New peripheries in the making? Analysing German cities from the lens of a suburbanisation of poverty

In this paper we discuss evidence on new socio-spatial patterns in the relation between urban and suburban areas due to the involuntary outward residential mobility of low-income households from central urban areas. Our findings build upon research in three German cities from a comparative perspective, with interviews with local experts providing the main source of information. For analysing decentralisation or suburbanisation of poverty processes we focus on the role of the urban housing market, residential settlement changes, and the role of social security systems. We cannot confirm the displacement of low-income households beyond city limits as a general trend. Instead, our findings reveal a local and context-dependent variation of socio-spatial shifts in the three investigated cities, related to city-specific housing market and urban development factors. As a general factor, existing social security systems continue to work towards keeping low-income households inside cities, seemingly preventing a large-scale suburbanisation of poverty. There is, however, a worrying trend in the concentration of low-income households in ever-decreasing segments of affordable housing, alongside increasing burdens on low-income households coping with the current challenges of a housing affordability crisis. City-regional monitoring and further research are needed.


Introduction
In Anglophone countries in particular, the increase in low-income households in suburban areas has been discussed for years (Allard & Paisner, 2016;Bailey & Minton, 2018;Cooke & Denton, 2015;Covington, 2015;Kavanagh et al., 2016;Kneebone & Nadeau, 2015).Suburban communities, traditionally composed of the middle class, are becoming increasingly diverse in terms of class (Allard, 2017, p. 15).This can be due either to opportuneness among low-income households, thus a voluntary decision (for example for a greener environment, or for better schools), or it can reflect the fact that lower-income households are being pushed out of the more central urban neighbourhoods, or have limited or no access to more centrally located housing, due to direct, indirect or exclusionary displacement (Marcuse, 1985).
This article sets out to explore urban inequality shifts in German cities from the analytical lens of the suburbanisation of poverty debate.The research focuses on 1) investigating whether a decentralisation or suburbanisation of poverty can be identified as an emerging, general trend across German cities, and on 2) understanding the factors promoting or limiting this process generally, and in different local contexts (between cities).Our focus is on the factors influencing the ability of low-income households1 to set up, retain and maintain adequate housing in the city in the context of an urban housing affordability crisis.We make use of qualitative expert interviews in the three chosen cities and interpretive analysis to answer the research questions.The focus and added value of this article is to analyse how local housing experts, social and urban planners in cities understand and interpret current patterns of (de)concentration of low-income households in the three study regions.It is important to note that this research makes no attempt to explain the individual residential mobility decisions of low-income households, as data protection restrictions hinder any analysis of residential moves at the level of individual households.For our study, we thus operationalise our understanding of a suburbanisation of poverty in a simplified way, asking whether suburban poverty concentrations are growing due to increasingly limited possibilities for low-income households to find adequate housing in more central urban neighbourhoods.This includes examining evidence of low-income groups moving away from such neighbourhoods to more peripheral suburban neighbourhoods.
While there is some evidence of displacement of low-income households beyond the city limits for individual, highly dynamic metropolitan areas such as Berlin or Munich, no insights into the dynamics found in smaller, 'more ordinary' German cities are as yet available.This led the authors to research the trends and processes in the three German cityregions of Aachen, Karlsruhe and Leipzig from a comparative perspective.The question whether there is evidence of a suburbanisation of poverty in Germany-and thus a need for greater policy attention-motivated this study.Our study aims to add evidence to the international debate on the suburbanisation of poverty from a German perspective, asking whether we can view this process as a universal phenomenon, or instead a city-specific and contingent process (Cooke & Denton, 2015, p. 300;Maloutas, 2018).

The suburbanisation of poverty as a new urban development trend
The suburbanisation of poverty can now be considered a general phenomenon in US cities (Allard, 2017;Cooke & Denton, 2015;Covington, 2015;Kneebone & Holmes, 2015;Kneebone & Nadeau, 2015), with poverty on the rise in suburban locations in almost all metropolitan regions (Cooke & Denton, 2015).While the percentage of the population that is poor remains higher in urban parts of metropolitan areas (Allard & Paisner, 2016, p. 1), increased poverty in the suburbs nevertheless marks a break in urban development trends.
Studies from the UK (Bailey & Minton, 2018;Kavanagh et al., 2016) see similar trends, although their dynamics and dimensions are not comparable to US metropolitan regions.In their analysis of the distribution patterns of poverty, Bailey and Minton (2018) map decentralisation processes in the 25 largest city-regions in England and Scotland over the period between 2004 and 2015/16, using aggregated small-scale data from administrative sources on area deprivation (Indices of Multiple Deprivation).These decentralisation processes are most evident in the London metropolitan region (Bailey & Minton, 2018, p. 910).Turning to the Netherlands, studies of residential moves in Amsterdam, Rotterdam and the Utrecht city-region, using household-level income data, confirm that poverty is being decentralised in these cities (Hochstenbach & Musterd, 2018, 2021).
