Coping with the Global Economic Crisis: The Regional Political Economy of Emergency Social Shock Absorbers in Italy

ABSTRACT This article contributes to the debate on governing the global financial crisis, focusing on the regional governance of emergency social shock absorbers in Italy. The article seeks to make two related contributions. First, it argues that subnational governments have been the main drivers of change in labour market policies. Second, it shows that state–local governance elicited a path-altering system by ‘patching up’ a hybrid administrative structure and by ‘converting’ the traditional goals of social shock absorbers from income maintenance to welfare-to-work. The article provides qualitative evidence on the changing organizational bases of the labour markets of two large Italian regions: Lombardy and Emilia-Romagna. Evidence suggests that administrative innovation and path dependence intertwined in the governance of the global economic crisis in Italy, mitigating the entrenched distortions of labour market policies.


Introduction
The outbreak of the global economic crisis in 2007-2008 severely hit South European welfare states, with rising unemployment and poverty rates (Gutiérez, 2014;Petmesidou and Guillén, 2014). Most countries introduced emergency measures to fill in the gaps of existing coverage systems by adopting short-time work policy schemes (Hijzen and Venn, 2011). The purpose of anti-crisis packages was to maintain employment relationships while subsidizing a reduction in work time. In line with this trend, the Italian government relaxed regulations by extending coverage and duration of benefits (Mulé and Di Stefano, 2014). Policymakers were able to quickly implement emergency social shock absorbers (ESSA), aimed at preventing a total collapse of the economy. 1 Analysts assert that the policy of ESSA "clearly constitutes a major institutional innovation in Italian labour market policy" (Sacchi et al., 2011: 478). The reason is that the policy instrument that makes up the crisis-related 'emergency social shock absorbers' differs from the 'social shock absorbers' established before the crisis to mitigate the effects of job loss and industrial restructuring (the wages guarantee fund, job security agreements, early retirement and mobility schemes). The novelty introduced by ESSA lies in the mandatory link between passive and active policy instruments. Reform of the traditional social shock absorbers had been on the agenda of national governments since the late 1990s, with an emphasis on the need to transform them from passive to active instruments (Gualmini, 1998). This article argues that the global economic crisis of 2008 finally encouraged the policy shift in the shape of 'emergency social shock absorbers'. Consistent with theoretical developments in the literature on institutional analysis, economic shocks may represent a critical juncture triggering path-altering policy trajectories (Mahoney and Thelen, 2015).
Yet, despite their importance, scholars have paid scant attention to ESSA. There are two reasons for such lack of interest. On the one hand, there seems to be an ambiguity regarding the drivers of policy change; on the other hand, it is unclear whether the policy change is incremental, with the patching up of administrative structures, or innovative, implying a conversion of policy goals. In Heritier's terms, patching up is an incremental change that consists of adding new elements with new underlying principles to old structures ruled by old principles. "The result of 'patching up' existing structures as a response to required changes are hybrid administrative structures consisting of various and diverse administrative principles and policy instruments" (Heritier, 2001: 54). It should be noted that as a mode of institutional change patching up differs from 'layering'. The former refers to hybrid administrative structures, coexisting without necessarily being incoherent, while the latter features institutions with a "highly disjointed pattern" and "incoherence" characterized by a mechanism of differential growth, where new institutions eventually crowd out the traditional system (Streeck and Thelen, 2009: 119). While both patching up and layering are incremental modes of change, patching up neither sets in motion differential growth, nor is it intended to supplant the traditional system. Patching up refers to the manufacturing of hybrid administrative structures.
The second mode of institutional change brought about by the implementation of ESSA is "conversion". In Streeck and Thelen's terms, conversion is a different mechanism of change from patching up because institutions are "redirected to new goals, functions, or purposes" (Streeck and Thelen, 2009: 121, emphasis in original). Policymakers deploy existing institutional resources to new ends, altering the core objectives of a policy programme.
This article argues that patching up and conversion are different facets of state-regional labour market policies in the aftermath of the global financial crisis. Our work shows that Italian policymakers responded to the economic downturn by orchestrating a 'hybrid system' of ESSA, characterized by patching up of central and regional administrative structures and by conversion of policy goals from income maintenance to welfare-to-work. It should be noted that earlier social shock absorbers provided income support with no conditionality. Thus, regionalization coupled with policy conversion have established institutional and cultural frameworks which may have paved the way for the recently issued Jobs Act (law 183/2014), the Renzi government labour market policy reform.
