JTBD RE Pulse Week 7/31/26
Inflation cooled in June. Your mortgage rate went up anyway.
The buyer of the bond changed, so the high rate is here to stay, and in Massachusetts it holds prices up while the market slows underneath them.
The reason that happened matters for your next deal. The people who buy US government debt have changed. The foreign governments that used to hold it no matter what are stepping back, and the leveraged traders who took their place only lend if you pay them more, all while the government floods the same market with a record wall of new borrowing. So the rate will not drop just because inflation does. The money pouring into artificial intelligence works on that same rate from both ends. Its borrowing piles onto the shrunken pool of buyers and pushes the rate up further, and the jobs it cuts decide who can still clear that rate. It is also the piece most likely to snap. Massachusetts is where the durable rate and the widening split both land, and the market here is slowing underneath prices that still hold.
- The buyer of the bond changed. The foreign governments that held US debt no matter the price have been replaced by traders who only hold it if they get paid more, so the rate will not fall just because inflation cools.
- AI works on that same rate from two sides. The money building it is borrowing so heavily that a Federal Reserve bank now measures that borrowing against the government's own, and the technology is cutting the entry-level jobs that feed the rental market.
- The market slows instead of clearing. Most owners are locked into cheap mortgages, so at a rate in the 6s they will not sell, and prices hold while sales slow.
- Massachusetts feels it first. The high-paying office jobs are shrinking here while coastal rents hold the line, so the question on every building is which side of that gap its tenants are on.
One rate, and who can clear it
One thing sets up your next deal, and it bites twice. The buyer of US government debt changed, which sets a floor under your mortgage the Fed cannot lower. The money rushing into artificial intelligence works on that same floor from both ends, adding to the borrowing that holds the rate up and splitting the job market that decides who can meet it. Read the chain from top to bottom to see where it lands on a Massachusetts closing table.
| The link | What is happening | Where it lands |
|---|---|---|
| The bond buyer changed | Leveraged traders replaced foreign governments, into a record wall of new supply | A rate the Fed cannot lower by cutting |
| AI adds to it | The same boom borrows heavily and cuts entry-level jobs | It lifts the rate and splits the economy |
| Inflation cooled, the rate did not | Prices eased in June while the ten-year rose | Cooling inflation alone will not lower the rate |
| The market slows | Owners on cheap mortgages will not sell into a rate in the 6s | It reprices without clearing |
| The economy splits | Pay, rent, stocks, and spending pull apart, top from bottom | Which side a tenant is on decides if you transact |
| Massachusetts lands it | Office jobs shrink while coastal rents hold | Prices hold, the market slows, the bottom thins |
Who is actually buying the bond
The buyer who held no matter what is backing away
Foreign governments used to buy US debt for reasons that had nothing to do with the return, which made them the steadiest hand in the market. They are backing away. Foreign investors hold about a third of the Treasury market, down from nearly half in 2008. China is down to about $659 billion, close to an eighteen-year low, and Japan, still the biggest foreign holder, sold roughly $66.7 billion of Treasuries in a single month this spring to prop up a sinking yen, its biggest sale in more than three years. What replaced them is not another government. It is the hedge funds registered in the Cayman Islands that borrow money to trade bonds. A Federal Reserve study finds those funds held about $1.8 trillion of Treasuries by the end of 2024 and absorbed roughly $1.2 trillion of the government’s net new borrowing since 2022. The holder that stayed no matter what is leaving, and the one that took its place only stays if it gets paid more. That extra pay is the return climbing on your mortgage.
Why the rate has a floor, and how AI helps hold it up
The extra return lenders now demand just to hold a long bond, which had sat near zero for years, is positive again, about 0.78%. You can see it in long-term bonds, where the 30-year Treasury has held above 5% for the longest stretch since 2007. It is why the ten-year rate climbed in July even as inflation fell. Two waves of new borrowing are hitting a shrinking pool of buyers at the same time. The government has to refinance something like eight trillion dollars of old debt over the next year, close to a record, much of it borrowed cheap and now coming due at today's higher rates. And the money building AI data centers is borrowing so heavily that the Dallas Fed puts it at about an eighth of the government's own long-term bond issuance, which Apollo economist Torsten Sløk says is likely to push rates up further. The same money building AI is borrowing at a scale a Federal Reserve bank measures against the government itself, so the AI boom that is cutting jobs in Section 4 is also, through its borrowing, part of what keeps your mortgage rate high.
The market has just started to doubt that boom, which is the honest other side of it. In late July the AI-chip trade cracked. The Nasdaq slid toward a correction over fears about AI spending, after a Chinese memory-chip maker called CXMT surged 466% on its Shanghai debut and threatened the fat profit margins the sector was built on, with Intel already down about 21% earlier in the month. If that selloff keeps going it cuts both ways. Less AI borrowing would take some pressure off rates, but a fast unwind in the AI trade is its own danger, the same weakness sitting under the borrowed-up bond traders from 1.1. The boom that is helping hold your rate up is also the thing most likely to break it.
