbourhood satisfaction. In contrast, older residents and those who have a strong local network are more likely to show a greater satisfaction and attachment to their neighbourhood.

At the neighbourhood level, physical factors as well as social features of the community are primary factors in the determination of residential satisfaction (Spain, 1988) Features that positively affect neighbourhood satisfaction include the proximity of friends and relatives, satisfaction with racial homogeneity and feelings of safety. Features impacting satisfaction negatively are among others crime rates, poverty levels and perceived problems in the neighbourhood. Lower-income households are more at risk for neighbourhood satisfaction due to their housing options being limited to neighbourhoods with higher levels of noise, pollution, crime and poverty.

Residential satisfaction and place attachment are also expected to be strongly related to tenure, as mentioned previously. According to Rohe and Stewart (1996), homeownership is expected to result in positive outcomes regarding place attachment and residential satisfaction because owner-occupiers are economically motivated to protect their property value by being good neighbours and investing in social capital. Furthermore, owner-occupiers are more likely to actively participate in the community, build local relationships and live in their homes longer than renters. These are factors which positively impact residential satisfaction and attachment. Nowok et al. (2016) examine the relationship of life (dis)satisfaction with residential mobility and moving intentions, whereby the difference between movers and stayers is looked at. A number of key findings come forward from their research. Firstly, movers are significantly less satisfied with their lives in general than stayers. Further, movers’ social life and local ties are less important towards their overall life satisfaction than it is for stayers. Having completed a move, it becomes apparent that life satisfaction improves significantly.

The formation of local ties and creation of social capital are also mentioned as important factors with regards to place attachment (Clark et al., 2017). The authors state that place attachment is

19 a result of the length of residence of households in a certain area, and that place attachment plays a determining role in households’ decision to remain living in a certain area. They conclude that the attachment to life space and the activity sphere of households play a central role in the decision of households to stay locally.

The theories discussed above form the foundation for the formulation of expectations and the execution of statistical analyses to test these expectations. The following sections will introduce the data, methods and approach of this piece of research, after which the analyses will be conducted.

20

4 METHODOLOGY

The following section will describe the research methodology and data that have been applied for this thesis. Firstly, the research question will be elaborated upon. Subsequently, the various elements of the research design will be discussed.

4.1 RESEARCH QUESTION

The research question for this piece of research will be the following:

- To what extent are the moving intentions of the Dutch population influenced by the life course, and how do these differ for the young population?

In order to be able to answer this question, a series of analyses will be carried out regarding various characteristics relating to moving intentions. Following the insights gained from the theoretical overview, a key distinction that will be made regarding the determinants of moving intentions, is between individual and dwelling characteristics, household characteristics, dwelling & neighbourhood characteristics, and municipal characteristics.

Personal characteristics such as ethnicity, educational attainment and age are expected to have a significant effect on the moving intentions of an individual. A key part of the analysis will consist of determining to what extent personal characteristics, combined with household characteristics such as income and household composition, influence Dutch individuals’ moving intentions and how they are related to age and the life course.

Besides individual and household characteristics, there are certain other factors expected to have a significant effect on the moving intentions of individuals in the Netherlands. Contextual factors such as dwelling, neighbourhood and municipal characteristics are likely to have an effect on mobility intentions as well as personal and household characteristics. The analysis will focus on the extent to which the different variables have an effect on the moving intentions of the Dutch population.

After the statistical analyses have taken place, the results will be discussed in relation to the theoretical framework and conclusions will be formed regarding the outcome. Finally, the limitations of this piece of research will be discussed, and recommendations will be made for further avenues of research.

21

4.2 RESEARCH DESIGN

4.2.1 Research approach

This piece of research will be carried out by utilising a strictly quantitative approach. The key reason for choosing a quantitative method is that a large-scale statistical analysis can be considered to be an effective way of approaching the research question, because the scope of the research comprises the entire population of the Netherlands. With a large-scale statistical analysis it becomes possible to analyse the population at a high level of scale, as opposed to a qualitative analysis, which would have a very specific focus and be less generalizable to a larger population.

Due to the nature of the research question and the theme of this piece of research, a quantitative analysis is expected to provide more generalizable results than a focused qualitative approach. As the mobility intentions of the Dutch population are being analysed, a dataset of over 60,000 respondents can be expected to provide statistically relevant results, upon which general conclusions can be formed regarding the population of the Netherlands.

