Introduction
In low- and middle-income countries, cash transfers (CTs) have been paid to vulnerable households to provide a social safety net to cover the costs of basic necessities such as food, health care or education. Cash payments have been successful in reducing poverty and improving health and social outcomes [
1,
2]. The mechanisms through which CTs have been found to reduce poverty are by improving food security, offering economic autonomy and self-sufficiency, and strengthening households by increasing resilience, enhancing human capital, and consequently reducing intergenerational cycles of poverty [
3‐
6].
CTs have generally been paid to female caregivers rather than male household members based on the evidence that cash controlled by women is likely to benefit the entire household, particularly children [
7]. Evaluations of these programmes have linked CTs to increased child immunisation [
8] and school enrolment and attendance [
9,
10]. Recently, in sub-Saharan Africa, trials assessing CTs for HIV prevention have experimented giving the cash directly to adolescents [
11‐
13]. The trials have either paid CTs solely to the adolescent [
13] or paid a portion to the adolescent and a portion to the parent or guardian [
11,
12].
Studies examining the effect of CTs on HIV outcomes have documented mixed results. Most studies show that there are benefits to both conditional and unconditional CTs for adolescent girls and young women (AGYW). The Zomba trial in Malawi paid AGYW up to USD 5 and their parents up to USD 10 monthly to assess the effect on HIV prevalence among AGYW [
11]. Findings showed that AGYW who received CTs had lower HIV prevalence. They were less likely to engage in risky sexual behaviour and less likely to have to teenage pregnancies and early marriage [
11]. In South Africa, HPTN 068, a randomised controlled trial (RCT) assessing the effect of a CT conditional on school attendance on HIV incidence among AGYW, paid USD 10 to AGYW and USD 20 monthly to their caregivers. In this study, adolescent recipients were less likely to report risky sexual behaviour and intimate partner violence than were non-recipients. However, no impact was found on HIV incidence [
12]. Qualitative studies forming part of these trials show that CT payment to adolescents increased their autonomy, peer status, self-esteem, and decision-making abilities [
14,
15].
Evaluation studies suggest that the payment of CTs to individuals in vulnerable households also has a broader unintended impact on social relationships [
16‐
18]. Research carried out on national social grant programmes in Eastern and Southern Africa has shown that in addition to reducing poverty, the grants also strengthened social networks and promoted participation in community events and social cohesion [
18‐
21]. In Kenya, the CT provision to female caregivers encouraged sharing, borrowing, eating together, and an increased ability to partake in community events [
22]. However, these studies also showed that targeting specific households and excluding others can have divisive consequences in close-knit, poverty-stricken communities [
23,
24]. Such targeting potentially leads to tension, jealousy, and conflict [
16,
25]. Even the eligibility criteria used in CT programmes to identify recipients could threaten social ties [
26]. Targeting specific individuals or households and excluding others may induce stigma and resentment between recipients and non-recipients [
16,
25,
27,
28].
These findings are important to bear in mind when considering benefits and drawbacks to the payment of CTs to adolescents. Adolescence is a unique developmental stage in which peer relationships play a significant role in one’s development and well-being [
29]. A sense of belonging is essential among adolescents, demonstrated by their heightened need for acceptance and social inclusion, and the desire to identify with and participate in events with their peers [
30]. Adolescents spend time together, discussing issues such as sex, alcohol, fashion, and hairstyles, and a significant component to peer acceptance is looking and dressing well [
31]. The decisions, behaviour and values of individual adolescents tend to be heavily influenced by peer group norms, for example, choosing the same clothes and hairstyles as their friends through a variety of mechanisms, including persuasion, information exchange, modelling, and social interactions [
32]. In this way, peers contribute to adolescents’ well-being and perception of themselves and have the potential to become a primary source of support [
33,
34].
This background raises several questions about cash payments directed to adolescents. What does the provision of CTs to a sub-group of adolescents mean for peer relations within the larger group? Specifically, what are the unplanned social consequences of cash payments to AGYW on their peer relationships? Answering these questions is critical. There is currently a limited understanding of how direct CT payments to AGYW affect their relationships with peers and others in the broader community. Recent studies that explored the CTs’ influence on adolescent recipients’ interpersonal relationships focused on intra-household relationships [
35], and on intimate and platonic relationships with members of the opposite sex [
36]. However, a pilot study of a CCT paid directly to AGYW prior to the full implementation of HPTN 068 found that social relationships between young women and their peers remained unchanged. The study showed limited negative impacts on social relationships, such as jealousy between recipients and non-recipients, although the duration of the study may have been too short to observe social impact over time [
37]. Furthermore, it was hypothesised that any negative effect might dissolve once people better understood the intervention.
