settings, as signaled by the recent high-level commitments at the World Humanitarian Summit, the. Grand Bargain .... ben
Can transfers increase social cohesion in refugee-hosting settings? Experimental evidence from Colombian refugees in northern Ecuador
Elsa Valli, Amber Peterman and Melissa Hidrobo1
Abstract: There is increasing interest in understanding if social protection has the ability to foster social cohesion, particularly between refugees and host communities. Using an experimental evaluation of transfers, including cash, food and food vouchers to Colombian refugees and poor Ecuadorians in urban and peri-urban areas we examine if transfers resulted in changes in social cohesion measures. The evaluation was a cluster-randomized control trial (cRCT) examining a short-term program implemented over six months by the World Food Programme. We examine six aggregate dimensions of social cohesion, derived from 34 individual indicators, in addition to an overall index of social cohesion. Overall results suggest that the program contributed to integration between Colombians and the hosting community: through the increases in personal agency, confidence in institutions and social participation for Colombians specifically and increasing attitudes accepting diversity for all participants. Importantly, it appears there were no negative impacts of the program on indicators or domains analyzed. Although we are not able to specifically identify mechanisms, we hypothesize that these impacts are driven by joint targeting, messaging around social inclusion and through interaction between nationalities at mandated monthly nutrition trainings.
Key words: Social transfers, Colombian refugees, social cohesion
1
Elsa Valli and Amber Peterman are with the Social and Economic Policy Unit at the UNICEF Office for Research—Innocenti, Melissa Hidrobo is with the Poverty, Health and Nutrition Division of the International Food Policy Research Institute. Corresponding authors: Elsa Valli and Amber Peterman, Piazza SS Anunnziata 12, Florence, Italy 50142:
[email protected];
[email protected] We are grateful to the Centro de Estudios de Poblacion y Desarrollo Social (CEPAR) for data collection and management, and to the World Food Programme (Quito and Rome) for collaboration and program implementation. We thank colleagues at the International Food Policy Research Institute during the design and implementation of the impact evaluation including John Hoddinott, Amy Margolies and Vanessa Moreira and participants at the International Conference on Social Protection in Contexts of Fragility and Forced Displacement in Brussels, September 2017 for helpful comments and feedback.
1
1. Introduction According to the United Nations High Commissioner for Refugees (UNHCR), an estimated 65 million individuals were forcibly displaced in 2015, an increase of nearly six million individuals from the previous year (UNHCR 2016). In response, global actors have committed to exploring policies and interventions to mitigate against health, economic and social consequences for both displaced populations and populations living in fragility in sending countries. Social protection, including the use of social transfers, has been identified as a key intervention for vulnerable populations in these settings, as signaled by the recent high-level commitments at the World Humanitarian Summit, the Grand Bargain, and the High Level Panel on Humanitarian Cash Transfers (ODI and CGD 2015; WHS 2016). As the use of social protection schemes among refugee populations to diverse ethnic groups within fragile settings increases, the effects of these policies on social cohesion is increasingly questioned. Social protection has potential to increase (directly or indirectly) measures of social cohesion by creating good will, feelings of equal treatment (both within and between groups), trust in institutions, and social capital through interaction with other beneficiaries during program-related activities (e.g. trainings, community meetings) (Leites et al. 2017). However, it is equally possible that implementation of social protection could generate feelings of resentment and jealousy towards recipients and trigger or exacerbate intra-community or intra-ethnic tensions. The concept of social cohesion has been used to describe social relations, including cooperation and solidarity between groups and individuals in a society, and their interrelation with broader economic, social and political outcomes (Babajanian, 2012). Because of the variety of definitions that have been produced, varying by discipline, context, approach used, or issues it is being used to frame, it has been described as a ‘quasi-concept’ (Beauvais and Jenson, 2002; Green et al., 2009). Definitions often respond to policy needs and foci of agencies and institutions. For instance, the Council of Europe, among the most active promoters of the concept, has broadly referred to social cohesion as “the capacity of a society to ensure the welfare of all its members, minimising disparities and avoiding polarizations (Council of Europe, 2004, p. 3).”
