Rethinking universalism: Older-age international migrants and social pensions in Latin America and the Caribbean

Social protection

Social protection, or social security, is considered a social development practice as it raises the household income while developing beneficiaries’ human capital by encouraging school attendance and raising health outcomes (Midgley, 2014). Therefore, social protection involves the transference of cash benefits to individuals and the provision of benefits in kind (e.g. public education, healthcare, housing). This article pays attention to cash benefits in the older-age population.

There are two broad available cash benefit options in the social policy menu: social insurance and social assistance programmes. Social insurance is a cash transfer based on consumption smoothing and collective risk sharing. It is, generally, a contributory programme where employers and employees contribute shares of the wages to fund cash transfers of workers at specific times of vulnerability (e.g. unemployment, disability, reaching retirement age). On the contrary, social assistance programmes are generally non-contributory programmes ⁴ using benefits to level up individuals living below a societal minimum standard (Cruz-Martinez, 2019b; Barr, 2012).

One of the ways in which the older-age population copes with non-random disadvantages accumulated across their life is pensions. Pensions are cash benefits that can be classified as social insurance (contributory pensions), social assistance (social pensions) or private schemes (individual capitalisation funds), which differs according to the funding sources and the ideology supporting them. Contributory pensions are funded via contributions or taxes, while the state usually finances social pensions, and individuals fund private pensions. Welfare residualism and (neo)liberalism could be considered as the ideologies supporting private pensions, while Keynesianism and the social security Bismarckian approach stands by contributory pensions. Social pensions can follow a residual approach if benefits are targeted to those considered as the ‘truly-deservers’, or can take a social democratic approach if benefits are guaranteed to all older-age individuals. The following subsection briefly describes the main difference between universalism and targeting to be able to define different types of social pensions and distinguish how specific eligibility requirements can discriminate between citizens and particular categories of immigrants.

Targeting versus universalism

Cash benefits such as social pensions can be assigned to all individuals (e.g. universal social pensions) or reserved to a group of individuals with specific characteristics (targeted social pensions). Targeting implies a range of techniques to identify those considered as the neediest and ‘deserving poor’. Following Gilbert (2000: 9–11), at least two ‘pure’ types of targeting mechanisms are identified: categorical targeting and targeting by means.

Categorical targeting uses an easily identifiable category as a proxy to identify a group in the population. Categories include a range of criteria such as age, employment status and geographical location. If categorical targeting is used as the only targeting mechanism, beneficiaries are expected to receive a universal flat-rate benefit. Nonetheless, as this article shows for social pensions in LAC, categorical targeting is usually used in conjunction with other types of targeting mechanisms (e.g. targeting by means).

Targeting by means involves means-testing to identify poor-income households. Means-testing might involve the corroboration of incomes (e.g. wages, contributory pensions) or assets (e.g. properties, capital investments). If targeting by means is used as the only targeting mechanism, beneficiaries are expected to receive a flat-rate benefit/service (e.g. public healthcare) or an incremental benefit conditional on the relative degree of need.

By definition, social pensions are targeted social assistance programmes. The reason for this is that – at least – age is used as an eligibility criterion by the state to identify the beneficiary group. However, once targeting by age is performed, a social pension can still be universal if every older-age individual could become a beneficiary. This is what we call universal targeting (Cruz-Martinez, 2019a; Cruz-Martinez, 2018). If the social protection programme demands other eligibility requirements, such as residency or citizenship, then the social pension could not be considered a universal cash transfer because not every older-age individual could become a beneficiary. However, when discussing these concepts and policy tools, scholars have traditionally connected universalism with the notion of citizenship (Kildal and Kuhnle, 2008; Koehler and Rabi, 2017; Willmore, 2007). Following Marshall (1950), universal social policy entails that citizens in a nation-state benefit from social protection as a social citizen right. Welfare benefits are thus considered as birth rights (Marshall, 1950). Likewise, the Economic Commission for Latin America and the Caribbean classifies a social policy as universal if it is ‘available for the citizenry as a whole’ (Cecchini et al., 2015).

