When the Bill Beats the Building

Start with the arithmetic, because it is the least interesting part and people spend the most time on it. A capitalization rate is net operating income divided by price. A mobile home park throwing off $700,000 of NOI that sells for $10,000,000 trades at a 7 cap. That is a 7 percent first-year cash yield on an unlevered dollar.

Nothing about that number is a quality score. It is not a grade for the park, the operator, or the market. It is a yield — the same species of number as the interest rate on a savings account or the coupon on a bond. And yields only mean something in comparison.

This is the first thing most students get backwards. They learn to compute a cap rate, then treat it as a property characteristic, like lot count or road frontage. It is not. It is a price, and it is set by people who are not in the room and have never seen the park.

Your park’s cap rate is determined in the Treasury market as much as in your leasing office.

Invert the formula and the stakes appear. Price equals NOI divided by the cap rate. The numerator is what you control — rents, expenses, occupancy, infill, the slow work of running a place well. The denominator is what the capital markets hand you. You can grow NOI by five percent, have the denominator move against you by a hundred basis points, and be poorer than you were.

Every risky asset in the world is priced against one benchmark: the yield on United States government debt. A Treasury bill is a promise from the entity that prints the currency the promise is denominated in. It has no tenants, no roof, no vacancy, no eviction docket, no deferred capex, and no closing costs. You can sell it in seconds at a price you can see on a screen.

Whatever that instrument pays is the price of doing nothing. It is the return available to an investor who takes no risk, does no work, and hires no one. On August 19, 2026, the ten-year Treasury closed at 4.65 percent and the three-month constant-maturity bill at 3.86 percent.

Everything else must clear that bar and then pay you extra for the trouble. That extra is the risk premium, and it compensates you for illiquidity, for capital expenditure, for the possibility that a tenant stops paying, for the fact that a park cannot be sold on a Tuesday afternoon at a price you can look up. So:

Cap rate ≈ risk-free rate + risk premium.

Which means a cap rate quoted without the risk-free rate beside it is not information. A 7 cap against a 1.5 percent ten-year is a fat, generous yield. The same 7 cap against a 6 percent ten-year is barely worth the paperwork. Same park, same income, same number — opposite conclusions.

The vocabulary that matters

The difference between the two is the spread, quoted in basis points, where one basis point is one hundredth of a percentage point. A cap rate of 5.9 percent against a 4.65 percent ten-year is a spread of 125 basis points. That single number tells you more about the state of the market than any cap rate ever will.

Here is where the lesson usually surprises people. Between 1991 and 2019, the average spread between commercial property cap rates and the ten-year Treasury ran about 342 basis points. That is roughly three and a half percentage points of compensation for choosing a building over a bond. Today that spread sits near 172 basis points across property types — the twenty-fourth percentile of its range since 1965.

Now look at what that reframes. The famous record-low cap rates of 2021 were not, on this measure, unusual at all.

In 2021 the cap rate was at an all-time low of 4.8 percent, but the ten-year was at 1.45 percent, so buyers were still collecting 335 basis points over the risk-free rate — almost exactly the long-run norm. The pricing was aggressive in absolute terms and completely ordinary in relative terms.

Today the cap rate is higher and the deal is worse. Cap rates rose about 110 basis points; the ten-year rose more than 300. The compensation for owning real estate instead of a bond has been cut by roughly two thirds.

That gap is not an academic curiosity. It is a force.

Capital is not sentimental and it is not small. Pension funds, insurers, sovereign wealth funds, endowments, and the private funds that serve them allocate in the trillions, and they allocate by comparison. Every one of them runs the same silent calculation every quarter.

Multiply that meeting by every institution on earth and you have the migration. It does not look like a stampede. It looks like an absence — bids that never arrive, brokers who cannot find a second offer, deals that die quietly in due diligence for reasons nobody can name.

And the seller experiences the whole thing as a price cut. That is the part worth sitting with: the income did not change. The park did not get worse. The alternative got better, and the discount rate moved.

Run your eye down that chart slowly. Six bars, one income stream, a spread of $6.4 million between the top and the bottom. Not one dollar of that difference has anything to do with the property.

This is not a thought experiment. It is the recent history of American commercial real estate, and it is worth walking through with real numbers because the popular version of the story is consistently overstated.

Two disciplines fall out of that timeline, and both are worth more than the headline.

The first is that indices measure different things and disagree by a factor of two. If someone tells you commercial real estate fell twenty percent, ask which index. Green Street says twenty-one; MSCI RCA says eleven. Neither is lying. They are measuring appraisals, repeat sales, and spot pricing, which respond to the same shock at different speeds.

