How to Verify a Fund Manager's Track Record — and What a Loan Tape Should Show

Anyone can claim a track record. Here is what a real loan tape contains, how to test the claims in it against public records, and what an incomplete tape is hiding.

Apurva Sanghavi · October 2, 2026 · 15 min read

A track record is a claim until you test it

Every deck in this business has a slide with a number on it. Loans funded. Principal deployed. Zero losses. Years in business. None of those numbers is evidence of anything until you know the denominator, the date range and what was excluded.

The useful thing about private lending, compared with almost any other private asset class, is that the underlying transactions are recorded in public. A first-lien mortgage loan on Texas real property leaves a deed of trust in the county clerk's records, and when it is repaid it leaves a release of lien. You cannot verify a manager's returns from public records. You can verify that the loans exist, roughly when they were made, against what kind of property, and whether they were released. That is enough to catch most of what goes wrong.

The sequence below is: get the tape, test the tape, check the tape against the county, then check the people.

What a complete loan tape contains

Every loan. Not the closed ones, not the good ones, not a representative selection. Every loan the manager has made since the date the track record claims to start, including ones still outstanding, ones in default, and ones written off.

Each row should carry at least these fields.

Field	Why it matters
Loan number or internal ID	Lets you reconcile counts and spot gaps in a sequence
Funding date	Vintage. Performance by vintage is more informative than the aggregate
Original principal	The denominator for every loss figure
Current or final balance	Distinguishes paydowns from payoffs
Collateral type	Single family, multi-family, commercial, construction, land. Risk differs enormously
Market or ZIP code	Tests whether the manager lends where it says it lends
Lien position	A first lien and a second lien are different businesses
Loan-to-value at origination, and the basis	As-is or as-completed. The basis is where optimism hides
Valuation source and date	Appraisal, broker opinion, or the manager's own view
Interest rate and points	Feeds the reconciliation of fund yield to loan yield
Original term and any extensions	Extensions are the earliest visible sign of stress
Status	Current, late, in default, in foreclosure, repaid, foreclosed, written off
Final outcome for closed loans	Repaid in full, repaid after workout, foreclosed with full recovery, foreclosed with loss, deed in lieu
Principal lost	In dollars, per loan, not as an aggregate percentage
Interest actually collected	Separates accrued from received
Repeat borrower flag	Concentration and, read the other way, evidence of underwriting

Borrower names and street addresses can reasonably be redacted. Dates, amounts, ZIP codes, loan-to-value ratios and outcomes cannot, because redacting those removes the ability to test anything.

The three tests

Completeness. Add up the rows. Does the loan count match the number in the deck? Does the sum of original principal match the "principal deployed" figure? Does the date of the earliest loan match the claimed start of the record? Look for gaps in an internal numbering sequence. If the tape has 48 rows and the website says 60 loans, the question is not rhetorical: ask what the other twelve were.

Survivorship. Are unresolved loans in the file? A tape consisting entirely of closed, repaid loans is not a track record, it is a highlight reel. Three specific moves to look for. Loans that were moved to an affiliate entity before they went bad, which shows up as a tape whose loan count is lower than the count of deeds of trust recorded in the county. A start date chosen to begin after an earlier fund or an earlier vehicle ended. And a defaulted loan recorded as "modified" and given a new loan number, so a single problem appears in the file twice as two performing loans.

Realisation. Distinguish cash from marks. Interest accrued on a non-performing loan is not income; it is a number in a spreadsheet that may never arrive. A loan carried at par while the borrower has not paid in seven months is a valuation judgement made by the person whose fee depends on it. Ask for interest collected in cash, per loan, and compare it to interest accrued. Ask, for any loan carried at full value, when the borrower last paid.

Those three tests fail differently. Completeness failures are usually carelessness. Survivorship failures are usually deliberate. Realisation failures are usually optimism, and optimism is the one that has cost investors the most money in private credit.

Verifying against public records

Take five loans from the tape at random, not from the top, and check them.

In Texas, a loan secured by real property is documented by a deed of trust, which is recorded with the county clerk in the county where the property is located. The borrower is the grantor. The lender is the beneficiary, and appears in the index as a grantee. When the loan is paid off, a release of lien is recorded, with the lender as grantor. When a foreclosure is completed, a substitute trustee's deed is recorded. All of it is public.

