RESEARCH DOSSIER
Maritime Collision Liability and Insurance Apportionment
Abstract. Maritime collision liability arises from the interaction of navigational duties, factual causation, comparative fault, legally recoverable damage, limitation of liability, and overlapping contracts of marine insurance. This monograph reconstructs that interaction through the 1910 Collision Convention, the International Regulations for Preventing Collisions at Sea (COLREGs), English maritime statutes, United States admiralty decisions, original judgments in Ever Smart and Alexandra 1, representative hull clauses, Nordic market terms, protection and indemnity (P&I) arrangements, and maritime safety-investigation materials. Its principal analytical finding is that fault apportionment, cross-liability calculation, statutory limitation, and insurance response are sequential but distinct operations. The same collision can therefore generate several internally correct percentages with different denominators and legal meanings. Comparative study shows why proportional fault depends on rule applicability and causally relevant conduct; why personal injury, pollution, cargo, and wreck liabilities require separate classification; and why a P&I pooling tower does not establish the contractual cover of an individual shipowner. A transparent hypothetical example and reconstruction of reported damages in the Ever Smart litigation demonstrate the monetary sensitivity of fault findings without treating illustrative netting as a judicial payment order. The evidence is documentary and doctrinal, supplemented by reproducible arithmetic. The work makes an original integrative analytical synthesis; it proposes no new theory and derives no empirical collision frequency or causal treatment effect from unavailable claims data.
Keywords: Maritime collision; collision liability; COLREGs; navigational fault; comparative negligence; collision apportionment; admiralty law; marine hull insurance; running down clause; protection and indemnity; P&I clubs; cross liabilities; joint and several liability; maritime limitation; subrogation; collision investigation; loss adjustment; marine insurance.
Introduction
The economic meaning of a ship collision is rarely exhausted by the repair cost of either vessel. A collision can affect hulls, machinery, cargo, harbor facilities, passengers, crews, coastal property, the environment, and subsequent commercial operations. It can also create rival claims between shipowners and claims by persons who never participated in the navigation. Determining who caused the event answers a different question from determining which insurer bears a resulting loss. The former concerns conduct, causation, and legally recognized responsibility; the latter requires an insured interest, a particular policy wording, coverage period, exclusions, deductibles, attachment limits, and sometimes contractual recourse against other insurers.
The international reference point is the 1910 collision convention, whose proportional-fault approach helped displace older all-or-nothing and equal-division practices. Operational duties are specified principally by the IMO collision-prevention regulations. These texts interact with national maritime law; neither document by itself supplies a globally uniform insurance contract. The resulting problem is one of ordered legal and evidentiary interpretation. This monograph examines that order through British and American legal materials and comparative marine insurance terms, with international institutions providing a wider frame.
The method is close reading of primary legal instruments and judicial decisions, comparison of identifiable contract architectures, analysis of official casualty investigations, and explicit arithmetic under disclosed assumptions. Jurisdiction and clause edition are tracked for every specific rule. The study does not infer accident probabilities from case reports, treat loss-prevention reports as negligence judgments, or identify a single worldwide split between hull insurers and P&I clubs. The thesis guiding the analysis is that a sound allocation must preserve its denominator and doctrinal source at every step. The percentage of navigational fault, the percentage of an adversary's recoverable loss, the fraction covered by hull insurance, and an individual club's retained reinsurance layer are not interchangeable quantities.
Scientific Concepts and Historical Origins
The collision as a sequence of legal determinations
A collision, in this inquiry, is a maritime incident in which the interaction of vessels gives rise to alleged navigational liability, including consequential physical loss and associated third-party claims. Physical contact between two vessels is the paradigm case. Closely related “no-contact” navigation damage, striking of fixed objects, tug-and-tow incidents, and contractual indemnities may engage different rules or cover clauses, and are therefore identified separately in the claim schedule. The COLREG scope and definitions explain the navigational obligations but do not settle those coverage classifications.
A complete liability determination moves through identifiable inquiries: which rule governed the encounter; which acts or omissions departed from the legal standard; how each departure contributed to the casualty or specific head of loss; whether the claimant can recover that loss; what percentage or other liability rule applies; whether liability can be limited; and what remains recoverable against the responsible party. Insurance coverage begins with its own inquiry, including whether the claimant or shipowner is an insured, the nature of the liability, and the governing wording. An insured's legal liability can be wider than the sum insured; conversely, a hull insurer can respond to the insured's own physical damage even when that ship principally caused the collision.
In practical claims accounting, at least four ledgers should be retained: each vessel's own property loss; liabilities to the other vessel and its property interests; liabilities to non-vessel claimants; and the insurance and recourse transactions. An agreed net settlement can combine several ledgers for payment convenience. It cannot by itself explain which costs each policy was intended to indemnify. A defensible evidentiary record must preserve both gross and net values.
Historical movement toward proportionate fault
The international Collision Convention text (1910) distinguishes accidents or force majeure from unilateral and mutual fault. Its Article 4 apportions liability in proportion to the respective degree of fault when two or more vessels contribute, with equal division when comparative degrees cannot be established. The Convention also separately addresses joint liability for damages caused by death or personal injuries, together with recourse between vessels. Its applicability depends on treaty conditions and the domestic law giving it effect; it is not a blanket answer to every modern collision claim or insurer dispute.
