The pool is clean because someone cleans it

I took the tour once.

Free breakfast. Ninety minutes, they said. A folder with my name already printed on the front. I sat there in a windowless room off a hotel lobby telling myself I would be polite and leave.

Two hours later a man was sketching a monthly payment on the back of a brochure.

I walked out. Plenty of people don't. The industry's own trade group counts about ten million American households that own one of these things, and $10.7 billion of sales in 2025.

But the vacation is not the business. The loan is.

Roughly two-thirds of Hilton Grand Vacations buyers borrow the purchase money from Hilton Grand Vacations. The average rate on that book runs 14.4%. So the company sells you the week, lends you the money, and then charges you a fee every year to maintain the place.

Then it does one more thing. It gathers thousands of those loans into a pool, and sells the pool to bond investors. Insurance companies buy them. Pension funds buy them. The company gets its cash back on day one instead of over ten years, and lends it out again.

Those bonds have a spotless record. In June the company sold a $300 million pool and, by its own finance chief's account on the earnings call, orders came in around nine times the size of the deal, at the tightest spread the timeshare market has seen since January 2022.

So the loans must be fine. That is the reasonable conclusion, and I held it too until I opened the quarterly filing.

The filing splits the company's own loan book in two. Loans that sit inside the bond pools. Loans that sit outside them. Same company. Same product. Same kind of buyer, sold in the same room.

Inside the pools, 0.7% of the balance is more than 120 days late. Outside them, 25.2% is.

Some of that gap is screening. Only healthy loans get into a pool in the first place. But loans go bad after they are inside a pool too, and the pools still print 0.7%.

The filing explains it in one sentence. When a loan inside a pool defaults, the company often swaps a good loan in or buys the bad one back at full principal. A hundred cents on the dollar.

The bad loan moves back onto the company's own books. The pool stays clean. The bond investors stay happy.

The pool is clean because someone keeps cleaning it. That is not the same as the loans being good.

None of this is hidden. The sentence has sat in the filings since at least 2021. The repurchases are optional and probably good business, because cheap funding is worth paying for.

What worries me is narrower. Investors treat the performance of these pools as evidence about the borrowers. It isn't. It is evidence about the sponsor's willingness to keep writing cheques.

Those are two very different things, and they only look the same while the sponsor is healthy.

I have no idea when that stops being true. Nobody does. But I know which number I would rather be reading.

Start underneath the bonds, with the people making the payments.

Hilton Grand Vacations discloses an annual default rate in its yearly report. It ran 5.14% in 2019. It ran 10.77% in 2024 and 9.86% in 2025.

The lending rules did not loosen. The filings have described the same minimum ten percent down payment since 2021. So the rules held and the people passing them got weaker.

The June filing adds two more. Loans the company has stopped accruing interest on grew to $514 million, from $430 million six months earlier. And the reserve it holds against its own originated book stood at $1,116 million at the end of March, up from $1,078 million at the start of the year, after writing off $55 million in a single quarter.

Contract sales fell 3% in the June quarter to $810 million. Tours rose 6%. More families are sitting through the presentation and fewer of them are saying yes.

The stock fell about a tenth on the day. Management called it an execution problem in Orlando and Myrtle Beach. That may be right about the selling. It explains nothing about the loans.

And here's where it spreads.

Fitch grades these deals and tracks them afterwards. Its surveillance of the 2018 through 2020 pools puts the share of loans bought out or swapped by the sponsor at roughly 2.4% to 3.5% of the original pool size.

Fitch also records the outcome. No deal in the programme has produced a loss for bondholders, and Fitch credits the buyouts and substitutions for that record.

So the rating agency has written down the mechanism. The bond market reads the outcome and stops there.

You can see it in the pricing. A $300 million deal in June drew orders around nine times its size and cleared at the tightest spread this corner of the market has seen since January 2022, according to the company's finance chief.

One number never appears anywhere. The filings do not disclose how many dollars of defaulted loans get bought back each year. That cost sits inside the general loss lines, without a line of its own.

