Solvent schemes and IBNR -is the BAIC decision right?

By Peter Fidler and Mark Allen*

This article was first published in Insurance Run-Offs Newsletter, 24th March 2006, issue 86.

Solvent schemes appear, on the surface, to have recovered from the shock caused by the refusal to sanction the BAIC scheme in the summer of 2005, and a significant number of schemes have been sanctioned since then. Efforts have been made to address a number of the concerns raised by the BAIC judgement and to make sure schemes continue to provide benefits for both insurer and insured.

From an actuarial point of view, the BAIC decision has created a number of questions for the industry. The reasons for putting the Incurred But Not Reported ("IBNR") claims into the separate class in solvent schemes of arrangement are not sound, as they are based on a misunderstanding of the process involved in estimating reserves, and the estimation of IBNR claims is not always less certain than the estimation of notified claims. In particular the irregular position of IBNER highlights the problems with distinction.

So was the BAIC decision right? It certainly changed the way in which solvent schemes have been designed but the authors disagree that IBNR claims are fundamentally more difficult to value than notified claims and disagree that this is a reason for concluding that creditors cannot properly consult together in one class. They also dispute the assertion that the latent claim types cited as "unprojectible" in BAIC are, in fact, so difficult to project that actuaries and other claim specialists cannot produce sensible and robust estimates. This view is supported by the fact that many creditors with so-called "difficult" latents have been able to agree commutations outside any formal scheme of arrangement.

Who is right? There are a number of schemes shortly to be proposed which will have issues similar to those encountered in BAIC. We will see.

Full Article

Schemes of arrangement have, in practice, been used for over 100 years and have been an essential tool in corporate reorganisation and reconstruction over that period. Whilst they have been used in non-life insurance insolvencies for almost 20 years, the 2-3 years up to the summer of 2005 had also seen solvent schemes of arrangement become one of the best and most efficient solutions for a non-life insurer wishing to close a run-off book or portfolio. They were efficient, effective and produced finality. Insurers loved them and creditors voted for them in overwhelming majorities because of the benefits, for example, provided by early payment and the payment of risk premiums.

Then, in July 2005, the world of solvent schemes appeared to change overnight when the Court refused to sanction the British Aviation Insurance Company Limited ("BAIC") scheme. Specifically, the Judge in BAIC agreed with a small number of creditors’ objections to the BAIC scheme that the classes were not correctly constituted in that the scheme creditors voted in one class rather than two. So was that the end of solvent schemes? Some commentators thought so but, eight months later, solvent schemes are well and truly alive. More than 40 were sanctioned before BAIC and more than 20 have been sanctioned since July 2005. So was the BAIC decision well-founded or was it decided on a set of facts which distinguished it from other cases?

We consider that some of the points emerging from BAIC were helpful in clarifying the appropriate legal structure for schemes, and we note that solvent schemes post BAIC have already adapted to accommodate these points. But we also argue that schemes for non-life insurance business should not distinguish between classes in the same way that was decided in BAIC and we beg to differ on two of the key fairness points that emerged from the BAIC decision.

In refusing to sanction the BAIC scheme, the Judge concluded that the rights of scheme creditors with Incurred But Not Reported ("IBNR") claims were substantially different from the rights of scheme creditors with notified claims. In his view, the scheme creditors could not properly constitute a single class since the rights of these two sets of creditors were too dissimilar. In reaching this conclusion, he made a clear distinction between notified or known claims which he called "accrued claims" and IBNR or unknown claims which he sometimes called "future claims".

That conclusion appeared to be founded on the perception that the liabilities of creditors with IBNR claims were fundamentally more uncertain or difficult to value than the liabilities of creditors with notified claims. This is, in part, because creditors with IBNR claims have to estimate both the size of future claims and the likelihood of the claim occurring in the first place. Put another way, it is unknown how many IBNR claims will occur or what the precise nature of those IBNR claims will be (because they have not yet been notified). It would seem to follow then that estimating IBNR claims must then surely be much more difficult than estimating the cost of notified claims, where the claims are already known and there is at least some available information on each claim?

