Download Briefing
The cost of claims arising from sexual abuse in the UK is on the increase and likely to peak in the coming decade. Paul Murray of EMB discusses the actions that insurers and reinsurers can take to anticipate the consequences.
Sexual abuse carries such a stigma, especially when it involves children, that it is often hidden from view. Those involved – the perpetrators, their victims, employers and insurers – all prefer to keep it quiet. Victims can be very slow to come forward, if they do so at all. Cases are generally settled before they come to court in order to avoid the publicity and resulting damage to reputation.
It is impossible to say with accuracy just how much abuse claims are costing liability insurers and reinsurers because they, like everyone else, are reluctant to share the information. All the evidence, both in the UK and elsewhere, suggests that there has been a surge in losses, and that the trend may well continue for perhaps the next ten years.
Sexual abuse claims have much in common with asbestos-related illness claims in that they have a long latency period and insurers must, therefore, pay out for incidents that took place many years ago. As with asbestos, recent court cases have tended to favour the victim over the compensator.
And, just as new buildings are now much safer meaning that asbestos-related losses will eventually reduce, the world has woken up to the dangers of paedophiles and the enormous damage they do. Organisations such as churches, schools, councils and clubs have taken steps to protect young people, and the introduction of a Sex Offenders Register will also assist. As a result, this type of claim should eventually become much less frequent. In the meantime, insurers must prepare for an increase in pay-outs.
In recent years, abuse claims have become much more common as victims pluck up the courage to come forward following media coverage of other (sometimes related) cases and after detailed police investigations into certain care homes and communities. In the current environment where it is easier (though undoubtedly still traumatic) for victims to come forward, more claims are likely.
The graph below shows future claim notifications for a hypothetical insurer where the peak of exposure to possible abuse claims was in the early-to-mid 1980s. The top line assumes that claims notifications peak approximately thirty years after the main exposures, whilst the bottom line takes a more optimistic view (that claim numbers will reduce after the recent publicity related claims surge). Regardless of whether or not the number of notifications rises, we expect severity to increase. Insurers and reinsurers would be well advised to adopt a cautious view.
Let’s now consider the nature and extent of the problem and some of the legal ramifications before moving on to how insurers and reinsurers can anticipate claims trends. As we shall see, they can and should be modelled like any other significant type of loss.
Various studies (see the excellent 2005 Moran, Nicholls and Konstantinidis paper) indicate that around 10%-15% of children have historically suffered from sexual abuse. Tragically, most of these incidents occur among family and friends. There are more likely to be insurance claims in cases where people are abused by those employed to look after them; these amount to perhaps 2% of all children, although only some will seek compensation.
The legal position surrounding abuse cases can be complex, one of the main issues being the extent to which the passage of time protects abusers, their employers and insurers from the threat of legal action. In the UK, there is a strict six-year limitation on assault cases (usually starting from the date of abuse or on reaching the age of 18, whichever comes later) and a three-year limitation where negligence is alleged. In the latter case, however, the courts have discretion to extend the period.
For negligence, the three-year period begins at “date of knowledge” – a vague notion clarified by the Appeal Court. In KR & Others v Bryn Alyn (2003), which involved abuse at children’s homes, the Court ruled that “date of knowledge” is when the victim “would reasonably turn his mind to litigation.” Generally speaking the judiciary seem to be moving to interpret this definition in a way that favours the victim.
The limitation in assault cases, meanwhile, faces unprecedented legal scrutiny. In Mrs A v Iorworth Hoare (known in the tabloids as “the lotto rapist”), the Court of Appeal referred to the House of Lords due to “very serious deficiencies and incoherencies in the law.” The law lords heard this case at the end of 2007 and a written ruling is due. Most legal experts expect it to make it easier for victims to claim damages by extending the six-year limitation. This would remove apparent anomalies in the law that in 1996 allowed a girl to successfully take her mother to court for negligence when aware of abuse by the girl’s father, while the six-year limitation on assault prevented proceedings against the abusive father.
Other areas of the law are moving in the victims’ favour, too. The 2001 case of Lister v Helsey Hall was the first to allow an employer to be held vicariously liable for sexual abuse by an employee, depending on their duties. (The judge made the distinction that an employer would not be liable for abuse caused by say a gardener but where the employment is close enough to cause an obvious risk, such as ensuring that children go to bed, then they can be liable).
As the number of claims against insurers starts to mount some may wish to try to aggregate these claims to reduce the number of reinsurance deductibles they suffer. To do this, they would need to argue that one man abusing several children is one event. Realistically though, it is difficult to see how abuses carried out at different times on different children could qualify. Even a lack of supervision causing a number of claims is much more likely to be considered as a ‘state of affairs’ and so aggregation would not be possible.
Insurers and reinsurers alike can, however, anticipate and prepare for future claims. In doing so, it is important to understand the possible impact of big losses. While for some the main concern will be a large number of claims in the tens of thousands of pounds, potential recoveries on the largest claims may still be crucial, especially for reinsurers.
The biggest court settlement for abuse in the UK so far was £621,444 in the case of “A” v Archbishop of Birmingham in 2005. Most of the award was for loss of future earnings. The victim, who had been abused by a priest between the ages of 7 and 18, was deemed unlikely ever to gain full-time employment. Taking into account claims inflation, it is only a matter of time before the largest payouts exceed £1 million.
That is only part of the story, though. As the legal landscape changes key judgements tend to cause step changes so that it would not be surprising to see the largest awards increase by material amounts every few years.
The most realistic way for insurers and reinsurers to calculate potential losses is by creating stochastic models – tools that calculate the financial impact of a vast number of scenarios taking into account their likelihood. Lack of data, which is sometimes used as a reason not to use models for abuse claims, actually makes them essential.
Stochastic models are specifically designed to understand better the uncertainty surrounding any estimates. When your largest abuse claim could easily be twenty times (or more) the size of your average such claim, it is crucial to use a model that considers all possibilities. Looking at the current average claim size as the base for any estimate will be virtually useless.
This all means that (re)insurers will have to make assumptions based on an analysis of their exposures and estimates of their eventual cost. In addition to firms’ own internal data, they can use the growing number of cases that are in the public domain.
As with all exercises of this kind, they will need to estimate the number of future claim notifications, the claim size distribution and future claims inflation using a variety of different criteria. In addition, they will have to take into account notification delays, payment delays and the length of the period of abuse. A complication here is that, again reminiscent of asbestos, abuse may have been committed over a number of policy years involving different insurers, leaving questions as to how claims are allocated. Other policy specifics that may need to be considered are the (underlying) original maximum policy limits, any stability clauses and whether costs are in addition.
With any stochastic exercise, there will be a range of possible outcomes with probabilities attached. The resulting figures will only be as accurate as the assumptions upon which they are based, but these can be sense checked and refined in light of experience. The more insurers and reinsurers get to understand their exposure to these horrific crimes, the better prepared they will be for the inevitable fall-out.
This article first appeared in Insurance Day
