Housing Regime Typologies and Their Discontents: A Systematic Literature Review

ABSTRACT

Housing regime typologies represent a key conceptual construct in the comparative housing research literature, whose classificatory and explanatory capacity is still the subject of lively debate. The variety of methods and results amassed in the field, alternatively dismissing or supporting their usefulness, makes it difficult to discern the classificatory and explanatory power of various housing regime approaches. To this aim, this study offers a systematic literature review of housing regime theorization and its applications. In doing so, we make a case for the use of systematic literature reviews in housing studies, still rare in the field. We thus map, compare and evaluate the use of such typologies by looking at how they were used in clustering national housing systems, and in empirical research testing the regime-related hypotheses. Our results point to the enduring relevance of housing regime theory in comparative housing research, although from a historically grounded perspective.

KEYWORDS:

• Housing Regimes
• Varieties of residential capitalism
• Comparative housing analysis
• Housing theory
• Systematic review
• Scoping review

1. Introduction

The merits and limits of regime typologies are subject to a longstanding debate in housing studies, a field in which the development of cross-country classifications for comparative research found particularly fertile ground. The theoretical construction of classificatory typologies and models represents a fundamental theme in comparative social research more broadly. Ideal types are supposed to grasp underlying regular patterns and social mechanisms across multiple geographical and time settings (Aspalter 2012; Bengtsson and Hertting 2014), making it a useful analytical heuristic. The identification of cross-country classifications, like Welfare State regimes (Esping-Andersen 1990) and Varieties of Capitalism (Soskice and Hall 2001), arguably gave a crucial impetus to the development of Comparative Welfare and International Political Economy studies.

Similarly, housing regime (HR) typologies fuelled the debates around the same rationale and goals of comparative housing research, influencing the different theoretical and empirical trajectories in the field. Such ideal types have been used to organize the empirical material around a set of hypothesized causal factors and explanatory mechanisms (Ruonavaara 2020, 6–7). The analytical and explanatory capacity of regime typologies, however, is a recurrent bone of contention. The field has amassed a significant number of studies, based on different methods and leading to different results, alternatively dismissing or supporting the validity and usefulness of regime approaches.

A still unattempted way to assess the use of regime typologies in housing can be found in the systematic literature review methodology, aiming to systematically map and assess all the relevant literature in a specific field as a premise for developing evidence-based and rigorous research. Systematic review studies are receiving a growing interest in the comparative housing field (Soaita et al. 2020), emerging as a promising way to strengthen its methodological foundations.

This article presents a systematic literature review of the HR typologies in comparative housing analysis. Our aim is twofold. First, we make a case for the use of systematic literature reviews in housing studies, which are still rare in the field, by using the PRISMA standardized framework for mixed-methods systematic review. Second, our review identifies the most frequently used HR approaches in the literature and assesses their use, both as classificatory and explanatory frameworks. To this aim, we provide a mapping of the most diffused HR classifications within the comparative housing literature in a broad timeframe (1990–2021). Then, we compare and evaluate the use of such typologies from two central angles. We first assess the degree of overlap and disagreement regarding the clustering of national housing systems in different studies to expose the prevailing patterns and problems in the classificatory exercise. Then, we look at how HR types have been empirically tested against a broad set of housing-related outcomes and with what results.

The results of this review point to the enduring relevance of HR typologies in comparative housing research, although under precise conditions. Rather than offering an exhaustive picture of housing systems, regime typologies prove helpful as a tool to track cross-country trends, continuities, and changes in housing provision, policy and markets, allowing scholars to test specific explanatory hypotheses on the relationships between welfare, financial and housing systems.

2. Methodology

Systematic approaches to literature reviews aim at providing transparent and potentially replicable mapping, reporting of trends, and meta-level analysis of findings and gaps in a given research domain, to establish a solid ground for further theory development (Denyer and Tranfield 2009, 672). Yet, the definition of uniform criteria in the review work remains a paramount concern for qualitative and mixed-methods studies in social and political science (Dacombe 2018, 154; Wallace et al. 2004). Building on a preliminary mapping of systematic review methods on housing issues in Social and Political Science (see Appendix 2), we adopt the PRISMA protocol, as the standardized guidelines for mixed-method systematic reviews most widely used and recognized by scholars (Moher 2009). In addition, we conduct a systematic qualitative synthesis to provide a comprehensive interpretative analysis of our literature sample (Saini and Shlonsky 2012, 84–87). It is important to note that our review mapping and analysis do not entail any claims about the truth or validity of the HR theories.

