The symbolic power of policy instruments

(1)

The symbolic power of policy instruments

Master thesis

Jelke Bosma

Student number: 6044654

Email: jelkebosma@gmail.com

University of Amsterdam

Research Master Urban Studies

Supervisors: dr. J.L. Uitermark & dr. C. Hochstenbach

Second reader: prof. dr. E.R. Engelen

(2)

Acknowledgements

First of all, I would like to thank my supervisors Justus Uitermark and Cody Hochstenbach for their encouragement and feedback, and especially their flex-ibility and willingness to read my work on a short notice. Moreover, Justus’ work was important in sparking my interest in (urban) sociology. From both Cody’s work as well as the courses in which he thought I have learned a lot about housing.

I also found great support, an excellent study rhythm, and illuminating discussions in the P.C. Hoofthuis with my study buddies Lieven Heeremans and Bram Harkema. The countless hours I spent in the library would have been more boring and less interesting without your company. A special thanks to Lieven as well for reading and discussing some earlier drafts of my thesis. This might also be a good place to express my contentment about my friends putting up my tent at Lowlands Festival at the moment of writing as well as Marco for inviting me as his “roadie” this weekend (I’ll be done soon — but really!)

My parents’ loving stimulation and financial support over the past eight years have allowed me to follow the not-so-straightforward study path. Their home repeatedly offered me a welcome respite from Amsterdam when I needed some extra focus and less distraction. For all of this and more, I am extremely grateful.

Lastly, without Dianne’s confidence in me and endless support I would probably not have been able to complete this thesis. I am looking forward to spending more weekends together the coming months, without some distant stress creeping on me because of some deadline I am worried about.

(3)

Preface

This research project has followed a rather long and windy road. I initially set off, somewhat hesitatingly, with an inquiry into the effects of the finan-cial crisis on finanfinan-cialisation of sofinan-cial housing in Amsterdam. Later, I added a comparative element by including Berlin as well. During my fieldwork there, I realised how common notions, problems and ideas that were central to the housing market in Amsterdam did not exist or were understood in a different way. Obviously, in different cities one will find different problems. However, over time I kept on wondering why some notions that appeared central to the debate on housing in Berlin were absent in Amsterdam — and vice versa. In the media in Berlin rents are usually expressed in the form of the Nettokaltmi-ete, i.e. the rent per square meter without heating. In Amsterdam, contrast-ingly, rents are commonly expressed in total rent per month, with or without heating and other costs.

Simultaneously, I got intrigued by the techniques housing providers use to manage their stock and set out their strategy. Portfolio management was one of these techniques that was repeatedly mentioned by respondents. I will not bother the reader here with the details of these techniques, which also did not end up in this thesis, but they made me wonder how one’s home, an actual space where someone lives, gets translated to something that is susceptible for such techniques. Oxford Dictionary of English explains the ‘portfolio’ comes from Italian portafogli, which in turn is a combination of portare ‘carry’ and foglio ‘leaf, sheet’. Thus, how is a dwelling made portable? In looking for answers for these sets of questions I repeatedly stumbled upon the systems that are used to regulate rent levels, which eventually led my shift of focus in their direction.

The process of writing this thesis was not always easy, with hindsight maybe because of a lack of discipline — that does not mean a I did not to put enough time or effort in it (which I hope shows in the end result), it refers to my lack of a singular disciplinary focus. Mirroring my own course of study the past eight years, this project I think shows some philosophical aspects and some hints of questions of planning. It shows affinity with the fields of ge-ography, and has its roots in urban studies. And lastly, it partially takes a sociological approach. It also made it a bit more complicated sometimes to decide what to focus on. In the end, given the extensive attention to theory, I also decided to use some more words than the article I originally planned to write would have contained.

What I wanted to do with this thesis works on several levels. First, there is quite a practical level: the housing market and cities are becoming increas-ingly inaccessible for lots of people. City dwellers who do own a home have to spend more and more on it while they live under increasingly precarious or ‘flexible’ conditions. Second and related, I am interested in the ways govern-ments govern. Why does the state have the authority to make collective deci-sions and how does it doe so? Last and third, I have been looking to connect abstract theories with empirical research. All of this made this an interesting project for me to work on; which I hope I can transfer to the reader.

(4)

1 Introduction

In February 2017, during a discussion on the Rent Index in the Urban devel-opment and Housing committee of the Berliner parliament, a representative of private landlords in Berlin explained his opinion about the Rent Index:

The construction of the Rent Index is an endless chain of compromises. . . . But compromises are the enemy of science! In science there are no com-promises, it is where truth is sought, where yes or no are sought, and that’s why I think there can be no scientific, no qualified Rent Index.

(Berlin House of Representatives 2017: 10)

Without going into details of the context, this quote brings us to the core of the construction of a policy instrument, which involves a search for the truth in a context of compromises. This thesis explores the construction, controversies and consequences of policy instruments for rent regulation.

Policy instruments for rent regulation play an important role in the simul-taneous processes of gathering knowledge and making policy on housing. Due to the way they are constructed and the particular way they render hous-ing and the houshous-ing market, rent control instruments exert a specific form of symbolic power over the field of housing. Even though the construction and calculation process for these instruments might be debated, they limit the possibilities for political debate on more fundamental aspects of housing by cloaking the power relations inherent to the housing system. Thus, this article argues that policy research should pay specific attention to the policy instru-ments that are used in constructing policy objects. It does so on the basis of a comparative case study of policy instruments in the context of the hous-ing markets in Amsterdam and Berlin, respectively focushous-ing on the Houshous-ing Evaluation System Woningwaarderingsstelsel and Rent Index Mietspiegel. These instruments become central to the understanding of housing by both policy makers as well as the general public. Whereas they originally are intended to serve a specific goal, they partly start to live a life on their own. As such, they influence the strategic practices of housing providers and guide the imagina-tion of the future development of housing and housing policy.

Housing has an obvious material aspect; a dwelling is a physical object. Thinking or imagining a single dwelling is relatively easy — some chambers, an entrance, windows and a kitchen. An apartment building or group of dwellings is still graspable, as a multiplication of the aforementioned, joined together by galleries, an elevator, and staircases. But above that scale, let’s say at the level of neighbourhoods, thinking housing becomes more problem-atic. At this level, housing tends to take shape in the form of indicators, such as ‘share of a type of housing’, which require categorisation, statistical de-scription. This becomes even more important when one tries to think about housing on the scale of a city or a housing ‘portfolio’, which is the stock a housing company owns. It requires some kind of abstraction to be able to say something about housing at this level, as an aggregated object.

The state, leaving the definition of this relative elusive concept open for now, has to gather such abstracted knowledge of this type on the objects of its policy, as to be able to govern them. To be able to make something into an ob-ject of government it first has to be stabilised and rendered mobile. Inspired by

(6)

Bruno Latour’s description of ‘action at a distance’, Peter Miller and Nikolas Rose conceptualise this as central steps required in governing objects that are not within immediate reach (2008, also see Uitermark 2005). To be able to in-stall housing policy, for example, one needs stable knowledge or information about housing at a neighbourhood, city or national level. Ultimately, policies are purposed to have an effect on the material reality of housing. However, before it can do so, its object has to be defined, assembled, or constructed. There is no singular, concrete thing called housing out there that policies can aim at. Thus knowledge has to be collected, compiled, and made sense of.

