The Maturity Wall: The CRE Debt Coming Due Right Now

Balloons, Not Mortgages

A homeowner's mortgage amortizes down to zero over thirty years month by month until the debt simply disappears. A commercial real estate loan almost never works that way. The standard structure lasts five seven or ten years with payments that are interest-only or only partially amortizable and then a balloon: The entire remaining balance due in a single payment on a single day. No one actually plans to write that check. The plan always is to refinance with a new loan or sell the building before the due date arrives. For a decade that worked wonderfully because credit was cheap and values kept rising and the industry made loan after loan without much drama. maturity wall It's what analysts call an unusually large group of these balloons landing at the same time. I think the current one is built on two stacked layers: the huge volume of loans made at rates of 3 and 4 percent between 2016 and 2021 plus a second layer of loans that were already due in 2023 2024 and 2025 and that were extended rather than resolved piling up on everything that lies ahead

The Wall by the Numbers

Industry estimates put roughly $875 billion of commercial and multifamily mortgage debt due in 2026 alone. It's one of the largest refinancing waves ever recorded and it's not slowing as expected. Deferrals continue to advance old balances so rating agencies now expect the wall to persist through 2027 rather than disappear. Volume alone isn't really the problem. The problem is the gap between the worldin which these loans were taken out and the world in which they now have to mature

Two rate points don't sound catastrophic on their own until you remember that the value of a commercial property is just its income divided by a yield the cap rate a relationship our separate cap rate explainer explains in detail. If that yield is raised the price a buyer or lender can justify drops immediately and arithmetically with no sentiment involved. Office values are down about a third on top of that for reasons our Office reset piece covers alone. IfPut the two together the maturity wall is actually a problem of rates and a problem of values that collide at the same time

Why One Rate Can't Refinance Into Another

This is the part I found really confusing when I first tried to model a maturing loan so let me go slower. A commercial mortgage does not depend on the borrower's credit score or a general idea of how healthy the property looks. It is calculated from two mechanical tests and the lender accepts whichever answer is lower

The first test is the loan to value. The lender chooses a maximum percentage commonly between 60 and 70 percent for a currently stabilized property and multiplies it by the appraised value of the property. The second test is the debt service coverage ratio or DSCR which asks how much of the property's net operating income is left after the mortgage payment. A lender sets a minimum DSCR say 1.25 or 1.35 meaning the income must exceed the annual mortgage payment by that multiple. Divide the maximum debt service coverage ratio the coverage ratio allows by the annual cost of borrowing one dollar today call it the mortgage constant (more on that below) and you get the largest loan that will support the proof ofcoverage. Lenders call the lower of the two numbers the testing procedures and it is the most important mechanism in this entire story

When a loan is originated both tests are applied at whatever rate and value prevails that day. Between 2016 and 2021 rates were low enough that the coverage test rarely affected. Servicing a dollar of debt cost so little that most stabilized properties could carry much more debt than a 65 percent-to-value loan would require so LTV was almost always the binding constraint and loans approachedThat's the silent reason why much of this debt was issued with high leverage. The hedging math just didn't get in the way

If you move the calendar to a refinance in 2026 both tests will move against the borrower at once. The mortgage constant the annual cost of servicing a dollar of new debt is much higher because rates are much higher.The borrower is not being punished for anything he did. He is asked to pass two tests simultaneously and both became more difficult using income that mostly did not grow to match

The Refinancing Gap, Worked

Here's a concrete case and it's worth looking at because every number in it does real work. Take an office tower purchased in 2019 for $100 million financed with a $65 million loan with interest at 4.25 percent. The building earns $5.5 million a year in net operating income the interest bill is $2.8 million and back then no one cared about any of that. Now it's 2026 andthe loan comes due. After the office revaluation the building is worth perhaps $65 million well below the purchase price. A new lender will advance about 60 percent of the current value or $39 million or about 6.5 percent. The borrower owes $65 million on the maturing loan and can raise $39 million on the same building. That $26 million hole is the refinancing gapAny way to close it is painful: pumping fresh capital into a building worth less than the price paid for it negotiating yet another extension selling in a distressed market or handing the keys over to the lender. The property is still collecting rent through all of this. No one stopped renting apartments. The maturity wall isn't really about buildings failing as businesses. It's about capital structures failing like arithmetic

A Second Pass: Sizing Off Coverage, Not Just Value

The tower above turned out to be limited by the loan-to-value the 60 percent advance rate. It won't always be the loan-to-value test that is more difficult and I think that case is actually more instructive so let me construct a completely illustrative example made up numbers for the exercise to show that the coverage ratio causes the damage

Suppose a different property call it a mixed-use building that no one actually owns has an overdue loan balance of $50 million. Today it earns $3.5 million a year in net operating income and its appraisal is $70 million a cap rate of 5 percent 3.5 divided by 70. The new lender will advance up to 60 percent of that value which is $42 million 0.60 times70. In loan-to-value alone the borrower would be only $8 million short of the $50 million balance. It's not fun but you can survive it