Interventions that seek to improve sexual health outcomes in adolescents should also monitor and assess broader social and inter-personal impacts, which could mitigate or strengthen the impact of these interventions [
38]. To contribute to our understanding of the broader effects of CTs directly paid to AGYW on peer relationships and social norms, we analysed qualitative data from the HPTN 068 study (also known as Swa Koteka) in South Africa. This article advances findings from the HPTN 068 pilot study by examining how peer interactions of both female recipients and non-recipients were impacted over a period of three years.
Methods
Study setting
The study was located in Agincourt, South Africa, a sub-district of Bushbuckridge, Mpumalanga province, situated near South Africa’s border with Mozambique. This location is characterised by poverty, unemployment, poor infrastructure and temporary local migration, with family members of working age typically migrating to Johannesburg or nearby cities for work and visiting home infrequently [
39]. HIV prevalence in this area among those older than 15 years is estimated at 19.4%, with a prevalence of 23.9% among women and 10.6% among men [
40].
The HPTN 068 trial
We report findings from a qualitative study nested within the Swa Koteka trial. Swa Koteka was a Phase III randomised control trial assessing the effect on HIV incidence of CTs conditioned on 80% school attendance among South African young women (
n = 2448) aged 13–20 years. The trial provided a monthly CCT (R100 ≈ USD 10) to AGYW and their caregivers (parents/guardians) (R200 ≈ USD 20). They received cash deposited into their bank account every month during which they had met the school attendance criteria, for up to 3 years. At enrolment, participants underwent pre-test counselling, sample collection for HIV and herpes simplex virus (HSV)-2 testing, and post-test risk reduction counselling. They also completed an Audio Computer-Assisted Self-Interview (ACASI), a tool used to collect demographics and behavioural quantitative data. Participants had annual study follow-up visits at 12, 24, and 36 months until the study completion date or their planned high-school completion date, whichever came first. Follow up procedures were the same as enrolment [
41]. School attendance was collected from local schools monthly. The young women in the control arm of the trial underwent the same procedures, except for receipt of CCT. Social harms, such as experiences of violence due to receipt of CCT or participation in the study, were assessed at each follow-up visit by a counsellor (see Pettifor et al., 2016 for more details).
Qualitative study design and data collection
A qualitative study was nested within the trial to explore the acceptability of the intervention, the social and relational impact of CTs, and the meanings attached thereto.
Participant selection and recruitment
From a sample of 2448 AGYW in the trial, we purposively recruited a sub-sample of 39 AGYW for participation in up to six serial in-depth interviews (IDIs) conducted 6-monthly. We used purposive sampling to ensure that sexually active AGYW who reported missing school and those reporting engagement in transactional sex were represented in the sample [
35]. Using the trial contact database, potential participants were contacted telephonically and invited to participate in IDIs. A home visit was scheduled where additional information about the study was provided. Participants were asked to provide written informed assent and/or consent. Caregivers provided written informed consent for AGYW who were minors to participate in the qualitative study.
We also randomly selected approximately 120 HPTN 068 CT recipients and non-recipients, for participation in focus group discussions (FGDs). Those who had participated in the IDIs were not eligible to take part in the FGDs. This decision allowed for a larger overall sample and enabled us to capture a more diverse range of perspectives from AGYW in the trial Potential participants were contacted by telephone to invite them to participate in the planned FGDs. Home visits with interested participants were scheduled to obtain written assent and/or consent, as described above.
Data collection
IDIs and FGDs took place between 2012 and 2015. IDI participants were interviewed twice a year throughout their participation in the trial, providing verbal re-consent at each interview. The IDIs explored perceptions of the CCT programme, use and impact of CCT, exclusion from the programme, and social relationships in households, schools, and communities. All interviews were conducted at participants’ households. The duration of the interviews was 30–60 minutes.
Six FGDs were conducted annually for three years. FGDs explored participants’ thoughts about the CCTs and the impact of the cash on peer relationships among recipients and between recipients and non-recipients. The FGDs were conducted on weekends and during school holidays at local schools. Each FGD comprised 10–12 participants and lasted approximately 60 to 90 minutes.
The IDIs and FGDs were facilitated by four trained interviewers, all of whom were women, local residents, fluent in XiTsonga, and with a minimum Grade 12 education. The study was approved by the Human Research Ethics Committee of the University of the Witwatersrand, the Institutional Review Board of the University of North Carolina, and the Provincial Department of Basic Education in Mpumalanga Province.
