Discussion Document 105 (v 3) was approved as a Position Paper by Steering Committee on 12 September

Transcription

1 Solvency Assessment and Management: Pillar 1Sub Committee Capital Requirements Task Group Position Paper (v 3) Market Risk SCR Structure and Correlations EXECUTIVE SUMMARY This document discusses the structure of the Market Risk Module of the standard formula SCR and the calibration of the correlation matrix relating thereto being proposed for SAM. Note that this document does not discuss the approach to be taken for risk aggregation under SAM. This is discussed in Discussion Document 48 (SCR Standard Formula Correlations). The task group recommends the following: The correlation matrix aggregation approach recommended in Discussion Document 48 is considered suitable for the Market Risk module. This allows for flexibility in allowing for dependencies between risks, and is consistent with the way other risk modules and submodules within the Standard Formula are treated. The structure of the Market Risk module as tested under SA QIS3 seems more relevant and suited to South African circumstances than the Solvency II structure. Input is required from the credit risk working group as to the structure of risk mitigation impairments (reference to credit risk discussion document); from the technical provisions task group in order to assess the need for and nature of an illiquidity premium sub-module (reference to the relevant discussion document); and in progress (Discussion Document 106) with regard to inclusion of interest rate and equity implied volatility stresses. A separate working group is also looking at the treatment of participations (Discussion document 53). The structure of and correlations within the spread risk/credit default risk module are outside the scope of this document. Position Paper 47 discusses the need for a separate SA equity category and correlations between the different categories. The correlation parameters tested under SA QIS2 seem reasonable, and whilst data is unavailable or scarce, the overall diversification effect does not seem out of line with European experience. The task group is of the opinion that correlation of 0 between concentration risk and other market risk sub-modules is appropriate since concentration risk is targeted at nonsystematic risk in contrast to other market risk events. The counter-cyclical premium is included for Solvency 2 where this forms part of the technical provisions. Since this is not relevant to SAM, it will not be considered further and is to be removed from the recommended SAM correlation matrix. There is the potential to include a two-sided currency correlation (if there is an equity fall it is probably more likely that there will be a currency depreciation than an appreciation). From QIS 2 results it became apparent that around 20% of respondents to QIS 2 were exposed to the risk of the Rand depreciating. This is a significant proportion of insurers and hence the task group proposes a two-sided correlation in this instance. 1 Discussion Document 105 (v 3) was approved as a Position Paper by Steering Committee on 12 September 2014.

2 A default (after a look-through and mandate based methodology of classifying assets in collective investment schemes) to Other Equity is preferred rather than a methodology allocating it to either SA equity or Other equity to maximise SCR because: it is not deemed appropriate for an instrument or scheme to switch between classifications from one valuation to the next, and it shows complicated or unknown instruments risks in one place. 1. INTRODUCTION AND PURPOSE This document sets out the recommendations of the Capital Requirements task group with respect to the structure of the Market Risk module and the calibration of its correlation matrix within the Standard Formula SCR. 2. INTERNATIONAL STANDARDS: IAIS ICPs IAIS is the international standards setting body for insurance supervisors. The standards are principles based and set out high level guidance on the setting of solvency capital requirements. There is no reference to detailed capital requirements below the level of market risk. However, the following are relevant within the broad framework of the capital requirements, of which market risk forms part (reference: Insurance Core Principles, Standards, Guidance and Assessment Methodology 1 October 2011 ): ICP 17 Capital Adequacy The supervisor establishes capital adequacy requirements for solvency purposes so that insurers can absorb significant unforeseen losses and to provide for degrees of supervisory intervention. Some sub-points in this standard that should be considered includes: 17.1 The supervisor requires that a total balance sheet approach is used in the assessment of solvency to recognise the interdependence between assets, liabilities, regulatory capital requirements and capital resources and to ensure that risks are appropriately recognised The supervisor addresses all relevant and material categories of risk and are explicit as to where risks are addressed, whether solely in technical provisions, solely in regulatory capital requirements or if addressed in both, as to the extent to which the risks are addressed in each. The requirements are also explicit as to how risks and their aggregation are reflected in regulatory capital requirements. Types of risks to be addressed The supervisor should address all relevant and material categories of risk - including as a minimum underwriting risk, credit risk, market risk, operational risk and liquidity risk.. Dependencies and interrelations between risks The assessment of the overall risk that an insurer is exposed to should address the dependencies and interrelationships between risk categories (for example, between underwriting risk and market risk) as well as within a risk category (for example, between equity risk and interest rate risk). This should include an assessment of potential reinforcing effects between different risk types as well as potential second order effects, i.e. indirect effects to an insurer s exposure caused by an adverse event or a change in economic or financial markets conditions. It should also consider that dependencies between different risks may vary as general market conditions change, Page 2 of 21