A qualitative analysis regarding mobility intentions could potentially be interesting, however it would require units of analysis on a significantly smaller scale, for example regarding a specific neighbourhood or demographic group. Furthermore, it would most likely not be possible to generalise the results to a greater scale, such as the entire Dutch population.

The intention of this piece of research is to provide insight regarding moving intentions for policy makers and business leaders, therefore a large-scale quantitative approach has been deemed to be a more relevant approach for this instance.

4.2.2 Data collection

The statistical analysis at the heart of this piece of research will be carried out based on the Netherlands’ Housing Survey (WoON 2015) data. The purpose of the WoON 2015 survey is gathering information regarding the housing situation of the Dutch population and their living requirements and needs. Among many others, the survey covers household composition, the dwelling and living environment, housing costs and personal characteristics. The WoON 2015 data has been connected to a Statistics Netherlands (CBS) database containing municipal characteristics.

The WoON 2015 data collection process was coordinated and conducted by Statistics Netherlands (CBS) as part of a large scale fieldwork project. The survey methods that were applied comprised of telephone, face to face, and online surveys. The dataset contains a total of 62,668 respondents.

Following the data collection phase, the data was collated and processed by ABF Research. The work carried out by ABF Research comprised of preparing a suitable statistical database, ready to analyse and stripped of unnecessary information. The data processing consisted of the following stages: coding of open questions, carrying out consistency checks, imputation of non-response, determination of inferred variables, determination of additional living expenses, weighting, plausibility checks and non-response analysis.

22 As the parties involved in the data collection and processing are professional parties with numerous quality management protocols in place, it is assumed for this analysis that the subject data is free of significant errors and inconsistencies.

4.2.3 Units of analysis

The units of analysis for this piece of research will be formed by the Dutch population over 18 years of age and living in a private household, i.e. not in a prison, hospital or any such institution. The units of observation in this case are the household units in the aforementioned population. In order to form conclusions regarding different personal and household characteristics, each individual will be coded with a certain code regarding characteristics such as level of education, income, or stage in the life course.

As the dataset has been produced anonymously by Statistics Netherlands (CBS), ethical constraints regarding anonymization and privacy are less applicable in the preparation of this piece of research, as the processes of anonymization have already taken place.

When classifying respondents into categories such as level of income or educational attainment, problems could arise in relation to bias of the researcher regarding these themes. To combat this, the classifications that are used will be based on the existing classifications within the WoON 2015 dataset. If the WoON 2015 dataset does not provide a classification, a statistics-based classification will be applied by the researcher and explained accordingly.

In the following section, the classification of each variable used in the analysis is introduced. Furthermore, a frequency table of each variable is included.

23 The following section will provide an overview of all the variables that have been utilised in this piece of research. Each variable is introduced, whereby the categorisation and frequency are both addressed. The dependent variable, moving intention, is a nominal variable which has been categorised into two categories.

4.3 DEPENDENT VARIABLE

The dependent variable in this piece of research is the intention to move house. This is defined as the respondents’ intention to move to different dwelling within the next two years. The relevant term is “moving intentions”. The WoON 2015 survey includes a specific question regarding this issue. The question whether respondents intend to move within the next two years will be categorized as follows:

Moving intention Answers Observations (62,668) % of total

Yes Yes, I have already an

alternative dwelling Yes, definitely

I would like to, but cannot find anything

Possibly yes, maybe

24,055 38.38%

No Definitely not 38,613 61.62%

Table 4.1: Classification and frequency of Moving intention.

The dependent variable will be named “Moving intention”. It is the key variable in this piece of research, and will form the foundation of the statistical analyses. Regarding the answer “Possibly yes, maybe”, a decision was made to include this in the categorisation of “Yes” due to these respondents holding an inclination to move, and the fact that the answer was given in very few cases and thus does not warrant its own category.

A respondent’s moving intention is used in this piece of research as an indicator of whether the respondent household will move house in the next two years. As various authors have stated (Dommermuth & Klüsener, 2017; Coulter et al., 2011; De Groot et al., 2011) moving intentions are strong indicators of actual moving behaviour, and therefore it is deemed possible to apply the stated moving intentions as an indicator of the expectation of actual moving behaviour in the near future.

