Assessing the Impact of Reinsurance on Insurers Solvency under Different Regulatory Regimes

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Working Paper Assessing the Impact of Reinsurance on Insurers Solvency under Different Regulatory Regimes Eugene N. Gurenko, Lead Insurance Specialist Alexander Itigin, Actuarial Consultant Draft GCMNB June 9, 2010

2 Background This paper aims to identify discrepancies between an economic risk-based view of insurers solvency net of reinsurance and the regulatory treatment of reinsurance under three most common regulatory solvency regimes: Solvency I, Solvency II and Swiss Solvency Test (SST). The main rationale behind this comparison is to pinpoint potential shortfalls in the regulatory treatment of reinsurance for the purposes of solvency calculations. The paper also makes practical recommendations on how to best utilize reinsurance for the purposes of risk management and increasing profitability which may be of interest for insurers and insurance regulators. To build the argument, the paper utilizes numerous reinsurance scenarios in which we test the effects of most common forms of reinsurance (e.g. Quota Share (QS), and Excess of Loss (XL)) on the solvency position of a hypothetical insurance company specialized in writing catastrophe risk under different regulatory regimes compared to the risk-based economic view of solvency. The selection of a specialized catastrophe risk insurer has been driven mainly by a more pronounced discrepancy arising from a comparison of the regulatory and economic views of solvency net of reinsurance for portfolios of catastrophe risk. The proposed work builds on the previous research by the authors on the accuracy and reliability of risk-based solvency measures in approximating the level of risk retained by insurance and reinsurance entities specialized in writing catastrophe risk. Methodological Approach To demonstrate the differences between the regulatory and economic views of minimum riskbased capital required for a given portfolio of insurance risk for different levels of risk retention, we subject the modeled portfolio of catastrophe risk to a series of reinsurance scenarios that vary by the extent and type of risk transfer. For each test, we then compute the solvency margin required under Solvency I, Solvency II and SST and for an economic risk-based view of solvency. The portfolio of insurance risk subjected to the solvency tests consists of stand-alone catastrophe insurance policies covering earthquake related property damages of residential dwellings in four countries of South-Eastern Europe. For simplification purposes, we assumed that the company has no investment risk (i.e. invests all its surplus and reserves into cash and short-term highly rated government paper) hence the risk capital requirements for the assets were set at zero. Again, for simplification reasons, we also assumed away the operational risk, whose quantification is required under Solvency II. Event data set

3 To carry out our analyses, we used a stochastic data set of insured property losses in South East Europe. The data was provided by RMSI (2008), a risk modeling consultancy. The data set approximates losses that would be sustained by an insurance portfolio consisting exclusively of earthquake risk related liabilities, e.g. a portfolio of stand-alone earthquake insurance policies for residential property in four countries Croatia, Bosnia and Herzegovina, Macedonia and Albania. A deductible of 3 percent was assumed for each insured risk. For each country, the RMSI provided an estimate of insured residential loss in the capital city and in the rest of country. Insurance penetration was assumed to be 8 percent in capital cities and 6 percent outside. The probability of occurrence for each loss event and its severity were estimated by RMSI based on the underlying physical hazard and building vulnerability data. We assumed that the event frequency is distributed according to the Poisson distribution. From the original event data set described above, we derived the expected frequency (Poisson lambda) of A customized distribution for loss severity results directly from the original event data set. Through a Monte-Carlo simulation for these frequency-severity parameters, we produced 100,000 data records of annual residential earthquake insured losses. Our sensitivity analysis was based on this derived data set. The data set specifies both (i) the total annual loss for the total residential portfolio in four countries and (ii) the total annual loss for each country separately. Further, for each country the loss in the capital city and outside is specified. Each of the 100,000 annual insured losses can occur with the entry probability of 1/100,000 = percent. Measures of required solvency capital net of reinsurance For the above described portfolio of insurance risk we calculated the insurer s solvency margin capital required under Solvency I, Solvency II and the SST regimes. The mathematical formulas used for these calculations are described below. Solvency I Required Solvency Capital = max (Premium Index, Claims Index) Premium Index = Retention Rate * (0.18 * min (Gross Premium, 50m) * max (Gross Premium - 50m, 0)) Claims Index = Retention Rate * ( * min (Mean Gross Loss last 3 yrs, 35m) * max (Mean Gross Loss last 3 yrs - 35m, 0)) Where Retention Rate = Total Retained Loss for last 3 yrs / Total Gross Loss for last 3 yrs As we assume that there hasn't been any loss and the retention rate was held steady over the last 3 years, the above formula for the Retention Rate has been substituted with the one below: Retention Rate = Total Retained Premium / Total Gross Premium Solvency II 2

