Assessing displacement in a tight housing market: findings from Berlin

Displacement from one’s home is a contested issue, not only in actual urban politics but also in urban research. Empirical studies rely on rather different notions of displacement, which makes a coherent picture of the phenomenon difficult to obtain. In this article, we first deal with the conceptual ambiguities in the debate on displacement so as to carve out an empirically viable definition of direct displacement that focuses on the decision-making process before a relocation. We then present and discuss our own empirical findings on displacement. Finally, we reflect on possible conclusions one could draw from these findings in relation to housing policy. Our empirical results come from a survey we conducted of more than 2,000 tenants who had recently moved from their homes in Berlin, Germany. We found that more than 15% of the respondents had experienced direct displacement. A rent increase after refurbishment or the selling of the property proved to be the most common triggers of displacement. Addressing these particular issues therefore appears to be critical to curbing displacement in tight housing markets.

Assessing displacement in a tight housing market: findings from Berlin

Fabian Beran and Henning Nuissl
Displacement from one's home is a contested issue, not only in actual urban politics but also in urban research.Empirical studies rely on rather different notions of displacement, which makes a coherent picture of the phenomenon difficult to obtain.In this article, we first deal with the conceptual ambiguities in the debate on displacement so as to carve out an empirically viable definition of direct displacement that focuses on the decision-making process before a relocation.We then present and discuss our own empirical findings on displacement.Finally, we reflect on possible conclusions one could draw from these findings in relation to housing policy.Our empirical results come from a survey we conducted of more than 2,000 tenants who had recently moved from their homes in Berlin, Germany.We found that more than 15% of the respondents had experienced direct displacement.A rent increase after refurbishment or the selling of the property proved to be the most common triggers of displacement.Addressing these particular issues therefore appears to be critical to curbing displacement in tight housing markets.

T
he housing market has become increasingly tight in many urban areas of the global north.Rising rents and a decreasing availability of rental properties have led to the reemergence of the provision of housing as a topical question in urban politics.At the same time, tenants are organising to protest for affordable housing and against displacement in many European cities.In Berlin, the case study for this article, citywide protests have clearly helped the gearing of housing policy toward regulative instruments that go far beyond established market-and supply-oriented strategies.These included the introduction of a 5-year rent cap in 2020, which, among other things, froze existing rents and even allowed for rent reductions (up to defined limits).However, the Federal Constitutional Court of Germany decided in April 2021 that the federal state (land) of Berlin lacks the authority to set its own rent cap because the legislation on tenement and rents, according to the Court, has to be set on the federal level.This annulation of the renter-friendly legislation of the city-state increased the momentum of civil society-based initiatives.For example, a tenants' initiative is calling for the expropriation of real estate companies that own more than 3000 flats in Berlin in favour of municipal ownership.The initiative, supported by trade unions and left-oriented political parties, launched a petition for a referendum to pass a law to implement its goal (Richter and Humphrey 2021), which was successful in September 2021.The Berlin state government is now considering a new law on the socialisation of housing.
The displacement of tenants receives particular attention in the reinvigorated political debates on housing because it is probably the most significant among the unwanted consequences of tight housing markets.However, there is also a long tradition of dissent among both politicians and urban scholars regarding the magnitude and the very nature of the problem of displacement.Slater (2009Slater ( , 2010) ) and Hamnett (2009) have provided a prominent example of this dissent with their debate in this journal.Whereas the former saw displacement as part of a class struggle for urban space in the context of gentrification, the latter characterised it as a somewhat negligible by-product of desperately needed inner city revitalisation.
Scholarly disagreement about the role, scope, and importance of displacement in urban development certainly reflects different political positions authors take in relation to present urban conflicts.However, it is also due to the generally shallow empirical and conceptual basis of the debate.There is a lack of substantial research that has examined the magnitude and patterns of displacement, which can be seen as the somewhat inevitable result of a lack of conceptual clarity as to what displacement exactly means (Lukens 2020, 2).Moreover, the notion of displacement (as a specific incident) is often tightly linked to the concept of gentrification (as a more general phenomenon or syndrome of capitalist urban development), in particular in critical urban research, which adds to the conceptual confusion around displacement.Gentrification usually signifies the upgrading of an urban neighbourhood or area in terms of both real property value and social status.'Social upgrading' is usually interpreted as a result of the displacement of low-income households.Therefore, displacement is a central-and, for many authors, an even constitutive-element of gentrification (e.g.Davidson andLees 2005, 1187;Slater 2006, 748;Morris 2019).However, whereas the plausibility of a tight relation between gentrification and displacement is indisputable, the actual interrelation of the two concepts often remains blurred, which is why their conceptual nexus has been questioned recently (e.g.Gillespie, Hardy, and Watts 2021).In this vein, Lukens (2020, 16) called for a disentanglement of the two concepts: 'If displacement is to remain central to gentrification research, it is imperative that gentrification' s displacement impacts be disaggregated and examined carefully' .Hence, it seems necessary to first understand and quantify displacement in its own right before it is possible to acquire a full understanding of gentrification dynamics and to identify remedies against displacement.Accordingly, Elliot-Cooper, Hubbard, and Lees (2020, 504) claimed 'that more robust data are needed' on displacement phenomena.Likewise, Zuk et al. (2018, 35) asserted that 'advanced tools to define and measure these analytically distinct phenomena' are required.Empirical operationalisation of and investigation into displacement faces a particularly tricky methodological challenge: Displaced people do not live at the locations from which they were displaced, so they are difficult to find.Atkinson (2000, 163) described this challenge as 'measuring the invisible' .
In this article, we present an approach to conceptualise and measure displacement.This approach is based on a study we conducted together with the Wüstenrot Foundation (Beran and Nuissl 2019).First, we show that empirical research on displacement usually applies rather different-and sometimes even imprecise and inappropriate-concepts and methods, thus leading to a rather inconsistent overall picture.We then move on to our own research on displacement in Berlin, which is based on a representative survey of movers: We sketch the conceptual framework of our study and present what we found out about the extent of displacement and the factors that cause it.On the basis of these findings, we discuss how housing policy can help prevent displacement.