24

4.4 INDEPENDENT VARIABLES

The independent variables in this piece of research will be briefly outlined below. The independent variables will be divided into personal and household characteristics, dwelling & neighbourhood characteristics and municipal characteristics. These categories have been chosen because they offer a clear framework in which to place the variables. All the variables that are described below were chosen after reviewing the theoretical framework, and are aimed at providing results that reflect the various theories.

4.4.1 Personal and household characteristics

The following variables will be included in the statistical analysis, to provide insight in to the extent to which individual and household characteristics have an effect on moving intentions. Each table provides an overview of the categorisations that have been applied in case of any variable transformations, as well as the (relative) frequency of the variables.

4.4.1.1 Educational attainment

This variable will be recoded into three categories according to the highest level of education that the respondent has completed. The categorisation that has been applied is based on general categorisations used by CBS with regards to educational attainment.

Educational attainment Answers Observations (62,668) % of total

Low Primary school, Lower secondary

school, basic vocational training

19,633 31.33%

Medium Upper secondary school, skilled vocational training

22,711 36.24%

High University education 19,234 30.69%

Other Other 1,090 1.74%

Table 4.2: Classification and frequency of Educational attainment.

4.4.1.2 Age

The variable age is split into the following four categories. These categories have been chosen due to the stages that they represent in the life course, such as starting and completing higher education, family formation and retirement.

Age Observations (62,668) % of total

18-25 9,079 14.49%

26-40 12,969 20.69%

41-65 26,568 42.39%

65+ 14,052 22.42%

25

4.4.1.3 Ethnicity

The categorisation of the variable ethnicity is taken directly from the dataset, as more in-depth information relating to the ethnicity of the respondent is not available. The standard CBS categorisation applies, whereby a difference is made among immigrants between western and non-western immigrants.

Ethnicity Observations (62,668) % of total

Dutch 51,997 82.97%

Non-western 5,021 8.01%

Western 5,650 9.02%

Table 4.4: Classification and frequency of Ethnicity.

4.4.1.4 Household income

The independent variable income will be coded in to three categories. The income that is measured is the disposable household income. For single-person households, the variable is already aggregated to the level of household in the dataset. This remediates the issue of imbalance between single and dual income households. This variable has been categorised by taking half a standard deviation from the mean. This approach has been chosen, due to the representative categorisation that it creates.

Income Categorisation Observations % of total

Low < € 25,016 20,161 32.17%

Medium € 25,017 – € 52,018 29,316 46.78%

High > € 52.019 13,191 21.05%

Table 4.5: Classification and frequency of Disposable income.

4.4.1.5 Household composition

The variable household composition is coded into the following three categories:

Household composition Observations (62,668) % of total

Single-person household 18,047 28.8%

Multi-person household with children 15,675 25.01%

Multi-person household, no children 28,946 46.19%

Table 4.6: Classification and frequency of Household composition.

It must be noted that in selecting a household composition variable, a deliberate choice was made for a variable with the three-way categorisation above. The dataset contains several variables regarding household composition with differing categorisations, however after having reviewed these alternatives it became apparent that the various categorisations did not add any relevance or additional information to the analysis.

26 4.4.2 Dwelling & neighbourhood characteristics

The following variables are related to the characteristics of the respondents’ current dwelling and neighbourhood, as well as the respondents’ perception of the neighbourhood.

4.4.2.1 Tenure

This variable concerns the tenure situation of the respondents dwelling. The variable regarding residential ownership situation has been applied from the WoON 2015 dataset. It shows the tenure situation and provides a categorisation of both social and privately rented dwellings.

Tenure Observations (62,668) % of total

Social rental 15,338 24,48%

Private rental 5,618 8.96%

Owner-occupied 34,022 61.61%

Unknown 7,690 12.27%

Table 4.7: Classification and frequency of Tenure.

4.4.2.2 Type of dwelling

This variable specifies the type of dwelling the respondent currently resides in. The typology single-family dwelling means a house, as opposed to an apartment which is denoted as a multi-family dwelling in the dataset. A difference is applied to detached houses and “other” houses, such as semi-detached and terraced houses.

