Better Use of Individual Capital Assessments
Insurers and reinsurers have invested enormous amounts of time and money in their Individual Capital Assessments, so why don’t they put them to better use? Richard Rodriguez of EMB explains how they can provide competitive advantage.
Marco Polo recounts how, as guest of honour at a pig roast in Asia, he was horrified to discover that they ate only the crackling. The meat itself was thrown away or given to servants. Why go to all that trouble just to eat the skin, he asked.
I am occasionally reminded of this story by the way some insurers and reinsurers use the Individual Capital Assessment. Being a new regulatory requirement of the FSA this year, all UK-based underwriting entities have been producing their own ICAs. As anyone who has been through an ICA will tell you, it is an expensive and massively time-consuming process. Yet too many organisations do not even try to exploit its full potential value.
An ICA amounts to a root-and-branch review of the entire business – one that involves just about every department and every senior executive. It requires analysis, and usually modelling, of all your business processes. As a result you acquire data and understanding of your internal dynamics that could have far-reaching applications. They should be put to many uses other than just keeping the regulator at bay. In fact, regulators have made it clear that the ICA analysis has to be integrated within the business management processes if any reliance is to be placed on the analysis and results.
The information you gather, the analysis you do and the overall modelling undertaken can provide an insight into how you manage your capital. Hence the output can underpin the analysis required for many core strategic activities. These include:
• Allocating Capital;
• Reinsurance purchase;
• Product and pricing strategy;
• Business expansion;
• Mergers and acquisitions;
• Alternative Risk Transfer mechanisms;
By way of illustration, let’s take a look at the second of these, reinsurance purchase. This will demonstrate how the analysis acquired during an ICA can be used to improve corporate performance.
What is an Individual Capital Assessment?
An Individual Capital Assessment (ICA) involves quantifying the risks inherent in your insurance or reinsurance business and then deciding the amount of capital you need. The FSA requires you to be able to demonstrate at least 99.5% probability of solvency over a one year period.
An ICA takes into account Operational Risk, Group Risk, Insurance Risk, Market Risk, Credit Risk and Liquidity Risk. Within these categories, there are many smaller types of risk to be analysed.
As well as aggregating the capital requirements of all the risks, the FSA demands that you demonstrate an understanding of how they interrelate. For example, a big insurance event could trigger a fall in equities (as after 9/11) and so accentuate the strain on your capital.
In practice, all but the smallest insurers find that the best or only realistic way to conduct this exercise is through the use of stochastic models. These are financial models that enable you to stress test your business through the use of simulation and hence allow you to consider any number of possible scenarios and their outcomes.
Reinsurance – how to identify a better strategy
Buying reinsurance is one of the most basic functions of any insurance company, and often one of the areas where a company’s analysis is weak. The programme’s effectiveness is bound to have repercussions for your whole financial performance, including the return on capital. Yet historically there have been many instances where insurers and reinsurers have not truly understood their reinsurance programme.
The modelling platform from the ICA process will not, on its own, give you the answers. However, it will, if built in an appropriate way, provide you with the analytical tool necessary to analyse your reinsurance programme in more detail.
Although no two companies or syndicates are the same, the overall process required to look at the reinsurance programme is similar. At the heart of any reinsurance modelling analysis and indeed the ICA modelling process itself are good quality gross loss profiles for all classes and subclasses of business written.
This kind of information has often been hard to acquire because of the way that claims have traditionally been logged. The ICA requirement has, however, acted as a catalyst for the better storage and retrieval of data and in the future it should not be too laborious to extract what you need. Gross loss profiles should preferably be organised on a class-by-class basis taking into account the characteristics of the losses, such as claim type and claim size.
The existence of high quality gross loss profiles provides the necessary platform, but not the entire picture. It is sometimes forgotten that we need to capture all the risks inherent in the business so that any reinsurance option can then be tested with regard to its impact on overall company performance and capital. This is where the ICA work comes in as it is essentially a model of all a company’s risk with the output being the capital required.
So, the next step is to use the ICA modelling platform as the basis for a series of calculations to test the outcomes of any number of different reinsurance strategies. They could be quite simple such as just assuming an increase in excess points. Alternatively, the analysis might turn into a wholesale review of reinsurance strategy looking at, for example, combinations of changing aggregate deductibles, excesses, limits, reinstatements and placement percentages with the outputs from each scenario being a number of key statistics.
The most appropriate scenarios to consider will vary from organisation to organisation and will depend on the complexity of any ICA modelling work undertaken. In all cases, the crucial question is, what do you want your reinsurance programme to do for you? To put it another way, what are your priorities?
To make an informed decision, therefore, senior management needs to determine what it regards as the most important statistics. Possibilities include:
• Return on capital;
• Capital required;
• Expected net profit;
• Probability of failure to meet profit targets;
• Probability of falling below the regulatory minimum capital;
• Probability of failure to meet dividend targets;
• Probability of failure to meet certain rating targets;
• Likely effects on perceived solvency/ratings.
