Abstract
Do policy feedbacks interact with transnational policy ideas? How do they impact domestic policy development? Using a qualitative research method based on a comparison of two poverty alleviation strategies in Latin America, this article asserts that in all likelihood, if a previously implemented policy initiative was ‘locked-in’ through domestic policy experimentation and resultant state-building and interest group policy effects, it will not easily be replaced by alternative, transnational policy ideas that appear on the radar screen of national policymakers at a later date. However, if transnational policy ideas and models can draw on or build upon already established ideational and symbolic beliefs, they can actually be used to further the motivations of key political actors and societal interests – here, diffusion opens up great reform opportunities and opens the door to further policy institutionalization.
It has long been established, since the publication of Hall’s eminent article on policy paradigms and social learning (1993), that the state does not autonomously evoke policy change, and that ideas play a critical role in politics. Increasingly, in the era of globalization, many of the circulating policy ideas and resultant innovations are no longer a uniquely domestic affair. This has led to a renewed interest in the literature focusing on policy transfer (Dolowitz and Marsh, 1996) and policy diffusion (Simmons et al., 2008; Weyland, 2007) and the role of international institutions in influencing policy change (Finnemore, 1996; Hall, 2007; Keck and Sikkink, 1998; Mahon and McBride, 2008). A broad consensus has generally been reached that transnational policy ideas affect domestic policy development, yet the debate over exactly how, when, and why remains. Using the area of conditional cash transfer (CCT) programme development in Latin America and its subsequent development and performance in two countries, this article suggests that the literature on policy diffusion could be strengthened if it interacted more closely with the policy feedback scholarship.
The basic concept of policy feedback refers to the impact of existing policies on politics and policy development (Pierson, 1993). Traditionally, the focus was on state-building effects, interest groups and lock-in effects. Newer streams of feedback scholarship have focused on the interaction between policy feedback and political behaviour, the relationship between public and private policies, and the role of ideational and symbolic policy legacies – all of which will continue to improve our understanding of politics and policy development (Béland, 2010: 569). However, it is now commonly accepted that ‘policy learning’ (Heclo, 1974), the idea upon which much of the traditional policy feedback literature is based, occurs not just through the ways in which actors draw lessons from previous policies, but also from transnational actors and the policy experiments of neighbouring countries – both of which inform the domestic decision-making process. The literature on policy diffusion has, importantly, contributed to our understanding of how, and by whom, transnational ideas are spread, but it does not provide us much insight into why ideas can get locked in and why they can become ‘unlocked’ in the national setting (Skogstad and Schmidt, 2011). Do the ideas and symbols of existing or past domestic policies influence the interpretation, framing, administrative processes and eventual outcomes of an internationally backed policy model such as the integrated CCT model? 1 Which types of domestic policy feedbacks impede future policy development, and how do they effect the institutionalization of transnational policy ideas? When do travelling policy ideas open up reform opportunities? These are the central questions this article will address.
By comparing the evolution and outcomes of CCTs designed to alleviate poverty in two neighbouring countries, Brazil and Argentina, this article shows how a transnational policy model impacted domestic policy development in both countries, yet, certain policy feedback effects facilitated the implementation and performance of Brazil’s Bolsa Familia (BF) in contrast with Argentina’s Plan/Programa Familias (PF), a social programme that has now been discontinued with a social insurance initiative. Although political variables are present in each story, I argue that more importantly, the ideas behind what can be described today as a consolidated global policy paradigm (or interpretive framework) in poverty reduction were not anomalous to Brazil prior to the implementation of its well-documented CCT programme, BF. By comparison with Argentina, powerful state-building effects and interest groups concentrated within the powerful sector of ‘labour’ impacted the development and ‘locking-in’ of the by then international CCT model designed to alleviate poverty through human capital investments. By tracing the evolution of poverty alleviation strategies over time in a cross-country comparative approach, this article suggests that if previous domestic policy experimentation resulted in certain powerful feedback effects, it will not easily be replaced by an alternative, transnational policy idea at a later date. If, however, transnational policy ideas and models build upon certain policy feedbacks, they can actually be used to further motivate key political elites and societal interests. Here, policy diffusion opens up great reform opportunities and opens the door to further policy institutionalization. The key lesson here is that different types of policy feedbacks will impact how transnational ideas are interpreted, and will reorientate national policy development in a variety of ways.
Research design
In order to show how domestic policy feedbacks impacted CCT policy development and outcomes, I first process-trace the development of CCTs in Brazil and Argentina. Second, I compare the various types of policy feedbacks produced by two previously established poverty alleviation initiatives in Brazil and Argentina, showing how they impacted the performance of the two integrated CCT programmes that were later implemented. In 2003, both countries launched, with international financing, CCT programmes that integrated health, education and nutrition conditions – the now dominant CCT model in the region. This integrated approach was broadly based on Mexico’s Progresa-Opportunidades (Aguiar and Araújo, 2002: 70; Sugiyama, 2012: 18).
The cases of Brazil and Argentina were chosen for focused comparison because of the observed variation on the outcome variable: Brazil’s BF successfully dismantled, reformed and transformed Bolsa Escola (BE), and continues today under President Dilma Rouseff (2011–2015), whereas Argentina’s PF (2002–2009) was not able to successfully dismantle, reform, or transform the Programme of Unemployed Heads of Households (PJJHD) as intended and was discontinued in 2009. These two programmes are excellent cases to explore how diffused policy ideas and domestic policy feedbacks interact and affect future policy development. Although important research on how CCTs diffused horizontally within Brazil, as well as vertically throughout the region, has been published (Sugiyama, 2008, 2011, 2012), the literature on policy diffusion does not provide many answers about why these ideas can get ‘locked-in’ at the national level as they did in Brazil, or why they can become ‘unlocked’ as they did in Argentina. As I will show in the Argentine case, policy ideas can compete with one another even while one is internationally hegemonic (Skogstad and Schmidt, 2011: 9). This research is based on in-depth qualitative case studies and elite interview material obtained from each country by the author. All interview material was triangulated across a variety of agencies at multiple levels of government.
