The Operator Bought The Building Back. Two Weeks Earlier, Another One Sold It.

Two trades, opposite directions, thirteen days apart

On September 1, 2026, CareTrust REIT closed $400 million for 2,622 licensed skilled nursing beds in the southwest and leased every one of them straight back to the operator already running them — triple-net, annual inflation-based escalators, multiple renewal options, an expected stabilized yield of approximately 8.6% (CareTrust REIT press release, September 15, 2026). The operator kept the business and gave up the real estate.

On September 3, 2026, The Pennant Group did the exact opposite. A Pennant affiliate has operated Mainplace Senior Living in Orange, California under a triple-net lease since the company's 2019 spin-off. Effective that date, Pennant bought the building out from under its own lease (The Pennant Group press release via GlobeNewswire, September 16, 2026). The operator kept the business and took the real estate back.

Both of these are rational. That is the interesting part. The same structure — an operator in a building it does not own — was worth dissolving in one case and worth creating in the other, in the same fortnight, in the same asset class. Anyone advising an operator on whether to sell the dirt needs to be able to say which situation they are in, and the answer is not a matter of taste.

What a triple-net lease actually prices

A sale-leaseback converts an owned asset into two things: a pile of cash today, and a fixed obligation forever. The rent is not a cost of occupancy. It is the amortization of that cash at the landlord's required return, and in senior housing that return is currently visible: CareTrust's own disclosed expectation on the September portfolio is approximately 8.6% stabilized, and the company reported a blended stabilized yield of approximately 8.7% across roughly $1.9 billion of 2026 investments in two dozen transactions (CareTrust REIT press release, September 15, 2026).

Call that the going rate for converting senior care real estate into cash through a REIT in 2026: high-8s. An operator doing a sale-leaseback at that level is borrowing at an effective 8.6% against an asset, permanently, with no amortization in its favor and with the residual value transferred away. That is not a criticism. For an operator whose capital compounds faster than 8.6% inside the operating business — through acquisitions, through census and rate growth, through new markets — it is straightforwardly accretive. The cash gets redeployed at a higher return than the rent costs.

The trouble starts when the operating business stops out-earning the rent.

Why the benchmark matters more than it used to

The 10-year Treasury crossed 5.00% in the week of September 14, 2026 and held there — the first sustained move above that level in nearly three years (CNBC, September 16, 2026; Levin Associates, September 15, 2026). CareTrust's high-8s stabilized yield sits roughly 360 basis points over that benchmark.

That spread is the whole argument. When the benchmark was 2%, an 8.6% lease looked expensive relative to almost any other financing an operator could arrange, and sale-leasebacks were a last resort for sponsors who could not get bank paper. At a 5% benchmark, an 8.6% permanent obligation with inflation-linked escalators stops looking like distress financing and starts looking like a market-clearing cost of capital — which is precisely why CareTrust is deploying $1.9 billion into it rather than fighting for assets.

But the same math cuts the other way for a tenant. A lease struck years ago, against a very different curve, is a fixed obligation that did not reprice when the world did. Rent set with escalators in a 2019 deal has been compounding upward through a period when the operator's own alternatives changed completely. That is the situation Pennant's Mainplace lease had been sitting in since the 2019 spin-off, and buying the fee interest is the cleanest way to stop the compounding.

The supply picture is what makes ownership defensible

An operator buying real estate is making an implicit claim: this building will be worth owning. In most of commercial real estate in 2026 that claim needs an argument. In senior housing the argument is already made, and it is made by the construction pipeline.

Units under construction across NIC MAP's Primary and Secondary Markets have fallen below 24,000 — the lowest level since mid-2012 — and construction starts continue to hover at historically low levels with no visible sign of acceleration (NIC MAP, Q2 2026). Against that, occupancy across those same top 99 metros reached 90.1% in the second quarter of 2026, with 57 of the 99 markets individually above 90% and 23 above 92% (NIC MAP, Q2 2026).

An operator looking at those two numbers together is looking at a building whose replacement is not being built and whose census is tightening. Owning that is a different proposition from owning an asset class where supply can answer demand within eighteen months. The residual value an operator gives away in a sale-leaseback is worth more in a market with a fourteen-year construction low than it was when the lease was signed.

The screen: four questions before selling the dirt

1. Does the operating business reliably out-earn the rent? Not in a good year — reliably, through a bad one. If the answer is yes, a sale-leaseback converts a slow asset into fast capital and the structure works. If the answer is only sometimes, the operator has converted a flexible asset into an inflexible obligation that must be paid in the bad year too.

2. Is the cash going somewhere specific? Sale-leaseback proceeds with a named use — an identified acquisition, a known market entry — are capital allocation. Proceeds with no named use are a balance-sheet event that leaves the operator paying 8.6% forever for money it has parked.

3. How long is the operating history in the building? Pennant's affiliate had run Mainplace since 2019 before buying it. Seven years of operating data is the best possible underwriting on a single asset, and it is information no third-party buyer has. An operator with that history who chooses to own is exercising an informational edge; an operator without it is speculating on real estate while running a business.

4. What does the escalator do over the remaining term? Annual inflation-based escalators of the kind CareTrust disclosed are benign when rents start low and punishing when they start high. The question is not the rent today, it is the rent in year ten against a census forecast the operator actually believes.

What this means for a sale process

For brokers, the practical consequence is that the tenant is now a bidder. In any asset where an operator holds a long lease and a long operating history, the operator is not merely an encumbrance on the sale — it is the one party underwriting with complete information, and it may be the clearing bid. Pennant's Mainplace purchase had no broker and no competing process because none was needed; the buyer was already inside the building.

That has two implications when advising a landlord. First, a leased asset marketed without a conversation with the sitting tenant may be marketed to the wrong room. Second, the tenant's willingness to buy is itself a valuation signal — an operator that declines to buy the building it knows best is telling you something about the rent, the census, or both, and that information is worth more than another broker opinion.

The honest limits of this read

Neither transaction disclosed a price. Pennant did not publish what it paid for Mainplace, did not disclose the unit count, and did not name the seller; its only characterization of value was "attractive pricing," which is a public company declining to give you the cap rate. CareTrust did not name the operator, the states, the facility count, or the price allocation across the portfolio. So the spread analysis above rests on CareTrust's disclosed yield expectation and the published benchmark, not on two comparable prints — and it should be read as a framework rather than as a comp. What is verifiable is the structure of each deal and the direction each party chose. That is enough to build a screen on. It is not enough to price one.