Skip to content

Note no. 3

The nail is gone

By

This note: 6,559 words, counted in pebbles

Decision

Yes on 186 million euros of works if every loan states before departure where it returns and who gets paid; otherwise only the 132 million whose return is written

Published
5 October 2026
Review due
11 October 2027
  1. AYes, subject to information(the note's choice)
  2. BYes for works returning to Kyiv, no for the rest
  3. CYes, on the engine's terms

The diagram · the loan that does not come back

The lenders

four Crimean museums

Amsterdam, February 2014

the Allard Pierson museum shows 565 loaned objects

The lender's location changes sovereignty

March 2014

Russia annexes Crimea while the objects are in the Netherlands

Kyiv, November 2023

after the district court (2016), the court of appeal (2021) and the Supreme Court (2023)

A loan of a few months became a custody of nine years, and the original nail was never found.

This note became a decision on October 1, 2026. It builds the file, sets out what a rating grid would have returned, and proposed three complete decisions. Option A, recommended by the author, was adopted: it appears in section 6 with its reason, and the other two stay on the record.

Reviewed on October 1, 2026, before publication, at senior underwriter level. Four additions, no substance removed: the exposure bounded lot by lot (section 1), the wording of the sanctions clause and what French law has made of it (section 3), the accumulation and the scenario that would hurt a portfolio (section 3), and the dated thresholds that would change the decision before the works leave (section 6). The decision is unchanged; the added facts are checked in section 9, facts 37 to 42.

Every fact was checked on September 27, 2026, or removed. The first draft was written from memory. The list in section 9 gives, for each fact, its exact reference and what became of it. Web access was partial: a search engine answered, but the pages themselves (Légifrance, EUR-Lex, the courts' and museums' sites) were blocked from the reviewing machine. "Verified" therefore means: the exact reference is identified, and the fact is confirmed by its summary or by two concurring sources; no ruling was read in full.

Proposed review date: October 11, 2027. On that date the note will be read again against what actually happened, and the review will be published at the top, whether it proves the decision right or wrong.

The cases are public; the insured is fabricated. The court rulings, statutes and state measures cited here are real and dated. The foundation seeking cover does not exist: it combines configurations common in the exhibition market. The note says what an underwriter would decide; it tells no one what to buy or where, and it makes no recommendation on coverage, limits, deductibles or premium.

0. Why this case

Notes No. 1 and No. 2 were cyber. This one opens the third line added to the site's charter on September 26, 2026: fine art. It stays on the junction: an act of state that, without touching a single work, changes what a fine art policy actually covers.

Nail-to-nail cover is the simplest contract in the market: it follows the work from the lender's wall to its return to that same wall. Everything rests on an assumption no one writes down because it seems obvious: the nail will still be there at the end of the loan, and it will still belong to the same lender. Two public cases have shown that a state can undo that assumption while a loan is under way, once through annexation and once through sanctions. Neither damaged a single object.

1. The risk

The public record

The Crimean gold in Amsterdam. In early February 2014, the Allard Pierson Museum of the University of Amsterdam opened an exhibition on ancient Crimea, made up mostly of 565 objects lent by four Crimean museums, and in part of objects lent by museums in the rest of Ukraine. In March 2014, Russia annexed Crimea while the objects were in the Netherlands. The borrowing museum faced two conflicting demands for return: one from the Crimean museums, which had lent the objects and wanted them back, and one from the Ukrainian state, which claimed them as national heritage. In September 2014 it returned the objects from museums in the rest of Ukraine, and held the Crimean ones pending a court ruling. The Amsterdam District Court held on December 14, 2016, that the objects should go to Ukraine; the Amsterdam Court of Appeal affirmed on October 26, 2021; and the Supreme Court of the Netherlands dismissed the Crimean museums' appeal on June 9, 2023. The objects arrived in Kyiv, at the National Museum of the History of Ukraine, in November 2023. A loan of a few months had become custody of nine years, and the original nail was never found again: the objects did not go back to the museums that had lent them.