This article contributes to the literature on political responses to the global economic crisis. Much of this literature centres on responses at national level and our work brings a much-needed subnational focus. It builds on Pérez and Rhodes' discussion of the impact of the crisis on Italy's "southern European social model" (Pérez and Rhodes, 2015). Exploring hypotheses from the literature on government crisis response and the impact of reform on social policy, they argue that Italy has benefitted from being able to fall back on institutional legacies to help manage the crisis but, significantly, they provide no examples of this. They argue that the recent decentralization of social policy competences to local government has empowered existing local social networks. This article reinforces and develops their argument on institutional transformation in crisis response empirically, by showing that the path-altering trajectory of social shock absorbers consolidated local identities by bolstering pre-existing network coordination policies.
Network coordination is due to central governments entrusting regional policymakers with discretionary power in setting eligibility and entitlement rules to ESSA. This twist in the governance of social shock absorbers means that regions should be conceived of as 'spaces for politics and policies' with distinct logics and strategies of action in the field of social policy. 2 Such process of decentralization is in line with European trends towards devolution of social policies to the subnational level, transforming the welfare 'state' into welfare 'regions' (Ferrera, 2008;Kazepov, 2010;Wright, 2014).
Recent research findings add weight to previous work showing that the success of regional involvement in Italy stems from the quality of subnational institutions (Fargion and Gualmini, 2012). This observation raises the question of how to select the regions for comparative analysis given that, qualitative comparative work, involves purposeful, rather than random, selection of the cases. One answer is to adopt a comparative strategy based on the 'most similar systems design', which compares similar cases that are as similar as possible to each other. The logic behind the 'most similar systems design' is that those factors held constant through the selection of cases cannot explain observed variations, thus eliminating a number of possible explanations. By holding some variables constant, researchers more productively isolate the factors responsible for variation between the case studies (Sartori, 1991).
Yet, it is important to acknowledge the limits of the chosen research design. The 'most similar systems design' does not allow us to control whether the argument holds across contrasting cases, thus the researcher may extrapolate regional differences that are not so sharp nationwide. However, the case selection in our empirical work does not aspire to offer a representative sample of Italian regions. It is well known and documented that Italian regions are characterized by significantly different political, institutional and economic performance (Vassallo, 2013). Against this background, it is unlikely that any region could be selected as a 'representative case'.
Our work follows a different strategy by adopting the 'most similar systems design' in order to keep at least some crucial variables constant. Such variables in our two case studies are political stability and high levels of economic performance, enjoyed by both Lombardy and Emilia-Romagna. Between 2009 and 2012 about a third of total ESSA benefits (28.5%) were given to workers employed in micro and small firms located in these two regions (Institute for the development of workers' professional skills-Istituto per la formazione professionale dei lavoratori ISFOL, 2014: 37), and in 2009 a total of 84 610 ESSA benefits were received by workers employed in these two regions (ISFOL, 2012, Tab. A.5).
We believe that for the purposes of our empirical work the advantages of the chosen comparative method outweigh its limits for two reasons. Firstly, regional political stability means that the implementation of ESSA cannot be explained by changes that may have occurred in the party affiliation of regional political leaders. Second, the similarity in political stability between Lombardy and Emilia-Romagna adds simplicity to our research efforts as does comparing similar levels of economic performance.
The data we draw upon consist of qualitative research methods that integrate analyses of national and regional documents, reports and laws, with in-depth semi-structured interviews of experts selected by relevance, i.e. according to their role in implementing the ESSA. 3 We interviewed leaders of the social partners, Trade Unions and Associations of Employers, as well as regional administrative elites.

Insiders, Outsiders and 'Mid-Siders'
The unemployment benefit system in Italy is well known for its weaknesses, such as the dual labour market, with some sectors protected and others left with no safety nets. One reason for such weakness is that unemployment benefits have never been part of a 'system' of benefits but have instead been extended and rearranged, if and when, economic crises urged government intervention.
Like other Bismarckian welfare regimes, in Italy unemployment benefits are tied to occupational status (Esping-Andersen, 1990;Ferrera, 1993;Gualmini and Rizza, 2013). One consequence is the formation of a dual labour market characterized by the lucky 'insiders' and the unlucky 'outsiders'. Insiders are core workers (public employees and workers employed in medium-large firms) who benefit from a high level of job protection and are also entitled to generous income replacement. By contrast, 'outsiders' have low levels of job protection and cannot rely on any kind of income protection. This duality between insiders and outsiders has been central to labour market policy research. Recently, however, Jessoula et al. have argued that there is evidence of a third category of workers, the 'mid-siders', "who do not have access to the same (high) level of protection as the insiders, but, on the other hand, are employed and can be entitled (to a certain extent and with variations) to some contributory unemployment" (Jessoula et al., 2010: 563-4). The authors divide the category of 'mid-siders' into two subgroups: workers employed in micro and small firms, with lower levels of job protection and much less generous unemployment benefits than the insiders, and the heterogeneous world of 'atypical workers '. 4 This article focuses exclusively on 'mid-siders' of the first subgroup, i.e. workers employed in micro and small firms, because ESSA mostly affected this subcategory. Jessoula et al. maintain that the numerical relevance of these workers in the Italian case supports the adoption of a new analytical category. Table 1 illustrates that in 2009 the average firm size was 3.9 employees and 94.8% of firms had fewer than 10 employees.
