Picture the following scenario. A policy is placed with a direct insurer, for a multinational in the US, generally on a comprehensive general liability (CGL) basis. The insurer subscribes either alone, or more likely as part of a market. To protect his ultimate net loss (UNL), the insurer purchases excess of loss reinsurance.
There is a claim and the insurer becomes involved in say, a pollution action in the US in which his insured seeks indemnification for clean-up costs. The insurer's attorneys suggest that although the insured's case is questionable and although the coverage may appear not to respond, the insurer will elect to pay a calculated share of the overall claim – which the insurer's attorneys will probably have already recommended. In this way the insurer will avoid a bad faith claim or likely punitive awards that could have been incurred had the action proceeded before a probably partisan jury in a hostile jurisdiction. Again assume that the coverage was wide enough to incorporate cover for punitive damages – a more than likely case since in the US in general, punitive damages are insurable contingent on them being applicable at all, depending on which state has jurisdiction over the action.
At this point the insurer breathes a sigh of relief, at least for the moment. He won't now be involved in a protracted legal dispute, he won't be called upon to make any punitive damages payment and he has no further costs involved. Of course the deal accepted by the attorneys on his behalf was subject to both client attorney privilege and secrecy until it was completed in case, somehow, the attorneys' assessment of how good the case against him was leaked and fell into the hands of the other side's attorneys. If that happens, the story goes, the other side would see for themselves how weak the insured's case was perceived to be, by their own attorneys, and ‘up the ante' on the action. This in turn would increase the cost of the settlement for the insurer and anyone else involved through the chain of reinsurance.
Of course if the reinsurer participates on a proportional arrangement it's likely he will follow those fortunes (settlements). This is, of course, subject to the existence of an effective ‘follow the settlements' clause, though unfortunately things are often not that straightforward. What happens, though, when the gross account is protected by excess of loss reinsurance, with no follow clause? Are such reinsurers liable?
Defences can be raised, the first of which is the claims co-operation clause – should it exist. The first the reinsurers knew of this loss was when they were asked to pay their share. They had no opportunity to avail themselves of defences open to them. Secondly, is the claim actually valid under the insurance and/or the reinsurance contract? If liability is not established against the insurer, then how can liability be established against the reinsurer? The argument goes that if the reinsured agrees a claim with questionable exposure under the cover, then that is the responsibility of the reinsured, not the reinsurer, who can afford the luxury of this argument. (After all, the reinsurer now knows what the maximum liability to him could be.)
What to do?
The reinsurer will often simply refuse to provide cover unless and until the case against the insurer was litigated to a conclusion, thereby proving the loss. (This, of course, could never happen in this example, as the likely adverse verdict was the basis for the settlement in the first place.) From this position of strength the reinsurer may or may not compromise the claim. What has to be realised here, though, is that the reinsurer may also assume liability on this loss via other channels, whether they be retrocessions of subscribing reinsurers, reinsurances of co-insurers on the original risk, pro rata reinsurances and so on. Moreover, the reinsurer may be aware that he may not be able to recover this part of his UNL from his own retrocessionaires should he admit the claim, for the very same reasons he is questioning the loss against him.
There are two additional factors to take into account here; the type of coverage afforded and the position of the particular reinsurer on the layering of the cover. These two may be linked.
The reinsurance cover may be arranged in layers and although generally arranged on a costs inclusive basis in the London market, may be on a costs-in-addition basis. Each of the excess of loss layers may be in a different and unique position depending on how far up the gross loss goes. For example, participants on the first layer would be able to argue that they were being called upon to make payment much more quickly than otherwise would be required, whereas a reinsurer on a higher layer may realise that he would probably have been called upon to provide a higher level of indemnity had the case proceeded and the claim been proved.
