The Apartment Recovery Is Real.
Welcome to the Skeptical Investor blog, right here on BP! A frank, hopefully insightful, dive into real estate and financial markets. From one real estate investor to another.
----
This week, we're talkin' apartments. The apartment market has turned the corner. But the doomer headlines say Nashville is missing the party. The problem: the headlines are counting the wrong buildings. I'll show you what the rent roll says instead. Plus, I'll show you why two of the biggest data providers in real estate can look at the same city and come back with opposite answers about next year.
Let's get into it.
----
Today's Interest Rate: 7.20%
(☝️ .08% from this time last week, 30-yr mortgage) (Mortgage News Daily)
They hiked, and the bond market had already done it for them: the 10-year touched 5.04%, the highest since 2007, before Warsh ever spoke (CNBC). Now oil is falling for a third straight session (Trading Economics). The question for an investor isn't whether the Fed was right; it's whether they just tightened into the top of an energy shock that's already unwinding.
The 30-year touched 7.24% on decision day and settled back to 7.20% by Friday, which tells you the mortgage market had priced this in weeks ago (MND). More in the news below.
----
The Weekly 3 in News:
- -The Fed hiked. First time since 2023, and they're signaling more. The FOMC raised the target range a quarter point to 3.75% to 4.00% on Wednesday in a 12 to 0 vote, citing inflation that energy prices keep feeding (CNBC, Sept 16; Schwab). Futures now put the odds of another hike in October somewhere around 56% to 60% (CME FedWatch). I'll own it: back in June I wrote that the next move would be a cut, and last week I said they'd hold. Wrong twice. What I underestimated was how sticky this conflict would remain, plus the Saudi oil pipeline disruption. Perfect storm for $100 oil. However, crude just broke back under $100 for the first time since the Red Sea escalation, at $97.56 (Trading Economics), and Truflation's real-time index now reads 2.33% against the BLS headline of 3.4% (Truflation; BLS). The Fed may have just tightened into the top. We'll see.
- -AI has created about a million American jobs. Five for every one it hit. The Economist ran the numbers and found roughly 1 million jobs created by the AI buildout against something like 200,000 AI-attributed layoffs (The Economist; Semafor). Two details matter for us. First, the professions closest to the boom (engineers, developers, data scientists) added roughly 730,000 jobs above trend since 2022. Second, and this is the one for anyone who's tried to hire a contractor lately, data-center-adjacent industries added roughly 320,000 jobs above trend since 2023, and that hiring spilled straight into electricians, HVAC techs, plumbers, utility crews and commercial construction. I wrote in June that construction was short 349,000 workers. The AI boom appears to be competing for the same people who renovate your units. Your subcontractor quotes may be the first place the AI economy shows up in your P&L.
- -Manufacturing profits are booming. U.S. manufacturers earned $359.4 billion after tax in the second quarter, up $143.8 billion from a year ago, when the figure was $215.6 billion (Census Bureau, Quarterly Financial Report). That's a 66.7% increase in profit in twelve months (calculated). Sales grew too, to $2.32 trillion, 17.4% (calculated). The economy's mood is terrible (consumer sentiment sits at 47.8, near a record low, per University of Michigan), but the economy's numbers are not. Vibes are mostly political. Remember this.
Fun Things Happening in Nashville This Week:
- Touch of Brews, Saturday, Sept 26, 11 a.m. to 9 p.m., Two Rivers Mansion in Donelson. Craft beer, food trucks, bands channeling the Grateful Dead, and a kid zone, on the grounds of one of the prettiest antebellum houses in the county. (Nashville Guru; Touch of Brews)
- Kacey Musgraves, Sunday and Monday, Sept 27 and 28, 7:30 p.m., Bridgestone Arena. Two nights of the Middle of Nowhere Tour with Flatland Cavalry opening; doors at 6, tickets from $64. If you only do one show this month, this is the one. (Bridgestone Arena)
----
Apartment Demand Just Beat Supply by 45,000 Units
Let's start with the number that should be on the front page of every real estate section.
