The Shape of the Tax: Transaction Shifting at New York City's 2019 Transfer-Tax Notches and the Mortgage Recording Tax

New York's 2019 reform introduced seven buyer-paid transfer-tax notches. At the three thresholds with enough sales to study, recorded prices moved below the threshold, compared with sales at the same prices before the reform and with round prices that were never taxed. The USD 1 million notch, which the reform left unchanged, gives a further comparison. The responses across thresholds reject a proportional benchmark, but the markets at each threshold also differ, so this is not a rate experiment. If the extra sales below each threshold moved there from above it, buyers avoided USD 2–3 million in tax a year. Loan records also give a benchmark for the gross mortgage recording charge and a static costing of continuous replacement schedules. Neither exercise identifies who bears the tax or what the optimal tax would be. Files: the preprint, the supplemental appendix (unchanged since 2.6), the source archive and a short summary (paper_in_brief.md). The replication package is at 10.5281/zenodo.22880953. See revision_notes.md for earlier changes.