Type of dwelling Observations (55,225) % of total

Detached house 7,887 14.28%

Other house 30,258 54.79%

Apartment 16,166 29.27%

Non-residential dwelling 914 1.66%

Table 4.8: Classification and frequency of Type of dwelling.

4.4.2.3 Crowding

Following the type of dwelling, as mentioned above, a second variable has been constructed to make it possible to measure the suitability of the type of dwelling in relation to the inhabitants. This variable was calculated by dividing the number of rooms by the number of persons in the household. This aims to show the suitability of the current dwelling in terms of the stage of the life course that a respondent is currently in. The categorisation that was applied is shown in the table below. The categorisation is done as De Groot et al. (2011).

Crowding No. of rooms per person Observations (55,095) % of total

Crowded <1 6,383 11.59%

Neutral 1-2 25,250 45.83%

Spacious 2> 23,462 42.58%

27

4.4.2.4 Attachment to current dwelling

This variable measures the extent to which a respondent feels an emotional attachment to his or her current dwelling. The statement put to the respondents was “I feel attached to my home”. The answer categories “no attachment” and “absolutely no attachment” have been combined to form a single category.

Dwelling attachment Answer Observations (55,095) % of total

Very Strong Very strongly attached 17,663 32.06%

Attached Strongly attached 26,277 47.69%

Absent Not attached

Absolutely not attached

11,155 20.25%

Table 4.10: Classification and frequency of Attachment to dwelling.

4.4.2.5 Sense of community

The question relating to this variable is posed as a statement to the respondent, to which they must indicate whether they agree or disagree: “My neighbourhood is gezellig, a place where people help each other and do things together”. This has been translated as the extent to which there is a sense of community in the neighbourhood, and recoded into three categories. Neighbourhood sense of

community

Answer Observations (62,668) % of total

Strong Strongly agree

Agree

29,779 47.52%

Neutral Neither agree nor

disagree

19,207 30.65%

Absent Disagree

Strongly disagree

13,682 21.83%

28 4.4.3 Municipal characteristics

4.4.3.1 Degree of urbanization

This variable measures the degree of urbanization of the respondents’ place of residence. As Bootsma (1998) states, the degree of urbanization can be an important factor in determining moving intentions. The degree of urbanisation is measured by CBS as the address density per square kilometre. In ascending order, the address density per square km that determines the degree urbanization is fewer than 500, 500 – 1000, 1000 – 1500, 1500 – 2500 and more than 2500. The variable has been categorized into three categories, shown in the table below. Degree of urbanization Answer Observations (62,668) % of total

Rural Not urbanized

Moderately urbanized

17,720 28.28%

Medium Reasonably urbanized

Highly urbanized

31,235 49.84%

High Very highly urbanized 13,713 21.88%

Table 4.12: Classification and frequency of Degree of urbanization.

4.4.3.2 Grey pressure of municipality

The grey pressure of an area is an indicator of the ratio of elderly population compared to population aged between 20 and 65 years old. The grey pressure taken into account in this analysis is taken from the CBS municipal statistics database. Therefore, the value is applicable to the municipality that the respondent lives in. This variable has been categorised by taking half a standard deviation from the mean. This approach has been chosen, due to the representative categorisation that it creates.

Grey Pressure Value Observations (62,668) % of total

Low < 26.5% 18,369 29.31%

Medium 26.6% - 34.2% 25,255 40.30%

High > 34.3% 19,045 30.39%

29

4.5 HYPOTHESES

The statistical analysis that will be performed in the following part of the thesis will take place based on a number of expectations that have been formed as a result of the literature review. Instead of forming hard hypotheses, a more narrative approach has been chosen. The expectations as based on the literature are formulated below. One could interpret these as ‘soft’ hypotheses.

The statistical analysis will be carried out in order to gain insight into whether the existing theories regarding moving intentions can provide confirmation of the existing literature, or alternative explanations of the factors that influence people’s intention to move.

Having reviewed the literature, certain expectations have been formed regarding the determinants of the moving intentions of the Dutch population. Characteristics relating to one’s educational attainment are expected to have an effect on the moving intentions. According to the literature reviewed in the previous section of this thesis, the following has become apparent and will be analysed in relation to the statistical analysis.

