1 CBFA Belgian Q&A QIS4 - List of Methodological Issues Raised by participants Peter De Vos 6/26/2008 General Disclaimer The answers given below are not official CBFA positions but tentative answers referring to QIS4 only. The document contains questions asked by the Belgian participants of the QIS4 exercise for which an answer was given by CBFA. For some of these questions the CEIOPS - QIS4 Task Force provided the answer. Other questions still need to be answered because additional information was requested and was not yet received. Stakeholders are free to comment on this document. Please use the following address to that end:
2 Table of contents 1. General 2. Section 1: Valuation of Assets and Liabilities 3. Technical provisions 4. Annex 1: IFRS-Accounting/Solvency adjustments for the valuation of assets and other liabilities under QIS4/ and Annex 2: Proxies 5. Section 2: Own Funds 6. Section 3: Solvency Capital requirement: the standard formula/ SCR general remarks 7. SCR Risk Mitigation 8. SCR Calculation Structure 9. SCR Market risk module 10. SCR Counterparty risk module 11. SCR Life underwriting risk module 12. SCR Health underwriting risk module 13. SCR Non-Life underwriting risk module 14. Section 4: Solvency capital requirement: Internal models 15. Section 5: Minimum Capital requirement 16. Section 6: Groups 17. Annexes The questions below have been regrouped according to the chapters of the table of contents.
3 CHAPTER 1 Question 1.1.: Op de CEIOPS-site wordt er een excel met helper tabs ter beschikking gesteld. Volgen hier ook nog verdere instructies over wat en hoe in deze tabs juist dient ingevuld te worden? Answer 1.1.: The instructions are included in the spreadsheet instructions, starting on page 114. Question 1.2.: The difference between the QIS4 balance sheet calculations and the IFRS or local GAAP calculations will in most cases highlight a substantial release in buffer reserves. Does this available capital (locked in current accounting provisions) need to be reported under I. General tab row 152 "Valuation adjustments to liabilities). If this is the case, should this include a tax reduction? Answer 1.2.: Regarding cel F152, the spread sheet instructions refer to cells E46 to E54 which include insurance liabilities. Question 1.3.: I m writing you to know if for each release and update from CEIOPS concerning the Solo spreadsheet, an automatic update related to the patching tool is provided or we have to use the only one we had up to date for all the further update. Answer 1.3.: The patching tool for the solo spreadsheet works with any solo spreadsheet version.a little background info: The patching tool is designed in such a way that it uses the dataset structure of the new spreadsheet to find out what inputs it shall copy from its predecessor s input sheet. For the differently designed other workbooks however, we have to provide a different solution. CHAPTER 2 Question 2.1.: Helper RM Non Life : wat is definitie average duration? formule simplified MVM maakt steeds gebruik van intrest termstructure 2007 en Dit is toch afwijkend van de formule in TS.II.C.25 Answer 2.1.: Question 2.2: Wordt de afwikkeling van driehoeken met chain Ladder methode op basis van Paid en Incurred als 2 verschillende methoden gezien (paragraaf TS.II.E.20)? Answer 2.2.: Our understanding is that the method is identical but that the data used are different. However, the exact content of the quoted paragraph in our view does not imply that undertakings are obliged to use two different methods (..some general principles are suggested:...) although this is highly recommended. Question 2.3.: Waar moet de reserves voor Gemeenschappelijk Motorwaarborgfonds opgenomen worden in QIS 4? Dient dit bij overige technische voorzieningen opgenomen te worden? Moet deze reserve meegenomen worden bij de bepaling van de Best Estimate van de schadereserves BA Auto? Answer 2.3.: The provision " Gemeenschappelijk Motorwaarborgfonds" is a claims provision and should be included in the determination of the best estimate. Question 2.4.: Worden de te gebruiken indexen voor inflatie per LOB gepubliceerd of opgegeven of dient elke onderneming apart een inschatting te maken van de te gebruiken index? Answer 2.4.: Indexen voor inflatie worden niet gepubliceerd. Mocht inflatie een rol spelen dan moet
