Understanding Limitations on Welfare Policy Innovation: The Case of National Guaranteed Minimum Income All Italiana

At the beginning of 2016, Italy remained one of the very few Eurozone countries without a national-level Guaranteed Minimum Income (GMI), arguably a major lacuna in the armoury of the welfare state at a time of severe crisis and significant public spending cuts. Drawing on personal communications, parliamentary archives and secondary sources, an actor-centred approach emphasizing the role of domestic actors such as parties, trade unions and religious organizations is utilized to discuss the chronicle of the Italian GMI experience; especially the experimental national GMI in the late 1990s and the early 2000s. The experiment met with an early demise due to a meagre support base, in conjunction with the dominance of northern interests at the ministry responsible for the design and implementation of a scheme that mainly benefited the poorer south. A policy targeting ‘outsiders’ and introduced without broad consensus among key policy actors is unlikely to be durable. The recent economic crisis has revived the debate, as old and new policy actors express support for national-level GMI.


Introduction
The Guaranteed Minimum Income (GMI) is an instrument of the 'third generation' of social assistance schemes. These are distinguished mainly by the combination of monetary allowances with social insertion measures (Kazepov, 2011:106). Payments usually cover the difference between a household's actual income and the resources it needs, according to a defined standard for living in a specific region or country. Recipients are required to participate in social insertion programmes meant to facilitate job-market integration, such as vocational training programmes (Matsaganis, 2013: 16, 21-22).
EEC recommendations issued in 1992 advocated common criteria on sufficient resources and social assistance and convergence of social protection objectives and policies. Following this, the political and intellectual climate in most EEC member states turned toward the establishment of national-level GMI schemes.
At the beginning of 2016 1 , Italy remained one of the very few Eurozone countries still lacking a permanent, national GMI.
Italy launched a national GMI experimentally in 1998, only to abandon it a few years later. Other national, regional and local programmes against poverty have been developed, but do not substitute for a national GMI. A relatively delayed and unclear framework of social assistance decentralization has largely contributed to extremely uneven development and non-generalized adoption of these schemes by Italian regions and municipalities. Additionally, the sub-national schemes depend on resource availability and do not constitute individual enforceable rights (Lalioti, 2013: 126, 133-135;Madama, Jessoula and Natili, 2014: 16-18, 25, 27).
Notorious for high spending on old-age benefits 2 , the Italian welfare state is known for paying limited attention to 'outsiders'. Aggregate spending on the two major social assistance benefits on the national level, the civil disability pension (pensione di invalidità civile) and the social pension (pensione sociale), persistently amounts only to about 3-4% of total social spending, according to the Bank of Italy.
At the territorial level, data from the Italian National Institute of Statistics indicate standards and criteria for access to social assistance differ greatly among regions and municipalities (Lalioti, 2013: 59-61).
This article draws on personal communications, parliamentary archives and secondary sources to record the story of the Italian GMI experience, especially the story of the experimental national GMI in the late 1990s and the early 2000s. In the absence of 'functional equivalents', above all of a coherent and inclusive social safety net of last resort, and given relatively high poverty 3 , the history of the GMI experiment is striking. An actor-centred approach is employed, focusing mainly on governmental actors, while also acknowledging the policy effects of actions (or inactions) by other key players, such as organized labour and religious organizations.
From the viewpoint of political economy, GMI beneficiaries are outsiders with limited political representation, forced to rely on other actors to promote their interests (Huber and Stephens, 2001: 18).
Policy actors' attitudes towards GMI are here assumed to reflect the internal dynamics of given time periods and specific national contexts. Exogenous influences on domestic welfare agendas, such as from 'soft-law' instruments like European recommendations, are relatively mild. Innovations and changes in welfare policy do not follow deterministically, not even from extensive economic developments such as those linked to severe crises (Maino and Neri, 2012: 117). Influences, innovations and changes all depend on discretion, i.e. the interests and ideological preferences of domestic policy actors.
