Read untitled text version

Solvency II Revealed

October 2011

Contents

4

An Optimal Insurer in a Post-Solvency II World Changing the Landscape of Insurance Asset Strategy Capital Relief Through Reinsurance Natural Catastrophe Capital Requirement Under Solvency II

27

Boosting Knowledge of Life Catastrophe Risk Rating Agencies and Solvency II Reinsurance Assets: Aggregate or Individual? Risk and Capital Modelling for Solvency II: A Pillar of Strength

10

32 35

16 21

41

Aon Benfield

Solvency II Revealed

The insurance industry has seen an extraordinary rise to prominence for the proposed Solvency II regulation which will bring fundamental reform of insurance supervision across Europe. As the deadline for implementation approaches, and the economic environment continues to present significant challenges, the key is in understanding how and where to prioritise resource to not only achieve compliance but also make the most of the business opportunities that Solvency II offers. Aon Benfield, in collaboration with its clients, has made enormous progress in understanding the practical implications of the new regulatory landscape. Solvency II Revealed explores new ways of thinking about the regulatory challenges and practically addresses these through a series of in-depth articles and case studies. Aon Benfield identifies where Solvency II will have most impact on the industry, with advice on how best to plan ahead for all roles involved in managing the regulation from CFO and CRO to actuaries and catastrophe modellers. Seven key themes are explored in the report: After Solvency II implementation, what will a capital-efficient insurer look like? Taking a long-term approach, the article predicts how an insurer would structure its business to maximise capital efficiency under the Solvency II rules, considering both the asset and liability sides of the balance sheet. The insurance industry faces significant challenges transitioning to an economic framework for investments. The article reveals how Solvency II will impact insurance asset strategy and identifies the key considerations for the CFO and CIO in repositioning their portfolio to achieve capital efficiency and sidestep possible dislocations in the financial markets. This case study examines the potential impact of a non-proportional retention protection reinsurance on the non-life solvency capital requirements (SCR) for a notional company under the Standard Formula. The study demonstrates how such a contract can substantially reduce capital requirements as an additional benefit of the reinsurance protection. Internal models, despite requiring a significant investment, result in more accurate reinsurance recoveries and, consequently, net capital requirements, than the Standard Formula. The article delves into the alternatives to calculate insurers' natural catastrophe SCR. Assessment of terrorism and pandemic risks has been taken to a new level through innovative developments in partial internal models. Case studies are used to illustrate the risk-mitigating effect of reinsurance. Not only are regulators focused on Solvency II, but to no surprise, so are the rating agencies. Reviewing recent feedback from rating agencies, the article helps reinsurers prepare for the potential impacts of rating agencies' calculations of capital requirements. Calculating the fair value of a reinsurer's share of technical non-life liabilities could be a challenging task if the reinsurance programme has changed in recent years. This article examines Solvency II's framework directive requirements and presents two different approaches from a practical perspective. a pillar of strength: Companies using internal models need to ensure they satisfy each pillar of Solvency II. The article highlights how an internal model can deliver tangible benefits when completing the Own Risk and Solvency Assessment (ORSA) and help companies prove to regulators that risk is being effectively managed. Aon Benfield is helping insurers prepare for all pillars under Solvency II by identifying cost effective means of improving capital efficiency, by assisting with modelling of asset and underwriting risks and by validating (partial) internal models and ESGs. The firm offers expertise on both sides of the balance sheet and is advising clients on designing optimal insurance and asset strategies under Solvency II. Our Solvency II capabilities comprise asset management, reinsurance and capital market solutions covering life, non-life and health insurance. As the industry rapidly approaches implementation, Solvency II Revealed aims to provide insurers with a fresh view of Solvency II and empower firms to achieve both regulatory compliance and a level of capital efficiency that exceeds investors' expectations.

3

Solvency II Revealed

An Optimal Insurer in a Post-Solvency II World

Uniting the management of insurance and asset risk provides a valuable opportunity for insurers to implement better management practices by viewing risk and capital holistically. This approach targets the overall balance sheet risk rather than insurance or investment risk as a silo. By leveraging the internal model framework, insurers can optimise business strategy across insurance and investment to improve both shareholder return and economic efficiency.

Solvency II is changing the way regulatory capital is assessed for European insurers. The results of the latest impact assessment study, QIS 5, suggest that the average solvency ratio for non-life European insurers will drop from over 200% to around 165% (overall ratio since non-life ratio is not available separately). Additionally, unlike the existing Solvency I regime, Solvency II uses a risk-based approach to set the level of each individual insurer's solvency capital -- thus requiring more capital to be held for riskier insurance and investment activities. This means that insurers that take a higher level of risk, as measured by Solvency II, will suffer a far greater fall in solvency ratio than those with less risky portfolios, whose solvency ratio may even improve. Despite presenting clear challenges, Solvency II also offers insurers the opportunity to improve their business strategy through better allocation of risk and capital to target opportunities that provide the highest rate of return per unit of risk. Solvency II encourages firms to view risk, capital and value from a top-down perspective, rather than from a silo based approach. Insurers must set strategy in accordance with two sets of constraints simultaneously: the capital constraints imposed by regulators, and the economic constraints imposed by stakeholders, including shareholders, policyholders and management. To maximise performance, insurers must pursue a combined strategy for both sides of the balance sheet -- a strategy that comprehends the potential dependence between insurance and asset risk behaviour. To date, very few organisations have optimised their allocation of risk and capital across both insurance and asset risk under a consistent measurement framework. In practice, the two sides of the balance sheet have been managed by separate business functions and strategies are set without a full understanding of the impact on the overall level risk and capital. For example, credit insurance losses are highly correlated to economic risks and setting asset strategy without consideration of insurance risks may result in a strategy that actually increases overall risk for the firm.

New approach for optimisation

Aon Benfield has developed an optimisation process for setting consistent strategy across asset and liability risks, recognising all relevant economic and capital constraints. This process will support insurers to better manage their risk and capital under Solvency II. In a post-Solvency II world, those insurers that can transform their business to maximise economic and capital efficiency will enjoy competitive advantages and improved shareholder returns. The process is outlined below, with sample exhibits for a hypothetical insurer "Multi-line Plc". Figure 1 illustrates the process for optimising the firm's overall business strategy across insurance and asset risk. The initial strategy of Multi-line Plc is derived from the insurance and asset strategy of the average non-life company in Europe. The risk assessment of the existing business strategy has been carried out using the Aon Benfield Insurance Risk Study assumptions for 2011 for underwriting volatility and correlations and Aon Hewitt Capital Market Assumptions 2011 for asset risk. Additional assumptions for reserve risk and underwriting performance have been assessed using industry data. The Aon Benfield business strategy optimisation process for Multi-line Plc is illustrated on the next page:

4

Aon Benfield

Figure 1: Aon Benfield Process for Insurance and Asset Strategy Optimisation

Insurance Asset Strategy Optimisation

1

Risk Risk Capital Capital

Articulate the firm's overall risk appetite, capital target and driver of shareholder value

2

Insurance Classes Insurance Classes Insurance Constraints Insurance Constraints

Identify optimal allocation of insurance risk under selected risk and capital measure

3

Asset Asset & Classes Classes & Constraints Constraints Insurance Insurance & Asset Risk & Asset Risk Model Model Optimise Optimise Asset Asset Strategy Strategy

Utilise remaining risk and capital budget to develop optimal investment strategy

risks characteristics

firm constraints

e.g. 150% Solvency II ratio model

current strategy

asset risk

Value Value earnings volatility combined

Optimise Optimise Insurance Insurance Strategy Strategy

to identify portfolios that budget and capital budget

Source: Aon Benfield

1

Risk Appetite, Capital Target and Drivers of Value Overall risk for the firm will be quantified as the volatility of surplus, i.e. assets less liabilities, from all sources of insurance and asset risks. An internal model of the full balance sheet will be utilised to measure surplus volatility and to consistently allocate risk between insurance and asset risks. Aon Benfield's price-to-book regression study points to a volatility measure of risk as best capturing investor risk tolerances. The binding capital metric for Multi-line Plc is the Solvency II capital requirement (SCR) under the Standard Formula. Capital utilisation for insurance and asset risks will be measured by the contribution to the overall SCR from non-life insurance and market risk. This assumes that shareholders are motivated by optimal exposure to insurance risks, careful management of balance sheet volatility and attractive returns on equity. Following consultation with Multi-line Plc's management, the overall risk appetite for the company has been articulated as 10.0% surplus volatility across both insurance and asset risk, hence maintaining its existing level of overall balance sheet risk. Its existing Solvency II ratio of 165% has been judged an appropriate long term position and the strategy review should maintain this level of capital adequacy.

2

Identify Optimal Allocation of Insurance Risk The firm wants to improve its insurance strategy to generate improved returns for shareholders and achieve better risk characteristics. Shareholders of non-life insurers typically seek firms that offer exposure to a carefully selected portfolio of insurance risks, with an asset strategy that supports their liabilities and enhances risk-adjusted return within the overall risk and capital budget. Therefore, when optimising the insurance and asset strategy of a non-life insurer, the first stage is to optimise the insurance portfolio. Once the optimal allocation of insurance risks has been selected, the remaining risk and capital budget can be deployed to enhance shareholder returns through the asset strategy. Multi-line Plc has defined upper and lower bounds for premium volume by class of business and agreed that the total premium volume can vary between 85% and 100% of the current level (100m): by optimising the risk allocation it may be possible to achieve higher profitability at a lower premium volume.