3 and may significantly increase during periods of stress or when extreme events occur. "Wrong way risk", which is defined as the risk that occurs when exposure to counterparties, such as financial guarantors, is adversely correlated to the credit quality of those counterparties should also be considered as a potential source of significant loss e.g. in connection with derivative transactions. Where the determination of an overall capital requirement takes into account diversification effects between different risk types, the insurer should be able to explain the allowance for these effects and ensure that it considers how dependencies may increase under stressed circumstances The supervisor sets appropriate target criteria for the calculation of regulatory capital requirements, which underlie the calibration of a standardised approach The level at which regulatory capital requirements are set will reflect the risk tolerance of the supervisor. Reflecting the IAIS s principles-based approach, this ICP does not prescribe any specific methods for determining regulatory capital requirements Calibration and measurement error The risk of measurement error inherent in any approach used to determine capital requirements should be considered. This is especially important where there is a lack of sufficient statistical data or market information to assess the tail of the underlying risk distribution. To mitigate model error, quantitative risk calculations should be blended with qualitative assessments, and, where practicable, multiple risk measurement tools should be used. To help assess the economic appropriateness of risk-based capital requirements, information should be sought on the nature, degree, and sources of the uncertainty surrounding the determination of capital requirements in relation to the established target criteria The degree of measurement error inherent, in particular, in a standardised approach depends on the degree of sophistication and granularity of the methodology used. A more sophisticated standardised approach has the potential to be aligned more closely to the true distribution of risks across insurers. However, increasing the sophistication of the standardised approach is likely to imply higher compliance costs for insurers and more intensive use of supervisory resources (for example, in validating the calculations). The calibration of the standardised approach therefore needs to balance the trade-off between risk sensitivity and implementation costs. 3. EU DIRECTIVE ON SOLVENCY II: PRINCIPLES(LEVEL 1) Relevant extracts from the Solvency II level 1 principles are provided below. As is the case with the IAIS core principles, these requirements are in nature of a higher level than required for the establishment of detailed requirements for the Market Risk module within the capital requirements. However, it provides the broad framework within which these requirements are to be considered. Article 101: Calculation of the Solvency Capital Requirement The Solvency Capital Requirement shall be calibrated so as to ensure that all quantifiable risks to which an insurance or reinsurance undertaking is exposed are taken into account. It shall cover existing business, as well as the new business expected to be written over the following 12 months. With respect to existing business, it shall cover only unexpected losses. It shall correspond to the Value-at-Risk of the basic own funds of an insurance or reinsurance undertaking subject to a confidence level of 99,5 % over a one-year period The Solvency Capital Requirement shall cover at least the following risks: (a) non-life underwriting risk; Page 3 of 21

4 (b) life underwriting risk; (c) health underwriting risk; (d) market risk; (e) credit risk; (f) operational risk. Article 104: Design of the Basic Solvency Capital Requirement The Basic Solvency Capital Requirement shall comprise individual risk modules, which are aggregated in accordance with point (1) of Annex IV. It shall consist of at least the following risk modules: (a) non-life underwriting risk; (b) life underwriting risk; (c) health underwriting risk; (d) market risk; (e) counterparty default risk The correlation coefficients for the aggregation of the risk modules referred to in paragraph 1, as well as the calibration of the capital requirements for each risk module, shall result in an overall Solvency Capital Requirement which complies with the principles set out in Article Each of the risk modules referred to in paragraph 1 shall be calibrated using a Valueat-Risk measure, with a 99,5 % confidence level, over a one-year period. Where appropriate, diversification effects shall be taken into account in the design of each risk module The same design and specifications for the risk modules shall be used for all insurance and reinsurance undertakings, both with respect to the Basic Solvency Capital Requirement and to any simplified calculations as laid down in Article 109. Article 105: Calculation of the Basic Solvency Capital Requirement 1. The Basic Solvency Capital Requirement shall be calculated in accordance with paragraphs 2 to The market risk module shall reflect the risk arising from the level or volatility of market prices of financial instruments which have an impact upon the value of the assets and liabilities of the undertaking. It shall properly reflect the structural mismatch between assets and liabilities, in particular with respect to the duration thereof. It shall be calculated, in accordance with point (4) of Annex IV, as a combination of the capital requirements for at least the following sub-modules: (a) the sensitivity of the values of assets, liabilities and financial instruments to changes in the term structure of interest rates, or in the volatility of interest rates (interest rate risk); (b) the sensitivity of the values of assets, liabilities and financial instruments to changes in the level or in the volatility of market prices of equities (equity risk); (c) the sensitivity of the values of assets, liabilities and financial instruments to changes in the level or in the volatility of market prices of real estate (property risk); Page 4 of 21

