Year 2000 and date recognition problems

ISSUES FOR REINSURERS

Anthony Hobkinson

October 1998

1.YEAR 2000 AND IMPLICATIONS FOR REINSURERS

INTRODUCTION
The Year 2000 problem has arisen as a direct result of computer technology developed in the 1970s which saved valuable memory by designating dates by the use of two digits rather than four. The result is that many computer systems and embedded software cannot differentiate between the Year 2000 and 1900. Although everyone is aware of the problem in general terms, surprisingly there is still a widescale lack of appreciation of the specifics. For example, many people mistakenly believe that:

-the problem only affects computers

-it will only strike at midnight on December 31st 1999.

-if they bought a computer or electronic product in the past five years it

must be Year 2000 compliant

-they can defer taking remedial action until the middle of 1999.

The reality is of course that an indefinable amount of equipment may be affected, from the most sophisticated multinational computer network, down to a basic word processor, fax machine, telephone, or video recorder. Just think for a moment how many safety systems, such as thermal protection switches, rely on software to function.

Losses have already occurred due to system malfunctions in processors relating to date sensitive information, for example the miscalculation of a leap year in 1996 caused the breakdown of an aluminium smelter process in New Zealand causing US$1 million of damage. In Europe, drugs were destroyed by a manufacturer where a computer system believed the drugs to be out of date, the computers having interpreted the Year 2000 as 1900. Elsewhere, credit cards with an expiry date of "00" have been rejected because the card reading machinery believed the validity period ended in 1900. Such is society's reliance on technology that the potential for disruption is frightening. The result is that such exposures are apparently incapable of quantification.

It follows from the above that the insurance industry, and therefore the reinsurance industry, face a unique challenge. While the industry knows the problem is here, and its impact is in a very general sense foreseeable, the precise impact of Year 2000 cannot be predicted. This is not only because of the all pervasive nature of electronic technology, but also because none of the experts can agree how many companies will be able to avoid problems by achieving Year 2000 compliance over the coming months. Additionally, as it is a new phenomenon, there are no claims trends which would enable an actuarial analysis to be completed.

While reinsurers have the option (at least in theory in the current soft market) to exclude Year 2000 losses, the pattern beginning to emerge is that most reinsurers will not do so, albeit that they may protect themselves by way of aggregate limits (see later). To some extent, therefore, reinsurers' futures will rest with those of the direct market, and it logically follows that the starting point in considering reinsurers' exposure is an analysis of the problems which the direct market faces.

2.PRELIMINARY COMMENTS IN RELATION TO THE YEAR 2000

Are Year 2000 losses fortuitous?

The nature of the Year 2000 problem, if not the degree, is foreseeable and has been known about (at least by certain IT professionals) for some years.

Further, in most cases it is within the power of businesses either to prevent the impact of Year 2000 on them entirely or at least to minimise the impact by replacing non-compliant systems. It will, however, be impossible to eradicate risk completely even if the business is Year 2000 compliant given its dependency on other business partners who may not be compliant.

The question arises as to whether losses resulting from Year 2000 problems are losses to which primary insurance contracts should respond at all, their being apparently foreseeable and potentially preventable.

Are Year 2000 losses in fact insurable events? The broad rule is that insurance responds to risks and not certainties. In Prudential Insurance -v- Inland Revenue Commissioners 1904 2KB 658, It was said that "there must be either uncertainty whether the event will happen or not, or, if the event is one which must happen at some time, there must be uncertainty as to the time at which it will happen". The Court of Appeal decision of Soya -v- GmbH (CA 1982 1LLR122) made it clear however that the question of uncertainty must be judged from the perspective of what the insured knew or should have known.

Although the use of non-compliant computer systems will almost certainly lead to disruption of some sort, the timing of the disruption and its consequences will not be so predictable. For example, an insured may simply not appreciate that a piece of office machinery is reliant upon date sensitive data to function.

If the only logical consequence of system malfunction is the corruption of data in the computer, then that "damage" may be inevitable. The damage would arguably not be fortuitous and would therefore not be covered by insurance policy in force at the time. However, it is likely that the result of Year 2000 malfunction will be damage or loss to other property consequential upon the system malfunction. It does not necessarily follow that the consequential damage which occurs will have been inevitable and the loss may then be covered by the policy, it being accidental.

