2016
The ‘skewed’ tenants of
Amsterdam’s social housing
On which factors influence moving intentions of
scheefwoners to private housing
Raymond Frederiks
Studentnumber: 10280758
Raymond.frederiks@hotmail.com
Master Urban & Regional planning
Supervisor: Dhr. Dr. Christian Lennartz
Second Reader: -
2
Abstract
Seldom has Amsterdam seen such a demand for housing. The popularity of the city as a place to work, live and recreate draws households from all kinds of nature. The tight housing market of Amsterdam has led to the need for housing to be distributed more effectively. Cheap ‘Scheefwonen’ is one of the aspects which are said to hinder the functionality of the housing market. The decreasing amount of social housing could be used more effectively, if scheefwoners would move to private sector housing. Measures like income restrictions and the income dependent rent increase have already attempted to decrease the amount of scheefwoners. These measures however, pass the fact that residential mobility of scheefwoners to the private sector in tight housing markets is lower. This thesis has investigated the conditions under which the scheefwoners would be likely to move. According to scheefwoners themselves, offering proper alternatives would create conditions for them to move, so profiling the socio-demographic conditions and wishes of the scheefwoner would contribute to creating. The scheefwoner is more often a household with two breadwinners compared to the regular tenants in social housing. 1/3rd doesn’t earn more than 39.000 euro’s a year and 2/3rd doesn’t earn more than 44.000 euro’s a year. They are more often higher educated and inner-city stadsdelen have larger relative amounts of scheefwoners compared to regular tenants. Most of the scheefwoners still prefer their next dwelling to be a social rented dwelling. Only 20% consider moving out of Amsterdam, while stadsdelen West, Zuid and Oost are popular destinations within Amsterdam. Zuidoost, Noord and Nieuw-west are considerably less popular destinations. In terms of dwelling size, scheefwoners barely prefer dwellings with the size of 50 square meters. In terms of what certain households are willing to pay, the amount of scheefwoners who consider paying more than the limit between regulated housing and private housing limits itself to households who already pay a rent near that limit, and with incomes at least exceeding 44.000. The scheefwoners with relatively low rents but with higher incomes however, very often consider rents above that limit. Private housing in Amsterdam will need to account for these preferences and characteristics to create the alternative for scheefwoners. The scheefwoners who are dissatisfied with their current conditions in social housing, still perceive themselves to be well off and stated to have postponed or altered a move. Private renting is considered too expensive, which could explain the low amount of stated preferences for private renting. On top of that, households who are confronted with uncertainties like varying incomes, health conditions or relationship issues decreases their intention to move. In these cases, they fear that they would be unable to return to social housing because of the long waiting lists. Generally negative attitudes towards moving to private housing are seen among scheefwoners who argue that not scheefwonen but housing associations selling their stock are the reason for housing shortages in social housing. Those who already consider themselves paying reasonable rents, often doubt the affordability of social housing itself. According to scheefwoners in the focus groups, the income dependent rent increase doesn’t affect an intention to move. Ultimately, offering the alternatives which satisfy the housing needs of the scheefwoner would create the conditions for scheefwoners to move ‘naturally’. Under current market circumstances, scheefwoners will be unlikely to move, due to the majority of scheefwoners even with an intention to move still preferring social rented housing. If not, scheefwoners will likely have to alter preferences like moving out of Amsterdam or paying more rent than initially stated.
3 Structure Abstract ... 2 Preface ... 5 Introduction ... 6 2: Literature review ... 9 2.1: What is scheefwonen? ... 9
Policy change regarding ‘scheefwonen’... 9
2.2: Moving intentions ... 12
2.2.1: Theory of planned behavior ... 12
2.2.2: Life course theory ... 13
2.2.3: Predictors of moving intentions ... 13
2.2.4: Preferences ... 15
2.2.5: Tenure choice ... 16
3: Methodology ... 17
3.1: Conceptual model ... 17
3.2: Mixed method approach ... 20
3.3: The Stadgenoot objective ... 21
3.4: Limitations and justifications of the study ... 23
4: The changing structure of the Amsterdam housing market ... 23
5: Who, where, what: the cheap scheefwoner of Amsterdam ... 27
5.1: Who is the scheefwoner? ... 27
Income ... 28
Household & Age ... 29
Location ... 31
Education ... 32
Logistic regression of scheefwonen ... 32
5.2: Preferences ... 34
Tenure ... 34
Location ... 37
Affordability ... 38
Dwelling size ... 39
5.3: Scheefwoners with an intention to move ... 40
6: Why us? Perceptions and attitudes of scheefwoners ... 43
Policy change and residential dissatisfaction ... 44
Where and how? ... 44
The importance of context: Perceived control of behavior ... 45
4
The income dependent rent increase ... 48
7: Conclusion: The scheefwoner of Amsterdam ... 51
Reflection ... 53
Literature ... 54
5
Preface
It seems just a little while ago that I started writing this thesis about the scheefwoner in Amsterdam. Initially, the first feed of this thesis was built around middle income housing in Amsterdam, and very much oriented on the housing market. Housing association Stadgenoot wondered if offering middle income housing would be legitimized. Under current circumstances is the Amsterdam housing market, offering middle income to housing to scheefwoners would be most likely for housing associations to offer those dwellings to.
But scheefwonen lived more than I expected. In-depth research to this group caused me to rethink the very problem of scheewonen, especially in the context of Amsterdam. On different occasions, I heard stories about scheefwoners profiting from cheap housing with very high wages, and even I had that prejudice. This is the first time doing social research has caused me to not only preform a research objective, but even rethink the problem of a social phenomenon itself, namely scheefwonen. This has literally meant my coming of age as a researcher.
The ultimate goal of this research is to offer useful insight for housing associations, private landlords and policymakers into the group of scheefwoners. For the functionality of the housing market of Amsterdam, knowing this group could very well contribute to better conditions for this group. I hope that the results of this study reach them, and that this study could even contribute in policymaking regarding scheefwonen in Amsterdam.
Finally, I would very much like to thank Stadgenoot, De Alliantie and Rochdale, for providing me with a unique and large dataset. At Stadgenoot, especially my supervisors Willy de Looper and Perry Hoetjes. Their assistance helped me especially in practical matters, and allowed me to learn more about social housing and scheefwonen. I would like to thank Jeroen van der Veer from the AFWC as well, for allowing me to present my research proposal to other housing associations as well. And concluding, I would like to thank my thesis supervisor from the University of Amsterdam, Christian Lennartz, for his pleasant guidance throughout the process.