English law expresses a comparable proportional idea in Merchant Shipping Act section 187. For a collision caused by the fault of two or more vessels, liability for damage or loss is in proportion to their degrees of fault, with equal proportions where different degrees cannot be established. English rules for personal injuries and contribution are addressed independently. Statutory proportionality offers a basis for allocating fault; the legal classification and quantum of each claim remain essential.
United States admiralty law reached modern comparative allocation through judicial development. The Supreme Court's Reliable Transfer decision (1975) rejected the longstanding equal-division rule for maritime property damage and directed proportional allocation according to comparative causative fault, reserving equal division for cases in which parties are equally at fault or meaningful apportionment is impossible. This historical transition matters when reading older cases and insurance clauses: the existence of an equal-division example in an older judgment does not establish a current default for all American collision cases.
The comparison shows convergent principles without uniform procedural machinery. Collision treaties, domestic substantive law, burdens of proof, limitation procedures, and policy contracts operate within their own regimes. A source applying to a British-registered vessel in English litigation cannot automatically be extended to a foreign flag, foreign waters, or a different insurance governing law. Conflict-of-laws analysis, contractual jurisdiction, and the forum's characterization of claims must therefore precede any calculation that purports to bind parties.
Actors, claims, recoveries, and governing law
The parties associated with a collision include registered owners, bareboat charterers, operators, managers, masters and watchkeepers, pilots, cargo interests, terminal operators, injured individuals, public authorities, hull underwriters, P&I clubs, and reinsurers. These roles overlap in practice but create separate legal capacities. Vessel ownership does not by itself establish whether a charterer is contractually indemnified. Likewise, the party that brings a subrogated action may be a property insurer exercising a derivative right rather than an additional owner of the physical loss.
Under English law, the Marine Insurance Act subrogation rules provide an important historical statutory foundation for recovery after indemnification. Recoveries from the party responsible for the damage cannot be counted as a second uncompensated loss. Proper classification distinguishes an insurer's reimbursement, its contractual contribution rights, and a claimant's outstanding uninsured interest. Double recovery is a practical accounting risk when insured values, salvage payments, and third-party damages are booked under separate systems.
Liability Apportionment
International and domestic fault allocation
Comparative apportionment assigns shares of responsibility to vessels whose causative fault has been established. The 1910 Convention's Article 4 and English proportional-fault provision both support allocation by degree of fault and equal allocation where comparison cannot reasonably be made. The Reliable Transfer holding provides the American comparative-fault counterpart for maritime property damage. A set of findings such as 70%/30% means that relative causative responsibility has been assessed between the specified wrongdoers for the specified claims; it does not indicate that one vessel's hull insurer must necessarily pay 70% of its own repair bill.
Where more than two vessels, facilities, or third parties contribute, the denominator must identify all legally relevant responsible parties and the effect of a settlement or absent party under the governing law. The American McDermott v. AmClyde decision illustrates a proportionate-share settlement approach in admiralty involving construction equipment. It is a useful analogy for partial settlement analysis but was not a two-ship collision holding. Any transposition to a particular collision must check the forum's applicable substantive and procedural law, especially contribution and the treatment of nonsettling parties.
A claim's recoverable amount is independently determined. A tribunal may exclude speculative consequential losses, require causation and mitigation proof, or recognize a legally different loss base for cargo owners and other claimants. As a result, a single fault fraction may multiply several different eligible bases. Without recording the eligible base and the payor–payee direction, a percentage is insufficient for accounting or insurance recovery.
Comparative calculations and cross-liabilities
Assume vessel A is 70% causatively at fault and vessel B 30% at fault. Suppose the independently established, legally recoverable physical damage is USD 4 million to A and USD 10 million to B, with both owners entitled to make property claims and no independent limitation or contractual change. On a gross cross-liability representation, A owes B 70% of B's USD 10 million loss, or USD 7 million. B owes A 30% of A's USD 4 million loss, or USD 1.2 million. A bilateral setoff, if permitted and actually agreed, produces a single net amount of USD 5.8 million payable from A to B. The net amount is a payment simplification after the underlying liabilities have been derived; it does not make the gross claims disappear.
This calculation follows the proportional principle described in the Collision Convention. It assumes both losses are recoverable under one law, that causative percentages apply symmetrically to the two vessel property claims, and that there are no contractual exclusions, settlement compromises, deductions, or third-party claims. Those assumptions are methodological choices for arithmetic illustration, not findings about a real casualty. A vessel's damaged hull is still a first-party claim under its own hull policy, even when the opposing vessel also owes damages that may later be collected through subrogation.
