Escaping Poverty Neighbourhoods: Theory, Practice and the Role of Policy in the Netherlands

 

Gideon Bolt and Ronald van Kempen[1]

 

Urban Research Centre Utrecht

Faculty of Geographical Sciences

Utrecht University

PO Box 80.115

3508 TC Utrecht

The Netherlands

Email: G.Bolt@geog.uu.nl or R.vanKempen@geog.uu.nl

 

Draft (not to be cited without permission of the authors)

 

Abstract . While patterns of spatial segregation and concentration of ethnic minority groups in European cities have been documented very well, more dynamic figures on migration have been more scarce until now. We especially do not seem to know much about the possibilities of different kinds of households to escape from poverty areas. Escaping these areas can be important in order to escape unwanted housing and living conditions. The analyses in this paper show that in the Netherlands the chance for escaping such poverty neighbourhoods partly depends on ethnic descent: some immi­grant groups are less able to escape than Dutch households. The policy implications of this result will be outlined in the last part of the paper.

 

 

1.         Introduction

 

A long tradition exists in the field of research into processes of spatial segregation and concentration of minority ethnic groups in American and European cities. In most cases it is concluded that segregation and concentration of minority ethnic groups do exist. This is a more or less logical conclusion, because it would be a mere coincidence if groups (ethnic minorities, low-income households, but also blue-eyed men and brown-haired women) would be spread out exactly evenly over an urban area. Assessing the seriousness of the patterns found is another, and also a different story. Segregation indices may be high, for example compared to other countries, but this does not necessarily mean that problems are attached to these figures. A high spatial concen­tration of a certain group may very well be the result of a deliberate choice of many group members who want to live in proximity to each other and who profit from each other in all kinds of ways. However, in general, high segregation and concentration figures are related to problematic situations in many publications.

Changes in patterns of segregation and concentration are often detected by comparing two or more points in time. On the basis of such a comparison it can be concluded that segregation and concen­tration has increased or not and that some areas have become more important for housing members from different groups, while others have become less important (among many examples of these kinds of analysis, we can for example mention the following recent articles: Peach (1998) for London, Brevik (2001) for Oslo and Van Kempen and Van Weesep (1997) and Bolt et al. (2002) for the Netherlands, Ladanyi (2001) for Budapest, Giffinger and Reeger (1998) for Vienna; Kemper (1998) for Berlin, Murdie and Borgegard (1998) for Stockholm). Some­times international comparisons are made (see for example Kesteloot and Cortie (1998) for Belgium and the Nether­lands and Musterd et al. (1998) for a few European coun­tries and Canada). The tradition of this kind of research goes back to the Chicago School in the 1920s.

Comparing patterns of segregation and concentration over time does however not say anything about the people involved. Declining figures of segregation for a specific group may indicate that a lot of people belonging to this group have been able to move away from one area to other areas, but this does not necessarily have to be the case. Likewise, stable figures of segregation and concentration may indicate stable patterns, but they might as well hide enormous dynamics. When in an area 1,000 people of a group move out and 1,000 people belonging to the same group move in, concentration figures may very well be constant (depending of the in- and out-movement of other groups), but in the meantime we might have to talk about a total new population. Surprisingly, in most segregation research, there is no attention paid to the role of residential mobility. With the exception of a few studies (cf. Massey et al. 1994; South & Crowder 1997a, 1997b), there is no attention for the residential moves between neighbourhoods that bring about changes in the patterns of residential segregation.

To our opinion it is important to pay attention to dynamics. It makes a lot of difference if people are able to move out of some neighbourhoods or not. Especially the possibility to escape from unattractive neighbourhoods can be very important for the well-being of a person. Having to live in areas that are characterised by all kinds of negative aspects, such as high criminality, noise, pollution, lack of amenities, and the presence of people that have different or even competing lifestyles may impede people from living as they like to live. Having the possibility to escape from these areas, may lead to a better housing situation and to a better life in general.

The main question addressed in this paper is to which extent ethnic minority members in the Netherlands succeed in moving out of poor neighbourhoods. We believe this question is relevant for different reasons. Especially in the American literature residential mobility out of poor neighbourhoods is believed to enhance employment prospects, to reduce exposure to crime, and to increase access to a variety of services and facilities (South & Crowder 1997a). Although social-spatial cleavages in the Nether­lands are much smaller, there is evidence that living in a poor neighbourhood in the Netherlands heightens the risk of violent victimisation (Wittebrood 2000) and that the quality of life in many poor neighbourhoods is under pressure (Hooimeijer & Van Kempen 2000).

A study of the residential mobility of ethnic minorities not only sheds light on the extent to which they succeed to improve their life chances, but also offers insight into the process of integration into the host society. From the assimilation perspective, minority groups are expected to improve their neighbourhood environments as their social economic status and their acculturation increases. Since the largest ethnic groups in the Netherlands have arrived relatively in the Netherlands relatively recently, it is to be questioned how far the assimilation process has advanced.

In section 2 of this paper we will briefly focus on some theoretical aspects that relate to the main question of this paper. Section 3 presents some basic segregation data of the Netherlands and especially the four largest cities. This section is added to give a general picture of the Dutch situation with respect to segregation and concentration. It shows that segregation and concentration do exist, but that figures are not comparable to the United States (from which country much of the literature on the topic has been derived). Section 4 gives a brief overview of the data that are used for the analysis of the escape from poverty neighbourhoods. Section 5 focuses on the analysis itself. On the basis of crosstables and logistic analysis we will try to formulate an answer to our main question. Section 6 gives a discussion of the results and relates them to present urban and housing policies in the Netherlands. The conclusions are presented in section 7.

 

2.          Theoretical backgrounds

 

The question on the possibility to escape from poverty neighbourhoods is first and foremost a question on residential mobility. Who moves and who does not move? An enormous amount of literature exists on the importance of life-cycle factors as critical determinants of the decision to move. Household composition and age of the head of the household can be seen as the most important variables here (see Rossi 1955). While these variables can be linked to the overall probability of moving (young households move more often than older households, family households more often stay put than young singles), it is generally less clear how they affect the choice of a new location or housing situation. Families with children are probably more prone to find a safe, quiet and green area than young singles, who might choose more often for an urban environment with a lot of urban facilities, while taking the liabilities for granted, at least for a certain period of their life.

