30 July 2024
SASSA says its biometric verification system for the Social Relief of Distress grant is designed to prevent fraud. But it is irrational to expect applicants who only qualify if they live on less than R624 a month, to have access to a smartphone and the internet, say the writers. Graphic: Lisa Nelson
New barriers in the Social Relief of Distress (SRD) grant system have prevented people from receiving the grants to which they are entitled.
The SA Social Security Agency (SASSA) says the changes are to prevent fraud but has provided no evidence of systematic fraud. In fact, the grant is only reaching about half of those eligible, suggesting that more people are being excluded in error than are being included in error.
Late in June, only a week before payment dates, many beneficiaries of the R370-a-month SRD grant received notifications from SASSA that they had been placed on “referred status” and needed to complete a “biometric identity verification” in order to get their grant.
As a result, many of these beneficiaries did not receive their grants for June. SASSA gave no notice and no prior explanation for this new practice. Afterwards it emerged that people who received this message were suspected of having committed, or been the victim of, “fraud”.
In order to reactivate their grants, beneficiaries who were “referred” needed to provide a high-quality photograph of themselves—via a multi-step online process—which can be matched to a photograph held in government databases by facial recognition software. This appears straightforward enough, but for many, it is an insurmountable obstacle.
People need to be living on less than R624 a month in order to qualify for the SRD grant. This is well below the R760 a month level determined by Statistics SA as necessary just to cover basic food needs, without having any money left over for other essentials including rent, electricity and transport. However, in order to navigate biometric identification, “referred” beneficiaries need to have a very high level of online connectivity. They must have access to a smartphone with a high-quality camera, a stable internet connection, their own contact number on which they can continuously receive communications, and advanced digital and English literacy. It is beyond irrational to expect a group that by the government’s own criteria cannot afford food, let alone other expenses, to have a smartphone and internet access.
We know that many SRD grant applicants need to borrow other people’s phones and numbers to apply for the grant. This creates a serious problem when SASSA sends them a biometric verification link, or a request for further information, as they often do not receive it. However, the government has recently given itself the right, in an amendment to the SRD regulations, to cancel the grants of beneficiaries who are deemed “untraceable”. This is deeply unfair, because it is precisely the government’s own decision to make the system reliant on smartphones and internet access that results in untraceability.
SASSA has recently begun taking steps to address this specific issue, by launching a pilot scheme in Nelson Mandela Bay, in which beneficiaries can go to SASSA offices to access wifi or laptops with webcams, in order to complete biometric verification. While we welcome this action, it remains very limited at this stage.
Compounding the problem is that SASSA has now stated that anyone wishing to change or update their contact or banking details must also undergo biometric verification. This means that people who borrowed phones and numbers in order to apply for the grant, must have access to the phone they borrowed in order to receive a link, as well as to a number of their own in order to complete the verification process and remain traceable.
People who used the now defunct cash or Postbank payment options also have to undergo biometric verification in order to provide banking details to receive payment. Many of them have had to open a bank account and pay bank fees. Those affected by these changes are the most vulnerable of an extremely vulnerable group, or those least likely to have the resources to navigate the biometric verification system.
What evidence does SASSA have to back up its claims that there is fraud? Well, a mysterious “risk score”, seemingly determined by an algorithm, based on data provided by what SASSA calls “industry stakeholders” (like banks). We do not know what data is being used or how the level of risk is calculated.
Overseas, scandals have followed the use of opaque algorithmic systems to profile welfare recipients. In 2021 the city of Rotterdam was forced to pause the use of a machine-learning system that classified and ranked welfare beneficiaries for fraud risk. It was revealed that factors including being a parent, a woman, young, not fluent in Dutch, or struggling to find work, all contributed to a higher risk score.
Also in the Netherlands, thousands of lives were ruined, after the Dutch Tax Authority wrongly accused 20,000 families of committing childcare benefits fraud, based on an algorithm that assessed risk factors some of which seemed very arbitrary. These included holding a dual nationality or having a low income. One thousand children were taken into state care.
In Australia, a successful class action lawsuit stopped the use of an automated system which accused hundreds of thousands of welfare beneficiaries (in many cases wrongfully) of under-reporting their income, on which basis they were deemed to be in debt to the state, and sometimes handed over to private debt collectors. In Jordan, a World Bank-backed algorithm aimed at targeting welfare grants to the poorest households excluded many people in need.