The above argument is, however, based on a misunderstanding of the process involved in the estimation of the reserves needed to settle all unpaid claims, whether the estimation is done by the policyholder or by the insurer. First, it is often difficult, in the reserving process, to distinguish between notified and IBNR claims as they share many of the same characteristics. Reserve estimates will often be formulated at an overall level with no distinction made between the two different types of unpaid claim. Indeed, the IBNR reserve will often be defined simply as the "difference" between the overall reserve and the aggregate value of the current estimates on notified claims (known as case estimates).

Second, it is not always the case that the estimation of IBNR claims is less certain than the estimation of notified claims. Indeed, some IBNR claims are easier to estimate than notified claims regardless of the nature of the book of business under review, whether it is insurance, reinsurance or a mixed book.

Whilst the estimation of IBNR might be less certain than the estimation of notified claims in the majority of situations, there are numerous examples that are very common to insurance portfolios where the estimation of IBNR claims may be equally certain or indeed more certain than the estimation of notified claims. Thus, the underlying rationale that IBNR claims are fundamentally more difficult to estimate than accrued claims so that IBNR and accrued creditors cannot properly consult together does not hold true in practice.

One of the distinctions drawn in BAIC was that the estimation of IBNR claims involved an assessment of whether the claims were going to occur, whereas this was not required for notified claims. Strange as it might seem, it is not always known whether even notified claims will actually give rise to a real claim against an insurer. For example, consider a reinsurance contract where the reinsured is required to notify the reinsurer if any direct claim rises to a level of more than 80% of the point where the reinsurance cover kicks in. In this case, it is often not clear whether a notified claim will turn out to be a real claim against the reinsurer or not. Notified claims, as well as IBNR claims, can therefore involve an assessment of whether or not a claim will ultimately arise.

This is not the only point that highlights how difficult it can be to distinguish between notified claims and IBNR claims. The IBNR reserve includes not only an amount ("pure IBNR") relating to claims that have not yet been notified but also an Incurred but Not Enough Reported ("IBNER") reserve relating to any deficiency or redundancy in total case estimates on claims already notified. The IBNER reserve therefore straddles the distinction drawn in BAIC between notified claims and IBNR claims because it clearly relates to notified claims but is normally calculated as part of an overall IBNR reserve.

The separation of pure IBNR and IBNER is not usually a problem for actuaries or reserving specialists because they are generally concerned only with identifying reasonable reserve estimates for all unpaid claims in total, rather than attempting to split that figure into its component parts. Indeed, the different components of reserves are simply different categorisations, at different times, of the amounts required over the lifetime of a claim. These categories may be described as follows:

When the event or occurrence giving rise to a claim has happened but has not yet been reported, it is pure IBNR.

When the claim has been reported, it is given a case estimate but case estimates in aggregate may not be sufficient and so the insurer also books an IBNER reserve at an aggregate level (that is not normally allocated down to an individual claim level).

As the case develops, further information becomes available. When the case estimate is revised to take account of all known information, just before settlement, the notional IBNER in relation to that claim effectively reduces to zero.

Once the claim is paid, all of the reserve components reduce to zero.

An example may help to illustrate how reserves including both pure IBNR and IBNER are estimated in practice. Take the case of an insurance company that has written business for a particular policyholder on which there are ten unsettled bodily injury claims, each of which carries a case estimate of £1m which is "correct" on the basis of all known information so that the sum of case estimates is £10m. These claims are likely to be considered as part of an overall projection that includes an allowance for new notifications or pure IBNR. That projection might come to an overall reserve figure of say £15m, of which the pure IBNR might be £2m and the ultimate cost of the ten notified claims might be £13m, including an IBNER of £3m in excess of the £10m sum of case estimates. The £15m overall estimate will, however, have been derived as a whole without reference to the component parts. It does not matter to an insurer how the reserve splits down into the different parts and, indeed, the reserve often can’t be split down any further as there won’t be sufficient data to do so.

