My collateral turned out to be the borrower's LLC interests, not the houses

Four months framed and open in the southeast means the sheathing got wet and somebody paid twice for it later. The draw schedule is the part that bothers me most here. On a six house program the draws should be tied to inspected completion percentages per address, not to a calendar. If the sponsor can draw on a calendar he has no reason to finish anything fast, and your money funds the slow ones while the fast ones refi out and disappear from the pool. Did the offering say who signed off on draws?

A pledge of membership interests and a mortgage are different collateral, and the difference is the whole story here.

A mortgage or deed of trust puts a lien on the real property itself. Enforcement runs through that state's foreclosure process, judicial in some states and non-judicial in others, and the lien position is recorded so you can see who is ahead of you. A pledge of the equity in the borrowing entity gives the lender a security interest in the LLC interests, usually enforced under Article 9 of the UCC as adopted in that state, generally by a public or private sale of those interests. What a buyer gets at that sale is the company, subject to everything the company already owes. Senior construction debt, mechanics liens, unpaid taxes and accrued fees all survive, and the equity holder sits behind all of it.

Two things drive whether an equity pledge is worth anything. The first is how much property level debt sits in front of the entity and how it amortizes. The second is whether the operating agreement lets the interests transfer on default without member consent, since a pledge over interests the company can block from transferring is close to decorative. Separately, control of the draws matters as much as the collateral, because a borrower who can draw ahead of completed work can strip value from the collateral before anyone declares a default.

The assumption your outcome depended on was that senior debt would never sit between the note and the houses. Nothing in the collateral language rules that out. Whether a specific pledge is enforceable and what remedies are actually available depends on state law and the documents, so a creditors' rights attorney should read the security package before you fund, not after.

The distinction has finally landed for me after months of reading about notes. So when the offering says secured, I need to ask secured by what and recorded where. Is there any way to check the recording myself before funding, or do you only get the address list after you commit?

Sorry that happened. The part I keep rereading is that the extension was the sponsor's election. That never occurred to me as a term you could just have.

@harrow is right about the draws but I'd go further. Six houses under one entity with one draw schedule means the sponsor can rob the good ones to feed the bad ones and the reporting will still look fine at the pool level. On my sites the bank sends an inspector per address per draw and the release is per address. Portals almost never buy that, because the inspection cost eats the spread on a $30k slice. That is the honest reason it isn't there.

Who actually did the workout? Was the special servicer part of the platform or a third party they hired at that point?

11.5% for gap capital behind bank construction debt on 1960s auction stock is not a mispricing you found, it's the market telling you where you sat. Equity behind a construction line on flips should be looking for a mid to high teens number with real controls, or it shouldn't fund. The yield was the disclosure. Everything else in the doc was decoration around that number. I don't think the pledge language is the failure, arbor. The failure is that the coupon was consistent with senior secured paper and the position was not, and that gap is checkable in about ninety seconds once you know to check it.

Running your cash flows: 30,000 out at t0, 287.50 in monthly for 11 months, 8,900 at month 34, 2,340 at month 41. That's 14,402 back on 30,000, so a 52% loss of principal-equivalent, and annualized you're somewhere around negative 19% depending on how you date the recoveries. The month 34 and month 41 timing costs you more than people assume. If the same $14,402 had come back at month 14 you'd be near negative 30% total but only about negative 26% annualized on a much shorter clock, and you'd have had the money back to redeploy. Slow recoveries are their own loss.

Half a year of reading and I still have not opened a single security agreement. Going to fix that this week.

On flip crews the subs are always last in line and always surprised about it. Turns out capital works the same way.

Asking the embarrassing one so somebody else doesn't have to. If I own the membership interests of an LLC that owns a house with a bank loan on it, and the bank forecloses, I get nothing and I also own an empty company. Is that right? And is that different from just being unsecured, or is it worse because it looks secured on the page?

@compass it's the same recovery as unsecured in most bad outcomes, and the paper looks better, which is the actual harm.

arbor, the phrase I'd hunt for in your docs is in the security agreement, usually a sentence letting the collateral agent release or subordinate collateral without noteholder consent, sometimes worded as "in its sole and absolute discretion." Pair that with an extension option held by the borrower and you have a note where every date and every piece of collateral is optional to everyone except you. I've passed on two portal debt offerings for exactly that pair of clauses. The second one funded in an hour anyway, so my caution bought me nothing except sleep.

Receiverships on scattered single family are miserable. Six houses in different neighborhoods, each needing utilities on, lawn cut, and someone to chase the copper thieves, and the receiver bills hourly for all of it. That is a real part of where your recovery went between month 26 and 34.

What did the workout cost you in fees? Was it netted out of the 8,900 or billed separately somewhere in the statements?

I'll disagree with @sextant a little. The coupon was low for the risk, agreed. But the thing that actually killed the recovery was that all six houses sat under one borrower entity with one sponsor. Six separate notes, six separate borrowers, and two bad houses don't take the other four down with them. arbor bought sponsor concentration and called it diversification because there were six addresses in the deck. I do this due diligence for a living and I have made exactly that mistake with a five property pool, so I'm not scoring points.

This is the kind of thread that makes me analyze for another six months, which I'm told is my problem.

The sponsor drew on schedule while houses sat open. That's the sentence I'd put on the wall.

I keep landing in the same place after reading these. If I can't tell in one pass whether there is a recorded lien in my favor and what is in front of it, that's a pass for me, even if the yield is lower somewhere else. Boring is fine. I'd rather take 7% on something stabilized and understand where I sit.

arbor, if you saw the same offering tomorrow, what specifically would you ask the platform, and did they answer questions like that at the time or just point you back at the PPM?

Slow reader here, so two questions. When the extension fired at month 12, did the interest keep accruing at 11.5% or did it step up? And when the note went into workout, did you keep getting monthly statements with a value on them, or did the position just go dark until month 34? I ask because the reporting gap is the part I'd handle worst.

@meridian for what it's worth, the two questions I now send before funding any portal debt slice are: send me the recorded instrument number and county for the security, and send me the current payoff on everything senior to it. If the answer takes more than a business day or comes back as "the collateral structure is described in the offering materials," I'm out. That filter has cost me some deals I'd probably have been fine in. It has also kept me out of one that looked exactly like arbor's.

Same structure shows up on land and entitlement paper constantly, and it's worse there, because the entity holds dirt that produces no income while the clock runs. At least arbor's collateral had houses that somebody eventually wanted.

Reading this next to everything I've been learning about tenant screening. Same shape. The document tells you what happens when things go wrong, and that's the only part worth reading closely.

41 months of dead capital is the loss people underrate. The 15,600 you never got back hurts once. The 30,000 that couldn't do anything else for three and a half years hurts the whole time.

@orchard that's the piece I tried to put a number on above and it doesn't fit neatly in an IRR. If arbor's alternative was a boring 6% somewhere liquid, the opportunity cost over 41 months on 30,000 is roughly another 6,900 of forgone earnings. Add it to the realized loss and the deal cost about 22,500 against that baseline. I don't love stacking hypotheticals on top of real losses, but it's the right frame when you're deciding whether to fund the next one.

Twice through this and I'll probably go back to the expert reply a third time. The idea that "first priority" can be technically true and still put you last is not something I would have caught on my own.

Thanks for posting the actual statement dates, arbor. Most loss posts stop at "it went bad" and the month 26 to month 41 stretch is exactly the part I had no picture of.