The Morozov collection in Paris. Starting on September 22, 2021, the Fondation Louis Vuitton showed some two hundred works from the collection of the Morozov brothers, lent mostly by Russian state museums and in small part by private collectors. As French law has allowed since 1994 (Article 61 of Law No. 94-679 of August 8, 1994), a ministerial order had declared the works lent by the museums immune from seizure during their stay in France. That article only covers property lent by a foreign power, a public body or a foreign cultural institution: works lent by private owners were not eligible. Russia's invasion of Ukraine came while the show was running; it was extended to April 3, 2022. On February 28, 2022, the European Union added Petr Aven to the list of persons whose assets are frozen under Regulation (EU) No 269/2014, through Implementing Regulation (EU) 2022/336. When the show closed, the works from state museums were sent back to Russia; three works stayed in France, two of which show exactly what this file fears. On April 9, 2022, France held a self-portrait by Pyotr Konchalovsky belonging to Petr Aven, for as long as the freeze lasts. A portrait of Margarita Morozova belonging to the art museum of Dnipro, in Ukraine, stayed at the Fondation, at the request of the Ukrainian authorities, until it could be returned safely. The private lender had no immunity to set against the freeze, and the Ukrainian museum's work could not go home.

The two cases do not say the same thing, which is why this note takes them together. The first shows that a change of sovereignty over the lender's location makes return impossible without a court ruling. The second shows that a measure aimed at the lender as a person turns the borrower into the custodian of a frozen asset, and that a war at the place of return is enough to keep a Ukrainian museum's work in France.

The insured, fabricated

A private modern art foundation with its headquarters and exhibition space in Lyon. It is preparing, for spring 2027, from March 12 to July 11, an exhibition of 74 works of the Ukrainian avant-garde from the 1910s to the 1930s.

  • 61 works are lent by three Ukrainian public museums. Two are in Kyiv and lend 25 works. The third is a regional museum in a city in the southeast of the country, about sixty kilometers from the front line; its storerooms were evacuated in 2022 and its works are now kept in Kyiv, but the loan agreement for its 36 works names the regional museum as lender and its home city as the place of return.
  • 13 works are lent by two private collectors. One is an individual living in Switzerland. The other is a holding company registered in Cyprus, whose list of beneficial owners the foundation has not received.

Total agreed value is 186 million euros, of which 41 million covers the thirteen private works. The outbound leg runs by climate-controlled truck, with couriers, from Kyiv to the Polish border, then by road to Lyon. An order of immunity from seizure has been requested but not yet published. The Lyon facility report is good, the foundation has had no losses on its six previous exhibitions, and the loan agreements follow the template used by major European museums.

The request is for nail-to-nail cover on all loans, from the storerooms in Kyiv until return.

The exposure, bounded

The file reads as four lots, not one. The amounts belong to the fabricated insured; they reconcile with the totals in sections 1 and 5.

  1. The two Kyiv museums: 25 works, 114 million euros, returning to Kyiv. Exposure bounded by the contract period, about five months.
  2. The regional museum: 36 works, 31 million euros, returning to a city sixty kilometres from the front. Bounded in value, unbounded in time: if return becomes impossible, the policy runs until it becomes possible again or a court names another place. Amsterdam gives the order of magnitude, nine years instead of a few months: about twenty times the period priced, on the same sum insured.
  3. The Swiss collector: 6 works, 18 million euros, beneficiary known. Ordinary exposure.
  4. The Cypriot holding company: 7 works, 23 million euros, beneficial owners unknown. Bounded in value, but payment of a claim is not certain to be possible: if a beneficial owner is designated, the indemnity cannot be paid to them for as long as the measure lasts (section 3, the wording that matters).

The accumulation at one location is 186 million euros, in a single gallery in Lyon, for four months: the ordinary maximum loss of an exhibition, fire or water damage, and nothing specific to this file. The accumulation in transit depends on the number of trucks and the value loaded on each, which the file does not state; I ask for it with the route. What is specific to this file is the 54 million euros of lots 2 and 4, one of which may run without end and the other may not be payable.

2. What is missing

The actual place of return for each work, and who can change it. The regional museum's loan agreement says "return to lender." If the lender's city is occupied during the loan, who decides where the work goes: the museum, Ukraine's Ministry of Culture, or a judge? As long as the agreement does not say, the policy has no end date. This is the most important piece of the file.

The beneficial owners of the Cypriot holding company. Without them, I do not know whether one of the lenders already is, or could become during the loan, a person whose assets are frozen. If so, an indemnity for damage to its works could not be paid to it, and the work itself could not be handed back.

