South Asia’s unprotected poor: a systematic review of why social protection programs fail to reach their potential

Structural Flaws in Program Theory

Our review indicates that most programs had flawed program theory which outlines how an initiative will produce its intended outcomes (i.e., the program design) [31]. Program theories frequently rely on oversimplified linear models of cause and effect and presuppose that their one-dimensional initiatives will produce a chain of consequential impacts. They fail to consider the multi-dimensionality of poverty and how it is rooted in the broader social, economic, and political structures that adversely shape incorporation into society. This is best exemplified by Bangladesh’s Challenging the Frontiers of Poverty Reduction: Targeting the Ultra-Poor initiative, which aimed to redirect the destitute from low-skill occupations through education, income generation training, and microfinance loan services [32,33]. However, evaluations of the program showed that the landless households experienced negligible gains from participation [34]. Entrepreneurial activity increased in the short- and medium-term following program graduation, but the results were not sustainable in the long term [33,35]. Most beneficiaries returned to their baseline occupations, often as beggars and maids [33,35]. We argue this happened because the program’s design assumed that poverty is transient and that a one-time offer of micro-entrepreneurship support was sufficient to matriculate from penury [36]. It ignored the multiple forces operating to keep the poor poor, including their systematic exclusion from markets and the social relationships between and within networks that underpin entrepreneurial activities.

Another example of program theory deficiencies is the Nepal Food Corporation, which provides subsidised rice for the poor. The program required the beneficiaries collect the rice from central depots, which are located at significant distances from the remote villages where the target beneficiaries live [37]. Since the destitute lack the financial capital to travel to the central depots, the gap in service is filled by individuals with the means to purchase and transport large quantities of grains for the entire village. These middlemen sell the government-subsidized rice for a hefty profit - citing ‘transportation costs’ - forcing the poor to purchase their entitled rice quotas through high-interest credit. Consequently, the poor, low-caste Dalits only have 2% of their food needs met through the social protection program [37]. This crucial flaw in program design minimizes the program’s impact on food insecurity, instead nurturing profits for the rich and poverty for the poor.

Other program theories held built-in assumptions that inherently required that beneficiaries have a certain degree of wealth. The Community Midwife Program in Pakistan serves as an illustrative example. This program aimed to ensure skilled birth attendance for impoverished rural women. However, its placement within the private sector proved to be a fundamental flaw [38]. Cognizant of the dissonance between providing a service for the poor, while locating the midwives in the private sector, program managers suggested midwives charge a modest fee of USD 0.66 USD for childbirth. Yet, to maintain their practices, the midwives needed to charge fees 10 to 20 times higher than recommended. This private-sector-based approach ultimately failed to meet its objectives, imposing unrealistic financial burdens on both the impoverished women it aimed to serve and the midwives themselves. Over-estimation of the poor’s assets was also evident in Nepal’s Food-for-Work initiative, which has three-fold objectives: (1) improve agricultural output; (2) build road infrastructure to increase market accessibility; (3) provide poor households with employment opportunities by working on the farms or building infrastructure, for which they are paid in-kind with rice [37]. In the district of Humla, the program only benefitted the high-caste, rich landowners since they got cheap agricultural labour and road access from their lands to markets. Meanwhile, the landless, chronically poor - the key target group - got nothing more than an insufficient amount of rice, meeting only 15% of a household’s food requirements [37]. The initiative is essentially futile and trivial for the poor, for it failed to achieve its stated objectives of reducing food insecurity and poverty.

Elite Capture of Social Protection Programs

The second theme emerging in our review was the concentration of power in the hands of the South Asian elite and how this impeded program reach to the poorest of the poor. In this context, the denomination of the elite is based on their status as high-caste, with its associated opportunities, privileges, and entitlements [37,39–42]. This includes land ownership, wealth, and membership in networks that sustain their dominant social and political identities.