Whatever your choice, you must bear in mind that these statistics are inter-related and no one measure can be viewed in isolation. For example, it is difficult to imagine being able to maximise your return on capital without giving some consideration to the actual capital you hold, the reaction of the rating agencies and the regulator.
In the diagram below the return on capital is compared to the required capital from the output of an ICA model when run with ten different reinsurance structures. The current return on capital on the portfolio is 18% with a capital required of £200m. However, the diagram illustrates that we can improve on the return on capital and reduce the capital required by changing our reinsurance structure, in particular, we find that we can improve our return on capital to 27% and reduce our capital required to £190m under one of the scenarios.
The human angle
Another reason why companies will vary in their responses to this type of analysis is that the decisions you take may have HR, psychological and organisational implications. For example, the most cost-effective way to buy reinsurance may well be at an organisational level that rises above individual classes or territories.
Underwriters who have always bought reinsurance to guarantee the viability of their own operations may find, therefore, that they lose control over this aspect of their activities. A direct consequence of this may be that there is an increase in volatility at a local level, which may not be acceptable to some underwriting entities. They may prefer to adopt a reinsurance model that is, at least on paper, slightly less efficient.
The role of the broker
Whatever the outcome, making full use of your ICA platform will shape relations with your broker. In particular, the buyer should be able to take more control of the process, making decisions with respect to the whole organisation rather than just some low level pieces. Yes, brokers will need to understand the thought processes and, of course, procure the cover. Their strategic advisory role will, nonetheless, diminish.
Informed choices
All this helps to demonstrate how a thorough overhaul of your reinsurance buying, based on an understanding gained during an ICA, can involve a radical review of your operations and involve virtually every aspect of an underwriting operation. And remember, reinsurance is just one of many areas of decision-making where the ICA can give you the platform to make well-informed decisions.
As I have tried to explain, this analytical approach will never be a substitute for the subtleties of management. Even something as numbers-driven as reinsurance goes well beyond the realm of the actuary, but the approach described above asks fundamental questions. And it ensures that the decisions are taken on the best possible advice.
Case Study
An insurer has an expected return on capital of 18% on £200m of capital, however, it wishes to examine its reinsurance programme to determine if it can gain a higher return on capital without changing its capital base. To determine if this can be achieved the insurer uses its ICA model to examine, at a reinsurance contract level, both the effectiveness of its current reinsurance structure and a number of alternatives using both graphical and tabular exhibits.
The graphs below are examples of the typical graphical output that the insurer examines from a stochastic ICA model to help determine the appropriateness of the reinsurance programme.
Graph 1
The graph below shows the gross versus net underwriting result for a class of business. From the graph it can be seen that when a there is a high gross underwriting profit there is a lower net one as the reinsurance programme is being paid for with few, if any,reinsurance recoveries. Conversely, if there is a gross underwriting loss the net result is better due to the operation of the reinsurance programme. 
Graph 2
Additionally, for each reinsurance contract that has been included explicitly in the ICA stochastic model, a graph of the reinsurance premium versus the reinsurance recoveries can be examined. 
The graph above shows that for nearly 90% of the time there is no reinsurance recovery from the contract, however, when there is a claim it is likely that the recoveries will exceed the premium paid including reinstatements. Further, it can also been seen that the reinsurance limit of £50m is breached a small percentage of the time. Using graphical means such as these the insurer is able to quickly determine if a reinsurance contract is providing benefit.
In addition to the graphs, analysis by reinsurance contract can be produced from the ICA model, an example of which is in the table below.
| Reinsurance Output by Contract | |||||||
| Rein Contract | Expected Premium | Expected Recoveries | Std Dev of Recoveries | Expected Reinsurers' Profit | Reinsurers' Probability of Loss | Capital | Cost of Capital |
| 1 | 5,289 | 2,926 | 5,834 | 2,363 | 21% | (8,481) | 28% |
| 2 | 7,177 | 4,041 | 9,729 | 3,135 | 16% | (20,041) | 16% |
| 3 | 8,238 | 5,189 | 15,393 | 3,048 | 11% | (45,456) | 7% |
From the table above, it can be seen that given the insurers desire to increase return on capital it is reinsurance contract 1 that should be reviewed, for although it provides some £8.5m of capital relief this comes at a cost of 28%, some 10% greater than the insurers overall return.
Having performed all of the above analysis and examined a variety of graphical output for all of its reinsurance contracts individually and in aggregate, the insurer decides to re-run its ICA model assuming all of its low levels of reinsurance do not exist i.e. it decides to have a higher net retention. Unfortunately, this increases the capital required with only a marginal increase in the overall return on capital. To counteract the increase in capital the insurer decides to purchase higher layers of reinsurance instead which has the impact of producing a return on capital of 22% with the total capital back to £200m.
This is not the end of the study because the insurer has analysed only one of many potential changes to the reinsurance programme. Therefore, the above process is repeated many times until the insurer has a number satisfactory reinsurance structures to consider. It is the outcome of this repeated analysis that is then taken to the next stage of the decision making process, which includes testing the various reinsurance proposals on the reinsurance market.
This article appeared in Insurance Insider in July 2005