The National Association of Realtors reported on Tuesday that its Pending Home Sales Index fell 7.3 percent in September to 66.0. That is the lowest reading since the series began in 2001, down 11.3 percent from a year earlier. Economists had expected a decline of about half a percent. The miss was not close.
A pending sale is a contract signed, not a closing. The index leads completed sales by a month or two, which is why it gets attention out of proportion to its sample. A one-month drop can be noise. A record low after a year of high rates is a signal about how the market is clearing, and the answer is that it is not clearing through price. It is clearing through volume.
That distinction is the useful one. The headline says demand collapsed. The regional detail and the rate data say something narrower: buyers who need a mortgage cannot make the payment at 7.4 percent, and sellers who already have a cheap mortgage do not need to sell. Both sides are waiting. The transactions that disappear are the ones that would have required one of them to move first.
The number
The index stood at 66.0 in September, down from 71.2 in August. The year-over-year decline was 11.3 percent. The August reading had risen 0.3 percent, which makes the September drop look like a reversal rather than a continuation. In fact it is both: the market had a small summer bounce as rates dipped, and the bounce ended when rates turned back up.
Every region fell. The Northeast dropped 4.2 percent to 59.0 and was down 9.4 percent from a year earlier. The Midwest fell 6.5 percent to 67.3 and was down 8.7 percent. The South fell 8.2 percent to 79.2 and was down 12.0 percent. The West fell 9.2 percent to 49.3, the weakest reading of the four, and was down 15.0 percent year over year.
The West is the tell. It has the highest prices, the longest commutes and the most rate-sensitive buyers. When the national index falls this hard, the West usually falls harder, and it did. The South's decline is also notable because it has been the country's fastest-growing housing region. A record low there is not a coastal story.
The index is constructed from contract signings, and its sample is smaller than the completed-sales series that gets the headlines. That makes the monthly reading noisy and the year-over-year comparison more reliable. On that basis, the September figure is not a one-month fluke. It is the continuation of a decline that has now lasted more than two years, punctuated by brief rallies whenever rates fell enough to pull a few buyers off the sidelines. The pattern is a market that responds to rate moves quickly and then gives the gains back just as quickly when rates turn.
What the rate did
The 30-year fixed mortgage rate averaged 7.40 percent in Freddie Mac's survey for the week of October 15, the second consecutive week at that level. In late August the same survey put the rate at 6.66 percent. That is a move of about 74 basis points in seven weeks, which is large for a market where a quarter-point change can alter a monthly payment by enough to disqualify a buyer.
The payment math is unforgiving. On a $400,000 loan, the difference between 6.66 percent and 7.40 percent is roughly $190 a month, or about $2,300 a year, before taxes and insurance. For a buyer already near the limit of what a lender will approve, that is not a preference. It is a different house, or no house.
The rate move has a cause, and the cause matters for what happens next. Long-term yields rose as oil prices and inflation expectations climbed, and the mortgage market followed. That is not a story about the Federal Reserve cutting or holding. It is a story about the long end of the curve, which the Fed does not control. A buyer waiting for a Fed cut may be waiting for the wrong event.
The spread between the 10-year Treasury yield and the 30-year mortgage rate is the other half of the story. Mortgage rates are set by the bond market, not by the Fed's policy rate, and the spread over Treasuries compensates lenders for prepayment risk and credit risk. When that spread widens, mortgage rates can stay elevated even if Treasury yields fall. It is one reason a buyer can watch the Fed hold or cut and still find that the rate on the loan has not moved in their favor.
Why the drop is worse than it looks
The consensus forecast was for a decline of about half a percent. The actual decline was 7.3 percent. That gap is unusual, and it suggests that the summer bounce was more fragile than the monthly data showed. A market that loses 7 percent of its contracts in one month is not adjusting gradually. It is repricing.
