ECTraining HK

Joint Regional Seminars Economic capital training Presented by: Supported by: © 2011 Towers Watson. All rights reserved...

0 downloads 212 Views 2MB Size
Joint Regional Seminars Economic capital training Presented by: Supported by:

© 2011 Towers Watson. All rights reserved.

Greg Solomon, Swiss Re and Paul Carrett, UBS John Rowland, Towers Watson and Penny Fosker, Towers Watson

Objectives 

Hands on introduction to economic capital 



Background – what are risk based capital techniques and why these techniques were introduced How economic capital measures are used in business – why they are relevant to you



Defining the framework – the key areas of focus



Historical trends – how approaches have changed over time



Practical implementation – stepping through the development phases



Operational issues – some of the key hurdles you need to address

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

2

towerswatson.com

Why should economic capital matter to you? Globally, accounting and regulatory regimes are moving towards principles based approaches using economic measures But also from a management perspective… …company performance will be driven by the underlying economics in the long term Recognising this is leading many companies to make strategic decisions which are informed by economic capital metrics © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

3

towerswatson.com

What is capital?

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

4

towerswatson.com

Objectives 

Introducing and defining capital



Different approaches used for different purposes



What are risk based capital techniques



Why these techniques have been introduced

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

5

towerswatson.com

What is capital?  

Free capital

Required capital

Available capital



Assets

Capital = assets – liabilities Required capital = capital required to support the business written 

Regulatory required capital



Rating agency required capital



Internal required capital

Available capital = capital available to support the business written

Technical provisions

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

6

towerswatson.com

Why do companies hold capital? 

Holding capital to support the business is a regulatory requirement…



…and it’s also essential for managing the business

Capital

Risk





Risk exposure determines capital needs Required capital is a function of the tail of the risk distribution

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

7

towerswatson.com

How much capital is enough? How much is too much?





The market price of risk is a key driver of value creation

Value

Price depends on overall characteristics of the risk distribution



Capital utilisation has a cost and hence reduces value creation

Capital

Risk





Risk exposure determines capital needs Required capital is a function of the tail of the risk distribution © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

8

towerswatson.com

Capital = assets - liabilities 

But what do we mean by “assets” and “liabilities”?



Different definitions arise from different accounting conventions 

Inclusion/exclusion of specific assets or liabilities — Exclude intangible assets



Application of different methodologies — Book value versus market value methodology — Inclusion of prudent margins in liabilities

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

9

towerswatson.com

Different definitions – a sample overview Solvency 1

Risk based capital

Economic capital

Assets

Mix of book value and market value

Market value

Market value

Technical Provisions

Including a margin for prudence

Best estimate liability plus provision for adverse deviations

Best estimate liability plus risk margin for non-hedgeable risks

Net premium valuation using prudent valuation interest rate

Discount at risk free rate

Discount at risk free rate

No explicit allowance for value of options and guarantees

Explicit allowance for value of options and guarantees

Simple factor based approach – no direct link to risks

Sum of capital charges across specified risks

Change in net asset value under stressed conditions

Cap on allowance for reinsurance

Capital charge for each risk category calculated as risk factor x risk driver

Parameter driven stress tests or stochastic derivation, dependent on purpose and level of sophistication

No explicit allowance for value of options and guarantees Capital

Static risk factors based on industry level analysis at a selected probability of remaining solvent over a given time period

Allow for diversification within and across risk categories

Covariance factor / formula applied to allow for diversification across risk categories © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

10

towerswatson.com

Economic capital

Risk based capital

Solvency 1

Comparing approaches… Prudent reserves Simple factor based Weak link to underlying risks

Prudent reserves and fair value assets Industry factors applied to individual risk drivers

MVA and BEL (excl TVoG) Industry factors applied to individual risk drivers

MVA and BEL (incl TVoG) Industry factors applied to individual risk drivers

MVA and BEL (incl TVoG) Industry stress tests applied across risk categories

MVA and BEL (incl TVoG) Company specific stress tests applied across risk categories

MVA and BEL (incl TVoG) Internal model approach

Increasing levels of sophistication © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

11

towerswatson.com

Recap… 

Insurers accept risk from policyholders in return for a premium



These risks need to be managed so that





Policyholder claims are paid as they fall due



The company makes a profit

Capital provides a cushion that 



Reduces the likelihood that obligations to policyholders cannot be met in adverse circumstances Allows the company to continue to grow and develop

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

12

towerswatson.com

Why have risk based capital techniques been developed? 

