Reinsurance Chapter 1 PDF

# The Buyer's Reinsurance Needs ## Introduction The decision of an insurer to purchase reinsurance involves the consideration of many important issues. Chapter 1 of coursebook 785: Principles of Reinsurance examines the purposes of reinsurance and the nature of the different types of contract. I...

# The Buyer's Reinsurance Needs ## Introduction The decision of an insurer to purchase reinsurance involves the consideration of many important issues. Chapter 1 of coursebook 785: Principles of Reinsurance examines the purposes of reinsurance and the nature of the different types of contract. In essence, reinsurance provides the buyer (the insurer) with protection from the financial instability faced as a result of an unacceptable level of losses, as well as the capacity to grow and the benefit of the reinsurer's expertise. Chapters 1, 2, and 3 of this coursebook are concerned with how the needs of the buyer are identified and met. In this first chapter, we will look at the decision to purchase reinsurance from a strategic perspective and will consider the following interrelated issues: * How much risk to retain. The concept of retention is fundamental to arriving at the decision to reinsure, we shall examine the factors which need to be taken into account, and how they interrelate. * **The implications from the buyer's strategic and financial point of view in retaining some risk and of reinsurance the remainder. Reinsurance is a complicated financial tool which needs to be understood and used in relation to the buyer's overall financial and corporate strategy. We will consider what influences may be brought to bear on the decision to buy.** * The market environment at any given time which may affect the decision to buy reinsurance. As we shall see, these concepts are relevant to the purchase of appropriate, affordable, and secure reinsurance. They are interrelated and need to be discussed by senior personnel in an organisation. Needless to say, the outcome of the decisions must be kept under constant review, and alternative action pursued as and when circumstances change. ## Retention In this section, we will examine the main factors influencing or limiting retention. Fundamental to the decision to reinsure is the fixing of the amount which the insurer (i.e. the prospective reinsured) will retain. We will consider the philosophy behind the concept of reinsurance and the considerations which an insurer should take into account when deciding to effect reinsurance contracts either individually or as part of a program. Let us remind ourselves of the definition of retention: the retention of an insurer is the limit of liability, usually expressed as a monetary amount, which the insurer retains for its own net account after reinsurance is ceded. The retention may apply to a single risk or to a series of risks, or to a single loss or a series of losses. Deciding how much to retain is an important matter: by retaining too much of a risk or a loss, and having to pay for them by itself, the insured's financial strength may be weakened and its asset base impaired. On the other hand, by not retaining enough, the insured may be giving away premiums for reinsurance unnecessarily and, therefore, running the account at less than optimum profitability. The decision on how much to retain is a complex one involving, as we shall see, several factors. Although the decisions are likely to be taken by higher underwriting and corporate management, reinsurance managers must play an important role, while reinsurers and reinsurance brokers must appreciate the needs of the original insurer if they are to fulfil their role. The responsibility for fixing retentions will vary from company to company, but usually the general management at board level will take the major decisions. Retentions are usually fixed on each class of business separately (fire, accident, marine, aviation, etc.) with classes subject to further subdivision, such as marine hull and marine cargo. In addition, there may be an overall retention over the combined portfolios of the various classes. An examination of the levels of retention of a number of insurers in most classes reveals a wide variety of levels, without obvious reasons for the variations. Therefore, it is clear that the fixing of retentions follows few technical roles. Although the factors which influence their decisions are common to all, each insurer attaches different weight to each factor and therefore reaches different decisions. As a result, there is an optimal retention for each insurer rather than correct retention. The optimal retention will provide the insurer with the framework to achieve its corporate aims. Even if two insurers have similar portfolios, their corporate objectives are unlikely to be identical, and their retentions may be quite different. The actual retention will usually be the result of a compromise between that which would best satisfy the overall aims and wishes of the prospective reinsured, and the demands of reinsurers. Ultimately, the level of retention also depends upon what