In the second quarter, nationally we absorbed roughly 164,000 apartment units and delivered roughly 118,000. Demand beat new supply by more than 45,000 units in ninety days. Absorption was up 13% from a year ago. Deliveries were down 22% (CoStar, Aug 6).
Source: CoStar / Apartments.com, Q2 2026 multifamily vacancy update (Aug 6, 2026).
National vacancy fell to 8.2%, down 26 basis points on the quarter and 14 from a year earlier. Grant Montgomery, who runs multifamily analytics at CoStar, put it plainly: "the national supply-demand balance has become increasingly favorable over the past year" (CoStar). A week later CoStar raised its rent growth forecast, lifting the third-quarter number from 0.8% to 1.4% and the fourth-quarter number from 0.5% to 1.9% (CoStar, Aug 12).
Apartment management software and data company Yardi Matrix has the national advertised rent at $1,773, up $2 in August, with year-over-year growth at 0.4%, which it calls "the highest rate of increase in nearly a year." August was the sixth straight monthly increase (Yardi Matrix, August 2026). Similarly, RealPage logged its first positive year-over-year effective rent growth in twelve months in July. Apartment List has vacancy improving five months running, ending a streak that goes back to late 2021 (Jay Parsons, Aug 20).
Four trackers, one direction. Up.
Back in July I called it "the apartment glut going keto," and the numbers since support my thesis.
But here's the thing you should notice before you get excited…
Those same four trackers cannot agree on where we actually are.
CoStar says national vacancy is 8.2%. Yardi says national occupancy is 94.8%, which works out to roughly 5.2% vacant (calculated). That's a 300 basis point gap, on the same country, in the same quarter. And while CoStar says vacancy is improving, Yardi's occupancy figure actually slipped 20 basis points in August even as its rents firmed (Yardi Matrix).
Sources: CoStar / Apartments.com Q2 2026 vacancy update; Yardi Matrix Multifamily National Report, August 2026. Implied vacancy calculated as 100% minus 94.8% occupancy.
Neither of them is lying. They're counting different buildings. And once you understand which buildings, the rest of this issue falls into place, because it's the same reason the "Nashville rents are falling" headline may be describing a market you don't own.
----
The Recovery Is Real. Is It Yours?
Strip away the persiflage about an apartment recovery and there is exactly one question that costs you money: what rent growth do you put in the model for next year?
If you own in Nashville, like me, the headlines sound scary. Every major tracker has Nashville rents negative year over year. Yardi has us at −1.1%. CoStar has us at −0.6%. Apartment List has the city median at −2.9% (Yardi Matrix; Matthews, CoStar data; Apartment List). Nashville rents fell half a percent in August alone, one of the four weakest months among Yardi's top 30 metros, while the nation as a whole rose a tenth of a percent (Yardi Matrix).
So the country turned and Nashville didn't.
But that's the story you'll read elsewhere.
I think that narrative measures the wrong thing. Let me show you why, and then let me show you what I think is actually happening, which is that Nashville may be the market coming out of the supply wave in the best shape of anyone in the Sun Belt.
Pick your data
Start with the fact that the two largest data providers in commercial real estate disagree about Nashville's next twelve months by an amount that should make you nervous about all forecasts.
CoStar, via Matthews' quarterly report, says Nashville rent growth turns positive and finishes the year at +2.2% (Matthews). Yardi Matrix, in its August national report, forecasts Nashville at −1.3% for year-end 2026 (Yardi Matrix).
Sources: Matthews, Nashville Multifamily Market Report Q2 2026 (CoStar data); Yardi Matrix Multifamily National Report, August 2026, p. 11.
That's 350 basis points apart (calculated), on the same metro, for the same twelve months. One says we inflect positive by December. The other says we're still falling.
Let's clear this up.
