I’ve been thinking about debt a lot lately. Seems like everyone else has too, whether it’s debt on your home, for your business or for the country. When my dad, George, started our development company back in the late 1970s, he had one core strategy for the business: find great corners, lease to high credit tenants on long term leases, and use long term fixed rate debt to optimize cash flow. I’m proud to say we’ve been on that exact path as we now head towards our fiftieth year of business. But the last part of the strategy has been the most difficult as of late. Long term fixed rate debt is no longer attractive.
The Old Math
For many years, when we opened the doors on a new shopping center, long term debt was accretive to cash flow on day one. We built a reasonable spread between our yield on cost going into the deal versus our anticipated market capitalization rate at stabilization. We assumed rates and values would remain relatively constant by the time we got our certificate of occupancy. That discipline helped keep our pencil sharp: the deal had to underwrite based on its own merits, not an uncontrollable change in macroeconomic conditions which would help (or hurt) value. And in past years, we were blessed with double good news at completion: the value of the income stream we created had risen and interest rates had fallen. This allowed us to refinance out a good chunk of initial equity and produce cash flow in excess of our pro forma projections. Refinancing the construction debt with permanent fixed rate debt, usually with institutional lenders, became a meaningful event to pull forward our value creation. In what may be our final execution of this strategy in February 2022, we signed a term sheet to refinance Church Street Commons, a Publix-anchored center we had just completed in Burlington, North Carolina. The lender was an insurance company that offered $20.5 million at 71% LTV, a 3.99% fixed rate, on a ten-year term (five years interest-only, then amortizing on a thirty year schedule). That gave us over 200 basis points of positive spread on day one against our modeled capitalization rate.
We locked this rate in late February 2022, about three weeks before the Fed’s first hike since 2018. As a matter of practice, we rarely locked in a rate prior to closing. Debt markets had been stable and predictable for a long period, and there was little to gain by locking early (in fact most times, the rate moved a few basis points in our favor between term sheet execution and closing). But we had a feeling the world was changing quickly in early 2022, so we paid the early rate lock deposit (not a small amount on a $20.5 million loan), and secured our financing. Our spread was 205 basis points over the ten-year, which was sitting near 1.94% at the time. By the middle of June 2022, when we closed, the ten-year alone had reached 3.49%. The risk-free rate had climbed to within fifty basis points of what we were paying on a leveraged shopping center. I didn’t fully realize it at the time, but the value creation at refinance that we had enjoyed uninterrupted for decades had just finished its last full cycle.
What Broke
It would be easy to say rates went up and leave the story there, but that misses what changed in the business.
For most of my working career spanning the last 25 years, the annual cost of carrying a permanent loan sat below the property’s capitalization rate. That gap is the entire reason borrowing made us money instead of costing us money. Every dollar of debt we placed was accretive to leveraged cash flow, and the wider that spread, the more the loan did for us. At Church Street Commons, we created real value by creating a Publix anchored center at a great corner. But there’s no question that the refinancing super charged the return.
Today the gap is gone. On the long-term debt we are quoted now, once you account for amortization, the annual cost of debt service lands at or above where these assets are valued. Long-term borrowing no longer is the default strategy. In some cases, it actually undermines our strategy. And that’s assuming your original proforma wasn’t predicated on value creation at refinancing in the first place.
The refinance used to be the “automatic” for the value creation in real estate development, the moment when the current and future value we had built got recognized immediately and some of our equity came back to us to go build the next project. Now the refinance is the cleanup. It retires the construction loan (and hopefully the recourse risk as well), it takes out the preferred equity that was patient enough to wait, and more often than not it asks us to write a check rather than receive one. Today, this event is certainly not the cherry on the sundae it once was.
This predicament is not unique to shopping center development. Something on the order of
$875 billion in commercial mortgage debt matures this year and another $652 billion next year, much of it written when money cost 3.0% – 4.0% and now facing 6.0 % – 7.0% or higher. The industry is spending a great deal of energy on whether those borrowers can hold on until conditions improve.
Which is the right question, but not the one I find myself asking. I want to know whether conditions are going to actually improve.