The same numbers with A 80% at fault and B 20% yield cross-claims of USD 8 million and USD 0.8 million respectively, or USD 7.2 million net from A to B. The ten-percentage-point shift increases the net by USD 1.4 million. Algebraically, when both property losses are fixed and both gross counterclaims remain enforceable, the derivative of the net A-to-B payment with respect to A's fault fraction equals the sum of the two recognized losses. This is an accounting identity under specified assumptions, not a new liability model. Its usefulness is to identify exactly why allocation uncertainty can be financially large even without a change in the repair evidence.
Personal injury, cargo, pollution, and wreck costs
Personal injury can engage legal rules different from vessel property damage. The English personal-injury provision addresses joint and several liability in qualifying collision cases, and section 189 contribution addresses the relationship between liable parties. This distinction matters where an injured seafarer may recover from one defendant notwithstanding an internal comparative allocation between vessels. The other defendant's contribution exposure and the first defendant's eventual insurer recovery require separate entries. The precise cause of action, claimant status, and statutory application must be verified before using these provisions.
Cargo damage also requires attention to the bill of lading, carrier identity, cargo ownership, contractual exceptions, and any separate cargo insurance. A collision may cause damage to both a vessel and goods aboard it, yet the carrier's liability to cargo interests may differ from an opponent's tort responsibility. Cargo insurers may pursue their own recoveries, and the liability policies of carriers or charterers can contain specialized terms. Charging every cargo loss to the same hull liability account simply because the collision was navigationally caused is a category error.
Oil pollution provides a striking example of specialized liability. The IMO's Civil Liability Convention overview describes the dedicated international framework for certain tanker-source persistent oil pollution claims and compulsory insurance. Other pollution events, including bunker spills from ships outside particular tanker regimes, can implicate different treaty systems and domestic law. A collision can trigger pollution liability, but a 70% navigation-fault allocation does not itself settle the applicability of a strict liability regime, its defenses, compulsory security, direct action, or the ultimate insurer.
Wreck removal, harbor property, salvage, and loss-of-use claims similarly require legal and contractual classification. A harbor authority may incur legally recoverable expense even though neither ship suffers a repair loss proportionate to it. Applicable legislation, administrative action, limitation instruments, and P&I rules determine whether that expenditure is indemnified. The rigorous unit of analysis is the individual claim, not a single undifferentiated “collision loss.”
Limitation of liability and settlement effects
Limitation law may cap a shipowner's exposure even after fault and the gross recoverable loss have been adjudicated. The IMO's LLMC Convention account describes the 1976 Convention, the 1996 Protocol, and amendments increasing limits with effect from June 2015. The applicable regime depends on ratification, domestic implementation, relevant date, eligible claim and defendant, gross tonnage, and circumstances affecting the right to limit. It must not be assumed that an LLMC limit is an insurance policy limit or that an insurer's contractual deductible reduces the statutory fund.
For an illustrative 10,000-gross-ton vessel under the post-2015 1996-Protocol property-claim scale described by the IMO, the amount is 1.51 million SDR for the first 2,000 GT plus 604 SDR for each of the remaining 8,000 GT, or 6.342 million SDR. The calculation is conditional on that instrument and category applying, and does not convert SDR to dollars or claim the fund is always established. A liability share of USD 7 million cannot simply be compared directly with 6.342 million SDR without exchange-date, eligible-claims, and procedure analysis. Multiple claims may compete for a limitation fund; a limitation analysis should preserve claim classes and distributions.
Settlement creates another distinct allocation layer. The choice to settle at a discount can reflect proof risk, procedural costs, insolvency risk, interest, or commercial relationships rather than a new factual fault percentage. A settlement may include releases, contribution protection, subrogation assignments, or insurer consent conditions. Treating a negotiated percentage as the court's finding would distort a comparative case series. The McDermott settlement analysis reinforces the importance of identifying how partial settlements alter claims among different parties, without importing that decision as a universal solution to collision contribution.
The Ever Smart and Alexandra 1 litigation
The collision between the container vessel Ever Smart and tanker Alexandra 1 occurred near Jebel Ali in February 2015. The initial Admiralty Court decision allocated fault 80% to Ever Smart and 20% to Alexandra 1. After appellate litigation, the UK Supreme Court's 2021 crossing-rule decision required reconsideration of the applicable rule framework. The subsequent Admiralty Court judgment reported that the crossing rule applied and redetermined the proportions at 70% and 30%, respectively. The 2022 judgment on remittal is the primary record for both procedural history and the revised apportionment.
The same judgment reports assessed recoverable loss of USD 9,308,594.71 for Alexandra 1 and USD 2,531,373.71 for Ever Smart. Applying the revised 70%/30% allocation to those stated figures for a simplified gross cross-liability illustration gives Ever Smart a liability of USD 6,516,016.30 toward the tanker loss, and Alexandra 1 a liability of USD 759,412.11 toward the container ship loss. The resulting hypothetical bilateral net, before any exclusions, interest, legal costs, enforcement, cover questions, or other claims, is USD 5,756,604.18 from Ever Smart to Alexandra 1. Repeating the same exercise using the earlier 80%/20% shares yields USD 6,940,601.03 net. The difference before individually rounding the net values is USD 1,183,996.842 (approximately USD 1.184 million), or ten percent of the two assessed losses combined. Independently rounded net amounts may differ by one cent when subtracted. This is an analytical counterfactual based on the court-reported figures, not a statement that the court ordered, collected, or insured either net amount.
