Journal / Essay · Debt

The Math of Being Credit Invisible

For a decade the number was 45 million Americans with no credit score. In 2025 the agency that counted them found a data error and cut the estimate roughly in half. The 32 million people left in the corrected count still can't rent an apartment, insure a car, or get a phone plan without paying extra for being unreadable.

And Then You're Dead · August 2026 · 10 min read

A credit score isn't a report card on whether you're responsible with money. It's a prediction, built entirely out of borrowing history: credit cards, auto loans, mortgages, the accounts that report to Equifax, Experian, and TransUnion. If you've never opened one of those accounts, or opened one years ago and haven't touched it since, there's no history to predict from. The bureaus call that credit invisible if there's no file at all, or unscored if the file is too thin or too old to generate a number. Either way, the result is the same: no FICO score, no VantageScore, nothing a landlord, lender, or insurer can pull up. This is a specific, countable population, and the federal government has been measuring it since 2015 — badly, it turns out, for most of that time.

The Number That Was Wrong for a Decade

The figure repeated everywhere for ten years was 45 million: 26 million Americans with no credit history at all, plus another 19 million with files too thin or stale to score, drawn from a 2015 CFPB report using 2010 data. It became the standard citation in congressional testimony, bank press releases, and fintech pitch decks — the number an entire "credit invisible" industry got built around.

In June 2025, the CFPB published a technical correction to its own report. One of the three nationwide bureaus had, for years, been submitting a sample that quietly excluded whole categories of records: files containing only deferred student loans, only collections, or only closed accounts. Added back in, the 2010 estimate of credit-invisible adults dropped from 25.9 million (11.0%) to 13.5 million (5.8%) — cut by more than half, with most of the difference reclassified into the unscored category instead of vanishing.

Add those two corrected 2020 figures together and it's about 32.3 million adults — roughly one in eight — who cannot produce a credit score. A Federal Reserve publication from October 2025 lands on the same rounded number, describing "approximately 32 million American adults" as effectively locked out of traditional credit scoring. The underlying decade-long trend was real progress, not just a math error: the share of adults with a scored file rose from 81.6% to 87.5% between 2010 and 2020. The invisible population shrank. It just didn't shrink to zero, and the number everyone quoted for it was never right to begin with.

That's the part worth sitting with. Whether the honest figure was 45 million or 32 million, nobody who was actually locked out of the system for lack of a file got a rebate when the count changed. The correction fixed a spreadsheet. It didn't get anyone a lease, a loan, or a lower insurance quote.

Who Actually Ends Up Invisible

The CFPB's original 2015 breakdown — the one whose national totals were later revised — remains the most detailed public look at who ends up in this group, and the disparities are too large to be an artifact of the data error alone.

Race tracks the same pattern. About 15% of Black and Hispanic adults are credit invisible, versus 9% of white and Asian adults, and another 12-13% of Black and Hispanic adults have unscored files versus 7% of white adults — close to three in ten Black or Hispanic adults can't produce a score, roughly double the white and Asian rate. The CFPB notes the gap opens in consumers' late teens and mostly persists rather than closing with age.

The model doesn't read "unknown" as neutral. In practice, it prices unknown the same as risky.

None of this requires anyone to have done anything wrong. A file goes thin because someone paid off their only credit card and stopped using it, moved to the U.S. with no domestic history to import, or has simply gotten by on cash and a debit card. A file goes invisible because a person is 19, or has spent years proud of never borrowing a dollar. The credit system doesn't distinguish "never needed to" from "never got the chance." Both come back blank.

What No File Actually Costs

Landlords, insurers, and lenders don't treat a blank file as neutral information. Most rely on it as a signal, and a missing signal tends to get filled in with the worst-case assumption.

Renting is the clearest version of this. Tenant-screening reports typically bundle credit history in with eviction records and criminal background, and a thin or missing credit file routinely triggers the same response as a bad one: a higher security deposit, a demand for a co-signer, or an outright denial, decided by an algorithm the applicant never gets to see the logic of. The National Consumer Law Center's 2023 survey of tenant-screening practices found the process is opaque enough that renters frequently can't tell which factor sank their application, and that the burden falls hardest on Black and Latino applicants regardless of their actual payment history.

Insurance works on a separate but parallel track. In states that allow it, roughly 95% of auto insurers and 85% of home insurers use a credit-based insurance score to help set premiums, according to FICO figures cited by the National Association of Insurance Commissioners. Regulators generally bar insurers from using that score as the sole reason to deny or cancel a policy — but they can still use it to price one. When an insurer gets no hit at all, the FTC's 2007 report to Congress on credit-based insurance scoring found those "no-hit" consumers frequently pay more, not less, than the average scored customer — and that the scores correlate closely enough with race that using them predictably raises average premiums for Black and Hispanic drivers. Separately, rate-comparison analysis puts the gap between excellent and poor credit at roughly 67% for full-coverage auto insurance, about $1,581 more a year. A blank file doesn't get sorted into "excellent" by default. It gets treated like the worse end of that range, because the model has no evidence to put it anywhere else.

Then there's the workaround economy a missing score pushes people into. A cell carrier or utility that can't verify your history will typically ask for a deposit of a few hundred dollars just to turn service on. A secured credit card — often the fastest legitimate way to start building a file — requires putting up $200 to $500 of your own cash as collateral before it extends a dollar of credit back. And a share of this population never enters the banking system at all, instead relying on check-cashing counters and money orders — convenient, immediate, and structurally incapable of ever generating a credit file, because none of it gets reported anywhere.

The Fix Has Its Own Bill

The system knows this population exists and has started building workarounds — just not free ones. Since March 2023, the Federal Housing Administration has required lenders to factor positive rental payment history into its mortgage-underwriting algorithm, and in 2022 Fannie Mae and Freddie Mac both announced plans to incorporate cash-flow data like bank-account history into their own underwriting, specifically to reach borrowers with thin or missing files. Rent-reporting services, which submit a tenant's monthly payments to the bureaus so they start counting toward a score, exist for the same reason.

Each fix carries a catch a person with no cushion can least afford. The National Consumer Law Center's 2022 analysis of rent-reporting programs found that many charge the tenant a monthly fee to report payments already being made on time for free, require a formal lease and often a bank account most credit-invisible renters don't have, and — worst case — can convert a single late payment into a new negative mark on a file that previously had no mark at all. A secured card works the same way in miniature: the deposit has to come from somewhere, and a missed payment on it can leave someone worse off than the blank file they started with.

The Federal Reserve's own October 2025 research points at the irony directly. Alternative data — cash flow, rent, utility payments — can identify what some researchers now call "invisible primes": people inside the unscored and invisible population who would actually qualify as low-risk borrowers if a lender had the data to see it. The 32 million adults with no score aren't a monolith of bad risks the system correctly avoids; some real share are exactly the customers a lender wants and simply can't find. The system's default price for "we don't know" is identical to its price for "we know, and it's bad" — and proving you're the former is itself a transaction with a fee attached.

And Then You're Dead

The number was wrong for ten years, and correcting it changed nothing for anyone actually standing at a leasing office with a blank file. Thirty-two million adults still get quoted the deposit, the premium, and the cosigner requirement that "unproven" and "risky" share by default — and the tools sold to fix that come with their own fee attached, payable up front, by people who by definition don't have much up front to give.

That's it. That's the whole thing.

Read about the cost of having no bank account either · or see every essay