Concurring urbanizations? Understanding the simultaneity of sub- and re-urbanization trends with the help of migration figures in Berlin

Abstract Most classical models of urban development conceptualize suburbanization and reurbanization as opposing processes of in- and out-migration. The article at hand starts from the understanding that this linear interpretation of migration figures falls short of the mark. Drawing on different approaches to operationalize reurbanization, case-specific migration-data from Berlin illustrates a population development where decline and revival do not exclusively happen at either the urban core or the city’s fringes, but can be observed throughout the city region. The paper concludes that seemingly contradicting urban dynamics, such as population gains and losses in different parts of a diversified spatial system such as the city region result from the same land and housing market processes which favour and disadvantage different social groups and neighbourhoods, respectively. This observation calls for a dismissal of mono-linear policy approaches in urban governance and planning, so as to address both macro- and micro-scale disparities in a fair manner and to prevent the neglect of the latter.


INTRODUCTION
Recent literature on the urban widely states a trend towards re-urbanization and a revival of central cities after decades of urban decline and population losses in favour of the suburban realm. At the same time, research on suburbanization finds substantial changes in what this century's (post)suburbia is made of. In the context of tense housing markets, conflicts in land use and heightened inter-urban mobility, what dominantly characterizes suburbia today is shifting from the once-imagined ideal of single-family houses towards a place of necessity for those who cannot afford to stay in the urban centres. As debates on re-and suburbanization Regional StudieS, Regional Science rarely get involved with the respective other, these findings are often overlooked. Hence, this paper tries to put both phenomena into perspective. By uncovering intersections of re-and suburbanization processes for the case of Berlin, Germany, it questions how far an exclusively numerical trend of population growth or decline can represent a cohesive image of urban development. By analysing quantitative migration data for the city and its surroundings and taking qualitative findings into account, it concludes that Berlin is not merely re-urbanizing, but will rather be characterized by the concurrence of re-and suburbanization in the years to come.

