In 1909, a group of Franco-American Catholics in Manchester, New Hampshire opened St. Mary's Cooperative Credit Association — the first credit union in the United States. They were not untrustworthy borrowers. They were unreadable ones. Creditworthiness, in that era, ran almost entirely through social and institutional networks like church standing and existing banking relationships. Traditional banks had no framework for assessing risk outside those networks. The Franco-American community existed entirely outside them. So they didn't appeal for inclusion. They built their own system — one that could use the information banks couldn't see.
The model was direct: pool capital within the community, lend based on proximity, shared context, and mutual accountability. Not charity. Infrastructure.
That was 1909. The infrastructure has updated but the context problem hasn't.
According to the Consumer Financial Protection Bureau, roughly 45 million American adults (nearly 1 in 5) either have no credit record or one that can't be scored by standard models.¹ They aren't necessarily high-risk borrowers. They're borrowers whose financial lives don't generate the right kind of legible data.
The CFPB term is "credit invisible." The description is precise and from the standpoint of modern underwriting, these people functionally don't exist.
The barrier has changed shape. In 1908, the obstacle was cultural familiarity because banks had no tools or trust for communities outside their frame of reference. Today the obstacle is algorithmic. Underwriting models require a specific kind of financial history: consistent, formal, documented. If your income is irregular, gig-based, or thin-file, the system doesn't flag you as risky, it registers you as unreadable. Different barrier. Same outcome.
That's where the roughly 11.9 million Americans who worked as independent contractors in 2023 enter the picture — about 7.4% of the workforce, counting only those for whom contracting was the primary job.² Count full-time independents by the broader MBO Partners methodology, and the number reaches 27.7 million in 2024 — more than double the 13.6 million recorded in 2020.³ That curve predates the full arrival of AI-driven automation, which is expected to accelerate both voluntary and involuntary exits from traditional employment.
These workers are, in structural terms, what the Manchester immigrants were in demographic ones: a population whose financial life doesn't fit the standard template. Volatile income, irregular cash flow, no employer-verified W-2. Not high-risk, just unreadable.
The original credit union worked because proximity created information that formal systems couldn't. Members knew each other. Accountability was social before it was financial. What a bank saw as opacity, the community read as context.
Why This Matters
The barrier is now invisible too. The exclusion is no longer overt, but the mechanism is similar. Credit scoring models gate access based on data legibility, not creditworthiness. A borrower with consistent rent payments, utility history, and a clean track record may remain invisible to standard underwriting — because none of those signals are captured in the model.
Volatile isn't the same as unreliable. The gig economy has produced a structural class that is systematically underserved by financial products built for W-2 stability. The infrastructure of mortgages, business credit, and personal loans was engineered for predictability. It performs poorly on income that is volatile but not unreliable.
The data exists but the relationship may not. With the proliferation of data across all industries and infiltrating every aspect of life (see: Burger King's "Patty"), fintech, community lending, or alternative data sources should be able to do what proximity once did but may also remove something essential when the relationship is engineered out of the equation.
The first deposit at the North American credit union — organized by Alphonse Desjardins in Quebec in 1900, before he helped bring the model to New Hampshire — was ten cents.⁴ That's not a metaphor for humble beginnings or a nod to Scrooge McDuck's lucky #1 dime. It's a data point about who the system was built for and who decided to build something else.
Today's unreadable class is larger and less visible. Tools designed to reach them are emerging such as cash-flow underwriting that reads bank account activity rather than credit history, alternative data models that incorporate rent and utility payments, community development financial institutions extending credit where traditional banks won't. Whether these close the gap in a meaningful way, or reproduce a more sophisticated version of the same blind spot, remains to be seen.
Sources
[1] Consumer Financial Protection Bureau, "Who Are the Credit Invisibles?" December 2016. cfpb.gov [2] U.S. Bureau of Labor Statistics, Contingent and Alternative Employment Arrangements, July 2023. bls.gov [3] MBO Partners, State of Independence in America, 2024. mbopartners.com [4] National Credit Union Administration, Historical Timeline. ncua.gov