The Warranty of Liabilities (Garantie de Passif)

When you buy the shares of a French SARL, you buy the company as it stands - including any liability lurking in its past that has not yet surfaced in the accounts. The law gives a share buyer little protection against that risk: an unexpected tax reassessment or an old dispute that lands after completion can wipe out the value you paid for, with no automatic remedy. That is why buyers insist on a warranty of liabilities (garantie de passif) - a contractual promise by the seller to cover liabilities that predate the sale but come to light afterwards. This guide sets out how the garantie de passif works: what it protects against, the two main forms it takes, who benefits from it, the drafting and the information procedure that make or break a claim, and the tax treatment of a payment under it.

Why a warranty of liabilities matters

The legal actions open to a disappointed share buyer are limited, especially where the shares turn out to be worth less than expected. A buyer who discovers, after completion, that the company owed a tax debt, faced an undisclosed claim, or had overstated its assets, generally cannot unwind the deal on the general law alone. So a careful buyer requires a contractual guarantee - the warranty of liabilities - to fill that gap.

The term is generic: it covers several kinds of undertaking that bind the seller more or less tightly, but the aim is always the same - to protect the buyer against the revelation of a hidden liability, or against certain listed risks. It is the single most negotiated document in most SARL share deals, precisely because it decides who carries the risk of the unknown. In practice, a warranty agreement usually opens with a long set of seller's representations about the company's accounting, legal, financial, employment and environmental position. These representations have real legal force: they guarantee the accuracy of the seller's statements and of the annexes to the agreement, and a false representation is itself a route to a claim.

One boundary matters from the outset. A seller's undertaking to take on "all liabilities" is only a warranty of liabilities - it does not amount to a warranty of the net assets. The two are different promises, and which one the buyer has depends on the words used, not on a general sense that the seller "stood behind" the company. Getting the scope right in the drafting is where the protection is truly won or lost.

The two types of warranty: reconstitution and price revision

Warranty clauses group around two distinct mechanisms, and the difference decides who gets paid and how much.

Under a liability warranty proper (or reconstitution warranty), the seller undertakes to pay off creditors revealed later, or to repay the company the amount of debts disclosed after the sale, so that the company clears the liability. A clause by which the seller agrees to pay back into the company's accounts the net impoverishment suffered gives the company a direct, personal right against the seller - a stipulation for the benefit of a third party - letting the company itself sue for the sums matching that net loss. The compensation is tied to the reduction in net assets, whether from a rise in liabilities or a fall in assets, and it can exceed the sale price unless a cap is freely negotiated - but it benefits the company or its creditors, not the buyer's own pocket.

Under a price-revision clause (or value warranty), the seller instead repays the buyer directly the difference in value caused by the revealed liability, reducing the purchase price. This works best where the price is paid in instalments, allowing a set-off between the unpaid balance and the liability discovered. But it has firm limits: the restitution cannot exceed the sale price, the clause benefits only the buyer (even one who is no longer a shareholder), and it operates in proportion to the shares transferred. Pushed too far - reducing the price below a valid determinable figure - a price-revision clause can even risk the validity of the sale itself.

Naming the mechanism is not a formality: for some courts, the identity of the beneficiary is the decisive criterion that distinguishes a true warranty from a price-revision clause. So the clause has to say, clearly, who is paid and on what basis.

Who benefits from the warranty

Identifying the beneficiary is one of the most important - and most litigated - points in a warranty. In principle the beneficiary is the buyer. The deed can make the acquired company the beneficiary, exclusively or alongside the buyer, but only if it contains a clear stipulation to that effect; absent such wording, only the buyer of the shares can claim payment under the warranty from the seller.

This has practical consequences. Where a SARL is the intended beneficiary, the buyer's action to enforce the warranty is admissible only if the buyer acts for the benefit of the company, not for themselves. A company that later merges with the acquired beneficiary company can validly invoke the warranty. And a sub-buyer - someone who later buys the shares from the original buyer - is generally not covered: a value warranty given to the buyers "or their assignees" does not reach a sub-buyer who was neither a party to the deed nor an assignee of a party. The original beneficiary can, though, assign the warranty to a sub-buyer by a clause transmitting the whole of the seller's commitments on the resale - unless the contract was concluded intuitu personae, tied to the person of the original buyer.

The warranty is also legally welded to the sale. A warranty of liabilities forms a whole with the reciprocal deed of transfer into which it is integrated. And, importantly, even where the sellers give it jointly and severally, it is not subject to the validity conditions of a guarantee (cautionnement) - it is a warranty in the sale, not a suretyship, so the formalism that governs personal guarantees does not apply to it.