The published immunity order. I want it in the Journal officiel before the works arrive, with the exact list of works it covers. It protects against creditors and private claims; it protects against no measure taken by the French state, and it can only cover the museums' 61 works: the 13 private works are not eligible.

The return route, and who bears war risk on the road. The return to Kyiv crosses a country at war. As a rule the market does not cover war risk on land for goods in transit (the so-called Waterborne Agreement rule), and fine art policies generally exclude war, a practice I have not checked here against a policy wording. I need to know who carries that risk on the final leg: the lender, the Ukrainian state, a public indemnity, or no one.

What the loan agreements provide if return becomes unlawful or impossible. Storage at a designated location, for how long, at what cost, and who pays. This is what was missing in Amsterdam, and it left the borrowing museum holding the objects for nine years.

3. The signals

Retained

A nail-to-nail policy has no end if the nail is gone. This is the heart of the note. An exhibition policy is priced on a duration, here about five months of transit and display. If return becomes impossible, the work stays in custody with the borrower or in a warehouse, and the policy runs until it goes home. The exposure is not damage; it is time, and that time is no longer set by the contract but by a war or a court. Amsterdam put a number on it: nine years instead of a few months.

The lender, its city and its jurisdiction are three different things. The regional museum is a sound public lender whose works are now in Kyiv. But its contractual address is in a city near the front. In Crimea, the lenders still existed after the annexation, still had their buildings, and wanted their objects back: that is exactly what produced two conflicting claims. The risk is not that the lender disappears; it is that the lender changes state.

An opaque private lender is a sanctions risk, not an integrity risk. I do not suspect the Cypriot holding company. I note that I cannot tell whether it will escape a freeze during a four-month loan, and that in 2022 a listing came in the middle of an exhibition, against a lender subject to no EU asset freeze when the show opened.

Immunity from seizure reassures the grid, and it does not cover what matters here. It is built for an heir's claim or a creditor of a lending state. Both public cases in this note are state measures, or a deadlock only a court could break: the order would have changed nothing in Amsterdam, and it prevented nothing in Paris. It also covers only public loans: the 41 million euros of private works in this file are exactly the share it leaves out, and the share exposed to a freeze.

Set aside

The lenders' nationality, on its own. Ukrainian museums are not a risk because they are Ukrainian. They become one through the place of return named in the agreement. The two Kyiv museums, whose works return to Kyiv, do not raise the question the regional museum raises. I look at the return address, not the passport.

War as a risk of damage during the exhibition. The works will be in Lyon, in a building with a good facility report. The war risk that concerns me is not a missile on the gallery; it is the last leg of the return, and it is a question of routing and exclusions, not of display.

The 2022 looting in Kherson. It is documented, it is serious, and it involved works left in an occupied city. Here the works have been evacuated. I keep it in mind as proof that returning to an exposed city is no formality; I do not put it in the file as a risk of the Lyon exhibition.

The value of the works. 186 million euros is an average exhibition for loans from national museums. Value sets the amount at stake; it says nothing about duration.

The wording that matters: the sanctions clause

The grid ticks "sanctions" as an exclusion. The most widely used market wording, LMA 3100 of September 15, 2010, says something else: "No (re)insurer shall be deemed to provide cover and no (re)insurer shall be liable to pay any claim or provide any benefit hereunder to the extent that the provision of such cover, payment of such claim or provision of such benefit would expose that (re)insurer to any sanction, prohibition or restriction under United Nations resolutions or the trade or economic sanctions, laws or regulations of the European Union, United Kingdom or United States of America." Three phrases carry the whole file.

  1. "to the extent that": the clause does not remove cover, it suspends it for as long as payment would expose the insurer. The Commercial Court held so on October 12, 2018 (Mamancochet Mining v Aegis, [2018] EWHC 2643 (Comm)), on a clause of the same make: liability is suspended, not extinguished. During a freeze the policy therefore keeps running: that is exactly lot 4.
  2. "would expose": the insurer must show that payment would be prohibited, not merely risky. The prohibition exists without the clause: Article 2(2) of Regulation (EU) No 269/2014 forbids making funds available, directly or indirectly, to any person listed in its Annex I. The clause does not create the inability to pay; it records it.
  3. The law that reads the clause. The foundation is in Lyon, and its policy will most likely be governed by French law. On June 21, 2022, the Paris Court of Appeal characterised a sanctions clause as an exclusion, in a directors' and officers' policy, and set it aside for failing to state with enough precision which sanctions it covered, when an exclusion must be "formal and limited" (Article L. 113-1 of the French Insurance Code). The Lloyd's Market Association answered on October 5, 2023 with LMA 3100A, the same text under a title that no longer says "exclusion", and LMA 3200, drafted for civil law jurisdictions. Under French law, what the grid ticks is therefore not a given; what is a given is the 2014 regulation, which prohibits payment whatever becomes of the clause.

