Subnational Social Investment in Three European Cities: An Exploratory Comparison

Abstract

With the evolution towards more service-intensive social investment welfare states across Europe, research on the institutional capacities of subnational welfare provision is increasingly relevant. Based on a comparative case analysis of three post-industrial municipalities in Europe, this article harbors a two-pronged objective: first, empirically, to show how regional and local governance capabilities are crucial to effective SI policy delivery; second, more positively, to bring out the proficiency of vertical coordination between national administration and subnational layers, alongside the critical role of horizontal policy discretion at the local level to align social benefits and capacitating services for the success of SI delivery; and, by implication, the overall responsiveness of national welfare systems to the changing nature of 21st century socioeconomic risks.

1. Introduction

Since the turn of the century the notion of social investment (SI) has gained considerable traction as a novel welfare edifice to address post-industrial economic and social change in an integrated fashion (Morel et al., 2012; Hemerijck, 2017) and, in the aftermath of the Great Recession, it has garnered augmented supranational political credit as a prerequisite for strategies of ‘inclusive growth’ (European Commission, 2013; World Bank, 2016; OECD, 2018). Since then, various efforts have been made to qualitatively trace and quantitatively assess SI reforms across OECD countries (among others, Vandenbroucke and Vleminckx, 2011; Deeming and Smyth, 2015; Busemeyer et al., 2019). This increasing amount of literature is painting a broad picture of the diffusion of SI policies across many countries and related outcomes in terms of human capital, family and employment services, individual well-being, inclusion of vulnerable groups and poverty mitigation. Yet, a profoundly relevant piece of the puzzle, pertaining to the governance nexus between national social security policy and local service provision, remains under-researched. This does not come as a surprise, as mainstream comparative welfare state research has its roots in national systems of social security, pensions and sickness benefits (Esping-Andersen, 1990). However, to the extent that the 21st-century SI welfare state conjures up a service-intensive welfare state, any assessment of national welfare state proficiency today can no longer ignore that subnational – local and regional – social service delivery is the pivoltal point for national social security.

This article addresses the subnational governance in SI research by shifting the focus from national social security administration to local institutional capacities in welfare service delivery. As such, the aim of the article is twofold: first, empirically, to expose the growing importance of institutional capacity and local mechanisms of governance (Crouch et al., 2004) for SI welfare provision; second, to underscore how vertical coordination between national and subnational levels of governance, together with horizontal policy complementarity in service provision in the local context, affect the success of SI policies and, by implication, the overall responsiveness of 21st-century welfare states.

Our explorative analysis highlights that even in the most centralised systems of public administration, such as France and Sweden, where redistributive and employment policies are delineated nationally, a wide range of measures – such as childcare services, active labour market policies (ALMPs) and even minimum income benefits – are provided, if not designed, locally. Subnational discretion and autonomy, if nested in a national administrative framework that facilitates vertical coordination, while allowing subnational actors and institutions to enhance tailor-made policy complementarity and synergies across benefits and services, boost SI delivery (Scalise 2020). As such, the ultimate success of national SI reform is crucially contingent on local delivery and governance capabilities.

The next Section discusses the SI approach and the rescaling process and administrative devolution which have amplified the local responsibility in social policy. Section three introduces the research design. Section four traces different experiences of subnational SI policy delivery in Gothenburg, Barcelona and Lyon, explaining how the three municipalities have pursued SI strategies over the past decade. In conclusion, we expand the comparative gaze on subnational SI and provide suggestions for future research.

2. Adjusting the social investment (SI) turn to welfare rescaling

SI has gradually become one of the most influential and discussed paradigms in the academic and political debates on welfare recalibration in post-industrial societies. The new emphasis on prevention and the adoption of a life course perspective in human capital development and female employment promotion represent a crucial turning point in the field of social policies, which triggered both endosements and critics. Some scholars define SI as the policy paradigm which counterbalances the neoliberal market-oriented welfare model (Morel et al., 2012) and acknowledge the returns of its comprehensive policy mix, based on the interdependence between education, labour market policies, poverty alleviation and work-life balance measures (Hemerijck, 2017; Kuitto, 2016; Plavgo and Hemerijck, 2021). Others, instead, have focused on the biases, shortcomings and unintended consequences of SI, criticising the superiority of ‘active’ SI over ‘passive’ social insurance spending (Nolan, 2013), questioning its political viability (Häusermann, 2010; Streeck and Mertens, 2011; Breunig and Busemeyer, 2012) and drawing attention to the potential ‘Matthew Effects’ inherent in certain SI policies, which favour the middle class and leave behind the most vulnerable groups (Jenson, 2009; Cantillon and Van Lancker, 2013).