The transnational ‘ideas’ behind CCTs
The article frames this new policy paradigm as transnational because the programmatic ideas, policy goals and discourses that led to pioneering policy experimentation in Latin America – and the eventual formulation of coherent policy instruments to reduce poverty (anti-poverty human-capital based CCTs) – have their roots beyond regional actors and arose from multiple sources. As Béland and Orenstein assert in this issue of GSP, ‘beyond the role of financial loans and material interests in policy diffusion and transnational influence, ideational processes play a major role in shaping domestic policy development’. In the case of cash transfer programmes, Brazilian ex-President Fernando Henrique Cardoso clearly states in a televised interview that, although the actual subnational and eventual national programme experiments of BE were Brazilian, the basic idea of combining short-term poverty alleviation with long-term human capital investment came from within the World Bank during the 1980s following the failure of a neoliberal discourse. 2 This earlier discourse objected to almost any state intervention in development policies; its eventual failure came with the legitimation of a newer discourse centred on ideas of human capital development. These latter ideas were foundationally based on the work of key development economists, such as Amartya Sen.
The new interpretative framework that was used to understand poverty was heavily supported in the developing world by international institutions such as the World Bank, the Inter-American Development Bank, as well as a host of UN bodies like UNICEF and UNESCO. These agencies reached a broad international consensus during the 1990s over what the World Bank termed a ‘two-pronged approach’ to social protection: economic growth combined with investment in poor people’s human capital, which would be expanded through investments in basic health care and primary education – well-targeted safety nets provided by the state (Lipton and Ravaillion, 1995: 2571). Therefore, it is feasible to claim, using Hall’s term, that ‘the first order change’, or ‘the overarching goals that guide policy in a particular field’ (1993: 278), that motivated Latin American states to experiment with targeted cash transfer programmes in the mid-1990s was a broadly conceived transnational policy paradigm based on the idea of targeted cash transfers and investing in human capital to alleviate poverty.
Broadly stated, CCTs are social programmes that transfer small amounts of cash to socially vulnerable families who are usually selected through some form of means testing. In return, programme recipients must meet certain conditions that are linked to fulfilling social development goals in the areas of health, education and nutrition. This means meeting the conditions that are intended to break the intergenerational transmission of poverty, rather than the actual transfer of cash. CCTs are a form of non-contributory social protection, a policy area that, in theory, resides within the jurisdiction of local governments because of its association to social assistance; yet, in practice, CCTs can, and have operated at all three levels of government in both countries.
The development of CCT programmes in Latin America
Brazil and Mexico were pioneers in developing what Hall (1993: 278) called ‘the second order change’ – the techniques or policy instruments used to develop targeted social safety nets. The original CCT programmes PROGRESA (Mexico) and BE (Brazil) can be considered indigenous programmes, in the sense that they were initially designed and financed without the help of development banks (Handa and Davis, 2006: 514). As already stated, however, both the motivating foundational idea (to target poverty with human capital investments) and the subsequent country-level developments in Mexico in 2002, from rural to urban coverage, and in Brazil in 2003, from a compartmentalized (education only) to an integrative approach (health, education and nutrition), were heavily influenced by transnational policy actors and networks, backed by international financing, and ‘globalized’ through international media proliferation. 3 As the World Development Report 2012: Gender Equality and Development shows, this policy paradigm has become so entrenched that highly influential development agencies have (somewhat alarmingly) reduced social policy and social protection discussions almost exclusively to CCTs, in particular with a focus on Sub-Saharan Africa.
Following the ‘municipal era’ of CCT development in Brazil (1995–2001), many other countries in the region – Mexico (1997), Honduras (1998), Colombia, Costa Rica and Nicaragua (2000), Brazil and Jamaica (2001) Argentina and Chile (2002), Ecuador (2003), Dominican Republic, El Salvador, Paraguay, Peru and Uruguay (2005) – created national programmes based on the idea of using cash transfer programmes to combat poverty. 4 In 1995, this policy instrument was first experimented with in two Brazilian cities, Brasilia and Campinas. This article does not seek to uncover either the causal mechanisms that propelled the spread of CCTs throughout the region because it has already been well documented (Sugiyama, 2008, 2011, 2012), nor will it examine the empirical impact of such programmes on combating poverty, also well documented (Handa and Davis, 2006; Hoddinot et al., 2000; Parker et al., 2008; Soares et al., 2010). Rather, the article starts from the assumption that the idea of using CCTs as an effective poverty-alleviating policy instrument has reached a broad consensus among international organizations and national governments (Barrientos, 2004, Handa and Davis, 2006; World Bank, 2012). This consensus and the rate of emulation and spread of this policy are largely attributed to the early and rapid successes of national-level CCTs in Mexico (1997) and Brazil (2001) and their established positive impact. Beyond providing high promise, these successes garnered international support for this policy instrument. According to Weyland’s theory of policy diffusion (2007: 8), ‘the representativeness heuristic induces policy-makers to jump on the bandwagon of a diffusion process’. Beyond Weyland’s idea, which essentially means that CCTs represent an available solution to a pressing problem, the normative idea behind CCTs is symbolically embedded in a rights-based framework that is appealing to a far-reaching international community. Although commonly described by words such as ‘targeting efficiency’ and ‘programme effectiveness’, CCTs are also framed using both the discourses of citizenship and social justice. These normative frameworks have enabled the idea of CCTs to be both nationally appealing and internationally legitimate.
Brazil
The Brazilian transformation of Bolsa Escola (BE) (2001–2003) into Bolsa Familia (BF) (2003–present) is about ‘politics as social learning’ combined with ‘politics as power’. The unique political and historical circumstances that facilitated the original policy experimentation at the subnational level in Brazil during the early 1990s were created out of both necessity and opportunity. The post-1988 period of radical decentralization left many cities in need in terms of social services that were no longer nationally provided, as well as with new political and administrative opportunities to provide these services emanating from the National Constitution, which was completely rewritten in 1988. Out of this period of Brazilian history, from 1988 to 1995, a plethora of influential policy ideas were generated.
