A 6 month interest reserve funded from loan proceeds buys the borrower time but costs the lender a warning signal

@jordan_ellis2 yes, exactly that. @wendy_park48 take the position of a borrower carrying two loans at once, one funded by a reserve and one paid from pocket every month. The reserve loan tends to feel comfortable. The one paid out of pocket is the one that gets lost sleep and priority, because if a wall gets hit, the lender making noise gets paid first, and a reserve lender by definition isn't making noise. Structuring a loan around a silent reserve can put a lender in second place in a borrower's attention. The unit completion trigger fixes some of that. It also means that when the borrower misses the trigger in month 4, there's a decision to make in month 4 instead of month 7, and it's worth deciding in advance what that decision would be, because most lenders in that spot do nothing at all.

Your stabilized number is carrying the whole loan and I don't think it holds. $1,150 across 6 units is $82,800 gross annual. Take 5 percent vacancy and credit loss, call it $78,660, and if this is a tired brick walkup your expense ratio is probably 42 to 48 percent, not 35. At 45 percent you're at $43,263 NOI. At a 7 cap that's $618k, at a 7.5 it's $577k. You said $640k, which requires a 6.75 cap and a 42 percent expense load on a building with turnover you're funding right now. So call stabilized $590k. Your $412k is 70 percent of that, not 64. Add the $23,690 reserve, which is real principal outstanding, and the ratio is what it is. That's still a fundable loan. It just isn't the loan you described. The part that would actually move me is the exit. A DSCR refi at $412k plus costs needs the appraiser and the lender to agree with your $590k and needs the rents documented on leases with seasoning. What's his contingency if the refi comes back at $540k?

The reserve is a documents problem before it's an underwriting problem. Who owns that $23,690 while it sits. If it's characterized as his funds held by you, it may be reachable by his other creditors or swept into a bankruptcy estate, and how that shakes out depends on state law and on how the reserve agreement is drafted. If it's undisbursed loan proceeds you control, you're in a different and generally better position. That distinction lives in one paragraph of the reserve agreement and I've read plenty where it isn't addressed at all. Have a lawyer in that state draft it rather than pulling a form. On the cross-collateral, a separate deed of trust on the duplex is cleaner than adding it to the same instrument, and get its own title work and its own insurance certificate. Also check whether the duplex is truly free and clear or free and clear of mortgages with three years of unpaid taxes on it.

I'd take the cross-collateral off the table entirely. A $145k duplex he values himself, unverified, on a loan where your real problem is timing and information. It adds a second title file, a second insurance policy to monitor, a second set of taxes, and in a workout it gives him an argument that you have to proceed against both or neither depending on how your state handles it. You'd be buying $145k of claimed value with a pile of new operational surface.

What I'd want instead is the thing that costs him nothing if he's telling the truth. Monthly rent roll and bank statements by the 10th, and a right for you or your inspector to walk any unit with 48 hours notice. If he won't sign the reporting covenant, that tells you more than the appraisal does.

Milestone-based release, and the milestone should be leased, not turned. A finished unit with no tenant is a photo. A signed lease with a deposit cleared into the operating account is a fact that can be verified from the bank statement already being requested. A structure that works on this shape: three months of reserve at closing, then each subsequent month released only if trailing leased units meet the schedule, with a cure window. If the borrower falls behind, they pay from their own pocket while catching up. That's the whole point, the reserve stops covering them at the exact moment the plan stops working. @hannah_west29's expense ratio is the right fight to have. Worth adding that an as-is loan-to-value around 82 percent is the number that decides this loan, because a stall in month five means taking back a building worth roughly what it was on day one with half the units in pieces. That's not the as-is value anymore, that's the as-is value minus the demo that got started and not finished.

Genuine question from the cheap seats. Every mitigant in this thread adds work for the lender. Monthly statements, inspections, milestone tracking, two title files. At what point is this no longer the passive income it gets sold as?

@priya_n21 I've been wondering the same thing reading it. It looks like the passive part is the money, and the active part is everything that keeps the money from disappearing.

@priya_n21 that's the trade as far as I can tell. Lending through a fund hands the monitoring to someone else and you take a lower yield for it, direct lending pays you more and the difference is your labor. @wendy_park48 is doing the work here that a fund manager would do and keeping the spread. @wendy_park48, on your actual question, the rate doesn't look like it's pricing the reserve risk at all. 11.5 and 3 points is roughly what this borrower would pay without a reserve. So the lender is being paid the same for a loan with worse information. Keeping the reserve is worth a point more or fifty basis points, and if he says no to that he's told you what he thinks of the risk.

@chriswalton8 that last line is the clearest thing in the thread for me. If the borrower won't pay for the feature he's asking for, the feature is worth something to him and he'd rather you fund it.