4 daarmee rekening worden gehouden in de modellisering van de cash flows. CHAPTER 3 Question 3.1.: Fiscale gevolgen en behandeling vrijval reserves door verschil boekhoudkundige reserves en marktconforme waardering reserves (BE + RM) (belastingslatentie)? Answer 3.1.: In principle, Solvency II should be tax neutral. The supervisory authorities are not competent with tax matters and therefore for this kind of questions the fiscal authorities should be seized. Question 3.2.: Hoe gebeurt de bepaling van BE en RM van de vergrijzingsreserves bij individuele hospitalisatie? Answer 3.2.: As the best estimate is the present value of future cash flows, in principle the ageing provision will no longer be required as it will automatically be part of the overall best estimate plus risk margin. Question 3.3.: Behandeling toekomstige premies in niet-leven meer bepaald levenslange dekking in individuele hospitalisatie : tacit renewals = inschatting maken van percentage van portefeuille dat komende jaar gaat vernieuwen of enkel degenen die niet meer kunnen opzeggen (3 maanden voor vervaldag)? Answer 3.3.: It may be helpful to clarify the treatment of future renewal premiums by reference to the broader narrative included within the CFO Forum Elaborated Principles ( EP, the relevant extract is included in an annex to this Q&A document for your information) which sought to distinguish recurring premiums (future premiums which form part of the existing contract) and renewal premiums (premiums paid for a new contract). It is the intention to include the former and exclude the later. According to CFO Forum Elaborated Principles the expected cash flows from a contract should be taken into account beyond the first year if the insurer does not have the ability to freely and independently re-price and re-underwrite this contract on an annual basis. However, in the case of a contract that is re-priced and re-underwritten annually, no future renewal premiums would be taken into account since there is no stand ready obligation beyond 1 year in respect of this particular contract with a term of only 1 year. For further details on the treatment of future premiums, please refer to annex 1 of this document. Annex 1: Extract from CFO Forum Elaborated Principles D. Policyholder behaviour, including recurring premiums and lapses, should be reflected in the measurement of liabilities. Renewal options or provisions that obligate the insurer to continue to provide coverage should be recognised to the extent they are included in the contract. BC45) There are two different types of renewal premiums. Future premiums that form part of a contract (recurring premiums) and premiums paid for a new contract (renewal premiums e.g. on a motor policy). In addition, renewal options occur where the insured has the right to renew the current contract with or without underwriting or re-pricing. BC46) Both renewal options and recurring premiums need to be reflected in the measurement of liabilities. However, renewal premiums for a new contract, which can be underwritten and re-priced, should not. BC47) The occurrence and amount of different payments (claim, maturity, surrender, paid up)
5 are contingent on the payment of premiums. Equally, the payment of recurring premiums is contingent on the non-occurrence or occurrence of insured events or policyholder lapses. It is not possible to consider one element of the cash flows without considering the impact on all other elements. BC48) Further, for some risks, notably life and health cover, the risk of a claim increases over time, while the recurring annual premium is usually kept level throughout the term. As a result, in the later stages of a contract, the policyholder continues to renew because it would not be possible to obtain the same cover for the same premium under a new contract. In this manner, the payment of recurring premiums is related to the level of cover yet to be provided. Recurring premiums EP35) The cash flows included in the estimate of the insurance liability should only include cash flows associated with the current insurance contract and any existing ongoing obligation to service policyholders. This should not include expected renewals that are not included within the current insurance contract EP36) Recurring premiums should be included in the determination of future cash flows, with an assessment of the future persistency based on actual experience and anticipated future experience. BC49) Policyholders decisions to lapse policies or take renewal options are based on a wide range of factors, such as personal circumstances and preferences, or institutional factors (e.g. changes in tax treatment, changes in regulations and legal changes) and the financial consequences of the decision being taken. Although it may be in all policyholders financial interest, under certain economic conditions, to lapse their policy and purchase a new one, in practice many policyholders will continue with their existing policies, perhaps because they would not be able to obtain a similar policy on the same or better terms, for example, due to impaired health. Similarly, policyholders decisions on whether to take up a guaranteed renewal option will depend on not only the financial aspects of the guarantee but also