Involving different national contexts, these findings indicate a break in urban development trends with a wider socio-political and structural relevance (Allard & Paisner, 2016, p. 2;Kavanagh et al., 2016Kavanagh et al., , p. 1287)).They raise questions about changing opportunity structures for low-income or poor households in suburban regions and a potential "spatial mismatch" between the location of low-income households and the location of jobs, relevant services and infrastructure necessary to get ahead (Allard & Paisner, 2016;Kavanagh et al., 2016Kavanagh et al., , p. 1287;;Kneebone & Holmes, 2015;Kneebone & Nadeau, 2015).Living in a peripheral location can affect households' access to public transport, social or health services, or job opportunities.This is particularly true for low-income households restricted in their mobility (Rozynek et al., 2022).
Gained mostly in the context of studies on urban gentrification and displacement, evidence of a suburbanisation of poverty in German cities is rather fragmented.In their study of displacement in selected inner-city districts in Berlin, Beran and Nuissl however note how "peripherally located neighbourhoods (.) have disproportionately high numbers of displaced tenants moving in" (2022, p. 103).In her qualitative study of low-income households' residential location decisions in the Munich region, Sterzer (2019) reports on several households ending up, involuntarily, in areas outside city limits, after being forced to look for new rental accommodation.There is thus some evidence on low-income households being squeezed out of central locations in German city-regions (see also Schipper, 2022).Considering deregulation and marketisation reforms, and the sharp rise in rents and real estate prices over the past years in German cities (Holm et al., 2021), it seems safe to believe what is implied by this evidence and to expect a new quality of displacement of low-income households in ever-tighter urban housing markets.
Drawing on the emerging body of research on the suburbanisation of poverty in critical geography and housing research, specifically in the Anglo-American context, our research focuses on three different, albeit interlinked perspectives: the role of the urban housing market (categories of housing providers, the owner-tenant mix, housing typologies, rent levels, etc.), changing patterns of residential settlement and mobility (such as reurbanisation or gentrification processes), and the role of social security and welfare systems, from the national to the local level (including social transfers, social policies, etc.).While the three perspectives are interlinked and relate to wider urban restructuring processes such as urban labour market polarisation, economic segregation, or the commodification of housing in recent decades (Bailey & Minton, 2018), they are separated here for analytical reasons.Each of the three perspectives allows us to identify causal or filtering factors which play a role in explaining trends and processes in the decentralisation or suburbanisation of poverty (Allard & Paisner, 2016).
The promotion of homeownership in cities with a view to attracting and retaining middle or higher-income groups, urban renewal and social mixing projects as well as the conversion of social housing to free market housing have all reduced affordable rental housing in cities (Hochstenbach & Musterd, 2018, pp. 33-34;Bailey & Minton, 2018;Kadarik & Kährik, 2022).Housing has become a tradeable commodity in cities across the globe, raising concerns about housing becoming detached from its local context and its social function as a human right (Pattillo, 2013).The commodification of housing and (neoliberal) market-oriented housing policies are important factors for understanding (decreasing) housing opportunities for low-income households (Kadarik & Kährik, 2022).For a long time, the German housing market was considered, due to its long-term stability, as a "counter-example of the more liberal and financialized US housing market" (Wijburg & Aalbers, 2017, p. 968).However, as Wijborg and Aalbers (2017) show, the German housing market has experienced waves of financialisation in the past decades, including large-scale privatisations in the social and rental housing sector, and the increased entry of financial investors.The past few years have seen a rapid increase in housing prices and rent levels in German city-regions (Glatter & Mießner, 2022;Holm et al., 2021).
Understanding residential settlement in the city-region and mobility patterns between urban and suburban areas constitutes a second important analytical perspective.A key element here is the renewed interest of middle-or higher income groups in staying in or moving back to the city.While suburbanisation of these groups is still a dominant trend, specifically in the transition to parenthood, staying in the city and re-urbanisation are current countertrends, albeit differing in their extent and relevance between cities (Booi & Boterman, 2020).Alongside changing lifestyles, the wide and varied range of urban amenities and the socio-cultural infrastructure are important pull factors explaining the renaissance of inner-city living (Hochstenbach & Musterd, 2021;Glatter & Mießner, 2022).These changes in residential settlement patterns have been boosted by urban policymaking in many countries, including Germany, in favour of strong investments to strengthen urban labour and business markets and cater for highly qualified professionals (Bailey & Minton, 2018;Wijburg & Aalbers, 2017).The preference for urban living of middle-class fractions, as well as the (partly state-led) upgrading of inner-city neighbourhoods, resulting in changing neighbourhood composition patterns, thus contribute to explaining the displacement of lower-income residents from central urban areas (Hochstenbach & Musterd, 2021).Yet displacement is not always linked to direct residential dislocation.Inner-city neighbourhoods may become inaccessible to low-income households, with the latter no longer able to find adequate accommodation there (exclusionary displacement).In addition, physical and socio-cultural changes in the neighbourhood may generate displacement pressure more indirectly (Marcuse, 1985: 205).Rent levels tend to be lower in suburban commuter sheds (beyond the suburbs of wealthier population groups) than in more central urban areas in general (ILS, 2022).This difference in rent levels and access to lower-priced housing on the outskirts of a city might foster the decentralisation of low-income households.