Consequence for the decision: the protection for lot 4 is not in the insurance contract, it lies in knowing the beneficial owners before the works leave. That is condition 2 of option A, and this wording makes it necessary rather than merely prudent.

The accumulation, and the scenario that would hurt

A book of exhibitions does not accumulate this file's damage, it accumulates its durations. Since 2022, Ukrainian museums have lent across Europe to keep their collections visible. A fine art underwriter who writes several of these exhibitions carries the same peril on each: the place of return.

The scenario that would hurt fits in one sentence: the city of a lending museum falls under occupation while three of its groups of works are on show in three European countries. No work is touched; three policies priced for a few months run together for years, on three sums insured, with three borrowers asking whom to return to. The portfolio's catastrophe model does not see it, because it counts damage by location, and here the loss is a calendar.

The second scenario is a wave of designations: an update of the freeze lists that reaches several lending collectors on the same date, as on February 28, 2022 (fact 22). It accumulates no losses; it accumulates works held that can neither be returned nor indemnified.

What the underwriter does with this accumulation is not for this note, which sets no sum insured: it says only that the accumulation in this file is measured by place of return, not by place of display.

4. What an engine would have decided

The grid

An ordinary appetite grid for temporary exhibitions would have read this:

borrower             private foundation, 6 shows          compliant
facility             favorable facility report            compliant
lenders              public museums (82%)                 favorable
agreed value         EUR 186M                             in range
transit              climate-controlled trucks, couriers  compliant
duration             about 5 months nail to nail          standard
immunity             requested (private works             favorable
                     not distinguished)
loss history         none over six shows                  favorable
------------------------------------------------------------
result               accept, standard exclusions:
                     war, confiscation, sanctions
borrower
private foundation, 6 shows
compliant
facility
favorable facility report
compliant
lenders
public museums (82%)
favorable
agreed value
EUR 186M
in range
transit
climate-controlled trucks, couriers
compliant
duration
about 5 months nail to nail
standard
immunity
requested (private works not distinguished)
favorable
loss history
none over six shows
favorable
result
accept, standard exclusions: war, confiscation, sanctions

The engine accepts, and it has good reasons to. On the data it reads, this is a clean file: institutional lenders, a well-run venue, museum-grade transit, and a short duration.

Where judgment departs, and why

In one place only, and it is precise: the line "duration: about 5 months nail to nail." The engine reads it as a given. I read it as a duration that depends, for 61 of 74 works, on sovereignty over a city and, for 13 others, on a sanctions list, two things that can change during the loan and already have, in comparable loans, in 2014 and in 2022.

Three reasons make this more than a scruple:

  1. The engine believes the exclusions close off the risk. War, confiscation, sanctions: it ticks them and treats the exposure as bounded. None of those exclusions says when the policy ends. In Amsterdam, no object was destroyed, confiscated or seized; the objects stayed in custody, insurable and needing insurance, for nine years. An exclusion removes a loss; it does not give back a duration.
  2. The engine reads "public lender" as a favorable signal. It is, for solvency, conservation and contract discipline. It turns into the opposite when the lender is a state institution in an area whose sovereignty is contested: then a state is the claimant, and sometimes two.
  3. The engine does not see who would receive the indemnity. The policy pays the lender. If the lender is frozen on the day of the loss, the sanctions clause the grid ticks forbids payment: the damage is real, the work is harmed, and no one can be indemnified while the measure lasts.

What the review will settle. On October 11, 2027, we will know whether the front line has moved toward a city that hosts a lending museum of this kind, whether new art lenders have been added to European lists, and whether loans of Ukrainian works in Europe went home on schedule. If nothing happened, the grid was right for the period, and judgment will have to say whether its departure was justified caution or wasted cost.