Although the facts giving rise to each loss will have to be looked at carefully to consider whether the insured knew that such a loss would occur, to rely solely on lack of fortuity on the basis that the Year 2000 problem as a whole is broadly foreseeable, is unwise. It would seem that this is now accepted by the market who have employed other options in seeking to reduce their exposure to such claims.

3.CLASSES OF DIRECT INSURANCE POTENTIALLY TRIGGERED BY YEAR 2000 LOSSES

There are significant potential exposures for London and International insurers and reinsurers from Year 2000 claims. These exposures arise from a wide variety of classes of direct business including Property, Directors and Officers, Professional Indemnity, Product Liability and Business Interruption.

Insurers face greatest exposure form policy holders in industries such as manufacturing, financial services, utilities, healthcare facilities and software where reliance on embedded chips is significantly greater than other low tech industries.

Research has also shown that insurers and indeed their reinsurers will face heightened exposure if the policy holder is located in certain "high risk" countries such as France, Germany and Japan as well as the Philippines, Russia and China.

Set out below are examples of types of insurance policy likely to respond to Year 2000 losses. It should be stressed that this is by no means an exhaustive list. The examples are given principally to show that this is not simply an issue for software houses, computer consultants, and product manufacturers.

All Risks Policies

Where the insuring clause provides for "Accidental loss, destruction or damage to insured property as described in the schedule to the policy"

There is some uncertainty as to whether corrupted computer programmes will constitute damage to property. There is Court of Appeal authority to the effect that the reconfiguration of magnetic particles on a metallic disk can be damage to property (R -v- Whiteley 1993 FSR 168), but it is submitted that this outcome was reached on the particular facts of the case and is not support for the general proposition that corruption of data is damage. However, any physical damage arising from a Year 2000 malfunction and any consequential financial loss would prima facie fall to be covered under a typical All Risks (Business Interruption) policy.

Product Liability

Insureds under product liability policies are covered for liabilities arising from damage caused by their products. Where an insured supplies a computer controlled device, for example a piece of machinery, to another party and where that device is affected by Year 2000 and causes injury or property damage as a result, a product liability claim may lie against the insured and thus against his insurance in respect of those losses.

It should be emphasised however that most policies will not cover either damage to the product itself or to the cost of recalling it should that prove necessary.

Professional Indemnity Insurance

Such policies customarily provide cover for breach of professional duty arising from neglect, error or omission by the insured. The words "neglect, error or omission" were examined in the case of Wimpy -v- Poole (1984) in which it was found that those words encompassed negligence but additionally some omissions and errors occurring in the absence of negligence, in particular breach of contract.

The most exposed class is that of computer consultants who have written or advised clients in relation to the choice of software. They may be vulnerable to claims both in negligence and breach of contract. Of particular relevance in claims brought in negligence will be the date on which the computer consultants became aware or should have become aware of the Year 2000 problems and have advised clients on the implications thereof.

A more indirectly exposed group is the legal profession: solicitors may face liability where a systems failure arising out of date recognition has resulted in a missed limitation deadline. They may also face liability where they have advised on software contracts in relation to the purchase of a business, without drawing the implications of the Year 2000 issue to their client's attention.

Directors and Officer's Liability

A typical insuring clause to be found in a London market Directors and Officer's liability insurance policy will provide cover for a "wrongful act". A "wrongful act" will be defined within the policy wording but will often include wording such as "breach of duty, neglect, error or omission", encompassing both breach of contract and negligence. If a company suffers financial loss caused by a failure of its own computer systems resulting from the Year 2000 problem, directors may face liability for breach of fiduciary duty (e.g. to shareholders to protect their investments) to which such a policy is likely to respond.

There seems to be a consensus emerging in the market that D&O and E&O policies will have to respond to most Year 2000 claims for negligence or wrongful acts and certain claims for property and business interruption losses will also be covered.

An alternative way of looking at the problem is to divide classes of insurance up into those that are directly exposed in that the purchasers of those policies deal directly with software for example, and those where the exposure is indirect.

