On 8 August 2026 New York City carried 3,717,235 open violations across its three enforcement systems. 1,955,528 of them — 52.6%, more than half — were more than five years old. 413,868 were issued before the year 2000.
"Open" is a status in a database, not a statement about a building today. This is a count of how long that status has persisted, which is a different and less flattering question than how many violations the city writes.
The clock had a hole in it
The obvious way to age an HPD violation is the date the notice was issued, novissueddate. That column is empty on 241,093 of HPD's open records — 8.4% of them — and the emptiness is not random: 238,087 of those 241,093, or 98.8%, are class I, the registration class. They are cases where the paperwork failed and no notice was ever served, so there is no issue date to have.
Every open HPD record does carry an inspectiondate — all 2,860,496 of them. So that is the clock used here, and the difference is worth showing, because the hole does not just add noise: it makes the answer depend on a choice nobody would think to declare. Age on the issue date and drop the 241,093 records that have none, and 48.4% look older than five years. Age on the issue date but keep those records in the denominator, and it is 44.3%. Age on the inspection date, which covers all 2,860,496, and it is 47.8%. A four-point spread, produced by nothing but how one missing column is handled.
By agency
| System | Open | Older than 5 years | Older than 10 | Older than 20 |
|---|---|---|---|---|
| HPD — housing maintenance | 2,860,496 | 1,367,254 (47.8%) | 873,738 (30.5%) | 461,932 (16.1%) |
| DOB — buildings | 579,356 | 412,685 (71.2%) | 311,614 (53.8%) | 184,387 (31.8%) |
| ECB — the hearings board | 277,383 | 175,589 (63.3%) | 122,574 (44.2%) | 53,547 (19.3%) |
| All three | 3,717,235 | 1,955,528 (52.6%) | 1,307,926 (35.2%) | 699,866 (18.8%) |
The three systems age very differently. HPD's open file is the youngest, and it is the one attached to conditions inside apartments — heat, hot water, leaks — which get re-inspected and certified. A DOB violation is the one that sits. Seven in ten open DOB violations are more than five years old and nearly a third are older than twenty; the median open DOB violation was issued in 2015, against 2022 for HPD and 2018 for ECB. Across all three the median lands in 2021.
The most serious close fastest
HPD grades what it finds: A non-hazardous, B hazardous, C immediately hazardous, and I for a registration failure. If the grade means anything operationally, the serious ones should clear faster — and they do.
| Class | Open | Older than 5 years | Share |
|---|---|---|---|
| C — immediately hazardous | 578,695 | 227,644 | 39.3% |
| A — non-hazardous | 676,316 | 291,795 | 43.1% |
| B — hazardous | 1,366,587 | 765,145 | 56.0% |
| I — registration | 238,898 | 82,670 | 34.6% |
Class C is the freshest of the three hazard grades, which is what a working enforcement system looks like. The number underneath that is still large: 227,644 violations the city classed as immediately hazardous have been open for more than five years. Class B is the outlier in the other direction — the middle grade, the biggest pile, and the one most likely to sit. What the classes mean is worth reading before drawing conclusions from the shape of that table; class I in particular is not a fourth severity level.
By borough, and the Bronx surprise
HPD violations, by where they are:
| Borough | Open (HPD) | Older than 5 years | Share |
|---|---|---|---|
| Brooklyn | 1,203,930 | 645,597 | 53.6% |
| The Bronx | 647,865 | 212,546 | 32.8% |
| Queens | 495,438 | 248,452 | 50.1% |
| Manhattan | 445,009 | 225,313 | 50.6% |
| Staten Island | 68,254 | 35,346 | 51.8% |
Four boroughs sit between 50% and 54%. The Bronx sits at 32.8% — twenty points below the rest, and the only borough where most of the open file is recent.
That is a striking number and it supports two opposite readings, so it is worth stating both. It can mean the Bronx's open violations are being resolved faster, leaving a younger file. It can equally mean the Bronx is receiving new violations faster than the others, which pushes the same ratio down without anything being fixed. This data cannot tell those apart — a closed violation leaves the open file either way — and anyone who tells you which it is from this table alone is guessing.
DOB's file is older everywhere and much flatter across the map:
| Borough | Open (DOB) | Older than 5 years | Share |
|---|---|---|---|
| Brooklyn | 236,694 | 172,422 | 72.8% |
| Manhattan | 132,600 | 95,656 | 72.1% |
| Queens | 118,825 | 83,112 | 69.9% |
| The Bronx | 72,481 | 48,176 | 66.5% |
| Staten Island | 18,696 | 13,316 | 71.2% |
Every borough is between 66% and 73%. Whatever makes a DOB violation sit for years is a property of the system rather than of the neighbourhood it is in.
The tail
413,868 open violations were issued before the year 2000 — 272,717 at HPD, 116,551 at DOB, 24,600 at ECB. The 1990s alone account for 116,240 open HPD violations and 109,519 open DOB ones.
Past that the record thins out fast and stops being believable: 15,015 open HPD records carry an inspection date before 1980, and DOB has 114 and ECB 14. The very earliest carry dates in the 1910s, which is not a plausible inspection and is a data-entry artefact rather than a finding. They are left in the totals because removing rows for looking wrong is how a dataset gets quietly edited into agreement with expectations, and 129 records out of 3.7 million move nothing.
What an old open violation is, and is not
It is not proof that a condition still exists. A violation stays open until somebody certifies it corrected, and certification is a filing an owner has to make — so a hazard fixed the week it was written stays on the record for twenty years if the paperwork never followed. Some of the tail above is exactly that.
Nor is it nothing. The certification is the owner's obligation, and an open record is the city's own statement that it has not been told the problem was solved. For a class C violation — the grade the city gives to no heat and no hot water — five years without that statement is a fact about the building's management, whatever the apartment looks like today.
The practical consequence for a reader is narrow and useful: read the age, not just the count. Forty open violations issued last month and forty issued in 2009 are very different buildings, and the totals people quote do not distinguish them. Every building page on this site lists each violation with its date.
How this was computed
Three datasets, counted in full rather than sampled: HPD Housing Maintenance Code Violations (wvxf-dwi5) where violationstatus='Open'; DOB Violations (3h2n-5cm9) where violation_category contains ACTIVE; and ECB Violations (6bgk-3dad) where ecb_violation_status='ACTIVE'. Ages are measured from inspectiondate for HPD and from issue_date for DOB and ECB, against 8 August 2026.
Medians are the year in which the running total crosses half, computed from a count of open violations per year — a median year rather than a median number of days, which is the honest precision for data where the day of month is not always meaningful.
DOB records the borough as a digit on most rows and as a legacy letter on a few, so M, B, Q and S are folded into Manhattan, the Bronx, Queens and Staten Island respectively; 60 rows carry a blank or zero borough and are in the citywide totals but in no borough row. The class and borough tables each sum exactly to the citywide HPD total of 2,860,496, which is the check that the folding lost nothing.
These are counts of records, not of buildings or of people. One building can hold hundreds of open violations, and nothing here is attributed to any owner, contractor or tenant.