SI has also raised theoretical and methodological questions on how to capture policy change and performance, as well as the interaction effects of the broad mix of policy instruments, which cut across serveral fields ranging from childcare and education to employment regulation, family and social services, and income transfers. While some perspectives adopt a narrow definition, excluding the compensatory component of the welfare state from the SI framework (Busemeyer and Garritzmann, 2019), other approaches (Eichhorst et al., 2021; Scalise, 2020) refer to Hemerijck’s (2017) stock-flow-buffers theory, which encompasses (1) measures in the arena of education and training, to raise and maintain the ‘stock’ of human capital and capabilities throughout the life course; (2) measures to ease the ‘flow’ of labour market and life-course transitions; and (3) inclusive safety nets as income protection and economic stabilization ‘buffers’.

This article builds on the latter definition of SI and on the literature which stresses that the success of this paradigm lies in the interplay between its comprehensive policy mix and the institutional and structural conditions and capability (Kazepov and Ranci, 2016). Retrieving the concept of institutional complementarity (Hall and Soskice, 2001), we argue that certain institutional combinations reinforce each other and contribute to improving or hindering the functioning and coherence of SI developments. In such institutional configurations we need to include territorial articulations and capacitating services. SI cannot function without both solid national social protection foundations and well-coordinated service provision, that are, most often than not, provided locally. Within a multilevel governance structure of education, labour market and social policy, established by the central state, the regions and the municipalities, SI policy delivery needs an integrated institutional capacity to effectively align stocks, flows and buffers, and to ensure consistent and coordinated policies. This is why, to understand concrete SI functioning, it is indispensable to take into account the subnational levels of regulation and local provision of welfare services.

While a relevant number of studies have traced and measured national SI-oriented welfare reform processes across OECD countries, a systematic study of the multi-level governance of SI and of its subnational regulation and delivery is missing from research. Some disjointed attempts to trace the implementation of specific initiatives and programmes, which report single experiences from local communities, have been made (Baines et al., 2019) and case studies on local social innovation practices illustrate governance dynamics and welfare mixes for poverty reduction in different contexts (Oosterlynck et al., 2019). However, a gap in the literature concerns how SI is shaped by subnational regulation and local policy delivery. This is increasingly problematic for understanding current European welfare states, since they are the result of policy rescaling (Kazepov, 2010), administrative devolution and reforms of service delivery that – since the 1990s – have started to transform the territorial organisation of social policies and the public and private actors involved in their design, management, implementation and funding (Gingrich, 2011).

The downward transfer of regulatory authority to regions and cities (Hooghe et al., 2010), with horizontal and vertical coordination and partnership-working, was encouraged by the European Open Method of Coordination and incentivized through the European Social Fund (Zeitlin et al., 2005). In two decades, territorial institutions have gone from providing residual social assistance and last resort schemes, to designing employment and social policies and services which complement national instruments and state contributory and insurance-based unemployment protection, losing their residual connotation. ¹ The surge of atypical contracts and tightened elegibility in benefit access, which excluded an increasing part of the population and those out of work from unemployment protection, has made non-contributory and means-tested schemes, for which municipalities are often responsible, ever more prominent (Natili, 2019). At the same time, the promotion of personalized and targeted interventions in social policy created spaces for local experimentation and bottom-up capacitating social services (Sabel and Zeitlin, 2017).

Vertical and horizontal coordination are key dimensions for ensuring coherence and consistency of standards in policy design and implementation within and across levels of government (Bouckaert et al., 2010; Rodrigo et al., 2009) and higher levels of coordination guarantee better quality policies and robust and resilient public administration (Christensen and Lægreid, 2006). According to Scharpf (1997: 133), an elementary political requirement for successful positive coordination across ministries is the recognition of the existence and legitimacy of transversal policy problems, in terms of both the distribution of costs and benefits and with respect to a problem-solving commitment to search for ‘positive-sum’ solutions. Progress in SI requires (a) inter-ministerial alignment of the policy functions of stocks, flows, and buffers, as they generally fall under the jurisdiction of several ministries; (b) vertical commitment to facilitate subnational discretion in local welfare provision; (c) semi-permanent ‘learning-by-monitoring’ policy feedback mechanisms, for the central level to keep an eye on output and performance and for subnational actors to be exposed to the full array of local SI initiatives, inspiring them to improve policy delivery. To the extent that a learning-by-monitoring arrangement allows for in-depth discussion of cause-and-effect relations, what works and what does not, also as regards feasibility and normative concerns, institutional trust across ministries and between national and subnational layers of government can strengthen transversal problem-solving (Sabel, 1994). These institutional prerequisites for effective SI may appear unrealistic in a world of self-serving rational choice, but as we will reveal below empirically, with respect to SI, horizontal inter-ministerial coordination, multi-level administrative cooperation and discretion, facilitated by effective policy feedback and learning mechanisms, in the common interest is not per se anathema.