In a globalised world with large numbers of individuals relocating to countries different from their place of birth, the social policy literature should question the normative prerequisite of citizenship for a human being to be considered a deserving individual. Citizenship status and immigration law invalidate the universal access of welfare benefits and hierarchise rights to welfare programmes. For instance, Könönen (2018) confirms the differential inclusion of non-citizens in Finland, one of the so-called universalistic or social democratic welfare states. Könönen (2018) demonstrates the existing legal hierarchies between non-citizens’ entitlements: ‘Non-citizens’ social entitlements differ depending on the nationality, the type of legal status and the form of employment, affecting their position in the labour markets and in the society’ (p. 53). Hayes de Kalaf (2019) evidences the role played by civil registries in the Dominican Republic to exclude ‘racialised others’ from social citizenship rights. Casey (2018) pays particular attention to the discriminatory attacks against the rights of ‘racialised’ others (i.e. immigrants, minority groups, refugees). Hence, citizenship and residency, along with other types of targeting mechanisms, act as powerful devices of exclusion and inclusion of social welfare programmes.

Citizenship refers to the recognition of being born in a specific nation-state by the so-called lottery of birth (i.e. the status of being a citizen from a particular country). Therefore, nation-centric and rigid social policy architectures with social citizenship rights discriminate among individuals because of the chance deriving from the lottery of birth. ‘States typically conceptualise [. . .] social contracts vis-à-vis citizens, rather than with migrants or other visitors’ (Noy and Voorend, 2016: 606). These authors differentiate citizen rights – those extended to every citizen – and human rights – those granted to every individual, citizen or not. Migration is transforming the traditional notion of citizenship and social citizen rights (Bauböck, 2006), signalling the gap between members with full rights to social welfare programmes (i.e. citizens and holders of long-term residence permits) and those living in the nation-state without rights to social welfare (i.e. short-term stayers, migrants with irregular status; Bosniak, 2000).

The literature presents the basic universalism as the next step in the Latin American emerging welfare states (Montero, 2008). Basic universalism, according to Molina (2006), is achieved through a combination of universal coverage of essential services, quality transfers for everyone and with the state as the leading provider. Filgueira et al. (2006: 21) agree with Molina and explicitly call for a basic universalism grounded on universal coverage of welfare programmes coping social risks with uniform standards for everyone and following the social-citizenship right approach. The limitation of the basic universalism proposal is that the social-citizen approach implies the provision of welfare benefits with a rights-based approach but only for citizens. Citizens represent the ‘everyone’; therefore, basic universalism proposes rights-based benefits only for those who meet the citizenship – and in several cases, residency – eligibility requirement.

More recently, the literature has defined universalism as a social policy output achievable in multiple ways. Countries can deliver universal policy outputs through a diverse set of instruments (e.g. social insurance and social assistance) and with welfare providers across different levels of government (Moreno, 2007; Pribble, 2013). Pribble (2013) proposed an operational measure to track the progress towards universalism using Latin America as a case study. The four criteria of a universal model according to her are universal coverage, little or lack of segmentation in the services and transfers, sustainable and equitable financing, and non-discretionary transfers. According to Cecchini et al. (2015: 41):

A pure universal model would be one with full universal coverage (e.g. all families with children, all older adults, all unemployed, the whole population in the case of health-care access) based on objective criteria and underpinned by laws that guaranteed basic rights, adequate and homogeneous service quality, little or no stratification of transfers and sustainable financing based on general revenues (with a progressive tax base) or contributory models whose architecture, combined with provision, generated progressive effects.

Martínez Franzoni and Sánchez-Ancochea (2016) argues that the welfare system secures universal social policy outputs if the entire population living inside the national boundaries has access to generous transfer and services regardless of their class, citizenship, gender and status. Therefore, massive coverage of population needs to be combined with generous benefits – in terms of level and quality – and even distribution (i.e. equity in coverage and benefits). Martínez Franzoni and Sánchez-Ancochea call this the triangle of coverage, generosity and equity.

