Rent regulation in 21st century Europe. Comparative perspectives

Abstract The general housing policy trend in Europe has been towards neoliberalization meaning less state involvement in housing market and less government support for housing production. However, private rental markets are still regulated in many European countries. Here, we classify 33 European countries based on rent regulation system and welfare state regime. There seems to be some but not too much correspondence between the welfare state regime and whether rents are controlled. However, it seems that the role of rent regulation depends on the context and one should take a closer look at specific cases. We look at Nordic welfare states that are similar in that all represent the social democratic welfare model but different in their housing regimes by which we mean the basic principles of how housing provision in the country is organized.


Introduction
Housing researchers are quite unanimous about a general neoliberal trend in European housing policies (e.g. Bengtsson et al 2013;Aalbers 2015;Dewilde & De Decker 2016;Clapham 2019). Governments are withdrawing from large scale intervention in the workings of the housing markets, financial support for housing production is cut to very little or nothing, support for housing consumption is targeted only to those most in need, and housing markets are deregulatedalso in the countries that have been thought to have de-commodified housing the most (see Clark & Johnson 2009;Christophers 2013). How this trend has affected the regulation of private rental markets (PRM), a policy that seems to be quite out of tune of the present trends in public policy? Rent control was a common policy in Europe during the first half of 20 th century and especially during wartime crises. After that there has been great variation in the use of this policy instrument. Some countries wanted to get rid of rent control as soon as possible (e.g. Finland), sometimes taking it up again later temporarily or in milder versions (e.g. Ireland), in other countries (e.g. Sweden) regulation of rents became more persistent feature of the housing system. This paper attempts to answer three questions. The first question is in what ways European countries practice regulation of rents today? Assuming the general trend towards free markets we could expect that regulation of rents would be disappearing as a policy. We will see that there is quite a lot of variation in this respect between European societies. Perhaps practicing regulation of rents is linked to the way how welfare is provided in a country. Therefore, our second question is: to what extent there is an association between a country belonging to a distinct type of welfare regime and its rent regulation policy? As we know that housing does not play a particularly central role in welfare regime theory, we might employ a regime concept that is more specifically targeted to housing. The third question of this paper is: can the type of a country's housing regime account for its rent regulation policy?
This paper is a policy review drawing on information about rent regulation in European housing policy systems from mainly secondary sources. The information on individual countries has been retrieved from TENLAWproject's country reports and various other published sources (research articles, government documents, etc.). We have defined rent regulation as a type of tenure legislation that states limitations to rent setting and rent increases by private (i.e. for-profit) landlords. We have classified 33 European countries and the 'hardness' of their rent regulation system based on the idea of generations of rent control. Countries that regulate rents or rent increases have been classified as having a second-or third-generation rent regulation system. Countries with no rent regulation have been classified as having free rental markets.

Regulation of Rents 1 as Housing Policy
Housing differs from other market commodities in the sense that it is a necessity, and it requires most households to spend a considerable amount of their income (Arnott 2015). Housing policies often aim to provide everybody good quality housing at a reasonable price. How this goal is pursued varies between countries, but it usually requires the state to take a hand in the housing market. According to Doling (1997) the state has six options to intervene in the housing market: non-action, exhortation, regulation, taxation, subsidy and provision. In practice these options are often implemented in various combinations that interplay with each other. For example, the introduction of housing allowances can affect rent regulation systems (Turner & Elsinga 2005). Both housing market and housing policy are based on tenures. The most widespread tenure forms are owner-occupation, social rentals and private rentals, but tenure forms have different social connotations depending on the socio-economic context (Haffner et al. 2018). The state may structure markets in different ways, and the consequence of the varying decisions taken by national governments is that housing markets differ in their structures and their consequent outcomes between countries and over time (Clapham 2019). In this paper we examine these variations in rent regulation and their association with types of welfare regimes in European context and housing regimes in Nordic context. Our approach can be described as social constructionism as we emphasize the significance of social, political and economic context in researching housing policy (Jacobs et al. 2004). Haffner et al. (2018) distinguish between social and private rental sector based on landlord type, allocation and subsidies. In private rental sector (PRS) landlords are typically institutions or private persons, housing is allocated based on the market and there are typically no subsidies. Still, private renting has a legal structure that determines the duties and responsibilities of landlords and tenants. Common areas of regulation include the length of tenancies, rent levels and the security of renters. In countries with strong regulation, there is a framework of national law that seeks to define responsibilities and protect the tenant as it is assumed that there is a power imbalance between the two parties. In countries, where there is not much regulation, the contract between the landlord and tenant is very important (Clapham 2019).