3. Objectives, case-study design and methodology

Our exploratory empirical analysis aims at investigating SI local governance and delivery within different welfare state regimes and national institutional architectures. We opt for an empirical view of the SI local governance embedded in different national welfare states, and our approach is exploratory rather than hypotheses-testing. We follow a most dissimilar country selection with respect to national welfare systems but similar with respect to the local context. To this end, we have selected three regional cases in Western Europe: west Sweden, Catalonia and Rhône-Alpes ² with a strong focus on the three comparable post-industrial metropolitan cities of Gothenburg, Barcelona and Lyon.

More specifically, to investigate the role of the local level of regulation in developing context-specific stock-flow-buffer measures, the analysis is guided by the following exploratory research questions:

1. (1) How does the national institutional architecture affect the local SI governance?

2. (2) What margins of autonomy are granted to the local administration to deliver its SI strategy?

3. (3) What kind of complementarity exists across local SI measures?

4. (4) What is the role of the actors involved in the local SI governance?

5. (5) Can we identify a specific local capacitating and skills development strategy, which shows the local commitment to SI and an integrated approach?

To reply to these questions, both the regional (the county in Sweden) and the metropolitan levels have been taken into account, because the division of responsibilities in the welfare domain between these government levels is partially different in the three countries.

To operationalize the stock-flow-buffer architecture and adopt it as an analytical framework for the empirical research at local level, we focused on four policy areas (in which the regulatory competence is shared at different territorial levels) that have been analysed in a semi-structured manner:

1. (1) childcare policies (flow);

2. (2) secondary and adult education (stock);

3. (3) active labour market policies (ALMPs) (stock);

4. (4) income support schemes (buffers).

We decided to cover several policy arenas – instead of focusing on one and gaining more analytical depth and detail – since we are interested in exploring and collecting information on horizontal policy synergy between different fields which could represent an integrated SI local approach, and vertical coordination between levels of regulation which cuts across several policy fields (Table 1).

Table 1. Investigated features

The research is based on a combination of qualitative tools: in-depth process tracing and analysis of regional and municipal policy programmes, measures, schemes and regulations in the four policy arenas mentioned above; and content analysis with MaxQda software of 34 semi-structured interviews with regional and municipal policymakers, public servants, stakeholders and other relevant key informants (social partners, employment service and regional municipal staff, local scholars and experts) conducted in Gothenburg, Barcelona and Lyon between November 2015 and October 2018 (see list of interviews in Annex, Supplementary Materials) ³ . The collection of original qualitative information has also been combined with the analysis of OECD and Eurostat macro data on regional and city levels.

The three selected contexts are exposed to similar SI pressures, especially in terms of rising female employment and high-skilled labour demand, thus offering relatively favourable socioeconomic preconditions to SI. High levels of employment and equity and low child poverty make Sweden a SI vanguard country; France, the largest welfare state in terms of social protection spending, with a pro-natalist tradition, would seem to make it an obvious candidate for SI innovation; finally, Spain embarked on the road of SI reform in the early 2000s introducing ALMPs and improving childcare services, but when the Euro crisis intensified was forced to pursue austerity reform and welfare retrenchment.

According to Eurostat (2020), for all three regional contexts, total, youth and female employment rates are above national levels with unemployment below national averages; the population aged 25 to 64 has a higher educational attainment level compared to the national average (or the same as the national average, such as in west Sweden). Catalonia is characterized by a higher number of those not in education, employment, or training (NEET) (13.7 per cent in 2017) compared to Rhône-Alpes (8.9 per cent) and west Sweden (5.8 per cent) and has been more seriously affected by the consequences of the 2008 recession and for longer than the other two territories. At the time of the fieldwork, the three selected second-capital cities were governed by centre-left coalitions, with assertive social justice and equal opportunity normative orientations.

The distribution of competences between central and subnational levels in the three countries denotes subnational administrative autonomy with ample spending margins. According to OECD data, in 2016 subnational expenditure represented 25 per cent of the GDP in Sweden, 20.8 per cent in Spain (both above EU and OECD countries’ averages), and 11.1 per cent in France. As a unitary and decentralised state, the Swedish central government holds exclusive legislative powers, but the constitution recognizes local self-government. Counties have competence in the field of growth and development and municipalities hold powers in the fields of social welfare, economic development and education. In Spain the central state has competence in labour legislation and social insurance, but the Catalan autonomous community manages education and has full legislative power on social assistance, non-contributory pensions and unemployment benefits. The competences of the municipality of Barcelona are mainly related to the promotion of social reintegration and education. On the contrary, French subnational institutions do not have legislative power but exercise their functions by means of regulations and through the execution of their budget. Regions, departments and municipalities share responsibility in education, vocational training and apprenticeship, regional economic planning, inclusion and social welfare. In all cases these activities are financed through municipal taxes, government grants and charges: how financial resources are to be distributed is a political decision at local level.

