JTBD RE Pulse Week 8/21/26
Supply outran the buyers.
Then it outran the borrowers.
The world bought more American debt this year and still fell behind, because Washington issued faster than anyone would absorb. That gap is what sets the mortgage rate now, and in Massachusetts it has already taken a rung off the bottom of the buyer ladder.
- Issuance outran foreign demand. Foreign holdings of Treasuries rose over the past twelve months and still lost ground as a share of the debt, because issuance grew twelve times faster than they did. Central banks alone cut $114 billion.
- The Fed's own committee is arguing the other way. Three members dissented on July 29 in favor of a hike, and the two-year note trades forty-four basis points above the top of the target range. Neither is a signal that mortgage rates fall from here.
- The inflation that matters to a building runs through fuel. Headline inflation is running almost a point above core, and every bit of that wedge is energy. Diesel is up 46.9% in a year, which arrives as a delivery surcharge on every material and appliance in a renovation.
- Massachusetts buyers thinned while prices held flat. Within every price band, average sale prices are flat. The statewide median moved only because six hundred and sixteen fewer households closed under $500,000, and FHA-financed sales fell four times faster than the market.
One chain, from the auction desk to the closing table
Issue 5 asked which buyers could clear a single rate. Two months on, the question has moved upstream. The rate itself is no longer a Fed decision in any practical sense, because the Treasury has to place roughly $9.7 trillion of maturing paper this fiscal year into a world whose appetite is growing at a fraction of that pace. When supply outruns demand in any market, price is the release valve, and in the Treasury market price means yield. Every link below is that one fact working its way downhill.
| The link | What is actually happening | Where it hands off |
|---|---|---|
| Issuance outruns absorption | Foreign holdings grew; the debt grew twelve times faster | The marginal buyer becomes domestic and private, and prices accordingly |
| Treasury crowds the short end | Bill issuance sits above the share its own advisory committee recommends | A third of the debt reprices within twelve months |
| The energy shock refuses to clear | Hormuz flows 4.9 mb/d against 21.6 before the conflict | Headline inflation stays above core and refuses to converge |
| The Fed loses the argument internally | Held 3.50 to 3.75% on a 9 to 3 vote, all three dissents to raise | Rate relief stops being a base case |
| Long rates hold their level | The ten-year and the thirty-year mortgage both sit above where they stood a year ago | The payment a buyer can carry stops improving |
| Massachusetts loses its entry buyer | 616 fewer closings under $500,000; FHA volume down 7.7% | Inventory builds at the bottom while prices stand still |
The buyers are still there and the paper keeps outrunning them
The world bought more Treasuries this year and lost ground doing it
The popular version of this story says foreign buyers have walked away, and the data does not support it. Treasury's own capital-flow tables for June, published on August 17, show total foreign holdings of $9.299 trillion, up $205 billion over twelve months and up almost exactly $1 trillion over twenty-four. The appetite is there; the ratio is what moved. Marketable debt outstanding grew $2.43 trillion over the same twelve months, so foreign holders financed roughly eight cents of every new dollar and their share of the whole fell from 31.7% to 29.9%. The gross national debt crossed $40 trillion on August 18, three days before this issue. The number that sets the mortgage rate is the $31.1 trillion of marketable debt inside it, because that is the part which has to find a buyer at a price.
Underneath that average sits the division that matters. Foreign official holdings, meaning central banks and sovereign funds, fell $114 billion over the year and are down to 12.2% of marketable debt from 14.2% two years ago. Private foreign holdings rose $1.05 trillion over the same two years. Governments are stepping back and money managers are stepping in, and money managers price for risk in a way that a central bank recycling a trade surplus never had to. That substitution is a large part of why the ten-year has held its level through a cutting cycle.
One correction worth carrying, because the wrong version circulates weekly. Japan remains the largest foreign holder at $1.117 trillion, the United Kingdom is second at $939.9 billion, and China is third at $633.4 billion. China has not been in second place for some time. And the monthly declines that get reported as sales are holdings measured at market value, which blends actual selling with the mark-to-market effect of a rising-yield month. Treasury's own flow table in the same release records net foreign purchases of long-term U.S. securities of all types of $207.1 billion in June.
