How We Pay (2/7): Who Is Actually Excluded

In the first article of this series I left a question open: what, in the design of today's systems, keeps cash rational? The usual answer is that part of the population does not want to enter the financial system. Twenty years in the field have convinced me of the opposite. Financial exclusion is almost never a refusal. It is an outcome. It is produced by a series of thresholds, each of which can be crossed, and each of which is an engineering or regulatory decision taken elsewhere.

Let us start by not talking about the excluded in general. The category is too broad to be useful. We are talking about a market trader who takes two hundred small payments a day, a driver paid fare by fare, a farmer whose income arrives twice a year, an employee paid in cash by an employer who does not declare them. These people handle money constantly. They are not outside the economy. They are outside its record.

The first threshold is identity. A payment institution is not allowed to open an account for someone it cannot identify: that is what KYC means, and it is not administrative friction but a regulatory obligation on which its licence depends. The World Bank estimates that roughly 850 million people worldwide have no official proof of identity, close to half of them in Sub-Saharan Africa. For a significant share of the population, financial exclusion therefore does not begin at the bank. It begins at the civil registry.

The second threshold is price. Transfer and withdrawal tariffs are built in bands, and within its interval a band behaves like a flat fee. The result is mechanical: the smaller the amount, the heavier the fee in proportion. Whoever moves the smallest sums pays proportionally the most. For that person, withdrawing everything at once and staying in cash is not distrust. It is the correct calculation.

The third threshold is whether the service is actually available. An agent network is measured on a map, but it is experienced during the day. An agent who has run out of cash cannot serve a withdrawal; an agent who has run out of electronic value cannot take a deposit. Managing that liquidity is one of the most stubborn operational difficulties in the business, and it is invisible in coverage statistics. To the user, an agent with an empty till is a closed branch — and two failures are enough to send them back to cash for good.

The fourth threshold is proof. Informal employment accounts for a very large majority of jobs on the continent. That means real income with no documentation: no payslip, no financial statements, no transaction history. When that person applies for credit, they are not turned down because they are insolvent. They are turned down because the file cannot be assembled. The distinction matters: the system is not measuring risk, it is measuring available documentation, and it treats the absence of documents as risk.

These four thresholds are not my interpretation. They come out of the surveys. When the World Bank asks unbanked adults directly why they have no account, the dominant answers are always the same: I do not have enough money for it to be worth it, it costs too much, it is too far, I lack the paperwork, I do not trust it. "I do not want one" never comes first. Willingness is not the binding constraint.

There is a fifth threshold, quieter than the others, and it is the one we are directly responsible for as designers. An onboarding flow built for a recent smartphone excludes whoever has a basic handset. A form written in administrative French excludes whoever does not read it fluently. A support line reachable during office hours excludes whoever is working precisely at those hours. None of these decisions was taken against anyone. All of them produce the same effect as a prohibition.

That is what makes the subject tractable. A regulatory threshold, a tariff table, a network architecture, a form: these are built objects, and built objects can be changed. Where we say "people prefer cash", we should read "we have not yet made the alternative better for them". The reformulation changes everything, because it points to work to be done rather than a cultural inevitability to be observed.

One serious objection remains. Even with those thresholds lowered one by one, a user who has to choose between three networks that do not talk to each other still faces a system more complicated than their wallet. That is the question of interoperability, and it is the subject of the next article.

Sources: Global Findex Database, World Bank — self-reported reasons for not holding an account; ID4D Global Dataset, World Bank — estimate of people without official proof of identity; International Labour Organization, Women and Men in the Informal Economy — share of informal employment in Africa; GSMA, State of the Industry Report on Mobile Money — agent liquidity management.