Good Faith Purchase and the Geography of Title

Two systems, and the whole difference between them

Common law starts from nemo dat quod non habet: a thief has no title and can convey none, so a buyer in good faith from a thief takes nothing, however honest and however much was paid. Civil law systems start from the opposite instinct, protecting the security of commercial transactions, and give a good faith acquirer of a movable a route to ownership even where the seller had none. Every difficult restitution case in the last thirty years is a variation on which of those two starting points the court reaches for.

Germany illustrates that the systems are not as far apart as the slogans suggest, because the civil law route is usually conditional. Section 935(1) of the Buergerliches Gesetzbuch excludes good faith acquisition of stolen property, so the German starting point looks like the English one. Three qualifications then reverse much of it. A person who acquires a stolen work in good faith and keeps possession for at least ten years, remaining in good faith throughout, acquires valid title by prescription. Section 935(2) transfers title to a good faith acquirer where the work is sold at public auction. And under section 197(1) number 1, the owner's claim for return is time-barred thirty years from the day of loss, so that even a bad faith possessor, who can never acquire title, becomes unassailable in practice once the period has run.

The practitioner's conclusion follows directly. A German museum defending a 1938 loss does not have to argue that it owns the object. It only has to point at the calendar. This is exactly why the German mechanism for Nazi-era claims had to be built outside the ordinary courts, and why its awards rest on an agreed assessment framework rather than on the civil code.

Spain, and the three years that ended a twenty-year case

In 1939 Lilly Neubauer was compelled to sell a Camille Pissarro to a Berlin dealer for 900 Reichsmarks in order to obtain an exit visa to England. After the war the Federal Republic settled with her for 120,000 Deutschmarks, a sum the Ninth Circuit put at roughly USD 250,000 in current terms, a conversion that is itself date-dependent and should be treated as the court's derived figure rather than a fact about 1958. The painting moved through the trade, and Baron Hans Heinrich Thyssen-Bornemisza bought it in 1976 for USD 275,000. In 1993 Spain acquired the collection, for USD 350 million, through the Thyssen-Bornemisza Collection Foundation, which put the picture on display in Madrid. Claude Cassirer, Lilly's heir and a California resident, located it in 2000 and sued in 2005.

Twenty years of litigation reduced to a single question: which law decides who owns it. In 2022 the United States Supreme Court held that a federal court hearing a Foreign Sovereign Immunities Act case must apply the forum state's choice-of-law rules, which sent the case back to the Ninth Circuit to redo the analysis under California's governmental interest test. On 9 January 2024 that court found a true conflict between Spanish and Californian property law, held that Spain's interests would be more impaired by the application of California law than California's by the application of Spanish law, and applied Spanish law. Under article 1955 of the Spanish Civil Code, ownership of a movable vests after three years of uninterrupted good faith possession, or six years without good faith. The court found that the foundation had not known the painting was stolen when it acquired it, and gave judgment for the foundation.

Judge Callahan concurred and said plainly that Spain, having reaffirmed its commitment to the Washington Principles, should have relinquished the painting voluntarily, and that the court was nonetheless bound to apply the law. That is the clearest judicial statement available of the gap between a political commitment and a rule of decision. Had the same facts been litigated under Californian law, the outcome would have been the reverse, because a thief cannot pass good title there.

The American clock: demand, refusal, and then a statute that stopped it

New York, where most of the relevant American transactions happen, follows the demand and refusal rule. In Solomon R. Guggenheim Foundation v. Lubell the Court of Appeals dealt with a Chagall gouache stolen from the museum by a mailroom employee in the late 1960s and bought by the Lubells from a gallery in 1967 for USD 17,000. The museum learned where it was in 1985, demanded its return in January 1986, was refused, and sued in September 1987. The court held that against a good faith purchaser possession is not wrongful until demand has been made and refused, so the limitation period does not begin before then. It declined to impose a duty of reasonable diligence on the dispossessed owner, observing that owners of stolen property should not all be expected to behave in the same way, and it criticised the Second Circuit for having imposed such a duty without certifying the question. It preserved laches as a defence, so the possessor's prejudice from delay remained live at trial.