THE RE-URBANIZATION DEBATE AND THE CONTINUING RELEVANCE OF SUBURBANIZATION
In the late 1970s, a debate on the new attractiveness of cities commenced in Europe and the United States, where the decades after the Second World War had predominantly been characterized by the disregard of dense urban structures, the esteem of green living and concomitant pronounced suburbanization. Ever since, urban scholars have been focusing with increasing enthusiasm on a range of intersecting topics such as urban revitalization and urban renaissance schemes, new urban industries, living habits in post-Fordist societies, or international migration patterns occurring in the wake of globalization. While these contributions juggle with terms ranging from 'rebirth', over 'resurgence' to the 'revival of cities' (Rink, Haase, Grossmann, Couch, & Cocks, 2012), re-urbanization has emerged as a core term with which many authors strive to comprehend the various processes and trends that work towards a revaluation of 'the urban'. Being located within the highly complex triangle of research on migration, mobility and urban development, however, many of the publications and case studies that have contributed to this vibrant debate fall short of a comprehensive definition as well as a unified and undivided understanding of the phenomenon itself (Kabisch, Haase, & Haase, 2010, p. 969).
In current debates on re-urbanization, the heuristic model of urban development presented by van den Berg, Drewett, Klaassen, Rossi, and Vijveberg (1982) in their study on growth and decline in urban Europe more than 30 years ago is still an important point of reference -maybe less so because it provides an accurate account of the urban reality, but rather because it is so clear-cut and easy to translate in statistical analysis. It differentiates four phases of urban development: urbanization, suburbanization, desurbanization and re-urbanization -with the latter being defined (just like urbanization) as an increase in the core city's share of the entire urban region's population. In their attempt to understand the recent revival of many cities, several authors have refined this straightforward approach by considering demographic developments (e.g., Cheshire, 1995;Lever, 1993), migration figures (e.g., Herfert & Osterhage, 2012;Hirschle & Schürt, 2010;Kabisch et al., 2010), or the interrelation of built environment and population density (Wolff, Haase, Haase, & Kabisch, 2016). The analysis of these aspects has sometimes been combined with more qualitative features such as the diversity of urban population and housing (e.g., Buzar et al., 2007b;van Criekingen, 2010;Haase et al., 2010;Ogden & Hall, 2000;Seo, 2002). A variety of empirical findings show that often rather ambiguous intra-and inter-urban migration patterns drive population change in an urban region. Therefore, such diverse trends as resurgence, gentrification and degradation in different urban neighbourhoods as well as suburban growth and suburban decline in different areas of the urban fringe are common for many urban regions today (cf. Buzar et al., 2007b, p. 671;Deas & Hincks, 2014, p. 2578Jessen, Siedentop, & Zakrzewski, 2012, p. 206;Rérat, 2012Rérat, , p. 1122.
Few studies deal with re-urbanization at the European scale. In a quantitative analysis of population trends in several European cities, Buzar, Hall, and Ogden (2007a) highlight the growing importance of international and interregional migration for the cities' trajectories and reveal changing household demographics in inner-city areas that could be interpreted as a kind Regional StudieS, Regional Science of re-urbanization. In a similar vein, Rae argues 'that a population explosion in many of the inner cities was contributing greatly to population growth across the core cities as a whole' (Rae, 2013, p. 97). Looking at individual case studies, Buzar et al. (2007b) as well as Haase et al. (2010) describe re-urbanization trends in four European cities and identify both their drivers and obstacles. However, they also conclude that differences in re-urbanization trends call for a much more differentiated analysis -not only between individual cities but also between particular areas within these cities. In addition to these studies, a host of papers focus on instances of (presumed) re-urbanization by discussing urban development trends in individual countries. In the UK, for instance, case studies engage with significant population growth in revitalized city centres (Couch, Sykes, & Börstinghaus, 2011) or causes and characteristics of urban shrinkage and growth (Rink et al., 2012). Focusing on Switzerland, Rérat (2012) recognizes international migration to be the driving force of re-urbanization. Interestingly, he finds a concurrency of phenomena such as urban sprawl, gentrification and suburbanization at the macro and micro scales (Rérat, 2012(Rérat, , p. 1122. For the German context, especially Häußermann and Siebel (1987) have triggered a debate on 'new urbanity'. In this vein, Brühl, Echter, Frölich von Bodelschwingh, and Jekel (2005) discuss whether the most recent changes in urban migration and population figures actually indicate the end of inner-city decline due to suburbanization. Sander (2014, pp. 233ff.) highlights the rediscovered attractiveness of urban living. She finds that young adults continue to move to cities, while many young families do not want to move to the urban fringes anymore. In a case study in the East German city of Leipzig, Wolff et al. (2016, p. 1) analyse the complex relationship between the built environment and housing demand by focusing on the influence of population density. They find that regrowth in Leipzig is marked by a very heterogeneous spatial pattern and must therefore be understood as both a distinctive and fluid process.
However, only very recently the main focus of urban debates has generally shifted from a concern about an ongoing expansion and sprawl of urban areas towards re-urbanization and concomitant gentrification and displacement of poor households due to an increasingly dense housing market (Adam & Sturm, 2014;Holm, 2012). Focusing on Berlin's branding as a 'creative city', Colomb (2012, p. 147) identifies clear areas of conflict regarding temporary uses of vacant lots and spaces in the city centre and their redevelopment for economic and marketing purposes.
As presented above, current debates on re-urbanization cover a vast spectrum of topics, including highly contested phenomena such as gentrification and displacement. However, as a rule they concentrate more or less exclusively on dynamics of growth, densification and increased competition in the land and housing market. Thereby, they hardly acknowledge simultaneous processes of deconcentration and peripheralization (in both spatial and social terms) and tend to overlook trending conceptual changes in research on the (post-)suburban. At the same time, suburbia remains a highly relevant part of urban development. According to Ekers, Hamel, and Keil (2015, p. 19), 'suburbanization will be the defining urban-regional process for at least a generation to come' with wide implications that touch upon intersecting domains and raise critical questions about social segregation and justice, sustainable forms of land use, and uneven development.
In its classical understanding, suburbia is mostly depicted as a place of longing. In accordance with idealized images of a peaceful, quiet and green living environment close enough to the vibrant city, moving to the suburbs has been linked to individual economic prosperity and the wish for a socially homogeneous neighbourhood (Häußermann, 2009). Today, the majority of the global urban population does not live in city centres, but on the edges, the fringes, the so-called periphery -and many have not made their residential choice out of privilege, but rather moved out of the centre as they were forced to find affordable housing options elsewhere (Hochstenbach & Musterd, 2017). Post-Fordist suburbia broaches the issue of revised spatial identities and is shifting its image from a place of longing towards a place of everyday negotiations of belonging. Notwithstanding the observations of re-urbanization tendencies, what happens in the cities' suburban realm continues to be highly important not only for suburbia itself, but also regarding its connections to and dependencies Regional StudieS, Regional Science with the urban core. Against this background, the following section will analyse re-urbanization processes in the Berlin urban region in a way that connects it to concurrent suburbanization dynamics, i.e., peripheralization and polarization.