One of the more successful ideas to arise out of this period of Brazilian history was the use of non-contributory cash transfers to invest in basic education, an idea that has now been internationally diffused to countries as far away as Bangladesh and Indonesia. Even though Mexico’s national experience, based on the work of Santiago Levy’s 5 (1991) paper, ‘Extreme poverty in Mexico: A policy proposal’, pre-dated Brazil’s by three-and-a-half years, in the world’s perception, the fame, success and sheer size of BF have forever tied CCTs to Brazil. Although Brazilian CCT experts assert that Bolsa Escola was not heavily influenced by Mexico’s experiments (interviews E Suplicy, São Paulo, 2006, and A Fonseca, Brasilia, 2006), we know that ‘learning from foreign models such as the inclusion of health monitoring as part of conditionality, [did] occur after the national program was well-established and reframed under [BF]’ (Sugiyama, 2012: 42). Most authors argue that both countries’ pioneering experiences contributed to the development of what is today considered a consolidated CCT model; however, Mexican policymakers are given credit for the development of the first integrated model that included health and nutrition monitoring as part of its conditionality (Aguilar and Araújo, 2002; Sugiyama, 2012; World Bank, 2001). Transnational policy ideas thus contributed to the consolidation of a domestic policy paradigm in poverty alleviation, but were conditioned by a variety of policy feedback effects from earlier initiatives indigenous to Brazil.
Early Brazilian strategies for alleviating poverty circled around two principal policy ideas. The first, which can be traced to two Brazilian economists trained in the United States during the 1970s, is the guaranteed minimum income proposal. Relying on the ideas of citizenship and social justice, minimum income proposals are based on the assumption that the guarantee of a subsistence-based threshold is a basic social right of citizenship and should thus be guaranteed by the state. This policy idea became further embedded in Brazil thorough the long-term efforts of São Paulo Senator Eduardo Suplicy of the Brazilian Workers’ Party (Partido dos Trabalhadores), who first proposed it to congress as ‘Law Project 80’ in 1991. In terms of concrete policies, the first cash transfer programme revolving around the idea of minimum income was launched in Campinas, São Paulo, in 1995. The Campinas programme, Renda Minima, provided a small cash amount per capita, but it was not conditional. Minimum income programmes focus on the rights of individual citizens, not families. The programme was implemented by a PSDB (Brazilian Social Democrat Party) affiliated Mayor, Jose Malgalhães Teixeira, an important policy entrepreneur who, along with Suplicy, propelled the idea of basic income onto the Brazilian social policy agenda.
The second domestic policy idea developed to combat poverty and promote social inclusion in Brazil entered into policy discussions around 1987, following the circulation of Cristovam Buarque’s unpublished policy document, ‘An agenda for Brazil’ (Aguiar and Araújo, 2002: 39). Buarque’s idea was to develop a minimum income programme that would be conditional on educational attainment, with the goal of increasing school attendance among the poor. By 1994, the idea of using cash transfers for the poor attached to concrete policy goals in education was clearly linked to the platforms of the Workers’ Party (PT) of Brazil. During 1994 presidential election campaigns, the idea was included in the platforms of PT candidate Luis ‘Lula’ Ignacio da Silva. Following the election of Buarque (PT) as the governor of Brasilia, the federal capital of Brazil, BE was launched in that jurisdiction.
From this point forward, BE spread rapidly across municipalities in Brazil, and by 2001, more than 200 cities had a local version of this programme (Sugiyama, 2008: 198). According to Sugiyama’s empirical analysis (2008), electoral competition did not spur the diffusion of local programmes; rather, ideological and sociological motivations compelled local elites to emulate Buarque’s initiative. Such initiatives were also institutionally back by the National Secretary of Social Assistance, which had been created that same year, and further promoted the idea of social inclusion as a social right.
In 1998, under a left-of-centre PSDB administration, a further fiscal incentive was created to propel the diffusion of BE at the subnational level. This National Minimum Income Guarantee Programme provided fiscal transfers to poor municipalities by covering 50% of their costs in providing CCTs linked to school attendance. Eventually, this same government, under the leadership of President Fernando Henrique Cardoso, created the national social programme BE. Cash transfers from this CCT were designated to mothers of households with dependent children and were also conditional on records of school attendance. Additionally, it consolidated several innovations, such as delivering cash benefits through electronic bankcards. Early successes from this national programme attracted a lot of attention from international organizations such as UNESCO, UNICEF and the International Labour Organization. By 2002, BE peaked, with over 5 million households included (Ministério do Desenvolvimento Social e Combate à Fome [MDS]).
The nationalization of BE consolidated the ideational and symbolic legacies of the two main ideas designed to alleviate poverty that had been circulating in Brazilian policy circles and within international epistemic communities. It also had the dubious challenge of creating new administrative capacities at both the national and municipal levels, which were responsible for selecting and registering eligible families. Still, many organizational difficulties surrounded the implementation of the programme. Of these, a consolidated national database of socially vulnerable families who might meet the programme criteria was not yet fully operational. Particularly difficult was the coexistence of programmes across a variety of social policy sectors. It was suggested, in various interviews with policy experts, that the causes of weak coordination during this period, which initially extended into the next administration, were policy-oriented and driven by intra-ministerial conflict, rather than partisanship (interviews A Fonseca, Brasilia, 2006, and M Pochmann, Campinas, 2006).
Negative feedbacks, like positive feedbacks, can explain major shifts in policy regimes (Weaver, 2010: 138). Although not specifically spelled out in Weaver’s language, negative learning from the BE experience has been cited as the frequent impetus for the administrative reforms evident in the much more successful Bolsa Familia, launched by President Lula in 2003 (Fenwick, 2009; Hall, 2006). In line with Weaver’s theoretical contribution (2010), which is focused on advanced industrial countries, BE’s negative feedback effects created policy challenges related to concerns about its efficiency, in particular around the issue of targeting and monitoring. As Hall (2006) and Fenwick (2009) both show, this created opportunities for the incremental reforms of BF, such as streamlining application processes and the completion of the Cadastro Unico (unified database) project – all of which contributed to the massive expansion of CCTs in Brazil under President Lula (2003–2010).
A ‘politics as power’ story also contributed to explaining Brazil’s policy shift in 2003. Lula was motivated to transform and expand BE into a bigger and better CCT, because he was elected in 2002 with a clear majority of 61.3%, on the basis of his promises to improve the living conditions of ordinary Brazilians though raising the minimum wage and alleviating hunger, misery and poverty. Moreover, according to a credible public opinion survey taken in 2002 (IBOPE, 2002: #558), of the 52 campaign promises Lula made prior to his election, the three that the public believed he could accomplish in his four-year term were: alleviating hunger, misery and poverty (24%); creating jobs (17%); and raising the minimum salary (10%). On the basis of public expectations, Lula’s government had a clear mandate to prioritize poverty alleviation. Before he even took office, this issue was on the federal agenda.