5

Solvency II Revealed

Table 1: Insurance Allocation Constraints

LOB Initial 33% 18% 4% 30% General Liability 12% 4%

Source: Aon Benfield

Allocation Min 28.0% 15.5% 2.5% 25.5% 8.5% 2.5% Max 38.0% 21.0% 4.5% 34.5% 14.5% 4.5%

An internal model of the insurer is created and, using Aon Benfield's proprietary optimisation framework, determines the economic and capital efficient frontiers of the insurance portfolios that provide the maximum expected profit for a specified level of economic volatility or Solvency II capital utilisation, respectively. The Solvency II capital efficient frontier differs from the economic frontier, and capital allocations that are efficient under the proposed Standard Formula can be suboptimal from an economic perspective. This is because the proposed Standard Formula assesses capital based on prescribed volatilities and correlations for non-catastrophe underwriting risk and prescribed events for natural and man-made catastrophes -- these prescribed factors are not based on economic best estimates and are often conservative. The goal of optimisation is to identify portfolios of insurance risk that are efficient from both the economic and capital perspective. The identification of jointly efficient portfolios is achieved by comparing the economic and capital efficient frontier and seeking portfolios that lie on the intersection of the two frontiers. Figure 2 plots the economic and capital efficient frontiers on a single graph in volatility-return space. The economic and capital efficient frontiers coincide in locations A and B, which provide jointly optimal allocations. In order to decide which mix of insurance risk is preferable, it is necessary to consider performance metrics at the two candidate portfolios.

Figure 2: Superimposition of Economic and Capital Efficient Frontiers

4.1 3.9 3.7 3.5

Economic Capital

B

Profit

3.3 3.1 2.9 2.7 2.5 8.0% 8.1% 8.2% 8.3% 8.4% 8.5% 8.6% 8.7% 8.8%

A

Initial Portfolio

Economic Volatility

Source: Aon Benfield

In Figure 3, the insurance portfolio composition is shown along the economic efficient frontier and two performance metrics: the economic Sharpe ratio and return on capital. In comparing options A and B, the highest ratio of profit to risk and capital is sought: this will achieve the optimal allocation of insurance risk. The selected portfolio is at the left most point of the intersection of the economic and capital efficient frontier in region B, where both the economic Sharpe ratio and return on capital is higher than at region A, leading to greater profit per unit of risk and capital. This portfolio provides the best combination of economic and capital efficiency.

6

Aon Benfield

Figure 2: Optimisation of Insurance Risk Under Economic Risk Measure

100% 90% 80% 70% 40% 35% 30%

Allocation

60% 50% 40% 30% 20% 10% 0% 8.04% 8.10%

A

B

25% 20% 15% 10% 5% 0%

8.20%

8.30%

8.35%

8.40%

8.45%

8.50%

8.55%

8.60%

8.65%

8.70%

8.75%

Economic Volatility

Credit and Suretyship Motor and Vehicle Liability Marine, Aviation and Transport Motor, Other Classes General Liability Sharpe Ratio Fire and Other Damage to Property Return on Capital

4.5 4.0 3.5 3.0

Optimal Portfolio

Initial Portfolio

Profit

2.5 2.0 1.5 1.0 0.5 0.0 8.0% 8.2% 8.3% 8.4% 8.5% 8.6% 8.7%

B

8.8%

8.9%

Economic Volatility

3

Optimisation of Asset Strategy Having selected the optimal insurance portfolio, the next stage is to investigate how the asset strategy can be improved within the remaining risk and capital budget for the firm. Overall, the level of insurance volatility has increased by 3 bps and the non-life Solvency II capital has increased by EUR0.1m. As the total budget for risk is 10.0% and the firm is targeting a 165% Solvency II ratio, this imposes two constraints on the asset portfolio: volatility must be such that overall surplus volatility does not exceed 10.0%. This will be computed as part of the optimisation as it is dependent on economic liability correlations be such that the overall SCR remains at the current level of EUR69.56m: through calculation this implies that the SCR_Mkt should be EUR29.86m In assessing the overall level of surplus volatility for the optimised insurance strategy alongside candidate asset portfolios, it is important to consider the impact that liability volatility has upon economic risk.

7

Solvency II Revealed

Figure 2b: Optimisation of Insurance Risk Under Economic Risk Measure

Portfolio Initial 1 2 3 4 5 6

8.62% Profit Statistics 3.3 56.8 10.2% 5.9% 33.2% 18.0% Allocation 3.7% 30.1% General Liability 11.5% 3.5% Total

Source: Aon Benfield

8.0% 2.6 54.1 22.3% 4.8% 28.0% 18.0% 2.5% 25.5% 8.5% 2.5% 85.0%

8.2% 3.2 55.0 30.6% 5.9% 28.0% 15.5% 4.5% 25.5% 8.5% 4.2% 97.3%

8.4% 3.4 55.7 31.1% 6.2% 28.0% 15.5% 4.5% 26.9% 10.5% 4.5% 89.9%

8.5% 3.6 56.3 31.1% 6.5% 28.0% 15.5% 4.5% 29.0% 12.1% 4.5% 93.6%

8.65% 3.8 56.86 31.0% 6.7% 28.0% 15.5% 4.5% 31.1% 13.7% 4.5% 97.3%

8.77% 4.0 57.3 31.0% 6.9% 28.0% 15.5% 4.5% 33.0% 14.5% 4.5% 100.0%

100.0%

In performing the asset strategy optimisation in the context of the overall insurance balance sheet, the following key characteristics are incorporated into the asset liability model: Interaction of liability uncertainty with economic risk: unavoidable market risk arises due to liability volatility interactions with interest rate risk. For example, if the liabilities increase by 50% then impact of interest movement will also increase by 50%. Economic liability correlation: some insurance risks, such as credit and surety, are highly correlated to the economy. Ignoring economic liability, correlation understates true level of overall risk, leading to incorrect allocations Using the asset liability model under our optimisation framework, a constrained efficient frontier of asset portfolios is determined for which the Solvency II market risk capital requirement does

not exceed EUR29.86m. The optimal asset portfolio for the company is then determined by the portfolio lying on the efficient frontier that achieves an overall surplus volatility of 10.0%. However, while this portfolio will provide optimal return characteristics within the risk and capital budget, it lacks a number of desirable qualitative features. The asset strategy is refined by overlaying qualitative constraints for insurance: liquidity purposes (e.g. cat events) years at each key rate duration generating assets of 10% This will help ensure that the asset portfolio is robust during economic downturns and has good asset liability characteristics for non-life insurance.

8

Aon Benfield

Figure 4: Optimal Asset Portfolio Composition

100% 90% 80% 2.6% 2.5% 2.4% 2.3% 2.2% 2.1% 2.0% 1.9%

Portfolio Initial 10.0% 29.86 2.12% Statistics 3.54% 30.00 32.68% 17.41% 0.96 2.50% 3.46% 29.86 43.06% 20.02% 0.95

Excess Return

Optimal 20% ±1 year 10% 10%

Allocation

70% 60% 50% 40% 30% 20% 10% 0%

Initial

Em Eq FoHF (Hedged) Real Estate A Credit 5 AA Credit 5 Gov Bonds 3 Cash Private Equity

Optimal

Equities A Credit 10 AA Credit 10 Gov Bonds 5

Gov Bonds 1 Excess Return

Source: Aon Benfield

The optimal asset strategy shown in Figure 4 has provided a 38 bps increase in return compared to the initial portfolio, while meeting quantitative constraints for risk and capital budgeting and insurance.

Conclusion

Multi-line Plc's economic and financial characteristics have been transformed under Aon Benfield's proprietary optimisation framework. Underwriting performance has been significantly enhanced by optimising premium volumes across each class of business, while applying realistic constraints to limit significant deviation from the initial underwriting strategy. The recommended insurance strategy was selected as the portfolio of insurance risks that: efficient frontier return on capital

This portfolio provides the greatest profit per unit of risk and capital among all possible allocations of insurance risk within the specified constraints. After allocating risk and capital to the optimal insurance strategy, the remaining risk and capital budget was allocated to an optimal asset strategy. The selected asset strategy fully utilises the remainder of the risk and capital budgets and provides optimal return while meeting bespoke qualitative constraints specific to insurance. As shown in Figure 5, the overall financial and economic impact of the business strategy optimisation is an increase to expected profit of EUR1.4m, an improvement of shareholder return from 13.3% to 14.5%. In addition, there has been no increase in volatility or capital requirement under Solvency II.

Figure 5: Overall Comparison of Initial and Optimal Business Strategy

Profit Initial Liabilities 3.81 Initial Assets 12.77 Initial 15.23 9.97% 16.59

Source: Aon Benfield

Vol. 8.62% 8.65% 3.43% 3.46%

SCR 64.37 64.46 30.00 29.86 69.56

Sharpe Ratio 24.45% 31.03% 32.17% 43.06% 29.15% 36.07%

17.00

ROE 2.91% 3.32% 10.36% 11.13% 13.27% 14.45%

14.00 8.00%

Optimised

Expected Profit

16.00

ROE of 14.45%

3.34

11.89

15.00

Initial

ROE of 13.27%

9.00%

10.00%

11.00%

12.00%

Volatility

9

Solvency II Revealed

Insurance Asset Strategy

Solvency II will change the investment behaviour of insurance companies. It introduces an economic balance sheet and capital charges for assets that reflect the degree of asset risk and asset liability matching. Under the Standard Formula, the calibration of some capital charges is inconsistent with an economic view of risk. It is important to understand what potential market dislocations could occur if a significant number of insurers choose to alter their investment strategy accordingly. Solvency II also encourages a holistic view of risk and capital across insurance and investment. Allocating risk and capital across underwriting and investment more dynamically provides an opportunity to deliver a more stable return to shareholders through the underwriting cycle.

Introduction

Solvency II is a major catalyst for insurance companies to revisit their asset strategy, driven by capital requirements that reflect the riskiness of each asset class and how well assets and liabilities are matched in a fair value accounting world. This is in contrast to the current state of play under Solvency I, where the same level of capital is required whether assets are held in cash or private equity and no consideration is given to the sensitivity of the firm's valuation to movements in economic variables such as interest rates and credit spreads. The capital charge under Solvency II for asset and economic risks is called the Market Risk Solvency Capital Requirement and represents the potential deterioration in the net asset value of the firm following a 1 in 200 year event over a one year time horizon across all asset and economic risks. This includes the potential loss in asset values and increase in liabilities due to changes in the discount rate. The market risk charge is decomposed into contributions from each underlying economic risk that drives changes in asset and liability valuation: this consists of a number of sub-modules that are described in Table 1. The capital charge for each sub-module is calibrated to the 1 in 200 year return period over a 1 year time horizon. The overall market risk charge is computed by aggregating together each sub-module using a prescribed correlation matrix to provide the 1 in 200 level of loss across all sources of market risk.