In line with PRISMA reporting recommendations, we set out the research questions, the search criteria and sources, together with the resulting sample used in the subsequent meta-synthesis. Our review aims at mapping, comparing, and evaluating the use of HR typologies in cross-country comparative housing policy analysis. We address three research questions:

• RQ1 - What are the most used HR approaches and typologies in the literature?

• RQ2 - How are they used to classify countries from a comparative perspective, and with what results?

• RQ3 - How are they tested empirically against housing outcomes, and with what results?

To include all the relevant cross-country comparative literature, we used the following search terms: “housing*”, “regim*”, “system*”, “model*”, “typ*”, “residential capitalism*”. We combined these terms using the Boolean operators according to the different databases’ search engines. Although a conceptual distinction has been made in the housing literature between “regime” and “system” (Kemeny 2001, 59), we put these terms on the same level, so as not to miss any relevant study. Thus, “housing regime” refers here broadly to the defining features distinguishing how housing provision and consumption are organized across different national settings.

We conducted our search on three broadly used databases in Social and Political Science: Scopus, Web of Science, and the International Bibliography of Social Sciences (IBSS). To ensure broad coverage of the debate in the Social Science domain, while looking at the documents’ quality and potential reach, we adopted the following exclusion criteria: 1) documents not in English; 2) publications not pertaining to the Social and Political Science fields; and 3) non-peer-reviewed articles and “grey” literature (like conference proceedings and PhD theses). Our search covers the years from 1990 to 23 November 2021. The base year 1990 was chosen as the year of publication of The Three Worlds of Welfare Capitalism (Esping-Andersen 1990), whose welfare regime typologies opened the debate on housing regime classifications (Kemeny [1995] 2002).

We further screened the resulting sample according to additional eligibility criteria, retaining studies containing: 1) a comparison of at least two countries; 2) typologies conceived for more than one country (so avoiding national-focused types, e.g. that of a “Norwegian” regime). Lastly, we operated a “snowballing” check of eligible documents which possibly escaped our database search, by looking at the references contained in our 117 items. We obtained then our final sample (n = 127, see Appendix 1), as noted in Figure 1.

Figure 1. PRISMA flow Chart.

Next, we coded the documents’ full text. In our coding strategy, we looked at 1) the HR approaches and typologies used; 2) the explanatory mechanisms associated with them; 3) the HR-related country classifications; 4) the dependent and independent variables; 5) the methodology adopted. Our units of analysis included both single terms (like the country names) and core sentences containing the needed information (like explanatory mechanisms). In the first coding session, we inductively identified the definition and variety of the above-mentioned themes. After having finalized our codebook, we rechecked the whole literature sample by revising and integrating the codes where needed. The different stages of the literature selection, screening, and coding have been cross-checked independently by both authors.

3. Results

3.1. Mapping the Debate

As Figure 2 shows, the majority of the documents in our sample were published after 2010 (84 out of 127), potentially spurred by the renewed interest in housing finance after the global financial crisis and the Varieties of Residential Capitalism (VoRC) approach (Schwartz and Seabrooke 2008). The debate primarily took place within the housing field (n = 77 items), but HR studies also gained importance in Political Science (n = 12) and other Social Science academic outlets (n = 17) from 2008.

Figure 2. Retrieved documents per year.

Table 1 offers a summary of the most frequently applied HR approaches and their associated typologies, main variables and hypothesized mechanisms. Here, we outline five fundamental dimensions shaping the debate on HRs, as identified through our mapping.

Table 1. Most diffused HR approaches, typologies, dimensions and mechanisms.