Moreover, according to the ISFOL report of 2014, workers employed in micro and small firms constitute about 40% of total employment. Between 1 January 2009 and 31 December 2012 there were 1 339 253 ESSA benefits with a total of 824 000 beneficiaries of which 310 000 (37%) benefitted from more than one ESSA. In the same period 156 000 firms were authorized to use ESSA (ISFOL, 2014: 34), enabling firms to continue their productive activity without closing down. ESSA preserved human capital because it allowed the reduction of working hours or the suspension of the worker for a period. In addition, ESSA beneficiaries were able to re-qualify and retrain. These observations indicate that ESSA was indeed a relevant measure against the hardship caused to small firms and mid-siders by the global economic crisis.
Traditionally 'mid-siders' had not been eligible for benefits perhaps because it was thought that the economic system would be strong and flexible enough to re-employ small numbers of workers if and when such firms went bust. But the unemployment insurance scheme that is still the core institution of Italian labour market policies was built in quite a different era. The underlying assumption was that periods of unemployment would result from short-lived economic downturns. This assumption was largely accurate in the post-war period of economic boom and full employment. However, under radically different economic conditions the unemployment insurance scheme in its conventional form proved to be ill suited to performing the same objectives.
For this reason, regional governments resorted to ESSA specifically for unemployed 'mid-siders'. 5 It should be noted that the traditional unemployment protection system in Italy was established only for permanent workers employed in firms with more than 15 employees. Hence, coverage was chiefly dependent on firm size. In this scenario, regional governments introduced ESSA with a dual aim: to provide income maintenance to unemployed or suspended workers employed in firms with fewer than 15 employees, but also to help micro and small enterprises overcome a period of turbulent economic crisis. Both these aims were crucial to reduce uncertainty and protect the economic system from collapsing. In so doing, as one interviewee stressed, "ESSA prevented a social revolution by maintaining acceptable levels of social stability and cohesion." Protecting 'mid-siders' meant that regional governments moved into what Lauria et al. (2007) have dubbed "no man's land". Evidence suggests that 'no man's land' was dramatically hit by the economic crisis. A member of the trade union association in Lombardy described the situation as follows: "the crisis was a Tsunami … the number of firms filing requests for ESSA was growing by the day." 6 ESSA beneficiaries rose from 282 480 in 2009 to 398 944 in 2010 and dropped slightly to 352 013 in 2011 (ISFOL, 2014, tab 2.1). In three years the number of workers covered by the measure increased by 25%. 7 ESSA regulations allowed Italian regions to sidestep the strict eligibility limits and entitlement rules set by the ordinary unemployment benefit system. In a nutshell, ESSA endowed regions with discretionary power, which enabled regional policymakers and their social partners to select beneficiary firms of different sizes and work contracts of diverse types as well as to vary benefit duration (Barbieri, 2009: 84). In this way it was possible to mitigate social unrest that might have occurred had thousands of firms suddenly gone bust and thousands of workers found themselves unemployed. 8

Patching Up and Conversion
A first step in evaluating changes in regional labour market policies is to get a clear picture of what has changed. There is little doubt that the centrepiece of policy change between the regional governments and their social partners was the 2009 Accord signed by the central and regional governments (Conferenza delle regioni e province autonome, 12 February 2009). The Accord rested on three criteria: co-funding between state and regions; mandatory link between active and passive labour market policies; concertation among regional policymakers and their social partners, trade unions and associations of employers.
Following the Accord, regions became the main players because the administrative responsibility for ESSA shifted from central to regional governments. Regional policymakers became the key actors, independently setting the criteria for selecting the target population during the territorial negotiations, which were followed by the local Accord (accordo quadro). What is more, administrative devolution went hand in hand with changes in the sources of funding. Until 2009 the central government was solely responsible for the unemployment benefit system, but after the Accord the sources of funding for ESSA were split between national and regional governments (70% and 30% respectively). The central government funded 5250 million euros, whereas the regional contribution was 2650 million euros. In 2011 the regional contribution was raised to 40%. The Accord established that most of the regional financial resources for ESSA would come from the Structural Funds (Conferenza Stato-Regioni, Accordo 2/2009: 1).
