I've Never Owned an Apartment Building. In 2027 I'm Going to Buy One.

In July I wrote that the Midwest was where the numbers worked and the Sun Belt was still correcting. I stand by the reasoning. I'm about to argue something that sounds like the opposite, so let me be clear about what changed and what didn't.

What didn't change: the numbers have to work on the day you buy.

What changed: I spent October reading about private credit — the non-bank lending that financed most small commercial real estate between 2020 and 2022 — and the picture it paints is specific enough to act on. The distress in this cycle isn't in houses. It's in 5-to-10 unit buildings, and it's arriving in 2027 on a schedule you can read in advance.

I've never owned one. Thirty years of single-family and condos, two foreclosures bought in 2008, and not one apartment building. That's about to change, and this is the homework.

The loans, not the buildings

Start with the thing most market commentary gets backwards. Small apartment buildings are not in trouble. The Arbor Small Multifamily Price Index moved -0.3% year over year in Q2 2026 — essentially flat — with financed-property occupancy at 96.3%. There is no fire sale.

The trouble is in the debt.

Between 2020 and 2022, a great deal of small multifamily was bought with floating-rate bridge loans — short-term, interest-only, taken out at peak pricing on the assumption that rents would keep climbing and rates would stay low. Neither happened. Those loans are now maturing.

KBRA published a study in June 2026 tracking 6,919 loans in CRE CLOs — the securitised pools these bridge loans end up in. The vintage numbers are the whole story:

Origination year	Default rate	Distress rate
2022	18.1%	44.2%
2021	13.6%	33.9%
Study average	10.4%	22.9%

Nearly half the 2022 loans are in distress. Not nearly half of apartment buildings — nearly half of the loans written in one particular year at one particular moment of optimism.

And the losses are small. KBRA found only 33 loans that took a principal loss, across 19 deals. The buildings are fine. The borrowers are fine. The financing is the problem.

Maturity default is the one to understand

Here's the distinction I'd never properly thought about, and it's the one that creates the seller.

A payment default is a borrower who stopped paying. A maturity default is a borrower who paid every month for five years and then couldn't refinance when the loan came due.

In KBRA's study the cumulative payment default rate was 4.9%. The combined rate including maturity defaults was 10.4%. Most of the trouble is the second kind.

Think about what that person looks like. They bought in 2021, they've run the building competently, the tenants pay, and the loan is due. Their lender wants more equity than they have, at a higher rate than they modelled, on a building whose rents are flat instead of up 6% a year. They're not distressed in the sense of being bad at this. They're out of time.

That's who sells to you. Not a bank, not a trustee's sale — a person with a date on a calendar.

The Mortgage Bankers Association put roughly $162 billion of multifamily mortgage debt maturing in 2026 and $168 billion in 2027. Banks and thrifts face $396 billion of total commercial maturities in 2026 alone, 21% of their commercial mortgage balance. These are not evenly distributed — they're concentrated in exactly the vintages above.

Why five units changes everything

Everything about this gets easier at five units, which sounds backwards until you look at who you're bidding against.

One to four units qualifies for residential financing. Thirty-year fixed, conventional or FHA, possibly owner-occupied. That means you compete with first-time buyers, house hackers, DSCR investors and every retail landlord in the metro.

At five units, residential financing stops. You need commercial debt, underwritten on the building's income rather than your pay stubs. That's too small for institutions — a fund isn't buying an eight-unit — and too complicated for most retail buyers, who take one look at a three-to-seven-year balloon and go back to single-family.

That gap is the opportunity. Fewer bidders is worth more than a lower price.

It is also where the risk sits, and I want to be blunt about it because it's the thing that would actually hurt me. A thirty-year fixed mortgage on a rental house is a problem you solve once. A five-year commercial loan is a problem you solve again in five years, in whatever market exists then. The sellers I'm describing are people for whom that went wrong. You don't get to buy their building and pretend the same clock isn't now running for you.

What I learned borrowing commercially

I took out my first DSCR loan last year — a loan underwritten on the property's income rather than mine. It's not the same as a commercial multifamily loan, but it's the closest I've been, and four things surprised me.

The rate was higher than a conventional mortgage, which I expected. The prepayment penalty ran five years, which I hadn't priced in — that's five years where selling or refinancing costs you real money, and it meaningfully changes what "the numbers work" means.