Direct appropriation of benefits by the elite

The first pattern of elite capture is a traditional form of corruption in which the non-poor participate in welfare programs meant for the poor. For instance, in their sample, Sinha [43] found that 47% of households enrolled in India’s Rashtriya Swasthya Bima Yojana health insurance scheme were non-poor. In Nepal and Bangladesh, the elite captured, or in simpler terms stole, subsidized grains meant for the food-insecure poor. Through secret transactions with program managers in Nepal’s Food Corporation program, large quantities of rice meant for the poor were diverted to local hotel owners to brew alcohol [37]. Bangladesh’s Food-for-Education program encountered a similar appropriation of grains. Despite providing medium-quality rice to disincentivize wealthy households, the rich still captured the resource - not for their consumption, but rather as a payment source to their poor servants [44]. The landed rich also captured India’s Mahatma Gandhi National Rural Employment Guarantee Scheme, a program designed to employ the rural poor. Those in the highest wealth quintile not only took over the employment opportunities meant for the landless, the illiterate, and the poor, but they also earned a higher daily wage [39,41,45–48]. Over 25% of poor households who sought work did not receive it despite being ‘guaranteed’ 100 days of work [45].

Elite’s capture of resources through their power over program implementation

The elite also capture resources via the power they hold as social protection program implementers. Most program managers are the local elite since they meet the educational requirements and have the social capital needed to access such job opportunities. This role gives the traditional, informally powerful elite formal power. It simultaneously situates the poor at the mercy of the very people responsible for their subordination. In Nepal, personnel responsible for distributing cash transfers to the poor under the Dalit Child Grant delayed payments, instead redirecting funds into their personal accounts to collect interest [49]. Likewise, the elite captured the implementation of India’s Mahatma Gandhi National Rural Employment Guarantee Scheme. The program aimed to build village infrastructure and simultaneously provide employment opportunities for rural men and women [41,42]. The program planners designated the Gram Panchayat, a village-level governing council, as the implementing body responsible for all program tasks, including employment of labourers and allocation of work hours [41,42]. Gram Panchayat are traditionally made up of land affluent dominant classes. Many evaluations noted that the Gram Panchayat members captured the government-funded labour sources for their personal farms and other private projects [41,42]. For example, Pattenden [41] found that in the village of Panchnagaram, the appropriation of labour led to a third of the local program funding being allocated towards projects on the private lands of the Gram Panchayat members. These elites also disincentivized the poor from applying to the program by delaying and withholding wages, sharing misinformation, requiring hidden fees, hiring through nepotism, and, in some cases, preventing registration by withholding applications [39,41,42,46,48]. Ultimately, rather than reducing poverty, the program reproduced and even strengthened the dynamics of traditional inequity [41].

The necessity of elite patronage & their gatekeeping role

The literature highlights how the elite often act as gatekeepers, controlling access to social protection for the poor. Here lies the catch-22 of social protection: it is supposed to offer an opportunity for the destitute to lower their dependency and potentially disengage from abusive relationships with the elite, but program design ensures it is precisely these relationships that the poor require to access the programs. Ideally, program accessibility should be based on state-citizen social contracts, but in South Asia, this ideal reigns inferior to the informal contracts with local elites [37,39]. The politically and socially powerful groups dictate the distribution of resources based on caste and biraderi (kin) relationships [38,39,50–52]. Due to their high decision-making entitlement and low accountability, they act as the poor’s primary route to access social protection [39,53]. In India, these bonds between the powerful and poor are so integral to societal structure that the practice of their formation has been given the special name of bhav-vyavhar, or the emotional and behavioural relationship building with the elite [39]. Bhav-vyavhar occurs through social transactions such as the poor offering free physical labour, emotional support, and even financial transfers, including a portion of their pension.

In India and Pakistan, some program designs allocated the task of identifying poor households to municipal-level officials and the village elite (often the same group). These officials registered their non-poor family and friends instead of the destitute [54,55]. In India, the poor perceived India’s National Old Age Pension Scheme as unequal and inaccessible since they lacked the social linkages necessary for participation in the program [56]. This is best reflected in the finding that 74% of deserving recipients were not receiving the Pension [56]. Gatekeeping was also operative in Pakistan’s Zakat initiative, in which Local Zakat Committees, comprised of the community’s well-off, are responsible for identifying the poor and eligible households [57]. An evaluation by Sajid et al. [58] showed that in 2019, 40% of the beneficiaries were non-poor.