Examples of directly exposed classes would be:

• Computer consultants
• Management consultants
• Product manufacturers (but see above as to the need for "damage")

Examples of indirectly exposed classes will be:

• Directors and officers
• Credit Insurance
• Accountants
• Solicitors
• Insurance Brokers

Even in the absence of exclusion clauses, it will not always be the case that Year 2000 losses will be covered by such policies. The direct insurers' liability will be determined by a host of other factors including whether the loss was foreseeable at the time the contract was made, the wording of the operative clause, existing general exclusions, compliance with warranties and obligations of disclosure and also the impact of limitation issues on any claims.

Subject to what is said above, however, on the basis that Year 2000 losses are covered by the primary policy, reinsurers (whether proportional or non-proportional) of carriers writing the above classes of business also therefore face potential Year 2000 losses.

4.REINSURERS' RESPONSE

4.1Faced with these issues, the reinsurance market has contemplated, and in some instances adopted, various options. These include:

Specific exclusions in reinsurance contracts. These can be either blanket (eg NMA 2800) or include some element of write-back (eg NMA 2801)

Aggregate limits for Year 2000 losses. This can give rise to difficulties in practice particularly where the Year 2000 failure is any one of several proximate causes.

The imposition of underwriting guidelines on the reinsured - for example a requirement that the reinsured takes all reasonable steps to ascertain the Year 2000 compliance of the insured.

4.2Of all these issues, it is perhaps those of exclusion clauses and the perceived difficulties in applying existing aggregation clauses to Year 2000 losses which are causing the biggest debate.

5.EXCLUSION CLAUSES

5.1The use of these clauses may be unattractive in the soft market that exists today and in the absence of a market trend to do so. It may however be the only way of containing exposure where reinsurers are not satisfied that the underlying insured business is likely to be compliant.

5.2In terms of the drafting of any Year 2000 exclusion clause, it is important, given the nature of the types of Year 2000 related losses which are to be excluded, many of which will be remote, that the clause is clear and unambiguous and covers all direct and indirect losses arising not only from a date change from 31st December 1999 to 1st January 2000 but also out of the fact that the Year 2000 is a leap year and in relation to other potentially troublesome dates. Reference should be made to recommended industry/ market wordings, but each will need to be tailored to fit the specific wording to which it attaches.

5.3The scope of such exclusions clauses are likely to be the subject of much litigation in order to determine the extent of cover. Given the contra proferentem rule, it is vital that reinsurers include very clear wording.

5.4A major difficulty in applying any form of Year 2000 exclusion clause will however be to identify those losses that relate to Year 2000 incidents. This will be the case where an originating Year 2000 malfunction leads to a typical physical damage type claim. Matters will be complicated where human error contributes or operates in conjunction with the original malfunction to cause the loss. It is likely that expert evidence from the IT industry will be required to determine such causation issues.

5.5This also raises a practical problem for reinsurers impacting on the effectiveness of exclusion clauses in relation to claims processing. Reinsurers should be alive to the need for proper claims processing and/or reporting to ensure that the cause of each loss is very clearly identified in any bordereau. This will avoid the possibility of the acceptance of losses which have the appearance of typical claims covered by the policy, but are in fact consequential losses arising out of year 2000 excluded malfunction.

6.AGGREGATION

6.1Certain insurers of industries heavily reliant on machinery incorporating embedded chips, for example, may face a large number of claims resulting from Year 2000 malfunctions in the period of the policy. In certain circumstances, the reinsured may wish to aggregate these losses for the purpose of presenting a claim to its reinsurers. This is likely to be the case if the deductibles and upper limits per event/occurrence are high. Entitlement of the reinsured to aggregate will depend on the nature of the reinsuring clause and the presence or otherwise of an aggregate extension clause.

6.2The majority of London market excess of loss treaties contain clauses providing for recovery in excess of a certain amount up to a certain limit "for each and every loss". The definition of each and every loss will determine the reinsured's ability to aggregate.

6.3Events Based Clauses

A variety of aggregating words are used including "event", "accident" and "occurrence". The text below focuses on the use of "event" in this context, in light of a number of important decisions on that definition.

6.3.1Caudle-v- Sharp [1995] LRLR 80, reversed [1995] 1 LRLR 433

6.3.1.1The facts of this case will no doubt be familiar to most reinsurers but for completeness, are repeated here.