6
Introduction
Certainly, housing in Amsterdam is one of the hottest topics currently embedded in the media in and around Amsterdam. The attractiveness of the city as a place to live, work and recreate is widely known and increases the demand for housing in the city. Gentrification, the process of revitalization of ‘former’ poor neighborhoods, is seen widespread throughout Amsterdam following the increased demand among richer and younger residents. While Amsterdam as a whole has thus become a mix of increasingly different incomes, opinions on how the structure of the Amsterdam housing market should be, diverge. The ambition of Amsterdam as the economic engine of the Netherlands and one of the five urban
powerhouses in Europe (Gemeente Amsterdam, 2016) seem to contrast the ideal of the ‘undivided city’, a city with a mix of all kinds of incomes and ethnicities. Newspaper ‘Het Parool’ (Damen, T. 2016) wrote that Amsterdam is even facing conditions similar to London. The quest for affordable housing in
Amsterdam is presumably becoming more difficult with each passing day, given that for example, housing prices in Amsterdam are rising while the amount of transactions is decreasing (Wanders, J. 2016).
Figure 1: “How do I find a dwelling in Amsterdam?”
The housing stock is changing as well. The recent publication ‘Living in Amsterdam 2015’ of the Municipality of Amsterdam (2016) shows these transitions. In 2011, 29.4% of the housing stock were owner-occupied dwellings. The housing associations had 48.1% of the total stock, and the remaining 22.5% were private rented dwellings. In 2015, the assets of the housing associations have decreased to 45.5%, while the amount of privately owned1 housing has grown (Gemeente Amsterdam, 2015). Scientific research at the University of Amsterdam in the field of Urban Georgaphy has already shown the
implications of this development towards an increasingly deregulated housing market. Amsterdam has seen a decrease in the amount of low- to middle income and unemployed households, while the surrounding region has seen an increase in these households over the period of 2004 until 2013 (Hochstenbach & Musterd, 2016).
Lower middle incomes are seemingly negatively affected by these developments. The recent publication ‘Wonen in Amsterdam’ (Gemeente Amsterdam, 2016) has shown a decrease in the amount of households with incomes ranging from 35.000 to 45.000 a year. So why is the amount of low- to middle income households decreasing, in comparison to the rise of the amount of the rich and the ‘working poor’? According to Hochstenbach, problems arise when these households have an urgency to move (Remie, 2016). Households who currently rent a social-rented dwelling who are inclined to move usually have a few options. When households within social housing have a lower income than 35.000 euro’s a year, it is likely that they will move to another social rented dwelling. In addition, housing associations are allowed to offer 10% of the social housing stock to households with incomes within the range of 35.000 – 39.000. Households with incomes above roughly 39.000 will have to move to private sector housing2, because income restrictions in social housing prevents them from moving to another social rented dwelling. But what happens when the difference in rents and housing prices in the private sector differ too much from
1 Owned by private landbords excluding housing associations or owner occupied.
7 social housing? Sven Heinen, chairman of the Union of housing brokers in Amsterdam, sees a disturbingly small amount of ‘affordable’ private rented housing. The amount of dwellings with rents between 700 and 1000 euro’s a month seems to be lacking (Duin, 2016).
Parallel to this development is the amount and usage of social housing. Long waiting lists for social rented housing, ranging from 13 years in Amsterdam Zuidoost to 17 years in Amsterdam Centrum in 2016 for example (RTL Nieuws, 2016), fuel the discussion about what households should be in social housing. The new housing law (Rijksoverheid, 2015) inclines that housing associations should almost only provide for households with the lowest of incomes, and thus providing extra pressure on liberalization on the housing market of Amsterdam. Schilder & Conijn (2015) argued that shares of the current social housing stock should be liberalized to able to meet the needs of middle incomes, which applies to middle income households who live in but are no longer eligible for social housing as well. But housing associations providing only for the lowest of incomes results in two key developments. The first one is that housing associations will to a lesser extent provide housing with rents between 710 and 1000 euro’s a month. The second one is that the current stock would have to fit the targeted population more efficiently, meaning that even middle incomes who earn above the 35.000 no longer belong to the target group of social housing, and implying that these households earn too much. These households are called ‘cheap scheefwoners’, and will further on in this research continuously be referred to as scheefwoners. The households with incomes beneath that limit in a social rented dwelling will be referred to as regular tenants.
These scheefwoners, usually middle incomes with incomes ranging between 35.000 and 45.000, seem to be confronted with a problem once they decide to move, as Hochstenbach (2016) argued. According to juridical terms, they earn too much for social housing, but in case of moving, they might find themselves trapped due to the high rents and housing prices in the private sector within certain parts of Amsterdam. The lack of private rented housing with rents ranging from 710-1000 euro’s a month seems to limit the accessibility of housing for middle incomes, and limits the amount of households moving from social- towards private rented housing. The position of scheefwoners on the Amsterdam housing market thus seems to be uncertain. On the one hand, social housing is not suited fort scheefwoners, while some scheefwoners in turn find private sector housing to be unsuitable (Stols, 2015).
The income dependent rent increase, introduced by national government of the Netherlands, is one of the instruments to encourage scheefwoners to move towards private sector housing. Pressure on social housing in general makes scheefwonen an unwanted situation (Poulus & Blije, 2015), and policies like the income dependent rent increase support that image. While the income dependent rent increase might increase the attractiveness of moving to the private sector, high rents and housing prices decrease that attractiveness. These interventions aim to force a flow of households from social to private housing, yet the effects of this measure remains to be questioned due to the gap between social and private sector housing. In addition, financial motivations only play a certain part in a decision to move. While the effectiveness of the income dependent rent increase on moving intentions of scheefwoners remains unclear, offering alternatives to scheefwoners would seem a measure which could clearly affect moving behavior from social to private housing.
The presumed natural flow due to ‘natural’ causes to move, like household expansion or rises in income, are thus be influenced by the structure of the market. The current structure of the Amsterdam housing market limits moving intentions of the cheap scheefwoners. The current supply of housing in Amsterdam will need to adjust to meet their needs, so this thesis should be seen in light of policy change to counter scheefwonen. In order to understand what conditions would be suitable, a clear explanation of
scheefwoners and their housing needs is necessary. This thesis aims to provide a better understanding of who the scheefwoners is, what their wishes are and what kind of households within that population are
8 more inclined to move in the context of Amsterdam. The scheefwoners will be compared to regular tenants, in order to provide insight in what characteristics and preferences are unique for scheefwoners. In addition, being labeled as a scheefwoner, and receiving income dependent rent increases, could affect moving intentions as well. By answering these questions, this thesis hopes to provide some useful insight for stakeholders who are interested in providing the wanted alternatives of scheefwoners, and provide insight in how recent measures to counter scheefwonen are received. This is done in order to explore to what extent these measures contribute to conditions for scheefwoners to move to private sector housing. This has led to the following main research question:
“Under what conditions would the scheefwoner most likely have an intention to move to private sector housing?”