State of the research: approaches to define and quantify displacement
The term displacement Recent urban research has dealt with displacement from various perspectives, thus yielding a variety of terms that are linked to different phenomenological and methodical dimensions but that fail to establish a coherent understanding of exactly what displacement is.Elliot-Cooper, Hubbard and Lees (2020, 493) perceived a 'pliability of the term' and even warned of displacement 'becoming a classic "chaotic" concept' .Against this background, it seems worthwhile to recall the seminal definition that Georg and Eunice Grier suggested as early as 1978: Displacement occurs when any household is forced to move from its residence by conditions which affect the dwelling or its immediate surroundings, and which: 1. are beyond the household' s reasonable ability to control or prevent; 2. occur despite the household' s having met all previously-imposed conditions of occupancy; and 3. make continued occupancy by that household impossible, hazardous, or unaffordable.(Grier and Grier 1980, 256) Building on this definition, Peter Marcuse (1985, 204ff.)developed a muchcited typology of four types of displacement.The first type, last resident displacement (i), is equal to what Grier and Grier described in their definition and is yet further divided into physical displacement (caused, e.g. by the owner turning off the heating) and economic displacement (usually caused by a rent increase).If this kind of displacement occurs repeatedly in the same house or tenement, we see an instance of 'chain displacement' (ii).Marcuse complemented these two types of direct displacement with an additional two, more indirect, types: 'exclusionary displacement' (iii), which occurs when a housing unit cannot be occupied again by a household with comparable socioeconomic characteristics after the previous tenants have moved out because the unit is now offered at a higher price, or left vacant, and 'pressure of displacement' (iv), which arises from circumstances outside of one' s own home, such as the replacement of existing retail and service infrastructures in the neighbourhood by new ones that do not meet the demand of the existing residents.This latter displacement type covers phenomena that have also been called cultural displacement (e.g.Cahill 2007;Valli 2015) or displacement due to alienation (e.g.Hyra 2015).Marcuse emphasised that all four types of displacement have to be taken into account at the same time in order to be able to describe and explain the entire issue of displacement.
Marcuse' s displacement typology offers two essential methodological pointers for the conceptualisation and measurement of displacement.First, regarding the unit of analysis, he suggested investigating direct displacement and displacement chains at the level of persons or households, exclusionary displacement at the level of dwellings and displacement pressure at both levels (Marcuse 1985, 206ff.).Second, one can deduce from Marcuse' s typology that different forms of displacement require different times of study: Direct and indirect forms of displacement can be observed at a particular point in time, but displacement chains can be observed only via longitudinal analysis.Marcuse' s (1985) and Grier and Grier' s (1980) concepts of displacement involve at least two definitional elements to distinguish it from other forms of relocation: (1) it is induced by particular causes (physical, emotional, or economic pressure), and (2) it affects only tenant households.The former criterion allows a differentiation between different forms of displacement.As regards the latter criterion, it is important to note that it is not necessarily constitutive of the displacement notion in that there are also a few studies about the displacement of property owners (e.g.due to rising property taxes; e.g.Martin and Beck 2018).