Regulation of PRM can take many forms, but in this paper, we are mainly interested in the regulation of rents. Rent regulation means that the rents of rental dwellings are somehow restricted. Rent regulation empowers a democratically accountable government body to regulate prices of rental housing in a specific jurisdiction (cf. Zimmer 2017). Rent regulation can be done by regulating either the level of rents or rent increases or combining both (Turner & Malpezzi 2003). Arnott (1995Arnott ( , 2003 classifies rent regulation in three generations of rent control. First-generation rent control means 'hard' control in which rents are frozen at a certain level. Second-generation rent control is more 'soft' regulation in which a set of regulations govern rents and rent increases, but usually a 'reasonable' rate of return is ensured to landlords. There is more flexibility in the design of a second-generation rent control package than in the design of a first-generation rent control package (ibid., Lind 2001). In third-generation rent control system rent increases are controlled within a tenancy but are unrestricted between tenancies. This classification of rent control generations is used here to make distinctions between different types of rent regulation and to underline the development of rent regulation systems into a more market-oriented course. We will show in this paper, where European countries lie at present in terms of the generations of rent control and in relation to welfare state regimes. We will also demonstrate later in more detail, how housing regimes have affected the evolution of rent regulation in Nordic context.
First-generation rent control started as wartime rent freezes during the World Wars in Europe and continued long after the Second World War had ended. The era of second-generation rent control started in Western Europe in the 1970s when milder regulation measures were introduced in many European jurisdictions. (Haffner et al. 2012, Arnott 1995. In Eastern Europe the history of private renting was quite different under the socialist regime where housing was more or less state controlled and private renting existed as a 'crack' in the system, typically in the form of forced private renting or subletting (Heged€ us & Puzanov 2018). Since then, rent regulation systems in Europe have continued to evolve in correspondence with economic and political turns. There has been a move towards third-generation rent control (Arnott 2003) and deregulation of PRM. Rent regulation is usually employed to protect what is considered as the weaker party of the tenancy agreement, the tenant. Rent regulation can be described as balancing between the interests of tenants and the interests of landlords 2 . The landlord-tenant relationship is inherently unequal as the landlord has the discretionary power to withhold access to a good that is central to a tenant's wellbeing (housing) (Zimmer 2017), and therefore rent regulation might be introduced to protect the tenant from this power asymmetry. But if the regulation is too strong, landlords may exit the market in the long run (Haffner et al. 2008).
According to Jenkins (2009), the general view among economists is that rent regulation has many negative and market-distorting side effects such as misallocation, rental housing shortage and under maintenance (see also Arnott 1995Arnott & 2003. This is the case especially concerning first-generation rent controls that no longer exist in Europe (as is demonstrated later in this paper). Despite these critical views, rent regulation has persisted and it still has advocates. Haffner et al. (2012) offer rationales for rent regulation based on welfare economics: housing market inefficiency, uneven income distribution or social segregation might justify rent regulation policy if disadvantages for tenants take centre stage in housing policy. Clapham (2019) argues that the effectiveness of housing policies should be judged based on how they impact 'dwellers', those who are or could be housed. Therefore, 'a good housing policy is one that improves individual and collective well-being through increasing the affordances that dwellers achieve from their housing' (ibid. p. 23). On the other hand, empirical research in Europe has shown that the impacts of rent regulation vary greatly between countries and regions and are therefore specific to housing market situations (Arnott 1995;Haffner et al. 2012;Turner & Malpezzi 2003). Whitehead et al. (2012) 3 found no clear link between the level of regulation and the size of the PRS in European countries.
Even when intended as temporary, rent regulation programmes seem to hold on (Arnott 2003). The contemporary PRM reflect specific regional and national traditions, laws, housing policies, practices and power relations (Haffner et al. 2018). Maintaining the system can be an active choice or a result of non-decision-making (Kettunen 2018).

Is Rent Regulation Disappearing in Europe?
In the beginning, we posed the question of what ways European countries practice rent regulation today. Table 1 summarizes our analysis of tenancy laws in 33 European countries in alphabetical order. It is based on material retrieved from project TENLAW's national reports and previous research on rent regulation and tenancy laws. We analyzed the material to investigate if there is legislation on rent regulation in a country and how it is targeted in terms of rents and rent increases. We defined rent regulation as any type of tenancy legislation that states limitations to rent setting and/or rent increases in rental contracts in the PRS, namely rental sector with private landlords and market based allocation. We aimed at making the classification as clear as possible, so even countries with very mildalmost market oriented regulation were classified as having a rent regulation system. Rents or rent increases tied to an index were classified as rent regulation, as well as regulation limited to certain areas or parts of the housing stock.
