Moving to and from enclaves. Housing choices of ethnic minorities in Denmark

Transcription

1 Moving to and from enclaves. Housing choices of ethnic minorities in Denmark Hans Skifter Andersen, Danish Building Research Institute, Dr. Neergaardsvej 15, Denmark-2970 Hoersholm, Denmark. Tlph.:

2 Moving to and from enclaves. Housing choices of ethnic minorities in Denmark ABSTRACT In most European countries ethnic minorities has had a tendency to settle in certain parts of the cities and often in social housing together with countrymen so-called enclaves. An explanation for this could be low income combined with lack of knowledge of the housing market and discrimination, which limits the possibilities of ethnic minorities. Another, that immigrants for different reasons choose to settle in ethnic enclaves where they can find an ethnic network, which can support them in their new country. In traditional research literature about immigration it has been shown that living in enclaves for many immigrants has been a temporary situation. The 'spatial assimilation theory' says that this situation ends when the family has been more integrated in the new society and then moves to other parts of the city. This paper reports the results of a Danish study on moves and housing choice among ethnic minorities in Denmark based on public registers and a survey among moving households. The study confirms both explanations of why ethnic minorities moves to enclaves. It shows that there is a considerable higher probability that an ethnic minority family will move to a social housing area, if the number of residents in the area coming from the same country exceeds 50 persons. The study also to some extent confirms the spatial assimilation theory, which says that enclaves are not a permanent living place for immigrants. Many ethnic minorities move out of the enclaves and these families have stayed longer in the country, are more often in work and have Danish citizenship than families moving into or inside the enclaves. Introduction Housing preferences of ethnic minorities in European countries can to a great extent be expected to have the same explanations as for other citizens and be dependent on family situation, economic resources and local housing market possibilities. The research question here is to what extent housing preferences and housing choices among ethnic minorities differ from preferences and choices among people of national origin, and if there are other factors that influence or constrain housing choices of ethnic minorities. Evidence shows that the housing situation for ethnic minorities in most countries diverges. In most countries in Northern Europe there has in recent years been a growth in the number of immigrants and it has been a tendency that these families have settled in certain parts of the housing market and in limited parts of the cities (Musterd et. al. 1998) - in most countries in the least attractive parts of social housing. In this way parts of the cities have obtained a large share of ethnic minorities and have been transformed to what we call multiethnic neighbourhoods, where citizens of national origin have become a minority. Some researchers are of the opinion that the housing situation of ethnic minorities mostly can be explained by their lack of resources and by discrimination. Not only economic resources are important but also cognitive, political and social resources (Van Kempen 2003), which especially ethnic minorities often lack. In parts of the housing market good contacts to important persons or institutions are decisive for access to dwellings. This concerns especially private landlords. It is important to have good knowledge on the possibilities and rules on the housing market, which also often demands good language skills or good access to advisers. A study in Sweden (cited in Van Kempen and Özükren 1998) showed for example that some Turkish immigrants in social housing did not know there was a private rented sector in Sweden. Some other studies (Aalbers 2002, Andersson 1998) points to discriminative practices on the housing market, where social and private landlords to some extent exclude ethnic minorities from their housing. There could also be discriminative practices among banks or institutions providing capital for purchase of housing where ethnic minorities because of prejudices are seen as less solvent customers. Because of all these factors ethnic minorities are restricted to the least attractive parts of the housing stock, which often are located in certain parts of the cities. Alternative or supplementing explanations are connected to the notion that some ethnic minorities have special housing preferences or behaviour that are culturally conditioned or connected to their special situation as immigrants. Some studies (Zavodney 1998, Jaeger 2000, Bartel 1989, all referred in Damm 2002) shows that it has importance for immigrants' housing choice if there are many other residents of the same origin and ethnic social networks in the neighbourhood. In the opinion of some researchers (Musterd et. al. 1998, 181) this is only a parallel to a known phenomenon among all house hunters that people want to live together with someone who have a similar social status and cultural background. In the opinion of others (Wacquant 1997, Peach 1998, Murdi 2002) moving to neighbourhoods with many countrymen called ethnic enclaves - for 2