The transition into Solvency II raises a number of important questions for insurance companies and this article reveals how these can be addressed: market risk relative to underwriting and reserve risk? strategy under Solvency II to achieve capital efficiency whilst targeting attractive returns? markets and what steps could be taken to avoid potential market dislocations?

Capital Allocation at the Enterprise Level

Historically non-life insurance companies have tended to manage underwriting risk and investment risk as silo activities under the Chief Underwriting Officer and the Chief Investment Officer without joined up measurement of risk and capital for the purposes of setting strategy. Solvency II changes the way insurers think about risk, capital, volatility and value generation through unified risk management processes. Many companies are introducing the role of Chief Risk Officer who is responsible for managing the overall level of risk and capital utilisation in the organisation across both sides of the balance sheet. Having a more holistic view of risk and capital provides the non-life insurance industry an opportunity to achieve more consistent levels of return on equity

10

Aon Benfield

throughout the underwriting cycle by dynamically allocating capital between underwriting risk and investment risk. As illustrated in Figure 1, by continually monitoring and forecasting the pricing cycle, business plans can be adjusted to target business classes that provide maximum profit per unit of risk and capital. During soft markets, underwriting strategy can be more cautious and premium volumes reduced temporarily for less profitable lines. This will free-up risk

bearing capacity that can be redeployed to support higher yielding investment strategies until the market turns, at which point investments can be de-risked and a more aggressive underwriting strategy can be followed. From this perspective, the objective of investment strategy for non-life insurance should be to enhance the firm's return on equity within the risk and capital budget remaining after following the optimal underwriting strategy.

Table 1: Standard Formula: Overview of Market Risk Capital Changes

Risk Capital Change Implications

capital change

Property Spread income assets Duration Factor 0.9% 1.1% 1.4% 2.5% 4.5% 7.5% 3.0% Capital Charge by Duration 1 0.9% 1.1% 1.4% 2.5% 4.5% 7.5% 3.0% 3 2.7% 3.3% 4.2% 7.5% 13.5% 22.5% 9.0% 5 4.5% 5.5% 7.0% 12.5% 22.5% 37.5% 15.0% 10 9.0% 11.0% 14.0% 25.0% 45.0% 60.0% 30.0% bonds may increase Solvency II

Rating AAA AA A BBB BB B or lower Unrated

premium markets

concentration threshold assets for credit ratings A or above and 1.5% of total assets

Source: Aon Benfield

11

Solvency II Revealed

Figure 1: Dynamic Risk and Capital Allocation

Source: Aon Benfield

Capital Efficiency of Investment Strategies

The capital requirements for different asset classes under Solvency II vary considerably and are not always set in line with economic principles. This creates inconsistencies between optimal strategies as viewed from an economic risk measure and the Solvency II Standard Formula capital requirements. It is important therefore to develop a framework for setting investment strategy that can incorporate the management's own view of risk alongside the constraint of the Solvency II capital requirements: while achieving capital efficiency is important, it should not override the importance of careful risk management. Where significant disparities exist between the Standard Formula and economic principles, one option is to develop a partial internal model covering market risk or specific asset classes where greater risk granularity is desired. For example, the Standard Formula assigns a capital charge of 49%

to Other Equity which includes a wide range of alternative assets. In the case of risky flavours of private equity such as venture capital this is quite sensible but for a diversified fund of hedge funds, this would be overly cautious: hedge funds have historical levels of volatility significantly lower than listed equity. In setting investment strategy it is instructive to understand the relative capital efficiency of different asset classes. One approach for comparing the capital efficiency is to consider the return on capital achieved for each investment under current market conditions from a silo perspective (i.e. ignoring its contribution to diversification). This comparison can be helpful in identifying whether the existing strategy is overweight in less capital efficient assets. In Figure 2 key asset classes' return on capital under the Standard Formula is compared to the economic view replicated using an internal model view (based on current market conditions).

12

Aon Benfield

Figure 2: Comparison of Return on Capital Across Asset Classes

6.0%

5.0%

4.0%

ROC

3.0%

2.0%

1.0%

0.0%

sh

s5

s7

0

3

5

7

10

s1

10

3

5

7

s3

3

5

7

10

ty

s

s1

tie

ta

ui

it

it

it

it

it

it

it

it

it

Ca

nd

nd

nd

nd

ge ed

nd

ed

ed

ed

Bo

Bo

Bo

Bo

Cr

Cr

Cr

Cr

Cr

Cr

Cr

Cr

Cr

Eq

Bo

al

Cr

Cr

Cr

e

ov

ov

ov

ov

AA

AA

AA

Re

A

A

A

B

B

B

ov

AA

BB

BB

BB

iv

A

B

at

BB

G

Pr

Asset Class

ROC (Standard Forumla) ROC (Economic)

Source: Aon Benfield, Aon Hewitt

The Standard Formula significantly overstates the risk bearing capital required for longer duration credit and hedge funds. It is also noteworthy that the riskiest asset classes provide the highest return on capital, despite having relatively high capital charges. Return generating assets can still make an important contribution to return on equity despite the 49% capital charge -- the impact of the additional expected yield is greater than the marginal increase in capital relative to less risky asset classes. For non-life insurers an important consideration under Solvency II is the capital charges for interest rate risk. Currently most categories of non-life insurance liabilities are accounted for on an undiscounted basis. This means that the management and investors of

many non-life firms are focused on achieving positive investment return in their income statement, rather than considering the asset return achieved relative to the return of the liabilities. Solvency II is encouraging insurers to think about the economic balance sheet and has significant capital charges for interest rate duration mismatching. However, until IFRS 4 Phase 2 is implemented, the general accounting view will continue to be based on undiscounted liabilities. An important consideration will therefore be the trade-off between capital efficiency and managing earnings volatility on current accounting principles: will non-life insurance investors understand that negative investment returns do not necessarily represent an economic loss when assets and liabilities are matched?

Fo

HF

G

G

G

G

(H

Em

it

it

ed

ed

ed

ed

ed

ed

ed

ed

ed

it

ui

Eq

Es

Eq

d)

te

13

Solvency II Revealed

Impact of Solvency II on the Investment Market

An important question is whether the new regulatory framework could itself have an impact on the investment market through changing the investment behaviour of the insurance industry. We have already seen that in many cases there is a disconnection between the basis on which the capital charges have been set under the Standard Formula and economic reality. The insurance industry plays a significant role in institutional investment and is a major participant in European bond markets. Changes in investment behaviour attributable to Solvency II may originate from a number of sources including: 1. Matching the components of the liability discount rate to reduce balance sheet volatility 2. Capital constrained insurers who need to improve their Solvency II ratio 3. Insurers who target an investment strategy that maximises their return on equity under the Solvency II Standard Formula for market risk 1. The Liability Discount Rate Under the current proposals, liabilities are discounted using a rate derived from the risk-free rate plus an illiquidity premium. The risk-free rate is swap based with an adjustment for credit risk and the illiquidity premium is variable depending on the level of illiquidity implicit in the liabilities: for annuities in payment which cannot be altered this will normally be 100% of the illiquidity premium and other more liquid liabilities will have a low percentage applied. The illiquidity premium itself is based on the observed spread between a basket of corporate bonds (using the iBoxx index) and the swap rate that cannot be explained by credit risk. Some insurers will be motivated to invest the assets backing their technical provisions more closely to the liability discount rate under Solvency II as this will help to stabilise their Solvency II ratio.

2. Capital Constrained Insurers Under Solvency II, the capital constrained insurer's concern will be primarily to take steps to reduce the capital requirement. This means reducing exposure to return generating assets that attract the 39% and 49% charges and any other assets that have large capital charges. As illustrated in Table 1, long duration corporate bonds are capital intensive and for credit rating BBB or lower are in line with return generating assets. It is therefore likely that insurance companies will reduce their exposure to equities and longer duration bonds rated BBB or lower. In addition, there is an interesting secondary effect for life insurers. Currently, life companies will invest in long duration corporate bonds to match the duration of their liabilities (typically 10 ­ 12 years). This works well as the strategy provides a good yield that enables competitive annuity pricing and the liability discount rate is usually asset based so there is no additional balance sheet volatility. As noted previously, under Solvency II the liquidity premium component of the discount rate is based on a basket of corporate bonds, which supports investment in matching bonds. However, under QIS 5, the calculation of the spread risk stress test has been disconnected to the illiquidity premium stress so assets and liability valuations are stressed separately. While an implicit link between spread risk and illiquidity risk has been maintained through a negative 50% correlation in the aggregation calculation, this acts as a disincentive to match the spread duration of the liabilities. The more capital efficient strategy is to invest in shorter duration corporate bonds which have a lower spread duration, and hence capital charge, and to utilise an interest rate swap to increase the rate duration of the assets to that of the liabilities.

14

Aon Benfield

3. Maximising Return on Equity As discussed earlier, the Standard Formula is not consistent with an economic perspective which means that firms aiming to maximise return on equity may design an investment strategy that differs substantially to their current asset allocation. In particular, Figure 2 shows that long duration credit BBB or lower is less capital efficient and hedge funds also do not achieve a good return on equity under the Standard Formula. In general, the Standard Formula will encourage holding assets classes that provide maximum yield for the capital category they fall into: for example, within Other Equity the most capital efficient assets will be risky forms of private equity investments. While many insurers are not expected to focus purely on capital efficiency, it is likely to be a consideration that will tilt the average insurance asset allocation away from less capital efficient asset classes such as high yield debt that feature low quality credit exposures.