A powerful tool is the elite’s control of knowledge about social programs’ existence and benefits. Most program designs place the responsibility of disseminating information in the hands of the elite, like the Local Zakat Committees in Pakistan. An evaluation [58] found that the Committees were withholding this knowledge as 80% of eligible beneficiaries in their sample were unaware of the benefits they were entitled to. Das et al. [46] also reported censorship of information in their evaluation of India’s Mahatma Gandhi National Rural Employment Guarantee Scheme. The elite forming the Gram Panchayats withheld information to ensure local villagers remained unaware of the basic guidelines and operations of the program. Consequently, the poor beneficiaries were underpaid and underemployed.

Targeting Errors

The third theme in our review was the insensitivity of program design when targeting beneficiaries. The literature identified three targeting strategies traditionally used to reach recipients: geographic targeting, self-targeting, and data-based targeting. Geographic targeting purposively selects regions based on poverty or health indicators. Self-targeting (or self-selection) places the onus of identifying eligibility on the poor themselves. Data-based targeting relies on a pre-existing national dataset, such as India’s Below Poverty Line classification. Our review showed that all three mechanisms failed to target their ideal beneficiaries sufficiently and were instead more likely to reach undeserving households.

The Janani Suraksha Yojana program in India illustrates the insensitivity of geographic targeting. This initiative specifically focused on states with low rates of institutional deliveries, introducing a cash incentive scheme to encourage women to deliver in healthcare facilities. [59]. An evaluation in nine participating states revealed that all socioeconomic groups experienced a similar average increase in institutional birth. However, the reduction in maternal mortality was four times greater for the richest quintile compared to the poorest [59]. Thus, despite a principal program objective being to lessen financial barriers for poor women, the broad geographic targeting mechanism disproportionately benefited the rich living in these localities.

India’s Mahatma Gandhi National Rural Employment Guarantee Scheme serves as a prime example of self-targeting. The program provides employment opportunities for unskilled labourers [47]. An evaluation [47] showed self-targeting permitted uptake by all socioeconomic groups, irrespective of need. In the state of Tamil Nadu, only 40% of the participants were truly poor. Likewise, Pakistan’s Livelihood Support Cash Grant implemented a similarly ineffective form of self-targeting. Applications for the Grant were open to all, but applicants were subject to a screening process based on the government’s predetermined eligibility criteria [52]. Although the program enrolled nearly 700,000 families as beneficiaries, a community-based survey showed that half of the eligible families were excluded, while half of the ineligible rich were provided benefits [52].

A widely recognized static targeting strategy in India is based on classifying households as Below or Above Poverty Line through a state-level census. This classification is not infallible as an evaluation found that 70% of ineligible households in the state of Karnataka possessed Below Poverty Line cards, while 13% of eligible households did not [55]. The downstream consequences of this targeting flaw become evident in various social programs’ failure to reach the poor [55,60]. One example is the failure of Swarnajayanti Gram Swarozgar Yojana microfinance program for the destitute. An evaluation [61] found that none of the beneficiaries in their sample were extremely poor, while over half of the program participants were from high-income households falsely possessing Below Poverty Line cards. Another example is the National Old Age Pension Scheme, in which non-poor Below Poverty Line cardholders had a higher likelihood of receiving the social allowance than poor cardholders [56]. In fact, Mishra & Kar [62] found that 50% of the Pension recipients were from the top two income quartiles. A third example can be found in India’s Targeted Public Distribution System, a food ration initiative. Five evaluations [50,54,60,63,64] found that a large fraction of the wrongly classified and wealthy Below Poverty Line cardholders were receiving food rations, while a substantial proportion of the poorest households were not covered by the program. These are worrying findings given that India’s poor lead a dangerous “hand-to-mouth existence” [50].