retention the reinsurers are prepared to agree. Although theory, mathematical formulae, and computer models can assist the reinsured in arriving at a desirable level of retention, in practice well-tried market custom and usage plays a role. The final negotiation of retention's depends upon the unpredictable and subjective views of insurers and reinsurers regarding varying circumstances which have uncertain effects. As a result, retentions remain primarily a matter of judgment. The first step in protecting the insurer's assets is to design the intended net portfolio of accepted risks so that the retention in any one risk or the retention for all losses caused by the same event (for example, an earthquake) will bear an acceptable relationship to the premium income. The aim will be that no single loss or accumulation of losses can unduly affect the overall underwriting results. To a large extent, this depends on the type of business in which the insurer operates, the type of risk sought and offered, the extent of the exposures to them and the mix of business. The arrangement of retentions and reinsurance for individual risks or types of risk will be discussed in chapter 2. ## Assets, Capital and Free Reserves and Solvency Shareholders who have invested money in a publicly-owned insurance enterprise do so in the expectation of making a profit over time, and certainly not of losing their Investment. Any anxiety on the part of the investors could cause a withdrawal of their investment, with a subsequent decline in the value of the company. From the point of view of the shareholders, protection is necessary in respect of: * Their original capital investment, * Remaining shareholders' funds, namely the free reserves built up from profits and not required for technical reserves; and * Profits, to the extent necessary to service their capital. As we have already noted, the asset base and the value of the company could be severely damaged if the claims arising out of insurance losses are greater or more frequent than expected. Therefore, the management of the company will fix its retentions and arrange its reinsurance programme to ensure that the shareholders' interests are protected. Similar considerations apply to Lloyd's and mutual companies. Although their financial and organisational structures are different from proprietary companies, the protection of the 'owners' is an important factor in the purchase of reinsurance. In most countries regulatory authorities require insurers to comply with specified solvency or surplus requirements which set minimum ratios between the assets (capital and free reserves) and premium income (solvency is discussed in chapter 7). As a result, regardless of the wishes or requirements of shareholders or other providers of capital, an insurer has to maintain a minimum capital base. An insurer may hold assets in excess of the minimum statutory requirements to allow a margin for possible misrepresentations in the valuation of assets due to external factors (such as a falling stock market, or a major fluctuation in currency values) and to allow for the development of business. Since solvency requirements are geared to premium income, only substantial increases in the portfolio and a resulting increase in premium income require an increase in free reserves or assets. It may not always be possible or easy to do this, for a number or reasons. In the best-ordered underwriting environment, insurers would have close estimations of expected loss ratios on any specific classes of business and so be able to load premiums to the extent necessary for sufficient reserves to be developed from the balance of premium over claims. However, in a realistic economic environment, it is not always easy to assess the expected average loss ratios within acceptable bounds. Competition does not always allow premiums to be loaded to the extent required. If the development in a particular line of business coincides with an increased loss ratio as a result of, for example, reducing premium rates as part of a competitive strategy, it may not be possible to increase the free reserves from the technical balances. A failure to hold a suitable margin of assets may also obstruct development. Therefore, in the light of its corporate strategy and regulatory requirements, the insurer will decide on the level of capital and free reserves or shareholders' funds (previous profits not distributed, to finance growth or strengthen the company's operations) it wishes to hold. The insurer should then take the first step In fixing its retentions by deciding to what extent it can afford to risk the loss of its free reserves in any one year (free reserves are lost when then have to be transferred to increase or replace the technical reserves). The period of twelve months is chosen because supervisory authorities and shareholders would view the account over this period. The primary and most important factor when deciding the level of capital and free reserves to be retained is to consider the expected underwriting results over a period of two or more subsequent