3 new immigrants is part of a strategy for survival and integration in their new country. Some of the arguments for this strategy are that immigrants often have family or friends in the enclaves who they want to live close to. It has been shown in some studies (Portes 1998; Sassen 1995; Damm and Rosholm, 2005)) that an ethnic network in the enclave can improve the ability of the members of the group to find a job. Often there are also local shops which purchase consumer goods from the homeland. Moreover, that can reduce the costs of using ethnic goods and services (Chiswick and Miller 1995), Moreover, the feeling of security and safety in a well known social and cultural environment can be important. Seen in this perspective the preferences for ethnic minorities to move to enclaves can depend on to what extent they are integrated in the new society. A hypothesis can be formulated that new immigrants and less integrated ethnic minorities have a greater need of the support they can get from networks in the enclave, which influence their housing choice. On the other side residents in enclaves that during the course of time get a stronger position in the new country could change their preferences in favour of moving away from the enclave. Studies of enclaves (Skifter Andersen 2006a, Wacquant 1997, Peach 1998) show that even if the share of ethnic minorities remains constant or increases there are many ethnic minorities moving out of the neighbourhoods and replaced by others. Earlier a theory, called spatial assimilation (e.g. Massey and Mullan 1984, Massey 1985), stemming from the earlier research of the Chicago School (Park 1925), has been formulated that immigrants often start their career in the new country by moving to enclaves, but that they often, after some time, move out again to housing that is more in accordance with their resources and needs. In this theory enclaves are only preserved because new immigrants all the time are coming into the country. Competing theories to the spatial assimilation theory are called the 'Ethnic resources' theory (Portes and Bach 1985) or the 'Cultural Preference' theory (Bolt 2006). They both say that access to ethnic resources and mobility possibilities inside the enclaves will motivate ethnic minorities to do their housing career inside the same neighbourhood, which will increase segregation. Research questions for the Danish study In the Danish study, which is referred below, a number of research questions have been in focus: To what extent can housing choices of moving ethnic minorities not be explained by their economic resources, housing needs and local housing market conditions? Are there parts of the housing market that are less accessible for ethnic minorities? Do ethnic minorities want to move to enclaves and why? What importance has the number of countrymen for which neighbourhoods they move to? Can Danish experience on the movements of ethnic minorities into, inside and out of enclaves support the spatial assimilation theory or the ethnic resource/cultural preference theories or both? The Danish study The study of housing choices among ethnic minorities in Denmark is based on two sets of data: 1. A database on all moving households in a certain year (2002) based on data from public registers 2. A survey among a sample of moving families of foreign ethnic origin and among households moving to and from three selected multiethnic neighbourhoods. The database on moving families contains, among others, data on the age, incomes, employment and place of birth of persons in every family. Moreover, there are data on their dwelling both before and after they have moved, based on the Danish Building and Housing Register (BBR), which is quite unique compared to other countries. Data on social housing estates and their residents have been selected from another database established by the Danish Building Research Institute. In this way it has been possible to analyse the composition of residents on the social housing estates that ethnic minorities have moved to and from. The survey has been conducted as interviews through telephone among a sample of households from the database. It contains questions on the background of the households, on their housing preferences, reasons for moving, reasons for choice of new dwelling and on their satisfaction with the new housing area. Some of the results of the study have already been published in Børresen 2006 (Review of European research literature on housing choice of ethnic minorities) and Skifter Andersen 2006a (Moves and housing choice of ethnic minorities in Denmark). 3

4 Ethnic minorities in Denmark The Danish housing market As can be seen from Table 2 a little more than half of Danish housing is owner-occupied. 26 per cent is either private renting or private co-operatives and about 20 per cent is social housing. Private renting and co-operatives are highly regulated sectors with rents and prices well below market levels especially in the larger cities. This means that there is a queue to these dwellings where people with good contacts to landlords and residents in co-operatives has a great advantage. In Danish social housing there is a system of waiting lists from which new tenants have to be taken. Sometimes the lists could be quite long, which means that households coming from other regions have difficult to get into social housing. Moreover, residents inside the sector has a special and much shorter waiting lists which means that is much easier for them to move to a new dwelling inside the sector than for outsiders. Ethnic minorities Ethnic minorities have in this study been defined as people coming from other countries than EU, other countries in Western Europe, North America, Australia and New Zealand. Included are immigrants born in these countries and their children born in Denmark, but only when both parents are immigrants. The defined ethnic minorities sum up to 5,5 per cent of the population. The study is based on analyses of households, which is more complicated because some households are mixed. Only 3,3 per cent of the Danish households contains only ethnic minorities. Moreover, 4 per cent are mixed households with ethnic minorities. Only in 0,5 percent of these the person in the household with the highest income belongs to ethnic minorities. In Table 1 the number of ethnic minority households is distributed on different groups of nationalities. It can be seen that compared to some other European countries the immigrants in Denmark are coming from many different places. The largest groups are Turks, Palestinians, Iraqis, Iranians, Somalis and Bosnians. (Table 1) The settlement of ethnic minorities in Denmark The majority of ethnic minority households in Denmark more than 60 per cent - are living in social housing. Only 20 per cent of all households are. We have in the study divided social housing in two groups: 1. housing estates where more than 40 per cent of the residents are ethnic minorities called multiethnic neighbourhoods and 2. estates where less than 40 per cent are. In Table 2 households with ethnic minorities is distributed on housing type and compared with all households in Only 2 per cent of all households are living in multiethnic neighbourhoods in social housing while 22 per cent of minority households are. The calculated figures of over-representation show that ethnic minorities are living more than ten times so often in multiethnic neighbourhoods and more than two times so often in other social housing compared to all households. (Table 2) The table shows that ethnic minorities in Denmark seldom live in owner occupied housing, especially in detached houses. Some of them have settled in private renting or co-operative housing, but less often than all households. Differences between ethnic groups There are large differences between different ethnic groups concerning housing and the extent to which they are concentrated in social housing and in multiethnic neighbourhoods. In Table 3 ethnic minorities from different groups of countries are distributed on type of housing. The over-representation of each group in types of housing is calculated in the last part of the table. Moreover, an index of segregation is calculated, which shows to what extent the distribution of a group between types of housing deviates from all households. The index of segregation shows that all ethnic groups are very unequally distributed in Danish housing but some groups more than others. The Somalis are the most segregated group followed by Arabs and Afghanis. (Table 3) 4