BERLIN AS A RE-URBANIZING CITY -A BOLD AND SUPERFICIAL DIAGNOSIS?
For the second half of the 20th century, Berlin's development was marked by a declining number of inhabitants. After the unification of Germany and, also, Berlin, the city's population continued to decline and dropped by around 100,000 during the 1990s. This process was mainly driven by removals of Berlin households to the urban surroundings (i.e., suburbanization) and peaked in 1998, when Berlin lost over 30,000 inhabitants to its suburban fringe within only a year (Bluth, 2004, p. 95). Afterwards, the negative population trend tailed off, and since 2004 Berlin has shown continuous growth rates, resulting in about 3.5 million inhabitants today. Moreover, in 2015, the Berlin city administration expected an additional increase in population of at least 250,000 until 2030 (Senatsverwaltung für Stadtentwicklung und Umwelt, 2016) -and due to the vast number of refugees who have currently been moving to Germany it is very likely that this forecast is even too moderate. While these trends have often been interpreted as a transition from suburbanization towards the much-invoked re-urbanization, it is worthwhile to take a closer look at the data underlying this diagnosis.
In his study on re-urbanization in Berlin, Brake (2012, p. 282) detects a clear trend towards re-urbanization as a result of a noticeable increase of the inner city -in comparison with the outer city -in attractiveness for housing and creative economies as well as the establishment of an urban milieu of consumption since 1990. In a similar vein, a previous study by Hentschel (2008, p. 6) had found decreasing suburbanization trends in Berlin, but did not interpret this finding as an indication of re-urbanization. Insofar, both studies and their (seemingly) contradictory results hint at the necessity to reflect thoroughly on possible interpretations of quantitative data on urban population development before taking them as a proof for major trends in current urbanization. The present paper, therefore, (re-)investigates population data from the Berlin urban region and thereby subjects the well-established interpretation of these data as a signal of re-urbanization to critical scrutiny. It will highlight that the complexity of demographic transformations usually does not display one trend of urban development, such as re-urbanization or suburbanization, exclusively.
This paper therefore strives to get an analytical grip on current dynamics of urbanization by analysing current demographic data from the Berlin city-region that covers the period between 2006 and 2013. In order to take Berlin's size as well as its heterogeneity into account, the analysis not only will differentiate between the city proper and its surroundings, but also will consider population change at the neighbourhood level within the urban core. The small-scale analysis of population and migration data draws to the 447 Planungsräume (planning units, i.e., neighbourhoods) of Berlin, which have been delineated for statistical, monitoring and planning purposes on the basis of the Berliners' (assumed) life-world-based perception of neighbourhood areas. The availability of data for these units makes it possible to differentiate between the inner and outer city within the city proper. Berlin's inner city is roughly delimited by the inner urban railway circle (though at places it extends beyond this circle) and is mainly characterized by perimeter development of the (late 19th-early 20th centuries) Wilhelmine period and high population density. On the contrary, the outer city shows a diverse building structure and, inter alia, dominantly consists of cooperative housing estates of the inter-war period, large social housing developments of the post-war decades and vast areas of single-family homes.
Regional StudieS, Regional Science Re-urbanization as increased significance of the core city as compared with its surroundings In a first step, the critical analysis and interpretation of population data from Berlin draws on three concepts of measurement which have been introduced by Osterhage (2011Osterhage ( , p. 1358 to quantify and operationalize re-urbanization. The comparison of the core city with its suburban surroundings lies at the heart of the concepts according to which re-urbanization can be defined either as: • a positive population trend (growth) in the urban core; • a relatively favourable population trend in the urban core as compared with its surroundings; or • migration gains of the urban core from its surroundings. 1 The following sections will now discuss re-urbanization processes in Berlin with regard to each of Osterhage's three concepts and then offer alternative modes of interpretation, which suggest that re-urbanization needs to be analysed in connection with suburbanization to lead to a productive debate on regional policy implications in Berlin.