These cases must provide important lessons for South Africa on the risks inherent in an uncritical use of automated decision-making.
In the case of the SRD grant, providing data about beneficiaries to third parties to calculate fraud risk also raises red flags. The SRD grant regulations essentially allow for the unchecked sharing of information about applicants, who must consent to the sharing of “any information” about them with “any government or private institution considered necessary.” This calls into question the extent to which SRD grant beneficiaries are protected by the Protection of Personal Information Act (POPIA), which protects people’s personal information from being abused by companies and government agencies.
It does not appear that beneficiaries are informed about what kind of information is being shared about them, or whether they have the ability to request, or challenge this information. This introduces yet more risks of undue surveillance, as well as inaccuracy and a lack of accountability in decisions affecting peoples’ basic rights and dignity.
It is also unclear whether private companies benefit in some way from the surveillance and handling of data, and what safeguards are in place to protect beneficiaries from predatory practices. Concerns have been raised in the past about the involvement of private actors in the provision of social grants. Sensitive banking or personal information may be shared between the private sector and the government, with little transparency about how it is managed and stored, who can use it, and for what purpose.
In addition, there is a danger that profit-driven companies may use their access to information about beneficiaries to target them for promotions or offer advances on grants. The Net1 scandal in 2018 provided a shocking example of this, when a company contracted by the government to manage the payment of social grants used its access to target beneficiaries for predatory loans, using their social grants as collateral. Evidence collected by Black Sash—which is a member of the Universal Basic Income Coalition— and its partners played an important role in revealing this scandal.
Finally, of serious concern is that the use of facial recognition software for biometric verification discriminates against holders of green ID books. Officially, the smart ID simply makes the process easier but in fact that the biometric verification process effectively does not work for holders of green ID books. This seems to be because Home Affairs is unlikely to be able to automatically match them to a photo in their database—seemingly because the photos are too old.
This is part of the government’s ongoing migration to a digital ID system. But the SASSA biometric verification system is an example of the government essentially forcing people to migrate to the smart ID card by stealth, in order to retain access to their basic rights. In Uganda, the government is being sued for a very similar policy which, according to its civil society opponents, has excluded over a third of the population from life-saving benefits. The High Court is being asked to order the government to allow alternative forms of identification.
This is because, in Uganda, as in South Africa, there are barriers to obtaining a digital ID or smart ID card, which disproportionately impact the most disadvantaged. Getting a smart ID in South Africa is costly. The application fee is R140, then there’s the cost of transport, and the time spent queueing, and waiting for the card to be issued. These costs and wait periods have dire impacts on peoples’ access to social grants. Even if they are able to apply, delays in processing mean beneficiaries may miss out on their grant for one or more months, or even have their grants suspended.
We are concerned that the SRD grant may be being used as a testing ground for digitisation of other grants. At the moment, the SRD grant is managed differently to all other SASSA grants. While other grant applicants can apply at SASSA offices, the application system for the SRD grant is exclusively online. While other grant applicants can their income in an affidavit, SRD grant applicants are subjected to (inaccurate) automated bank checks each month. We believe that ultimately government plans to roll out more digitalisation, bank surveillance and biometrics across the grants system, to other vulnerable groups—such as caregivers of children, older persons and those with disabilities.
The new Minister of Social Development has stated that a policy to introduce basic income support will be tabled in Cabinet in this financial year. This is long overdue, and we eagerly await the opportunity to review and comment on the policy. However, if the basic income grant is to meet the objectives of tackling hunger and economic exclusion, and build on the base of the SRD grant as the government has repeatedly indicated, the Minister must take urgent action to dismantle the barriers which block access to the grant.
We urge government to redirect the resources and effort being spent on stopping people from getting the grant, towards enabling the grant to do what it is meant to do - protecting people from starvation.
Kelle Dawson, from the Institute for Economic Justice, Hoodah Abrahams-Fayker from the Black Sash, Vayda Megannon, from The Family Care-giving Programme at UCT, and General Alfred Moyo from Pay The Grants, write on behalf of the Universal Basic Income Coalition.