Seen in this light, one could think that the split from national to regional financial responsibility is explained more by the rules of the EU, which require that regional governments manage ESF, than by any deliberate rescaling strategy at the national level. It is well known that the requirements of the European Commission are that benefits financed through the ESF must be tied to workers' participation in programmes and must systematically be monitored and evaluated. Yet as one interviewee declared "Regions decided to allocate their share of ESF to fund emergency social shock absorbers." What this means is that subnational actors devolved ESF to ESSA rather than to other policies and projects. In addition, and consistently with ESF rules, the Accord established that regional resources would be allocated to promote and implement active labour market policies, while central funds would pay for social security contributions as well as a large part of income maintenance (Accord: 1. Art. 2). Although regional governments appeared proactive in using ESF funds, ESSA could be considered an example of successful multi-level governance among the European Union, the national and subnational governments.
The split in funding responsibilities between central and regional governments had important implications for the regional governance of ESSA. According to European rules, ESF are meant for the promotion and implementation of active labour market policies. This means that using ESF was tantamount to making the first move towards an 'activation turn', which hinged on the improvement of educational and consultation systems as well as on training and consolidation of skills.
To be sure, activation is not new in Italian labour market policy since promotion of labour market re-entry had been attempted for decades (Gualmini, 1998;Sacchi and Vesan, 2011). But conditionality for beneficiaries had never been mandatory and, not surprisingly, the Italian spending profile in active labour market policies was one of the lowest in Europe between 1990and 2005(Liso, 2009Bonoli, 2012). In Italy unemployment benefit schemes were typically conceived as passive measures, and to some extent also 'paternalistic', owing to the absence of conditionality and low investments in active labour market policies (Madama et al., 2013: 62). Therefore in Italy the European Commission's requirement of conditionality represented an important novelty, propelling policy conversion to welfare-to-work.
Patching up and policy conversion set in motion institutional, organizational as well as cultural transformations. New institutional frameworks created multi-level governance; new organizational settings boosted a system of territorial employment centres; the centrality of activation shaped cultural attitudes, shifting the logic of labour market policies from passive measures to welfare-to-work. The Accord initiated something new because benefits were now tied to workers' willingness to enrol in retraining programmes.
In brief, as a response to the global economic crisis regional governments 'patched up' the existing unemployment benefit system, creating a hybrid central-local administrative structure where conditionality is a new underlying principle added to old structures ruled by old principles, i.e. income maintenance. Besides the patching up of administrative structures, another crucial feature of the Accord was conversion, the redirection of labour policy goals from mere replacement income to welfare-to-work with the aim of moving unemployed workers from assistance into employment (Mulé, 2013).
The results of this policy effort are illustrated in Table 2. The data show that between 2009 and 2012 a total of 653 000 ESSA beneficiaries received information services from employment services, that is 98.6% of ESSA beneficiaries (ISFOL, 2014: 52).
In addition, 99% of those persons receiving information services obtained individual case management (643 998) and 12 000 beneficiaries were trained or retrained at the workplace. The beneficiaries in the subcategories reported in Table 2 overlap, since the categories are not mutually exclusive and therefore beneficiaries may fall into more than one category. The data published in the ISFOL report of 2014 show that 68% of beneficiaries were aged between 26 and 49, while 26% were aged 50 or over. Between 2009 and 2012 the number of employment services delivered was 3 million, with an average of 3.8 services per beneficiary. This is an impressive effort, especially in light of Italy's well-known 'weak' administrative capacity.
Patching up administrative structures and conversion of policy goals spawned a virtuous cycle that led to convergence of interests and 'political exchange' between regional governments and their social partners. Regional governments thought ESSA allowed them to prevent 'a social revolution', as one interviewee observed. For employers, instead, the advantage of ESSA was to train workers. The representatives of the regional employers' association put it clearly: "we were favourable to emergency social shock absorbers because they were linked to active labour market policies" … "We believe that active labour market policies are crucial … because it has been shown that they exert important psychological effects on workers" … "Active labour market policies had never been implemented in a mandatory manner and this does change the way we perceive unemployment benefits, we no longer see them as a 'parking spot' but as a way to retrain workers." These observations help explain why active labour market policies weakened employers' resistance to extending the coverage of ESSA.
Trade unions were sympathetic to ESSA as they provided coverage to the previously excluded 'mid-siders'. Moreover, as a trade union member declared, regional governments acknowledged the power of trade unions because "we knew what was going on in the small firms". Trade unions became privileged contact points of the regional government together with employers' associations. Such remarkable convergence of interests between the regional associations of workers and employers springs from the elasticity of the 'regional political space' introduced by the Accord. Regions and their social partners enjoyed ample room for manoeuvre in selecting the criteria for ESSA's target population.