Congress then went further. The Holocaust Expropriated Art Recovery Act of 2016 gave claimants six years from actual discovery of both the identity and location of the work and the claimant's possessory interest in it, for losses in the covered period from 1 January 1933 to 31 December 1945, and made that period supersede all other statutes of limitations. It carried an exception for claimants who had the necessary knowledge after 1 January 1999 and let six years pass, and it was to expire on 1 January 2027.

On 13 April 2026 the successor statute, the Holocaust Expropriated Art Recovery Act of 2025, was signed into law, described by the White House as permanently extending and expanding judicial authority under the 2016 Act. Practitioner analysis of the enacted text reports that the sunset is removed and that courts are precluded from relying on laches, the act of state doctrine, forum non conveniens and international comity, that the statute addresses foreign acquisitive prescription of the kind applied in Cassirer, and that it reaches claims pending at enactment at district and appellate level. The Association of Art Museum Directors opposed the expansion, warning that removing traditional defences sets a precedent and carries foreign relations consequences. No appellate court has yet construed the new Act, and how far the preclusion of non-merits defences extends is untested.

The treaties do less than their reputation suggests

The 1970 UNESCO Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property was adopted on 14 November 1970 and entered into force on 24 April 1972. Its best known operative provision, article 7(b), obliges states to prevent museums from acquiring cultural property illegally removed from another state party after entry into force, and to take appropriate steps to recover and return stolen inventoried property at the request of the state of origin. It then adds a sentence that surprises most people who cite the convention approvingly: the requesting state shall pay just compensation to an innocent purchaser or to a person who has valid title. The United States entered its ratification on the footing that the convention is neither self-executing nor retroactive, and article 13(d) is directed at objects removed after entry into force for the states concerned.

The UNIDROIT Convention on Stolen or Illegally Exported Cultural Objects, adopted in Rome on 24 June 1995, is the sharper instrument and the less widely ratified. Article 3 states flatly that the possessor of a cultural object which has been stolen shall return it. Claims must be brought within three years of the claimant knowing the location of the object and the identity of its possessor, and in any case within fifty years of the theft, extended for objects from public collections and for sacred or communally important objects to three years from knowledge with no absolute outer limit, and states may declare a seventy-five year period or longer. Article 4 entitles a possessor who neither knew nor ought reasonably to have known that the object was stolen, and who can prove due diligence, to fair and reasonable compensation on restitution. The convention expressly does not legitimise transactions concluded before its entry into force, and it is not retroactive.

Neither treaty reaches a 1939 forced sale in Berlin. Both are prospective, both operate between states rather than between a claimant and a possessor, and both place a payment obligation on the recovering side. They are import controls with a restitution clause attached, and they belong in a discussion of antiquities and export offences far more than in a discussion of Holocaust-era title.

What a buyer is actually pricing

The practical consequence of all this is that an object's legal history is a function of its geography. A picture that spent 1970 to 2000 in Madrid is a different asset from the same picture in the same condition that spent those years in New York, and the difference is not detectable by looking at it. Two questions therefore belong at the front of any diligence file. Where has the object physically been, and in whose hands, at each moment when title might have passed. And which forum would a claimant realistically use, since the forum's choice-of-law rules will select the substantive law, as the Supreme Court's 2022 ruling in Cassirer made explicit.

Documented diligence is worth buying for a second reason that is often missed. In the civil law systems it is what makes good faith provable, and good faith is the condition on which German prescription and Spanish article 1955 both depend. Under UNIDROIT article 4 it is what converts a possessor who must give the object back into a possessor who must be paid for it. In a nemo dat jurisdiction it does not save the title, and the buyer's remedy lies against the seller on the contract.

One further shift deserves watching. The American legislative direction since 2016 has been to strip out the doctrines that decide cases without reaching the facts, and the 2025 Act pushes that considerably further by naming laches, comity, act of state and forum non conveniens. If that survives construction, the United States becomes a forum in which Holocaust-era claims are decided on provenance evidence rather than on procedure, at the same time as European claims are moving into arbitral tribunals with express presumptions. Both trends point the same way: the burden is shifting back onto the holder to explain the gap.