7 | SOCIAL POLICY AND THE TENSION BETWEEN IDENTIFICATION AND SEGREGATION WITHIN SOCIAL ORDERING AND DEVELOPMENT
The dilemma of all approaches to the study of poverty, including social exclusion in terms of how it is conventionally conceived, can be summed up as a tension between identification and segregation. Poverty approaches derive from a fundamental concern with the wellbeing of the poor and are designed to evaluate their situation and/or to provide information for guiding actual policy-making, which boils down to an exercise of identification. However, the institutionalisation of this exercise tends to lead to a vision and practice of segregating the poor, both in terms of how their place in society is conceived, and in terms of how policy and social provisioning is directed towards them. Even if the intentions are otherwise, the inertia of the logic consistently tugs in this latter direction.
This tension is usually framed in terms of universalism versus targeting. The debate essentially refers to the challenges of treating people the same while at the same time differentiating them in order to redistribute income and wealth across society, and to address disadvantage and discrimination. This classic debate, however, often tends to get somewhat convoluted and easily dismissed, if only because some targeting is usually required in most policy-making, especially in areas of poverty reduction and social assistance.
There has also been a subtle shift in the way that universalism has been referred to in recent years, towards a connotation of universal coverage or access, such as sending all children to school, regardless of how such schooling is provided or financed, or else everyone having some sort of health insurance, regardless of whether the insurance offers complete coverage or whether health care provisioning is unified or segregated. Martinez and Sanchez-Ancochea (2016) refer to this as a minimalist understanding of universalism and give examples of its prominence, particularly among various organs of the UN system, such as the ILO, that are in part guilty for the normalisation of this minimalist meaning.
However, the strongest influence on this shift of meaning probably comes from the World Bank, which explicitly takes the position that universalism is achieved as long as everyone has access to something, regardless of how this is provided. In other words, even if the middle and upper classes access health and education services privately, this is considered to be part of a ‘universal’ system.1 This interpretation is also implicit in the special issue of The Economist (2018) on universal health care. It thereby includes the vast swath of private provisioning in health and education to the unsecured middle 50–60 per cent of populations in most lower- and middle-income countries as part of a ‘universal’ provisioning system. Given that almost all systems, even in the poorest countries, usually have at least some very rudimentary public health and schooling systems for the poorest, alongside private provisioning for the rich, this makes it very easy to pre-maturely celebrate the achievement of universal health and education for all, as has become a common platitude.2 Such sleights of narrative are impressive in the way that they coopt or hijack progressive concepts. Indeed, this is similar to the way that ‘pro-poor growth’ was defined by the World Bank in the early 2000s as growth with any absolute poverty reduction, which effectively includes most growth episodes as pro-poor.
In the midst of this confusion, a certain reframing is necessary. The fundamental problem is not necessarily with targeting per se, but with the increasing normalisation of a segregationist approach to policy-making and social organisation, whereby poor people are not necessarily un-serviced but are treated differently, and often quite poorly. Much of the criticism of targeting in social policy scholarship is implicitly about this aspect of segregation. It is discussed in terms of how targeting severs cross-class solidarities and interests, thereby reducing political support for maintaining service quality to the poor and often resulting in a worsening quality of social provisioning to the poor, and also how it encourages stigmatism of welfare recipients or else perverse incentives. Indeed, even though the argument of perverse incentives has typically been made from the right as a criticism of universal welfare, the irony is that its logic actually applies to targeting – the least universalistic aspects of welfare systems – given that the perversion is the result of the use of thresholds and means-testing, not welfare per se. Nonetheless, the challenge remains with regard to how to treat poor people the same as anyone else, without stigmatisation or segregation, while at the same time differentially directing resources towards them in a way that ideally generates a degree of redistribution from rich to poor and otherwise improves their prospects in society. The dilemma encapsulates classical concerns of both liberalism and socialism.
The problem with the de-politicisation of poverty debates, as elaborated in Chapter 2, is that it allows for orthodox agendas to frame the narrative about the optimal ways of solving this dilemma, and about the possibilities and constraints of public action. This has been evident in both the MDGs and the SDGs, neither of which address the policy means that should or could be used to achieve their proliferating goals and indicators. While some might argue that these policy means should not be specified, the void nonetheless gets avidly filled by the most powerful and assertive voices.
Since the beginning of the so-called neoliberal era, from the early 1980s onwards, these orthodox voices have strongly adhered to the idea of targeting in social policy as an overall institutional modality of social provisioning, rather than as just a particular policy tool within a larger profile of more comprehensive programmes and interventions. As pointed out by Mkandawire (2005), this ironically contradicts their opposition to targeting in trade or production policy. The emphasis in turn has tended to reinforce fragmentation and segregation in social provisioning systems, in many cases already existing but in others created anew. The irony is that the ‘liberal’ in neoliberalism would, in principle, be against creating systems of differentiated citizenship rights, but it appears that this is where neoliberalism shows its reactionary bias. Of course, care must again be taken with the twists of word play, because the World Bank also puts much emphasis on overcoming fragmentation, although by this they do not refer to social fragmentation across classes and different social groups, but sectoral fragmentation, such as with respect to health, education and social assistance services that are targeted to the poor in a fragmented rather than coordinated manner.3
A good example of this mainstream position is found in the various World Development Reports (WDR) of the World Bank. I have discussed the 2003 WDR on services in Fischer (2010a, 2013a) and the 2013 WDR on jobs in Fischer (2012) alongside The Economist’s position on inequality, as discussed in Chapter 2. With regard to the latter WDR (World Bank 2012), besides the regular advocacy of macroeconomic ‘fundamentals’, trade liberalisation, deregulation and the virtues of special economic zones and foreign takeovers of domestic firms, the position of the WDR on matters related to social policy or poverty alleviation remain remarkably in line with the targeting approach promoted by the World Bank for decades. This typically enters as a corollary to labour market deregulation and social security reforms, with social protection to support the most vulnerable and least able to adjust to the dislocations wrought by these policies (e.g., see WB 2012, p. 257 and pp. 307–309). Indeed, a similar view is expressed in the recent IMF Framework on Social Spending (IMF 2017), which created widespread reaction among the academic and policy communities, specifically in relation to the priority given to fiscal concerns over the rights to social security in the IMF suggestions for reforms of the latter.4
The absences in the 2013 WDR are also telling, such as on matters related to industrial policy or other forms of government intervention into production or employment.5 Similarly, the role of universal public provisioning of health care and education is absent from the discussion on human capital. The only brief discussions of ‘universalism’ address the role of social protection in inducing formalisation of labour forces, although the report advises caution in this regard (e.g., pp. 210–212 and 276). That both state-led industrial policy and universal provisioning of basic health care and education were keys to the employment and growth successes in countries such as South Korea and Taiwan is generally ignored (except one qualified mention of universal education in South Korea on p. 177).
The reasoning of these policy positions remains clearly orthodox. As discussed in Chapter 2, it is founded on the deductive principle that perfect markets – if ever these could exist – lead to the most efficient and pareto-optimal outcome possible. In this sense, there is nothing new under the Chicago sun. This logic has been at the core of the dominant policy paradigm that has arguably induced much of the rising inequality over the last 30 years. In their defence, proponents of this paradigm would retort that it has been the continued obstruction of markets by government policies or other collective actors such as trade unions that has been preventing the outcomes predicted by theory. This is the reason why government spending and policy are repeatedly emphasised in The Economist (2012) as the big drivers of inequalities today. To paraphrase Karl Polanyi’s criticism of the laissez-faire utopia of economic liberalism: because the ideal free market is never attainable as a reality in a complex modern economy, it can always be argued that the failures of economic liberalism are not because markets were liberalised, but that they were not liberalised enough.