Now run the coverage test alongside it. Let's say the lender requires a minimum debt service coverage ratio of 1.35 which means that the net operating income must be at least 1.35 times the annual mortgage payment. The maximum annual payment the lender will allow is 3.5 million divided by 1.35 which is equal to about $2.59 million. Next comes the mortgage constant the annual payment per dollar borrowed at the interest rate and calendarcurrent amortization rates. A loan that amortizes over 30 years at 6.5 percent carries a constant mortgage of about 7.6 percent meaning each dollar borrowed costs about 7.6 cents a year in combined service principal and interest. Divide the maximum payment allowed by that constant: 2.59 million divided by 0.076 is about $34.1 million. This is the most this property can borrow justunder the coverage test almost $8 million less than the loan-to-value ratio would have allowed

the testing procedures Take the lower of the two answers and here that's the restricted coverage 34.1 million not the restricted value 42 million. Against a balance of $50 million the actual gap is 50 minus 34.1 or about $15.9 million almost double what the loan would have suggested to value the math alone. This is the part I think is overlooked in the informal coverage of the wall of maturity. Two properties can show exactlythe same appraised value and exactly the same loan-to-value ratio and still refinance very differently because one of them comfortably passes the coverage test and the other does not. Income not just value decides how much debt survives the refinance

The End of Extend and Pretend

Extend and pretend is the industry nickname for a lender that gives more time to a maturing loan rather than forcing a resolution effectively agreeing to pretend that the old appraised value still holds. From 2022 to 2025 this was the dominant strategy and honestly it wasn't stupid. Foreclosure immediately crystallizes a loss against the bank's capital. The extension preserves the possibility that falling rates or rising revenues will quietly repair the math before anyone hasto admit something. The problem is that the bet only pays off if rates really go down or revenues really go up before the extension expires. Three years later none of that has happened on the scale the bet called for

The tone has changed. Recent extensions are being drafted in months not years. Special servicing rates on office loans reached 15.8 percent by the end of 2025 nearly one office loan in six and servicers are increasingly resolving rather than renewing them through discounted payments note sales and foreclosures. Distressed office sales at discounts of 50 to 70 percent off previous values are doing the trick.grim and useful work of setting real prices that the entire market can finally see

Who Is Holding the Bag

Banks own about half of all commercial real estate debt and that exposure skews heavily toward regional and community banks for whom CRE is often the largest loan category on the books. That's exactly why New York Community Bancorp's early 2024 scare shook the markets so much. It foresaw how quickly a CRE provision can eat into a small bank's profits in a quarter.Delinquencies give the clearest danger signal in real time life insurers which typically lent conservatively with low leverage and by most accounts are sleeping well and a rapidly growing army of debt funds that raised billions of dollars precisely for this moment. Those funds supply mezzanine debt and preferred equity an expensive equity spread that sits between what a new senior lender will advance and what the old loan actually requires. I'm not going to explain here how that layer is structured. It's a deep topic and deserves an article on its own

Case Study: 1740 Broadway

The clearest real-world example of this exact mechanism in which a building's value and its debt-service capacity are wiped out at the same time is 1740 Broadway an office tower in Manhattan that Blackstone bought in 2014 in a deal valued at just over $600 million. The purchase was financed with a mortgage that the press at the time put at about half that price. For years the building performed well as a well-leased office tower.in Midtown and with reasonable debt

Then two things happened at once which is the point of this article. Interest rates rose sharply from their previous lows so any refinancing would have to overcome a much tighter mortgage constant than the original loan. Office values across Manhattan also fell sharply driven by the same remote work shift and the same capped rate expansion elsewhere on this site so the appraised value supporting the building also fell. When the loan came dueReports at the time described an appraised value that had fallen to a fraction of what Blackstone paid for the tower less than a decade earlier well below the loan balance still owed. Instead of injecting fresh capital to close that gap Blackstone decided to let the loan default and return the building to its lender rather than pay off the debt on an asset now worth less than it was owed

I want to be precise about what this case shows and what it doesn't show. It's not proof that Blackstone manages distressed buildings. It's one of the largest and most sophisticated real estate investors in the world and that's exactly why the case is important. A company with so much analytical power did the math on refinancing a real building and concluded that paying the difference was worse than giving up equity. If the math isn't made clear for such a homeowner it won't be made clear by dint of optimism for a smaller less capitalized homeowner facingthe same maturity

Why This Is Not 2008

The comparison everyone turns to is the financial crisis and in most cases it doesn't fit. That crisis was due to household mortgages securitized subprime debt and overnight financing that could disappear in a single day so it unfolded at the speed of a bank run. The CRE restart today looks different by design. It's institutional secured and spread over years of staggered maturities rather than a weekend of panic. Herethe first loss is suffered by shareholders not depositors. Banks have been provisioning for these losses for three years now and the most affected asset class offices is a known quantity whose price changes daily in REIT share prices and CMBS spreads rather than a hidden amount buried on someone's balance sheet. The honest case for risk is narrower than a systemic crisis: a handful of overexposed regional banks and municipal budgets that lean heavily on downtown property taxes. Mybase case is not an accident. It is a grueling multi-year transfer of buildings from old capital to new money at reset prices loan after loan extension after extension