Solvency 1 style capital requirements 

Factors applied to limited risk driver proxies



…and risk management and mitigation techniques are not recognised



Lack of transparency over the actual level of solvency of the company — Due to prudence in reserving bases





…and inconsistencies between reserving strength and capital requirements — Life business:

more prudent reserves = higher capital requirement

— Non-life business:

more profitable premiums = higher capital requirement

Not responsive to changing risk profile

Risk Based Capital The amount of capital a company should hold, based on an assessment of the risks it faces © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

13

towerswatson.com

Risk based capital versus economic capital Practice

Theory Risk based capital



Risk based capital synonymous with regulatory capital calculations 





Applying industry factors to pre-defined risk drivers Allowing for diversification via application of factor or formula

Economic capital synonymous with scenario and probability distribution based approaches 

Stressing the balance sheet to understand the change in the net asset value

Economic capital © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

14

towerswatson.com

Risk based capital versus economic capital 

The Solvency II Solvency Capital Requirement (SCR) is a form of risk based capital calculated on an economic basis 

But it is not the only definition of economic capital Risk based capital Solvency II SCR

Economic capital © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

15

towerswatson.com

Economic capital is measured by “stressing” the economic balance sheet Asset stresses include equity market declines, changes in interest rates and credit spreads, defaults

Stressing the balance sheet produces an economic loss of capital

Liability stresses include adverse reserve development, product underpricing, catastrophe, and uncollectible reinsurance, as well as asset effects

Economic Loss Stressed Capital Baseline Assets

Stressed Assets

Stressed Liabilities

Baseline Capital

Baseline Liabilities

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

16

towerswatson.com

The basic concept of economic capital is straight forward + 0 Losses from “extreme” events are covered by economic capital up to the defined risk tolerance level



Risk tolerance level

Ranked outcomes – each equally likely

What level of assets is needed today to provide reasonable assurance that future obligations can be met? 

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

17

towerswatson.com

Different stakeholders have different risk tolerances 1 in ∞

1 in 1 in 2000 200

1 in 10

Mean

Ideally, a view of the whole distribution will be available 

Policyholder focus Rating agency focus

Regulator focus

Investor focus

Earnings/change in value/surplus

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

18

towerswatson.com

Current regulatory requirements in Asia Pacific Country

Assets

Technical provisions

Capital

Australia

Market

Best estimate plus margins – VIR varies with the economic environment

RBC

China

Book/market

Prudent – VIR not greater than min (7.5%, pricing rate)

Solvency 1

Hong Kong

Market

Prudent – VIR varies with the economic environment

Solvency 1

India

Book/market

Prudent – VIR varies with the economic environment

Solvency 1

Indonesia

Book/market

Prudent – VIR max 9% (IDR), 5% (USD)

RBC

Japan

Book/market

Prudent – VIR varies by policy issue year

RBC

Malaysia

Market

Best estimate plus PRADs – VIR varies with the economic environment

RBC

New Zealand

Market

Best estimate plus margins – VIR varies with the economic environment

None*

Philippines

Book

Prudent – VIR < 6%

RBC

Singapore

Market

Best estimate plus PADs – VIR varies with the economic environment

RBC

South Korea

Book/market

Prudent – VIR locked in at pricing

RBC

Taiwan

Book/market

Prudent – VIR locked in at at pricing

RBC

Thailand

Book/market

Prudent – VIR < 5%, locked in at pricing

Solvency 1

Vietnam

Book/market

Prudent – VIR < 80% of 10yr Government Bond yield, locked in at pricing

Solvency 1

* Institute of Actuaries New Zealand recommends an approach equivalent to Australian requirements. © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

19

towerswatson.com

Future regulatory requirements in Asia Pacific… China Regulator investigating RBC and EC