From the perspective of the dilemma mentioned above, the important point here is that these orthodox positions do present themselves – and possibly honestly see themselves – as the progressive option. Despite decades of critique, targeting is still presented as a logical win-win for redistribution; transfers will have much more poverty-reducing effect if concentrated on the poor rather than being dispersed across whole populations and among many non-poor who do not need the transfers or should not have them. Indeed, it appears as a straightforward Robin Hood scenario of taking from the rich and giving to the poor, which is reinforced with the impression that universal-type policies also lead to state largesse as well as corruption, and therefore need to be tightly reigned in and managed. The position appears impervious to decades of critique of targeting, which at best have merely dented its armour in battles with greater demons from the neoliberal imaginary. The critique has also become entangled in confusions surrounding the meaning of universalism, as mentioned above, which facilitates the orthodox dismissal of this critique.
In this respect, there is a necessity to insert a holistic social policy perspective into these discussions, regarding the key role of universalistic modes of social policy in both rich and poor countries as some of our most powerful policy tools to date for dealing simultaneously with poverty and inequality, especially in combination with broader developmentalist agendas. This endeavour, which I already elaborated in Fischer (2010a and 2012), joins similar efforts made by Mkandawire (2001, 2005), as well as others such as Martinez and Sanchez-Ancochea (2014, 2016), and institutionally by the United Nations Research Institute for Social Development (e.g., UNRISD 2010).
The contention here is that shifts towards more universalistic principles in social policy (which may include elements of targeting) are crucial to bring about more egalitarian and equitable processes of social integration and citizenship. Of particular importance is the degree to which poorer and middle social strata,6 or various other social groupings, are integrated (or segregated), given the centrality of this dimension to the longer-term social and political sustainability of any poverty- and/or inequality-reduction strategy. In addition to this common insight from the field of social policy, another problem with the focus on poverty and even some inequality indicators is that these do not necessarily reflect a wide range of adverse social processes occurring across middle social strata and that can have important consequences on mobility, stratification and cohesion, as discussed at the end of Chapter 5 on social exclusion. Indeed, a lack of attention to middle strata can inadvertently reinforce a tendency to attribute various inequality-induced social disorders to poverty, thereby reinforcing conservative phobic impulses to segregate the poor from other social strata, even though the opposite is arguably required in order to move towards lasting socially inclusive development.
However, such issues are fundamentally political, not merely technical. Hence, they require a politicised engagement within current development agendas in order to create the space for serious deliberation of these possibilities, rather than relying on the apparently apolitical moral ground of goals and indicators. The risk of not explicitly anchoring future development agendas within politicised policy debates is that these agendas can be (and are often being) subverted towards policy agendas that possibly undermine inequality reduction and/or fragment citizenship rights in many contexts.
This is argued in three sections. The first presents a background to conceptions of social policy, particularly with a context of development and with respect to criticisms of targeting. The second offers an attempt to conceptualise universalism in social policy in a manner that bears more general and practical applicability beyond ideal-type Northern welfare states, and that clarifies much of the conceptual confusion surrounding the term, particularly with respect to its application to developing countries or development goals. Accordingly, universalism in social policy needs to be understood as an umbrella term reflecting a set of guiding principles along three dimensions: provisioning modalities, which includes issues of access and coverage; costing and pricing, which relate to commodification; and financing, which involves the principle of (social) insurance. Within each dimension, we can think in terms of a spectrum from strong to weak (or absent) universalistic principles underpinning various institutional systems of social provisioning. In the third section, this understanding is then brought back into the question of placing poverty and inequality within broader questions of social integration and social needs within development. The conclusion offers some reflections on the necessity of politicising the policy debates along these lines.
Social policy and social ordering in development
The idea of social policy lies at the core of our understanding of modern institutional processes of social ordering and inequality. Social policy generally refers to the range of publicly or collectively provided, funded and/or regulated forms of services and interventions in a society, such as schooling, health care and social protection. The function or purpose is not only to provide these public or social goods, but also to affect various social outcomes through such provisioning (such as learning or health) or else to influence the access to and the incidence of adequate and secure livelihoods and income (as per Mkandawire 2004, p. 1).7 It is also generally considered to include several other areas such as housing policy, as well as more legal or regulatory aspects such as in child protection or labour market regulation.
Within this broad understanding, social protection has received most of the recent attention as a more narrow view of social policy, whereas it is properly understood as a subset of social policy. It has been commonly divided according to three categories: social insurance; social assistance (i.e., welfare); and standards and regulations (formal and/or informal), such as labour standards and regulations, or child protection.8 However, health and education provisioning play a central role in both social policy and inequality dynamics. These two social services usually constitute the largest shares of government expenditure in most countries (education is usually the largest, more than three times health care spending in the case of China, for instance). They involve potentially large impacts on household expenditures (particularly if commoditised), for both poor and non-poor, however defined, and thereby have a large bearing on poverty (especially health). Education has huge implications for social mobility and the structuring of education systems goes to the heart of social stratification and the social reproduction of inequality. Both health and education systems touch a core nerve of social politics because they structure the ways that various social groups and classes might come into contact with each other in moments of intimacy and vulnerability.
From the political economy perspective of the previous chapter, social policy can be understood as playing an important redistributive (or circulative) role in an economy, particularly through education and health spending. It has served as the primary policy realm in which most direct public action on poverty reduction is implemented (poverty reduction via growth can be considered indirect, i.e., trickle-down via the employment and demand effects of growth). Social policy also plays an important role in regulating distributive outcomes via its effect on wages and other aspects of employment (such as the role of child care and early school provisioning on women’s labour force participation).9 In the productive realm, social policy and associated policy approaches (such as fiscal policy) also play synergistic roles with productionist development strategies such as industrial policy. Its redistributive function is important in this respect given the exclusionary tendencies of such development strategies (e.g., see Sumner 2017). However, it also plays a direct role in productionist strategies in terms of providing skilled labour inputs, and also a political economy role in terms of forging cross-class alliances in support of such strategies. In all of these senses, social policies are fundamentally political given that they serve as the basis for defining and instituting citizenship rights, distributing public goods, redistributing wealth and articulating some of the main mechanisms of integration and segregation within societies.
Indeed, Mkandawire (2005) highlights the paradox that, historically, poverty alleviation was most successful when it was not necessarily the primary focus of social policies, as opposed to other priorities such as late industrialisation, state consolidation, demand stabilisation, political cohesion or else sheer survival. The key to this paradox is found within the choice of provisioning modalities, broadly characterised as universalism versus targeting.
Trials of targeting
In terms of institutional modalities, the formative debates in the field of social policy have been between targeting and universalism. While targeting often plays a role within universalistic systems,10 targeting as an institutional modality refers more generally to the abandonment of the guiding principles of universalism and towards more selectivity in publicly funded provisioning and benefits (e.g., means-tested welfare). It is characteristic of the main approach to social policy under the dominant neoliberal economic policy paradigm from the 1980s onwards, in response to the void created by welfare state retrenchment in the North and fiscal crises in the South, and as epitomised by the targeted social safety net approach of the World Bank. According to this approach, publicly funded and/or provided services or benefits are to be targeted selectively to the (identified) needy, such as means-tested welfare or means-tested free education and health care. In practice, selectivity is usually implemented through differentiated provisioning systems, based on segmentation and even segregation between different social groups (typically, between the poor and the middle classes), such that state-subsidised services or assistance are provided separately from privately or publicly funded systems servicing middle and upper social strata. Indeed, segregation often constitutes much of the political appeal of targeting for middle- and upper-class people, who seek to obtain their own privileged access without needing to rub shoulders with lower classes (and often ethnicised classes – increasingly so in Europe and the US).