                    Decisions to move are in most cases made on the basis of preferences (although we should not forgot that sometimes decisions to move can have a very forced character, for example in the case of fire, urban renewal or divorces). Preferences can be defined as more or less concrete transformations of the general goals people have in a certain period of their life (Mulder 1993). Different groups, even different households have divergent ideas of what constitutes a desirable housing situation. Each household has its own subjective hierarchy, which is not static. Research in the behavioural tradition focuses on these preferences.

            Decisions to move can never be made on the basis of preferences only. They are made in a context of constraints and perceived oppor­tunities. Con­straints can be seen as pressures or obstacles. They prohibit people from regarding parts of the housing market as opportunities, as possibilities to move to (Mulder 1993). Constraints can arise from factors such as absolute shortages in the housing market or a competition of many households for the same type of housing (for example in­expensive social rented dwellings). Choices can be constrained by the number of dwellings that is available and affordable and by allocation rules.

                    Individual resources are tools the households use to overcome the constraints. People are distinguished from one another by their strength in the housing market (Rex 1968). The strength is determined by the resources available to the household. Four  kinds of individual resources can be discerned: material, cognitive, political and social resources (see also Van Kempen & Özüekren 1998). Especially the role of socio-economic differences should be stressed as an important variable that distinguishes between those who are able to move to non-poor areas and those who either remain behind or move to other poor areas.

Opportunities are the options that are open to individuals or households. Oppor­tunities can also be seen as the . choice set. of available alternatives (Mulder 1993). Potential movers have the possibility to accept or reject the available opportunity (Gordon & Vickerman 1982). Opportunities emerge from an intricate interplay between individual preferences and resources on the one hand and external constraints on the other hand. Moreover, macro-level factors and developments can be influential. Individuals and households do not act in a vacuum, but operate within the economic, demographic, socio-cultural and political context of their countries, their regions, their cities and their neighbourhoods. There is, in other words, a close relationship between individual preferences and resources, constraints and opportunities and the wider social structure (Phillips & Karn 1992).

        Finally, the present housing condition can be a key resource for the future housing situation. This is clear with regard to the ownership of a house, but it can relate to rental dwellings as well. In some housing markets, the inhabitants of certain types of rental units have allocation priority for other dwellings. In other cases, a more desirable housing situation can be obtained by swapping rented dwellings. In general, households living in owner-occupied dwellings are less prone to move than renters (see, e.g. Clark and Dieleman 1996).

 

The aspects mentioned above apply to al kinds of households and individuals, they may hold for ethnic minorities, as well as for nationals. In the literature some specific aspects have been mentioned, however, that specifically relate to minority ethnic groups. Especially in the UK, a so-called ethnic-cultural approach has been developed that can be seen as a special form of the behavioural approach. The general argument within the ethnic-cultural approach runs thus: housing conditions and residential patterns differ between groups, and these differences can be attributed to cultural differences between these groups. There is a clear element of choice in this approach (see e.g., Peach and Smith 1981, Robinson 1981, Bowes et al. 1997).

 

All these aspects relate to movements in general. With respect to movements between areas some other notions are important. They can be summarised in two models: the spatial assimilation model and the place stratification model. Both models have been frequently mentioned in the American literature.

                    The spatial assimilation model states, following the work of Park (1925) and Gordon (1964), that residential mobility of minority group members into better neighbourhoods is a part of a more general process of socio-economic and cultural assimilation (Logan and Alba 1993; South and Crowder 1997a; Fong and Wilkes 1999; Freeman 2000). Immigrants often arrive as chain migrants and tend to choose for an ethnic enclave where they take advantage of the help of their fellow countrymen and where they can maintain the lifestyle they had in the country of origin. The fact that the socio-economic status and the neighbourhood quality of these enclaves are usually low, is compensated by the advantages of living in an ethnic community. However, as the degree of acculturation rises, the neighbourhood preferences of immigrants tend to change and they will, like the members of the majority, opt for a neighbourhood with the highest economic status they can achieve. As immigrants move up the social ladder in the course of time, they will be better able to afford dwellings in neighbourhoods with better amenities. Summarising, acculturation provides the desire and social mobility the means for immigrants to achieve spatial assimilation (Massey and Denton 1985). The implication of this is that second generation immigrants will generally succeed better to move from poverty neighbourhoods than those belonging to the first generation of immigrants.

            The place stratification model emphasises the role of discrimination. The most discriminated immigrant groups are not able to match their social economic status with that of their neighbourhood. As a consequence of continuing discrimination they are less able to move into desirable neighbourhoods than other groups, even when socio-economic status is controlled for. According to this model, differences between ethnic groups will not disappear when socio-economic variables are controlled for. The effects of individual economic resources are weaker for minority groups than for majority groups; housing discrimination reduces the locational . return. minority group members receive for their human capital (see, e.g., South and Crowder 1997a). The role of discrimination can also be found in the literature in which the role of . gatekeepers. is stressed. Gatekeepers can also influence the possibilities minority ethnic households have. The crucial role of these . managers. is stressed in the work of Pahl (1975; 1977) and Lipsky (1980).

 

Ethnic minority groups are almost always over-represented in the larger cities of a country. Generally most poor neighbourhoods can also be found here. This implies that this variable should be taken into account when analysing the possible moves out of poverty neighbourhoods to other areas.

 

3.          Segregation and spatial concentration in Dutch cities

 

Before we continue with the analysis of the differential possibilities to escape from poverty neighbourhoods, we will give a brief overview on segregation patterns in the Netherlands. In the Netherlands, the large ethnic minority groups are very strongly concentrated in the big cities. In 1992, 57 per cent of the Surinamese, 48 per cent of the Moroccans, 37 per cent of the Turks and 31 per cent of the Antilleans lived in the four big cities. The share of the four big cities in the total population was only 13 per cent. The concentration of the ethnic minority­ groups in the four big cities has remained more or less stable in the last decade (Bolt et al. 2002).