Approach	Typologies	Housing Dimensions	Core Mechanisms
Housing-Welfare (n=79) [primarily theory-driven]	Tenure forms (Kemeny 1995) - Unitary - Dualist - Integrated Liberal/Corporatist (Kemeny 1995, 2006) - Liberal (dualist) - Corporatist (unitary/integrated) Capital-led Non-profit influenced Balanced Non-profit led Labour-led Non-profit dominated	Housing tenures Rental housing providers, markets and regulations Political system (degree of fragmentation) Balance between capital and labour interests	Trade-off: inverse relationship between homeownership rates and role of welfare state and pension provision. The growth of homeownership favours negative attitudes towards welfare state and pension contributions. Unitary/Dualist dynamics: Dualist regimes lead to residual social rental housing and homeownership as the dominant form. The unitary rental system, on the contrary, may develop into an integrated rental sector, with the non-profit sector successfully competing with the for-profit sector, promoting cost-rental housing. Divergence trend: housing systems’ divergence; however, policy responses to not-for-profit housing maturation could lead to convergence towards dualist systems
Varieties of Residential Capitalism (VoRC) (n=17) [primarily theory-driven]	VoRC typologies (Schwartz and Seabrooke 2008) - Liberal Market - Corporatist Market - Statist developmentalist (financially repressed) - Familial Financialization patterns (Dewilde and De Decker 2016) - Financialized/commodified - Less financialized/de-commodified Financialization Trajectories (Fernandez and Aalbers 2016) − 4 scales, from low to high financialization	Homeownership rates Mortgage market indicators (debt to GDP, securitization) Housing finance providers (State, banking systems, financial markets)	Trade-off: Housing finance markets, equity and wealth accumulation negatively impact welfare state and pension provisions (asset-based Welfare) High housing financialization and poor public/social housing provision lead to high social stratification, together with housing markets’ insider/outsider and inter-generational inequalities Convergence trend: convergence towards housing financialization, leading to interlinked divergent housing provision patterns
Housing Provision (n=21) [primarily analytical framework]	Housing Production and Supply (Barlow and Duncan 1994; Arbaci 2007) Structure of Housing Provision (Harloe 1995) - Residualization - Mass-model Housing Provision Chain (Doling 1999) -Liberal -European Mainland/Corporatist -Little Tigers (South-East Asian Powers)	housing development (funding sources and public/private provisions) housing construction (regulations and main real estate agents) housing consumption (tenure forms and housing access factors)	Analytical framework, looking at the relationships between state and market in each stage of the housing life-cycle [no actual theoretical mechanisms defined] Convergence trend: housing as a marketized form of welfare provision, linked to capitalist accumulation processes. Residualization of social housing as the ‘normal’ form in capitalist economies (Harloe 1995)
Southern European/Familial (n=32) [mixed, theory-driven/analytical framework]	- Southern European/Familial (Allen etal. 2004) - Familial (VoRC)	Role of familial welfare Weak State housing provision intergenerational transfers Fragmented labour markets	Familial regime dynamics: 1) Inverse relationships between housing affordability and quality; 2) Reproduction of labour-market fragmentation and gender-based inequalities; 3) Inequalities in independent housing access for young adults and market outsiders Divergence trend: financialization processes dividing Southern countries (e.g. Iberic countries vs Italy and Greece)
Central-Eastern European (n=25) [mixed, theory-driven/analytical framework]	Eastern European Housing model (Hegedüs and Tosics 1992) - Classical - Reformist Post-socialist trajectories (Mandič 2010) - Central Eastern European - Southern Eastern European - Northern Eastern European	Trajectories of post-socialist privatization of housing Role of the State in housing provision Role of family and informal housing Restitution policies	Transition regime dynamics: 1) Privatization of former public/social housing stock in post-socialist countries; 2) Super-homeownership ‘trap’: privatization of public-social housing stock and restitution policies leading to the marginalization of other housing forms (private and social rental); 3) Financialization/commodification leading to increasing inter-generational and insider/outsider housing-related inequalities Divergence trend: commodification trajectories shaped by previous socialist housing models

Theory-driven/Analytical approaches (first column, Table 1). Regime typologies have been used either as part of theory-driven explanatory mechanisms linking together housing and other social/political factors (n = 62), or as a tool to investigate housing systems (n = 56) with an “analytical approach”. The former option was opened by Kemeny’s work in his re-elaboration of Esping-Andersen’s welfare regime framework (Kemeny 1995). According to his “Housing-Welfare approach” (Stephens 2016), key housing dimensions (like tenure forms or funding sources) deeply affect broader social and political domains (like welfare policies). Testable hypotheses are thus provided to explain the system-wide impact of the housing sector. In addition to the Housing-Welfare approach, VoRC-related HRs emerged as the second main theory-driven approach. Both of these can be contrasted with the analytical approaches, which focus on the factors and actors defining a housing provision and consumption system without elaborating specific housing-based explanatory mechanisms (e.g. Ball and Harloe 1992, 13–14; Barlow and Duncan 1994). This perspective lacks a coherent and distinctive theorization of the housing-social system relations while providing an analytical framework to be applied with “external” macro-level social and political theories. Recent works has called for more analytically rich and multi-level HR theory-building (Hegedüs 2020; Stephens 2016, 2020a, 2020b). On the other side, the analytical perspective remained a baseline reference for theory-led HR studies and received a new impulse from the VoRC strand and its linked approaches to housing finance systems and financialization (Johnston and Kurzer 2020). Indeed, the two alternative perspectives are increasingly intertwined in the development of regional-oriented studies (like the Southern and Central-Eastern HRs) and mixed-method research.