The corollary is that the higher the claim goes, assuming the coverage is costs in addition, the lower the overall (indemnity plus costs) retention of the reinsured can become. This is simply because costs in such cases are expressed as a percentage of the relevant indemnity. For example, should the loss result in the reinsured retaining only 10% of an indemnity settlement, as opposed to 20%, because the loss goes higher in the first case then reinsurers might argue that a high, non-contentious settlement would be in the reinsured's favour. In other words, the reinsured's retention is limited to a finite value that would, in all probability, be exceeded by the claim, so the further up the layers, the lower the retained loss. Looking at this another way, if this were the case then the reinsured may as well fight the action ‘tooth and nail' anyway (up to the extent of his vertical reinsurance cover), since then he would have established liability against him. However, in these circumstances an alternative argument of the reinsured not mitigating his loss professionally could be levelled. Of course, this analysis is based on one claim (or policy buy-back to be more precise). What happens if the reinsured should wish to commute a whole portfolio or account with an individual party? In that case similar arguments might apply. Would the excess of loss reinsurers support these attempts? Would they agree to the allocation of the commutation to individual events? What would happen if there was a dispute over policy response? They may argue for a discount based on accelerated payments coming their way, or at worst they may refuse to comply. In these cases, however, the reinsured (more likely a retrocedant) may be able to enlist reinsurers' (retrocessionaires') general support first. The reinsured could even provide numbers and different scenarios before agreeing the assumed proposition, thereby garnering support. Whether or not as part of this exercise, and within the terms of the original commutation deal, he may be able to divulge the identity of the other party is less clear. For example, the assumed commutation may be as a result of some disagreement over ultimate liability. Would the original insured or reinsured want that information implied or disclosed? What effect may that have on the rest of his original market?
There are various methods of allocation of a commuted payment to excess of loss reinsurers. Without going into detail, these range from basic vertical event coverage (with perhaps a deduction for present value discounting) via pro rata to ultimate loss or reserves to actuarial allocation. For an excellent explanation of this subject the reader may want to review page 107 onwards of the recent Insurance Institute of London's ‘Developments in Excess of Loss Reinsurance' book.
The market has avenues, including various committees and interested parties, to explore these issues and come up with an overall formula. The current position – ranging from ad hoc agreements which may not be enforceable somewhere down the retrocession chain, to refusal of support – leads to uncertainty and ultimately more expense. In the run-off market in particular, such situations only serve to defeat the real reason for the crystallisation of losses in the first place: to shorten the tail. In these cases reinsurers may be arguing against affording coverage on the one hand and then find themselves on the receiving end on another case/book of business.
Reverting to the single loss example, we could speculate if there would be a consensus. The problem in finding a consensus is that any such agreement would be against the current precedental case law and, as such, each party within the market may decide not to adopt it. This is particularly true of companies that have no interest in current business, and may wish to avoid liability at all costs. For them to agree to such a proposition, it would probably need to be to their net financial benefit.
These companies may feel that they would need to make an assessment of all participations in that loss occurrence. However, it's still not that straightforward. Such participations may be through retrocessions for which the retrocessionaire may have little idea whether the retrocedant would participate in the original agreement. If the retrocedant does participate then clearly his loss may be more or less than that advised depending on the basis of his advice.
What about the other losses?
If the loss is then put through the market, and contingent on the standard adopted, this may free up some coverage. Effectively this is a moot point on a paid basis. If the cover was not exhausted incorporating the figures for the disputed loss, then other losses would be collected on it anyway. If the cover was exhausted (ex the disputed loss), then the reinsured may already be arguing – perhaps justifiably – that the contract should be paid in full since if he took out the disputed loss then the contract limits would be recoverable. Furthermore, the replacement of those losses may result in increasing values of the UNLs on other losses, and in different accumulations of other market losses. There would clearly be additional considerations if the cover was on a costs-in-addition basis.