Conclusion

Solvency II is driving non-life insurers to think holistically about risk, capital, volatility and value generation across insurance and investment. We believe that bringing together the management of insurance and investment risk through the Chief Risk Officer provides a valuable opportunity for insurers to: risk and capital holistically across both insurance and asset risk underwriting and investment more dynamically throughout the underwriting cycle to provide a more stable return to shareholders There are many challenges for European insurers during the transitional period to Solvency II and beyond to the new international accounting standards. To achieve attractive returns on equity under the new capital regime for market risk, significant changes to investment strategy will be required to manage asset liability risk. Moving to an economic view of the balance sheet has significant implications for companies who report on an undiscounted basis and careful communication with senior management and investors is required to carefully manage this transition. Finally, in the current draft of the Standard Formula, there are many areas of economic disconnect that could have broader implications for the investment market. Until the Standard Formula is finalised, it is difficult to judge at what point insurers will start to switch their portfolios, but it is important to be aware of the potential market dislocations and consider how to position your firm's investment portfolio to minimise the impact of the new regulatory framework.

15

Solvency II Revealed

Capital Relief Through Reinsurance

Insurers do not necessarily have to choose between a reinsurance programme which makes business sense and one which reduces capital requirements -- even if the company appears thinly capitalised under QIS 5. Non-proportional reinsurance often provides the best solution for the business by removing frequency risk and tail risks at a cost that makes economic sense. Under the Standard Formula significant capital relief can be obtained for non-proportional reinsurance, making it an attractive solution from both a business and capital perspective.

Introduction

Standard Formula risk mitigation techniques under Solvency II have already become a hot topic for actuaries, CROs and CFOs. For most companies, regulatory capital requirements have historically played a relatively small role in the decision to purchase a particular reinsurance contract, where managing the volatility of shareholder returns and economic and rating agency capital requirements have typically had the upper hand. However, QIS 5 indicated that regulatory capital requirements under Solvency II are likely to increase significantly for most non-life insurers. Although unlikely to become the dominant factor in a reinsurance purchasing decision, the impact of the reinsurance on Solvency II capital should be explored. In the Solvency II framework, an insurer can choose to use the Standard Formula or its own internal model to estimate its Solvency Capital Requirement (SCR). The Standard Formula is a non entity-specific risk-based formula designed by EIOPA, the European insurance regulator. Alternatively, the undertaking can build an internal model and submit it for approval by the regulator to determine their SCR. For the vast majority of companies the investment required to develop and submit an internal model is too great and so a good understanding of how the Standard Formula recognises the risk mitigation effect of reinsurance is essential. This case study uses the QIS 5 version of the Standard Formula to estimate the impact of a specific reinsurance structure on a notional company's non-life

The non-life underwriting risk module comprises three

This case study considers the impact of the reinsurance structure on both Premium Risk and Cat Risk which must be calculated separately and then combined together using a prescribed correlation coefficient. Under the Standard Formula, the premium risk by class of business is calculated as the product of a prescribed underwriting volatility and the company's premium volume. For proportional reinsurance, such as quota share, the capital relief can be easily determined by multiplying by the ceded percentage. For nonproportional reinsurance, the capital saving effect is less immediately apparent under Standard Formula. However, as this case study will demonstrate, nonproportional reinsurance can offer significant capital savings without requiring a partial or full internal model to be developed.

Notional Insurer and Reinsurance Structure

The case study is based on a notional Swedish monoline company whose property portfolio has a premium volume of EUR130m and is protected by existing risk attachment of a 12 year return period. The focus of the study is the capital benefit of adding an aggregate protection to the existing retention. Since the company has existing risk and catastrophe large individual and catastrophe claims, the additional

16

Aon Benfield

structure they are interested in purchasing is a risk and catastrophe aggregate reinsurance to provide more sideways protection on their retention. With this new structure in place, the insurer is protected by the following reinsurance contracts:

Table 1 shows the combined effect of these reinsurance protections for an example set of large individual claims:

Table 1: Combined effect of reinsurance protections

EURm Gross Loss Net of Risk XL Loss below Retention Presented to Aggregate Recovery from Aggregate Overall Net

30

10

10

9

0

10

80

20

10

9

1.5

18.5

40

10

10

9

9

1

5

5

5

4

4

1

Source: Aon Benfield

Undertaking Specific Parameters

Under QIS 5, companies have the choice of two methods to estimate the impact of non-proportional reinsurance on their non-life insurance risk (on top of the effect of the reduced premium volume) -- both involve customisation of the volatility factor. The first approach is the Non-Proportional Adjustment Factor for reinsurance. Most insurers failed to apply this adjustment in their QIS5 submissions for a number of reasons, including: (1) it can only be applied for standard limits or deductibles are excluded, (2) the assumption made by the adjustment calculation that individual large loss severities follow a lognormal distribution may be of questionable appropriateness giving results which are hard to believe.

The second method is known as Undertaking Specific Parameters (USPs). Non-life premium risk USPs allow insurers to determine the volatilities to apply in the premium risk calculation using their own historical losses and one of three prescribed methods. The final premium risk USPs are weighted averages of the insurer's calculation and the Standard Formula where the credibility weights depend primarily on the number of years of available data and the line of business. For example, for property (fire) the weighting is 100% for 10 years or more of data. To apply one of the USP methods, the historical losses are first adjusted for elements such as inflation and then the annual losses on an as-if basis. After all of these adjustments have been made, the volatility to use for the premium risk calculation is derived. The appeal of this method is that the impact of any reinsurance structure can be taken into account. Also, in comparison with approval by the supervisors.

17

Solvency II Revealed

Historical Loss Data

A credibility mechanism should be used when applying USPs. The credibility factors to be applied should be chosen according to the length of historical loss data. In this case study of USP on the Premium Risk, USP Method 3 will be applied to 10 years of historical loss data from the notional Swedish company.

Figure 1: Effect of Reinsurance

5.5 5.0 4.5 4.0

Premium Risk Results

By applying the reinsurance programme to the 10 years of historical data, the USP method can estimate both the gross and net volatilities as a percentage of gross and net premiums respectively. Figure 1 shows the average large loss for each year before and after the reinsurance programme.

Millions

3.5 3.0 2.5 2.0 1.5 1.0 2001

Gross

B

2002

Net of XL

2003

2004

2005

2006

2007

2008

2009

2010

Net of XL & AGG

Source: Aon Benfield

The volatility of the losses decreases significantly after the reinsurance programmes are applied. Table 2 shows the premium volatility and premium SCR charges obtained from applying the USP method to the loss

Table 2: Premium Volatility and SCR Charges Using USP Method

Net of XL USP 9.1% Net of XL & Agg 8.5%

32

Source: Aon Benfield

29

Both of these volatilities are lower than the 10% prescribed for property (fire) premium risk under QIS 5. Since, in this case, there is a property line with 10 years of data, a credibility weighting of 100% can be applied to the insurer's volatility calculation. The premium SCR decreases by approximately 9.5% due to the Aggregate Protection.

18

Aon Benfield

Cat Risk Model

Property catastrophe exposure in Sweden is purely Natural Catastrophe of which 100% is windstorm. In this case study the Catastrophe SCR is estimated based on real company data using Cat Method 1 of QIS 5 for this property exposure in Sweden. Using the Crestazone gross exposure data, the QIS 5 formula determines the 1 in 200 year event loss for the peril. To arrive at the 1 in 200 total peril loss (the CAT SCR) (Table 5), the Standard Formula requires two alternative hypothetical years to be created (Table 3). This is first done on a gross basis and each is then netted down for reinsurance, after which the maximum net annual total of the two is taken (Table 4). For the windstorm peril, the two hypothetical years are: 80% 20% of the 1 in 200 year event.

Table 3: Two Hypothetical Years for Cat Scenario

Cat Scenario (EURm) 1 72 2 89

36

18

Table 4: The Cat Reinsurance

Reinsurance (EURm) Attachment XL Agg

15

16.5

Limit

75

20

Table 5: The SCR Cat Results

Before Agg 30 After Agg 18.5

Non-Life SCR Result

Figure 2 shows the result after the aggregation of the Cat risk and Premium risk by using the QIS 5 correlation matrix.

Figure 2: Aggregation of Cat and Premium Risks

50 40 30 20

49 38 30 19 32 29

Millions

10 0 -10 -20

-9 CAT SCR

Before AGG After AGG

-13 Diversification Non Life SCR

Premium SCR

Source: Aon Benfield

19

Solvency II Revealed

Since the aggregate contract protects both premium risk and cat risk, some thought should be given to how the aggregate deductibles and limits are shared between the two risk categories. By applying the aggregate conditions separately to the premium risk calculation and cat risk calculation, as has been done in this study, a conservative assumption has been made due to the aggregate deductible being imposed twice. Conversely, allowing for the full aggregate limit in both calculations is an overly-generous assumption and therefore, a further condition is imposed: the total reduction in SCR contract, in this case EUR20m. A reduction of only no such cap is required. Therefore the aggregate reinsurance programme decreases the capital charge for Cat Risk by 38% and Premium Risk charge by 9% under the USP method. After diversification, the total Non-life SCR decreases by 22% from EUR49m to EUR38m.

Conclusion

This case study clearly demonstrates that, even for a reinsurance contract that is structured primarily to achieve very specific business benefits -- such as the sideways retention protection -- the risk-mitigation significant even under the Standard Formula where it may not be immediately apparent at first glance. As for proportional reinsurance, non-proportional reinsurance programmes with tailored characteristics can also significantly reduce the Solvency II Non-life underwriting SCR. This is due to a combination of reduction effects including the ability to fully recognise reinsurance in the calculation of non-life catastrophe risk, as well as the ability to capture the actual volatility of the company's net premium risk using the USP method. For companies for whom Solvency II capital is a key constraint, the reinsurance programme could be designed from more of a capital management perspective. This first requires a company to define its risk appetite for insurance underwriting risk, upon which an optimal reinsurance programme can be structured to help the company to meet these objectives, whilst achieving other desirable outcomes such as reducing ceded profit and retained volatility. Reinsurance has always been a valuable risk-mitigation instrument. Different reinsurance programmes provide different business and capital benefits, and this case study demonstrates that the two can go hand in hand under Solvency II without necessarily using an internal model.

20

Aon Benfield

Natural Catastrophe Capital Requirement Under Solvency II

A magnitude of difference can exist between the Standard Formula and an internal model to calculate solvency capital for catastrophe risk. The discrepancies are revealed in the case study which stresses the importance of making a strategic management leading Cat risk mitigation tool and proves to be a cost effective source of capital, which is now recognised by Solvency II.

