With an unemployment rate of 33% and more than 25% of its residents living at the extreme poverty level, South Africa maintains the title of “most unequal country in the world.” In the wake of economic turmoil catapulted by the coronavirus pandemic, rural-dwelling South Africans increasingly depend on government-issued social grants for survival.

However, significant programs like the Social Relief of Distress (SRD) R370, a grant in which eligible South African citizens and refugees receive a stipend of $21.2 per month, are increasingly administered by grant algorithms that deny qualified recipients grants for basic goods and welfare accommodations.

Inception of the Social Relief of Distress Grant

In a first attempt to remedy this, the South African Social Security Agency (SASSA) responded to the pandemic’s tumultuous economic conditions by launching an Economic Reconstruction and Recovery plan. The plan included measures to address the social distress caused by the pandemic.

By 2023, social grants served as the primary form of income for working-age South Africans living at the extreme poverty level, individuals who would otherwise collect a maximum of $40 per month in means. The stipend is available to any unemployed South African of working age. However, many eligible residents encounter demographic obstacles that make acquisition nearly impossible.

Barriers To Grant Approval

For one, the platform, which largely benefits inhabitants of rural areas populated by native speakers, is strictly available in English. Applications are only received digitally, requiring impoverished individuals to acquire a device for submission.

To receive the SRD grant, an individual’s maximum income may not exceed 624 ZAR ($35.21), significantly below the food poverty threshold. The SRD grant is administered by a digital grant algorithm that scans bank accounts and flags any indication of income.

The Automated Means Test

This process, known as the Automated Means Test, rejects qualified applicants whose accounts possess erroneous means that may not reflect sustainable income. One such miscontextualization of funds perpetually influences Brenda Mtshali, a widow and tomato vendor who scarcely makes enough to support her six children.

In her case, the grant algorithm disqualified her from benefaction, falsely arguing that her account contained a loan. In addition to this invasion of personal and financial data, the detection software misinterprets miscellaneous funds as “means” that exceed an already underrepresentative poverty threshold.

Since the inception of the Automated Means Test, the number of grant beneficiaries has decreased from 10.9 million to eight million, despite an existing eligibility pool of approximately 17 to 18 million people. In analyzing this disparity, the Institute for Economic Justice (IEJ) identified an exclusion rate of 89.7%. Whereas failures on the Automated Means Test cause millions of rejections, the IEJ concluded that only 24% of cases should be eliminated in ethical circumstances.

Benefits of the Social Relief of Distress Grant

On the contrary, the beneficiaries of the stipend report significant increases in quality of living. “Mind the People,” a short film directed by Mozilla Africa Mradi (and available for streaming on YouTube), dissects the disparate qualification process by collecting testimony from individuals who reside in the rural Mountain View and Eldorado regions of South Africa.

Ntombizodwa, an individual who testifies in Mradi’s film, describes how the SRD grant allows her to live a “much better life,” providing access to toiletries, food and electricity that she previously lacked. Nonetheless, significant improvements in accessibility must be made to improve the efficacy of the program.

Researcher Response to Algorithmic Inequity

According to the IEJ Report, applicants should be permitted to submit documents supporting their petition for a grant and that means tests should be conducted over a longitudinal period to eliminate algorithmic error. To improve accessibility, the report suggests shifting to a hybrid and multilingual application model.

A 2024 document published by the European Union–Agence Française de Développement (EU-AFD) Research Facility on Inequalities proposed a new structure to improve the efficacy and sustainability of the former SRD model. The document recommended that SASSA use self-reported data to assess eligibility, increase grant amounts and establish grant permanence.

Conclusion

Ethical access to social grants in South Africa is not only an issue of socioeconomic disparity, but also a matter of social equity and justice. Whereas recent projects have advanced critical conversation about broken South African grant algorithms, many people remain excluded from the precise benefits that might salvage them from tragedy. 

– Talia Gitlin

Talia is based in Natick, MA, USA and focuses on Good News for The Borgen Project.

Photo: Flickr