This post uses hypothetical scenarios for illustrative purposes only. It does not describe any actual client, transaction, or representation, and is not legal advice.
The seller’s CFO read the indemnification section out loud, in the conference room, the day before signing. “Cap is ten percent of purchase price,” she said. “We’re fine.” The cap was, in fact, ten percent of purchase price — for one category of indemnification claim. The agreement also had a separate cap for breaches of the fundamental representations (one hundred percent of purchase price), a separate cap for tax-related claims (the full tax liability, uncapped), a separate cap for breaches of the pre-closing covenants (capped at purchase price plus interest), and a fraud carve-out that displaced every other cap entirely. Eighteen months after closing, when the IRS audit produced a deficiency notice for pre-closing periods, the seller learned that the ten-percent cap had nothing to do with the dollars she was now writing checks to satisfy. The cap she had read was real. The other three were also real, and were larger, and were the ones that mattered.
The architecture of indemnification caps in a modern middle-market purchase agreement is the single most misread piece of the document on the sell side. Buyers’ counsel draft the section so that the most prominent number — usually titled “Cap” or “Indemnification Cap” — is a low percentage of purchase price. That number is the one the seller’s CFO sees. The other caps, scattered through subsections and cross-references, are the ones that determine how much money is actually at stake.
The four caps
The first cap is the general indemnification cap. This is the one labeled “Cap” in most agreements. It governs breaches of the operational reps and warranties — the ones about contracts, employees, intellectual property, and so on — and it is the cap that gets discussed at the negotiation table. In a typical deal it runs ten to fifteen percent of purchase price, occasionally lower in a rep-and-warranty-insured transaction.
The second cap is the fundamental rep cap. The fundamental reps are the ones the buyer regards as foundational — typically due authority, capitalization, title to assets or stock, no broker, and sometimes no liens. Breaches of these reps are not subject to the general cap. They are subject to a much larger cap, usually one hundred percent of purchase price or sometimes some defined multiple of the cash component. The theory is that a breach of a fundamental rep goes to the seller’s core delivery of what the seller was selling, and the buyer should not be limited by the general cap when it gets less than what it paid for.
The third cap is the tax-indemnification cap. Tax claims are usually pulled out of the general indemnification framework entirely and placed in a separate tax-indemnification article. The cap there is often the full tax liability, sometimes capped at purchase price, often uncapped through the applicable statute of limitations. The mechanics of the tax indemnification — what counts as a covered tax, who controls the tax contest, how the indemnification gets paid — are negotiated separately from the rep-and-warranty indemnification, and the dollars at stake can be very large in deals with pre-closing tax exposure.
The fourth cap is the fraud carve-out. Most agreements provide that indemnification claims based on fraud — variously defined — are not subject to any cap, basket, or survival period. The fraud carve-out is the seller’s true tail-risk, because the buyer can recharacterize a contested indemnification claim as a fraud claim and litigate around the otherwise-binding cap. Delaware’s increasingly narrow definition of “fraud” for these purposes — generally limited to knowing, intentional misstatements in the express reps — has constrained this risk over the past decade, but the carve-out remains the single largest seller exposure in most deals.
How the caps stack — and what that means in practice
The agreement frequently provides that the four caps stack independently. A single set of facts can produce a claim under the general cap, a claim under the fundamental cap, a claim under the tax cap, and a claim under the fraud carve-out, all at once. The buyer can elect which category to pursue. The cap that applies is the cap of the category the buyer elects — which means, in practice, the buyer chooses the cap that produces the largest recovery, and the seller is bound by that choice unless the agreement specifically requires the buyer to pursue the categories in some defined order.
A simple example: the target had an undisclosed wage-and-hour exposure of $4 million. The general cap is ten percent of a $50 million purchase price, or $5 million. The wage-and-hour exposure is also a breach of the “no undisclosed liabilities” rep (general cap) and the “compliance with law” rep (general cap), and arguably a breach of the “fundamental” tax rep to the extent the unpaid wages produced unpaid payroll taxes. The buyer’s claim under the general cap would be capped at $5 million minus prior claims, leaving the seller fully exposed to the $4 million. The buyer’s claim, recharacterized as a tax claim under the tax indemnification article, might be uncapped — and the seller’s exposure becomes $4 million plus interest and penalties.
The buyer’s lawyers will run the numbers and elect the most favorable category. The seller’s lawyers, if they did not negotiate either a stacking limitation (a single aggregate cap across all categories) or a most-favored-cap election (the seller can elect the lowest cap), are stuck with the buyer’s choice. Most agreements do not contain either limitation by default, and most seller-side negotiations do not focus on the stacking question, because the stacking question only matters when something actually goes wrong and by then it is too late to redraft the agreement.
The interaction with rep-and-warranty insurance
Rep-and-warranty insurance has reshaped the cap discussion in ways that make the architecture even more important to understand. In a typical RWI deal, the policy provides coverage for breaches of the operational reps up to a defined policy limit, and the seller’s indemnification cap shrinks to a token amount — often the deductible, sometimes one-half percent of purchase price, sometimes literally zero. The seller’s exposure for general-rep breaches has, in those deals, been transferred to the carrier.