The official world is quietly leaving the dollar too
The private buyer changing has a public twin. In the World Gold Council's June survey, a record 45% of central banks said they plan to buy more gold this year, and 74% expect the dollar's share of the world's reserves to keep shrinking over the next five years. That is the same message as the changing bond buyer, just from the government side: the world's reserve managers are moving out of US debt at exactly the moment the government needs them most.
Why a rate cut cannot reach your mortgage
The Fed can lower the short rate. It cannot lower your mortgage.
June inflation cooled, which in a normal year would give the Fed room to cut. But your mortgage is priced off the ten-year Treasury, and the ten-year answers to who is buying the bond, not to the Fed's short-term rate. So even a cut leaves the ten-year, and the mortgage sitting on top of it, about where they are. A Fed cut lowers what your savings account earns. It does not lower the rate on your next purchase loan.
What the Fed did on July 29, and the tell in it
The Fed held the rate at 3.50 to 3.75% on July 29, but the vote was 9 to 3, the first time three members dissented since 2016, and all three wanted to raise, not cut. Chair Kevin Warsh would not call it a pause, and he made clear the committee is still fighting over above-target inflation that a renewed Middle East conflict and a fresh jump in oil have made harder. The market's response is the part that matters to you. Even though the Fed did not move, the ten-year rose on the day and the 30-year climbed to about 5.21%, its highest since 2007, while the two-year, the rate that actually follows the Fed, fell after Warsh spoke. The long end went up without the Fed touching it. Warsh welcomed exactly that, saying investors are "learning to play the ball, not the referee." The chair of the Federal Reserve just told you the market sets the long end. Your mortgage answers to the buyer of the bond, and the Fed is fine with it.
The one honest way your mortgage eases
There is a way your mortgage comes down even if the ten-year does not, and it is the real exit. The gap between the 30-year mortgage and the ten-year Treasury is wide right now, about 1.9 points at the July 24 reading, because the Fed stopped buying mortgage bonds after years of doing so. If that gap narrows, the mortgage could drift into the low 6s while the ten-year stays put. Fannie Mae and Freddie Mac began buying agency mortgage bonds in 2026 under a federal directive meant to narrow it, though the effect is contested and the rising ten-year has swamped it so far. This is the one path where mortgages fall without the Fed first beating inflation, and it is worth watching more closely than the short rate.
The same split, in four places
Read these four together as one point. The economy is splitting into a top and a bottom, and that same split decides who can still clear a mortgage in the 6s and who can cover your rent. It is the one divide showing up in four places.
| Where the split shows up | The read | Signal |
|---|---|---|
| Stocks | The biggest 7 tech names are near 32% of the S&P 500 (top 10 at a record share, close to 40%) | Top of the split |
| Spending | The top 20% of earners drive about 60% of all spending | Tied to stock prices |
| Rent | Top-tier (Class A) rents still rising while bottom-tier (Class C) keeps falling, a split several quarters deep | The rent split |
| Car loans | Lower-credit borrowers 60 days late about 6.8%, near an all-time high; top-credit under 0.5% | Two-tier borrower |
Where the split reaches the closing table
Bring the whole thing down to one deal. Here is an illustration, not a live quote. At 6.58%, a buyer putting 20% down on the typical Massachusetts multifamily at $755,000 borrows about $604,000 and pays roughly $3,850 a month in principal and interest. At 3%, where rates sat in 2021, that same loan cost about $2,546. That $1,300 a month is the higher rate made personal, and because owners who locked in the low rate will not give it up, the market does not clear. It slows.
A market that reprices without selling
Across the country, half of all mortgages carry a rate under 4% and two-thirds are under 5%, against a market rate in the 6s. So most owners have every reason to stay put. The result is a market that reprices without selling. The national median price hit a record $440,600 in June even as the number of homes sold ran near one of the slowest paces in decades, even as June sales ticked up about 3% from a year earlier. Massachusetts shows the same pattern in the live MLS numbers. Multifamily still sells and the median held, but homes took 24 days to go under agreement this year against 21 last year, they took 13 days to draw an offer against 10, sellers got about 99 cents on the dollar of asking against roughly 100 last year, and only about a dozen multifamily listings statewide are lined up in Coming Soon to hit the market. In Boston, multifamily is selling at a $1.2 million median, right at asking, in about 26 days. Prices hold while the market slows underneath them, which is what a high rate does: it dries up the sales without cutting the price.
