Financial Inclusivity Is a Myth

I've been hearing "financial inclusivity" for a decade. The FDIC numbers haven't moved.

Neobanks were going to bank the unbanked. AI lenders were going to score the unscorable. BNPL was going to democratize credit. And every year, the FDIC publishes its household survey, and the same numbers come back. 5.6 million households are still unbanked. 19 million are underbanked. The products changed. The outcomes didn't.

I have a bone to pick with the inclusivity crowd, and it starts with a structural claim: the systems that fintech operates around, the credit bureaus, the banking regulations, the risk models, are designed to benefit people who already have money. Not by accident or neglect. By design. And no amount of AI or mobile-first UX changes that design.

Forty-five million people the system can't see

The credit system can't see 45 million Americans.

Twenty-six million Americans have no credit file at any bureau. Another 19 million have files too thin to generate a score. That's 45 million adults, roughly one in five, locked out of the primary system lenders use to decide who gets capital.

Fifteen percent of Black consumers and 12% of Hispanic consumers are credit invisible, compared to 9% of white consumers. Median credit scores in majority-Black communities sit at 582. Majority-Hispanic communities, 644.

A Harvard study from Opportunity Insights, published in July 2025, found that Black borrowers score 15 points lower than white peers with identical repayment histories. Same behavior, different score. The scoring model bakes in the gap.

Over 80% of 18-19 year olds are credit invisible or unscorable. The system excludes anyone who hasn't already been inside it.

The math that kills inclusivity

So build a better model. Use alternative data. Deploy machine learning. Score the unscorable.

This is the Upstart pitch. AI that looks beyond FICO to assess creditworthiness. More data, better decisions, more people approved. The NAACP Legal Defense Fund and Student Borrower Protection Center raised concerns that Upstart's use of educational criteria, which college you attended, what degree you hold, produced discriminatory lending outcomes through a different door. The CFPB terminated Upstart's no-action letter.

But even if the model were perfect, the math still breaks. Subprime default rates run 5 to 10 times higher than prime. A lender serving this population has to charge rates high enough to cover losses, build reserves, and return capital to investors. The Federal Reserve's own interagency guidance states the quiet part plainly: subprime loans can be profitable "provided the price charged is sufficient to cover higher loan-loss rates and overhead costs."

Sufficient to cover higher loss rates. That means APRs in the 20s and 30s for personal loans. Payday-adjacent rates for smaller amounts.

The default rate is the constraint, and it reflects the borrower's economic reality, which the loan itself can't change.

So the lender either charges predatory rates and calls it inclusion, or offers reasonable rates and loses money. There is no third option within the existing risk framework.

Products that wear inclusivity as a costume

BNPL was the most recent version of this pattern. Marketed as financial inclusion for people without credit cards. Pay in four installments, no interest, no credit check.

Forty-one percent of BNPL users made at least one late payment in the past year, up from 34% the year before. The Kansas City Fed found that 96% of BNPL users who made late payments are financially constrained. BNPL users are 26 percentage points more likely to have overdrafted their bank account and hold $11,981 less in non-retirement savings than non-users.

Who uses BNPL most? Black, Hispanic, and female consumers, according to the CFPB's own 2023 data. The product reaches the people traditional credit locks out, then charges them when they slip.

The CFPB issued a rule in 2024 classifying BNPL providers as credit card issuers under the Truth in Lending Act. Dispute rights, refund rights, periodic statements. The same consumer protections credit cards have carried for decades. But the current CFPB has permanently dismissed 23 enforcement actions against predatory lenders. The regulatory backstop is being dismantled while the products keep selling.

The price of exclusion

Being locked out of the banking system costs money. Every day.

The FDIC's 2023 survey found that 42.3% of unbanked households cite the same reason: they don't have enough money to meet minimum balance requirements. The system excludes them for being poor, then charges them for being excluded.

Unbanked Americans spent $189 billion in fees and interest on alternative financial products in a single year. Check cashing, money orders, prepaid cards, payday loans. Roughly $3,000 per household per year, paid by the people who can least afford it.

Brookings calculated the lifetime cost: $40,000 in check-cashing fees alone. That money, invested over 40 years at market returns, would be worth $360,000.

One in five African-American households is unbanked. One in three is underbanked. Twenty-two percent of adults earning below $25,000 have no bank account, compared to 1% of those earning above $100,000. The companies cashing their checks and selling them money orders are the ones profiting from the gap.

The gate that stays locked

Opening a bank account in the United States requires a government-issued photo ID, proof of address, and a Social Security number. The Patriot Act's Customer Identification Program mandates it.

A person leaving prison often has no photo ID. A person experiencing homelessness has no proof of address. An undocumented immigrant has no SSN, and a young person just entering the workforce may not have any of the three.

These requirements exist for legitimate anti-money-laundering reasons. They also systematically exclude the populations that "financial inclusivity" claims to serve.

Congress chose these requirements. Better technology doesn't remove them. Every fintech product built on traditional banking rails inherits the same gate. A neobank with no minimum balance still needs KYC. A mobile-first savings account still needs a Social Security number. A better onboarding screen doesn't help someone without a Social Security number.

What comes next

If the systems are designed to exclude, can anything built within them be inclusive?

I'm not sure. But the question itself points somewhere specific: away from the existing rails entirely. Toward infrastructure where participation doesn't require a credit score, a bank account, or a government ID to start.

That infrastructure is being built. Stablecoins are already functioning as savings accounts in emerging markets, with two-thirds of the $280 billion stablecoin supply held by individuals in countries where the local banking system failed them first.

But tokenized finance faces the same three problems: the on-ramp still requires KYC and a bank account, institutional versions are permissioned walled gardens, and US adoption skews toward people who already have traditional accounts. Different rails, same gatekeepers.