4 Required Solvency Capital = + Scenario per occurrence loss with return period 200 yrs Expected Annual Net Retained Loss - RI Recoverables where Scenario per occurrence loss with a return period 200 yrs denotes the event loss which can be expected not to be exceeded more often than once in 200 years. This value can be estimated directly from the annual per occurrence probability loss exceedence curve for the modeled portfolio of risk provided in Table I below. RI Recoverables stands for reinsurance recoverables that will be received from reinsurers under a company s reinsurance program. Please notice that in our definition of Required Solvency Capital (RSC), we follow the Standard Model approach rather than the Internal Model approach. According to the Internal Model philosophy, RSC is calculated based on the Annual Net Retained Loss VaR (99.5%) instead of a scenario loss with return period 200 yrs. Swiss Solvency Test Table 1. Occurrence Exceedence Probability Curve ( ) Severity OEP REP Conf level 315,397, % 262,694, % 221,425, % 196,232, % 150,292, % 97,640, % 86,330, % 70,192, % 53,958, % 27,387, % 9,501, % Required Solvency Capital = + Annual Net Retained Loss TVaR(99.0%) - Expected Annual Net Retained Loss where Annual Net Retained Loss TVaR (99.0%) denotes the Annual Net Retained Loss Tail Value at Risk (also often called the Expected Shortfall) at the confidence level of 99.0 percent. This is the average annual aggregate net retained loss calculated from the data set values that exceed those in excess of 99.0 percentile of the annual aggregate net retained loss distribution. As opposed to VaR which looks at losses up to a predefined percentile (thus disregarding distributions tails), TVaR approximates the average value of the tails. Table 2 below presents the comparison of the VaR and TVAR values of the annual aggregate loss exceedence curve for the modeled risk portfolio. 3

5 Table 2. Annual Aggregate Net Retained Loss Exceedence Curve ( ) Conf level Value at Risk Tail Value at Risk 99.00% 159,253, ,180, % 165,545, ,407, % 175,110, ,467, % 184,185, ,707, % 195,865, ,085, % 211,438, ,398, % 222,773, ,926, % 247,751, ,518, % 282,934, ,248, % 320,521, ,236,285 The differences between the VaR and TVAR approaches are further illustrated in Figure 1. Figure 1 Simulations In this Section we demonstrate that for the same portfolio of insurance risk protected by a given reinsurance program insurers operating in three different jurisdictions (say, in EU countries in 2012, in Switzerland in 2011, and now in all EU countries) will end up with three very different regulatory solvency requirements. Each reinsurance scenario differs in some way from the previous one in terms of the type of reinsurance treaties used (XL vs. QS) and the level of overall risk retention by the insurer. We then calculate the required solvency capital for each of three regulatory solvency regimes and compare them with an economic risk-based solvency requirement. The latter is defined as the risk-based capital needed to support a portfolio of retained risk (e.g. net of reinsurance recoverables) for loss scenarios with a return period of up to 500 years on a per occurrence basis 4