Washington moved its borrowing to the part of the curve that reprices fastest
Facing that arithmetic, Treasury has done what any borrower does when the long end gets expensive, which is to borrow short and hope. Bills now make up 22.2% of marketable debt outstanding. The Treasury Borrowing Advisory Committee, the panel of dealers and investors Treasury itself convenes, put the sensible long-run share at around 20% with 15% as the floor for healthy market function. The government is financing itself above the long-run average its own advisors recommended, and every point of overshoot is debt that has to find a buyer again within the year.
The scale of that is easy to underestimate. The Government Accountability Office reports that Treasury refinanced $9.1 trillion of maturing debt in fiscal 2025 and must refinance $9.7 trillion in fiscal 2026, and that 33% of all outstanding debt now matures inside twelve months against 24% in 2014. A borrower with a third of its balance rolling annually has no insulation from the rate. It gets the market's opinion, monthly, in full. That is the mechanism by which a Fed cut can coexist with a mortgage rate that does not move.
The dollar's managers spent the end of July defending a foreign currency
On July 31 the United States joined Japan in buying yen. Japan's Ministry of Finance confirmed it in an official statement on August 3, describing an operation conducted "in coordination with the U.S. Department of the Treasury" against "excessive volatility and disorderly movements in the Japanese yen," and saying Tokyo is willing to intervene jointly again. Secretary Bessent confirmed American participation in a post the same morning, calling the operation a response to "disorderly yen movements." Neither government has disclosed a size, an instrument, or a counterparty currency.
Why a Treasury Secretary would care about the yen at all is the connection worth making. Japan holds $1.117 trillion of Treasuries. A finance ministry defending its currency with its own reserves is a finance ministry that may eventually need to sell some of that pile, and a sale of that scale into a market already absorbing that volume of refinancing would push the long end further. Helping Tokyo hold the yen is, in that light, a way of protecting the American bond market from a Japanese decision. The widely reported detail that the United States funded its side by selling euros comes from press reporting and from a letter Senator Elizabeth Warren sent Bessent on August 13, not from any official disclosure. It remains an allegation.
Two figures bound how large this could have been. The Exchange Stabilization Fund held $18.9 billion of foreign currency assets, in euro and yen, at June 30, its last statement before the operation, which caps anything Treasury could have funded from it. And the Federal Reserve's standing repo line for foreign central banks, the facility Tokyo said on August 3 it "plans to utilize" in future, stood at zero on the August 19 balance sheet and peaked at $3 billion all year. The pressure valve has been built, announced and left untouched, which tells you the stress is anticipated rather than present.
China holds a fertilizer reserve, India holds a subsidy, and the United States holds neither
The same conflict that emptied the oil chokepoint runs through the fertilizer trade, and it produced the clearest demonstration this year of how fast a chokepoint transmits. Urea, the benchmark nitrogen fertilizer, went from $415 a tonne in January to above $850 in April as Hormuz closed, then collapsed back to $400 by July after the June 17 memorandum reopened traffic. It doubled in three months and gave all of it back in three. That memorandum has since lapsed. It expired on August 18 without a permanent deal, and transits through the strait have fallen back to single digits, which puts the fertilizer chokepoint back where it was before June rather than behind us. Potash sits 29.5% below its 2022 peak and phosphate is up 1.4% on the year. The fragility is real and the price today is not elevated, and saying both is what separates this from the headline pile.
What survives the round trip is structural. China operates a codified state fertilizer reserve under measures issued jointly by its planning commission and finance ministry, holding physical nitrogen, phosphate and potash with release authorized centrally. India runs a different defense, holding 18.0 million tonnes of stock and fixing the retail price of urea by subsidy, a price it has held flat for years. The United States announced its own response on July 1, a $500 million program whose stated purpose is building domestic production capacity. Plants take years. Two of the three largest agricultural economies can release supply into a shock and the third can only start construction. American exposure runs mostly through potash, of which the country produces under 5% of its own need, and that supply comes through Canada, Russia and Belarus rather than the Gulf.
