Replication package: When a Deed Is Not a Market Sale: Foreclosure Transfers, Statutory Consideration, and Repeat-Sales House-Price Measurement

Abstract

Version 1.0.1 corrects one bibliography entry: the companion paper, Loschi (2026), Who Gets the House? Financing-Associated Price Gaps in New York City, is now cited as published on Zenodo (version 2.3, 10.5281/zenodo.22921991) rather than as unpublished. The code zip and manuscript PDF were rebuilt accordingly; the README and citation file now carry the record's DOIs. Data, code, results and all other text are unchanged from version 1.0. Replication package and working-draft manuscript for When a Deed Is Not a Market Sale: Foreclosure Transfers, Statutory Consideration, and Repeat-Sales House-Price Measurement, by Pablo Loschi (independent researcher, Berlin). The paper studies lender take-title deeds in New York City's ACRIS land registry: referee's foreclosure deeds to lenders, deeds in lieu, and lender-to-lender transfers. State and city transfer-tax rules can base consideration for these events on a foreclosure judgment, a bid, or outstanding debt rather than a negotiated sale price; their detailed calculations differ. The package identifies 16,655 such deed-family documents among 1.47 million recorded in Manhattan, the Bronx, Brooklyn and Queens from 2003 to 2025. Standard screens retain many of them, and 94–95% of screened legal-rule deeds appear in the Department of Finance sales file at the recorded amount. In a Bailey–Muth–Nourse repeat-sales index of 1–3 family houses, removing pairs with a lender take-title endpoint changes the quarterly index by up to 3.7 log points (95% CI 2.5–4.9). In this sample and estimator, the cumulative peak-to-trough decline is within 0.6 log points of the uncleaned series. The maximum reaches 10.9 log points for Bronx houses and is negligible for condominiums. Files. ACRIS_CONSIDERATION_v1.0.1.zip: the code, the classifier, the manuscript (source, PDF and Word), tables, figures, documentation and the blind document-review sheet. ACRIS_CONSIDERATION_v1.0_DATA.zip: the frozen NYC Open Data pulls and the derived data. ACRIS_CONSIDERATION_v1.0.1_MANUSCRIPT.pdf: the standalone working draft. The README explains how to reproduce the results from the frozen snapshot. Validation status. The numerical cross-check passes 102/102, and an independent reconstruction reproduced the headline index correction. The 400-deed blind probability sample awaits human review, so classifier precision and population recall are not yet established. A non-random remark sample gave an indicative recall range of 89–94%, based on AI readings of buyer names. It is not a population error-rate estimate or human validation. Source data are public NYC Open Data records; see the included LICENSE for the source-data terms.