years. If a period of underwriting loss is envisaged, the insurer can afford to risk a smaller proportion of its assets (and therefore will retain less and reinsure to a greater extent) than if a profitable period were expected. For example, a company decides that it can afford to loose up to £5m out of free assets, but foresees three years of underwriting losses. It will not be prepared to risk losing more than one-third of the £5m, and probably less, in a single year: therefore, it will arrange a suitable reinsurance program. Conversely, if it expected the following years to be profitable, the company could afford to risk losing a greater proportion in one year. When depicted as percentages of capital and free reserves (total assets), the retentions of different companies in any class deviate from the average to such and extent that there can be no acceptable percentage. As a general 'rule of thumb', 'per risk' retentions in the main classes of fire, motor, accident and marine, usually lie between 0.5% and 2% of free reserves. However, retentions in respect of a series of losses from one event (one common cause) are generally not more than 2.5 times the individual risk retentions. One might expect retentions to increase proportionally to increases in capital, so long as similar solvency margins are attained. In practice, this is not the case, particularly when a certain size is attained. There are two possible reasons for this: * An increase in capital may have been required in order to match an increasing premium income, and to maintain solvency standards. If an immediate increase in business k not expected, retentions would remain unaltered. Alternatively, capital may have been increased to allow further development and an increase in retentions would result. * Any increase on 'per risk' retentions of a certain size will produce an increase in the retained account over a narrowing field since there are fewer higher-valued risks than those or a lower value Increases in retentions are limited so as not to disturb the balance or the account Small insurers apply retentions which represent a larger proportion of their capital than larger insurers. One would expect that a merger of two companies with a larger combined capital base would result in a higher retention. Indeed, this factor (and the savings in costs which would result) may have been conducive to the merger. However, the decision to retain more would depend, to a large extent, on the nature of the new portfolio. ## Size and Nature of the Portfolio; Premium Income and Profitability Retention levels will also be determined by the size and nature of the book of business written by the reinsured. A 'profile' of risks by type and size (for example, sum insured for a property portfolio) needs to be established. If all risks were the same, there would be little difficulty in arriving at retention. It would be an exceptional portfolio, however, which comprised entirely homogeneous risks. What the insurer needs to be concerned about in arriving at retentions is the 'homogeneity' of the risks he is writing; i.e. the extent to which they resemble one another. In addition, the insurer needs to take into account the rate at which it is growing and its claims experience. As a portfolio develops with generally similar business, it can be assumed that the probability of fluctuation in experience also reduces. In any portfolio, the results in a period are affected by the number of losses and their average size. The expected results can be estimated by an examination of previous statistics, but fluctuations in the results may arise from an increase in the frequency or severity of loss, or the occurrence of a large lose or losses. In arriving at retentions, the insurer must determine the extent of volatility which can be safely absorbed. The larger the number of homogeneous risks of a similar value and content in a portfolio, the lower the probability of fluctuation in the expected claims experience. Professor R L Carter in his book, Reinsurance, demonstrates this point by comparing the relative variability of claims cost of two homogeneous life portfolios, varying only in the number of insured risks. The relative variability is measured by the coefficient of variation V: $ = standard deviation. V=s/E(x) E(x) = expected (i.e. average) claims cost. Taking the mortality rate as 0.007%, the coefficient of variation measured in terms of the number of claims on a portfolio of 1,000 risks would be: (.993x.007x1000) V = 7 2.6365 7 = 0.3766 whereas on a portfolio of 2,000 risks, it would be: V= (.993x.007 x 2000) = 14 3.7285 14 = 0.2663 Retention is also influenced by the growth of the portfolio. As a portfolio grows, it achieves a better balance; provided the quality of content is similar, the larger the insurer's portfolio, the larger the retention which can be safely carried. In theory, an insurer with an increasing portfolio could increase the retention to a disproportionate extent and carry the same probability of fluctuation. In practice, however, the insurer will wish to gain from the improved