# Stolen Art Databases and the Limits of a Clean Search
> A clean search proves an object was not matched against one organisation's records on one day. It is not a title check, and a register of thefts cannot record a sale made under duress.
> The Study - 2026-08-02
> Canonical HTML: https://iconotheca.com/study/stolen-art-databases-and-the-limits-of-a-clean-search/
## Key points
- The Art Loss Register is a private company with more than 700,000 items and about 450,000 searches a year, populated by police, insurers, the trade and the public, which makes its coverage a map of what was reported and insured.
- Interpol's Stolen Works of Art database holds almost 57,000 items and accepts entries only from National Central Bureaus and named partner organisations, and only for fully identifiable objects.
- A theft register structurally cannot record a forced sale, which is the characteristic Nazi-era loss; the German Lost Art Database exists precisely because that gap could not be closed by police data.
- The German Court of Arbitration decides Nazi-era claims on indirect and circumstantial evidence to a standard of high probability, with a presumption that a sale loss resulted from persecution, so the absence of a database hit proves very little to the tribunal now hearing these cases.
## Three registers that are not doing the same job
## A theft register cannot record a forced sale
## Registration is voluntary, and the gaps are systematic
## What a clean certificate says, and what it is worth
## The tribunal now hearing these claims starts from the gap
A clean database search is a statement about one organisation's records on one day, and buyers routinely read it as a statement about the object. The losses that generate the largest and most durable claims were never thefts in the sense any of these registers can record, which is why the German Lost Art Database had to invent a report for holders who merely suspect a gap, and why the tribunal now deciding German claims begins from the presumption that the gap is real and that a sale loss followed persecution. Treat the search as the cheapest and least informative step in the file, complete it, keep the certificate for the good faith it evidences, and then do the archival work that actually establishes where the object was.
## Questions
**Does an Art Loss Register certificate give a buyer good title?**
No. It records that the object was not matched against the records held by that organisation at the date of the search, using the identifiers supplied. Title is determined by the law of the relevant jurisdiction, and a search result is evidence of diligence rather than a warranty. Its practical value is in proving good faith, which matters in civil law systems and under UNIDROIT article 4.
**Why would a Nazi-era loss not appear in Interpol's database?**
Interpol accepts entries only from National Central Bureaus and named partner organisations, and only for fully identifiable objects. Losses through forced sale, confiscation under colour of law or emigration levies were never reported as thefts to any police force, and much of the material was never photographed to a standard that would support an entry.
**What does an entry in the German Lost Art Database prove?**
By itself, nothing about title. A search request records that someone is looking for an object; a found-object report records that a holder considers its 1933 to 1945 provenance to indicate or suggest a seizure. The foundation directs users to a separate legal FAQ rather than attaching legal consequences to publication, and both entry types are best read as research signals.
**How many databases should a serious diligence check cover?**
More than one, and none of them is sufficient. The theft registers, the Einsatzstab Reichsleiter Rosenberg database for occupied France and Belgium, the Central Registry, the German Lost Art Database and museum-level provenance publications cover different periods, geographies and object types, do not share identifiers and do not cross-query. The American Alliance of Museums portal, in use from 2003 to 2024, is no longer maintained.
## Sources
- [1] Art Loss Register, home page (more than 700,000 items; approximately 450,000 searches annually; who may register).
- [2] Art Loss Register, 'About us' (established London 1990; origin in the IFAR art theft archive of 1976; company registration number 02455350; services free to law enforcement and nation states).
- [3] INTERPOL, 'Stolen Works of Art Database' (almost 57,000 items; entries via National Central Bureaus, UNESCO, ICOM and ICCROM; fully identifiable objects; authorised user access and image matching).
- [4] German Lost Art Foundation, 'Lost Art Database' (operator, purpose, search requests and found-object reports, coverage of Nazi-looted property and wartime losses, reference to a separate legal FAQ).
- [5] Claims Conference, Holocaust Looted Art and Cultural Property Initiative, 'Research Databases' (ERR database, Central Registry, Lost Art restitution figure as of January 2022, NEPIP 2003-2024, Legacy Explorer launched December 2025).
- [6] Center for Art Law, '25 Years of the Washington Principles' (Nazi-Era Provenance Internet Portal: approximately 30,000 objects across 179 museums).
- [7] German Lost Art Foundation, 'Court of Arbitration for Nazi-looted Cultural Property launched' (operative from 1 December 2025; binding awards; reduced burden of proof and presumptive rules).
- [8] Koch Boes Rechtsanwaelte, 'Germany's New Arbitration Court for Nazi-Looted Cultural Property: Key Procedural Features' (indirect and circumstantial evidence, prima facie reasoning, high degree of probability, presumption that a sale loss resulted from persecution).
- [9] Cassirer v. Thyssen-Bornemisza Collection Foundation, US Court of Appeals for the Ninth Circuit, 9 January 2024, FindLaw (1939 forced sale for 900 Reichsmarks; post-war settlement with the Federal Republic).
- [10] UNIDROIT Convention on Stolen or Illegally Exported Cultural Objects, Rome, 24 June 1995 (article 4: compensation for a possessor who neither knew nor ought to have known and can prove due diligence).