Conversely, a universalistic modality essentially (or ideally) implies that all are serviced through the same or similar provisioning systems (usually publicly funded). Martinez and Sanchez-Ancochea (2016, ch. 1), for instance, elaborate what they call a maximalist vision of universalism, as opposed to the minimalist visions described in the introduction of this chapter. These include three dimensions: coverage, generosity and equity. Coverage refers to ‘massive coverage’, meaning that most people in a given category are covered or have access (depending, of course, on how citizenship rights define who has the right to be covered or have access – Martinez and Sanchez-Ancochea suggest that non-national residents should be included even if this is difficult to sustain politically). Generosity refers to the level and quality of benefits, with universality being associated with more comprehensive and better-quality benefits. By equity, they refer to the distribution of coverage and generosity across beneficiaries (although perhaps this would be better termed as equality, especially since the 2006 WDR framed ‘equity’ in terms of equality of opportunity, whereas the implication here is more than that). Hence, countries could secure massive access with uneven generosity, as is typical in minimalist versions of universalism, whereas more universalistic systems would have more evenly distributed benefits.
The contention in the seminal work by Mkandawire (2004, 2005) on synthesising debates in social policy with parallel debates in the field of development studies is that universalistic approaches have been much more successful at poverty reduction (and inequality reduction) than targeted approaches. Targeting, he explains, has been variously advocated since the 1970s on grounds of efficiency, expediency and even equity. The equity argument is based on the logic that if the expenditure used to provide universal benefits is focused on the poor instead, the poor will receive more and the non-poor less, hence producing a more redistributive and equalising outcome than if all receive the same, regardless of need.11
Drawing from work by Korpi and Palme (1998) and Rothstein (2001), Mkandawire (2005, p. 6) points out that many of these arguments on the greater potential for inequality reduction in targeted versus universalistic policies are partial and misconceived. They tend to focus only on the partial transfer effect of social spending (holding all else constant) rather than also considering the inequality impact of how such spending is funded, and the dynamic political economy interactions between these two as shifts are made towards more targeted approaches. Hence, the fact that universalistic systems are generally funded by progressive taxation makes them particularly effective at reducing inequality, given that even perfect equity in transfers (i.e., when all people are provided with the same benefit) is more than compensated by redistribution in taxation. Indeed, as demonstrated by Rothstein (2001), even a flat-rate tax funding an equal transfer payment to all results in a substantial reduction in income inequality. The inequality reduction effect would be even greater with progressive taxation.
A poignant example of similar misconceptions can be demonstrated, for instance, in an article by Appleton, Song and Xia (2010) with respect to China. The authors inadvertently make a critique of targeted welfare, albeit only after their argument is correctly re-interpreted. They argue that government anti-poverty programmes in China had little impact on urban poverty between 1988 and 2002. Instead, through an econometric analysis of urban household survey data, they contend that urban poverty had fallen almost entirely due to overall economic growth rather than ‘redistribution.’ The misconstrued element in their argument is that they refer not so much to redistribution but, more specifically, to targeting, given that China’s urban anti-poverty programmes were heavily oriented towards means-tested targeting within an overall retreat from more universalistic principles over the 1980s and 1990s, including the rapid erosion of most pre-existing redistributive and/or social security systems, and a notably regressive shift in the burden of taxation (e.g., see Khan and Riskin 2001). In this context, it can hardly be said that targeted poverty-reduction programmes constituted a strong case for redistribution. Rather, targeted social assistance probably represented one of the few marginal factors compensating an overall regressive shift in the social policy regime of China, as well as rising inequality. Hence, it comes as no surprise that most poverty reduction could be shown through econometrics to have come from growth, although this can hardly be used as a case against redistributive polices. With this corrective in mind, the argument of these authors otherwise corroborates well with the insights from the social policy literature discussed here.
Obviously, according to this logic, the redistributive potential for targeting could be strong if integrated within a broadly universalistic social policy regime, including progressive taxation. Indeed, this is the logic behind the idea of ‘targeting within universalism’ (Skocpol 1991), both in terms of its redistributive potential as well as its ability to address diversity, disadvantage and special needs without eroding broader universalistic principles. However, as Mkandawire (2005, p. 7) points out, targeting and selectivity without this universalistic basis usually undermines political support for both progressivity in taxation and for maintaining the size of transfers directed towards the poor. Hence, according to the ‘paradox of redistribution’ (Korpi and Palme 1998, p. 681), the more benefits are targeted to the poor rather than being equally distributed to all, the less likely poverty and inequality will be reduced.
Mkandawire (2005) also synthesises how targeting can lead to a variety of perverse outcomes. The errors of under-coverage (missing the legitimate poor, also referred to as ‘exclusion errors’) and leakage (including the non-poor, or ‘inclusion errors’) are the most commonly discussed in the literature. Indeed, precision in poverty measurement becomes ever more crucial precisely when social policy regimes shift towards greater selectivity, such as under the social safety net approach promoted by the World Bank in the 1980s and 1990s. Such shifts usually involve the erosion of more generalised social security provisions that could cover people in the event that they are not effectively identified by a means-testing approach (i.e., in the event that targeting fails). For instance, some estimates of exclusion rates in the Bolsa Familia, the flagship cash transfer programme in Brazil, a country with relatively strong administrative state capacity, were as high as 46 per cent in the 2006 data and 59 per cent in earlier data (Kerstenetzky 2011). Exclusion errors were apparently even higher in the Oportunidades programme in Mexico, another flagship cash transfer programme that arguably originated the fad (ibid.).
These problems of accuracy become all the more acute in contexts of rapid socio-economic change, such as urbanisation, whereby many of the more traditional rural-based social security systems – which are often taken for granted by urban policy-makers and economists – become ineffective or break down, and where states might not even have accurate or complete civil registers of the population. As Mkandawire notes, poor countries also have among the least administrative capacity to be able to target precisely, especially when shifts to targeting occur in parallel with economic crises and severe fiscal retrenchment, as has been the case in many poor countries since the successive debt crises and structural adjustment programmes of the early 1980s onwards.
Other perversities of targeting reside in its political and social implications. Because targeting usually entrenches segmentation and segregation in provisioning systems, this tends to reinforce social and economic stratification by separating the middle classes from the services accessed by the poor. As a result, the political voice of the middle classes is also removed from these services as well, as was seminally laid out with respect to public policy more generally by Hirschman (1970) in his classic Exit, Voice and Loyalty.12 In the best of pro-poor times this can lead to short-term bouts of poverty reduction and even inequality reduction, as occurred in Brazil under the various targeted poverty-alleviation programmes of the Lula and Dilma Administrations. However, sustaining these gains requires strong political commitment and leadership in order to maintain funding, supply and quality within these provisioning systems servicing the poor. Yet, this condition for sustainability is undermined precisely by the institutionalised segregation of provisioning systems encouraged by these targeting approaches. Sustainability under such settings is particularly fragile once politics turn less pro-poor, as has happened in most of the erstwhile New Left governments in Latin America that have been replaced by the Right in recent years, or else where the shift to the right has taken place within the New Left government itself, as in Ecuador, for instance. The recent exception of Mexico will be very interesting to observe in this respect. The resulting political economy paradox was best expressed by Richard Titmuss (1968) – although often attributed to Amartya Sen – that the targeting of services to the poor usually results in poor services.