Catastrophe risk is a key driver for capital under Solvency II, with the benchmark to withstand a 1-in-200 year event for natural and man-made disasters. There is a basic calculation method that insurers can use to determine their Solvency Capital Requirement. However the methodology for the standardised scenarios for natural catastrophe modelling overlooks key data features. As such, natural catastrophe (Nat Cat) calculations are ignoring 15 years of critical evolution under the currently proposed Solvency II Standard Formula, which could lead to higher capital requirements for insurers when the regulation comes into force. Insurers need to choose between the Standard Formula and a partial internal model to assign a more appropriate capital charge. The article reveals the different outcomes through a detailed case study.

Figure 1: Process Options to Calculate Nat Cat SCR

NatCat SCR

Standard Formula Standardised scenarios are defined per European country and peril. The Standard Formula approach is designed to be applicable to the majority of companies and will be a practical solution for smaller companies as internal models can be costly and require a complex regulatory approval process. Standard Formula parameters, such as damage factors and correlations, as well as peril selection depending on a country hazard profile, are based on the Catastrophe Task Force (CTF) guidance. The CTF is a working group which includes regulators, (re)insurance industry participants and catastrophe modelling agencies. The Nat Cat Standard Formula approach is currently under review after some criticism following the QIS 5 industry exercise. The factor-based method is used where standardised scenario is unavailable or non-applicable, including:

Standard Formula

(Partial) Internal Model

Method 1: Standardised Scenarios

Source: Aon Benfield

Internal model

Method 2: Factor based Use of Cat Models

Internal models based on catastrophe modelling software output better reflect the risk profile of a company, which is particularly critical in producing results that reflect the company's potential exposure to Nat Cat risk.

21

Solvency II Revealed

Table 1 outlines the differences between the data requirements and therefore data quality impacting risk sensitivity of the possible approaches used in QIS 5 for the Nat Cat SCR.

Table 1: Impact of the Different Data Requirements for Nat Cat SCR

Parameters and metrics Standardised scenario Factor based Internal model (cat model)

Perils Subsidence

Gross Written Premium

Geographic resolution

All levels of resolution

Property coverage Interruption

Line of business split

other secondary characteristics

Loss scenario

Single event

and Subsidence

Loss calibration

Source: Aon Benfield

Due to significant differences in data granularity between Standard Formula and partial internal model, the output SCR will inevitably differ, with the former yielding a higher SCR in the majority of cases. Therefore companies will base their choice of method on the approach which provides the more accurate representation of their risk in their view. Significantly, an internal model offers several risk management applications in addition to the calculation of a Solvency II SCR, and provides the opportunity to fully recognise the benefit of complex mitigation structures.

22

Aon Benfield

Natural Catastrophe Reinsurance Under Solvency II

Both proportional and non-proportional reinsurance are reasonably taken into account as mitigation methods under QIS 5. The QIS 5 technical specifications do not prescribe a specific method to apply reinsurance to the proposed Nat Cat scenarios, because a single method is unlikely to be appropriate for all reinsurance programmes. Instead companies are asked to apply their reinsurance programme using an appropriate methodology which is then explained to the regulator. Since most internal models fully capture the details of reinsurance programmes, they should provide a more accurate picture of the company's net position at a 1 in 200 year level. The Standard Formula suggests that companies assume two events for windstorm, flood and hail, which are composed in such a way to test the adequacy of reinsurance protection: a combination of a large and small event for vertical cover, and two smaller events for horizontal protection. This allows the Standard programmes with a reinstatement by providing capital relief from the second limit. Internal models will of course recognise all reinstatements.

Case study

We examine below the differences between Standard Formula and internal model for a UK company writing property business. The reinsurance protection in place (proportional 1) Surplus share one with occurrence limit GBP80m

@ 100%

23

Solvency II Revealed

All types of reinsurance in this case are adequately recognised by both the Standard Formula and internal model. The steps for calculating Nat Cat SCR are described in Table 2:

Table 2: Process for Applying Reinsurance

Steps Standardised scenario Internal model

combined

loss per peril

Loss for all perils correlated

Source: Aon Benfield

24

Aon Benfield

Table 3 compares Nat Cat SCR using the Standard Formula and internal model for the case study example.

Table 3: Nat Cat SCR Comparisons Between the Standard Formula and Internal Model

GBPm Standard formula Internal model

Wind + Surge

Wind + Surge

340.90 409.08 500.65

185.41 203.95

243.40 261.45 304.78

153.03 164.51

SS 1 SS 2

107.02 58.11 7.27 7.10 148.37 327.87

76.74 32.17 3.37 2.75 54.09 169.12

75.41 35.60 4.09 3.40 119.70 328.59

54.86 29.74 3.48 2.30

81.21 96.04 adjustment **

Source: Aon Benfield

34.83

57.04 57.04 0.00

+14.58

** In the Standard Formula, there is a risk of double

being separate per peril and then correlating the net results. Correlation mitigates double counting to some reinsurance recoveries, especially on lower layers in the case of exposure to multiple perils. In our case the mitigation effect of the bottom layer exceeded the total limit available by GBP14.58m after combining perils with the effect of correlation.

This case study illustrates the magnitude of difference that can exist between solvency capital calculated using the Standard Formula and an internal model. If, as in this case study, the company chooses the Standard Formula route, it will need to maintain 94% more Nat Cat capital to meet the SII requirement than if it uses a partial internal model. This difference originates from the gross calculation. This might hint at a calibration issue with the Standard Formula. However, the Standard Formula gives higher capital relief due to the two-event scenarios, and allows recoveries from reinstatements.

25

Solvency II Revealed

Figure 5: Comparison of Nat Cat SCR Using Standard Formula and Internal Model

500 400 300 200 100 0 -100 -200

reinsurance is assessed using CatMetrica, a template developed in ReMetrica. In addition to capital relief, CatMetrica provides measures of efficiency of reinsurance such as the Ceded RoE (calculated in this example as the ratio of reinsurance margin to capital relief before diversification). In Table 4, CatMetrica capital at a Ceded RoE of 4.94% under an internal model. Under the Standard Formula, the capital relief for this programme is 60% at a ceded RoE of 3.59%.

Gross SCR

Standard formula Internal model

Net SCR

Difference

Conclusion

As demonstrated, reinsurance is a competitively priced source of capital, which is expected to have full recognition under the Solvency II regime.

Source: Aon Benfield

Table 4: CatMetrica Reinsurance Evaluation

Premium Expected Recovery Total Reinsurer Reinsurer St Dev of Margin to Margin Recoveries Std. Dev. Premium Multiple Probability of Attach/Exhaust 1st Limit 2nd Limit Economic Capital at 200 yr Capital Benefit Ceded ROE SII Standard Formula (QIS 5) Capital Benefit Ceded ROE

0.6 2.3

[email protected]%, 100% placed 3.33%

11.7 1.8

2.51%

0.49% 30.53% 4.08

1.37% 0.49% 4.79% 1.39% 15.90% 4.93% 0.01% 0.00% 0.07% 0.01% 0.83% 0.13%

0.6

0.82%

5.8

46.6

3.78%

67.7

2.60%

3.4

[email protected]%, 100% placed 6.77% 4.3 17.39%

1.3

2.65% 2.3 9.01%

2.1

4.12% 2.1 8.37%

7.3

28.36%

2.55

41.7

4.93%

50.9

4.04%

[email protected]%, 100% placed

6.7 8.8 28.3

31.31%

1.93

31.4

6.67%

46.0

4.55%

10.7 15.4 10.1 4.2 11.2 5.9

176.7 35.15% 2.42

15.90% 0.49% 0.83% 0.00%

275.2 4.94% 164.6 110.6 3.59%

16.8 15.2

119.7 57.0

Source: Aon Benfield *Expected reinsurance recoveries are based on catastrophe model output.

26

Aon Benfield

Boosting Knowledge of

Pandemic and terrorism risks incur the largest catastrophe capital charges under requirement of Solvency II, catastrophe models are evolving. Impact Forecasting's UK terrorism model incorporates input from counter terrorism experts on elements such as credible attack types and damage profiles. In addition, the average cost of a catastrophe mortality events other than pandemic.

Introduction

hold capital for the impact of 1 in 200 year extreme mortality events. To grasp what this definition means, in essence, this translates to not only considering one specific scenario with the probability of 0.5%, but thinking of holding enough capital to withstand the cost of extreme events at the 99.5th percentile and therefore consider multiple tail events. Such events include, but are not limited to, pandemics, terrorist attacks and natural catastrophes. The extent to which a company is exposed to catastrophe risk depends on various features of the underlying portfolio such as demographic profile, geographical location and product types.

Pandemic risk

Pandemic risk has many different definitions but the common feature is the spread of infectious disease across a large geographic region. The three major pandemic influenzas in the last 100 years, namely the Spanish caused over 50 million of deaths in total. The latest pandemic recognised by the WHO was the 2009 H1N1 which, although generating widespread public concern, was not as lethal as feared. In fact, with a reported death toll of less than 20,000 cases, it fell far short of WHO's expectation of 250,000-500,000 annual deaths arising from seasonal influenza. The impact of Spanish flu, if the history was to repeat itself on today's insured portfolio, has the potential to threaten the survival of many otherwise well capitalised insurers. Many may argue that an exact repeat of a Spanish flu is unlikely due to lessons learnt from the past such as the use of quarantine to contain the spread of

followed by a much smaller share of terrorism risk. Impact Forecasting, Aon Benfield's model development centre of excellence, has created models for pandemic, however the article focuses on pandemic and terrorism risks as these incur the largest catastrophe capital charge of a combined 90%. In the QIS5 technical specification, the Standard Formula additional deaths at 1.5 per thousand lives insured. A case study will demonstrate how an internal model can the standard formula.

facilitating rapid and effective communication of disease better access to medical assistance. On the other hand, some may argue the prevalence of international travelling, urbanisation and increased population density in city centres provide a favourable environment for the spread of infectious disease. In addition to the changing landscape, we are constantly under the threat of emerging pandemics with mutation of pathogens or resurgence of past virus strains. The outbreak of E-coli in Germany in 2011 had the world situation was quickly contained. It is not a matter of if, but when, the next major pandemic will strike.