The knowledge-producing and sense-making aspect of the policy making process shows clear commonalities with the production of knowledge in sci-ence. Here, conceptualisation plays an important role:

[A]ny conceptualisation involves a reduction . . . of the fullness of empiri-cal reality. This is a somewhat pragmatic observation (a concept per thing would be a clumsy intellectual world), but also is productive of building

shared understandings. (Robinson 2016: 306)

As empirical reality consists of a myriad of objects, details, and aspects, scien-tists abstract from particular impressions and form concepts to cover repeat-ing impressions, processes, and mechanisms. Understandrepeat-ing of reality comes into being through repeated testing, validation and use of concepts. To collec-tively construct new scientific knowledge some common ground is necessary first. Similarly, policy makers cannot do without a shared understanding of what their object is. In the case of housing, policy makers need a shared un-derstanding of what the housing market looks like and what problems are to be solved.

Common understandings of the policy process hold that first a problem is defined, possible solutions are sought, which are then implemented in poli-cies, aiming to ease or solve the problem. In designing the polipoli-cies, instru-ments and measures are designed to make the problem at hand legible, to be able to steer policy interventions, and to measure their effects. As such, these instruments simultaneously serve the goals of a policy as well as produce knowledge on the problem at hand. Rent regulation instruments, such as the Woningwaarderingsstelsel in Amsterdam or the Mietspiegel in Berlin, function in a similar manner. They form the foundation for rent control measures (re-lating to housing as a policy-object) and at the same time render housing in a particular way, construct knowledge about it, and form a system for building shared understandings. However, this process works the other way around as well. The available policy instruments, as derived from existing policies, strongly influence the thinkable, (acceptable, reasonable, perceivable, poten-tial) policy responses at a given moment. The ways the policy object is known and quantified determines the conceivable potential policy interventions.

The policy instruments are not neutral. Their construction, as I will show, involves a struggle between various actors, in which they try to shape the instrument in a manner that benefits their interests. When this process is com-pleted and the instrument becomes codified in policies or laws, these struggles and the power relations inherent to them become cloaked, or at least partly hidden from view. In other words, the institutional structures that arise from

(7)

a particular set of policy interventions are over time accepted as given and become largely uncontested.

(8)

2 Theoretical framework and methods

In the following chapter I set out the theoretical approach in this thesis. The first section sets a Bourdieusian understanding of social reality, with a specific focus on the role of the state. The next section provides a concise overview of more theoretical literature on statistics. By confronting this ontological frame-work and understanding of statistics with insights from the field of Science and Technology Studies, I theorise how policy instruments might affect so-cial reality. I end this chapter with an overview of recent literature in Urban Studies that resonates with this framework.

2.1 Reality as double

Departing from a Bourdieusian perspective in the approach of my research object, I understand social reality as neither fully objective nor subjective, but existing in a double sense. Bourdieu understood material world as consti-tuting the “objectivity of the first order” (1992: 7), where physical structures have their place. The corresponding mental structures constitute “objectivity of the second order” (Bourdieu & Wacquant 1992: 7). Even though the latter are mental, they are just as important as these objectivities mutually construct each other. The distribution of resources – economic capital, but importantly also other types of capital, especially cultural capital – over social beings in social reality, and thus their position in the social system determines the way they interpret the world. Simultaneously, the mental categories of percep-tion and classificapercep-tion make and constitute the objective social world. Thus, Bourdieu combines structuralist and phenomenological perspectives on social reality. Subjective, internal agency and objective, external structures exist in a dialectical relationship. To fully understand social reality we should look both at the material reality and the way in which social actors make sense of this reality.

The subjective, individual element of social reality Bourdieu defines with the concept of habitus. This is the ‘system of dispositions social agents acquire over time, through an implicit or explicit learning process, that functions as a system of generative schemes’ (Bourdieu 1989 [1976]: 177). The habitus functions as a set of ‘symbolic templates for the practical activities—conduct, thoughts, feelings, and judgments—of social agents’ (Bourdieu & Wacquant 1992: 11). These templates or dispositions are accumulated over time through the social interaction of an individual with others. An agent’s habitus thus brings his or her history into play at the present and constitutes the link be-tween the collective and the individual.

The collective, external side of the equation that constitutes social reality Bourdieu conceptualises in the form of fields. A field is ‘a network, or a con-figuration, of objective relations between positions’ (Bourdieu & Wacquant 1992: 97). These are ‘buried’ or hidden meso-level structures through which a wide range of parts of society are organised by determining a space of possi-ble actions (Bourdieu & Wacquant 1992). Within a society multiple fields exist simultaneously and often partly overlapping; the fields of art, science,

(9)

econ-omy for example. The positions in a field are taken by the agents, depending on the capital an agent has. Besides economic capital, Bourdieu recognises several other types such as cultural, social, and symbolic capital. The latter provides a player with the capability to make sense of the field in their own terms.

Objective structures such as fields are reproduced by the strategies and practices an agent develops on the basis of his/her habitus. Thus, habitus is one of the mechanisms that ensures the reproduction of fields:

Habitus is a structuring mechanism that operates from within agents, though it is neither strictly individual nor in itself fully determinative of

conduct. (Bourdieu & Wacquant 1992: 18)

On the other hand, an agent’s involvement in various fields structures his habitus. Thus, a dialectical relationship exists between habitus as an internal structure and fields as external social structures. This relationality is funda-mental in Bourdieu’s sociology. Instead of granting either subject or object, agency or structure, ontological priority, he insists on the‘primacy of relations’ (Bourdieu & Wacquant 1992: 15). This appears in both the concept of fields -where positions are always defined in relation to other positions, in relation to the field - as well as in habitus, which is determined by the relations one accumulates over time. Instead of fields, objects or agents having an intrinsic identity or meaning, this can only be constructed of constituted through the interplay and connectivity with other elements.

So far, we have a social space composed of various fields. Within these fields, agents have a position determined by the capital they possess. Mirror-ing their social position, agents develop their habitus, the ‘mental structures through which they apprehend the social world’ and understand their own place therein (Bourdieu 1989: 18). Classical categories through which social reality is understood are for example class, gender and race. The formation of such categories is effected through the use of symbolic power, ‘the power to impose and inculcate systems of classification that effect the naturalization of structures of domination’ (Bourdieu & Wacquant 1992: 13). These cate-gories are thus not natural, not self-evident, but shaped through relations of power. The production and reification of the categories and classifications is part of the struggle in the field and helps to stabilise it, benefitting the dom-inant agents. Thus, ‘symbolic systems are not simply instruments of knowl-edge, they are also instruments of domination’ (Bourdieu & Wacquant 1992: 13). Classifying objects or people is an activity that always needs a subject, a classifier so to say. The classifier or the symbolic system cannot be disin-terested, i.e. he/she would not go against their own interests. In effect, the classification serves the need of the classifying subject.

The agents that are active in a field take part in a struggle on a central stake and constantly try to improve their position. By taking part in this struggle or ‘game’, agents agree to the doxa or ‘rules of the game’. Doxa is ‘that which is beyond question’ (Bourdieu 1994: 163) in a field, a set of common beliefs and views that are taken for granted by the players. These are not fixed, however. Players’ strategies in the game can be put on a spectrum ranging from ortho-dox to heteroortho-dox. Orthoortho-dox strategies are generally pursued by the dominant

(10)

players in the game, who try to maintain the doxa as it is and thereby main-tain their position. Contrastingly, heterodox strategies try to undermine the prevailing doxa and change the rules of the game. Doing so, the specific cap-ital that is recognised in the field changes in favour of the dominated player who thus improves his position.

