The Market Prices Evidence, Not Ambition

A durable hope, in every market, is that time or momentum will lift an ambitious price into range. We tested that hope against the full record of Arizona’s $2 million-plus market — 37,688 resolved listings from 2008 to 2026 — and it does not hold. The market is generous to sellers in exactly one circumstance, and it is not the one most expect.

One · The fail factory is structural, not seasonal

Of 37,688 resolved $2 million-plus listings, 56% never sold — cancelled or expired without a transaction. That is not a bad-market number; it is the base rate. And it barely moves with the market’s temperature. Through the stable years of 2012–2019 the failure rate ran about 62%; it fell to 32% only in the extraordinary frenzy of 2020–2021, and rose to 80% only in the depth of the 2008–2010 crash. Between those rare extremes — which is to say almost always — roughly three of every five luxury listings failed regardless of whether the market was called hot or cold. Failure in this tier is the normal outcome, not the exception, and the market’s mood does little to change it.

Two · Even the winners are corrected

Look only at the homes that did sell — 16,488 of them — and the original asking price still rarely survived contact with the market. The median sale came in 6.7% below the original ask. Only 30% of successful sellers got within 3% of their opening number; 36% sold 10% or more below their original ask. So even success usually meant a correction. The opening price, across the tier, is an aspiration the market negotiates down — and that is among the sellers who got the good outcome.

Three · Raising an unproven price fails again

Some homes fail, leave the market, and return later at a higher asking price than the one that already didn’t work — often reasoning that conditions have since improved. We identified 3,498 such cases across the record: a listing that failed, then relisted more than 2% above its own failed ask. Sixty percent failed again. Raising a number the market had already declined, on a home no buyer had agreed to, did not overcome the market. In the majority of cases it simply repeated the outcome — sometimes across several cycles and several years, the same home climbing in price and failing each time.

Four · Raising a proven price works

Now the mirror image, and the finding that gives the study its name. Consider homes that actually sold, then resold within three years — 1,615 such pairs. Here the price also rose. But 81% resold higher, at a median gain of 13%. The distinction between this group and the last is the entire point:

The seller who raises the ask on an unsold home asserts a value no buyer has agreed to. The seller who reprices a home that has already sold once builds on a value a buyer has already paid. One number is anchored to a completed transaction; the other to hope. The market treats them as opposites, and it is right to.

The effect is not an artifact of renovation between sales. Homes that resold in under six months — too fast for meaningful rebuilding — still rose 78% of the time, and the appreciation curve barely steepens over longer holds, as it would if construction were driving it. Renovation cannot be fully observed from listing data and we do not claim to observe it; the sub-six-month cohort is offered as the cleanest evidence that the effect is price, not construction.

And the two things that do not work

Two common responses to a home that will not sell both fare poorly, and the data is blunt about it. Raising the ask fails 60% of the time, as above. Simply waiting is no better: of 4,072 listings that sat beyond 400 days, only 27% ever sold at all — and among the minority that did, those whose final asking price sat below their original outsold those who never moved off it. Patience, in the record, is not a pricing strategy. It is usually a slower version of the same mistake.

What this means — and what it does not

It does not mean price low. It means price to the evidence. The market appears to punish not high prices but unvalidated ones — numbers no completed transaction supports. A defensible price, anchored to real comparables, is the closest thing eighteen years of data offers to a reliable outcome, in any market. An ambitious price, anchored to nothing, is close to a coin flip against the seller — and no market on record, hot or cold, reliably rescued it. The single most useful question a seller can ask of a number is not “is the market strong enough to reach it?” but “has a real buyer ever paid something that supports it?”

Method

Universe. ARMLS residential listings with an original list price of $2,000,000–$60,000,000, 2008–2026 (a $1,000,000 floor was used only when linking resale pairs, to capture homes whose first sale preceded the tier). 37,688 resolved listings; 16,488 closed sales; 1,615 same-parcel resale pairs within three years; 3,498 serial-relist cases; 4,072 listings exceeding 400 days.

Definitions. A listing failed if it ended Cancelled or Expired with no recorded sale, and sold if it closed. A relist is “higher” if its original list price exceeded the prior failed listing’s by more than 2% on the same parcel. Discount is measured from original list price to sold price. Resale gain is measured between recorded sold prices on the same parcel.

Price data. Each listing record carries three price points — the original list price, the final list price, and (if it closed) the sold price. The intermediate path is not observed: a home that was reduced several times appears only as its opening and closing figures. Every discount and repricing measure here is therefore an endpoint comparison — original versus final, or original versus sold — not a count of how many times or when a price moved. This is sufficient for the questions asked and is stated so no more is inferred from it.

Limits, stated plainly. Resale figures describe homes that transacted more than once and do not generalize to all homes. Renovation between sales is not fully observable from listing data; the sub-six-month cohort is the control for it. The 400-day figure is reported over all such listings, sold and unsold, precisely to avoid the survivorship trap of counting only the eventual sales. Figures describe historical outcomes and are not a prediction for any individual property.

More research →  ·  The weekly index →  ·  One Shot: the 25-year listing study →