stability which the increased portfolio offers and k, therefore, unlikely to increase the retention, even proportionately. An insurer carrying a maximum retention of £250,000 on a net retained premium income of £10 million would probably not carry more than £300,000 on a premium income of £15 million. Retentions must be reviewed in times of inflation. Growth in premium income may be illusory insofar as it does not represent 'real' growth. The retention can only be increased proportionately to reflect the same probability of fluctuation. Moreover, if an increase in portfolio and premium income has been achieved by competitive reductions in rating, the comparison will be distorted since the portfolio will not be earning the same margin. It may be that no increase in retention should be carried. The retention should bear a relationshipp to the profitability of the portfolio and the trends for the future, measured by the loss ratio and compared to that of others in the market. Generally, as a portfolio becomes increasingly more profitable, the insurer has the option to increase retentions (although, as we noted above, it would not necessarily do so). If the portfolio shows signs of being less profitable, the insurer needs to reassess its strategy. This may, of necessity, force the insurer to retain more risk or for the account, at the insistence of the reinsurers or, alternatively, retain less, as part of a re-structuring strategy. The reinsurance programme would also have to be re-assessed. ## Type of Risks and Pattern of Losses The underwriting capacity provided by a given sum of capital or assets depends on the margin of premium over losses. In practice, the actual retention will be decided by the insurer, according to its estimates of the possible loss ratio after allowing for probable fluctuations in the frequency and size of losses. The overall losses must not exceed the sum of the total retained premiums and the insurer's assets. If they did, the insurer would be bankrupt. The risk content and the claims pattern should be studied because in the different classes, and in different portfolios in the same class, the pattern of claims varies according to the frequency with which losses occur and according to the size of losses. This depends not only on the sums insured or limits of the original policies, but also on the possibility of accumulation of claims under several policies from losses which arise from a single 'event', accident or occurrence. Thus, as we will see in the different scenarios below, retentions will be affected depending on the composition of the portfolio and the associated probability of their claims experiences. ### Similar Risks Insured Under Identical Policies In its simplest form, an insurance portfolio would comprise a number of identical policies covering similar insured objects or persons. Such a portfolio would be improbable, but the claims pattern would be simple and consideration of such a portfolio is useful. We will consider a portfolio of accident policies covering loss of life only, each wih the same fixed sum insured. There would be no fluctuation in the amount of the individual loss since the same claim amount will be paid on each insured fatality. However, the expected number of fatalities will have been assessed, principally from past statistical records, by taking the average number of deaths over a period of years. The number of fatalities is likely to differ from the average to a greater of lesser extent in any one year. There will, therefore, be a fluctuation in the frequency of claims for which allowance must be made. The probable fluctuations in frequency of loss may be calculated by statistical methods, and the probability of overall loss on the portfolio in one year can be calculated at a given level of premium. The probability that the premium will be exceeded by the claims in any one year can be assessed according to the Poisson probability distribution formula: P(x) = e.a x! e= a constant equal to 2.7183. X = the number of claims, which may assume any integral value from 0 to infinity. a = the average or expected number of claims in the portfolio. The expected number of claims in each portfolio is 0.5 per mille of 20,000) = 10, and 0.5 per mille of 40,000 = 20. In arriving at a suitable level of retention, the insurer must consider the possibility of two or more insured lives becoming involved in a single accident such as an air crash or natural catastrophe. Assessment of this possibility is a matter of judgment based on knowledge of the portfolio. Points to consider are as follows: * Any common link between the lives insured which may cause them to be involved in the same accident or catastrophe. Members of the same family are likely to travel in the same vehicle or aircraft, as are groups of people who are members of one firm or association. * To what extent the portfolio is involved in particular geographical areas which are prone to natural catastrophes or other similar factors. * Such cumulative losses will be rare and may not show themselves in the loss record, unless over