Language matters and shapes our thinking. Based on the social policy architecture framework, specifically on the eligibility criteria, this article argues that social protection programmes could be labelled as universal only if every individual living in a country is eligible. If citizenship, residency or other attributes act as a discriminatory mechanism, then at most, we could talk about universalism for citizens-residents, but we must not refer to universalism. By doing so, researchers and policymakers are excluding non-citizens, legal residents and non-legal residents from the category of human beings.

Re-examining the concept of universalism: citizenship and residency as discriminatory targeting mechanisms

‘Eligibility rules set out who is entitled to a programme and can exclude certain groups from social protection. These can include rules set in relation to targeting and conditionality’ (Hopkins et al., 2016: 4). One of the groups that can be excluded by eligibility requirements are non-citizens or at least different categories of immigrants such as skilled workers, long-term legal residents, refugees, undocumented migrants.

Sainsbury (2012) argues that it is necessary to have a multidimensional approach when understanding the social rights of immigrants in the host country. She argues that to truly understand the variation in the quality of immigrants’ social rights, researchers need to go beyond the four conditions ⁵ used by Esping-Andersen (1990) to develop the typology of welfare regimes. Specifically, Sainsbury proposes to include two additional complementary conditions: the immigration policy regime (Boucher and Gest, 2015) or incorporation regime (Soysal, 1994), and the form of immigration (Morris, 2002).

The analysis of welfare regimes in shaping the degree of inclusion or exclusion of older-age adults in Latin American social pensions is out of the scope of the article, as it focuses on the structural, programmatic requirements of social pensions which includes or excludes older-age adults based on eligibility requirements. The novelty is the incorporation of citizenship and residency as targeting mechanisms; thus, the relevance to consider the form of immigration and the incorporation regime.

The incorporation regime ‘consists of rules and norms that govern immigrants’ possibilities to become a citizen, to acquire permanent residence, and to participate in economic, cultural and political life’ (Sainsbury, 2012: 16). Here, we are not interested in examining these inclusionary measures but the inclusion of individuals – citizens and non-citizens – into social pensions.

The form of immigration refers to the different ‘entry categories’ of non-citizens in a nation-state (e.g. legal residents, refugees, asylum seekers, ethnic citizens). ⁶ Sainsbury (2012) and Perna (2019) among other international migration scholars highlight the need to update the welfare regime literature taking into consideration the particular rights and restrictions to social rights depending on the category or type of immigrant. The literature mentions several categories: refugees, labour migrants, co-ethnic immigrants, family members and asylum seekers, among others. As Sainsbury (2012: 16) notes, ‘entry categories stratify the social rights of immigrants’. This article examines if citizenship and residency status stratify the social rights of older-age immigrants relocating to LAC.

International mobility of migrants is motivated by several reasons, including the insecurity created by conflicts, the search for upward mobility, family life and pleasure among others. Faist (2013) argues that in fact, spatial mobility has significant explanatory power on social mobility and social inequality. However, as Sheller and Urry (2006) claim, one of the social sciences’ scholarship limitations has been its ‘a-mobile’ or static perception of society. Following this argument of the new mobility paradigm, not considering citizenship-testing and residency-testing as targeting mechanisms in social protection programmes could be exposed as a ‘sedentarist’ limitation of the social policy literature.

According to Sheller and Urry (2006: 208–209) sedentarism, derived from the work of Heidegger (2002), ‘treats as normal stability, meaning, and place, and treats as abnormal distance, change, and placelessness’. This sedentarism and territorial nationalism in social policy could be used as an explanation to understand why residents and citizens are portrayed as the natural beneficiaries of social protection programmes, while irregular migrants – which does not meet citizenship or residency eligibility requirements – are not seen as customary deserving recipients.

A more inclusive society would incorporate non-citizens and would entitle them with more social rights than an exclusive society. While the previous sentence is valid and useful to merge the concepts of incorporation regime and the form of immigration, we must conduct a critical reflexion on what conditions a social programme must have to be considered universal.