2
Better understanding of dynamics between social protection and social cohesion is justified not only by the fact that social protection is viewed as a crucial policy to promote social justice, of which social cohesion and inclusion are integral part (Devereux et al., 2011). In addition, social cohesion plays a role in fostering economic and development outcomes, for instance via its influence on quality of institutions and, in turn, on implementation of pro-growth policies (Ritzen et al., 2000; World Bank, 2005; Easterly et al., 2006; Ferroni et al., 2008; Hayami, 2009). Social cohesion is also widely viewed as crucial in relation to peacebuilding and conflict prevention (King et al., 2010; Colletta and Cullen, 2000). While the beneficial development impacts of social protection has been widely studied, the contribution of social protection to social cohesion and state-building have been primarily assumed based on theory (Babajanian, 2012). For instance, research and policy institutions highlight the role of social protection in fostering social cohesion by “strengthening the ‘contract’ between citizens and the state, promoting social inclusion, integration and greater accountability” (DFID, 2011, p. 42; OECD, 2009; ODI, 2012). This topic is increasingly on the political agenda as the volume and visibility of refugee and displaced persons have placed host communities and governments under stress (ILO, 2011). Despite this positive implied role of social protection, programs may also have social costs, which could entail the deterioration of community cohesiveness and the potential erosion of informal support networks (Devereux et al, 2017). On the micro-level, the evidence of the impact of social protection on various social cohesion outcomes is limited and inconclusive, with few examples in refugee hosting or fragile settings. In Lebanon the cash component of a winterization2 program targeting Syrian refugees as well as vulnerable Lebanese households had a positive impact on social relations between beneficiaries and other community members (Lehmann and Masterson, 2014). In particular, the treatment group was found to be statistically more likely to receive help from Lebanese community members, to provide help to Lebanese community members and less likely to be insulted by them. Familias en Accion, Colombia’s flagship conditional cash transfer (CCT) was found to increase social capital, proxied by the willingness to cooperate among community members during a public good game (Attanasio et al.
“Winterization” is a term used by the humanitarian community to refer to the process of assisting beneficiaries in staying warm, dry, and healthy during winter months (Lehmann and Masterson, 2014). 2
3
2015). In Lesotho, the Child Grant Program was found to strengthen informal sharing arrangements in the community, with beneficiary households being more likely to receive informal support from family members, friends and neighbors and to provide support to the rest of the community (Pellerano et al., 2014). However, there is also some evidence suggesting potential for adverse effects of social protection on social cohesion. In Indonesia, poor targeting of the Bantuan Lansung Tunai (BLT), a national unconditional cash transfer (UCT) aimed at compensating the poor from the increase in the kerosene price (as a consequence of the reduction of fuel subsidies and which erroneously included a large number of better-off households), resulted in deteriorated social capital, as measured by the number of community groups individuals’ participate in, as well as increases in crime rates3 (Cameron and Shah, 2012). Qualitative studies also report a number of instances in which negative consequences mostly deriving from discontent around targeting. For instance, in Zimbabwe, the dissatisfaction linked to targeting procedures of an emergency pilot program was found to worsen social cohesion, although this dynamic was found only for cash transfers which caused community tensions, as opposed to food transfers which did not have the same effect (Kardan et al., 2010). In Yemen and Kenya, feelings of jealousy around targeting led to a considerable degree of tension between beneficiaries and non-beneficiaries of an unconditional cash transfer program. Envy, jealousy and resentment were reported in West Bank and Gaza for similar reasons (Pavanello et al., 2016). Overall, evidence suggests that context and program design specifics (e.g. targeting, complementary activities, program framing) are crucial in explaining diverse findings, however a critical mass of research is not yet available to make these types of generalizations. Using an experimental evaluation of transfers, including cash, food and food vouchers to Colombian refugees and poor Ecuadorians in urban and peri-urban areas we examine if transfers resulted in changes in social cohesion measures. The evaluation was a cluster-randomized control trial (cRCT) examining a short-term program implemented over six months by the World Food
3
The definition of crime used by the authors include theft, looting, pillaging, assault, arson, rape, misuse of drugs, illegal drugs, murder, the sale of children and other. These categories were predefined and collected in the Indonesian Village Census, one of the data sources used in the analysis.