a long period. ### Similar Risks Insured for Varying Sums Insured If the portfolio were to comprise varying sums insured, but was limited to policies covering accidental loss of life, the claims pattern would change. The claims cost would be affected by the distribution of the sums insured. The expected frequency of deaths would not have changed, and the probable fluctuations in the claim cost would be assessed taking account of the proportions of the portfolio according to sums insured. The probability of overall loss would be greater since the portfolio is less balanced. ### Similar Risks Insured for Varying Amounts and Conditions The portfolio might by changed by amending the policies to include personal injury, subject to a scale of indemnities for various injuries (such as loss of a limb or eye). It follows that the claims pattern would change again, since the frequency of accidents involving such injuries would be greater while the average claim amount would alter, and the probability of overall loss on the portfolio would again be different. This probability would be smaller than before, as outlined below: * The statistical base would be larger, as the number of partial injury claims always exceeds the number of fatalities. * The two parts of the portfolio, partial injury and fatality, will tend to balance each other. ### General Portfolios In practice, a portfolio will be more complex and varied than the examples given above, and retentions may be more difficult to determine. There will be many variations between the individual risks which will have to be considered in order to obtain the most balanced net portfolio, that is, one in which each retention holds the same quantum of risk. This perfect met portfolio is barely attainable, because individual risks vary and several risks may be exposed to a single catastrophe. Nevertheless, the best balance must be achieved. It is important to appreciate how differences in the risks in a portfolio arise. Dividing the portfolio into sections comprising similar risks (in respect of size and frequency of losses) will assist when assessing the risks. Unless this is done, the statistical limitations of past results may be overlooked, and retentions should be based on statistical information. In the past, the table of retentions of insurers on their property account filled a substantial volume in which the risks might be divided into 200 or more classifications. The retentions varied not only according to the classification, but also according to the type of construction, premium, extent of fire protection and many other factors. The overall retention plan will be a compromise. Thus, insurers have developed a classification system which fixes a retention for all those properties with similar characteristics, such as woodworkers' premises. The classification and retention may also be applied to other properties which present the same risk of fire, such as a printer's workshop. However, it is important that the content of the classifications should be maintained for each type of property, and statistics kept on the basis of the original classification. The need to examine the content of the wider classifications cam be demonstrated simply and clearly through the marine account. The risks accepted can be divided into the following: * **Insurance for time on:** * Hull and machinery of vessels on full conditions * Other shipowner interests insured primarily against total loss of the vessel only * Primary liability risks * Liability risks * Offshore oil risks * **Insurance for voyage on:** * Vessels in tow * Cargo, and so on. The claims pattern on each differs greatly. Furthermore, particularly in categories (i), (v) and (vii), there are varying risks (vessels of various ages and types operating in different conditions and under different national flags) and many different cargoes being shipped to and from ports, each with widely varying conditions of handling and security. Variations in the individual risks will occur even in a portfolio of simple risks. Over a large account, these variations are important, as the statistical records will not show distortions on account of such brutal variations. However, the more complex or specialised risks become, the greater the effect of important variations between risks. Therefore, less reliance can be placed on general past statistics in assessing the frequency of major claims. As major claims increase, even if their frequency reduces, the uncertainty of their effect is greater. In an industrial fire account, for example, the results are heavily effected by the incidence of large losses. On a marine hull account, major losses represent a smaller proportional of overall losses than on an industrial fire account. The size and frequency of such losses can, and do, have an important effect out marine hull portfolios. An aviation account is primarily affected by major losses, and any changes in the frequency of major losses can have a serious effect. When the portfolio is vary unbalanced, or uncertain in outcome, retentions should be fixed with care. Statistical