Why This Isn’t a Cycle
The 10-Year Treasury now hovers around at 4.74%. That is seventy-five basis points higher than the rate we locked in for a decade in 2022. The federal government now borrows long term money at a substantially higher cost than a Publix-anchored shopping center in Burlington, NC did four years ago.
Ask “how did we get here” to anyone who analyzes interest rates and bond yields for a living, and I think you mostly get one of four answers, depending on who you ask.
The first, and from what I find, most common, is inflation, which has been sitting above the Fed’s 2.0% target long enough that the market has stopped assuming it will drift back down on its own. The second is the Fed itself as an institution, where a new chair and an open question about how hard the committee intends to fight inflation has left investors unsure what they are pricing. The third is corporate borrowing. American companies have issued nearly $1.7 trillion in bonds so far this year, up 27% from the same period last year and more than all of 2025 combined, most of it to build AI-focused data centers. The fourth is the volume of federal debt, which crossed $40 trillion this month.
All four are real, but three of these are temporary. One is not.
Inflation will likely ease from one of two directions. Either the supply shock eases, as the disruptions in the Gulf that have kept energy prices elevated this year eventually work themselves out, or demand gives way under the weight of the rates being used to restrain it. The market will adapt to this new Fed leadership, and either way, Fed governors have term limits attached. And the data center build-out, for all its scale, is, well, a build-out. Trees do not grow to the sky, and at some point, AI infrastructure will be sized to the demand it serves, the spending will slow, and the cost of that capital will revert toward a mean. We have watched this movie in our own industry more than once. Every asset class that has ever been undersupplied eventually stops being undersupplied, and the capital that rushed in to fix it goes looking for the next thing.
Government debt, though, is different in my view. It is not a build out and it does not have a completion date. Nobody is going to announce that the deficit has been sized to demand. That is the one input on the list with no natural stopping point, and the one I have been thinking about most lately. If there is one constant over the last several U.S. administrations, it’s that politicians on either side of the aisle don’t seem remotely constrained by borrowing costs. Which means, one way or the other, more debt is in our future regardless of which party is in power.
In that reality, how does the Treasury, the foundation of our borrowing, retain its special status?
I’m reading a lot on the convenience yield. For most of modern financial history, investors paid a premium to own Treasuries. Not in price exactly, but in yield given up. Treasuries were the safest and most liquid instrument on earth, and buyers accepted a lower return for the privilege of holding them.
A paper published this spring by Wenxin Du at Harvard, with Ritt Keerati and Jesse Schreger, takes the measure of that premium and finds it has largely disappeared1. Their central finding is that the dollar and the Treasury, long treated as two expressions of the same American advantage, have come apart. The dollar is still special. The Treasury is not. On a currency hedged basis, our government now pays more to borrow than several of its peers do, which is close to the opposite of how the world worked for most of my father’s career.
Their explanation will sound familiar to anyone who has ever tried to lease up an overbuilt submarket. Outstanding Treasury debt went from roughly $5 trillion in 2008 to about $29 trillion in 2025. Du puts it about as plainly as an academic can: flood the market with an instrument and it loses its specialness, and the price reflects that. Scarcity was always part of what made Treasuries worth a premium. There is nothing scarce about $29 trillion.
And when you consider the US has an aging population and a shrinking workforce, the projection on debt load gets gloomier.
This is the part of the forecast nobody really argues about. The Congressional Budget Office says its projections of how many people will draw Social Security and Medicare are among the most certain it makes, for the simple reason that everyone who will be sixty five in 2040 has already been born.2 Right now there are about 2.7 working age Americans for every person over sixty five. Thirty years from now there will be 2.2.3 The Social Security retirement trust fund is projected to run dry in 2032 and Medicare’s hospital fund around 2040. Spending on Social Security and the major health programs runs 11.2% of GDP this year and heads toward 14% by mid-century, while interest on the debt itself climbs from 3.3% of GDP to nearly 7%. CBO’s baseline has federal debt held by the public reaching 120% of GDP within a decade, against a fifty year average near 51%.4
That’s why I keep coming back to the convenience yield. What drives it is not confidence, not credit quality, not the dollar’s role in trade. It’s supply, pure and simple. The relative volume of government paper outstanding is primarily what determines whether Treasuries command a premium, and the premium eroded as that volume grew from $5 trillion to $29 trillion.
