What the Early Evidence Can—and Cannot—Prove
Open finance is often described through impressive counts: participating countries, linked accounts, API calls, new providers, or customer consents. The harder question is what those counts demonstrate about competition, choice, credit, and inclusion. This chapter interprets the early international evidence at the level of the claim it can support. It separates system activity from customer outcomes, descriptive growth from studies that examine policy changes, and a detailed Brazilian case from universal conclusions. The aim is not to downplay the gains, but to make their strength and boundaries visible.
From adoption signals to market effects
The first measurement problem is that “adoption” is not one variable. A registered user tells us that someone entered a system. A linked account tells us that an institution has been connected. An active consent suggests that a customer has authorized a data flow. An API call indicates that a request was made, while a successful API call says something about the interface’s ability to respond. A completed payment, new lending relationship, or useful financial-management service is closer to an outcome for the customer. These measures can all be valuable, but they occupy different points in the chain from permission to benefit.
The BIS review records open-banking or open-finance policy frameworks in 95 jurisdictions by 2024. That is evidence of broad institutional interest, not evidence of a common level of use. Frameworks differ in their objectives, product scope, participation rules, technical maturity, and definition of a customer. Some are in early exploration; others have live services and substantial activity. Treating the count as a ranking would hide the very variation that determines whether a permission is useful.
The report’s jurisdictional examples show why denominators matter. Korea had 36 million subscribers and 194 million registered accounts as of December 2023. The subscriber figure represented nearly 70% of the population, but the report warns that the number may include double counting because a person can subscribe to services or hold accounts with multiple institutions. Brazil recorded 55 API calls per user per month in March 2025. That is an intensity measure at a different date, not a directly comparable measure of reach. One set of figures describes the scale of registered relationships; the other describes the frequency of system activity. Neither alone tells us whether users received lower prices, better advice, or fairer access.
The shape of the service matters as well. Read-only open banking may appeal mainly to a financially sophisticated group that wants a consolidated view. Payment initiation can place the data-sharing capability inside an everyday transaction, giving a customer a reason to return. That may help explain faster use in some settings, but it does not mean that write access is inherently better. A payment service creates a more immediate benefit and a more immediate risk. High activity can reflect convenience, automation, or repeated use; it can also reflect a narrow service with limited distribution.
The market-entry evidence is more tangible than a raw consent count. The BIS review reports that more than 4,500 third-party providers were active in EU and UK retail financial services by December 2024. Many operated across borders through the European Union’s passporting arrangements. The figure supports a claim that the frameworks helped create a market in which third-party providers could exist and operate. It does not, by itself, show that all those providers were equally successful, that their products were better than incumbents’ products, or that customers in every segment benefited.
Switching provides another way to test contestability. In the United Kingdom, a related Current Account Switch Service had helped more than 11 million people switch current accounts among more than 50 banks and building societies.