5. The three possible decisions

Each is complete and defensible. None touches the terms of the insurance contract: they say whether to accept, for which works, and on what information.

Option AYes, subject to information

Accept the whole exhibition, and turn the uncertain duration into a written one. Before the works leave, the foundation provides:

  1. for each loan, the place of return and the authority that can change it, written into the loan agreement, with a designated fallback location if return becomes unlawful or impossible, and the person who decides;
  2. the list of beneficial owners of the Cypriot holding company, and screening of every private lender against European, British and American lists, at signing and again fifteen days before return;
  3. *the immunity order published in the Journal officiel***, covering the museums' 61 works, the only ones the law allows it to list;
  4. the return route, and a written statement of who bears war risk on the final leg: the lender, the Ukrainian state, or a public indemnity.

If any one is missing, the decision becomes option B. These are not recommendations to the insured; they are the facts without which I do not know when my policy ends, or whom it would pay.

Option BYes for works returning to Kyiv, no for the rest

Accept the works of the two Kyiv museums and the Swiss collector, whose return and beneficiary are known, 132 million euros of the 186, and decline the works of the regional museum and the holding company until their place of return and their beneficial owners are in writing. It is the easiest decision to defend before a committee. It carries a real cost for the insured, which must find cover elsewhere for 54 million euros of works, or leave them out of the show.

Option CYes, on the engine's terms

Accept as the grid does, with standard exclusions. The lenders are public and sound, the works have been evacuated since 2022 and are kept in Kyiv, the immunity order will be published for the museums' works, and the Amsterdam case came from a lightning annexation no exhibition could have foreseen. Declining or conditioning Ukrainian loans because Ukraine is at war deprives these museums of the exhibitions that make their collections known when they need it most, over a risk the market already handles through its exclusions.

6. What would prove each option wrong

A would be wrong if the conditions were met on paper and useless in fact: a written fallback location that no one can enforce once two authorities claim the work, as in Amsterdam, where it took three courts to say to whom the objects should go. The conditions would then have measured drafting, not an outcome. It would also be wrong if the foundation could not obtain the holding company's beneficial owners: that would mean the lender does not want to disclose them, and option A should have been B from the start.

B would be wrong if, by October 11, 2027, every loan of Ukrainian works in Europe had gone home on time, the front line had reached no city hosting a lending museum, and no private art lender had been listed in Europe in the middle of a loan. We would have charged an exhibition, and museums at war, the cost of a risk that did not materialize.

C would be wrong if, during the period, a city hosting a lending museum fell under occupation while some of its works were on show abroad; or if a private art lender in Europe were frozen in the middle of a loan; or if a European borrower had to keep lent works past their due date because of a state measure or pending a court ruling. In each case, the policy the grid treated as bounded would have kept running over a duration no one priced, or could not have paid anyone.

What would change my mind before the works leave

Option A is a conditional decision; here, dated, are the facts that would tip it before the works leave Kyiv, planned for early February 2027.

  1. By January 15, 2027, the Cypriot holding company has not disclosed its beneficial owners: lot 4 (23 million euros) moves to option B, declined, without further discussion.
  2. A beneficial owner, or a relative the list names as an associate, appears on an EU, UK or US freeze list at the screening on signature or the one fifteen days before return: lot 4 is declined, not conditioned, because no wording makes payable an indemnity the regulation prohibits.
  3. By January 15, 2027, the regional museum's loan agreement names neither Kyiv nor a fallback place, nor the authority that decides: lot 2 (31 million euros) moves to option B.
  4. The regional museum's city falls under occupation, or under ordered evacuation, before departure: lot 2 is declined, because the place of return written in the agreement no longer exists.
  5. *The immunity from seizure order is not published in the Journal officiel by the date of departure*: condition 3 is not met, and the decision becomes option B, as section 5 provides.

These five thresholds will be reviewed on October 11, 2027 alongside the five facts in section 7: if any was crossed, the review will say whether the note kept its own rule.

The author's recommendation

Option A. Reason: it is the only one that addresses what the underwriter actually needs to know, when the policy ends and whom it pays, and both are facts the insured can obtain before the works leave, since they sit in the loan agreements and in a beneficial ownership register. B declines works whose only flaw is a missing line in a contract; C treats the exclusions as a boundary when Amsterdam showed that none of them bounds the duration.

