Litigation pending at closing forms a third category, and it must be set out clearly because the first two do not accommodate it. It is known, so it falls out of the warranty policy like everything known. But neither is it a question of legal interpretation resting on facts everyone admits, which was the logic of the tax policy: here the facts themselves are disputed, there is an opponent who acts, evidence that can move, and an outcome spread across a range rather than binary. Contingency insurance is the product that takes that risk, and its economics resemble neither of the previous two.
What the insurer buys deserves to be named precisely, because it explains the price and the thinness of the market. It is not buying a legal conclusion, it is buying a distribution of possible outcomes, and above all the future behavior of people who have not yet decided. Will the claimant hold out financially, will a witness be credible at the hearing, will the bench change, will a settlement be offered at the moment the insured is worn down. Underwriting is therefore closer to a bet on a process than to the pricing of an opinion, it requires underwriters who can read a litigation file, and there are few of them.
There is also a difficulty of access to information that exists in no other product in this field. The party that knows the file best is the one seeking cover, which is the textbook case of adverse selection. The insurer's protection is to require the complete litigation file, including counsel's analytical notes and their frank assessment of the weaknesses. Yet handing privileged material to a third party can, in some jurisdictions, waive that privilege as against the opponent. The disclosure the insurer needs therefore risks compromising the defense it is insuring. This is handled through common interest arrangements and through careful staging of what is shared and when, but it is a genuine constraint rather than a formality, and it lengthens placement.
Moral hazard is sharper here than elsewhere, because a fully covered insured has no reason left to defend well or to settle sensibly. Policies answer with a retention, with a share left to the insured above it, and above all with the insurer's control of or approval over the conduct of the case and over any settlement. A classic clause requires the insured to continue defending as though uninsured, and it is not boilerplate: it is the counterpart to the fact that no insurer can plead in the insured's place or examine its witnesses.
One must however see something those protections do not do, and it is this module's contribution. A retention disciplines only the party that bears it AND decides. In a transaction those two roles regularly separate: the seller keeps conduct of the case because it knows the facts and the people, while the retention and residual exposure sit with the buyer, or the reverse. As soon as the person deciding is no longer the one paying the first euros, the mechanism meant to prevent moral hazard stops working, and it stops working silently, without any clause having been breached. That is the most expensive defect in these structures, and it becomes visible only when a settlement is offered.
The product's value in a deal nonetheless remains considerable, and it is measured by the deadlock it clears. Facing pending litigation, the buyer values at the worst case and the seller at the best, each rationally, and the gap cannot be bridged by negotiation since neither is wrong. The policy takes that gap. One structural difference distinguishes it from the tax policy: the outcome being spread over a range rather than binary, cover frequently attaches above a level the parties agree is likely, the insurer taking the tail of the distribution rather than the first euro. The result is that the buyer carries the expected outcome, which it can provision for, and protects itself against what it could not absorb.
That leaves the question of time, which decides more files than people think. A first instance judgment arrives years after closing, an appeal years after that, and that horizon exceeds the survival period of the warranties and sometimes the life of the selling fund. The policy period must therefore outlast the litigation, and cover expiring before judgment is worth nothing whatever its amount. Cover for the appellate stage is a distinct question that arises at underwriting and not when the judgment lands. Finally one must say what no policy of this kind does: it indemnifies money. A risk of being enjoined, of a product being withdrawn or an activity stopped, is not a money risk, and it is handled in the structure of the deal rather than in a cover.
An acquirer buys a manufacturer at an enterprise value of 210 million euros. A competitor claims 45 million against it for patent infringement, proceedings issued in 2023, first instance hearing expected in 2026, with a risk of an injunction on a product line representing 30% of revenue. Counsel considers the defense strong and the expected exposure between 8 and 12 million. The seller, a founder who stays on as a minority shareholder and manager, refuses to indemnify beyond 10 million. A contingency policy is placed: it attaches at 12 million, carries a 33 million limit, the insurer approves any settlement, and the seller keeps conduct of the case because he knows the technology. In 2026, mid-hearing, the claimant offers to settle at 20 million with a cross-license. The buyer wants to accept, and so does the insurer. The founder refuses: he wants a judgment that publicly clears him. Who decides?
The founder decides, and the structure willed it rather than it being an accident. The policy gives the insurer approval over settlements, that is a veto over what the insured proposes, and not the power to impose a settlement on the party conducting the case. As long as the founder refuses, nobody can accept in his place. The tempting reasoning, and it is wrong, is that whoever pays must be able to decide: that is not automatic, it must be written, and nobody wrote it. One must then see why his refusal is not irrational, because that is where the structure reveals itself. He is exposed only to 10 million; the policy attaches at 12. Beyond his personal cap he is playing with the insurer's money and the buyer's business, and he is pursuing an interest of his own, his reputation and the other products he sells. The retention was meant to prevent exactly that behavior, and it fails here because it sits with one party while the decision belongs to another: a retention disciplines only the party that bears it and decides at the same time. That leaves the point the question does not ask and which is the most serious. A risk of injunction over 30% of revenue is not a money risk, and no 33 million limit restores a product line. That risk should never have been handled by a policy: it was to be handled in the deal, through a license negotiated before closing or by carving the line out of the perimeter, and the insurance served to create the impression that it was covered.
- 01Pending litigation is a third category: known, so outside the warranty policy, but the facts are disputed and the outcome is spread, not binary.
- 02The insurer buys a distribution of outcomes and the future behavior of the parties, which prices less easily than a legal opinion.
- 03The disclosure the insurer needs can waive the privilege protecting the defense it insures: this is staged, not a formality.
- 04A retention disciplines only the party that bears it and decides: conduct with one and retention with the other, and moral hazard returns intact.
- 05A risk of injunction is not a money risk: no policy restores a product line, and that is handled in the deal.