27

Solvency II Revealed

Figure 1: SIR Model

Susceptible

Latent

Infectious Asymptomatic

1000

Hospitalized

Dead

Recovered

Source: Aon Benfield

In order to understand the impact of pandemic risk on insurers' portfolios, Aon Benfield has developed an influenza pandemic model based on the SusceptibleInfectious-Recovered (SIR) methodology, commonly used by epidemiologists to model transmission of infectious diseases. The model accounts for the transition of healthy population (Susceptible) coming into contact with the disease, developing symptoms, requiring hospitalisation, recovering from illness or dying. In turn this enables monitoring of: the cumulative number of people becoming ill, admitted with a given insured portfolio. As population density transmission rate is therefore adjusted. For each simulation, three different mortality profiles (W-shaped in Figure 2, U-shaped in Figure 3 and uniform distribution) are considered, hence producing three different estimates.

Figure 2: W-shaped Excess Mortality

3000

1913-1917 1918

Figure 3: U-shaped Excess Mortality

1952-1956 1957

Flu Mortality (per 100,000 population)

800

600

400

200

0 <1 1-4 5-14 15-24 25-34 35-44 45-54 55-64 65-74 75-84 >84

Age (years)

A probabilistic approach allows the financial impact to be analysed by different return periods to form the basis of reinsurance optimisation strategy. Standard catastrophe cover is designed to remediate the tail risk of portfolio but does not provide appropriate protection against pandemic risk. This is mostly because the "hours clause" attached to a catastrophe hours. However, based on what has been observed in the past, the full force of a pandemic can last for months if not years. For this reason, the industry has developed a range of specialist solutions including traditional stop loss cover, pandemic reinsurance and pandemic bond in the capital market. At the height of H1N1 pandemic, some reinsurers were quoting pandemic cover in excess of 10% rate on line, which was prohibitively expensive. However, in today's market, without the over sensationalised threat of pandemic, prices are more competitive and below the cost of internal capital for insurers. Similar excluded medical conditions, pandemic risk management is best considered at a time without immediate major threat.

Flu Mortality (per 100,000 population)

2500 2000 1500 1000 500 0 <1 1-4 5-14 15-24 25-34 35-44 45-54 55-64 65-74 75-84 >84

Age (years)

Source:

28

Aon Benfield

Terrorism risk

Terrorism has become an increasingly recognised risk factor over the last two decades. In the wake of the

probabilistic terrorism model. Firstly, a list of potential terrorism targets is identified, along with different types of attacks ranging from nuclear devices to a shooting rampage of a single gunman. For each type of attack, hit zone is analysed. The probability of occurrence for each attack or weapon type is also assigned.

Exhibit 2), insurers took a large hit on their balance sheets but most survived by risk spreading through have focused more on their management of concentration risk. Terrorism is a major catastrophe risk event for Group business providers, especially if their portfolios have concentrated exposure in high risk areas. Similar to other catastrophe events, terrorism is difficult to predict when and where it will next strike. However, the act of terror is usually carried out to maximise why the World Trade Centre and the Pentagon were In determining the exposure to terrorism risk, Aon Benfield has drawn on the knowledge and experience of Aon's counter terrorism experts to develop a

Another major piece of the puzzle is in determining the frequency of terrorist attacks. Historical records have shown that attack frequency has increased over upon in this rapidly changing environment with reference to a range of frequencies indicated by the Impact Forecasting database. As the model has been calibrated at postcode level, client exposure is illustrated by postcode to evaluate terrorism risk at various return periods. The terrorism model is helping insurers to better understand their exposure of terrorism risk and possible financial impact. In addition it can be adopted b as an integral part of an internal capital model and as an enterprise risk management (ERM) tool to negotiate optimal reinsurance terms and demonstrate their ERM strength to rating agencies.

Figure 4: Heat Map of Terrorism Risk Exposure

Attack in the City of London Attack at Canary Wharf Concentric Distance from Impact

(exposure) for client data

overlaid with two possible event attacks.

Source: Impact Forecasting

29

Solvency II Revealed

Case study

To highlight the benefits of pandemic modelling to

For this portfolio, the models suggest that the 1 in 200 year influenza pandemic loss is GBP8.3m and terrorism cost is GBP6.2m. The pandemic and terrorism events are assumed independent of one another i.e. zero correlation. Technical Specifications, the combined 1 in 200 year catastrophe loss will be GBP10.4m (=SQRT(8.3^2+6.2^2)), which is consistent with the capital requirement under the Standard Formula approach.

Table 2: Modelled Pandemic Losses in London

Pandemic PML Points Return Period (£000) Modelled Losses 10,742 8,316 5,643 1,980 144

Table 1:

Table 1: Hypothetical LTA Portfolio

Age Group 35-44 Lives insured 20,000 Sum Insured £3,000,000,000

45-54

20,000

£4,000,000,000

Total

Source: Aon Benfield

40,000

£7,000,000,000

400 200 100 50 Average Annual Loss

Source: Aon Benfield

Based on the Solvency II Standard Formula, the the portfolio Sum at Risk (SAR). For the age profile of this particular portfolio, best estimate reserve is capital requirement is therefore calculated at GBP10.4million. Assuming lives insured are evenly distributed across cost is expected to be GBP10.3million for the next 12 months (based on TMC00 table).

Table 3: Modelled Terrorism Losses in London

Terrorism PML Points Return Period 400 200 100 50 Average Annual Loss

Source: Aon Benfield

(£000) Modelled Losses 14,347 6,193 4,774 2,016 1,131

Figure 5: PML of Pandemic losses in London

2.5%

2.0%

Exceeded Probability

1.5%

B

1.0%

0.5%

0.0% 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000

Modelled Losses (£000)

Source: Aon Benfield

30

Aon Benfield

A typical structure of pandemic cover is considered with the following features: mortality loss mortality loss i.e. 90% excess 110%

Table 4: Summary of London and Edinburgh Portfolios

London 10.4 6.3 4.1 0.37 9% Edinburgh 7.7 1.6 6.1 0.37 6%

This results in a pandemic cover with an attachment

Source: Aon Benfield

Figure 6: Illustration of Pandemic cover

year for this example). The loss retained by the cedant before the pandemic cover is triggered is GBP1m so

Losses (£millions)

25

20

200% of annual mortality loss Limit £9.27m Buffer Expected annual mortality loss £10.3m Attachment point (110% of annual mortality loss)

(=SQRT(1^2+6.2^2)) resulting in a capital saving ofGBP4.1m compared to no pandemic cover. favourably to an average cost of capital of 10% to 11%. Ceded RoE is the cost of reinsurance divided by capital relief from buying reinsurance. Accretive risk transfer implies that ceded RoE is less than internal cost of capital. Assuming the same portfolio of lives insured is in

15

10

5

0

Source: Aon Benfield

Terrorism risk can also be mitigated through Benfield's Global Death and Disability catastrophe benchmarking study examines the catastrophe reinsurance purchasing pattern between countries. The

arising from a terrorist attack is significantly less in If the Edinburgh portfolio has the same cost, attachment point and limit for the pandemic cover as becomes GBP1.6m (=SQRT(1^2+1.2^2)) with a capital it is clear, for this particular example, that pandemic reinsurance is a cost effective risk mitigation solution compared to holding capital. factors such as attachment point and the relative risk a very cost effective method to mitigate extreme mortality events other than pandemic. In addition, the cost of ceding the catastrophe risk to reinsurers is well below the internal cost of capital of insurers.

Conclusion

Despite the relatively small capital requirement on average for life cat risk in the overall SCR, reinsuring life catastrophe risk with effective terms and conditions is a very cost efficient tool to achieve a reduction.

31

Solvency II Revealed

Rating Agencies and Solvency II

Ratings agencies are unlikely to change their core ratings processes, including capital requirements and risk management expectations, as a result of Solvency II. However, the new regulation will change how companies think about risk. Insurers that can demonstrate an effective internal modelling process are likely to be at an advantage and may achieve favourable capital adjustments under the rating agency models over time. This could mean a reduction in the required capital needed to support their existing rating.

Rating agencies are just as focused on Solvency II as regulators and regulated companies. However, a rating looks beyond a simple quantification of solvency (Table 1), considering ongoing financial performance, viability of management and strategy, and other operating issues that support capital growth and sustainability over time. Solvency II will significantly increase the prominence of regulatory capital but, ultimately, companies that wish to maintain a strong rating and compete in the global marketplace will need to keep a focus on ratings and the underlying capital considerations, beyond those of Solvency II.

Table 1: Drivers to Solvency II Versus Those of a Rating

Solvency II Standard & Poor's*

Investments

Source: Aon Benfield *In Standard & Poor's ratings rationale an opinion of strength is provided for each of the first seven areas listed

Capital Implications

With the implementation date for Solvency II approaching and results from QIS 5 published, rating agencies believe Solvency II will have a significant capital impact on the insurance industry. Standard and Poor's (S&P) report entitled Solvency II Uncertainty after Fifth Quantitative Impact Study published in April 2011 states that insurers are likely to maintain a material buffer of 20% above the SCR. Based

on the QIS 5 results, this would lead to a further 8% of participants unable to meet their solvency target. S&P stated that its overall concern is that approximately one quarter of European insurers will see their capital position challenged under Solvency II. With the macro-economic uncertainty and softening underwriting cycle, risk-mitigation has become increasingly important. Insurers will need to rationalise the link between risk tolerance, capacity and reward in order to enhance business strategy.

32

Aon Benfield

Fitch states in its June 2011 report, Solvency II Set to Reshape Asset Allocation and Capital Markets, that insurers will make significant changes to asset portfolios in order to enhance their capital position. Fitch anticipates a shift from long-term to shorter-term debt and a migration towards higher-rated corporate debt and government bonds. Ultimately, insurers would consider adopting lower risk investment policies in order to reduce their exposure to volatile assets.