In addition to the 'objective' cause(s) of a move, several studies have taken into account the 'subjective' perception of the move in order to conceptualise displacement: Usually, they define displacement as an involuntary move (e.g.Kearns and Mason 2013;Watt 2018, 71).Häußermann, Holm, and Zunzer (2002, 219f.)went even further, suggesting that we need to classify any relocation of a household along a scale that spans the poles of (forced) displacement and an entirely voluntary decision to move.In their view, direct displacement is tantamount to coercion to move, whereas indirect forms of displacement can be perceived as both voluntary and involuntary.Other authors generally have refused to use the individual perception of a move to conceptualise displacement, for two primary reasons.First, the individual perception of an issue is contingent; that is, it can be based on factors (e.g.emotions, life circumstances) that have little or no connection with this issue, and second, psychological processes, such as (reduction of) cognitive dissonance, can substantially distort an individual' s perception of an event, in particular if this event, as is the case with relocation, has profound consequences for the individual or household concerned (Betancourt 2018, 248;Grier and Grier 1980, 253).
There are more criteria that can be used to define displacement.For Kearns and Mason (2013, 196f.),only moves that have adverse consequences for the persons concerned constitute displacement.For example, they defined a move that leads to a loss of social contacts as social displacement (Kearns and Mason 2013, 184).However, with regard to empirical research, defining displacement by means of its consequences has the same drawbacks as defining it by means of individual perception because the evaluation of these consequences will usually be possible only on the basis of what the person concerned says about the consequences of their move.Thus, there is a considerable risk of underestimating the negative consequences of a move (that is, displacement), because people tend to evaluate their current housing situation positively, regardless of the actual circumstances (Häußermann and Siebel 2000, 218).
In addition to the conceptual approaches we have just sketched, some authors have proposed to extend the displacement notion to particular occurrences in the life of tenants that do not involve a change of address at all.Blasius (1993, 235) described households that cannot afford to move after a rent increase (because offers on the housing market are even more expensive) and therefore have to lower their other expenses as being displaced from their standard of living.For Davidson (2009, 228) and Elliot-Cooper, Hubbard, and Lees (2020, 498) the alienation of tenants from their living environment (because of changes in the latter) is also a kind of displacement.Hence, the notions of displacement found in the literature also differ with regard to whether they include the possibility of displacement without a change of location, that is, with regard to the spatial reference of the phenomenon.Moreover, displacement concepts for which a change of location is not constitutive are usually based on an extended time reference: Displacement is seen as a process rather than a singular event (e.g.Sakizlioğlu 2014, 16f;Elliott-Cooper, Hubbard, and Lees 2020, 501f.).In this vein, Watt (2018) showed how recurrent displacement (including a change of location) constitutes a specific issue, experienced in particular by the most vulnerable members of society.Hirsh, Eizenberg, and Jabareen (2020, 399) also stressed that displacement refers to a timeframe.In addition, they took a particularly theoretical stance and related displacement to concepts such as power, positionality, eligibility, and resistance, thereby embedding displacement in its particular societal, economic, and cultural context.However, although these contributions have the merit of locating the issue of displacement in broader urban theory, it is hardly possible to translate their approach empirically, to actually measure and map displacement.
Overall, displacement is a multidimensional concept.The deconstruction of some of the most eminent displacement concepts found in the literature reveals a range of theoretical dimensions (variables) of displacement along which these concepts are defined.Each displacement concept adopts at least one of the specifications (values) within these dimensions, which are thus constitutive of the respective concept (Table 1).