The level and extent of urban segregation and social exclusion patterns are mediated by social security systems in the wider context of welfare state arrangements (Hochstenbach & Musterd, 2018).In this respect, Germany is characterised by a subsidiarity system which gives local authorities a certain amount of leeway and autonomy in the German federal system of government, allowing them to play an active role in local housing, social and economic policy (Glatter & Mießner, 2022).In the context of our research, housing and social benefits play a key role.Housing benefit covers part of the rent/the total housing costs and is meant to ensure that low-income households can afford adequate housing.The full cost of housing is covered by the municipality when households receive longterm unemployment benefit under national Social Security Code II regulations (SGB II) (Federal Ministry of Labour & Social Affairs, 2020, p. 45).Rent increases are thus internalised in local authority budgets, or (in the case of housing benefits) partly refunded by the Federal Government.Due to tighter municipal budgets in recent years, the financial capacity and steering power of municipalities have however become limited.Moreover, national and local institutional environments and policy arrangements are shifting towards putting more emphasis on individual responsibilities in a situation featuring social exclusion, commodification of housing, and state austerity policies guided by neoliberal thinking, in Germany as elsewhere (Bailey & Minton, 2018;Wijburg & Aalbers, 2017;Kadarik & Kährik, 2022).Expecting consequences from these shifts for low-income residents and their access to affordable housing in central urban areas, we conducted an in-depth investigation in three structurally different cities.

Methodology and data
The research presented here is part of a wider study including quantitative and qualitative analyses.Data availability in Germany poses limitations to quantitative research approaches across city-regions.The lack of individual-level census data or aggregated small-scale data for comparative or longitudinal analyses across city-regions in Germany thus led to a mixed-method study for exploring the suburbanisation of poverty.In a first step, a quantitative analysis was carried out for the 33 largest German city-regions, where we used 'purchasing power of households' as a suitable proxy to display relative poverty (and relative wealth) on a city-regional scale.City-regions are defined as a core city with more than 200,000 inhabitants and 100,000 employees, including the commuter sheds beyond city limits (ILS, 2022).Quantitative analysis provided key information on the cityregions (such as the share of the population receiving social welfare benefits, population development, purchasing power and rent levels), and led to maps being drawn up indicating spatial concentrations of low-income households in German city-regions.Based on this data we selected three city-regions for in-depth analysis.Cities located within polycentric areas were excluded, as their surrounding areas overlap, generating fuzzy correlations between these areas.We also excluded cities with over one million inhabitants, deliberately focusing on smaller German city-regions to investigate to what extent the suburbanisation of poverty can be identified as a general trend across German city-regions (beyond evidence from high-priced large city-regions such as Berlin, Hamburg, Frankfurt or Munich, see Sterzer, 2019;Beran & Nuissl, 2022;Schipper, 2022).These two exclusionary factors reduced the list from 33 to 17 city-regions.We then sorted the city-regions into three groups, from which we selected one case each.East German city-regions made up one group, as there are structural differences in the German housing market between East German and West German city-regions (see Holm et al., 2021, p. 45).As for the two West German city-regions, we differentiated between those with a high share of persons receiving welfare benefits (above 10 percent for the core cities) and those with a lower share.To select the final candidates from each group, we analysed the data on low-income concentrations in the city-regions (using purchasing power as our proxy).We chose those three cases where our data pointed to an increased deviation of the purchasing power of commuter sheds from the city-regional average between 2009 and 2017, indicating increasing inequalities within the city-regions.Chosen as representatives for West German cities, Aachen and Karlsruhe differ in terms of socio-economic characteristics.For example, the share of households receiving social welfare benefits is twice as high in Aachen as in Karlsruhe.The East German city of Leipzig has experienced a significant increase in population in the past as well as a strong decline in the share of residents receiving social welfare benefits, despite remaining well above the national average.All three cities are university cities.
