In Chapter 2, we explained that insurance is sometimes conceptualized almost entirely in terms of its function of spreading losses; and that this function is often understood, under the influence of an ‘actuarial model’, in terms of spreading losses across a discrete risk pool. But, as we began to explain in Chapter 2, the nature and extent of the loss-spreading effected by market insurance in practice, and the mechanisms through which it is achieved, do not resemble this picture.
In the present chapter, we focus on the means by which loss-spreading by insurance is achieved, and most particularly on issues of private law which arise in this process. We identify the involvement of a complex series of party transactions and relations and show that losses are spread much further and more widely through insurance than is typically appreciated. One of the key operating factors here is reinsurance, and some core elements of law relating to reinsurance are introduced in 6.4. The issues addressed in the chapter illustrate that risks are by no means spread simply within defined risk pools; and that the pattern of loss-spreading does not necessarily serve, or even match, any overarching purpose or method. The compatibility of loss-spreading with private law concepts is thus enhanced, but at the same time, this raises questions of cost, control, and substantive desirability (and the need for regulation) which are not apparent from the presuppositions of the actuarial model, which de-emphasizes private risk-shifting activity.
While insurance does indeed achieve loss-spreading on a significant scale, it spreads losses not just beyond like risk pools, but also beyond national borders and into different markets. This is one reason why the supposition that loss-spreading goals taken to their logical conclusion must imply that the state should spread losses is misconceived, and we explore some examples of market-state interaction in the final two sections. In this chapter, we touch on the allocation of losses within one or another insurance fund, and the intermingling of losses in different funds. This process becomes stronger as the number of insurer participants increases, and bears little resemblance to insurance in its widely (mis)understood sense of a payment in return for an indemnity from a like risk pool. Indeed, we seek to show that this image is generally speaking (and outside the realms of genuine mutuality) something of a fantasy.
Some tort scholars in particular have approached the risk-spreading effects of insurance in a restricted way, which overstates the tension between insurance and private law duties, and understates the intimate connections between them.1 It has been said that insurance is a closed system which confines benefits to contributors;2 that insurance creates a homogenous risk pool which divides members into groups and charges them risk-related premiums, so that in principle in the long run each member would recover what it has paid in—described as embodying a principle of ‘insured pays’, and thus in effect (in comparison with tort), ‘victim pays’;3 and that each pool is self-financing and does not carry the potential for imposing financial penalties on outsiders.4
The difficulty is that none of this accurately reflects the way that the insurance market operates to spread risks.5 Risk-spreading is much less bounded, and at the same time much more open to creating incentives, than this implies. Both first-party and third-party insurance policies confer benefits on other persons: first-party policies are often designed to protect a range of persons, including associated companies and other contracting parties; and third-party policies rarely make payment to the assured but rather payment is made directly to the assured’s victim, a process required by statute where the assured is insolvent6 or under various compulsory insurance statutes.7 Insurance rarely consists of a homogenous risk pool, in that policies cover a variety of first- and third-party risks and there is no legal requirement, other than in respect of the separation of assets and liabilities attributable to life (long-term business) and nonlife (general business), for premiums to be allocated to classes of risk.8 Further, risks are to some or other extent inevitably reinsured, and reinsurance is made necessary by the fact that individual insurers cannot hope to gain the full benefit of the ‘law of large numbers’. Even though it is acknowledged as the basis of modern underwriting and insurance pricing, the law of large numbers alone is not enough to ensure the survival of the insurer,9 and it is a regulatory requirement in most jurisdictions that an insurer has adequate reinsurance. In practice, for risks to be spread, diversification is required. Even leaving aside the need for diversification rather than homogeneity in risk, the notion that members would in the long run ultimately receive benefits reflecting their premiums in the insurance risk pool is not borne out in practice, in that most policies are short-term; the identity of insured persons changes from year to year; and insurance funds are often used up from large losses affecting a small number of persons.
From the point of view of the relationship between insurance and the law of obligations, perhaps the most important misunderstanding is the assumption that insurance does not impose financial penalties on outsiders and thus—unlike tort, and indeed, contract—cannot operate as a deterrent. A person suffering loss may decide to sue the wrongdoer, and if the claimant happens to be insured then his or her indemnity insurers will do it in their place (indeed, may be more likely to do it than the assured party acting on his or her own account).10 First-party property claims in this way become third-party liability claims; and we dealt with the need to regulate this process and to moderate the way that insurers may seek responsible parties, in the previous chapter. Liability insurers for their part are able to exert a degree of control over policyholders by increasing premiums or, ultimately, withdrawing cover, sanctions which plainly have a significant deterrent effect; and may also involve themselves in the risk management process of their clients. As we saw in Chapter 2, ‘moral hazard’ is an idea which originated in insurance, and reflects the longevity of the idea that insurance must be accompanied by responsibility. Last but not least, it can also be argued that an uninsured person who has nothing to lose if he or she faces liability has less incentive to take care than an insured person who cannot continue their activities without insurance. Insurance, in principle, may be a powerful regulator.11
Contribution exists in a variety of legal contexts,12 and has been explained as a principle of fairness or equity: where many parties owe the same debt, it is inequitable for one to pay without contribution from the others.13 In the context of insurance, it operates to allocate the burden of any loss to all insurers whose policies respond to that loss. The role played in the development of the principles of the law of contribution by cases to which insurers are parties is notable,14 and the application of contribution to overlapping marine insurance policies can be seen as early as 1763.15 Today, insurance contribution takes its place as a particular illustration of the wider equitable principle of contribution, recently restated by the High Court of Australia in Friend v Brooker,16 in the context of ensuring equity as between mutual debtors of the same creditor. Although subrogation and contribution have common equitable roots, a key difference is that insurers make and defend contribution claims in their own right, and thus in their own names.