                   At the same time, the sub­urbanisation of Turks and Moroccans remains particular­ly low in comparison with the indigenous population. The limited suburbanisation can be partly explained by the poor income position of many ethnic minority groups. Their housing market position appears to be weaker than that of comparable indigenous groups, so that they have fewer oppor­tunities to leave the city (Bootsma 1998).

 

How can the development of ethnic segregation in the four largest Dutch cities be characterised? Segregation is markedly stronger in Rotterdam and The Hague than in Amsterdam or Utrecht. It is also clear that in general Turks and Moroccans display a stronger segregation than other groups, such as the Surinamese and Antilleans. Between 1980 and the first half of the 1990s few large shifts occurred in the scale of the segregation (table 1).

         Apparently it is not the case that the various groups spread over the city to an increasing extent. Only in The Hague (often called the most segregated city of the Netherlands) was there any sign of a decline in the segregation of the different groups. In the other three big cities there is evidence of a fairly stable level of segregation. As already said in the introduction to this paper, this does not necessarily mean that no changes have occurred. Dynamic processes are not displayed in these figures and the same value of the segregation index might hide different patterns of spatial concentration. Bolt et al. (2002) and Van Kempen and Van Weesep (1997) have indicated that the fairly stable segregation figures for example do not show that the most important concentration areas for Turks and Moroccans in Amsterdam and Utrecht are now the areas built in the period 1945-1960 and in the 1960s, while the relative importance of the pre-war neighbourhoods near the inner city is declining.

Table 1      Segregation index¹ of ethnic minority groups² in the four big cities, 1980, 1986, 1995 and 1998

 

	Turks	Moroccans	Surinamese	Antilleans	Minorities3
Amsterdam  1980  1986  1995  1998	37.3 38.8 40.7 42.3	38.8 36.9 39.1 41.2	27.8 33.7 34.8 34.2	26.2 33.0 34.9 35.1	30.0 31.5 31.7 33.1
Rotterdam  1980  1986  1995  1998	. . 51.7 49.9	. . 46.8 44.2	. . 28.6 26.2	. . 28.5 25.7	46.5 43.0 43.0 40.8
The Hague  1980  1986  1995  1998	66.4 65.1 54.6 53.0	64.7 57.3 49.9 48.6	. 46.4 40.2 38.7	. 26.9 25.5 25.9	55.5 52.7 43.1 41.8
Utrecht  1980  1986  1995  1998	. 46.4 43.2 43.1	. 37.2 42.2 44.5	. . 24.0 23.4	. . 22.5 19.5	37.1 33.2 32.9 34.8

1. The segregation index indicates the share of the population ­category which would have to move to obtain a distribution over the municipality which is equivalent to that of the rest of the population. The maximum value is 100, the minimum value is 0.

2. A member of an ethnic minority is someone born outside the Netherlands, or having one parent born outside the Netherlands.

3. Turks, Moroccans, Surinamese, Antilleans, South-Europeans; for 1995 and 1998 also . other non industrialised countries. .

Source: Tesser et al. (1995); Bolt and Van Kempen (2000); Bolt et al. (2002)

 

 

4.          Escaping poverty neighbourhoods: data

 

The Dutch Housing Demand Survey (HDS) is a large-sample survey conducted every four years. The analyses in this paper are based on the sample of 1998. The sample is person-based and representative of the Dutch population aged 18 and over, not living in institutions. The survey is intended to provide insight into the composition of house­holds, their housing situation and housing preferences. The data set contains not only information about the present housing situation, but also about the residential mobility since 1994. For those who moved since 1994, the previous housing situation and the previous postal code is known.

                    The data set contains geographical identifiers in the form of postal code area. Four-digit postal code areas roughly correspond to the level of a neighbourhood. The average number of addresses for a postal code area is about 2000. Therefore, postal code areas are smaller in urban than in rural areas. To make a distinction between poor and non-poor neighbourhoods income data at the level of the postal code area data from the Internal Revenue Service are used. Because of the scale of this sample (5.5 million persons) the data on the level of the postal code area are reliable. In this paper a neighbourhood is classified as poor when at least 30 percent of the households living in that neighbourhood can be found in the lowest quintile of the income distribution in 1998. It is assumed that there is a considerable stability in the poverty rate over the period from 1994 to 1998.

 

5.            Escaping poverty neighbourhoods: analysis

 

Before we focus on the possibility to escape from poverty areas, it is necessary to know how many people actually live in poverty areas. From table 2 it can be see that in 1994 only 9.8 per cent of the native Dutch respondents lived in a poor neighbourhood or moved out of such a neighbourhood since then.[2] The comparable figures for Surinamese and Turks/Moroccans are around 50 per cent.[3] This indicates a big difference between ethnic groups. From earlier research we also know that at least income is a determining factor for these differences: Surinamese, Turks and Moroccans have on average much lower incomes than the native Dutch (see, e.g., Van Kempen 1997).

 

Table 2:      Proportion per ethnic group that lived in a poverty neighbourhood in 1994 (or moved out of such a neighbourhood since 1994)

 

	Native Dutch	Surinamese	Turks and Moroccans
Poverty neighbourhood	9.8	48.9	51.5
Other neighbourhoods	90.2	51.1	48.5
Total (unweighted)	87,618	909	988

 

Source: Housing Demand Survey 1998, CBS/VROM; WMD 1998/RIO/reworked by ABF

 

Table 3 gives some descriptive statistics of the native Dutch, the Surinamese and the Turks and Moroccans involved in the analysis (i.e. all people belonging to these groups and living in poverty neighbourhoods in 1994). It can be seen that especially the Surinamese can be found in the largest cities of the Netherlands, but also more than 70 per cent of the Turks and Moroccans were living there in 1994. The native Dutch are far less oriented to the largest urban areas. This gives already one clue for the explanation of the differences between the groups. Because the immigrant groups are much more concentrated in urban areas and because poverty areas can be found disproportionately in urban areas, the change for the Surinamese and the Turks and Moroccans is theoretically bigger to end up in a poverty area than for the native Dutch.