This article aspires to contribute to studies on the Italian GMI experience such as those by Benassi and Mingione (2003), Ranci Ortigosa (2007a), Sacchi (2007), and Madama, Jessoula and Natili (2014). The following sections argue that the early demise of the experimental scheme in the early 2000s followed from a relatively weak support base for a national GMI, in conjunction with the dominance of northern interests at the ministry responsible for the design and implementation of a scheme that mainly benefited the poorer south. The crisis period since 2008 has seen a more intense debate, as national GMI has met support among old and new policy actors.
The concluding section synthesizes the main lessons to be learned from the analysis.

First steps towards a GMI
The coming to power of the Prodi centre-left coalition in May 1996 raised hopes for policy changes that would integrate the weakest socio-economic strata and strengthen the country's potential for a stable European future. The context was set partly by EEC initiatives calling for policy innovation, such as the 1992 recommendations 92/441/EEC and 92/442/EEC, by a broad need to conform with EEC-imposed macroeconomic standards, and by Italy's own deep political crisis.
Changing the scenery in social assistance policy was a means of approaching these targets, also helping Prodi to raise his coalition's profile as a modernizing government.
Prodi in January 1997 appointed the expert Onofri Commission, tasked with assessing whether ambitious goals for welfare system reform were compatible with macroeconomic adjustments underway in preparation for European Monetary Union (personal communication with commission chair Paolo Onofri, 11 February 2012).
The commission proposed a consistent reform strategy that would include institutionalization of a national GMI 4 (Commissione per l'analisi delle compatibilitá macroeconomiche della spesa sociale, 1997). The idea was advertised both as a means of facilitating labour market integration for outsiders (Carpo, 1997) The most prominent opponent to GMI, however, was inside the Prodi government itself. Minister of Labour and former Christian Democrat Tiziano Treu compared offering GMI to the young to dispensing methadone (Lupoli, 1997 The GMI being a social assistance scheme which does not promote job creation does not make it the best policy we (the unions) could imagine... an experimental phase will definitely help, however, in the sense of mitigating the risks of making the GMI permanent... we will see whether it helps beneficiaries get a job... instead of ending up as another sort of social assistance programme (Esposito, 1998, 6).
Objections to GMI and defences of the existing welfare system were related to a general fear of restructuring, given that the national GMI proposal was linked to the gradual abolition of other social assistance categories. Labour confederations feared that a GMI might trigger an attack on unemployment benefits, be used as a minimum wage threshold, or replace active labour market policies. Insofar as GMI typically targets outsiders, it might mean a smaller pie for insiders (Lalioti, 2013: 103-104).
Meanwhile Innocenzo Cipolletta, general director of the employers' federation Confindustria, claimed the scheme would function as a disincentive against labour (Il Sole 24 Ore, 5 March 1997).
The religious community displayed two clear trends. First, the church hierarchy took no public stance, following the non-alignment policy embraced by the Vatican through part of the 1990s, and feeling secure after their successful defence in 1988 of the principle of subsidiarity and the resulting consolidation of Catholic religious organizations' role in the social assistance field. This lack of involvement contrasts with areas where the church felt more threatened, for instance in family policy (Lalioti, 2013: 91, 93, 104;Chianura, 1999). Second, however, organizations like Caritas and the Christian Associations of Italian Workers (Associazioni Cristiane Lavoratori Italiani, ACLI) believed GMI aligned with their organizational principles, seeing it as a significant step towards recognition of social citizenship for people at society's margins. Caritas' support did not translate into formal advocacy in public documents, however. Francesco Marsico, vice-director of Caritas Italiana (personal communication, 20 March 2012) attributes this to its identity as the pastoral organization of the Catholic Church.
The initial national GMI debate was thus shaped by division in the government, scepticism or hostility from parties, interest groups and labour, and only 'unofficial' support from Catholic grass-roots groups. Given this lack of a strong constituency, let alone a coalition in support, Prodi went ahead with a limited experiment as a way to demonstrate poverty was on the agenda, postponing actual financial and policy decisions until after the experiment. RMI recipients were defined as persons whose income did not exceed the 1998 poverty line of 500,000 lire (258.23 euros) monthly for an individual living alone (Legislative Decree 237/1998, Article 6). The monetary allowance equalled the difference between the poverty line and actual income. The amount was set nationally, without consideration of regional differences in cost of living.