The state-field The state performs a special role with regards to the

struc-tures of understanding of social reality. It exert’s power not just over single field, but is involved in all fields that society is composed of. It does so in various ways. First, in Weber’s classical definition of the state it is primar-ily the holder of the legitimate monopoly of physical violence (1999 [1919]). Secondly, as Jessop (1990) points out, Weber recognised that the state should regulate the relations between its citizens through legislation and the juridical system. Here however I want to shift attention to a more elusive aspect of gov-erning, namely the way the state influences the way we see and understand the world. Bourdieu thus plays on Weber’s classical definition when he con-ceives the state ‘as the holder of a monopoly over legitimate symbolic violence (Bourdieu & Wacquant 1991: 100, cited in Bourdieu & Wacquant 1992: 112). Symbolic violence might be understood as the symbolic power a state exer-cises over a nation1. For Bourdieu, importantly, symbolic power is accepted by the dominated, who is complicit (1992: 167). It appears as natural given that agent’s grew up in the social conditions, imprinting it in their habitus.

The state is not a coherent, contain ‘myriad social orders whose relations can be as conflictual and constraining as any other fields’ (Fligstein & McAdam 2011: 8). For Bourdieu, the state is just as well a field, with largely the same characteristics as other fields. Thus, it is not a hegemonic, unitary institution, but is the stage of internal struggles as well. These struggles have a double character.

First, struggles revolve around the question what particular projects and interests the state should focus on. Even though ideally the state serves the general will or common interest of society, its power and resources are not unlimited (Jessop 1990). Jessop coined the concept strategic selectivity to de-note this necessity of making choices, as it is impossible to define a univer-sal a general will: ‘the common interest or general will is always asymmetri-cal, marginalizing or defining some interests at the same time as it privileges other. There is never a general interest which embraces all possible particu-lar interests’ (1990: 342). Following Bob Jessop’s strategic-relational approach of the state, I assume it always has to select what projects and interests are privileged. As Uitermark (2005) argues this begs the question, first, how past processes and strategies have shaped the selectivity of the state, and secondly, how the state as a strategic terrain privileges some strategies over others.

Secondly, the agents that are active within the state follow their own terests and try to maintain or improve their position. State officials and in-stitutions will not just serve the general interest but will — acknowledged or

1_{“the power to constitute and to impose as universal and universally applicable within a}

given ‘nation,’ that is, within the boundaries of a given territory, a common set of coercive norms.” (Bourdieu & Wacquant 1992: 112)

(11)

unacknowledged — also try to guarantee their own position and existence. This adds a layer of complexity to the struggle within the state, in which it is not always straightforward whose interests are being served. Moreover, as Uitermark (2005) convincingly argues, the state is not centralised but operates on various scales. The central state or government, where policy is often for-mulated, requires cooperation from local actors and institutions to implement policies. In addition, the local functions as a site for the formulation of policy problems and experimentation with policy interventions. These various rela-tions show how various scales cannot do without each other, central and local aspects of (urban) policy are co-constitutive.

To finish this section, some attention should be given to the differences and correspondences between the state and politics. As Justus Uitermark writes in his analysis of the dynamics of power in the integration debate:

Politics does not consist exclusively of the ordering and processing of en-dogenous preferences but is also about the interpretation of reality, the demarcation of symbolic boundaries and the mobilization of sentiments.

(2012: 17)

Following this, it should be acknowledged the state is not the exclusive do-main of politics. Perspectives on and interpretations of reality are just as well shaped by discourse and debate outside of the formalised institutions of the state. The state however has a particular characteristic in that it has an in-tricate relation to society as a whole, as a collective. Here I follow Jessop in defining the state: ‘the core of the state apparatus comprises a distinct ensem-ble of institutions and organizations whose socially accepted function is to define and enforce collectively binding decisions on the members of a society in the name of their common interest or general will’ (Jessop 1990: 341).

The formulated approach to social reality and the role of the state have remained on a fairly abstract level up to this point. In the following section attention shifts to one of the practices that might be understood as a means of exercising symbolic power.

2.2 Knowledge, statistics and the state

If symbolic power allows the state ‘to impose and inculcate systems of classifi-cation’, how are these classifications created in practice? A branch of literature has long argued that statistics play an important role in this. In his 1977-1978 lecture series at the Collège de France Michel Foucault argued that the devel-opment of statistics marked a fundamental shift in the way states are ruled. In the preceding period this rested on the fundamental principle of sovereignty, the supreme and unassailable power of the sovereign king. This power was exercised over and applied to its objects as individuals. With the emergence of methods to collect, store, and calculate large amounts of data the relation between individuals and the collective altered. From the statistics appeared certain constants that did not apply to individual cases but to the population. This formed the condition of possibility for the ‘art of government’ (Foucault 2007: 104) that took the population as it’s object. The population emerged as a social entity with properties that were not reducible to the either the

(12)

individ-uals or households it was composed of. These properties remained constant over the years, making them recognised as ‘a set of processes to be managed at the level and on the basis of what is natural in these processses’ (Foucault 2007: 70, emphasis added). Within these processes a norm appears, denoting a group of subjects showing ‘normal’ behaviour or characteristics . Concomi-tantly, an ‘abnormal’ category emerges. Thus statistics, which originally had the meaning science of state (Foucault 2007: 101), allows the state to gather new types of knowledge on and classifications of its population.

Similar arguments appear in Ian Hacking’s The taming of chance (1990). The increasing availability of enumerated data on social phenomena showed that social laws existed, similar to natural laws. However, social laws were driven by chance rather than determinism, suggesting that deterministic un-derstandings of the world no longer held up. These came together with the idea of normalcy, shifting psychology its focus from human nature to ‘nor-mal people’. However, the categories of understanding surrounding what was deemed ‘normal’ were not a priori given. Rather, ‘categories had to be in-vented into which people could conveniently fall in order to be counted’ and Hacking concludes, the systematic collection of data as such ‘has profoundly transformed what we choose to do, who we try to be, and what we think of ourselves’ (1990: 3).

A recent report by the Netherlands Scientific Council for Government Pol-icy (WRR) provides a recent critique of the classificatory aspects of polPol-icy making (Bovens et al. 2016). The report argues in favour of replacing the categories ‘autochtonen’ and ‘allochtonen’, which have their roots in sociol-ogy, were popularised through the use in policy documents and statistical data, but have since then become common in public discourse. The reduc-tionist quality of such terms was criticised, as well as their persistence even when their original reference (in this case: not being dutch) does not apply anymore. Classification and quantification involves a transformation of peo-ple, it ‘means studying people in classes, abstracting away their individuality’ (Porter 1995: 77). This is not intrinsically bad, but one should be aware of this necessary condition of statistics.

Bourdieu and instruments Bourdieu has spent only limited attention to the

importance of instruments, technologies or ‘things’ in power struggles. Ex-cept for his comments on the power of the passport, he has limited attention for how instruments might exert symbolic power over a field. As Schinkel (2007) notes, even relatively material things with high symbolical meaning such as pieces of art are treated by Bourdieu as relatively meaningless in them-selves. Instead he regards them as mere reflections of the structural position of artists in the art field and the struggles to maintain or improve that position.