There were no countries with rent freezes typical to first-generation rent control. The distinction between second-and third-generation rent control was made based on how the regulation was targeted. If both initial rents and rent increases were regulated, it was classified as second-generation rent control. If only rent increases were regulated, it was classified as third-generation rent control. The distinction between second-and third-generation rent control was not clear-cut in all cases. There are some countries that have legislation where rent regulation depends on e.g. the age or location of the premises or the year that the rental contract is conducted. (There might also be other national and local exceptions that we were not able to retrieve from the material.) Regarding these cases, we determined the generation of rent control based on the restrictiveness of current legislation: if initial rents are regulated even in parts of the housing stock, it is classified as second-generation rent control. In some countries, old strict laws are still applied to rental contracts dated before the 1990s, but this legislation was excluded from the classification. However, this shows the stickiness of rent regulation in some cases. The share of private rental housing in each country is included in Table 1, but we have no data on the coverage of regulatory measures (i.e. how many households are affected).
There are several interesting points to be made about the information in Table 1. At present, there is still a considerable amount of rent regulation in Europe, though in most countries, altogether 17, a free market reigns in PRM. There are ten countries with third-generation rent control system and six countries with second-generation rent control system. The generations typology is used to describe the 'hardness' of the control system, as third-generation rent control is usually a lighter form of regulation than second-generation rent control. However, in practice both second-generation measures and third-generation measures (as defined in this paper) consist of restrictions of different scale. Figure 1 displays the countries on a map. As we are analyzing policies rather than outcomes, the restrictiveness of certain rent regulation system (in practice) is outside the scope of this paper. If we look at the share of private rental housing, we see that there is great variation. This variation is not consistent with the generation of rent control. In countries with second-generation rent control, the share of PRS is between 8 percent and 41 percent. In countries with third-generation rent control and free market renting the numbers are 2,9 percent to 52 percent and 1,7 percent to 19,8 percent, respectively (to see an overview of the share of different tenures, see appendix 1). All the 33 countries listed above have had some type of rent regulation systems at some point in their past (Whitehead et al. 2012;Van der Heijden & Haffner 2000;Haffner et al. 2008;Baldini & Poggio 2012;Crook & Kemp 2014; TENLAW National reports 2017). Coming to this day, some of them have stayed in the path of rent regulation, albeit adjusting the system along the way, and some have decided to free PRM from rent regulation. As we know that housing policies evolve in interplay with general economic and social processes along with national political forces and demographic and economic factors (Haffner et al. 2018), we come across with the idea presented in the beginning of the paper: could practicing regulation of rents be linked to the way how welfare is provided in a country?

Type of Welfare Regime and Rent Regulation
Housing policy is part of the welfare state, albeit a peculiar one. Housing researchers know that housing does not fit in well with the other pillars of welfare, pensions, education, health care and social aid, due to its ambiguous place between the market and the state (see e.g. Torgersen 1987;Kemeny 2001). Practically everywhere the market actors take care of the production and a part of the distribution of housing at the same time as the state carries the final responsibility for housing of the population. Welfare state policies have been successfully analyzed with welfare state typologies that group countries with similar policies into distinct welfare regimes. The most famous of these is the three worlds of welfare capitalism model developed by Gøsta Esping-Andersen (1990). The welfare regime approach has interested housing researchers quite a lot, even though housing is not central in the original theory (Doling 1999;Fahey & Norris 2011;Hoekstra 2003;Kemeny 2001;Matznetter & Mundt 2012;Stamsø 2010;van Gent 2010;Zhou & Ronald 2017).
Esping-Andersen's original welfare state typology distinguishes three types of welfare state regimes based on three interrelated criteria: de-commodification, stratification and the arrangements between the state, market and family. Esping-Andersen's regime types are social democratic, corporatist and liberal welfare state (Esping-Andersen 1990). Subsequent research has added South European welfare state model to this list (e.g. Allen 2006;Ferrera 1996;Rhodes 1996), as well as a post-socialist welfare model (see, e.g. Aidukaite 2004;Stephens et al. 2015;Stunding 1996). Table 2  groups 33 European countries according to these five welfare state regime types. The countries within the same regime type allegedly have similar principles concerning the provision of welfare, as well as a same kind of historical development in the field of welfare. Therefore, it can hypothetically be expected that their housing policies would also be similar. Thus, we are interested in looking at the countries under the same welfare state regime to see if they have similar rent regulation policies. According to Clapham (2019, p. 51), when it come s to private renting, liberal housing regimes seem to give priority to owner-occupation rather than private renting and social-democratic regimes seem to give priority to public rather than private renting. Only in corporatist regimes regulated private rented sector seems to flourish. In all the countries belonging to the corporatist welfare regime, there is a system of rent regulation. The systems are mostly classified as third-generation rent control, but there are also second-generation systems. Some of these countries, like Germany and Switzerland, have a large PRS whereas in others the sector is rather small, like in the Netherlands, where the rental housing sector is exceptionally large, but it is dominated by rental housing by private non-profit corporations, not classified as private renting. In Jim Kemeny's typology of rental housing systems, most of these countries belong to the group of unitary rental markets where non-profit rental housing companies operate on the open market influencing crucially the whole rental market. In theory there should be a tendency in these countries towards phasing out of public subsidies and regulation (see Kemeny et al. 2005). This seems not to have happened.