The main research question will be answered after the secondary research questions. These are:
Who is the scheefwoner in the city of Amsterdam in socio-demographic terms compared to regular tenants?
What are the stated preferences of the scheefwoners compared to regular tenants’?
What scheefwoners are more likely to have an intention to move to a private rental dwelling?
Of the scheefwoners, what socio-demographic characteristics and preferences influence having an intention to move?
How does the perceived behavioral control of scheefwoners within the context of Amsterdam affect their moving intentions?
How do legislative changes, being the income dependent rent increase and the DAEB / niet-DAEB policy affect their moving intentions?
The theoretical framework will elaborate on the most important concepts of the questions above. The first element is the concept of scheefwonen3. The second of these elements is the concept of moving intentions, and the variables predicting them. The concept moving intentions will be explained, and what kind of predictors influence that concept. The predictors which influence moving intentions include socio-demographic characteristics of scheefwoners, their behavrioal control, and attitude towards moving to private housing. The last concept which influences the conditions of scheefwoners for having moving intentions is that of the structure of the Amsterdam housing market, and will be elaborated on separate from the theory. That chapter will provide an overview of the current structure of the Amsterdam housing market and developments in recent years.. The aim is here to provide an overview of recent changes in the structure of the housing market of Amsterdam, and to understand to what extent these developments could influence moving intentions of the scheefwoner.
Further on, the research methods that are used will be explained, and why quantitative as well as qualitative data are used. The main part of this thesis contains of analysis following a quantitative survey which was set out among all tenants of housing associations Stadgenoot, De Alliantie and Rochdale, including regular tenants. This survey is used to discover general patterns in socio-demographic
characteristics and preferences of scheefwoners, and how these characteristics and preferences influence their moving intentions. Next to the survey results, two focus groups were hosted in addition to the survey. This method was mainly used to gather in-depth and context related information of individual households on why they would or would not move to private sector housing, as well as finding out what they think about the concept of scheefwonen itself. The conclusion will summarize and connect these
3 Scheefwonen literally means paying rents which either fall short or exceed the amount that tenants are
able to pay, but legally households are considered scheefwoners when they earn above a set limit of 35000 euro’s gross a year and rent a social rented dwelling.
9 results and provide an answer to the research question. This thesis will conclude with a reflection based on the results of the survey and focus groups on the one hand, and the structure of the Amsterdam housing market and [policy changes regarding scheefwonen on the other hand.
2: Literature review
2.1: What is scheefwonen?Scheefwonen is a phenomenon which, at least in literature unfortunately, has not been directly translated to English yet. The definition of scheefwonen is however, that the traits of a dwelling don’t match the traits of an household. RIGO (2010) has made a distinction between scheefwonen in terms of traits of the dwelling, and in terms of financial capability. A family who lives in an apartment which can be considered way to big, is scheefwonen in terms of dwelling size. Scheefwonen in financial terms can be distinguished into two categories, called ‘cheap scheefwonen’ and ‘expensive scheefwonen’. The latter can be described as households with rent subsidies, who rent dwellings with a rent above the capping limit of 628,76 euro’s a month (Poulus & Blije, 2015). This research will not focus on this group, but on the ‘cheap
scheefwoners’. These are households with a moderate or high income who rent a social rental dwelling with a basic rent usually beneath the ‘liberalization limit’, which entails the maximum rent for a social rented dwelling (Philipsen & Blije, 2013), currently 711 euro’s a month. In this research, the households with a social rented dwelling with rents above 711 euro’s due to the income dependent rent increase, are still included in the population of ‘scheefwoners’.
Scheefwonen had no direct consequences before 2011 for the scheefwoners themselves, although the phenomenon has been deemed as an unwanted situation in 1989 by Enneüs Heerma, state secretary of public housing, because of the size of state subsidiaries being spent on public housing. In 2009, The European Commission proposed the ‘DAEB- vs niet-DAEB’ policy, which determined that housing associations should only offer housing to households with incomes which do not exceed the indexed 33.000 euro’s a year (European Commission, 2009). 90 % of the total stock of housing associations would then be allocated to these households, with 10% remaining to be distributed among households with ‘other elements of social prioritization’, with incomes above that threshold but still in need of assistance, such as big families. Dwellings which are allocated for this target group fall under the name of DAEB, (Dienst algemeen economisch belang), meaning these dwellings can be financed with state subsidiaries. Housing associations are still allowed to preform commercial activities, but only when commercial parties are not interested.
Policy change regarding ‘scheefwonen’
Households with incomes currently exceeding around 35.000 euro’s a year (2016 index) are not eligible for social housing, meaning these households earn too much for social housing. Until 2020 however, housing associations are still allowed to offer 10% of all housing allocations to households with incomes between 35.000 and 39.000 euro’s a year (Woonbond, 2015). With the income dependent rent increase which was introduced in 1 july 2013, households who earn above the limit set by the EU commission and rent a social rented dwelling, received rent increases based on their income. Households with incomes above 44.000 receive an higher rent increase. For the scheefwoner, this means that they will most likely be unable to move to another social rented dwelling, and receive rent increases as well until a certain limit is
achieved.
The effects of these policies to prevent state subsidies for housing not ending up at middle incomes and thus ‘scheefwoners’ under the current term, have been simulated by PBL (Eskinasi et al, 2012). These estimated effects are significant for this research because they provide an explanation for postponing or
10 not being able to realize moving intentions. The PBL calculated the likely added search time comparing income categories to one another. The PBL came to a certain set of conclusions on the effect of this measure.
Table 2.1: Waiting time in years, before and after income restrictions for social housing Search time
in years
Difference Search time in years
Difference Income (x 1000) With or without
income limits National Amsterdam Low (<33) With 2.2 -19% 5.3 -14% Without 2.7 6.2 Low to middle (33 – 38) With 3.2 52% 6.8 27% Without 2.1 5.3 Middle – middle (38 - 43) With 2.7 34% 5.9 22% Without 2.0 4.8 High middle (43 – 62) With 2.1 18% 4.4 17% Without 1.8 3.8 Highest (>62) With 1.4 5% 2.2 5% Without 1.4 2.1 Source: PBL (2012)
While the search time for households with a low income decreases, the search time for all households with higher incomes increase, especially among low to middle incomes. dwelling , comparing different income categories and different contexts, namely national average vs. the average of Amsterdam. In Amsterdam, all income categories show an higher waiting time in years compared to the national average. Although the waiting time decreases significantly for low incomes, the middle incomes all face an increase in waiting time. As a direct consequence of income restrictions for social housing, the likelihood of realizing a moving intention within a certain time frame for middle incomes will likely decrease.The latest report of ‘Living in Amsterdam 2015’ shows that the relative amount of dwellings of housing associations in Amsterdam as part of the total amount is decreasing (Gemeente Amsterdam, 2015). This means that fewer social rented dwellings are becoming available, and scheefwonen becomes an increasingly unwanted situation for the functionality of the housing market. Fewer available social housing means that fewer dwellings would be available to a certain target group, or the target group would have to be reduced. Scheefwonen is thus considered a problem when households who could afford private sector dwellings occupy social housing, because they no longer belong to the target group. In addition to that, the amount of these cheap scheefwoners seems to be decreasing, as seen in the table below.