The high-paying jobs are shrinking here first
Massachusetts feels the job side of this first, because its economy runs on exactly the kind of white-collar work AI is squeezing. Tech and information jobs in the state are down about 10% from their recent peak, professional and business jobs off about 22,400 from their 2022 peak, about 3.5%, and finance down about 3% from its 2023 high. Those losses land in exactly the high-paying tiers, even when a given month's hiring looks broad. The headline number hides it. Massachusetts unemployment is 4.4%, still above the national jobless rate of 4.2%, because a falling overall rate can sit right on top of the high-paying tier thinning out. Nationally the top is the safe side of the split, but Massachusetts inverts that: the white-collar top is exactly what AI squeezes, so the people who pay premium rents and buy three-deckers are the ones losing ground here, and they are losing it first.
Rents are high, but the scarcity is ahead, not here
Boston is still one of the most expensive rental markets in the country, fourth in the nation, with a typical rent around $3,210 and roughly $128,000 of income needed to carry it without stretching, so only about half the listings in town sit within a middle-income household's reach. But rent growth has stalled to near zero, and landlords are starting to offer concessions, because Greater Boston just absorbed a five-year high of new apartments. About 9,300 units were delivered in 2025, with another 10,000 or more still under construction, and that supply has pushed vacancy up toward 7%. The scarcity story is real, but it is ahead of you, not behind you. Statewide permits just fell to their lowest since 2012, so the new-supply pipeline thins after this wave clears, and quality near jobs, where a submarket like the Seaport already runs near 3.7% vacancy, is where pricing power comes back first.
Two things that moved since last issue
Two things changed that a Massachusetts owner should note. The rent-control question I told you to plan for both ways is off the table. The state's high court struck it from the November ballot on June 23 because the petition included a religious exemption, which the state constitution bars from ballot initiatives, so the question is gone for now but can come back in a cleaner form. And the lender behind most small multifamily is quietly shakier than it looks. The Federal Reserve finds that fast-growing small and regional banks now originate close to 40% of new commercial real estate loans while holding only about 35% of the assets among the banks it studied, and their lending dries up fast if local deposits pull out. The rent-control question went your way this cycle, but the local bank that finances your triple-decker is the pressure point to watch: if its deposits flee, its lending dries up, refinances fail, and forced sales are the one thing that finally cracks the price floor this whole issue rests on.
What confirms this, and what breaks it
A view you can stand behind names what would prove it wrong. Here are the signals over the next few weeks that test the core claim: even if inflation keeps cooling, the changed buyer keeps the rate up.
The Fed held 9 to 3, with three members pushing to hike, the most dissents since 2016. There is no meeting in August, so September is the next decision, and two more inflation prints plus the renewed Middle East oil shock will settle whether a hike lands.
The late-July drop in AI chips is the other side of the story. If it deepens, less AI borrowing could ease the rate, but a fast unwind is its own risk. Watch whether AI spending holds or cracks.
The traders who now set the price of Treasuries at the margin are the weak point under the whole rate. Any stress in their funding, or a forced unwind, would move rates fast, either way.
The one clean exit. If the gap between the mortgage and the ten-year narrows as Fannie and Freddie buy agency mortgage bonds, your rate could ease into the low 6s even with the ten-year stuck.
We will carry each of these into Issue 6 in August.
If you are tracking a deal, a refinance, or a development timeline, these are the points that move your math.
Buy the sort, not the cycle
If the split, not the up-or-down cycle, is what decides a deal, the old playbook that waited for prices to lift every deal has to be rebuilt around which side of that split a building is on. Here is what changes in your next deal.
Underwrite a rate a Fed cut will not lower
Take the Fed rescue out of your model. Your mortgage sits on the ten-year, and the ten-year sits on a buyer problem, so a rate cut lowers what your savings earn, not what your loan costs. If a deal only works at 5.5%, or on a refinance you are counting on rather than underwriting, you are buying a bet the bond market is pricing against. Underwrite the rate in the 6s as the floor, and treat the mortgage-gap exit as upside, and underwrite as if it never comes.
Favor quality near jobs, where supply is tight
The pay gap decides who can cover your rent, and supply decides whether you can raise it. Favor good units near jobs and transit for the supply, not the demand: that tenant base thins first here, but new apartments are not landing in those submarkets, and with permits at a decade-plus low the pipeline behind them is drying up. Little new product competes with you once the current wave clears, so rent growth holds up there even as the high-paying job base shrinks.
Give yourself more room on the soft side
Plain, lower-end product is on the side that softens. Bottom-tier rents are already falling, the lower-end tenant is the one whose credit is fraying, and value-add plays that lean on pushing those rents need more room for vacancy and slower rent growth than the last few years taught. The discount you demand to buy on that side should be bigger than it used to be.
Watch the weak point, not just the level
The stuck rate is being held up by borrowed-up bond traders and an AI borrowing boom that just started to wobble. The risk now is speed as much as level. If either one unwinds, rates move fast. Keep cash ready and your financing flexible, because the same thing that could snap the rate down could also spike it, and the owner who can move when the AI trade or the bond trade cracks is the one who buys when others are forced to sell instead of getting caught in it.
Stop asking if the market is up or down. Ask which side of the split it is on.
The deals built for the sort are the ones still standing when the rate moves.