Post-socialist welfare states are another group with a rather consistent association between the welfare state regime and rent regulation. During the socialist years the idea of housing policy was to attain general and equal accessibility of housing. In practice this meant strong, though never total or cross-nationally uniform (see Tosics 2012, Soaita & Dewilde 2019) state control over different aspects of housing policy. Private renting was usually legal under certain conditions but landlord's property rights were extremely limited (Heged€ us & Puzanov 2018). After the collapse of communism, the liberation of housing markets and privatization of rental housing to individual tenants alongside with tradition of self-provision led to a very high levels of homeownership and low levels of mortgage debt (Stephens et al. 2015). Some of the housing restituted to past owners was transformed into private rental housing as sitting tenants were protected against evictions (Lux et al. 2018) Soaita & Dewilde (2019) point out that the differences that these countries had in their housing provision during communist era, are still reflected in the contemporary housing quality. At present, owner-occupation is the dominating tenure in post-socialist countries and PRS has been a neglected area of housing policy and this has led to widespread informality of the sector ). There are twelve countries classified in this group, and third-generation rent control is professed in only two of them, Poland and Croatia. In both countries, as well as in all but one of these countries, PRS is quite insignificant.
In all the countries 5 classified in the liberal welfare regime we would also expect to find a free PRM. This does not seem to be so, as two of these countries, Ireland and Scotland, practice rent regulation, some introduced quite recently. In Ireland, there is a very mild second-generation rent control system targeted at private rental housing. However, Ireland has placed new restrictions on PRM lately, as a time-bound (up to 3 years) system of rent pressure zones restricting rent increases was introduced in December 2016 (see Government of Ireland 2016). Lately, also Scotland has introduced a mild third-generation type of rent regulation system that Robertson and Young (2018) call 'almost free market approach to rent regulation'. In this system, rent increases are only permitted annually and tenants can challenge rent increases by referring the rent rise demand to a Rent Officer, who then determines the going market rent. In addition, Scotland has introduced rent pressure zones similar to Ireland (ibid.; Private Housing (Tenancies) (Scotland) Act 2016). Due to new legislation, Scotland is classified as having third-generation rent control. However, this classification is debateable since the requirements for rent regulation or rent pressure zones are set very high for them to actually take effect.
Most of the countries classified as belonging to the South European welfare regime also have a free market in PRM. However, Spain and Cyprus have a somewhat different form of third-generation rent control. In Spain, rent increases in all rental housing are controlled for the first 3 to 5 years, whereas in Cyprus, regulation of rent increase concerns only statutory housing. According to Allen (2006) the development of housing systems in South European countries differs from the northern Europe. After Second World War there was much less damage to the housing stock and most of the population was still living in rural areas. There was no need to develop extensive social rented sectors thus housing policies mainly supported owner occupation. The extended family was an important provider of cash and other resources for newly forming households.
The group of countries classified as social democratic welfare states might be thought to favour the regulation of PRM. In three of the five countries, Denmark, Norway and Sweden, some form of regulation is, in fact, professed. In two of the countries, Finland and Iceland, there is a free market in PRM. We will look more closely at these countries in the next section since they represent an interesting mix of state involvement in the private rental sector. The discussion concerning them will add some (more) complexity to the issue of rent regulation.
To conclude our first two questions, there still is rent regulation in Europe and there are even countries that have lately taken measures to strengthen rent regulation. Apart Ireland and Scotland, also Germany introduced a 'rental brake' for new lettings in areas of high demand in 2015 (Deschermeier et al. 2016). All these legislations are targeted to areas of high demand where market rents are expected to increase excessively. This could also be an attempt to update rent regulation to tackle present day rental housing market problems. Looking at the different approaches taken by different European countries concerning PRM, we come back to the question, why some countries preserve or even return to rent regulation while others abolish the system. It seems that a country's belonging to a specific welfare regime will not provide an answer to this.