The example establishes a more specific conclusion than “rules matter.” The legal interpretation of encounter geometry can move a recoverable claim by an amount driven jointly by both vessels' losses. A claims system that updates only the liability percentage and neglects the provenance of the underlying quantified losses can produce internally inconsistent results. A historical analysis must preserve the date and court level of the original finding and the replacement judgment. The Ever Smart record is also a caution against treating an initial trial assessment as a final, timeless empirical datum.
Insured Allocation
Hull physical damage and collision liability
Hull and machinery insurance typically begins with the insured vessel's physical interests. The hull insurer's contribution to repairs therefore depends on insured perils, proximate cause, policy valuation, limits, deductible, and exclusions. The same collision may create the shipowner's liability to another vessel, which might be covered by a collision-liability extension to hull insurance, by P&I, or by some combination. Ownership of a loss and legal fault are separate dimensions: a substantially at-fault owner may still collect first-party hull indemnity, subject to its contract.
The historic Institute Time Clauses Hulls (1983) contain a running-down clause addressing a specified fraction of collision liability, traditionally three fourths, with detailed limitations, exceptions, and cross-liability accounting. This wording is a historical comparative reference, not a statement that every current hull policy adopts the 1983 clauses. Particular hull covers may have four-fourths collision liability or bespoke collision provisions; some fleets buy standalone or additional cover. Exact issue date, attached clauses, governing law, and endorsements are necessary evidence.
On the liability side, the insured interest is generally the shipowner's legal obligation to indemnify another claimant, rather than the opposing vessel's property as such. A property insurer settling the other vessel's damage can become a subrogated claimant. The defending shipowner's collision liability coverage must then be evaluated against the legally established obligation. Liability coverage does not necessarily attach to an uninsured, voluntary, or ex gratia settlement merely because a club or hull underwriter considers it commercially useful.
Running down clauses and P&I interfaces
The familiar “three-fourths hull, one-fourth P&I” description is conditional. It describes one allocation architecture for qualifying collision liability where a hull running-down clause covers three fourths and P&I rules respond to the balance. Whether a club responds also depends on the nature of the liability, exclusions, prior consent, cover conditions, and whether the liability falls under the club's specified class of risk. The UK P&I 2026 conditions offer an identifiable example of collision-risk clauses and refer to one fourth or such other proportion as may be agreed. These are edition-specific terms whose exact applicability requires the member's actual entry and endorsements.
Under the USD 7 million gross collision liability in the earlier hypothetical example, a simple three-fourths hull collision clause would allocate USD 5.25 million to that clause and USD 1.75 million to the balance potentially falling to P&I, before any separate policy limit, deductible, exclusions, recoveries, or claim-specific conditions. This split is an illustration of contract coverage shares of a liability amount, not a new 75%/25% assignment of navigational fault. The opponent vessel's USD 10 million own repair loss remains its own first-party claim unless and until insurers or owners recover the legally attributable portion. When comparing hull and P&I payments, the analyst must distinguish covered versus actual paid amounts and avoid assuming all eligible sums have been reimbursed.
Not all consequences of collision are captured by the running-down clause. Personnel injury, carried cargo, pollution, removal costs, and liabilities excluded from hull collision terms may fall under other insurances, club rules, contracts, or uninsured exposure. P&I may respond to particular liabilities even where the hull has no three-fourths clause, and collision liability can be fully covered in hull under certain systems. Therefore, the market shorthand cannot replace the operative risk definitions. A good coverage matrix records the claim type, exact clause, fraction, limit, and exclusion for each entry.
Nordic coverage and divergent market wordings
The Nordic Marine Insurance Plan's Chapter 13 supplies a different comparative approach to shipowner collision and striking liability. The relevant provisions and official Nordic commentary discuss liability protection and excluded categories in a coordinated plan framework, rather than presupposing a universal three-fourths hull convention. In broad terms, collision liability attached to the insured vessel can be treated on a four-fourths basis within the plan's terms and cover limits, while other liabilities are allocated or excluded by specific clauses. The precise version in force and contractual modifications control the result.
A market comparison should proceed by matching the same loss items across the two sets of wording. A broad hull collision clause might absorb a larger share of vessel-to-vessel property claims while still excluding crew injury or pollution. A P&I club might assume a greater collision fraction where a vessel has less or no hull collision cover, but entry terms and insured responsibilities still matter. Different market structures can produce similar total insurance recoveries for a simple two-hull collision while producing materially different deductibles, claims control, subrogation rights, and protection for exceptional losses. Conversely, an apparent four-fourths promise cannot be treated as a guarantee of coverage for every associated third-party claim.
The scientific comparison must preserve the plan's edition and the actual risk inception date. The existence of a current online plan index does not prove that a later version retrospectively amends a policy written in an earlier year. Legal and contractual reconstruction should archive the relevant edition or cite its actual provision, identify any endorsed deviation, and separate published model wording from signed insurance terms.