Table 2.2: Changes in absolute amount of scheefwoners in the Netherlands Amount of expensive scheefwoners Amount of cheap scheefwoners 2011 230.000 782.000 2013 367.000 685.000 2015 528.000 518.000
Source: Ministerie van BZK (2016)
According to the report Wonen in Beweging (2016) by the Dutch ministry of internal affairs, the decline in cheap scheefwoners and the rise of expensive scheefwoners was mainly seen at the housing
11 increase. In addition, EU regulation prevents the influx of higher incomes in the social rented sector, thus restricting a large part of inflow towards the social rented sector of the housing market. The new housing law largely facilitates this development, by providing the legal framework for EU law and the income dependent rent increase. The graph below shows that the amount of residents moving from a regulated dwelling to a non-regulated dwelling has risen in all income categories, especial those with higher incomes. Kromhout et al. (2016) concluded that the income restrictions are the primary cause of this development, while the effect of the income dependent rent increase remains benign.
Graph 2.1: Relative amount of residents moving from regulated towards non-regulated housing
Source: Kromhout et al. (2016)
Since the EU legislation and the income dependent rent increase was institutionalized, different professionals within the housing sector in the Netherlands have criticized these legislative changes and predicted certain consequences of this change in policy for middle incomes in particular. Pararius (2010) argued before the implementation of the DAEB/Niet-DAEB legislation, that especially in larger cities with tight housing markets, middle incomes would mostly be unable to move due to high rents in the private sector. Being confronted with income dependent rent increases as well decreases their position as well. Legislation implies that they should not be in social housing, ‘scheefwoners’ have mostly lived there over longer periods of time. In 2012, Peter van den Heuvel, director of housing association Woonlinie which operates in the towns of Waldrichem and Zaltbommel, argued that this legislative changes would have negative consequences for the position of scheefwoners (Brabands Dagblad, 2012). He doesn’t answer the question how to solve the problem of scheefwonen, but asks if scheefwonen is even a problem itself. He argues that lower incomes receive lower rents and rent allowances, while home buyers receive deductions on the interest paid over their mortgages. Households with incomes between 33.000 and 60.000 euro’s a year and unable to buy a house, are instead confronted with rent increases.
Financial economic journalist Peter Hendriks (Hendriks, 2016) compares housing and the concept of scheefwonen in the Netherlands with other countries. In the United Kingdom, social housing is provided by law for the lowest incomes to the intermediate middle class, meaning those with jobs like a teacher or a police officer. In Austria, the social rented dwellings are allocated according to income criteria, but also household size. Larger households can rent a social dwelling even while earning as much as 80.000 euro’s a year. In France, the income limits are similar. He argues that scheefwonen is a concept which was
16% 29% 32% 46% 51% 60% 66% 76% 87% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2009 2012 2015 Low (<33.614) Middle (33614 - 43000) High (>43000)
12 created in the Netherlands itself. However, England and France have an entirely different structure of the housing market, with most of the social housing stock concentrated in highly urbanized areas in France and a very limited stock of social housing in England (Whitehead & Scanlon, 2007), meaning that the amount of middle incomes living in social housing is significantly lower than in the Netherlands.
2.2: Moving intentions
2.2.1: Theory of planned behavior
The theory of planned behavior, as proposed by Ipek Ajzen (1991), is a groundwork theory which is used as a base for predicting social behavior in many fields of research. As seen in the model below, certain background factors determine attitudes, norms and behavioral control towards an intention, and
eventually, behavior. Translating this to moving intentions and the research question of this thesis, moving intentions are presumably influenced by attitudes, norms and behavioral control. The intention here are plans to move from a social rented dwelling to a dwelling in the private sector. The attitude entails the favorable or unfavorable evaluation of the behavior in question. Some scheefwoners might for example appraise a move to the private sector because it could be the natural thing to do. The perceived norms entail the degree to which a certain group an individual household belongs to approve or don’t approve a certain behavior. Scheefwoners could for example, have certain norms regarding moving to the private sector. The third factor is the perceived control on behavior, which entails the perceived difficulty of performing the behavior. For example, given a certain income or household, what are their perceived odds of having an intention to move?
Figure 2.1: Theory of planned behavior as proposed by Ajzen
Source: Ajzen, 1991
Ajzen (1991) states that as a general rule, the more favorable the attitude and subjective norm with respect to behavior, and the greater the perceived behavior control, the stronger should be an individual’s
intention to perform the behavior under consideration. Translating this to moving as a scheefwoner, a move seems increasingly likely once first of all, a norm arises among groups, even among scheefwoners themselves, to indeed move to the private sector. Furthermore, when there is a positive attitude towards this move and the desired move is within perceived possibilities, an actual plan to move arises. Once an household acts on an intention, it could be that the actual control may be somewhat different from the perceived behavioral control. The actual control, as seen in the later stage in the model, intervenes between intentions and behavior. A possible move may backfire, once an household isn’t actually able to realize that move due to factors like a tight housing market, or the inability to find anything. The
13 difference between perceived and actual control is that perceptions may be inaccurate. For example, an household may deem themselves able to move, but the structure of the housing market may pose additional and unexpected restrictions. This feedback-loop is described in the conceptual model above. These restrictions influence the perceived behavioral control themselves. There are however, numerous reasons why an household could have an intention, not just because the attitudes, norms and perceptions condition an intention.
2.2.2: Life course theory
The theory of planned behavior is thus most suitable in explaining if a certain behavior takes place, in the context of pre-behavioral considerations, like attitudes, norms and perceptions. It does not however, fully explain the reasons why and under what conditions households move in the first place. Coulter & Scott (2015) argue that most of the theory about residential mobility and why a move takes place has been written from a life-course perspective. Why people move is the key question which was answered through this perspective, when it comes to researching and explaining residential mobility. In 1955, Rossi
introduced the ‘life-cycle’ theory in his book ‘Why families move’. The life-cycle theory assumes that the choice for housing and thus moving behaviour is largely determined by changes in household formation. Changes in the household fuels the desire to move to a more fitting dwelling for the newly formed household. Wolpert (1965) and Speare (1974) built on this theory by introducing the concept of dissatisfaction and place utility (Wolpert, 1966). With these concepts, these authors argue that an household is inclined to move when significantly dissatisfied with their current living situation. This fuels the formation of certain preferences, and shapes the image of a preferred destination to move to. In this sense, moving functions as an opportunity for households to improve their housing satisfaction. The model below is a simplified overview of the arguments proposed by Wolpert & Speare.