The reason this reads as a crisis in most coverage is worth stating plainly, because it is a case study in what this letter exists to correct. J.P. Morgan has published on it twice, and both times hedged. Global Research put out a ripple-effects piece on April 7 in which senior global economist Nora Szentivanyi projected that the fertilizer disruption "could lift global food inflation temporarily to 4-5%," with the impact arriving at a considerable lag. A climate advisory note from Dr. Sarah Kapnick followed on May 21, mapping a conditional risk and observing that over 36% of global urea is imported from the Persian Gulf and that restarting damaged capacity takes one to four years. Both use the word "could," and both put the effect out at a lag. What reached the reader by August was a dated food-crisis call routed through a third-party summary rather than through either original, with the condition and the lag stripped out of it. Meanwhile the FAO food price index sits at 131.1, up 1.0% on the year and 18% below its 2022 peak, and the USDA's August supply estimates show the second largest American corn harvest on record.
The argument inside the committee is about hikes
Three members voted to raise at the last meeting
The July 29 decision held the target range at 3.50 to 3.75%, which reads as a pause until the vote is examined. It was nine to three, and the statement records that Beth Hammack, Neel Kashkari and Lorie Logan "preferred to raise the target range for the federal funds rate by 1/4 percentage point at this meeting." Three of twelve voters wanted tightening in a summer that most commentary has spent forecasting cuts. The next meeting is September 15 and 16 and carries a full set of projections, which makes it the first date this autumn where the committee has to put numbers behind whichever of those two positions it holds.
The two-year note is trading above the policy rate the Fed just held
The two-year Treasury is the cleanest read on where the market thinks the funds rate averages over the next two years, and at the August 20 close it yielded 4.19%. The top of the current target range is 3.75%. The market is pricing the policy rate to average forty-four basis points above where the Fed just set it, which is a bet on tightening rather than easing. The two-year has added fifteen basis points over three months, though it has come down eighteen from its July 23 peak of 4.37%, so this is a firm signal rather than an accelerating one.
Set that against the inflation the committee is actually looking at. July CPI came in at 3.4% headline and 2.5% core, and the entire ninety-basis-point gap is energy, up 14.7% on the year. Shelter, the component that dominates core, rose only 0.1% on the month. A central bank facing 2.5% core would normally be comfortable. A central bank facing 3.4% headline, an oil chokepoint the EIA does not expect to clear until early 2027, and three of its own members voting to hike, is a central bank with very little room to deliver the relief the housing market has been waiting for.
The balance sheet turned back up at the end of 2025 and has been growing since
Quantitative tightening is over. Total Federal Reserve assets stood at $6.746 trillion on the August 19 balance sheet, up $132 billion over six months and $127 billion over the year, having troughed in late 2025. The July implementation note authorizes the New York desk to buy Treasury bills "to maintain an ample level of reserves." The direction is genuinely upward and the mechanism is reserve management rather than stimulus, and both halves of that sentence matter. This is the central bank keeping the plumbing full so the funds rate stays inside its range, and it is also, functionally, the domestic buyer stepping in where the foreign official buyer stepped out. The most recent single week was down $14 billion, so this is a trend, not a headline.
The six reads that carry the argument
| Indicator | Current read | Signal |
|---|---|---|
| 10-yr Treasury (Aug 20 close) | 4.69% (30-yr 5.23%) | Long end will not fall |
| Retail diesel (Aug 17 survey) | $5.454/gal, +46.9% y/y | Delivered into every cost |
| Retail gasoline (Aug 17 survey) | $4.049/gal, +29.6% y/y | Household squeeze |
| Strategic Petroleum Reserve | 293.4M bbl, down 110.0M y/y | Buffer spent |
| MA retail electricity (May) | 25.12¢/kWh vs 13.83¢ national | Structural operating cost |
| Real average hourly earnings | −0.2% y/y (weekly +0.1%) | Stalled, not falling |
New England is the exception the national electricity story keeps missing
Massachusetts pays 25.12 cents a kilowatt hour against a national average of 13.83, which is the number every operator here already knows. The part worth adding is the direction. National residential electricity is forecast up 5.6% this year on data-centre load, while New England is forecast up 0.2% over the same year. The region is expensive and, for now, the slowest-rising piece of the national picture, largely because power was already dear enough here that new load has gone elsewhere first. An underwriting model that imports the national escalation rate into a Massachusetts pro forma will overstate the next two years of operating cost.