The P&I club, pool, reinsurance, and insured
Protection and indemnity clubs are mutual liability insurers or arrangements governed by membership rules, calls or premiums, and defined covered risks. International Group pooling distributes exceptionally large eligible claims among participating clubs and reinsurers. The International Group agreements describe the annually governed pooling arrangement, risk eligibility and exclusions, contribution principles, and collective purchasing of reinsurance. None of these features means that the underlying shipowner is automatically entitled to every value shown in a reinsurance tower. Coverage arises first under the member's club rules, certificate, entry conditions, and relevant defenses.
For the 2026/27 policy year, the Group's pool and GXL structure reports an individual club retention of USD 10 million per claim, a shared Pool from USD 10 million to USD 100 million, and a Group Excess of Loss (GXL) program attaching above USD 100 million. Its structure includes contractual layers, captive participation through Hydra, specific aggregate limits for named risks, and higher overspill arrangements. These are annual reinsurance financing parameters, not claimant priority rules, a statutory maritime limitation fund, or a USD 10 million deductible imposed on every shipowner. The Japan P&I 2026/27 circular supplies a second primary statement of the year's structure and particular limitations.
For example, an eligible USD 17 million covered P&I claim does not imply that the member must fund the first USD 10 million personally. The club may cover the loss under its rules less the member's contractual deductible, while the first USD 10 million of the eligible club claim lies inside the club retention for reinsurance purposes and the remainder may enter the Pool according to its agreement. These two money flows answer different questions: who owes indemnity to the member, and how a club finances its indemnity. A club's reinsurance recoveries are also exposed to conditions, limits, aggregate provisions, and exclusions that may have no identical expression in the insured's original entry.
Subrogation, contribution, deductibles, and claims practice
Subrogation and contribution are necessary to complete the allocation. After paying the insured's own hull damage, a hull insurer may pursue the opposite shipowner within the limits of the insured's rights under applicable law. The English Marine Insurance Act provisions provide statutory illustrations of subrogation and contribution. An insurer defending the insured's liability may control settlement or seek recovery from another responsible party under a different legal instrument. The same gross collision creates both outgoing liability payments and incoming recovery possibilities.
Consider A's own USD 4 million damage in the hypothetical example. If A's hull insurer pays that claim, it may have a potential claim derived from A's USD 1.2 million recovery against B, subject to deductibles, policy wording, priorities, and the insured's residual interest. This recovery is not an additional USD 1.2 million of first-party damage. At the same time, A's USD 7 million collision liability toward B could be indemnified through hull collision and P&I cover according to the contract's applicable fractions, even though bilateral owner-level setoff yields USD 5.8 million. Correct claims accounting tracks the gross liability and the corresponding receivables or offsets, then allocates recoveries without compensating the same loss twice.
An operational workflow should preserve documentary evidence of notice, coverage reservations, guarantees and security, surveyors' reports, quantum verification, club and hull underwriter correspondence, collision settlement agreements, and payment currency and date. A deductible can differ between own hull damage, collision liability, and P&I; erosion of a hull sum insured and the applicable collision-liability maximum may be calculated independently. Currency changes, interest awards, security costs and litigation expenses can make nominal claims with the same fault share economically different across years. Net claims data are consequently insufficient for robust inference unless gross loss, recoveries, policy layers, and deductibles are recorded separately.
Research Evidence Quality and Sensitivity
Source authority, versions, and jurisdiction
Primary legal texts and judgments have higher authority than insurer summaries for the legal rules they state. Official treaty and statute material identify formal scope and amendments; judicial decisions establish governing interpretations for their forum and facts. Insurance terms establish coverage only when they are the terms actually incorporated into the insured contract. P&I clubs' published explanations document institutional practice, but an explanation on a website is not itself a signed entry. Safety investigation reports supply technical causal findings and are kept separate from negligence adjudication.
For this dossier, each legal and institutional source is linked to the claim it supports and mapped to an evidence ledger with source type, jurisdiction, verification scope, and date. Historic sources such as the 1910 Convention and the 1983 Institute Hull Clauses are analyzed as historic documents, while 2026/27 pooling arrangements are dated to that policy year. The main comparative inference remains valid only at the level of a source-defined legal or contractual architecture. No passage is intended to represent the current law of every flag state, nor does a general text replace a statute's precise effective version on an accident date.
The UK Supreme Court case record and the 2022 remittal judgment illustrate the importance of maintaining judicial version history. The first apportionment, the intermediate appeal, the Supreme Court's rule-application decision, and the reallocation are distinct dated evidentiary events. An original incident reconstruction can be correct as a description of what a trial court held in 2017 while being incorrect as a statement of the final apportionment after 2022. Research and claims records must preserve this provenance rather than silently replacing earlier decisions without a date.
Reconstruction of disputed collision facts
Independent incident reconstruction begins with time synchronization and map reference frames. A vessel track derived from AIS may use reported positions and courses at irregular intervals, while radar plots and a voyage data recorder may capture different samples. Analysts should retain the raw observation, instrument source, signal quality and any interpolation. The IMO's VDR safety requirements support use of recorded bridge data, but not an assumption that all vessels or accident dates provided identical recording channels.