Figure 2.2: Lineair overview of the process preceding moving behavior
Source: Wolpert (1965) & Speare (1974) (Own modification)
According to Coulter & Scott (2015), separately researching specific characteristics or preferences which results in moving has one limitation, namely that not all reasons to move have an equal effect. By uncovering the variations of intentions among different life courses, residential moves can be better explained, instead of using separate revealed preferences as main factors explaining moving behavior.
2.2.3: Predictors of moving intentions
Several authors have stressed the importance of the relationship between current and desired housing consumption. Then a disequilibrium exists, the desired dwelling could be consumed to satisfy housing needs. This is a very basic representation of the process of residential mobility, because changes in household don’t necessarily lead to dissatisfaction, and moving behavior depends on other predictors as well instead of just household changes. Such predictors are for example, the macro-economic situation on the housing market, meaning the available supply of housing. These considerations all precede the
formation of an intention to move. Individual predictors
Macro- and microlevel constraints and restrictions influence the expectancy to move. Microlevel predictors are predictors on household level. Those are factors like income, household type, age,
education level, labor market status and ethnicity. Changes in income for example, may fuel the desire to
household
Change in
Dissatisfaction
with current
dwelling
Formation of
alternate
preferences
Moving to
other dwelling
14 move because rises in income open up new housing possibilities or a fall in income could precede a move from private to social housing. The income itself determines the amount of dwellings that are within reach, so a higher income is likely to facilitate a move (Priemus, 1984; Mulder & Hooimeijer, 1999). Ageing may fuel the desire to move, but limits the amount of dwellings that are within reach. Upstairs apartments may not be a valid option for the elderly with health conditions. These limitations make that only a certain amount of dwellings is accessible. As was argued by Coulter et all. (2011), household ties and commitments can limit the expectancy to move. Having to move the kids to other schools can be a constraint for an household to acting on the desire to move. Young single households for example, have less ties to take into account, making the realization of a moving intention easier (Mulder & Hooimeijer, 1999). Also other factors, like human capital can extend the amount of possibilities one has, because the amount of connections and knowledge one has can influence the ability to realize a moving intention (Coulter et all. 2011).
According to Lu (1998), tenure, income and age have significant effects on actual behavior, but are also channeled through their intentions. Fang (2006) argues that limited economic ability and a lack of choice in the housing market constrain actual moving behavior, meaning that individual characteristics and housing market opportunities interact. This could interplay influences the formation of intentions as well, because when households expect hindrances because of a lack of opportunities and resources, a favorable attitude towards moving might result in letting go of an intention to move. (De Groot, 2011). De Groot also found that households with higher incomes and a strong desire to move have a higher probability to move in comparison to other households. A study by Kan (1999) found that decreases or increases in family size, job change, changes in income, retiring or having more rooms than needed al have positive effects on the mobility expectation These factors influence the possibility that a household will actually move, even though the preferences might still be the same
Macro-economic barriers
When it comes to research on a macro-level, the context of the local housing market and the general economic situation are presumed to influence the housing choice as well. According to Strassman (1991), mobility models often treat macro-level variables in housing as an exogenous factor in residential choice, purportedly because complex government interventions in land use and in finance, construction and pricing of housing constrain the supply of (new) housing. On a microlevel, factors such as age and income can influence housing choice, and are integrated in research as starting variables, since age and income shape certain preferences (Coolen et al, 2002). The amount of resources one has can thus create possibilities but also pose restrictions. A discrepancy arises when stated preferences, and thus certain values, can’t be realised due to the structure of local housing markets. Previous studies by Murie (1976) and Jones (1979) have shown that the allocation of dwellings can and should also be understood in terms of preferences and in market constraints. Kendig (1984) argued that the market structure impacts the propensity to move as well. A lower propensity to move is seen among owner-occupancy than rental dwellings, because renting is considered more flexible. Kendig concludes his paper with the following statement: ‘Market conditions constrain the choices of migrants, forced movers, and households being formed or dissolved ; they accelerate or postpone the advancement of households in the early stages of a housing career’.
Several examples around the world show that market conditions shape expectations and even influence preferences. In Hong Kong, severe increases in demand in the centre has led to larger area’s within the city to become unaffordable for the general public (The Proto City, 2015). This connects with findings that moving is considered more difficult in the current marco-economic context of housing busts and higher unemployment rates (Coulter et all., 2011). In 2008, the Dutch Planbureau voor de Leefomgeving (PBL) has done a nationwide research which investigated the stated preferences and residential choice
15 behaviour in the Netherlands. It found that in the urban conglomeration of the Randstad, households have less possibilities to move due to the high demand and low supply on the housing market compared to other regions. This finding is complementary to the research of Kearns & Parkes (2003), which argued that housing shortages create significant perceived barriers.
Intentions and actual behavior
De Groot et all. (2011) argue that the relationship between intentions and actual moving behavior has not been sufficiently explored. The study of actual behavior gives more information about the interplay between constraints, triggers, resources and opportunities. Coulter et al (2011) therefore argue that assessing moving expectations are more likely to predict actual moving behaviour. Also, differences in the intention and the expectation to move may arise from macro- or microlevel constraints which influences the expectations. Another limitation of this model put forward by Coulter et al. (2011) is that in this case, residential mobility seems a rather linear process. However, sudden lifetime events like the dissolution of partnerships, or the death of one’s partner, may trigger a sudden and undesired propensity to move. In this sense, another category is added, namely those who don’t intend to move but expect a move anyhow. Figure 2.3: Intentions vs expectations
Source: Coulter et all, 2011 (own modification)
The study of Coulter et al. (2011) compared intentions and expectations with variables such as age, income and education level. An intention to move is considered much more a pre-move thought which is related to dwelling or neighbourhood dissatisfaction. The expectancy to realize this intention includes an individual’s analysis of housing market constraints. In this study, the expectation to move is much higher among highly educated and spatially flexible individuals. The study also revealed that older individuals, lower incomes and social renters express desires to move but the expectations are generally low. Coulter et all. (2011) also argue that households with greater levels of household ties and commitments have a lower chance of actually acting upon their intention and an higher risk of abandoning their intention. Coulter et all. (2011) argue that several concepts could be used to measure the relationship between intention and behavrior. An intention could be a wish, a desire, or an expectation to move.
