With two debt funds at the same 9 percent net, which single piece of diligence deserves the time

Tape is the right vote, though it's worth being precise about what one actually looks like. It's a spreadsheet with a row per loan, not a summary. A one page document with average LTV and a pie chart by state is marketing, not a loan tape, and any fund that sends only that in response to a diligence request is telling you something.

Tape, and it isn't close. A workout record is a story the manager chose to tell. A tape is a set of numbers that either reconciles to the NAV statement or doesn't. What matters per loan: original balance, current balance, lien position, origination date, maturity date, extensions taken, appraised value with the basis stated, days past due, and interest reserve remaining. That last column is the one people skip and it's where a bridge book hides its problems, because a loan funded with twelve months of reserve is current by construction until the reserve runs out. Sort by maturity, then by reserve balance, and the next four quarters become visible before the manager writes about them.

I'd fold vintage into the tape rather than treat it separately. A book originated against 2021 and early 2022 values carries a different cushion than one written in the last eighteen months at repriced values, even at identical stated LTV, and the origination date column tells you that for free. Geography does the same job. Two funds at 65 percent average LTV where one is spread across secondary markets with slow price movement and the other is concentrated in three metros that ran hard and gave it back are not the same fund.

Documents. The tape describes a moment. The LPA describes every moment after it. The clauses to read first are what constitutes a modification versus an extension, whether the manager can extend without it being called a modification and how many times, who has discretion to suspend redemptions and on what test, and whether the promote is calculated on accrued or collected interest. A fund can have a spotless tape and a document set that lets the manager hold capital for two years while the promote keeps accruing. Verity, ask both managers for the current LPA and the most recent amendment, and read the amendment first.

Some definitions, since a few of these terms are doing heavy work in the thread.

A loan tape is the loan-by-loan schedule of everything the fund holds, one row per loan. What vellum described, the one page with averages and a pie chart, is a portfolio summary. Averages hide the shape of a book. A fund with an average LTV of 65 percent might hold everything between 60 and 70, or it might hold a lot at 55 and a handful at 85, and only the row-level file tells you which.

Non-accrual means the fund has stopped booking interest income on a loan because it no longer expects to collect it. Until a loan is moved to non-accrual, interest can keep showing up in the fund's reported yield even though no cash arrived.

A gate is a provision letting the manager limit or suspend redemptions, usually when requests in a quarter exceed a stated share of net assets. It exists so the fund isn't forced to sell loans cheaply to pay exiting investors. It also means your exit date is a request rather than a certainty.

On the choice in the poll, the chapter this room is built on puts it plainly: the primary risk in debt investing is borrower distress exceeding the collateral cushion, which makes the vintage and concentration of the loan book the thing to examine. That points at the underlying portfolio. The fund documents decide what happens to you once distress shows up. Both funds should give you the loan schedule and the current governing documents in writing, and anything in those documents you're relying on should be reviewed by an attorney, since fund terms and how they're enforced vary.

The tape is worth sorting three different ways before anything else matters. Weighted average LTV by origination year, dollar concentration by borrower, and dollar concentration by property type. Sextant is right that averages hide dispersion, and the fix is cheap. Run the standard deviation of LTV across the book. Two funds quoting 65 percent where one has a 6 point spread and the other a 19 point spread are running different businesses. That's a fair first screen, and only after it does the manager question earn attention.

On any construction or heavy rehab loan the LTV column on the tape is close to fiction without knowing how the manager runs draws. Two approaches show up in practice. One lender sends an inspector before every draw, holds retainage, checks lien waivers. The other funds off photos and a spreadsheet from the borrower. Same loan, same stated 65 percent, completely different actual exposure, because the looser process has already funded work that isn't in the ground. For a book with construction in it, the manager is the deciding factor, and the question worth asking is who inspects and how often.

Dumb question from the back. If the loan pays every month, does any of this matter to me, or is all of it about what happens when it stops?

It's all about what happens when a loan stops performing, and that's exactly why the manager side deserves the vote. Originating a loan at 65 percent is not hard. Collecting on it is a specific operational skill involving servicers, counsel in whatever state the property sits in, tax and insurance advances, and someone willing to finish or sell a half-done building. Recovery rates across the same LTV band vary enormously by who is doing the work. Ask for completed workouts only, with dates. Anything still open is a forecast.

There's a check size problem nobody's said out loud. At 25k I'm not getting a loan tape from anyone. Two of the three funds I asked told me the schedule goes to institutional LPs and I get the quarterly report. So for someone at my size the poll collapses to documents, because the PPM and the LPA are the only things they're obligated to hand over.

Judge the manager, not the pitch, and do it by watching the shape of the answer to a hard question. Ask about their worst loan of the last three years. A manager who gives a specific, address free story with real numbers, including what they got wrong, is telling you something useful. One who answers with a paragraph about disciplined underwriting culture is telling you they would rather not. That single question reveals more than the spreadsheet ever will.

What happens if the tape they send is three months stale? The one I got was dated at the prior quarter end and by then they'd originated another sixteen loans. Am I diligencing a book that no longer exists, and is it reasonable to ask for a current one or does that mark me as difficult?