Mr Outhwaite was the active underwriter of two Lloyd's syndicates and between 1980 and 1982, he wrote 32 reinsurance contracts covering the run-off of other Lloyd's syndicates and insurance companies. The reinsurance contracts were disastrous, largely as a result of the fact that the risks reinsured included US long tail asbestosis claims. Names on the two syndicates concerned brought proceedings against Mr Outhwaite, the managing agency and certain members' agents for negligent underwriting. The proceedings were settled in 1992 and the settlement was paid in part by the E&O underwriters of Mr Outhwaite and the underwriting agency. This E&O cover was procured from other Lloyd's syndicates including that of Mr Sharp. E&O insurers were reinsured for the relevant year with other Lloyd's syndicates including that represented by Mr Caudle under 4 excess of loss policies in substantially the same form. The excess of loss policy provided cover for "each and every loss" defined as meaning "each and every occurrence, catastrophe, disaster or calamity arising out of one event".

6.3.1.2The reinsurers required Mr Sharp to bear a retention, the cover being triggered by losses in excess of that retention. The problem for Mr Sharp was however that the liability sustained by him in respect of the 32 contracts were below the retention level whereas when aggregated, they exceeded the retention level. This therefore begged the question of the exact nature of the "event" as identified losses that could be aggregated.

6.3.1.3At first instance, the "event" for the purpose of professional liability cover was continuing failure by Mr Outhwaite to take proper steps to investigate the underlying problem of asbestosis (which had been characterised previously in the arbitration as Mr Outhwaite's "blind spot"), on that basis, the 32 losses were capable of aggregation and there was one event out of which the losses arose, for the purposes of the reinsurance.

6.3.1.4The first instance decision was overturned in the Court of Appeal in which Evans LJ gave the leading judgement. He held that there were 3 requirements that needed to be satisfied before there can be said to be an "event" under this type of Excess of Loss contract:

(i) There must be a common factor can be properly described as an "event" under the policy. It was implicit in Evan's reasoning that in order to qualify as an event, the common factor must be limited in time and space. Mr Outhwaite's "blind spot" could not be an event as it would have been happening without a beginning or an end. His failure to carry out an investigation into the problem of asbestosis was an omission capable of constituting negligence but did not actually become a negligent omission (to which the underlying E&O policy responded) until he took the decision to underwrite each of the 32 policies. Mr Outhwaite's "blind spot" did not fall within the ordinary meaning of the word "event", except by reference to each and every occasion on which he entered into an insurance contract. Each decision to underwrite had to be the event in this case.

(ii) The common factor had to satisfy the test of causation. A "pragmatic" approach was adopted by the Court of Appeal in this regard. The words "arising out of" were not to be construed as referring to the direct and proximate cause of the loss in the strict sense (as that test applied to occurrences and not events of which occurrences formed a part). If the "event" in this case was to be regarded as the negligent underwriting, it was arguable on the facts that the losses arising out of each of the 32 contracts were caused by this "common factor".

(iii) However, the broad test applied on causation was cut back by the last requirement which was that the event must not be too remote from the losses. In this case, if the initial omission (i.e. instance of negligent underwriting) was regarded as the "event" the next 31 contracts underwritten by Mr Outhwaite and causing loss were too remote to be regarded as "arising out of" that event.

6.3.2Axa Reinsurance (UK) Plc -v- Field [1996] 2 Lloyd's Rep 233 (HL)

6.3.2.1The definition of event was indirectly confirmed in this case by the House of Lords where Lord Mustill stated that "in ordinary speech, an event is something which happens at a particular time, at a particular place in a particular way"

6.3.3American Centennial Insurance Company -v- INSCO Ltd 1996, unreported

6.3.3.1This case concerned the collapse of a Savings and Loans Association arising mainly out of the financial mismanagement by 14 officers of the Association. The Association had procured liability cover for its officers on which claims were subsequently made. The insurer paid 14 different claims in relation to each of the 14 officers as did the reinsurer. The reinsurer then claimed to be entitled to aggregate all of the 14 losses for the purposes of a claim under its excess of loss retrocession cover on the basis that all arose out of one event i.e. the collapse of the Savings and Loans Association. It was held that the relevant liability was incurred as a result of the officers' individual acts and omissions during their conduct in office. The collapse did not cause the claims to be made and could not be treated as an event. Each act and omission of each of the 14 officers was an event in its own right and therefore no aggregation was allowed of what was 14 events.

6.4Year 2000: an Event?