5 The three groups most concentrated in multiethnic neighbourhoods are Somalis, Turks and Arabs. More than 30 per cent of these groups are living in multiethnic neighbourhoods. The Pakistanis are also overrepresented, but on a somewhat lower level. The other ethnic groups are more seldom located in multiethnic neighbourhoods compared to all ethnic minorities, but much more often compared to the Danes. In the other parts of social housing there are relatively many Iranians and Afghanis. In total the Somalis is the group that is most concentrated in social housing. It is mostly immigrants from Asiatic countries (Vietnam, Sri Lanka etc.) that have spread to other parts of the hosing market than social housing. Pakistanis and immigrants from 'Other countries' (mostly South America) are more often living in owner occupied flats. Explanations for housing choices of ethnic minorities the importance of preferences for enclaves Reasons for choosing enclaves or multiethnic neighbourhoods To what extent and how do ethnic minorities show preferences for living in enclaves? Do they choose these neighbourhoods because they prefer the type of housing found in multiethnic neighbourhoods? In the survey we have asked for housing preferences among moving ethnic minorities and for the reasons why they chose the specific neighbourhood to which they moved. In Table 4 is shown some of the answers for both a sample of all moving ethnic minorities in 2004 and for some ethnic minorities moving into three specific multiethnic neighbourhoods with a very high share of ethnic minorities. (Table 4) The results from the Danish study points to that ethnic minorities do not choose multiethnic neighbourhoods because they prefer the type of housing found in them. Earlier Danish studies (SBi and AKF 2001) have shown that about 70 per cent of the Danish population wants to live in owner-occupied housing and 83 per cent to live in a detached house. The table shows that preferences among ethnic minorities are quite similar. But actually much fewer of them live in these types of housing. In the three selected multiethnic neighbourhoods more than 90 per cent of the dwellings are located in multi-storey blocks. These results show that ethnic minorities for different reasons move to other kinds of housing than they actually wants. Another conclusion that can be drawn fro the table is that a considerable share of ethnic minorities chooses their neighbourhood because they want to live close to family and friends. More than half of the respondents in both groups had for different reasons specifically chosen to live in this neighbourhood (and no others). There was no difference between all movers and movers to the multiethnic neighbourhoods in this respect. For a little less than half of all movers the main reason for this choice were that they had family and friends in the area. For only one third it was important that there were many countrymen. This points to that the close social network has much more importance than ethnic network in general. It is somewhat surprising that these reasons had much less importance among movers to the three multiethnic neighbourhoods where the probability for finding ethnic networks is much stronger. This points to that there are many other reasons for moving to multiethnic neighbourhoods, which could be much easier to get access to. The table also shows that it is preferences for enclaves and not preferences for multiethnic neighbourhoods with few Danes that are important. Very few of the respondents want to live in neighbourhoods where the majority of residents are ethnic minorities. But one third of them find that this does not matter. The preferences are somewhat different in the three selected multiethnic neighbourhoods Among in movers to these neighbourhoods only 18 per cent want to live in neighbourhoods where ethnic minorities constitutes less than one third, while more than 40 per cent did not care about it. Factors explaining moves to different types of housing how do ethnic minorities differ from Danes? To determine to what extent there are 'immigrant' or 'ethnic' explanations for ethnic minorities moving to social housing and multiethnic neighbourhoods we have used statistical analyses, where we try to eliminate the importance of other variables than ethnic origin. The population is all moving household I Denmark in Variables on the households, on the local housing market and social composition of geographic area (municipality) are included. The independent variables in the statistical analyses (logistic regression) are: 5