What has not been transferred is the fundamental-rep cap, the tax cap, or the fraud carve-out. The RWI policy typically excludes pre-closing taxes (or covers them only under a separately negotiated tax-liability endorsement), excludes fraud, and either excludes fundamental-rep breaches above the operational-rep cap or covers them only up to the policy limit. The result is that even in an RWI deal, the seller’s exposure under the fundamental, tax, and fraud categories is undiminished. The seller’s CFO who sees “Cap: $0” or “Cap: 0.5% of PP” on the term sheet often does not realize that the other three categories of exposure are independent of the RWI policy and remain fully on the seller’s books.
I wrote earlier about where the deal sits on the seller-friendly to buyer-friendly spectrum, and the cap architecture is one of the cleanest places to see how the spectrum actually operates. A “seller-friendly” cap is not a cap that is low in nominal terms; it is a cap that is structured to bound the seller’s total exposure across all categories. A “buyer-friendly” cap structure can have a very low nominal general cap and still leave the seller with substantial real exposure through the other three categories. The labels and the substance are often inversely related.
The drafting moves that actually matter
The first move is to insist on a single aggregate cap across all categories. The provision reads roughly that “the aggregate liability of the seller for all claims under this Agreement, whether under the general indemnification, the fundamental representations, the tax indemnification, or otherwise (other than claims arising from actual fraud), shall not exceed [X]% of purchase price.” Buyers’ counsel will resist, on the theory that the fundamental reps and the tax indemnification need to be unlimited. The seller’s response is that the fundamental reps’ value is not protected by an unlimited cap if the seller is judgment-proof above some threshold; a meaningful aggregate cap forces the buyer to think about real recovery rather than nominal exposure.
The second move is to tighten the definition of “fundamental representations.” The buyer’s first draft will list every rep that has ever been called fundamental in any Practical Law form — title to assets, capitalization, due authority, no broker, no liens, no insolvency, compliance with anti-corruption laws, taxes, employees, IP ownership, environmental, and sometimes a long-tail list. The seller’s negotiated list should be three or four reps — title, due authority, capitalization, and arguably no broker — and the others should be subject to the general cap. Each rep that is downgraded from fundamental to general cuts the seller’s tail exposure by a large amount.
The third move is to time-limit and dollar-limit the tax indemnification. The default in many forms is the full tax liability for the full statute of limitations period, including any extensions. The seller’s negotiated version is the actual tax liability capped at some percentage of purchase price (often the same as the general cap, sometimes higher), with the survival period limited to the regular statute of limitations and explicitly excluding any extensions the buyer might consent to post-closing. The buyer’s tax counsel will insist on the unlimited version, but the negotiation usually lands somewhere in the middle, and the middle is far better for the seller than the default.
The fourth move is to define “fraud” tightly. The seller wants “fraud” defined as a knowing and intentional misrepresentation of an express representation, made with intent to deceive, with reliance and damages. The seller does not want “fraud” to include constructive fraud, negligent misrepresentation, equitable fraud, or fraud in the inducement of the deal generally. A tight fraud definition keeps the fraud carve-out as the narrow exception it was meant to be rather than as a back door around every cap in the document. The Chancery Court has been narrowing the operative definition of fraud for indemnification-carve-out purposes over the past several years, and the seller’s counsel should track the doctrinal development and put the narrow definition in the agreement explicitly rather than rely on the default.
What this looks like when it is done right
A well-drafted indemnification architecture from the seller’s perspective looks like this: a single aggregate cap across all categories at a defined percentage of purchase price; a short list of fundamental reps with clear boundaries; a tax indemnification that is folded into the aggregate cap rather than carved out of it; and a tightly defined fraud carve-out that limits the buyer’s ability to litigate around the cap. The general cap can be modest if the architecture is right. The general cap is meaningless if the architecture lets the buyer recharacterize claims through the other three categories.
From the buyer’s perspective, the architecture looks different. The buyer wants the cap categories to remain independent, the fundamental reps to be a long list, the tax indemnification to be unlimited, and the fraud definition to be expansive. The buyer’s negotiated objective is not the headline cap number — that is the number the seller is focused on — but the structural ability to reach beyond the headline cap when something goes wrong post-closing. The buyer’s counsel will frequently concede on the headline number to win the structural points, because the structural points are where the actual recoveries happen.
This is a place where the deal-table discussion and the deal-document substance diverge sharply. The CEOs and the bankers talk about “the cap” as if it were a single number. The deal lawyers know it is four numbers, that the four numbers interact through layered cross-references, and that the seller’s total exposure is determined by the layering rather than by any single number. Sell-side representation on this point is, in significant part, the work of translating the four-cap architecture into a one-number-equivalent that the seller actually understands before signing.
If you are negotiating a purchase agreement and trying to figure out whether the indemnification cap structure protects the exposure you think it does, feel free to reach out to my firm manager, Magda, at Magda@montague.law, or fill out our contact form. Mention you read this post.
— John
