How to Read a Non-Performing Loan Tape: Fields, Red Flags and Formats

A loan tape is the spreadsheet a seller sends prospective buyers with one row per loan for sale: balance, lien position, borrower, property address, payment dates and status. Reading one means finding the seven fields you need to price a loan, then working out what each seller's column labels really mean, because no two sellers build a tape the same way. On a top-20 bank's tape I worked with, the unpaid balance on a pool of mostly second liens sat in a column labeled "First Principal Balance."

What Is on a Loan Tape, and Why Does Every Seller's Look Different?

The name is a holdover from when loan data was stored on magnetic tape. Today it is most often an Excel spreadsheet, and in the deals I see it comes in one of three shapes: a raw export from a bank's loan system, a cleaned-up spreadsheet from a trader who resells loans, or a single-loan listing on a retail marketplace.

What US whole-loan sellers don't share is a format. A bank selling non-performing loans is usually running a loss mitigation exercise, not a trading desk, so you get whatever its system outputs, with its own headers, abbreviations and internal codes. Europe has gone the other way: since 2023, an EU regulation has set standard templates "for the provision to buyers of information on their credit exposures," which banks use when they sell non-performing loans (Commission Implementing Regulation (EU) 2023/2083). The European Banking Authority drafted them.

On a US tape, the translation is your job. Before you price anything, find the column that really holds the unpaid principal balance, confirm the lien position of every loan, and check that names and addresses are clean enough to send to your vendors. Reading the tape opens the first phase of our mortgage due diligence checklist, and the rest of the checklist depends on getting it right.

The Seven Fields You Need Before You Can Price a Loan

Without these columns you can't price a loan or start due diligence. If one is missing, request it from the seller or gather it yourself before you bid.

With all seven populated, you can start preliminary pricing. Everything else either speeds up the work or sharpens the number.

Fields That Speed Up Pricing, and How to Get Them Yourself

Professional sellers often include these. When they don't, every one of them can be gathered through your own vendors and public records.

Two of those deserve a closer look because state law turns them into pricing inputs. Property taxes first: in Florida, taxes are "a first lien, superior to all other liens" on the property (Fla. Stat. §197.122), so a delinquent tax bill comes out of your equity before your mortgage does. Then maturity: in Florida, the lien of a mortgage terminates "5 years after the date of maturity" when the final maturity is ascertainable from the record (Fla. Stat. §95.281). A secured status scrub has traps of its own, and our guide to secured vs. unsecured mortgage debt walks through them.

Check the Tape Before You Check the Loans

Bad data usually shows up at the file level before it shows up on a loan. A few minutes on these checks saves you from pricing a pool off a broken spreadsheet:

• Find the as-of date. Balances, delinquency and credit data are snapshots, and the older the snapshot, the more the numbers you price on may have moved.
• Tie the totals. The loan count and total UPB on the tape should match the seller's email or bid instructions. If they don't, ask which is right before you build anything on it.
• Look for duplicates and blanks. Sort by loan number and by property address. A repeated loan number, or two loans on one address with no lien position, needs an answer from the seller.
• Test the date logic. A next due date earlier than the last payment date, or an origination date after the maturity date, means a column is mislabeled or the export broke.
• Ask for the data dictionary. Some sellers define their codes on a second tab of the spreadsheet. If yours didn't, ask for one in a single email rather than ten.

The full scrub, from cleaning property addresses to building a seller request list, is in our guide to analyzing a loan tape step by step.

Three Real Tape Formats, Zero Standardization

In 2022 I compared three real tapes side by side: one from a top-20 bank, one from the secondary market and one retail listing. The first two cover the same portfolio. My client bought loans off that bank's tape, and my team rebuilt the data to resell part of the portfolio to local note investors, so the comparison shows the same assets before and after a cleanup.

A top-20 bank's tape

The bank's tape was a raw export from its loan system, running out to column AM. Reading it left to right:

• Internal fields first. An investor ID that identifies the beneficiary inside the bank and is no use to a buyer, an "LOB" code whose meaning you would have to confirm with the bank, and a notation tied to a successor by merger, meaning the bank had acquired a smaller bank and these loans came over with it.
• A vendor scrub under an opaque label. A column called "PROLI results" turned out to be a secured-status scrub. "Subject lien not matched" on a loan means either the mortgage isn't recorded on title or the vendor couldn't identify it. Research that loan before you price it.
• A mislabeled balance. The UPB sat under "First Principal Balance" even though most of the loans were second liens. I knew that code covered both first and second liens only because I had worked with that bank before. A first-time buyer could read it as the senior balance and get every equity calculation wrong.
• Gaps that look like data. Some interest rates were zero, probably on adjustable-rate loans. FICO scores had no date. Borrower and co-borrower names sat in single fields that you have to split into first, middle and last names before most vendors will take them.
• Status codes worth reading. There were no loans in active bankruptcy, which matched that bank's practice of leaving them out of its sales. Foreclosure and litigation codes were blank, and a loss mitigation status code of "R" needed the bank's definition.
• What was missing. Property value, senior lien status and balance, and property taxes: the inputs for every equity calculation.