Methods of quantifying displacement
Before we present our own study, in which we assessed the actual share of displaced tenants among all moving tenants in a particular case study area, we take a closer look at similar studies that endeavoured to measure displacement.These studies either rest on a survey of moving households or tenants, or they used a secondary analysis of demographic and migration data.

Survey of (potentially displaced) movers
Asking people who have moved about their reasons for moving is considered the most valid method for empirical research in all forms of displacement Source: Slightly modified from Beran and Nuissl (2019, 73).
that imply a change of housing (Carlson 2020, 577;Diller 2014, 29).However, this method requires that one locates the people to be interviewed (by means of the official register of residents) and-if the results of the survey are to be generalised-ensures their representativeness.Primary surveys that meet these requirements, and thus allow a quantification of displacement, are available in only small numbers (e.g.Blasius 1993;Desmond and Shollenberger 2015;Schill and Nathan 1983).Secondary analysis of survey data, which contain information on the housing situation and moves of the respondents but that do not directly target the problem of displacement, are more frequent.However, these studies are usually not able to provide a valid estimate (let alone a measurement) of displacement.For example, the representativeness of the New York City Housing and Vacancy Survey used by Freeman and Braconi (2002) and Newman and Wyly (2006) seems questionable because it does not include people who moved in with relatives or friends after a displacement (Newman and Wyly 2006, 29).Moreover, studies of displacement that are based on survey data (be it data from other research contexts or primary surveys) can often be criticised for failing to operationalise displacement convincingly because they usually use only one indicator to identify displaced tenants, namely the-most important-reason for moving given by the respondents.Thus, they do not adequately reflect the complexity of the decision to move (e.g.Freeman and Braconi 2002;Newman and Wyly 2006;Schill and Nathan 1983).If, for example, a tenant takes the impending modernisation of their flat and the foreseeable increase in rent as an opportunity to move closer to their workplace, one cannot rule out the possibility that they will cite in a survey the latter as the most important reason for moving, so that the actual (displacing) nature of their move remains in the dark (Freeman, Cassola, and Cai 2016, 6;Lees, Slater, and Wyly 2010, 319).

Analysis of secondary data
There are quite a few studies that analyse displacement empirically by means of existing data sets.Usually, they operationalise displacement as a departure from an (urban) area that can be attributed to upgrading or gentrification.In a first step, these studies select gentrification areas where, in a second step, they examine relocation data in order to estimate the number of displacement cases (Holm and Schulz 2018, 258f.).With regard to the first step, several authors (e.g.Easton et al. 2020, 288) have pointed out that the conceptualisation of gentrification underlying the selection of areas is often questionable.Holm and Schulz (2018, 259) noted that many studies use only sociodemographic data to identify gentrifying areas and fail to allow for indicators on real estate economic upgrading.Barton (2016, 108) criticised the lack of control areas in many studies that focus on gentrification in urban neighbourhoods.Preis et al. (2021) showed that four different approaches to determining gentrification, which they took from other studies and tested in relation to Boston, Massachusetts, lead to very different results as to which areas of Boston are gentrifying and face displacement pressure.With regard to the estimation of the extent of displacement, secondary analytical studies take different paths and, as a consequence, yield different results.Some studies (e.g.Van Criekingen 2009, 840f.;Atkinson 2000, 158ff.)focus solely on vulnerable population groups and interpret their departure from a neighbourhood as displacement; however, this operationalisation appears to underestimate the complexity of displacement because it associates displacement only with particular sociodemographic groups and loses sight of the reasons for moving (Easton et al. 2020, 292).Others try to establish the effect of gentrification on mobility and to extract the share of displacement in all moves by means of regression analyses (e.g.Ding, Hwang, and Divringi 2016;Freeman, Cassola, and Cai 2016).In this vein, Fransham' s (2020, 10) 'logistic regression models of the propensity to move out of an area show that there is no evidence of increased outward mobility of income-poor individuals from gentrifying compared with nongentrifying areas' .Studies of this kind can be criticised in at least two ways.First, the mobility rate of poor people is dependent not only on the development of the area where the people live but also on the general availability of affordable housing.In other words: the problem of displacement might be concealed because to a considerable extent it occurs as what Marcuse (1985) called either exclusionary displacement or displacement pressure.Second, apart from a few exceptions (e.g.Schulz 2019), these studies usually fail to include meaningful indicators of real estate upgrading in their regression models, so their explanatory power is limited.
Case study: conceptualising and measuring displacement in inner city Berlin, Germany

