Redfin puts 14% of July's home-sale agreements as falling out of contract

A coordinator's file count and a coordinator's income are the same number only if every file closes. Redfin's July data puts the gap between them at about one in seven.

In its monthly housing market report published 12 August 2026, the brokerage reported that 14% of July's home-sale agreements fell through, up 0.3 percentage points from June and 0.7 points from a year earlier, and described it as the highest share since 2023. In the same table Redfin recorded 335,051 pending sales and 285,312 closed sales for July, a spread of roughly 50,000 transactions between the two.

Why this figure matters more to a coordinator than to an agent

An agent working on commission and a coordinator working on a flat per-file fee both lose money when a contract dies, but they lose it differently. The agent's loss is proportional to the deal and is priced into a commission structure that assumes a failure rate. A coordinator's loss is the full cost of the work already performed on that file: the opening, the disclosures, the deadline calendar, the vendor coordination, most of which happens in the first two weeks after a contract is signed.

That is the part of the transaction where a coordinator's effort is front-loaded. A file that dies at the inspection has consumed most of its coordination cost and produced no closing. At a 14% failure rate, a desk running 200 signed files a year is doing the opening work on roughly 28 that never reach a settlement statement.

What is causing contracts to fail

Redfin attributes the July level to affordability and economic uncertainty, noting the median sale price rose 3.2% year over year to $407,730, the highest July level in its records, while the monthly average 30-year mortgage rate rose to 6.54%. Chen Zhao, Redfin's head of economics research, described 'would-be buyers grappling with record-high home prices, increasing mortgage rates and growing financial insecurity.'

NAR's Pending Home Sales methodology names the mechanisms without putting a number on them, stating that variations between contract and closing 'can be caused by issues such as buyer difficulties with obtaining mortgage financing, home inspection problems, or appraisal issues.' Those three are the classic contingency exits, and all three are the coordinator's chase list.

The measurement problem

This publication's standing rule is that vendor statistics run beside independent corroboration. On this measure that is not possible, and the reason is worth stating plainly rather than papering over.

No federal statistical series measures the share of real estate contracts that fail. Census counts new-construction sales at contract signing and does not track subsequent cancellation as a published rate. NAR publishes contract signings and closings separately but no failure rate connecting them. FHFA's mortgage database records loans that originated, which by definition excludes transactions that never reached closing. The brokerages and portals that can see both ends of a transaction in their own books are the only parties publishing a number, and each of them is describing its own book.

The number to keep yourself

1. Track your own fall-through rate by month. It is the one figure in this story that is knowable with certainty, and most desks do not record it.
2. Track it by referring agent as well as in total. Fall-through concentrates, and a single agent's pipeline can carry most of it.
3. Track where in the file it happens: inspection, financing, appraisal, or buyer withdrawal. Each has a different point of no return and a different amount of consumed work.
4. Compare your own rate to the direction of the published figures rather than to their level. A brokerage's book is not your book.
5. Revisit the engagement letter when your own rate moves, not when a national figure does.