Which buildings get counted
Yardi's August report includes a finding that explains: rent growth, in its data, "is highly correlated with new properties in the lease-up phase" (Yardi Matrix). Markets with a big share of stock still in lease-up show weak rents. Markets that have worked through it show firming rents. Simple.
And Nashville still has a lot of stock in lease-up. About 8.9% of it, the fourth-highest share among Yardi's top 30 metros, behind only Charlotte at 11.6%, Austin at 10.9% and Phoenix at 9.8% (Yardi Matrix).
Source: Yardi Matrix Multifamily National Report, August 2026.
Think about what that means mechanically.
Nearly a tenth of the apartments in this metro are a brand-new building with a banner out front offering two months free. When you compute a metro-average rent, those units drag the number down, and they drag it down harder in an advertised-rent index (Yardi) than in an effective-rent index that leans toward stabilized product (CoStar). That's the 350 basis points. It's also the 300 basis points in the national vacancy gap. Same fingerprint. You can even see it in the construction counts: for Nashville, CoStar counts 11,486 units under construction and Yardi counts 16,686 (Matthews; Yardi Matrix). A 45% difference (calculated) in what's in the pipeline, because they draw the boundary differently.
Now, here's what the bears would say. CoStar titled its own second-quarter analysis "As Most Renters Focus on Newest Apartments, Existing Properties Face Elevated Vacancy Risk" (CoStar). Read that as an owner of a 1985 building. The recovery may be real, and it may be flowing to the newest product, because a building in lease-up can buy a tenant with concessions that a 40-year-old asset can't match.
I'll come back to that. It's the best counterargument. But first, the case for Nashville, because I think it's stronger than the case against.
Nashville's Supply Glut is Over
Here's the number I'd put in front of anyone quoting doom about this market.
CoStar puts apartment vacancy in the South at 10.8%. That's the worst region in the country. The national figure is 8.2%. Austin, the market Nashville always gets compared to, is at 12.3% (CoStar; Matthews).
Nashville is at 5.40%.
Source: CoStar / Apartments.com Q2 2026 vacancy update (national, regional); Matthews' CoStar-based Q2 2026 market reports (Nashville, Austin).
Half the regional rate. Exactly half, as it happens (calculated). Nearly three full points below the nation. Sitting next to the Northeast and the Pacific, the two tightest regions in America, rather than next to Austin. Nashville absorbed the biggest apartment building boom in its history, more than 35,000 units since 2023, and came out with a vacancy rate that looks like Boston's.
That is not a market that's struggling.
That's a market that cleared.
Why?
Demand.
Nashville absorbed 2,755 units in the second quarter, up 14.7% from a year ago, against 1,189 delivered. Demand ran at roughly 2.3 times new supply (calculated), in a quarter when inventory was still arriving. Vacancy fell from 6.21% to 5.40% in ninety days (Matthews).
Meanwhile the supply wave, which peaked last summer, has fallen off a cliff. Trailing twelve-month completions are down 27.9% to 6,988 units, against 12,597 delivered in 2024 (Matthews). That's a 44% drop from the 2024 pace (calculated). Units under construction have fallen to 11,486, which is 5.51% of inventory, down from a peak of 15.87% in the third quarter of 2022. Completions are forecast to slow again to 5,164 by year-end (Matthews). And that lease-up share I mentioned, the drag on the average? It peaked at 14.2% in March 2024. It's 8.9% now (Yardi Matrix). Down by more than a third (calculated).
Source: Matthews, Nashville Multifamily Market Report Q2 2026 (CoStar data); Yardi Matrix, August 2026 (lease-up share).
Every one of those series points the same direction, and none of them reverses quickly. You cannot start a building you didn't finance two years ago, and nobody was financing apartment buildings in Nashville two years ago at 7% money.