What the recommendation is not: a decision. If the decision chooses B or C, that choice prevails, and the published note will carry both, disagreement included.


Decision: Option A, yes subject to information, following the author's recommendation, under the delegation rule of September 29, 2026: absent a contrary decision before publication, the author's recommendation applies.

Reason in one sentence: It bears on the two facts that decide this file, when the policy stops and whom it pays, and the insured can obtain both before the works travel, from the loan agreements and from a register of beneficial owners.

Date of decision: October 1, 2026

Review date: October 11, 2027 (proposed, and open to change)


7. The review

On October 11, 2027, or on the date that has been set. Since the insured is fabricated, the review looks only at public facts. We will look for five verifiable facts, in this order:

  1. A Ukrainian city hosting a public art museum that fell under occupation between October 11, 2026, and the review date. Yes or no, with its name and date.
  2. A listing, on a European, British or American freeze list, of a person known to lend works to exhibitions in Europe, during the same period.
  3. A European borrower forced to keep lent works past their due date, because of a state measure or pending a court ruling, whatever its nationality.
  4. *Immunity orders published in the Journal officiel for works lent by Ukrainian museums*, and how many: a measure of how large these loans have become in France.
  5. A published court ruling on the return of lent works after a change of sovereignty or a sanction, in Europe, after the Dutch Supreme Court's ruling of June 9, 2023.

The review will say which occurred, what it means for the option chosen, and who was right, the grid or judgment. If none of the five occurred, it will say so too, and option C will be credited with the win.

8. Objections

Adversarial review, played by a political risk reviewer in the role of a senior political risk underwriter in the London market. Its three hardest objections, and the answers.

Objection 1 · "Amsterdam will not happen again."

"The annexation of Crimea took a few weeks, almost without fighting, over territory no one thought was contested. It was a black swan. A city sixty kilometers from the front is a risk everyone can see, and the lenders handle it themselves by evacuating their storerooms, as this one did in 2022. You are building a note on the exception."

Answer. The objection is right about probability, and the note does not claim the city will be occupied. It claims something else: that the cost does not scale with the probability, because it is measured in years, not in damage. Evacuation protects the works; it does not change the place of return written into the agreement, and it was that place, not where the objects physically sat, that created the Amsterdam dispute. The visible risk is precisely the one that can be handled with a line in the contract before the works leave. The failure would be to see it and not write it down.

Objection 2 · "The freeze is not your problem."

"If a lender is frozen, the work stays in France, in the borrower's custody, in perfect condition. There is no loss. The sanctions clause does exactly what it should: it prevents payment to a listed person. You are presenting the normal operation of the policy as a flaw."

Answer. It is right about the clause, which is why option A leaves the clause alone. The problem lies elsewhere: during the freeze the policy keeps running, the work stays exposed to the ordinary risks of custody, and the only possible payee of an indemnity cannot be paid. An underwriter who accepts that must know it on the day of acceptance, not discover it midway through the loan, as happened in Paris in 2022. Condition 2 does not try to prevent a freeze; it tries to establish, before the works leave, whether we are knowingly taking custody of a work we may be able neither to return nor to indemnify.

Objection 3 · "You are making Ukrainian museums pay for the war."

"Exhibitions abroad are one of the few ways these museums have to show their collections and raise funds. Every condition you set delays a loan, and your option B cuts out a third of the show. An underwriter who only covers loans from museums at peace is doing politics, not underwriting."

Answer. This is the most serious objection, and it is why the note recommends A, not B. None of A's conditions concerns any quality of Ukrainian museums: three of them are documents one would ask of any lender, and the fourth, war risk on the final leg, applies only to the return to Kyiv. None declines a work; all of them say where the work goes home and to whom. The objection does, however, perform a precise service: it forces the note to state that if the regional museum names Kyiv as the place of return and the ministry as the deciding authority, the gap between A and C becomes almost nil for its 36 works, which is why condition 1 is the first document to read.


9. The facts, checked

Review of September 27, 2026. Thirty-six facts: 25 kept, 8 requalified, 2 removed, 1 added. Web access partial, as stated at the top of the note. Set out as a list, since note pages do not render tables.