Moreover, poverty was still a pressing issue in Brazil. The proportion of Brazilians living below the poverty line had fallen 18.47 points from 1993 to 1995 because of macroeconomic stabilization, but from 1995 onwards, it remained stable (FGV, 2006). The 1999 devaluation and the Argentine fiscal crisis of 2001 led to renewed fears of rising poverty throughout Brazilian society. BF signified a shift towards a broader policy approach that included health, nutrition, education, as well as including families in extreme poverty who did not have dependent school-age children. It essentially departed from the narrow emphasis on educational attainment and children developed nationally in Brazil since 2001, moving towards a model more closely resembling that of Mexico’s integrated approach.
Argentina
The Argentine story of CCT evolution differs greatly from Brazil’s. In contrast with the rest of the region, it was a late adopter of this policy model; however, many important lessons can be found within its experiences for other nations currently experimenting with CCT implementation. Essentially, it offers a story about how policy feedbacks impacted the domestic policy development of a transnational idea, and how two policy ideas simultaneously competed with one another until one failed.
It is well documented in the literature that Argentina is a country that has confronted serious setbacks and difficulties in producing effective public policy (see, for example, Spiller and Tommasi, 2008). One of the sectors with notable deficiencies is social protection. Although many social protection initiatives existed prior to the Argentine fiscal crisis that unfolded in late 2001, because of a severe rise in poverty rates post-2002, poverty alleviation became a federal priority. When President N Kirchner of the Peronist Party (PJ) was elected in 2003, he attempted to deliver desperately needed public goods to socially vulnerable families by ending a previously implemented emergency workfare programme called Plan Jefes y Jefas de Hogar Desocupados (PJJHD), 6 replacing it with Programa Familias (PF), an integrated CCT programme. Unfortunately, the federal administration was unable to successfully discontinue the previous workfare programme and transfer its beneficiaries to the new CCT programme.
The outcomes of Argentina’s PF were largely unsuccessful. It was replaced in 2009 with a more formalized social protection initiative linked to the National Administration of Social Security (ANSES). In 2006, three years after the inception of PF, only 21% of Argentine municipalities were included in this CCT, delivering benefits to just 372,000 families. PJJHD, the previously implemented workfare programme that was intended as a short-term emergency initiative, continued to make over 1.3 million payments per month, even though the actual registry of the programme had been closed since May 2002. 7
Context of the PJJHD
Context is important to understanding the history of PJJHD. On 18 and 19 December 2001, social mobilization exploded in violent protests on the streets of Buenos Aires against the freezing of citizens’ cash assets. These protests quickly formed diverse social movements organized by factions of piqueteros (groups of unemployed labourers) who, armed with rocks and sticks, contributed to the worsening of political instability. On 20 December 2001, democratically elected President Fernando de la Rúa of the Alianza, an electoral coalition of the centre-left party FREPASO and the centrist Unión Cívica Radical (UCR), resigned amid rampant protests and looting on the streets Buenos Aires and its conurbation. Beyond the well-cited details of Argentina’s fiscal crisis, the federal government’s weak leverage over the 23 provinces, which had fallen in tandem with Menem’s popularity in 1999, contributed to the depth and severity of the 2001–2002 political and economic crisis. According to Levitsky and Murillo (2003: 153), Menem left three legacies: a dramatic growth in public debt, an expensive privatized social security system and widespread social exclusion. In essence, PJJHD was President Duhalde’s (appointed after de la Rúa’s resignation) politically ingenious solution to remedy the legacies that were associated with his political party’s collapse.
In 2002, PJJHD, a non-contributory cash transfer programme that was conditional on community-based work contributions, replaced all previously existing national and provincial workfare programmes. Its introduction was facilitated by the existence of previous workfare programmes of the Menem era, particularly the national Plan Trabajar (1996–2000). Given the short time frame and chaotic nature in which PJJHD was implemented, Plan Trabajar undoubtedly facilitated the administrative implementation of this second national workfare programme. Another programme established at this time, which remained largely unknown, was PF. Designed as a CCT to invest in human capital development, it was supported by the IDB (Inter-American Development Bank). Its continuation was an integral part of the negotiations for the continued external financing of PJJHD in 2003.
The political power inherent in PJJHD is a large part of this programme’s ‘stickiness’. When Kirchner won the 2003 presidential elections with only 22.4% of total votes in the first round, it was in large part because of the electoral support from the province of Buenos Aires, which had been engineered by the outgoing president, Duhalde. Although, as noted above, PJJHD’s registration had been officially closed in May 2002, cross-provincial fluctuations in its allocation continued until they peaked in May 2003. 8 As the presidential elections were on 27 April 2003, Kirchner inherited PJJHD, its powerful ‘organized’ groups of unemployed workers and the powerful Ministry of Labour that administered the programme (a bureaucracy that was loyal to the outgoing president). To foster the necessary vertical party alignment required to govern Argentina, Kirchner had to forge his leadership as distinct from his predecessors and legitimize his power by creating new political bases within the party. In 2004, he discontinued PJJHD and re-launched an existing but inactive programme under the new name Programa Familias.
By October 2004, the government’s official plan was to restructure PJJHD and build an integrative CCT model within 180 days (Decree 1506/04). The policy idea – and its solution to pressing problems existent in Argentina – came from Mexico (interviews MDS and PNUD, 2006), and was supported by the IDB in Argentina. According to interviews and official documentation, the intended strategy was to transfer, by February 2005, former PJJHD recipients who were eligible for one of three new programmes, el Hambre mas Urgente (a food programme), Manos a la Obra (a new workfare programme) and PF.