Walking through Harris County, which covers most of Houston:

Go to the Harris County Clerk's website at cclerk.hctx.net and open the real property records search. The County Clerk, not the District Clerk, holds real property records. Basic index searching is free; copies of imaged documents carry a small per-page fee.
Search by party name. Enter the lender's entity name exactly, for example the fund's legal name or the manager's legal name. Name indexing is literal, so run the variants: with and without "LLC", with and without commas, and any d/b/a the manager uses.
Set a date range matching the track record you were given, then filter by document type. "Deed of Trust" shows loans originated. "Release of Lien" shows loans paid off. "Substitute Trustee's Deed" or "Trustee's Deed" shows completed foreclosures. "Assignment" shows loans sold or transferred to another party.
Read a result. The index gives the recording date, the file number, the grantor and grantee names, and the legal description or address. Open the imaged document and you will typically see the original principal amount, the date of the note, the maturity date and the property description. Recording usually follows funding closely, so the recording date is a reasonable proxy for the funding date.
Compare. Pick your five tape rows and find each one: does a deed of trust exist, recorded at about the date claimed, for about the amount claimed, on the property type claimed? Then run the search in reverse. Count all deeds of trust recorded in the lender's name during the period and compare that count to the tape. If the county shows substantially more loans than the tape does, ask which ones are missing and why.
Check the releases. For each loan the tape shows as repaid, a release of lien should exist and should be recorded at roughly the payoff date. A repaid loan with no release recorded, years later, is a housekeeping failure at best.
Check for trustee's deeds. Search the lender's name for substitute trustee's deeds. Each one is a completed foreclosure. Count them and compare to the number of foreclosures the manager disclosed. This is the single sharpest test in the list, because a foreclosure is very difficult to hide and very tempting to omit.

Other Texas counties run their own systems with the same document types under different interfaces: Fort Bend, Montgomery, Brazoria, Galveston, Bexar, Travis, Dallas and Tarrant all offer online party-name searching, with varying amounts behind a fee. Do the same search wherever the manager says it lends.

Two honest limits. Property held in a single-purpose entity or a land trust can make borrower-side identification difficult, though the lender side is what you are searching. And a lender who originates through multiple affiliate entities will not be found under one name, which is exactly why you ask for the list of entities before you search rather than after.

What to ask about every loss

The question is never whether there were losses. It is what happened in each one.

For each workout in the tape, ask: What was the loan-to-value at origination and what was the basis for the value? When did the borrower first miss a payment, and what did the manager do in the following 30 days? When was a current valuation ordered? How long from default to resolution? What were the total costs, meaning legal fees, taxes paid to protect the lien, force-placed insurance, securing and carrying the property, and brokerage on resale? What was the final principal recovery? And what changed in the underwriting afterwards?

The last question is the one that separates managers. A loss that produced a documented change in policy is a manager learning. A loss described entirely as bad luck, a bad contractor or a bad market is a manager who will produce the same loss again.

Listen also for specificity. A manager who has genuinely worked out a file can tell you which month the copper disappeared. One who has not will describe the process in general terms.

Manager background checks

Four searches, none of which takes long.

State entity records. Confirm formation date, standing and registered agent. In Texas the Comptroller's Taxable Entity Search at mycpa.cpa.state.tx.us/coa is free and shows whether the entity's right to transact business is active or forfeited; SOSDirect at direct.sos.state.tx.us charges a small fee and gives formation date, governing persons and filing history. Compare the formation date against the claimed record. Entities formed long after the record supposedly began need an explanation, which is often a legitimate one about restructuring.

EDGAR. Company search at sec.gov/cgi-bin/browse-edgar with filing type "D" for the fund's Form D, which lists the related persons and the amount sold to date. Then full-text search at sec.gov/edgar/search for the principals' names, which sometimes surfaces prior offerings.

NMLS Consumer Access, at nmlsconsumeraccess.org, searchable by company and individual. Read this one carefully: business-purpose lending against non-owner-occupied investment property generally sits outside consumer mortgage licensing, so most legitimate private lenders in Texas do not appear there at all. An absence proves nothing. A revocation, or a claimed licence that does not exist, proves a great deal.

Litigation. County district clerk civil records by party name, for the entities and the individuals. Harris County's is at hcdistrictclerk.com and is free to search. Then PACER at pacer.uscourts.gov for federal civil and bankruptcy matters. Interpret the result rather than counting it. A lender with no litigation history has probably never foreclosed. What matters is whether investors have sued, whether fraud has been alleged, and whether there are personal judgments or bankruptcies against the principals.

Also run the SEC Action Lookup for Individuals at sec.gov/litigations/sec-action-look-up and FINRA BrokerCheck at brokercheck.finra.org for each principal.