5 (d) the sensitivity of the values of assets, liabilities and financial instruments to changes in the level or in the volatility of credit spreads over the risk-free interest rate term structure (spread risk); (e) the sensitivity of the values of assets, liabilities and financial instruments to changes in the level or in the volatility of currency exchange rates (currency risk); (f) additional risks to an insurance or reinsurance undertaking stemming either from lack of diversification in the asset portfolio or from large exposure to default risk by a single issuer of securities or a group of related issuers (market risk concentrations). Annex IV (4): Calculation of the Market risk module The market risk module, set out in Article 105(5) shall be equal to the following: SCR Market = where SCRi denotes the sub-module i and SCRj denotes the sub-module j, and where i,j means that the sum of the different terms should cover all possible combinations of i and j. In the calculation, SCRi and SCRj are replaced by the following: SCR interest rate denotes the interest rate risk sub-module, SCR equity denotes the equity risk sub-module, SCR property denotes the property risk sub-module, SCR spread denotes the spread risk sub-module, SCR concentration denotes the market risk concentrations sub-module, SCR currency denotes the currency risk sub-module Note the correlation matrix (Corr i,j ) for the life underwriting module is not specified in Annex IV (4), although it is specified for the BSCR correlation matrix in Annex IV (1). 4. MAPPING ANY PRINCIPLE (LEVEL 1) DIFFERENCES BETWEEN IAIS ICP & EU DIRECTIVE No differences the EU Directive is in line with the IAIS core principles. 5. STANDARDS AND GUIDANCE (LEVELS 2 & 3) 5.1 IAISstandards and guidance papers This was covered in section 2 above CEIOPS consultation papers and Level 2 advice The following CEIOPS (replaced by EIOPA in 2011) paper was consulted for this discussion document: 2 The IAIS Insurance Core Principles, Standards, Guidance and Assessment Methodology issued October 2011 has superseded previous Standards and Guidance (in this case Standard No and Guidance paper No on the structure of regulatory capital requirements). Page 5 of 21

6 CEIOPS-SEC-40-10: Solvency II Calibration Paper ( QIS5 Calibration Paper ) QIS 5 Calibration Paper: Market risk CEIOPS has carried out extensive both qualitative and quantitative analysis to revise the correlation parameters of the market risk model in line with the 1:200 VaR target level for the calculation of the SCR In its draft advice, CEIOPS set out a qualitative assessment of the lessons learned from the financial crises with regards to this issue; and has proposed to increase the correlation coefficients for a number of risk pairs In the consultation, stakeholders have commented that such qualitative analysis in itself would be insufficient to derive an appropriate a revision of the factors. They took the view that it would not be sound from a statistical perspective if a calibration of correlations should be based exclusively on the observations derived from the current crises, and suggested that CEIOPS should undertake a more thorough statistical analyses based on historic data from a longer period of time CEIOPS acknowledges these concerns and has undertaken such further statistical analysis on basis of a methodology which is consistent with the general aims of setting correlation parameters. This intended to determine the overall level of diversification implied by the correlation matrix proposed, and to assess its appropriateness; and to statistically assess the correlation between individual pairs of risks in the market risk module using historical data The following sets out the results of CEIOPS analysis on this issue. More detailed background information on the statistical quantitative analysis undertaken is provided further below. General considerations and lessons learned from the financial crises The current financial and economic crisis provided further strong empirical evidence that the dependence structure of market risk changes in stressed situations. Risk factors that have not revealed a significant correlation during ordinary market conditions showed a strong dependence in the crisis. It could also be observed that the risks had a reinforcing effect on each other For all risks that are covered by the market risk module a strong simultaneous change in market parameters was observed: Credit spreads widened in an unprecedented manner. The market price for equity fell stronger than during the crises in 1973 or at the beginning of the century. The MSCI world index dropped by 40% and the STOXX 600 by 46% in Interest rates fell sharply, for example for German 10 year Government bonds by 30% in the second half of The key interest rates of the U.S. Federal Reserve System and the Bank of England were set to historic lows. Property prices in some markets strongly decreased. In the United States the Case Shiller Index dropped by 19% in Similar declines could be observed in some European markets. Exchange rates were also quite volatile. For instance, the British Pound lost 24% against the Euro in Also the currencies of Iceland and some other states outside of the Euro zone came under pressure. Page 6 of 21

7 This simultaneous adverse change across a range of market risk drivers left only limited scope for diversification, i.e. it was very difficult for market participants to offset losses with respect to one risk category with gains in other risk categories. Only where risks have a two-sided nature like interest rate risk or currency risk, market participants were able to offset risks if they were on the right side (for example short in Icelandic króna or short in interest rates) CEIOPS considers that for the calibration of the correlation parameters in the market risk module of the SCR standard formula the empirical evidence provided by the current crisis on the existence of a significant degree of tail correlations between different market risk drivers should not be ignored. In line with its general observations on the calibration of correlation parameters, CEIOPS has reflected this tail dependency in its statistical analysis for setting the correlation parameters in order to ensure that the aggregated capital requirements are in line with the 99.5% confidence level for the calculation of the SCR. Overall diversification benefit implied by proposed correlation matrix To test the overall appropriateness of the correlation matrix proposed in its draft advice, CEIOPS has carried out a statistical top down modelling analysis to assess whether the overall diversification benefit implied by the matrix is consistent with the 1:200 year confidence level targeted for the determination of the capital charge for market risk as a whole The diversification benefit implied by the matrix can be measured as where SCR mkt denotes the capital charge for market risk, Mkt r denote the capital charges for the individual market risks, and where is derived from the capital charges for the individual sub-risks by using the proposed correlation matrix CorrMkt r,c This diversification benefit as implied by the aggregation matrix is consistent with the targeted confidence level of 99.5% for market risk if it coincides with the risk-theoretic diversification benefit which is given as where VaR mkt denotes the Value-at-Risk 99.5% capital charge for market risk as a whole and VaR r denote the Value-at-Risk capital charges for the individual sub-risks of market risk Assuming that the calculation of the capital charges Mkt r of the individual sub-risks are commensurate with the 99.5% Value-at-Risk confidence level, it follows that the diversification benefit implied by the matrix is consistent with the 99.5% confidence level if the capital charge SCR mkt derived from aggregating the individual charges with the correlation matrix coincides with the risk-theoretic 99.5% Value-at-Risk capital charge VaR mkt for market risk as whole, i.e. if the aggregation error Page 7 of 21