4. Varieties of stock, flow and buffer welfare provision in post-industrial cities

OECD data on subnational spending give us a preliminary indication of the differing logic behind social policy investments in the three contexts. In 2016 subnational government expenditure in education represented 5.1 per cent of GDP in Sweden, 3.9 per cent in Spain and 1.7 per cent in France; subnational government expenditure in social protection was 6.6 per cent of GDP in Sweden, 2.2 per cent in France and 1.5 per cent in Spain (Table 2). In Sweden, social expenditure as percentage of total transaction of the same level of governance in the same year was 16.3 per cent; 13.6 per cent in Spain; and 10.5 per cent in France. However, if we focus only on subsidies, we observe that expenditure for this type of buffer was 15.1 per cent of the total subnational spending in France, 14.2 per cent in Spain and only 4.4 per cent in Sweden (Table 3).

Table 2. Subnational government expenditure by function (% of GDP) in 2016

Source: OECD.

Table 3. Subnational expenditure in 2016 (as % of total transaction, same level of government)

Source: OECD.

The following analysis confirms that SI progress is not simply a matter of how much of their resources governments devote to various measures, but that the institutional architecture and governance logic behind SI allocation, together with local actor engagement, are essential to create effective synergies across complementary policy measures.

4.1 Gothenburg’s integrated social investment strategy

Within the overarching national commitment to inclusive welfare provision, Gothenburg municipality delivers its own strategy to facilitate a favourable economic life cycle and well-being returns. This is made explicit by one of the civil servants interviewed in the municipality: ‘Here the main strategy to avoid poverty and social exclusion is to invest in education, training and services more than in subsidies and to provide policies, such as childcare and parental benefit, that facilitate parents’ labour market participation (…) and establishing a strategic policy framework helps ensure that individual policies are consistent with the government’s goals and priorities’ (Gothenburg City, April 2016).

Embedded in the national active welfare state tradition, Gothenburg’s governance approach combines different national and local policy tools to deliver high childcare takeup, guarantee near universal secondary youth education, and boost general skills expansion.

With regard to childcare and education, the city integrates national funding (1.6 per cent of GDP, the highest in the EU) with municipal fees – established on income levels – in order to guarantee the right to childcare from a very early age (12 months in Sweden, 3 years old in France and Spain). The municipality ensures a nearly universal enrolment and a low child-to-staff ratio (in formal day care services for 0 to 3 years old is around 5:1 and for preschool around 11:1, lower than in most other OECD countries) and provides child and school day care on a full-time basis, allowing parents to work full-time or extended part-time.

The school system, which was decentralised in the early 1990s, allows families to choose a school within the city, and municipal after-school care is offered to all children up to the age of 12. The municipality is the main provider of adult education, which is grant aided and free of charge. As explained by one of the interviewees within the municipality staff, the particular local commitment in this field is due to the fact that ‘The high educational threshold to enter the labour market has been identified as one of the major problems in Gothenburg’ (Gothenburg City, April 2016). This results in high educational attainments and participation rates (fig. 1).

FIGURE 1. Population aged 25–64 by educational attainment (%, 2017)

Source: Eurostat

Policy complementarity is particularly developed in two arenas, adult education and labour market policies. Since 2014, the city has decided to keep these two areas together, developing a new division, the municipal Labour Market and Adult Education Unit, to improve and exploit synergies between the two policy fields and coordination between the different actors involved, ‘adapting tertiary education to the needs of local economy and working life and reducing the differences in educational choices and attainments between genders’ (Gothenburg City, April 2016). The unit coordinates the different actors involved in these fields at different levels: city districts, the National Public Employment Service, local social partners, universities and other education providers, the Västsverige Region, and the Västra Götaland County. As for ALMPs, parallel with the governmental labor market policy system, municipalities develop a wide range of local work-related activation programs for unemployed social assistance recipients, and are granted rather extensive freedom to determine both the organization and the delivery of activation measures, which range from mandatory work requirements to training and education. For instance, as labour market integration of newly arrived refugees and migrants is one of the major local concerns, the municipality has developed a comprehensive targeted pilot project to improve migrant social and economic inclusion since 2016.