Dimensions of housing provision and regulation (second, third and fourth columns, Table 1). In general, theory-driven approaches look at a few dimensions and related variables to account for housing-welfare mechanisms, while analytical ones, by definition, consider a broader range of factors to examine a housing system. The Housing-Welfare approach (n = 79) brought out the centrality of rental housing funding and regulation, shaped by capital-labour relationships and political party systems, as the key factor shaping the whole housing tenure system and, through it, households’ attitudes on public spending and welfare policies (e.g. Kemeny 1995, 2006; Dewilde 2017). Policies promoting the dualization between non- and for-profit rental housing are hypothesized to shrink social housing and push up private rental prices, thus making homeownership a preferable alternative for the middle classes. In turn, homeownership costs are assumed to favour negative attitudes towards governmental taxes and spending undermining financial stability, while the acquired housing wealth slackens the demands for higher pensions (and propensity to pay for pension contributions). On the contrary, in unitary regimes, state subsidies to cost-rental housing and regulation of private rents lead to more affordable and quality rental housing, lowering the demands for owner-occupation.

The centrality of homeownership and the financial markets’ role in shaping housing systems are the focus of VoRC-related studies (n = 17). Here, mortgaged access to owner-occupation, housing equity accumulation and investors’ penetration into the real estate sector are nodal points linking global financial markets to households. Increasing financialization of housing provision has been linked to rising socioeconomic inequalities along class and inter-generational dimensions, the spread of asset-based welfare policies and households’ exposure to financial markets’ risks and speculation (e.g. Fernandez and Aalbers 2016; Johnston and Kurzer 2020; Schwartz and Seabrooke 2008).

On the other side, the dynamics of non-financialized access to owner-occupation in the context of poor public intervention and fragmented labour markets have been investigated in the Southern European/Familial regimes (n = 32), whereby housing provision depends mostly on familial transfers and inherited wealth. Family-based housing tends thus to reproduce and exacerbate long-term economic, labour and gender inequalities, delaying or hindering independent housing for young adults, particularly for those deprived of parental financial support (e.g. Allen et al. 2004; Norris and Domański 2009). While sharing the attention to familial resources, analysts of Central-Eastern European housing systems (n = 25) looked more at State privatization and financial liberalization policies as specific historical factors leading to the rapid creation of private housing markets out of the former public housing provision (e.g. Hegedüs and Tosics 1992; Hegedüs, Lux, and Horváth 2018).

Core/periphery (second column, Table 1). While the first wave of HR studies focused on a few Western countries, significant revisions to HR theorization came from scholars who looked at European peripheries. Such regional-oriented approaches dealt in particular with Southern and Central-Eastern European countries. These studies built on the main theory-driven and analytical HR methods to further elaborate new typologies and hypotheses. Recent developments point to a further broadening of the geographical perspective, particularly towards East Asian countries (n = 8) (e.g. Ronald and Doling 2012; 2014; Lee 2018), together with a reorientation of comparative housing analysis from the national to the local dimension (Hoekstra 2020), and the adoption of a global core-periphery perspective (Kováts 2021; Tulumello 2021).

Convergence/divergence (fourth column, Table 1). Kemeny first traced a convergence/divergence dichotomy when he criticized housing scholars’ assumptions about the development of national housing systems towards a common model (convergence thesis). Against that, scholars advanced a “middle-range” theorization looking at the divergence of housing paths (Kemeny and Lowe 1998). The unitary/dualist typology aimed to grasp the fundamental cleavage shaping housing-welfare trajectories in modern societies. The divergence thesis then came to dominate the debate (Bengtsson and Ruonavaara 2010). However, from the 2000s onwards, such a dichotomy gradually gave way to a less rigid understanding of the tensions between convergence and divergence. In fact, Kemeny recognized the possible decline of unitary systems towards convergent patterns of residualization (Kemeny 2001, 2006). From an International Political Economy view, VoRC then focused on the globalization forces brought by housing finance markets and their translation into distinctive domestic financialization trajectories (Aalbers 2015; Fernandez and Aalbers 2016; Johnston and Kurzer 2020). A redefinition of the convergence/divergence problem thus surfaces from those empirical studies that investigate the interactions among entrenched housing paths and cross-border trends leading to increasing housing-related inequalities (see below). Such a changing perspective is mirrored in our literature sample, in which convergence (n = 36) is investigated more than divergence (n = 20).

Methodology. The majority of reviewed studies make use of qualitative research approaches (n = 84), most of which employ historical and cross-country comparative analyses (n = 65). A minor share of works explicitly relies on path dependency (n = 27), which represents a widespread model in housing studies (Bengtsson and Kohl 2020). On the other hand, quantitative techniques are less numerous (n = 50). Among the latter, the use of various statistical models prevails (n = 23), followed by quantitative descriptive analysis (n = 21), particularly in combination with historical and cross-country qualitative comparisons. Quite surprisingly, cluster analysis (n = 14) and factor analysis (n = 4) appear relatively poorly exploited in the HR literature, despite their respective relevance to empirically test HR groupings and relevant variables. Similarly, works building on mixed qualitative and quantitative methodologies represent a minority in our sample (n = 19).

