Why the Average Barely Moves

The state-wide credit market is the subject here, but the useful focus is the gap between its average and its distribution. The New York Fed study finds only a small aggregate decline after the consumer rate caps even though the riskiest borrowers lose substantial measured credit. Seeing how those results fit together helps readers avoid the common mistake of treating a quiet average as proof that a policy had no meaningful effect.

The average is a mixture, not a typical borrower

At the state level, the study estimates that total household debt balances fall by about 0.9 percent after the limits take effect, while the number of open accounts falls by about 0.6 percent. It finds no change in delinquencies for the average borrower. Those numbers describe a comparatively calm market-wide surface. They do not imply that every borrower experienced a change close to zero.

An average is formed from households with different exposure to the policy. The state laws targeted selected high-cost consumer products rather than every form of borrowing. Many households may have had no relationship with a payday, installment, or auto-title lender and therefore little reason for their measured credit to respond directly to a 36 percent cap. Their stable balances and accounts are part of the state-wide total. They can dilute the effect on the households that depended more heavily on the affected market.

This is the first reason the average borrower is a poor proxy for the borrowers most exposed to a high-cost rate limit. The study’s credit-bureau sample includes households across the risk distribution, not only people seeking the regulated products. A household with conventional credit access can appear in the same state and quarter as a household for whom a capped loan was one of few available options. The state-level estimate gives both households equal presence in the aggregate calculation even though the policy’s economic force is very different for them.

The second reason is offsetting movement. The study’s high-risk analysis finds a decline in debt and accounts for the bottom risk decile. Its reallocation analysis finds increased measured credit for some of the second through fifth deciles. When a group loses credit and another group gains credit, the two changes partly cancel in a state-wide total. A small net change can therefore coexist with a large change in who receives credit.

The cancellation need not be exact, and it need not involve the same lender. Imagine a market in which the highest-risk group becomes less attractive to serve under the cap while a somewhat safer group becomes more attractive. The first group’s accounts can close or fail to renew; the second group’s accounts can expand. If the first group is smaller than the rest of the market, or if the second group’s gains offset part of its loss, total accounts may move only slightly. The distribution of access has changed even though the headline total has not moved much.

The same logic applies to balances. A reduction in the amount owed by high-risk households can be offset by larger balances among moderately safer households, or by stable balances among households outside the affected market. The state-level debt estimate is not a ledger showing that each dollar lost by one group reappeared in another. It is a net outcome across groups with different exposure and different responses.

There is simple arithmetic behind the intuition.