Housing Regimes and Rent Regulation -the Nordic Countries
Our focus on the persistence and forms of regulation of rents is interesting in a sense that deregulation and market solutions are a major trend in many countries in Europe but it is deliberately narrow (as we are not looking at other aspects of the rental housing relation, like security of tenure). However, a deeper look at our topic would need a more case-based approach and contextualization. Tenure forms are not independent of each other. They are alternative institutional arrangements through which consumers gain possession of housing. They form a system in the sense that developments concerning one tenure form may have repercussions for others. For example, increasing supply of social rental housing may push rents lower in the PRM, or, substantial strengthening of security of tenure in renting may diminish people's interest in becoming owner-occupiers. Therefore, the significance of rent regulation as a policy is dependent on the institutional system of possession of housing, a significant part of the country's housing regime.
The concept of 'housing regime' is often used in comparative housing researchbut without consensus on how it should be defined (see e.g. Kemeny 2006, Dewilde & De Decker 2016, Clapham 2019). By 'housing regime' we refer to the set of fundamental principles according to which housing provision is operating in some defined area, municipality, region, state (Ruonavaara, forthcoming). Such principles are thought to be embodied in the institutional arrangements concerning housing provision, in political interventions in housing issues, as well as in discourses through which housing issues are customarily comprehended.
In welfare regime typologies, the Nordic countries represent the same social democratic welfare state regime creating an expectation that also their housing regimes would be quite similar. Lujanen singles out the following similarities: Nordic countries 'seek to provide housing of a decent standard for the whole of their population', there is a division of labour between the state and municipalities in housing policy and the private sector builds housing and 'is responsible for the areas where this sector functions effectively' (Lujanen 2004, p. 15). These similarities are, however, quite general and it is doubtful whether they single out Nordic housing regimes from most others in welfare capitalist societies. A good case can be made that in Nordic countries there are larger differences between housing regimes than welfare state regimes (see Bengtsson et al 2013).
We agree with Clapham that '[e]very housing regime is unique' (Clapham 2019, p. 39)but for analytical purposes they need to be classified into types. Here we use the typology by Bengtsson et al (2013) in their study of Nordic housing policies. The book contains formulations of the core characteristics of the post-war housing systems in five Nordic countries. However, when comparing the housing regimes, the different characteristics are reduced to two dimensions, the orientation of housing policy and its institutional basis. The first dimension refers to whether the policy is universal (general), targeted to all households, or selective, targeted to those who need support the most. The latter refers to what housing tenures the policy uses as its main instrument. We start from this typology and then bring in other features of the national housing regimes. Table 3 shows the Nordic countries according to the type of housing regime they have. Note that this table is a description of housing regimes at their 'high point', sometimes around 1980. Changes have happened as the housing regimes built to maturity in the post-war period are challenged by a trend towards marketization and housing policy retrenchment.
Danish, Swedish and Norwegian housing policies have been described as 'universal' in the sense that they are directed towards all types of households and all segments of the housing market. Finnish and Icelandic housing policies have been labelled as 'selective', meaning that they are directed more directly towards households of lesser means based on individual means-testing (Bengtsson & Ruonavaara 2010). Also, the role of different tenures varies between Nordic countries. In Denmark, the main instrument for housing provision has been rental housing in the so-called third sector that consists of housing associations that are self-governing units (Nielsen 2010). In Finland, housing policy has in the past operated through both renting and homeownership, as state subsidies have been targeted for both tenures. More recently, the focus has shifted to social rental housing in line with the Finnish housing policy's selective orientation. Housing policy has mainly been understood as social policy aiming to provide housing for those who cannot acquire it in the market. In Iceland, the orientation of the country's selective housing policy has been towards self-build home ownership. In Norway, housing policy has been institutionally based on individual and cooperative ownership. The social rental sector represents only a marginal percentage, and the role of private renting has historically been restricted. In Sweden, the municipal rental housing sector has been the central housing policy instrument. However, the long-term policy objective of the Swedish housing policy has been tenure neutrality in the distribution of state subsidies (Bengtsson & Ruonavaara 2010;Ruonavaara 2012; see also Bengtsson et al. 2013). Table 4 summarizes the present rent regulation policies in the Nordic countries. The two countries with a selective housing policy, Finland and Iceland, are the only ones with a free market in PRM. Finland has a long history of rent regulation. During the first years of independence (1917)(1918)(1919)(1920)(1921), several short-term rent control acts were introduced, but the tenancy law of 1925 had no rent control. During the Second World War, rent control measures were