Pricing tells the same story. Nashville apartments are trading at $219,173 per unit, just 5.4% below the 2022 peak of $231,654, while most peer Sun Belt markets remain 10% to 18% below their highs. Cap rates have held essentially flat near 5.44% for two years. And second-quarter investment volume more than doubled, to $404.4 million from $200.1 million a year earlier (Matthews). Values held, cap rates held, and the buyers came back, through the worst of the supply wave.
I wrote about the East Bank pipeline in June and about the rent turn in August. What's new this month is that the numbers behind both pieces have gotten better, not worse, but the doomer narrative didn't notice.
What the rent roll says
So why is the metro rent number negative if all of that is true?
Because it's not your rent number.
Every index I've quoted, Yardi's −1.1%, CoStar's −0.6%, Apartment List's −2.9%, measures advertised, asking, or new-lease rent. That's the price of the marginal unit in a concession war. It's the number on the banner. It is not what happens to the resident who already lives in your building and decides to stay.
That resident is measured separately, and the national number looks like this. Renewal lease rent growth is running at about 3.5% so far in 2026, per RealPage. Move-in lease rents, the ones the headlines track, are essentially flat year over year. And 56% of residents whose leases expired in the first half of 2026 signed a renewal, against a 2010 to 2019 average of 51.5% (RealPage data, via LeaseLock chief economist Greg Willett, July 30, 2026; RealPage Analytics). Historically strong retention. Turnover is down. And two details from the same data that matter for what comes next: renewal increases show "very little difference" across regions or asset quality levels, and the retention stats exclude most renters behind on rent, so distress isn't propping up the number (LeaseLock; RealPage). Renter mobility has been sliding for a decade as the typical renter ages and the for-sale market stays out of reach at 7.2% mortgage rates.
Sources: RealPage data via LeaseLock (July 30, 2026); CoStar via Matthews; Yardi Matrix (Aug 2026); Apartment List (Sept 2026). Class B renewal figure is the author's own portfolio.
Now for some operator evidence.
Our renewals on the Class B apartments we own and manage in Nashville are up 3% this year. Tenants are staying put and absorbing a small bump, about the same as general inflation.
A 3% renewal bump on stabilized Class B is not a lot, but it's not flat and definitely not negative. It's within half a point of the national renewal number, and the national number barely varies by asset class.
But the gap between "Nashville rents fell 1.1%" and "our renewals are up 3%" is not a disagreement about facts. It's two different lease types being reported as though they were one number. The headline describes the building with the banner. The rent roll describes the resident who renewed.
This is also, incidentally, the answer to CoStar's warning about existing properties facing elevated vacancy risk. It's a real risk on the day a lease expires and the tenant walks across the street to the new building with two months free. But 56% of them aren't walking, and the ones who stay are paying 3.5% more. The concession war is a war for the move-in, and the move-in is a smaller share of your revenue than the renewal.
In my book, The 5 Ways Real Estate Investors Make Money and Build Wealth, I treat cash flow and appreciation as separate engines because they respond to different forces. This is a clean example. The headline rent index is an appreciation signal; it tells you what a buyer might pay for the next building. The renewal is the cash flow engine; it tells you what your building pays you. Right now those two engines are reading different gauges, and the one that pays the mortgage is the one that's green.
A Free
Want help running numbers on your next deal (or an existing property you have)?
I use the Deal Lab free real estate deal analyzer: DealLab, the deal analyzer.
Run your next deal through it. Bookmark it. It will come in handy.
The US Consumer/Tenant is Strong
The renewal bump only holds if the renter can afford it. So let's ask the people who can actually see the bank accounts: Bank CEOs.
The three largest consumer banks in the country, which between them hold the deposits and card balances of a very large share of American households, are saying roughly the same thing.
Brian Moynihan, who runs Bank of America, said this month that "credit statistics are as good as they've been for years." The bank's data shows customers spent about 4% more in August than in August of last year, a pace he described as consistent with a growing economy. Asked about the scary headlines on record card balances, his answer was that "the economy is 40% bigger," so "what you see is a recovery back to trend" (CNBC, Sept 9; Axios, July 15).