6 (e.g. VAR with 99.8% confidence level) 1 minus annual expected loss. The regulatory solvency margin requirements are then expressed as a percentage of the minimum risk-based economic capital. Base Case Scenario Under the base case scenario the insurer assumes all the risk, i.e. transfers no risk to reinsurers. As a result the insurers keeps the entire premium - 45 mm, see Table 3. Table 3. Reinsurance Premium vs. Retained Premium Total Reinsurance Premium 0 Total Retained Premium 45,069,552 The estimates of the required regulatory solvency capital (RSC) under three different solvency regulations are presented in Table 5, which also compares how these amounts correspond to the economic capital required by the S&P for an AA rated company, which for the purposes of this paper has been defined as follows: EC = OEP 500 year loss RI Recoverable Annual Net Retained Expected Loss= 252,286,069 Where OEP 500 year loss is defined as a loss with the exceedence probability of (100%- 99.8%=0.2%) per individual occurrence for a given portfolio of catastrophic risk. The values of OEP Loss Exceedence Costs can be obtained from Table 1. RI Recoverable stands for reinsurance recoverables, e.g. the maximum amount that can be recovered from reinsurers in case of a catastrophic loss covered by the insurer. In the case of the Base Scenario no risk is transferred to reinsurers, hence this value is set at zero. Annual Net Retained Expected Loss (ANREL) is the annual long-term average expected payout covered out of the insurer s own financial resources (e.g. not covered by reinsurers). The ANREL value can be found in Table 4 below. Table 4. Annual Net Retained Loss Annual Net Retained Expected Loss 10,408, To compare how the regulatory treatment of reinsurance (for the purposes of solvency calculation) differs from the economic effect of reinsurance on the required minimum economic risk capital (EC) we introduced the ratio of required Regulatory Solvency Capital (RSC) to Economic Capital (EC) for three different solvency regimes. As can be seen from Table 5, the values of the ratio significantly vary from one regulatory solvency measure to another. The RSC/EC ratio for Solvency I is the lowest of all, which demonstrates how inadequate the measure of Solvency I is for mono-line insurers involved in writing volatile types of risk (e.g. catastrophe, aviation, credit risk insurance). Although the ratios for two other measures of risk- 1 While we acknowledge a certain degree of arbitrariness in the selection of the return period for our definition of economic solvency capital, we would like to point out that in its definition of a minimum risk-based capital for a AA rated insurer, the S&P proposes a similar definition. The readers however are invited to reset the return period to either a higher or lower number, based on their own views of what constitutes a prudent level of financial resilience for an insurance/reinsurance entity. 5

7 based solvency Solvency II and SST are different as well, they are considerably closer to the economic risk capital required for the utilized portfolio of risk. The difference between the two ratios can be attributed to the mathematical definitions of risk chosen 99.5% VAR on a per occurrence basis for Solvency II and TVAR for a 99.0% VAR on an aggregate annual risk basis for SST. Table 5. Regulatory Solvency Capital Requirements vs. Economic Capital Solvency Capital EC* RSC/EC Solvency I 8,112, ,286, % Solvency II 185,824, ,286, % Swiss Solvency Test 217,771, ,286, % As can be seen from Table 5, standing at 86.3% the RSC/EC ratio for the SST demonstrates that it is the most stringent of three solvency measures due to the use of the 99% TVAR approach to risk measurement. The RSC/EC ratio for the SST is closely followed by that of Solvency II (73.7%), which by definition disregards all risk scenarios beyond a 200 year return period on a per occurrence basis. Scenario 1: Reinsurance - 20% Cession to a Quota Share Reinsurance Treaty Under this scenario, the insurer cedes 20% of the risk and premium to a reinsurer on a proportional quota-share basis. See Table 6 below. Table 6. Retained and Reinsured Premium Total Reinsurance Premium 9,013,910 Total Retained Premium 36,055,641 We assume that the treaty has no aggregate reinsured limit and thus provides for a complete risk transfer regardless of the ultimate size of insured loss. In Table 7 below we provide the outcomes of this simulation. Table 7. Regulatory Solvency Capital vs. Economic Capital Regulatory Solvency Capital 6 EC Delta to base case scenario RSC/EC Solvency I 6,490, ,828, % 3.2% Solvency II 148,659, ,828, % 73.7% Swiss Solvency Test 174,217, ,828, % 86.3% Notes: *ANREL = 8,326,474. As expected, a 20% cession of the risk across the whole portfolio results in a 20% reduction for both RSC and EC, which can be observed by comparing the values of RSC/EC ratio to those obtained under the Base Case Scenario the ratios stayed the same. Since the values of the RSC/EC ratio changed proportionately by 20%, one can see that in this specific case the regulatory treatment of reinsurance for the purpose of providing solvency relief fully corresponds to the economic reality a 20% risk reduction in the required risk capital achieved through a risk transfer to a third party.

