Figures from the New York Fed are balances outstanding, not sale prices. Bands cited from DebtMarket’s pricing study are conversation ranges, not offers. This note is education for operating companies and creditors. It is not legal advice.
Buyers change a bid, or they pass, for a short list of reasons. The reasons show up in the tape and in a small sample. They do not show up in the adjective a seller uses for the file.
The five
| Signal | What the buyer does |
|---|---|
| Charge-off dates missing on a large share of rows | Treats the window as unknown and prices the worst vintage, or passes. |
| Media flag fails the sample | Bids the sample percent, not the flag. |
| Two products in one face amount | Asks for a split, or bids only the product in the box. |
| Seller name absent from the assignments | Stops until chain of title is visible. |
| No single cut-off date | Refuses to treat balances as final. Payments in between become a dispute. |
None of these is cured by a higher asking price. A seller who fixes the signal before the bid keeps the percent on a face both sides can see. The columns are on the data tape page. A pre-tape read of the snapshot is a valuation review.
Fix the signal or accept the lower percent
A seller who can fill the blank charge-off dates from the system of record should do that before the bid. A seller who cannot should expect the bid to assume the oldest window in the file. Shipping a corrected tape after the award is a new deal. Say so, and reprice, rather than calling the correction a true-up.
Compute the blank rates before the sample
Five signals move a bid: missing dates, a media flag the sample does not support, more than one product in the face amount, a chain of title that does not tie to the IDs, and a cut-off date that is missing or inconsistent. You can measure the first two before anyone opens a contract. Divide the blank rows by the row count.
| Column | Blank rate that changes the conversation |
|---|---|
| Charge-off date | Any material share blank. The bid then has to assume a window |
| Last payment date | Blank on a large share of rows. Recent payers and silent accounts are mixed |
| State | Blank rows cannot be matched to a buyer’s box |
| Originator | Blank originator stops a resale file and weakens a first sale |
| Media flag | A flag of yes on every row, with no percent, is unread until the pull |
Put the failed pull in the bid as a fraction
If 40 statements are requested and 22 open, the bid says 22 of 40. A seller who can fill charge-off dates from the system of record should do that before the award. A corrected tape after the award is a new delivery and gets a new look, including a new price if the window changed. The columns themselves are listed on the data tape page. How those signals turn into a percent is what moves the price.
A tape with the five rates written on it
Two thousand rows. Charge-off date blank on 240 rows, which is 12 percent. Last payment blank on 500 rows, which is 25 percent. State blank on 30 rows. Originator filled on every row, and the originator is the seller. Media flag set to yes on all 2,000, and a pull of 40 opens 19 statements. The bid email can say: window is incomplete on 12 percent of rows, last-payment is missing on a quarter of the file, 30 rows have no state, title is a first sale from the originator, media opened on 19 of 40. That note is more useful than a revised asking price. The seller can fill the dates from the system of record and send a new tape. The new tape is sampled again. The columns are the ones on the data tape page.