Every month, roughly 18 million South Africans receive a social grant. It is one of the largest welfare programmes on the continent by headcount, and for years the story we told ourselves about digitising it was a conveniently tidy one. Think faster payments, fewer ghost beneficiaries and less cash sloshing around to be skimmed at the till. The Daily Maverick's Open Secrets investigation is starting to unpick that version of events. Digitisation was sold as efficiency. It has also, as it turns out, been sold as a revenue stream. The reporting focuses on the plumbing behind the Social Relief of Distress grant- the R370 payment (which the Pretoria High Court has since ordered raised to R624) that keeps food on the table for millions of unemployed South Africans. Sassa doesn't verify eligibility itself; rather, it contracts the work out. Absa, African Bank, FNB, Nedbank, Standard Bank and Tyme Bank are all on the roster, contracted from March 2025 through July 2027 to run monthly income checks against applicants' bank accounts. Prosense Technology built the web portal, and Vodacom runs the zero-rated USSD line. GovChat, on the other hand, handles the WhatsApp channel. Every layer between the beneficiary and the payment is now a private counterparty with a contract. That is not, on its own, scandalous. Governments outsource plumbing all the time. The problem starts when the plumbing is mislaid, and the wrong tap runs dry. In January 2025, Judge Leonard Twala declared the bank-based income verification unconstitutional. That ruling is suspended pending appeal, and the Supreme Court of Appeal hears the matter on 25 August. The reason it got that far is straightforward. The bank check cannot tell the difference between a burial society contribution and a salary. It cannot smooth for income that arrives lumpily, which is how most poor households actually receive money. Sassa itself concedes, through executive Brenton van Vrede, that "a lot of people got excluded that should not have been excluded." One beneficiary, per the Daily Maverick, had her grant declined because R1,000 toward her mother's funeral registered as income. Set the constitutional argument aside, and the investable reality is simpler. Anyone sitting between the state and the beneficiary collects something on the way through. Sometimes it is an explicit fee. Sometimes it is float - i.e. the days between Treasury releasing the money and the beneficiary drawing it. Sometimes it is data. A monthly, verified snapshot of the transactional behaviour of the country's poorest 18 million people is not a trivial asset in a market where the big four banks are all racing each other into the lower-LSM segment. Multiply any of that across a monthly disbursement cycle, and you have a structurally attractive annuity sitting on top of a fiscal line item. Grant spending runs at roughly R270bn a year. It is a bigger book than most listed South African financials would turn down. The beneficiary bears the cost as a deduction they usually cannot itemise, and almost never contest. Sassa's own complaints workflow, per the investigation, is that a call-centre agent forwards the issue to Sassa IT, which forwards it to the bank's IT team, who checks whether the bot got it wrong. That is not an audit trail; it is a queue. The consensus in fintech circles is that digitising grants is unambiguously progressive. Faster, cheaper, less corruptible than cash. Most of that is true. Digitisation has genuinely reduced one form of leakage- the cash-theft, ghost-beneficiary, envelope-in-a-drawer variety that hollowed out earlier programmes and put Cash Paymaster Services on the front pages a decade ago. What it can also do is institutionalise another kind of leakage, such as platform fees and data monetisation. Verification errors that re-price the welfare transfer downward before it even reaches the household. The net welfare gain is thus not a given. It depends entirely on who sets the fee architecture and whether Sassa has the capacity to audit it. Treasury's response to the Daily Maverick was that many of the questions "fall outside its mandate." Sassa's was that journalists should file a PAIA request. On current evidence, the answer to both is: not really. For SA-focused allocators, there are two things worth carrying out of this. Any listed or pre-IPO fintech or payment processor with material Sassa exposure carries a regulatory-risk premium that is not currently in the multiple. The big-four banks can absorb an adverse SCA ruling without flinching, given they earn the bulk of their money from net interest margin (NIM)- i.e. the spread between what they pay depositors and what they charge borrowers. Group NIM sits in the 4.5% to 5% range across the majors, which translates into tens of billions of rand a year in net interest income. Standard Bank's net interest income alone runs north of R90bn. Against that number, the SRD verification contract is small. The smaller players, however, cannot walk away as easily. Further journalistic investigation, a parliamentary inquiry, or an unfavourable ruling on 25 August could reprice that overnight. This is not a call to short anything, but more a call to price the tail. The second is broader, and it connects to two other stories from the same news cycle. Project Syndicate this week argued that West Africa's monetary union is facing its own institutional stress test, with Senegal's debt problem exposing a governance gap that a shared currency cannot paper over. RMB, meanwhile, has just launched a digital trade finance platform for African corridors. Different stories, but they essentially boil down to the same truth: digital infrastructure is being laid across the continent faster than the governance layer can keep up. Whilst value accrues in that gap, so does the leakage. Zimbabwe serves as a timeous reminder. Digital rails compound in value only when the institutions above them hold up. When they don't, the rails still work. They just work for someone else. Eighteen million people get paid once a month, and a handful of intermediaries decide how much of it actually arrives. Two decades into the post-apartheid welfare state, that is the version of financial inclusion we have actually built. It is worth being honest about who captured the value the digitisation created - the recipients, or the counterparties standing between them and the money? | 