on their personal circumstances, for example, whether they need that type of policy and whether they would be able to obtain an alternative policy. BC50) Most current insurance accounting approaches include all cash flows from the current contract in the measurement of the liability. As a contractual relationship exists between the insurer and the policyholder, the contract will normally be amended if premiums are not paid, so it is appropriate to include all future cash flows expected to arise during the current contract term. Renewal options EP37) Where a contract includes options or guarantees that provide rights under which the policyholder can obtain a further contract on favourable terms (for example, renewal with restrictions on re-pricing or further underwriting) then the value of these guarantees and options should be included in the evaluation of the insurance liability arising under the existing contract. Where no such restrictions on re-pricing or underwriting exist, there is no ongoing obligation to service policyholders. BC51) Insurers issue some contracts that include the option to increase or renew insurance coverage. Under these contracts, the value of the option varies in line with the insurer s ability to re-underwrite the risk or change the premium. Under other contracts, the insurer is constrained in its ability to adjust the terms of the option. Where an insurer has the ability to re-underwrite risks fully, on a policy-by-policy basis, it is likely that the option will have zero value. Where there is limited ability to re-price, for example, the current rate for new business must be charged for renewals, there will be some potential cost. Options to effect additional insurance on fixed terms are likely to give rise to the most onerous additional cost. BC52) Such renewal options should be valued as a cost arising during the term of the current
6 contract and included in the measurement of the liability. The valuation approach may involve projecting the additional future cash flow arising from the option. However, this projection is aiming to establish a cost, not bring forward cash flows from future contracts. Question 3.4.: Tijdens de opleidings-sessie van levensverzekeringen werd gesteld dat igv opeenvolgende koopsommen enkel die koopsommen moeten meegenomen worden waar vandaag reeds de voorwaarden (oa garantierente) gekend zijn. Indien dit niet zo is, moeten deze koopsommen niet meegenomen worden bij de berekeningen. Als ik deze redenering doortrek voor Universal Life produkten waar voor de toekomstige premies het tarief niet vastligt (garantierente niet, maar ook sterftetarief en kosten in het tarief kunnen gewijzigd worden voor lopende contracten), dan betekent dit dat voor dergelijke produkten nooit toekomstige premiestortingen moeten beschouwd worden. VRAAG: Is bovenstaande stelling correct? Answer 3.4.: See question 3.3. Question 3.5.: aansluitend op vorige vraag Indien zuivere overlijdensverzekeringen worden aangeboden zonder tariefgarantie (dwz dat voor bestaande contracten het tarief en dus ook de toekomstige premies kunnen worden aangepast), dan zou ook daar geen toekomstige premies in rekening gebracht moeten worden. Echter, indien geen toekomstige premies in rekening gebracht moeten worden, zal de overlijdensdekking niet voor de ganse contractperiode verzekerd kunnen blijven. VRAAG: Mag dan bij de bepaling van de toekomstige kasstromen gesteld worden dat van zodra de reserve niet meer toereikend is, de polis gestopt wordt (wat met zekerheid voor de einddatum van het contract gesitueerd is, als nog niet alle premies betaald werden)? Answer 3.5: In this particular case only the cash flows for the coming year are to be taken into consideration. CHAPTER 4 Question 4.1.: Zullen gemiddelde durations van verschillende LOB s voor de Belgische markt gepubliceerd worden (voor eventuele toepassing van discounting proxy)? Answer 4.1.: De durations zijn gepubliceerd op de CBFA-website tesamen met de marktontwikkelingspatronen per tak van verzekering. CHAPTER 5 Question 5.1.: Vrijwillig aangelegde egalisatievoorziening valt weg binnen IFRS en bijgevolg ook hier? Answer 5.1.: Egalisation provisions will in principle no longer exist under solvency II as unforeseen losses are to be covered by capital. Under QIS4 they can entirely be taken up as own funds. With regard to the final solvency II regime and the possibility of having egalisation provisions, we need further reflexion on the subject. Question 5.2.: Wat gebeurt er met de wettelijk aan te leggen egalisatievoorziening (bvb Brand)? Komen die in de Balance sheet economic vision onder onder other liabilities, fully loss absorbent? Answer 5.2.: In our view part of the provisions may go into the best estimate, the remainder becoming own funds. See also question 7.