Following quantitative analysis and case study selection, qualitative case study research started with a document analysis and literature review for the respective three city-regions, covering the topics of housing, urban development and segregation, urban and housing policies.Interviews with local experts in the cities took place in September and October 2021.A research team of three of the authors conducted all expert interviews, combined with on-site visits to the mapped low-income neighbourhoods in the city-region (the outcome of quantitative analysis) and their photographic documentation.The research team conducted a total of 12 interviews with 24 local experts (see Table 1), the majority of which were conducted in-presence (only two were conducted digitally).2Interview partners were carefully selected to ensure that the topics housing, social planning and urban planning were covered by well-informed experts and discussed to a similar degree in all three cities in order to ensure comparability.The expert interviews lasted on average one hour, ranging from a minimum of 50 up to 85 min.All experts were provided beforehand with our maps displaying identified poverty concentrations within the city-region, and were thus familiar with the findings of the quantitative analysis.The interviews then focused on the experts' interpretation and explanation of the mapped patterns.The interviews were recorded, fully transcribed and the whole corpus systematically coded using Maxqda, a software for qualitative analysis (Mayring, 2014).For the sake of this paper, we use the codings related to housing market, segregation, suburbanisation, residential mobility and settlement dynamics, gentrification and suburbanisation of poverty.

Characteristics of the housing market in the three cities
We rely mainly on collective understandings, perceptions and interpretations of the housing-market-related characteristics of the three cities according to interviewed experts, alongside statistical data.Grasping these city-specific characteristics allows us to understand what limits or promotes the decentralisation or suburbanisation of poverty, as discussed in the following section.
The three housing markets differ in terms of inner-city development potential, perceived housing pressure, and housing market structure (owner-tenant mix, co-operative or municipal housing segments, building typologies).According to local experts, further development potential in Aachen and Karlsruhe was limited by a lack of building land and by geographical and administrative boundaries (Interviews A1, A3, K4).By contrast, Leipzig continued to have sufficient development potential including vacant buildings and derelict land, despite an increase in population of nearly 100,000 between 2010 and 2020 (Interview L2).Rental prices have increased significantly in all three cities in the past averaging + 40/41% (Karlsruhe, Aachen) and + 52% (Leipzig) between 2009 and 2018, measured in terms of quoted rents in Euro/sqm per month.In the cities of Aachen and Karlsruhe, quoted rents have risen on average just as much as in the surrounding commuter sheds, whereas in Leipzig those in the city have climbed 52%, compared to 43% in the surrounding region.Consequently, the housing markets of Aachen and Karlsruhe were considered overheated by the interviewed experts, while that of Leipzig was considered to be moving in this direction.Information on the rent burden ratio for the three cities supports the statements of the interview partners (Holm et al., 2021, p. 178-179; see also Table 2).One factor contributing to lower housing pressure in Leipzig is the fact that around 24% of its housing stock is owned by co-operatives and the municipal housing corporation LWB (Jacobs & Diez, 2021, p. 10).This is comparable to Karlsruhe, where 25% of the rental housing market is owned by the municipal housing association Volkswohnung (the city's largest landlord) and further co-operatives-a comparatively high rate in the German context (Interview K1).This situation contrasts with that in Aachen with its high share of private owners, a small stock of co-operative housing (Duikers et al., 2020, p. 94) and a housing association in which the municipality has a 61% stake (gewoge, 2022).
As discussed above and confirmed by the experts, wide-scale privatisations of the social and rental housing sector have occurred over the past decades in Germany, leading to a decrease in social housing in all three cities (Duikers et al., 2020, p. 94;Interviews K2, L2).This raises the question as to how those housing market segments accessible to lowincome households have developed in recent years.
According to analyses conducted by the Aachen city administration (Interview A1), about 65% of all quoted rents in 2010 were classified in the affordable rent segment of 'below 7 euros/sqm per month', for instance in former regeneration areas.This market segment has since disappeared completely.Furthermore, the share of social housing dropped to 7.1% of the total stock in 2019.While social housing is virtually the only segment offering affordable housing in Aachen, the stock is already insufficient to provide housing for those entitled to it (Duikers et al., 2020, p. 57).The Aachen experts further confirmed that the number of social housing units was expected to decrease even further in the years to come and that residential areas previously providing 'islands of affordable housing' for low-income households, such as former urban regeneration areas, had experienced the sharpest rent increases in recent years (Interviews A1, A-2).
Affordable housing was still provided in Karlsruhe and Leipzig through co-operative and municipal housing associations.These offered apartments at below-average rents and tried to keep rent increases at very moderate levels (Interviews K1, K3, L2).Furthermore, according to the housing market experts, more than 95% of the rental apartments offered by Volkswohnung were in a rent segment where rents were paid by the city for households receiving social benefits.Though not social housing stricto sensu, the housing association usually maintains the social housing rent level after expiry of the lock-in period for the building subsidy. 3Since 2014, Karlsruhe has also managed to restore its social housing stock (KaWoF-Programme), after years of decline.However, the waiting list for Volkswohnung apartments has been growing over the past few years, indicating an increasing gap between demand and supply (Interview K3).In Leipzig on the other hand, there is a housing supply surplus, meaning that affordable housing continues to be available for low-income households.However, it is mostly located in city districts characterised by large housing estates built prior to German reunification (the so-called Plattenbausiedlungen) or in areas characterised by old buildings in need of refurbishment (Interview L2).