Data analysis
Interviews were audio-recorded and subsequently transcribed and translated into English by the interviewers. Transcripts were checked for quality by one of the authors (MNK) through comparisons with the audio files. The transcripts were subsequently imported into Atlas.ti. Transcripts were coded using a coding frame that was both inductively and deductively generated by two of the authors (CM and MNK), following a framework approach [
42]. This approach entailed the generation of themes from
a priori hypotheses and questions that guided the objectives and aims of the study (deductive), and issues derived from the data (inductive). Coding was conducted by five trained researchers, who were also involved in data collection. Coders sought to attain acceptable intercoder reliability (ICR) scores on early coded transcripts [
43]. A random sample of 10% of the transcripts was double-coded throughout the study to ensure that ICR remained acceptable.
Discussion
Current studies on CTs geared towards AGYW for HIV prevention have primarily focused on education and HIV outcomes, with little attention paid to impacts on peer relationships. Yet CCTs for HIV prevention may have impacts beyond HIV outcomes; they potentially have spill-over implications for other aspects of recipients’ lives. This study contributes to the growing body of research that seeks to understand the broader social implications of CTs for HIV prevention.
Our findings demonstrate that CCTs provided to encourage school attendance with the primary goal to reduce HIV in AGYW also had unintended consequences for peer social relationships. From the recipients’ viewpoints, CCTs improved their social standing within their peer groups, facilitated peer identity, and promoted social connectedness among AGYW receiving the CCTs. Receipt of CCTs enabled AGYW to look and act like their peers who had money, and in so doing, to diminish visible signs of their poverty. This gave young women a sense of belonging derived from conformity to a set of positive – and shared – social norms. CCTs also facilitated social interactions, information sharing and instrumental social support between AGYW. CCT recipients experienced an increased ability to network, share, and reciprocate with others, which in turn potentially increased their social capital. Evidence shows that improved social connectedness among peers is significant. Social interaction, resource exchange practices, and reciprocity form an integral part of adolescent development and social capital generally [
44]. Social ties provide a sense of interdependence and belonging, which are often sought after by adolescents and associated with improved personal and social well-being [
44,
45].
Paradoxically, for the non-recipients, the CCT trial disrupted peer norms of ‘homogeneity’, with the recipients being resented for seeming to be ‘better’ than their peers who were non-CCT recipients. Giving CCTs to one group of young women and excluding others elicited negative emotions among non-recipients, and triggered some gossip, teasing, and rumours. However, these were slight disruptions overall and did not last long, with most young women reporting that they had maintained relations with their existing friends. Similar findings were observed in our pilot of HPTN 068, that showed that jealousy, teasing, gossip and tension should be anticipated in CCT programmes involving adolescents [
35].
Consistent with the literature, our findings illustrate how the psychosocial experiences of poverty were negative for young women. Being different from peers, exclusion, and failure to meet social expectations generated feelings of distress, indignity and humiliation. As a result, poverty and social exclusion could lead to AGYW engaging in negative coping strategies, including engaging in sexual risk behaviours [
46,
47]. In this context, CTs paid to vulnerable AGYW may reduce vulnerability by improving their economic well-being and social standing. In the context of HIV prevention, these results are significant as evidence shows that social connectedness to peers has crucial health implications for adolescent development [
48,
49]. Strong social connections with helpful peers and sharing of health information can support positive health behaviour and buffer health risk-taking [
48‐
50]. Studies have shown that interventions that build social ties are more likely to influence HIV related outcomes in the long run. However, this impact depends on the content and resources available from social connections [
47,
51,
52]. The current study findings underscore the building of social capital among peers as a potential pathway by which CCT could reduce HIV risk, depending on whether these interactions encourage negative or positive behaviour.
Surprisingly, cash recipients made instinctive associations between receipt of CCT and positive/ good behaviours. They shared narratives about shedding bad behaviour since the trial had started and distancing themselves from the friends of ‘bad’ influence; for example, friends who skipped classes or those who hung out in taverns. Recipients reported that instead they focused on their schoolwork and church, and spent the money wisely. However, it is unclear what informed this behaviour as trial organisers did not prescribe how AGYW should use the CCT, aside from encouraging school attendance. These findings suggest the trial and conditionality of the CT, school attendance, as well as other trial procedures such as HIV testing might have led to the CCT being interpreted as promoting ‘good behaviour’ in general. While these findings require further exploration, they highlight another possible way a CCT may reduce risky behaviour among AGYW recipients.