6.4.4.1Take for example a firm of computer consultants who have advised various clients over the years in relation to the choice of computer software/hardware. Would it be possible for a professional liability insurer of that computer consultancy to aggregate claims arising out of the negligent advice of one consultant to a number of different clients in relation to the Year 2000 issues, for the purpose of presenting a claim to its excess of loss reinsurers? (I am assuming here that the aggregation wording in the reinsuring clause uses the word "event".) Applying the three tests laid down by Evans LJ, such losses are unlikely to be capable of aggregation on this basis.

6.4.4.2As with Mr Outhwaite's "blind spot" the consultant's negligent approach to Year 2000 IT advice is unlikely to be considered an event given that it constitutes a continuing state of affairs, lacking a specific point in time and space as is required. On this basis, each instance of negligent advice concerning Year 2000 is likely to be the event for this purpose. Such an omission may be capable of constituting negligence but was not in fact a negligent omission until the advice was given.

6.4.4.3Under the broad test imposed by Evans LJ as to causation, it probably could be said that if the negligent advice was the "event" (as the insurers would argue), the losses sustained in relation to the various claims of negligent Year 2000 advice, could be said to arise out of the "event". However, even if the first instance of negligent advice could be regarded as the "event" the next and following instances would be too remote to be regarded as "arising out of that event" for the purposes of aggregation (different clients, different contracts, different dates).

6.5Cause Based Aggregation Clauses

6.5.1Some LMX reinsurance contracts provide for losses to be aggregated by reference to a common cause rather than by reference to a common event. These wordings offer more scope for aggregation of Year 2000 losses than event based wording, "event" being limited in time and space whereas "cause" is not necessarily so limited. Cause based wordings are however comparatively rare in contracts provided by LMX underwriters or brokers operating in the LMX market but they are not unknown. They are possibly more common in contracts drafted by US reinsurers and brokers.

6.5.2It is instructive for reinsurers to consider the judgment of Phillips J in Cox -v- Bankside (1995 CLC 180) which considered the E&O policy of a members' agency for its liability in negligence to Names placed on certain Gooda Walker syndicates. The insurance provided a maximum sum for "any one occurrence or series of occurrences arising out of one originating cause". It was held that the approach to underwriting of each of the three Gooda Walker underwriters during the years in issue was one originating cause and, accordingly, the court looked at each underwriter separately but aggregated the acts of each into one originating cause (NB this case was decided after the first instance Clarke J decision in Caudle -v- Sharp but before it was overturned on appeal).

6.5.3Phillips J's view seems to have been given some support in Axa -v- Field which concerned excess of loss reinsurance of the direct policy in Cox -v- Bankside. The reinsurance allowed for aggregation of losses "arising from one event" in contrast to the direct policy which allowed for aggregation arising out of one "originating cause". In Axa, the reinsured argued that "event" should be given the same meaning as originating cause in order to secure back to back coverage thereby depriving the word "event" of the meaning given to it in Caudle -v- Sharp. Whilst the Court of Appeal was prepared to do so, the House of Lords was not, on the basis that there was no assumption of back to back cover in relation to non proportional excess of loss reinsurance and the underlying direct policy, the underwriting considerations for each being very different.

6.5.4For present purposes, this meant that the court did not have to reconsider what was meant by "originating cause". Lord Mustill however drew the following helpful distinction between event and originating cause when he commented:

"In my opinion these expressions are not at all the same for two reasons. In ordinary speech, an event is something which happens at a particular time, at a particular place, in a particular way. A cause is to my mind something altogether less constricted. It can make continuing state of affairs ..... the word "originating" was in my view consciously chosen to open up the widest search for a unifying factor in the history of losses which it is sought to aggregate".

6.5.5Lord Mustill's comments on the meaning of "originating cause" were applied in the case of Municipal Mutual Insurance Limited -v- Sea Insurance Limited (26th March 1998, unreported). This case concerned excess of loss reinsurance contracts covering the reinsured's interest in legal liability policies which it had issued to the Port of Sunderland. The conditions of the reinsurances were expressed to be the same as those of the original policies and the original policy applied a single limited indemnity to "all compensation payable to any claimant or any number of claimants in respect of or arising out of all occurrences of a series consequent on or attributable to one source or original cause". The Port was held liable to compensate the owners of two drag line excavators which had suffered damage as a result of a succession of individual acts of pilferage and vandalism carried out by a number of individuals or groups probably acting independently of each other.