Vertical coordination with the national level of regulation does not prevent the municipality having extensive discretion in deciding over benefit levels, duration and incentives to get beneficiaries into employment, and it has the legal right to impose sanctions (i.e. reducing social assistance payments). Interviewees refer to ‘strict eligibility criteria and exacting activation requirements’ (Gothenburg City, April 2016). Indeed, the Labour Market and Adult Education Unit also acts as an interface between public employment service (PES) and the local social welfare agencies to manage, administratively and financially, the social assistance programme. This targets people in households that lack sufficient means to support themselves, not just from work income, but also via access to contributory social insurance benefits. This municipal buffer policy secures adequate household and individual income during risky transitions and has unlimited duration. However, in accordance with the national welfare logic, activation prevails over the passive supports, and cash and care are closely intertwined, aimed at identifying and treating the underlying causes of economic hardship. This implies that ‘social assistance benefits are intimately tied to the local activation programmes’ (Gothenburg City, April 2016).

The development of capacitating provisions is rooted in corporative governance practices reproduced at local level. Social trilateral negotiation allows local government to identify the needs of the local potential workforce and to design tailored interventions based on participants’ goals, which include personalized assistance for coaching and training, guidance and counselling, job matching and practice: ‘A personal supervisor gives people in need support entering the labour market, checks skills and shows appropriate projects or opportunities to improve skills and meet employers’ needs, help people with reduced working capacity or disability or who have been unemployed for a long time or been away from working life because of illness to find the best solution for their cases.’ (Gothenburg City, April 2016)

Local policy makers’ and stakeholders’ collaboration also concerns the review and monitoring of policy performance, creating monitoring and reporting databases to track the experience of individual participants, and to codify best practices: ‘Implementation procedures and monitoring mechanisms are designed to ensure that policies can be adjusted in the light of progress, new information, and changing circumstances.’ (Policy advisor, Gothenburg University, April 2016)

Notwithstanding this integrated local approach to SI, recent reforms in welfare and education are somehow undermining important components of the Gothenburg model. Following the 2008 economic crisis and the exceptional influx of migrants to Sweden in 2015, conditionality criteria were tightened in Gothenburg to contrast the increased caseloads and social assistance expenditures. The activation requirement was used to limit social assistance claims and make labour market inclusion faster, ‘increasing also the pressure on individual obligations and the potential risk for unstable job positions.’ (LO Swedish Trade Union Confederation, Gothenburg, April 2016) As some interviewees confirmed, socioeconomic and housing segregation and ethnic-based disparities in terms of living conditions, education and income have increased in the city. Inequality is reproduced among children of immigrants, who are more social and economically vulnerable compared to native children (Gothenburg City 2015). The introduction of free school choice in the municipality has amplified these disparities and the differences in terms of school results, which in the long run may provoke intensified inter-generational disadvantage.

4.2. Barcelona’s assertive social investment ambitions dashed by institutional fragmentation and the eurocrisis

Spain has achieved substantial welfare expansion since the end of the 1980s, embracing in recent decades the SI logic. Its deep transformation includes ground-breaking legislative initiatives that have been able to concretely promote female labour force expansion (risen by 50 per cent between 1999–2007) work-family life balance, and ALMPs (Guillen and Leon 2011). The strong decentralisation constituted both an opportunity and a constraint for local governance renewal. On the one hand, it creates opportunities for local policy innovation and activism in the welfare arena. On the other, where taken too far, devolution could hamper effective policy coordination and territorial solidarity.

Indeed, two of the greatest limits to effective SI delivery in Spain are the lack of coordination between central and subnational initiatives and cross-sectoral policy fragmentation, which result in a patchwork of programmes that have been inadequate to respond to the 2008 crisis-led rise of unemployment and poverty. Additionally, the fiscal consolidation since 2010 has impacted drastically on welfare provisions, undermining the continuation of delivering social services and ALMPs.

Childcare represents a clear example of SI ambition hampered by both financial constraints and administrative fragmentation. In 2004 the Catalan region launched a programme to create 30,000 childcare places over the period 2004–2008. In 2008 the central government provided an additional 50 per cent in co-funding to promote early childcare (‘Educa3’ programme). ‘These measures allowed the city of Barcelona to start increasing the number of public nurseries and to rebalance between private and public services.’ (Ajuntament de Barcelona, October 2018) However, the national programme was cancelled in 2012 and the coverage of public childcare service has never reached the level necessary to meet local demand. ‘In Barcelona competition for places in nurseries is high,’ (Ajuntament de Barcelona, October 2018) and the child-staff ratio is higher than the EU average (1:9 for children under one year old, 1:20 for two- to three-year-olds). Priority to single parents and low-income families is given for places in public early childcare centres (where children are accepted from five months, although standard maternity leave lasts four months and paternity leave is fifteen days). Non-profit subsidized childcare, for-profit services, and informal childcare provided by relatives supplement formal childcare, with relevant differences in terms of costs, affordability and quality.