What the warranty covers, and when it bites

The reach of a warranty is set by its wording, and the case law has drawn several lines that recur in practice.

The buyer's knowledge is usually irrelevant. A clause covering any prior, unprovisioned liability bites as soon as such a liability surfaces after the sale - it does not matter that the buyer knew about the lawsuit that generated it. Likewise, where the seller failed to provision a known dispute contrary to the clause, they must cover its consequences whether or not the buyer knew of it. The point of the warranty is to allocate the risk of pre-sale liabilities to the seller, not to reward the seller for the buyer's awareness.

The date and cause of the liability control. Where the clause covers liabilities "existing, or having a cause or triggering event" at the sale, a court cannot instead fix on the balance-sheet date the parties happened to reference - the wording about the origin of the liability governs. A tax or social liability whose origin predates the sale is caught even if it crystallises later.

Whether loss is required depends on the words. Where the warranty covers a net accounting position, a loss not reflected in that position triggers the warranty even without actual damage - unless the clause expressly requires a loss. But where the clause promises to indemnify "any loss or harm suffered", the parties have made damage a condition, so the warranty does not bite if the company's turnover holds up or grows, whatever the cause. A management fault by the seller before the sale, revealed later, is not enough on its own: the buyer must still show a fall in assets or a rise in liabilities.

The information procedure that makes or breaks a claim

Most warranties oblige the buyer to inform the seller, within a set period, of any claim, event or fact that could trigger the warranty - and this notification duty is where many otherwise good claims are lost. How strictly it is enforced turns entirely on the clause.

Where the deed is silent on the sanction for a late or missing notification, the trial judges can decide that the buyer's failure to inform, by itself, bars them from invoking the warranty at all. Where the clause sets a notice period but no sanction for overrunning it, the seller still owes the warranty even if the buyer is late - the seller's only remedy is damages if the delay caused them harm. And where the clause fixes a sanction - for example, that the buyer "loses all rights under the warranty" for a late or non-compliant notice - that sanction is applied. Timing details matter too: where the clause requires warning within a number of days, and is silent on how the period runs, it is the date the seller receives the letter that counts, not the date it was sent.

The practical lesson is twofold. A seller wants a clear notification obligation with a real sanction, because it is their best defence against stale claims. A buyer wants either no strict sanction or a comfortable, clearly-triggered deadline - and, whatever the clause says, should notify early, in writing, with the supporting documents, by a method that proves receipt. A model warranty clause typically fixes the scope, the buyer's information obligation and its sanction, the mechanics of payment, a floor (a franchise below which no claim lies), a cap, and a duration - often extended in tax, customs, social and criminal matters until after the relevant limitation period.

Floors, caps, duration and the limits of a warranty

A warranty is bounded by the figures and dates the parties negotiate. A franchise (a floor) is common: the buyer can only call on the warranty once the total claimed passes the floor, filtering out trivial items. A cap limits the seller's total exposure, and a value warranty is in any event capped at the sale price. A duration ends the warranty on a fixed date, with liabilities revealed or notified after it excluded - though the parties often carve out tax, customs, social and criminal matters, keeping the warranty alive until some days after the legal limitation period expires, because those are exactly the areas where old liabilities surface late.

Two further limits are worth knowing. If the buyer approved the accounting methods used to draw up the reference accounts in the deed, they cannot then use the warranty to challenge those methods. And enforcing a warranty does not let the seller turn on the company's accountant: a seller who obtained an inflated price because of the accountant's errors cannot make the accountant refund the part of the price they had to give back to the buyer. There is, however, an escape above the cap: a buyer deceived by the seller - for example, by an accounting treatment so flagrantly wrong the seller could not have been unaware of it - can recover more than the warranty's ceiling, because fraud unravels the negotiated limits.

The tax treatment of a warranty payment

A payment under a warranty has tax consequences that differ by who pays, who receives, and which kind of clause it is.

For companies

For a corporate buyer, an indemnity received from the seller under the warranty does not reduce the cost of the shares carried as an asset; it is included in taxable profit at the ordinary rate. For the acquired company, a sum paid to it directly under a repair obligation is a taxable receipt where the loss or charge it compensates is itself deductible - but where it compensates a charge that is not tax-deductible, it is not a taxable product. By contrast, sums paid under a price-revision clause follow the capital-gains regime rather than being treated as income.