While housing market policies in Aachen and Karlsruhe have consistently fostered the construction of social housing in response to this development, this has only been the case in Leipzig since 2017(Stadt Leipzig, 2022), as the city had faced an oversupply of housing and increased vacancy rates in the years before.A range of policies, adapted to local housing market situations, are applied by the cities to develop further affordable housing.Such measures include, for example, the quota arrangement applied in Aachen.This requires that 40% of new construction projects linked to a development plan must be subsidised social housing (Interview A1).In Leipzig, social conservation areas serve to protect longestablished population groups from high rent increases by imposing restrictions on owners and landlords (Interview L2).Among the three cities, Karlsruhe has taken the most active measures to increase the social housing stock in the last years (KaWoF-Programme) (Stadt Karlsruhe, 2021).Similarly, densification and upsizing of the existing affordable housing stock serves as an effective measure, as implemented by Karlsruhe's municipal housing association which has also undertaken to achieve a 60% quota for subsidised housing in its new construction projects (Interview K3).

Is poverty being suburbanised?
In response to our key interest, namely to interpret (de)concentration patterns in the cityregion (including a potential suburbanisation of poverty), and to discuss the underlying factors, several distinct discourses and explanations came up in the interviews.These were related to residential settlement patterns, namely suburbanisation (who moves out?), as well as gentrification and segregation.We analyse the findings city by city, ending the section with a comparative perspective across the city-specific findings.

Aachen: displacement of low-income populations beyond city limits as a likely future scenario
In Aachen, the suburbanisation has been seen as an ongoing process for the last 20 to 30 years.In the perception of experts, residential suburbanisation remained a middle-/ higher-income population phenomenon (Interviews A1, A2), with middle-class families moving out as the costs of renting or buying properties in the surrounding commuter sheds were comparatively lower than in Aachen.Such moves gave families more floorspace for their money, or allowed them to buy a house, something no longer "feasible in Aachen, either in terms of space or in terms of costs", even for middle-class families (Interview A1).The housing market in Aachen no longer allowed even middle-income families, let alone low-income families, adequate access to housing, as one expert explained: "But I think that in Aachen […] it is now not only leading to deconcentration just among poorer households, but overall.It is reaching, or has long since reached, the middle classes.[…].The markets are overheated.And I believe that in addition to those who seek out the islands of affordability in the region, those who want to buy something and who realise that it simply doesn't work in Aachen, are also being pulled out."(Interview A2).
In response to the question of whether poverty is being suburbanised due to increasingly limited opportunities for low-income households to access housing within city limits, experts saw this is as a likely process.But they also mentioned the lack of data on residential relocations across municipal boundaries in general, and more specifically when it came to low-income populations.A recent Aachen housing market report does however see a population displacement to lower-priced markets such as the municipalities in the surrounding area as a likely scenario (Duikers et al., 2020, pp. 56-60).While there is no official data to prove residential relocation or displacement of low-income households away from Aachen, the interviewed experts assumed relocation processes of low-income households to surrounding municipalities.
"In response to the basic question of whether we believe it is possible that segregation processes will take place in the surrounding areas, I would say yes.[…] Given this development of the housing market in the last ten years, which has brought significant price increases with it, especially when a household is forced to move […], many people on low incomes are actually forced out into the city-region […].We hardly have any apartments in Aachen, especially if there are more than two people in the household.At least not in those privately financed housing market segments, which come into question for households whose rents […] get paid by the [municipality].That's another point that perhaps plays a role."(Interview A1) The experts referred here to the situation of households receiving social benefits (see Sect. 2).Local authorities cover 'reasonable' accommodation and heating costs for these households (Federal Ministry of Labour & Social Affairs, 2020) up to a set maximum reflecting household size.However, as social and other affordable housing segments are decreasing in absolute numbers in Aachen, low-income households face increasing competition for such apartments in the city.Faced with no alternative, households tend to stay put in too small apartments (Interview A1).
Explaining decreasing affordable housing segments, experts in Aachen also referred to the modernisation of housing stock in neighbourhood regeneration areas and gentrification.While such areas still offer affordable housing, urban redevelopment measures have led to their upgrading-especially those with an attractive Wilhelminian/turn-of-the-century building structure -, making them interesting for higher income groups.This confirms the role of urban gentrification for understanding decreasing housing opportunities for lowincome households (Hochstenbach & Musterd, 2021).