8 is zero To carry out the analysis, a model of a typical European insurer as described in QIS4 was created with a standalone capital for market risks of 100. This is made up of: Interest Rates Equity Property 8.39 Spread Currency 5.22 Concentration Using 12 years of historical data for year on year falls in indices relating to each of the market risks, a simulated empirical calculation of the Value-at-Risk capital charges for the individual market sub-risks as well as for market risk as a whole was undertaken. This empirical simulation exercise then allowed a comparison of the risk-theoretic diversification benefit with the diversification benefit implied by the proposed correlation matrix Under this empirical analysis, the risk-theoretic diversification benefit for the aggregated market risk (in relation to its sub-risks) was determined as 17.3%, whereas according to the analysis the diversification benefit implied by the proposed correlation matrix was measured as 16%. This indicates that the correlation matrix proposed in CEIOPS draft advice provides overall capital figures which are broadly consistent with the targeted 99.5% Value-at-Risk confidence level. The analysis also included sensitivity testing of key assumptions, which indicated that the results of the analysis are relatively robust Notwithstanding this overall indication that its proposed pre-consultation correlation matrix for market risk appears to be broadly adequate,ceiops has undertaken further analysis on specific correlation pairs, as explained below. The revised proposal for a correlation matrix as set out in this paper is expected to lead to higher diversification benefits than estimated above For further detail on the analysis we refer to Annex A of CEIOPS-DOC-70/10. Statistical analysis on correlation between specific sub-risks For the setting of correlation parameters between specific pairs of subrisks in the market risk module, CEIOPS has complemented its qualitative assessment set out in its draft advice by a quantitative statistical analysis. This was based on an analysis of historical data onthe year on year percentage changes in the underlying risk drivers For example, to consider the correlation between interest rate risk and equity risk, the analysis was based on the MSCI world equity index from1970, compared with UK 10 year spot As was noted above, in view of the assumed tail dependence of market risks in stressed situations the correlation analysis was based on cutting out adequate subsets of data pairs in order Page 8 of 21

9 to obtain a measure of the tail correlation, as well as a measure of the weight in the tail as opposed to that expected by a simulated Gaussian copula Typically this involved a cut along various percentiles in each of the two variables. For example, the red boxes in the diagram below represent the data in the 99th percentile for equity and interest, the 95 th percentile for both, and the 90th percentile for both: Overall dependence between market sub-risks in stressed situations A strong fall of equity prices as reflected in the equity sub-module (-45%/-60%) does not leave the other market parameters unaffected. A drastic change in equity prices of this scale indicates an economic recession and a severe reduction of undertakings expected profit. Such a situation is usually accompanied with an increase of risk-aversion and higher default probabilities. Therefore, credit spreads can be expected to increase sharply as well. For the same reason the demand for property, in particular commercial property, can be expected to decrease leading to vacancies and lower property prices. On the other hand, if credit spreads widen as greatly as in the spread risk submodule, it signals an increased risk aversion and higher default probabilities. These circumstances would certainly affect directly or indirectly the expected profits and the market value of stock corporations in a relevant way, causing a fall in equity prices. Similar arguments apply to property risk. All three risks are intrinsically connected via the economic conditions, so that in extreme situations, they relate to each other in a similar way as in a causal relationship. These considerations indicate that a higher correlation factors between these risks might be appropriate Concerning the dependence between interest rate risk and other sub.-risks, we note that the monetary policy of the relevant central banks usually reacts to an economic downturn (and in particular to a fall in equity markets) by lowering the key interest rates. This can be observed for example in the downturn where the ECB changed the key interest rate for the euro from 4.75% to 2% or the current crisis where rates fell from 4.25% to 1%. Similar reactions took place in the UK (6% to 3.5% and 5.75% to 0.5% resp.) and the US (6.5% to 1% and 5.25 to 0.25% / 0% resp.). These are direct reactions to the adverse movements of the market parameters which are addressed in the market risk module, such as equity prices, credit spreads, property prices and exchange rates. The central banks attempt to flood the market with cheap money in order to mitigate the worsening of these parameters. If key interest rates fall sharply in economic crisis situations, then so do the risk-free interest rates. Therefore a high correlation of a fall in interest rates with an adverse change in the other market risks can be appropriate. Correlation between interest rate risk and equity risk Page 9 of 21

