How Much Liability Limit Is Enough? A Benchmarking Method for D&O, Public and Product Liability in India 2026

Most Indian corporates set their liability limits the way they set them last year: take the expiring number, apply a small uplift, and move on. That habit is now the single most common source of underinsurance in the casualty book, because the severity of third-party awards has moved faster than most renewal files.

Three developments through 2025 and into 2026 make a fresh adequacy test necessary rather than optional. First, compensation under the Motor Vehicles (Amendment) Act, 2019 and the way Motor Accident Claims Tribunals now compute future prospects and dependency have pushed large-vehicle and premises-adjacent awards materially higher. Second, the product liability chapter of the Consumer Protection Act, 2019 (Sections 82 to 87) has given claimants a statutory route against manufacturers, sellers and service providers, and the National and State Commissions are awarding on it. Third, tariff-driven trade friction has increased the share of Indian output that touches US and EU buyers, where damages behave very differently from Indian courts.

Award severity, not frequency, is the driver. A liability programme rarely fails because there are too many claims. It fails because one claim, in one adverse forum, exceeds the tower. A limit that looked generous against a 2021 loss history can be a fraction of a single 2026 judgement in a US product suit or a mass-tort premises event.

The rest of this note sets out a repeatable method: size from exposure, benchmark against peers, load for foreign jurisdiction, and build the excess tower so the last rupee of cover sits where directors and the balance sheet actually need it.

The weakest way to size a public or product liability limit is a turnover multiple picked by convention. Turnover tells you premium-rating exposure; it does not tell you what a bad day costs. The defensible method starts from loss scenarios and works back to a limit.

Build three to five realistic severe scenarios specific to the operation. For a chemicals or manufacturing site, that includes a fire or release affecting neighbours, third-party bodily injury to visitors and contractors, and downstream property damage. For a consumer product manufacturer, it includes a defect batch causing injury across a distribution region, plus the legal defence cost of a contested product suit. For each scenario, estimate the components a court would actually award: bodily injury and death compensation on current tribunal and commission multipliers, third-party property damage, claimant legal costs, and your own defence costs, which in a contested casualty matter can run into a significant share of the settlement.

Defence costs inside versus outside the limit matters here. In many Indian liability wordings, defence costs erode the limit rather than sitting above it. If your policy is costs-inclusive, the indemnity available to pay the claimant is the limit minus what you spent defending, so the working limit is smaller than the headline number.

Work the worst credible scenario to a rupee figure, then set the limit above it with margin for aggregation, because a single event can trigger multiple claimants under one occurrence. Note where statute already fixes a floor: the Public Liability Insurance Act, 1991 mandates no-fault relief cover for entities handling hazardous substances above threshold quantities, but the relief amounts are modest and do not extinguish common-law claims, so the Act sets a compliance minimum, not an adequacy target.

Exposure sizing gives you a floor. Peer benchmarking tells you whether that floor is in line with how comparable companies actually buy, which matters both for board defensibility and for renewal negotiation. The question brokers hear most often, how much D&O cover is enough, is really a benchmarking question.

For D&O, the relevant peer set is companies of similar revenue, similar listing status, similar sector litigation intensity, and similar overseas exposure. A domestically listed mid-cap industrial with no US securities exposure and a clean regulatory record sits at one end. A newly listed technology company with an aggressive growth story, retail shareholders, and SEBI and class-action attention sits far higher. SEBI (Listing Obligations and Disclosure Requirements) Regulations, 2015 require D&O insurance for independent directors of the top 1000 listed entities by market capitalisation, so for a large listed company the presence of cover is not discretionary, and the debate is limit size, not whether to buy.

Benchmark on several axes at once, not just revenue. Two companies with identical turnover can warrant very different limits if one has US-listed securities, an active acquisition programme, a regulated-sector exposure such as financial services or pharmaceuticals, or a history of shareholder disputes. Sector matters because litigation propensity is not uniform across the economy.

For public and product liability, benchmark against peers in the same manufacturing class and export profile. A domestic-only food processor and an exporter of the same product to the US carry different limits precisely because the forum of a likely claim differs.

Use benchmarking to challenge outliers in both directions. A limit far below the peer band is an underinsurance flag the board should see. A limit far above the band without a specific exposure to justify it is spend that could buy a higher attachment or a Side-A layer instead. The output is a limit that sits inside a defensible range for the risk, with the deviations explained on the file.

A limit that is adequate for an India-only claim can be badly short for the same event litigated in the United States or the European Union. This is the loading that Indian exporters and internationally listed groups most often underestimate, and 2026 trade patterns have widened the number of companies exposed to it.

The US differs on several dimensions that all push the required limit up. Damages can include punitive elements that are unavailable in India, contingency-fee plaintiff bars make suits economically viable for claimants, class-action and mass-tort mechanisms aggregate claimants, and jury awards carry a variance that Indian bench-decided matters do not. For a product liability exporter, a single defect suit in a US district court can dwarf the entire Indian product liability programme. The EU adds strict producer liability under its revised product liability regime and active regulatory recall enforcement, so the exposure there is severe even if the award mathematics are less extreme than the US.