4
Programme (WFP). Overall results suggest that the program contributed to integration between Colombians and the hosting community: through the increases in personal agency, confidence in institutions and social participation for Colombians specifically and increasing attitudes accepting diversity for all participants. These impacts appear to be largely independent of the type of transfer given, and appear to accrue to all Colombian nationals, regardless of their motivation for migration (e.g. economic versus political or personal). Importantly, it appears there were no negative impacts of the program on indicators or domains analyzed. Although we are not able to specifically identify mechanisms, we hypothesize that these impacts are driven by joint targeting, messaging around social inclusion and through interaction between nationalities at mandated monthly nutrition trainings. This manuscript contributes to the literature in a number of important ways. First, it adds to the scarce literature around social cohesion impacts in refugee hosting settings. It is the first experimental evidence of quantitative impacts in refugee hosting settings which we are aware of, and the first evaluation of any type in the Latin American setting. Secondly, the paper contributes to discussion around measurement, giving the most comprehensive operalization of social cohesion thus far and suggesting areas where measurement could be improved in future studies. Finally, we discuss potential design components and mechanisms through which social protection has potential to affect measures of social cohesion, an important area of further study with direct policy and program implementation implications.
2. Framework and Literature Originally developed by Durkheim (1893), the concept of social cohesion was used in relation to the effects that modernization and industrialization had on forms of solidarity (King et al., 2010). More recently, the multi-dimensional concept has regained attention as national policies seek to contribute to factors including stability, inclusion, participation, belonging and trust. As noted by Jenson (1998), the attention to issues related to social order and cohesion tend to arise at times of rapid social change. As Durkheim and others’ work on social cohesion flourished in response to the transition to industrial capitalism in Europe, the more recent attention to social cohesion can be linked to the rapid globalization and its effects (Green et al., 2009). 5
Different authors and institutions have suggested different approaches to analyze or measure social cohesion. Beauvais and Jenson (2002), for instance, suggest a framework that distinguishes between those dimensions that capture social bonds and associational activity (often referred as social capital) and those which emphasize solidarity and equity. Similarly, Babajanian (2012) reviews social protection and its contribution to social cohesion and state building. In doing so, he distinguishes between distributional and relational dimensions of social cohesion, the former referring to “the patterns and the extent of distribution of resources and opportunities in a society’ and the latter to ‘the nature and quality of interpersonal and social relations (p. 13).” For the purpose of this analysis, we follow the approach theorized by Babajanian (2012) focusing on a sub-set of the relational components proposed including the domains of “empowerment, community cooperation and solidarity [and] social participation (p.8).” We complement this definition with two domains suggested by Green et al. (2009) who reviews regimes and definitions of social cohesion, specifically 1) tolerance and respect for other individuals and cultures, and 2) interpersonal and institutional trust. Due to data availability and appropriateness of the domain in relation to social protection, we do not include all components suggested by the authors (e.g. conflict and stability as suggested by Babajanian (2012) or shared values and goals as suggested by Green et al. (2009)). As pointed out by Green et al. (2009), social cohesion definitions do not necessarily need to include all these characteristics. Therefore, our operational definition of social cohesion is based on six aggregated indicators which include the following dimensions: 1) trust in individuals and social connectedness, 2) personal agency, 3) attitudes accepting diversity, 4) freedom from discrimination, 5) confidence in institutions, and 6) social participation. We discuss these domains in more detail in the methodology. Social protection can, on theoretical grounds, influence social cohesion in a number of ways, and different design features and characteristics of social protection can have direct or indirect, intended or unintended effects on a number of dimensions of social cohesion. We briefly summarize key features identified in the literature which arise important design considerations, including targeting, communication (messaging and framing around implementation), implementer, size of benefits provided and complementary activities or system linkages.