I read the demographic projections as what they are, a supply forecast. The line does not flatten. It steepens, and it steepens for reasons that are already locked in. Whatever you believe about the political will to address deficits, the beneficiaries are born, the workers funding them are fewer, and the interest bill compounds on top. That is more paper coming to market every year, into a market the research already says has more than it wants.
And that is only the supply half. The demand side has its own problem, which surfaced in public this month. Japan holds well over a trillion dollars in reserves, much of it in Treasuries, and a weakening yen creates pressure to sell some of it to raise dollars. Washington has been working to prevent exactly that. The Treasury intervened in currency markets, asked the Fed to expand the facility that lets foreign central banks borrow against their Treasury holdings rather than sell them, and doubled its own long maturity buybacks. Read that sequence for what it is. The largest foreign holder of our government’s debt may need to sell it, and our government is going to considerable lengths to arrange that they do not have to.5 Whether the actions are successful or not, it’s clear additional Treasury demand from our major current holders looks unlikely.
All of this matters greatly to us as real estate borrowers, because none of the debt we seek is ever priced in isolation. When an insurance company quotes a permanent loan on a stabilized center, it starts with the Treasury and adds a spread for the risk of the asset, the tenant, and the sponsor. Going back to Church Street Publix, if you unpack the 3.99% rate we achieved, 205 basis points was a judgment about Publix, about Burlington, and about us. The other 194 basis points had nothing to do with any of it. That was simply what the world charged the United States government to borrow for ten years, and for four decades the world charged remarkably little. Every fixed rate loan our company ever placed for a real estate asset was built on top of that cheap Treasury.
Half of what we paid in rate measured our work. The other half measured somebody else’s willingness to lend cheaply to Washington. The first half is roughly where it always was, and possibly even better today for certain retail real estate assets and sponsors. The second half has more than doubled, and I believe that cost is not changing anytime soon.
That is why I do not read a ten year fixed rate quote today as a temporarily bad price on a good product. I read it as the correct price for something I no longer want to buy.
Borrowing Short
Last October we took a term sheet from a bank to refinance Beaufort Station, a large grocery anchored center we developed in Beaufort, South Carolina. Thirty-five million dollars, five years, priced at one month Term SOFR plus 210 basis points with a swap for the full term and the full amount. All in, roughly 5.60% fixed.
Put that next to Church Street Commons and the number worth looking at is not the rate. It is the spread.
In February 2022, an insurance company quoted us 205 basis points over the 10-year Treasury on a Publix-anchored center in North Carolina. In October 2025, a bank quoted us 210 basis points over SOFR on a grocery anchored center in South Carolina. Five basis points apart, three and a half years and the largest repricing of money in four decades in between.
Two large grocery-anchored centers in the southeast, developed by the same company, financed by two institutional lenders in two completely different markets. And the market’s judgment of the real estate, the tenants and the sponsor came back almost exactly the same. If anything the underwriting on the more recent loan was a touch friendlier than the first.
That is the whole point. Nothing about what we do as developers got worse. Our corners did not decline, our anchors did not weaken, and our track record did not suffer. The part of the cost of money that reflects our work is where it has always been. What changed sits underneath it, in the price of the government’s own borrowing, and I have already made my case for why I think that part stays elevated.
So why sign up for five years rather than ten, if I believe long money is expensive for a long time?
Because a ten year quote today charges me for all four of concerns the market has today
(government debt, Fed leadership, AI build out and inflation) at once. But only one of them in my mind is more-or-less permanent. I would be paying a decade of premium for a situation I think could look a lot different in 3-5 years. Shorter term borrowing lets me pay for the part of the yield I can see and decline to prepay for the part that I want to see play out.
