# Corporate-Family Resolution Is Not a String-Matching Problem: A Public Benchmark Stratified by Name Visibility

arXiv: https://arxiv.org/abs/2609.04269 · HTML: https://mlsanity.com/paper/2609.04269/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 50 references that are in the map.
- First posted on arXiv 2 Sep 2026.

## Builds on (most cited first, 21 of 21)

- [BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents](https://mlsanity.com/paper/2504.12516/) (2025)
- [Measuring short-form factuality in large language models](https://mlsanity.com/paper/2411.04368/) (2024)
- [BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agent](https://mlsanity.com/paper/2508.06600/) (2025)
- [Deep Entity Matching with Pre-Trained Language Models](https://mlsanity.com/paper/2004.00584/) (2020)
- [Neural Networks for Entity Matching: A Survey](https://mlsanity.com/paper/2010.11075/) (2020)
- [Entity Matching using Large Language Models](https://mlsanity.com/paper/2310.11244/) (2023)
- [Using ChatGPT for Entity Matching](https://mlsanity.com/paper/2305.03423/) (2023)
- [WDC Products: A Multi-Dimensional Entity Matching Benchmark](https://mlsanity.com/paper/2301.09521/) (2023)
- [Machamp: A Generalized Entity Matching Benchmark](https://mlsanity.com/paper/2106.08455/) (2021)
- [Pre-trained Embeddings for Entity Resolution: An Experimental Analysis [Experiment, Analysis & Benchmark]](https://mlsanity.com/paper/2304.12329/) (2023)
- [Bridging the Gap between Reality and Ideality of Entity Matching: A Revisiting and Benchmark Re-Construction](https://mlsanity.com/paper/2205.05889/) (2022)
- [Link Prediction using Graph Neural Networks for Master Data Management](https://mlsanity.com/paper/2003.04732/) (2020)
- [Alaska: A Flexible Benchmark for Data Integration Tasks](https://mlsanity.com/paper/2101.11259/) (2021)
- [CompanyKG: A Large-Scale Heterogeneous Graph for Company Similarity Quantification](https://mlsanity.com/paper/2306.10649/) (2023)
- [Evaluating Blocking Biases in Entity Matching](https://mlsanity.com/paper/2409.16410/) (2024)
- [How to reduce the search space of Entity Resolution: with Blocking or Nearest Neighbor search?](https://mlsanity.com/paper/2202.12521/) (2022)
- [Graph-based hierarchical record clustering for unsupervised entity resolution](https://mlsanity.com/paper/2112.06331/) (2021)
- [CompanyName2Vec: Company Entity Matching Based on Job Ads](https://mlsanity.com/paper/2201.04687/) (2022)
- [Transformer-based Entity Legal Form Classification](https://mlsanity.com/paper/2310.12766/) (2023)
- [xLP: Explainable Link Prediction for Master Data Management](https://mlsanity.com/paper/2403.09806/) (2024)
- [MaDI-Bench: An End-to-End Data Integration Benchmark](https://mlsanity.com/paper/2606.30371/) (2026)

## Abstract (arXiv)

Deciding whether two supplier records belong to the same corporate family is a prerequisite for spend consolidation, credit exposure aggregation and sanctions screening. It is usually treated as entity matching, but the tasks differ: a family link connects records that are deliberately different entities, and the evidence often appears in neither record. We introduce CorpFam, a public benchmark of 54,864 candidate pairs over 10,307 corporate families, derived from 6,638,350 US federal award records in which every supplier self-reports its ultimate parent to a government registry. Pairs are stratified by name visibility: whether the names are identical after normalisation, share a distinctive token, or share none. Because strata have positive rates from 10.2% to 97.3%, we report per-stratum recall, base-rate invariant, rather than F1, which is not. The strongest of 5 matchers recovers 100.0% of identical pairs and 4.2% of invisible ones; no method exceeds 4.7% on the latter. The failure begins before matching. Blocking decides which pairs a matcher sees, and we evaluate 7 schemes spanning phonetic keys, attribute keys that ignore the name, and semantic nearest neighbours. None reaches three percent on invisible pairs, and their union recovers 6.8%. 93.2% of these links never enter the candidate set, so no matching-stage improvement can reach them. The links are real: against SEC Exhibit 21 subsidiary schedules, which share no provenance with procurement registration, 64.2% of invisible links are corroborated, against 0.16% under permuted parents and 0.41% against the same parent's wrong exhibit: two unrelated nulls agreeing to within 0.25 points. Corporate-family resolution is a retrieval problem misfiled as a matching problem; the intervention point is candidate generation, not ranking. The benchmark, adjudication log, and code reproducing every number are released.