A 'stillborn' GMI experiment
Beneficiaries were also expected to participate in customized plans for social integration, such as training courses and care services to be devised locally (Sacchi, 2006: 874-875). RMI provided territorial actors (local governments) with an opportunity to confront extreme poverty, assess the quality of services provided by local managers in the fight against poverty, and increase their popularity at low financial cost.
While criticisms continued, inter alia proving the use of the term 'experiment' ambiguous (Lalioti, 2013, 107-108), the main official reason for the pilot RMI was to identify operational difficulties territorial actors faced in providing benefits and managing insertion programmes, thereby testing in different contexts the scheme's effectiveness in overcoming economic need and marginalization (Legislative Decree 237/1998, Article 2). The plan was to evaluate the experiment, then decide whether to convert it into a permanent national programme at the end of 2000. By then the programme had 86,000 recipients. About 220 million euros were spent during the first two years (IRS, Fondazione Zancan, and CLES, 2001;Sacchi and Bastagli, 2005: 117-118), less than one-thousandth of the Italian state's overall social expenditure of 282,617 million euros according to the Bank of Italy (Lalioti, 2013: 106). Centre-left parties opted to prioritize alternative policies, particularly the integration of minimum pensions, rather than using the same resources for a reform that would introduce GMI… Labour confederations exerted similar pressures. They were reluctant to accept a dilution of the Italian welfare system's contributory character, which secured for them a direct interventionist role in the system, for the sake of a more universalistic system.
Italian media busily publicized the gap between the needs and capacities of territorial actors in the north as against the south (Magni and Sala, 2000;Sala, 2000;Arachi, 2002). This should be viewed within the broader fracture between north and south, a long-standing historical phenomenon that widened during the first phase of industrial development in the early twentieth century (Putnam, Leonardi and Nanetti, 1993;Fargion, 2005: 138).
While the government could not easily ignore these factors and make the scheme permanent, it should also be recalled that the decision to extend the experiment came only months before the May 2001 general elections. Under pressure from southern territorial actors, some cadre of the centre-left coalition, mainly from the south and connected with Catholic third-sector organizations, supported the extension of RMI in their electoral districts (Lalioti, 2013: 110-111).
This dissension reflected yet again how fragmentary the scheme's support was. With the centre-left always divided and other key actors sceptical or hostile, it would prove impossible for the politically fragile initiative's few ardent supporters to ensure its survival against future changes in government.
It has furthermore been argued (Sacchi and Bastagli, 2005: 85;Sacchi, 2006: 876) that termination of the GMI experiment was enabled by a 2001 constitutional reform depriving the central government of the power to issue planning instruments in the social assistance field (as Law 328/2000 had provided). After losing regional elections in 2000, Olive Tree adopted more devolution policy to compete with the Northern League (Lega Nord, LN) and undermine the centre-right (Bull and Pasquino, 2007)  Cadre from Berlusconi's Forza Italia (FI) and the Lega Nord soon showed their disapproval of RMI. Although Naples was among the municipalities that the programme had benefited most, it was denounced as pure welfare dependency by the FI provincial coordinator, Antonio Cuomo, and by the head of FI in the Naples regional council, Francesco Bianco (Fuccillo, 2004).
Lega Nord, increasingly eurosceptic, was even more adamant in its opposition.
During the election campaign its cadre, sounding like libertarians, claimed the free market offered effective protection to citizens and argued that public funds should be redirected from welfare benefits ('state handouts' that hindered economic development) to initiatives against unemployment (Greene, 2003:199). RMI reminded them of 1993, when protests broke out in northern Italy against a new minimum tax opponents believed would drain northern resources to support a dependent south (Gold, 2003: 94).
Lega Nord had long accused southern Italy of developing 'a welfaredependent culture addicted to the guarantees of the Italian bureaucracy' (Bossi and Vimercati, 1993: 42), diverting federal resources that might otherwise have been invested in the private sector to foster growth. Welfare benefits for unworthy recipients in southern Italy would cause further deterioration of a country already operating at two speeds, 'that of the producer... and that of the parasites' (Bossi and Vimercati, 1993: 20).