India Regulator introduced EC as a supplementary measure

Singapore Review of RBC announced

Australia Introduction of ICAAP requirements from 1/1/13

Japan Revised RBC requirement effective from 31/3/12

Hong Kong Regulator investigating RBC and EC

Thailand RBC effective from 30/9/11

New Zealand Required capital effective from 31/12/11

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

20

towerswatson.com

The role of economic capital in business

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

21

towerswatson.com

Objectives 

Economic capital and its role in risk management



Integration of economic capital into business processes



Use of economic capital in risk based pricing

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

22

towerswatson.com

The role of economic capital in business Economic capital and its role in risk management

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

23

towerswatson.com

Improving business performance through ERM and economic capital

ERM goes far beyond the calculation of a number © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

24

towerswatson.com

Gearing up for an effective ERM framework Strategy/Policy • Risk strategy

Risk processes and tools

Monitoring/Reporting • Data structure

• Risk appetite

• Risk identification

• Dashboards

• Risk policies

• Capital management

• “What if” analysis

• Economic capital • ALM/hedging What types and levels of risk support strategic objectives?

• Risk-based product design • Risk-based pricing

What risk information is needed by decision makers?

• Risk optimisation • Technology • Performance management

Organisation

What data and analysis are needed to evaluate risk/reward?

Management

• Governance

• Accept risk / no action

• Roles and responsibilities

• Mitigate risk

• Risk culture

• Transfer risk • Exploit uncertainty to create value

What structure best supports effective decision making? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

25 25

towerswatson.com

An economic perspective links risk, capital and value Maximise value by relating a firm’s decisions on the risks it takes to the decisions on the capital it uses to finance its business Economic capital is the unexpected impact to value at some confidence interval over some time horizon

Value Creation

Return on Risk

Portfolio of Enterprise Risks Risk Structure

Capital Costs

Value Management

Capital Adequacy

Portfolio of Capital Resources

Risk and Capital Management How much capital do I need?

What type of capital do I need?

Capital Structure

Economic Capital © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

26

towerswatson.com

Uses fall into different categories Calculation of capital requirements (including analysis of movement)  Monitoring risk appetite, limits and controls Risk management 

Capital allocation and strategic planning  Business planning 

Reinsurance programme  Product design, underwriting and pricing  M&A, re-organisations  Performance management and compensation  ALM, hedging

Planning



Business processes

Other reporting and key performance indicators: IFRS, regulatory, GAAP, rating agency Wider reporting 

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

27

towerswatson.com

The role of economic capital in business Integration of economic capital into business processes

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

28

towerswatson.com

Questions are being asked about integrated, enterpriselevel risk analysis and decision-making  

How much capital do we need and what return should we get on it? How do we identify the key and emerging risks within our organisation, and how do we manage them?



How can we maximise our return on capital, given our risk appetite?



How do we select our growth strategies, given our risk environment?



 

How do we best invest our assets, given the structure of our exposures? How much, and on what terms, should we insure and hedge? How should we report our risk management results and communicate with external audiences about our risk management programs?



How do we build a risk culture within the organisation?



How do we coordinate all of this? And how do we get started?

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

29

towerswatson.com

Company management is faced with a bewildering array of metrics Economic Capital

Traditional Embedded Value

Target Credit Ratings

Local Statutory Capital

Key Operating Ratios MarketConsistent Embedded Value Value of New Business

Book Capital

IAS

Local Accounting Rules

Internal Rate of Return



Inconsistency



Confusion



Solvency II Capital

RORAC/ RAROC

Information overload

ROE

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

30

towerswatson.com

Economic capital is increasingly becoming a useful tool in influencing significant management decisions Capital Adequacy Assessment/Capital Management Asset/Investment Strategy (Including Hedging)

46%

Reinsurance Purchasing

46%

M&A and Divestiture

24%

Performance Measurement

23%

Incentive Compensation

21%

33%

28%

26%

24%

39%

37%

Product Design and Pricing

13%

35%

52%

Annual Business Planning

18%

28%

54%

Strategic Planning and Capital Allocation

48%

28%

35%

42% 31%

15%

Currently using EC

12%

26%

62%

Plan to use EC in the next 24 months

54%

Do not use EC and no plans to use

Towers Watson global ERM survey 2010 Q.32. (Select all that apply.) © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

31

towerswatson.com

An economic capital framework allows risk based decisions to be put on an equal footing Traditional approach