Given the number of PJJHD beneficiaries and their unwillingness to transfer voluntarily to the new initiative, a strange anomaly emerged for the targeting of PF. The formula targeted municipalities with the highest number of PJJHD recipients. This would mean that the previous programme’s informal and locally controlled distribution, which was rapidly implemented in the context of a social emergency, was now embedded in the distribution of this new programme. According to the office responsible for administering the new CCT, the formula used to target certain municipalities was based 10% on population, 30% on households with dependent children below the poverty line and 60% on existing PJJHD beneficiaries. Although this last variable can be justified on the basis of the domestic policy goal to discontinue PJJHD, it makes little sense from an international development perspective. Moreover, even though the new CCT officially included PJJHD recipients, from 2003 to 2006, PJJHD coverage only decreased by 33%, and of the households no longer receiving PJJHD, only 15% transferred to the new CCT. Therefore, the bulk of PF beneficiaries were new recipients, and informally, PJJHD continued to operate as a parallel programme until 2009, when the Asignación Universal por Hijo (AUH) was introduced based on the Universal Family Benefits system (see Table 1).
Discontinuation of PJJHD relative to the expansion of PF and AUH (numbers of benefit recipients).
Source: Ministry of Labour and Office of PF, Ministry of Social Development 2003–2006; Repetto and Langou (2010) 2007–2010.
The performance of PF
Why did Argentina’s first attempt at implementing an integrated CCT programme perform so poorly from 2003 to 2009? Beyond the obvious constraints placed on PF – in terms of having to dismantle the organized groups who had politically benefited from PJJHD, but would not benefit from PF because it was designed without local control (interview La Matanza, 2006) – overcoming a decade of Peronist workfare programmes was an administrative challenge that could not be easily overcome in a country with notoriously weak policy-making processes. Nevertheless, several administrative mechanisms were created to facilitate the transition from the Ministry of Labour’s back-to-work programme towards an integrated CCT designed to alleviate longer-term poverty.
First, the National Coordinative Council of Social Policy was given responsibility for overseeing the programme. Created in 2005, the Council’s mandate is to give economic, logistical and technical support to social policy implementation and monitoring to both the Ministry of Labour and the Ministry of Social Development. Beyond providing technical support, this Council was also responsible for crosschecking multiple databases to ensure that recipients were not ‘double dipping’, and were in fact eligible for support, based on each programme’s criteria.
The data crosscheck was designed to provide this information using SINTYS, a national IT system that, by law, crosschecks in-house inter-sector databases every three months. 9 Technical sources from the PF team and in the Ministry of Economy who were responsible for aggregating consolidated social expenditure data confirmed to the author that this well-intended monitoring tool was seldom used. In 2005, only one crosscheck was reported as completed, and, within the first nine months of 2006, there had been none. 10 The proliferation of bureaucratically embedded pre-existing programmes, each with its own database located in a separate bureaucracy, together with the creation of a new programme within a Ministry with a minor role in Argentine social policy history, and the fact that its oversight was tied closely to the Casa Rosada (the presidential office), facilitated the prevalence of executive authority and administrative chaos in this policy area.
At the local level, from 2004 to 2009, the administration of PF worked as follows. Based on criteria cited earlier in the article, the federal government selected each province, and assigned each a team in both the national and provincial programme offices. The provincial team was selected by shortlists (ternas), where the province shortlisted three candidates for each position to be filled and the national programme office then made the final selection. Some provinces, however, selected their teams autonomously, rejecting the use of ternas. In such cases, the federal office either denied the programme to that province, or accepted the provincially selected candidates. The same process existed for municipalities, without provincial intermediation. All employees, office equipment and infrastructure were paid for, and supplied by, the federal office. The local office was accountable to a counter-member of both the provincial and national offices, making it organizationally hierarchical. It was reported that there was frequent tension and conflict with the way the programme was administered. 11 All communication between the three levels of government was also hierarchical. The national office would contact the provincial office about a discrepancy, and the provincial office would then send someone out to the local office, that would then contact the family directly. This structure impeded intergovernmental collaboration by excluding local authorities from participating in the programme’s implementation.
Several factors have impeded the success of PF in Argentina and the discontinuation of the previous initiative, PJJHD. First, confusion surrounding both the former programme’s discontinuation and the creation of a new policy framework led to a proliferation of uncoordinated programmes with no clear national policy objective. 12 Second, the centralized manner in which this CCT was implemented limited local authorities from participating and generated local dissent. This stands in stark contrast to the informal opportunities that PJJHD provided to local officials, which, indirectly, contributed to the programme’s popularity at the municipal level.
Third, the ambiguous discontinuation of PJJHD and the creation of PF as a new programme intended to replace it within two completely separate bureaucratic entities made this transition extremely obtuse. An official of PJJHD in the Ministry of Labour stated that the goal of PF was to ‘absorb’ the recipients of the previous workfare programme, which had been ‘discontinued’. 13 This was confirmed by a senior official from the Ministry of Economy, who stated that, ‘formally yes, but, informally, nothing has occurred over three years and I cannot see that it will’. 14 In the Ministry of Social Development (MDS), the goals of the programme were reported differently.
The head of the MDS office responsible for PF stated that, in her estimation, there were three choices for federal policymakers in 2004: (1) leave each programme running; (2) unify them; or (3) reformulate them (interview, 2006). She stated that they opted to reformulate. In her words, ‘the transfer of eligible beneficiaries is being coordinated with all the municipalities; we meet with each mayor in order to coordinate it mutually’. 15 This statement was contradicted, however, by a senior representative from the municipality of La Matanza, who told the author that ‘the municipality has nothing to do with [PF] and has never been approached; it is a centralized program’. 16 The Minister of Social Development’s then right-hand official stated that the PF was intended, in the long term, to be the only CCT in Argentina, and that it would be conceived as a ‘right’. He further stated, in 2006, that it was his own, and the minister’s vision for PF to become an extension of the ‘family grants’ that are part of Argentina’s social security system. Over time, this would prove to be the government’s actual agenda, even though none of the other senior officials interviewed had any idea that PF was already in decline. 17 Evident during these interviews, which were carefully triangulated across multiple levels of government and offices, is the lack of any explicit policy goals, and, moreover, very little mention of the new CCT programme’s merit or goals, which they were all actively coordinated in implementing.
