A credit score is a prediction, and everything in the file is there because someone believed it improved the prediction. Medical collections were the rare item that failed on its own terms. They arrived through billing disputes and coverage errors as often as through inability to pay, they were wildly inconsistent in how and whether they were reported, and they turned out to be a poor guide to whether a borrower would repay a car loan.
Removing them was therefore not generosity. It was a correction, and its effect on households has been real: scores rose, some by enough to change the price of credit, and the increase was concentrated among people whose only blemish was a hospital bill they were still arguing about.
The signal was doing two jobs
Here is the complication. Medical debt predicted repayment badly, but it did correlate with something lenders care about a great deal: a household under sustained financial strain. Stripping it out improved the fairness of the file and removed a piece of information that was, in aggregate, telling lenders something true — just not the thing it was formally measuring.
Lenders responded the way they always do, by finding it elsewhere. Cash-flow underwriting reads the bank account directly, and a bank account shows a hospital payment plan whether or not a bureau records it. The information did not leave the system; it moved to a place with less regulatory scrutiny, no dispute process comparable to the one governing credit files, and far less visibility to the consumer being assessed.
This is a pattern the consumer credit system keeps repeating. Buy now, pay later grew up and got reported for the opposite reason — invisible obligations were distorting the picture — and both episodes make the same point: the file is an approximation, and pressure applied at one point relocates rather than resolves.
The underlying cost has not moved at all, which is the part households experience. Medical debt is a symptom of an expense structure that keeps rising alongside every other non-discretionary line, and a family absorbing higher premiums and deductibles is in the same position it was in before, with a better score. Set beside insurance behaving like a second inflation and a property tax bill catching up with the housing boom, the improvement is real and narrow: the household is not paying more for credit because of a hospital bill, and is still paying the hospital bill.
For lenders the practical problem is a period of worse information rather than better. Scores rose for millions of people without their circumstances changing, which compresses the distribution and makes the score less discriminating at exactly the point where it is used most heavily. Underwriters describe recalibrating against a population that suddenly looks more creditworthy than it did, and being unsure how much of the improvement is a correction and how much is noise.
The reasonable expectation is not that risk assessment got kinder. It is that it got less visible, and that the next argument will be about the data nobody has a right to inspect.
Topics moneycredithousehold finance