Decision-theoretical migration research as a conceptual starting point
Studies that use a comparatively precise definition of displacement rarely succeed in adequately capturing the complexity of relocation decisions.To address this conceptual flaw, we draw on decision-theoretical migration research and distinguish two (ideal-typical) phases in the process of consideration and decision making that precedes a move (Föbker 2008, 50).In the first phase, a household develops a desire to move; in the second phase, it looks for a new place to live.In the first phase, a discrepancy occurs between the housing demands and the housing situation of a household, creating a problem pressure, that is, 'stress' (Brown and Moore 1970, 2).This discrepancy can be due to changing desires regarding housing, for example, after a baby has been born, but it also can be due to changes in the housing or rental situation, such as a rent increase.The latter, so-called external changes, together with the reactions that can be taken in face of them, are key to the conceptualisation of displacement.We distinguish three sets of actions from which a household can choose in order to reduce the problem pressure that are mainly in line with the work of Moritz Rinn (2021): opposition, passivity, or migration (Kecskes 1994, 130f.). 1 For instance, if they enter an objection against the announced modernisation of their house, they go for opposition; if they do not do anything (but instead wait to see if things change for the better), or adapt their aspirations to a reduced financial budget, they go for passiveness (Steinführer 2004, 23f., 28-32).If they decide to move, the external factors that brought about the perceived problem pressure can be interpreted as displacement factors.In the second phase, a household that wants to move looks for and then chooses a new place to live.If they are unsuccessful, they return to the first phase and must weigh the three sets of action again.

The concept of displacement from the perspective of decision theory
The definition of displacement we propose in this article is based on the phenomenological dimensions presented in Table 1 and the arguments of decision-theoretical migration research explained in the previous section.
The focus is on direct displacement caused by physical or economic pressure.
In addition, displacement is understood as the departure of tenants from a flat and as a singular event that is preceded by a conscious decision.Table 2 illustrates how the concept of displacement chosen in this article relates to the dimensions of the displacement notion introduced in Table 1.It does not consider the question of how the moving tenants themselves perceive their relocation because, from a decision-theoretical perspective, only the motives for an action are relevant, not the ex post assessment of this action or its consequences.
According to the cornerstones of our concept of displacement listed in Table 2, we can now specify the conditions for classifying a certain event as displacement: We define displacement as moving out of a rented dwelling due to changes that • affect the tenancy of the moving household; • cannot be controlled or avoided by the person(s) moving (so that they can occur, e.g. even if all the obligations prescribed in the tenancy agreement are fulfilled); and • have significantly contributed to the decision to move.This definition strongly resembles Grier and Grier' s (1980) but modifies it in such a way that it corresponds to the phase model of decision making in regard to relocation.It can be operationalised empirically by classifying moving tenants as either displaced or not displaced on the basis of the reasons they give for moving (without considering whether they feel displaced).In the section after next, we explain what external changes that can trigger a move (displacement factors) can be used to classify moves as displacement.

Research design
To gain a quantitative insight into the problem of displacement, we conducted a written survey in 2015 to find out the reasons why movers had changed their flat and whether they had been displaced.We thus examined displacement at the level of individuals and as a point-in-time study.
Our study area consisted of two inner city districts in Berlin: Friedrichshain-Kreuzberg and Mitte, which have been severely affected by a tightening of the housing market.The basic population of this study was defined as all tenants over age 18 years (born before 1 September 1995) who had moved from their flat in the study area (primary residence) to another house within Berlin between 1 September 2013 and 31 August 2015.A random sample of 10,000 people was drawn from this population, using the Berlin population register, which included all necessary information.
The (largely) standardised questionnaire was sent to the selected 10,000 persons by post.To increase the response rate and to minimise bias regarding the participating population groups, the survey was based on the Tailored Design Method (Dillman, Smyth, and Christian 2014).The persons to be interviewed were contacted three times, and the questionnaire was available online in German, English, and Turkish.1170 persons could not be reached (presumably because of outdated entries in the population register).2466 questionnaires were sent back, 2082 of which could be analysed (incomplete questionnaires and questionnaires from persons who did not correspond to the basic population were excluded).In reference to the adjusted sample of 8759 cases, the study thus arrives at a utilisation rate of 23.8%.This sample has been checked with population statistics for the Berlin inner city and has proven largely representative. 2