The need for the IBNER reserve of £3m in this example is caused by the historical experience showing that the case estimates on a certain percentage of notified liability claims will need to increase in the future, for example because of late medical reports. All of this underlying development is inherent in the available data and is taken into account in the projection of overall reserve requirements. Thus, whilst it is accepted that each of the ten £1m case estimates is "correct" on an individual basis, on the basis of current knowledge, it can be equally evident that the total cost of these ten claims will ultimately be more than £10m, although it will be not be known which of the ten claims will actually cost more than £1m in practice.

Scheme estimations at a policyholder level will often use similar types of approach and, in most cases, the projection of pure IBNR reserves and IBNER reserves is done as part of the same exercise. IBNER represents an amount added on to case estimates to cover further development on notified claims and pure IBNR represents an amount related to claims that may be reported in the future. These subdivisions are, however, generally not relevant to the overall projection which will be conducted at an aggregate level.

To take another example, consider the Dutch Aviation Pool ("DAP") schemes which were modified before the DAP convening hearing, in the immediate aftermath of BAIC, and which were sanctioned on 28 September 2005. In this particular instance, very careful consideration was given to the precise definition of IBNR reserves which was as follows:

"Incurred but Not Reported" Scheme Liabilities owed to a Scheme Creditor in respect of losses which have been:

1. incurred but have not been reported to the Scheme Creditor or by the Scheme Creditor to a Scheme Company; or
2. reported to the Scheme Creditor or by the Scheme Creditor to a Scheme Company if and to the extent that an adjustment to the Notified Outstanding Losses for such claims is appropriate in order to reflect uncertainty as to the ultimate amount of such losses".

The category of IBNR claims described in (b) in the above definition was intended to capture IBNER. The view taken by DAP was that an amount added on to the sum of case estimates to cover further development of notified claims was a matter to be included within IBNR. This is consistent with the normal reserving approach described above and, in the DAP schemes, creditors with IBNER claims were grouped together in the same class as creditors with Pure IBNR claims. This is significant since claims in the former category relate to losses that have already been notified, whilst only claims in the pure IBNR category can truly be described as "unaccrued" or "future" in the sense used in the BAIC judgement.

All of this commentary highlights the first real problem in interpreting the BAIC decision, namely where does IBNER go?. The BAIC decision did not focus on IBNER at all but instead referred exclusively to accrued claims and to IBNR or future claims, without recognising that IBNER reserves (a key component of IBNR) actually relate directly to accrued or notified claims.

Notwithstanding the above issues relating to IBNER, the BAIC decision appears to have depended more crucially on the conclusion that the estimated reserves relating to IBNR claims were much more uncertain than estimated reserves relating to notified claims. Consequently, there was held to be a fundamental difference between notified claims and IBNR claims leading to the conclusion that BAIC creditors should have been divided into two classes rather than one.

It is probably true that, more often than not, the level of uncertainty that applies to an estimate of future claim payments on IBNR claims is greater than the uncertainty that applies to an estimate of future payments on notified claims. The difference in this respect is, however, a matter of degree and circumstance, rather than there being any fundamental difference. Indeed, there are hundreds of examples, all of them very relevant to run-off books of business, where the estimation of IBNR amounts is relatively easy and subject to very little uncertainty; equally there are many circumstances where the estimation of the reserve required for notified claims is exceedingly difficult. This is best illustrated by a few examples:

Insured is a large chemical manufacturing company that has bought product liability cover for products claims (maximum of £20m for any one occurrence) with a per-claim deductible

Product liability cover has one unsettled notified claim against it which relates to claims made against the manufacturer by a number of users of its chemical products

Loss adjusters have assessed the claim at £8m, which is the current case estimate

There is no possibility of future notifications to the insurer because the terms of the product liability coverage preclude the notification of new claims beyond a certain date

Insured has also bought public liability cover for non-products claims

Public liability policy has a number of notified bodily injury claims on its public liability cover.