6 Independent variables: Ethnic minorities? Is the household belonging to the group of ethnic minorities as defined above? Age/10 The age of the person with the highest income divided by 10 Children? Are there children in the household? Couple? Is it a couple? Household income/ (DKK) Total income for all persons divided by (DKK) Employed? Is the person with the highest income employed? On welfare? Is the person with the highest income on welfare? In Greater Copenhagen? Does the family live in Greater Copenhagen? Size of city Code for city size (1-13) - higher values are smaller cities Moving inside same city? Was the old dwelling located in the same city? Moving between regions? Was the old dwelling located in another region Per cent owner-occupied housing in mun./10 Share of dwellings (divided by 10) in the municipality that are owner-occupied housing Per cent social housing in mun./10 Share of dwellings (divided by 10) in the municipality that are social housing Per cent ethnic minorities in mun./10 Share of inhabitants (divided by 10) in the municipality that are ethnic minorities Per cent unemployed in mun./10 Share of inhabitants (divided by 10) in the municipality that are not in work, under education or pensioners Moves from dwell. in social housing? Has the household moved from a dwelling in social housing Size of dwelling before move/10 Size in square meter of former dwelling The dependent variables in the five analyses are if households have moved to a specific type of housing. The results of the analyses are shown in Table 5 as the calculated increase in probability of moving to a certain type of housing by increasing the independent variables by one unit. They are based on the odds ratios. The established models can to a reasonable extent explain the variation in housing choice (Nagelkerke R2). (Table 5) It is shown that ethnic minorities have an increased probability of 125 per cent for moving to social housing and 408 per cent for moving to multiethnic neighbourhoods. The extensive moving of ethnic minorities to this housing thus cannot be explained by other factors as income and employment. Other explanations have to be found. If we look at the results for moving to social housing it can be seen that low income, unemployment and welfare recipients as expected are important reasons for both Danes and ethnic minorities. Families with children also more often move to social housing, but mostly among the lower income groups. The local housing market has some importance. The more social housing available the more often people move to social housing. It has also importance if the rest of the housing marked consists of owner-occupied housing or private renting, which to some extent competes with social housing. Taking these factors in account it does not have much importance which part of the country it is and the size of the city. The type and location of the former dwelling has some importance however. In Danish social housing there is a system of waiting lists from which new tenants have to be taken. Sometimes the lists could be quite long, which means that households coming from other regions have difficulties in getting into social housing. Moreover, residents inside the sector has a special and much shorter waiting list, which means that is much easier for them to move to a new dwelling inside the sector than for outsiders. It can be seen from the table that the probability for moving to social housing increases very much if the households are coming from other dwellings in social housing. There are also some interesting differences between the results for moves to multiethnic neighbourhoods and to all social housing. Movers to multiethnic neighbourhoods are much more often welfare recipients and have children. The households come more often from other regions and less often from the same city and from other social housing. They are as could be expected more often located in municipalities with many ethnic minorities. An explanation of why ethnic minorities move to social housing could be that they have difficulties in getting a dwelling in other parts of the housing market. The most obvious choice for ethnic minorities with low incomes is private renting and private co-operatives, which constitute about 25 per cent of the housing market. Both sectors have a somewhat rigid price regulation, which means that they are often cheaper than newer social housing - and especially the multiethnic neighbourhoods - and much cheaper than owneroccupied housing in the larger cities and towns. This results in a surplus demand. The access to these 6

7 dwellings especially to larger dwellings is therefore dependent on to what extent one has good connections to landlords or to residents in co-operatives. Ethnic minorities have seldom these kinds of social network and as can be seen from the table they have a reduced probability of 57 per cent for moving to this part of the housing market. It is thus less than half so often that ethnic minorities move to private renting and co-operative. Instead some of them move to owner-occupied flats where the access is only determined by your ability to pay the price. The ethnic minorities have an increased probability of 61 per cent for moving to owner-occupied flats. Statistical analyses of the connection between choice of neighbourhood in social housing and the ethnic composition of the chosen neighbourhood As mentioned in the introduction the literature on immigration shows that immigrants for different reasons and to a different extent moves to neighbourhoods were they can live close to their family, to friends or to fellow-countrymen. Based on the data on moving ethnic minorities we have conducted statistical analyses of the connection between the choice of neighbourhood, among households from different ethnic groups, and the ethnic composition of the chosen neighbourhood. The dependent variable in the models is if households, who move to a social housing estate, are immigrants from a specific country or group of countries. Eight logistic regression models are constructed in three versions. In the first version the share of residents of the same origin in the chosen neighbourhood is an independent variable. In Version 2 is instead used the actual number of residents from the same group (divided by 10). In Version 3 is used a logical variable are there more than 50 residents from the same ethnic group in the neighbourhood? Some of the same independent household variables are used as in Table 5. Moreover the size of the neighbourhood (divided by 100) is used. For the two last versions only the variables on the ethnic composition of the neighbourhood is shown. (Table 6) It can be seen from Table 6 that for all ethnic groups the probability of moving into a social housing estate increases with the share of residents that belong to the same group. This shows that there is a tendency for ethnic minorities to move into neighbourhoods with countrymen. But the effect is not very strong. If the share of Turks in the area e.g. increases with 10 per cent there is only a 5 per cent increase in the probability that a family that moves in is from Turkey. The groups with the highest increase in probability are Iranians, Pakistanis and Africans, who all are groups that not so often live in social housing and in multiethnic neighbourhoods. An explanation could be that it is not the share of residents that have importance but the actual number. In the second version of the models this variable has been used. The table shows that this variable in some cases has a stronger effect but again for the 'small' ethnic groups while it is not very strong for more dominant groups in social housing like Turks and Arabs. The explanation could be that these groups in many neighbourhoods have reached a size that is big enough to create a viable ethnic society and that an increase in the number of people in the group does not matter much. These groups can choose between many more neighbourhoods if they want to live together with countrymen. For the small groups there could be fewer neighbourhoods with a viable ethnic milieu why the number of countrymen on the estate has a much greater importance. We have therefore constructed a version 3 of the models where the independent variable is if the number of residents from the same group in the area exceeds a certain number in the chosen version 50 residents. The table shows that this variable has a very strong effect and that it also concerns some of the dominant groups. The probability of that a family that moves in is from Turkey thus increases more than three times when there are more than 50 Turks on the estate. All groups have a high increase in probability but this especially concerns the Somalis and immigrants from Balkan. Spatial assimilation? Comparison of ethnic minorities moving to and from multiethnic neighbourhoods The theory of 'spatial assimilation' for immigrants assumes that new immigrants often move into enclaves with many countrymen but that some of them only stay there for a couple of years and move out when they have become more integrated in the new society. To partially test this hypothesis we have compared households moving to and from multiethnic neighbourhoods with more than 40 per cent residents belonging to ethnic minorities. A simple comparison 7