3.2. Country Comparisons

A crucial testbench for HR approaches lies in the degree of consistency and stability of their proposed country classifications (Ball 2020, 39–40; Dewilde 2017). In this section, we compare the classification of countries according to the HR typologies identified in Table 1. We consider a sample composed of the most cited countries in our literature sample, presented according to three geographical and linguistic areas: the English-speaking core countries, Western/Northern and Southern Europe. To better check different classificatory exercises, we thus focus on the most frequently examined countries, keeping in mind that the HR literature shows a tendency towards Western countries. Thus, we look at the studies which grouped national housing systems by exploring the relationships between key regime characteristics with different data and variables. By analysing the agreements and disagreements in the HR country classifications, we scope out their degree of consistency and stability.

Figure 3. HR classification of selected English-speaking countries.

Source: authors’ elaborationNote: The “VoRC” column combines all VoRC-related HR typologies identified in Table 1.

The English-speaking countries, shown in Figure 3, broadly share regime categorizations associated with dualist housing tenures, liberal market arrangements, and financialized systems. However, significant disagreements surface in the UK and Irish cases. While the British housing system from the 1980s has been considered a typical case of a liberal regime, the inheritance of a former strong public housing sector (Kemeny 1995, 106–111; Stephens 2020a, 538–40) still makes the UK a dualist system with strong social housing. As of the year 1991, Doling (1999) calculated a higher score in his de-commodification index for the “dualist” UK than for the “unitary” Germany or Sweden: the opposite of what was expected according to Kemeny’s theory. Similarly, by looking at affordable and quality housing for low-to-moderate-income households, other scholars clustered the UK from the end of the 1990s as a country with a dualist “but fairly large social sector” (André and Chalaux 2018, 32; Dewilde 2017, 393, 397; Dewilde and De Keulenaer 2003, 147). In his criticism, Ball showed how the high level of outright homeownership and intergenerational financial transfers would pull the UK next to Southern European housing systems (Ball 2020, 39).

Sitting somehow between the Anglo-Saxon financialized regime and the de-commodified familial one, Ireland shows more conflicting categorizations. Schwartz and Seabrooke framed Ireland as a case of a “familial” regime by capturing only the initial process of its financialization (referring to 1992–2002). The subsequent mortgage market boom and collapse radically changed its housing system profile (Bohle 2018, 214; Fernandez and Aalbers 2016; Johnston and Kurzer 2020, 285–86; Norris and Byrne 2021).

Figure 4. HR classification of Western-Northern European Countries.

Source: authors’ elaborationNote: The “VoRC” column combines all VoRC-related HR typologies identified in Table 1.

The classification of European Western-Northern countries, represented in Figure 4, is more challenging, mirroring the transformations and inner tensions in their domestic housing systems. Critical examples of unitary and integrated rental systems until the 1990s, such as Sweden, Denmark, and the Netherlands, were later classified as dualist/commodified. According to some, these three countries experienced shifts towards a liberal market expansion of homeownership and mortgage markets (Wind, Lersch, and Dewilde 2017, 630). Sweden represented Kemeny’s prototype of unitary (Kemeny 1995, 116; Balchin 1996, 69) and integrated regime (Andersson et al. 2007, 154; Kemeny 2006, 14; Kemeny, Kersloot, and Thalmann 2005, 864–6). Similarly, according to the VoRC classification, referring to data from the 1990s, Sweden appeared between the statist-developmentalist and the corporatist-market types (Schwartz and Seabrooke 2008). Yet, other studies delineate the shift of the Swedish system into an increasingly dualist and residualized system in the 1990s (Doling 1999, 163), providing a case against Kemeny’s expected resilience of integrated rental systems (Stephens 2020a, 543; Whitehead 2020, 576). Recent studies frame Sweden as a hybridized financialized/unitary regime (André, Dewilde, and Luijkx 2017, 248; Dewilde 2017, 396; 2020, 90; Elsinga and Lind 2013, 960; Musterd et al. 2017, 1067–8) and highlight its Neoliberal turn (Clapham 2019).