10 CEIOPS has carried out additional statistical analysis on the correlation between interest rate risk and equity risk as described above The results of this analysis indicate that the proposed correlation of 50% does not appear unreasonable. There is clearly a positive correlation between equity and interest rates On the other hand, a distinction should be drawn between correlations between a fall in interest rates and a fall in equity prices on the one hand, and between a rise in interest rates and a fall in equity prices on the other hand. Whereas there is clear statistical evidence of a positive correlation (in the range of 40% to 50%) between the first, much less data is available to support an analysis of the correlation between a rise in interest rates and a fall in equity prices In light of these conclusions, a two-sided correlation between interest rate risk and equity risk in the standard formula is introduced: In case the insurer is exposed to a fall in interest rate risk, a correlation parameter of 50% between interest rate risk and equity risk should be applied to aggregate the respective capital charges; In case the insurer is exposed to a rise in interest rate risk, a correlation parameter of 0% between interest rate risk and equity risk should apply. The correlation parameter then results from the decisive risk for the undertaking. Therefore the application of the two-sided correlation depends on whether a fall or rise in interest rates is the crucial factor. Correlation between interest rate risk and property risk The results of this analysis indicate that a correlation of 50% does not appear unreasonable. It could even be argued that the data at the 80 th and 85th percentile indicates that the correlations between property and interest rates should be closer to 75% than to 50% On the other hand, as in the case for the correlation between interest rate risk and equity risk, a distinction should be drawn between correlations between a fall in interest rates and a fall in property prices on the one hand, and between a rise in interest rates and a fall in property prices on the other hand. Whereas there is clear statistical evidence of a positive correlation between the first, this is less strong in the case of a correlation between rising interest rates and falling property prices, where in some instances even a negative correlation can be observed In light of these conclusions, a two-sided correlation between interest rate risk and property risk in the standard formula is introduced: In case the insurer is exposed to a fall in interest rate risk, a correlation parameter of 50% between interest rate risk and property risk should be applied to aggregate the respective capital charges; In case the insurer is exposed to a rise in interest rate risk, a correlation parameter of 0% between interest rate risk and property risk should apply. The correlation parameter then results from the decisive risk for the undertaking. Therefore the application of the two-sided correlation depends on whether a fall or rise in interest rates is the crucial factor. Correlation between interest rate risk and spread risk The results of this analysis indicate a correlation of 50% does not appear unreasonable, especially in view of an increased dependence in the tail of the distributions As was the case for the correlation between interest rate risk and equity risk or property risk, the analysis indicates that there is stronger support for a positive correlation in case of falling interest rates than in the case of rising interest rates. Page 10 of 21

11 In light of these conclusions, a two-sided correlation between interest rate risk and spread risk in the standard formula is introduced: In case the insurer is exposed to a fall in interest rate risk, a correlation parameter of 50% between interest rate risk and spread risk should be applied to aggregate the respective capital charges; In case the insurer is exposed to a rise in interest rate risk, a correlation parameter of 0% between interest rate risk and spread risk should apply. The correlation parameter then results from the decisive risk for the undertaking. Therefore the application of the two-sided correlation depends on whether a fall or rise in interest rates is the crucial factor. Correlation between equity risk and spread risk In the analysis it was observed that year on year changes to credit spreads tend to be relatively stable, except for a few events (two in the last 12 years), where they jump rapidly. It seems plausible that such a jump would be seen in a general 1:200 year event (such as 2008) Hence in the analysis of empirical correlations between equity and spreads were assessed at higher percentiles, and on condition of extreme movements in credit spreads. The results of this analysis indicate that empirical correlation between equity risk and property risk rises rapidly in the tail Given this tendency for very high correlations during periods of market stress, we can conclude that a correlation factor of 75% is reasonable. Correlation between property risk and spread risk In the analysis similar anomalies between spread risk and property risk were observed as between spread risk and equity risk. Correlations between spread and property approach 50% in the 95th percentile Given these results, CEIOPS proposes to apply a correlation factor of 50% (rather than as 75% as suggested in the draft advice) between property risk and spread risk. Correlation between equity risk and property risk The statistical analysis indicates that the correlation of 75% as proposed in the draft advice would seem justified. Correlation between currency risk and other risk types If these drastic changes in key market parameters take place it is likely that not all markets are affected in the same way and that currency exchange rates between the markets become volatile. On the other hand, strong movements in the exchange rates of main currencies can cause or reinforce the movements of other market parameters. These connections can be observed in the 1973 dollar crisis, the 1997 Asian crisis or the current financial crisis. Therefore, high correlation factors between currency risks and the other market risks can be adequate. Onthe other hand, currency risk is a twosided risk. Depending on the currency mismatch, a fall in a currency exchange rate can cause a loss or a profit in the balance sheet of an undertaking. Taking this nature of currency risk into account, a medium correlation factor seems to be justified. Correlation between concentration risk and other risk types The correlation factors of concentration risk in relation to equity risk, spread risk and property risk depend on the definition of concentration risk. The concentration risk sub-module covers the additional loss (compared to a well-diversified portfolio) that the undertaking may incur if concentrations in the equity, bond or property portfolio in respect to a single counterparty exist. Therefore, because of the definition of the concentration risk sub-module, the correlation factors Page 11 of 21