There are possibly further bodily injury claims that have not yet been reported.

Claims history is available in terms of the development of bodily injury claims over time.

In estimating reserves for this policyholder, the scheme company or the policyholder will first consider the products cover against which there is just one notified claim. The maximum amount that could be payable in respect of this claim is the limit of the cover (£20m) but there may be a doubt about the liability of the manufacturer to the original claimants. Furthermore, it may not be clear whether the individual claims can be aggregated when making a claim against the insurer. If they cannot, then the products liability cover may not be impacted at all, as the per-claim deductible may not be exceeded.

Past claims experience for the policy in question is of no assistance in this case because we are considering a one-off claim that has never previously arisen and no historical data or trends in data will assist in producing a reasonable estimate. It will therefore be necessary to study opinions received from legal advisers and loss adjusters regarding the circumstances of the claim, liability and aggregation. This will permit a single estimate to be assessed or, alternatively, a range of different future possibilities to be identified to which different probabilities might be assigned (again leading to a single overall estimate). This might, for example, amount to £12m which can be expressed as the case estimate of £8m plus an IBNR reserve of £4m, all of which is IBNER in this instance. Although this estimate obviously relates to a notified or known claim, it would be very clear that the overall estimated reserve of £12m was particularly uncertain, because of the way that it was derived.

For the public liability cover, case estimates are applied to the notified bodily injury claims and overall reserve requirements are projected, allowing for the future development of these case estimates and for pure IBNR notifications. In this case, the likelihood of IBNR notifications may be relatively easy to predict because they may generally be small events connected with prior claims being reopened for the payment of small fee amounts, so the estimation of IBNR claims is somewhat easier and more certain than the estimation of the quite large claims that have already been notified and where significant uncertainty may exist.

In summary, the reserve relating to the single notified claim on the product liability cover would be very uncertain and much less certain than the reserve relating to notified claims on the public liability cover which would itself be less certain than the reserve relating to IBNR claims on the public liability cover. It would seem, therefore, very anomalous that this creditor should have claims in two different classes if the rationale for the class division is that IBNR claims are much more difficult to value than notified claims. As can be seen in this case, and in many other similar cases, IBNR claims can be easier to value than notified claims.

Indeed, it is clearly true to say that both the notified and IBNR claims in this and other examples are uncertain in terms of their ultimate value, and sometimes in their existence as a valid claim, because they have not yet been paid. They can, in this sense, be considered to be alike and the only difference between them is that the notified claims can be identified explicitly. Given that the notified and the IBNR claims are alike in all other respects, this suggests very strongly that they should be in one class rather than two classes for scheme purposes.

Three US asbestos producing companies, all insured by the same insurer

Company A has had many asbestos claims filed against it; the total estimated costs of notified claims to date have not yet reached the insurer’s layers but:

Future asbestos claims filings are very likely to exhaust coverage

There is virtually no chance that Company A will ultimately remain within the limits of its cover, even if optimistic assumptions are made regarding the average costs of future claim notifications.

Company B has also had many asbestos claims filed against it; the total estimated costs of notified claims to date may have exhausted about half of the insurer’s layers but:

The estimated costs to Company B of claims already notified to it are difficult to identify accurately and the insurer’s liability to Company B on these notified claims will therefore be similarly uncertain.

There will be a wide range of claimants with differing degrees of injury, cases could come up in different US states, and therefore the courts could take different views both of those cases for which Company B is liable and of those for which it is not.

Nevertheless, it may again be easy to estimate numbers of future asbestos claims filings and there may again be virtually no chance that Company B will ultimately remain within the limits of its cover, even if optimistic assumptions are made regarding the average costs of future claim notifications.

Company C has had only a few claims notified to it to date and the total estimated costs of notified claims to date have exceeded the attachment point of the insurer’s layers but:

The insurer’s coverage of Company C is very significant, with very high vertical limits.