For individuals

For an individual seller, where the sale carried a warranty and the seller has to repay sums under it, they can claim, by way of a contentious claim, the discharge or reduction of the capital-gains tax they paid, up to the sums repaid. For an individual buyer, sums received under a warranty of liabilities or net-asset warranty reduce the acquisition cost of the shares to be used when the buyer later sells them, for computing the gain. So a €50,000 warranty payment on shares bought for €180,000 brings the effective acquisition cost down to €130,000 - €1,300 per share across 100 shares - which increases the taxable gain on a future sale. The warranty, in other words, is not tax-neutral for either side, and the treatment should be modelled alongside the commercial terms.

Two refinements complete the picture. The distinction between a price-variation clause and a warranty of liabilities is not only commercial but fiscal: sums under a price-revision clause follow the capital-gains regime, while an indemnity under a true warranty of liabilities is treated on its own footing, so mislabelling a clause can change the tax outcome as well as who is paid. And where a warranty payment is made directly to the acquired company, its character in that company's hands - taxable receipt or not - tracks whether the loss it repairs was itself deductible: a payment repairing a deductible charge is a taxable product, while one repairing a non-deductible charge is not. Because the same payment can be taxed differently depending on how the clause routes it and what it compensates, the tax analysis belongs in the drafting stage, not after a claim is paid.

The seller's own guarantees survive the sale

A warranty of liabilities protects the buyer against the company's past - but the seller has their own exposure to manage on the way out, and it is easy to miss. Guarantees the seller gave for the company's debts - a personal guarantee to a bank, for example - do not fall away merely because the shares are sold, unless the seller made their status as shareholder the determining condition of the guarantee. It is up to the seller to terminate the guarantees they gave to banks or other creditors; if they do not, they can be pursued by the beneficiary of the guarantee for company debts arising after the sale.

The trap deepens where the buyer agrees to take over the seller's guarantees. The fact that the buyer undertook to substitute themselves in the guarantees given by the seller does not discharge a seller who failed to terminate their own guarantee commitment. And a buyer's promise to take on "all the guarantees the seller may have given" has been held too indeterminate to enforce where the deed did not identify the guarantees, their nature and their extent. The practical rule is to identify each guarantee in the deed, name its beneficiary and the amount secured, and have the seller formally ask the creditor to release them - a substitution clause alone is not enough.

There is a symmetry worth noting. Where a buyer needs the warranty to reach backwards into the company's past liabilities, a seller needs to close off their own forward exposure on guarantees. A well-run share sale deals with both at once: the buyer's warranty of liabilities, and the seller's release from the personal commitments they gave while in charge. Handling one without the other leaves a party exposed after a deal they thought was closed.

Warranty versus earnout, and dispute clauses

A warranty of liabilities should not be confused with a price-adjustment or earnout clause, which does the opposite job. A variation-of-price clause is usually set up to let the seller share in the company's future results - a price supplement paid by the buyer if the business performs. A warranty of liabilities, by contrast, protects the buyer against the revelation of a hidden past liability. The two point in different directions, and their tax treatment differs: a price supplement paid under an indexation clause is taxable in the seller's hands, under the securities capital-gains regime, in the year it is received, whatever the time elapsed since the sale.

Getting the label right therefore matters for both risk and tax. A clause headed "price revision - net-asset guarantee", providing that a contradictory set of accounts at the sale date would reduce the price on any fall in net assets, has been treated as a price-revision clause whose effect the parties did not intend to exceed the sale price - in one case bringing the price down to zero. Whether a clause is a warranty, a value guarantee or an earnout is decided by its substance and its named beneficiary, not its heading.

Share-sale deals also commonly carry a dispute-resolution clause - an arbitration or conciliation clause governing disagreements over the interpretation or performance of the agreement, including the warranty. A conciliation clause in a share-transfer deed makes a court action on the covered disputes inadmissible until conciliation is attempted, and setting it in motion suspends the running of the limitation period. Insurance is a further option: the seller's warranties can be backed by a policy covering the financial consequences of a breach or an inaccuracy in what was warranted, which can bridge the gap where a buyer doubts the seller's ability to pay a large claim years later.

Frequently asked questions about the garantie de passif

This article states general principles of French law as at its date of publication and is provided for information only. It does not constitute legal or tax advice and creates no lawyer-client relationship. Figures, rates and thresholds evolve; verify them against the texts in force before acting, and take advice on your specific situation.