Same gross coupon, 130bp apart on net, all of it the warehouse line

Who services the loans in each one? If A is self-servicing 118 loans with modifications running through it, that's a different operation than B outsourcing 64. I ask because I've watched a servicing shop drop the ball on escrow and nobody upstream knew for two quarters.

19 second extensions on 2021-22 multifamily bridge is the whole story to me. I've been the borrower on that shape of loan. A second extension usually means the renovation budget blew through, the rent bump didn't land, and the exit refi doesn't pencil at today's rates. The lender extends because the alternative is owning it.

Doesn't automatically mean loss. Extension fees are real money and the sponsor is still paying. But I'd want to know how many of those 19 are current on cash interest versus accruing, because those are very different animals and both get reported as performing.

Dumb question, what is a repo facility in this context? I understand a fund borrowing, I don't understand why the lender gets to determine marks.

Following. The bit about asking twice and getting a summary page would be my whole answer honestly.

What does a 5% quarterly gate mean in practice? If everyone wants out at once do I just get 5% of my money back per quarter until it runs out?

I don't think 0.85x is the problem, sextant. Advance rates on seasoned first mortgage collateral aren't wild and plenty of credit shops run more. The thing I'd chase is the margin call trigger. Is it a mark on the loan, a mark on the collateral, or a covenant on the fund's NAV? Those three fail at completely different moments.

If they won't send the facility docs I'd read that as the answer and take B.

Five weeks is a long time for a document that already exists.

Has either fund ever actually gated? Not the policy, the history. I'd ask A for quarterly gross redemption requests versus redemptions paid for every quarter since inception, and I'd ask for it as a table, not a paragraph. If they've never had a stressed quarter the gate language is theory.

Also the 90 days notice on a quarterly window means your real exit could be five months out from the day you decide. That matters more to me than 130 basis points.

Splitting 50/50 feels like it solves the discomfort rather than the question. You end up owning the tape you don't trust, just less of it.

@bramble that's fair but the other read is that A's 41% multifamily bridge concentration and B's 64-loan book both have concentration problems, just in different directions. 64 loans at $310m means average five million a loan. Lose two and you've eaten a year of coupon.

sextant, do you have B's largest-loan-as-percent-of-NAV? I'd want top ten exposure on both.

Nobody's asked about geography. 118 loans could be 118 loans in four metros. Where are they?

If any of that book is ground-up construction the draw administration is the risk that never shows up on a tape. I've been on the receiving end of a lender whose inspector came out every 45 days and signed off on work that wasn't done. That fund's loan reads as 62% LTC and the building is 40% complete.

Ask who does the draw inspections and whether it's an employee or a third party.

@harrow is describing me. I bought small multifamily in 2021 with a bridge loan I was sure I'd refi out of by 2023. I didn't. I got an extension, paid the fee, and I'm still current because I put more of my own cash in. From the fund's side I look like a performing loan. From my side I'm one bad quarter from a conversation.

What's the fee load on each? You gave net targets but not what's coming out to get there.

Running a small team taught me that when the answer takes three weeks it's usually because someone has to ask permission. Doesn't mean bad, means the document says something they'd rather explain over the phone.

Went through this twice and I'm still only catching maybe half. Saving it.

@rigging same. The part I did follow is that the higher return isn't coming from better loans, it's coming from borrowing at the fund. That reframed the whole thing for me.

I went through something similar last year and the thing that ended up separating the two managers wasn't the tape, it was whether they'd send one at all at loan level. One did, unredacted except for borrower names. One sent a chart with buckets. I scored each loan on vintage, extension count, and cash-pay status and the chart fund couldn't be scored, so it was out by default.

sextant, if A gives you 118 rows you can do the same in a weekend.

The extension fee income running through the return is the line I'd stare at longest.

Waiting five weeks costs you nothing but five weeks of B's 7.2 on part of the money. That's a small price for seeing the facility docs. I'd hold the whole thing rather than commit into a document you haven't read.

Any land or entitlement loans in either book? Those get bucketed as bridge sometimes and the collateral behaves nothing like an occupied building. No cash flow, no cushion, and the exit depends on an approval that may not come.

Asking the thing I'd be embarrassed to ask out loud. The 5.5% index figure is gross of fees and A's 8.5% is net, so A is claiming to beat that index by a lot more than 300 basis points once you gross it back up. Is that plausible from 0.85x leverage alone, or is something else in there?

@compass leverage plus extension fees plus origination points retained at the fund. Points can add a lot in a year when you're churning short paper. It's plausible. It's also the kind of income that stops the second originations slow down.

@flint that's the part that worries me about the whole segment right now. Fundraising was $51bn in final closes last year, highest since 2021. All that money has to get placed. Underwriting standards move when there's a queue of capital behind you.

Go back to the LPA on the leverage limit. "Up to 1.5x" measured against what? If it's 1.5x NAV and NAV falls because loans get marked, the ratio moves against you without the manager borrowing another dollar. Some documents measure against contributed capital instead, which behaves completely differently in a drawdown.

Also check whether the limit is a hard covenant or a target the manager can exceed with advisory committee consent. I've read three debt fund LPAs where the number everyone quotes in the deck is a target and the actual ceiling is a consent away.

Ask both managers for realized losses and cumulative charge-offs by origination year since inception, plus every loan that went into REO and what it recovered at. A manager with a clean answer sends it in a day. I've had two send it in a day and four send me a marketing deck.

The 2022 and 2023 vintages are the useful ones because that's when everyone was underwriting to exit cap rates that didn't happen.

Trying to get the order of who loses money. The repo lender is ahead of the fund's own equity, so if the collateral goes bad the lender takes the loans and the LPs eat it first? That's what I'm reading between the lines here.

@sable roughly, yes, and that's exactly why the 130 basis points exists.

sextant, on sizing, I'd separate two decisions you're currently making as one. Whether A is investable is a diligence question with an answer. How much of your allocation goes to any single manager is a policy question you should have settled before you looked at either. If your own cap is 25% per manager the 50/50 option was never on the table anyway.

Starting from nothing here and this thread is the most useful thing I've read all week. The idea that a fund can look identical on the surface and be a totally different risk underneath is new to me.

So the thing I'm still stuck on. Both funds hold paper. If A's borrowers keep paying, the leverage never matters and A wins by 130 every year. If enough of them stop, A loses more than B loses. Is there any version where A's structure helps in the bad case, or is the leverage strictly a fair-weather trade that you're paid for taking?