Charlie Scharf at Wells Fargo: "Consumer trends and resilience have been very strong. It's hard to see that changing in the short term." Wells reports spending still rising, delinquencies and charge-offs low, savings and investment balances growing, net loan charge-offs down 10 basis points from a year ago, and deposits up 10% to $1.5 trillion (Wells Fargo, Q2 2026; Yahoo Finance).
Jamie Dimon at JPMorgan: "customers are still in really good shape from our perspective," with the economy "resilient" and "consumers continue to spend" (Investing.com). On JPMorgan's second-quarter call, CFO Jeremy Barnum reported that "consumers and small businesses continue to show resilience despite elevated gas prices and inflation," and the bank actually lowered its expected card charge-off rate to about 3.2%, "reflecting better-than-expected consumer credit performance" (JPMorganChase, 2Q26 transcript).
Add this to the recent economic data: Jobs up 162,000 in August against a 56,000 estimate, with unemployment steady at 4.1% (BLS). Manufacturing profits up two-thirds. Card charge-offs coming in better than expected. Deposits growing double digits. Consumer sentiment at 47.8. The mood and the money are pointing in opposite directions, and I wrote in August that in this economy you should follow the money. Again, vibes are political.
The renter is the consumer. If the consumer's credit is clean and their balances are growing, the renter renews instead of moving, the operator gets a 3% bump on a stabilized Class B asset, and none of it shows up in an advertised-rent index that's busy measuring the lease-up building down the street.
--
The Steelman: The Case That I'm Wrong About Nashville
I'm bullish on this market and I've just spent two thousand words telling you why. So here's the best version of the other side, and I want to be honest that parts of it are strong.
One: the vacancy number may be flattering. CoStar's 5.40% for Nashville leans toward stabilized product. Yardi's occupancy figure for Nashville is 93.7%, and its non-stabilized occupancy, the lease-up buildings, is 91.9% (Yardi Matrix). So the "half its region" chart is real, but it's the flattering cut. On Yardi's numbers Nashville is around 6.3% vacant (calculated), which is still well under the region and the nation, but it's not Boston.
Two: capital paying near-peak prices is not the same as capital being right. The 2022 peak was set at a 3% cost of debt. Today agency multifamily debt is quoting around 5.6% to 6.2% (Select Commercial), which is above the 5.44% cap rate. That's negative leverage. It can mean buyers see the supply cliff and are paying for what's coming. It can also mean pricing is stale rather than resilient, and the repricing hasn't happened yet because sellers haven't been forced. I don't know which. Nobody does.
Three: the renewal argument has a ceiling. Renewals at 3.5% while move-ins are flat is a spread, and spreads close. Either move-in rents recover and the spread closes from below, which is the bull case, or renewal pricing power fades as tenants realize they can get two months free across the street, and the spread closes from above. RealPage's own data shows renewals outpacing move-ins because move-ins are weak, not because renewals are strong in absolute terms. And the move-in weakness is not evenly spread: stabilized upper-tier product is getting about 1.9% on new leases, Class B move-ins are flat, and Class C move-ins are down 2% (LeaseLock, July 30, 2026). So a Class B owner is living on the renewal book right now, and RealPage expects rents to "back up a little" seasonally in late 2026, though less than last year's cuts. If you own the 1985 building and the 2024 building next door finishes its lease-up and starts competing on renewals too, your 3% may become 1%.
Four: the consumer is fine until the day it isn't. Every bank CEO I quoted was describing a loan book as of the second quarter. Since then the Fed has hiked, the 10-year has touched 5.04%, and the dot plot says more is coming. The renter's resilience was measured before the tightening, not after. And the same oil shock that's now unwinding could re-escalate on a single headline out of the Red Sea.
