1Until the early 1990s, social protection was marginal to mainstream understandings of development, primarily due to the association of the concept with either the social security of wealthy nations or contributory social insurance programmes for workers in the modern sector. For its part, the International Labour Organization (ILO) – the key international organisation operating in this field – , continued its efforts to extend social coverage to workers, but did not incorporate populations in the informal sector. The idea of extending non-contribution based social security to non-salaried populations was considered both prohibitively expensive and likely to reinforce a “culture of poverty.” This critique was taken even further during the economic liberalisation of the 1980s. The World Bank rejected social protection programmes for workers as economically harmful and socially unjust. Only very minimal safety nets, reserved for the poorest of socially vulnerable populations, were considered acceptable.
2In the late 1990s, however, driven by the disenchantment with the performance of economic adjustment programmes, the 1997 Asian economic crisis and a heightened awareness of the negative impact of global poverty, the dominant paradigm changed. Social protection thus became a preferred instrument of the Millennium Development Goals, while the World Bank promoted social protection as a key component of international poverty reduction strategies (social risk management) (World Bank, 2001). The ILO took the initiative of mounting a global campaign to extend social security to developing nations, the Social Protection Floor Initiative (ILO and WHO, 2009). The United Nations Development Programme (UNDP) underscored the vital role of social protection in development policy. In the United Kingdom, the Department for International Development (DfID) placed social protection at the centre of its policies. Several major international conferences (such as Livingstone in 2003, Arusha in 2007 and Dakar in 2008) centred on the theme of social protection and development were initiated by or in cooperation with the World Bank, the DfID and the United Nations (UN). Successful social protection programmes developed in the Global South – such as Brazilian and South African social pension schemes and conditional cash transfers (CCT) established in Mexico and Brazil – were adopted as model programmes at the global level.
3This consensus in favour of social protection represents a fundamental paradigm shift. Social protection in developing countries is no longer perceived as a short-term means of ameliorating economic shocks, but rather as a global policy (Voipio, 2007) combining cash transfer programmes for extremely vulnerable populations, new programmes incorporating a social investment perspective into social transfer policies (Jenson, 2008), and both public and private social insurance programs for formal sector workers. This consensus, however, is not unanimous, and is vulnerable to political manipulation. Despite a degree of political alignment, perspectives on risk management, social needs and social rights continue to clash (Voipio, 2007). For this reason, it is vital to step back from the soft consensus currently dominating thought on social protection. Social protection choices are not a simple technical matter, but require the gradual implementation of social learning in order to enable the creation of increasingly judicious policies (Barrientos and Hulme, 2008). Social protection discourse is still a subject of discussion in some epistemic communities and institutions (Merrien and Mendy, 2010). Overall, the discourse reflects a relatively coherent set of values and stable analytic framework through which social insecurity issues are evaluated and policy responses are devised.
4This article aims to analyse the emergence of social protection as a legitimate concern in the field of development policy, discussing the issues –actual and symbolic – inherent to the international debate regarding the role and nature of social protection in the fight against poverty. It is from this perspective that the paper examines the various types of programmes promoted by the international community, with an emphasis on conditional cash transfer (CCT) programmes. It concludes with an assessment of the relative appropriateness of social protection policies for developing countries.
5In general, social protection policies were first introduced in developing countries following the Second World War. According to the terms of the ILO Convention, 1952 (No. 102), social protection encompasses social security policies aimed at protecting workers from social risks. The convention identifies nine areas in which must be included in the provision of social security: medical care, sickness, unemployment, old age, employment injury, family, maternity, invalidity and survivors’ benefits. It also establishes the minimum level of benefits to be provided.
6The dissemination of Bismarckian, Beveridgean and 'liberal' models of social protection to independent Latin American and Asian states and the colonised countries of Africa and Asia essentially aimed to cover employees in the public sector and the so-called “modern” private sector (Bailey, 2004; Gough and Wood, 2004; Merrien et al., 2005). The ILO thus came to play a seminal role in producing and disseminating international social protection ideas, values and standards (Strang and Chang, 1993). During the phase of industrialisation by import substitution, and under the influence of the former colonial powers and the ILO, modernising elites in developing states began to understand the extension of social security as a functional necessity (Collier and Messick, 1975). Government authorities sought to ally themselves with the work force, the spearhead of modernisation. Social protection was primarily associated with the universal and corporate contributory social insurance programmes inherent to modernisation. Such programmes currently cover approximately 40 per cent of the working population in wealthier nations such as Argentina and Brazil, but less than 10 per cent in sub-Saharan Africa. Those working in the rural or informal sectors are not included, and remain under the protective wing of traditional solidarities (Box 1).