It seems very unlikely that the costs of all past and future notifications will come near to exhausting the cover as a whole.

The assessment of reserves is very straightforward in the first two cases above. Insurance coverage will be exhausted in both cases and the overall reserve will be equal to the total amount of the cover available less amounts paid to date. In the first case, the IBNR reserve is equal to the whole of that cover and is very certain in amount. In the second case, the insurer’s liability on notified claims is uncertain but the IBNR reserve removes all of the uncertainty because the overall reserve, consisting of both notified claims and the IBNR reserve, will exhaust the available cover and is therefore very certain in amount.

It is only in the case of Company C that the insurer’s liability in relation to IBNR claims is probably more uncertain than the insurer’s liability in relation to notified claims, because the IBNR is not constrained by policy limits, unlike Companies A and B. So the circumstances of Company C have some parallel with the rationale in BAIC, but exactly the opposite circumstances apply to companies A and B.

To summarise the conclusions from this example, Company A is a pure IBNR claimant with virtual certainty about its claim amount. Company B is a claimant with both notified claims and IBNR claims with a lot of uncertainty regarding the cost to the insurer of its notified claims but that uncertainty is removed by the IBNR because the total reserve required by the insurer will exhaust the available cover and is very certain in amount. Company C is again a claimant with both notified claims and IBNR claims and, in this case, the cost of notified claims is uncertain but the cost of IBNR claims is more difficult to estimate and therefore more uncertain than the cost of notified claims.

In this example, it does not seem reasonable that these three creditors should each have claims in two different classes if the rationale for the class division is that IBNR claims are much more difficult to value than notified claims. This conclusion is clearly not correct for Companies A and B. Whilst it may be correct for Company C, this is only to a degree and the difference is not fundamental in the way that the BAIC decision appeared to suggest.

In particular, it is not reasonable to draw a distinction between notified and IBNR claims on the basis that the latter are more difficult to value because again this is often simply not true, as is illustrated in this example. Indeed, it can be seen that the only real uncertainty in estimation applying to each of these three companies surrounds the uncertainty applying to the reserves as a whole, which include both notified and IBNR claims together. Such uncertainty as there is, however, in the case of companies A and B, depends on whether total liabilities from past and future notifications will reach the limits of the policies or not. Any distinction between notified and IBNR claims is positively unhelpful in understanding the overall position in this example.

The conclusion to be drawn is that the difference between estimating liabilities on notified claims and on IBNR claims is a matter of degree rather than a matter of principle. Hundreds of other examples could be given, all of which suggest that there are, in fact, very significant similarities between notified and IBNR claims rather than any fundamental difference between them.

If then a division into separate classes on the basis of notified and IBNR claims is not a viable distinction, the question arises whether there is any other sound basis for a division into classes. It can hardly be seriously suggested that differences based on the type of business written or the underwriting years concerned are suitable bases for subdivision. If differences of this type between creditors were sufficient to require different voting classes because, for example, different assumptions needed to be applied in estimating the reserves of different business classes or underwriting years, it would be easy to envisage a situation where there would soon be as many voting classes as there are policies! This surely cannot be a sensible conclusion.

In this respect, in the case of Re Hawk Insurance Company Limited, Chadwick LJ said that it was necessary "…that those whose rights are sufficiently similar to the rights of others that they can properly consult together should be required to do so lest by ordering separate meetings the court gives a veto to a minority group. …[It is] important [that the] test should not be applied in such a way that it becomes an instrument of oppression by a minority"

Any proper criterion for a division into separate classes must focus on differences in the rights of scheme creditors. The actuarial view would be that, for both notified claims and IBNR claims, the scheme creditors’ right, if the run-off were to continue as a solvent run-off, would in each case be to wait until the liability has been fixed and then to be paid under the terms of the insurance policy in place. Specifically, those with unsettled notified claims and those with IBNR claims are in the same position as each other. The supposed difference in estimation suggested in BAIC is not at all fundamental and is, at most, a matter of degree rather than a difference of principle. It should not therefore form the basis of a class division.