Operationalisation of direct displacement
In line with the definition of displacement as a move that occurred because of external changes affecting the tenancy (see above), we assigned the respondents to two groups on the basis of the reasons they gave for moving: displaced and nondisplaced.We did this by determining whether or not the reasons the respondents gave can be considered as factors of displacement.We defined seven potentially relocation-inducing circumstances as displacement factors: (1) structural upgrading (modernisation and energetic rehabilitation), (2) disturbances of living due to construction noise, scaffolding and/or flat inspections, (3) sales of houses or flats, (4) rent increase (net rent or/and operating/ancillary costs), (5) lease terminations (that were not due to tenant misconduct), ( 6) landlord pressure, and (7) structural decay.In the case of three displacement factors-structural upgrading, rent increase, and sale-we regarded not only the actual presence of the relevant circumstance but also its announcement by the landlord, a potential factor of displacement because it can demonstrably make people quit their home.The seven displacement factors correspond to the definition of direct displacement introduced earlier in that they concern the issue of tenancy, imply a change in a preexisting situation (e.g. the absolute rent level, which was already too high for those affected without an increase, is not a change and thus not a displacement factor), and are not controllable by those affected (e.g. a termination due to rent arrears is not assessed as displacement relevant).Whether the displacement factors we considered include all relevant circumstances that could induce displacement is debatable.In other studies, for example, the absolute rent level is also considered a displacement factor (e.g.Freeman and Braconi 2002, 2;Newman and Wyly 2006, 30), which significantly increases the number of moves to be classified as displacement in those studies.
To find out whether one or more of the displacement factors mentioned above prompted the respondents to move, we applied a three-step research design: Step 1: The interviewed person was asked which of the displacement factors (and which other factors potentially triggering relocation) they had experienced in the time before their move.
Step 2: The respondents were able to select, from the items they had ticked, those that were actually decisive for moving.In addition, they had the opportunity to supplement this selection with other reasons for their relocation; these free-text mentions were then summarised by inductive categorisation and checked to see whether they corresponded to the definition of displacement.It turned out that the numerous freetext mentions could all be assigned to the seven predefined displacement factors.
Step 3: respondents were asked to indicate their three most important reasons for moving (if they had mentioned more than three reasons) and thus the factors that had significantly contributed to their decision to move.Only respondents who named at least one displacement factor among these three major reasons for moving were classified as displaced; all others were classified as not displaced.

Quantifying direct displacement
Fifty-four of the 2082 questionnaires in the sample contained insufficient information to establish whether the interviewed person had been displaced or not.Of the remaining 2028 respondents, 53 (2.6%) moved only because of displacement factors.In contrast, 1606 respondents (79.2%) did not mention any displacement factor as a reason for moving.369 respondents (18.2%) cited both displacement factors and other reasons for moving; 260 of them (12.8%)indicated at least one displacement factor among their three most important reasons to move and were thus classified as displaced.In total, 313 of the 2028 moves covered in the questionnaires were classified as (direct) displacement, which corresponds to a share of 15.4%.
When interpreting this rate of direct displacement that was determined for a large part of the inner city of Berlin, it is important to bear in mind that it is based on a comparatively strict operationalisation-not least in relation to other empirical studies of displacement.In other words, the presented figure is likely to underestimate, rather than overestimate, the extent of the problem of direct displacement.Instead of operationalising displacement by asking for the three most important reasons for moving only, it would also make sense to consider all reasons for moving (because even the most marginal deterioration in the individual housing situation can increase the problem pressure to such an extent that it is precisely this change that is decisive for a move).In our study, this alternative approach would yield a rate of direct displacement of almost 21%.However, this approach has the drawback of somewhat blurring the difference between displaced and nondisplaced persons.