Box 1 – Social Protection in Africa Prior to the Introduction of Adjustment Plans
Despite many commonalities, a range of different social policy trajectories are evident among African states (Bailey, 2004). The countries of North Africa, for instance, have experienced the broadest extension of Bismarckian social insurance – relating to protections against old age, sickness and employment injury, as well as the provision of family allowance – initially under the colonial regime and continuing post-independence. Countries with commodity- and, particularly, energy-based economies have developed fairly generous social protection systems aimed at creating a clientelistic relationship between the authorities and workers. By exploiting extractive resources, these states have been able to provide basic public services (health, assistance, etc.), generally free of charge. Social insurance is also relatively advanced, arising strictly from the integration of workers into an atypically large public sector. Public employees thus benefit from free – or partially free – medical care, employment-injury, old-age and family benefits, whereas private-sector employees do not. The sustainability of these systems, however, is strongly dependent on the growth of their primarily rent-based economies.
In French sub-Saharan Africa, colonial France introduced the region’s first social insurance programmes, initially covering work accidents and later extended to include maternity and family allowances. These programmes were maintained by the post-independence governments, which expanded mandatory social security to provide old-age benefits for modern-sector employees between 1960 and 1965. As was common in other jurisdictions, such benefits were set according to salary and the period of time during which contributions were made.
In former British colonies, employment-injury benefits were prioritized over other forms of social protection. Social insurance was less sophisticated and, when available, was paid out in a basic lump sum. Family benefits, such as those developed in the former French colonies, were not implemented states previously colonised by the British. Government officers were generally the only recipients of social insurance. National health services and community social policies, however, were more extensive. Some countries created national pension systems based on savings (provident funds) (Charlton and McKinnon, 2001). Somewhat similarly, former Portuguese colonies had only rudimentary forms of social insurance.
Without question, despite the large gap in successful implementation and proportion of the population served between African states and the Global North, social protection policies in former French or British colonies nevertheless constitute two distinct systems that strongly reflected their origins. These policies, which covered only a very small portion of the African population (less than 10 per cent), make sense only when considered in light of the then-dominant assumption that the African continent was in the process of a gradual industrialisation-salarisation that would eventually incorporate the majority of the workforce. In the decades following the Second World War, African political elites largely conformed to this perspective. Moreover, just as in Latin America, government authorities tried to align themselves with wage earners. With respect to those not employed in the public sector, as members of peasant societies, the majority of the population continued to operate according to traditional solidarities (Vuarin, 2000) and to receive meagre government health benefits.
7At this time, then, the extension of social security beyond the formal sector was completely absent form the international agenda. The notion of social protection in the absence of financial contribution, moreover, was practically taboo, as the concept of assistance had been subject to puritan-inspired understandings of productivity virtually since the inception of development policies. At the micro level, the adage “give a man a fish and you feed him for a day; teach a man to fish and you feed him for a lifetime” clashes with assistance policies designed for the poor. At the macro level, similarly, was characterized by a consensus that highlighted the imperative of economic growth which, it was believed, would eventually produce trickle-down improvements in the population’s standard of living as they gradually entered the modern sector.
8This two-tiered system, comprised of a protected formal sector and an unprotected informal sector, collapsed during the 1980s under the effects of adjustment plans, economic and social crises and the rise of neoliberal thought. According to the tenets of neoliberalism, social security was no longer to be considered a reasonable objective, but an obstacle to the development of market economy. The ensuing division was dramatic, as social security policies came under fire for unfairly favouring salaried workers, to the detriment of the poor in the informal sector. Not unrelatedly, the 1980s also marked the relative marginalisation of the UN and its institutions in Geneva, including the ILO.
9Shortly after their initial implementation, however, financial organizations were compelled to add a social dimension to the purely economic aspects of adjustment. A report by the United Nations Children’s Fund (UNICEF) underscored the disastrous social effects of structural adjustment programmes and called for “adjustment with a human face” (Cornia et al., 1987). During this period, short-term safety net programmes, reserved for the poorest of the poor and primarily intended to absorb some of the shock of the economic crisis following adjustment, became key social protection programmes (Mkandawire, 2004).