The practical method is to size the export exposure separately, apply a jurisdiction loading to the worst-credible scenario for US-facing product, and confirm the wording grants US and Canada jurisdiction cover rather than excluding it. For internationally listed groups, the D&O tower must contemplate US securities litigation and regulatory action, which is a different and larger number than domestic SEBI exposure alone. Detailed treatment of the export-specific product exposure sits in our note on global product liability coverage for Indian pharma and manufacturers.

Once the required total limit is set, the next decision is how to build it. A high limit is rarely bought as a single policy. It is assembled as a tower: a primary layer that pays first, then excess layers stacked above it, each attaching where the one below exhausts. Getting the tower structure right is what separates a limit that pays cleanly from one that argues internally during a claim.

The economics favour a tower because the cost of cover falls sharply with height. The primary layer is the most expensive rupee of limit because it is the most likely to be hit. Each excess layer above it is cheaper per rupee because it is less likely to be reached, which is why buying the last, say, high tranche of a tower is far cheaper than the first. This is the argument for commercial umbrella and excess liability cover: the marginal limit at the top of the tower is inexpensive relative to the balance-sheet protection it buys.

Watch three structural traps in a multi-insurer tower. First, follow-form integrity: each excess layer should follow the terms of the primary, because a narrower excess wording can create a gap where the primary would have paid. Second, exhaustion language: an excess layer should attach once the underlying limit is exhausted by payment, and disputes arise when an underlying insurer settles below its limit. Third, defence-cost treatment must be consistent up the tower, so an event does not erode layers inconsistently.

The attachment points should reflect the exposure sizing, not round numbers picked for neatness. Set the primary limit where the bulk of frequency losses fall, then build excess capacity to cover the severe tail identified in the scenario work. A well-designed tower puts inexpensive capacity exactly where the balance sheet is exposed and avoids paying premium-heavy primary rates for limit that a claim will rarely touch.

For D&O specifically, total limit is not the whole story. The structure of the tower determines whether an individual director is actually protected when it matters most, which is precisely when the company cannot or will not indemnify.

D&O cover is written in three insuring agreements. Side A protects individual directors and officers where the company does not indemnify them. Side B reimburses the company when it does indemnify. Side C covers the entity itself for securities claims. In a severe event, Side B and Side C claims by the company can consume the shared limit, leaving individual directors exposed on the Side A cover they personally depend on. Insolvency makes this acute: if the company is insolvent it cannot indemnify, so Side A is the only cover standing, and the shared limit may already be eroded by entity claims.

A dedicated Side-A DIC layer is the standard answer. Side-A Difference-in-Conditions cover sits above the main programme, is ring-fenced for the exclusive benefit of individuals, and drops down to respond where the underlying programme is exhausted, rescinded, or fails to pay. It gives independent and non-executive directors a layer that cannot be consumed by the company's own securities claims. For listed companies, and especially those recruiting independent directors under Schedule IV of the Companies Act, 2013, the presence of a ring-fenced Side-A layer is increasingly a condition of accepting the seat.

Sizing the Side-A layer follows the same logic as the main tower but weighted to the personal-exposure scenarios: regulatory action against named individuals, shareholder claims naming directors, and the defence costs of a multi-year investigation. The premium on the D&O programme, including the Side-A layer, is not treated as director remuneration under Section 197 of the Companies Act where the policy is bought for the company's officers generally, which removes a common internal objection to buying adequate personal-protection limit.

The method comes together as a short, repeatable annual exercise the risk function runs before every casualty renewal. Run it deliberately rather than defaulting to the expiring limit, and the file will defend itself later.

The sequence is straightforward. Rebuild the worst-credible loss scenarios with current award multipliers and current defence-cost assumptions. Re-benchmark the limits against the current peer band for revenue, sector and listing status. Apply the US and EU jurisdiction loading to any export or foreign-listing exposure and confirm the wording actually grants that jurisdiction rather than excluding it. Re-examine the tower for follow-form and exhaustion integrity, and confirm the Side-A layer on the D&O programme is sized and ring-fenced. Where a limit sits below the exposure-based floor or the peer band, record it as an underinsurance flag and take it to the board rather than absorbing it silently.

The hardest part of this exercise is not the arithmetic. It is confirming that the wording delivers what the limit promises: the defence-cost treatment, the jurisdiction grant, the exhaustion and drop-down language, and the exclusions that could quietly remove the very scenario you sized for. A limit is only as good as the clauses that govern when and how it pays.

This is where a searchable view of insurer policy wordings changes the exercise. Sarvada lets risk managers and brokers compare the actual liability wordings across insurers, side by side, to check jurisdiction clauses, defence-cost basis, exhaustion language and the exclusions that decide whether the tower responds. Instead of taking a limit on trust, you can verify that the wording behind it holds under the scenario you are pricing for. To stress-test your liability programme against the wordings that will govern a real claim, Request Access to the Sarvada platform.