            The number of households with children is especially large for the Turks and Moroccans, while the majority of the native Dutch households living in poverty areas is part of a household without children. The level of education is especially low for the Turks and Moroccans. Dutch households live relatively more often in owner-occupied dwellings. Most immigrants still belong to the first generation. These are the people that, for different reasons, migrated from the home country towards the Netherlands. The overwhelming majority of first generation Turks and Moroccans came to the Netherlands as guest workers. During their stay here, they decided not to go back to their home countries and have their families came over. Surinamese came here for more political reasons: especially just before 1975, when the colony of Surinam became independent from the Netherlands, many Surinamese migrated to the Netherlands, which at that moment was still relatively easy.

 

Table 3:             Descriptive statistics for predictors of mobility out of poor neighbourhoods

 

	Native Dutch	Surinamese	Turks and Moroccans
City Size   < 50.000 50 . 100.000 inhabitants  100 - 150.000 inhabitants  150 - 250.000 inhabitants    > 250.000 inhabitants	18.7 9.8 16.1 10.7 44.8	2.9 3.8 5.0 3.8 84.4	4.3 6.3 9.8 8.4 71.1
Household structure   one person household    more person household with                         children   more person household without                         children	27.7 23.5   48.8	21.7 50.7   27.6	8.6 73.3   18.1
Level of education  Low   Intermediate  High	52.5 26.5 21.0	59.5 27.8 12.7	82.7 12.9 4.3
Tenure  Owner occupied  Rented	30.8 69.2	16.3 83.7	10.4 89.6
Age Age square  Annual income (x 1000 guilders)	48.8 2790.0 47.1	40.4 4289.1 41.2	37.4 942.3 36.8
Generation   First generation (arrived <1980)    First generation (arrived > 1980)   Second generation		51.8 29.4 18.8	40.5 25.8 33.8
Involved in political discussions  Actively  Passively  Not involved	44.4 37.3 18.3	30.3 45.2 24.4	18.7 32.0 49.3
Total	8,041	442	509

 

Source: Housing Demand Survey 1998, CBS/VROM; WMD 1998/RIO/reworked by ABF

 

Who have been able to escape from poverty neighbourhoods between 1994 and 1998? The native Dutch have been considerably more successful than the Surinamese and (especially) the Turks and Moroccans. Moreover, while the Surinamese and the Turks and Moroccans have been more mobile than the native Dutch in the research period, their moves were relatively often between one and the other poverty neighbourhood (table 4). This probably indicates that the three immigrant groups have been able to improve their housing situation, but in general only in much smaller steps than the native Dutch. These small changes have been detected in other investigations among these groups. Bolt (2001) and Bolt and Van Kempen (2002) have labelled them as sideways moves (in contrast to a more normal pattern of upward moves).

 

How can these differences be explained? We used a logistic regression analysis to find out which variables are important for the explanation of the chance to move out of poor neighbourhoods. The dependent variables in the model are, of course, ethnic group, city size (because we expect differences between large and small cities), age (because younger people are generally more mobile than older people), household structure (because we expect that households with children are more prone to leave poverty neighbourhoods), income (for the obvious reason that a larger income may provide more opportunities to move to a better place) and tenure (it is known from the literature that owner-occupiers are generally less mobile than renters). The results of the analysis can be found in table 5. In this analysis, every variable contains a zero value for one reference value or category (for example the native Dutch have the zero-value for the variable ethnic groups. The B-values in the table can then be read as deviations from this zero-value. A B-value lower than zero indicates a lower chance to move out of a poverty neighbourhood, a B-value higher than zero indicates a higher chance of moving out of such a neighbourhood.

 

Table 4:    Escaping poverty neighbourhoods

 

	Native Dutch	Surinamese	Turks and Moroccans
Did not move at all	64.7	53.7	53.8
Moved to another poverty neighbourhood	15.9	33.1	36.9
Moved out of a poverty neighbourhoud	19.5	13.2	9.3
Total (unweighted)	8,041	442	509

 

Source: Housing Demand Survey 1998, CBS/VROM; WMD 1998/RIO/reworked by ABF

 

The first column confirms what we already know from table 4: Surinamese and Turks and Moroccans have been less able to move from poverty neighbourhoods. In the second column, city size is added as a variable. The effect of this variable is as expected. Living in a large city decreases the chance of moving out of a poverty neighbourhood. The inclusion of city size in the model makes the minus-sign of the Surinamese change into a plus-sign, while the B-value for Turks and Moroccans becomes considerably lower compared to the first column. In other words: the strong concentration of Surinamese in the larger cities fully accounts for their smaller chance to leave the poverty neighbourhood. The same, but to a slightly lesser extent, can be said for the Turks and Moroccans. The implication from this might be that living in a less urbanised area may increase the chances to move out of a poverty neighbourhood. However, as has been before, until now it has proved to be very difficult for immigrants to find affordable housing outside of the (large) cities, for example in suburban areas (see also Bootsma 1998).

            When the other variables are included in the analysis, the difference between Turks and Moroccans on the one hand and the native Dutch on the other hand is still highly significant. This means that even when controlled for city size, age of the head of the household, household structure, household income, education of the head of the household and tenure, Turks and Moroccans are still less capable to move out of poverty neighbourhoods. Although it might be argued that this result could indicate that Turks and Moroccans stay in poverty areas, because they would like to live there, our interpretation is that impediments to move might be important. Probably, these impediments to move can be found in the lack of adequate supply of dwellings. Maybe the system of housing distribution can be important too. It is known that immigrants might suffer from specific inadequacies of different housing distribution models (Van Kempen and Idamir 2002).[4]

 

Table 5:     Logistic regression of residential mobility out of poor neighbourhoods

 