10The 1990 World Development Report (World Bank, 1990) legitimised social safety net programmes, which were intended to protect people against two forms of economic adversity: the chronic inability to work/earn an income and/or a reduction of this ability during times of economic, political or environmental upheaval. Generally speaking, safety nets are based on monetary transfers or the provision of food supplies (Gentilini, 2005). When possible, such initiatives endeavour to avoid the “trap of dependency” by combining safety nets with recovery programmes (such as public works programmes). In the 1990s, the development community often used the term “social safety net” to describe the concept of social protection. The World Bank, in the name of fiscal and institutional realism, supported this minimalist and pragmatic approach to protection (Devereux and Sabates‑Wheeler, 2007).
11The latter half of the 1990s, against a backdrop of economic and social crisis, witnessed a more dramatic change of direction. The Asian crisis of 1997 revealed the urgency of finding new means of protecting populations from adverse events. The 1995 World Summit for Social Development, organised by the UN in Copenhagen, followed by the various processes leading to the adoption of poverty reduction strategies by the World Bank and the International Monetary Fund (IMF) in 1999 and the international community’s adoption of the UN’s MDGs the next year, created a new agenda for action. The concerted effort to fight poverty lent new impetus to social protection.
12With the adoption of the MDGs in 2000, social protection was no longer sidelined as a minor debate within the international development community. The need to extend social protection was a newfound matter of consensus (sometimes referred to as the “post-Washington consensus”), whose apparent accord obscured differences of opinion about what the extension of social protection would mean in practice.
13While the World Bank and the DfID were the major contributors to the conceptualisation of social protection, they were by no means alone. They were joined by the Economic and Social Council (ECOSOC) of the United Nations, UNICEF, the UNDP, the Organisation for Economic Co‑operation and Development (OECD) – with its Development Assistance Committee (DAC) and its specialised arm, the Poverty Reduction Network (POVNET) – the ILO and the International Social Security Association (ISSA), along with a host of international foundations and organisations, such as Oxfam and HelpAge International, as well as bilateral development organisations, including the German agency Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). Lastly, rising powers like Brazil and South Africa were also becoming increasingly active in the field of social protection. The general tenor of discussion centred more on peaceful accommodation than confrontation. That said, two general approaches may nonetheless be identified (Molyneux, 2008; Devereux and Sabates-Wheeler, 2007): one expressed in primarily terms of efficiency (orthodox or instrumental), championed by the World Bank, and another promoted by DfID, centred on social rights and advocacy.
14According to the first approach, promoted by what may be referred to as the instrumentalists, inequality, vulnerability and the risks inherent to extreme poverty impede the achievement of the MDGs. On the basis of this observation, the World Bank created a Social Protection and Labor division, which reformulated the concept of social policy as “social risk management” (SRM) (Holzmann and Jørgensen, 2000). This new concept provided an expanded view of social protection, taking it in a new direction. The essence of SRM lies in reducing extreme poverty through better risk management. Significantly, the concept of risk is defined inclusively, in so far as it encompasses social, economic, political and environmental risks, including labour market-related risks, such as unemployment and non-employment (Holzmann and Jørgensen, 2000). The SRM program also targets those individuals who are “most vulnerable”. As stated by Robert Holzmann and Steen Jørgensen (2000, 1), not only are the chronic poor the “most vulnerable” to risk, but they also, as a group, typically lack appropriate risk management instruments. The World Bank thus emphasized “the double role of risk management instruments – protecting basic livelihood as well as promoting risk taking” (Holzmann and Jørgensen, 2000). The SRM strategy is based on three key pillars: 1) risk reduction through labour market management; 2) risk mitigation through diversification of resources or the introduction of community and/or informal social protection mechanisms; and 3) risk coping, or the bolstering capacity to respond to risk effectively, through the introduction of specific measures such as social transfers or public projects.