Prices held still while the buyer pool thinned
Start with what a rate that refuses to fall actually does to a household. Illustrative: a buyer who can carry $3,000 a month in principal and interest supports a loan of about $467,000 at today's 6.65%. At 5.50%, the rate much of the market spent last year forecasting, the same payment supports about $528,000. The gap is roughly $61,000 of purchasing power, and it lands on the household that was already at the edge of qualifying rather than on the middle of the market. Everything in the Massachusetts data this year is that $61,000 showing up as an absence. Every Massachusetts figure in this section is a live MLS PIN pull as of August 20, 2026, the day before publication.
Inventory is up eighteen percent and it is building at the bottom
Against 6,066 on August 20 last year, a rise of 18.0%. Same calendar date, so no seasonal adjustment is doing any work here.
From 1,119 listings to 1,435. Above $1 million the rise is 7.5%, so the build is concentrated where the buyer left.
Down from $799,900 a year ago, which is a change in what is being listed rather than a cut in what anyone is asking.
The listings arriving to market this summer are cheaper than the listings that arrived last summer, and there are far more of them at the low end. That is the supply side of the same $61,000. Sellers at the bottom are still coming to market on last year's expectations and the buyer who would have absorbed them has been repriced out.
Six hundred and sixteen households stopped closing under half a million
Closed single-family sales for the year to August 20 came in at 23,244 against 23,645 in the same window last year, a decline of 1.7% that looks like nothing. The composition underneath it is not nothing. Sales under $500,000 fell from 6,106 to 5,490, down 10.1%, while sales above $1 million rose from 5,389 to 5,600, up 3.9%. The market did not shrink so much as it lost its bottom and grew at its top. Condominium sales tell the same story from a different angle, rising 1.1% in count while taking fifty days to sell against forty-three a year ago.
The FHA borrower is leaving four times faster than the market
Filtering the same closed sales by financing type gives the sharpest number in this issue. FHA-financed single-family sales fell from 1,968 to 1,817, down 7.7%, against a market down 1.7%. FHA's share of Massachusetts single-family closings dropped from 8.3% to 7.8%, and FHA deals now take forty-six days to sell against forty-one for the market as a whole. The low-down-payment buyer is the rung that came off.
That matters beyond Massachusetts because of what is happening inside the national FHA book. Serious delinquency on FHA loans has gone from 4.30% to 6.48% over the year, concentrated in the 2022 and 2023 origination vintages, where the 2023 cohort runs near 11%. And since October 1 last year, borrowers have been limited to one home-retention workout every twenty-four months rather than eighteen, a change HUD made because redefaults on the pandemic-era recovery options kept rising. That valve is now narrower, and FHA foreclosure starts have gone from 6,352 last June to 9,590 this June, a rise of 51%, while new ninety-day delinquencies sit about nine percent above where they were a year ago. Fewer borrowers are being cured, so more of them reach the first legal step. Massachusetts is well insulated by national standards, with FHA at 7.2% of originations against 13.3% nationally and foreclosure filings up 3.2% in the first half against 21% for the country. The exposure here is concentrated rather than absent. Hampden County, which contains Springfield, runs 14.9% FHA and Bristol County, which contains New Bedford, runs 13.1%, so the places most exposed carry a national-level share inside a state that carries half of one.
Prices did not move, and the median rose anyway
This is the finding most likely to be misreported elsewhere, so it is worth being exact. The statewide single-family median rose from $675,000 to $695,000, about 3%. Within price bands, average sale prices are flat to within a few tenths of a percent: the $500,000 to $600,000 band went from $547,074 to $546,628, the $600,000 to $700,000 band from $644,126 to $645,935, the million to $1.5 million band from $1,211,762 to $1,212,416. Massachusetts homes held their price this year, and the median moved because of who stopped buying.
The same arithmetic trap sits inside the Gateway Cities, and it is worth showing because it is the exact error a careful reader would otherwise catch. Taken together, Brockton, Lawrence, New Bedford and Springfield show a median down 2.5%. Taken separately, Springfield is up 2.7% and the other three are up 0.4%. The combined figure falls because Springfield, at a $305,000 median against $485,000 for the rest, grew its share of that group's sales from 45.7% to 50.7% while sales in the other three fell 7.8%. Prices rose in both groups. The cheapest city in the cohort absorbed the transactions the others lost. Buyers are migrating down the price ladder rather than off it, and the ladder itself is what is getting shorter.