I would argue though that the (policy) instruments that are used in a field exert a strong symbolic power. Similar to law, a policy instrument ‘freezes a certain state of power relations’ by fixing it (Bourdieu 1984: 480). Moreover, as I will show in later sections, policy instruments describe reality - they mea-sure, quantify, and categorise - and doing so simultaneously create or make up understandings of reality. Like statistics involve a kind of ‘making up

(13)

peo-ple’ (Hacking 1990: 3), have a ‘creative power’ (Porter 1995: 37), and metrics are productive (Beer 2015), policy instruments might form the foundation of the classifications we apply to the world.

2.3 Policy instruments

Recently there has been increasing interest urban studies in the politics and construction of instruments. This branch of literature looks at how and why such instruments are designed and used in urban politics and policy. Wilde & Franssen (2016), for example, trace back how the NAP, an index used in policy making for deprived neighbourhoods, has been constructed and subsequently used in practiced. They show how the index uses semiotic, statistical and visual techniques to produce ‘hard data’ about neighbourhoods, i.e. quantitative data that suggests to be objective and incontestable. As such it transforms the qualification ‘deprived’ to a quantitative measure, which strips its of a set of problematic arrangements that ‘deprivation’ connotes.

Similarly, Uitermark et al. (2017) explore the political life of the Liveability-o-meter, a statistical instrument used to measure ‘liveability’ in Dutch neigh-bourhoods. They show how the instrument is either ignored or used as a ground for the instalment of exceptional policy measures (excluding popula-tions from disadvantaged neighbourhoods) that are part of the Rotterdamwet. In some cases they are employed as objective basis for specific controversial measures. Here the statistics are argued to speak for themselves, their extraor-dinary values are argued to show that exceptional measures are necessary in targeted neighbourhoods. In other cases however, the importance of statis-tics is downplayed. In these cases it is argued ‘reality on the ground’ is more complicated and not fully reflected in the statistics. As such, statistics are em-ployed a inconsequent and politically opportunistic manner. Thus, statistics turn out not have a universalising, depoliticising effect as their incontestable truths are regarded as they speak for themselves. Instead, the way statistics are embedded in a particular political context determines what their effect is.

Both articles show the importance of approaching public policy not only by looking at organisation, actors and representations, but also take instru-mentation into account, as Lascoumes & Le Galès (2007) call for. As they ar-gue, instruments perform specific functions in public policy that are often not accounted for. By implicitly theorising the relation between governing and governed parties, they embody specific representations of the issue at hand. Moreover, they argue that instruments are not neutral, ‘they produce spe-cific effects, independently of the object pursued (the aims ascribed to them), which structure public policy according to their own logic’ (Lascoumes & Le Galès 2007: 3).

The effect of instruments on power and inequality By focusing on the

policy instruments into which power relations tend to condense and around which public policy subsequently evolves, the paper shows how these instru-ments tend to conceal the interests served and the political decisions that have produced them by presenting the output of instrument in a quantified,

(14)

uni-form manner. However, the knowledge produced by these regulatory instru-ments does not ‘mirror’ some objective, unquestionable reality, but a continu-ously reproduced set of aims, interests, and social relations. Thus, they tend to reinforce spatial inequalities, as well as existent inequal power relations that fed into them during the construction of these instruments.

2.4 Epistemic objects and cultures

Even though science is the primary knowledge-producing institution, as Knorr Cetina explains her choice of subject, the production of knowledge is not lim-ited to science alone. Her own work has dealt intensively with the epistemic cultures and practices in financial trading (Knorr Cetina & Bruegger 2002) and economic transactions (Knorr Cetina & Preda 2001). Heath (2012) uses the concept of epistemic cultures to analyse how public policy for marriage pro-motion became accepted through the use and propro-motion of particular sources of knowledge by proponents of such policies.

Epistemic cultures are ‘cultures of creating and warranting knowledge’ (Knorr Cetina & Reichmann 2015). The concept emerged from the field of science and technology studies, which has its roots in anthropology and so-ciology of science. Instead of focusing on the contents of the output of sci-ence, scholars such as Bruno Latour, Werner Reichmann and Karin Knorr-Cetina shifted attention to the practice of doing science. This meant they went into the laboratories to see what scientists do to be able to produce scientific knowledge. They found that the practices of fact finding, or rather fact con-struction, diverged widely in different scientific fields. In Epistemic cultures: How the sciences make knowledge (2009) Karin Knorr Cetina developed the no-tion of epistemic cultures to explain these divergences. A more detailed def-inition describes epistemic cultures as ‘those amalgams of arrangements and mechanisms-bonded through affinity, necessity, and historical co-incidence which, in a given field, make up how we know what we know’ (Knorr Cetina 2009: 1). Knowledge does not arise spontaneously, it does not emerge from objects simply by looking at them, and is not recorded automatically. Instead, it requires the combination of various structures, cultures and instruments to produce knowledge and validate it, i.e. the construction of relations between them. In the construction of knowledge not only epistemic cultures are rel-evant, the instruments, things, and techniques that are used to do so play a significant role as well. A clear example of the importance of instruments is the Large Hadron Collider, which has been indispensable in recent research in the natural sciences.

Similarly, policy objects are not just ‘out there’. One cannot start from scratch and design, let’s say, housing policy by looking at ‘the housing market’ and then designing something to change it. Instead, policy objects in general and the housing market in specific are what (Knorr Cetina & Bruegger 2002) calls epistemic objects or knowledge constructs. Designing policies to deal with such objects are ‘epistemic embedded’, they refer to a reality that is ‘purpose-fully assembled and unfolded by professional knowledge workers and whole technological systems which provide the frames of reference and the means for experience and transactions to take place’ (Knorr Cetina & Preda 2001:

(15)

30). The nature of the objects of research makes that they cannot become ‘real’ without the agents that work on them:

The defining characteristic of this kind of object [the epistemic object], from a theoretical point of view, is its lack of “objectivity” and complete-ness of being, and its non-identity with itself. Since objects of knowledge are always in the process of being materially defined, they continually acquire new properties and change the ones they have.

We can now turn back to markets as these are confronted by traders, and ask whether the characterization just given of knowledge objects applies to them? The answer will be yes; traders are engaged in a process of continually defining the market not only in the sense of trying to read and understand it, but also in the sense or “making” or articulating it, by testing it, moving it, and manipulating it.

(Knorr Cetina & Bruegger 2000: 149)

This begs the question how policy makers construct the object their policies aim at. How do they assemble and unfold the housing market?

2.5 Methods

Departing from the theoretical approach outlined in the previous section, this chapter sets out my methodological approach. First, I define the main research question. Next, I describe the methods I use to answer these questions. Third, I describe the case selection process and shortly introduce both cases.

This thesis deals with the relation between the housing market and policy instruments for rent regulation. These instruments maybe obviously aim to regulate rents, but here I try to expand the perspective and look beyond this observation. What I try to do in this thesis might be what Schinkel has called a sociological Von Münchhausen-manoeuvre, which requires the inquirer ‘to be a border inhabitant, to be schizophrenic ‘within’ and ‘outside’ the social’ (2008: 455). While my research does not take the social as a whole as an object, a par-allel could be made with housing. This study tries to go beyond the first-order observations of the field of housing, referring to the way the field describes itself, and aims at second-order observations, which take the common under-standings of the field as its object and asks what the conditions of possibility are for observing the field in such a way.

The central question in this article is How are policy instruments in the hous-ing market constructed and how do policy instruments construct the houshous-ing market? Thus, I assume a dual relationship between the housing market and these in-struments.

