There is a version of the data-sharing debate that sounds reasonable on paper. Dominant firms agree to participate in industry-led data pools. Smaller competitors gain access to the insights they need. Markets open up. Innovation spreads. Everyone benefits.
Jens Prüfer and Paul de Bijl, economists at Tilburg University and the Netherlands Authority for Consumers and Markets, respectively, spend considerable time in a new paper explaining why that version is a fantasy. Their argument has direct implications for banking, payments and any financial services sector where data drives product quality.
In markets where improving a product requires learning from customer behavior, and where that learning data is owned exclusively by the firm collecting it, dominant players do not just win. They lock in. A modest early lead grows over time because the leading firm can copy rivals’ innovations more cheaply, using its data advantage. Smaller firms, seeing what they are up against, stop investing. Some exit. Some never start.
This is not about anyone breaking the law. It is about how the market is built.
The financial services application is explicit. Research cited by the authors found that Apple Pay and Google Wallet combine transaction data, location signals, biometric authentication and search behavior in ways that traditional banks structurally cannot match. Banks are not locked out by contract or regulation. They are locked out by the architecture of the data ecosystem itself. The U.K.’s Financial Conduct Authority acknowledged this risk in financial services and responded with monitoring and pilot programs. The authors are direct: that is not enough. Once a market tips toward a single dominant player, reversing it becomes prohibitively expensive.
The same dynamic runs through credit scoring, fraud detection, insurance underwriting, payments routing and product recommendations. Any financial services market where customer behavior generates proprietary learning data is at risk.
Advertisement: Scroll to Continue
Why Voluntary Sharing Cannot Work
Industry-led data pools, collaborative sharing arrangements, and voluntary initiatives all rest on an assumption that dominant firms have some reason to participate genuinely. They do not, the authors argue.
“Dominant data-rich firms have no incentive to share because they already profit tremendously from their data advantage,” Prüfer and de Bijl write. A firm with a commanding market position faces a straightforward choice: share data and watch rivals improve, or keep the data and keep collecting outsized profits. The rational choice is obvious.
Voluntary pools attract data-poor firms looking for access and data-rich firms that participate in name only, submitting low-quality or outdated data. The cooperation falls apart under its own logic.
The authors also examine more elaborate schemes that use complex formulas to calculate each participant’s fair share of a shared data pool. These fail in practice too. Working out each firm’s fair contribution requires knowing the value of every possible data combination, now and in the future, plus an impartial referee to enforce the math.
In reality, the value of data depends heavily on context and changes over time. Disputes are inevitable, referees have their own interests, and the cost of managing the whole arrangement eliminates whatever benefit the sharing was supposed to create.
3 Requirements for a Solution That Works
The authors argue that data sharing is fundamentally an enforcement problem. Three features are required for any solution to hold:
- Real penalties, not social pressure. Private arbitration works when parties have roughly equal power and reputations at stake. Neither applies here. A firm earning billions from data advantages is not deterred by the threat of being excluded from a voluntary sharing pool. Enforcement needs legal authority and financial consequences.
- Separated responsibilities. The body that decides whether sharing is required should be different from the body that monitors compliance. National competition authorities, with economists and lawyers who understand these markets, should set the rules. A separate public agency should run the shared data infrastructure and conduct technical audits to verify that what firms submit is genuine.
- Public accountability. Firms are more likely to comply when the enforcer answers to elected officials and is subject to court review. Private governance bodies, however well designed, lack that legitimacy.
What This Means for Financial Services Now
The European Union’s Digital Markets Act was a necessary start, but the authors identify gaps that matter directly for financial services. Its thresholds miss mid-sized firms that are already dominant in their markets. Its obligations stop at designated platforms and do not follow data advantages across supply chains. And its interaction with privacy rules creates delays that undermine the sharing it was designed to require.
The authors recommend that financial institutions above 30% market share should be required to share customer behavioral data with authorized competitors. That threshold is set deliberately, to intervene before a market tips completely rather than after.
For banks and payments firms, the message is straightforward. Data-sharing rules are coming. The question is whether regulators will build the enforcement infrastructure to make those rules real. According to this research, rules without enforcement are not rules. They are suggestions, and firms with dominant data positions have no reason to follow them.