15Although highly influential on the world scene, the model proposed by the World Bank came under heavy criticism (McKinnon, 2004; Kabeer, 2004) for its failure to consider inequality and structural poverty. In contrast to this instrumental concept of social protection based on the principles of economism, a number of organisations, led by the DfID and the Institute of Development Studies (IDS), advocate the adoption of a social-rights centred approach. In this model, the emphasis is placed on 1) the universality of human rights; 2) their social dimension; and 3) the development of the means to improve the fulfillment of these universal social rights at the international level (DfID, 2000). The “social rights for the poor” school of thought, correspondingly, denounces extreme poverty and inequality as symptoms of social injustice and structural inequality. Generally, its supporters seek to institute a comprehensive approach incorporating social protection into development policy. The DfID and the IDS exercise significant influence over both relevant international organisations, such as the World Bank, and the international community as a whole, through the production of publications and involvement in various think-tank networks throughout academia and the development community. The “transformative social protection” framework championed by the IDS is a paradigmatic example of this approach. According to Stephen Devereux and Rachel Sabates-Wheeler (2004), social protection policies must extend beyond the economic aspects of risk and vulnerability to include legislative (i.e. the creation of status and rights) and institutional measures. This approach includes four broad categories of social protection measures: protective, preventive, transformative and incentive. Social protection itself must be considered a fundamental right of individuals. Work by the OECD’s DAC, informed by states of the global North, takes a similar approach, wherein social protection is considered to be “pro-poor” programme expansion, while simultaneously constituting a fundamental right (Voipio, 2007). For its part, the ILO follows similar lines, advocating for recognition of the universal right to social security and corresponding creation of a social protection floor (ILO and WHO, 2009; Ginneken, 2007).
16Finally, in the 2000s, a compromise was reached between the formerly incompatible positions of “social risk management” and “transformative social protection”. Social protection is no longer seen as an expenditure, but as a means of strengthening social and human capital. Accordingly, it quickly became a key element of development policies.
17The compromise between the supporters of an orthodox economic approach and those in favour of an advocacy-based approach to social security was facilitated by the international promotion of innovative, model social programmes (Leisering, 2009). The newfound favour enjoyed by social protection within the development community therefore owes a great deal to large-scale programmes instituted in Southern countries, such as conditional cash transfers (CCT) and social pensions.
18CCT programmes, introduced in Mexico in 1997 and Brazil in 1996 and 2003, represent a new form of social protection programme, typified by the narrow delineation of its target group – the poor or the poorest of the poor. Such programmes use a range of methods, such as demand, geographic targeting, means testing, proxy-means testing and community-based selection (see Box 2). In doing so, they combine financial assistance for poor families with a series of initiatives designed to strengthen their human capital. To receive benefits, recipient households must make quid pro quo commitments in the areas of health, education and nutrition.
Box 2 – Targeting of Social Transfer Recipients
CCT programs are characterised by the specific targeting of recipient populations. By definition, targeting is intended to achieve greater efficiency in the fight against poverty by focusing efforts on poor households rather than the population at large (universality). In developing countries, however, identifying recipients is particularly difficult, due to the increased risk of two types of error: inclusion error, the incorporation of households that should not be eligible for the program or policy and, conversely, exclusion error, which falsely deprives recipients of their rights. States have used a variety of methods to effectively target recipient populations. Such methods can be divided into two main categories: 1) those that examine the validity of applicant requests based on existing data, which may then be supplemented by questionnaires and/or interviews; and 2) those that endeavour to classify individuals and households as poor or not poor.
The first approach involves asking individuals and households to apply to local offices. Brazilian Bolsa Família allowance applications, for example, are made at social assistance centres. Applications are then reviewed and sorted by administrative authorities on the basis of the data collected and data derived from other sources.
The second approach includes a series of varied methodologies. The most theoretically sound consist of conducting questionnaire-based surveys of candidate populations to select households or individuals on the basis of their income or consumption level, a form of means testing. In poor countries, however, it can be extremely difficult and prohibitively expensive to collect and accurately measure means-related data (e.g. income, transfers, donations and assistance). For that reason, many experts have recommended another method, that of proxy-means testing. This form of targeting involves the development of a numeric score derived from the addition and weighting of a limited number of variables related to household living conditions. These variables generally include housing quality and occupancy status; ownership of durable goods; household demographic structure; and the status, sector of activity or level of education of household members (Grosh, 1994). Proxy-means testing is strongly recommended by World Bank experts. However, states in the global South tend to favour one measure over the other, in accordance with their administrative traditions. For example, Brazil uses the first method (application by recipients), and Mexico the second.