I will answer this question on the basis of empirical data from a compar-ative case study of the housing market in Amsterdam and Berlin. To do so, several steps are taken. First, I set out how bourdieusian field theory and a constructivist focus on epistemology and instruments might be combined. This has been explored above in the theoretical chapter. Next, for both Am-sterdam and Berlin the origins of rent regulation instruments – respectively the Woningwaarderingsstelsel and Mietspiegel – are described and their devel-opment over time is described. A focus is here on how on the one hand sup-porting coalitions are formed around these instruments, and on the other how

(16)

and by whom they are contested. Next, the current uses, contestation, and interpretations of these instruments are described. I embed this in an explo-ration of the wider epistemic culture within the field of social housing. Finally, the development and use of policy instruments in both cases is compared and used to explore how they influence the political debate on social housing in both cities.

The empirical contribution of this article is a story of how rent regula-tion instruments came into existence, developed over time, and the ways they are currently used and interpreted. The theoretical contributions are twofold. First, it combines two theoretical strands of literature in a novel way. A bour-dieusian approach to the power relations in (policy) fields and the state is combined with a constructivist understanding of the epistemic aspects of pol-icy making, derived from the field of Science and Technology Studies. Com-bining these approaches allows for the explanation of how power relations are obfuscated in and through instruments. Secondly, it brings into the field of policy studies questions of how policy objects are constructed and the role knowledge plays in this.

Two characteristics of the Amsterdam and Berlin housing market make them particularly well suited as a case. First, in both Amsterdam and Berlin, historically as well as contemporarily, there is a relatively strong state involve-ment in the housing market. In both cases, a share of the stock is owned by or is closely related to the state (see table 1). In Berlin, the municipality owns about 14 percent of the stock. In Amsterdam, around 45 percent of the stock is owned by housing associations, which are bound by special state regula-tions. Second, the regulation of rents tends to become more urgent in situa-tions where the market is under pressure and rents tend to rise. As such, I ex-pect the limits of these instruments to become apparent, prompting increased discussion about their suitability. The pressure on the housing market is ap-parent in various indicators. In Amsterdam over the past five years the aver-age price of homes has increased with more than 45 percent, frome275 509 in the second quarter of 2012 toe402 067 in the second quarter of 2017 (CBS, 2017). Regarding social housing, starters who for the first time move into so-cial housing have on average been registered for 9.7 years (AFWC 2017). Sim-ilarly, in Berlin rents and housing prices have been increasing rapidly over the past decade (CBRE & Berlin Hyp 2016).

Some attention should be given here to scale. The Housing Valuation Sys-tem covers the whole of the Netherlands and is defined at the national level. The city of Berlin has it’s own Mietspiegel, designed by the local authorities using national-federal guidelines. The policy documents studied in this case derive from the state level. The municipality and the State of Berlin coincide, gives extra weight to the local level in this case. Contrastingly, there are no local policy documents on the design of the WWS. Instead, the policy docu-ments studied derive from the national parliament primarily.

This might raise the question if these cases are comparable. I would argue they are, and even that the different scales at which the instruments are devel-oped offers extra insights. As set out in the theoretical section, I understand the state as a field, meaning struggle take place within the field as well. It

(17)

Amsterdam Berlin n % n % Owner-occupied 110.084 27,8 286.648 14,9 Social rental 185.125 46,8 471.333 24,5 ‘Genossenschaften’ 194.305 10,1 Municipal associations 277.029 14,4 Private rental 100.666 25,4 1.165.829 60,6 395.875 1.923.810

Table 1: Tenure structure in Amsterdam and Berlin (2012). Source: AFWC, 2015 & CBRE, 2015

shows how central and local authorities have to cooperate, are dependent on each other, and thus their relation can also be the source of controversies. The scale of application of policy instruments matters for what they do and create, how they work, what they see.

To answer the research question formulated above, several methods of data collection and analysis are employed. First, to develop an understanding of how actors perceive the field of housing in Amsterdam and Berlin, a series of expert interviews was conducted with various agents in the field. Table 2 provides an overview of the respondents. In these semi-structured interviews special attention was given to the strategies pursued by housing companies and agents perspective of their companies role and place in the field.

Second, to gain insight into the (historical) development of the policy in-struments under consideration, I analysed various types of documents from various sources. First of all, I studied official documents, including Acts and their Explanatory Memoranda to explore adjustments of the instruments since they were introduced. I combined this with the official and public minutes of meetings on both local and national governmental levels to not just get an overview of actual changes but also trace proposed adjustments that were not implemented and the discussion about them. Thirdly, I analysed academic and commissioned studies and advices on the policy instruments. Fourthly, I studied news items from both local and national newspapers on the policy in-struments. These news items added new perspectives to the aforementioned sources as they show how policy changes were received. The news items were retrieved through both LexisNexis and newspapers’ online archives.

Thirdly, this research benefitted from participant observation, primarily in the case of Amsterdam. Between 2013 and 2016 I held two positions closely related to the field, namely for the Student Association (ASVA) and for the Student Services department of the HvA/UvA. This gave me an up close look of the political as well as administrative processes in housing. In particu-lar, the paragraphs on shared housing have benefitted from this. Besides this formal involvement, I visited several debates and events in the early period of the 2018 municipal election campaign. On 6 March 2017 I participated in an event described as ‘Scenariogame’, organised by the housing association Stadgenoot. One of the debates took place in debate center de Balie in

(18)

Am-ID Case Date Function

1 Amsterdam 16/3/2016 Senior strategic and policy advisor at a housing association

2 Amsterdam 17/3/2016 Chairman of the board of directors at a housing association

3 Amsterdam 18/3/2016 Policy advisor at the umbrella organisation for housing associations in Amsterdam

4 Amsterdam 31/3/2016 Advisor of Strategy and Policy at a housing organisation

5 Amsterdam 1/4/2016 Director of Strategy and Policy at a housing association

6 Amsterdam 6/4/2016 Former chairman of the board of directors at a housing association

7 Berlin 6/12/16 Manager of housing sector and politics at the

umbrella organisation for public housing companies

8 Berlin 19/12/16 Employee of the controlling department of

communal housing companies at the

Senatsverwaltung für Stadtentwicklung und Umwelt

9 Berlin 17/1/17 Former chairman at a public housing company

10 Berlin 30/1/17 Managing director at a tenant organisation

11 Berlin 10/2/17 Managing director at a public housing company

12 Berlin 16/2/17 Managing director at a public housing company

Table 2: Overview of conducted interviews.

sterdam at 9 May 2017, the other at debate center Pakhuis de Zwijger on 13 June 2017.

(19)

3 Instruments

So far, the theoretical as well as methodological approach of this thesis are set out. In the following section I first set out the results of my the two case stud-ies. The first focuses on the policy instrument Mietspiegel in the context of the Berlin housing market. Second, I turn towards the Housing Valuation System in the context of the housing market in Amsterdam. For both instruments the historical origin and development is sketched, with a specific focus on contro-versies surrounding the design and implementation over the decades these instruments have been in use.

3.1 Mietspiegel

Origin and historical development

The Mietspiegel, Comparative Rent Index (Uffer 2011: 194), or just Rent Index, has its legal roots in the Anti-Eviction Act (German: Wohnraumkündigungsge-setz). This law was installed in 1971 to increase the protection of tenants. The problem the bill responds to is formulated as follows:

In big parts of the housing market demand is still larger than supply. This market situation is intensified due to the increasing building costs. These developments lead to a weakening of the position of renters/tenants and a considerable increase in rents. This bill should increase tenants’ protection for excesses and hardship of the imbalanced housing market.