8 Scenario 2. Reinsurance 80% to a Quota Share Reinsurance Treaty In this scenario, we continue to test the impact of a proportional reinsurance treaty on the changes in the regulatory measures of the required solvency capital compared to those for the required economic capital. In this case, the insurer transfers 80% of risk and premium to a reinsurer. As shown in Table 8 below, although one would expect the risk transfer to result in an 80% reduction for all three measures of RSC, this is not the case for Solvency I, where only a 50% reduction is achieved. This less than expected drop in the RSC under Solvency I can be explained by the fact that insurers cannot realize a solvency relief when their risk retention drops below the minimum prescribed by the Solvency I formula. While in the previous simulation this was not a constraint due to a very insignificant amount of risk transfer, in this case the constraint became binding resulting in a disproportionate reduction in the required solvency margin. We must point out that this discrepancy between economic and regulatory outcomes of risk transfer in the case of Solvency I solvency regulations can be viewed as distortionary as it creates strong incentives for companies to retain more risk, which in the case of mono-line writers can be highly counterproductive from the risk management perspective. However, as shown by the RSC/EC ratio the two other RSC measurers dropped proportionately to the risk transfer, e.g. by 80%. Similarly, we can observe a proportional drop in the amount of premium retained by the insurer see Table 9. Table 8. Regulatory Solvency Capital Requirements vs. Economic Capital Solvency Capital Delta to base case scenario EC* RSC/EC Solvency I 4,056, % 50,457, % Solvency II 37,164, % 50,457, % Swiss Solvency Test 43,554, % 50,457, % Notes: *ANREL = 2,081,618 Table 9. Reinsurance vs. Retained Premium Total Reinsurance Premium 36,055,641 Total Retained Premium 9,013,910 Scenario 3. Reinsurance: 40% cession to QS; CAT XL: 85m xs. 5m (up to REP 100 yrs) In this simulation we introduce a more complex reinsurance program whereby an insurer transfers 40% of total risk to a proportional Quota Share treaty while capping the remaining 60% of the risk with an 85 million Excess of Loss Catastrophe Risk reinsurance treaty (CAT XL), which attaches at 5million the insurer s own minimum risk retention. The overall effects of this risk transfer are summarized below. 7

9 Table 10. Regulatory Solvency Capital Requirements vs. Economic Capital Solvency Capital Delta to gross portf. EC* RSC/EC Solvency I 4,056, % 70,470, % Solvency II 30,593, % 70,470, % Swiss Solvency Test 45,146, % 70,470, % Notes: *ANREL = 2,145,885 The main finding of this simulation is a considerable deterioration in the ratio of the required regulatory capital to the economic capital relative to the previous case scenario. Despite the fact that the economic capital requirements increase from $50.4 million in the previous case to $70.4 in this scenario (due to a lesser amount of economic risk transfer provided by this program), we can observe that Solvency I capital requirements remain the same, while those under Solvency II even drop from $37.16 million to $30.6 million. This is due to the fact that the Solvency II measure of RSC by design does not account for any risk in excess of a 200 year return period on an annual per occurrence basis. Hence, even though the $85 million layer of reinsurance in excess of $5 million in addition to a 40% proportional cession (e.g. $104.8 m) did a better job reducing the risk from an annual event with a 200 year return, it did not transfer any risk beyond that point. The SST was only solvency measure that captured the lack of risk transfer beyond that point by increasing the amount of regulatory capital required from $43.5 to $45.1 million see Figure 2. With the RSC/EC ratios being 5.8% and 43.4%, respectively, Solvency I and Solvency II measures of RSC clearly fall short of approximating the true economic effect of this risk transfer on the solvency requirements e.g. while the economic quality of this risk transfer is inferior to that in the previous scenario, the Solvency I and Solvency II minimum capital requirements remained the same or dropped. Another interesting observation is a considerable increase in the retained premium for the insurer from $9 million in the previous case to $21.7 million in this scenario due to the reduced amount of available reinsurance protection see Table 11. An insurer operating under the Solvency II regulations in a competitive market is likely to have no other choice but to structure its reinsurance coverage in a way which would enable to achieve the minimum reinsurance cost while meeting the regulatory solvency requirements. Table 11. Reinsurance vs. Retained Premium Total Reinsurance Premium 23,336,734 Total Retained Premium 21,732,818 8

10 Figure II. Economic vs. Regulatory Capital $262 m X QS $196 m 40% 500 OEP REP 200 OEP REP $85 m XL $5 m Scenario 4. Reinsurance: 40% cession to QS; 112 m xs. 5m (up to REP 200yrs) to CAT XL In this Scenario, similarly to Scenario 3, the insurer employs a combination of proportional and non-proportional (excess of loss reinsurance). Keeping the percentage of risk/premium ceded to the QS treaty at 40%, the insurer increases the overall amount of coverage provided by the CAT XL treaty from 85 to 112 million a net increase of 27 million in reinsurance coverage. The results of the simulation are summarized below. Table 12. Solvency Capital vs. Economic Capital Solvency Capital Delta to base case EC* SC/EC Solvency I 4,056, % 43,668, % Solvency II 3,791, % 43,668, % Swiss Solvency Test 28,394, % 43,668, % Notes: *ANREL = 1,948,137 Table 13. Reinsurance vs. Retained Premium Total Reinsurance Premium 23,631,259 Total Retained Premium 21,438,292 9

