A FairPermit civic report · 2026

The State of NYC Permitting

What 16 million filings reveal about how the city builds, who it leaves behind, and how to fix it.

New York City files about 160,000 building permits a year, and the number is growing. More than half crawl through plan examination, where roughly one in twenty draws an objection and the rest face weeks of unpredictable delay. 82% of filers already pay an expediter whose business is a black box, and whose largest player owns barely 2% of the market. The result is slower housing, slower small-business openings, and a gatekept process that taxes the New Yorkers least able to absorb the delay. We mapped exactly where the system breaks. The fix is not more bureaucracy. It is transparency the city can hand to every New Yorker for free.

~160k
permits filed per year
52%
stuck in the slow lane
100×
higher objection rate, slow vs fast lane
82%
of filers pay an expediter
2.3%
market share of the single largest firm
~52k
jobs a year needlessly in the slow lane

1. How the system actually works

NYC building permits move down one of two lanes:

The same job type often qualifies for either lane. Within Alterations (85% of all volume), 58% already self-certify, but 42% go through plan examination anyway for jobs their own peers self-certify. That is roughly 52,000 jobs a year sitting in the slow lane that demonstrably did not have to be there.

Every job in the slow lane is a unit of housing, a storefront, or a family's renovation waiting longer than it needs to. Moving even a fraction of those 52,000 jobs into the fast lane is faster housing and faster small-business openings with zero change to safety.

2. The objection tax faster housing & small business

A plan-exam objection is not a safety win. It is a delay. The data shows 13,255 jobs drew objections since January 2025, of which 3,251 are Manhattan and Brooklyn alterations alone. Each objection cycle is weeks of unbillable schedule slip for a contractor and weeks of a closed storefront or an unfinished apartment for everyone downstream.

The objections are also predictable. A risk model trained on 2 million realized outcomes scores a job's objection probability before it is filed at AUC 0.82, validated out of time on 567,773 filings the model never saw. The city is absorbing weeks of avoidable delay on objections that were foreseeable and fixable at the desk.

3. The affordability gap affordability & equity

Permitting in NYC is a paid, gatekept market. 82% of all filings, about 123,000 a year, already go through a paid filing rep. The market is worth on the order of $120 million a year in pure filing-service spend, before city fees.

And it is brutally fragmented and opaque. The single largest filing-rep firm handles only about 2.3% of the market. The top eight together are about 12%. There is no brand, no published price, no quality signal. You call an expediter and you hope.

The New Yorkers who lose are the ones with the least slack: small general contractors, bodega owners doing a build-out, families renovating a brownstone. Among small residential jobs (1 to 3 family, 23% of all filings), 32% are filed do-it-yourself because the household cannot or will not pay an expediter, and those are exactly the filers most exposed to a surprise objection they do not know how to clear. The cost of opacity falls hardest on the people who can least afford the delay.

4. The transparency black hole open government

The city already publishes the data. It is just not legible. A homeowner cannot look up their own job and learn, in plain English, whether it is at risk, why an objection was raised, or what the fastest legal path is. The knowledge is gatekept inside a few thousand expediters' heads. Turning the city's own public record into a plain-English answer any New Yorker can read is a pure open-government win, and it costs the city nothing.

5. The hidden opportunity: good jobs distributed work across the boroughs

The supply side is as fragmented as the demand side. There are about 2,820 active filing reps, with a median of just 9 jobs over two years, and 74% of them file fewer than 50 jobs over two years. Only 276 file at any real volume. This is a long tail of small, under-tooled operators, spread across all five boroughs, who would do far more work if the hard 90% were done for them.

Separately, of about 4,190 self-certifying professionals, the median did only 5 self-certs in two years, and 66% are "light" users. Most licensed pros are too cautious to self-certify at volume because the audit risk feels personal and unmeasured. Give them measured, audit-safe confidence and the cautious majority become the supply that powers faster approvals, and a source of flexible, well-paid filing work distributed across every borough, not concentrated in a handful of midtown firms.

6. What the city could do, and where FairPermit fits

None of this requires new bureaucracy. It requires legibility.

FairPermit is the transparency layer the city has not had the tooling to build. The licensed human stays in the loop, so it is compliant by construction. It is decision support, not a substitute for the professional. And the free public tier is a genuine public good: faster housing, cheaper permits for working New Yorkers, fewer stalled sites, and a more honest, more open city government.

7. What to do, by who and by when

Concrete, low-regret moves. Each is grounded in a number above, and most cost little or nothing.

Department of Buildings the regulator

Mayor's Office of Operations and the housing team the executive

City Council oversight

Filing reps and design professionals the market

Civic partners and FairPermit the public good


Methodology and limits. Figures are queried from the full public DOB record: about 2.7M legacy BIS job filings, 0.9M DOB NOW filings, and 0.9M approved permits, with the full 62.7M-row, 120-dataset build maintained on a daily refresh. The objection proxy is filing status containing "objection" or "disapproved," treated as directional and consistent with the backtested risk model. The risk model is a hierarchical empirical-Bayes model over job type, professional-certification status, borough, and building type, validated out of time. Borough, lane, and rate figures are point-in-time and refresh with the daily ingest. This report makes no safety claims: the fast lane already carries a lower objection rate than the slow lane, so lane conversion is a speed and cost win, not a safety tradeoff.