(German Parliament 1970: 1, my translation)

The bill protects tenants against rent increases by limiting the possibilities for rent increases. These were only allowed when the new rents did not exceed the rents for similar dwellings in the same or similar areas, the so called ortsübliche Vergleichsmiete, with more than a specified percentage. What dwellings were deemed similar, had to be determined on the basis of five char-acteristics: the type, surface, configuration, quality, and type of location of the dwelling. However, these characteristics were not further operationalised andwhat should be considered ‘a comparable rent’ remained unclear. In 1974 the concept was thus further specified in the Second Anti-Eviction Act (Zweites Wohnraumkündigungsgesetz). It allowed municipalities to set up a database of rents to determine comparable rents. The wide possibilities of interpretation lead to a minimal effect of these rules. This database could then function as a ground for rent increases, accessible to both tenants and landlords, instead of having to search for comparable dwellings each time.

The possibilities for setting up a Mietspiegel are provided by national law. Even though the national government provides guidelines (see Federal Min-istry of Transport, Building and Urban Development 2002), the Mietspiegel it-self has to be constructed by municipalities, who are free to decide whether they do this and in what form. Over time, the Mietspiegel turned out to be effectively in some regards, but the differences in the way it was used in var-ious Länder made it unreliable. To strengthen the position of the instrument, a more tightly defined form was designed: the qualified rent index (German

(20)

Parliament 2001). To be designated as qualified, a Mietspiegel should be es-tablished on the basis of ‘scientific standards’. Scientific standards refer to the representativity of the collected data and the use of scientific methods in the analysis of the data. The exact proceedings of both the data collection and analysis process have to be documented as to make sure they can be checked afterwards (Federal Ministry of Transport, Building and Urban Development 2002).

The initial Anti-Eviction Act applied to the Bundesrepublik Deutschland (BRD), which included the western parts of Germany as well as West-Berlin. Within the BRD, Berlin had a peculiar position due to its geographical location within the eastern Germany, the DDR, and was isolated from the rest of West-Germany. The isolation had severe effects on the housing market and economy of West-Berlin (Häußermann & Kapphan 2014). Construction material as well as capi-tal to finance building was hard to get by. Moreover, demand for housing was relatively low. As such, housing was not a profitable business and private investors largely ignored Berlin. To maintain housing quality and availabil-ity, the state of Berlin heavily subsidised housing construction. Both private as well as public construction companies were eligible for these subsidies as long as they limited the rents for the resultant dwellings. Simultaneously, the Berlin population was relatively poor, making them unable to pay higher rents.

Contemporary construction and use

Now I have sketched the origin and development of the Mietspiegel over the past decades, I will explain in more detail how the Mietspiegel currently works in practice and what functions it serves.

The Mietspiegel takes into account five housing characteristics. These are defined in the German CivilCode (German: Bürgerliches Gesetzbuch, §558, part 2) and further specified in the Instructions for the Making of Rent Indexes (2002), devised by the Federal Institute for Construction, Urban, and Spatial Research: • Type (Art). Takes into account the type of dwellings as well as the build-ing. Includes age (alt vs. neubau), whether the building is a one-, two-or multiple-family building, apartment two-or ‘Mehrzimmerwohnung’. • Surface (Größe).

• Configuration (Ausstattung). Includes type of heating, type and avail-able amenities in the bathroom, floor covering

• Quality (Beschaffenheit). Includes construction quality of the dwelling, usually based on year of construction as a proxy.

• Area (Wohnlage). The quality of the residential area can be operationalised in various ways, for example by the connection to transport, the avail-ability of public transport, the distance to shops, green areas, and the city centre, as well as the image of the area. Figure 1 gives an overview of the spatial distribution of residential areas in Berlin.

(21)

Berliner Mietspiegel 2017 | www.berlin.de/mietspiegel 7.1 Wohnlagenkarte Berliner Mietspiegel 2017 überwiegend einfache Wohnlage überwiegend mittlere Wohnlage überwiegend gute Wohnlage Gebiete ohne betroffenen Wohnraum

Achtung: Die Karte soll lediglich

eine erste Orientierung über die mögliche Wohnlagezuordnung ge-ben. Über die genaue Einordnung Ihres Wohnhauses in die zutreffen-de Wohnlage gibt Ihnen auch Ihr Bezirksamt oder die Senatsverwal-tung für Stadtentwicklung und Wohnen Auskunft. Anschriften und Telefonnummern finden Sie auf den Seiten 26/27.

Kartengrundlage:

Übersichtskarte von Berlin 1 : 50 000 (ÜK 50), Stand 6/2016 Herausgegeben von der Senatsverwaltung für Stadtent-wicklung und Wohnen III

10

Figure 1: Wohnlagenkarte Berlin 2017. Source: Senatorial Administrative Of-fice for Urban Development and Housing 2017. Note that even though this map gives an idea of the most common residential area qualities, the actual assessment is made for each dwelling individually

The Mietspiegel is calculated on the basis of a sample of rents paid over the previous four years. In this sample, only contracts that had a starting date within the previous four years are taken into account. Moreover, several types of dwellings are excluded, in particular publicly subsidised housing or dwellings with type of rent control. The data as well as the Mietspiegel itself has to be adjusted to market movements every two years, every four years the data has to be collected anew. There are two types of Rent Indexes: a regular or simple one, and a qualified Rent Index. The qualified rent is used most commonly, except in some smaller municipalities. It is more demanding than a regular one, in terms of data collection and analysis, which have to be empir-ically representative and follow scientific standards, as well as recognition. To be accepted as qualified, a Rent Index has to be approved by representatives of both tenants and landlords. A qualified Rent Index has a stronger juridical status and can be used in court to determine comparable rents (Koch 2005).

For qualified Mietspiegel, there are two generally accepted ways of analysing the data sample: the table method and the regression method (Federal Min-istry of Transport, Building and Urban Development 2002). For the table method, cross tables are constructed using the characteristics that are pre-scribed for the Mietspiegel. For each cell at least a sample of 30 dwellings has to be taken into account. Based on this sample, three values are calculated: a lower, middle and high rent. The middle rent level is taken as the basis, which

(22)

can be adjusted within the given range on the basis of extraordinary charac-teristics, such as high quality floor covering or luxurious sanitary. Figure 2 gives an impression of the resulting table. For Mietspiegels of the regression type, two types of regressions are used (Kauermann & Windmann 2016). The first is parametric regression, in which for example a exponential relation be-tween surface and rent is assumed. Based on this assumption a function is calculated that matches the sample data as good as possible. Alternatively, the Mietspiegel can be based on a nonparametric regression. In this case, the available data is used to estimate a function that predicts the rent value given the surface and year of construction of a dwelling, after which corrections can be made for the other relevant variables in the Mietspiegel.