12 should properly describe the dependence between the risk of concentrations with respect to counterparty exposure, and the equity, spread and property risk. The correlation factors of concentration risk in relation to equity, spread and property risk should allow for diversification between property and equity/spread risk. For example, there is diversification between equity risk and property concentration risk or between property risk and the risk of concentration in names. Hence the correlations factors of concentration risk in relation to equity risk, spread risk and property risk should be The correlation factors of concentration risk in relation to the other risks,namely interest rate risk and currency risk, should be set in a consistent manner, reflecting the dependence of these risks and the triple consisting of equity risk, spread risk and property risk Based on the analysis above, the correlation factors for market risk should be chosen as follows: It should be noted that in many cases (and nearly all cases between the four largest risks of interest, equity, property, and spread), the CEIOPSanalysis gives a proposal consistent with the recent paper published bythe CRO forum. 5.3 Level 2 Implementing Measures (Draft 31 Oct 2011) In the newest level 2 implementing measures: The currency correlations with other risk factors as above were changed to 0.25 except for correlation between currency and concentration risk which was set at 0; and the concentration correlations were changed to 0 vs the risk factors as above; and Correlations for the counter cyclical premium versus other market risk sub-risks were set at 0; Giving the following correlation matrix: Page 12 of 21

13 The task group is of the opinion that correlation of 0 between concentration risk and other market risk elements is appropriate since concentration risk is targeted at nonsystematic risk in contrast to other market risk events. The counter-cyclical premium is included for Solvency 2 where this forms part of the technical provisions. Since this is not relevant to SAM, it will not be considered further and is to be removed from the recommended SAM correlation matrix. There is the potential to include a two-sided currency correlation (if there is an equity fall it is probably more likely that there will be a currency depreciation than an appreciation). The task group is of the opinion that very few (re)insurers will be exposed to a depreciation in the South African rand. As such it is expected that in the majority of cases the correlation between an appreciation in the South African rand and a fall in equity markets would be used. A low correlation of 0 is considered appropriate in this instance. 5.4 Other relevant jurisdictions (e.g. OSFI, APRA) It seems that the regulations under development for Canada would not make any allowance for diversification or concentration between risks until more is known about their behaviour in extreme circumstances. For the Australian regulations under development it seems that Equity, Property, fixed interest and index linked interest risk are assumed to be independent in determining the resilience reserve component of the capital requirement. 5.5 Mapping of differences between above approaches (Level 2 and 3) There is no contradiction between EIOPA Level 2 and 3 and the IAIS guidance. The solvency 2 approach seems to be the most realistic and comprehensive approach and is preferred. Page 13 of 21

14 6. ASSESSMENT OF AVAILABLE APPROACHES GIVEN THE SOUTH AFRICAN CONTEXT 6.1 Discussion of inherent advantages and disadvantages of each approach Discussion Document 48 recommends that the correlation matrix approach be adopted to define the dependency structure for all risk modules and sub-modules of the standard SCR model. Additionally, where causal links exist between policyholder behaviour and the risk in question, these interrelationships should be reflected within the specific module or sub-module where it arises. For the Market Risk module, this approach is deemed appropriate given the structure and aggregation methodology used for the rest of the Standard Formula. Regarding the structure (or sub-module components) of the Market Risk module itself, the following is noted: It is proposed to allow for the separate SA equity category within equity risk as per position Paper 47 It is proposed to allow for volatility risk as per Discussion Document 106. When a look-through and mandate basis of stressing assets within collective scheme is not possible, the third option allows the scheme to be treated as "other equity". It is possible that in some circumstances, a higher capital requirement would have been achieved if it is instead allocated to SA listed equity. The arguments for not including a test to see which (of other or SA) would maximise SCR and then using this are as follows: it is not deemed appropriate for an instrument or scheme to switch between classifications from one valuation to the next, and it shows complicated or unknown instruments risks in one place. 6.2 Impact of the approaches on EU 3rd country equivalence It is not expected that any of the recommendations in this document would impact third country equivalence. 6.3 Comparison of the approaches with the prevailing legislative framework Life insurers: Currently, for life insurers, CAR is calculated for market risks using two stressed scenario re-calculations in which a number of market risks are assumed to occur simultaneously instead of a separate calculation and aggregation, except for credit risk which is calculated separately and aggregated with market risk using some assumptions on the correlations as per SAP104. The difference between the two scenarios is that in one, interest rates are stressed upwards and in the other Page 14 of 21