Several social pensions have been widely classified as universal by scholars, think tanks and international organisations (e.g. Bolivia’s Renta Dignidad), even though cash benefits are not available unconditionally to all who meet the age eligibility (Cruz-Martinez (2019a); ECLAC, n.d.; HelpAge International, n.d.; Kidd, 2009; Ortiz et al., 2015; Willmore, 2007). This article argues that social protection programmes cannot be catalogued as ‘universal’ if citizenship and residency status acts as stratified mechanisms targeting the targeted older-age individuals. The article adopts the coverage dimension in Martínez Franzoni and Sánchez-Ancochea’s (2016) definition of universalism. Rather than massive coverage, universalism is understood as the total coverage of the population. The generosity and equity dimensions were not included in the empirical analysis and are left for a future research paper. Therefore, a universal social pension is a cash-transfer programme for the entire population above the age-eligibility requirement living inside the national boundaries regardless of their class, citizenship, gender and status.

As it was noted in the ‘Introduction’, higher numbers of older-age individuals are relocating to countries different from their place of birth. So what happens to an older-age individual if she/he migrates to LAC? Do immigrants have to meet the same categorical and means-tested targeting mechanisms as citizens? Is there any additional eligibility requirements limiting their inclusion as beneficiaries of a social pension scheme?

Inclusion and exclusion of beneficiaries via targeting mechanisms

To answer the research questions and describe the level of inclusion or exclusion of social pensions in LAC towards older-age immigrants, five targeting mechanisms are used as explanatory factors in the qualitative comparative analysis, summarised in Table 1.

OPEN IN VIEWER

Table 1.

Description of targeting mechanisms.

Type of targeting	Requisite to become a beneficiary
Citizenship	Be a citizen of the nation-state providing the cash transfer
Residency status	Be a legal resident of the nation-state providing the cash transfer
Geographical location	Living in a specific municipality or province
Access to any other kind of pension	Not being a beneficiary of contributory, service-related or private pensions
Means-testing	Have incomes/assets below the threshold defined in the social pension programme

The first two targeting mechanisms are citizenship and residency status. Both are categorical targeting mechanisms – as it was previously mentioned – even though they are not usually considered as such in the social pension’s literature. Citizenship ‘generally denotes full membership in a political community, comprising a set of rights and/or political obligations’ (Craig, 2002). An individual is considered a citizen of a nation-state if one or both of their parents are citizens (jus sanguinis), by the place of birth of the individual (jus soli), or by years of legal residency (naturalisation). Social pensions targeting citizenship will exclude all different categories of immigrants residing in the nation-state – regardless of their status.

Meanwhile, residency refers to the current place of residence and the legal status of the individual. Therefore, social pensions targeting residency will exclude citizens with a legal permanent residence in another nation-state as well as immigrants without legal residency status. Generally, a specific amount of years of legal residency prior to the application of the social pension is required.

Geographical targeting is also another type of categorical targeting. This mechanism ‘selects different districts, villages or regions – often on the basis of [income] poverty data where [deprivation] is known to be chronic, or on the basis of other criteria (such as proneness to earthquakes, flooding [or other] natural disasters’ (Slater and Farrington, 2009: 9). This type of targeting will include as beneficiaries only those living at predetermined specific locations inside the nation-state.

Pension-testing is a categorical targeting. Beneficiaries are targeted, taking into account if the individual is a pensioner (category) or not of a contributory, foreign or private pension. Therefore, pension-testing excludes individuals entitled/receiving any other kind of pension. ‘Pension’ is the label used for this type of targeting in Figure 1.

OPEN IN VIEWER

Figure 1.

Targeting older-age adults in Latin America and the Caribbean: examining eligibility requirements and immigrant-friendly social pensions in a five-level diagram after csQCA.

Finally, means-testing identifies the beneficiaries based on assessments of income, assets or wealth. Generally, those with incomes below the poverty line – or alternatives monetary thresholds – are included as beneficiaries.

Research technique: csQCA

The purpose of the QCA is to transform information from data into a truth-table where different combinations of conditions producing a specific result can be identified. In this research, QCA is used to determine if it is possible to identify a universal or an inclusive immigrant-friendly social pension programme by examining five categorical and means-tested targeting mechanisms as causal conditions.

