ICAs - A problem or opportunity?

UK run-offs will have to embrace Individual Capital Assessments as the FSA takes a growing interest in the sector. They should welcome the opportunity with open arms, says Paul Murray of EMB.

Many run-off managers would rather not think about Individual Capital Assessments (ICAs), seeing them as little more than a regulatory chore. In this respect, the UK run-off market is at a similar stage of development to their ‘live’ brethren a few years ago when the FSA made ICAs an integral part of their switch to risk-based regulation.

Since then more and more companies have come to see this imposition as a rare instance where additional regulation is actually a business benefit – one that is now seen to have given the UK industry a competitive advantage. This change of heart has come about because two things have become apparent. Firstly, ICAs are not as daunting as they might seem. Secondly, they are a valuable management tool in understanding the risks in your business and making the most of your capital.

The run-off community has just started to move along the same learning curve, from reluctant compliance to enthusiastic adoption. It has taken them longer because the FSA has kindly adopted a pragmatic attitude to run-offs as long as they seem to be trouble-free. This will not last; the regulator is already raising the ante, as this statement demonstrates.

“A point that we like to repeat and repeat, is that we consider that we regulate run-offs to at least the same standard as live companies” – Paul Taylor, Manager of Wholesale Firms, FSA

For most run-offs it is no longer a case of if but when they will have to develop their own ICAs. Rather than an imposition, better to see it as an opportunity: those that adopt the process in a positive manner will derive the maximum benefit.

What is an Individual Capital Assessment?
An ICA is a method for identifying, quantifying and managing all the risks inherent in your business and, through this process, calculating the capital that your business needs. It is already well established for live companies and is used by all Lloyd’s run-offs and a minority of other run-offs, especially those that are part of groups with live operations.

For all but the smallest and simplest of concerns, a best practice ICA demands the use of a stochastic or Dynamic Financial Analysis (DFA) model. Derived from the ancient Greek word for ‘random’, a stochastic model enables you to feed in any number of possible scenarios (e.g. reinsurer default, adverse claims development) and measure the potential impact on your business. You can then devise a strategy and make informed decisions to maximise profitability whilst reducing the probability of default to acceptable levels.

Can you avoid ICAs?
You hear many reasons not to use ICAs. Here are some of the main ones, which we then consider in a bit more detail:
- The FSA may not ask to see an ICA
- We’re doing fine, Thanks!
- It would cost too much/we do not have the resources
- Our data is not good enough
- We don’t understand how it will help

‘The FSA may not ask to see an ICA’
ICAs are actually a requirement for any insurer regulated in the UK. In practice, the regulator has backed off from pressing this rule on run-offs where things appear to be going smoothly. Do not, however, expect this relaxed approach to prevail indefinitely. As Paul Taylor of the FSA made clear at the ARC congress in February, limited access to capital can make run-offs a higher risk and therefore in greater need of this kind of control.

Taylor identified six primary run-off focus issues:
• Reserve Adequacy
• Reinsurance exposures & recoveries
• Governance structures (A risk register is fundamental – and this is an essential part of documenting your ICA)
• Quality/availability of skilled staff
• Claim Handling, Management & Data
• “Walk away” Risk

The combination of the managers’ knowledge of their businesses and the techniques of DFA are capable of producing very powerful solutions that address all these issues.

In any event, there are already circumstances where the FSA require ICAs or their equivalent: for example, when buying a run-off or bringing in a Part VII transfer (An ICA automatically addresses the question, posed by the courts and the FSA, of whether the receiving company can pay claims in full in distressed circumstances after the transfer).

‘We’re doing fine, thanks’
You may be ‘doing fine’, but could you do even better and are you doing as well as you think? Many run-offs are companies that once appeared to be successful, but then learnt that they had underestimated the risk in their businesses.

‘It would cost too much/we don’t have the resources’
Some run-offs, especially the smallest ones and those nearing the ends of their lives, probably could not justify the investment in a full ICA. For them, perhaps, a simplified spreadsheet compromise would suffice. They are, though, a minority.