The second is that real estate does not reprice on a screen. Sellers anchor to what they paid. Buyers underwrite to what capital costs today. The gap between those two beliefs does not resolve into a price — it resolves into no transactions at all, sometimes for years. Then a loan matures, the bank asks a question the owner cannot answer with optimism, and a whole vintage reprices at once.

Everything so far assumed you paid cash. Almost nobody pays cash, and leverage does not simply amplify the outcome — it changes its sign.

The number that governs this is the loan constant: annual debt service divided by the loan balance. It is not the interest rate. It includes principal amortization, so it is always higher than the coupon. A 6.10 percent loan amortized over thirty years produces a constant of about 7.27 percent. That is the real annual cash cost of a borrowed dollar.

If the cap rate is above the loan constant, debt lifts your return. If it is below, every borrowed dollar drags it down.

That second condition has a name: negative leverage. And on today’s numbers it is not a hypothetical.

Loan constants computed on thirty-year amortization. Rates reflect quoted stabilized manufactured housing agency pricing, which currently runs roughly 6.0 to 6.75 percent; the 6.10 percent figure is a teaser quote, not a market average.

Notice what negative leverage does to an underwriting model. It means the deal cannot pencil on income alone. It has to pencil on something happening — rents rising, expenses falling, lots filling, or a buyer at a lower cap rate five years out. Every one of those is a forecast. When the spread was 335 basis points, the income did the work and the forecast was upside. When the spread is 125 and leverage is negative, the forecast is the deal.

This is the single most useful thing a student can learn from the current market: the structure of a deal quietly tells you how much of your return is a fact and how much is a hope.

There is one more mechanism, and it is the one that turns a slow repricing into an event.

Lenders do not size a loan off value. They size it off coverage — net operating income divided by annual debt service, the debt service coverage ratio, typically floored around 1.25 times. Rearrange that and maximum loan proceeds equal NOI, divided by the required coverage, divided by the loan constant. Every term on the right is set by someone other than the borrower.

So take a park with $700,000 of NOI. At a 3.60 percent rate, the 5.46 percent constant supports about $10.3 million of debt. At 6.10 percent, the 7.27 percent constant supports about $7.7 million. Identical income, identical park, $2.6 million less debt available — because the constant moved.

An owner who borrowed $9 million in 2021 and comes due in 2026 does not have a valuation problem. They have a cash problem, and it arrives on a specific date with a specific number attached. They write a check for the difference, sell, extend on the lender’s terms, or hand over the keys.

That is what the $875 billion maturing in 2026 represents. Not a prediction — a schedule.

Already true

The ten-year sits at 4.65 percent, manufactured housing transaction cap rates near 5.9 percent, and agency debt near 6.1 percent. That combination is negative leverage on arithmetic alone, and $875 billion comes due this year.

What has to happen

Either cap rates expand toward the loan constant, or the Treasury curve falls enough to pull debt costs down with it, or NOI grows fast enough to make the coverage test irrelevant. One of the three has to give, and only the third is under an owner’s control.

Where I am probably wrong

I may be too gloomy on manufactured housing specifically. Cap rates in this sector already compressed 40 basis points off their 2024 peak while office kept falling, which means capital is treating manufactured housing as a defensive asset rather than a distressed one. If that holds, the sector could clear its maturities through refinancing and extension without the price discovery I am describing — and the lesson would survive as arithmetic while failing as a forecast.

Strip away the sector detail and five things remain true in any market, in any year.

1. A cap rate is a price, not a property statistic. It is set by capital markets, and it applies to your park whether or not anyone consults you.
2. Never quote a cap rate without the risk-free rate beside it. The level means nothing. The spread means everything.
3. When bills out-yield buildings, cap rates must rise. Prices fall until the spread pays someone for the risk again. This is not a market opinion; it is what has to happen for capital to come back.
4. The loan constant, not the interest rate, decides whether debt helps you. Above the cap rate, leverage subtracts. Below it, leverage adds. Compute it before you fall in love with a deal.
5. Value can collapse with rising income. If you only ever remember one line from this, make it that one. It explains almost every confusing thing that happened to real estate between 2022 and 2025.

The instinct to treat a cap rate as a verdict on a property is natural and completely wrong. It is a verdict on the alternatives — a running tally of what the rest of the world will pay you to do nothing. When doing nothing pays 3.86 percent, the bar for doing something is high, and the price of everything that requires effort has to come down until the effort is worth it.

Which is why the most important number in a real estate deal is one that appears nowhere in the offering memorandum.