11 Several interesting observations can be made based on the results of this simulation. First, the amount of solvency capital required under Solvency II is below that required under Solvency I. The outcome can be attributed to the fact that while the solvency margin under Solvency I remains the same until a certain minimum risk retention level has been reached, the RSC under Solvency II is phased out to zero as the amount of reinsurance protection approaches the expected probable loss from a 200 year event - the maximum amount of risk capital required at this confidence level. The effect is clearly demonstrated in Table 12 which presents the values of the RSC/EC ratio. The biggest discrepancy between the economic capital and that required by the regulators arises for Solvency I and Solvency II measures, 9.3% and 8.7%, respectively. This case is of particular interest as it shows that a risk-based measure of solvency such as Solvency II can result in a larger discrepancy with the economic measure of solvency than Solvency I, a nonrisk based measure of solvency. Another important finding is the almost negligible additional premium cost required to obtain a very considerable solvency relief achieved under Solvency II compared to the previous scenario. If in Scenario 3, an insurer would retain million in premium and would be required to put up 30.6 million in solvency margin requirements; under the current scenario, an insurer would keep million in premium but would be required to set aside only 3.8 million in solvency capital. A simple calculation demonstrates that by forgoing only 0.3 mm the insurer would be realizing a capital cost saving of 4.02 million 2 - quite a stellar return on the investment! This discovery strongly underscores the cost effectiveness of XL type coverage for insurers operating under Solvency II. While being inferior in quality to proportional reinsurance treaties (with no aggregate limits) XL coverage allows insurers to achieve regulatory solvency capital relief at a much lower reinsurance cost. We also observe that the solvency relief provided under the SST in response to a 27 million increase in XL coverage amounts to only million, which raises interesting questions about the adequacy of this risk-based solvency measure when it comes to estimating the economic effects of XL reinsurance coverage on insurers capital position. We are going to explore this specific issue in more detail in the next three simulations. Scenario 5. No Proportional Cession to QS; XL - 190m xs 5m (up to REP 200yrs) In this simulation we explore the effects of an XL reinsurance program 190 million in excess of 5 million on the regulatory solvency requirements as well as the economic capital of the insurer. The results of the simulation are summarized below. Table 14. Regulatory Solvency Capital vs. Economic Capital Requirements Solvency Capital Delta to base case scenario EC RSC/EC Solvency I 6,258, % 70,260, % Solvency II 3,799, % 70,260, % Swiss Solvency Test 42,005, % 70,260, % 2 The calculation assumes the insurer s cost of capital of 15%. 10

12 Notes: *ANREL = 2,433,819 Table 15. Reinsurance vs. Retained Premium Total Reinsurance Premium 10,297,915 Total Retained Premium 34,771,637 The first most obvious finding of this simulation is that compared to the previous simulations the insurer achieves the lowest required RSC under Solvency II at the lowest reinsurance cost which presents a tempting opportunity for the regulatory arbitrage. At the same time, while achieving the maximum regulatory capital relief at the lowest cost possible, the insurer is compromising on the quality of its reinsurance protection as shown by a very low RSC/EC ratio of 5.4%. Hence, the finding demonstrates considerable economic benefits for the insurer of strictly complying with the Solvency II capital requirements without considering the impact of the reinsurance program on its economic capital the company can now retain 34.8 of the premium out of the total 45.1 written while, at the same time, obtaining an almost complete capital relief. As opposed to the previous scenario, by substituting a 40% of QS 3 reinsurance coverage with the additional 78 million of XL coverage, the insurer sees its RSC requirements under the SST increased from 28.4 million to about 40 million, which demonstrates a strong regulatory bias for QS coverage rather than CAT XL for the purposes of solvency relief. Scenario 6. No Proportional Cession to QS; XL 225 m xs 5m (up to TVAR of gross portfolio) In this and the following simulation (Scenario 7) we will focus our attention exclusively on the effect of XL reinsurance on the regulatory solvency requirements under SST. We demonstrate that the marginal impact of additional layers of XL reinsurance on the company s regulatory solvency requirements gets less and less pronounced as we move out over and above the level of protection required for a 200 year event. We also show that despite this effect the insurer is considerably economically better off employing the XL reinsurance coverage rather than QS type treaties or its own capital for meeting the minimum SST requirements. As can be seen from Table 16, compared to Scenario 5, in this case the insurer bought 35 million more of XL protection, which results in the reduction of the regulatory solvency capital under the SST by only about 13 million clearly a highly disproportionate solvency capital relief. Table 16. Regulatory Solvency vs. Economic Capital Solvency Capital Delta to base case EC RSC/EC Solvency I 6,219, % 35,402, % Solvency II 2,708, % 7.6% 3 Depending upon the definition of the event return period, the amount of reinsurance coverage to be afforded under a 40% QS treaty may vary from 80 million (for a 200 year loss) to 200 million for a 500 year loss. 11