2.2.4: Preferences
In other housing literature, the concept of preference is put forward, mostly as a tool to investigate why certain residential choices are made and to be able to predict mobility towards housing with a certain set of characteristics. Mapping the stated preferences can be a useful tool to develop an understanding of
Population
Intends to
move
Doesn't
expect to
move
Expects to
move
Doesn’t
intend to
move
Expects to
move
Doesn't
expext to
move
16 what kind of dwellings would suffice for households who intend to move. To measure preferences, a distinction can be made between algebraic models and non-algebraic models. Zinas and Jusan (2012) built the following conceptual model of housing preference research, based on the theories of Timmermans et al. (1994); Coolen & Hoekstra (2001).
Figure 2.4: Housing preference research in practice
Source: Zinas & Jusan, 2012 (own modification)
The algebraic approach contains quantitative methods to investigate preferences, while the non-algebraïc approach contains qualitative methods. Typical methods within the application of the algebraic approach are surveys where respondents are asked to rate the preferred attributes of a dwelling. The most
prominent example of qualitative methods to present data is the use of decision nets. This was first introcued by Witlox(1995) an alternative to quantitative strategies which aim to map intentions and choice behaviour among an aggregated population. This method, decision tables, maps the residential choice of individual households in a matrix. Timmermans (1994) argues that some approaches don’t take context into account, and focusses too much on the functionality of outcome variables. Compositional methods explore housing preferences by recording separately and explicitly how people evaluate housing attributes, while decompositional methods evaluate dwelling profiles as a whole, and statistical methods are applied to estimate the contributions of attributes and attribute levels (Middelkoop & Bouwmeester, 2014). Compositional models are still widely used however, because the outcome variable aims to discover the importance of predictor variables.
2.2.5: Tenure choice
As argued in the theory above, tenure choice is influenced by household characteristics, the structure of the housing market and government policy to encourage or discourage a certain tenure choice. Tenure choice can be seen in terms of renting versus owner-occupancy, but distinctions between tenure can be made within the segment of renting as well. In renting, social and private renting can be distinguished. This study will focus among others, on the tenure choice of households in social housing, comparing the tenure preference of scheefwoners against regular tenants. This because the propensity to move of the scheefwoner in social housing towards the private sector renting is investigated. The key question is what kind of predictors influence tenure choice. A previous study by Coolen et al. (2002) found the variables age, current tenure, income and household composition account for 91% of the variance of tenure choice. For scheefwoners however, no distinction can be made in current tenure
Regarding owner-occupancy, accessibility is largely determined by mortgage interest rates, the types of mortgages, the loan to income ratio and tax regulations, in addition to income (Van Middelkoop & Boumeester, 2014). These restrictions make owner-occupancy more inaccessible in comparison to the rental market and will likely pose significant restrictions for households with lower incomes. As a
consequence, lower incomes tend to live in or move to social rented dwellings, while higher incomes tend to own a dwelling. A study by Ioannides & Kan (1996) found that the residential mobility among
Housing
preference
Revealed
Actual choices
Stated
Intended housing
choices
Non-algebraic
approach
Algebraic
approach
17 homeowners is lower than renters, and that household characteristics play an important part in predicting tenure choice.. Coulson & Fischer (2001) argued that even among renters, residential mobility differs. Households with lower incomes and longer spells of unemployment are less mobile than other renters. On an individual level, different factors determine the tenure choice of households. Drew (2014) argued that economic determinants play a dominant role in tenure choice theory, by summarizing that “In the case of housing tenure decisions, this theory suggests that decisions about owning versus renting housing are determined by the combination of individual demands for attributes associated with different kinds of tenure, and constraints on an individual’s ability to access the desired kind of tenure”. Also, the life-course theory which was explained before plays an important role in tenure choice literature. Demographic factors such as household expansion plays an important role in tenure choice (Drew, 2014). The term ‘housing-ladder’ is frequently put forward in literature, which implies that households continuously tend to move ‘up’ towards better housing. When it comes to tenure choice in general, albeit owning vs renting or social- vs private renting, Drew (2014) recognized three dimensions to examine tenure choice general:
1. Economic determinants: rationally choosing the most advantageous dwelling 2. Socio-demographic: characteristics; housing careers, moving up the ladder 3. Behavioral factors: desired dwellings by individual households
18
3: Methodology
3.1: Conceptual model
The quantitative part of this thesis focusses firstly on the socio-demographic characteristics of scheefwoners. The compositional model is the most suitable here, because not the dwelling itself but individual attributes are investigated. Characteristics form the predictors in order to investigate what the most prominent traits of scheefwoners are. Providing an image of the scheefwoners in terms of socio-demographic traits is furthermore useful to explain the preferences of scheefwoners. Differences between scheefwoners and regular tenants are investigated among others, to discover to what extent their
preferences differ. Socio-demographic charateristics is the central concepts investigated as predictor in this thesis. The concept of scheefwonen is used as outcome variable in the first subquestion, while (not) intending to move is used as an outcome variable in the latter. The preferences a household has are determined by household changes. Scheefwonen is, strictly speaking, a predictor value based on income. In this thesis, the preferences of the research population are continuously explained by comparison in the descriptives, to give a clear image of the preferences of the scheefwoner.
Like Fang (2006) argued, a specific context affects the possibilities for individual households to act upon their intentions. In Amsterdam, it is assumed that the scheefwoner has limited possibilities due to higher prices in the private sector. The possibilities to act upon intentions are thus limited. This creates a problem in itself, as was suggested in the introduction. In a perfect market situation, the possibility to act upon an intention to move is the highest. The functionality of the Amsterdam housing market can improve, if the housing stock matches the intentions of the scheefwoner. The regulatory changes, namely the income dependent rent increase and the Daeb/ Niet-daeb policy scheefwoners not being able to move to other social rented dwellings, means that the supply of private sector housing will need to adjust to the preferences of scheefwoners.
There are two ways in which could be predicted whether private sector housing in Amsterdam is able to meet the needs of the scheefwoner. Studying actual moving behavior as an outcome variable could provide an image of the scheefwoner who was able to realize a moving intention, and those who didn’t. this requires however, longitudinal data. The other way is to investigate pre-moving thoughts and preferences. Investigating the stated preferences of the scheefwoner is not a predictor of actual behavior, but could offer some very useful insight in future residential moving behavior. By comparing the current supply of housing with the socio-demographic characteristics and preferences of scheefwoners,
predictions can be made whether scheefwoners would be able to realize their intentions, given those current market circumstances. In this thesis, there was chosen for the second option for logistic reasons. Because this thesis was written with the help of several housing associations, it was easier to approach current scheefwoners then scheefwoners who have already moved.