The danger of reinforcing stratifying and subordinating tendencies can be illustrated by the current agenda of social protection, in particular conditional cash transfers (CCTs), which are now strongly promoted by the World Bank and constitute a major component of the World Bank response to inequality and social inclusion. Besides the problematic moral issues related to the use of conditionalities to manage the behavioural dispositions of ‘the poor’ (e.g., see Rodger 2012), which are usually directed towards women and children and often imply considerable degrees of unpaid time commitments, CCTs are usually targeted through segmented systems of social provisioning. Evaluations of such programmes usually show a poverty reducing impact via increased consumption at the very least (which is a fairly obvious result to all but the staunchest opponents of welfare), as well as some increase in health care and schooling (although this is also often related to simultaneous improvements in the provisioning of these services). There is also some evidence that the expansion of CCTs can bring about quick reductions in inequality largely through their transfer effect on the bottom end of the income distribution. However, evaluations of the longer-term employment or development impacts are less clear, which is arguably due in part to the fact that such programmes are mostly implemented through often poor-quality and poorly funded systems of segregated social provisioning, aimed at servicing the poor. As a result, they tend to have little transformative power on broader social relations or social mobility, although they do reinforce institutional segregation and policy practices of targeting, in lieu of other potentially more transformative strategies of social policy.
This has precisely been the dilemma of the much-lauded Progresa/Oportunidades conditional cash transfer programme in Mexico. The programme has shown some degree of success in raising consumption levels, certain health outcomes and school attendance and enrolment rates (e.g., see earlier studies in Skoufias 2005). These results were obtained with relatively low operational expenses, in large part because the programme was implemented through an already well-established network of clinics and schools servicing the targeted rural populations (as distinct from the subsidised network servicing the urban middle classes). However, even its proponents such as Levy (2006) admit that increased coverage was achieved at the cost of lower quality within this overstretched and segregated network. At that time the programme had no impact on the academic performance of students or on their later employment prospects. Thus, while it had a positive impact on absolute human development indicators, it did so at the cost of entrenching the segmentation of provisioning systems and probably reinforced social stratification as a result. Similar outcomes have been noted with the Bolsa Familia in Brazil (e.g., see Kerstenetzky 2011). Several more recent literature surveys of cash transfers have also consistently noted the problem of a lack of improving school performance or employment prospects with cash transfers (e.g., see DFID 2011).
The potential for targeting to bring about marginal improvements in poverty, education and/or health is not necessarily in question (if, that is, targeting actually results in an increase in resources transferred to the poor, which sometimes it does not). If a poor person is given ten dollars, it should be no surprise that his or her income would be ten dollars higher by the end of the year than it otherwise would have been without such a transfer. The hope or assumption underlying many cash transfer schemes is that the person’s income would increase by more than ten dollars due to some sort of micro-multiplier effect unleashed by the transfer, such as when the extra cash allows the person to overcome other obstacles to increasing their productivity or returns to labour (aka the poverty trap, as in Banerjee and Duflo 2011). The fear of those warning of the perverse incentives induced by welfare is that the income of the person receiving the transfer is less than ten dollars at the end of the year, due to the substitution of some work for the welfare received, to the extreme that the income would not have changed at all due to a perverse threshold effect. This means that the beneficiary would be just below the targeting line and would therefore avoid working so as not to pass above the line. (Of course, as noted in Chapters 3 and 4, it is unlikely that poor people would have any comprehension of such poverty line measures in any case, especially with proxy means testing, which has become the dominant policy tool for targeting in such programmes; e.g., see Brown et al. 2016.) With respect to informality, the fear is also that welfare will create incentives to remain informal, as a means of hiding income in order to continue qualifying for the transfer (aka welfare cheats or queens).13
These narrow questions and debates underlie the obsessive attempts to measure the impacts of cash transfer schemes on the overall income of beneficiaries, which have dominated impact evaluations of these programmes.14 However, that the increment in income might bring the person above a poverty line – say, if their income was previously five dollars below this poverty line and their income increases by ten dollars as a result of the transfer – is so obvious that it is not even interesting. Unless, of course, one is obsessed with the perverse incentives contention, which unfortunately is the case forced upon many policy-makers given the demands to address such concerns among increasingly right-wing electorates and governments.
Rather, the broader concern needs to be the stratifying, segregating and subordinating trajectories brought into play by the institutional modalities used to enact such marginal improvements. These are potentially very counterproductive for any long-term strategy of poverty reduction, particularly if and when resource transfers to the poor decrease rather than increase as a longer-term consequence of targeting, as has often been the case.15
The much-noted case of Brazil is also worth highlighting in this respect. There are debates in Brazil concerning the extent to which the Bolsa Familia – a targeted conditional cash transfer programme – has been the main cause of inequality reduction in recent years, versus other policies such as minimum wage legislation, which might have contributed significantly more.16 Soares et al. (2010, p. 41) calculate that between 1999 and 2009 – during which the Gini coefficient in Brazil fell from 0.591 to 0.538 – reduction in labour income inequality accounted for 59 per cent of the Gini reduction, whereas the Bolsa accounted for only 16 per cent and the non-contributory indexed (targeted) pension system for 15 per cent. In personal communications with Fabio Veras from the International Poverty Centre in Brazil, where many of these studies have been piloted, he commented that this is no surprise given that labour income accounts for around 70 per cent of total income. Moreover, it remains an open question whether the labour income effect was due to increases in minimum wage, reductions in returns to education (with respect to top incomes), a tighter labour market (the unemployment rate had never been so low), or to the increased formalisation of employment (email communication, 5 July 2011). However, the fact that so much inequality reduction occurred within labour income is, in itself, a notable achievement, given that this source has more usually been dis-equalising rather than equalising in the past.
Hence, this particular pattern of inequality reduction would appear to validate the labourist and developmentalist agenda of the Lula administration, in combination with a favourable climate for primary commodity exports and the impact of Chinese investment and demand over these years (and despite tight monetary policies of high interest rates), much more so than the targeted social protection programmes. The contribution of the latter has nonetheless been significant, even if minor, especially if one would make the argument that cash transfers – conditional or otherwise – raise the reservation wage rates of the recipients and reduce labour supply, thereby tightening the labour market and putting upward pressure on wages for low-skilled labour. (However, care must be taken with the logic of such arguments because they easily slip into a discourse of perverse incentives.)
Commenting on earlier inequality decomposition analyses by Soares et al., Kerstenetzky (2011, p. 5) also notes that the inequality-reduction impact of the Bolsa Familia in those years is considerable given that it accounts for a much smaller share of (average) household incomes than its share in inequality reduction. This is explained by the fact that the transfers of the Bolsa are targeted to the poorest households; their appearance in average household incomes would be small even though their weight in the incomes of the poorest households would be large. The immediate impact on inequality is therefore brought about through raising this lower tail-end of the income distribution through the cash transfers (this would be the case even if these poorest households would not necessarily be lifted above the poverty line as a result, which Kerstenetsky notes is often the case). In contrast, the other factors effecting inequality (such as those occurring in the labour market) would be felt much more broadly throughout the income distribution.