A collision reconstruction needs at least two competing causal narratives when the underlying evidence plausibly permits them: for instance, an allegedly late avoidance maneuver, an earlier improper course, or a mistaken assessment of relative motion. Every narrative should be tested against observed times, speed, geometry, maneuvering characteristics, collision point, and witness testimony. Alternative legal findings can be generated by a change in rule applicability even when the instrument data stay fixed. This is the central lesson of the Ever Smart litigation and a useful distinction between measurement uncertainty and doctrinal uncertainty.
The MAIB investigation of City of Rotterdam illustrates a further source of uncertainty: crew perception may be distorted by structural features in ways that a post-event route plot alone does not show. Independent nautical technical evidence therefore complements, but does not substitute for, witness assessment or judicial factfinding. The research value of an incident report lies in its documented mechanism and recommendations, not an inferred nationwide distribution of blame.
Hypothetical loss calculation and scenario tests
The following figures are invented for transparent sensitivity testing and represent neither a real insurer's loss portfolio nor an actuarially fitted estimate. Let A's recoverable physical damage be USD 4 million, B's USD 10 million, and A's fault share be a variable p between zero and one. Then A's gross liability to B is p × USD 10 million, while B's gross liability to A is (1 − p) × USD 4 million. On the assumptions of two enforceable mutual property claims and legally permitted bilateral setoff, net payment by A is USD 14 million × p − USD 4 million, for the range in which the expression remains positive. The expression is a bookkeeping identity derived directly from two gross claim equations.
At A fault shares of 60%, 70%, and 80%, the corresponding hypothetical net amounts from A to B are USD 4.4 million, USD 5.8 million, and USD 7.2 million. A change of one percentage point changes net payment by USD 140,000 given fixed recognized losses and cross-claim rights. This sensitivity isolates one assumption: fault. It does not address uncertainty about whether the losses are recoverable, whether a joint-liability regime applies, whether one defendant can limit, or whether insurers can enforce a setoff. If B's loss valuation were revised upward, both A's gross liability and the sensitivity to p would change.
For an insurer-layer sensitivity example only, apply the three-fourths collision clause to A's USD 7 million gross liability. Hull collision cover has a nominal share of USD 5.25 million and the nominal residual share is USD 1.75 million. If an otherwise comparable four-fourths hull clause applies to the same eligible liability, the nominal hull share becomes USD 7 million and there may be no residual for P&I on that particular head of loss. These are conditional contractual arithmetic scenarios, not coverage opinions. In both scenarios, liability apportionment remains 70%/30%; the policies redistribute the same gross insured obligation after fault is determined.
A further sensitivity arises from statutory limitation. The IMO's LLMC amendments overview specifies SDR limits whose governing versions depend on date and jurisdiction. The illustrative 10,000-GT property-claim calculation of 6.342 million SDR cannot be transformed into a cash insurer payment without a defined claims fund, claimant population, SDR conversion date, and evidence that the limiting party is entitled to invoke the regime. Mixing currency and limitation levels would introduce false numerical certainty.
Inference limits, data quality, and legal uncertainty
Collision litigation is selected for dispute severity, complexity, insurance structure, and the parties' willingness to contest. Published judgments therefore cannot be treated as a representative random sample of maritime navigational behavior. The severity of a reported court case cannot be converted into an empirical expected-loss rate without a denominator of exposed vessel-time or voyages and a systematic account of unreported and settled collisions. The EMSA annual casualty information provides incident reporting in a defined European maritime-interest framework. Its aggregate casualty counts do not themselves establish collision-specific insured frequencies, and revisions or definitional differences must be reconciled before cross-year comparison.
Additional limitations concern incomplete policy access, cross-border conflicts, variable casualty investigation availability, and source accessibility. The relevant running-down clauses, Nordic plan, and club terms are representative published versions rather than actual policies in the judicial case studies. This monograph has neither a complete fleet-level insurance dataset nor direct access to confidential claims files, reserves, or member entries. Its comparative findings are therefore doctrinal and institutional, supported by transparent illustrative calculations; they do not warrant new global frequencies, causal estimates, or pricing relativities.
Integrated Scientific Synthesis
A claim-level reconstruction method
A complete collision file should start with the identification of the incident and its parties and then record navigational acts, causation findings, each claimant's legally recoverable damage, and the applicable comparative-fault determination. The legal record then applies any special category rule for injury, cargo, pollution, wreck costs, or limitation. The financial record preserves the gross direction and amount of each obligation before any lawful setoff, contribution or settlement. The insurance record maps each surviving insured liability or property loss to the exact policy wording, insured capacity, deductible, limit, exclusion, and recovery pathway. This is a disciplined sequence of ordinary legal and claims operations; it is offered as an analytical presentation of existing practice rather than a novel theoretical framework.