Contribution between insurers is particularly pertinent to our analysis of loss-spreading, and allows liability relating to a single loss to be divided and the shares to be shifted to different risk pools. There can, for example, be contribution between a motor insurer and an insurer of domestic goods if they happen to cover the same subject matter. Contribution is not the only process through which losses, translated into liabilities, are divided and spread between insurers, but it has a significant role in division and dispersion. Insurers typically seek to avoid the consequences of contribution by express policy terms.
6.3.1 Double insurance
The trigger for contribution is double insurance, a concept which may be defined as any situation in which two or more policies cover a single claim. The practice of taking out overlapping policies developed in the early years of insurance for good and bad reasons. The good reason was the fear that one underwriter might not be able to pay to the full limit of his subscription, so that a second policy might be taken out by way of security. The bad reason was furthering a fraud by deliberately scuttling a vessel and claiming from each insurer, a practice which led at the end of the seventeenth century to the compulsory registration of policies in the Chamber of Assurances in order to pick up cases of double insurance. The growth of prudential supervision has minimized the risk of insurer insolvency, and today double insurance tends to arise ‘incidentally’, or by accident. ‘Incidental’ double insurance occurs where two policies are different in scope but happen to converge in specific circumstances, as where a motorist is insured under his own policy to drive any vehicle, and borrows a vehicle from a friend whose own policy covers any person driving with the friend’s consent. ‘Accidental’ double insurance occurs where the assured is unaware that he has a perfectly good policy and obtains another. This can occur where, for example, a person is unaware that his household policy provides litigation funding cover (before the event, BTE insurance) and takes out a litigation funding policy (after the event, ATE) in respect of a claim that he wishes to bring. It can also occur where travel cover is provided by the assured’s bank account and also by his employers.17
The common law abandoned any objection to double insurance in the middle of the eighteenth century,18 and the marine authorities were codified by the Marine Insurance Act 1906 s 32. The effect of the section is that the assured is entitled to insure as many times as he wishes, and he can claim against one or more of the insurers in such proportions as he wishes, but the principle of indemnity dictates that the total amount recovered can never exceed his actual loss. There is no implied term in a policy that an insurer is only liable for its own proportion of any loss.19
All of this is, however, subject to express agreement. Insurers remain wary of double insurance and seek to control it by contract provisions of three main types. First, the insurers may seek to come off risk entirely in the event that other cover is taken out: there may be a condition precedent to cover that no other insurance exists at the time, and there may be a cancellation clause if some other policy is subsequently taken out. Such terms are known as ‘escape’ clauses. Second, the insurers may adopt a lesser provision under which their liability is postponed to that of other insurers, so that if at the time of the loss there is another insurance in place, that insurance has to pay up to the limits of its cover and only if there remains uncompensated loss does the policy respond. The policy thus becomes an excess layer policy, responding only where other policies are exhausted. Third, insurers may insert a rateable proportion clause, under which the insurer is only liable for a proportion of the loss rather than its full amount.
These clauses work without undue difficulty where only one policy contains them,20 but give rise to intriguing and difficult questions where each of the policies seeks to ‘escape’, to postpone liability or to apportion liability. It suffices here to say that: concurrent ‘escape’ clauses cancel each other out, as otherwise there would be the absurd position that the assured could not recover from either insurer;21 concurrent excess clauses are similarly self-cancelling, as it is plainly not possible for each policy to operate in excess of the others;22 and an escape or excess clause in policy A trumps a rateable proportion clause in policy B, because the former negatives cover entirely under policy A whereas the rateable proportion clause in policy B operates only where policy A remains in force.23
6.3.2 The contribution principle
The Marine Insurance Act 1906 s 80, which codified the common law rules on contribution between insurers,24 provides as follows:
80.—(1) Where the assured is over-insured by double insurance, each insurer is bound, as between himself and the other insurers, to contribute rateably to the loss in proportion to the amount for which he is liable under his contract.
(2) If any insurer pays more than his proportion of the loss, he is entitled to maintain an action for contribution against the other insurers, and is entitled to the like remedies as a surety who has paid more than his proportion of the debt.