In any case, both traditional means testing and proxy-means testing have one major drawback: they require administrative capacities that are simply absent in many poor countries. These methods are also very expensive, and may lead to the stigmatisation of recipients. Instead of targeting individuals or households on the basis of income per unit, it may be simpler and more effective to target geographic areas where poverty is most heavily concentrated. It has been shown, for instance, that in many poor nations, place of residence, rather than other household traits, is often the most reliable indicator of poverty (Ravallion and Wodon, 1997). This observation has led many policy-makers and poverty reduction programme administrators to select recipients according to their geographic area of residence, while distinguishing poorer from more affluent regions. The effectiveness of geographic targeting, however, is highly dependent upon the concentration of poor persons in the areas in question. If the population is fairly heterogeneous in terms of poverty, as in Malawi or Mozambique, this form of targeting generates both inclusion errors (areas identified as poor but containing a large proportion of non-poor persons) and exclusion errors (poor persons living in areas identified as non-poor) (Bigman and Fofack, 2000; Ellis and Marchetta, 2009).
Alternatively, recipient targeting may also be aimed at particularly vulnerable segments of the population, defined on the basis of gender (women), age (children and the elderly) or ethnic background (minorities). The ILO, UNICEF and large NGOs, such as HelpAge International, promote this method. Although it contains the drawback of making transfers available to individuals or households that may not actually be poor, the selection of recipients on the basis of vulnerability nevertheless offers significant advantages. Targeting by category is less costly and complex to implement than targeting systems based on strict income-related criteria, reduces stigmatisation, and enjoys greater public support. It also produces positive externalities (Duflo, 2000).
For two decades, the international financial organisations and agencies of the global North have favoured targeting the poorest of the poor. In the interim, however, many studies have highlighted the extreme implementation challenges and perverse effects of the insistence on highly targeted programmes in the poorest nations (Ulriksen, 2012). A growing number of experts and organisations now favour broader programmes based on simple category-related criteria, such as universal child benefits or non-contributory social pensions for the elderly (Hanlon et al., 2010).
19The general philosophy underlying the provision of cash transfers conditional on recipients' participation in human capital strengthening programmes represents a departure from the former safety net programmes, often issued in kind (i.e. food aid or subsidies for basic products) and relatively free of conditionality (Rawlings, 2005). In short, CCTs simultaneously aim to relieve immediate poverty while also preparing for the future by breaking the intergenerational cycle of poverty. Since the early 2000s, CCT programmes, initially disseminated widely in Latin America on the basis of initiatives developed in Mexico in 1997 (Box 3), Colombia in 2001 and Brazil in 1996 and 2003, have become model programmes for anti-poverty policy (Adato, Hoddinott, 2007; Lautier, 2006).
Box 3 – The Progresa CCT Programme in Mexico
The Progresa programme is considered the original CCT model. It was established in 1997, during a time of severe economic crisis, as a substitute for agricultural products subsidies (such as the tortilla subsidy) considered to be ineffective and poorly targeted. Progresa was primarily intended for all poor households from rural areas. The selected households receive a benefit subject to their compliance with certain conditions (quid pro quo) related to education (e.g. regular school attendance), nutrition and health. The benefit, paid twice a month, has two components: a lump sum to cover food and a school allowance based on the gender and education level of the children. It is paid to mothers, on the assumption that they are most capable of managing the funds in an altruistic manner that prioritises the interests of children. To avoid inadvertently increasing fertility rates by incentivising large family units, the payout is subject to a ceiling and only children aged seven years and older are eligible.
Recipient targeting is based on household means testing through the use of three complementary methods. First, population census data is used to identify regions containing high concentrations of vulnerable individuals. To that end, the Mexican Department of Social Development (Secretaría de Desarrollo Social) developed a proxy-means test based on the following criteria: illiteracy rates, access to water, access to electricity and the percentage of the population working in the primary sector. This process permitted the identification target regions and verification of the existence of local schools and health centres. Second, the Department of Social Development conducted a detailed survey in order to determine which households subsisted under the benchmark poverty line, assigning each household a numeric score Third, meetings were organised to present the list of recipients to the communities, thus allowing for community consultation and debate and the possibility of reaching a subsequent consensus. Lastly, programme evaluation was entrusted to the International Food Policy Research Institute (IFPRI), which determined the programme to be extremely effective as early as 2000. By the end of 1999, Progresa benefited 2.6 million households, or 40 per cent of rural families and 11 per cent of total families. Despite policy changes in 2000, the programme continues to operate, albeit under a different name: Oportunidades. Buoyed by the positive evaluations and support from international organisations (i.e. the World Bank and the Inter-American Development Bank [IDB]), Progresa has been emulated across Latin America and around the world (including Turkey, Indonesia and the Philippines).