It should be noted that in Italy the regional politics space is buttressed by two important institutional reforms implemented in the past 20 years (Vassallo, 2013). The process of region-building was strengthened by administrative devolution in the health system since 1991 (Toth and Carboni, 2012). This reform was followed by amendments to article 117 of the Italian constitution in 2001, which established the possibility for regional governments to interact directly with the European Union. Such administrative and constitutional reforms accelerated regional devolution in the field of social policy and stimulated endogenous activity (Mulé, 2003;Baldi, 2011). 9 The second institutional reform concerns the direct election of the region's president in 1995, its highest political authority. Direct legitimacy means regional governments have strong incentives for credit claiming policies and for bolstering regional social capital. Credit claiming in the field of labour market policy is all the more important for regional governments because, contrary to expectations, recent research findings show that labour market risk is shared by both high-skilled and low-skilled workers (Häuserman et al., 2014).
The regional governance of unemployment benefits through ESSA was always intended as a temporary measure (see Accord: 1). ESSA provided coverage to thousands of workers and helped thousands of micro and small firms to survive the global financial crisis. Arguably, however, there are at least three features of ESSA that may have had profound and lasting effects on the Italian welfare state. Firstly, ESSA extended coverage, and this extension was included and further extended in two recent labour market policy reforms, the Fornero Law of 2012, under the Monti government, and the Jobs Act of 2014, under the Renzi government. The Jobs Act not only further extended coverage, it made active labour market policies the centrepiece of reform. As noted by the OECD Report of 2015, a key strategic element of this reform is the creation of the Unique Agency for Active Policies (ANPAL) and the attribution to the new agency of a central role in the coordination of regional actions regarding active labour market policies. Hence it looks as if current Italian policies have incorporated the new path-altering paradigm set in motion by regional devolution.
Second, ESSA elicited multi-level governance in the fields of unemployment benefits, among the state, the regions and their social partners. In turn, this multi-level framework unleashed a dynamic inter-institutional coordination mechanism that today nourishes a programmatic ability to implement active labour market policies. ESSA buttressed the provision of active labour market policies services by aligning regional and local policies and by involving major public and private actors. This extraordinary institutional effort improved the quality and quantity of services necessary to respond to the spiralling number of beneficiaries. Training programmes were set up, revamped or modified to meet the different needs of micro and small firms as well as of their workers.
ESSA were a temporary measure and not a structural one. Yet it is likely that its diffusion set forth some of the conditions for the subsequent reforms of labour market policies, both the Fornero Law and the Jobs Act. These two reforms have reduced and will progressively abolish emergency measures. But the regional governance of ESSA probably will have deep and lasting effects because it shaped cultural attitudes and created the institutional and organizational settings for recent reforms of labour market policies.

Regional Network Coordination
The consolidation of the regional politics space offers a unique opportunity to examine how regions resolve the coordination problems that emerge with new configurations of actors. The labour market policy orientation induced efforts at institutional coordination because activities are now channelled through agreements at the regional level, where a tripartite committee sets out the basic rules Di Stefano, 2012, 2014). Hence regional coordination is one of the main ingredients of the ESSA scheme. The new institutional set-up of the active labour market policies and steering system strengthens the role of the regional social partners. It is absolutely crucial to the implementation of this new structure that coordination between regional enterprises, regional authorities, local authorities (provinces and municipalities), unemployment offices, training institutions and social partners be efficient.
Like other European countries, such as Denmark (Dahl et al., 2002), Italian regions have experienced changes in the organizational bases of the labour market system. Regional labour market policies rest on network coordination between the regional governments and their social partners as well as on horizontal subsidiarity. Against this background, it is not surprising that regional policymakers have embarked on a number of policy experiments. The following sections examine one such experiment, i.e. how regions coordinate passive and active labour market policies, comparing the strategies pursued by the regional governments of Lombardy and Emilia-Romagna.
Both Lombardy and Emilia-Romagna have enjoyed comparatively high levels of economic performance and remarkable political stability in the past 20 years. Right-wing majorities ran the Lombardy government from 1995 to 2011, while left-wing political parties have governed Emilia-Romagna uninterruptedly since 1970, when regions were established in Italy. In the period under investigation political leaders of the two regional governments had radically different political preferences. Scrutinizing the impact of political preferences on public policy-making is beyond the scope of this work, chiefly because it entails a different research question.
Economic performance and political stability mean that Lombardy and Emilia-Romagna are endowed with similar background conditions. Thus, Lombardy and Emilia-Romagna represent two excellent case studies to examine and compare the regional strategies enacted to manage coordination problems unleashed by the Accord.