This research restricts itself only to intentions, and not actual behavior. The intention, being the antecedent of actual moving behavior of the scheefwoner, is influenced by changes in household and residential satisfaction, with preferences as the outcome variable. Certain preferences, like the amount of rent an household is willing to pay, or a preference to buy or rent, influences In this thesis, it is assumed that preferences follow after an intention to move arises. Those with higher incomes might be more positive due to increased possibilities, while households with lower incomes could be less positive. ‘Scheefwoners’ might even estimate their odds lower, because they are not eligible for social housing, while the private sector is too expensive. Preferences result in a certain estimation, and when that estimation becomes negative around certain preferences, the lower the odds will be that the Amsterdam
19 housing market meets the demands of these households. Household characteristics are variables like income, age, location and household composition. These characteristics each influence preferences to a certain degree, while some preferences function largely independent. Those who prefer to live in the inner city might have the same income, age, and household characteristics as those who prefer to move outside of Amsterdam. Most of these variables regarding household characteristics resurface in research about moving intentions. De Groot et al (2011), Clark & Dielemans (1996) and Fang (2006) use similar characteristics and preferences when developing an understanding of what variables influence moving behavior of scheefwoners. An example is given in the table below.
Table 3.1: Overview of variables
Household Characteristics Preferences
Income Tenure
Source of income Preferred costs of dwelling
Age Location
Education Dwelling size
Location Gender
Household type
The causal relationship between all variables is seen in figure 3.1. Based on the basic theories of moving behavior (Rossi, 1955; Speare, 1965), a change in household could initiate residential dissatisfaction on the one hand, and a change in resources on the other hand. A rise in income for example, could initiate dissatisfaction due to the fact that with this higher income, a dwelling of higher quality could be achieved. At the same time, initial income restrictions to tread into a certain segment of the housing market could fade, increasing the perceived behavioral control within a certain context. When the attitude of moving to a private sector dwelling is also favorable, an intention to move will likely arise. After formulating an intention to move, which is here defined as a wish to move, preferences will be developed. In this thesis, preferences of households without an intention to move are researched as well, even though their preferences aren’t developed after an intention to move has been stated. In that sense, those preferences resemble their current housing needs and not the needs of the dwelling they want to move to. These preferences are however, always developed within a certain context. Macro-economic factors influence the housing possibilities one has. A family who lives in Amsterdam might for example, settle for less floor space compared to families in rural areas. The actual control a household has over moving behavior differs from perceived behavioral control, in the sense that actual control is the degree of control after one has attempted to perform that behavior (Ajzen, 1991). When preferences match an available dwelling and the actual control to perform moving behavior is seen, a move will likely take place.
However, a difference between perceived and actual control can result in households adjusting their preferences to a specific context, postpone or even cancel a move.. This could result in major frustrations in the housing market. The attitude itself, as seen in the conceptual model of Ajzen (1991), might even change. It are exactly these attitudes which are rich qualitative data as well. Attitudes towards the social rented – and the private sector result from the feedback cycle from actual to perceived behavioral control. In the tight housing market of Amsterdam, it would seem likely that more households have canceled or postponed a move in the latter stages, or did not form an intention at all because they would deem their resources inadequate for a move in the first stage. That is why focus groups were organized, to explore the role of these perceptions, and to create an image of attitudes towards moving to private sector housing.
20 Figure 3.1: Conceptual model of the process of moving
Source: Own modification
3.2: Mixed method approach
The goal of this thesis is to uncover which scheefwoners and under what circumstances a scheefwoner is likely to be having an intention to move to private sector housing. A mixed method approach is used to uncover the role of specific variables in the process of moving behavior. A survey will give a clear, quantitative image of who the scheefwoner is, and what their preferences are, both compared to regular tenants in social housing. This is done from the thought of providing ‘alternatives’ as a condition. This research is rather deductive, because it assumes that there are certain variables which influence moving intentions of the scheefwoner, and will give a clear image of the scheefwoner who wants to move to the private sector. The likelihood of a scheefwoner actually moving to the private sector could thereafter, but not in this thesis, be investigated by analyzing the current stock of housing in Amsterdam. Looking at figure 3.1, the survey investigates (changes in) household characteristics, intentions and preferences. The purpose of the qualitative part of this thesis is, referring to the theory of planned behavior, how perceived behavioral control varies and attitudes of scheefwoners influence moving intentions to private sector housing. Also, to find out to what extent the income dependent rent increase has already
conditioned individual households to have an intention to move to private housing. Being a scheefwoner could for example, form a psychological motivation to be wanting to move to the private sector. In some cases, it might soften an intention to move because they might deem themselves lucky or well off given the context of a tight housing market. In light of recent policy change regarding scheefwonen, the qualitative part of this thesis aims to uncover the role of these policies as well, and to what extent they influence attitudes towards moving intentions. Through qualitative methods, the role of these attitudes and perceived behavioral control on moving intentions can be better understood.
It can present a more dynamic image of the thought process of scheefwoners when it comes to moving, and how they have, for example, altered their preferences due to macroeconomic circumstances, while
21 quantitative methods show a static, momentary image of moving behavior. The qualitative part contains two focus groups, both hosted at the office of Stadgenoot, and qualitative data from the survey. One of the risks of such a focus group is that the moderator has less control over what is said compared to one-on one-one interviewing (Morgan, 1996). In this research, an open dialogue was chosen over one-one one-on one-one interviewing because it would be possible to see if there would be shared attitudes of these scheefwoners towards moving to a private sector dwelling, or if they could differ. Also, the limited timeframe of this research made numerous one- on one interviewing unlikely.