Berliner Mietspiegel 2015 |

www.stadtentwicklung.berlin.de/wohnen/mietspiegel

9. Berliner Mietspiegeltabelle 2017

Ortsübliche Vergleichsmieten (Stichtag 01.09.2016) Netto-Kaltmiete in Euro je Quadratmeter monatlich

www.berlin.de/mietspiegel 12 Altbau Neubau Bezugsfertig bis 1918 1919 - 1949 1950 - 1964 1965 - 1972 1973 - 1990 West a 1973 - 1990 _Osta mit Wendewohnungen 1991 - 2002 ohne Wendewohnungen 2003 - 2015 Wohnfläche Ausstattung Wohnlage mit Sammelheizung (SH), Bad und WC in der Wohnung (IWC)

mit SH, Bad und IWC mit SH, Bad und IWC mit SH, Bad und IWC mit SH, Bad und IWC mit SH, Bad und IWC mit SH, Bad und IWC mit SH, Bad und IWC 1 2 3 4 5 6 7 8 bis unter 40 m2 einfach A _{5,44 - 10,00}7,45 _{5,60 - 7,17}6,45 _{5,18 - 8,80}6,17 _{5,78 - 7,74}6,58 _{5,33 - 8,15}7,68 _{6,13 - 7,63}6,81 _{7,50 - 9,51}7,50* mittel B _{5,25 - 10,11}8,01 _{6,82 - 8,40}7,37 _{5,89 - 8,00}6,64 _{5,71 - 8,36}6,57 _{7,18 - 12,00}8,08* _{6,40 - 7,94}6,77 _{13,85 - 15,72}14,19* gut C 6,91 - 12,528,31** 6,04 - 10,007,04* 5,87 - 10,207,48 7,70 - 9,848,83 6,93 - 10,118,59* 7,09 - 9,148,04 40 m2 bis unter 60 m2 einfach D 4,81 - 9,226,51 5,30 - 8,095,93 5,16 - 7,295,71 5,13 - 7,015,75 5,80 - 8,436,77 5,30 - 6,675,61 6,21 - 9,177,43* 6,90 - 12,508,50 mittel E 5,11 - 10,066,61 5,52 - 7,656,34 5,43 - 7,486,11 5,39 - 7,205,72 6,65 - 7,977,39 5,37 - 6,755,83 6,84 - 9,378,18 8,40 - 13,9410,10 gut F 5,81 - 8,757,03 5,71 - 8,886,88 5,18 - 8,986,66 6,17 - 9,007,39 7,30 - 9,128,00 6,11 - 8,296,84 7,48 - 10,349,00 8,00 - 11,309,44 60 m2 bis unter 90 m2 einfach G 4,40 - 8,626,00 4,90 - 7,565,65 4,74 - 6,685,48 4,76 - 6,305,22 5,08 - 7,596,12 4,72 - 5,715,07 5,76 - 8,237,10 6,90 - 12,9910,13 mittel H _{4,75 - 10,27}6,78 _{5,18 - 7,62}6,09 _{5,18 - 7,28}6,00 _{5,05 - 6,59}5,49 _{6,16 - 8,76}7,50 _{4,68 - 5,78}5,11 _{6,52 - 8,96}7,59 _{8,61 - 13,93}10,31 gut I _{5,81 - 10,00}7,32 _{5,52 - 9,20}7,10 _{5,58 - 8,82}6,81 _{6,09 - 8,59}6,90 _{7,07 - 9,69}8,24 _{5,37 - 7,00}6,00 _{7,31 - 10,54}8,77 _{8,44 - 11,07}9,01 90 m2 und mehr einfach J _{4,22 - 8,41}5,72 _{4,29 - 7,51}5,80 _{4,29 - 6,60}4,87** _{4,80 - 6,14}5,29 _{5,39 - 8,05}6,49 _{4,24 - 5,60}4,70 _{5,95 - 8,47}7,18 _{8,60 - 13,56}10,70 mittel K _{4,38 - 10,18}6,39 _{4,97 - 7,10}5,63 _{5,77 - 8,03}6,46 _{4,91 - 7,50}5,53 _{5,17 - 8,80}7,48 _{4,76 - 5,60}5,11 _{6,30 - 9,27}8,00 _{8,69 - 13,10}10,19 gut L 5,44 - 10,007,25 5,35 - 9,506,65 6,77 - 9,618,26 6,99 - 8,548,05 7,60 - 10,498,83 5,24 - 6,925,63 7,18 - 11,139,24 8,40 - 12,5010,00 Spalte Zeile

Für Altbauwohnungen (bezugsfertig bis 1918 und 1919 bis 1949) ohne Sammelheizung, ohne Bad, mit WC in der Wohnung (IWC) liegt

die ortsübliche Vergleichsmiete je Quadrat-meter Wohnfläche monatlich 0,87 Euro unter

den Beträgen der ausgewiesenen Spalte 1 für Bezugsfertigkeit bis 1918*** bzw. der ausge- wiesenen Spalte 2 für Bezugsfertigkeit 1919 bis 1949***.

Für Altbauwohnungen (bezugsfertig bis 1918) mit Sammelheizung oder mit Bad, mit WC in der Wohnung (IWC) liegt die ortsübliche

Vergleichsmiete je Quadratmeter Wohnfläche monatlich 1,34 Euro unter den Beträgen der

ausgewiesenen Spalte 1***. Für Altbauwohnungen (bezugsfertig 1919 bis 1949) mit Sammelheizung oder mit Bad, mit WC in der Wohnung (IWC) liegt die ortsübliche

ausgewiesenen Spalte 2***. Für Neubauwohnungen (bezugsfertig 1950 bis 1964) mit Sammelheizung oder mit Bad, mit WC in der Wohnung (IWC) liegt die ortsübliche

ausgewiesenen Spalte 3***.

a Die Zuordnung West-Staakens basiert auf dem Gebietsstand 02.10.1990. Die Zuordnung der Bezirke basiert auf dem Gebietsstand 31.12.2000 vor der Gebietsreform (siehe Erläuterung unter Nr. 3).

Bei Leerfeldern lag für eine verlässliche Aussage keine

genügende Zahl von Mietwerten vor (unter 10 Mietwerte). Die mit * und ** versehenen Daten haben wegen geringer Zahl erhobener Mietwerte nur bedingte Aussagekraft (* = 15 - 29 Mietwerte, ** = 10 - 14 Mietwerte). Die mit *** versehenen Angaben haben wegen geringer Zahl erhobener Mietwerte nur bedingte Aussagekraft. In den Tabellenfeldern werden der jeweilige Mittelwert (Median) sowie die 3/4-Spanne dargestellt.

Figure 2: Berliner Mietspiegeltabelle 2017. Source: Senatorial Administrative Office for Urban Development and Housing 2017

The Mietspiegel is recognised to currently have several functions. The cen-tral aim is to simultaneously protect tenants against unjustified rent increases and keep the development and exploitation of residential property economi-cally viable. It is assumed to do so by creating and improving market trans-parency (Koch 2005: 117). In turn, as an effect of the market transtrans-parency, two important functions are mentioned. First, the Mietspiegel should make the rent increase process simpler and more objective (German Parliament 2000: 36). Secondly, it reduces potential (juridical) battles on rents, and as such reliefs pressure from courts (German Parliament 2000: 2).

The aim of protecting tenants was reinforced in 2015 with the introduc-tion of the Rentbrake (Mietpreisbremse), installed to stabilise the growth of rent increases in various German cities. Lastly, housing providers use the

(23)

Miet-spiegel as a basis for determining their rent levels. As one of my respondents explained:

In our rent politics, we orient ourselves on the comparable rents [Ort-süblichkeit der Vergleichsmiete]. And we think, that is sufficient [hinre-icht] and socially acceptable, and is the right way for the company to carry

out a sustainable management. (Interview,

ID6, managing director at a communal housing company, my translation)

These functions are not unquestioned though. There have been several contro-versies surrounding the Rent Index in Berlin over the past decades, to which the next section shifts attention.

Reflecting which reality?