7 Question 5.3.: Aangezien er impact van egalisatiereserves in health on dutch market is op standaarddeviatie premium & reserving risk is er dan ook geen impact voor België? In de laatste versie van TS onder TS.XII.C.10 als voetnoot nl. For the calculation of the premium and reserve risk in health business subject to the risk mitigating effects of a market-wide and mandatory equalisation system, as exists in the Dutch market, a specific treatment would be justified. For example, the Dutch market proposes to calibrate (M, prem,lob) and (res,lob) for QIS4 purposes to 1%. Note that the shock is contingent on the status of the equalisation system and may change accordingly in the future. For more information, see Annex SCR5 - TS.XVIIG Answer 5.3.: The treatment proposed for the Dutch health insurance is based on a thorough study by the industry on how the mitigating effect of the equalisation provisions impacts on volatities. To our knowledge, no such study was done for the Belgian insurance market. It follows that it is very difficult for CBFA to propose an alternative for the standard methodology given that simply taking over what is proposed for the Dutch market is not an option because of the divergencies in markets. CHAPTER 6 Question 6.1.: Existe-t-il un moyen dans le cadre du QIS4 de considérer le business non modélisé. En effet, une partie de notre business (moins de 1%) n est pas modélisé dans notre logiciel actuariel de projection. Est-il possible d effectuer les chocs et autres résultats demandés uniquement sur le business modélisé et de renseigner la partie résiduelle d une autre manière? Answer 6.1.: The non modelised part of the business could be considered to be not material and therefore even be left out of the QIS4 reporting. This is however not the prefered option. We refer to paragraphs TS.I.B.7 to TS.I.B.13 that explain what a participant needs to do when the conditions of paragraph TS.I.B.12 are fulfilled. Question 6.2.: Quels sont les produits qui doivent être regroupés sous le vocable 'Health'? (adaptation au marché belge)? Est-ce tous les produits baptisés 'Health' dans les spécifications du QIS4 (ou dans le projet de directive) doivent être traités avec le module 'Health underwriting risk module' ou est-ce qu'on peut avoir des produits 'Health' traités dans les module Vie et/ou nonvie? (si oui, lesquels?). Answer 6.2.: We refer to the chapter on segmentation of risks for non-life and health insurance business (paragraphs TS.VI.B) that clearly states how to deal with the health risk. Further, we also refer to question 2 of chapter 3 and question 1 of chapter 12 of the Q&A document that you can find on the CEIOPS-website. Question 6.3.: Connexe tarieven Voor wat betreft het mortaliteitsrisico en de daarbijbehorende schok wordt er gevraagd om de dekking leven van de dekking overlijden te scheiden onder bepaalde voorwaarden. Er is sprake van een contract. Volgens de economische logica zou mijn inziens een connex tarief als 1 contract behandeld moeten worden. Volgens de letter van de tekst volgens mij als aparte contracten. Wat is uw mening hieromtrent? Answer 6.3.: We refer to the chapter on segmentation of risks for non-life and health insurance business (paragraphs TS.VI.B) that clearly states how to deal with the health risk. Further, we also refer to question 2 of chapter 3 and question 1 of chapter 12 of the Q&A document that you can find on the CEIOPS-website.
8 Question 6.4.: Hoe gebeurt de berekening van de deferred taxes in de praktijk? Is hierover meer info beschikbaar? Answer 6.4.: We refer to the content of the specification (TS.VI.I.). As to more practical details, deferred taxes are treated exactly as FPS. We refer to paragraph TS.VIII.A.2. CHAPTER 7 Question 7.1.: If during the projected 12 months period ceded reinsurance treaties are expiring before the end of this 12 months period (for example when the SCR is calculated at 31/03, 30/06, etc.), can we assume: that these treaties will be renewed as expiring, or should we take into account a potential change in the conditions, or should we calculate a capital charge as if there will be no ceded reinsurance anymore after the expiration date? Answer 7.1.: Since under QIS4 there is no capital charge taking into account a potential change in the conditions so far, it will be difficult to accept those renewals as they are. On the other hand, if there is already a cover note for the renewal, the treaty will probably be renewed, back to the potential change in conditions. CHAPTER 8 Question 8.1.: Pourriez-vous nous donner plus d informations pratiques sur le calcul du ADJ(deferred taxes) à prendre en compte dans le calcul du SCR global? Answer 8.1.: We refer to the content of the specification (TS.VI.I.). As to more practical details, deferred taxes are treated exactly as FPS. We refer to paragraph TS.VIII.A.2. Question 8.2.: