How a 9.3 percent mortgage REIT yield gives back 7.5k in 22 months, worked as a case.

The 9.3 percent was the disclosure. Everything else was the footnotes.

I'm asking because I want to know, not to score points. If you'd bought the equity REIT version, an owner of the same kind of buildings rather than the lender against them, how would that have gone over those 22 months? I'd guess worse, because the lender at least sits ahead of the owner's equity and gets paid first. Which makes your outcome more interesting to me, since being senior in the stack still didn't save the position.

Running the numbers. $59,947 in, $52,443 out, so negative $7,504, call it negative 12.5 percent over 22 months, roughly negative 7 percent annualized before tax. After ordinary income treatment on $9,170 of distributions at whatever the marginal rate is, the after tax gap is meaningfully wider than the headline. At a 24 percent marginal rate that is another $2,200, which puts the real figure closer to negative $9,700. The cause was concentration. Office will take most of the blame in this thread and that is half right, because the identical thesis spread across six names would have made the office exposure a bad quarter instead of the whole outcome. A single position funded by the entire proceeds of a property sale is the decision that turned a sector problem into a personal one. @diane_kowalski the seniority point is real and it still would not have helped, because seniority protects principal at resolution and does nothing for the dividend in the meantime. Nonaccrual stops the cash flow long before anybody finds out what the collateral is worth.

@hannah_west29 I'd push back. Six names in commercial mortgage REITs during that stretch would have given you six versions of the same office exposure, because the sector correlated hard. Diversification across names inside one narrow credit sector isn't diversification, it's the same trade with more tickers. The cause was buying an office-heavy bridge book at a point when office collateral values were already visibly moving, and no amount of spreading it out fixes that.

@laurenkim53 did you look at the collateral mix at all before buying, or only the dividend?

Slowing this down. Nonaccrual means the borrower stopped paying and the lender stops recording the interest as income even though the loan technically still owes it. A risk rating of 4 or 5 is the lender's own internal opinion that the loan is in trouble, published by the lender but reflecting its own judgment rather than an outside audit. If that reading holds, then a rating that moved from 8 to 24 percent is management grading its own homework, which is a strange thing to lean on and also usually the most honest number in the filing.

@raj_mehta14 that's essentially right, and the grading is internal, which is why the migration matters more than the level. A book that sits at 8 percent for six quarters and then goes to 15 is telling you something a static number won't. One addition to @laurenkim53's list: the financing side. In a bridge lender's supplemental there's usually a breakdown of what portion of the borrowings are mark to market and what portion aren't. Non mark to market financing means the lender's own lender can't demand more collateral just because the market price of the loans moved. That percentage is the difference between a company that gets to work its problem loans out over three years and one that gets forced into selling into a bad bid. Two companies can hold nearly identical loan books and land in completely different outcomes because of it. @laurenkim53 what did that split look like on the one you were tracking?

The detail that stands out is selling at $10.90 when book value per share at that point was likely well above it, which is exactly the setup that often gets described as a buying opportunity. Take a name bought at a 25 percent discount to book, on the premise that a dollar is being bought for 75 cents. Sometimes the discount is right, and book comes down to meet the price over the next year. The discount was the market pricing in reserve builds that had not been taken yet.

What did you do on management before you bought? Not the bio page. Whether the same people ran a book through 2008 or 2020, what their realized loss history looks like on prior vintages, whether the CEO's own money is in the shares. That's a couple of hours of reading and it's the part of due diligence I've found actually separates these companies, because the loan documents are broadly similar across the sector and the workout behavior isn't.

I'm not saying it would have saved you. I'm asking because the answer tends to be nothing, and it was nothing for me too the first time.

Worth separating two things that got merged in this thread. A loan going to nonaccrual is a cash flow event for a shareholder. A loan taking a realized loss is a capital event. They happen at different times, and often the second one is much smaller than the reserve suggested, because bridge lenders on office frequently modify, extend, take a paydown from the sponsor, and get out whole eighteen months later. Which means a seller may have exited into the reserve build and missed the resolution. That is not a criticism of selling. Holding a name where the dividend fell from $0.35 to $0.16 while relying on the income is a perfectly good reason to leave. But a $7,500 gap in a case like that is partly a mark, and marks on loans that eventually pay off tend to come back. The underwriting question worth asking about a book like that is what the sponsors' equity looked like at origination. A 65 percent loan to a sponsor who put in real cash behaves completely differently from a 75 percent loan where the equity was structured. That detail rarely shows up in any table that can be screened for.

Dumb question from the tools side. Bridge loan on an office building means the borrower was renovating it or repositioning it, right? So the loans that went bad were ones where the plan was to fix the building and lease it up, and then nobody leased it up. That's just my world with more zeros.

A similar position exits about eleven months earlier in cases like this, and it usually isn't skill. On an earnings call, someone asks management whether distributable earnings cover the dividend, and when the answer takes ninety seconds and never lands on yes, that's the signal to sell that week. The coverage ratio is printed in the supplemental every quarter and can be checked directly rather than waiting to hear it dodged live. @laurenkim53 the piece worth arguing with is the framing that the yield was the market telling you the risk. Sometimes it is. Sometimes a name trades at 12 percent because it's small and nobody covers it. The yield alone doesn't tell you which, and either way it sends you to the loan book, which is the same conclusion reached from the other direction.

Threads like this are worth moving mortgage REITs from the column marked easy income to the column marked credit analysis with a stock price attached. @laurenkim53 for anyone putting money back into the space now, would agency be the better route? My understanding is the agency names take rate and prepayment risk with a government guarantee behind the paper, so the credit rating table problem doesn't exist there. Different risk, but at least it's a risk that can be named.

Saving this one. I understood maybe half of it, and the half I understood was expensive for you and free for me.