Table 3.2 Summary of research strategy used
Research method Data collection Theoretical approach Analysis
Mixed-method research Survey (Quantitative part) and 2 focus groups (Qualitative part)
Deductive SPSS: descriptive data analysis & logistic regression analysis. Narrative analysis An email with an invitation to fill in the survey was sent to all tenants in a social rented dwelling of the housing associations Stadgenoot, De Alliantie and Rochdale, even though the scheefwoners are the population which is primarily researched. One of the reasons was to avoid scheefwoners dropping out of the survey, because addressing just scheefwoners with a survey on moving to the private sector might anger the scheefwoners. In total, 27.534 tenants were approached, of which 3616 completely filled in the survey, with a response rate of 13%. Additional housing associations were asked to participate in this thesis, to acquire a more representative population of cheap scheefwoners renting from housing associations. Scheefwoners are compared to regular tenants with household incomes beneath 35.000 euro’s gross every year. This is done because this study does not compare cases over time, scale or area. Instead, it’s investigated how preferences and socio-demographic characteristics differ between these populations, given a specific geographical context (Amsterdam) and timeframe (2016). Housing
associations do have income data of their tenants, what makes it theoretically possible to send invitations to scheefwoners only, but that data has to be destroyed by housing associations once the income
dependent rent increase has been calculated. This resulted in a generally larger amount of respondents with an income below the EU limit, and not specifically scheefwoners. The population of scheefwoners which is primarily investigated is only 1/3rd of the entire population. The survey is added as appendix 1. The tenants who filled in the survey had the option to take part in the focus groups. Due to the
considerable sample size, diversity within the focus groups could be achieved. Respondents were invited to the session after controlling for income, age and household size. Open questions in the survey are used in addition to the findings from the focus groups. Attitudes and perceived behavioral control towards housing might have changed due to changes in the structure of the housing market and policy change. While legislative changes now imply that social housing is not meant for those with middle incomes like the scheefwoner, the attitudes among that group might differ. Initiatives to stop the selling off of housing stock of housing associations have followed dissatisfaction among tenants in social housing in Amsterdam (Damen, 2015). And exactly these sentiments within tenants of social housing might impact the attitudes towards moving to the private sector. Factors like the income dependent rent increase and housing associations selling of stock could in turn influence moving intentions themselves, even though these attitudes don’t influence preferences themselves.
3.3: The Stadgenoot objective
22 Source:
One of the main reasons of the subject chosen in the master thesis come from the interest of Amsterdam housing association Stadgenoot on this specific subject. Stadgenoot board president Marien de Langen argued in the newspaper ‘Financieel dagblad’ on 8th of January 2016, that housing associations could build private rental housing with rents between 700 and 1000 euro’s a month, because presumably real estate investors lack the financial incentives to build dwellings in that price range, especially larger dwellings. Stadgenoot is looking for a scientific motivation which could help them in their argument why they should build dwellings with rents between 700 and 1000 for middle incomes who are currently ‘trapped’ on the Amsterdam housing market. According to the report WIA 2015, the amount of rental dwellings with prices between 700 and 971 euro’s is 5.6% of the total housing stock (Gemeente Amsterdam, 2016). The question is to what extent that share suffices for the target group Stadgenoot has in mind.
The question Stadgenoot asked was, how to improve the legitimacy to build these dwellings. This is in light of the new housing law restricting private sector activities for housing associations, the DAEB/niet-DAEB policy (EU Commission, 2009). This thesis, which was written at Stadgenoot, will reflect later on if building these dwellings is legitimized for housing associations. With the new housing law, housing associations are only permitted to build these homes with prices between 700 and 1000 euro’s when ‘private developers are not interested’ (Rijksoverheid, Z.J.). The ‘markttoets’ was introduced as a tool in the new housing law of 2015, literally meaning market check, which is a check to determine whether there are private developers interested in the commercial activity a housing association would want to engage in. On the other hand, pressure on the social housing stock, and the legislative revisualization of the social housing stock increases the urge for households to move from social to private housing. For Stadgenoot, the legitimization might follow when their scheefwoners, being middle incomes, limitedly move to the private sector due to housing shortages in the private sector. The middle income families are the target group. Boterman (2012) supports this presumption of Stadgenoot by stating that the large social rented sector and the scarcity of larger dwellings makes it difficult for middle incomes to find suitable family housing. The markttoets could prove to be a useful tool for Stadgenoot if scheefwoners are not provided for by private landlords.
This interest has multiple effects on the way this thesis is written. First of all, Stadgenoot has interest in quantitative data to support their theory that options to move as a middle income household from social to private housing are limited. Second, the interest of Stadgenoot also opens up doors and possibilities. Because of writing at Stadgenoot and a former internship at the AFWC, the Amsterdam Federation of housing associations, the research proposal was introduced other housing associations in an early stage. The survey was put out under all tenants of Stadgenoot, and later on also among tenants of two other housing associations, De Alliantie and Rochdale, increasing the amount of respondents in this survey. For them, it is interesting to find out more about the preferences of the ‘scheefwoner’, and what kind of
23 dwellings might be needed to increase the amount of ‘scheefwoners’ leaving cheaper housing stock behind.
3.4: Limitations and justifications of the study
The three criteria for assessing the quality of social research are, as defined by Bryman (2008), reliability, replication and validity. Reliability can be achieved by increasing consistency in measurement and making sure the concepts are clearly defined, making the study repeatable. The quantitative part of this thesis clearly separates the concepts of socio-demographic characteristics, preferences and moving intentions. The replication in this thesis arises from building a survey which could also be used in a different time or different context. Additional help came from experts within Stadgenoot in order to construct a survey which was understandable and user-friendly for tenants in social housing. In building a reliable survey, thesis supervisor Dr. C. Lennartz provided a similar survey as an example which he had used when investigating tenure preference. The WIA 2015 survey is a survey which was held to investigate residential satisfaction and housing preference among inhabitants of Amsterdam, and was used as an example as well, to make sure that the survey met the practical standards which are necessary to conduct solid and reliable research. A lack of use of SPSS and advanced statistical analyses in the bachelor and master of Urban planning has however, increased the difficulty for the researcher to apply multiple logistic regression analyses.
One of the limitations of this studies replication is that it operates in one specific context. Comparisons between cheap scheefwoners and regular tenants would probably be answered differently in larger cities like Amsterdam then smaller towns elsewhere across the Netherlands. The influence of constraints posed by the market is therefore different in Amsterdam than other cities. However, in this thesis the choice was made not to compare the context of Amsterdam with others. The goal of this research is to give a reliable impression of the scheefwoners and their preferences within Amsterdam. The current shortage of housing on the Amsterdam housing market is known and comparing cases will not provide the results wanted, except confirm suspicions that scheefwoners in Amsterdam do indeed have different possibilities as well as perceptions, while even household characteristics and preferences will differ. This study operates within a specific context, and this survey could be held in other locations or times to discover the impact of context. As a suggestion for further research, suitable private sector housing is the end mean to accomplish moving behavior of scheefwoners, and it would be interesting to see in what way scheefwoners differ from other households who are currently designated to private sector housing. The last limitation of this thesis has met is that socio-demographic characteristics, preferences and intentions are analyzed, and not actual moving behavior. Preferences themselves are not a solid predictor of actual moving behavior, because it assumes that an household might move once they have an urgency to move and the structure of the preferred housing market could meet their needs. Investigating
preferences and the way they influence moving intentions limit the actual predictability of actual moving behavior. The results of this study should be interpreted as a study which offers insight in the direction in which these households wish to move, but not the actual odds to move. The actual odds to move could be investigated through researching realized moving intentions among ‘scheefwoners’, or comparing the results of this study with the structure of the housing market, to see to what extent that structure is able to facilitate moving intentions of ‘scheefwoners’.