Progress and regress also characterize the field of education, which has experienced significant transformation over the last 20 years: ‘The region concretely encouraged especially tertiary education, which has grown almost seven-fold, but the regional action is not included in a national systemic logic … We lack an administrative culture that promotes cooperation and a systematic dialogue between different policy communities and contributes to the strengthening of policy coherence.’ (Policy advisor and professor, Barcelona, November 2015) Yet a large percentage of adults have no upper secondary education (see figure 1) and the region is affected by high numbers of early leavers from education. Education still receives low investment in comparative terms: regional expenditure is below the Spanish and EU averages (in 2016 it was 3.9 per cent of GDP in Catalonia, 4.4 per cent in Spain and 5.1 per cent in the EU) and investment per student has fallen between 2010 and 2015 (Vilalta and Comas 2021).

The link between education and labour market policies is weak and incomplete and uncoordinated reformism permeates ALMPs as well. At the end of the 1990s, ‘The flexibility of the labour market was complemented by national and local activation measures … and benefit conditionality was introduced to try to incentivize skill enhancement and employability … but these different measures do not seem to be working efficiently … and are not integrated with each other.’ (Spanish Confederation of Employers’ Organisations, October 2018) Indeed, different programmes in this field are managed by a variety of actors – national PES, regional employment services and the municipality through its ‘Barcelona Activa’ agency – who do not cooperate effectively and thus deliver disjointed income support and upskilling measures.

The outbreak of the financial crisis exposed this fragmented governance system. To respond to the increased unemployment level, after 2008 a more intrusive workforce activation strategy prevailed. Local policies prioritize supply-side measures, focusing more on employment assistance and placement services than upskilling and prevention, increasing job seekers’ vulnerability: ‘Merely moving into work did not lead to improvement. The stories of people shifting from benefits into low-grade, temporary work and back to benefits again are widespread here.’ (Confederaciòn Sindacal de Comisiones Obreras, Barcelona, November 2015) At the same time, local employment and social services could not enforce conditionality, given the massive hike in unemployment, although the requirement to actively seek work was written into the national unemployment benefit statute. ‘Benefits have already been curtailed by austerity measures, sanctions are not applied. We also have a relevant problem with the lack of personnel in the public sector. It is not possible to provide real support to every unemployed person nor to apply these requirements.’ (Ajuntament de Barcelona, November 2015) According to interviewees, this requirement was interpreted by local public servants as a way of increasing the risk of social exclusion and stigmatization of non-compliant recipients. Moreover, an insufficient number of public servants in local administration made it difficult to guarantee efficient matching between training and job offers for the unemployed, and hindered effective control over requirements. Despite the various programmes offered, the participation rate in training programmes is still very low: according to Eurostat, in 2017 only 7.7 per cent of the working population in Catalonia took part in training in the previous four weeks, against 31.3 per cent in west Sweden.

Measures aimed at favouring female employment and supporting low-income families with children have experienced a similar fate. National and regional cash transfers, set up between 2005 and 2010 (i.e. tax relief for working mothers of children aged 0–3 years old and a one-off universal allowance at the birth of a child), were reduced due to consolidation measures and the bodies created to promote gender equality were downgraded (Salazar Benitez, 2016). The lack of coherent and integrated policies deeply affected the SI buffer function. ‘Low-intensity and highly fragmented forms of buffers which last for a limited number of months and are not renewable’ (Confederaciòn Sindacal de Comisiones Obreras, Barcelona, November 2015) have not been effective to prevent a deterioration in the living conditions of large parts of the population during the crisis.

All in all, institutional fragmentation has adversely affected the efficacy of SI governance in Barcelona. Additionally, while in the 2000s the Catalan region was characterized by strong social partnership, which promoted measures based on enabling ALMPs, the 2008 crisis triggered a conflict between the local government and unions. The latter lost legitimacy and bargaining power and the local administration reduced the room for coordination. The marginality of social partners in local governance exacerbated the lack of coordination among key players for policy delivery and restrained monitoring.

4.3. Lyon’s failure to transcend the French siloed welfare policy legacy

Comprehensive social security provision, historically grounded in a pro-natalist policy legacy with ample investments in education, in combination with female employment above the EU average, should together constitute a fertile environment for SI progress. Yet, in terms of institutional and governance organisation, a rigid, and extremely centralised, system of public administration (despite various reform attempts), and the relative generous compensatory benefits, have made it extremely difficult for Lyon to adapt its local welfare governance to the SI imperatives of post-industrial cities. Contrary to the west Sweden and Catalan cases of ample administrative discretion, the pervasive role of the French state at all territorial levels leaves little space for autonomous action to subnational government to deliver innovative social services and respond to localized social problems.