Previous research on the labour markets of Lombardy and Emilia-Romagna highlighted two distinct features, namely the development of a flexible labour market in Lombardy and a strategy of requalification in Emilia-Romagna (Gualmini and Alti, 2001). This dichotomy is similar to what Barbier (2005) defines as the two opposite poles of liberal activation, marked on one side by strong work incentives and on the other by human capital investment and extensive training. Such distinctions are useful starting points in analysing regional activation models.
The literature adopts a narrow and a broader definition of the term 'activation'. Narrowly, it involves developing links between unemployment benefits and active labour market policies making benefits conditional upon participation in job-search programmes. More broadly, activation refers to policies and instruments that increase labour market entry. With the term 'activation' we mean social and labour market services necessary to promote the active inclusion of people of working age. Such services in Italy have traditionally been particularly scarce (Sacchi and Vesan, 2011).

The Lombardy Model: Liberal Market Economy and Competitive Subsidiarity
Between 1995 and 2012 the right-wing government in Lombardy implemented a complex system of reforms in the field of social policy, initially focused exclusively on health policy (Regional law n. 31, 1997) but later extended to include ESSA. Following the Regional action plan, the Regional law n. 22/2006 introduced an important innovation by triggering 'quasimarket' mechanisms. In 'quasi-market systems' the government finances and monitors services, but does not necessarily provide them. Provision is offered by independent actors competing for customers in 'quasi-market' settings (Le Grand, 2013). 10 The main feature of the Lombardy model is a peculiar type of horizontal subsidiarity driven by competition between public and private providers of social services (Colombo, 2012;Colombo and O'Sullivan, 2012). For example, local centres for employment (centri per l'impiego) compete with private providers for employment. Horizontal subsidiarity has been at the core of the regional government's electoral appeal and is at the heart of a series of major regional reforms that have restructured welfare state services by introducing quasi-market mechanisms. Like other liberal market economies, the Lombardy government provides less support for non-market forms of relations. 11 The Lombardy government reformed labour market policy by functional differentiation. It retained regulatory, programming and financing functions while delegating the management and delivery of services to independent bodies. Moreover, the users of public services are able to choose the provider. In such a system the regional government "plays the role of rule-maker without being responsible for the provision of services" (Colombo, 2008: 188). Users' freedom to choose between private or public service providers fosters an "administered competition" (Pavolini, 2004).
One distinct feature of the Lombardy model is the 'dowry' system (Dote). In the dowry system users are given a voucher so that they can choose among service providers; public and private providers compete for customers. The dowry system was well under way in Lombardy before the introduction of ESSA, therefore there is no innovation in using it in this region.
The novelty lies in the link between ESSA and the labour market dowry. As explained by the Lombardy region the ESSA dowry is a new type of dowry … that allows ESSA beneficiaries to enjoy personalized services aimed at helping them to return to work and to improve their skills. In this manner there is integration between active labour market policies and income maintenance. Income maintenance and ESSA dowry are inseparable. Whoever turns down the services offered by the dowry will lose the economic benefits. 12 As one interviewee declared, "the centrepiece of this system is individual responsibility. Workers must stop conceiving of unemployment benefits as mere income support but instead should take the opportunity to improve their skills and qualifications." The competitive system of horizontal subsidiarity built in the Lombardy model aims to empower individuals by stressing their freedom of choice.
It goes without saying that the Lombardy model suffers from the same criticisms levelled at quasi-market operations. It can work properly only if a number of provisions are put in place. To begin with, pluralism must be ensured to prevent the advent of dominant or monopolistic providers. In addition, the market should offer an effective network of providers. Moreover, service providers should adopt transparent procedures that can be understood by all individuals who are in need and therefore vulnerable. However, these traits are not easy to achieve. So much so that the ISFOL report of 2012 finds that, at the time of publication, the Lombard government had not released performance indicators of service providers. This delay weakens incentives for competition because users have no information concerning the efficiency and efficacy, and thus the quality, of the services.

The Emilia-Romagna Model: Coordinated Market Economy and Integrative Subsidiarity
In stark contrast to the Lombardy liberal market political economy, labour market policy in Emilia-Romagna rests on coordinated market economy. Policymakers deliberately retain public control and governance of the services provided for the unemployed workers. Like other coordinated market economies, Emilia-Romagna relies more heavily on non-market modes of coordination. These entail more extensive relational interactions, network monitoring and more reliance on collaborative rather than competitive relationships (Hall and Soskice, 2001). Regional policy actors have explicitly declared their resistance in devolving public functions to private providers and are sceptical regarding the virtues of competitive drives between private and public providers. Regional policymakers conceive of the 'public' mission as being significantly different from the 'market' mission. They criticize the competitive model centred on Adam Smith's 'invisible hand' allegedly able to choose among service providers. Instead, they believe that public authorities ought to both protect the unemployed workers and monitor the quality of service providers according to specific standards.