Voor de berekeningen van de SCR is het noodzakelijk om voor het winstdelings-gedeelte een BE te bepalen van het gedeelte mbt de reeds in het verleden toegekende WD en mbt de toekomstige WD. Dit betekent dat alle toekomstige kasstromen strikt gescheiden moeten kunnen bepaald worden. In produkten waarbij het overlijdenskapitaal gelijk is aan het maximum van een vooraf vastgesteld bedrag en de totale reserve (contractueel + WD), is het niet éénvoudig te bepalen welk deel van het uit te keren overlijdenskapitaal toe te schrijven is aan de toekomstige WD. Zeker zolang er een risicokapitaal bestaat, dient met verdeelsleutels bepaald te worden welke deel naar WD (toekomst en verleden apart) gaat en welk deel naar het contractuele deel. VRAAG: Mag hier een verdeling gebruikt worden welke ook gehanteerd wordt voor de bepaling van de fiscale afhoudingen (waarbij ook het deel WD en contractueel moet vastgelegd worden)? Answer 8.2.: Question 8.3.: Le premier calcul du SCR de chaque module est effectué sur la base des prévisions de versements de PB cohérentes avec les pratiques actuelles : cela suppose-t-il bien de maintenir un taux constant de taux de PB (en pourcentage des produits financiers, et techniques), ou plutôt un montant constant (en pourcentage des provisions revalorisées), donc aucune absorption de perte ni même de diminution du rendement des actifs financiers? Puis le calcul des nscr s effectue o (a) suivant les intentions du management o (b) pour le lower boundary nscr au taux minimum prévu par la
9 réglementation et les contrats. Answer 8.3.: The specification does not prescribe the way in which the profit sharing is to be modelized, hence it is perfectly possible that the undertaking assumes a fixed profit sharing percentagen or a fix amount or something else. CHAPTER 9 Question 9.1.: Dans le calcul du risque de concentration, doit-on inclure le cash? Answer 9.1.: The second part of TS.IX.G.4 provides an answer to this question: "Bank deposits with a term of less than 3 months terms, of up to 3 million Euros, in a bank that has a minimum credit rating of AA are also exempted from an application of this module." CHAPTER 10 Question 10.1.: For internal reinsurance we use a Bermuda based entity. It is an unrated entity with a solvency ratio of 297%, based on Bermudan solvency rules. Is it possible to assume a AAA rating for the entity as it suggests in the QIS4 guidance? The Solvency margin calculation is not risk based and is almost the same as the current Solvency I basis that we currently use. The guidance is slightly unclear as the ratio table does not say whether it should be on a Solvency II basis or the local basis. The entity is used solely for internal reinsurance throughout the world. Answer 10.1.: For QIS4 purposes, participants are requested to assess the solvency ratios based on Solvency II rules. Question 10.2.: Is it correct that mortgage loan portfolios, aside from the interest rate module, also have to be considered under the counterparty default module and treated as an unrated exposure? This treatment is considered disproportionate as the stress factor that has to be used is rather high given the possible risk selection. Could you please confirm this point of view and comment the rather severe treatment under counterparty default. Answer 10.2.: It is correct that mortgage loans should be considered under counterparty default risk as other credit exposure. Concerning disproportionate treatment, this might be alleviated by considering the collateral effect of a mortgage. Contrary to exposures to reinsurers or in financial derivatives, the counterparty default risk does not explicitly takes into account collateral for exposures to intermediaries or for other credit exposures. Instead participants might include the collateral implicitly as a risk mitigation instrument in the calculation of the best estimate of the credit exposure, given that the risk mitigation principles outlined in TS.VII apply. Especially with respect to mortgage loans it might usually be assumed that the principles are fulfilled. Question 10.3.: We do not think that we can perform a Solvency II calculation for our unrated non-eea reinsurer which we use for internal reinsurance only, so we will not be able to apply the method in TS.X.A.19. How should we then assess the counterparty default risk for this firm? Answer 10.3.: For QIS4 purposes, a treatment according to TS.X.A.19 requires the calculation of an SCR of the intragroup reinsurer following the methodology laid out in the Technical Specifications. If such a calculation is not possible, TS.X.A.11 applies in order to determine the probability of default of the reinsurer (probability of default for an unrated reinsurer not subject to Solvency II regulation: 30.41%). Participants are invited to comment on this approach when answering the qualitative questionnaire.