From this perspective, the relatively minor contribution of these targeted cash transfers must then be balanced with their potential perverse impacts on social integration. For instance, Lavinas (2006, p. 103) has argued that, despite a law approving the right of a basic income, ‘Brazilian social policies are increasingly focused on increasing the number of means-tested income programmes while making them conditional on a proven lack of resources and targeting only the very poorest segments of society for a limited period.’ In order to counteract this tendency, she proposed that the government should move from means-tested programmes such as the Bolsa to basic income programmes such as the adoption of a universal child benefit scheme. Lavinas (2013, 2017) subsequently has made much more incisive and critical evaluations of what actually happened under the PT government, more or less confirming her initial fears, compounded by alarm at the degree to which the Bolsa had been coopted by a neoliberal financialisation agenda.
Similarly, Kerstenetzky (2009) presciently argued that less selectivity in the scheme, and hence less separation between those who pay for the scheme and the beneficiaries, as well as paradoxically higher expenditures, would ensure wider adoption and political support for the scheme, whereas the-then current emphasis on selectivity in the scheme would result in financial and political constraints to its expansion. In Kerstenetzky (2011, pp. 9–11), she noted that, to a certain respect, political processes undermining the Bolsa had already come into play as early as 2007 when the Brazilian Senate rejected a government proposal for maintaining one of the main sources of funding for social programmes such as the Bolsa, which was supported by a concerted attack on the Bolsa in the media. The rejection ‘represented a concrete threat of stagnation for existing social programs, for it impaired the planned and announced expansions, leaving to the executive the conception of alternatives.’ She thus suggests that ‘we must inevitably ask if the conflict over the [proposal] would in some way anticipate a reversion or saturation of the solidarity indirectly revealed by opinion polls and would constitute a permanent challenge for social programs (especially Bolsa Família), their continuity and necessary expansion.’
Notably, in contexts where the political conditions for maintaining support for redistribution are not as optimal as in Brazil under Lulu and Dilma, such erosions can even be built into cash transfer schemes. For instance, in some cases funding is only mandated for a temporary duration (say, five years). In many others, transfers are not indexed to inflation (this is the case of most cash transfer programmes that I have researched in the course of my recent fieldwork). Or else, a strong emphasis might be put on the ‘graduation’ of recipients from the schemes, as occurred, for instance, in the case of Ecuador from 2013 onwards (see Palacio 2017).
Depoliticisation redux
Viewed in this way, it is clear how the de-politicising allure promised by various global development goals can obfuscate these very political (and politicised) choices about the ways that societies provide public goods and social security to their citizens. The seductively technocratic appeal of the MDGs and SDGs, for instance, with clear goals and indicators grounded in an authoritative and scientifically informed battery of poverty measures, potentially serves to depoliticise these choices among the general lay public, if only through the force of incomprehension and resulting deference to presumably better-informed experts (who are often poorly informed in these wider social policy debates). De-politicisation serves to veil underlying agendas, particularly with respect to the normalisation of targeting and segmentation within social provisioning systems.
Moreover, these agendas cannot be said to be unbiased towards these choices. They are well-served by targeting given their focus on absolute indicators (whether in income, health or education) rather than relative (i.e., inequality) indicators, as well as by the immediacy they compel. For instance, it is relatively easy to raise school enrolments, but much more difficult to raise the quality of schooling, particularly in ways that would significantly alter the employment trajectories of students. This is especially the case when increased enrolments are achieved in stigmatised and lower-quality underfunded schools designated for poor people within a segmented education system. Indeed, if the quantity and quality of employment and the level of wages faced by the poor are not addressed by poverty reduction strategies, the expectations raised by educational improvements might lead to frustration and alienation in the medium term.17 This has precisely been the dilemma of the much-lauded Progresa/Oportunidades conditional cash transfer programme in Mexico, which, as noted above, has had little impact on the academic performance of students or on their later employment prospects. Thus, while it had a positive impact on absolute human development indicators, it did so at the cost of entrenching the segmentation of provisioning systems and probably accentuated social stratification as a result.
While human rights-based approaches are less obviously associated to this tendency of depoliticisation to veer towards orthodoxy, it is important to caution that they can also carry a similar propensity to be coopted due to their ambiguity on a variety of policy fronts, as already discussed in Chapter 2. Human rights-based approaches are arguably founded on a universalistic agenda of social provisioning and social security, as pointed out by Langford (2009). However, in the quest to operationalise these approaches, a degree of ambiguity enters into the translation from ethics to practice. For instance, does the principle of non-discrimination imply universalism (i.e., the same treatment for all) or targeting? Universalism has been criticised, in that purportedly universalistic policies have often reflected fundamental underlying societal biases, such as racial or gender biases. In turn, this implies that a degree of selectivity is required in order to allow for the practice of affirmative/positive action and other forms of preferentiality for disadvantaged or discriminated groups.
Similarly, in the good programming practices specified in the UN Common Understanding (see UNDG 2003), the principle that programmes should focus on marginalised, disadvantaged and excluded groups can be easily construed as a rationale for targeting, particularly when asserted in absence of any substantive discussion of policy. The emphasis on reducing disparity does not, in itself, resolve debates between targeting and universalism given that targeting has been posed by its proponents as more equalising than universalism, as noted above. The principle that people should be recognised as key actors in their own development, rather than passive recipients of commodities and services, can also be attributed as a rationale for using conditionalities in cash transfers, for labour market activation policies or for other means of restricting welfare more generally. This logic – that welfare renders people as ‘passive recipients’ – has been typical in right-wing political attacks against universalism over the past decades. The point is that all of these policy options are fundamentally political in the choices that they elicit; the ethical principles offered by human rights-based approaches do not necessarily resolve these politicised choices along any predictable path.
Indeed, many advocates of human rights-based approaches appear to acquiesce to targeting as the de facto status quo.18 Although the potential fragmentation of citizenship rights that targeting can entail should be a major bone of contention for human rights advocates, this specific but powerful dimension rarely figures in many of their analyses, albeit with some exceptions.19 While such avoidance – perhaps by oversight – provides the appearance of politically neutral moral authority, it also misses a valuable opportunity to explicitly re-embed human rights-based approaches into their earlier postwar association with welfare states and Keynesian commitments to full employment. Instead, human rights approaches risk being subverted as monitoring and disciplinary devices for policy agendas that are otherwise quite antithetical to the spirit of universal rights.20
In light of this confusion, in which the potentials for cooptation under the guise of depoliticisation are great, it is important to elaborate on the meaning of universalism in social policy and how it might apply to the context of most developing countries. As noted previously, universalism has been at the core of modern social policy since it emerged in the nineteenth century alongside modern nation-states and modern notions of rights.21 Universalism has also been central to development studies given that the first countries to move towards more universalistic principles in social policy were not industrial leaders such as the UK but industrial latecomers such as Germany, Japan and Sweden, and universalistic social policy played a key role in their strategies of late industrialisation, as classically discussed by Gerschenkron (1962). Universalistic principles in health and education were also central to the development strategies of South Korea and Taiwan from the outset of their industrialisation efforts in the 1950s, even whilst their social protection systems remained very familial-based up until recently.22 Indeed, it is misguided – at least from historical experience – to state that poor countries cannot afford universal social policy given that successful industrialisers have relied on it in various ways from very poor starting points. Instead, there is arguably an historical precedent that the later an industrialiser, the greater the imperative to innovate shifts towards more universalistic forms of social provisioning earlier than would be predicted by the past experiences of more industrialised countries.23
However, the basic principle of universalism – that all are treated the same – quickly becomes complicated in practice, especially in polarised societies, as noted in the previous section. Indeed, the classical dichotomy between targeting and universalism has led to a certain degree of confusion in the recent policy literature, in particular the recent social protection literature. For instance, targeting often plays a role within universalistic systems, as classically articulated by Skocpol (1991) with reference to the targeting of special needs or disadvantages, or else to allow for diversity within integrated universalistic systems of provisioning, or what she calls targeting within universalism. Moreover, some degree of targeting is almost always required for certain aspects of social protection, such as food relief, and even some of the most universalistic systems include important elements of means-tested targeting, such as welfare in Canada or Sweden. However, this understanding of targeting as a specific policy instrument is different from the broader understanding of targeting as an institutional modality, whereby an entire system of publicly funded provisioning is organised along the lines of selectivity, as discussed in the previous section. It is this latter meaning that has been the focus of criticism by authors such as Mkandawire (2005).
