Manhattan's average rent hit a record $6,655. The 1.49% vacancy rate quoted beside it is June's number, and it is not the city's.

On August 13, 2026, Corcoran published its July Manhattan rental report. The average rent was $6,655, an all-time high, up 10% from a year earlier. In the coverage that followed, that figure travelled alongside a second one: a vacancy rate of 1.49%, described as the tightest New York has seen since 1968. Both numbers are real. They come from two different reports, they measure two different things, and the 1.49% is from June.

What Corcoran actually published.

The July report is specific. Median Manhattan rent was $5,295, unchanged from June and up 6% year over year. The studio average reached $4,088 and the one-bedroom average reached $5,486, both records. Two-bedroom averages rose 13% and three-bedroom averages rose 12%; the New York Post reported those levels at $8,054 and $12,228. Active listings fell to 5,198, down 22% from a year earlier, the sixth consecutive month of double-digit annual declines and the lowest July total since Corcoran began tracking the series in 2019.

The vacancy figure in that same report is 1.56%. Corcoran states it plainly: the Manhattan vacancy rate rose in July from the 1.49% recorded in June. The 1.49% appears in the June report, which also notes that May was 1.57%. So the vacancy number quoted alongside the record rent is the prior month's reading, and the direction it moved in the month the record was set was up. Corcoran's series also begins in 2019. It cannot reach 1968.

The 1968 comparison belongs to a different survey.

The measurement that reaches back to 1968 is the New York City Housing and Vacancy Survey, conducted roughly every three years since 1965 and sponsored by the Department of Housing Preservation and Development. It is not a market-colour statistic. It carries a statutory job. Under NYC Administrative Code § 26-501, rent stabilization continues in New York City only while a housing emergency exists, and the vacancy rate has to sit at or below 5% for that finding to hold.

The 2023 survey put the citywide net rental vacancy rate at 1.41%, down from 4.54% in 2021, the lowest since 1968. That translates to roughly 33,210 vacant units available for rent out of nearly 2.4 million rental units. Those figures, and the borough detail below, are reproduced in the Rent Guidelines Board's 2026 Housing Supply Report of May 21, 2026.

The 2023 survey is three years old, and it remains the most recent authoritative citywide measurement. The 2026 NYCHVS is in the field now, conducted for the first time by the University of Michigan's Institute for Social Research with CUNY students as field interviewers, under a partnership HPD announced in 2025. Its results are not out. Every argument being made this month about citywide scarcity is being made against a 2023 baseline.

By the city's own survey, Manhattan is the loosest rental borough in New York.

This is the part that does not survive the transfer. The 2023 NYCHVS did not find a uniform 1.41% across the five boroughs. It found a range, and Manhattan sat at the top of it.

Manhattan's rate was the highest of the four boroughs that could be measured, close to three times the Bronx's. The tightest markets in New York City are the ones where the rents are lowest, and the borough generating the $6,655 headline is the one with the most slack in it. Rent-stabilized units were vacant at roughly half the rate of market-rate units, which is the opposite of what a warehousing story predicts.

Two brokerages, one month, opposite conclusions.

Corcoran is not the only firm that published a July 2026 Manhattan report. Douglas Elliman published one prepared by Miller Samuel. The two describe the same borough in the same month and do not agree on the size of the market or the direction of leasing volume. Elliman's figures below are as reported by Brick Underground.

A $349 gap in the average and a $295 gap in the median are large enough to move any affordability calculation built on them. The leasing figures point in opposite directions entirely. Neither series is audited by any agency. Both are counts of what moved through a brokerage panel, and the panels are not the same. When a number has no custodian, every party to the argument gets to pick the version that supports the position they already hold.

The 57,421 vacant stabilized apartments are a third definition.

The state's Division of Homes and Community Renewal reported to the Rent Guidelines Board that 57,421 rent-stabilized apartments were registered vacant as of April 1, 2025, up 7,995 from the 49,426 registered a year earlier, roughly 6% of about a million stabilized units. The 2023 NYCHVS put the stabilized vacancy rate at 0.98%. Two measurements of the same housing stock, six times apart.

The gap is definitional rather than contradictory. The survey counts units available for rent. The registration file counts every unit an owner recorded as vacant on a single date, including new buildings not yet leased up and units mid-turnover. HCR Deputy Commissioner Anthony Tatano told Gothamist the registration "does not indicate the cause of vacancy, condition of units or whether the units are transitioning between tenancies." Jay Martin of the New York Apartment Association reads the same file as suppressed supply. Judith Goldiner of the Legal Aid Society reads it as ordinary churn in a million-unit stock. The file does not settle it, because it was never built to.

The one audit that has been run on the question came from the Comptroller's office in March 2024. Using the 2023 survey, it found units that were vacant and unavailable fell from 42,860 in 2021 to 26,310 in 2023, and units classed dilapidated or uninhabitable fell from about 11,500 to just over 3,000. It reported no evidence that the 2019 rent law drove an increase in vacant or distressed stabilized units.

What the historical record shows about supply.

Here the numbers are unusually good, and they cut against the story being told with them. In 2025 an estimated 38,691 residential units in new buildings were completed citywide, up 14.3%, the most in a single year since 1965 and the second consecutive year above 30,000, which last happened in 1966. That is the strongest two-year production run in six decades.

