New York clears 1,210 pied-a-terre letters it mailed before reading the 2025 returns

The supplemental roll that started all of this was an unfiltered list of nearly a million properties, promptly branded a rich hit list online, and it pulled in thousands of legitimate primary residences ^[19]. That is the base the surcharge is being assessed against, and it is being narrowed by discovery rather than by measurement.

The audit only makes sense read backward, from the correction to the cause. 1,210 clearances against roughly 17,000 first-wave notices is a 7.1% error rate ^[1], and DOF's own screening did not surface it; Albany moved its calendar ^[6], releasing preliminary records that Mamdani said are typically unavailable until February of the following year ^[8], the same mayor having acknowledged that the letters went out before the agency had access to the 2025 filings at all ^[7]. Of the cleared owners, 630 came off on a single clean match, a 2025 return listing the property as the primary home, while 580 required 2025 extension filings stitched to 2024 returns ^[5], which means 48% of the corrections took more than one document to make ^[2].

Fifty-nine percent of the new wave rests on that same gap ^[11][3]: entity-owned homes where the city says it does not have enough information on who actually lives there, with the burden of proof routed to the owner, who has 55 days from the arrival of the state data to file ^[6], or an appeal to the tax commission early next year if that date passes ^[17]. Randy Mastro, the attorney suing over the rollout, reads the retractions as proof that there are many thousands fewer owners subject to the surcharge than the administration claimed when it sent out 17,000 notices ^[9]. His read on the direction may hold up even if the number he is citing turns out to be premature.

Several endings are still open here. The clearances read as a functioning audit, the exemption window absorbs the remainder, and the annual take lands where City Hall says it will. Or the sequencing argument prevails and DOF has to re-derive the roll from state records before it mails anything else, which pushes collection into a later window. Or the 6,400 entity-held households produce exemptions in volume and the payer list contracts a third time, at which point the revenue estimate written before any of this data existed becomes the document worth reading.

This desk's read, and it could be wrong: the false positives are a rounding error, but the sequence is the finding, because a levy the legislature passed on May 27 and the governor signed the next day ^[18] was still assembling its payer list in late August, and a first-year revenue line built on a list of that vintage gets booked on inputs that arrived after it. The thesis fails if October's exemption filings come in at or below the 1,210 already cleared, which would mean the roll was roughly right and Albany's early hand-off caught essentially the whole error in one pass.