Bank sellers notice which buyers make their lives easy. Work out what you can on your own (an unexplained code sitting next to the origination fields can be checked against the collateral documents) and save your questions for genuine ambiguity.

The same loans after a secondary-market cleanup

The resale tape ran out to column BN. What the extra columns added:

• Three identifiers per loan: asset ID, bank loan number and servicer loan number, which makes cross-referencing vendor data faster.
• Five color-coded pools, three junior-lien and two senior-lien, so buyers could go straight to the niche they buy.
• Credit data updated in July 2022: FICO, senior lien status (current, unknown or 120 days late), senior balance, past-due amount and pay string, with the prior senior status alongside for comparison.
• Value from more than one source, plus a rent estimate for modeling an REO exit, the property type and HOA dues.
• Vendor-ready borrower data: names split into first, middle and last, and a mailing address alongside the property address, so you can see whether the borrower lives in the house.
• Tax balances with links to the tax records, HELOC or conventional amortization, full origination details, and the charge-off amount, last payment date, next due date and charge-off date.

With the bank's tape your job is gathering and cleaning data; with this one it's verifying it, which is faster and cheaper. Even this tape carried the line "information deemed accurate but not guaranteed," and that is why you still check every material figure before you fund. For the credit columns in particular, our borrower credit report guide shows how to read the pay string.

A retail marketplace listing

Retail platforms show one loan per screen instead of a spreadsheet row. The listing I reviewed, on Paperstac, was a non-performing first lien with the LTV already calculated, a narrative from the seller and a payment history. Two details did most of the work. The total payoff was about $80,000, roughly $14,000 of it arrears and late fees on top of the principal. And the payment history showed the seller had bought the loan as a performer in 2018. It had been listed for 26 days without selling, and my read was that the seller's pricing guidance was steep for a loan that had stopped paying.

Retail listings suit buying one loan at a time. They don't scale to pools.

Red Flags to Screen on Any Tape

Whatever the format, these patterns call for a closer look, or a pass, before you spend money on research.

On the bankruptcy row, the rule is statutory: a petition operates as a stay of "any act to create, perfect, or enforce any lien against property of the estate" (11 U.S.C. §362(a)).

What the Tape Can't Tell You

A tape says a loan isn't paying. It doesn't say why, what shape the title or the house is in, or how long the resolution will take. Two loans my team bought and resolved show the gap between the inputs on the tape and the outcome:

On the second lien, the equity on the tape made a payoff likely sooner or later, even with a silent borrower. On the land loan the dollars worked, but nothing on the tape showed the clock, and three years of holding is what cut the return. The other blind spots are just as real: the borrower's situation comes from outreach through your servicer, title problems such as a broken chain of title surface only in a title search, and property condition needs a BPO or an inspection.

A Seven-Step Workflow From Tape to Bid

1. Run the file-level checks. As-of date, totals, duplicates, date logic and the data dictionary.
2. Confirm the seven pricing fields on every loan and flag the gaps for one seller request.
3. Split by lien position. First and second liens need different pricing models, so give each its own working tab.
4. Calculate equity coverage. For a first lien, subtract taxes and anything else ahead of you from the value. For a junior lien, also subtract the senior lien payoff.
5. Screen the red flags, cheapest first. The pre-bid waterfall orders the checks so free eliminations come before paid reports.
6. Decide what can wait. Some missing data has to be filled before you bid; the rest belongs in the due diligence window that opens when your letter of intent is accepted.
7. Price and submit. For pools, cursory research and indicative pricing gets a defensible number on every loan; for valuation methods, see how to value NPLs to make an offer.

The rule I use: a loan gets paid research only after it passes the file checks, carries all seven fields and clears the red flags. Everything else on the tape is a claim to verify, and if a claim turns out to be wrong after closing, your recourse lives in the representations and warranties of your loan purchase agreement.