The Rise of Fraud Concerns in South Africa's Social Grant System
South Africa's social grant system, administered by the South African Social Security Agency (SASSA), provides essential support to millions of vulnerable citizens. These grants include the old age pension, child support grant (CSG), disability grant, and the Social Relief of Distress (SRD) grant, which gained prominence during the COVID-19 pandemic. With over 28 million recipients in 2025, the system represents a lifeline amid high unemployment rates exceeding 33% and widespread poverty. Public perception of rampant fraud has intensified, prompting government action. However, recent research questions whether the response targets the right culprits.
Key Findings from the Institute for Economic Justice Research
The Institute for Economic Justice (IEJ) released UBIG Factsheet 13, 'Going after the gogos,' in early 2026, analyzing fraud cases from 2014 to 2025. The study reveals that the ongoing crackdown disproportionately burdens legitimate beneficiaries rather than addressing root causes. Drawing from parliamentary reports, news, and beneficiary accounts, it highlights systemic flaws in verification processes.
Who is Really Committing Social Grant Fraud?
Contrary to popular narratives, IEJ data shows 75% of 3,455 reported fraud cases (2,658 instances) involved government employees, contractors, or officials. Examples include ghost beneficiary syndicates defrauding R260 million in 2025 and manipulated accounts stealing R10.2 million in 2020-2021. Beneficiaries accounted for only 17% (590 cases), averaging 54 per year—a tiny fraction of 28 million recipients. Predatory third parties, like loan sharks and unauthorized deductors, contributed another 6% (207 cases).
- Government insiders: Ghost cards, duplicate payments, inflated grants (at least R354.88 million stolen since 2014).
- Private actors: Unauthorised insurance deductions, data-selling scams mimicking SASSA sites.
- Beneficiaries: Rare intentional misrepresentations, like undeclared income or ghost children (R61,256 repaid 2012-2025).
Outsourcing to private firms like Net1/CPS has raised data privacy risks, enabling predatory marketing.
Government's Anti-Fraud Measures Explained
The National Treasury mandated enhanced verification in the 2026 budget, requiring SASSA quarterly reports on reviews, suspensions, and savings. Measures include:
- Bank account monitoring to detect 'income' above thresholds.
- Biometric verification (fingerprints, eKYC).
- Cross-checks with SARS, UIF, Home Affairs, credit bureaus, correctional services, and payroll databases.
- Algorithmic risk detection flagging suspicious activity.
- SMS notifications for in-person reviews within 30 days, or grant cancellation.
By February 2026, 291,000-300,000 beneficiaries were flagged, with 34,600 grants cancelled, yielding projected R3 billion savings. SASSA's biometric rollout saved R300 million by early 2026.
Disproportionate Impacts on Vulnerable Beneficiaries
The IEJ warns these tools are blunt instruments. Bank checks misinterpret remittances, child maintenance, loans, or stokvel contributions as income—54-76% of SRD rejections erroneous. Biometrics fail on elderly with worn fingerprints ('gogos') or disabled individuals. Rural poor face transport barriers, long queues from 3am, and limited office capacity (25,000 reviews/month vs. 210,000 flagged in 2025).
Most cancellations stem from non-attendance, not proven fraud. This 'double victimization'—first by fraudsters, then exclusion—deepens poverty, forcing asset sales or medical crises.IEJ Factsheet 2 details cases.
Real-World Cases and Beneficiary Struggles
Elderly women queue overnight in rain for gold card replacements. Disabled caregivers juggle reviews with dependents. Rural recipients travel hours, incurring costs exceeding grant value. One 2025 case: 5 SASSA employees dismissed in R260m ghost card fraud, yet thousands of grannies penalized. Privacy erosion via data-sharing consents lacks transparency, with algorithms opaque.
Civil society reports scams via fake SASSA sites stealing data, loan sharks holding cards hostage.
Government and SASSA Perspectives
SASSA urges SMS responses to avoid suspension, emphasizing reviews verify eligibility and prevent fraud. Minister Sisisi Tolashe highlights biometric savings (R300m). Treasury's Enoch Godongwana notes R536m lost to fraud in prior years, justifying R3bn projected savings. Grants increased inflationarily in 2026 (e.g., old age to R2,315/month), with SRD extended.SASSA statement.
Yet IEJ's Dr. Kelle Howson argues: 'Resources policing beneficiaries are disproportionate... consequences fall on the vulnerable.'
Stakeholder Views and Broader Implications
Open Secrets' Abigail May: 'Fraud exists... but implementation targets wrong people.' GroundUp echoes insider dominance. Implications: Eroded trust, deepened inequality, policy debates on universal basic income (UBIG) to reduce errors. For researchers and policymakers, this underscores data quality needs.GroundUp analysis.
In higher education, studies like this highlight social policy's research role. Explore academic career advice for policy researchers.
Recommendations for Balanced Fraud Prevention
IEJ proposes:
- Prioritize official/third-party probes.
- Human oversight for automated flags, transparent appeals.
- Non-digital options for elderly/disabled.
- Reduce documentation burdens.
- Advocate UBIG to minimize targeting errors.
SASSA could expand capacity, refine algorithms. Long-term: Robust anti-corruption in public sector.
Photo by Markus Winkler on Unsplash
Future Outlook and Policy Pathways
As grants rise to R292bn in 2026/27, balancing integrity with access is key. Ongoing reviews may save billions but risk humanitarian costs if unrefined. Researchers urge rights-based approaches protecting the poor. For South Africans, stay informed via SASSA updates. Policy experts can contribute via research jobs or SA academic opportunities. Engage at Rate My Professor or pursue higher ed career advice.
This research spotlights evidence-driven reform needs, ensuring social grants uplift, not exclude.