Five: I own the asset I'm defending. I manage Class B in this market. You may want to discount my read accordingly, the same way I told you to discount the bank CEOs. I've tried to build this piece on CoStar, Yardi, RealPage and the Census rather than on my own book, and I've shown you the calculations, but we all have bias. Though I do put my money where my mouth is, I'm actively looking for my next multifamily deal here in TN (and if you have one, email me! I pay a generous referral fee).
Here's what I'd say to all five.
The bear case is a case about timing and about which buildings. It is not a case about direction. Supply is down by two-thirds from its peak and cannot come back quickly. Absorption is running at more than twice deliveries. Vacancy, on either vendor's cut, is well below the region. Renewals are positive on the national data and on mine. The bears may be right that the next two quarters are choppy, and I'd take that under advisement.
I don't think they're right that Nashville is the laggard in this recovery. I think Nashville is the market that already did the hard part.
Key Takeaway For The Investor
If you own stabilized product in Nashville: your rent roll is a better indicator than the metro index, and your renewal book is the number to watch. Concessions are a move-in problem. Retention is your revenue.
If you're buying: underwrite both vendors' forecasts and price the spread. A deal that works at −1.3% rent growth and sings at +2.2% is a deal. A deal that only works at +2.2% is a data subscription.
If you're in Franklin or Brentwood, where 12.5% of inventory is still under construction, you're in a different market than South Nashville, where it's 1.1% (Matthews). East Nashville led the metro in rent growth at +1.6% year over year, the only submarket above 1%; Central Nashville lagged at −3.6%. The metro number is an average of markets that are years apart in their cycle.
And if you're reading the doom: ask yourself which buildings they counted.
One more thing, since this whole issue is about two credible sources reaching opposite conclusions from the same market. The book I hand people who want to get better at that is Nate Silver's The Signal and the Noise. It's about why forecasts fail, and the short answer is that the forecaster usually knows less about what's being measured than they let on. Every chart in this issue is an example. If you read one thing on how to think about a vendor's number before you put it in your model, make it that.
My Skeptical Take:
The apartment market turned. Four independent trackers say so and I believe them.
The apartment market also can't agree on where it turned from, by three percentage points nationally and three and a half in my own city. That should bother you more than it bothers the people who publish the numbers.
Here's what I take from it as an owner.
The metro rent index is measuring the marginal new lease in a concession war. It's a real number. It's the number a lease-up building lives and dies by. It is not the number my Class B tenant pays when she renews, and 56% of tenants nationally renewed in the first half of this year at rents 3.5% higher than before. Mine did too, at 3%.
Nashville's supply wave peaked last summer. Completions have fallen 44% from the 2024 pace. Construction is down two-thirds from its peak as a share of inventory. The lease-up overhang that drags our average has fallen by a third. And through all of it, vacancy here is half the regional rate, absorption is more than double deliveries, and pricing is within 5% of its all-time high.
To say it plainly: Nashville is the market coming out of the Sun Belt supply cycle in the best shape of any of its peers.
The bears' best point is that the recovery is flowing to the newest buildings first. Fine. But it always does. Then the newest buildings stabilize, stop giving away rent, and become the comps that lift everybody else. That's not a threat to a Class B owner. That's the sequence.
So when someone shows you a rent number, ask which buildings they counted and which lease type they measured. Those two questions would have saved more people from bad underwriting than any cap rate ever has.
The sociologist William Bruce Cameron wrote the line that gets misattributed to Einstein, which is itself a nice reminder to check the source:
"Not everything that can be counted counts, and not everything that counts can be counted."
— William Bruce Cameron, Informal Sociology (1963)
The advertised rent can be counted. The tenant who stayed is what counts.
Don't believe the doomer news.
Read the rent roll.
Until next time. Stay Curious. Stay Skeptical.
Herzliche Grüße,
-The Skeptical Investor
-P.S. If this newsletter has been useful to you, share it! I promise you 24 hours of good joo joo.
Comments