Asignación Universal por Hijo
Argentina’s most recent social protection programme, Asignación Universal por Hijo (AUH), has effectively replaced both PJJHD and PF. Implemented in 2010, this programme delivers a benefit of 220 pesos to each dependent child under the age of 18. Although it is still too early to thoroughly analyse, AUH stands out from other CCTs in Argentina and the region for several reasons. First, the programme ‘de facto’ is not based on conditionality, nor is it explicitly linked to human capital investments. Eighty percent of the benefit is transferred automatically to the family via the already existing family benefits system, which is a formal component of the social security system in Argentina and is controlled by the National Administration of Social Security (ANSES). Therefore, it is intended for families already officially registered within Argentina’s social security system whose income falls below the government-set benchmark. Moreover, the money to finance the AUH comes from formal company-based pension contributions.
Second, although purportedly non-contributory and universal in nature, recent media accounts attest that it is neither. Due to the public administration system that it relies on, the AUH does not include families that have never been registered formally within the social security system or those whose children have never attended school. In 2011, the World Bank pledged a US$3.6 million loan to increase the coverage of AUH, but no mention of strengthening either the conditional aspects of the programme or creating a system for their enforcement was made. It is widely considered, simply, a cash transfer. For these reasons, available analyses of its preliminary results have reported no changes in variables such as school attendance or vaccination rates (La Nación, 14 April 2012). The World Bank’s Argentine office claims that, in 2012, there were 3.6 million beneficiaries of the AUH; however, because of Argentina’s currently limited production and manipulation of official government statistics, the newspaper El Clarín (21 May 2012) currently claims there are no official numbers from ANSES regarding how many payments have been made per month during the last eight months. Moreover, with a currently estimated inflation rate of 22% and a set non-indexed minimum salary threshold required to receive the benefit, $5200 pesos, it has been reported that over 1 million families are no longer able to claim this benefit because of salary inflation. Much further research is required to evaluate this programme and to establish whether or not it can actually be considered a CCT programme.
Comparative analysis
What lessons can be drawn from a comparative analysis of these cases? Did the ideas and symbols of existing or past domestic policies influence the interpretation, framing, administrative processes and eventual outcomes of BF in Brazil or PF in Argentina? Did domestic policy feedbacks impede the impact of transnational policy ideas? Why was the effect different in each case? First, in both cases, an argument can be made about how previous initiatives created policy feedbacks, both positive and negative, shaping future policy development. The predominant type of feedback effect produced however differed in each case. Therefore, so does its impact on future policy development.
All factors are also present to make a diffusion-based argument about why both countries chose in 2003 to reform and shift their previous policies: both President Lula and President Kirchner had utilitarian goals (politics as power); there was broad public support at the time for social policy reform in this issue area; both countries had a pressing pre-existing problem (poverty and inequality); and both had an available alternative policy model, based on a broad integrated CCT which had been empirically established by 2003, to impact simultaneously on human capital development in the areas of health, nutrition and education. A diffusion-based argument, however, cannot fully explain the variation in outcomes of these two programmes, or why this now globally accepted policy instrument is locked-in in Brazil, but not in Argentina. It is true for these two cases and many others in the region that transnational policy ideas were entering Brazilian and Argentine policy space and reorienting their domestic policies, but much more attention needs to be given to the notion of path dependency and whether domestic policy feedbacks shape the development of these ubiquitous programmes.
Types of feedback effects
It is not new that similar policy models function differently in different contexts (Skogstad and Schmidt, 2011); yet, we still need to understand how and why they do so. In order for feedback research to continue improving our understanding of politics and policy development within the context of an increasingly interconnected and globalized world, this scholarship must specify which types of feedback impede transnational ideas from impacting upon domestic policy development, and how they do operate. Along the trajectory of CCT development and its programmatic shifts in Argentina and Brazil, many types of policy feedbacks can be identified operating in a variety of ways, all of which uniquely impacted the institutionalization of anti-poverty programmes. Policy development is not always a result of policy feedbacks, but some types of feedback will impede the ability of transnational ideas to impact upon domestic policy development more than others.
State-building effects
State-building effects refer to how previously existing policies have the power to transform and expand the capacities of the state by changing the administrative possibilities for future policy initiatives (Skocpol, 1992). Within the poverty alleviation shifts of both Brazil and Argentina, state-building effects that became impediments to future initiatives are most visible in the Argentine case. The state-building effects from PJJHD impeded the eventual success of PF (2003–2009) by blocking the required inter-ministerial cooperation and limiting the share of information required for effective targeting.
This transfer of knowledge from the more experienced Ministry of Labour to the Ministry of Social Development was also limited by divided political loyalties among the ministries themselves. In contrast with Brazil, which has been governed since 1995 by a presidential-led coalition government, inter-ministerial conflict in Argentina at the federal level is partisan-based. This is largely attributed to the majoritarian nature of Argentina’s political dynamics and the high levels of intra-party fragmentation within the Justicialist Party (PJ). Although the old Peronist leaderships of Menem and Duhalde are long-time enemies, both controlled traditional sectors of the PJ that were embedded in the Ministry of Labour, which played a strategic role in administering PJ workfare programmes.
The administrative capacities surrounding workfare programmes concentrated within the Ministry of Labour had been cumulative since the implementation of President Menem’s Plan Trabajar (1995–2000). These same administrative capacities and bureaucratic actors played key roles in implementing interim President Duhalde’s emergency workfare programme, PJJHD (2002–2009). The ability to transfer the administrative capacities and knowledge, including the willingness of ministry bureaucrats to do so, over to President Kirchner’s bureaucratic terrain in the Ministry of Social Development headed by his sister, was paramount to PF’s continuity. Unfortunately, the state-building effects of PJJHD limited the transfer of its recipients over to the newly created programme and compromised its effectiveness through both competition and duplication.
The replacement of Brazil’s BE to BF required a transfer of policy authority from the Ministry of Education to Lula’s newly created Ministry of Social Development. Many of the previous administration’s coalition members were also members of the governing coalition that Lula headed in 2003, facilitating the willingness of many actors to cooperate with the intended reforms. Lula’s first initiative however, Fome Zero, actually led to the creation of a fourth CCT in Brazil linked to nutrition, and further added to the complex administrative chaos behind CCTs that had been inherited from President Cardoso’s administration (1995–2002). It was, in fact, some of the negative state-building effects of Cardoso’s three CCTs that Lula then made four, which contributed to the improved administrative model of BF.