Similarly, the terminology of universalism is often used to refer to an ideal type located in one world region at one point in time (e.g., the UK circa 1970 or Sweden circa 1980) and hence it is argued that it is not applicable to poor countries today. Alternatively, it has come to be used in certain contexts in such a generic way that its meaning has been rendered nebulous. For instance, as noted in the introduction and first section of this chapter, there has been a subtle shift in the implied meaning of universalism towards a narrower connotation of universal coverage or access, such as all children attending school, regardless of how such schooling is provided or financed, or all people accessing health insurance, regardless of whether this insurance covers the totality of health care expenditure needs.
In this regard, debates about universalism and targeting in social policy have exhausted themselves, at least in the realm of social protection, despite efforts to revive criticisms of targeting in the 2000s. The impasse in scholarship has been reached despite the fact that targeting is still very much en rigeur in actual social protection practice. If anything, despite these debates and the ample evidence of the perverse consequences of targeting, the emphasis in current policy agendas is on sharpening and refining targeting as the accepted best or second-best way forward in achieving so-called ‘universal’ social protection, as well as health or education. Indeed, part of the problem emerges from confusions about what, exactly, universalism entails. As noted previously, the World Bank, for instance, has coopted the term by including just about any sort of access to any sort of provisioning as falling under the umbrella of universal coverage, even if large parts of such access occur through segregated and privatised forms of provisioning. Even the agenda of universalism in various UN agencies has been whittled down to a minimalist understanding of universal coverage. However, these confusions have also been facilitated by the difficulties of establishing what would be the realistic and practical alternatives to targeting, especially in the field in social protection. There is more clarity about universalism in the policy sectors of education and health given that these are the policy sectors in which universalistic principles have been practised most comprehensively in strong cases of universalism, even though there are few prospects to achieve the same in most countries. This is in contrast to other areas such as public housing or various aspects of social security, where less achievement has been made towards universalistic principles even in ‘advanced’ countries.24
Indeed, there has been much recent confusion with regard to whether the Mahatma Gandhi National Rural Employment Guarantee Scheme in India should be considered a universal programme given that it does not impose means-testing on those who claim their right to receive employment. Many claim that it is universalistic for this reason, although in effect it represents a form of self-targeting – one of the classic forms of targeting. The degree to which universalism should be applied in the direct provisioning of employment is of course debatable – short of the state assuming a collectivist role in the labour market. This is the position of some, such as with the idea of a universal job guarantee programme, with the government acting as employer of last resort, as advocated by scholars at the Levy Institute in the US.25 Such contention is much less pronounced with respect to health and education, particularly if there is a strong political consensus that public provisioning in these sectors should not be commoditised (which is not the case in the US and in much of the Global South).
Similarly, it is not clear if or how the idea of universalism should be applied in the area of social assistance, which by definition is directed to those in need, not the general population. The position of Barrientos (2013) and Devereux (2016), both leading scholars on the subject of social protection, is that universalism does not apply in this policy sector. But if it would, what would it look like? Devereux (2016) claims that universalism in social protection is sometimes conflated with categorically targeted programmes, such as social pensions for all older persons, although a similar argument could be made with respect to education – the fact that it is categorically targeted towards children does not make the principles of universalism less relevant. But what would universalism be in the area of social protection? Universal basic income is one proposal, but as of yet there are no examples of its implementation beyond a few pilot programmes and there is still much to be answered in the modalities of its implementation – the devil remains in the details still to be revealed. More generally, that elements of targeting are also required even within universalistic systems of provisioning leads some to query what, in the end, is at issue. This leads to the position, in response to the critiques of targeting in otherwise-cherished social protection programmes, that targeting for the time being remains the only available and realistic contingency measure for addressing poverty in contexts of resource scarcity, high levels of informality, political obstacles to more ambitious redistributive agendas, and other constraints.
Some clarification is therefore in order. While universal coverage is obviously a necessary condition of universalism, it is not a sufficient condition. Rather, universalism is best understood as an umbrella term to reflect a set of guiding institutional principles. In an attempt to clarify and systematise these guiding principles in a manner that is attuned to policy-making across the globe, and building on the work by Mkandawire (2004, 2005), I have come to categorise these guiding principles along three dimensions: provisioning modalities, including access/coverage; costing/pricing; and financing. Within each dimension, we can think in terms of a spectrum from strong to weak (or absent) universalistic principles underpinning an institutional system of social provisioning. This approach is useful because it takes us away from the dichotomy of targeting versus universalism and towards a method of identifying shifts towards stronger or weaker universalistic principles, along with their equalising or disequalising potentials, as well as the institutional obstacles potentially blocking such shifts.
Martinez and Sanchez-Ancochea (2014, p. 6) argue that their approach of focusing on coverage, generosity and equity is better than the one I propose here (which draws from Fischer 2012) because mine conflates policy principles with the instruments to achieve these principles, whereas various instruments could be used to achieve these principles. While their point is a valid one and has caused me to reflect, at the same time, the argument tends to come down to splitting hairs about what is a principle and what is an instrument. Their own approach also runs the risk of tautology, that is, bringing results into the definition of causes. Analytically, we should not assume that various instruments associated with universalism would necessarily produce, for instance, greater coverage, generosity and equity/equality. Rather, this should be subjected to empirical inquiry, particularly given that these attributions of universalistic policies are contested by both the Left and Right (especially the Right). Hence, it is important to keep these causal empirical considerations of universalism open to such questions, rather than labelling outcomes as universalistic regardless of the means to achieve them. That being said, as they themselves note, our goals are similar: ‘to separate the definition of universalism from a specific historical experience, to establish the possibility of promoting universalism through different channels, and to consider different degrees of universalism’ (p. 6). It is for this reason that I have decided to maintain my own approach, in the interests of intellectual diversity and as a complement to their own approach.
Provisioning modalities
As stated above, the universalistic principle is not simply that all people access a social good or service (e.g., education or health care), but that this access is provided through integrated systems. In this manner, all people access the service through the same organisational channels or else different means of access are governed by a unified entity, through which needs and standards can be assessed and managed collectively within the system along principles of equity. Hence, a universalistic health system implies that everyone accesses the same or similar hospitals and clinics, wherein the same quality of service is provided to all without discrimination, and triage is organised according to needs rather than means (meaning the ability to pay or other financial considerations). This is not necessarily a public-versus-private-sector issue given that private-sector provisioning can (and often does) occur within such integrated systems (such as private schools or private clinics). Universalism within such systems is determined by the degree to which private providers are regulated and/or managed as an integrated part of the system and are accessible on the same terms regardless of their ownership status.
