How to score loan-level tapes when comparing residential bridge debt funds

The 1.1% versus 9% gap is the whole post and it's worth pushing on the recount. Counting every second extension as impaired is a blunt instrument. Some sponsors extend because the sale contract slipped 60 days, and that loan pays off in full with a fee. Applying the same rule to the 61-loan fund's three extended loans might well have produced a number that changed the conclusion. Small denominator, so three loans could easily be 8% of balance there too. Not arguing the conclusion. Arguing that the rule needs to be applied symmetrically or the comparison is doing less work than it looks like.

@hannah_west29 fair, though a first extension on a rehab loan is normal and a second one usually means the contractor is gone or the budget's blown. That pattern shows up consistently from the construction side. The distinction between one extension and two isn't arbitrary.

The recount is the right instinct and @hannah_west29's objection about symmetry is correct, so apply the rule both ways and see what survives.

On the specific gap you found: a reported default rate and an extension count measure different things and neither is wrong on its own. Default typically means a payment or covenant breach that the manager has declared. An extension granted before maturity is a contractual amendment, so the loan never breaches and never enters the default number. In a book of short-duration bridge paper written into 2021 and 2022 pricing, a meaningful share of extensions is the ordinary state of the world. The question your scoring should answer is whether the extended loans are still paying cash interest and whether the collateral value supporting them has been re-tested since origination. You did the second part by using the fund's current value opinion rather than the appraisal, which is the more useful of the two adjustments.

The 96 cent recovery is one data point on one small loan. It tells you the process worked once at a low balance in a liquid asset class. It says close to nothing about how a $4m stalled construction loan in the same book would resolve, and recovery timelines vary by state because foreclosure is judicial in some and non-judicial in others, which changes carrying cost before you get to price.

What would sharpen your next pass is realized loss by origination year across the manager's full history, not just the current book. The current tape only shows loans that haven't resolved yet. Everything that already went badly has left it.

How this income is treated in your own return depends on the fund's structure and your situation, and that's a question for your accountant.

Three evenings of spreadsheet work before putting money in is more diligence than I've done on anything in my life. Genuinely recalibrating.

The fund that says the tape is available after subscription is the detail worth rereading. Do managers actually expect that to land? A more reasonable answer, and one investors see often, is that the tape is available under NDA before subscription. That version respects diligence. After subscription isn't diligence, it's a reveal.

@sonia_alvarez I got the same line once and the explanation was borrower confidentiality. Then the other fund sent me borrower names redacted and everything else intact, which solved the same problem in about ten seconds.

What did giving up 170 basis points of net target cost in actual dollars over the 14 months? On 60k that's roughly $1,200 a year of forgone yield if the 9% had paid as advertised. Worth it for the sleep, but worth saying out loud that the safer fund isn't free.

Asking the beginner version. When you say recovered at 96 cents on the balance including accrued, does that mean the fund got back 96% of what it was owed and the LPs ate the other 4%, or did the fund's reserve absorb it and distributions never moved?

@jessreyes12 in most structures a shortfall that small gets absorbed against income for the quarter and would never show up in a distribution, which is exactly why realized loss history matters more than any single month's payment. What happens if the 60k needs to come back out? Closing windows going in were mentioned, but nothing about the path to getting out.

Move one variable. Say the 61-loan fund's top metro goes soft and that 14% concentration doubles in problem terms. Sixty-one loans is not a lot of diversification. Picking the cleaner book but the smaller one means the second thing partly undoes the first. That's still the right call to make. But it's worth knowing what the fund's plan is if two loans of that size go sideways in the same quarter.

The line worth taking away is that you never found out whether the chart fund was fine. That's an honest ending, and most people would have gone back and rationalized it.