6
A key consideration in social protection programs with implications for social cohesion is targeting. As previously discussed, the most commonly observed and reported negative effect concerns between-group tensions and feelings of resentment arising as a consequence of exclusion of vulnerable groups from programming (Babajanian, 2012; Pavanello et al., 2016; Leites et al. 2017). This is particularly the case for complex targeting which is not completely transparent. For example poverty targeting with proxy means tests, particularly in areas of widespread poverty will result in beneficiary lists which may not be easily distinguishable from excluded populations by community members and other stakeholders. Other targeting methods, for example categorical targeting which identify beneficiaries based on demographics, are quite transparent and are less prone to this type of dissatisfaction. On the other hand, it has been suggested that in some cases categorical targeting social groups deemed undeserving of assistance, i.e. ex-combatants in post-conflict contexts, might result in feelings of resentment by those excluded from the program thus impairing the process of peace building (Holmes, 2009). Further, some researchers suggest that community-based targeting has higher levels of acceptance due to involvement of key local stakeholders and community members as compared to methods like proxy means tests (Ellis, 2008; Alatas et al., 2012). Clear communication about program objectives and targeting and involvement of the community in its design and beneficiary selection can therefore significantly reduce the risks associated to intra-community tensions and negative state-society relations (Pavanello et al., 2016). Receipt of resources and services can foster social cohesion through enhancement of selfconfidence, agency and empowerment. Higher availability of resources, which allow beneficiaries to participate in ceremonial, cultural and other social activities, can help to strengthen social ties and break isolation (Pavanello et al., 2016). Children able to attend school can feel more accepted among their peers and feel higher level of achievement and self-worth, contributing to children and adolescents agency (Attah et al., 2016). Being selected for a regular transfer, particularly if it is a benefit from a state actor, can improve the confidence that an individual has about the future and can strengthen trust in institutions (Pavanello et al., 2016). These impacts can accrue not only from the receipt of financial resources, but also from the complementary services and system linkages layered onto or integrated into programs (Roelen et al., 2017). For example, numerous social protection 7
programs require or encourage the active social engagement of beneficiaries which have potential to increase social networks and social capital. These design features could include participation in training and other group activities or social events (whether conditional or not), which have been found to generate feelings of solidarity, mutual support and increased dignity (Pavanello et al., 2016). However, the creation of groups and activities associated with the program can also foster social divisions, particularly if beneficiaries are part of already socially marginalized groups (Skoufias, 2005). These design features and mechanisms may be particularly important within social protection programs targeting minorities or isolated and vulnerable groups, including refugees. For example, if programs exclusively select minority non-national citizen group, excluded local vulnerable individuals could perceive this as unfair, resulting in higher levels of social fragmentation and tension. On the other hand, social protection could foster solidarity among refugees and national beneficiaries, particularly if the social interaction is encouraged for the first time through the program. These effects can be categorized using definitions of Narayan (1999) and Putnam (2000), according to whom social protection can foster ‘bonding’ social capital (i.e. solidarity within members of the same group) and ‘bridging’ social capital (i.e. linking between different social groups). As documented by an extensive literature (reviewed by Bauer et al., 2016), individuals that have been exposed to higher levels of war or crime violence appear to have a more pro-social behavior and higher levels of civic and political engagement, particularly within groups of similarly affected individuals. For example, in Sierra Leone, individuals that experienced more violence were also more likely to report attending community meetings, to vote in elections, to join social and political groups, and to participate in school committees and “road brushing” (Bellows and Miguel, 2006; 2009). They were also to be altruistic and more inequality averse towards in-group members during lab-in-field experiments (Bauer et al., 2014). However, Bauer et al. (2016) hypothesize that the high levels of within-group cohesion are associated with low levels of social cohesion between groups, as a direct consequence of the parochial attitude. This in turn can have the potential for conflict cycles and return to violence. A social protection policy that targets refugees could therefore have potential for both positive and/or negative effects on both intra and inter-group social cohesion. 8
3. Program, evaluation design and methodology
3a. Program and context WFP implemented the ‘Cash, food and voucher’ program over six months from April 2011 to September 2011. The program was targeted to Colombian refugees and poor Ecuadorians across urban and peri-urban neighbourhoods in seven urban centers in the northern provinces of Carchi and Sucumbios. The three objectives of the program were to: 1) improve food consumption by facilitating access to more nutritious foods, 2) increase the role of women in household decision-making related to food consumption and 3) to reduce tensions between Colombians and host Ecuadorian populations. Neighbourhoods’ within the seven urban centers were selected for program implementation based on consultations with UNHCR, as both areas with high numbers of refugees and levels of poverty. Households qualified for the transfers if they met the poverty threshold as determined by a proxy means test, and if they were not currently receiving the government flagship cash transfer program. Transfers, which were equivalent to $40USD, were targeted to women and delivered monthly: cash transfers were delivered via pre-paid ATM cards, food comprised of rice, vegetable oil, lentils and canned sardines, and food vouchers were redeemable for pre-approved nutritious foods at local supermarkets. In addition to transfers, beneficiaries were required to attend monthly sessions on nutrition training. Previous analysis of the data indicates that transfers were successful in meeting food security objectives, increasing both quantity and quality of food consumed, and also resulted in decreases in intimate partner violence (Buller et al. 2016; Hidrobo et al. 2014; 2016). Although a main program objective, the implementation activities to support the goal of increasing social cohesion were not necessarily explicit. For example, in targeting areas with high populations of Colombians, and in delivering the same benefits to both Colombians and Ecuadorians in these areas, WFP aimed to create a feeling of equal treatment.4 In addition, through the nutrition
4
Originally, WFP had conceived of the program as targeted exclusively to Colombians. However, this was viewed as problematic with the potential to create social tensions, thus the scope was expanded to include both nationalities (excluding Ecuadorians already receiving the Government flagship cash transfer program, the Bono Desarrollo Humano).