Thus it came as no surprise when the government downplayed evaluators' positive remarks about the first RMI phase. The evaluators found that the scheme mobilized territorial actors and encouraged inter-institutional cooperation and mutual learning within the public sector, as well as between the public and non-profit sectors (Saraceno, 2006: 105) During this time the centre-right sought to distance itself from consultation (concertazione) and move to a sui generis social dialogue that required no agreement between government and other social actors (Pulignano, 2003: 6).
Under the Pact for Italy (Point 2.7), the RMI experiment proved 'the unfeasibility of legally identifying persons with the right to enter this social safety net', so that it was necessary to create 'a new instrument' called the 'Income of Last Maroni's attitude conforms with Lega Nord's ideological stance, in which solidarity is restricted towards northerners and directed against those considered to be strangers, i.e. southerners. The Lega's priority was to represent northern interests.
According to Ranci Ortigosa (personal communication, 25 January 2012), poverty is mainly present in the south, and the resources allocated for RMI were paid largely to southern municipalities. With political regionalism at its height, Lega Nord exploited growing discontent with the huge share of public expenditure fruitlessly earmarked for the south, calling RMI another example of the Italian state's 'southernization', an imposition of deviant 'southern' practices as opposed to 'northern' ones (Bull, 1994: 75  When I managed, threatening to go public, to de-freeze the Commission some months later, in order to complete the annual poverty report, even the computer files were no longer there.

The political dominance of northern interests at the Ministry of Labour and
Social Policy is also manifest in Fabrizio Spinnato's words. Asked for the Lega's position on RMI, Spinnato, in charge of social affairs at Lega Nord's Padania office, pointed to public, government documents such as the Pact for Italy (Point 2.7) and the White Book on Welfare (pp. 37-38, on which more below) (personal communication, despite continuing leaks to the media concerning the south's problematic performance, the report had not publicized or discussed in parliament as the law required (Lalioti, 2013: 115).
Thus the end was near, given a meagre support base for RMI, the antisouthern Lega's control of the relevant ministry, and the poorer south's gains from the scheme. Putatively high costs and ineffectiveness became the centre-right's official reasons for abolishing RMI. Yet neither was justified by the facts. Estimates for the annual cost of implementing national-level RMI ranged from two billion (Guerra and Toso, 2004: 2) to three billion euros (Sestito and Nigro, 2004). Even the latter For many southern municipalities, the RMI experiment was the first experience of a modern income support founded on the concept of citizenship as a right rather than a privilege (as had been the norm). The admitted difficulties in managing demand for associated services in some cases triggered clientelism and fraud, but these were isolated incidents (Benassi and Mingione, 2003: 43). Local variability in policy outcomes is, moreover, manifest even in systems believed superior to Italy's, for instance in France (Mingione, Oberti and Pereirinha, 2002: 61-77), without however being used as an excuse to abolish GMI.

Alea iacta est; allegations and truths
Finally, the experimental phase had been presented as a necessity for testing feasibility, locating weaknesses, and identifying corrective measures. The decision to cancel RMI precisely because of the expected unequal institutional performances among different territorial actors contradicts this. As Ranci Ortigosa (personal communication, 25 January 2012) asserts: The overall outcome was not discouraging. Especially the evaluation helped to articulate a clear set of guidelines for addressing the difficulties that arose within an appropriate time span. Thus, I do not think that there were any substantive reasons for the centre-left not to support the scheme or for the centre-right to abolish it.
As supposed evidence of failure, the government and some journalists even pointed to the relatively high percentage of families exiting the RMI. They tended not to mention that this was because they had overcome the initial state of need (Sacchi, 2006: 876). The second RMI evaluation report, this time conducted by four research centres, showed that the highest exit percentages were in municipalities exhibiting the lowest rates of unemployment, effectively arguing for the success of the RMI (Ministero della Solidarietà Sociale, 2007: 102 -104).