The appraisal came in below the comps I'd been working from. On a commercial deal that's worse, because value is a function of income rather than comparable sales, and a lender's cap rate assumption can move the number more than any sale down the street.

And the loan qualified on the building, not on me. I wasn't working at the time. For anyone whose income is irregular, that's the whole point of this financing — and it's why the underwriting is stricter elsewhere.

The Denver numbers, and the trap in them

I live here and I'll buy here, so this is the local read.

Metro Denver's stabilized apartment vacancy was 6.3% in Q3 2026, down from 7.5% earlier in the year. Average rent is $1,816, down 5% year over year — the lowest in three years.

Now the part that matters. Stabilized vacancy excludes buildings still in initial lease-up. Include the new unleased apartments and the Apartment Association of Metro Denver put the rate between 9.4% and 10.9%.

Both numbers are true. The first tells you how existing buildings are doing. The second tells you how much competition for tenants is about to arrive. If you read only the first you'd conclude Denver's apartment market is fine; if you read only the second you'd conclude it's flooded. The buyer's version is: existing buildings are holding occupancy by discounting, and the discount is the real rent.

A year ago I'd have quoted you 7.6% vacancy and called it a 16-year high. That was Q4 2025 and it's no longer true — this is exactly the kind of number that gets repeated for a year after it stops being accurate.

Where I'd actually look — and where I was wrong

My first instinct was Cheesman Park. Older brick walk-ups, central, a neighbourhood I like. Then I looked at vacancy by county:

County	Stabilized vacancy	Average rent
Denver	7.0% — highest	$1,798
Arapahoe (south Aurora, Englewood)	6.6%	$1,653
Adams (north Aurora)	6.0%	$1,633
Jefferson (Lakewood, Golden)	5.4% — tightest	$1,803
Douglas, Boulder/Broomfield	~5.3%	$1,924–1,980

I'd assumed the suburbs were where the softness was. They aren't. Jefferson County — Lakewood — is the tightest market in the metro. The vacancy pressure is in the city of Denver, because that's where the new buildings went up.

So the honest answer is that my instinct was closer than my first correction. Central Denver and Aurora are where the combination sits: older stock, softer occupancy, lowest rents in Aurora's case, and the submarkets where 2021-22 buyers were most active. Lakewood is a fine place to own a building. It isn't where the discount is.

The buy box

This is the part that doesn't move.

The numbers work on day one, or I don't buy. Not if rents recover in 2028. Not if rates come down. Not if I push the operating expenses down to what the broker's pro forma says. On the actual trailing twelve months, at today's rate, with a vacancy allowance that assumes the lease-up supply lands.

Specifically, what I'm looking for: an older building, five to ten units, eighty-five to ninety percent occupied with below-market rents, owned by someone who bought in 2021 or 2022 on a loan that comes due. Underwritten flat through 2028. Cash and a HELOC to start, and I'm open to seller financing or assuming a low-rate loan where the seller's position makes that the obvious trade.

Yardi has pushed its forecast for positive rent growth in high-supply markets out to the end of 2027. That's the argument for buying before it arrives — and it's also the thing to be careful about, because that forecast has already slipped once. Underwrite as though it slips again.

What I'm not saying

I bought two foreclosures in 2008 and my honest regret is that I didn't buy more. That tempts me to talk about this as the opportunity of a lifetime, and I'm not going to, for two reasons.

The first is that someone has to be on the other side of these trades, and they're mostly not villains — they're people who took a five-year loan in a year when that looked sensible. Buying well doesn't require being pleased about why it's available.

The second is that a forecast is not a plan. Everything above is a reason to go looking. None of it is a reason to buy a particular building. The only thing that decides that is whether the trailing numbers cover the debt service at today's rate, and that's arithmetic about one address, not a view on a market.

The 30-year fixed was 7.24% on 29 October 2026. Commercial money for a small building costs more than that, over a shorter term. If a deal works at those numbers, it works. If it needs 2028 to work, it isn't a deal — it's a bet with a mortgage attached.

I'll let you know what I buy.

This is how I think about my own investing. It isn't financial advice, and it isn't a recommendation about any specific property, market or loan. Talk to your own advisors.