So an investor cannot size the support. Which means the investor also cannot tell how much of it the sponsor could still afford in a bad year.

Meanwhile, look sideways at the competitors.

Marriott Vacations assumes in its own filings that 13.05% of its loan balances will eventually default. Travel + Leisure charged $141 million against expected losses in the June quarter, up from $128 million, on sales of $665 million.

That is roughly a fifth of the sale price set aside on the day the sale is made, and it is normal here. Nobody in this business is surprised by it.

The industry itself keeps growing. Its trade group reports $10.7 billion of US sales in 2025, higher every year since 2022. Average prices have slipped a little, from about $24,200 in 2023 to about $23,200 in 2024, so the growth is coming from volume.

And the bond desks want more of it. Timeshare deals sit inside the corner of securitization that investors call esoteric, next to aircraft leases and music royalties. KBRA expects total asset-backed issuance to reach roughly $385 billion this year, another post-crisis record.

There is one crack in the calm. Trade coverage in May reported that investors had begun questioning underwriting after the 2025 deals showed higher delinquencies than any vintage since 2021.

Questioning is not selling. But it is the first time in years anyone has asked.

Three more things worth keeping track of this week.

Taiwan's life insurers have cut their currency hedging to the lowest level since at least 2020, ahead of an accounting change that makes unhedged positions easier to carry. These firms hold roughly seventy percent of their portfolios in foreign assets, almost all of it in dollars. Their cushion against currency swings, the sector's foreign exchange volatility reserve, sat around NT$284 billion against more than NT$23 trillion of foreign investments. Fitch has the sector on a deteriorating outlook. A strong Taiwan dollar is the one thing that hurts them, and they just took off more of the coat.

American municipal credit has turned direction without turning ugly. Nearly 95% of the main municipal index is still rated A-minus or better, which is historically high. But S&P downgraded more issuers than it upgraded in five of the six months through April, reversing the pattern that held since the pandemic. Negative outlooks are rising at both S&P and Moody's. Ratings drift like this usually shows up two or three years before anything happens to prices.

And American shippers are paying far more to move less. The US Bank freight payment index put national shipment volumes down 2.8% in the second quarter against a year earlier, while spending on those shipments rose 28.1%. In the Midwest, volumes fell 3.7% from the prior quarter while spending still ran nearly 23% above last year. That is not demand pulling prices up. That is capacity leaving faster than freight does, which tightens the screw on every small carrier still running. We'll see.

My grandmother's neighbour bought a mobile home in the eighties.

He put almost nothing down. A dealer arranged the loan on the lot, the same afternoon. He was proud of it, and he had every reason to be. It was the first thing he ever owned.

A company in St. Paul called Green Tree Financial wrote a great many loans like that one. Then it did something new. It gathered thousands of them into pools and sold the pools to pension funds and insurance companies as bonds.

Green Tree kept servicing the loans afterwards. Its agreements with the trusts also let it substitute loans in and take loans out.

The bonds performed. The ratings held. Money kept arriving, and the servicing book grew from about $4.6 billion to more than $18 billion.

In 1998 an Indiana insurer named Conseco looked at that record and paid about $6 billion for the whole company.

They bought the record. The record was never the loans.

Two years later the arithmetic arrived. Conseco restated its 1999 earnings because losses on the securitized mobile-home loans came in larger than it had assumed.

Then it kept arriving. On December 17, 2002, Conseco filed for Chapter 11 listing $52.3 billion of assets. It was the third-largest bankruptcy in American history at the time, behind WorldCom and Enron.

The bonds had looked orderly for years while the borrowers underneath them were falling behind.

And the manufactured-housing bond market never came back to what it was. A funding channel that had carried tens of billions of dollars closed, and homes that had been easy to finance became almost impossible to finance.

My grandmother's neighbour kept his home. Many people did not, and the ones who lost theirs were the ones with the least behind them.

So when a lender tells me its pools are performing, I want to know one thing first. Who is holding those pools up, and what happens on the day they stop.

Clean pools are not the same as good loans.

We'll see.