6.5.6It was held by Hobhouse LJ that the reinsured was entitled to make an aggregate presentation of the loss to its reinsurers because there was "a clear unifying factor in the history of all the losses" which the owners of the goods had sustained. The Port Authority's want of care was found to be the consistent and necessary factor allowing the individual acts of pilferage and vandalism to occur, (it having no adequate system to protect the goods).

6.6Year 2000: an Originating Cause?

6.6.1Taking the example of the negligent computer consultant, and assuming the aggregating provision was in fact on an originating cause basis, would the reinsured be entitled to aggregate several losses arising from separate instances of negligence in relation to Year 2000 advice by one consultant for the purpose of presenting a reinsurance claim?

6.6.2Following the decision in Cox -v- Bankside, it might be argued successfully that the negligent approach to year 2000 advice of the individual IT consultant was the unifying factor and that all losses arising therefrom could be aggregated into one. This may mean that the reinsured would bear only one retention and the reinsurance cover would be penetrated, shifting the cost of Year 2000 losses to reinsurers. The more restrictive word "event" may therefore be preferable from a reinsurer's perspective.

6.7Conflicting aggregation provisions in insurance and excess of loss reinsurance

6.7.1To continue the example of the negligent computer consultant, what is the position on aggregation of year 2000 losses if the liability insurance aggregates on the basis of originating cause but the reinsurance aggregates per event?

6.7.2Reinsurers are afforded some degree of protection following the decision in Axa -v- Field which dealt with exactly this issue. It was held that there can be no application of the back to back presumption in construing words determining the method of aggregation in excess of loss reinsurance recognising that had the reinsurers intended the methods of aggregation to be identical, it was open to them to use identical wording, which they had failed to do. There was a manifest intention to achieve different ends with regard to aggregation.

6.7.3It should be noted that this decision concerned non-proportional excess of loss reinsurance which was in no real sense "back to back" with the underlying policy (different considerations being applied by insurers and reinsurers in fixing aggregate limits). With regard to proportional reinsurance, however, there must be a strong presumption of back to back cover since the nature of the reinsurances is such that the reinsurer is sharing the risk borne by the direct insurer. It was therefore likely that the treatment of multiple losses was meant to be the same.

6.8Discretion As To The Basis Of Aggregation: "Sole Judge" Clauses

6.8.1Some excess of loss covers, rather than leaving the determination of the aggregating event to the general law and thus potential dispute as to the correct basis of aggregation, empower the reinsurer to determine the basis of any aggregation.

6.8.2These clauses are known as "sole judge" clauses and have been held to be effective by the court of appeal in the case of RE Brown -v- GIO Insurance Limited (18th February 1998). This case concerned an excess of loss reinsurance contract which incorporated the following wording:

"The reinsured's definition of each and every loss and/or event shall be final and binding on the reinsurers hereon. The reinsured shall be the sole judge as to what constitutes each and every loss and/or one event"

6.8.3The reinsured argued that this clause enabled it to decide how losses should be aggregated. It was held that the clause had precisely that effect and that it was a matter of commercial expediency that a problem as difficult as aggregation should be determined by the reinsured alone and without the need for legal proceedings. The judge accepted that the reinsured's right was subject to fairness, reasonableness and bona fides and it remained open to the reinsured to challenge any determination by the reinsurer on these grounds.

6.8.4Clearly, a reinsured is likely to adopt an approach to aggregation which is more advantageous to himself than to reinsurers and to maximise reinsurance recoveries where possible. A number of problems arise out of this, particularly with regard to a reinsurer's right to challenge the reinsured's discretion. The case supports the proposition that the reinsured's exercise of discretion is only unreasonable where the reinsured's determination is plainly wrong as a matter of law. In view of the concerns arising out of aggregation of Year 2000 claims, and the unknown extent of exposure, it may be preferable not to employ "sole judge" clauses. Whilst in theory one could draft any such clause in reverse, naming the reinsurers as the sole judge as to what constitutes an event or originating clause, it is questionable whether such a clause is likely to be commercially viable.