Where This Breaks

I have set out the income test as if it were a rigid formula that always produces a gap and here I should argue against it myself because in practice it is not so clear

The first place where the model fails is revenue growth. All of the above assumes that net operating income remained more or less stable while rates rose but that's an office story not a universal one. Industrial and logistics buildings that freed up space at significantly higher market rents from 2021 or multifamily well located in a market with real rent growth may show a 20 or 30 percent NOI increase sinceits origin.If income increased enough the amount of the limited loan amount can actually grow even as the mortgage constant worsens because the numerator moved as much as the denominator.Not all properties on the wall are fighting the same battle as the office towers

The second is initial leverage. A loan originated at 50 percent loan-to-value the type life insurers tend to provide has so much cushion built in that even a significantly lower income test still wipes out the old balance with room to spare. The gap math above is most important for aggressively financed properties near the top of the cycle not those conservatively leveraged

The third is that lenders are not robots executing a formula without discretion. A bank with a long relationship and a healthy balance sheet may choose to extend the debt on easier terms than an outsider would accept accept a partial payment instead of a full and clean refinancing or restructure some of the debt in a way that does not require the borrower to find new cash today. The means test describes what a new lender would offer at arm's length. It does not describe what an existing lender invested in a tidy solution instead ofa cancellation I might actually accept

Fourth is timing. If rates drop significantly before a loan extension expires the entire gap can shrink or disappear as a lower mortgage constant directly increases the amount of the loan with limited coverage. That hasn't happened on a large scale so far but it's the only variable that could make several years of gloomy forecasts look overblown in retrospect. I don't think betting on it is a plan but I'd be dishonest if I said it can't happen

If I'm wrong about how painful this wall is it's probably one of four reasons: earnings surprised to the upside somewhere I wasn't watching leverage was lower than I assumed a lender showed more patience than the formula predicts or rates moved before extension time ran out

How I Actually Use This

My read is that the maturity wall is one of the most manageable things in the markets to model which is rare so this is the order I follow when looking at a REIT or CMBS deal with debt about to mature

I start with the maturity schedule which most REITs disclose in their earnings supplements down to the property or loan level. I want the original rate the original value of the loan and the maturity date because that tells me how far the mortgage constant needs to move before it starts to bind. Then I calculate today's NOI and today's cap rate for the property or borrow the REIT's disclosed cap rate assumption if it provides it and go back to a value ofappraisal. From there I perform both tests myself: loan-to-value at whatever current market advance rate and hedging at any DSCR a lender would possibly require for that type of property then I take the lowest answer exactly as a real lender would. Any gap that arises I compare it to the borrower's actual balance sheet because a REIT with cash on hand and access to the public equity markets can fill an $8 million gap without much drama while a single asset owner with no other liquidity cannot.can do it

The habit that changed the most for me is treating the coverage ratio with the same seriousness as the loan-to-value. I used to only check whether a property was worth more than its debt which is the intuitive question and it turns out to be the first wrong question for many of these properties. A building can be worth much more than its loan and still not manage to refinance the entire balance if the coverage calculations are not clear at the current exchange rate. I got this backwards the first time I tried it I calculated a value ofhypothetical refinance alone and was surprised when I redid it with coverage and got a lower number. That surprise is basically the entire thesis of this article

The other thing I pay attention to reading company language rather than doing the math is how a REIT describes its own extensions. "Extended on favorable terms" works a lot in a sentence like that and I've learned to ask what rate and what term is actually attached to it because a one-year extension at a market rate is a very different signal than a five-year extension at a below-market rate that a lender granted because it had no better option. None of this is a signal to buy or sell anything. It's away of reading a balance sheet that I think is underused because the arithmetic is really simple once you know which two numbers to divide

The Bottom Line

About $875 billion of real estate debt comes due in 2026 with amortization at about 4.3 percent and refinancing at about 6.2 percent versus office values falling by about a third. The gap between what a maturing loan owes and what a new lender will actually advance is not a feeling it is the mechanical result of two tests the loan-to-value ratio and debt service coverage and a lender accepts thehigher rates reduce the restricted amount of coverage directly lower values reduce the restricted amount of value separately and many properties are now failing to do both at once. The gap is being filled with new capital expensive surrender capital extensions that are now measured in months instead of years or keys and the era of extending and pretending to put off this reckoning is visibly ending. This is a slow observable stress test not a crash. The numbers it's worthworth tracking in real time are regional banks' CMBS delinquencies special servicing rates and CRE provisions and beneath all three is the same two-line calculation: what the coverage ratio allows what the appraisal allows and which is lower