It also keeps us able to move. The Burlington loan carried a two-year lockout and yield maintenance through year ten, which meant that if the world improved, we would have watched it improve and kept paying. The Beaufort loan can be prepaid at any time without penalty. If inflation settles, if the market makes its peace with this Fed, if the data center build-out finds its ceiling, we are positioned to capture that rather than read about it. We will still have the underlying government debt problem, but maybe we end up with a 10-year Treasury closer to 4.0% than 5.0%.
This is a much riskier bet on the long-term financing side than we played for many years, and we acknowledge that. But we’re buying time as much as anything else, and playing the hand in front of us is our business.
Holding Long
So, simply put, we are not getting paid the way we used to at the closing table. What we are still getting is the asset, and I would argue the asset is better positioned now than it was when money was cheap. The same rates that took the juice out of our refinance are keeping a great deal of new supply from breaking ground. Nobody is building a shopping center on a hard corner at these construction costs unless the deal is exceptional, which means the ones already standing get scarcer every year. If inflation runs warm, our rents move with it and our debt service does not. If it cools, we can refinance into the better debt market. Both of those roads run through owning the real estate and waiting, which is what our company has always strived to accomplish. The difference is that we used to realize value twice, once at refinance and again over time, and now for this group of assets, we likely only get paid the second way.
Which is why George’s strategy hasn’t aged badly at all. I think the tailwind we developed into has stopped blowing. Great corners, credit tenants, long leases and patience were never a bet on cheap debt. They were a calculated bet on the scarcity of land, desirable tenants and stable income, and cheap debt simply made that bet pay faster than it otherwise would have. Take it away and the thesis is intact, just slower to the finish line. For now, we are borrowing short to avoid paying premium interest rates while we wait out a repricing that may take several years to sort itself out, and we are holding long because that is where the value was always going to come from for hard assets. Fifty years later, that is still the core tenet of the business.
References
Cited in text
1 Du, Wenxin, Ritt Keerati, and Jesse Schreger. “Decoupling Dollar and Treasury Privilege.” NBER Working Paper No. 35000, March 2026. Also issued as Federal Reserve Board International Finance Discussion Paper No. 1427. https://www.nber.org/papers/w35000
2 Congressional Budget Office. “How Budgetary and Economic Outcomes Might Differ From CBO’s February 2026 Projections.” February 27, 2026. https://www.cbo.gov/publication/62184
3 Congressional Budget Office. “The Demographic Outlook: 2026 to 2056.”
https://www.cbo.gov/publication/61994
4 Congressional Budget Office. “The Budget and Economic Outlook: 2026 to 2036.” February 11, 2026. https://www.cbo.gov/publication/62105
5 Reporting on the U.S. Treasury’s August 2026 currency intervention, the expanded FIMA repurchase facility, and the accelerated long maturity buyback program, and on the subsequent move in yields. CNBC, August 19–21, 2026.
Market data referenced
Ten-year and thirty-year Treasury yields, August 2026. Board of Governors of the Federal Reserve System, H.15 Selected Interest Rates, via FRED series DGS10 and DGS30.
https://fred.stlouisfed.org/series/DGS10
Corporate bond issuance, 2026 year to date. Securities Industry and Financial Markets Association
(SIFMA), as reported by CNBC, “U.S. government debt yields are surging at a bad time,” August 18, 2026.
Federal debt crossing $40 trillion, August 2026. U.S. Department of the Treasury, as reported by CNBC, August 20, 2026.
Commercial and multifamily mortgage maturities of $875 billion in 2026 and $652 billion in 2027. Mortgage Bankers Association, Commercial Real Estate Loan Maturity Volumes survey.
Company loan terms are drawn from The Morgan Companies’ own term sheets: Reinsurance Group of America commercial loan term sheet dated February 10, 2022 (Church Street, Burlington, NC), and CIBC Bank USA summary of terms dated October 13, 2025 (Beaufort Station, Beaufort, SC).
Background reading not cited in text
Boyle, Patrick. “The Hidden Risk in the US-Japan Yen Rescue.” Patrick Boyle on Finance (video podcast), 2026. https://www.youtube.com/watch?v=yh18YXKMk3g — On the risk that Japanese intervention to support the yen forces sales from the largest foreign holder of US Treasuries. This discussion shaped the demand side argument above.
