Your financial data is not just a record of what you have done with money. It is a map of your life. Every salary credit tells the story of your employment. Every pharmacy payment hints at your health. Every school fee, every takeaway order at eleven o'clock at night, every loan repayment that came in three days late — they all say something about who you are, how you live, and what you are going through. Aggregated across millions of transactions, across millions of people, this map is one of the most commercially valuable assets in the contemporary economy.
South Africa's regulatory conversation about financial data has matured considerably in recent years. We talk with increasing sophistication about consent, about data minimisation, about the right to withdraw authorisation, about who should be permitted to access the data flows that open finance creates. These are important conversations, and I have been part of some of them. But almost without exception, they are conversations about access — about who is allowed to see the data, under what conditions, and with what protections.
The questions we are not asking — or not asking loudly enough — are about value. Who captures the economic rent generated by the patterns hidden in your transaction history? And if data is truly the 'new oil', why is the person who generates it the only party in the value chain excluded from the dividend?
The paradox of open finance
Open finance rests on a compelling premise: that consumers should have the right to port their financial data from incumbent institutions (like their primary bank) to third-party providers (like a budgeting app, a specialist lender, or an insurance comparator). The FSCA's 2024 Open Finance Policy Recommendations articulate this vision clearly, framing data portability as a catalyst for competition and financial inclusion.[1]
But there is a paradox embedded in the current open finance model. The rhetoric is one of consumer empowerment — 'your data belongs to you'. Yet the economic structure remains entirely asymmetrical. When you grant a third-party app access to your transaction history, you receive a service in return (a budget breakdown, a credit offer). The app provider, however, receives something of durably greater value: a permanent addition to their data asset base, which they use to refine their algorithms, cross-sell products, attract venture capital, and build enterprise value.
In this exchange, the consumer is positioned merely as the source of a raw material. We have built an elaborate regulatory infrastructure to ensure the consumer consents to the extraction of this material safely. But we have built almost nothing to ensure the consumer receives a fair share of the wealth that the refined material ultimately produces.
Data as labour, not just privacy
To understand why this is problematic, we need to change how we conceptualise data. The legal and regulatory framework in South Africa, anchored by the Protection of Personal Information Act (POPIA), treats data primarily through the lens of privacy and personality rights. Data is something that describes you, and therefore something you have a right to shield from unauthorised view.
But in a digital economy driven by artificial intelligence, data is also something else: it is labour. The economists Eric Posner and Glen Weyl have argued persuasively that the data users generate as they go about their digital lives is the essential work that trains machine learning models and generates commercial value.[2] When a bank's credit scoring algorithm becomes more accurate because it has processed your repayment patterns alongside millions of others, you have performed a form of unpaid work that improves the bank's intellectual property.
If we accept the premise that financial data is a form of value-generating labour, the conversation shifts. It is no longer just about whether a consumer ticked a box authorising data sharing. It is about whether the exchange was economically fair.
The models we could build
What would a financial system look like if it acknowledged the economic value of consumer data? We do not have to invent the answers from scratch; experiments in data valuation and collective bargaining are already emerging globally.
One model is the direct data dividend. In 2019, the Governor of California proposed that tech companies should pay a 'data dividend' to consumers in exchange for monetising their personal information.[3] Translated to financial services, this could involve institutions offering tangible financial incentives — fee reductions, preferential interest rates, or direct micropayments — explicitly priced against the continuous access they hold to a customer's broader financial dataset.
A more structural approach involves 'data trusts' or data cooperatives. In this model, individuals do not negotiate the value of their data one by one (a fight they will always lose due to asymmetry of information and bargaining power). Instead, they pool their data rights into a trust, governed by fiduciaries who negotiate data access and profit-sharing agreements with financial institutions on behalf of the collective.[4] For South Africa, where collective economic structures like stokvels already have deep cultural resonance, the concept of a financial data cooperative is theoretically potent.
There are also models emerging in the decentralised finance (DeFi) space that attempt to bake data remuneration into the protocol layer. Certain Web3 projects allow users to tokenise their verified financial credentials or transaction histories, earning yield when institutions pay to query that data for underwriting purposes. While these solutions remain nascent and carry significant regulatory complexities of their own, they represent a serious architectural attempt to solve the value capture problem.
The limits of current regulation
I raise these models not because any of them are ready for immediate deployment in the South African market, but to highlight the boundaries of our current regulatory imagination. As regulators, we are generally comfortable intervening when financial products are unfairly priced. The Conduct of Financial Institutions (COFI) framework is designed to ensure that institutions do not extract unfair value from consumers through opaque fees or unsuitable product design.
Yet we currently have no framework for assessing whether the value extracted from consumers in the form of data is fair. If a consumer receives a \"free\" budgeting tool in exchange for data that the provider subsequently uses to train an AI model that generates millions in revenue, has the consumer been treated fairly? Under current paradigms, as long as the privacy disclosures were clear, the answer is assumed to be yes. But as data becomes the primary driver of enterprise value in finance, this assumption will become increasingly difficult to defend.
As we design the rules for open finance and artificial intelligence in South Africa, we must ensure we are not merely regulating the safety of the extraction process while ignoring the destination of the wealth. It is time for consumers to start asking, with equal seriousness, what their data is worth — and whether any of that worth should be theirs.