Similar classificatory conflicts emerge in the other Continental housing regimes. The Netherlands switched from a unitary type to a financialized and increasingly dualist system (Dewilde and De Decker 2016, 136; Elsinga and Lind 2013, 968–9; Fernandez and Aalbers 2016, 83; Kemeny, Kersloot, and Thalmann 2005, 869; Wind and Dewilde 2019, 544). The difficult pigeonholing of some Continental systems is well shown in the case of Germany, which swings from a unitary type tending towards residualization (Angel 2021, 13–15; Dewilde and De Decker 2016, 136; Van der Heijden 2002, 338) to a dualist variety with significant social rental sector (Dewilde 2017) or a highly regulated private sector (Wind and Dewilde 2019, 544).

Figure 5. HR classification of Southern European countries.

Source: authors’ elaborationNote: The “VoRC” column combines all VoRC-related HR typologies identified in Table 1. An additional “Southern European” bar is added to this figure to illustrate the frequent use of this classification when studying Southern European countries.

On the contrary, Southern countries, shown in Figure 5, demonstrate less variegation. A cleavage nevertheless emerges along the commodification/familial axis. Some recent studies classified Spain and Portugal along with the liberal-market and financialized Western regimes (Alves 2017, 243–4; André and Chalaux 2018, 24; Fernandez and Aalbers 2016, 81; Johnston and Kurzer 2020, 286; Lennartz, Arundel, and Ronald 2016, 832; Wind and Dewilde 2019, 542). As these authors highlight, from the 1990s onwards, the Spanish and Portuguese governments underwent a radical turn in liberalizing their housing finance systems, pointing at a convergence dynamic. The consistency in the Italian and Greek cases suggests relative HR stability (Allen 2006; Allen et al. 2004; Dewilde 2017; Flynn 2020).

As shown, classificatory disagreements correspond to housing systems which underwent major structural changes from the 1990s. In other cases, characterized by more continuities or homogeneity, HR classifications are more consistent. As HR types “fix” particular housing characteristics in time, conflicting classifications may point to significant changes that occurred within and across housing systems. However, in other cases, like the UK or Germany, the problematic labelling into dichotomous categories would signal flaws in the underlying HR theory.

3.3. HRs and Housing Outcomes

The relationship between regime types and real-world expected outcomes represents the second main way to assess the use of HR approaches (Dewilde 2017; Flynn 2020; Stephens 2016, 26). This section looks at how the defining variables and mechanisms of theory-driven HR typologies have been tested against housing tenure, policy, and inequality outcomes. A table summarizing the relevant studies’ findings is available in Appendix 3.

The works that aim at empirically testing HR approaches represent about half of our literature sample (n = 62), primarily published after 2009 (n = 48). In terms of methodology, qualitative studies prevail (n = 35) compared to quantitative ones (n = 27). The housing outcomes in our sample concern mostly housing affordability and quality (n = 22), followed by housing tenure composition and developments (n = 15), the welfare state (n = 13), housing policy (n = 11), ethnic segregation (n = 2) and fertility (n = 1).

Housing-Welfare typologies are tested with the greatest frequency (n = 34). Significant evidence supporting housing-welfare hypotheses has been detected in 21 cases (58%), while five studies found no significant relationships. The remaining 8 records only partly confirmed their validity. Most criticisms point to the Housing-Welfare approaches’ underestimation, or neglect, of the role of housing finance systems and their developments, which are, on the contrary, the focus of VoRC. Significant associations between VoRC typologies and housing-related outcomes were found in half of the 14 analysed documents, with the remaining studies finding either non-significant relationships (n = 3) or problematic ones (n = 4). The original VoRC approach has been criticized for the excessive narrowness of its explanatory variables, to which the subsequent financialization studies, and particularly the housing provision and regional approaches’, counterpose a broader range of housing dimensions. However, the closer the latter move to an “analytical” framework, the less falsifiable their hypotheses become. Therefore, Southern/Familial HRs (n = 8) show only one problematic result (linked to VoRC financialization dynamics), together with the Housing Provision approach (n = 4), while Central-Eastern regime studies (n = 10) were generally associated with significant results. Similar characteristics are found in the remaining HR cases, like the “Homeownership-based” regime studies (n = 2, based on Mulder and Billari 2010) and other housing regime types (n = 3). Below we focus mainly on the Housing-Welfare and VoRC-linked regimes, as the two theory-driven approaches.

As for housing affordability and quality outcomes, the majority of studies found unitary/integrated housing systems had better housing conditions compared to dualist ones (Alves 2017, 244; André and Chalaux 2018, 32; Borg 2015, 90; Lux and Sunega 2014, 51; Dewilde 2017, 397–8; Hoekstra 2009, 58–9; Mandič and Cirman 2012, 789; Milligan, Dieleman, and van Kempen 2006, 252). This is in line with the Housing-Welfare expectations about escalating price dynamics and decreasing regulatory pressures in dualist rental systems leading to deteriorating affordability and quality outcomes. On the contrary, however, Norris and Winston (2011, 136), rejected hypotheses derived from Kemeny’s theory, having found no systematic differences in homeownership distribution and quality conditions for low-income owners between unitary and dualist regimes.