Compass, ask for two tapes, a quarter apart. The comparison is worth more than either one alone. It shows which loans moved to non-accrual, which took extensions, whether balances rolled or paid off, and how fast the book turned. A single tape is a photograph. Two tapes is the beginning of a direction. It also exposes whether reporting is consistent, since column definitions sometimes shift between quarters without anyone flagging it.

A common response from sponsors is to hand over a portfolio summary and cite borrower privacy instead of a full loan tape, which is worth taking with some skepticism. A workable middle version exists: anonymized rows with the borrower name stripped and the property described only by type and metro. That gives an investor everything the underwriting actually needs without naming anybody, and a sponsor who won't produce even that much is telling you something.

Anonymized rows is the standard ask, and most decent shops will do it. If a fund will not strip names and hand over the schedule, that is the answer.

The case for documents holds up whenever liquidity has a deadline attached to it, say an investor who needs redemption proceeds on a known date for a purchase. In that situation the gate provision is the term that decides whether the fund is usable at all. A great tape paired with a discretionary suspension clause still leaves that investor unable to plan. The redemption mechanics, notice period, queue treatment and any lockup are worth reading before the return figures, not after.

Adding to girder's point. Lender-placed insurance and how fast they catch a lapsed policy is a small operational thing that turns into a total loss if a property burns while uninsured. Nothing on the tape shows you whether the servicer is watching that. It's a manager question.

A comparison sheet across six funds tends to separate most on origination year, exactly as arbor said. Two funds writing most of their current book in 2021 and 2022 have been extending since, and the stated LTV hasn't moved because nobody's reappraised. That's a tape finding and a document finding at the same time, since the LPA decides how often collateral gets revalued.

Coming back after reading this, the documents matter more than the tape here. A gate is the thing that should scare an allocator most, because everything else in this thread is about a loss that might get taken and a gate is about not being able to leave at all. The definitions in this thread are the missing piece for anyone new to debt funds.

Sextant, agreed on one thing. A tape reconciles or it does not, and that is a real test the documents cannot give you. Where the disagreement holds is that the tape cannot tell you what the manager is permitted to do next. A current book at 12 percent office exposure means nothing if the LPA permits 35 and the manager has been drifting toward it. Read the hard caps and assume they will eventually be used.

LPAs in this space typically give the manager discretion on the gate rather than a mechanical trigger. The language is usually something like the general partner may suspend redemptions if it determines that satisfying them would be materially adverse to remaining investors. That's a judgment standard, and investors are relying on the manager's judgment at the exact moment their interests and the manager's diverge. The other clause worth flagging is the valuation policy for a loan in default. If the fund holds a non-accrual loan at par until a sale, the NAV redeemed at is overstated and the last people out eat it. Ask who sets the reserve, on what schedule, and whether anyone independent reviews it. Any conclusion about how these clauses would actually be enforced needs a lawyer, since it turns on the governing documents and the state they're written under.

Fathom and warrant have moved the argument part way. The valuation policy for defaulted loans is a document question the tape can't answer, and it's material to what an investor redeems at. Still, the main point holds. Document sets in this segment are more alike than different, drafted from a handful of templates, while the dispersion between two funds' actual loan books is enormous. If everyone's paperwork is roughly the same, the paperwork can't be what distinguishes them. The tape can.

Sextant's point about templates holds up broadly. PPMs across different sponsors are often structurally identical. The books behind them rarely are.

I do all four, and if forced to rank, the tape without servicing notes is thinner than sextant is allowing. A row that reads current tells you a payment arrived. It doesn't tell you the payment came from a reserve the fund itself funded at origination, or that the borrower called three weeks ago asking about a forbearance. So I ask for the tape and then ask for the last five completed workouts including the one that went worst, and I cross the two. If the workout list has no bad outcome in it across a three year book, either they've been lucky or they're not showing me everything, and both are worth a follow-up question.

Borrower quality shows up long before a missed payment, and property management is where you see it. Deferred maintenance, staff turnover, vendors not getting paid. A lender who never visits collateral finds out last.

Running numbers on the concentration point. Take a 60 loan book, equal weight, no fund leverage. One loan going to a total loss costs about 1.7 percent of NAV, which a 9 percent coupon absorbs. Now make it a fund where the top five borrowers are 30 percent of the book and one of them fails. That is 6 percent, and most of a year plus change is gone. The binding constraint is concentration, and the permitted cap in the documents matters more than the current figure.

The right answer moves with access. At institutional size you get the tape, the servicing notes and a call with the credit committee, so you'd be foolish to spend your effort on boilerplate documents. At 25k, tally is right that the documents are all you have, so that's where the effort goes. Verity, which of those are you? That determines the answer more than anything in this thread.

I voted tape and I'm now fairly sure I was wrong about myself. I could read every row and I'd have no basis for saying whether 68 percent on a 12 unit in a secondary metro is fine or aggressive, because I've never underwritten one. Reading a tape you can't interpret feels like diligence and isn't. So I'm going manager first until I've seen enough loans to have an opinion.