Re-urbanization as a positive population trend in the urban core
As Berlin has been continuously growing since 2004, it is now in a phase of re-urbanization according to the simplest understanding of the term. Additionally, the annual growth rate has strongly increased in the last years (Figure 1), so that a permanent positive development trend can be assumed here. This overall population gain is mainly a result of national as well as international migration. In the migration balance between Berlin and the 15 other federal states of Germany, the former was able to gain more than 150,000 inhabitants between 2006 and 2013. On the international scale, the city showed a positive migration balance of more than 93,000 people in the same period -and in particular since 2011 when immigration rates soared, following the economic crisis in many south-east European states. Overall, the number of new citizens coming from abroad has doubled since 2006, which conforms to Buzar et al.'s (2007a) aforementioned emphasis on the importance of international and interregional migration. Regional StudieS, Regional Science Re-urbanization as a relatively favourable population trend in the urban core as compared with its surroundings A comparison of the population development in Berlin and its surroundings reveals that the former clearly exceeds the latter in terms of annual growth rates since 2011 ( Figure 2). Therefore, re-urbanization can be stated from the point of view of the second of the measurement concepts presented above as well. However, this diagnosis only holds true for a very limited time span. Up to date, it is therefore hardly possible to confirm a general reversal of trends. Moreover, Figure 2 shows that since 2011 the surrounding area of Berlin has strongly grown in population, which might suggest another approximation of population gains between the core city and the suburban fringe.

Re-urbanization as migration gains of the urban core from its surroundings
Finally, it is especially the migration data on the city-region that raise doubts about the re-urbanization diagnosis for Berlin. In the considered period between 2006 and 2013, the city's migration balance with its surrounding area is throughout negative. Furthermore, just like the core city, the suburban fringe can register population growth. Although migration losses of the core city to the suburban realm decreased until 2010, they have increased again since then ( Figure 3). The current data hence contradict the hypothesis of (continuous) decline of urban-to-suburban migration which was recognized by Brake (2012). This again corroborates the assumption of an ongoing suburbanization in Berlin.

Re-urbanization as increased significance of the urban core (inner city) as compared with the outer city
Overall, the careful analysis of population and migration data justifies the diagnosis of re-urbanization for the case of Berlin only to a limited extent. There is certainly an absolute (and, more recently, also a relative) reinforcement of the core city as compared with its suburban surroundings. When it comes to population development, however, migration data indicate ongoing, though somewhat fading, suburbanization. Due to the sheer size of Berlin's core city, only a small-scale analysis can provide more precise information on the quantity and quality of those migration trends working in favour of re-urbanization. To this end, the analysis of population and migration data will now turn to the level of neighbourhoods and, in so doing, consider the differentiation between the inner and the outer city of Berlin. Figure 2. annual population growth/decline in the Berlin city-region, 2007-13 (%).

Regional StudieS, Regional Science
Berlin's population growth manifests itself in both the inner as well as the outer city, but is considerably more pronounced in the former (Figure 1). In the inner city, almost nine out of ten Planungsräume grew between 2007 and 2013, while this is only true for three out of four Planungsräume of the outer city (Figure 4). It is also notable that population losses can mainly be found on the urban edges. Just like the numerical comparison between the core city and its surroundings, the mere population development within Berlin shows a tendency for re-urbanization in the sense of disproportionate reinforcement of the central areas.
But -as was the case with the comparison of the city and its surroundings -these findings need to be put into perspective by looking at migration patterns. From 2006 to 2013, the inner city of Berlin continuously shows a negative migration balance with the outer city ( Figure 5). Migration losses even increased from 7.5 to 11.8 people per 1000 inhabitants. Hence, the dominant trend of internal migration within the Berlin city limits definitely has to be understood as deconcentration rather than concentration. This result provides an argument against the re-urbanization diagnosis and becomes particularly clear when examining the migration balances of the single Planungsräume ( Figure 6).
To conclude, the comparison of the city and its surrounding area as well as the comparison of inner and outer city yields the same result for the case of Berlin: with regard to population development, a tendency towards re-urbanization is visible -yet the process of intra-regional migration still follows a centrifugal pattern. A look at how migration patterns vary over all three areas of analysis (inner city, outer city, surrounding area) provides a more detailed insight. It shows that the inner city lost about 4.5 people per 1000 inhabitants to the suburban fringe between 2006 and 2013. This is a lot less than the outer city, which had to put up with population losses of 20.4 people per 1000 inhabitants. However, the general trend regarding these figures is the same for both the inner and the outer city: the losses to the suburban fringe decreased from 2006 to 2010 (when the inner city even achieved an almost compensated migration balance), but from 2011 to 2013, the losses were again rising. Figure 7 breaks down urban-to-suburban migration balances to the level of municipalities (Gemeinden) and Planungsräume. It demonstrates how the migration losses of the outer city to the suburban fringe are clearly exceeding those of the inner city. Surprisingly, they are most pronounced in the outer-city units far from the urban core and close to the suburban fringe, while the adjacent municipalities located directly outside Berlin's city boundaries are profiting most from this suburbanization.   Regional StudieS, Regional Science Altogether, the analysis of internal migration patterns in the Berlin city-region reveals a cascade-like removal chain -from the inner to the outer city and from there to the urban surroundings, i.e., the suburban fringe: national and international incomers apparently are searching for housing mainly in the inner city, while, when moving into a new home, established households tend to move in an outward (centrifugal) direction, away from the centre towards the periphery. A possible reason for this observation lies in the growing tension in the Berlin housing market. This is also underpinned by the general decrease of the internal migration volume: the number of removals of households within the Berlin city-region dropped by 13.4% since 2009, which relates to the tendency of people to try to stay in their apartment if they can. Even though in 2012 and 2013 the general migration volume in the inner city had risen again, those migration patterns point towards the growing relevance of affordable housing and its implications for individual relocation decisions as well as connections with debates on displacement and gentrification processes. It is important to note that the social groups who are the driving force behind current re-urbanization trends are in all possibility different from those that move outwards. Regional StudieS, Regional Science