In 2011 the regional government and its social partners signed the Pact for sustainable, intelligent and inclusive economic growth (ISFOL, 2012: 245), which set out the overall strategy as well as the operational procedures. The Regional Tripartite Committee and the Committee for Interinstitutional Coordination formulated the policies and monitored their implementation. Not surprisingly, ESSA's in-built discretionality proved an advantage to regional policymakers. The head of the Regional office for labour services (Servizio Lavoro) thinks that discretionality "is a very powerful instrument that enhances the efficacy of emergency social shock absorbers because it offers the flexibility to meet the needs and demands arising from below". Such needs and demands are filtered through an extensive network of services for the unemployed workers, which includes private providers and is administered and monitored in both formal and informal collaborations. Any outsourcing of services for active labour market policies is strictly connected to performance indicators and standards (ISFOL, 2012).
Regional policymakers believe that direct access to public funds is possible only if strong collaborative relationships between public and private service providers are in place. In this system public authorities control the steering wheel, monitoring and evaluating the results. The understanding of an efficient labour market policy in Emilia-Romagna is centred on integrative subsidiarity and institutional support encourages service providers to gravitate towards a coordinated political economy.
Close examination of active labour market policies in Emilia-Romagna shows that the region was well equipped from the viewpoint of policy initiative as well as policy implementation to become one of the most efficient Italian regions. Reasons for this success are straightforward. Firstly, the regional administrative procedures required very few modifications to fit the regulations of the Accord. Path dependence clearly played a role. In Emilia-Romagna the existing organizational structure facilitated the development and implementation of training and requalification programmes. A cooperative style of working relying on the integration of different competencies had been in place before the Accord, allowing for smooth adaptation. Most importantly, the problemsolving capacity of the regional network structure had been highly effective and flexible. A representative of the trade unions in the Regional Tripartite Committee (Commissione Tripartita Regionale) declared that "what distinguishes Emilia Romagna in the implementation of emergency social shock absorbers, with respect to other regions, is that we did not suddenly improvise concertation between the social partners." In short, Emilia-Romagna policymakers relied on traditional systems of concertation with their social partners (Pavolini, 2004;Fargion and Gualmini, 2012).
Our findings, however, not only confirm the persistence of concertation but, most importantly, contribute to the literature by highlighting the role of non-market modes of coordination in the field of labour market policies. The distinction we draw here between concertation and non-market modes of coordination is nuanced, but important. Concertation centres on cooperative interactions, characterized by exchange of resources, policy legitimacy and competence, built upon mutual trust rather than upon recourse to law. Non-market modes of coordination mean that the regional government does not rely exclusively on private providers but instead retains a strong monitoring role on public and private service providers. Private and public providers interact in a complementary fashion, integrating and not substituting public functions and services. Integration is achieved with targeted services for specific groups of unemployed workers or by adding new services to those run by public providers.
The picture that emerges shows that regional labour market policy is based on functional integration and cohesion among local actors. Integration and cohesion strengthen the efficiency, efficacy and quality of the services provided to customers. It warrants repeating: the coordination of active and passive labour market policies in Emilia-Romagna rests on integration between private and public service providers, a model which has probably become more important as the economic crisis deepens.
Before concluding, it is worth noting that the different ways Lombardy and Emilia-Romagna have chosen to implement ESSA may reflect their governments' political preferences and the political interests that underpin those preferences. As Thelen (2014) has persuasively argued, when inspecting contemporary changes in the political economy of institutions we should pay greater attention to the coalitional foundations on which those institutions rest. In Emilia-Romagna, the preference for the government to maintain control on service management may be inspired by a belief in the importance of the 'role of the public in policy-making'; it may also respond to the regional specific political coalition, marked by an enduring alliance with the cooperatives and the unions. As one interviewee declared, what is distinct about Emilia Romagna in the management of ESSA, and explains our achievements, is that we did not have to suddenly engage in concertation because we could rely on a well established network of local relationships and alliances.
By contrast, in Lombardy the preference for devolution in service delivery and the creation of quasi-market mechanisms seems to reflect the centre-right government ideological preference for the virtues of the market. As mentioned at the outset, a deeper understanding of the policy implications of ideological preferences is outside the scope of this work because it would have needed a different research question.