13 35,402,399 Swiss Solvency Test 28,970, % 35,402, % Notes: *ANREL= 2,291,764 Table 17. Reinsurance Premium vs. Retained Premium Total Reinsurance Premium 10,517,411 Total Retained Premium 34,552,141 The marginal cost of this capital relief however is rather small 0.22 million- which compares rather favorably with the insurer s own opportunity cost of financing 13 million in regulatory capital requirements in the absence of the XL risk transfer. Assuming the expected return on equity of 15%, the insurer would be spending 1.95 million in capital financing costs, which demonstrates the benefits of substituting the insurer s own risk capital with reinsurance in the upper layers of the risk transfer program. Scenario 7. No Proportional Cession to QS; XL 455 m xs 5m In this final simulation, we aim to further emphasize the inherent limitation of the SST which manifests itself in the unattainable full solvency relief no matter how much of XL reinsurance one can buy. In this case, the insurer more than doubles the limit of reinsurance protection acquired under the previous scenario by going from 225 to 455 million of XL protection. With this level of reinsurance protection, the insurer can comfortably cover itself against any loss that falls within a 1000 plus return period, e.g. well beyond the probable maximum loss chosen for our definition of the economic capital (a 500 year event or million loss). Yet, compared with the previous scenario, under the SST the additional reinsurance 230 million of reinsurance coverage bought by the company yields it a regulatory capital relief of only million see Table 18 below. This effect can be attributed to the TVAR measure utilized by the SST which produces the effect of an ever evading full solvency relief, e.g. the effect of a cat chasing its own tail. We also observe that since our definition of the economic risk capital has capped the maximum required at the level of the loss equivalent to a 1 in 500 year event, our EC and Solvency II RSC requirements have been more than fully satisfied. 4 Nevertheless, despite the disproportionately small marginal effect of the additional XL coverage on the solvency relief under the SST, the insurer may still find it beneficial to buy extra layers of XL reinsurance coverage at a relatively low cost. For instance, the additional 230 million of reinsurance coverage would cost the insurer only million extra vs. the opportunity cost of million in own risk financing costs to achieve the same relief in solvency margin. Hence, the practical implication of the above described drawback of the SST measure are rather limited. Table 18. Required Solvency Capital vs. Economic Capital Solvency Capital Delta to base case EC* SC/EC Solvency I 6,163, % % Solvency II 2,895, % % 4 Our calculation of required economic risk returns zero (rather than negative numbers) when the set capital limit is exceeded. 12

14 Swiss Solvency Test 11,778, % % *Notes: *ANREL = 2,104,031 An interesting paradox is also observed in the change of the solvency margin required under Solvency II despite the fact that compared to the previous scenario the insurer now has more than 200 million of additional reinsurance protection its solvency margin requirement goes up from million to million. Table 19. Reinsurance Premium vs. Retained Premium Total Reinsurance Premium 10,827,550 Total Retained Premium 34,242,002 Estimating the Impact of Counterparty Credit Risk on Reinsurer s Solvency In this Section we consider the impact of counterparty credit risk on the solvency requirements under the SST and the Solvency II regulations. There are no such requirements under the Solvency I regulations. The counterparty risk in this case is defined as the risk of reinsurers default on their obligations i.e. a failure to pay reinsurance claims in full to reinsured parties. Obviously, in case of such a default, the reinsured parties will be faced with reduced reinsurance recoverables and additional incremental increases in the required risk-based solvency margin both of which will lead to the immediate need for more capital. To provide for such adverse scenarios, the SST and the draft Solvency II regulations 5 require insurers to set aside additional capital related to the credit quality of reinsurers. A brief description of the methodologies used for computing these credit-risk related charges can be found in Annex I. In this Section we provide only the outcomes of credit-risk related calculations. Swiss Solvency Test Under the SST, the Required Capital for credit-risk related charges varies with the credit rating of reinsurers in accordance with the following formula: Required Capital = Exposure * Weighting Factor * 8% Where Exposure = Expected annual loss ceded to RI and the weighting factors for different ratings are presented in Table 20. Table 20. Reinsurance Credit Risk Weighting Factors S & P Rating Weighting Factor AAA 20% AA 20% 5 See Technical Specifications for QIS 5 draft. 13