For instance, an example of strong universalism within an education system would imply that schools are organised under an integrated and unified organisational structure (such as a ministry of education or school board), which regulates quantity and quality within whole system according to universal criteria applied as equally as possible to all. This does not imply that no streaming or targeting takes place within the system given that all school systems implicitly or explicitly stream students from fairly early ages (e.g., towards academic versus technical education), or they target learning disabilities, etc.The question is whether streaming or targeting is organised in an integrated manner within a unified institutional structure accessed by all students, wherein all students are evaluated according to the same standards and treated according to ability or need rather than status or monetary means.
Conversely, weaker universalism within an education system would imply that the school system is stratified or segmented, such that different school systems serve different categories of people according to different standards and that these parallel systems do not necessarily feed into each other, whether in principle or in practice. Hence, certain students (e.g., poor rural students) are streamed into a school system by virtue of their status (poor and rural), which, by consequence of the type and quality of schooling, locks these students into a segregated and subordinated stream that prevents most of them from entering other streams of education at a later stage (such as more academic streams), regardless of their ability. A strong universalistic principle would seek to correct this by providing mechanisms to correct for the disadvantages faced by schools in poorer localities (through funding and also human resources) and assuring that these schools remain integrated with the more advantaged parts of the system.
Targeting in this sense is not necessarily a useful concept when applied to a system of education provisioning (versus, say, social assistance to help poor people pay for tuition fees or to ‘condition’ them to send their children to school). Pure targeting as an exclusive institutional modality applied to education provisioning would imply that publicly funded schooling is only provided to certain means-tested categories of people (e.g., the poor) and all others would access schooling through private means, either in separate schools or else in the same schools. In this sense, pure targeting probably rarely exists within the organisation of education systems given that publicly funded schooling usually occurs at all social levels even if segregated. From an inequality perspective, it is more revealing to analyse how and at what levels public funding and/or provisioning occur.26 The more important principle within sectors such as education or health care is whether the system as a whole is integrated or stratified/segregated, not necessarily whether parts of the service are targeted towards specific groups of people.
This understanding also helps to clarify much of the confusion regarding targeting within universalism, as per Skocpol (1991). There is a world of difference between targeting special needs or disadvantages, such as maternal health needs or learning disabilities, within an integrated system where needs assessments of the population can be managed in a comprehensive and consistent manner, rather than in a fragmented system where different standards are already in force in different parts of the system. High degrees of relatively unregulated private provisioning, such as in the health and education systems of many developing countries and also the US, render such targeting especially problematic.
Moreover, targeting need not imply weak universalism, although in effect, selectivity is usually implemented through segmentation and even segregation of social provisioning systems between different social groups (typically, between poor and middle social strata, or also between different ethnic groups, such as the minority education system in China). Indeed, as noted above, segregation often constitutes much of the political appeal of targeting for middle and upper classes. Esping-Anderson (1990, p. 25) makes a similar point regarding what he calls flat-rate universalism and how this inadvertently promotes dualism within a context of growing working-class prosperity and rising new middle classes because better-off people start to turn to private insurance and to fringe-benefit bargaining to supplement the modest equality promoted in public systems.
Modes of regulation within social provisioning systems also help to distinguish stronger from weaker universalistic principles. Stronger forms of universalism tend to regulate of the quality of service provisioning as a regular part of managing the process of service provisioning (this is also made possible by integrated provisioning), whereas weaker forms of universalism tend to rely more on the measurement of outcomes whilst people exit the system, in part because of an inability to regulate quality within the provisioning. For instance, a high degree of fragmentation in an education system can prevent the ability to regulate and manage quality through the course of education provisioning, and hence such systems usually rely heavily or exclusively on standardised testing as a means to evaluate the quality of learning outcomes in the passage of students to higher levels of schooling.
With regard to social protection, universalism in this dimension mostly applies in the area of social insurance (such as with universal pension schemes or universal health insurance), but much less so with respect to social assistance, as noted above (although there are initiatives in this direction, such as the social protection floor recently mandated by the International Labour Organisation).27 Certain categories of social assistance are by their nature – often by definition – only destined for certain categories of people deemed to be in need and/or worthy of assistance. Hence, pure targeting is much more common as an exclusive institutional modality in social assistance, which is why it becomes an important issue in poverty-reduction policies. The differentiation between strong and weak universalistic principles within social assistance is more along the lines of whether the provisioning of assistance is based on rights-based criteria or means-tested criteria, with the recent trend of micro-conditionalities (as in conditional cash transfers) constituting a compounded form of means-testing.
The second dimension that can help to identify a spectrum from strong to weak universalistic principles in social policy is related to how the costs and prices of provisioning are determined within a system. A strong universalistic principle would imply that costing and pricing are decommodified, meaning that the prices of the provisioning are not determined by market intermediation as if the provisioning is a commodity (i.e., through an auctioning process between supply and demand, or by driving up prices as far as demand continues to allow). Instead, prices are managed through administrative means, and costs are thereby internalised and socialised within the system. As a corollary, users of a universalistic service are not usually faced with the effective price of the service at the time of use, which is discussed in the third dimension below.
Again, the education and health care sectors are the most relevant here, and to a lesser degree social insurance, whereas this dimension is less relevant for social assistance, which by definition is non-contributory and does not impose a monetary price on the recipient. With regard to health care and education, this dimension goes to the heart of the perversities of what Karl Polanyi called ‘fictitious’ commodification, in this case treating health or education as commodities even though they are not ‘produced’ for buying and selling on the market (unlike real commodities).28 Applying market intermediation to the pricing of health services, health insurance or education is problematic not only because markets in these services tend to be highly imperfect given the monopolistic practices that often occur within the provisioning of these goods (such as a locality being serviced by only one hospital, insurance provider or school). It is also problematic because health and education are not alienable or negotiable like commodities (hence the intellectual perversion of the now-dominant ‘human capital’ discourses). This is particularly the case with health, given that the inevitable human condition of ill health debilitates bargaining power precisely at a time of greatest need, leading to a stark asymmetry in bargaining between user and service provider. This asymmetry can become particularly perverse when applied to a market setting. Similar principles apply in education, particularly given the degree to which schooling can be crucial for reproducing class and privilege (or ‘social capital’ in the Bourdieusian sense). Hence, in the absence of quality alternatives due to an underfunded and undermined public education system (such as in many parts of the US), the price of private schooling can be well beyond reasonable levels for the norms of a particular community, thereby powerfully reinforcing social inequalities according to means rather than need or merit. To use economics terminology, because health care and quality schooling are to a large degree demand-inelastic, they carry a huge potential for rent-seeking from the cartel-like activities of private profit-seeking actors.
Conversely, a financially sustainable and affordable operation of a health insurance system, for instance, is predicated on a complementary control of costs within the associated health system. This point was noted, for instance, by President Obama himself during the debates over health care reform in the US during his first term, despite his subsequent inability to enforce such control on the US health care system. Similarly, health insurance programmes in India (such as micro-insurance programmes) or in China (such as the rural health insurance system) are only able to make minor dents on large catastrophic health expenditures by households, partly because of the inability to control costs within the effectively privatised health care systems of both countries – a problem that often faces huge institutional barriers to overcome given that it is usually entrenched in a variety of vested interests extending across providers and vying ministries.29
Financing
The third dimension of universalism relates to the modality of financing. This is closely related to the second issue of cost and price, although it is distinct because it addresses how users pay for the system, whether indirectly or directly. In strong forms of universalism, financing generally takes place indirectly (i.e., not at the time of need) through progressive forms of taxation (i.e., progressive income tax, corporate tax or capital gains tax). In weak forms of universalism, financing takes place directly (i.e., at the time of need), through often regressive forms of payment (such as out-of-pocket payments for health care, school tuition fees or user fees).