Childcare exemplifies the rigid institutional architecture in the French welfare state tradition. The national educational system splits its management between different public institutions ⁴ that co-fund and develop the regulations for the diverse types of centres where children are accepted from 3 months old, displaying a combination of subsidized centre-based and home-based arrangements. As in Barcelona, in Lyon demand exceeds availability of places in the public centres for children aged 0–3 years (co-run by the municipality, departments, and non-profit organisations). Alternative day care options are offered such as publicly subsidized home-based care by qualified nannies (assistantes maternelles) and non-profit nurseries run by parent associations (crèches parentales). ⁵ To respond to increasing inequality in terms of socioeconomic and territorial distribution of places, Lyon municipality tried to prioritize the increase of pre-school centres located in disadvantaged areas, where the childcare attendance rate is lower. However, local measures need to be coordinated with the national compensatory education policy, which directs the strategic orientation of the local administration, constraining the municipal scope for action. ‘Low income families and child poverty are concentrated in certain areas of the city … where job uncertainty and poverty are higher than the regional average, but competence constraints make it difficult to intervene with selective and prompt measures in these areas.’ (DIRECCTE Rhône-Alpes, Lyon, January 2016)

Fiscal and social redistribution for families has a critical re-balancing effect on income inequality. Despite redistribution, high expenditure in education and financial support for students, inequality is being reproduced. ‘In what we call ‘sensitive urban zones’ we find a larger proportion of school dropouts and remedial-year pupils. Entrance to the upper secondary pathway is heavily influenced by family economic and social background, which is reflected in the low upward intergenerational mobility in educational attainment for children of low educated parents.’ (Mission Emploi-Formation PRAO, Lyon, January 2016)

Increasing disparity in education is a priority issue in the political agenda and the local government is committed to expand the synergy between education and labour market. In the framework of the national ‘Territorialisation des politiques emploi-formation’, in 2015 the regional public service for orientation (PRAO) was set up and the region became the cornerstone for coordination of training policies and job assistance. Numerous local initiatives to improve employment access and training were enhanced ⁶ but interviewees explain that the complex institutional layering and alignment imperatives, which involve various actors with diverse responsibilities (i.e. de-localized national services, territorial branches of national agencies, region, departments, the Metropole de Lyon agglomeration) make it difficult to coordinate national and local initiatives and achieve the national goals. Interviewees reported redundancy of measures, low participation in programmes, little effectiveness in matching skills and jobs and weak capacity to approach unemployed people. Local stakeholders also complained of a lack of inclusion in local governance and unilateral decision-making: ‘Our involvement is a standardized practice …. We are called to meetings but we do not have a real influence on labour market and social programmes. The region and metropolitan area do not listen to us and duplicate state intervention, through the distribution of standardized aid to the territory.’ (Confédération générale du travail, Lyon, January 2016)

Activation and benefit conditionality have also been gradually introduced. The complex system of social protection (nine different income schemes for which the benefit amount varies depending on the applicants’ income and family composition) guarantees a relatively long benefit period and generous eligibility criteria. Already the first Minimum Insertion Income (RMI), launched in 1988 for working-age people who had no right to unemployment benefits, contained an insertion clause requiring claimants to formally engage in integration activities. However, the ‘reciprocal engagement’ between claimants and the state made the degree of obligation required of claimants rather ambiguous. The Law 758/2008 on Rights and Obligations for Jobseekers (Loi relative aux droits et aux devoirs des demandeurs d’emploi), and the Active Solidarity Income (RSA), which replaced the RMI in 2009, introduced the ‘reasonable work offers’ that claimants have to accept in exchange for income support, but ‘Income support beneficiaries are only formally obliged to undertake the actions necessary for a professional insertion; in common practice this is considered only a condition for access to financial support, not to put pressure on the unemployed … There is no sanctioning or compulsory work.’ (PLIE UNI-EST, Lyon, January 2016) Despite efforts to emphasize the potential of benefit conditionality, in practice it has a largely symbolic function. The contractual dimension is intended as an agreement that allows claimants to maintain their unconditional right to an income.

The logic of compensation prevails over prevention and activation at all territorial levels, as confirmed also by the response to increasing unemployment subsequent to the outbreak of the 2008 crisis, which was not only based on job matching, training and assistance, but also on extensive local state-aided job creation (i.e. Travaux d’Utilité Collective, Contrats Emploi-Solidarité, and Contrats Emploi Consolidé). New ‘supported contracts’ for inclusion in social life (CIVIS) and integration clauses in regional and municipal invitations to tender by the public sector stimulated the inclusion of the long-term unemployed, income benefit recipients, low skilled, and young inactive, providing an alternative to market employment through subsidized temporary jobs. ‘The prevailing objectives are still employment security, protection and compensation.’ (Conseil régional Rhône-Alpes, Lyon, January 2016)

The different out-of-work benefits combined with state-aided job programmes together guarantee ex-post compensation, but the limited vocational training capacity reveals the gap between active risk-prevention and passive risk-mitigation. However, weak local autonomy, poor inclusion of social actors in the local governance, increasing dualization of the welfare and labour market (Emmenegger et al. 2012), and social conflicts around the recent labour-market and pension reforms, are increasingly questioning the sustainability of the French model of social inclusion.