In the introduction to this article, we noted that Lombardy and Emilia-Romagna should not be considered representative of other geographical areas in Italy. One reason is that the academic and public debate over labour market reform persistently highlights the high degree of regional variation in institutional efficiency and efficacy (Gualmini and Rizza, 2013;Jin and Lenain, 2015). Former Minister of Labour Law, Elsa Fornero, stressed that the challenges in drafting the Labour Reform of 2012-the so-called Fornero Law-were the "widely different efficiency standards" among regions (Fornero, 2013: 6). She was aware that effective activation policies and quality employment services "are indeed a kind of chimera in many Italian regions". This variety raises the need for uniform policy instruments across the country. For this reason, a key strategic element of the Jobs Act implemented under the Renzi government is the attribution to the Unique Agency for Active Policies of a central role in the coordination of regional action on active labour market policies. Fana et al. (2015) argue that the Jobs Act is fundamentally neoliberal in nature, and this raises the question of which regional model was more influential, the Lombardy model or the Emilia-Romagna model? The answer seems to be a policy blend of both models. The Agency provides two policy instruments to all regions, unemployment benefits tied to active labour market policies and the so-called re-employment voucher (assegno di ricollocazione), which is similar to the Lombard dowry but more regulated by public intervention and management through the local centres for employment (centri per l'impiego) as in Emilia-Romagna. This sets a common denominator among different regional political economic institutions as well as socio-economic conditions that characterize the Italian regional landscape. Labour economist Leonardi (2016) maintains that the aim of the national agency is to work in synergy, and not in contrast, with the various Italian regional models.
However, the Jobs Act and the debate on the liberalization process raise important questions about the future role of regional governments in the political economy of social shock absorbers. Our findings suggest that different regional models might trace a variety of liberalization lines, underpinned both by path-dependent policies and by the coalitional foundations of regional political and economic institutions.

Conclusions
This article contributes to the debate on the governance of the Great Recession of 2008 by showing that Italian regionalization patched up administrative structures and converted policy goals. This process favoured the elasticity of the 'regional political space', allowing political exchange between regional governments and their social partners in the implementation of ESSA.
The comparative case studies of Lombardy and Emilia-Romagna show that social shock absorbers in Italy built upon regional path-dependent trajectories as well as innovative strategies. Lombardy extended the 'dowry' to include ESSA while Emilia-Romagna relied on well-oiled collaboration among the social partners. However, the article argues that the choices made by regional policymakers were not only path-dependent but innovative as well. Through a process of policy conversion, regional actors redirected the policy goal from income maintenance to welfare-to-work. Policy conversion hinged on the fact that benefits became inseparable from workers' participation in retraining and requalification programmes. Regions integrated this new policy goal in their pre-existing regional network coordination structures and consolidated their regional identity. Thus, policy innovation and path dependence intertwined in the regional governance of ESSA in the aftermath of the Great Recession in Italy. Future research will explore and establish whether the regionalization of ESSA represented a critical juncture propelling the longawaited transition to a universal unemployment benefits scheme in Italy.

Notes
1. These measures are also known as shock absorbers in derogation from the ordinary and extraordinary Wage Guarantee Fund, mobility and unemployment schemes (Cassa Integrazione Guadagni in Deroga) (EPSA, 2013). The measure is also called in the literature Short-Time-Working (STW) Scheme.
2. For arguments considering regions as spaces for politics, see the Special Issue of Regional and Federal Studies, Vol. 20, no. 3, 2010. 3. In total, 46 interviews were conducted in 2011. 4. In a footnote the authors remark that in a comprehensive discussion of the Italian labour market, the self-employed should also be considered among the 'mid-siders' (footnote 2, p. 581). 5. In the narrower denotation of the concept 'mid-siders' adopted in this work temporary and atypical workers were excluded from ESSA. 6. It is employers and not workers who file a request for the concession of emergency social shock absorbers benefits, which must be approved by the public authority. 7. ESSA benefit duration is maximum 12 months (Art. 7-ter, co. 3, Law n. 33/2009). 8. Interview with head of Work Protection and Re-employment Office (Tutela del lavoro e interventi per il reimpiego), Lombardy. 9. Since 1997 regions have enjoyed some autonomy because active labour market policies have been devolved to regions. 10. The general philosophy behind these changes can be traced back to the New Public Management School, which developed in the UK in the 1980s and then spread to other industrialized countries in the 1990s. See Hood (1991). 11. The distinction between liberal market economies and coordinated market economies is set out in Hall and Soskice (2001). 12. http://www.lavoro.regione.lombardia.it, 12 October 2015. The labour market dowry was initially targeted exclusively at ESSA beneficiaries but later extended to all unemployed workers who qualified (Regional Decree n. 20609/2012).