- 1 As observed by Nancy Birdsall, President of the Center for Global Development, “I think these prog (...)
20With the support of the World Bank, the UNDP, the ILO and many international and bilateral organisations (e.g. DfID, the European Union, etc.), initiatives based on Mexico’s ground-breaking Progresa programme, and similar programmers in Colombia (Familias en Acción) and Brazil (Bolsa Família), have been or are in the process of being implemented as pilot projects (McCord, 2009), in a growing number of states in virtually every region of the world: Latin America, the periphery of Europe (Turkey), North Africa (Tunisia), South-East Asia (Indonesia, the Philippines and Pakistan) (Lopes-Wohnlich et al., 2011) and sub-Saharan Africa (Ethiopia, Kenya and Mozambique). The dissemination of CCT programmes is justified primarily on the strength of their effectiveness in fighting poverty,1 as confirmed by the extremely positive evaluations arising from the Mexican (Skoufias and McClafferty, 2001; Rawlings and Rubio, 2003), Colombian and Brazilian experiences (Castañeda and Lindert, 2005). Additionally, as CCTs combine monetary benefits with conditionalities and targetings, they may be interpreted in both economic (creation of human capital) and social rights terms, thus appealing to orthodox economists and social advocates alike.
21Today, social pensions, which entail the payment of old-age benefits at a standard (non-contributory) rate to all elderly persons below a given income level, are extremely popular. The principle of ‘universal’ pensions, however, was long rejected for reasons related to cost, administrative difficulty and potential negative effects (Willmore, 2006). Prior to 2000, only Brazil, which provided rural pensions, and South Africa, of all developing states, offered this type of programme. In the 2000s, however, social pensions, in light of new considerations, were re-evaluated in a positive light. It was recognized, for instance, that due various crises affecting traditional solidarities, developing states were witnessing a sharp increase in poverty among the elderly. Similarly, analysts noted that in the countries most severely affected by the HIV-AIDS pandemic, particularly those in East and Southern Africa, the elderly are often responsible for orphaned dependents (Kakwani and Subbarao, 2005). This latter consideration provided an ethical rationale for these programmes.
22It is now commonly accepted that social pension programmes have a highly positive impact on poverty reduction and the maintenance of social solidarity. Studies show that under certain conditions, such as the existence of a minimal level of administrative and financial capacity and corresponding strength of political will, the creation of a universal old-age pension is an appropriate option for poor nations. Since the early 2000s, a number of countries in Southern Africa (Botswana, Lesotho and Swaziland), Latin America (Bolivia) and South Asia (India, Bangladesh and Nepal) have introduced programmes based on the South African model. Similarly, Argentina, Brazil, Chile and South Africa, which already had non-contributory pension systems, have strengthened them. With a pragmatic consensus emerging on the legitimacy and efficacy of these programmes, their global dissemination is receiving strong support from a coalition of international organisations (e.g. the UNDP’s International Policy Centre for Inclusive Growth [IPC-IG] and the World Bank), NGOs (e.g. HelpAge International and Save the Children) and bilateral agencies from the U.K., Germany, Brazil and South Africa.
- 2 With regard to this popularity, please refer to the article by Margaret Grosh, lead economist and (...)
23Over the course of just two decades, social protection discourse within development policy has changed radically. From an attitude of outright rejection stemming from moral economic objections in the global South, there has been a shift toward a ‘residual’, minimalist perspective wherein the safety net is regarded as an acceptable last resort. Since the early 2000s, social protection has come to be appreciated for its essential virtues, and the scope of such programmes has expanded (Grosh, 2012).2 This change is all the more remarkable considering that it has occurred in a context wherein players’ positions evolve not only according to their own changing interests and frames of reference, but also in light of the complex interactions between various organisations that are increasingly involved in the debate. The institutional ‘public policy forum’ (Jobert and Muller, 1987) on social protection includes international financial organisations like the World Bank, the IMF and regional development banks (primarily the Inter-American Development Bank [IDB] and the Asian Development Bank [ADB]), specialised UN agencies (particularly the ILO, UNICEF, the WHO and the UNDP), national development agencies (including the DfID, GIZ and the European Union), a number of Southern countries, such as South Africa, Brazil, India and Mexico, and NGOs, such as Save the Children and HelpAge International.
