UR E26, After the Mandate ③ — Through the Eyes of the Classification Society: Assurance…

UR E26, After the Mandate” ③ — Through the Eyes of the Classification Society: Assurance, Differentiation, and the Weight of Liability

The third installment of a six-part series — the second stakeholder (the classification society). Where Part 2 looked from the position of the party buying the boundary (the owner), this part looks from the opposite side of that boundary — the party that assures it. The aim here is neither to defend nor to criticize the classification society, but to map, through its own internal logic, the forces it actually moves between.

1. The Classification Society’s Dual Identity

A classification society wears two hats at once on the same matter. One is the obligation, as an IACS member, to assure the mandatory floor uniformly. Since the very reason Unified Requirements (UR) exist is to eliminate divergence, member societies bear a responsibility to apply that floor consistently. The other is the need, as a competing business, to differentiate above it. A classification society serves a non-profit public-interest function, yet it is also an organization that competes for customers.

These two identities often do not point in the same direction. Unification demands uniformity; competition demands difference. If the question the owner faced in Part 2 was “what to buy,” the classification society’s question is its mirror image — what to assure and how, how far to differentiate, and what liability to bear for the result. These three questions are intertwined, and to understand the knot, we must first look precisely at what a classification society actually sells.

2. What the Classification Society Actually Sells — Not a Guarantee, but a Verified Opinion

Let us first clear away a common misconception. A class certificate is not a guarantee of safety or seaworthiness. It is merely an attestation that the vessel conforms to that society’s rules. A classification society is not the party that guarantees safety at sea or the seaworthiness of the vessel; the non-delegable duty to ensure seaworthiness rests — legally — with the owner. A classification society does not design, build, operate, or maintain the vessel, nor does it control operations between periodic surveys.

And yet the entire market relies on this certificate. The flag state delegates statutory certification work to the classification society as a Recognized Organization; the port state uses class status in its passage decisions; and charterers, P&I clubs, and insurers trust it as an indicator of convention compliance. In other words, what a classification society sells is not a “safety guarantee” but trustworthy third-party verification — and it is precisely on that verification that the entire shipping system leans. This structure — depended upon by all, yet guaranteeing nothing — is the starting point that defines everything a classification society does.

3. Why Classification Societies Interpret Conservatively — The Structure of Liability

Let us turn to the question Part 2 promised: “what liability does the classification society bear?” The impression that classification societies often interpret conservatively is accurate, but the cause lies not in arbitrariness but in the asymmetry of the liability structure.

A classification society bears liability toward the owner, its contractual counterparty. It owes an implied duty to survey and class with reasonable care and skill, and if negligence is proven, liability may be found (U.S. cases such as Great American Insurance v. Bureau Veritas suggest this possibility). By contrast, legal liability toward third parties (cargo owners, vessel purchasers, and the like) has largely been limited or denied. In the English Nicholas H case, the court declined to recognize a duty of care owed by the classification society to cargo owners — the reasoning being that extending liability to third parties would upset the established (and limited) balance of liability between owner and cargo interests — and in the Morning Watch case, no duty to the purchaser was recognized either (though such holdings vary by jurisdiction).

But limited legal liability and small exposure are entirely different matters. Major pollution casualties such as the Erika (1999) and the Prestige (2002) showed that third-party claims against a classification society can overwhelm its net worth or insurance limits. And here lies a more fundamental asymmetry — price. In one U.S. case, the fee the classification society received was USD 85,000, while the damages claimed were USD 260 million; the court took the view that under such conditions the classification business itself could not survive. Because class fees are negligible relative to the value of the vessel or the potential loss, a classification society cannot price in unbounded risk.

The conservative interpretation of classification societies is therefore most accurately read as a rational response to an extremely asymmetric risk structure. In a structure where what is received is small and what can be lost is vast, the newer and more uncertain the domain, the stronger the conservatism. And cyber is precisely such a domain — threats evolve rapidly, the causation of incidents is complex, and accumulated case law and statistics remain thin. Much of the caution observed in E26 interpretation follows naturally from this structure.

4. The Force Pulling the Other Way — Competition

If liability pushes a classification society toward strictness, competition pushes the other way. And the tension between these two opposing forces is what produces a classification society’s actual position.

Let us face a historical fact squarely. There was a time when classification societies used being less strict as a competitive instrument. In a structure where the owner chooses the society and pays the fee, the society that interprets most strictly risks losing customers to a more flexible one. Forum shopping, discussed in Part 1 — the phenomenon of owners choosing a society in search of a more favorable interpretation — when flipped around to the society’s side, appears as competitive pressure discouraging any society from becoming the strictest interpreter.

So a classification society is caught between two forces — liability and reputation push toward strictness, competition toward leniency. The unification and quality systems of UR and IACS are precisely the mechanism for binding this tug-of-war at the floor. When all members agree to apply the same minimum line, the incentive for any individual society to be dragged down into a race to the bottom diminishes. In this respect, uniformity of the mandatory floor is not the owner’s interest alone but also a common interest of the classification societies as a whole. For if the floor wavers, the asset of trust on which the entire institution of classification depends is eroded.

5. Where Differentiation Draws Its Justification — A Reflection of Real Capability

If the two forces above act on the floor, it is on the upper storey that a classification society differentiates. And that differentiation, in large part, reflects differences in real capability — this is the key to understanding the classification society’s voluntary notation tiers.