Karlsruhe: Though the gap between supply and demand is increasing, there are still inner-city options
Similar to Aachen, experts in Karlsruhe saw a 'classical' suburbanisation process as the dominant process, i.e., middle-/higher-income populations moving out of the city into the surrounding region during the family formation phase, thus confirming the relevance of demographic factors (Booi & Boterman, 2020).Experts also saw some signs of reurbanisation-albeit a minor undercurrent -, with mostly older people moving back to the city.

"[I]t is mainly families […] who move out of Karlsruhe. And what we can certainly observe […]
is that suburbanisation is also taking place in the wider region.However, according to the interpretation of the data we have, it is not really a case of suburbanisation due to poverty, but rather that many families simply have motives for buying, which then bring them to the wider region.
There are just no more opportunities here [in the city]."(Interview K1) Later on in the interview, however, the expert pointed to a group of households they considered likely to move out of the city.These were households who-based on their income-were not eligible to claim social benefits.
"The only ones-and this is difficult to analyse-the only ones I could imagine are families whose incomes are just above the eligibility threshold for social benefits, i.e., the near-poverty stratum.For them, of course, it is very interesting: should I look for an expensive rented flat in Karlsruhe, or should I rather look in the surrounding area?Rents are just as high in the immediate suburbs, but they are somewhat lower in the wider [surrounding] area.And the supply is simply better.There are better-equipped flats for rent in the surrounding areas.[…] it's more the desire for a [better-equipped] flat that is still affordable."(Interview K1) The overall interpretation of experts on the question as to whether there was a potential suburbanisation of poverty was that low-income households still found alternative options within the city and were not being pushed to move away in search for affordable housing.
As a factor explaining the lower pressure-compared to Aachen-on low-income households to move outside city limits, experts repeatedly referred to the low segregation index in Karlsruhe (one of the lowest among German cities, see Helbig & Jähnen, 2018, pp. 139-140) (Interviews K1, K2, K3), and the fact that affordable housing segments were still available throughout the city (and not concentrated in specific neighbourhoods).They referred to the small-scale mixing of owner-occupied private housing, rental market housing and municipal or co-operative housing in all city districts as a factor allowing low-income households to find affordable alternatives close by, should they be forced to move out of their apartments.This, in turn, was seen as an outcome of Karlsruhe's active social policies of the last two decades."[…] I believe we are investing quite well in measures to combat poverty and develop social neighbourhoods.That's what I wanted to say earlier: the fact that Karlsruhe dismantled marginalised settlements so early on, was active early on in helping the homeless, and also spent money on the social labour market, these are all measures that had a positive effect later on, meaning that we don't have segregation on a massive scale."(Interview K2) However, for low-income newcomers to the city, or for those forced to move, the options to find affordable housing are becoming increasingly limited.The waiting list for an apartment with the largest landlord in the city is growing, and when it comes to social housing, it is a question of managing the shortage, as demand outstrips supply (Interviews K2, K3).

Leipzig: increased urban segregation instead of suburbanisation of poverty
Leipzig's population has grown enormously in the last few years (Interview L2).While the housing vacancy rate was 20% 20 years ago, it is now 3% (Interview L4).For a long time, the housing market was slack, and even today, experts saw sufficient development potential for new housing, even in inner urban areas.As is the case with other East German cities, the 'classical' family suburbanisation process started later in Leipzig than in Aachen and Karlsruhe.In contrast to the latter two cities, the patterns mapped from quantitative analysis for Leipzig showed larger concentrations of low-income households within the city.Experts referred to a segregated structure, with lower-income households living in specific neighbourhoods or housing estates.
"If you look at the city, there is a very clear distribution here: the lower-income households live in the areas close to the city centre and the higher-income households live in the outer areas, many of which were incorporated in the 90s, but which are still village-like.Where [new] owner-occupied houses predominate.Lowerincome households tend to be found living in the Wilhelminian inner city, or in the concrete high-rise Plattenbaugebiete." (Interview L1) There was a collective understanding among the interviewed experts that there was no suburbanisation of poverty.Boasting a sufficient reservoir of affordable housing for lowincome households within the city, most relocations constituted "internal migration of an inner-city nature."(Interview L2).
"Do we have any suburbanisation of poverty?No. What we have is segregation.[…] But we're not observing any suburbanisation, i.e., poor people moving out of the city to get cheap housing in a medium-sized city [in the surrounding region]."(Interview L2) However, experts referred to increasing inner-urban segregation and the concentration of low-income households in neighbourhoods with specific housing characteristics, specifically the larger housing estates.As rents in these housing estates (the Plattenbausiedlungen) are "about 5 euros/sqm per month" and among the lowest rents, people on low incomes, and especially if they are eligible for social benefits, move to these estates, even if they are not a very popular housing choice, due to the stigma connected to these areas (Interview L4) (Kadarik & Kährik, 2022).According to the calculations of the local administration, socio-economic indicators point to residential segregation increasing slightly within the city (Interview L1).Gentrification plays a role too, specifically in one inner-city neighbourhood (Leipziger Osten) where many low-income households continue to live.When apartment blocks are modernised, they immediately end up in a high-priced segment and thus do not match the demand of low-income households (Interviews L1, L3).This neighbourhood has, however, now been declared as a social conservation area, meaning that the local authority has taken measures to limit rent increases related to modernisation and upgrading measures.