8 (Skifter Andersen 2006a) confirms the hypothesis because it shows that newcomers that have been a shorter time in Denmark less often have work (see Table 9) and more seldom have Danish citizenship. In Table 7 is shown the results from a statistical analysis comparing newcomers and households moving out of the areas. The model tries to determine the variables explaining if a moving household is a newcomer to multiethnic neighbourhoods. The sample is only ethnic minority households moving to and from multiethnic neighbourhoods. (Table 7) The constructed model is not very successful in explaining the variation in differences between new comers and households moving away. This indicates that other important variables explaining why people move are lacking in the model, for example family and job changes. But there are some significant variables that tend to confirm the spatial assimilation theory. The newcomers have a higher probability of being refugees, they have more seldom obtained Danish citizenship, they tend to have lower incomes, are more often on welfare payments and more seldom in employment. They are younger and more often have children. Descendents of immigrants on the other hand tend to move away from the multiethnic neighbourhoods. But the fact that the ethnic minorities that move away from multiethnic neighbourhoods are more integrated than newcomers does not in itself prove the spatial assimilation hypothesis. Instead this result could just mirror a negative development in multiethnic neighbourhoods with an increasing concentration of less integrated ethnic minorities. To test if this is the case is used data from another study of the ethnic and social development in three multiethnic neighbourhoods. The ethnic development in recent years in three multiethnic neighbourhoods The study on three selected neighbourhoods with a high share of ethnic minorities among residents (Skifter Andersen 2006b) showed the same tendencies as for all multiethnic neighbourhoods: newcomers were less integrated that ethnic households moving away. To examine if this had negative consequences for the social development in the neighbourhoods the situation in 2004 was compared with 1998 (Table 8). (Table 8) A large share of residents in the three neighbourhoods were without employment, living on either early pension, welfare payments or unemployment benefits, both in 1998 and in For Vollsmose e.g. the total share in 1998 was 70 per cent of which 42 per cent were ethnic minorities and 28 per cent Danes. But it can be seen from the table that the total share actually decreased in the period. In the period all three estates had a decreasing population of Danes and an increasing number of residents belonging to ethnic minorities. Only in Vollsmose, which had a relatively large increase in the share of ethnic minorities there was also an increasing number of unemployed immigrants. In the two other neighbourhoods the number of unemployed immigrants was reduced in the period at the same time as the total number of immigrants increased. These figures do not point to an increasing number of unemployed residents in these multiethnic neighbourhoods in a period where newcomers were much less often employed than residents moving away from the estates. It can therefore be concluded that a number of immigrants in the neighbourhoods must have got employment in the period some of them moving away. This supports the spatial assimilation theory. Ethnic minorities staying in multiethnic neighbourhoods testing the Ethnic Resources and Cultural Preference Theories As mentioned in the introduction other theories have been put forward (Portes and Bach 1985; Bolt 2006) that the advantages for ethnic minorities to live in enclaves are so strong that they will tend to stay in the neighbourhoods and try to make their housing career inside the enclaves. In Danish social housing such a tendency is supported by the rules for access to dwellings. It is much easier to get a new dwelling in the same housing association because insiders have a much shorter waiting list. If changes in the family situation occur, which trigger of changes in housing needs, one can often much faster and easier get a new dwelling in the same neighbourhood. A study of this system (Skifter Andersen 2004) showed that the internal waiting lists more often were used by ethnic minorities and by households with low income. 8