Ragin’s (1987: 84) intention with this technique was to develop a strategy as a midpoint between the case-oriented approaches (qualitative) and variable-oriented approaches (quantitative), where the best features of both could be integrated into one method. Ragin (1987) rejects the argument that through QCA a single causality is determined, rather a number of different paths that might have allowed the corresponding final outcome to be determined. In his next books, Ragin (2000, 2008) calls for a justification of the use of crips (dichotomous) versus fuzzy sets. The main distinction between the two is the degree to which they satisfy membership criteria. With crisp-set QCA (csQCA) variables can only have two values, which have to be determined by the researcher based on a defined threshold. It distinguishes full-membership (1) versus non-membership (0) (i.e. in vs out, high vs low). On the contrary, with fuzzy-sets QCA (fsQCA), membership scores can vary from 0.0 to 1.0. This article relies on csQCA, as social pension programmes either incorporate an eligibility requirement as a targeting mechanism or not.

Therefore, the article examines the various paths of eligibility requirements in social pensions conditioning three specific outcomes: (1) access for every older-age individual living in the country (truly universal), (2) access for every category of immigrant (no targeting by citizenship or residency) and (3) access for older-age immigrants with legal resident status at the moment of application and who meet any other requirement (targeting by residency and no targeting by citizenship).

Variables in the csQCA are used to portray the stratifying mechanisms of eligibility requirements and to identify a potential universal or inclusive immigrant-friendly social pension. The csQCA tracks the set of causal conditions that must be met in each case to become a beneficiary. In addition, the csQCA allows identifying patterns of targeting mechanisms, tracing commonalities across social pension’s programmes in the region and could be even used to propose exploratory regimes of inclusionary social pensions.

Causal relationships between targeting mechanisms of eligibility requirements (explanatory factors) and the three outcomes are examined through Boolean algebra principles and comparative analysis. Qualitative data binary matrices (0, 1) are used to describe the presence or absence of specific characteristics or conditions. A case exhibits a ‘1’ in the explanatory variable presenting the favourable features for the desired outcome. Therefore, cases with a ‘0’ in the explanatory variable exhibit unfavourable features for the desired outcome. A social pension exhibits a ‘0’ in the csQCA if the programme has an eligibility requirement for any of the five targeting mechanisms. For example, a social pension with a ‘0’ in the citizenship factor denotes that only citizens can become beneficiaries, while a social pension with a ‘1’ denotes that both citizens and non-citizens can become beneficiaries.

The absence of the citizenship, legal residency, means, geographical and pension eligibility requirements lead to the first desired outcome: access to a social pension for every older-age individual living in the country (truly universal). The analysis of necessary conditions in the fsQCA software presents with an error, as expected, because all five tested conditions are necessary for the desired outcome.

The absence of at least citizenship and legal residency eligibility requirements leads to the second outcome being examined: access for every category of immigrant. The analysis of necessary conditions in the fsQCA software confirms, as expected, that citizenship and residency conditions have a consistency score of 1.0.

Third, the presence of the residency eligibility requirement and the absence of at least the citizenship eligibility requirement lead to the third outcome being examined: access for older-age immigrants with legal resident status. The next section unveils the need to distinguish, in addition, if immigrant-friendly social pensions require a specific amount of time of legal residence at the moment of application. The analysis of necessary conditions in the fsQCA software confirms, as expected, that the citizenship condition has a consistency score of 1.0 in the immigrant-friendly path without any specific time requirement of legal residence. However, the absence of means-testing also has a consistency score of 1.0, due to none of the cases following the third path target social pensions by means (see Ragin and Davey, 2016 for more details on the fsQCA analysis).

A multiple truth-table is constructed showing the favourable (1) and unfavourable (0) targeting conditions for a truly universal social pension (not shown here). The absence of targeting mechanisms is considered the favourable condition for a universal or an immigrant-friendly social pension. Instead of the truth-table, a five-level figure – one level for each targeting mechanism besides age – is built to display and trace more easily the combination of targeting mechanisms in the 30 social pensions. The figure reveals a binary map with the combination of conditions necessary for the social pension to be truly universal or – at least – open to granting access for older-age immigrants with legal residency status. Figure 1 not only helps to identify distinct targeting mechanism trajectories but also facilitates visualising the gaps in the access to social pensions for individuals with different immigration statuses.