Reading a track record from a different vehicle

This applies to us, so I will state it about us.

A manager's record from personal lending, from an affiliate's balance sheet or from a predecessor entity is relevant information. It tells you how this person underwrites, how they price, how they handle a default, and whether they have seen a bad outcome. It is not fund performance, and the distinction is not a technicality.

Fund performance is what a pooled vehicle returned to its investors after management fees, fund-level expenses, the cost of cash sitting undeployed, any leverage, and the waterfall. A manager's gross lending record has none of those drags in it. A 10% gross loan book is not a 10% net investor return in any fund, and a record built one loan at a time on a balance sheet did not have to manage subscriptions, redemptions or an undeployed balance.

The other difference is verification. Audited fund financial statements are prepared by an independent accountant to a standard. A manager's own record is a spreadsheet the manager maintains. County records let you test parts of it, which is the whole point of the section above, but testing is not auditing.

Our figures are the manager's own lending record and are not audited fund performance. Since October 2019 we have funded 60 Texas loans totalling $13.3 million of principal and earned $1.66 million of interest on them, with a 13-month average term and an 11-month median. Twenty of the 60 were to repeat borrowers. Fifty-five repaid on schedule. Five went to workout: four foreclosures and one deed in lieu. Three of those five returned principal in full and two did not. Total principal lost across all 60 loans is under $50,000, less than 0.4% of principal funded. Those loans were funded by Matador Lending LLC, Matador Note Fund 1 LLC and affiliates. None of it is a forecast, and past performance is not indicative of future results.

Our own tape

We will send the whole thing, redacted of borrower names and street addresses only, with ZIP codes, dates, amounts, collateral types, loan-to-value ratios, terms, statuses and outcomes intact, including the two loans that did not return principal in full.

Then apply the tests above. Reconcile the count to 60 and the principal to $13.3 million. Check that the earliest row is dated October 2019. Take five rows and look them up in the Harris County clerk's records using the walkthrough in this post, using all three entity names. Count the substitute trustee's deeds and compare them to the four foreclosures we have disclosed. Ask about both losses by loan number.

If the tape does not survive that, you should not invest, and I would rather you found out with a spreadsheet than with a subscription agreement.

Frequently asked questions

Q: What is a loan tape?
A: A loan tape is a line-by-line record of every loan a manager has made, not a selection of them. Each row should carry the funding date, principal amount, collateral type, market, loan-to-value, term, rate, current status and, for closed loans, the outcome and any principal lost. The row count and the principal total should reconcile to whatever headline numbers the marketing uses.

Q: How do I verify a real estate fund's track record?
A: Test it three ways. Completeness: do the loan count and the sum of principal reconcile to the claim? Survivorship: are unresolved and written-off loans in the file, or only closed good ones? Realisation: are the returns cash actually received, or values the manager marked. Then spot-check individual loans against county deed of trust records, which are public and free to search.

Q: Can I look up a private lender in county records?
A: Yes. Deeds of trust are recorded in the county where the property sits and are public. In Harris County, the County Clerk's real property search lets you search by party name, so you can enter a lender's entity name and see the deeds of trust naming it, plus the releases of lien that show which loans were paid off. Search several name spellings and affiliate entities.

Q: What is survivorship bias in a track record?
A: Survivorship bias is what happens when the loans that went wrong leave the record. It shows up as a tape containing only closed and repaid loans, as a start date that quietly excludes an earlier vehicle, or as loans transferred to an affiliate before they defaulted. The test is straightforward: ask for every loan since inception, including write-offs, and reconcile the count.

Request the loan tape. We send it as a spreadsheet, in full, with the losses in it, and we will sit on a call while you check it against the county records.

Important disclosures. This article is for informational purposes only and is not an offer to sell or a solicitation of an offer to buy any security. Any offer is made only through the Private Placement Memorandum, Operating Agreement and Subscription Agreement of Matador Note Fund 1 LLC, which govern in all cases and should be read in full, including the risk factors. Interests are offered only to accredited investors as defined in Rule 501 of Regulation D, pursuant to Rule 506(c); accredited status must be independently verified. Real estate debt investments involve significant risk, including possible loss of principal. The 8% preferred return is non-cumulative and is payable only from available distributable cash at the manager's discretion; it is not guaranteed. The 8–12% figure is a target objective, not a promise or projection of results. Targeted returns are not guarantees of future performance, and past performance is not indicative of future results. Figures describing prior lending are the manager's own lending record and are not audited fund performance. Nothing here is legal, tax or investment advice. Matador Lending LLC is not a registered investment adviser.