The positive state-building effects from BE, in particular its decentralized model, which required federal–municipal cooperation, and the single national database that was created in 2001 and is used for targeting all CCT programme beneficiaries, were easily transferred to the organizational architecture of the new BF programme. It was therefore partially because of the nature of Brazilian coalitional dynamics and the procedural decisions that were taken, that both the negative and positive state-building effects of BE contributed to the success of BF.
Interest group effects
According to Pierson (1994), social programmes favour the emergence of massive constituencies that actively support these programmes, making them difficult to reform. For both Skocpol (1992) and Pierson (1993), interest groups are primarily group-based. As already mentioned, three years after the initial implementation of PF, only 372,000 families were included, whereas 1.3 million payments continued to be made every month to workfare recipients who remained in PJJHD. Organized unemployed workers formed a powerful interest group in Argentina, which helps to explain part of the stickiness of PJJHD. In order to secure the programme continuance for their own benefit, highly organized unemployed workers and their leaders captured the programme early in its inception. As one ex-senior policy official in the Province of Buenos Aires told the author, ‘PJJHD was designed to be captured’ (interview PNUD, 2006).
Argentine workfare programmes are structured around locally initiated public works projects, a characteristic that had been pushed for and supported by mayors from around the country. 18 Although these programmes were intended to quell growing societal demands caused by rising poverty, economic stagnation and accusations of government corruption (Ronconi, 2002), they also had the non-intended effect of creating interest groups. Powerful and organized local groups, municipalities and social organizations presented community-based projects to the federal government to be financed by PJJHD. When the project was accepted, municipal authorities would distribute benefits to unemployed workers below the poverty line who had signed up to work on that project.
It is well known in the literature that Argentine workfare programmes produced many unintended consequences (Giraudy, 2007; Weitz-Shapiro, 2006). Of particular interest, the more informal distortions had a greater impact on interest group formation than the programme’s formal intentions. Informally, workfare programmes in Argentina were used for clientelistic ends, determined by partisanship or by levels of organized protest. Notwithstanding, the idea of centralized financing and local control impacted municipal governance. The ingenuity of decentralized workfare programmes was that they politically empowered municipalities from the top-down by generating political power from the bottom-up through community work projects. In this way, strategically located municipalities were plugged into a party-driven political machine.
It was the size and importance of this interest group, and Duhalde’s ability to manage the extremely delicate social situation of Argentina in 2001–2002, that made it so difficult for Kirchner to dismantle and discontinue the programme. PJJHD illustrates how the development of a specific policy can fuel interest group formation, which in turn impedes future policy development and policy-making. This is consistent with the policy feedback literature, which stresses the weight of existing programmes on future policy decisions.
The chief difference between the interest group formation of PJJHD and BE was that Argentina’s program was composed of formerly employed workers who had lost their jobs during the 2001–2002 crisis and were highly organized by union leaders, political activists and politicians, chiefly at the local level. When Brazil’s BE peaked in 2003, with around 5 million beneficiaries, these recipients were not formed into organized groups, and they had few links with other local organized groups. Education-sector policy experts, bureaucrats, NGOs and epistemic communities from within the educational sector had been heavily involved in BE; however, many of these same actors were also included in BF because the education component remained an integral part of the new programme. In fact, Cristovam Buarque, the chief policy entrepreneur behind BE, was Lula’s first Minister of Education. Moreover, the initial recipients did not change – all 5 million recipients from BE transferred voluntarily to BF. In return for accepting more conditions, there was a fiscal incentive of a higher benefit, so there was no need to change the structure of what was a relatively weak interest group between the two Brazilian programmes.
Ideational and symbolic legacies
The ideas and symbols of established policies can influence legislative outcomes while empowering the actors who draw on these ideas and categories (Béland, 2010: 580). These ideas and symbols, which are drawn from the language, framing and interpretive discourses surrounding certain programmes and the people to whom they are entitled, have a powerful impact on future policy development.
In Argentina, it is not easy to locate an interpretive discourse that was continuous from the era of workfare programmes (the 1990s) to the era of CCTs based on human capital investments (the 2000s). Workfare recipients had been framed as ‘victims’ of both economic neoliberalism and the Argentine fiscal crisis. The state employed a language of ‘social deficit’ that was initially invoked in the mainstream media to raise public support for PJJHD, and which helped to portray early government actions as positive. Within congressional debates, which always resulted in prolonging the financing of PJJHD, even though its registration was long-closed, this language was used to invoke feelings of national responsibility. According to a survey conducted by CEDLAS in collaboration with the World Bank, in 2007, the best-known national social programme remained PJJHD (Duhalde’s legacy), followed by Plan Trabajar (Menem’s legacy), with PF (Kirchner’s legacy) last (CEDLAS, 2007). It is incredible that PJJHD’s symbolic legacy was so prolonged, given the fact that ‘officially’, no new families had been able to join the programme since 2002.
The language used to frame PF did not succeed in achieving equivalent public visibility. Moreover, it was designed to be very different from the symbols and ideas of the previous programmes. The rights-based language that was associated with the rights of workers was used by those who fought for the PJJHD, but became, over time, discredited within the media and some social groups because of alleged links between these groups and government corruption (El Clarin, 22 May 2009), thus, a rights-based discourse was seldom employed to frame PF. Although this move was intended to demarcate a clear policy shift, human capital-based CCT programmes in Argentina had no ideational or symbolic legacies upon which to draw. By contrast, PJJHD had drawn on the legacies of ‘workers without shirts’ from the time of Evita and Juan Peron after the Great Depression. In the media, Duhalde expounded about how community work projects would put ‘Argentina on its feet and in peace’ – powerful nationalist symbols and sentiment.
Actors’ abilities to draw upon already established ideas and symbols in support of Brazilian CCT policy reforms were very different than the Argentine case. In Brazil, the symbols and ideas used by key actors, such as Senator Suplicy (PT), Senator Buraque (PT-PDT [Partido Democrático Trabalhista]), Patrus Ananias (Minister of Social Development 2003–2010) and President Lula (PT) to expand, reform and garner public and legislative support for BF were already firmly embedded in public discourse. The language of social protection in Brazil, in which the first two popular city-based programmes Bolsa Escola and Renda Minima were based, has become a powerful ideological tool in the politics of poverty alleviation, 17 years after it was enacted in Brasilia by Governor Buarque (PT) and in Campinas by Mayor Teixeira (PSDB). The language of ‘citizenship rights’, ‘social vulnerability’, ‘the deserving poor’, ‘gender’, ‘family’, ‘universalism’ and ‘children’ has strong symbolic and political meanings in official and public debates surrounding social policy reform in Brazil.