15 downwards, and the maximum of the two should be chosen. The confidence level and time period used for the calibration of CAR differs from that to be used for SAM. The CAR methodology has some advantages in that all compounding effects between different market risks can be allowed for, for example an equity market fall coupled with lower interest rates can be more severe as a result of potentially having lower management action available in future. It also circumvents the potential of double counting management actions. However, there are still quite a few assumptions necessary to calibrate such a scenario in the first place. Furthermore, no market risks are measured within CAR on funds exceeding the total liability plus CAR. General insurers (Board notice 169): Currently, stresses are to be applied to the value of total admissibleassets in each asset category as follows and summed together: This seems to be a proxy for a scenario stress similar to that used in CAR by Life Insurers. Given that the Solvency 2 methodology is being adopted and its calibration papers with respect to Market Risk are widely available, the Solvency 2/QIS 2 methodology is preferred (also refer to Discussion Document 48). 6.4 QIS5 and SA QIS1 reports (regarding structure and correlation of Market risks) There were some concerns with the treatment of participations in the SA QIS 1 report: There were some concerns over the treatment of participations in the equity risk capital calculation. In particular, there was a view that participations should be treated the same, whether they were seen as strategic or not. There was also a concern that a 22% stress on participations was not enough, and that participations should be stressed in line with other equity holdings. There is a separate working group looking at the treatment of participations going forward. Refer to discussion document 53. For QIS5, the diversification benefit within the market risk module for solo insurers was 36% of the diversified Market risk capital, while for SA QIS1 the diversification benefit was approximately 23%. This seems broadly consistent taking into account the relatively larger equity risk Page 15 of 21

16 component in South Africa. Below the three largest risks within Market risk (for QIS 5) (as % of diversified market risk SCR) is compared to the SA QIS1 situation (approximately) for solo insurers: QIS 5 SA QIS 1 Equity Risk 42% 77% Interest Rate Risk 28% 20% Spread Risk 30% 10% In SA QIS1, concentration risk was in fact a bit larger than spread risk. It is clear that market risk was dominated proportionally more by equity for South Africa and hence led to a lower diversification benefit within market risk. This was exacerbated by market risk sub-modules reflecting a situation gross of management action in QIS 5 and SA QIS1. Feedback from SA QIS 1 also included questions on more guidance on the treatment of certain instruments such as preference shares. Within QIS 3 this guidance was added, specifying that where instruments exhibit the traits of interest rate risky assets, they should be treated using the interest- and spread risk modules, otherwise as an equity. In general, characteristics of instruments need to be examined and it should be treated using "substance over form". This has also been added to the recommendation section of this document. 6.5 Adjustments to the Market Risk module for SA QIS2 Following feedback from SA QIS1 and other industry comments on the SAM forums, adjustments relating to the following were made for SA QIS2: A separate module for strategic participations was created, removing them from the equity risk sub-module This is currently under discussion in a separate working group, with some additional information discussed in section 6.4 above. 6.6 Correlation parameters In SA QIS 2 the same correlation parameters as in the latest level 2 draft implementing measures (October 2011) from Europe were used, except that an illiquidity premium risk module was retained from QIS 5/SA QIS1 and applied to the matching premium (in contrast to Solvency 2), and no counter-cyclical premium being applied for SA QIS2 meant that this sub-module was not necessary within the SCR (i.e. an illiquidity premium risk module replaced the counter-cyclical premium module within the draft level 2 implementing measures). The correlations of the illiquidity premium with other risks was assumed to be 0, except for the correlation with spread risk which was assumed to be The correlation matrix therefore looked as follows: Page 16 of 21

17 It is also worth noting that SA QIS2 required a net (of management action) calculation for each risk sub-module. 6.7 SA QIS2 feedback The SA QIS2 report did not highlight any specific industry comments or concerns around the correlation parameters used. For SA QIS2, the diversification benefit for solo life insurers was close to 40% of the diversified market risk capital and close to 35% for non-life solo insurers, compared to 36% for QIS5 and 23% for SA QIS1 considering all solo insurers. The removal of participations and net capital requirements seemed to have aligned the diversification benefits within the market risk module more closely with Europe. 6.8 SA QIS 3 feedback A question was raised as to the treatment of fixed assets. Property for own use should be stressed under the property risk sub-module and furniture and equipment do not attract an SCR stress. No changes to the recommendation is proposed Some work is under way to produced a mapping from the CIC table to the various market risk stresses. No changes to the recommendation in this document is proposed Many comments focussed on asset information difficult to obtain, difficulty in applying the look-through method etc. As several different methodologies are given in the event that the look-through method is not possible, and furthermore under the principle of proportionality simplifications/approximations where relevant can be applied, no changes to the recommendation are proposed. Some comments were made as to whether a two-sided correlation between currency and equity (0 if currency-down (appreciation) applies, 0.5 if currency-up (depreciation)) applies. From QIS 2 results it became apparent that around 20% of respondents to QIS 2 were exposed to the risk of the Rand depreciating. This is a significant proportion of insurers and hence the task group proposes a two-sided correlation in this instance. Page 17 of 21