9 In this study the internal movers in multiethnic neighbourhoods were compared with newcomers, all residents and households moving out. In Table 9 the social composition of ethnic minorities moving internal in multiethnic neighbourhoods is compared with the composition of newcomers, household leaving the estate and all residents. The ethnic minorities are divided into households, where the person with the highest income is either in work, without work (on early pension, welfare or unemployment benefits) or others (mostly old-age pensioners and students). (Table 9) It can be seen from the table that the internal movers are the group with the largest share of unemployed and the smallest in work. Especially they are less well of than the households moving out but also than the residents as a whole. This does not confirm the Ethnic Resources or Cultural Preference theories. If these theories were valid you would expect that a higher share of the internal movers were employed. Instead the figures point to that the weaker and less integrated immigrants try to make their housing career inside the multiethnic neighbourhoods while the more successful immigrants move away. Some of the employed immigrants, however, stay in the neighbourhood. In the study of three multiethnic neighbourhoods (Skifter Andersen 2006b, pp 47) it was shown that the social composition of households, who had stayed more than 10 years on the estate, was not very deviant from the average among residents with the same share of households in work. A question for further research is therefore why some ethnic minorities stay in the multiethnic neighbourhoods and others move away. A possible explanation could be the character of the social network, which the households have obtained in the neighbourhood. Close connections to family and friends could for some families be a crucial factor holding them back. There could also be 'ethnic' explanations. Some ethnic groups move away while others stay. Conclusions This study of housing choice among ethnic minorities showed that immigrants in Denmark to a great extent move to social housing and to neighbourhoods with many ethnic minorities among the residents called multiethnic neighbourhoods. Especially immigrants of Arab, Somali and Turk origin have gathered in multiethnic neighbourhoods. One of the reasons for this is shown to be that ethnic minorities have difficulties in getting access to the highly regulated private renting and co-operative sectors in Denmark. This points to that either discrimination or to lack of social network could be explanations. Moreover, ethnic minorities often are located in the largest and least attractive estates where most of the Danes do not want to live. Another explanation has been formulated as 'preferences for living in ethnic enclaves', which means neighbourhoods where a certain number of residents belong to the same ethnic group as the family in question. The Danish study supports this explanation. Ethnic minorities explicitly express preferences for living in neighbourhoods with many countrymen, but even more important for them is to live close to family and friends. Statistical analyses showed that the moves of ethnic minorities to multiethnic neighbourhoods cannot be explained by the usual variables as income, employment, family situation etc. The presence of countrymen in the chosen neighbourhood was showed to have a strong statistical effect in our analyses of housing choice. The analyses pointed to that the most important factor is if there are a certain number of countrymen in the neighbourhood to create a viable ethnic society. The total number of countrymen in the neighbourhood or the share of residents is not so important The study moreover tried to test two different theories on multiethnic enclaves. The 'Spatial Assimilation Theory' says that immigrants will leave the enclaves in the course of time as they get more integrated in their new society. This theory was to some extent confirmed by the Danish evidence as it was shown that ethnic households moving away from the multiethnic neighbourhoods had higher incomes and employment, more often had obtained Danish citizenship and had stayed longer in the country compared to newcomers at the same time as the employment level in the neighbourhoods did not change over the years. The study points to the fact that many of the immigrants moving to Danish multiethnic neighbourhoods are newly arrived and are not in work. Some of them move away again after some years when they have got employment and a better economic situation. Competing theories, called the 'Ethnic Resources Theory' or 'Cultural Preference theory' say that the advantages for ethnic minorities to live in enclaves are so strong that they will tend to stay in the neighbourhoods and try to make their housing career inside the enclaves. This theory was in the first time tested by comparing ethnic households moving inside the multiethnic neighbourhoods with newcomers, 9

10 households moving out and all residents. This evidence did not support the theory. It points to that only the weaker and less integrated immigrants try to make their housing career inside the multiethnic neighbourhoods while the more successful immigrants move away. It was, however, also shown that some of the more integrated ethnic minorities had stayed in the neighbourhood for more than 10 years. A possible explanation for this could be the strong importance ethnic minorities ascribe to live close to family and friends. Implications for integration and urban policies In Denmark and in many other countries there has for some years been a political focus on multiethnic neighbourhoods as a big problem in society. This focus has to a great extent been a result of increasing social problems in these areas with a high concentration of both Danes and ethnic minorities with low incomes and high unemployment. There is much evidence for that this focus is justified (Skifter Andersen 2003). But in Denmark in recent years there has been a growing conception of multiethnic neighbourhoods as an obstacle for integration of ethnic minorities. There has been created a picture of these neighbourhoods as a place where ethnic groups are accumulated and form 'parallel societies', which hamper or obstruct their integration in the Danish society and from which they will never escape. This conception has led to measures, which hampers the access for ethnic minorities to move into multiethnic neighbourhoods, and to proposals of demolition of estates. Our study has showed that this picture is not accurate. The neighbourhoods are far from being static and many immigrants use them as a temporary station on their way into the Danish society. Some are staying in the multiethnic neighbourhoods but we don't know if this is a problem. The literature on 'Neighbourhood Effects' in European multiethnic neighbourhoods does not have unambiguous conclusions about this (Schulz Larsen 2006, Damm et. al. 2006). References Aalbers, M. B. (2002). The neglected evidence of housing market discrimination in the Netherlands. Radical Statistics Journal, 79/80. Andersson, R. (1998). Socio-spatiel Dynamics: Ethnic Divisions of Mobility and Housing in Post-Palme Sweden. Urban Studies, 35 (3), Bartel, A. P. (1989). Where do the U.S. immigrants live? Journal of Labor Economics, 7(4): Bolt, G., van Ham, M. and van Kempen R. (2006) Immigrants on the housing market: spatial segregation and relocation dynamics. Paper for ENHR conference Ljubljana Børresen, S. K. (2006) Etniske minoriteters bosætning. Hvad viser forskningen (Housing choices of ethnic minorities. What can be concluded from the research literature?). SBi Hørsholm: Danish Building Research Institute. Chiswick, B. R. and Miller, P. W. (1995). The Endogenity between Language and Learning: International Analyses. Journal of Labour Economics 13(2), Damn, A. P. (2002). Etniske minoriteters bosætning og flytninger de seneste 20 års økonomisk kvantitative forskningsresultater (Moves and housing choices of ethnic minorities research from the last 20 years). AMID Working Paper Series 13/2002. Aalborg University Damm, A. P. and Rosholm, M. (2005). Employment Effects of Spatial Dispersal of Refugees. University of Copenhagen, Institute of Economics. CAM WP Damm, A. P., Schultz-Nielsen, M. L. and Tranæs, T. (2006). En befolkning deler sig op?. Copenhagen: Gyldendal Jaeger, D. A. (2000) Local Labor Markets, Admission Categories, and Immigrant Location Choice. Working Paper. Hunter College. Princeton University. 10