Invisible and Unacknowledged Barriers

Program designs failed to take into account a host of barriers handicapping the poor’s accessibility to social protection. These include gendered barriers, hidden costs, the poor’s lack of legal documentation, and the geographies of exclusion. We discuss each below.

Gendered barriers

Despite decades of advocacy, programs remain gender blind. In Pakistan, programs like the Lady Health Worker and Community Midwife are specifically designed to deliver domiciliary health care services to home-bound women in direct acknowledgment of the cultural practice of pardah (seclusion), and its’ associated value of male honor, limiting women’s unaccompanied movement [38,65,66]. The same gender values also require that the health workers be female.

Consequently, only women are recruited as Lady Health Workers and Midwives to provide home-based services. In doing so, the program design paradoxically expects the female health workers to transgress the very gender norms that created the need for their services in the first place. A reading of the papers suggests the program designers were aware of women’s gendered mobility restrictions but chose to ignore the implications of this reality for the female community health workers. This insensitivity and lack of consideration for the young Lady Health Workers and Midwives’ safety and respectability discouraged the workers from assisting households beyond a short walking distance or within their social geography [38,65,66]. The most vulnerable women remained underserved.

Similarly, in Nepal, a food security intervention reproduced gendered inequalities by erring in assuming female labour was a freely available commodity [67]. The program activities added to women’s existing time-consuming socially ascribed domestic duties. This led to female time poverty and subsequent selectivity in program participation [67]. The most effective development strategies were those that relieved gendered labour burdens through services such as the provision of school lunches [67].

Hidden costs for the poor

Another crucial structural flaw in program design is the hidden costs required to access the programs. One such cost is the time needed to travel and fulfill the often-onerous bureaucratic requirements, which often results in the loss of a full day’s wage [53,68]. An evaluation of India’s Targeted Public Distribution System in the sub-district of Sasan found that 64% of households without ration cards did not even apply out of concern that administrative delays would interfere with work schedules and subsequent earnings [64]. Given the poor’s fragility of income, these opportunity costs can considerably compound the depths of their poverty - resulting in a net financial loss rather than a benefit.

The assumption that time is trivial for the destitute was also exemplified in the design flaws within Pakistan’s Livelihood Support Cash Grants program [52]. To find out if they had been accepted as beneficiaries, poverty-stricken victims of a natural disaster were expected to make regular trips to district offices. If denied, families could only contest their eligibility status within the first seven days following the publication of recipient lists, which were released on an unpredictable schedule. Without regular travel to district offices, poor households could miss the opportunity to appeal - likely a contributing factor to the finding that half of the eligible beneficiaries were not reached by the program [52].

Another hidden cost is the solicitation of bribes. An overwhelmingly common phenomenon within South Asian social protection, corruption is often legitimized under the guise of registration fees and gratuities [39,46,50,69–71]. Taking the example of Pakistan’s Zakat cash transfers, the monetary value is reduced as beneficiaries highlight multiple hidden transactional fees from registration officers, peons, and even postmen [57]. In other cases, applicants were completely unaware that they were paying bribes - believing it to be a requisite for program participation [49].

Lack of legal documentation and literacy

One common but truly invisible barrier was the poor’s exclusion from documentation of their legal citizenship. Program designs nearly always require such documentation, often an ID card or birth certificate [50,53,56,57,71]. The poor lack these documents for a range of reasons including but not limited to the lack of a permanent address [72], lack of knowledge of the importance of documentation [73], the cost of obtaining documentation [73–75], internal displacement [76], and generational systemic exclusion due to gender [73,74] or caste [75]. These groups contribute to the ‘Invisible Billion’ worldwide who cannot prove their legal identity [74]. In India, the poor’s lack of appropriate documentation meant they could not apply for Below Poverty Line ration cards, which were needed to prove poverty status when registering for the Chief Minister’s Comprehensive Health Insurance Scheme [53]. Some applicants were also excluded from the country’s Rashtriya Swasthya Bima Yojana health insurance as their presented records did not match those of the company handling enrollment [43]. For Pakistan’s Zakat program, these requirements created a black market for illegal birth certificates [57].