N=8992 Initial . 2LL: 8558.6 Chi square improvement: df : Sign:	(1)     58.8   2   0.000	(2)   856.9 6 0.000	(3)   1951.1 14 0.000
	B	B	B	R (1)
Ethnic group Surinamese Turks/Moroccans (0=native Dutch)	- 0.48 *** - 0.99 ***	0.22 - 0.54 **	- 0.11 - 0.96 ***	0.05
City size 50 - 100.000 inhabitants  100 - 150.000 inhabitants  150 - 250.000 inhabitants    > 250.000 inhabitants    (0= < 50.000 inhabitants)		- 0.60 ***     - 0.91 ***      - 1.10 ***            - 2.06 ***	- 0.68 ***      - 1.08 ***    - 1.32  ***            - 2.20  ***	0.27
Age			- 0.05 ***	- 0.20
Age square (x 10 5 )			4.97 ***	0.07
Household structure   more person household with                         children   more person household without                         children   (0= one person household)			0.61 ***   0.59 ***	0.07
Annual income in guilders (x 10 5 )			0.65 ***	0.08
Level of education   Intermediate   high    (0= low)			0.27 0.68 ***	0.09
Tenure   Owner occupied   (0=rented)			- 0.64 ***	- 0.09
Constant	- 1.42 **	- 0.37 ***	1.02 ***

 

* = p < 0.05; ** = p < 0.01; *** = p < 0.001

^{1)         R is a statistic that is used to look at the partial correlation between the dependent variable and each of the independent variables. R can range in value from . 1 to +1. A positive value indicates that as the variable increases so does the likelihood of the event occurring.}

Source: Housing Demand Survey 1998, CBS/VROM; WMD 1998/RIO/reworked by ABF

 

^{Acculturation variables may be important in explaining the chance to move out of a poverty neighbourhood. This has already been stated in section 2 of this paper, where the spatial assimilation model was discussed briefly. The idea is that the more immi­grants are assimilated to the host society, the better their possibilities (and willing­ness) to move out of concentration areas. In table 6 the results are presented of an analysis in which we include two acculturation variables. First, the duration of stay in the Nether­lands is used. We have made a distinction between (1) first generation immi­grants who migrated to the Netherlands before 1981, (2) first generation immigrants who migrated after 1981 and (3) the second generation (those born in the Netherlands or who came with their parents, but were at the time of migration younger than 13 years). The expectation is that especially the early migrants will have been less able than the second generation to escape from poverty neighbourhoods, while the later first generation migrants may have a kind of intermediate position. This is, because the second generation is expected to be more integrated in Dutch society.}

^{The second indicator of acculturation is a variable that measures the interest in Dutch politics. Respondents could have been (1) not interested in Dutch politics, (2) interested in Dutch politics and (3) actively involved in discussions of Dutch policies. The expectation is that those actively involved are more integrated in Dutch society and thus have more chances to have moved out of a poverty neighbourhood.}

^{In the analysis we have distinguished between two groups: native Dutch on the one hand and Surinamese, Turks and Moroccans on the other hand. Relatively small numbers of immigrants in the database do not give us the possibility to have a multi­variate analysis of the groups separately (table 6).}

 

^{The main conclusions from this table can be summarised as follows:}

·        ^{In the multivariate analysis, no statistical significant relation can be detected between the duration of stay and residential mobility out of poor neighbourhoods. A bivariate analysis (not given here) would have indicated that such a relation would exist: the second generation immigrants is better able to escape from poverty neighbourhoods. In the multivariate analysis, however, the independent effect of generation disappears. The effect is totally explained away by age and education. In other words: the second generation moves more often from poverty neighbourhoods, because people belonging to this category are younger and better educated. Thus, compared to the second generation, the first generation} immigrants are not less able to utilise effectively their socio-economic resources in exchange for desirable neighbourhood environments.

^{·          The other acculturation variable, interest in Dutch politics, also seems not to be statistically significant. Also from a bivariate analysis (not given) it can be concluded that this variable does not have an influence on the chance to move out of a poverty neighbourhood.}

^{·          All in all, the conclusion must be that a significant effect of acculturation could not be detected.}

 

^{From table 6 another interesting conclusion can be drawn. Income and education have a more positive effect on the probability to move from a poor to a non-poor area for ethnic minority groups than for the native Dutch. As far as this aspect is concerned, the place stratification model is not valid for the Dutch situation. Place stratification would imply that socio-economic successful members of an ethnic minority group would hardly more often succeed in leaving the poverty neighbourhood than their socio-economic less successful counterparts, because of the fierce discrimination that affects them. There is no proof for such a situation from our analysis. In the Netherlands, escaping from poverty neighbourhoods is clearly more easy with a combination of a higher income and a better education.}

 

Table 6:          Logistic regression of residential mobility out of poor neighbourhoods,

                        including acculturation variables

 

N= Initial . 2LL: Chi square improvement: df : Sign:	Native Dutch   8041 7912.2   1789.4   12   0.000		Ethnic groups   951 652.3 134.1 17 0.000
	B	R	B	R
Ethnic group Surinamese (0=Turks/Moroccans)			0.85 ***	0.12
City size  50 - 100.000 inhabitants  100 - 150.000 inhabitants  150 - 250.000 inhabitants    > 250.000 inhabitants    (0= < 50.000 inhabitants)	- 0.70 ***      - 1.11 ***   - 1.34 ***            - 2.20 ***	0.28	- 0.41   - 0.53   - 0.95             - 2.19 ***	0.26
Age	- 0.05 ***	- 0.21	- 0.04 *	- 0.06
Age square (x 10 5 )	5.11 ***	0.07	0.17	0.00
Household structure   More person household with                         Children   More person household without                         Children   (0= one person household)	0.56 ***   0.60 ***	0.07	1.28 ***   0.08	0.13
Annual income in guilders (x 10 5 )	0.64 ***	0.08	1.25 *	0.08
Level of education   Intermediate   High    (0= low)	0.25 *** 0.65 ***	0.08	0.56 * 1.21 ***	0.11
Tenure   Owner occupied   (0=rented) Generation   First generation (arrived < 1980)   Second generation   (0 = first generation)	- 0.65 ***	- 0.10	0.04       0.06 0.29	0.00     0.00
Involved in political discussions    Actively    Passively   (0= not involved)			0.02 - 0.18	0.00
Constant	1.17 ***	- 0.37 ***	- 1.08