6.9Aggregate Extension Clauses

6.9.1These clauses originated in liability insurance and served the purpose of allowing the assured to aggregate a number of small and similar losses not flowing from a common event (each of which on its own would not be of sufficient size to exceed the deductible and to give rise to a claim). An insurer may decide to include an aggregate extension clause in relation to Year 2000 losses.

6.9.2If there was a Year 2000 computer malfunction in a factory causing personal injury and property damage, the various claims arising from the explosion would all be regarded as losses flowing from a single event and the aggregate extension clause would not be relevant.

6.9.3If a product incorporating an embedded chip which is non-compliant malfunctions and the product (being dangerous to users) gives rise to individual claims of small amounts, the claims will fall below the retention level. The individual claims in this case would not all be caused by a single "event" and each claim would be an event in its own right. In this situation the aggregate extension clause would operate to allow the manufacturers to add together all of those claims in order to exceed the retention.

6.9.4The aggregate extension clause is an important provision often seen in London market excess of loss reinsurance protections. It allows reinsureds to present claims in the aggregate to their reinsurers where liabilities are incurred by the reinsured for losses on risks covering on an aggregate basis - effectively allowing for the carrying through into the reinsurance protections of the same principle of aggregation as exists in the underlying policy.

6.9.5Given that reinsureds are likely to seek to aggregate Year 2000 losses, there is scope for abuse of these clauses where the underlying policy is not what is traditionally understood to be a risk covering on aggregate basis policy (i.e there is no aggregate extension clause). The Court of Appeal has recently heard two cases on what constitutes such a policy in the matters of Denby -v- English & Scottish Maritime Insurance Company Limited and others and Yasuda Fire & Marine Insurance Company of Europe -v- Lloyds Underwriting Syndicates Nos 229 and others (Court of Appeal March 5th 1998). These cases concern the use of aggregate extension clauses in relation to professional indemnity liability policies issued to firms of accountants under the Minet Professional Liability Lineslip. The Court of Appeal held that for a reinsured to take advantage of the AEC in its reinsurance contract and to present claims in the aggregate, the underlying cover must have been provided on an aggregate basis. An aggregate retention is a feature not conclusive of such a type of policy. Where there is no "each and every claim" provision or where there is another provision which provides for the aggregation of claims in one respect or another, this would probably lead to the conclusion that the policy provides cover on an aggregate basis. A per claim excess and limit is not indicative of an aggregate policy.

7.REINSTATEMENT CLAUSES

7.1Reinsurers may wish to specify that reinstatement does not apply to Year 2000 losses.

8.FOLLOW THE SETTLEMENTS

8.1The question also arises as to whether a reinsured's Year 2000 loss settlements are binding on its reinsurers. If the reinsurance contract is silent on the matter and does not contain a "Follow the Settlements" clause or indeed a "Follow the Fortunes" clause (the latter being more appropriate to proportional reinsurance and the former being more commonly found in excess of loss non-proportionate contracts) the reinsured will need to establish that the Year 2000 loss falls within the direct policy as well as the reinsurance policy.

8.2If the policy includes a straightforward unqualified "Follow the Settlements" clause, the reinsurer will be obliged to indemnify the reinsured where it has settled a claim, provided the claim accepted falls within the risks covered by the reinsurance policy as a matter of law and , in settling the claim, the reinsured has acted honestly and has taken all proper and business like steps in making the settlement (Insurance Company of Africa v Scor (UK) Reinsurance Limited).

8.3On the basis that the contract contains a "Follow the Settlements" clause which stipulates that the loss must be within the terms of the original policy and the reinsurance, reinsurers do not automatically have to follow a Year 2000 loss settlement. Rather, they would be entitled to question whether, as a matter of law, the original Year 2000 loss properly fell within the terms and conditions of the original policy. This is a vital protection for reinsurers upheld by the House of Lords in the decision of Hill -v- Mercantile & General Reinsurance Company Plc [1996] 1WOR 1239 (HL) This will enable reinsurers to ensure that their reinsureds have properly addressed all valid policy defences available on or in respect of Year 2000 claims rather than simply passing on the risk "on the nod" to the reinsurers. Additionally, if the reinsurance contains an exclusion clause in respect of Year 2000 claims, reinsurers will not be obliged to follow the settlements made by reinsurers of such claims.

9.YEAR 2000 AND TRIGGERS OF COVER

9.1How and when reinsurance policies are triggered by Year 2000 claims will affect the year of coverage (and thus the reinsurer liable), the number of occurrences and events (see the section on Aggregation above) and the effectiveness of any Year 2000 exclusion.