Two final questions arise from this: could these issues exist on current covers and how do insolvent estates manage? A reinsured would do well to assess his wordings and contact his broker to ensure his position relative to this issue is protected. Would reinsurers follow on compromise/ commutation settlements? Is there a
As for insolvent estates, we need to briefly review the case of Charter Re v Fagan. Here the argument turned on whether or not a company (Charter Re) could claim on its reinsurance even though it had not physically settled the losses. Since Charter Re was in liquidation, the estate would have been unable to secure recoveries based on assumed claims had they been unable to enforce the contract to include ‘paid' but not physically settled losses. Therefore the creditors would have, in effect, been penalised by both the company being put into liquidation (only payment as a creditor dividend) and by them being unable to participate in the reinsurer's protection (reducing the size of the estate to be distributed).
The interpretation of the clause had implications beyond insolvent estates because the result would be applicable to all parties to such contracts. The consequence was that the claims would be included in calculations of the UNL against the reinsurers, as long as the reinsured was liable to pay the loss rather than actually having paid it. As a result of this judgment a wordings clause was drafted that reinforced the intent of such contracts, stipulating that reinsurance recoveries could only be secured after claims had actually been physically settled, unless the reinsured was an insolvent company, in which case only the liability to such claims would need to be established.
Therefore for insolvent estates the question for their reinsurers to ask is always “has the liability been established against the insolvent estate (whatever its capacity) on such claims?” and not, post-Charter Re, “have they also physically paid it?”. This differs from the solvent companies post-Charter Re (assuming the new wordings clause has been incorporated) where proof is required that they were liable, and the claim values can only be incorporated into the UNL after physical settlement. Therefore insolvent estates are in the same position relative to this issue.
There is a great deal of expenditure in the aggregate reviewing compromise settlements (whether based on a single loss or across a portfolio) and determining their impact on excess of loss reinsurances. Perhaps one compromise settlement can be distinguished from another and different market UNL formulae applied, or perhaps a general formula can be devised. It does seem strange that an insurer appears to be encouraged to take a loss through the courts so that it can perfect a claim settlement against its reinsurers, when the alternative is to mitigate the loss. The simplest way would be for the reinsurers to participate in the decision-making process of their reinsureds. Since wholesale agreement to this would cause consternation to the reinsured's attorneys, particularly in a subscription market, it may be sensible for contracts to appoint underwriters' representatives. In those cases the representatives could liaise with the leader and determine whether or not the deal is good for the market. If agreed, the rest of the market is bound, and any argument is between the leader and followers only. However, what about the retrocessional market? Would retrocessionaires agree to this arrangement? Again, a market-wide agreement to all such cases may make sense. Clearly what is meant by the ‘retrocesional market' also needs to be examined. London market retrocessionaires may agree but then be prejudiced in their relations with overseas retrocessionaires.
This could be, with some innovation, extended to commutations. Here, however, the representatives would have to agree that the commutation is reasonable and the allocation would need to be determined separately by, for example, an independent consultant or actuary. To save money, the commutations may only require ratification if they exceed a finite value or a percentage of the overall paid values across the portfolio and/or if the allocation to individual losses is not based on a prescribed formula, etc.
As for losses currently in the market but in dispute, a market-wide agreement could be negotiated (for example, to agree to a percentage of the gross loss being applied). The alternative is either to unwind – refund – all such losses through the market, leaving the original insurer ‘high and dry', or to leave the position of these losses where they currently reside, i.e. to make them fully paid. Both of these are further fraught with practical difficulties.
Therefore as with many issues in reinsurance, the answer to whether commutations reduce the tail is ‘it depends'. A commutation (including policy buy-backs) will reduce the tail if:
In other circumstances the payment of compromise or contract commutations by reinsurers of the original liability party to the compromise or commutation is questionable. Furthermore, how such losses to ceded excess of loss contracts are allocated is uncertain. In such cases it is not axiomatic that an original compromise settlement or commutation will result in a shortening of the tail.
Craig Karlson is a reinsurance consultant with more than ten years' London market experience.