Table 2: Selected Differences in Capital Quantification

Solvency II -- Standard Formula S&P Risk Based Capital Model

Tier 1 and Tier 2 considerations Largely factor-based

Largely factor-based Largely factor-based

peril-based scenarios

Source: Aon Benfield

Transitional Arrangements

Proposed legislation project Omnibus II that will amend the original Solvency II directive will potentially reduce the risk of insurers being unable to meet capital requirements in the short term. Omnibus II outlines maximum transition periods of between 5 and 10 years for key aspects of Solvency II, such as meeting the full SCR and the treatment of hybrid instruments in solvency measures. In its March 2011 report entitled Weighing Solvency II's Impact on A.M. Best's Ratings, the firm explains that due to the transitional measures, the level of market disruption will be much less than if Solvency II was enforced in its entirety from the start date. Insurers are no longer at risk of having their market position abruptly challenged with fast-approaching deadlines, and transitional periods allow insurers a settling-in period for when Solvency II is put into operation. The proposed transitional periods may shift focus away from the quantitative capital impact that the regulation brings to the detailed, non-capital related specifics of the implementation. However, there is still debate on the ultimate content and outcome of the final terms of Omnibus II. Regulators must agree to the proposals and, if there is significant deviation from these then rating agencies may have new concerns for insurers with which to contend.

Future Rating Agency Capital Model Considerations

S&P and A.M. Best are the two ratings agencies with established capital models. Neither agency has suggested that their respective models will be changed in light of Solvency II. However, both have stated that considerations for a company's internal modelling process will be taken into account when assessing capital management and risk management. ERM Analysis: Methodology for Assessing Insurers' Economic Capital Models, confirmed it was refining its methodology for assessing an insurer's Economic 3 review, which is the next stage of its ERM focus. The review will concentrate on the quantitative and qualitative modelling considerations and specific risks that an insurer embeds into its ECM framework. meet certain criteria. An insurer must have an existing ECM which is incorporated into its decision making process and which is sufficiently documented, in order for the analysis to take place. It is also likely that only insurers with `excellent' or `strong' ERM after

33

Solvency II Revealed

review to have a significant impact on ratings. However the review can highlight risk management issues and also demonstrate whether the ECM adequately quantifies risks. This in turn can impact ERM and capital conclusions, as well as the assessment of an insurer's management and strategy. expected for those undergoing an internal model approval for Solvency II, including the following:

application

modelling process who can demonstrate a proficient Economic Capital Modelling process, is at an advantage over those who have not. Notably, those insurers whose internal capital modelling processes are viewed by S&P to be credible may achieve favourable capital adjustments within S&P's proprietary model up to one ratings category. This could mean reduction in the required capital needed to support their current rating. The question remains whether rating agency capital will remain more important than regulatory capital once Solvency II is effective? This is likely, but time will tell.

34

Aon Benfield

Reinsurance Assets: Individual vs. Aggregate Valuation Methods

Solvency II offers opportunities to include the effect of risk mitigation techniques in regulatory capital calculations and will likely generate new reinsurance techniques that fit within the Standard Formula. However a key step is accounting for the impact of historical risk mitigation and ensuring the true value of past investment is reflected in available capital. Insurers should look deeper into the merits of calculating the fair value of their reinsurance assets, as there may be an opportunity to unlock greater value than expected.

Introduction

Reinsurance assets are one of the most complicated to calculate within the Solvency II Fair Value Balance sheet. Nevertheless, if the reinsurance bought is material, correct calculations can have an important impact on the Net Available Assets. The complexity of past and present reinsurance programmes creates challenges to actuaries to deliver accurate estimates. The Solvency II directive has recognised the importance of a correct calculation of reinsurance assets by requiring a separate calculation for the gross best estimate claim provisions and the amounts recoverable from reinsurance In practice however, due to lack of time and knowledge or data constraints, many companies calculate the value of reinsurance assets using rules of thumb and relying on the gross results. Sometimes these give remarkably good results, but sometimes they can be very misleading. This article aims to reveal more accurate approaches, along with possible pitfalls that can easily be identified upfront, to indicate the accuracy of rule of thumb approaches.

Table 1: Sample Balance Sheet in Different Approaches

Assets 110 10 120 Liabilities 20 100 120 110 8,3 118,3 Assets

Case study

The case study presents an insurer's balance sheet that is represented from three different angles. The first approach (base scenario) is the current accounting view. Gross claim liabilities are valued on a case by case basis taking a conservative approach. The reinsurance assets are valued on an individual claim basis, based on the latest information available. Within Option 1, the actuary has valued the best estimate of the gross claim liabilities using actuarial techniques. Given the conservative approach each gross technical claims. For valuing the reinsurance assets, the actuary is lacking individual data and decides to use a net-to-gross ratio derived from accounting data. The reinsurance assets are hence valuation on both sides of the balance sheet, the Net Asset Value has changed from 20 to 35.3.

Liabilities 35,3 83 118,3 110 13 123

Assets

Liabilities 40 83 123

Source: Aon Benfield

35

Solvency II Revealed

Figure 1: Incurred Position Towards Its Last Known Incurred Value

250.00%

200.00%

150.00%

99.64% 95.17% 82.04% 73.42% 64.30% 88.49% 35.42% 100.00%

100.00%

50.00%

0.00% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Source: Aon Benfield

In Option 2, the actuary has analysed the large losses individually and discovered a trend that is seen amongst many insurance companies. Figure 1 explores the development of the incurred position towards its last known incurred value. The development periods are represented in the x-axis. The y-axis shows (using a box-plot) the spread of the observed ratios (current example the claims were reported if their last known incurred value exceeded EUR1.5m (note that the example is based on a European country's market motor book in Euros). After the first year of development, the median of all claims in the database have reached only 42% of their last known incurred value. This number increases over time to top 90% after eight years of development. After 13 years of development, the uncertainty drops and only a limited number of claims have an uncertain ultimate value. Based on this, the actuary decides to do a detailed analysis of the reinsurance assets. Applying a net-to-gross ratio as in Option 1 will undervalue these assets materially, hence leading to a calculated Net Asset Value that is understating the true available capital. The best estimate of the reinsurance assets is now calculated at 13, which is 30% higher than the current accounting value. The Net Asset Value increases further to 40.

36

Aon Benfield

Solvency II requirements

The Solvency II directive does not give any guidance on how these reinsurance assets need to be valued. They have to follow the same principles as the gross claim liability which can be summarised and detailed in the below table:

Table 2: Guidance on the Calculation of Reinsurance Assets

Directice Specific remarks for the calculation of Reinsurance Assets

"The best estimate shall correspond to the probability weighted average of future cash-flows, taking into

reinsurance program. Also the various clauses in the contract can increase the level can only be calculated from the probability weighted average.

rate term structure." deduction of the amounts recoverable from reinsurance contracts and special purpose vehicles. Those amounts shall be calculated separately, in accordance with Article 81" "The risk margin shall be such as to ensure that the value only be done once an accurate gross cash flow estimate

insurance and reinsurance obligations." be obtained if a separate calculation is done for both as the techniques, data and contract issues are different: gross calculations should reflect the insurance reinsurance policy conditions.

are protected by reinsurance, the risk mitigating effect of the reinsurance programme should be included in the calculation of the risk margin. Also the counterparty risk that arises should be included.

The timing of the payment of the reinsurance contract can deviate from the gross special purpose vehicles. "When calculating amounts recoverable from reinsurance contracts and special purpose vehicles, insurance and reinsurance undertakings shall take account of the time "The result from that calculation shall be adjusted to counterparty. That adjustment shall be based on an assessment of the probability of default of the counterparty and the average loss resulting there from (loss-given-

no longer than three months. The calculated recoverables should be reduced to reflect the credit position of reduction of assets due to expected counterparty default

Source: Aon Benfield

Within the scope of this document we will only focus on the calculation of the claim cash flows on a gross and net basis.

37

Solvency II Revealed

Aggregate methods

The most common approach to calculate the best estimate for reinsurance assets is based on aggregate triangle techniques. Historical triangles of paid and incurred claims data, gross of reinsurance, are modelled and projected to ultimate by using a variety of different mathematical curves. Assuming that past developments are a reasonable indicator of future developments the Bornhuetter-Ferguson (BF) method incorporates prior estimates of ultimate claims into the modelling process. It is particularly appropriate for recent years of account where the development factor modelling method may produce unreliable results. Unlike development factor methods, the BF method can additionally take into account collateral information, such as initial loss pricing models that may exist, and benchmark development patterns. The prior estimates are adjusted using development factor projection methods under a weighted average approach: the fewer the years of development, the higher the weighting placed on the prior estimate. Net results are then obtained by netting down the calculated gross results using net-to-gross ratios. These ratios are calculated from analysing the net incurred claims to the gross incurred claims or by a separate estimate of the ultimate retained percentage after examination of the reinsurance programme. When using such a method, one implicitly makes many assumptions: 1. Stable incurred patterns: When bringing the incurred values to ultimate, the development ratios show the release or increase of reserve surplus. In practice (as shown higher) large losses tend to be under-reserved whereas attritional losses carry the bulk of reserve surplus. This means that within the incurred triangle, two trends are aggregated.

2. Change in reinsurance programme attachment and limits: It is assumed that the reinsurance programmes are stable and future trends can be derived from past trends. Since claims triangles span many years, they can cover reinsurance cycles. These cycles have an effect on the pricing and this is compensated in the cedant book by retaining more or less risk by changing the priorities of the programme. Figure 2 details the payment pattern of European motor losses that exceed a threshold value. The higher the threshold value, the slower the payment pattern becomes.

Figure 2: Payment Pattern of European Motor Losses

100%

600 000 1 250 000 2 500 000

80%

60%

40%

20%

0% 1 2 3 4 5 6' 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Source: Aon Benfield

Therefore, changing the reinsurance retentions not only has an effect on the net-to-gross approach but also affects the payment pattern. 3. Overall change in reinsurance program: Historically lines of business can be protected by many reinsurance programmes: individual line of business protection by means of proportional or non-proportional treaties, portfolio protection by means of aggregate covers protecting the overall portfolio retention. In an incurred net triangle, these various programmes can have a material impact on the netting process and -- probably even more importantly -- are a result of recovery allocation rather than a true recovery of an individual loss.