For the vast majority of municipalities, RMI opened the way for rationalizing municipal services and integrating them into wider regional administrative schemes, also bringing about improvements in institutional and administrative capacities. Faced with zero or suspiciously low declared income, for example, several municipal governments assessed the living standards of RMI beneficiaries by visiting claimants' homes to scrutinize their lifestyle, or requiring recipients to participate in a social integration programme at strategic times during the working day so as to prevent them from having another job in the informal economy (Sacchi and Bastagli, 2005: 121-125).
In short, the evidence does not support that high costs and/or ineffectiveness were the real reasons for the RMI's abolition. A more credible explanation lies in domestic politics and the fragmented support or hostility of powerful constituencies and interest groups. This was reflected in the almost complete absence of protests against abolition.
For example, reactions from the labour confederations were mixed. A few Catholic grassroots organizations that had been involved in RMI activation measures, including Caritas, did express opposition when the White Book was presented to social policy actors by Maroni and the secretary of social affairs, the Forza deputy Grazia Sestini, on 20 February 2003. That opposition never took the form of more overt protest, however (Lalioti, 2013: 120).
The southern-dominated leadership of the National Association of Italian Communities (Associazione Nazionale Comuni Italiani, ANCI) was among the few representatives of territorial actors to organize a protest against abolition. The ANCI National Council was headed by the mayor of Foggia (Apulia) in southern Italy, Paolo Agostinacchio. He laid the blame for abolition on two members of Forza, Antonio Azzollini, chairman of the Permanent Parliamentary Committee on the Budget, and Alberto Zorzoli, the committee's secretary (Petrini, 2002).
As Ferrera (2003:635) put it, GMI schemes represent 'a sphere of asymmetrical solidarity, in other words public support based purely on considerations of need, which presupposes strong ties of we-ness'. These ties were fatally lacking. National Action Plan for Social Inclusion, the central government was required to cofund RUI in regions that decided to adopt it. RUI would cover up to 2.7% of Italian households with an average of 2,925 euros per household, while the funding needed was approximately 1.67 billion euros (Sestito, 2004: 3-4).

Post-RMI debate about a national-level GMI
The Ministry of Labour and Social Policy, however, avoided having the 2003 law state explicitly that RUI would be part of the Liveas that the government was supposed to ensure, or that it was exclusively regional (Strati, 2009: 5). Again, in the absence of these clarifications RUI paved the way for the central government to fund those territorial actors most able to implement the scheme, in other words, the wealthier, northern regions.
A final attack on RUI came in late 2004 from the judiciary. On a filing from several territorial actors who claimed the government's initiative infringed upon their autonomy in social assistance policy, the Constitutional Court ruled that Rome had no authority to co-finance RUI schemes. The court claimed that the 2001 constitutional reform gave the government authority only to define Liveas (Sacchi, 2006: 876 -877). Evidence from both the 1998 RMI experiment and more recent local experiences shows it would be relatively easy to design GMI so as to minimize fraud (Lalioti, 2013: 123). Moreover, scholars such as Monti and Pellizzari (2008) have indicated that a reasonably generous version of the scheme is financially viable. They agenda paid more attention to benefiting precarious workers, which Ferrero considered to be the solution to poverty (Lalioti, 2013: 124). Priority was also given to favouring pensioners; the government forged an agreement with labour unions to avoid a three-year increase in the retirement age, with a serious impact on the whole pension system (Madama, Jessoula and Natili, 2014: 12).
The laws for fiscal year 2008 (222/2007 and 244/2007) allocated no funding for RMI and the government did not re-introduce the scheme. In his own defence, however, Ferrero points to the early fall of the second Prodi government as the sole reason for inaction on the RMI front (personal communication with Vittore Luccio, 31 January 2012).
In the immediate pre-crisis years the major labour confederations, especially CGIL, had adopted a theoretically more positive attitude towards a national GMI. According to Eurostat data, the proportion of working poor in Italy increased from 8.8% in 2005 to 9.8% in 2007, a non-negligible rise of more than 10% in two years. This context arguably makes GMI a more attractive device for improving the income and living conditions of individuals who constituted the core of the union movement (Lalioti, 2013: 124-125).