As set out above, the Rent Index regulates rents on the basis of comparable rents in the housing market. Debates and controversies turn out often to re-volve around the question what the housing market looks like in reality. For example, a repeated discussion in the Berlin Parliament as well as local media pertains to which dwellings should be included in the construction of the Mi-etspiegel. Currently, only rents that in the four preceding years have changed or were set for the first time are taken into account. Excluded are rents of the existing stock (German: Bestandsmieten) that have not changed, as well as the rents of subsidised housing, which are contractually set at a reduced level. However, both real estate owners and members of parliament have argued that other types of contracts should be in- or excluded. Due to the exclusion of older, existing rental contracts, it has been argued that the Mi-etspiegel in fact does not reflect rents, but rent increases (Keicher 2014). The Left party (Die Linke) in Berlin has proposed to include more rents that were not recently changed (see Die Linke (2012)). They argue that the Mietspiegel will ‘push up prices if the share of dwellings considered in the sample does not reflect the real facts [German: realen Gegebenheiten]’ (Die Linke 2012: 3). The proposal was rejected in Berlin, but in the meantime the federal Min-istry of Justice and Consumer Protection has drafted new legislation to extend the period of which rents may be included to eight or ten years2(Immobilien Zeitung 2016: 1). Contrastingly, commercial housing providers argue that less dwellings should be considered for the Mietspiegel, focusing only on the most recent rents to reflect the actually offered rents (Angebotsmiete).

The discussion however does not only focus on the period from which rents that should be taken into account. In their coalition agreement, the local government in the state of Berlin proposed to limit the rent increases of all dwellings owned by the public housing companies to 2%. These opposed this measure, as they wanted to have the freedom to use varying rents increases (possibly with an overall increase of max. 2%). Moreover, as the Mietspiegel does not take into account contracts with rent control, they argue that their property will not be allowed into the Mietspiegel sample anymore, in effect raising the resulting Mietspiegel levels:

But now there is a very important question. If the government tells us now that we are only allowed to raise our rents a certain percentage, which has

(24)

nothing to do with the law, let’s say, which has something to do with the will of the government, then the big question is whether we can still be included in the Mietspiegel. And if we can’t, I think that will hurt everybody else who lives in a private apartment.

(Interview, ID12, managing director at a communal housing company)

Besides the question which dwellings should be included, there is also debate about how to assess them. Often this relates to the quality of the residential area (Wohnlage). Property manager Deutsche Wohnen questions if the division of all dwellings in three types of residential areas (Wohnlagen) - simple, mid-dle and good - is sufficient. Due to their broadness they would not display the actual market development. According to Deutsche Wohnen, the Wohnlage for many areas does not reflect the actual demand for dwellings in those places (see Berlin House of Representatives 2016 and Berlin House of Representa-tives 2017). As an example they mention the neighbourhood Kreuzkölln, on the edges of the districts of Neukölln and Kreuzberg in Berlin. This neighbour-hood has been gentrifying (Hochstenbach 2015) and the demand for dwellings is very high, but the common residential area quality is still assessed as simple (einfach, the lowest of the three). Similar calls come from statisticians, who ar-gue there can and should be more fine-grained categories (Bünnemeyer et al. 2016; Kauermann & Windmann 2016). Which rents should be included in the sample thus forms one of the central axes of debate.

Another axis is formed by the question what to do with the data from the sample. The problem here is how a model can be created that allows the extrapolation of the data to other dwellings, i.e. how a dwelling can be properly compared to the dwellings in the sample. Kauermann & Windmann (2016) argue the table method, as used in Berlin, is problematic because it re-sults in ‘jumps’ for dwellings that are on the edges of two categories. This could mean, for example, that two dwellings with nearly the same surface areas of 59,80 and 60,20 square meters fall within two categories of the Rent Spigel, resulting in quite different maximum rent levels for similar dwellings. The regression method, contrastingly, does not have this problem as it works without categories. In general they are negative about the statistical methods used in Berlin, which they deem to be invalid, outdated, and unreliable. Be-sides the jumps mentioned above, the method does not account sufficiently for geography. With the current method, a dwelling in Mitte in the city center might have the same maximum rent as a dwelling in Marzahn, on the eastern outskirts of the city, because the quality of the residential area is categorised the same. Instead, they propose to use a model that corrects for geography, raising rents for dwellings in the center and reducing them as they are further away from the center.

The table-method is criticised as well by Jan-Marco Luczak, member of the German federal parliament for the centre-right christian party CDU, who ar-gues the number of dwellings that form the basis for the cells are too low. To improve this, he wants better data as input as well as stricter national guide-lines on how the Rent Index should be calculated. This is needed to ensure the juridical validity of the Rent Index (see Luczak 2017).

(25)

Recognition

One of the purposed advantages of the Rent Index system is that it has to be accepted by the municipality, tenants’ representatives and landlord represen-tatives. If a Rent Index is not recognised by these parties it is no qualified Rent Index, which in effect makes in nearly useless as it has no juridical value any-more. Thus, to assure the Rent Index is accepted by all parties working groups are set up to negotiate them in advance of publication. The 2017 Berlin Rent Index, was officially recognised by the four different tenant organisations, as well as the BBU, the Verband Berlin-Brandenburgischer Wohnungsunternehmen. The latter is the umbrella organisation for the communal housing companies as well a series of smaller housing companies, in total representing about 35 % of the total housing stock in Berlin (BBU 2015). Besides these organisations is-sueing official recognition, representatives of commercial landlords and prop-erty owners have offered advice during the process (Senatorial Administra-tive Office for Urban Development and Housing 2017).

Not all parties in Berlin recognise the validity of the 2017 Rent Index though. Deutsche Wohnen, the biggest commercial housing company in Berlin, publicly criticised and rejected it. They doubt if it has been drawn up on the basis of sci-entific standards, as qualified Rent Indexes require. These doubts relate, first, to the way data has been collected through the survey. According to Deutsche Wohnen, too many rents from the public housing companies are taken into account compared to housing owned by commercial landlords. Since the lat-ter generally have higher rents, they argue the rent levels in the Mietspiegel are too low and do not reflect the actual average rents (see Berlin House of Representatives (2017)). Second, Deutsche Wohnen questions if the division of all dwellings in three types of residential areas (Wohnlagen) and thus did not recognise the Mietspiegel.

This brings us to a fundamental aspect of the Rent Index system, namely the way it aims to balance interests (German: Interessenausgleich) and the pre-vention of juridical battles (German Parliament 2000). It does so, by offer-ing an independent and objective basis for settoffer-ing rent levels. This aspect is stressed by the State Sexretary of Construction and Urban Development An-dreas Geisel, who declares ‘It [the Rent Index] is valid, and it is so in the in-terest of tenants, but also in the inin-terest of landlords, because it offers clarity, as it prevents juridical disputes (Berlin House of Representatives 2015: 6627)3. The Rent Index might not be so clear though. The controversies set out above testify to this, but also in court the 2015 Rent Index has been judged as not being based on scientific methods, and thus not recognised as qualified Rent Index (Gerichtsamt Charlottenburg 2015).

What is being sought after with the Rent Index, appears to be an objec-tive baseline that helps in the negotiation of interests. The state tries to assure this objectivity by referring to its scientific quality. However, the negotiation process that runs along with this makes the process fundamentally politically. This hints at a fundamental tension in the Rent Index system, namely between the political process, i.e. the struggle between the various agents to improve

3_{Original in German: ‘Er ist gültig, und zwar im Interesse der Mieterinnen und Mieter, aber}

The symbolic power of policy instruments