24Unfortunately, the development of social protection policies in the global South also raises many problems, particularly with reference to buy-in, design and ultimate purpose. Whilst emerging or middle-income countries may have adopted this type of programme without difficulty (Box 4)when they were not their own very impetus for reasons related to heightened effectiveness, electoral concerns or a desire to establish social peace. In ‘poor countries’, mostly located in sub-Saharan Africa (Box 5), similar programmes are not meeting with the same success, thereby highlighting the difficulties inherent to generating buy-in to new ideas.
Box 4 – New Forms of Social Protection in China
China’s first social protection system dates back to 1951, two years after the creation of the People’s Republic of China. As in other planned economies, a large number of social benefits gradually came to focus on companies or, as the Chinese refer to them, ‘work units’ (danwei). At the time, this network of ‘public provident companies’ and ‘people’s communes’, which provided various benefits supplied in the West by the welfare state, constituted the entirety of the Chinese social system (Merrien et al., 2005).
The economic liberalisation of the 1980s led to a gradual erosion in social protection. Company pensions and health services were gradually eliminated, while in the countryside the privatisation of land dealt a fatal blow to local social protection systems. Various categories of salaried, public and private-sector workers benefited from a range of social insurance systems and relatively extensive social coverage, but the situation of those rural and urban residents excluded from the formal social protection system quickly became critical. It thus soon became apparent to the Chinese government that the transformation of wage labour and rural economies had resulted in the emergence of new social problems with the potential to challenge the regime’s power and the state’s social cohesion.
The immediate response to these changing social conditions was to charge the people’s communes, with rebuilding a social protection system capable of delivering old-age benefits, health care and social assistance. It was the insistence of the central government, which was more sensitive to social pressure than local entities, however, that was behind the development of these new social protection programmes.
The first ‘minimum livelihood guarantee programme’ (dibao) for urban residents appeared against the backdrop of the social movements of the 1990s. This basic benefit, reserved for those who held a city work permit(hukou), was available to families living below the poverty line, poor persons unable to work and those whose unemployment benefits had run out or had fallen victim to industrial restructuring. The number of recipients increased quickly, from 0.85 million in 1996 to 2.66 million in 1999 and 23.3 million in 2008. Interestingly, this curve closely mirrors that of social discontent (Cho, 2010). Perhaps unsurprisingly, the rural dibao programme follows a similar curve, covering 66 million people in 2008. Finally, it should be noted that two health coverage systems were established in 2003 and 2007: the first for rural and the second for urban residents. In 2009 and 2011, China established minimum old-age benefit systems.
Since the 2008 crisis, the issue of social protection has occupied a central position on the agenda of the government and the Communist Party (Zhu, 2009). Social protection is no longer understood from a purely security-oriented standpoint, and is now taking a Keynesian turn. Like their international counterparts, Chinese experts are realising that social protection also acts as an effective buffer in times of crisis. In 2009, the report ‘Building a Social Welfare System Shared by All the People’ (Wei, 2010) discussed the increase in wages required to support consumer consumption, and the role that social protection can play in stimulating domestic demand. As a result, the development of universal social protection is now a national objective.
25Most programmes are initiated as pilot projects, at the instigation of multilateral or bilateral organisations (Chisinga, 2007). When international funding ceases, the experiment ends, indicating that social protection is not a priority of the governments of poorer states. Significantly, this apparent disinterest may also point to a lack of relevance of certain programmes – including CCTs – for some poor nations. This suggests that learning from foreign experiences, such as those of Latin America, requires knowing how to draw genuine lessons from these experiences, distinguishing between importable and non‑importable elements, and understanding the diversity of countries' economic, social, geographic and cultural circumstances.
Box 5 – The Dissemination of CCT Programmes in Sub-Saharan Africa
For the past several years, international organisations, experts and a growing number of donor countries and international development agencies have championed the creation of CCT programmes in sub-Saharan Africa, based on the Latin American model. As poverty remains endemic to the region, the reasons behind this push are readily apparent. CCT programmes, combined with a strong command of the technology required for their implementation, seem to produce positive and quantifiable results.