again introduced and they continued after the war due to a serious housing shortage (Juntto 1990). Rent controls were gradually eliminated by 1961 but reintroduced again in 1968. This system was found too rigid and was replaced with a second-generation rent regulation that lasted until the 1990s. The private rental market was gradually deregulated in a five-year period from 1991 to 1995 (Kettunen & Ruonavaara 2015;Kettunen 2018;Ruonavaara 2013). The story of rent regulation in Iceland is somewhat different. Like elsewhere, also in the early 20 th century Iceland most dwellings in urban centres were rented. A housing crisis in Reykjavik in 1917 led to introducing a rent control policy which was opposed by landlords and abolished in 1926. Rent control was introduced again during the Second World War and abolished in 1953 where after rental markets have been unregulated. However, the post-war housing policy was characterized by heavy promotion of self-built, privately owned family houses. By 1960 already 70 percent of housing was owner-occupied, and the rising trend continued until the 1990s making Iceland one of the foremost 'nations of homeowners' in Europe. The particularly strong emphasis on home ownership led to the marginalization of the PRS. (Sveinsson 2000.) Ruonavaara (2012) notes that one thing in common with the Nordic countries has been a certain policy bias against private renting. The housing policy regimes that developed in the Nordic countries after the Second World War tended to treat private renting as speculative activity that needed control and regulation for the protection of tenants. According to Sveinsson, 'the main force behind the growth of homeownership in Iceland -was not so much the 'pull' of homeownerships as it was the 'push' away from the early twentieth century dominant tenure, private landlordism' (Sveinsson 2000, p. 83). He adds that in a country like Iceland the escape from private renting was channelled to home ownership whereas in other countries non-profit rental housing was the solution.

Rent regulation in the Nordic countries
In the Finnish housing regime, the route of 'escape' was provided by subsidizing the growth of homeownership with production and consumption subsidies but also by creating a social rental housing sector. That sector grew from modest beginnings in the post-war years to become as big as the PRS. Rent regulation policies made private rental housing an unattractive investment and contributed to the decline of its share in the housing stock. The expansion of state involvement in housing provision ended by the end of 1980s and a period of retrenchment started. One of the most important signs of the shift towards retrenchment was the deregulation of PRM (Ruonavaara 2013).
Norway is a country that has for a long time had a universal orientation in housing policy and an emphasis on homeownership. The clearest and probably most famous example of the Nordic suspicion of PRM is the Norwegian social democratic leader Trygve Bratteli's statement in the parliament in 1951 that he cannot accept 'owning other people's homes as a field of private enterprise' (cited in Annaniassen 2013, p. 220; our translation). According to Torgersen, the housing policy priority in post-World War II Norway 'was to organize the exodus from the Egypt of private tenancy' to the Promised Land of homeownership (Torgersen 1987, p. 123). Norway has been and still is a 'social democratic home-owning society' (Annaniassen 2013) with a housing policy focussed on advancing ownership of housing, both in the form of homeownership and co-operative tenure (see e.g. Stamsø 2014 andSandlie &Gulbrandsen 2017). In contrast to Denmark, Finland and Sweden the public rental housing sector has been quite unimportant in Norway.
As many other countries in Europe also Norway introduced a first-generation rent control during the First World War. This system made way to milder second-generation rent control that lasted until 1935. After the Second World War rent control was again reinstated. Also, the evictions were made difficult, and the conversion of rental flats to owner-occupied homes was restricted. In the Norwegian system the regulation was enforced through local rent boards that landlords were supposed to involve in determining the rent of new dwellings. Whether decided jointly with the rent boards or independently the rent levels were to comply with the rent regulation rulesotherwise the tenant had the right to demanding lowering of the rent. In 1982 this system was abandoned, though regulations were left in force in dwellings built before the Second World War. The current rent regulation system is quite mild. The rent act of 1999 introduced a legislation that allowed market rents for new tenancies. Rent increases can be made once a year based on consumer price index. A larger revision is permitted every third year. The typical length for a fixed-term rental contract is three years in Norway. (Oust 2018). Though this system is quite close to unregulated market, we have categorized the Norwegian rent act as representing a very mild third-generation rent control since there is a limitation on rent increases, albeit the increases follow general price trends and the adjustments every third year secure market rents. Still, the limitations to rent increases reflect the state's involvement in the PRM.
In Norway the established housing regime with a traditionally universal orientation has not kept rent regulation strong in the housing system. But that regime is in transition due to the state retrenchment in housing policy. In recent times Norway has been moving towards a selective model of housing policy . Moreover, the Norwegian housing policyirrespective the government in powerhas been based more on homeownership and co-operative tenure than renting. The deregulation of prices in co-operative housing has been considered a more important milestone in housing policy than that of PRM (see e.g. Annaniassen 2013; Sørvoll 2013).