Four dates that settle the open questions
Each of these resolves a specific claim in this issue rather than adding a new one. Every one of them lands inside the next four weeks.
Japan's finance ministry publishes intervention amounts for the window containing July 31 around the end of this month. It is the first official number attached to an operation both governments have confirmed and neither has sized. A large figure makes the Treasury holdings question urgent; a token one makes it symbolic.
Beijing's suspension of phosphate fertilizer exports, announced last December on stated food-security grounds, runs through August. Whether it is renewed is the only genuinely forward-looking variable in the fertilizer story, and it is decided within days.
The test of whether the energy wedge is closing or widening. Watch the gap between headline and core rather than either number alone. A narrowing gap means the shock is passing through; a widening one means it is still arriving.
The first meeting where the committee has to put a rate path on paper since three members voted to hike. The dot plot, not the statement, is where the July dissent either spreads or gets absorbed.
What would break the argument in this issue: a durable replacement for the memorandum that lapsed on August 18, one that reopens the Strait for long enough to pull energy out of the headline number, a September projection showing the committee's centre moving toward cuts rather than toward the dissenters, or a Massachusetts autumn in which sub-$500,000 inventory clears rather than builds. Any of the three would mean the rung is being replaced rather than removed. The EIA's own forecast, which has Brent averaging $85 this quarter and falling to $69 next year, is the most credible case for that first off-ramp, and it is worth taking seriously rather than dismissing.
The bond market gets the last word on the mortgage rate, and it has not finished speaking.
NEXT ISSUE · SEPTEMBER 2026
What changes in the next deal
Four adjustments follow from this month's data, and each one is a consequence of the same chain rather than a separate idea.
Underwrite the rate that exists rather than the one in the forecast
A refinance assumption is the most expensive line in most Massachusetts pro formas right now, because it is usually the one that goes untested. The market is pricing the policy rate to average above where the Fed has set it, three sitting members voted to tighten, and the energy shock behind the headline number has an official normalization date in early 2027. A model that requires a rate cut to work is a model with a forecast in its exit assumptions and nothing in its downside case. Run the deal at today's rate held flat for the full hold period, and if it only works with relief, it is a rate bet wearing a building.
The value is under five hundred thousand, and so is the risk
Inventory under $500,000 grew 28.2% in a year while inventory above a million grew 7.5%, and sales in that same bottom band fell 10.1%. Supply is building precisely where demand thinned. For a buyer with cash or with financing that does not depend on the marginal borrower, that is leverage on price and on terms in a segment that spent five years offering neither. The risk sits alongside it: the reason the segment softened is that its buyer pool shrank, so an exit strategy that assumes a retail sale to the same FHA borrower who has been disappearing is the wrong exit. Buy where the buyer left only if the plan is to rent it or to sell it to someone who does not need the same loan.
Fuel and power belong in the operating model as separate lines
Diesel at $5.454 a gallon, up 46.9% in twelve months, reaches most owners as a surcharge on appliance delivery, on lumber and on every trade truck that quotes a job, rather than as a fuel bill, and it is one reason renovation budgets set twelve months ago no longer hold. Electricity is the opposite case and deserves the opposite treatment. Massachusetts pays roughly eighty percent above the national rate per kilowatt hour, and New England residential prices are forecast up 0.2% this year while the national average is forecast up 5.6%. Escalate the fuel-sensitive lines aggressively and the power line conservatively, because the national trend runs the wrong way for this region on one and the right way on the other.
Days on market is the number that actually moved this year
Prices held flat within every band, so the price data will tell an owner almost nothing about which way this market is heading. Time on market moved in every band of every product type, and it moved most in multi-family, where it widened from thirty-five days to forty-three, roughly twice the widening in single family once price mix is controlled for. Condominiums remain the slowest product in absolute terms in every band. Track marketing time rather than comparable sale price, because in a market where the buyer pool is thinning, duration reprices months before the comparables do. For a seller, that means pricing to the first three weeks. For a buyer, it means the listing sitting past day forty is the negotiation.
Price told you nothing this year. Duration told you everything.
JTBD RE PULSE · VOL. 1 ISSUE 6
