A reporter counted every condo in an LA tower. New York exempts what she found.

281 of the 822 condominium units at the Metropolis development in downtown Los Angeles share one owner, the original developer, and every one of those 281 addresses carries a U.S. Postal Service mail-vacancy designation. That is not an estimate. It is a count, assembled from county property records by a reporter working alone. New York City can run the same count on any of its 14,062 condo and co-op buildings, and on two separate tax instruments it has already flagged which units are somebody's primary residence. Both instruments exempt unsold sponsor units by name.

What she did, and what the records said.

Yoonj Kim is an independent journalist who covers housing and land. On August 17, 2026, she published a piece of records work on the Metropolis, a roughly $1 billion, 3.3-million-square-foot development built by Greenland USA and completed between 2016 and 2019. It contains a hotel, a rental tower, and two condominium towers holding 822 units. The local rumor she set out to test was that the luxury towers were empty.

Yoonj Kim, "I Traced Every Condo in This LA Skyscraper. The Rumor Is True," published August 17, 2026 on her own channel. If the embed does not load, watch it on YouTube. Her site: yoonjkim.com.

Her method used three public layers and stacked them. First, the county assessor's homeowners' exemption, which an owner claims only for a primary residence: 84 units, about 10 percent, had claimed it. Second, the assessor's owner field, sorted for repeats: 281 units resolved to a single owner, Greenland USA. Third, the Postal Service mail-vacancy designation, which marks an address that has not received mail in 90 days: all 281 carried it. She then went back at night and photographed the towers dark, which she is careful to describe as a corroborating data point rather than the finding. The fourth tower on the same block, the one leased as rentals, runs at 91 percent occupancy. Greenland USA did not respond to her request for comment, and the reason those units sit unsold is her stated inference, not an established fact.

Her closing point is the one that travels. Vancouver counts its empty homes. Paris counts. New York, she says, at least keeps track. Los Angeles, a city under a declared housing emergency, has never counted once. And the harder category, the third home used six weekends a year, hides better than a wholly empty tower.

New York flags primary residence twice, and exempts sponsors twice.

New York City is not short of the records Los Angeles lacked. It has two separate instruments that turn on whether a home is the owner's primary residence, and both of them create exactly the flag Kim used. Read them side by side and the same carve-out appears in both.

The statutory language in § 1351(g) is not ambiguous. Excluded property includes a residential condominium or cooperative unit that is subject to an offering plan required by General Business Law § 352-e where the unit "has not been sold" and no economic interest has been transferred by the person or entity that filed the plan. A New York developer holding 281 finished, unsold, empty units under its own offering plan is holding excluded property. The surcharge does not reach it, and the abatement roll never listed it.

The 959,710-record roll counts something else.

On July 24, 2026, the Department of Finance published two supplemental files flagging 959,710 property records as possibly subject to the new non-primary-residence surcharge, and mailed roughly 17,000 letters. We wrote about the arithmetic of that gap in "Taxed as a rental building" and about the owner-side deadline in "New York has no primary-residence registry". The point for this piece is narrower. That roll is the largest per-unit occupancy dataset the city has ever published, and it still cannot answer Kim's question, for two reasons.

1. Non-primary is not empty. A pied-à-terre occupied six weekends a year, a corporate apartment, and a unit nobody has entered since 2019 all land in the same bucket. The roll separates residents from non-residents. It does not separate used from unused.
2. The unsold-sponsor category is excluded upstream. Under § 1351(g) those units are not on the surcharge roll at all, so the one dataset built to identify occupancy at unit level is the dataset that omits the pattern the Los Angeles story is about.

The vacancy rate New York publishes does not measure this either.

The number quoted whenever New York is described as keeping track is the net rental vacancy rate from the NYC Housing and Vacancy Survey: 1.41 percent citywide in 2023, reported as the lowest since 1968. It is a real number produced on a real statutory schedule, and it is the number that legally matters, because NYC Admin. Code § 26-501 continues rent stabilization only while the vacancy rate sits at or below 5 percent. It is also a survey of the rental stock. It is triennial, it is a sample, and it counts units available for rent. An unsold condominium held by a sponsor is not available for rent and is not what that survey was built to find. Anyone reaching for the 1.41 percent figure to describe owned housing is using a rental instrument on a question it does not answer, and the borough detail makes the mismatch plain: in the same survey Manhattan reads 2.33 percent, the loosest of the four boroughs measured, not the tightest.

Where the sponsor number actually lives.

New York does collect the underlying fact, in a place almost nobody reads. A sponsor offering condominium or cooperative units must file an offering plan with the Attorney General under GBL § 352-e, and must file amendments under 13 NYCRR Part 20 for condominiums and Part 21 for cooperatives. Those amendments disclose the sponsor's position on unsold units, including whether the sponsor is current on common charges, taxes, reserve and working-capital payments, assessments, and financing obligations relating to unsold units. The information exists. It arrives as a PDF, building by building, filed with the Real Estate Finance Bureau, and no agency aggregates it into a count. There is no public figure for how many finished New York City condominium units are sitting unsold on a sponsor's books this morning, in any building, in any borough.

That is a records-publication gap rather than a records-collection gap, and it is the kind that is cheap to close. The city already runs a per-unit valuation file for condominiums and already administers a per-unit primary-residence certification under § 467-a. The Attorney General already receives the sponsor's unsold-unit disclosures on a filing schedule. What does not exist is the join, published, on a cadence, in a format a resident or a reporter can open.

Bottom line.

Kim's Los Angeles finding was possible because a county assessor publishes owner names and a primary-residence flag at unit level. New York City publishes more than that, and has just added the largest per-unit occupancy roll in its history. The category that story turned on, finished units still held by the developer, is written out of both of New York's primary-residence instruments in plain statutory text: exempt from the § 467-a abatement as a sponsor, and excluded property under Tax Law § 1351(g) until the unit is sold. Los Angeles has a counting problem. New York has a publication problem, which is the easier of the two to fix and the one nobody has been asked to fix.

Companion resources: Taxed as a rental building, exempt from every rental protection • New York has no primary-residence registry • The three numbers New York calls a vacancy rate • The LLC Transparency Act and condo sponsors • The 14,062-building master list • Issue index