Denmark and Sweden are the two Nordic countries with a universal housing policy and second-generation rent control. In Denmark, there exists four different rent regulation systems simultaneously, but, in general, all types of rented dwellings are subject to rent regulation. The primary type of regulation is cost-based rent. In other regulation systems, the rent is determined either through comparison with other tenancies or based on the value of the premises (i.e. determined through comparison with other similar tenancies in the same area) (TENLAW National report for Denmark 2017). However, in new dwellings built after 1991, there may be deviations from rent regulation. Also, municipalities have some latitude concerning the application of rent regulation, and a total of about 130 municipalities (out of 275) have no rent regulation for the PRM. It is estimated that non-rent regulated letting accounts for about ten percent of the Danish rental market (Skak & Bloze 2013). Rent increases can be made when the running costs of the apartment exceed the existing rent, when improvements have been made or when the value of the premises significantly exceeds the existing rent. Landlords and tenants can also agree on a stepwise rent increase. When rent regulation is not applied, rent increases may be based on a net retail price index or may be stepwise rent increases (TENLAW National report for Denmark 2017).
In Sweden, the rent regulations introduced during the Second World War were gradually replaced by the current rent negotiation system, between 1969 and 1978 (Andersson & S€ oderberg 2012). Sweden has its own unique system for setting rents and rent increases. Owner-occupied housing can be rented out freely, but this applies mostly to detached and row houses. Most of the rental housing consists of blocks of flats, and they are owned by municipal housing companies or private rental companies. Rents in these flats are bargained between tenants' and landlords' associations (Bengs 2015). Previously, private rents were tightly connected to public rents, but this connection has now been discarded. Regardless, private sector rents still must be shown to be comparable to those apartments of similar use-value (Christophers 2013). When setting rent for an apartment, other apartments with a similar utility value must be found. The rent is set based on the highest rents of the comparable apartments. If there is no comparative material, the rent tribunal will assess the rent (TENLAW national report for Sweden 2017). All rent increases are negotiated between landlord and individual tenants or between landlords and the Swedish Union of Tenants (f the landlord has a principal bargaining agreement with the Union and the tenants have a bargaining clause in their tenancy agreements). If there is no agreement between the landlord and the union (or the tenants), an application to the rent tribunal must be made (TENLAW national report for Sweden 2017).
Usually Sweden's rent negotiation system is classified as rent regulation-as is done here-but there are dissenting views about this. This system is sometimes called indirect rent control because rents are controlled through negotiations, rather than through an explicit direct form of rent control (Andersson & S€ oderberg 2012). Whether rents are regulated in Sweden or not is also partly a political issue. On one hand, those who generally favour deregulation are prone to consider the Swedish system as regulated, and this view is shared by many economists. Those who think that regulation of the housing market is good often do not consider the Swedish system as rent regulation. If the Swedish system is neither a rent regulation system nor a free market system, what is it?
According to Bengtsson (2016, p. 23-24), the Swedish system is a third type of rent setting system between regulation and free market. It is not rent regulation, as it is not based on public control but on bargaining between the landlord and tenant organizations. These negotiations are based on the assessment of the use value of the dwelling, and the qualities of housing according to which the assessment of the use value is made are such that they would have value in the market also. So, the negotiated rents are meant to reflect the potential market rent of the dwelling. Still, Andersson and S€ oderberg (2012) call the outcome of these rent negotiations 'surprisingly uniform ' and Bengs (2015) criticizes the centralized bargaining system for preventing rental markets from operating as open markets normally do.
Based on the brief survey of rent regulation in Nordic countries, can we say that there is a link between housing regime and rent regulation policy? The selective orientation in housing policy implies that most households are expected to provide housing for themselves by self-provision or through the private marketand public support is a special policy for those in greatest need. So, in such a regime the idea that the state should not interfere in the workings of the housing market seems quite natural. However, in Finland especially, rent regulation has in the past been rather persistent and its final abolishment is a part of the state retrenchment in housing policy. This may have to do with the fact that also 'type of housing regime' is an ideal construct that simplifies the complexities of real cases. Selective systems can have universal elements and vice versaand at times there might be attempts to change the policy orientation. Especially the cases of Denmark and Sweden, perhaps also that of Norway, seem to indicate that universalism in housing policy may coincide with persistence of rent regulationdespite the pressure towards retrenchment that affect all the Nordic countries. According to Christophers (2013), Sweden, the past model of universalist housing policy, has become a 'monstrous hybrid' where subsidies to housing provision have been cut down but rental market has not been deregulated.