NAR's chief economist, Lawrence Yun, attributed the weakness to fast-rising mortgage rates and said buyers and sellers may be holding back until rates move favorably. That is a polite way of describing a standoff. Buyers are waiting for affordability. Sellers are waiting for the market to prove that the summer was not the top. Neither side has an incentive to move first when the cost of waiting is mainly time.
The housing starts data released the same week reinforces the point. Starts fell 5.2 percent to a seasonally adjusted annual rate of 1.221 million. Builder sentiment, measured by the NAHB/Wells Fargo index, fell to 29, its lowest since June 2012. Builders are not immune to the same rate math. They respond by building fewer spec homes and offering more concessions, not by cutting list prices across the board.
The supply side is locked too
The standard explanation for a sales collapse is that buyers disappeared. The fuller explanation is that sellers disappeared too, and the sellers who remain are the ones with the least reason to cut. A homeowner with a 3 percent mortgage and a monthly payment that fits has no financial reason to sell into a 7.4 percent market. Moving would mean trading a cheap loan for an expensive one. Many homeowners would rather stay put, renovate, or rent out the old house and buy the next one only if they can carry both.
That behavior is the lock-in effect, and it is why prices have not fallen as fast as sales. The inventory that comes to market is skewed toward sellers who have to sell: estates, divorces, job relocations, distressed owners. Those sellers are price-sensitive, but they are also a small share of the market. The rest of the housing stock is held by people who are not listing.
The result is a market with fewer transactions and stickier prices. That is not a contradiction. It is what happens when the cost of moving rises for both sides at once. The volume adjusts, the price does not, and the adjustment shows up in time on market, concessions and the mix of homes that sell. A record-low pending index is consistent with stable prices if the supply of homes for sale is falling at the same time as demand.
What would unfreeze the market
There are three ways out. The first is a sustained fall in mortgage rates, enough to bring buyers back without forcing sellers to capitulate. The second is income growth that makes the same payment affordable at a higher rate. The third is a shock that forces more sellers into the market, which would raise inventory and put pressure on prices.
The first is the one the market is waiting for, and it is the one with the most ambiguous effect. A small decline in rates could bring buyers back before sellers, tightening inventory further and pushing prices up. A large decline could bring back sellers who have been waiting for a better market, increasing inventory and softening prices. The difference between a half-point and a point and a half is the difference between a squeeze and a thaw.
The industry's own forecast is not optimistic. Lisa Sturtevant, chief economist at Bright MLS, has projected that existing-home sales will finish 2026 below 4 million, which would leave the year among the weakest for resale volume in decades. That projection implies the freeze lasts through the year, not just through the autumn. It also implies that the adjustment will keep showing up in volume and in the mix of homes that sell, rather than in a headline price decline that buyers are waiting for.
The second is slower. Wages are rising, but not fast enough to offset a 74-basis-point move in a few weeks. The third is the one nobody wants to root for. A rise in unemployment or a wave of forced sales would raise inventory, but it would do so by inflicting the kind of stress that housing policy is supposed to prevent.
The number to watch next
The next pending-sales release is scheduled for November 18. By then the market will have had another month of rate data, another round of builder earnings and another reading on existing-home sales. The index is volatile, and one month does not make a trend. But the September reading is low enough that the burden of proof has shifted. The question is no longer whether high rates have hurt the housing market. It is whether the market can recover without a fall in rates, and the evidence so far says it cannot.
For buyers, the practical implication is that waiting has a cost and a benefit. The benefit is a possible lower rate. The cost is competition from other buyers if rates fall and the sellers who remain stay put. For sellers, the implication is that the market is thinner than the headline price suggests. The buyers who remain are the ones who can qualify at 7.4 percent, and they are not paying for the same house they would have bought at 6.6 percent.
The record low is a warning about volume, not a forecast of a crash. A market can freeze for a long time without collapsing. The longer it stays frozen, the more the adjustment happens through who can move rather than what things cost. That is a harder market to read, and a harder one to fix.