9
training sessions, it was possible that both nationalities would interact with each other, in some cases for the first time in a meaningful way. However, there was also explicit messaging during the start of the program, when sensitizing beneficiaries to the overall program components and guidelines of the program, as well as when presenting the program to local administration and governing stakeholders. For example, a first page of a program guide ‘booklet’ explains how Colombians have come to Ecuador to flee violence and are equally in need of assistance as Ecuadorians, and describes Ecuador as a ‘welcoming country’ (Figure 1). The information includes messages motivated by statistics in the region that poor nutrition is a concern for everyone and that no one should be left behind or left out. Therefore, although the program did not implement intensive activities to create impacts on social cohesion, there were implicit targeting decisions, socialization and messaging components which could have influenced social cohesion on an inter-personal level, in addition to the benefits of interacting with WFP and receiving economic transfers. It is important to note that although Ecuador is the largest refugee-hosting country in Latin America, hosting primarily Colombian’s fleeing from the decades-long conflict initiated by the Revolutionary Armed Forces of Colombia (FARC), we cannot explicitly identify refugees from asylum seekers from other types of migrants. According to UNHCR, in 2012 at the time of the study, Ecuador hosted approximately 55,500 refugees, nearly all of whom originated in Colombia (UNHCR 2012). However, due to an increasingly stringent application and approval processes, there were estimated to be approximately 68,300 people in ‘refugee-like’ status and 14,400 asylum seekers of Colombian nationality (ibid). UNHCR estimates that nearly 70% of refugees live in urban settings, thus there is attention from international actors to strengthen implementation of urban programming and assistance to these groups. There is evidence that Colombian refugees are subject to discrimination and stigma, with implications for economic and health outcomes (Echeverri Erk 2016; Shedlin et al. 2014). In some cases, this can be explicitly linked to ‘common perceptions’ that Colombian refugees are linked to violence and the drug trade, a belief that stakeholders have sought to address through media and other campaigns (UNHCR, 2012).
10
3b. Evaluation design The cRCT included four arms and 145 clusters randomized to cash, food, food voucher and control. Because of the close geographic proximity between clusters, the randomization was done in a twostaged process whereby first neighborhoods were randomized into treatment and control, and subsequently, clusters within treatment neighborhoods were randomized to treatment arms (cash, food and food voucher) (see Figure A1). The baseline survey was conducted in March-April 2011, before the first transfers and the endline survey approximately seven months later in October-November 2011. In total, 2,122 households were surveyed in both waves with an overall household attrition rate of approximately 10%. For this analysis, we restrict the sample to the panel of individuals who responded to the questionnaire at both baseline and midline (who were most often also the transfer beneficiary). Overall, the individual-level attrition rate was 20%, indicating that in approximately 10% of the baseline households, different individuals answered the questionnaire over time. Although the individual attrition rate is slightly higher in the control group, as reported in Table A1, the difference is not statistically significant. Despite the similar attrition rates between treatment and control, differential attrition could still threaten the internal validity of the study if individuals who left the panel are different than those who stayed, and these same characteristics influence outcomes of social cohesion. We therefore examine if individuals who attrited are different between treatment and control groups with respect to background characteristics and outcomes of social cohesion at baseline. Comparing the p-value of differences from tests of means (column 8), only two of 19 background characteristics and none of the seven aggregate outcomes indicators appear to have a statistical difference between control and treatment attritors at the p