10 Question 10.4.: I have a question regarding the loans and the scope of the spread/concentration/default modules. Is it correct to include the loans in the spread/concentration calculation. In fact both submodules are part of the market risk module. As loans are held to maturity and they are not exposed to the market spreads, it will be incorrect to cover them in the spread/concentration modules. Answer 10.4.: With regard to spread risk, I think we can agree that there is no need to include loans there as the value of the loan is not influenced by a change in creditworthiness of the borrower. We are not sure however about concentration risk where in our view the risk should be taken into account although it is probably close to impossible to have a capital charge coming thereoff. Question 10.5.: In TS.X.A.16 (Counterparty Default Risk), the formula for the calculation of Default of counterparty "i", regarding an implicit correlation R=1, is: Def(i)=LGD(i)*min(100*PD(i),1) This means that the probability of default applied for counterparty "i" is equal to min(100*pd(i),1). For the different solvency ratios described in page 158, using the formula described above means applying the following default probability (Probability Used): I understand that for R=1 (meaning a Herfindahl coefficient =1, which is possible when there is only 1 counterparty), we should be using the probability of Default that corresponds to that unique counterparty. I would like to know why the probability of default is increased by 100 times when it is less or equal than 0.24%. Answer 10.5.: Given the 99.5% calibration of the SCR, a 1-in-200-years event is assumed in the counterparty default risk module. Expected default probabilities (related to rating) are shocked in such a scenario: In stressed circumstances the rating goes down and therefore the probability of default increases. Putting the factor at 100 translates the assessment of the downgrading of the counterparty and is a calibration element which you might want to comment when answering the qualitative questionnaire. Question 10.6.: TS.X.A. SCR def counterparty default risk, it is about formula given for LGD (TS.X.A.3). It is stated the following: LGD = 50% max(recoverables + SCR gross u/w - SCR net u/w - Collateral;0) net I understand SCR u/w is the SCR calculated with the standard formula (net of reinsurance, for all counterparties - NOT by counterparty), and about SCR gross u/w : Is this gross of counterparty "i", meaning that we should quantify the effect of "not counting on" counterparty "i"? If this is true: how would you suggest to make this calculations? We find some difficulties in this part. If this gross means without reinsurance at all (without all counterparties), the difference will be much bigger and will be counted as "exposure" for each
11 counterparty, which increase real exposure. Answer 10.6.: The Technical Specifications are indeed referring to the calculation of SCR nl on a counterparty by counterparty basis, both gross and net, where gross means disregarding the reinsurance effected with that counterparty. A possible approximation for the counterparty risk calculations for non-life reinsurance could be to calculate LGD as follows LGD = 50% Max(Recoverables + NL pr /V * (Recoverables + gross earned premiums net earned premiums) + Additional capital charge relating to reinsurance recovery in event of CAT losses Collateral; 0) where NL pr is the normal risk capital charge for premium and reserve risk for all non-life business (net of reinsurance) and V is the corresponding volume measure for all non-life business net of reinsurance. Recoverables and premiums each refer to the counterparty in question. The 'Additional capital charge relating to the reinsurance recovery in event of CAT losses' would be the notional capital charge calculated (by the formulae in TS.XII.C.6 or TS.II.C.10 as appropriate) in respect of the potential recoveries from that reinsurer that would arise in the event of the CAT losses envisaged in the calculation of NL CAT. This would avoid the need to make multiple calculations of the SCR component for underwriting risk. CHAPTER 11 Question 11.1.: Il est demandé pour le choc «expenses» de corriger les frais de gestion de 10%, puis le taux d inflation de 1% par an. Pour un taux d inflation la première année de 2%, cela signifie-t-il - une inflation de 3% pendant toutes les années suivantes? - ou une inflation de 2,02% l an 2, 2,04% l an 3, 3,3% l an 50? Answer 11.1.: See CEIOPS Q & A. (antwoord = Belgisch scenario) Question 11.2.: Meer instructies omtrent niet CATNAT catastrofes zoals bvb pandemie :a% aantal verzekerden in hospitalisatie x gemiddelde schadelast terrorrisme : aandeel aantal verzekerden in hospitalisatie t.o.v. inwoners in België x b x gemiddelde schadelast x c a,b,c te bepalen door CBFA voor de Belgische markt a afleiden uit Spaanse griep b aantal mensen op een evenement of luchthaven c verhoging van gemiddelde schadelast gezien extreme situatie Indien je een portefeuille hospitalisatie hebt, speelt dit toch een rol en zeker als je een laag eigenbehoud hebt in CATNAT Answer 11.2.: Meer instructies van CBFA voor niet CATNAT catastrofes zijn niet voorzien voor de QIS4 oefening. Ik neem uw opmerking echter wel mee in de voorbereiding van volgende oefeningen. Voor de Belgische markt werd er geopteerd om de scenari van QIS3 te behouden daar de feedback die wij terzake hebben gekregen niet meteen aangaf dat het aangewezen was op dit punt wijzigingen aan te brengen. In termen van QIS4-specificaties worden de ondernemingen verzocht methode 2 toe te passen. Dit betekent evenwel niet dat de ondernemingen zich moeten beperken tot de voorgeschreven regionale scenari. Immers, de specificatie voorziet onder methode 3 in de mogelijkheid om zelf scenari te ontwikkelen zo die representatiever zijn voor de onderneming. U kan uiteraard van deze mogelijkheid gebruik maken.