The structural crisis of unitary models (and related theorization) unleashed by the housing financialization process has been captured mainly by VoRC-inspired studies. In line with the VoRC expectations about the inverse relationships between better housing conditions and financialization, Dewilde and De Decker (2016, 143–44) show how financialization can explain declining housing affordability in former unitary systems and its worsening in established dualist markets. Even along the gender and generational inequality dimension, Liberal/financialized regimes have been associated with lower homeownership access for divorced women than men (Dewilde and Stier 2014), and for young adults (Dewilde 2020; Lennartz, Arundel, and Ronald 2016). Conversely, contrary to the VoRC expectation, scholars detected a positive association between financialization and housing quality, with households in more financialized systems experiencing fewer housing quality problems (Dewilde and De Decker 2016, 145). Such a positive association between good quality conditions and financialized systems was already noticed by Norris and Domanski (2009, 404), who argued that broader homeownership access through mortgage markets improved dwellings’ quality and redirected state housing investment towards neighbourhood regeneration. Problematic results have been found in clustering country price developments (Springer and Wöhl 2020) and patterns of young adults’ housing access opportunities (Flynn 2020), which only partly correspond to the expected VoRC groupings, thus bringing out the need to revise them.

Clustering based on housing outcomes showed the distancing of Spain and Portugal from the former Southern familial HR (Dewilde 2017). As predicted, familial regimes have been related to unequal housing access conditions and standards (Dewilde 2017; Norris and Shiels 2007, Norris and Domanski 2009; Hoekstra 2005), and high residential segregation (Musterd et al. 2017). Similarly, transition countries experienced deteriorating housing access conditions, depending on the former soviet trajectories and privatization dynamics (André and Chalaux 2018; Mandič and Cirman 2012; Norris and Shiels 2007).

As regards to housing tenure outcomes, Housing-Welfare expectations on the dualist dynamics of the social rented sector’s residualization were confirmed (Angel 2021; Hoekstra 2009; Van der Heijden 2002). Similarly, focusing on price trends, unitary regimes were associated with stability and dualist ones with more volatility (Norris and Byrne 2018). On the other side, studies in transition countries criticized Kemeny’s approach for underestimating the policy factors shaping the residualization of the private rental sector in the European post-socialist systems (Lux et al. 2012, 156). However, one of the most comprehensive criticisms of Kemeny’s approach came from Stephens (2020a). Stephens’ historical account refuted the centrality of the cost- and for-profit rental sector dynamics in the case of the post-War UK. Moreover, it criticized Kemeny’s “Panglossian error” about the expected self-sustaining capacity of the mature integrated system in front of the Neoliberal turn and the related pressures on welfare state spending (Stephens 2020a, 541–2). Actually, Kemeny himself cast doubts on the survival of integrated housing systems facing maturation crisis by stressing the ultimate role of political decisions (Kemeny 2005, 71–3; Kemeny, Kersloot, and Thalmann 2005, 866–7). However, he still failed to properly acknowledge, and integrate into the theory, the convergent pressures stemming from the globalization of finance, something that Stephens stresses when calling for a revision of the Housing-Welfare approach. Scholars also advocate major reformulations for VoRC, whose focus has been considered either too narrow to account for the resilience of the social rented sector (Norris and Byrne 2021) and the emergence of highly financialized corporatist regimes (Anderson and Kurzer 2020) or unfit to explain owner-occupation rates (Cooper and Kurzer 2020) and their historical relations with mortgage debt (Blackwell and Kohl 2019). Greater attention to regional and historical contexts is claimed likewise by Central-Eastern HR approaches, which stressed the specificity of transition countries’ rental residualization processes (e.g. Hegedüs, Lux, and Horváth 2018).

Mixed findings also emerge in relation to housing policy outcomes. Dualist systems have been found to discourage adequate private rental regulatory standards due to lacking competition coming from the social rental sector (O’Sullivan and De Decker 2007). Yet, Kettunen and Ruonavaara (2021, 1462) contradicted this finding in their analysis covering Nordic European countries, where the authors found no relevant associations between housing tenure forms and rent regulation. Regarding households’ policy preferences, however, André and Dewilde explored a key implication of the trade-off hypothesis, finding mortgaged homeowners held more negative attitudes towards welfare redistribution than tenants across most of the analysed countries (except for 9 Scandinavian and CEE countries) (André and Dewilde 2016, 338). In addition, Liberal and homeownership regimes have been found to show the largest tenure gap in electoral turnout, with homeowners more likely to vote than tenants, compared to familial and unitary regimes (André, Dewilde, and Luijkx 2017).