Conclusion
Though it is quite true to say that the European housing systems have largely moved towards deregulation and neoliberalization, and that a close majority of countries have a free market in PRM, the policy of rent regulation has by no means disappeared from the scene. Sixteen of our 33 countries practiced second-or third-generation rent control, sometimes limited to particular parts of the housing stock. The example of rent regulation seems to indicate that the European housing provision systems are far from completely neoliberalized, and many systems could perhaps be considered as hybrid. However, our analysis shows that the generation of rent control approach has its limitation in classification of present day rent regulation systems. The range of second-generation rent control measures is very wide and the difference between second-and third-generation rent control is not always clear cut. In addition, new regulations in areas of high demand challenge old perceptions of rent regulation. The development of more nuanced classification could diversify comparative analysis in the future.
Belonging to a specific welfare regime type might be thought to predict whether rents are regulated or not. There is some convergence between countries in the same regime type, but there are also differences. Just like in some other aspects of housing (see Hoekstra 2005;Schwartz & Seabrooke 2008;Griggs & Kemp 2012) the regime typology does not seem to be the master key to explaining rent regulation policies. In order to take the analysis one step further, we took a closer look at the Nordic countries employing a typology of housing regimes. Nordic countries all belong to the Scandinavian welfare regime, but their housing regimes differ. To make comparison easier, a two-dimensional typology of housing regimes was constructed based on whether the policy is universal or selective and whether it uses as its institutional base renting or owner-occupation (or both). Only the policy orientation dimension seemed to be associated with regulation of rents. The concise case descriptions on the histories of rent regulation in the Nordic countries suggest that analysis on the level of types of housing regime need to be completed with more detailed studies of how national housing regimes have evolved taking into account contextual factors affecting social relations of private renting. This paper is a modest contribution to a comparative study of social relations of housing in the private rental sector. Regulation of private renting is, of course, more than just regulation of rents: it is also regulation of other aspects of the relationship between tenant and landlord, such as security of tenure (Haffner et al. 2018, Whitehead et al. 2012 or other nonprice regulation, e.g. letting agency regulation, registration of landlords, tenancy deposit schemes. The country reports of TENLAW project provide a rich material of the legal construction of private rental tenure in Europe. However, the legal framework is not the only aspect to be considered. Among other things, the position of private rental housing in housing provision as well as the degree to which the landlord and tenant interests are organized and represented politically have to do with how the landlord-tenant relation comes to be constructed in the PRS. We see comparative case studies as the most promising way of investigating the issue. Our analysis focuses on reviewing regulation policies. A further perspective on rental housing regulation would be to focus on housing outcomes. What are the consequences of the chosen regulation policy in terms of supply, housing quality and rent levels? A good regulation policy should guarantee that enough rental housing is provided for consumers, it is of good quality and affordable. An analysis with the focus on outcomes would require measures of supply, quality and affordability. A difficulty of such analysis would be that housing outcomes are not only a result of policies but also of the whole structure of social relations of housing as well as situational factors. The context-dependence makes it so that there is not necessarily one right policy that would guarantee the best outcomes. Notes 1. We prefer the term 'regulation of rents' over the term 'rent control' as a general concept referring to policies attempting to regulate the level of rents partly because in our national context the latter term refers to quite strict administrative control of rents with little concern for the return to the investment that the landlord might get. However, we are not entirely consistent with this usage as we will use the established terminology of 'generations of rent control' to refer to different forms that regulation of rents may take. 2. The power differential between landlord and tenant in an unregulated market is also the foundation of Tyler Zimmer's philosophical defense of rent regulation (see Zimmer 2017). Zimmer argues that tenants are relationally vulnerable in the landlord-tenant relationship. For example, the landlord can afford to walk away from a negotiation with a prospective tenant whereas the tenant cannot do this. Zimmer sees regulation of rents as one tool for balancing this power differential. 3. In the Appendix to the article there is a table where the division of tenures and types of rent regulation were cross-tabulated for our 33 countries. An earlier version of this paper included a section where we discussed whether there are some associations between tenure structures and rent regulation policies. About the only association there seemed to be was that countries with strict rent regulation systems tended to have high rental housing rates.
In this version of the paper this analysis is left out as it is a distraction from the general theme of regimes/rent regulation. However, the appended table is needed as a reference on the size of PRS. 4. Source EU SILC 2015 5. Our data separates England and Scotland that so far are part of one state, the United Kingdom. Our data does not cover Wales or Northern Ireland. 6. This category may include co-operative sector and other intermediate tenures as well as unknown tenure 7. In Sweden, co-operative apartments are counted as home ownership. Also, no distinction between private and public/social renting is made in legislation in Sweden. However, about half of the rental sector is owned by municipal housing companies. 8. In Norway, no distinction is made between private and public/social renting in legislation.
However, larger municipalities own a substantial proportion of rental dwellings.

Disclosure statement
No potential conflict of interest was reported by the author(s).