 

* = p < 0.05; ** = p < 0.01; *** = p < 0.001

Source: Housing Demand Survey 1998, CBS/VROM; WMD 1998/RIO/reworked by ABF

 

6.            Present policies in the Netherlands

 

In this paper we have focused our analysis on poverty neighbourhoods, which have been defined as areas in which at least 30 per cent of the households have an income in the lowest quintile of the income distribution in 1998. Of course, this income variable does not automatically say much about the character of the neighbourhoods. But it is clear from the literature on urban neighbourhoods that many of these poverty areas have to deal with a lot of social, economic and physical problems. In a large number of these neighbourhoods tension arises between the various population groups; there is little evidence of social cohesion, criminality figures are high, the physical environment becomes derelict, there is high unemployment and extensive dealing in and use of drugs (Tesser et al. 1995). To that can be added that the dwellings in these districts are usually small and have poor sound insulation, public space is unattractive or even considered as dangerous by most of the local inhabitants, the nuisance from immediate and other neighbours is substantial, as is the dissatisfaction with the dwelling and the residential environment (Bolt et al. 2002). In other words, these areas are not exactly nice places to live for a large number of households. Avoiding the disadvantages would necessitate the existence of opportunities elsewhere or improving the situation locally. Are these options feasible in the Netherlands?

                    To answer this question, we will briefly pay attention to three important policies: spatial policy, housing policy and urban policy. Although in practice these policies are more or less integrated in the Netherlands, they are treated here separately. Moreover, because of the available space in this article, we will only highlight some aspects of these policies here.

 

Spatial policy

In the beginning of the 1990s, the "Fourth Report on Spatial Planning Extra" was published by the Ministry of Housing, Spatial Planning and the Environment. One of the main items in this report was the assignment of building sites in or adjacent to the cities, in order to enhance proximity and to decrease car traffic between suburban living areas and cities. Housing production was not to be targeted primarily at the low-income groups. One of the main decisions for these new locations was that at least 70 per cent of the housing to be built there will consist of rather expensive dwellings in the rented or owner-occupied sector and only 30 per cent will be social housing, of which, by the way, only a fraction would be accessible for real low-income households. One of the main aims of this policy is to promote filtering: those with above-median incomes who now live in the inexpensive social rented stock are well able to live in more expensive dwellings. By offering them an attractive dwelling elsewhere, they are expected to vacate a unit in the social rented sector and make it available for a low-income household (Van Kempen and Priemus, 1999).

         The effects of this policy are expected not to generate many possibilities for low-income households to move from poverty neighbourhoods. Middle- and higher-income households will be able to move from older areas towards these new areas and this will eventually lead to an inflow of more low-income households in poverty neighbour­hoods. In other words: those who can afford it move to newer dwellings, while those with less money have to rely on vacancies in the existing stock. Relying on filtering does not create real new opportunities for those in disadvantaged situations. Their only alternatives are to stay put or to move to another not so very good neighbourhood.

 

Housing policy

Increasing spatial concentrations of low-income households were the reason for the Dutch government to take action and formulate a new policy of urban renewal, with restructuring the social rented housing stock as its main aim (Van Kempen and Priemus 2002). According to the Netherlands Ministry of Housing, there is only one way of preventing low-rent housing districts from becoming low-income areas in a situation where the occupancy of these dwellings by higher incomes is con­sidered undesirable: by a radical restructuring of the housing stock of these areas. Adding more expensive dwellings for higher-income households seems to be the best measure. That could involve the demolition, upgrading, or selling off of rented dwellings. The restructuring of the housing stock has become a hot topic since 1996 and was discussed in a white paper on urban renewal in 1997 (Ministerie van VROM, 1997).

When districts with a concentration of inexpensive (social) rented dwellings become restructured, an initial effect will indeed be that the area will end up with absolutely and relatively fewer social rented dwellings and at the same time absolutely and relatively more expensive dwellings, first and foremost in the owner-occupied sector. It can then be expected that in the future the restructured districts will house a larger share of households with a middle and probably even a relatively high income.

The restructured dwellings will mostly no longer form part of the social rented dwelling stock, while the dwellings which are not renovated, sold or demolished, will remain part of that segment. When substantial parts of the existing social rented dwellings have disappeared, it is evident that the low-income households will have a more restricted number of dwellings available, even if their own numbers fall. An important effect will be that the social rented sector will increasingly become the segment of the low-income households (Van Kempen and Priemus 2002).

Moreover, displacement effects might occur. Botman and Van Kempen (2001) have recently found some signs of these effects in Rotterdam. Urban restructuring in a poverty neighbourhood in the South of that city has resulted in migration of poor households to other neighbourhoods, because they were not able to find affordable homes in the original neighbourhood.

It might be expected that as a result of this policy the number of movements from or within poverty neighbourhoods will increase for two reasons. First, those with relatively high incomes will find new opportunities elsewhere in the neighbourhood and the city. Second, those with low incomes will be forced to move, because their present dwelling will be demolished or upgraded. Alternatives can be found in the parts of the cities and neighbourhoods that are not yet restructured. More moves from one poverty neighbourhood to another are probably a realistic expectation.

 

Urban policy

The Big Cities Policy focuses on cities. Initially only the four major cities (Amsterdam, Rotterdam, The Hague and Utrecht) were targeted, but over time many medium-sized towns have become part of the focus as well. The Big Cities Policy covers a wide variety of domains, such as local economy, employment, education, safety, care and . quality of living. (including housing) and the physical environment. The population groups who get special attention are . depending on the city . the unemployed, young people, ethnic minorities, (drug) addicts, (ex-) psychiatric patients, homeless people and criminals. Money is targeted to those areas that are characterised by multiple problems, the idea being that concrete area-based measures in the field of the physical environ­ment, the social infrastructure and in the (local) economy will eventually lead to better living conditions for everybody in the city (Doorstartconvenant 1998a; 1998b).