Looking at welfare-related outcomes, different studies validated Housing-Welfare HR mechanisms. Refining the trade-off arguments, Doling and Horsewood showed how house purchase prices, more than the percentage of homeonwership, accounts for the homeownership-welfare trade-off mechanism, with growing housing prices negatively related to the level of public pension provision (Doling and Horsewood 2011, 180). On the other side, Doling and Ronald (2010) showed how housing wealth’s capacity to reduce poverty risks is linked to national public welfare settings, thus casting doubt on asset-based welfare potentials to provide universal social security. Asset-based welfare has been found to lead to growing inequalities in case of economic crises (Ronald and Doling 2012). Linked to housing wealth accumulation, the promotion of a competitive social housing sector has been considered the critical condition to address the housing wealth gap increase between homeowners and tenants, more than asset-based welfare strategies (Wind and Dewilde 2019, 561–2). This conclusion resonates with Wind et al. (2017), who showed an increase in housing wealth inequalities to be pronounced in Liberal VoRC regimes. The enduring validity of the trade-off mechanism, however, has been questioned by Blackwell and Kohl (2019, 305–6). Through an analysis of bivariate relationships between homeownership rates and welfare expenditures over 40 years, the scholars show that Kemeny’s expected negative association would hold just until the beginning of the 1980s. From the income inequality perspective, on the contrary, Wood (2017) identified a negative effect of high mortgage stock on the wage share of GDP in VoRC Liberal regimes, supporting the related hypothesis about the disciplining role of debt on workers’ demands. Conversely, housing deprivation constitutes a core dimension of poverty in Familial countries like Belgium and Spain (Dewilde and De Keulenaer 2003).

4. Conclusions

This study systematically reviewed the use of housing regime typologies in comparative housing studies. By adopting the PRISMA guidelines, we supported the adoption of a transparent review methodology and qualitative meta-synthesis to map and assess the research trends, achievements, and challenges in one important subfield of comparative housing. Accordingly, we presented a comprehensive picture of the HR debate, teasing out the approaches, typologies, and related explanatory mechanisms. Their use has been assessed by analysing the way scholars classified national housing systems and then comparing the findings resulting from studies that engaged in HR empirical testing.

Overall, we showed how the diversity of HR country classifications and explanatory mechanisms face enduring challenges linked to the variables considered and the transformations of housing systems. Housing-Welfare and VoRC “theory-driven” approaches suffer mostly from a narrow selection of housing dimensions and explanatory variables, which hardly stand up to the test of a wider historical perspective or the critical socioeconomic changes occurring from the 1980s. Yet, the definition of regime typologies and causal mechanisms in comparative research always entails an unescapable theoretical trade-off. While the broadening of housing dimensions and variables enriches the empirical scope of such HR constructs, it runs counter to the definition of clear and testable hypotheses. For example, Stephen’s proposed multi-level “housing provision” framework (Stephens 2020b) would ensure a comprehensive focus on the housing dimensions, while leaving unspecified the formulation of clear explanatory mechanisms and testable hypotheses. Such an exclusive analytical perspective crucially discards the case for theory-based housing comparative analysis, including HR approaches.

While exposing the shortcomings and challenges of theory-laden housing regime approaches, this review points to the continuing relevance and future developments of such perspectives, complementing more empirically detailed HR analytical frameworks. We observed how most evidence has been collected on the enduring significance of the Housing-Welfare and VoRC-related theorizing, particularly on the deepening housing access and wealth inequalities in dualist/Liberal regimes and the most financialized unitary and familial ones. Scholars working on housing financialization and related inequalities indicated potential directions to revise and enrich Housing-Welfare and VoRC classificatory and explanatory mechanisms (Blackwell and Kohl 2019; Fernandez and Aalbers 2016; Flynn 2020; Johnston and Kurzer 2020; Norris and Byrne 2021).

Most notably, the shortcomings of HR theorization suggest a more accurate framing of its historical outlook. Rather than fixing ahistorical models and mechanisms, HR types help formulate historically situated hypotheses about continuity and change dynamics, which require testing and refining along different time frames. The HR synchronic classifications capture specific historical conjunctures of housing systems which are fundamental to assessing ruptures and path dependency, as well as convergence and divergence pressures. HR theories can contribute to the analysis of broader housing change dynamics, like those linked to financialization processes, by offering testable expectations to guide empirical, historical, and policy-oriented research.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14036096.2023.2265370

Additional information

Funding

This work was supported by the Luxembourg National Research Fund [14345912].