The Court refused to sanction the BAIC scheme because of the classes issue, but the Judge in that case also made a number of other points in relation to solvent schemes. Some of these points have already been taken on board by proponents of schemes and solvent schemes today look different from those of eight months ago.

There were, however, two other key points in the BAIC judgement that were arguably more damaging to the survival of solvent schemes than the classes issue, because they suggested that solvent schemes were not "fair" or even that they could not be "fair". In particular, the BAIC judgement referred to certain types of claims (e.g. asbestos claims possibly emerging from the aviation industry) which the Judge considered it was not fair to estimate. The Judge also said that direct policyholders "were not in the risk business" and suggested that, in his view, direct policyholders with IBNR should not be compelled to commute their cover (albeit for a sum of money) by reason of such a solvent scheme.

In BAIC there was a perceived problem over the treatment of aviation asbestos claims in the estimation process. Whilst there was no firm conclusion drawn as to whether they could or could not be estimated, this issue was clearly significant in terms of the case as a whole.

In this respect, it remains a fact that most creditors, even those with "more difficult" latent claims such as aviation asbestos liabilities, are often willing to enter commutation discussions with insurers. The implication of this is that they can value these claim types on the basis of similar types of models that they have themselves developed. How otherwise would they be able to decide whether they had or had not done a good deal in a commutation?

Furthermore, actuaries and other experts have been able to develop models for estimating almost all types of latent claim. In particular:

Some of these methodologies are very well established. Detailed models, some of them very sophisticated, have been used to conduct projections of future claims costs for all types of latent claim where there is a significant claims history. Examples include asbestos, pollution, agent orange, DES, hearing loss and there are many more.

Even where there is only limited claims history (e.g. asbestos in the aviation industry), but there is claims history in another industry (asbestos in motor manufacture), sensible and robust projections can be produced.

Indeed, even where there is very limited claims history and there is no comparable industry information, but no really difficult medical issues arise, sensible and robust projections can be produced. An example would be sexual abuse claims in religious institutions and schools which have emerged in the relatively recent past.

In relation to aviation asbestos claims that were a key issue in BAIC, the surge in aviation claims caused by the US plaintiff bar’s focus on aviation insureds in 2001/02 gave rise to some additional uncertainty in projections at the time the surge occurred. It should, however, also be noted that:

There has been a rapid return of claims filings to normal level after the initial surge in 2001 and 2002.

Developments in aviation asbestos (where claims are related to the tyres of aircraft) can be compared to those in tyre production for other industries such as the motor industry.

Asbestos exposures in the aviation industry are no earlier or later than "normal" US asbestos exposures so there is no particular difficulty caused by exposures being much later than and therefore more difficult to project than equivalent claims from other industries.

It will be clear from the above that the authors believe that the claim types at issue in BAIC (aviation asbestos) were, in fact, very capable of projection. That is not to say that all types of latent claims are capable of estimation. There are two types of claim that can cause problems in certain circumstances.

The first type consists of latent claims where there is little historical experience (in their own industry or in other comparable industries) and where the arguments over liability surround medical issues that are currently unknown; this differentiates them from the much more numerous types of latent claim that can be easily projected, as noted above. When both of these characteristics are present, it can sometimes be difficult to build robust ground up exposure models. There are, however, very, very few examples of this type of claim. The second type of latent claim would comprise potentially unknown types of latent claims on liability covers written in the relatively recent past.

Are these claim types a reason why solvent schemes should not succeed? The short answer to that question is no. The first type of claim does not normally cause any problem at all, because the very reasons why they are now problematic in estimation terms will also have been a problem for any insurers trying to price and underwrite the business in the first place. The very facts that there may be no claims history and that there may be difficult medical issues arising (perhaps because causation is not proven) will mean that insurance cover will not have been available to the relevant insureds against such claims for many years. The question of estimating such liabilities in solvent schemes has therefore not arisen.

