Most Social Grant Fraud Involves Officials, Not Beneficiaries, Report Says

Researchers warn that automated checks and inaccurate databases can lead to erroneous grant suspensions and cancellations.

• An analysis by the Institute for Economic Justice of reported cases of social grant fraud between 2014 and 2025 has found that about 75% involved government officials, employees or contractors.
• Yet recent anti-fraud measures by SASSA and the National Treasury have focused heavily on beneficiaries, the IEJ says.
• Researchers warn that automated checks and inaccurate databases can wrongly flag beneficiaries, and lead to grant suspensions or cancellations.
• Civil society groups have also raised concerns about outsourcing grant systems to private companies, which can give them access to sensitive beneficiary data.

In an effort to clamp down on fraud in the social grant system, the government is vetting beneficiaries. But according to recent analysis by the Institute for Economic Justice (IEJ), most of the fraud cases which have been reported publicly have involved officials and contractors.

The IEJ argues that the government’s anti-fraud measures risk unfairly excluding vulnerable people from the grants they rely on.

The analysis, which was discussed during a webinar hosted by the IEJ, drew on public data, government reports, news coverage and accounts from beneficiaries affected by fraud investigations.

The researchers found that between 2014 and 2025, about 75% of social grant fraud cases reported for investigation by the police – 2,658 in total – involved government employees, contractors or officials. Over the same period, only 590 beneficiaries were reported for fraud, about 54 people per year. These numbers are based on information supplied to Parliament by the DSD and the SA Social Security Agency (SASSA).

“In light of the fact that in 2025 approximately 28-million people received social grants, our factsheet argues that the resources being dedicated to policing beneficiary fraud are hugely disproportionate to the problem, and, especially given their adverse consequences, unjustifiable,” said IEJ researcher Dr Kelle Howson.

Recent anti-fraud efforts have focused on tightening verification processes for beneficiaries. New measures include stricter surveillance, biometric checks, bank income verification and large-scale database cross-checks. Beneficiaries flagged by these systems may be required to undergo in-person reviews at SASSA offices.

These measures form part of new conditions attached to SASSA’s budget by the National Treasury. The agency must now submit quarterly reports detailing how many grants have been reviewed, suspended or cancelled, and what savings have been generated.

In his budget speech, finance minister Enoch Godongwana said intensified fraud detection measures had led to the review of about 292,000 grants, with around 34,600 cancelled.

But researchers say a cancelled grant should not automatically be considered fraudulent.

“Fraud is used quite broadly to refer to ineligible people accessing social grants. But it’s a specific legal term,” Howson said. “To constitute fraud within the legal definition, there must be misrepresentation or deceit.”

She said this requires someone to knowingly misrepresent their circumstances in order to gain financially.

Beneficiaries flagged by automated systems must often travel to SASSA offices for verification, which can be difficult for elderly people, those with disabilities, or people living far from service centres.

“If you do show up and you’re found to not be eligible for a grant because you’re above the threshold, that doesn’t mean you’ve committed fraud,” Howson said.

The analysis suggests many grants are cancelled because beneficiaries fail to complete review processes, rather than because fraud has been proved.

The IEJ also raised concerns about the accuracy of the digital systems used to detect suspected fraud. Government databases are often outdated or inaccurate, they said, while bank checks cannot distinguish between different categories of money moving through an account, such as child maintenance payments, gifts or loans. These systems can therefore incorrectly flag beneficiaries as suspicious, Howson said.

Beyond beneficiary fraud, researchers highlighted other forms of wrongdoing within the system, including scams targeting beneficiaries, fake websites posing as SASSA to collect personal data, loan sharks using grants as collateral and unauthorised deductions for funeral policies.

Researcher Abby May at Open Secrets said the state has increasingly outsourced key parts of the infrastructure used to administer grants, including systems for applications, reviews and payments.

“A consequence of this is that much of the knowledge required on how to respond effectively to social grant recipients is no longer held by the state but by private companies that built the infrastructure. This is a risk.”

Private companies involved in these systems could gain access to large amounts of sensitive beneficiary data, she said.

“What Net1 and CPS showed is that CPS had complete access to information running throughout the grant system, which included ID information of the grant recipients, but more importantly, their financial situation and the transactions coming in and out of the recipients’ bank accounts,” May said.

She warned that access to this data can enable predatory marketing practices.

© 2026 GroundUp. This article is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.

You may republish this article, so long as you credit the authors and GroundUp, and do not change the text. Please include a link back to the original article.

We put an invisible pixel in the article so that we can count traffic to republishers. All analytics tools are solely on our servers. We do not give our logs to any third party. Logs are deleted after two weeks. We do not use any IP address identifying information except to count regional traffic. We are solely interested in counting hits, not tracking users. If you republish, please do not delete the invisible pixel.