11 Massey, D. S. and Mullan, B. P. (1984). Processes of Hispanic and Black Spatial Assimilation. American Journal of Sociology, 99: Massey, D. S. (1985). Ethnic residential segregation: A theoretical synthesis and empirical review. Sociology and Social Research, 69, Musterd, S. Ostendorf, W. and Breebaart, M. (1998) Multi-ethnic Metropolis: Paterns and Policies. Dordrecht: Kluwer Publications. Murdie, R. A. (2002). The Housing Careers of Polish and Somali Newcomers in Toronto s Rental Market. Housing Studies, 17 (3), Park, R.E. (1925) The city: suggestions for the investigation of human behavior in the urban environment. In R.E. Park & E.W. Burgess (Red.) The City (pp. 1-46). Chicago: University of Chicago Press. Peach, C. (1998). South Asian and Caribbean Ethnic Minority Housing Choice in Britain. Urban Studies, 35 (10), Portes, A. and Bach, R. L. (1985). Latin Journey:Cuban and Mexican Immigrants in the United States. Los Angeles: University of California. Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. Annual Review of Sociology 24, SBi and AKF (2001). Det danske boligmarked. Udvikling i boligforsyning og boligønsker (The Danish Housing market. Deverlopment in housing supply and demand.). Copenhagen: Byforum. Sassen, S. (1995) Immigration and local labour markets. In Portes, A. (Ed.) The Economic Sociology of Immigration. New York: Russel Sage, Schulz Larsen, T. (2006). Neighbourhood effects ansats til en kritisk refleksion. (Neighbourhood effects a critical evaluation). Working paper. Roskilde University Skifter Andersen, H. (2006a). Etniske minoriteters flytninger og boligvalg. (Moves and housing choices among ethnic minorities) SBi 2006:02. Hørsholm: Danish Building Research Institute. Skifter Andersen, H. (2006b). Undersøgelse af ti- og fraflytningen fra tre multietniske boligområder (A study of movers to and from three multiethnic neighbourhoods). Hørsholm: Danish Building Research Institute. Skifter Andersen, H. (2004). Virkninger af oprykningsretten i almene boliger. København: Erhvervs- og Boligstyrelsen Skifter Andersen, H. (2003). Urban sores. On the interaction between segregation urban decay and deprived neighbourhoods. Aldershot: Ashgate Wacquant, L. (1997). Marginalitet i storbyerne i det kommende årtusinde. Social Kritik, (52-53). van Kempen, R. (2003). Segregation and Housing Conditions of Immigrants in Western European Cities. Eurex Lecture 7 March 13th Located on: VanKempen.pdf van Kempen, R. & Özüekren, A. S. (1998). Ethnic Segregation in Cities: New Forms and Explanations in a Dynamic World. Urban Studies, 35 (10),

12 Zavodney, M. (1998) Welfare and the locational choices of new immigrants, Federal Reserve Bank of Dallas Economic Review, Second Quarter Åslund, Olof (2001): Now and forever? Initial and subsequent location choices of immigrants. (Working paper 2001:11). IFAU Office of Labour Market Policy Evaluation. 12

13 Table 1. Number and share of households in Denmark, where the person with the highest income is belonging to an ethnic minority group 2004 Number of households Share of all Turkey ,6% Pakistan ,2% Arabic countries ,8% Iran ,2% Afghanistan ,1% Somalia ,2% Balkan og Eastern Eur ,6% Other Asiatic countries ,7% Other African ,2% Other countries ,2% All ethnic minorities ,8% Table 2 Ethnic minority households distributed on type of housing compared to all households, and calculated over-representation in different types of housing 2004 (per cent) Ethnic minority All households Over-representation of ethnic minorities % households Social h. > 40 % ethn. min. 22% 2% 948 Social h. < 40 % ethn. min 39% 18% 117 Priv. rent.+ co-ops 22% 26% -15 Own.occ. flats 5% 5% -4 Own.occ. detached 10% 48% -79 Other 2% 2% 33 Total 100% 100% Note: Over-representation is calculated as the difference in per cent between the share of ethnic minorities living in a certain type of housing compared with the share of all households living in the same kind of housing 13