Major and minor forces of direct displacement
All in all, the 313 displaced respondents we identified in our survey named displacement factors 531 times as a reason for moving.Figure 1 shows the share of the seven displacement factors introduced above of these mentions.
With regard to the effectiveness of these factors (in terms of actual displacement), we observed the following: 38% of the displaced respondents indicated a rent increase as a significant reason for moving, which makes soaring rents the most important displacement factor.Of these 118 respondents, almost one-third moved after the announcement of a rent increase, and the other two-thirds moved after the rent increase was implemented.Structural decay proved to be the second most substantial displacement factor, with about 30% of those displaced citing it as a reason for moving.The third most important displacement factor was termination of the rent contract, which was experienced by more than one-fifth of those displaced.In an international comparison, the legal means of protection of tenants against termination are relatively elaborate in Germany.Basically, it is possible for a landlord to terminate a rent contract without much further ado only if he reclaims the flat for his own (or his family's) use.Accordingly, an eviction for the landlord's personal use was by far the most frequent type of termination observed in our study.Furthermore, some 15% of the displaced respondents had been pressured to move out by their former landlord.They pointed to, for example, 'insults' , 'being asked to move out multiple times' , or 'staggering discussions with their landlord' .In some cases, they were also subjected to physical pressure in the sense of Marcuse (1985), for example, by 'removing the gas pipes' or 'turning off the water and electricity' , which almost inevitably also triggers emotional stress.A little less than 15% of the displaced respondents identified the sale of the property they occupied to a new owner as a reason for moving.Disturbances of living due to construction noise, scaffolding, and/or flat inspections were listed as a reason for moving for just under 12% of those displaced.This displacement factor is usually either a side effect of refurbishment or modernisation or a consequence of a real estate transaction.Beran and Nuissl (2019, 127).
Renovation or modernisation measures were a reason to move for 11.5% of the displaced respondents.
In addition to displacement factors, 260 displaced respondents also cited other reasons for their decision to move, for example, the desire for another flat (mentioned by 41% of them) or a disturbance of the peace (18%).Although these issues do not represent genuine displacement factors, they can certainly be related to them (e.g.refurbishment, modernisation).Moreover, the observation that people who move often give both displacement factors and other circumstances as a reason for their move is not surprising from a decisiontheoretical perspective and must not be used to discard the operationalisation of displacement via individual reasons for moving, as is sometimes argued (Freeman, Cassola, and Cai 2016, 6).Instead, this result underlines that displacement rarely occurs as a single-cause phenomenon.Empirical research on displacement therefore must take into account the complexity or multicausality of decisions that are important in real life.
While on the one hand, the displaced respondents usually mentioned further issues, in addition to displacement factors, as reasons for their move, on the other hand they were often confronted with even more displacement factors that they did not identify as reasons for moving.On average, they named 3.3 displacement factors they were confronted with before their move, but listed only about half of these as reasons for moving.For example, only two-thirds of the 175 displaced persons whose rent had been increased, or who had been told that the rent was to be increased, had moved for this reason.Unsurprisingly, termination is the displacement factor that was most reliably perceived as a reason to move: This was the case for 92.2% of the 77 displaced respondents who had their previous tenancy agreement terminated.These findings indicate that most cases of direct displacement are not due to a single, independent cause (e.g. a rent increase or the sale of the flat) but result from the occurrence of several displacement factors.These factors interact with each other and add up until together they exert a displacement pressure intense enough to prompt a move.However, how strong this pressure has to be to actually induce a displacement-related move certainly depends to a large extent on whether a tenant has viable alternatives, which again very much depends on their social, cultural, and economic capital (Häußermann 2007, 175f.).A correlation analysis of the occurrence of displacement factors in our survey data supported this assumption and revealed two different patterns of displacement (see Table 3).
No significant positive correlation can be established between the occurrence of terminations or structural decay and other displacement factors.Thus, we can assert that these two factors typically constitute a simple form of displacement whereby the tenants' move can be attributed to only one factor.In contrast, complex forms of displacement are due to multiple displacement factors.Two types of complex displacement become apparent from the correlations in Table 3, each of which includes structural upgrades, disturbances of living, and pressure from the landlord: One has a rent increase at its heart, the other the sale of the flat or house.About 95% of the displaced respondents for whom either a rent increase or the sale of their flat or house was among the three most important reasons for moving cited at least one more displacement factor as a reason.We thus assume that there is a typical pattern of displacement that often starts with the sale of a house.The new owner then takes various measures to increase the return on their investment, such as modernising and/or energetic rehabilitation, which again lead to disturbances of living (i.e. a physical and psychological pressure to move out).Upon completion of reconstruction measures, the rent increases.Tenants react quite differently to the sketched situation and decide to move at different points in time-if they decide to move at all.

