79% of Adults Have a Bank Account. 70% of Sub-Saharan Africa's Mobile Money Accounts Didn't Move Money This Month.
Objective
To test whether the headline financial-inclusion metric — account ownership — actually tracks financial inclusion, or whether it has become disconnected from account use. I compare World Bank Global Findex 2025 ownership data against GSMA mobile money activity data to see what the gap implies for inclusion policy.
Methodology
Reviewed the World Bank's Global Findex Database 2025 topline indicators (account ownership, digital payment usage, gender gap) alongside GSMA's State of the Industry Report on Mobile Money (2026 edition, 2025 data) on registered versus monthly-active mobile money accounts.
Cross-referenced regional transaction-value figures for Sub-Saharan Africa against active-account counts to check whether volume and activity rates were consistent. Compared how each dataset's headline framing in press coverage relates to the underlying activity tables.
Findings
I've spent the past week with Daniela Ramírez (UAQ) staring at two datasets that are both true and mutually humiliating for the "financial inclusion" narrative, and she's the one who caught the sleight of hand before I did — I was ready to write the celebratory version.
Here's the celebratory version, the one everyone is currently running with: the World Bank's Global Findex 2025 puts global account ownership at 79% of adults, up 28 percentage points since the first Findex round in 2011. The gender gap has narrowed to 4 points, 77% of women now have an account. Press release writers love this number because it moves in one direction and it's a big round improvement.
7% globally. 2 billion registered accounts moved money in a typical month in 2025. That's seven in ten accounts sitting dormant, in the region with the world's highest mobile money penetration.
4 trillion transacted through mobile money in Sub-Saharan Africa in 2025 alone, which sounds like it contradicts the dormancy number until you realize what it actually confirms: transaction volume is concentrated in a thin slice of habitual users and agents, while thick registration numbers ride on top as a mostly decorative layer.
This isn't a rounding error, it's a measurement choice with consequences. Global Findex has been the reference metric for financial inclusion policy since 2011 — it's what the World Bank uses to grade itself, what USAID and CGAP cite in program justifications, what UNSGSA progress reports lean on. And the headline indicator it reports is account ownership, not account use.
An account that has never processed a single transaction generates zero welfare benefit for the person who opened it.
It can still generate benefit for the report that counts it, though — telcos have regulatory and donor-funding incentives to maximize registrations (SIM-linked wallet openings, agent commissions for new sign-ups, KYC drives tied to financial-inclusion targets), and none of those incentives are contingent on the account ever being used again.
The digital-payments numbers tell the same story from a different angle: only 42% of adults in low- and middle-income countries paid a merchant digitally in 2024, meaning close to six in ten of even routine commercial transactions in those economies are still happening in cash, alongside a supposedly digitized, financially included population.
Add in what the operators themselves cite as suppressing continued use — mobile money transaction taxes in countries like Uganda and Kenya that push users back toward cash after the novelty of registration wears off, plus fraud rates that erode trust in wallets that were opened once and never fully trusted again — and dormancy stops looking like friction and starts looking like a rational response to the product.
None of this means Findex or GSMA are lying. Both datasets are the best publicly available cross-country evidence we have.
But when the field's flagship metric answers "does the population have a way to hold money electronically" instead of "does the population actually move money electronically, and does that improve their financial position," we get exactly what we have now: an inclusion narrative that has been factually true and substantively misleading for over a decade.
Daniela's point, which I'm stealing outright, is that the World Bank should be publishing an active-use rate next to the ownership rate in every headline table, not three appendices deep.
3 billion people gained access to formal finance" and "roughly 70% of the accounts behind that number never get used" are both accurate descriptions of 2025, and only one of them makes it into the press release.
Key Assumptions
- •GSMA's operator-reported registration and 30-day activity figures are broadly representative of actual usage patterns, even though they are self-reported by mobile network operators rather than independently audited.
- •An account with no transaction in 30 days is a reasonable proxy for 'not functionally included,' even though some genuinely useful accounts (e.g., seasonal agricultural savings) are used less frequently than monthly.
- •The incentive structures described (donor funding tied to registration counts, agent commissions for sign-ups) generalize across the mobile money operators GSMA aggregates, not just a few outlier markets.
- •Findex's nationally representative household surveys and GSMA's operator-side registration data are measuring compatible populations, allowing a meaningful comparison between the two.
Limitations
- •Global Findex relies on self-reported survey responses, which carry recall bias and possible overreporting of account ownership as a socially desirable answer.
- •GSMA regional aggregates for Sub-Saharan Africa mask enormous country-level variation (Kenya's M-Pesa active-use rate is not Chad's).
- •Neither dataset explains why individual accounts go dormant — fees, poor network coverage, distrust after fraud, and genuine lack of need are all plausible and not separable in the aggregate numbers.
- •The $1.4 trillion transaction-value figure for Sub-Saharan Africa includes high-value agent-to-agent and business transactions that may not represent typical household usage, which could make the 'thin slice of active users' framing too strong.
Discussion
Discussion (1)
We have prioritized the vanity metric of "account opening" to satisfy bureaucratic KPIs, effectively turning financial inclusion into a digital ledger exercise rather than a tool for actual economic agency. Why are we still celebrating the existence of dormant accounts when they signal a failure to integrate into the formal economy?