12 Question 11.3.: TS.XI.F.3: Please find in the table below a possible outcome for the shock to be tested. Is it correct and if not how should the shock be calculated? Baseline Shock expenses +10,0% inflation 2,0% inflation 3,0% year expenses year expenses 1 100, , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,38 Answer 11.3.: The scenario outlined in the table mirrors the intentions of CEIOPS when designing the life expenses sub-module. CHAPTER 12 Question 12.1.: In Belgium long-term health products are sold with similar characteristics to the German ones (life long coverage, leveled premiums depending on the entry age and calculated with life insurance techniques, obligation to set up aging reserves, possibility to annually adjust premiums in order to maintain profitability, solvency margin calculated according to Life,...): should the underwriting risk for these products consequently be treated in the long-term health module, or in the life module? And what about guaranteed income (disability) and long-term care insurance, where we have leveled premiums and aging reserves as well? In what module should these be treated? Answer 12.1.: Firms should place these contracts in the most suitable module for the nature of the risks borne by the insurer in respect of these products, and explain their selection. Comments would also be welcome on the suitability of the approach described in Annex SCR6, or any other approach that firms may wish to propose. Question 12.2.: Waar moeten de Belgische individuele hospitalisatieverzekeringen, zoals gedefinieerd onder de nieuwe wet, ondergebracht worden? Onder short term (omwille van korte afwikkelingsperiode voor één voorvalsjaar) of long term (omwille van levenslang kararkter) health products? Indien onder long term graag extra toelichting hieromtrent. Answer 12.2.: De nieuwe wet op de ziekteverzekering verandert mijn inziens niets. Deze wet maakt een levenslange dekking voor de verzekeringnemer mogelijk maar verandert in wezen niets aan de aard van de prestaties die in de ziekteverzekering kunnen worden verzekerd. Het criterium
13 hetwelk wij in het QIS4 kader moeten hanteren is, de afwikkelingsperiode van de schadegevallen. I may add that there will of course be a difference regarding the treatment of future premiums because the premiums on these contracts qualify as recurring premiums in the sense of paragraph TS.II.B.33. See also question 14. Question 12.3.: Impact van vergrijzingsreserve op SCR (en MCR)? Bij Health short term wordt in de standaardberekening enkel met de premiereserves en schadereserves rekening gehouden. Bij Health long term wordt er m.i. geen rekening gehouden met schadereserves. Answer 12.3.: See question 12. For long term health it appears that under claims expenditure solely claims payments are included. Question 12.4.: Kan je mij zeggen wat een arbeidsongevallen verzekeraar moet invullen in de sheet I.Health and Non-Life, cel E 165? Ik neem aan dat dit enkel de reserves zijn die worden behandeld als "Niet Leven" gezien dat er op dit volume een reserve risico wordt berekend. De kapitaalsvereisten voor de "annuities" worden berekend aan de hand van scenario's. Answer 12.4.: Indeed, only the non-life provisions are to be included. Question 12.5.: In paragraph TS.XII.B.28, claims expenditure is mentioned. Is this payments only or does it also include the change in the technical provisions? Answer 12.5.: It is not intended that claims expenditure in the health epidemic and accumulation risk covers any allocations to the technical provisions. Included, however, is any profit or loss resulting from the run-off of claims provisions. Additionally, any expenses incurred with the settlement of claims should be included. CHAPTER 13 Question 13.1.: In the QIS4 methodology spreadsheet the premium risk is defined as "TS.XIII.B.2 Premium risk is understood to relate to future claims arising during and after the period until the time horizon for the solvency assessment. The risk is that expenses plus the volume of losses (incurred and to be incurred) for these claims (comprising both amounts paid during the period and provisions made at its end) is higher than the premiums received (or if allowance is made elsewhere for the expected profits or losses on the business, that the profitability will be less than expected). " If one wants to report the premium risk, using an internal model (which models + premiums - costs - losses), should