38

Aon Benfield

In many cases, when a closer examination of the (individual) claims data or the reinsurance policy wordings is made, even more practical issues arise: 1. Individual claims that have a substantial increase or decrease of reserves can have an important impact on the incurred development pattern. If the netting is done on aggregated data, the effect of such a change in incurred can lead to misleading results if it is assumed to be `normal' and hence projected forward. 2. How should insurers reflect the various clauses in the reinsurance programmes in such a modelling technique? 3. How should insurers deal with issues such as layering of reinsurance and clash covers?

Current industry practices include both deterministic and stochastic approaches to quantify the reinsurance asset. Due to the non-linear effects caused by reinsurance, the stochastic approach is preferred. In the case study described below the technique of 'nearest neighbour' has been applied to calculate the value of reinsurance assets. Individual models are not the magic trick that solves all problems when dealing with limited data, noise and a large volatility in outcomes. However, they do offer a much more transparent and robust analysis with the biggest advantage that each individual (large) claim can be discussed with the claims manager. This probably makes this technique much more defendable from a `use' perspective. For each individual claim ultimate paths can be created and discussed with the claims manager, reinsurance manager and legal department to assess the quality of the modelled results. Experience, non-quantitative data (e.g. medical reports, court uncertainty) can be used to discuss the feasibility of modelled results.

Individual models

In individual models, each claim that has a potential to create a reinsurance asset is projected to its ultimate position and gross and reinsurance recoverable cash flows are calculated. Such techniques have the major advantage that the actual reinsurance programme can be applied and hence all uncertainty of the results is due to the uncertainty of the gross (individual) model. Many assumptions are actually hidden in aggregate modelling (e.g. if one assumes that the past experience can be projected into the future), while in individual modelling many assumptions need to be defined explicitly. For example: selected and modelled to ultimate and how should these levels be changed to reflect claims inflation? (unreported or below the selected threshold) which have the potential to create a reinsurance asset when they are developed to ultimate?

Which approach to take?

The use of aggregate methods to calculate the value of reinsurance assets is defendable as a proxy method but only when all above considerations have been evaluated and in case the reinsurance assets are minimal compared to gross estimates. In other cases it is worth analysing individual methods and comparing the results between aggregate and individual models. Certainly when the impact on the Net Asset Value (and hence the solvency ratio) is material, it is advised to apply individual methods to have a more accurate result.

Case study

The case study was conducted on a European Motor book for a company which had an individual line of but where the priority changed every year. The analysis was done using a stochastic `nearest neighbour' method where each (potential) large loss was projected to ultimate and then sent to the various reinsurance programmes to calculate the netting effect. The modelling was done using ReMetrica, Aon Benfield's dynamic financial analysis modelling tool. No discount effect was introduced.

Applying such a modelling technique makes the results much more transparent and allows for individual stress testing on elements such as the impact of one individual claim and parameters.

39

Solvency II Revealed

Figure 3: Accounting for Reinsurance Assets

119.4 100.0

-9% +31% -12%

105.0

91.2

D. Impact of interest sharing clause added C. Ultimate effect added. The scenario's left and right were obtained by assuming fgu losses were 5% lower (left) or higher (right) then assumed in the best estimate. B. Accounting value of Reinsurance assets including effect of stabilisation clause A. Accounting value of Reinsurance Assets based on reported claims

Source: Aon Benfield

The first bar (A) reflects the current accounting situation of the reinsurance assets. Accounting valuation was done on paid recoverables + EOY estimate of the recoverable based on the incurred value as at EOY. If the company has valued the reinsurance assets using the gross incurred value but kept account of the effect of the stabilisation clause and the payment pattern of large losses, the value of the reinsurance assets would drop by 9%. This is reflected in bar B. In reality however, inflation will have an impact on the ultimate cash flows and -- in general -- large losses tend to increase over time due to incomplete information (as shown in one of the above graphs). After conducting a (stochastic) ultimate calculation of all individual gross losses that have the potential to breach the reinsurance programme, the best estimate is changed to 119.4 which is above the accounting value.

Two stress tests were conducted to calculate the effect on the value of the reinsurance asset in case From Finally the accounting value of reinsurance assets did not include any effect of a legal interest sharing clause which made the final best estimate drop to 105% (from the original 100%). The impact of clauses and individual claims data (and their relative position towards settlement) causes a roller coaster on the best estimate assumption. The example was chosen as each individual effect has a large impact and therefore demonstrates the importance of taking each into account. The fact that the final result was close to the initial accounting value should not be generalised since the ultimate is well below or above accounting value in many analysed portfolios.

40

Aon Benfield

Risk and Capital Modelling for Solvency II: A Pillar of Strength

Many insurers have chosen to reap the benefits of using an internal model for Pillar 1, albeit in consideration of the investment in time and resources. However, internal models can also play a positive role under Pillar II as part of the Own Risk and Solvency Assessment (ORSA) to demonstrate to regulators that risk is being effectively managed.

Introduction

Solvency II is putting the spotlight on capital modelling and creating both opportunities and challenges. There is significant growth in the adoption of partial or full internal models as insurers seek a more representative capital assessment than the Standard Formula. But there is also an increase in complexity as companies using internal models need to ensure their models can satisfy each pillar of Solvency II as well as existing and emerging accounting standards. In addition, they must ensure that these models are actually used in making key decisions about the business. So far in the Solvency II project, both regulators and insurers have focused much of their attention on Pillar I to calculate solvency capital requirements. Efforts are now well advanced, with firms having stepped through several Quantitative Impact Studies and have chosen either the prescriptive Standard Formula or a bespoke internal capital model using a Dynamic Financial Analysis tool such as Aon Benfield's ReMetrica.

For firms that have chosen to build partial or full internal capital models, the ORSA necessitates a longer term view of risk. Whereas Pillar I requires risks to be considered on a one-year time horizon, insurers prefer to model the total risk emerging in run-off when making risk management decisions. With this in mind, Aon Benfield updated ReMetrica to allow firms to generate both one year and longer-term views using a single model and thereby avoid duplication of effort. In addition, Pillar II requires firms with internal models to prove to the regulator that adequate model governance processes exist around the model and its data. To help address this requirement, ReMetrica's Enterprise Edition incorporates new components to help companies keep track of their models as they move one from iteration to the next. Enterprise Edition allows companies to control which users have access to the which parts of a model, record who has changed what and when, and compare one iteration of a model with another with reporting functions. The ORSA has sometimes, mistakenly, been viewed as a simple box-ticking exercise, supplementing the Standard Formula with a handful of template documents and forms. Although understandable given the lack of prescriptive guidance from EIOPA, this approach misses a principal aim of Pillar II ­ to prove to the regulator that a risk management culture permeates the organisation, and influences both everyday decision-making and longer-term business strategy. A regulator could impose a capital loading if they do not see sufficient evidence of a strong risk mitigation strategy.

Onus on the ORSA

Attention is now shifting towards demonstrating wider risk management and governance capabilities to the regulator to satisfy Pillar II and, in particular, the completion of the Own Risk and Solvency Assessment, regulator (EIOPA) due imminently. EIOPA has previously defined the ORSA only in broad, nonprescriptive terms, defining it as "the entirety of the processes and procedures employed to identify, assess, monitor, manage, and report the short and long term risks a (re)insurance undertaking faces or may face and to determine the own funds necessary to ensure that the undertaking's overall solvency needs are met at all times."

41

Solvency II Revealed

To this end, some firms using the Standard Formula for calculating capital are also turning to tools such as ReMetrica to model and analyse their key risks for Pillar II under a range of business scenarios and time frames. This particularly applies where risks, such as natural catastrophes, are not adequately captured by the Standard Formula assumptions or the risk interactions are complex. Modelling key risks in such cases allows firms to demonstrate these are being analysed, measured and monitored. In turn, this provides a tangible, quantitative output to inform business decisions, thereby allowing a firm to realise real business value from the investment in meeting the requirements of Solvency II. It also provides a less burdensome entry point into capital modelling, with some firms likely to evolve these models into a partial or full internal model.

Conclusion

Solvency II is increasing both the take-up of models and the range of decisions influenced by models. Tools such as ReMetrica are well-suited to the disciplined, analytical approach to risk management required for Pillar II, with a clear direction towards wider usage and acceptance of modelling key risks.

42

Aon Benfield

Contact Information

Gareth Haslip Head of Risk & Capital Strategy EMEA +44 20 7522 8137 [email protected] Marc Beckers Head of Aon Benfield Analytics EMEA +44 7931 472 999 [email protected] John Moore Head of International Analytics Aon Benfield Analytics +44 20 752 3973 [email protected]

Scan here to access Aon Benfield's thought leadership publications including the Solvency II Revealed report.

About Aon Benfield

Aon Benfield, a division of Aon Corporation (NYSE: AON), is the world's leading reinsurance intermediary and full-service capital advisor. We empower our clients to better understand, manage and transfer risk through innovative solutions and personalized access to all forms of global reinsurance capital across treaty, facultative and capital markets. As a trusted advocate, we deliver local reach to the world's markets, an unparalleled investment in innovative analytics, including catastrophe management, actuarial and rating agency advisory. Through our professionals' expertise and experience, we advise clients in making optimal capital choices that will empower results and improve operational effectiveness for their business. With more than 80 offices in 50 countries, our worldwide client base has access to the broadest portfolio of integrated capital solutions and services. To learn how Aon Benfield helps empower results, please visit aonbenfield.com.

This document is intended for general information purposes only and should not be construed as advice or opinions on any specific facts or circumstances. The comments in this summary are based upon Aon Benfield's preliminary analysis of publicly available information. The content of this document is made available on an "as is" basis, without warranty of any kind. Aon Benfield disclaims any legal liability to any person or organization for loss or damage caused by or resulting from any reliance placed on that content. Aon Benfield reserves all rights to the content of this document.

43

Information

untitled

44 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

1306582


You might also be interested in

BETA
IBM IAA Advanced Model Overview
2010-12-Current Conditions and Outlook for the U.S. and CT Economy.pub
Agribusiness planning