It is also finished. Permits were issued in 2025 for 17,673 units, and Manhattan's permits fell 14.3% to 1,617 units. The housing pipeline, meaning all active permits for buildings not yet complete, stood at 66,805 units on December 31, 2025, down 31.0% from 96,854 a year earlier. City Planning's own characterisation, quoted in the RGB report, is that the pipeline "represents a relative low compared to more recent years, and many neighborhoods, both in New York's low-density areas and centrally located parts of Manhattan, have few or no units in the housing pipeline." The record completions of 2024 and 2025 are the arrival of permits pulled before 421-a lapsed in June 2022, not evidence of a market that is currently building.

The condo and co-op pipeline moved the same way. The Attorney General accepted 202 residential co-op and condo plans in 2025, down 22.3% from 2024, containing 3,525 units, down 31.0%. Anyone waiting for ownership supply to relieve rental demand is waiting on a smaller queue than last year's.

What the pipeline arithmetic implies for 2027 through 2030.

The RGB states that a new permit typically becomes occupancy within about four years. Run the pipeline against that clock and the shape of the next several years is arithmetic rather than forecast. Assume the entire 66,805-unit pipeline completes, none of it is cancelled, and it lands evenly across four years. That is roughly 16,700 units a year, against 38,691 delivered in 2025. In Manhattan, 10,088 pipeline units across four years is roughly 2,520 a year, against 4,874 completed in 2025.

Those are ceilings, not estimates. Cancellations push the number down. New permits do not push it up inside the window, because the four-year lag means permits issued in 2026 arrive in 2030. The replenishment rate is the permit figure, and citywide that is 17,673 units against 38,691 completions. Manhattan's is 1,617 against 4,874. For context on the demand side, the city's net housing stock grew by roughly 60,000 units between the 2021 and 2023 surveys while about 275,000 new households formed, and the zoning package passed in December 2024 is projected at roughly 80,000 homes across fifteen years, or about 5,300 a year.

Nothing in that set of numbers produces a looser Manhattan rental market before 2030 on supply alone. The 485-x replacement for 421-a, the 467-m conversion incentive, and the accessory dwelling unit laws all sit inside that same four-year lag.

What political capital it has taken, and what it would take.

The record of the last twenty months is a fairly precise price list for housing policy in New York City.

• The rent freeze cost the least. The Rent Guidelines Board voted 7-1 on June 25, 2026 to set 0% adjustments for one- and two-year leases beginning October 1, 2026, covering about a million stabilized apartments. The Board is mayoral-appointed, so the mechanism was appointments rather than legislation. It reaches the 41.3% of occupied rentals that are stabilized. It does not reach the 48.1% that are unregulated, which is where a $6,655 average lives.
• The FARE Act cost a Council vote and a defence. The Council passed it 42-8; it took effect June 11, 2025. The Real Estate Board of New York sued in the Southern District, and Judge Ronnie Abrams declined to enjoin it. The law changed who pays the broker. It did not change the rent.
• The zoning package cost the most and delivers the slowest. City of Yes for Housing Opportunity passed 31-20 on December 5, 2024, scaled down from about 109,000 projected homes to roughly 80,000, with about $5 billion in accompanying commitments including $1 billion from the State. That is the going rate for roughly 5,300 units a year.

The 2025 primary answered the question of whether money is still the binding constraint. Citizens Union counted $46,092,604 in super PAC spending across the June primary, including $26.7 million supporting one mayoral candidate and $8.2 million opposing another, against just over $1 million spent supporting the candidate who won. Business super PAC money rose to $18.6 million from $5.2 million in 2021, with $1.2 million from the New York Apartment Association. Roughly $30 million of the total went to candidates who lost.

So the capital required is not primarily financial, and the city-level levers have already been pulled. What has not been touched is the set of instruments that actually price new rental supply, and every one of them sits in Albany: the tax-exemption programme that replaced 421-a, the 2019 rent law, and the Emergency Tenant Protection Act itself. That is the same jurisdiction where the reforms in our legislative graveyard have been held at committee level for nine consecutive sessions.

There is a cheaper reform available first, and it is the one this month makes the case for. No agency owns the number. The Rent Guidelines Board sets guidelines for a million apartments using a survey that is three years old and a registration file that does not record why an apartment is empty. The figures that reach the public between surveys are produced by brokerages measuring their own panels, unaudited, and they disagree with each other by $349 on the average rent and by 22 percentage points on leasing volume in the same month. A statutory obligation to publish one reconciled, methodologically disclosed vacancy and rent series would cost a fraction of what the zoning package cost, and it would make every subsequent argument checkable.

Bottom line.

Manhattan rents did set a record in July 2026, and inventory is genuinely scarce. The explanation attached to that record is assembled from three measurements that answer different questions: a brokerage listing rate for one borough, a Census-style citywide survey with a statutory job, and a landlord registration file that records a date rather than a condition. Quoting the June brokerage figure and the 1968 survey superlative in one sentence produces a claim neither source makes, and the survey that does reach 1968 puts Manhattan last in scarcity among the boroughs it could measure. This is the same structural finding we document in condo and co-op governance, where the managing agent needs no licence and no agency takes the complaint. The gap is not that New Yorkers disagree about the housing market. It is that nobody is required to keep the score.

Companion resources: The Extraction Economy · 485-x has a 99-unit floor · When your 421-a abatement expires · The pied-à-terre tax and your co-op board · The legislative graveyard · How we source and score · Write to your representative