The effects of this policy are not yet clear. Most likely the policy will affect the situation of those who are directly subsidised, like in the case of a newly generated job. Moreover, the projects initialised within the framework of this policy will probably improve the social, economic and physical circumstances in the targeted neighbour­hoods. Of course, it remains to be seen if the inhabitants in these areas will also think themselves that the neighbourhood has improved. Moreover, the principal disadvantage of an area-based policy is that some areas are not within the policy. This might mean that the worst areas will get the necessary attention, but those areas that are only slightly better will not be targeted. In the end this might that money invested in the targeted areas only resulted in the displacement of problems to other, often adjacent, areas.

 

7.            Conclusions

 

From the analysis in this paper it has become clear that for some people it is more difficult to move out of a poverty neighbourhood than for others. Especially households belonging to minority ethnic groups seem to have the biggest problems with escaping from such areas. Partly this difficulty can be explained by their concentration in large cities, where relatively and absolutely many poverty neighbourhoods are located and where better housing alternatives are sometimes difficult to find. Native Dutch seem to have better chances of escaping, because they seem to better able to find suitable dwellings. Their on average higher incomes explain at least part of this ability.

            Support for the spatial assimilation model could not be found in this paper. The ability of ethnic minority households to escape from poverty areas seems not to be related to some important acculturation variables we used in the analysis. Support for the place stratification model is also difficult to find. This model is said to be valid in situations where income and education are not relevant for attaining better positions on the housing market. From our analysis it has become clear that having a better education and (specifically) a higher income does generate better possibilities for ethnic groups.

            Present spatial, housing and urban policies will probably not increase the possibility for low-income households to move from poverty neighbourhoods. Spatial policy is aimed at producing more expensive dwellings, which theoretically might lead to filtering. But the most realistic effect is that dwellings become vacant in unattractive areas. This means that lower income households might have more possibilities to move, but probably there are not much more opportunities to move to much better areas. In addition, urban restructuring will lead to fewer affordable dwellings, which in the end must lead to increasing spatial concentrations of low-income households. Especially these low-income households will be forced to move, because their present dwellings might be demolished or upgraded.

            The Big Cities Policy, finally, will probably result in a general improvement of the poverty neighbourhoods that are targeted in this policy. A better living situation in these specific neighbourhoods will be feasible, which of course means that the inhabitants of this area can become satisfied and are prepared to stay in their neighbourhood. On the other hand, the area-based character of this policy might mean that other poverty neighbourhood will receive less attention, leading to all kinds of distressed situations.

            The production of new inexpensive dwellings in all kinds of urban areas, including the new areas at the margins of the cities, seems to be a necessary addition to the present policies. A smaller focus on area-based urban policies will probably also lead to better opportunities for low-income households to move from poverty neighbourhoods. It is probably not necessary to target housing or spatial policies specifically at the minority ethnic groups. It can only be hoped that, maybe with the help of the Big Cities Policy their labour market situation can be improved. Our expectation is that a better situation on the housing market will follow then more or less automatically (although not immediately).

 

 

 

Literature

 

Alba, R.D. & J.R. Logan (1993), Minority proximity to whites in suburbs: an individual-level analysis of segregation. In: American Journal of Sociology 98 (6), pp. 1388-1427.

Bolt, G. (2001), Wooncarrières van Turken en Marokkanen in ruimtelijk perspectief. Utrecht: Faculty of Geographical Sciences, Utrecht University.

Bolt, G., P. Hooimeijer and R. van Kempen (2002), Ethnic segregation in the Netherlands: new patterns, new policies? In: Tijdschrift voor Economische en Sociale Geografie (forthcoming).

Bolt, G. and R. van Kempen (2000), Concentratie en segregatie in Nederlandse steden. In: R. van Kem­pen, P. Hooi­meijer, G. Bolt, J. Burgers, S. Musterd, W. Ostendorf and E. Snel, Segregatie en con­cen­tratie in Nederlandse steden: mogelijke effecten en mogelijk beleid, pp. 13-34. Assen: Van Gorcum.

Bolt, G. and R. van Kempen (2002), Moving up or moving down? Housing careers of Turks and Moroccans in Utrecht, the Netherlands (forthcoming).

Bootsma, H.G. (1998), The Myth of Reurbanization. Amsterdam: Thesis.

Botman, S. & R. van Kempen (2001), The Spatial Dimensions of Urban Social Exclusion and Inte­gration: The Case of Rotterdam, The Netherlands. Amsterdam: Amsterdam Study Centre for the Metropolitan Environment (forthcoming).

Bowes, A., N. Dar and D. Sim (1997), Tenure preference and housing strategy: an exploration of Pakistani experiences. In: Housing Studies 12, pp. 63-84.

Brevik, I. (2001), Income inequalities and socio-economic segregation in Oslo: governance by the market. In: H.T. Andersen and R. van Kempen (eds.), Governing European Cities: Social Fragmentation, Social Exclusion and Urban Governance, pp. 211-232, Aldershot: Ashgate.

Clark, W.A.V. and F.M. Dieleman (1996), Households and Housing; Choice and Outcomes in the Housing Market. New Brunswick NJ, Center for Urban Policy Research.

Kesteloot, C. and C. Cortie (1998), Housing Turks and Moroccans in Brussels and Amsterdam: the difference between private and public markets. In: Urban  Studies 35 (10), pp. 1835-1853.

Doorstartconvenant Grotestedenbeleid Rijk G4 (1998a). Den Haag.

Doorstartconvenant Grotestedenbeleid Rijk G21 (1998b). Den Haag.

Fong, E. and R. Wilkes (1999), The spatial assimilation model reexamined: an assessment by Canadian data. In: International Migration Review 33 (3), pp. 594-620.

Freeman, L. (2000), Minority housing segregation: a test of three perspectives. In: Journal of Urban Affairs 22 (1), pp. 15-35.

Giffinger, R. & Reeger, U. (1997), Turks in Austria: backgrounds, geographical distribution and housing conditions. In: A.S. Özüekren and R. van Kempen (Eds.) Turks in European Cities: Housing and Urban Segregation, pp. 41-66. Utrecht: European Research Centre on Migration and Ethnic Relations.