Findings and discussion from a comparative perspective
In the perception of the housing market, urban and social planning experts from all three cities, residential suburbanisation processes remained above all a middle-and higher-income group phenomenon, linked to the family formation phase (Interviews A1-2, K1-2, L1-3) (Booi & Boterman, 2020).They further mentioned that municipalities in the surrounding region had made more building land available in the past years, targeting middle-class households and home ownership (Interviews K4, L1).No official data is available in any of the three cities on inter-municipality household relocations, either by class or income.While housing markets are thus becoming increasingly regionalised, only very few city-regions are beginning to build up data for monitoring regional housing markets (including the Aachen city-region).While there is thus no official data, according to the interviewed experts there is currently no quantifiable population displacement of lowincome households beyond city limits.Existing concentrations of low-income households in municipalities in the surrounding region were explained as being locally contingent, persistent and path-dependent (for instance, due to de-industrialisation), rather than newly emerging or growing due to an influx of low-income households moving out of a region's core city.However, experts in Aachen saw a suburbanisation of poverty, due to the involuntary outmigration of poor and low-income households to outside city limits, as a likely scenario for the future (Interview A1).Aachen presents the strongest evidence in this respect.In Leipzig, no such suburbanisation is likely to occur in the medium term, due to the housing supply surplus and large stocks of accessible housing for low-income households in housing estates (Interviews L1-4).Suburbanisation of poverty in German cities might thus indeed "still appear to be only a minor undercurrent when looking at the total population" (Hochstenbach & Musterd, 2018, p. 27).It should be taken into account, however, that our study did not investigate larger metropolitan cities, meaning that our findings may be related to the type of city chosen.For instance, Bailey and Minton (2018) find the strongest evidence for a decentralisation of poverty in larger British cities, and specifically in London.The same applies to the Dutch context, where spatial trends are more pronounced in the larger metropolitan region of Amsterdam (Hochstenbach & Musterd, 2021).This calls for further research in Germany, including more city-regions.A larger sample of cities might also bring insights into whether the case of Leipzig is representative of a wider group of cities.
Our findings do not imply, however, that there are no worrying urban inequality trends and developments.In explaining why the suburbanisation of poverty is less a phenomenon in our studied cities, experts referred to a set of underlying factors.First, in the current situation, low-income households were employing coping strategies.Faced with no suitable offer of affordable alternatives, households were having to downscale their living conditions in terms of floorspace and costs.They were staying put, accepting living in cramped or substandard conditions or cutting back on non-housing expenses in order to avoid having to move house: "Under these conditions of sharply rising prices, people are staying put.There is total persistence, as any move would lead to deterioration" (Interview L4, see also A2).These findings are in line with previous research on low-income households' residential location decisions (Beran & Nuissl, 2022;Sterzer, 2019).Second, and linked to this, experts noted that, although moving to the surrounding area might mean lower housing costs for low-income households, commuting costs could increase, leading to little improvement in living conditions (Interviews A4, L1) (see also Rozynek et al., 2022).The associated costs linked to moving outside city limits (in terms of access to transport, employment, social services) might outweigh the potential benefits (Interview L4) (Sterzer, 2019).When forced to look for new rental accommodation, these households would thus rather look for inner-city alternatives.However, in more segregated cities like Leipzig, households would be forced to move to stigmatised and deprived neighbourhoods where affordable housing is still available, including former or current urban renewal areas, or large housing developments, a process leading to increased urban segregation, and the concentration of low-income households in large housing estates (Kadarik & Kährik, 2022, p. 699).
One important factor explaining the findings on the suburbanisation of poverty in German cities relates to the German social security and welfare system, where the municipality covers the costs for appropriate housing for households eligible for social benefits.The experts from Karlsruhe also confirmed that the rental costs of households receiving social welfare benefits that exceeded the subsidized price segment remained covered by the city when there were no alternatives: "There are no social housing units left.The city of Karlsruhe allows tenants to stay in their apartments and keeps on paying the higher rent" (Interview K1).As numbers of affordable housing units (social housing or lower-rent segments provided by housing co-operatives or municipal housing associations) are decreasing in the cities, local authorities are thus internalising the costs of rising rents and the housing affordability crisis.