Displacement, gentrification, and policy
The approach to investigating displacement we have presented here focuses on the displacement of tenants rather than the gentrification of neighbourhoods.While many studies determine gentrification areas before deriving statements on displacement, we conceptualised and measured displacement as an independent issue.With this approach, we showed that displacement due to different displacement factors-among them many factors that are related to real estate upgrading (e.g.structural upgrading, rent increase)-is a highly relevant phenomenon in the tight Berlin housing market: At least 15% of the people who moved-and this is a conservative calculation-were directly displaced.This finding can be used to support tenant initiatives in their demand to policymakers and planners to continue in their efforts to ease the housing market and prevent displacement.It can also be used as the starting point of a spatially explicit analysis that looks for correlations between the occurrence of displacement and area-specific upgrading (i.e.gentrification) processes.
Our results also hint at ways in which displacement can be avoided.In this respect, it seems key to prevent rent increases.Policy instruments such as the rent cap recently imposed by the Berlin senate appear to be a sensible tool (even though, as mentioned, Berlin had to withdraw the law).In addition, it would be reasonable to skip, or at least limit the possibility of, allocating modernisation costs to the rent.A second starting point is the extension of protection against contract termination.In particular, termination of tenancy by the landlord reclaiming the house for his own use proved to be a frequent trigger for displacement.It is almost only private owners who propel this kind of displacement.Thus far, even profound measures to prevent displacement by counteracting the financialisation of the housing market, such as the expropriation of the housing stock of companies owning more than 3000 flats, as is currently being discussed in Berlin, will certainly not make displacement disappear entirely, even though it stands to reason that it will substantially decrease displacement pressure.It appears that keeping an eye on the housing stock in disperse ownership also is necessary.
Early warning systems on displacement could also help, in planning and political practice, to implement housing policies against displacement in good time.An exciting and demanding challenge can therefore be seen in further developing the existing approaches of secondary analytical monitoring of displacement phenomena (including displacement pressure) with the help of primary surveys such as the one we have presented here.In addition, longitudinal qualitative studies that follow the typical courses of 'displacement careers' and identify the effective factors can further help calibrate the envisaged monitoring system.A displacement monitoring system developed in this way should be able to estimate the actual strength of the interactions between gentrification and displacement-which, of course, presupposes that gentrification is conceptualised just as precisely as displacement.
Apart from the suggestions that can be derived from our study with regard to urban and housing policies, we hope that our approach also contributes to the methodological debate about how to tackle displacement (which literally is a 'moving' research target that is always difficult to study) in empirical research.Although our definition and operationalisation of displacement, which focuses on the reasons to move out of a flat (while neglecting what those moving feel in regard to their move), might be questioned for its scientific tenor (as well as its lack of empathy), it has the great advantage of allowing an unambiguous classification of empirical cases.This is made possible by the strictly individualistic conceptualisation of displacement which, on the one hand, leans on the benefits rational-choice theory has in terms of clarity but, on the other hand, does not dismiss the complexity of relocation decisions altogether because it takes into account the multiple causes inherent to these decisions by providing the possibility of naming a variety of reasons for moving when it comes to studying displacement empirically.While other approaches to the empirical research of displacement indisputably have their merits-for instance, as far as political economy approaches are concerned, with respect to the consideration of political issues or, as far as secondary analysis and modelling based on existing aggregate data is concerned, with respect to research efficiency-we hope our study provides some inspiration for further empirical endeavours in research on housing issues.

Figure 1 :
Figure 1: Importance of displacement factors as a reason for moving.Source: Slightly modified from Beran and Nuissl (2019, 127).

Table 1 :
Dimensions and specifications by means of which it is possible to distinguish different definitions of the term displacement (Note: Table can only be read horizontally).

Table 2 :
Dimensions of the definition and empirical operationalisation of the term displacement in this article.