Gordon, M.M. (1964), Assimilation in American Life. New York: Oxford University Press.

Gordon, I. and R. Vickerman (1982), Opportunity, preference and constraint: an approach to the analysis of metropolitan migration. In: Urban Studies 19 (3), pp. 247-261.

Hooimeijer, P. and R. van Kempen (2000), De mogelijke doelen van beleid. In: R. van Kempen, P. Hooi­meijer, G. Bolt, J. Burgers, S. Musterd, W. Ostendorf and E. Snel, Segregatie en concentratie in Nederlandse steden: mogelijke effecten en mogelijk beleid. Assen: Van Gorcum.

Kemper, F.-J. (1998), Restructuring of housing and ethnic segregation: recent developments in Berlin. In: Urban Studies 35 (10), pp. 1765-1789.

Ladányi, J. (2001), Residential segregation among social and ethnic groups in Budapest during the post-communist transition. In: P. Marcuse & R. van Kempen (Eds.), Of States and Cities: The Partitioning of Urban Space, pp. 170-182. Oxford: Oxford University Press.

Lipsky, M. (1980), Street-Level Bureaucracy: Dilemmas of the Individual in Public.

Massey, D.S., A.B. Gross and K. Shibuya (1994), Migration, segregation and the geographic concen­tration of poverty. In: American Sociological Review 59, pp. 425-445.

Massey, D.S. and N.A. Denton (1985), Spatial assimilation as a socioeconomic outcome. In: American Sociological Review 50, pp. 94-105.

Ministerie van VROM (1997), Nota Stedelijke Vernieuwing. Den Haag: Ministerie van VROM.

Mulder, C.H. (1993) Migration Dynamics: A Life Course Approach (Amsterdam, Thesis Publishers).

Murdie, R.A. and L.-E. Borgegard (1998), Immigration, spatial segregation and housing segmentation of immigrants in metropolitan Stockholm, 1960-95. In: Urban Studies 35 (10), pp. 1869-1888.

Musterd, S., Ostendorf, W. & Breebaart, M. (1998) Multi-Ethnic Metropolis: Patterns and Policies. Dordrecht: Kluwer Academic Publishers.

Pahl, R. (1975) Whose City? Harmondsworth: Penguin.

Pahl, R. (1977) Managers, technical experts and the state. In: M. Harloe (Ed.) Captive Cities, pp. 49-60 London: John Wiley.

Park. R.E. (1925/1974) The city: suggestions for the investigation of human behavior in the urban en­viron­­ment. In: R.E. Park & E.W. Burgess (eds.) The City, pp. 1-46. Chicago: University of Chicago Press.

Peach, C. (1998), South Asian and Carribean ethnic minority housing choice in Britain. In: Urban Studies 35 (10), pp. 1657-1680.

Peach, C. & Smith, S. (1981) Introduction. In: C. Peach, V. Robinson and S. Smith (Eds.) Ethnic Segregation in Cities, pp. 9-22. London: Croom Helm Ltd.

Phillips, D. and V. Karn (1992), Race and housing in a property owning democracy. In: New Community 18, pp. 355-369.

Rex, J. (1968), The sociology of a zone of transition. In: R.E. Pahl (Ed.) Readings in Urban Sociology, pp. 211-231. London: Pergamom.

Robinson, V. (1981), The development of South Asian settlement in Britain and the myth of return. In: C. Peach, V. Robinson and S. Smith (Eds.) Ethnic Segregation in Cities, pp. 149-169. London: Croom Helm.

Rossi, P.H. (1955), Why Families Move: a Study in the Social Psychology of Urban Residential Mobility. Glencoe: Free Press.

Sarre, P., D. Phillips & R. Skellington (1989), Ethnic Minority Housing: Explanations and Policies. Aldershot: Avebury.

South, S.J. and K.D. Crowder (1997a), Residential mobility between cities and suburbs: race, suburbanization, and back-to-the-city moves. In: Demography 34 (4), pp. 525-538.

South, S.J. and K.D. Crowder (1997b), Escaping distressed neighbourhoods: individual, community, and metropolitan influences. In: American Journal of Sociology 104 (4), pp. 1040-1084.

Tesser, P.T.M., C.S. van Praag, F.A. van  Dugteren, L.J.  Herweijer and H.C. van der Wouden (1995) Rapportage minderheden 1995: concentratrie en segregatie. Rijswijk: Sociaal en Cultureel Plan­bureau.

Van Kempen, R. (1997) Turks in the Netherlands: housing conditions and segregation in a developed welfare state, in: A.S. Özüekren and R. van Kempen (Eds.) Turks in European Cities: Housing and Urban Segregation, pp. 158‑190. Utrecht: European Research Centre on Migration and Ethnic Relations.

Van Kempen, R., P. Hooimeijer, G. Bolt, J. Burgers, S. Musterd, W. Ostendorf and E. Snel (2000) Segregatie en concentratie in Neder­landse steden: mogelijke effecten en mogelijk beleid. Assen: Van Gorcum.

Van Kempen, R. & M. Idamir (2002), Housing allocation and ethnic minority groups: the effects of different housing allocation models on Moroccan households in two Dutch cities (forthcoming article).

Van Kempen, R. & A.S. Özüekren (1998), Ethnic segregation in cities: new forms and explanations in a dynamic world. Urban Studies, pp. 1631-1656.

Van Kempen, R. & H. Priemus (1999), Undivided cities in the Netherlands: present situation and political rhetoric. In: Housing Studies 14, pp. 641-657.

Van Kempen, R. & H. Priemus (2002), Revolution in social housing in the Netherlands: possible effects of new housing policies. In: Urban Studies (forthcoming).

Van Kempen, R. & J. van Weesep, J.  (1997) Segregation, housing and ethnicity in Dutch cities. In: Tijd­schrift voor Economische en Sociale Geografie 88, pp. 188-195.

Wittebrood, K. (2000), Buurten en geweldscriminaliteit: een multilevel-analyse. Mens & Maatschappij 75 (2), pp. 92-109.

 

Metro.doc, 20 November 2001