Social protection programs also faltered in accommodating the close relationship between poverty and low literacy in rural South Asia [46,52]. Program resources were provided in written format, often in English. The poor’s subsequent inability to absorb the information reified relationships of dependency on the local elite and government staff [46,52].

Geography of exclusion

Geographic exclusion encompasses social and physical dimensions, an aspect that social protection programs often overlook in program designs. Despite its seemingly obvious role in accessibility, eight initiatives failed to consider how the lack of transportation and long distances to access points might limit uptake by the poor [34,38,43,53,68,71,77]. The program designs of India’s Targeted Public Distribution System and the Nepal Food Corporation required the destitute to travel to far, often centrally-located, ration shops at their own expense, incurring high transaction costs [37,54,63,64]. In Nepal, Gautam & Andersen [37] found that households with limited mobility resources would spend days walking through harsh terrain to procure their grain rations. Their travel burden was worsened by the uncertainty of whether or not depots would be open or even be capable of meeting their quotas. The program design ultimately penalized the most marginalized, the poor living in remote villages.

Social geography’s less obvious (but critical) role was overlooked in the design of Pakistan’s Lady Health Worker program [65]. Women living in bhattas (brick kilns), encompassing the poorest of the poor populations, were unreached by the program because the bhattas were considered no-go zones and not a part of human settlement. They were, consequently, beyond the borders of the workers’ catchment areas [70]. There was also the perception that it was “demeaning” to enter bhattas [70].

Gateways to Self-Exclusion by the Poor

Our final theme is the poor choosing to self-exclude from participating in social protection programs. We identified three possible gateways for this process: a desire to maintain dignity, a lack of capital, and a perception of programs as substandard.

The first form of self-exclusion is a desire to maintain a sense of dignity. This is best described by Axel Honneth [78], who says, “We owe our integrity… to the receipt of approval or recognition from other persons.” Nair [54] introduces the idea of aversion to social protection via the demonstration effect in which poor households emulate the patterns of the economically advantaged. This occurred following India’s Public Distribution System’s transition from a universal to a targeted program. As Above Poverty Line households were no longer eligible, low prestige was assigned to those who continued participating in the ration system. Participation stigma was also evident in India’s Mahatma Gandhi National Rural Employment Guarantee Scheme, in which both tribal and non-tribal families considered the work as below their dignity [51]. Beneficiaries of Pakistan’s Zakat program complained about the demeaning application process; they felt the government treated them like “beggars” to receive a mere 200 PKR a month [57]. The gravity of aid-affiliated shame is further demonstrated by a village in northern Punjab, Pakistan, in which some poor low-caste women actually preferred to remain invisible and untargeted by social protection to avoid attracting unnecessary attention to their already disgraced identities [70].

The second form of self-exclusion results from the poor lacking the financial, social, and political capital often required to access the resources ostensibly for them. A poverty-stricken household’s awareness of its externally - and sometimes internally - imposed limitations can restrict it from even attempting to access social protection programs. This is evident in the poor’s hesitancy to participate in India’s Swarnajayanti Gram Swarozgar Yojana microfinance initiative due to their fear of being unable to repay their loans [61]. Lack of capital also translated into the sometimes insurmountable challenges of seeking sunvaiyi (justice) from government officials [39]. Although the poor had warranted complaints of elite capture, receipt of justice required financial input and high social status to ensure that their grievances would not be disregarded.

The last gateway to self-exclusion is a consequence of the perception that programs are of low quality. For example, despite recognizing the benefits of skilled maternal care, Pakistan’s low-caste women continued to seek unskilled birth attendants to avoid experiencing inferior treatment - mitigating the value of government-funded healthcare [70]. In India’s Chief Minister’s Comprehensive Health Insurance Scheme, access to funds took up to 25 days, which delayed treatment even during emergencies [53]. Poor women instead subjected themselves to out-of-pocket payments to receive immediate and poor-quality care. Villagers’ perceptions of programs were also guided by bad experiences of seeking care shared within the community, thus deterring some ideal beneficiaries, and fabricating a form of collective exclusion [68].