For each claim, the minimum reproducible entry consists of claimant, defendant, interest, factual cause, legal cause of action, ruling or settlement date, jurisdiction, relevant COLREG duties, proportion of fault where applicable, currency, recoverable gross damages, limitation treatment, settlement treatment, applicable hull/P&I or other policy, coverage fraction, deductible, reimbursement, subrogation recovery, and a source pointer. Maintaining these fields exposes disagreement: experts may agree about a vessel's track but disagree about rule applicability; counsel may agree about a 70% fault share but disagree about cargo defenses; insurers may agree about an owner's liability but disagree about the clause covering it.
The integration must also recognize temporal dependencies. The policy edition is the version attached to the insured exposure; the governing legal instrument is the one applicable to the casualty or claim; the court finding belongs to its dated procedural stage; the reinsurance tower belongs to a specified policy year; and the eventual publication record concerns the actual release date. Treating these dates as a single “case year” loses information with material financial consequences.
Comparative institutional consequences
The first major finding is that navigational fault and indemnity are analytically orthogonal until they meet at a legally recoverable head of loss. The Reliable Transfer rule and English comparative statute can assign the same property-loss fault proportion while leaving different coverage questions under the parties' policies. A court's assessment of conduct should not be mechanically translated into a hull/P&I split. The former percentage refers to causative responsibility; the latter refers to an insured liability under a contract.
The second finding concerns doctrinal stability and financial sensitivity. The Ever Smart remittal changed the fault proportions following an authoritative determination of crossing-rule applicability. When the eligible damage amounts remain fixed, the resulting illustrative cross-liability change is exact and can be audited. Legal classification is therefore capable of altering a multimillion-dollar allocation even when no new repair survey is produced. The finding does not imply that all appeals change damages or that the arithmetic is the complete judgment debt.
The third finding is the importance of claim taxonomy. Personal injury contribution, pollution compensation, limitation funds, cargo carrier rights, and hull property recovery cannot safely be treated as a common loss pool with one uniform percentage. The personal-injury statute and CLC pollution regime demonstrate why claim type can change the legal basis before cover is tested. This differentiation reduces misclassification of liabilities as “collision damage” simply because they arose from the same incident.
The fourth finding concerns risk finance. An owner-facing P&I indemnity, a club's retained loss, a pooling contribution, and a reinsurer's participation occupy separate contractual levels. The 2026/27 Group reinsurance architecture makes these levels visible, but supplies no shortcut to the insured's terms. Hull and club arrangements should be modeled as a coverage matrix with explicit exclusions and deductibles; their financial relationship should be determined only after legal liability is stated and individual terms are read.
Research findings and agenda
This study establishes five evidence-backed conclusions within its documentary scope. First, an accurate collision analysis requires separate treatment of conduct rule, causation, fault proportion, recoverable damage, liability limitation, and insurance. Second, encounter-specific interpretation of COLREG duties can decisively alter apportionment, as the Ever Smart procedural sequence demonstrates. Third, gross cross-liabilities provide the correct starting point for explaining the effect of fault changes; bilateral netting is a contingent downstream operation. Fourth, a hull running-down clause and a P&I entry must be matched by actual wording and liability category, including possible three-fourths and four-fourths architectures. Fifth, mutual club pooling and reinsurance govern the financing of eligible claims and should not be confused with the insured's legal obligation or the club's initial coverage grant.
Further work can test these findings with anonymized, permissioned claim-level records combining time-stamped navigational evidence, adjudicated causation, gross and net recoveries, full policy wordings, and currency-consistent payments. Research must define exposure denominators and account for selection into litigation and insurance reporting. Controlled comparisons of contract editions and direct observation of changes in settlement structure would help separate true risk effects from changes in accounting presentation. The present monograph supplies the documentary classifications and transparent reasoning necessary to formulate such empirical work, without claiming that it has already been performed.
Conclusion
Maritime collision liability and insurance apportionment form a chain of related determinations. Navigation rules define the conduct expected of vessels; facts and causation determine which breaches matter; comparative-fault law allocates responsibility; the law of recoverable damage and limitation determines enforceable claims; and insurance contracts decide whether, how, and by whom those claims are indemnified. The legal fault percentage and the insurance coverage fraction can differ without contradiction because they refer to different objects.
The historical development from equal-division conventions toward proportionate maritime fault, the distinction between safety investigations and adjudication, the Ever Smart crossing-rule litigation, and the comparison between hull running-down clauses and P&I arrangements converge on a clear methodological conclusion: every percentage and monetary amount must carry its legal basis, denominator, policy edition, and date. An adjustment that preserves those identifiers can be explained and audited; an adjustment that merges them creates misleading claims histories and unsupported coverage assertions.
The research contributes an original integrated interpretation of established legal, nautical, and insurance materials, accompanied by fully disclosed hypothetical and judgment-based calculations. Its conclusions are confined to the verified documents and their jurisdictions. The appropriate next empirical step is a source-complete claims dataset rather than an ungrounded universal collision rate or a newly named theory. This dossier is complete as a scholarly documentary monograph; actual publication remains subject to separate authorization and production verification.