9.2In relation to "losses occurring during" excess of loss reinsurance policies, the liability of the reinsurer will depend on there having been a relevant loss occurrence during the period of the reinsurance as well as on the occurrence being of the type falling within the policy. Where, however, a reinsurance contract has been written on an "events occurring basis", the event must obviously occur during the policy period. The time of the event is decisive rather than the losses arising out of it as was illustrated in the case of Kelly -v- Norwich Union [1989] 2All ER 888.

9.3This case concerned an insurance policy providing cover against loss or damage to buildings from inter alia burst water mains and provided that Norwich Union would indemnify Kelly in respect of "events occurring during the period of insurance". The water mains supplying a house burst prior to the commencement of the policy. Leakage from the burst main however, caused damage to the house during the policy period. The Court of Appeal upheld Norwich Union's denial of liability on the basis that the word "event" did not refer to the damage but rather to the peril that gave rise to it.

9.4This could have important ramifications in relation to Year 2000 issues where it could be argued that the relevant reinsurance contract to respond is that in place at the date of installation of the defective software or component or hardware rather than the date of any resulting damage - this being the relevant "event". It therefore provides a means for reinsurers currently on risk to deny liability. The "losses occurring policy" may be triggered by Year 2000 losses in contrast. Such policy may respond in full or up to a proportion of the loss (es) which could properly be said to be allocable to the period of cover (See MMI -v- Sea Insurance Company (Court of Appeal 26th March 1996 unreported)

10.REINSURERS' POTENTIAL LIABILITY FOR NON-INDEMNITY PAYMENTS

10.1Potential exposure for the defence costs incurred by reinsureds from legal proceedings arising out of Year 2000 issues may be extensive. It will be recalled that very substantial legal costs were incurred in the defence of the declaratory judgment actions in the United States relating to asbestosis related claims between US insureds and insurers. Where Year 2000 coverage under primary policies is ambiguous, reinsurers may see an increase in the frequency of actions seeking declaratory relief in order to clarify policy obligations for Year 2000 claims.

10.2Whether or not and the extent to which defence costs are covered by the relevant reinsurance will obviously be determined by the precise wording of the policy in issue. If however the decision in Kelly referred to above is applied to Year 2000 exposures (re trigger of cover) it may be that the policy triggered is older and silent as to the issue of defence costs.

11.GOVERNING LAW AND JURISDICTION OF REINSURANCE CONTRACTS

11.1Reinsurance contracts potentially affected by Year 2000 claims are likely to be governed by the laws of a number of different jurisdictions. Year 2000 is a global problem. If the London market writes a supporting line on the reinsurance of a US insured led in the US, it is likely that the reinsurance is to be governed by the law of a particular US state (i.e US service of suit and arbitration clauses) The reinsurance of London Market cedants and indeed retrocessions led in London will however normally be governed by English law.

11.2The law governing the reinsurance or retrocession in question and the jurisdiction where it is agreed disputes are to be decided will have an important bearing on the outcome of any Year 2000 dispute including the enforceability or otherwise of exclusion clauses. Reinsurers should be alive to this factor. Further, insurers/retrocessionaires should ensure that there are no regulatory or statutory issues which may impact on the proposed use of such clauses.

12.ARBITRATION/ MEDIATION CLAUSES

12.1In order to avoid the cost of litigating Year 2000 related coverage disputes, reinsurers may wish to incorporate an arbitration clause or indeed a mediation clause which is triggered specifically by Year 2000 disputes.

13.CONCLUSION

Given the wholly unpredictable nature and scope of Year 2000 losses and the fact that, with the exception of facultative reinsurance, selective underwriting is probably an option for the reinsurance market, reinsurers' only realistic ways of minimising their exposure will involve one or more of the following:

• Identification of most exposed classes of insurance
• Selective application of exclusions
• Aggregate limits for Year 2000 losses
• Careful vetting of underwriting practices of cedants, possibly combined with specific warranties
• Realistic pricing.

Regrettably there is no "silver bullet" solution for the reinsurance industry any more than there is for the Year 2000 problem itself. New Year's Eve 1999 will be a time of trepidation rather than celebration.
; Anthony Hobkinson