data should only be relied on when the portfolio concerned relates to a balanced account of comparatively simple risks. This is demonstrated by the results of a gross portfolio over a period of ten years: | Year | Total loss ratio | Three largest losses as % of the loss ratio | |---|---|---| | 1 | 68.6% | 7.6% | | 2 | 73.2% | 8.4% | | 3 | 75.8% | 7.2% | | 4 | 99.5% | 20.5% | | 5 | 97.6% | 23.8% | | 6 | 67.5% | 6.3% | | 7 | 63.4% | 9.4% | | 8 | 56.4% | 7.8% | | 9 | 51.6% | 8.4% | | 10 | 86.3% | 21.3% | The fluctuations in the results of this portfolio are considerable and are clearly due to the effects of major claims. The size of the retention in the three largest losses in the years four, live and ten, influences the net results substantially. Over a short period, the statistical data might prove to be misleading. Finally, the geographical spread of risks in a portfolio must also be considered, as a number of individual risks may be affected by the same event, for example, conflagration, earthquake, flood, windstorm, major traffic accidents and so on. Therefore, the insurer needs to record its aggregate exposures to such events on an area basis, in order to determine the amount of reinsurance protection required to restrict its loss in any one event. This information will be required by the reinsurer to enable it to price the provided. ## Environmental Factors Affecting Loss Ratios Future loss ratios are influenced by changes in the economic, social, commercial, technological and legal climates. These changes cannot be measured in advance and may not be foreseen, but they must be considered when fixing retentions. For example: * Inflation may cause increased claim amounts, often without compensatory increases in sums insured. * Recession may result in reduction in business and premium income, and this could lead to an increasing claims frequency. * Competition may cause unplanned and unforeseen reductions in premium. * Changes in employment levels may affect crime. * Certain technological developments may improve risks, whereas others have the opposite effect. * Court awards for personal injury may be expected to increase, thus increasing liability claims. * Life expectancy has increased doe to medical science and technology. This could affect assessments for claims involving severe bodily injury. The subjective assessment of such factors, and the making of an allowance for unknown possibilities, partially explains the differences in retentions which are found amongst individual insurers who have broadly similar portfolios. ## Types of Reinsurance The Insurer chooses the form of reinsurance which its protect its portfolio, and does so with the intention of obtaining the optimum framework for its retained account. Decisions on retentions will depend on the following: * The availability and selection of different reinsurance contracts and their combination. Each form of reinsurance operates in different ways, and the extent to which the insurer will have to bear all or part of a claim affects the choice of contract and the level of retention. * Whether the retention is the same for each risk in a portfolio. * The insurer's choice will have to be acceptable to a reinsurer or reinsurers; * In the final analysis, the reinsurance will depend upon what reinsurers are willing to provide at an acceptable cost. The insurer, having effected a reinsurance arrangement, is now referred to as the reinsured, the ceding company or the cedant. Throughout this coursebook, we will use the term 'reinsured'. ### Proportional Reinsurance The reinsured may have purchased proportional reinsurance. In proportional reinsurance, all claims and premiums on risks falling within the reinsurance programme will be shared between the reinsured and the reinsurer. The claims pattern on the resulting retention will vary according to whether a quota share or surplus reinsurance treaty is used. #### Quota Share Reinsurance Quota share reinsurance, does not change the reinsured’s claim pattern, since the reinsurers pay the treaty’s fixed share of all losses falling within its framework. Quota share reduces the reinsured’s overall loss, and reduces or limits the loss on any one risk, without changing the balance of the portfolio: that is, the relative size of the potential claim fluctuations compared with.the retained premium income. #### Surplus Reinsurance Under surplus treaty reinsurances, the pattern of the reinsured’s retained losses changes, since each loss is shared according to the ratio of retention to the sum insured and the portion which is ceded. Losses arising on risks not ceded to the surplus treaty will be retained entirely. Surplus reinsurance reduces the net loss on larger risks as well as limiting the loss on any one risk, thereby providing a better balance for the reinsured’s net account. A comparison of quota share and surplus reinsurance shows the differences that have a bearing on the desired retention: * A surplus treaty provides the reinsured with a larger retained premium income for the same monetary retention limit on any one risk, and