18 6.9 Conclusions on preferred approach The correlation matrix aggregation approach recommended in Discussion Document 48 is considered suitable for the Market Risk module. This allows for flexibility in allowing for dependencies between risks, and is consistent with the way other risk modules and submodules within the Standard Formula are treated. The structure of the Market Risk module as tested under SA QIS2 seems more relevant and suited to South African circumstances than the Solvency II structure. Input is required from the credit risk working group as to the structure of risk mitigation impairments (reference to credit risk discussion document); from the technical provisions task group in order to assess the need for and nature of an illiquidity premium sub-module (reference to the relevant discussion document); and in progress (Discussion Document 106) with regard to inclusion of interest rate and equity implied volatility stresses A separate working group is also looking at the treatment of participations (Discussion document 53). The structure of and correlations within the spread risk/credit default risk module are outside the scope of this document. Position Paper 47 discusses the need for a separate SA equity category and correlations between the different categories. The correlation parameters tested under SA QIS2 seem reasonable, and whilst data is unavailable or scarce, the overall diversification effect does not seem out of line with European experience. The task group is of the opinion that correlation of 0 between concentration risk and other market risk sub-modules is appropriate since concentration risk is targeted at nonsystematic risk in contrast to other market risk events. The counter-cyclical premium is included for Solvency 2 where this forms part of the technical provisions. Since this is not relevant to SAM, it will not be considered further and is to be removed from the recommended SAM correlation matrix. There is the potential to include a two-sided currency correlation (if there is an equity fall it is probably more likely that there will be a currency depreciation than an appreciation). From QIS 2 results it became apparent that around 20% of respondents to QIS 2 were exposed to the risk of the Rand depreciating. This is a significant proportion of insurers and hence the task group proposes a two-sided correlation in this instance. A default (after a look-through and mandate based methodology of classifying assets in collective investment schemes) to Other Equity is preferred rather than a methodology allocating it to either SA equity or Other equity to maximise SCR because: it is not deemed appropriate for an instrument or scheme to switch between classifications from one valuation to the next, and it shows complicated or unknown instruments risks in one place. 7. RECOMMENDATION SCR Market Risk 7.1 Structure of the Market risk module Introduction Description Page 18 of 21

19 Market risk arises from the level or volatility of market prices of financial instruments. Exposure to market risk is measured by the impact of movements in the level of financial variables such as stock prices, interest rates, real estate prices and exchange rates. Undertakings should calculate the capital requirement for market risk separately. If an investment is subject to additional funding calls in the event of losses being incurred, these should be taken into account in the market risk calculations. The calculations of capital requirements in the market risk module are based on specified scenarios. In these scenarios, allowance should be made where there are causal links between policyholder behaviour and the risk factor concerned. Furthermore, appropriate allowance should be made for the credit default risk on risk mitigating contracts. Look-through approach In order to properly assess the market risk inherent in collective investment funds, it will be necessary to examine their economic substance. Wherever possible, this should be achieved by applying a lookthrough approach in order to assess the risks applying to the assets underlying the investment vehicle. Each of the underlying assets should then be subjected to the relevant sub-modules. The same look-through approach should also be applied for other indirect exposures, such as investments in entities functioning primarily as holding entities for underlying assets, except for participations in related undertakings. The look-through approach should not be applied to investments in listed equity, tradable securities or other financial instruments based on repackaged loans. Where a number of iterations of the look-through approach is required (e.g. where an investment fund is invested in other investment funds), the number of iterations should be sufficient to ensure that all material market risk is captured. The above recommendations should be applied to both passively and actively managed funds. Where a collective investment scheme is not sufficiently transparent to allow a reasonable allocation of the investments, reference should be made to the investment mandate of the scheme. It should be assumed that the scheme invests in accordance with its mandate in such a manner as to produce the maximum overall capital requirement. For example, it should be assumed that the scheme invests assets in each credit quality step, starting at the lowest category permitted by the mandate, to the maximum extent. If a scheme may invest in a range of assets exposed to the risks assessed under this module, then it should be assumed that the proportion of assets in each exposure category is such that the overall capital requirement is maximised. If the management of the assets representing the employees benefits liabilities has been outsourced, but the insurance undertaking, acting as a sponsor, is liable for any loss of value of these assets, then the outsourcing arrangement should be looked-through for the calculation of the market risk capital charge. As a third choice to the look-through and mandate-based methods, undertakings should consider the collective investment scheme as an Other Equity stress. The characteristics of investment instruments need to be considered when deciding on which market risk charges apply. For instance, commodities could be considered as similar to equities and therefore treated as Other Equity, while many preference shares exhibit properties more akin to interest rate risky assets in which case they should be treated within the interest rate and spread risk modules. Both of the above instruments should also form part of the concentration risk calculation. They should also form part of the currency risk calculation where relevant. Where instruments exhibit both the traits of equity and interest-rate risky assets, they should be treated as Other Equity. Page 19 of 21