The post-2008 crisis hit the Italian welfare state hard, causing severe cuts in health, social services and family policies and accelerated privatization of the pension system (Maino and Neri, 2012). Against this backdrop, despite Italian labour's claims of support for GMI, 'insiders', such as members of trade unions, may be incentivized to revert to traditional priorities and defend the status quo. The slice of the pie for insiders is shrinking. Policy measures that benefit the so-called working poor and extreme outsiders (those who usually form the GMI clientele) put a further squeeze on resources for the vast majority of those represented by the trade unions. Despite appearances to the contrary, unions (many of whose members are pensioners) are not ready in practice to support a re-balancing of social expenditures so as to make room for GMI (see also Ciccarelli, 2013 on the 'lavorismo' of the Italian labour movement).
All the same, the national GMI debate has revived. Social policy issues are increasingly salient for political party competition. GMI seems to have found support for the reform of the pension system that included a GMI for those over the age of 55 who were living in extreme poverty.
Is Italy closer to the establishment of a GMI? The future months may show whether the country will remain one of the very few European countries lacking a national GMI. Interestingly, some argue that the chances for a national guaranteed safety net in Italy depend on the success of other anti-poverty policies (Madama, Jessoula, and Natili, 2014: 15-16). Nonetheless, whether Italy will see a national GMI apparently also rests on the interests and preferences of major policy actors who traditionally respond to the pleas of the most powerful constituencies and interest groups, often at the expense of outsiders. A redistributive reform to include outsiders is not considered necessary, even given the crisis (Maino and Neri, 2012).
Endogenous dynamics appear, once again, to be more influential than external constraints.
Onofri's comment (personal communication, 11 February 2012), when asked why his commission promoted a 'selective universalism' instead of a more 'universalistic' model of social assistance, highlights another dimension of these dynamics: beyond the importance of budgetary constraints, Italian public opinion generally does not give much value to citizenship rights. Provisions based solely on financial need are not widely accepted.
According to a scientific survey of social media by the 'Voices from the Blogs' group at the State University of Milan, only 6.8% of individuals' opinions sampled online urged the Letta government to implement income support for families in any form, including establishment of a GMI (La Repubblica, 14 May 2013). The leading request instead was for government intervention in the labour market, with 23.2% of the sample in favour. A survey conducted by an Italian market research company 5 in January 2014 likewise showed just 6% of the sample favouring introduction of a GMI as a national priority, as opposed to 46% who prioritize labour reform (see also Vis, van Kersbergen and Hylands, 2012: 8, 10).

Conclusions
The twisted history of the RMI highlights a lack of political will within the centre-left, rejection from the centre-right, and the absence of a strong, coherent domestic coalition to promote establishment of a national GMI. The centre-left originally introduced the RMI but was divided about it. The centre-right was hostile, largely regarding GMI as benefiting the south at the expense of the north.
Trade union confederations were unprepared to accept restructuring of the entire welfare system for the sake of a permanent national GMI. The ecclesiastical hierarchy, secure in their institutional role in the social assistance field and at that time loyal to a policy of non-alignment, adopted no public stance. Catholic grassroots organizations were among the few to support the scheme because it fit the citizenship model they envisioned. This did not translate into formal advocacy in the public record, however, due to these organizations' 'pastoral' role.
Given a meagre support base, RMI was extremely vulnerable to a change of government. With Lega Nord as Berlusconi's indispensable ally at the time, RMI became a symbol for the danger the Italian state would be 'southernized', leading to the experimental scheme's early demise.
In more recent years, the impacts of the crisis, including increasing salience of social issues in elections, appear to have revived political parties' interest in national GMI, in theory. To 'old' proponents of the scheme such as Caritas, one may now add new formations such as the PD (at least parts of it), SEL, Grillo's M5S, or associations such as BIN-Italia. Against this backdrop, a recent proposal of reform of the pension system promises, inter alia, a GMI for people over the age of 55 who are living in extreme poverty. In a context characterized by the Italian public opinion's preference for other policies, the future months may thus show whether a national GMI mechanism will be established in Italy; at a time when such a scheme is more needed than ever.