The ideas and symbols behind this language, which have been used by key actors from different political parties over the past four presidential administrations to further state-centred poverty alleviation efforts, are grounded in the idea of transferring short amounts of cash to all families in Brazil under the poverty line because they are the deserving poor. An internal national debate that began in the 1970s over whether cash transfers should be conditional or not, does continue, but elements of both policy ideas can be found in official legislation and in the actual practices of BF. 19 Brazilian supporters of a more universal Minimum Income Scheme have never publicly or politically opposed the expansion of targeted CCTs; rather, they have argued they do not go far enough.
Policy feedback and political behaviour
A very recent area of feedback scholarship now focuses on how public policy may affect electoral and individually based political participation (Campbell, 2002), a feedback effect that, if established, will certainly have a future ‘lock-in’ effect on certain popular social policies. Research on Mexico’s pre-Opportunidades programmes and on Argentina’s workfare programmes has certainly focused on whether or not these programmes were manipulated for electoral ends. PJJHD has been proven to affect group behaviour (Giraudy, 2007), but this is related to interest group formation, not individual political behaviour.
The case of BF is quite unique, in that many scholars have intensely researched whether the success of CCTs in Brazil has established a polarization effect among the less- and more-developed regions of the country, an electoral shift which has previously never been witnessed in Brazil. Of great interest for this area of feedback research, it has been established that this programme did have an effect on electoral behaviour and political participation in the 2006 presidential elections (Licio et al., 2009; Hunter and Power, 2007; Nicolau and Peixoto, 2007; Soares and Terron, 2008; Zucco, 2008).
Lock-in effects
‘Lock-in’ effects refer to the well-established, path-dependent logic of public policy. Based on this logic, it is assumed that it is very difficult for a newly elected official to dismantle, replace, or transform certain policies because of the powerful effects of previously implemented policies (Béland, 2010). However, not all policies produce feedback effects that are resistant to change. BE produced important state-building effects that can be considered both positive and negative. It also produced important ideas and symbols that continue to be reproduced within the newest Brazilian President Dilma Rouseff’s initiative Sem Miseria, that includes BF, but also continues to expand and include more citizens. However, Lula was able to dismantle, replace and transform BE. Far from being path-dependent, he was able to use many of its feedback effects to go further, aided by the backing of transnational policy ideas that were highly influential at the global level. As I have tried to show, none of the ideas and symbols behind these transnational policy ideas, nor the required political institutions and administrative capacities, were absent in the Brazilian case. Therefore, is BF locked-in? It was evident in the most recent 2010 presidential elections that all the presidential candidates made an electoral promise to maintain and continue the programme, but this does not mean that new travelling policy ideas will not be invoked to bring about further reform through incremental change, consistently moving and evoking a familiar discourse.
Diffused policy ideas from both Brazil and Mexico may have influenced President Kirchner’s decision to attempt to dismantle, replace and transform the transitional goals of the PJJHD into a national CCT based on human capital investment. Consistent with the policy feedback literature however, PJJHD’s own success (1.7 million beneficiaries) had considerably deepened existent administrative procedures and bureaucratic capacities, as well as having created a powerful interest group that was antithetic to a CCT programme based on social development goals. Moreover, PJJHD further embedded the symbolic status of the Ministry of Labour and the legacy of Peronist-based workfare programmes within the collective imagination.
These powerful ideational and symbolic effects impeded the future development of a CCT designed to alleviate poverty not linked to Argentina’s traditional labour sector, even if this CCT had significant technical and financial backing from the IDB. Argentine president Cristina Fernandez de Kirchner’s latest poverty alleviation initiative, the AUH, has so far shown more promise. Under the control of ANSES, which is incidentally under the jurisdiction of the Ministry of Labour, the AUH draws upon the well-established ideational and symbolic legacies of Argentina’s social security system, which famously promoted its middle class. Although the AUH has been criticized by independent organizations and the media for not reaching the most vulnerable families, because of its attachment to a formal contributory system of social security (as well as its weak emphasis on ending the intergenerational transmission of poverty), within one year it has replaced both PJJHD and PF.
Conclusions
Policy feedbacks do not always ‘lock-in’ poverty alleviation programmes, and transnational policy ideas will not always trump the power of previously established domestic policies. Using the cases of CCT development in Brazil and Argentina, this article suggests that what matters most is which type of feedback effect intersects with transnational policy ideas, and how it operates. The important lesson is that some may be encouraging, while other may impede domestic impact.
In Brazil, the availability and international backing of an integrated CCT model opened up tremendous reform opportunities for key policy actors. These policy shifts, however, did not counteract with any of BE’s ideational and symbolic legacies, neither was it dependent on dismantling an established interest group, nor did the localized version which was far more decentralized than Mexico’s (and was based on Brazilian experiences) work against the administrative capacities and bureaucratic tendencies of Brazil.
In Argentina, all of the elements required to make a diffusion-based argument were present when Kirchner attempted to discontinue PJJHD and replace it with a Mexican-inspired CCT model. This policy shift was impeded, however, by the administrative capacities and bureaucratic tendencies built up during a decade of workfare programmes that were based on local involvement (and capture), by its lack of relevance to Argentine ideational and symbolic social policies legacies, by a powerful interest group of unemployed labourers and organized protesters, and by the lock-in effects of PJJHD – even though the architecture of PF had been localized to placate some of these effects by trying to include some former workfare recipients. It is worth noting, however, that Argentina’s latest poverty alleviation strategy is linked to Argentine ideational and symbolic legacies. It is therefore with great interest that scholars should follow its progression.
Footnotes
Acknowledgements
The author would like to thank Daniel Béland for his theoretical insights and generous support of this research. Thanks also to Anthony Hall for important comments on previous work, and to the anonymous reviewers of Global Social Policy for their valuable feedback.
Funding
This research was funded by the Johnson-Shoyama Graduate School of Public Policy, University of Saskatchewan and a fieldwork grant from the Latin American Centre of the University of Oxford.