therefore improves balance. For example, a reinsured elects to have a retained limit of £50,000. With a 50% quota share treaty, the reinsured would retain 50% of all risks with sum insured values or estimated maximum loss (EML) (if written on an EML basis) up to £100,000. * With a one line surplus treaty operating over a fixed retention of £50,000, the reinsured retains 100% of all risks (and so the premiums) with insured values, or EML, of up to £50,000. On all risks over £50,000, the premium is shared with the surplus reinsurers in proportion to the retained line and the number of lines ceded to the treaty. * On a surplus basis, the reinsured pays 100% of claims on smaller risks (those falling within its retention) with reducing proportions of claims to be retained as the risks increase in size. The frequency of losses may vary considerably between smaller and larger risks. * For reinsurers, the quota share treaty will be better balanced than the surplus treaty, so they will allow higher reinsurance commission on the former. This can provide the retained account with an additional risk-free premium loading, which is available to cover loss fluctuations. * Quota share reinsurance requires less administration, since all risks falling within its framework are shared equally. Under surplus reinsurance, the cessions to reinsurers vary from risk to risk according to the insured values. Therefore, the type of reinsurance selected will affect the expense ratio of the retained account, as administration will either increase or diminish. The reinsured has to consider whether the retention should be the same for each risk written, or 'graded' according to their features. Some types of risk are more hazardous than others and, as a type, produce more losses They normally attract a higher premium rate. Fireworks factories, for example, are charged a higher rate for fire insurance than museums. In general, the retention should reflect the reinsured’s view of the risk. The more hazardous and highly rated the risk, the lower the retention. On the other hand, the better the risk, the lower the rate, and the higher the retention should be. In property insurance, risks tend to be classified according to construction, location and occupancy, and individually underwritten accordingly. Therefore, the 'better' risks tend to have a higher retention than the 'poorer' risks, and the balance of the sum insured is reinsured to a surplus treaty if necessary. Thus, there will be a higher retention for museums than firework factories. Nowadays, retentions will also usually be based upon the EML (or the equivalent in other countries) rather than the insured values. ## Non-Proportional Reinsurance In non-proportional reinsurance the reinsured's retention is he deductible of the excess of loss programme. There is no sharing with the reinsurer of all claims of premiums belonging to the account protected. Losses which Individually or in the aggregarte do not exceed the chosen deductible remain within the retention. The excess of loss reinsurer then pays for all losses exceeding the deductible in consideration of a premium paid to it from the reinsured's retained income. Care should be taken to ensure that the reinsured purchases adequate cover under the excess of loss program, since losses which exhaust the limit fall back to the reinsured's retention. Non-proportional reinsurance changes the reinsured’s retained loss pattern and, therefore, retentions will be reflected in the type of non-proportional contract chosen. ### Risk Excess or Working Excess of Loss A risk excess protection safeguards individual risks and is an alternative to proportional reinsurance. As with proportional reinsurance, such protection cannot alter the frequency of claim. Its effect on the reinsured’s claims pattern is to limit the loss on any one risk and protect the retained account from an adverse fluctuation in individual claims amounts. However, as the reinsured bears all losses up to the underlying deductible in full, the retained premium income becomes larger and the relative loss fluctuations will he smaller than under a comparable surplus treaty. The retention (or deductible) and the number of times the reinsured would have to pay them during the year should be affordable, according to the reinsured’s past and expected loss experience. #### Catastrophe Excess of Loss Catastrophe excess of loss does not protect single individual risks, but is designed to respond when losses on two or more individual risks arise from the same event. It does not change the frequency of claims, nor the individual claims amounts attaching to the retention. Catastrophe excess of loss also limits or reduces the adverse effect of an accumulation of losses at one time in one place. The retention or deductible should take into account the reinsured’s commitments or exposure in a particular area for, say, earthquake, and the maximum amount that would be required to be paid in the event of a catastrophe. #### Stop Loss and Aggregate Excess of Loss Stop loss and aggregate excess of loss are not identical, but they