The evidence

The record behind the diagnosis.

Problem is the diagnosis. This page is the record room: the public sources, corrected estimates, caveats, and identity-boundary facts behind the thesis.

Sourced record

What the credit snapshot misses

The diagnosis on The Problem page rests on four questions. How many people are invisible to the file? How hard can a single shock hit? How much fraud can a clean file hide? Where does identity verification stop?

~32 million

U.S. adults cannot be scored, including people with no credit history and files too thin to read. This is the broad “no readable file” population. Federal Reserve.

13.5 million

The CFPB corrected the famous credit-invisible estimate to about 13.5 million for 2010 and 7.0 million for 2020. Use this as the narrower “no credit record” estimate, not the full thin-file population. CFPB.

171 points

A single 90-day-late student-loan payment cut strong borrowers’ scores by 171 points. The better the starting record, the harder the fall. New York Fed.

2.2 million

Borrowers saw scores drop more than 100 points in one quarter of 2025 after student-loan delinquencies returned. That is a population-level version of the same “fall frozen in the file” problem. New York Fed.

$6–$35 billion

Annual synthetic-identity losses are modeled industry estimates, not measured losses. The range is useful directionally, but it should never be presented as a counted ledger of losses. Boston Fed.

1.35 billion

Victim notices involving Social Security numbers were tied to 1,825 breaches in 2024. The identity surface is large even before an applicant reaches underwriting. Identity Theft Resource Center.

How to read it

The sources support the diagnosis, not a single magic statistic

These sources do not say one number explains the whole market. They show a pattern: the credit file can be empty, delayed, over-punitive after a shock, or clean in the wrong way.

That is why The Problem stays short. The page names the four applicant stories; this record keeps the caveats, corrected estimates, and boundaries visible.

The evidence does not argue that scores are useless. It argues that a score alone is too thin for the edge cases lenders most need to defend.

The gap

Verifying the number is not verifying the person

The government’s number-checking service can confirm a name-and-number match, but it explicitly does not verify identity. That leaves room for a file that looks clean while the life behind it is missing.

SSA eCBSV documents the boundary. Kenshiki is built for the missing underwriting question: is there a real life here, and does it fit the loan?

The old system checks the number. Kenshiki checks the person.