The social protection movement in Africa is clearly driven by agencies in the North.(Hickey et al., 2009). A plethora of pilot projects have been established in Eastern and Southern Africa and, less often, Western and Central Africa. In 2005, the United Nations Economic Commission for Africa recommended cash transfers as a means of fighting poverty; the following year, the African Union made plans to move forward in this direction.
The programmes implemented in Ethiopia, Kenya, Malawi and Zambia are often cited as examples. Apart from the Ethiopian Productive Safety Net Programme, which is subject to recipients’ participation in public projects, however, few programmes can be considered a success. Most occur at the impetus of donors and are limited in scope. They address only limited segments of the population and are heavily dependent on international financing and pressure.
Anna McCord (2009) cites six factors that impede the adoption of cash transfer programmes by African governments: 1) the view that CCTs are too costly to be funded within severely constrained domestic budgets; 2) the perception that these programmes are social expenditures made at the expense of productive investments; 3) the perceived reinforcement of a culture of dependence among recipients; 4) the low level of political capital to be gained from such programmes for elites in power; 5) a lack of alignment with donor preferences; and 6) difficulty in implementing targeting methods.
26Accordingly, while the dissemination of conditional programmes in Latin America has attracted special attention in sub-Saharan Africa, their application to the region remains problematic (Schubert and Slater, 2006). Conditionality operates as an effective policy instrument only if the problem truly derives from the demand side of social services. If there are too few schools or health centres, if teachers and health care personnel are scarce, or if school and/or health equipment is insufficient in number and quality, a conditional programme may not only be of limited interest, it may even result in additional problems. In a number of low-income countries, including in those of sub-Saharan Africa, issues related to improving human capital seem more closely related to insufficient supply than the stimulation of demand. Moreover, in countries characterised by weak administrative and financial capacity, the problems and management costs inherent to administering conditional programmes may prove prohibitive. Lastly, overly-conditional and too-finely targeted programmes have been demonstrated to be less effective in fighting poverty than other benefit programmes of broader scope (Hanlon et al., 2010; Ulriksen, 2012).
27Since the late 1990s and the early 2000s, social protection has gained legitimacy in international development policy circles. The contemporary consensus reflects the ILO's move to protect populations outside the salaried workforce, the World Bank’s step back from its once-exclusively neoliberal outlook, and interventions by development organisations such as the DfID and the UNDP in support of a so-called ‘third way’, wherein social protection is considered both a basic right and a social investment. That said, while these changes point to an obvious alignment of positions, a true international consensus remains elusive. Strong differences persist regarding matters such as the relative importance of the public and private sectors, the efficacy of targeting vs. universality, and the merits of conditional vs. unconditional access to assistance.
28It is also valid to question the reasonableness of the exceptionally positive expectations surrounding the potential scope of social protection programmes. The concept, after all, won general acceptance only after the medium- and long-term efficacy of these policies was emphasised. While assessments demonstrate incontrovertible short-term successes (school attendance, health of the monitored populations), there is little evidence of long-term effects. The higher employment rates and better qualifications predicted among children who participated in CCT programmes remain largely within the realm of political conjecture (Valencia Lomelí, 2008). Social protection remains a component of the fight against poverty, but it constitutes neither an employment nor a growth policy. Rising powers characterised by an increase in social inequality, such as Brazil or China, are understandably extremely sensitive to the need for social protection floors. Developing countries, however, are more interested in the implementation of growth strategies or programmes with a more attractive electoral payoff (Chisinga, 2007; Hickey et al., 2009).
29In any event, social protection programme funding is an issue that requires further examination. Some studies (ILO, 2006; 2008) estimate that the cost of social protection amounts to a small percentage (from 0.5 to 2.7 per cent) of poor states’ GDP. Nevertheless, in countries where international aid represents 35 to 60 per cent of the budget, even this small proportion of GDP is quite considerable, and the cost of social programmes are shouldered primarily by donors. The broader dissemination of social protection will not be possible unless the international community accepts its obligation to provide the long-term funding for a global universal social policy. With this aim in mind, social protection would then constitute a key component of the global social safety net envisaged by, among others, Jean-Michel Severino and Jean-Michel Debrat (2010).