Social insurance is most closely related to this dimension given that the principle operating behind insurance is about making the financing of a service more indirect (hence more universalistic), through socialising the costs of such financing for users. Indeed, private insurance is no different from public insurance in this respect, except that it operates according to a profit-motive, usually with less regulated prices and with a smaller pool in which to socialise costs, and for these reasons is often less efficient, more costly and more risk-prone than integrated public systems.
Financing is closely related to the cost/price dimension in that direct financing is usually predicated on commodified systems of costing/pricing, whereas decommodified systems mostly operate through indirect forms of financing as this is the most effective way to socialise costs in the widest possible manner. However, as research by van Doorslaer et al. (2005) has analysed with respect to health systems in Asia, the specific balance between these two extremes and along these two dimensions can have a huge bearing on the poverty and vulnerability impacts of out-of-pocket payments in various health systems. Generally, greater reliance on out-of-pocket payments combined with weaker control of costs and prices (or strong commodification) was strongly associated with a much greater poverty impact of catastrophic health expenditures in the cases they studied, circa the year 2000 (and controlling for level of GDP per capita).
Moreover, understanding the interplay of these two dimensions can also help to clarify much of the institutional tensions associated with reforming social policy systems, as mentioned above. Attempts to control costs within a health system that is largely financed directly usually leads to strong resistance to reform because it undermines the financing of provisioning units within the system. Or else shifts towards more progressive indirect forms of financing can be difficult to sustain in financial terms when prices are commoditised. The huge challenge in social policy reform towards more universalistic principles is found in this need for systemically coordinated changes, which at heart is a fundamentally political issue, not a technical one. For instance, while most people agree on the principle of sending all children to school or all people having access to health care, it is clear how the other dimensions of universalism are potentially rife with intense political dispute, as observed in the battles over health care reform in the US.
Additive versus substitutive changes within social provisioning systems
In addition to these three dimensions, for a full appraisal of universalism, it is also important to consider the more systemic interactions across different policies or sectors of provisioning because progressive expansions in some might be cancelled out by regressive retrenchment in others. I refer to this as a distinction between additive and substitutive dynamics within broader social policy systems. This runs against the general consensus in the welfare state and social policy scholarship that universalism should be judged in a sector- or even policy-specific manner, given that universalism might be identified in some policies or sectors but not in others. While this is true, a more systemic view is nonetheless important to assess the equalising or disequalising potentials of changes within a social policy system or regime and whether the overall tendency of change is towards or away from a more universalistic orientation.
The model reference for an additive approach is the case of the Bolsa Familia in Brazil, at least up until the end of the Dilma presidency. The programme was additive in the sense that the expansion of this programme in the formative years of the 2000s supplemented and reinforced other innovations in social, employment and wage policies and was part of a large increase in overall government spending in the social sector. As a result, it is difficult to distinguish the extent to which the positive outcomes associated with the programme were actually due to the programme itself or to other reinforcing and coordinated initiatives, in the absence of which the programme on its own might have had much less effect.
In contrast, the position of the IMF and the World Bank and many donors is explicitly or implicitly substitutive, in the sense that they advocate the expansion of cash transfers conjoined with retrenchment in other areas. Examples include reforming social security or other policies that have social protection functions, such as subsidies, on the basis that these are inefficient and that the money is better spent on transfers. For instance, this position was referred to earlier with respect to the recent IMF framework on social spending (IMF 2017) and was also common in World Bank and IMF positions on poverty reduction in the 2000s, which generally advocated that expansions in targeted transfers to the poor should be financed through fiscal savings in other areas. While it is possibly true that certain policies are inefficient, at least from a particular microeconomic perspective, the broader point is that such positions lead to no net redistributive gain, even though they might involve shifting groups of beneficiaries, given that spending increases in one programme are roughly cancelled out by spending reductions in others.
These questions are important from an evaluative perspective, especially with respect to more macro questions such as the demand effects of cash transfers on local economic activity. The broader question is not whether the income of the person receiving the transfer improved, but whether the net effect on the overall societal income improved as a result. Similarly, the educational or health effects of a cash transfer and its universalistic potential more generally are determined to a much greater extent by changes occurring in the education and health sectors than in the cash transfer per se.
Viewed from an institutionalist perspective of social policy, with particular emphasis on the principles of universalism and its dynamic interactions across policy sectors, it is clear how the de-politicising allure offered by various indicators of either poverty or inequality can obfuscate the very political (and politicised) choices about the ways that societies provide public goods and social security to themselves. The allure can also distract from the very political processes of social integration and stratification more generally, which go to the heart of distributive conflicts. The seductively technocratic appeal of the MDGs, for instance, with clear goals and indicators grounded in an authoritative and scientifically informed battery of poverty measures, arguably served to depoliticise these choices and processes among the general lay public, if only through the force of incomprehension and resulting deference to presumably better-informed experts. De-politicisation thus serves to veil underlying agendas, particularly with respect to the normalisation of targeting and segregation within social provisioning systems. This potential is equally present within a revised focus on inequality within the SDGs, if not more so than with poverty because inequality goes to the heart of power relations within and across societies, and yet measures of inequality are potentially less intuitive for general lay comprehension than poverty measures.
The challenge for contemporary development agendas lies in seriously re-engaging development debates with questions about how to create genuinely redistributive structures and institutions at national and global levels. These are political challenges given that they cannot be resolved through technocratic solutions, but require choices to be made about the types of societies we wish to inhabit and how we wish to treat each other within and across these societies. In this respect, development agendas would carry more political and policy salience for poverty and inequality reduction if they were explicitly anchored within a wider social policy lens. This includes a more comprehensive consideration of the meaning of genuine universalism within social policy as a holistic institutional guiding principle, rather than simply as an indicator to be achieved (such as enrolling all children in schools regardless of the quality of the schools). Given the paramount importance of these social policy modalities in determining patterns of social integration as well as citizenship rights, they should be central concerns in any broader inclusive development agenda. Conversely, insofar as we recognise high levels of inequality as problematic, the censor of universalistic social policy from mainstream agendas implies abandoning at the outset some of our most powerful policy tools to date for dealing with inequality and poverty simultaneously, particularly when combined with developmentalist policy strategies.
If the goal is to return to these more progressive and transformative development agendas, then a holistic understanding of universalism needs to be rehabilitated from decades of neoliberal obfuscation and explicitly presented as a viable policy goal. For this purpose, it would be much more effective to refer explicitly to these politicised policy modalities rather than to rely on the apparently apolitical moral ground of indicators, as has been the case with the MDGs and SDGs. Such explicit anchoring is necessary in order to render more transparent the political choices and institutional trajectories that such development agendas are often used to legitimate. The risk of not explicitly anchoring future development agendas in politicised policy debates is that these agendas can be (and are often being) subverted towards policy agendas that possibly undermine inequality reduction and/or fragment citizenship rights in many contexts, primarily by way of reinforcing social polarisation through processes of stratification, segregation and subordination.