  1. Exhibition opened at the Allard Pierson Museum in February 2014. Hoge Raad, press release of June 9, 2023 (February to August 2014). Verified, kept.
  2. Objects lent by four Crimean museums. Hoge Raad, June 9, 2023, ECLI:NL:HR:2023:865. Verified, kept.
  3. "and by a national museum in Kyiv", whose objects were returned. Interfax-Ukraine and press, 2023: objects from museums in the rest of Ukraine returned in September 2014. Requalified: "museums in the rest of Ukraine", returned in September 2014.
  4. "several hundred objects". NL Times, November 27, 2023: 565 objects. Requalified: 565 objects.
  5. Russia annexed Crimea in March 2014. Public record; Rechtbank Amsterdam, December 14, 2016, which treats it as a change of circumstances. Verified, kept.
  6. Two conflicting claims, Crimean museums and Ukrainian state. ECLI:NL:HR:2023:865. Verified, kept.
  7. The museum held the Crimean objects pending a ruling. Allard Pierson; NL Times, November 27, 2023. Verified, kept.
  8. Judgment of December 14, 2016: the objects go to Ukraine. Rechtbank Amsterdam, ECLI:NL:RBAMS:2016:8264, HA ZA 14-1179. Verified, kept.
  9. Appeal ruling of October 26, 2021, affirming. Gerechtshof Amsterdam, ECLI:NL:GHAMS:2021:3201. Verified, kept.
  10. Crimean museums' appeal dismissed on June 9, 2023. Hoge Raad, ECLI:NL:HR:2023:865, No. 22/00270. Verified, kept.
  11. Objects arrived in Kyiv in November 2023. NL Times and The Moscow Times, November 27, 2023: at the National Museum of the History of Ukraine; the day differs between sources (26 or 27). Verified, kept to the month, place added.
  12. "custody of nearly ten years". Allard Pierson: nine years in storage, 2014 to 2023. Requalified: nine years (four occurrences).
  13. The objects did not go back to the lending museums. ECLI:NL:HR:2023:865. Verified, kept.
  14. "three courts to say to whom" (section 6). Facts 8 to 10. Verified, kept.
  15. Morozov exhibition at the Fondation Louis Vuitton from September 2021. Alexis Fournol, lawyer, note of March 2, 2022: opened September 22, 2021. Verified, kept and dated.
  16. About two hundred works. The Art Newspaper, May 25, 2022. Verified, kept.
  17. Lent mostly by Russian state museums, in small part by private owners. The Art Newspaper, April 8 and May 25, 2022. Verified, kept.
  18. Immunity from seizure available since 1994, Article 61 of Law No. 94-679 of August 8, 1994. Légifrance, LEGIARTI000006698549. Verified, kept.
  19. "an order had declared the works immune from seizure". Article 61 only covers property lent by a foreign power, a public body or a foreign cultural institution; press of May 2022: the order did not apply to private lenders. Requalified: museum works only; date and list of the order not checked, so the note does not say whether the Dnipro portrait was on it.
  20. Exhibition extended to April 3, 2022. The Art Newspaper, May 25, 2022. Verified, kept.
  21. Invasion of Ukraine during the show. Public record, February 24, 2022. Verified, kept.
  22. Petr Aven listed on February 28, 2022 under Regulation (EU) No 269/2014. Council Implementing Regulation (EU) 2022/336 of February 28, 2022, OJ L 58. Verified, kept, reference added.
  23. Council Regulation (EU) No 269/2014 of March 17, 2014. EUR-Lex, cited by Regulation 2022/336. Verified, kept.
  24. Aven "subject to no measure" at opening. Regulation 2022/336 adds him to Annex I; earlier US or UK measures not checked. Requalified: "no EU asset freeze".
  25. State museum works sent back to Russia at closing. The Art Newspaper, May 25, 2022. Verified, kept.
  26. "at least one work" held. The Art Newspaper, May 25, 2022: three works. Requalified: three works.
  27. Konchalovsky self-portrait owned by Aven, held for as long as the freeze lasts. The National and CNews, April 9, 2022; decision of April 9, 2022. Verified, kept and dated.
  28. Portrait of Margarita Morozova, Dnipro art museum, kept in France at Ukraine's request until a safe return. The National, April 9, 2022; The Art Newspaper, May 25, 2022. Added, verified.
  29. Third work: Valentin Serov's portrait of Timofei Morozov, from a collection linked to Viatcheslav Kantor. The Art Newspaper, May 25, 2022; legal basis not checked. Verified, kept in the table, not used in the body.
  30. "Immunity did not protect the lender against the borrowing state itself". Aven's work, from a private lender, was not covered by the order (fact 19). removed: the sentence blamed the order for a limit that was not at stake; replaced by what facts 19, 27 and 28 establish.
  31. The market does not as a rule cover war on land for goods in transit. Waterborne Agreement of 1937, described by Atlas Magazine and brokers; text of the agreement not read. Verified at a secondary source, kept and named.
  32. "fine art policies exclude war". No policy wording checked. Requalified: "generally", stated as unchecked.
  33. Looting of the Kherson art museum in 2022. Human Rights Watch, December 20, 2022; press of November 2022 (October 31 to November 4). Verified, kept.
  34. Those works had stayed in an occupied city. Same source. Verified, kept.
  35. Objection 1: annexation "in three weeks, without fighting". Late February to March 18, 2014. Requalified: "a few weeks, almost without fighting".
  36. 2023 return "announced by Ukraine's Ministry of Culture". Only the museum's announcement and the press were found. removed from the sources.
  37. Added on October 1, 2026 (senior review): wording of LMA 3100, "Sanction Limitation and Exclusion Clause", dated September 15, 2010. LMA document blocked; sentence found word for word by exact search, in two insurers' reproductions of the clause and on Law Insider. Verified, added.
  38. Mamancochet Mining v Aegis Managing Agency, [2018] EWHC 2643 (Comm), October 12, 2018: the clause suspends liability, it does not extinguish it. BAILII record; commentaries by Hill Dickinson and Herbert Smith Freehills, October 2018. The clause at issue was the Joint Cargo Committee wording, of the same make as LMA 3100, hence "of the same make". Verified, added.
  39. Article 2(2) of Regulation (EU) No 269/2014: prohibition on making funds available to listed persons. Consolidated text (legislation.gov.uk, Wikisource); European Commission opinion of June 8, 2021 on that article. Verified, added.
  40. Paris Court of Appeal, June 21, 2022 (RG 20/10832): sanctions clause in a directors' and officers' policy characterised as an exclusion and set aside. Le Monde du droit; Noëlle Lenoir law firm; DAC Beachcroft (ground: lack of precision on the sanctions covered). Judgment not read in full; the link to Article L. 113-1 is the one French law draws for exclusions, not a quotation from the judgment. Verified, added.
  41. LMA 3100A and LMA 3200 published on October 5, 2023; 3100A repeats the 3100 text without the word "exclusion" in its title. LMA release of October 5, 2023; Browne Jacobson and Insurance Business, October 2023. Verified, added.
  42. Split of the 186 million euros into four lots (114, 31, 18 and 23). Fabricated insured: not a public fact, a detail of the file, consistent with the 41 million euros of private works (section 1) and the 132 and 54 million of option B. Added.