15 A 50% BBB 100% BB 100% B 150% CCC or lower 150% Based on the above formula and the weighting factors we have obtained the following results see Table 21 below. Table 21. Reinsurers Credit Risk Related Solvency Requirements under Solvency II Required RI Scenario Exp loss ceded to RI Required Capital AA RI Required Capital A RI Capital BBB RI 1 2,081,619 33,306 83, , ,326, , , , ,262, , , , ,459, , , , ,974, , , , ,116, , , , ,304, , , ,325 As can be seen from the table, in case Scenario 1, the credit risk charges under the SST can add in additional RSC requirements for an AA rated reinsurer to million for and BB rated reinsurer, respectively a 500% difference! However, if expressed as a percentage of the total solvency margin required net of credit risk charges, these amounts appear to be rather trivial well under 1%. Solvency II: RI Scenario Table 22. Reinsurance Credit Risk Related Solvency Requirements ( ) RSC_Crdt, AA RSC_Crdt AA/ RSC_UW RSC_Crdt, A 14 RSC_Crdt, A /RSC_UW RSC_Credit, BBB RSC_Credit, AA/RSC_UW 1 438, % 980, % 2,146, % 2 1,755, % 3,923, % 8,585, % 3 1,827, % 4,086, % 8,941, % 4 2,129, % 4,760, % 10,417, % 5 2,124, % 4,748, % 10,390, % 6 2,137, % 4,779, % 10,458, % 7 2,137, % 4,779, % 10,458, % Table 22 shows that in the case of Solvency II reinsurance related credit risk solvency requirements can be many times the solvency margin computed without the counterparty credit risk. For instance, in the case of Scenario 6, we notice that even for an AA rated reinsurer in case

16 of a small retention, an insurer would have to put up the additional 79% of original solvency capital in the form of credit quality related charges. This number goes to 386% in case of a nonrated reinsurer! Such a stringent approach to the counterparty credit risk is likely to considerably raise the capital costs of insurers seeking reinsurance protection hence pushing them to increase their own risk retentions or co-insure the risk with their own reinsurance captives located in the offshore zones. Conclusions Based on the presented simulations of reinsurance protection for a given portfolio of catastrophe risk under three different solvency regimes, we would like to draw the following conclusions: 1. Depending upon a regulatory framework, the same reinsurance protection can result in very different solvency margin requirements, which creates ample possibilities for regulatory arbitrage. 2. The upcoming Solvency II regulations appear to significantly underestimate the risk of extreme events by assuming away all potential losses which fall beyond the 99.5% confidence level. 3. The Swiss Solvency Test produces more conservative measures of solvency for monoline writers involved in highly volatile lines of business (such as catastrophe risk, aviation, etc.). 4. At the same time, for portfolios of risk with low frequency high severity characteristics, the SST does not provide adequate credit for XL reinsurance as it suffers from the ever diminishing effect of non-proportional reinsurance on the required solvency margin. In the extreme the SST produces a paradoxical situation for the insurer when the RSC can no longer be materially reduced by purchasing additional reinsurance. 5. The counterparty credit risk charges contemplated by the draft Solvency II regulations seem to make reinsurance protection uneconomical if purchased from below investment grade reinsurers and when the amount of risk transfer many times exceeds the risk retention. The effect is likely to favor larger insurers capable of retaining more risk and while co-insuring the rest of the risk with the group captives established in offshore zones. The latter can in turn reinsure the way they see fit without being penalized with additional credit risk related charges for risk transfer to third parties. 15