5. Discussion and conclusion

This exploratory work brings the analysis of local welfare policy into the heartland of SI to highlight the critical importance of effective governance structures to make SI work in practice. What this article demonstrates is that the local institutional capacity and territorial forms of governance affect how post-industrial metropolitan areas can adopt varying ‘fit for purpose’ SI strategies. Local government institutional capacity and autonomy are essential for effective SI delivery: although regions and municipalities are nested in the national policy framework and the central state’s coordinating function is crucial for ensuring overall coherence, local discretion and cross-sectoral alignment across the SI functions of stocks, flows, and buffers ultimately define the quality of SI delivery.

The Gothenburg case demonstrates how a calibrated balance between a functional vertical coordination and municipal autonomy with a strong focus on operational alignment foster greater capacity to deliver capacitating services which support citizens’ life course transitions. Gothenburg underscores the awareness of local actors of potential policy synergy, for which a transversal and extensive system of cooperation between local public authorities and stakeholders is essential. A horizontally joined-up governance regime, embedded in a vertical commitment to SI, based on solid national backing of local discretion in cross-sectoral collaboration and dialogue, allows Gothenburg’s administration to deliver services which aim at enhancing human capital, boosting productivity and employment quality by better understanding the skills that the local economy needs, increasing knowledge and job opportunities (Table 4).

Table 4. Local institutional capability and SI commitment

The case of Barcelona shows that an assertive political will to move towards SI at the local level is not sufficient. Poor vertical coordination between central state, the Catalan Autonomous Community and Barcelona municipality, hampers local SI delivery. High local autonomy, in a context of irregular administrative relations, generates fragmented and isolated policy interventions with poor performance, resulting from the many overlaps which hinder recipients of benefits in orienting themselves between the different measures. During the early 2000s, the local state promoted social trilateral negotiation to improve employment participation and job quality, supporting ALMPs and new local income support schemes. Then the 2008 crisis triggered a divergence of interests between political and social actors which undermined the negotiated activation strategy previously agreed, and the city returned to concentrate on providing benefit handouts, revoking its earlier commitment to SI.

The hierarchical coordination and siloed local policy complementarity characterizing the Lyon case study reveal that the centralistic French state tradition continues to stifle effective local SI delivery. In particular, we observe how the siloed organisation of stock, flow, and buffer policies at the national level is reproduced at the local level, disabling the emergence of a dynamic and innovative governance structure for integrated measures. Additionally, local governance, based on a formalized and standardised system of alliances composed mainly of public actors, leaves very little institutional space for discussion with local social actors and stakeholders.

The common thread that emerges from our study is the inter-organisational and multilayered governance required for any effective SI strategy. National political will is not a sufficient condition for SI delivery success. SI reform clearly requires national backing, but when it comes to delivery, building local capacities on the basis of ‘place-specific’ know-how, the national consensus should bolster – not undermine – political discretion at the local level to sustain social dialogue and public-private collaboration and experimentation, including feedback monitoring on both policy progress and the governance experience with all the stakeholders involved.

Our exploratory findings highlight a critical issue that should be a focus for future research. Despite the limitations of this study – which covers only Western European cities with common socioeconomic characteristics – it provides the foundations to open up a new research agenda for the bottom-up analysis of SI policy alignment and delivery that would be particularly significant to address growing territorial inequalities. The overall strength of comparative welfare research lies in its appreciation of deeply-ingrained national policy legacies. This strength harbours a weakness. The welfare state is changing in myriad ways. In the aftermath of the Great Recession, there is a growing understanding that the welfare state should capacitate people over the life course in the form of income buffers, but even more so with respect to service provision. With the general shift to welfare services, the subnational layer of welfare governance has gained in importance. Subnational variation across European countries in SI policy delivery thus deserves a prominent place in the comparative study of changing welfare states. Obviously, the architecture of national welfare regimes bears heavily on the design of subnational welfare governance. Yet, with the general diffusion of SI reform across Europe, there are strong empirical reasons for adding the subnational welfare layer of governance to comparative welfare research. The theoretical merit for such a research agenda bears on the (im-)possibility to align positive inter-ministerial coordination at the national level, together with support from national governments, to subnational SI discretion, facilitated by ‘learning-by-monitoring’ policy feedback arrangements to improve SI policy and delivery.