People who are highly educated are more likely to be intent upon moving house than those who have a low educational attainment (Coulter & Van Ham, 2013). The difference between highly educated people and those who are not highly educated, is that highly educated people are more likely to move in relation to their employment situation. Highly skilled professional jobs are often spatially concentrated in certain locations within a country, which means that highly educated people are more likely to need to move in relation to their job. An example of this type of spatial concentration within the Netherlands is the Zuidas in Amsterdam, which is the prime location in the Netherlands for professional service sector jobs. For people with a low level of education, i.e. suitable for blue collar and working class jobs, it is often not necessary to move house for employment opportunities as these types of non-skilled jobs are generally not clustered in specific locations, but available across the whole of the country. Coulter & Van Ham (2013) also state that young, highly educated people are more likely to be geographically mobile than older people due to their strong career focus and ambition. Furthermore, Coulter et al. (2011) state that higher-educated people often have a higher spatial awareness than whose with low level of education, which further enhances the development of moving intentions.

A characteristic that is also of significance regarding moving intentions, and is closely related to one’s educational attainment, is the income level (Helderman et al., 2004). There are various reasons for the level of income to be of importance in relation to moving intentions. It can be expected that those with a high income are more mobile in terms of their residential location, due to the fact they have less constraints regarding moving opportunities, as they have a large budget for housing and are more able to account for moving costs. Furthermore, those with a high income are more able to pursue goals of wealth and status compared to those with a low income.

Age is expected to be an important determining factor for moving intentions, as it can offer an indication of the stage in the life course that a person is currently in. People in the age category 41-65 are expected to be less intent on moving house than people in the category 18-25 and

26-30 40. People in the age category 18-25 are expected to be in the stage of the life course that involves leaving their parents’ houses and living independently. Furthermore, this age category includes people in the final stages of their education, which in many cases involves moving house. 26-40 year olds are likely to be in the process of forming relationships and families, as opposed to the older categories, who are likely to already have settled in.

Bernard et al. (2014) state that migration is an age selective process, and that moving intentions and behaviour are strongly linked to age and the stage in the life course that age represents. A key peak for moving intentions and migration takes place in the young adult years, which then decline as age increases. Kley & Mulder (2010) state that moving intentions and migration mainly take place in early adulthood. This is due to the issue of life course events being framed by institutions, such as education, and therefore being closely connected to certain phases in the life course and specific ages. It can therefore be expected that young, highly educated people have a higher probability of positive moving intentions than those with a low level of education or the older population.

Furthermore, Clark (2013) states that age serves as a proxy for life course events, and therefore captures many of these events early in the life course. Stockdale & Catney (2014) describe the relationship between urbanisation, age and life course events. It is expected that youth in rural areas are likely to intend to move, and older people are likely to intend to move if they reside in urban areas, in order to find more suitable housing befitting the stage in their life course (Bootsma, 1998) With regards to age, it is likely that young adults from 18-25 and 26-40 are significantly more likely to intend to move than the older population. The regression analysis will provide deeper understanding of the differences among young people controlling for other influences, such as the level of education or household composition.

Following the theories regarding household composition (Rossi, 1955; Lansing & Morgan, 1955; Helderman, 2004; Bernard et al., 2014; De Groot et al., 2011), one might expect household composition to be an important determinant of moving intentions. Family formation and having children are often key reasons for people to move house. However, the household composition and the suitability of the household composition are closely related with regards to age as well (Bernard et al., 2014). Therefore, the household composition is expected to be a strong determinant of moving intentions, differing per age category.

With respect to the relationship between ethnicity and moving intentions, it is expected that processes of spatial segregation and ethnic clustering (Zorlu, 2009) contribute towards ethnic minorities holding positive mobility intentions less often than the dominant ethnic group, in this case the Dutch population. Ethnic segregation takes place as a result of the emergence of ethnic networks within neighbourhoods. This can have the effect on ethnic minorities of wishing to remain part of these clusters, and therefore a lack of an intention to move.

The expectations mentioned above are relevant with regards to personal characteristics and the life course events theories described in the theoretical framework. Besides personal characteristics and life course events, various other characteristics are expected to be of importance regarding moving intentions.