FROM SUBURBANIZATION TO RE-URBANIZATION -TO BOTH?
The multifaceted character of Berlin's widely assumed trend towards re-urbanization has been highlighted above. The small-scale analysis of data from the Berlin urban region on population development and migration patterns confirms the fact that the German capital remains highly attractive for both national and international in-migration, which is reflected by its population growth. To go as far as to interpret this population growth as re-urbanization, however, does not take the overall migration patterns within the Berlin city-region into account. These still display patterns of suburbanization with a positive net-migration balance of the urban surroundings Regional StudieS, Regional Science compared with the core city. Moreover, a majority of the inner-city removals are headed towards the city's edges, not towards the centre. In sum, these findings shed light on the concurrence of several -and partly opposing -urban development trends, thereby calling for a differentiated and contextualized understanding of both contemporary re-as well as suburbanization. With regard to the often-opposing urban migration patterns presented above, we suggest to complement Osterhage's conceptualizations with an understanding of urban development that allows for concurrence to be a logical characteristic of migration flows. Especially when taking questions of tense housing markets, conflicts about the right to the city and the displacement of sensitive population groups into account, it seems inevitable to allow for the synchronicity of sub-and re-urbanization because these processes indicate conflicting processes on urban housing markets rather than the seemingly predestined effectiveness of principles of urbanization.
Overall, the findings from the case study presented in this paper ask for a more precise understanding of changing modes and characteristics of re-urbanization. In accordance with the recent scholarly interest on intersectionality and manifold modes of urban development, what has predominantly been considered part of a periodic and regular phase in van den Berg et al.'s urbanization model should be observed at a smaller scale of urban space so as to allow for the heterogeneity in urban development trends and to uncover non-cyclic, but nevertheless highly relevant, demographic trends and migration patterns in urban and suburban neighbourhoods.
The crucial point here is to think of suburbanization as a constant and steady migration phenomenon that might not always prevail as a numerical trend for the whole city, but which remains a reality for relevant numbers and groups of urbanites. In reference to claims forwarded by Hochstenbach and Musterd (2017, p. 1), a trend towards suburbanization and poverty could have serious and far-reaching implications for the socio-economic patterning of our cities. Interlaced with these findings, the solely quantitative trend towards re-urbanization in Berlin puts questions of urban inequality and social justice back on the agenda. Consequently, urban governance and planning modes need to be re-equipped to address specific conflicts that might arise from socio-economic segregation and new population dynamics on the city's fringes.
It must be considered, however, that this paper focuses exclusively on the demand side of residential mobility. Further research should therefore focus on a differentiation between migration triggers such as a possible displacement of old-established citizens due to the growing pressure on the inner-city housing market, or an ongoing, 'classic' suburbanization trend that is propelled by individual housing preferences. To meet further the challenges of grasping the complex and partly intersecting layers of urban growth and development in relation to mobility and migration, it would be desirable to know more about housing duration and satisfaction among dwellers in the core city and suburbia as well as on their willingness to relocate. This would take those people into account who did not move (yet) and therefore do not appear in migration statistics so far. The rationale behind this is that even the mere intention of changing or keeping a domicile could be interpreted as a micro-level indicator of or against an invoked re-urbanization.
Finally, if we shift towards a more integrated understanding of the city-region as a complex, connected and heterogeneous whole, we cannot continue to put aside micro-scale dynamics in supposedly less relevant units in favour of the dominant ones in our urban areas of concern. It is worth reflecting on heuristic models of urbanization to allow for a further opening of critical debates on intersecting and dynamic development in which phenomena of re-and suburbanization occur simultaneously.

NOTE
1. Note that the second of these concepts is identical with the definition of re-urbanization by van den Berg et al. (1982) and that only the first concept definitely excludes the possibility of re-urbanization in an overall shrinking urban core.