References
European Maritime Safety Agency (EMSA). (2026). Annual overview of marine casualties and incidents 2026: Annual casualty information and reporting scope. European Maritime Safety Agency. https://www.emsa.europa.eu/accident-investigation-publications/infographics-accident-investigation/download/8745/5844/23.html
International Group of P&I Clubs. (2026). 2026/27 Pool and Group Excess of Loss reinsurance contract structure. https://www.igpandi.org/reinsurance/
International Group of P&I Clubs. (n.d., accessed November 17, 2026). Group agreements: The pooling agreement. https://www.igpandi.org/article/group-agreements/
International Maritime Organization (IMO). (1972, as subsequently amended). Convention on the International Regulations for Preventing Collisions at Sea, 1972 (COLREGs). https://www.imo.org/en/about/conventions/pages/colreg.aspx
International Maritime Organization (IMO). (1976, 1996, and subsequent amendments). Convention on Limitation of Liability for Maritime Claims (LLMC). https://www.imo.org/en/about/conventions/pages/convention-on-limitation-of-liability-for-maritime-claims-%28llmc%29.aspx
International Maritime Organization (IMO). (1992 framework). International Convention on Civil Liability for Oil Pollution Damage (CLC). https://www.imo.org/en/about/conventions/pages/international-convention-on-civil-liability-for-oil-pollution-damage-%28clc%29.aspx
International Maritime Organization (IMO). (n.d., accessed November 17, 2026). Casualty investigation. https://www.imo.org/en/ourwork/iiis/pages/casualty.aspx
International Maritime Organization (IMO). (n.d., accessed November 17, 2026). Voyage data recorders. https://www.imo.org/en/ourwork/safety/pages/vdr.aspx
Institute of London Underwriters. (1983). Institute Time Clauses—Hulls, 1 October 1983 (historical market specimen). https://stream.law/wp-content/uploads/2021/07/Institute-Time-Clauses-Hull-1983.pdf
Japan P&I Club. (2025, December 18). International Group Reinsurance Programme for the 2026/27 Policy Year, Circular No. 25-018. https://www.piclub.or.jp/en/news/43244
Marine Accident Investigation Branch (MAIB). (2017). Collision between pure car carrier City of Rotterdam and ro-ro freight ferry Primula Seaways. UK Government. https://www.gov.uk/maib-reports/collision-between-pure-car-carrier-city-of-rotterdam-and-ro-ro-freight-ferry-primula-seaways
Marine Insurance Act 1906, 6 Edw. 7 c. 41, Part XIV, including subrogation and contribution provisions (United Kingdom). https://www.legislation.gov.uk/ukpga/Edw7/6/41/part/14?view=plain
McDermott, Inc. v. AmClyde, 511 U.S. 202 (1994) (United States Supreme Court). https://www.law.cornell.edu/supremecourt/text/511/202
Merchant Shipping Act 1995, c. 21, s. 187 (United Kingdom): Liability in case of collision. https://www.legislation.gov.uk/ukpga/1995/21/section/187
Merchant Shipping Act 1995, c. 21, s. 188 (United Kingdom): Personal injury liability following collision. https://www.legislation.gov.uk/ukpga/1995/21/section/188/2024-04-02
Merchant Shipping Act 1995, c. 21, s. 189 (United Kingdom): Right of contribution. https://www.legislation.gov.uk/ukpga/1995/21/section/189/2024-04-02
Nautical Challenge Ltd v. Evergreen Marine (UK) Ltd [2022] EWHC 206 (Admlty) (United Kingdom, Admiralty Court). https://www.bailii.org/ew/cases/EWHC/Admlty/2022/206.htm
Nordic Association of Marine Insurers (Cefor). (edition-specific online text, accessed November 17, 2026). Nordic Marine Insurance Plan, Chapter 13: Liability of the assured arising from collision or striking. https://nordicplan.org/the-plan/part-two/chapter-13/
Nordic Association of Marine Insurers (Cefor). (edition-specific online commentary, accessed November 17, 2026). Nordic Marine Insurance Plan Commentary, Chapter 13. https://nordicplan.org/commentary/part-two/chapter-13/
The Pennsylvania, 86 U.S. (19 Wall.) 125 (1873) (United States Supreme Court). https://www.law.cornell.edu/supremecourt/text/86/125
UK P&I Club. (2026). Fixed Premium Terms and Conditions 2026 (including collision liability provisions). https://www.ukpandi.com/fileadmin/uploads/ukpandi/00_Documents/Rules_and_Terms_and_Conditions/2026/Fixed_Premium_T_Cs_2026_01.pdf
United States v. Reliable Transfer Co., 421 U.S. 397 (1975) (United States Supreme Court). https://www.law.cornell.edu/supremecourt/text/421/397
United Kingdom Supreme Court. (2021, February 19). Evergreen Marine (UK) Ltd v. Nautical Challenge Ltd [2021] UKSC 6 (case record and judgment). https://supremecourt.uk/cases/uksc-2018-0216
United States Department of State, Office of the Historian. (1911). Convention for the Unification of Certain Rules in Regard to Collisions at Sea (1910), treaty correspondence and text. Foreign Relations of the United States, 1911. https://history.state.gov/historicaldocuments/frus1911/d24