14 Table 3 Ethnic minority households from different groups of countries distributed on type of housing. Over-representation in and index of segregation between types of housings 2004 (per cent) Arabic countri es Iran Afghani stan Turkey Pakistan Somalia Balkan Other Asiatic countries Other African Other countri es All dwellings Social h. > 40 % ethn. min. 30% 25% 32% 13% 13% 31% 19% 14% 13% 11% 2% Social h. < 40 % ethn. min 41% 31% 40% 51% 49% 49% 40% 33% 38% 27% 18% Priv. rent.+ co-ops 14% 25% 20% 18% 34% 17% 22% 24% 30% 35% 26% Own.occ. flats 6% 8% 2% 5% 1% 0% 5% 6% 5% 8% 5% Own.occ. detached 8% 10% 4% 11% 1% 1% 12% 20% 11% 18% 48% Other 1% 1% 2% 2% 2% 3% 2% 4% 4% 2% 2% Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Over-representation compared to all ethnic minorities Social h. > 40 % ethn. min Social h. < 40 % ethn. min Priv. rent.+ co-ops Own.occ. flats Own.occ. detached Other Index of segregation 45% 39% 52% 44% 51% 61% 39% 29% 37% 30% Note: Over-representation is calculated as the difference in per cent between the share of a group living in a certain type of housing compared with the share of all ethnic minorities living in the same kind of housing. Index of segregation is a measure of the difference between the distribution of the group on type of housings and the distribution of all households. Table 4. Preferences for kind of dwelling and for ethnic composition of neighbourhood and reasons for choosing the actual neighbourhood among a sample of all moving ethnic minorities and among ethnic minorities moving to three multiethnic neighbourhoods Preferences for dwelling Sample of all movers Detached house 74 % 84 % Owner-occupied 75 % 66 % Preferences for ethnic composition of residents in neighbourhood Less than one third ethnic minorities 29 % 18 % About half ethnic minorities 33 % 41 % Mostly ethnic minorities 2 % Does not matter 32 % 41 % Reasons for choosing the neighbourhood Did only want to live in this neighbourhood 58 % 59 % Because of family and friends 36 % 47 % Because of many countrymen 19 % 33 % Number of respondents Movers to three multiethnic neighbourhoods 14

15 Table 5 Results from five logistic regressions, which try to explain why households move to certain types of housing (increased probability by increasing independent variables by one unit) Social housing Multiethnic neighbourh oods Private renting etc. Owner occup. flats Owner occup. detached Ethnic minorities? 125% 408% -57% 61% -34% Age/10 8% -2%*) -8% 1% 16% Children? 68% 84% -28% -55% 74% Couple? 12% -8% -8% 200% Household income/ (DKK) -27% -19% -15% 6% 40% Employed? -15% -7%*) 18% 205% 155% On welfare? 40% 86% 12% -82% -86% In Greater Copenhagen? -17% -71% -32% 67% -21% Size of city -8% -16% -1% -21% 11% Moving inside same city? 19% -4% -9% -11% Moving between regions? -26% 12% 4% -7%*) -4%*) Per cent owner-occupied housing in mun./10 28% -22% -26% 3% 62% Per cent social housing in mun./10 77% -19% -39% -9% 34% Per cent ethnic minorities in mun./10 357% -8% 37% 3%*) Per cent unemployed in mun./10-4% 43% 5% 5%*) 5% Moves from dwell. in social housing? 288% 132% -55% -56% -37% Size of dwelling before moving/10-3% -5% 1% 2% 4% Nagelkerke R2 0,3 0,219 0,157 0,163 0,488 Note: The figures show the increased probability (per cent) for choosing a certain type of housing by increasing the independent variable by one unit (based on odds ratios). All shown results has a significance below 0,01 per cent except for the ones marked with *), which are between 0,01 and 0,05. Missing figures are not significant. 15

16 Table 6 Results of 8 logistic regression analyses in 3 versions that try to explain if an immigrant household moving into a social housing estate originates from a specific country or group of countries (increased probability by increasing independent variables by one unit). Arabic countries Iran Turkey Pakistan Somalia Balkan or Other Asiatic Eastern Eur. countries Version 1: Share of residents from same ethnic group (%) / 10 5% 8% 4% 8% 7% 6% 4% 8% Size of housing area (100 dwellings) 24% 32% 18% 27% Age/ 10-25% -8% 21% 23% 12% -10% Children? -43% 27% 24% -20% -17% Couples? 40% 36% 65% -55% -34% -69% Household income/ DKK -14% -13% 28% 21% Employed? 29% -17% -23% -32% 43% On welfare? -54% -64% 67% 87% -34% -48% -38% Greater Copenhagen? 305% 32% -37% -56% 54% City size code -6% -19% -15% -13% Nagelkerke R2 0,25 0,26 0,22 0,18 0,30 0,33 0,16 0,15 Version 2: Number of residents from same ethnic group in area/100 8% 24% 5% 206% 47% 84% 28% 294% Version 3: More than 50 residents from same ethnic group? 240% 156% 148% 189% 373% 305% 139% 114% Note: The figures show the increased probability (per cent) by increasing the independent variable by one unit (based on odds ratios). All shown results has a significance below 0,05 Other African Table 7. Results from a logistic analysis of the difference between newcomers and ethnic minority households moving away from multiethnic neighbourhoods - increased probability for being a newcomer. Increased probability Has Danish citizenship? -37% Descendent e.g. born in Denmark? -28% Refugee? 22% Age/10-10% Have children? 14% Household income/ DKK -14% On welfare? 34% In employment? -20%**) Nagelkerke R2 0,038 Note: As Table 6. **) This Result is only significant, when income is omittede from the model 16