For the rest, the benefits of making more efficient use of capital almost certainly outweigh the costs. Run-off managers are often surprised to find that ICAs and the modelling that goes with them are more intuitive and useful than they had expected.

DFA exercises are relatively manageable provided you adopt an incremental approach, starting with a simple self-contained model and then building on it once you are familiar with the principles. There are user-friendly software tools to assist.

‘Our data is not good enough’
Poor and inconsistent data has often been a problem for the entire insurance industry, but there is usually a cost-effective solution. In simple terms, you either repair your data or work within the limits of what you already have, treating the outcomes with appropriate caution.

‘We don’t understand how it will help’
The purpose of an ICA is to identify, quantify and manage the risk in your business and the impact it might have on your profitability under an almost limitless number of different scenarios generated and tested by DFA models. Once you have this information, it is much easier to devise strategies that maximise profitability without unduly increasing the risk of default. It becomes, for example, possible to construct optimal reinsurance programmes (whether a company is live or purchasing replacement or run-off protection), eliminating unnecessary cover and identifying areas where more is needed. It will support many other aspects including:
• Identifying the best run-off strategy for your company;
• Investments;
• Exit strategies;
• Reserving;
• Commutations;
• Claim submissions to Schemes of Arrangement.

In short, whilst an ICA is no substitute for the judgement of senior management, it will support almost every area of decision-making and give management a powerful extra weapon. Rather than seeing them as a regulatory imposition, ICAs will increasingly become part of the UK run-off sector’s competitive advantage.

DFA modelling – how it might work in practice. This illustration has been greatly simplified, but the principles are applied every day.

Scenario
You wish to evaluate an offer to accept a transfer of 10,000 identical policies. For simplicity, on each policy 999 times out of 1000 you can expect no claim. On the remaining 1 time out of 1,000 the claim is £1 million. The estimated average claim per policy, therefore, is £1,000.

For the purposes of this exercise, ignore discounting/investment income and all non-claims costs.

Questions to answer
What are the expected claims?
What additional risk premium do you need to cover the downside risk?

What are the expected claims?
If looking at only one policy, you may be tempted to say that usually you expect no claims. We know, however, that the average expected claim is £1,000 for each of the 10,000 policies. So, expected total claims = £1,000 x 10,000 = £10 million.

Unfortunately, the real world may not turn out to be ‘average’ in such circumstances. We need a better idea of the spread of possible outcomes. We therefore run, via a simple DFA model, one million different simulations of how the total claims might turn out.

By ordering these one million results we can produce percentiles. For example, the 90th percentile is the value where 90% of the time we expect the result to be below or equal to that value (i.e. only 10% of the time do we expect the value to be lower than the actual outcome).

The results are illustrated in the graph below. We can see that there is a clear risk of a major overrun, with a 5% probability of claims exceeding £15 million and 1% that they will exceed £18 million.



Calculating the Risk Premium
To do this you need three things: expected claims; capital held; and required return on capital. The capital held will depend on the confidence level used by the company (typically 97.5% for a company in run-off). Required return on capital will depend on the company’s aversion to risk and their view of how risky the particular business is (the more risky something is, the higher return one would usually require).

We already have Expected Claims (Loss Cost) = £10m
• For simplicity in the calculations that follow, assume the accepting company is determining their capital at the 95th percentile level, i.e. £15m.
• Assume the accepting company wants a return on capital of 25% p.a.
• Assume all claims paid in 12 months time
• Ignore discounting, expenses, etc
• Capital expected to be released in 12 months = £15m held now - £10m expected losses = £5m
• If we want a return on capital of 25%, we would need to employ £4m of capital now to get the £5m (which is 25% higher than £4m) in 12 months
• Premium for Transfer = Total Capital Needed – Capital Employed
= £15m - £4m = £11m
• Risk Premium = Premium for Transfer – Expected Claims Cost = £11m – £10m = £1m

Conclusion
From the above calculations, based on a simple model simulation, we decide that the transfer would require a risk premium of £1 million. 

This article appeared in Run Off Business Magazine in April 2007