Housing is the only market where we insist demand exceeds supply even as builders stack unsold inventory to 2009-era highs. What if the shortage story is true at yesterday’s prices and false at today’s? Demand is not a moral claim or a talking point. It is a function of price, income, and time. Shift the rate regime, and the demand curve moves. Build more high-spec houses for a shrinking pool of solvent buyers, and you do not solve scarcity. You reveal it.
Consider the paradox. Unsold homes climbed to roughly 121,000 in July, the most since the last housing crash. Pending home sales fell in June compared with a year earlier. Yet the median U.S. sale price hit a record near $400,000, still inching up. That is not healthy shortage. That is an affordability choke point. In markets like the Bay Area, years of job growth outpaced permits by a wide margin, which did create real scarcity. But the inventory spike today is not in the core job hubs. It is in places where new supply can get approved and built quickly, often far from the highest demand. Shortage in one location and price-insensitive supply in another can coexist. The average looks fine. Households do not live in the average.
Existing homeowners locked into 3 percent mortgages have little incentive to sell, so builders stepped in. Single-family starts were well above 2019 levels this spring, and new construction’s share of sales has risen from the mid-teens to over a fifth. Builders can buy down rates and dangle incentives. That pulls forward demand, but it does not expand it. You get a market split by finance: a frozen resale channel and a new-build channel juiced by incentives. The buyer pool at the price and monthly payment on offer is finite. Every rate tick that raises monthly cost shrinks it further. When incentives fade, absorption does too. The narrative says we underbuilt for a decade; the balance sheet says we are overbuilding for today’s purchasing power.
Housing looks slow and steady until it does not. Unsold homes are not just a number. They are carrying costs, covenants, and cancellation risk. Developers underwrite on assumed absorption, turn times, and stable financing. When pending sales stall and interest costs reset higher, inventory becomes a margin call in slow motion. We have seen versions of this movie. In India, a build-to-price-upmarket cycle left towers with lights off and balance sheets strained. In Canada, investor buying crowded out end users and made affordability worse, amplifying the eventual correction risk. In the U.S., the rise in unsold new homes back to crisis-era levels signals the same fragility. The longer units sit, the more aggressive the concessions, the more comps bend down. Price discovery wants to happen. Developers, lenders, and local tax bases want to deny it.
Each builder faces a Prisoner’s Dilemma. If everyone slows starts, margins hold. If one keeps building, they grab the scarce, rate-qualified buyer. The dominant short-term strategy is to build and push incentives. The group outcome is oversupply in the price bands with the least elasticity. Add zoning that blocks entry-level density in the job-rich core, and what gets built is larger, farther, and pricier than median incomes will support at current rates. This is why you can have both a shortage and a glut. There is a shortage of units at the payment a median household can carry within a tolerable commute. There is a glut of units priced for 2021 mortgage math, in locations optimized for permitting rather than utility. The spreadsheet approves; the household budget does not.
Households anchor to the last absurd. Sellers fixate on 2021 comps. Buyers anchor on a monthly payment that felt normal when the risk-free rate was near zero. Builders frame affordability around teaser rates and closing credits. These anchors are not laws of nature. They are leftovers from an era of suppressed volatility. When rates reset higher, the old price map is not a guide. It is a trap. In probability terms, the left tail matters again. A 1 percent drop in prices does not clear a market where the payment jumped 30 to 40 percent. Negotiation cannot overcome arithmetic. The system is fragile because it depends on keeping everyone convinced yesterday’s parameters will return any quarter now. That is not a base case. That is a bet.
Rate buydowns, down-payment assistance, and tax credits make for good press releases. They move contracts. They do not change the structural math of income, rates, and land. Build-to-rent is touted as a release valve. It can be, until rents flatten or financing costs outpace cap rates. Then it becomes another forced seller channel. Banks already carry meaningful exposure to construction and land loans, and municipalities depend on fees from new development. When absorption slows, the shock transmits: from developers to trade contractors, from local budgets to regional lenders. This is how a system looks stable and then breaks in clusters. Like a bridge that holds until the wrong combination of load and wind exposes a hidden crack, housing bears silent stresses until one more turn of the rate wrench snaps a joint.
Antifragile systems get better with volatility. Housing does not, because we designed it to be high leverage, long duration, and slow to adapt. An antifragile approach would do the opposite. Shorten cycle times. Favor modular and incremental builds that scale up or down as absorption changes. Permit more small, by-right infill where jobs already exist, so supply matches where demand is real. Stop subsidizing price and start deregulating production friction that blocks low-cost formats. Align product with local median incomes and transport realities. That is not as exciting as a megaproject ribbon-cutting, but it clears markets at actual prices rather than aspirational ones. It also disciplines balance sheets. If a developer cannot pencil a duplex on a modest lot without exotic finance, the problem is not just rates. It is that the unit is priced for the wrong era.
The headlines still talk about shortage. The data show a market trying to reprice while everyone resists. Unsold inventory climbing, pending sales slipping, prices inching higher only by selling to fewer, richer buyers with subsidies attached. That is not equilibrium. That is a dam under pressure. You can either open spillways to where demand actually lives, or wait and hope the wall holds. Builders are rational within their incentives, which is why policy and capital need to change the game, not the narrative. Build more where it matters and less where the spreadsheet tricks you. Price to the payment people can make without gimmicks. And accept that a decade of cheap money pulled purchases forward and fattened expectations. When the cost of capital rises, the market does not break. It reorders. The fragile parts are what we built on top of a story that stopped being true.