What the review changes in the decision. Nothing that brings it down; two things sharpen it. Fact 19 corrects an error in the first draft: French immunity can only cover the museums' 61 works, and condition 3 of option A asked for an order "covering all 74 works", which the law does not allow. It is rewritten. And facts 19, 27 and 28 strengthen the note rather than weaken it: in Paris itself, in 2022, a Ukrainian museum's work could not go home and a private lender had no protection against the freeze. Recommendation A holds, and holds better.

Public sources cited. Rechtbank Amsterdam, December 14, 2016, ECLI:NL:RBAMS:2016:8264; Gerechtshof Amsterdam, October 26, 2021, ECLI:NL:GHAMS:2021:3201; Hoge Raad der Nederlanden, June 9, 2023, ECLI:NL:HR:2023:865. Return of the objects to Kyiv in November 2023, announced by the Allard Pierson Museum. Law No. 94-679 of August 8, 1994, Article 61. Council Regulation (EU) No 269/2014 of March 17, 2014; Council Implementing Regulation (EU) 2022/336 of February 28, 2022. Fondation Louis Vuitton, The Morozov Collection: Icons of Modern Art, September 22, 2021 to April 3, 2022. The Art Newspaper, April 8 and May 25, 2022; The National, April 9, 2022. Human Rights Watch, December 20, 2022. Before publication, still to be read in full: the three Dutch rulings, and the immunity order for the Morozov exhibition, whose date was not checked.