So what of truly unanticipated latent claims, claims that nobody yet knows about? It should first be noted that the need to deal with such issues is very much the exception rather than the rule. Many of the books of business being proposed for schemes have been in run-off for many years and it is inconceivable, from a statistical standpoints that the first such claim of a new disease type could arise after a delay of ten or many more years.

Even for more recent books of business, sensible and pragmatic approaches to exclusion and/or to the estimation of unanticipated latent claims can produce equitable solutions for both insurer and insured alike. One possible approach to this issue reflects the fact that allowances for then unknown latent claims may have been made implicitly in the premiums originally charged and may still be part of existing reserve estimates. If the business is fairly recent and of a particular type, it might then be possible to estimate such claims reasonably fairly by reference to current allowances for such claims in reserves, and this might provide a reasonable deal for both insurer and insured.

In summary, then, the potentially problematic types of latent claim described above either don’t arise or can often be accommodated in solvent schemes in practice.

Finally, we turn to the remaining fairness point raised in BAIC. This last point, namely that it was unfair for direct policyholders with IBNR claims to be bound by schemes, because they were not in the "risk" business, was the most potentially damaging to the viability of future solvent schemes. In this respect, it clearly would be unfair for personal lines policyholders to be bound by a solvent scheme and it is not permissible, under UK law, for compulsory insurance covers to be schemed. The simple fact though is that the direct policyholders who are creditors in solvent schemes are generally substantial commercial companies. These companies are often well advised, by actuaries and others, and are very used to considering policy wordings carefully. Indeed, they will often need to estimate inwards claims for the purposes of their accounts or their regulatory returns and may have their own captives that retain some portion of the insurance risk in-house. They may even have their own risk managers who have a specific role in understanding the insurance coverage purchased and in assessing the level of recovery expected on different insurance policies.

For example, US corporations have to produce these figures at regular intervals for their accounts or SEC filings, which need to include inwards claims liabilities if they are at all material, with significant disclosure if the claims are non-standard. There are similar requirements for UK companies who need to comply with the Companies Act requirement that their accounts should represent a true and fair view of their affairs.

Such companies commonly commute their policies on an individual basis for cash and, accordingly, should be able to assess and participate in a solvent scheme which is no more than a global commutation with all policyholders.

Solvent schemes appear, on the surface, to have recovered from the shock caused by the refusal to sanction the BAIC scheme in the summer of 2005, and a significant number of schemes have been sanctioned since then. Efforts have been made to address a number of the concerns raised by the BAIC judgement and to make sure schemes continue to provide benefits for both insurer and insured. There may, however, be creditors who will continue to object to schemes that are proposed in the future on a number of grounds (eg due to loss of cover, as a negotiating tactic etc). Those creditors are likely to bring forward arguments similar to those raised at the BAIC sanction hearing.

So was the BAIC decision right? It certainly changed the way in which solvent schemes have been designed but the authors disagree that IBNR claims are fundamentally more difficult to value than notified claims and disagree that this is a reason for concluding that creditors cannot properly consult together in one class. They also dispute the assertion that the latent claim types cited as "unprojectible" in BAIC are, in fact, so difficult to project that actuaries and other claim specialists cannot produce sensible and robust estimates. This view is supported by the fact that many creditors with so-called "difficult" latents have been able to agree commutations outside any formal scheme of arrangement.

Who is right? There are a number of schemes shortly to be proposed which will have issues similar to those encountered in BAIC. We will see.

The article is reproduced from a newsletter published by Bannister Int

* Mark Allen is a senior actuary and partner of PricewaterhouseCoopers LLP with extensive experience in insolvent and solvent schemes of arrangement. Peter Fidler is a consultant specialising in run-off insurance at CMS Cameron McKenna LLP.

This article was written for Law-Now, CMS Cameron McKenna's free online information service. To register for Law-Now, please go to www.law-now.com/law-now/mondaq

The original publication date for this article was 25/04/2006.