Most cities never mark their building permits closed. We checked thirty of them.
September 5, 2026
An open building permit is supposed to mean something specific: work was authorised, started, and never signed off as finished. It is one of the few concrete red flags a homebuyer can check for free.
While connecting to thirty city permit systems, we found that what an open permit means depends enormously on which city issued it. In some, the status is meticulously maintained and an open permit is a real signal. In others, permits are marked issued and then the record is never touched again. In six, there is no status field at all.
The same word, in two different cities, can mean “work is happening right now” or “a permit was granted at some point in the last forty years.”
The finding
[CHECK: re-run the Las Vegas GROUP BY on STAT before publishing.] In Las Vegas, of roughly 1.27 million permits in the city's published record, about 95% are still marked “Issued” and under 3% are marked “Closed.” That ratio holds across an archive spanning decades.
Nearly every building in Las Vegas that has ever been permitted still carries a permit the public record shows as live. A homebuyer checking that record, applying the ordinary meaning of the word, would conclude that essentially the entire city is mid-renovation.
The likelier explanation is simpler: the department issues permits and does not routinely update the row afterward. That is a records practice, not a construction crisis. But from outside, the two are indistinguishable, and the distinction is the whole thing a buyer wants to know.
Three kinds of city
Cities whose records close
[CHECK: re-run each figure in this section.] Orlando reaches a terminal state on roughly 93% of its million-plus permits. Pierce County, Washington, reaches about 81% across forty years of records. Montgomery County, Maryland, finals just under 70% of residential permits.
In these jurisdictions an open permit is meaningful. The department demonstrably closes the ones that finish, so one that stayed open is genuinely unusual and worth asking about.
Cities whose records rarely close
Las Vegas is the extreme, but the pattern shows up in softer forms elsewhere. The diagnostic is the ratio: if a city has decades of history and almost nothing marked closed, the status column is recording issuance rather than completion.
Here the word “open” carries almost no information, and a risk score built on counting open permits would flag nearly every property in the city.
Cities that publish no status at all
[CHECK: confirm this list against your current config before publishing. As last verified it was Nashville, Denver, Buffalo, Baton Rouge, Prince George's County and Everett. If any of them has since added a status column, that is itself worth mentioning — it would be the first improvement we have seen.]
Six of the thirty systems publish permit records with no status field. You can see what was filed, when, by whom and for how much. You cannot see whether any of it was ever finished, because the city does not say.
The DC pattern: permits that expire rather than finish
[CHECK: re-run both DC cohort figures.] Washington DC publishes permits in one dataset per year, which makes it possible to watch a cohort age. Comparing the 2009 permits with the 2025 permits produced the most interesting result in the whole exercise.
- 01In the recent cohort, roughly 77% are marked issued, 20% completed, and only about 2% expired.
- 02In the cohort sixteen years older, expired has climbed to roughly 36%, while completed has barely moved at around 18%.
As DC permits age, they do not become completed. They become expired. The share that reach completion stays roughly flat while the expired share grows by a factor of eighteen.
Whatever the cause, the practical consequence for a DC buyer is that expiry, not the open state, is the signal worth watching, which is the opposite of the standard advice.
Stop-work orders are almost never published
A stop-work order is among the most serious things a building department issues: work was authorised, started, and then halted. As a signal of unpermitted or non-compliant construction it is far stronger than an ordinary expired permit.
[CHECK: confirm which cities publish a stop-work status value and the Montgomery County count.] Only three of the thirty systems publish it as a status value at all. In Montgomery County, Maryland, roughly 2,771 residential permits carried it as of our last check.
In the other twenty-seven jurisdictions, a stop-work order may well exist on a property and simply not appear in the public permit record. That is a gap worth knowing about before you treat a clean-looking record as an all-clear.
Why any of this matters to a buyer
The practical takeaway is that an open permit is not a fact, it is a claim made by a particular city's record-keeping, and the reliability of that claim varies enormously.
In a city with a well-maintained status column, an open permit standing years after the work justifies real concern. It is worth resolving before you close. In a city that never updates the row, the same open permit tells you almost nothing, and treating it as a red flag will have you renegotiating over an administrative artefact.
Neither situation is visible from the permit itself. You have to know how the city behaves.
What we do about it
This is why our reports withhold a risk score in some cities rather than producing one everywhere.
Where a city publishes no status, or where the vocabulary never reaches a closed state, there is no honest way to count open permits. So the report shows the full permit history and says plainly that a score cannot be issued, along with why. It is a worse-looking product in those cities and it is the only defensible one: the alternative is selling a green all-clear generated by a missing column, or flagging an entire city as high risk because of how its clerks file paperwork.
Method
Every figure comes from cities' own open data portals, queried directly through their public APIs. Counts are produced by grouping on each city's status column across its full published record, not a sample. Nothing is modelled, weighted or extrapolated.
Three limits worth stating plainly. First, we can only see what a city publishes; a permit closed in an internal system but not reflected in the open data feed looks open to us and to everyone else. Second, coverage windows differ enormously, from 1907 in Orlando to 2019 in Pittsburgh, so cross-city comparisons of raw volume are not meaningful and we have avoided making any. Third, some cities split permits across several datasets and we may not have connected to all of them.
[CHECK: state the date you re-ran these figures. A data study without a "figures checked on" date reads as unmaintained within a year, and reporters will not cite an undated number.]
If you want to reproduce any of this, the endpoints are public and the queries are ordinary aggregate counts. If you find an error, we would rather hear about it than not.
Frequently asked questions
What does it mean when a building permit is never closed?
It means the city's public record still shows the permit as issued or open, with no final inspection recorded. Sometimes the work genuinely was never finished. Often the work was completed and nobody updated the record. From outside the department, those two situations look identical.
Which cities publish building permit status?
Most do, but not all. Of the thirty systems we connected to, six publish no permit status column at all, meaning there is no way to tell from the public record whether any permit was ever closed out in those jurisdictions.
Why do some property reports show no risk score?
Because a score built on missing status data would be a guess. Where a city publishes no status, or where its status vocabulary never reaches a closed state, any risk number is invented. We show the permit history and withhold the score.
Is an open permit always a problem?
No. In cities where records are rarely updated after issue, an open permit is weak evidence of anything. In cities that close permits reliably, an open permit standing years after the work is a much stronger signal. The same word means different things depending on the city.
What is a stop-work order and how common is it?
It is a formal halt issued by a building department, usually for work that is unsafe or outside what was permitted. It is among the strongest signals a department publishes, and most cities do not publish it as a status value at all.
Where does this data come from?
Every figure comes from cities' own open data portals, queried directly. Nothing here is modelled, estimated or extrapolated. The method section explains exactly how each number was produced so anyone can reproduce it.
Check the permit history before your next showing
Search any address across our covered cities to see every permit on file, which are still open, and what the work actually was. Our free Chrome extension shows the same thing directly on Zillow and Redfin listings.
Search a propertyThis article is general information about the building permit records process. It is not legal, financial or real estate advice, and permit rules and enforcement practices vary by city. Verify anything you plan to act on with your agent, your attorney or the local building department.
All articles