Contribution between insurers operates independently of the terms of the contract of insurance (except to the extent that those terms vary the rights of the parties) and indeed of the statutory right of contribution between joint wrongdoers in the Civil Liability (Contribution) Act 1978. That Act entitles a paying defendant to recover from another person liable for ‘the same damage’. It has been suggested that the proper approach to contribution between insurers is now to treat them as subject to the 1978 Act on the basis that their indemnity to an assured is treated as a matter of law as remedying a failure to protect the insured from harm.25 However, it may be said that the 1978 Act was not designed to supersede existing equitable principles and appears to add nothing to them in the context of double insurance. In Bovis Construction Ltd v Commercial Union Insurance Co Ltd,26 David Steel J held (with only brief discussion) that the 1978 Act did not apply to a contribution claim by an architect against a property insurer: in this case, the property insurer was not liable for the same ‘damage’ as the architect, which had breached a duty of care, leading to ‘damage’ consisting of a propensity to flooding. The insurer only owed a duty to pay, not to avoid the damage: this was, therefore, not a liability for ‘the same’ damage. The House of Lords approved that reasoning in Royal Brompton NHS Trust v Hammond.27 In Greene Wood McClean LLP v Templeton Insurance Ltd,28 Cooke J held that a claim lay under the 1978 Act against a liability insurer whose failure to pay a claim against the assured for costs triggered the liability of the claimant solicitors under a guarantee in respect of the assured’s costs. However, the action was not between two insurers liable for the same loss, and the claim was in essence for an indemnity up to policy limits. It is of interest to note that Cooke J held that the same result was reached by the application of the equitable principle that a person who under a legal obligation discharges the debt of another is entitled to recover that sum from the primary obligee.
’Double insurance’—the trigger for contribution between insurers—requires that: the policies cover the claim, whether it be for damage to property or in respect of the assured’s liability; the policies exist concurrently rather than consecutively;29 the person making the claim is an assured under the policies; and the same interest is insured under the policies. Thus there is no double insurance where the owners of a cargo and the warehouse holding that cargo each insure it under separate policies; not only are the policyholders different people but also their interests in the cargo are quite distinct. In such a case, if the cargo is destroyed through the negligence of the warehouse, and the owners’ insurers are called upon to pay, they have a subrogation rather than a contribution action against the insurers of the warehouse.30 There can, however, be contribution even though the policies differ in nature, for example, as between a household policy on goods and a motor policy which happens to cover the same goods while in the assured’s vehicle,31 and differences in the amounts of deductible and financial limits do not affect the principle of contribution although they will affect its amount.
Contribution is only available as between policies which were both in force and enforceable at the date of the loss, so that if one of the insurers has a right to avoid cover or some defence under the terms of the policy, it is immune. The point here is that the assured must have been able to make a claim against an insurer at the date of the loss before contribution is possible. What remains controversial is whether events after the loss affect contribution claims so that contribution is possible only if all of the insurers are liable when contribution is actually sought. To give a simple example, suppose that the assured’s business premises are insured by insurer A and insurer B at the date of the loss, and that each policy contains a condition precedent that the assured notifies any claim within fourteen days of the occurrence of the insured peril. A fire then occurs and the assured notifies A, but not B, within fourteen days. Plainly A is liable to meet the claim, but if A pays and then brings a contribution action against B, B may assert that although it was liable when the fire occurred it ceased to be liable fourteen days later so that there is no basis for a contribution claim. The cases on the point are divided, some favouring the view that a contingent contribution claim arises as soon as the fire occurs and crystallizes when A makes payment,32 whereas others assert that the assured has to have a valid and subsisting claim against B at the date that A makes payment.33 The latter view can draw some analogy with the rules of subrogation and salvage in that the loss suffered by the assured does not from the outset give the insurers any contingent equitable right over subject matter which has been totally lost (salvage) or the assured’s cause of action against a third party for causing the loss (subrogation) and that those rights become vested only on payment,34 but the analogy with salvage at least is weak because if the assured disposes of the subject matter prior to payment his indemnity will be reduced by that amount. The former view disregards the independence of contribution from the policy itself, and recognizes that in many cases the assured will only claim against one insurer meaning that the insurer’s contribution will be defeated almost from the start unless there is a policy term35 which insists upon claims being made against all insurers.
On occasion, an insurer will pay a claim ex gratia in full or in part where another insurer which is liable for the loss is refusing to make any payment. This may typically happen where insurer A’s policy contains a rateable proportion clause so that it is liable for only a part of the claim, but insurer B has denied liability entirely: insurer A may then do the decent thing—pay the entire claim and seek a rateable contribution from insurer B. The difficulty with this approach is that insurer A, having paid as a volunteer, is, as a result of the maxim that ‘equity will not assist a volunteer’, unable to rely upon the equitable right of contribution. Fortunately, the most recent English decisions recognize that if insurer A unsuccessfully argues with insurer B to honour its obligations, insurer A cannot be described as a volunteer and is entitled to be indemnified.36 By way of extension it has been held that if insurer A faces no liability at all under its own policy, but nevertheless makes payment because it is the right thing to do, then insurer A has, by analogy with contribution, an equitable claim for recoupment against insurer B.37
There are at least three main methods of calculating the amount payable under a contribution claim. Which is used varies depending upon whether the policies are of the same type or only incidentally cover the same loss, but that which is most widely used for marine38 and liability39