To a lesser extent, our findings suggest that provision of CCT to selected AGYW maintained or widened the gap of socio-economic inequality among peers – AGYW who were lacking before the CCT remained in that position because they were not selected to receive the CCT, yet peers who were thought to be better off or those who were at the same level with them (non-recipients) before the CCT, progressed. These findings indicate that CCT given to certain young women and excluding others could potentially negatively impact on HIV outcomes. In line with evidence on the role of peer pressure on HIV risk behaviours, a desire for respectability, and a desire to appear ‘rich’ among non-recipients has the potential to drive AGYW towards risky sexual behaviours in order to achieve parity with their peers who are CCT recipients. In the HPTN 068 trial, AGYW were allocated to receive CCT through a randomisation process. In a real-life context, some of these issues may be less significant as CTs are generally allocated according to need. However, this is not to say these issues may not come up even in a needs-based approach, as previous studies have reported unhappiness with how decisions are made, and cut-off points and criteria used to define the needy [
16,
17]. This may be an issue particularly in communities where most people may be considered poor, despite the criteria used to allocate CTs.
Findings from the current study highlight the critical role peers can play as a support structure for AGYW, and their potential role in both mitigating and fuelling social vulnerabilities associated with HIV risk. These findings have important implications for research and CT programming for HIV prevention that seeks to optimise the impact of CTs’ HIV-related outcomes. Our results point to the need for researchers to recognise existing social connectedness and the opportunity it poses for CT programming targeting AGYW for HIV prevention. They should seek ways to leverage peer interactions and social ties, and use these as platforms to promote sexual health, build critical consciousness, and modify harmful norms to maximise HIV impact. CCT programming targeting AGYW should monitor peer interactions and assess these interactions as a potential pathway CCT reduce HIV risk in AGYW. This could provide valuable evidence given the literature that shows that social connectedness to peers (mutual feelings of trust, reciprocity, shared norms and identity and information and resource sharing) has crucial health implications for adolescence development. Strong social connections with helpful peers and sharing of health information can support positive health behaviour and could buffer health risk-taking [
49,
53]. Cross-sectional studies have shown that interventions that build social ties were more likely to influence HIV related outcomes depending on the content and resources available from the social connections [
52,
54].
CT programmes targeting AGYW should incorporate peer and community engagement processes such as peer-led strategies and social mobilisation to leverage these social ties. CT programmes should include education that ties together the social and health implications of behaviours and social structures. This may enhance the impact on HIV outcomes, given the evidence that effective HIV prevention interventions also consider social networks which can influence behaviour, communication and norms [
55]. These findings are complementary to evidence that underscores the need for multiple combined interventions for HIV prevention [
56], and CTs programming provides a platform for this work to occur.
Overall, the idea of paying CCT directly to AGYW seems to have been instrumental in shoring up the positive outcomes of this intervention overall. Adolescent recipients made their own arrangements for collecting the CCT, and devised their own spending plans. There is a reason to believe that if cash was paid to and controlled by caregivers only, it would have been limiting for adolescents. AGYW would not have had a chance to accompany one another to the post office to collect their CCT or engage in activities that allowed interactions between them. The findings indicate how paying CCT directly to AGYW contributes to their personal and social well-being.
One important contribution of this study is that it captured the voices of the non-recipients, which are often missing from studies of CCTs. These voices are significant as they represent the views of the broader community on the impact of CCTs to a specific group and illustrate something of the potential unintended consequences of this intervention. In our study, non-recipients’ concerns were expressed through negative emotions such as jealousy, as well as stigma, and rumours. Considering the evidence that shows the association between social networking, peer pressure, and HIV risk [
47,
53], CCT programmes need to find ways to monitor and mitigate the risk. We, therefore, suggest that in poverty-stricken communities, CCTs should be provided to all AGYW in a setting, and not only to a selected few. In South Africa, universal targeting used in social grants has proven to limit unintended adverse effects on community relationships [
18]. In instances where CCTs are offered to selected AGYW, the eligibility criteria should be clear, as imprecise and random eligibility criteria could lead to unintended negative consequences [
18].
This study was not without limitations. We only gathered perspectives of AGYW who were enrolled in the trial. Including voices of peers outside of the trial would have brought additional insights about the CCT trial and the experience of exclusion. The perspectives of male peers are also missing from this analysis, but are reported elsewhere [
36]. Furthermore, the analysis did not assess interactions over time; consequently, it is not possible to say whether interactions continued beyond the timeframe of the intervention. Lastly, while it is evident that some positive impacts resulted from receipt of CCT, it is also possible that some of this positive impact was influence by trial participation than CCT impact; however, positive impact was largely reported by CCT recipients as opposed to non-recipients which suggests CCT contributed to positive impacts.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.