the required capital (that will be compared with the QIS4 SCR) then be calculated as the difference between the 99,5% Var and the expected value (and hence the expected profit should be included somewhere in available capital), or should the required capital rather be defined as the difference between the 99,5% VaR and zero (and hence expected profit is used as first cushion for risk)? When an internal model is used to forecast the result of 1 year of (expected) earned premium, in most cases the technical P&L is modelled: + premium (comes in); - Costs and commission (are paid) - Claims (stochastic) are paid + Reinsurance recoveries are received (also stochastic) - Reinsurance premium is paid (also stochastic). = Technical P&L (discounted or not)
14 In this case the expected result (mean) equals the expected proft, whereas in a 1/200 VaR scenario, one could end up with a significant loss. Based on TS.XIII.B.2 there are two options given: The risk (=capital) is that the volume of losses exceed the premium => capital is defined as the difference between zero and the above obtained 99.5%VaR (which is negative) Or - if allowance is made elsewhere for the expected profit (which I assume should then be included in the available capital) - that the profitability will be less then expected (and hence capital is defined as the difference between the expected profit and the 99.5%VaR loss). Based on how premium risk is defined in QIS4 (TS.XIII.B.17), I assume that it is the first option (and hence the profit included in the premium acts as a first buffer for capital). But I would like confirmation on that. Otherwise it would be difficult for us to provide the results of internal models. Answer13.1.: CHAPTER 14 CHAPTER 15 CHAPTER 16 CHAPTER 17 Question 17.1.: When solely considering the definition (contractual or legal arrangement whereby part of the assets or eligible surplus of the company are strictly segregated from the rest of the company's investments or resources and can only be used to meet the insurance and/or resinsurance obligations with respect to which the ring-fenced fund has been established), we understand that the belgian segregated accounts should be considered as such structures. However, reading further on the specifications, we understand that the "non-transferability" of the own funds held within the ring-fenced structures should be the decision criterion. In our understanding, this is not the case for our segregated accounts. You invited us to consider own funds as they are defined within the SII framework which leads more or less (rather more than less) to identify segregated accounts as ring-fenced structures. Answer 17.1.: Contacts with concerned regulators abroad lead us to the conclusion that indeed the Belgian seggregated accounts can be considered ring-fenced structures. We recognize that insurers in Belgium operate with-profit business, of which some part of this business is managed as seggregated funds. However, we have no formal legislation that limits the transfer of assets or excess own funds of such funds to other businesses or entities within a group, and therefore we have not provided national guidance on the treatment of local withprofit business for the purpose of QIS4. On the other hand, we do expect that insurers contractually set restrictions and apply internally these restrictions since the assets are dedicated to seggregated fund policyholders only, and are not available to cover losses elsewhere. Therefore, no diversification effects can be recognized by pooling assets or liabilities of seggregated and non-seggregated funds. This limitation would also apply within a group context. Concretely, we would propose the following treatment for QIS4:
15 For local calculation: step 1: separation of individual seggregated funds from non-seggregated funds business step 2: calculate SCR for the non-seggregated funds business, taking into account diversification effects step 3: calculate for each seggregated fund the SCR step 4: sum all SCR's of seggregated funds step 5: add the sum of SCR's of seggregated funds to the SCR of non-seggregated funds business For group calculation: step 1: eliminate all local seggregated funds from the scope of consolidation step 2: use consolidated SCR calculation of the group, excluding the local seggregated funds step 3: add the sum of SCR's of seggregated funds (as in step 4 for local calculation) to the consolidated SCR (as in step 3)