Conclusions
The aim of the study was to understand recent poverty (de)concentration in German cityregions, responding to worrying accounts of an increasing proportion of poor households moving beyond city limits in an Anglo-American context (Allard & Paisner, 2016;Bailey & Minton, 2018;Covington, 2015;Kavanagh et al., 2016;Kneebone & Nadeau, 2015).Our findings contribute to the international debate on the suburbanisation of poverty in revealing the local and context-dependent variation of socio-spatial shifts in cities (Cooke & Denton, 2015;Maloutas, 2018).Among our case study cities, we find those in which the suburbanisation of poverty will in all probability not occur at all in the medium term (Leipzig), while in others it is seen as a likely scenario for the future (specifically for Aachen).
To analyse suburbanisation of poverty processes we focused on three interlinked perspectives: the role of the urban housing market, residential settlement changes, including gentrification, and the role of social security systems as part of the welfare regime.Experts in all three cities confirmed decreasing housing options for low-income households.While supra-local, national deregulations of the housing market play a role here (Wijburg & Aalbers, 2017), the comparative analysis shows the relevance of local policy choices and how they influence today's availability of accessible housing in the three cities.Thus, the available stock of accessible housing segments for low-income households varies substantially between Aachen and Karlsruhe.Furthermore, city-specific urban restructuring and historical transformation processes, as is the case in Leipzig, play an essential role.Gentrification was mentioned as an explanatory factor in the debate on new patterns of (de)concentration in Aachen and Leipzig, but not in Karlsruhe.Hence, specific local features outweigh generalisations across the investigated cities, confirming the call of Cooke and Denton (2015) for attention to locally contingent processes in the suburbanisation of poverty.The study confirms the role of welfare state transfers targeting households as a factor mediating and limiting the displacement of low-income households (see also Hochstenbach & Musterd, 2018;Bailey & Minton, 2018;Kadarik & Kährik, 2022).Those households receiving social benefits have their rent expenses covered by the municipality and are thus not automatically driven out of their homes by rising rents.The findings thus show the relevance of national welfare regulations in explaining our findings on the suburbanisation of poverty compared to those in more liberal welfare regimes, like the US.While Germany's welfare regime has thus obviously worked well in the past decades to prevent a suburbanisation of poverty and individual hardship, maintaining the system is becoming increasingly challenging in a context of ever tighter urban housing markets and municipal budgets, and at the same time, more important than ever.
However, we also see new shades and layers of urban inequality.In the current situation, with social and affordable housing units decreasing in absolute numbers in the study regions, low-income households are set to face increasing competition for affordable housing.Having to adjust accordingly, they resort to coping strategies to avoid relocation.Depending upon the locally specific spatial layout of accessible housing niches, there are signs of increased urban segregation, with households being forced to move to the last remaining pockets of accessible housing in the least attractive, stigmatised parts of a city.The response to the question as to whether there are new peripheries in the making is thus equivocal.They are not necessarily located beyond city limits, nor are all of them new.The costs of the housing affordability crisis in the three cities remain hidden or are internalised to some degree, as low-income households employ coping strategies, and the social security system still offers protection, with local authorities' covering rents for those eligible for social benefits.This is not the case for other low-income groups, i.e., those unable or unwilling to claim social welfare benefits and who are thus less protected against increasing rent levels: those with precarious work, self-employed workers on low incomes, or the working poor not entitled to state benefits.
Even if we cannot confirm the displacement of low-income households beyond city limits as a general trend for the three study regions, our findings underline the importance of monitoring current changes in the social geography of regions.Our exploratory quantitative approach (using purchasing power (de)concentrations in the city-regional context) created reliable results, as confirmed by the interviewed experts.Nevertheless, one limitation when studying the suburbanisation of poverty in Germany is the lack of data on residential moves across administrational boundaries.Complementing the perceptions of experts, such data is needed for further insights into new socio-spatial patterns in the relationship between urban and suburban areas in German city-regions.To gain deeper insights, individual-or household-level data and socio-economic data in finer spatial granularity are needed.It is no coincidence that evidence on the decentralisation or suburbanisation of poverty currently comes from those countries in Europe where such data is available in time series (the UK, Netherlands).Equally important are joint efforts to study current changes in the social geography of regions in cooperation with local authorities in city-regions, with a view to triggering political and public debate.While housing and labour markets are becoming increasingly regionalised, the monitoring of poverty concentrations-where it exists-remains scattered across different municipal administrations, thereby limiting the understanding of city-regional segregation processes and the debate on political measures.There is thus a need to develop city-regional monitoring and the accompanying research.Moreover, a strong political will for more decisive action to enhance housing options for low-income households is needed, such as social housing quotas in new housing developments, or the nurturing of providers of affordable housing (municipal housing companies, non-profit-housing cooperatives).
Funding Open Access funding enabled and organized by Projekt DEAL.This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) -Project number 451071476.

Table 1
List of expert interviewsFor reasons of anonymity of experts, no detailed information is given, for instance, on city administration department.Numbers in brackets indicate the number of interviewed experts per interview

Table 2
Characteristics of the housing markets of the three cities