both limit or reduce the aggregation of claims over a given period of time, normally on an annual basis. The reinsured’s total retained loss cost is contained within pre-determined boundaries. Under stop loss protection, the benefit of the cover is available to the reinsured when the deductible is exceeded due to an increased frequency of loss, from the size of individual losses, or a combination of the two. Aggregate protection is related to the sum of all losses in the cover period rather than single events. ### Retention Per Risk In seeking the optimum profile for the retained account, the reinsured’s eventual choice of types of reinsurance will depend upon the claims pattern of the class of insurance involved. The reinsured also needs to consider whether a fixed retention on every risk in each class should be taken, or graded retentions applied according to the nature of the risk. #### Retention Under Proportional Reinsurance As mentioned earlier in this section, the frequency of loss may vary between diferent types of risk in the same class. Such variations should be reflected in the premium rates applied to the original risks. In a perfect portfolio, each risk retention would bold he same quantum of risk. It can be argued that 'per risk' retentions should be graded according to the preinium applied to the risk, so that the retained premiuin on each retention is similar. There are limitations to this approach, since premium reflects two aspects of risk, namely the probability of a claim arising and the probable extent of such a claim. At the upper end, retentions must be limited so that a single claim, however low the probability of that claim, should not be allowed to affect the account to too great an extent. At the lower end, high raters of premium, if they reflect a high frequency of small to medium claims, may suggest that the particular section of the account is well balanced and merits a retention and, which although graded down, will be disproportionately higher than the premium would suggest. However, in general, the retention should reflect the reinsured’s view of the risk. The more hazardous and highly rated the risk, the lower the retention. On the other hand, the better the risk, the lower the rate, and the higher the retention should be. Premium is often influenced by competitive forces. However, retentions should never be related inflexibly to premium since that would result in retentions being increased while competition causes the rates to be reduced. Under proportional fire treaties, not only will retentions be graded, but they will also usually be based upon the EML rather than the insured values as EML assessments are themselves scientifically based judgements of what might be lost in a claim event. By basing retentions on EML., the reinsured is able to retain more risk and more premium. #### Retention Under Non-Proportional Reinsurance Under a non-proportional reinsurance contract, it is unusual to apply graded deductibles to the protected portfolio. There are considerable technical difficulties in fixing a fair premium on an excess of loss contract operating on such a basis. The difficulty lies in the fact that under excess of loss insurance, the reinsurance premium is related to the overall premium attaching to the protected account, rather than to the individual risks in the account. Changes in the content and mix of the protected portfolio (which under proportional reinsurance arrangements with grated retentions are directly reflected in the reinsurance premium ceded) may, under excess of loss arrangements with graded deductibles, increase (or reduce) the exposure of reinsurers without appropriate compensating adjustments to the reinsurance premium. Excess of loss reinsurance provides the reinsured with a larger retained premium incoine than that wjich is retained if proportional reinsurances are placed, since there is no ceding of premium income to the reinsurer on every risk in the portfolio. There is little need for grading, since greater fluctuations in claims cost can be absorbed in the retained account. Generally, therefore, the retention or deductible is the same for each risk written by the reinsured. Graded contracts a;re sometimes found in marine excess of loss programmes, particularly on a marine cargo account. Excess of loss protection at a working or exposed level should not absorb losses which have high frequency, particularly if there is little fluctuation in the frequency. As the frequency of losses of similar size increases, the portfolio shows a better balance less reinsurance is required. Unnecessary excess of loss reinsurance, particularly at the lower end, should be avoided. In a normal market, reinsurance premiums paid at the lower levels will always exceed claims recovered. This is to the extent of the loading which excess of loss reinsurers require to cover costs, profit, fluctuation reserve and, if placed through intermediaries, brokerage. Therefore, when fixing a retention per risk under excess of loss reinsurance, a reinsured must consider the acceptable level of fluctuation of