Cyber is a domain in which classification societies had invested for years before UR became mandatory. Building on its cyber security recommended practices (RPs) and its system and component type-approval programme (CP-0231), DNV issued its Cyber Secure notation in 2018, and built its own security profile scheme (SP1–SP5) mapped to the security levels (SL) of IEC 62443. Bureau Veritas has operated NR659, Rules on Cyber Security for the Classification of Marine Units; ABS, its CyberSafety programme; and ClassNK, its own guidelines. As seen in Part 1, this accumulation already existed before UR was mandated; if anything, UR aligned itself to it.

What this fact means is clear. Differentiation through higher notations is less about manufacturing and selling value that does not exist than about turning capability built up through earlier investment into a product. The ability to verify higher SLs, methodologies for assessing cyber governance at the operational and fleet level, and broader verification that extends to IT as well — these are products of R&D, and it is natural that their depth differs from one society to another. And this is precisely the space that IACS has codified in stating that “members are free to impose stricter requirements exceeding the minimum line.” The justification for differentiation is firmest when that differentiation is anchored in real capability.

6. The Tension the Classification Society Actually Manages — Performing Unification and Differentiation at Once

Here the classification society’s hardest task comes into view. A classification society must be a good IACS citizen at the floor (unification, uniformity) and a competitor on the storey above (differentiation). A single organization must perform both roles at once, on the same matter.

What makes this simultaneous performance tricky is that the two layers are not cleanly separated. If interpretation of the floor diverges from one society to another, the fragmentation seen in Part 1 recurs, and that, as shown above, erodes the asset of trust of the classification societies as a whole. So for the classification societies too, uniformity of the floor is a common interest worth defending. The problem is that the structural conditions that make such uniformity difficult — the interpretive latitude of a UR text written in goal-based terms, the lack of interpretive convergence in the early days of the mandate, and the competitive pressure seen in Section 4 — all operate at the same time.

Seen in this light, the moves by classification societies to overhaul their rules in late 2025 — for instance, DNV’s split of its class notation into Design and In-Operation — can also be read not as a mere product reshuffle but as an attempt to tidy up interpretation and clarify categories. Distinguishing what is a technical requirement at the design stage from what belongs to the domain of operational management makes the boundary between floor and upper storey that much sharper. To perform unification and differentiation at once, one must, paradoxically, draw the boundary between the two more clearly.

7. The Boundary, Seen from the Classification Society’s Position

In sum, a classification society balances between two layers. It assures the floor uniformly, differentiates on the storey above on the basis of real capability, and throughout manages an asymmetric liability structure. And running through all three is one principle — a classification society earns the greatest trust when it does not blur the boundary.

Keeping the floor clear and uniform is the common asset of trust of the classification societies, and the market rewards differentiation on the upper storey when it is anchored in real capability. Conversely, when the boundary blurs — when interpretation of the floor diverges, or when an upper-storey requirement is presented as if “mandatory” and seeps down into the floor — what is damaged is none other than the very trust on which the classification society itself depends. This is the point at which Part 1’s core proposition is confirmed anew from the classification society’s position. Boundary management is not a task that matters only to the owner; it is also in the classification society’s own interest. If what a classification society sells is trust, the surest way to protect that trust is to let the market see clearly what is assurance and what is differentiation.

8. Closing

If a classification society asks “what to assure and how far to differentiate,” there is a separate party that actually produces and submits the technical deliverables on which that assurance rests. The classification society verifies; someone produces. The party that draws Zone and Conduit Diagrams (ZCD) and design descriptions by hand, and must satisfy the differing expectations of multiple classification societies simultaneously on a single ship — the protagonist of the next installment: the shipyard.

The next installment moves to the other side of verification — the view of the shipyard (the system integrator). We will look at the burden of implementing the mandatory floor in actual drawings and documents, and the reality of multi-class handling — dealing with the differing interpretations of several classification societies entangled in a single ship.

Key Bases

• The express IACS principle that UR / Unified Requirements are minimum requirements and that “each member is free to impose stricter requirements.”
• Lineage of classification societies’ cyber capability: DNV (cyber security RPs, CP-0231, Cyber Secure 2018, security profiles SP1–5 ↔ IEC 62443 SL), BV NR659, ABS CyberSafety, ClassNK guidelines — accumulated before UR was mandated and aligned to UR.
• Classification society liability structure: the certificate is an attestation of rule compliance, not a guarantee of seaworthiness; the non-delegable duty to ensure seaworthiness rests with the owner; third-party liability is largely limited (English Nicholas H, Morning Watch), while negligence liability toward the owner is possible (U.S. Great American v. Bureau Veritas); the asymmetry of exposure relative to fees (e.g., USD 85,000 in fees against USD 260 million claimed) and the third-party claim exposure of Erika and Prestige — varying by jurisdiction.
• Statutory certification through the flag state’s delegation to the Recognized Organization (RO), and trends in classification society rule maintenance (e.g., DNV’s split of Design / In-Operation notations).

This series is a general analysis of the market structure surrounding IACS UR E26/E27 and is not a legal judgment or advice regarding any particular classification society, case, or jurisdiction. The statements concerning liability are general explanations based on publicly available case law and literature; specific application follows the law of the relevant jurisdiction and the individual facts.