The problem

Approach using economic capital

Pricing

Maximise ROE and/or IRR

GAAP equity subject to local accounting rules

Use economic capital as capital measure to standardise risk assessment

Diversification Benefits

Limited

“Eggs in one basket”

Standardisation of risk allows leveraging risk diversification opportunities

Capital Allocation

Local accounting

Management not informed on optimal deployment of capital

Risk transparency informs more optimal capital deployment

Hedging & Trade-off Analysis

Maximising earnings relative to capital impacts

Incents arbitrage of local accounting rules

Use economic capital to make true risk to value transparent

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

32

towerswatson.com

Analysis can transform economic capital into a powerful management tool 

Stress and sensitivity scenarios can provide valuable risk management information and facilitate discussions with senior management 



“What if” analysis can be help inform risk and strategic management decisions 



e.g. scenarios based on 1987 crash, Hurricane Katrina, 9/11, 2004 Indian Ocean tsunami

e.g. M&A, reinsurance strategies, asset allocation decisions

Ideally want the ability to see impact on results 

Across different probabilities



At several levels of granularity (e.g. by product, risk, geography)



With and without allocation of diversification benefits

With this capability, insurers can use economic capital to support a wide range of decisions and benefit from its full potential © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

33

towerswatson.com

Illustrative

Phased implementation depends on business benefit vs cost High

Capital requirements

Risk management Reinsurance programme

Deliv

Performance management and compensation Investment: ALM / Hedging

Business Benefit / Priority

Medium

Capital allocation

Business planning Management of participating business

Product design, underwriting and pricing Other reporting and KPIs: IFRS, Regulatory GAAP, rating agencies M&A, reorganisations

Low

Low

Medium

High

Difficulty / Cost © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

34

towerswatson.com

Using economic capital is decision making can result in a competitive advantage “Risk comes from not knowing what you're doing.” - Warren Buffett “Risk is like fire: If controlled it will help you; if uncontrolled it will rise up and destroy you.” - Theodore Roosevelt “Any time you take a chance you better be sure the rewards are worth the risk because they can put you away just as fast for a ten dollar heist as they can for a million dollar job.” - Stanley Kubrick © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

35

towerswatson.com

The role of economic capital in business Use of economic capital in risk based pricing

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

36

towerswatson.com

Risk based pricing can help maximise value created by new business 

Granular risk based pricing 



Focused targeting and “cherry picking” 



Calculation of the minimum price to charge for a product to more accurately reflect the risk and required return arising from product features, including leverage, asset risk, embedded financial options and the capital required

Identify and target areas where the company’s products are particularly profitable (and unprofitable), both in absolute terms and in their capital efficiency

Exploitation of market inefficiencies and risk synergies 

An effective and dynamic risk based pricing system will allow the company to exploit opportunities arising from temporary market pricing anomalies

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

37

towerswatson.com

Example: Stochastic pricing results show significant tail risk which leads to high economic capital 25.00% 20.00%

Expected Return

15.00% 10.00% 5.00% 0.00% -5.00% -10.00% -15.00%

Economic capital source

-20.00% Interest Rate Scenario 

Product has average return of 12% over set of stochastic scenarios



A small number of scenarios have negative returns



Poor scenarios have low projected interested rates © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

38

towerswatson.com

Example: After hedging required economic capital would be reduced 25.00% 20.00%

Expected Return

15.00% 10.00% 5.00% 0.00%

Economic capital source

-5.00% -10.00% -15.00% -20.00% Interest Rate Scenario



Average return is slightly lower over set of stochastic scenarios



Ends up being a cost for most scenarios as option expires with no value



Significantly improves tail scenarios © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

39

towerswatson.com

Implementation – from simple methods to best practice

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

40

towerswatson.com

Objectives 

The purpose of economic capital



Defining economic capital



Defining risk distributions and loss functions



An introduction to advanced modelling

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

41

towerswatson.com

Implementation – from simple methods to best practice What is economic capital for?

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

42

towerswatson.com

Why do insurers calculate economic capital? United Kingdom

Malaysia

6,000

Thailand 800

1,200

750

1,150 5,500

700

5,000

4,500

650 1,050 SET Index Level

Kuala Lumpur Composite Index Level

FTSE 100 Index Level

1,100

1,000

600

550

950 500 900 450

4,000 850

3,500

800

Date

350

Date

Taiwan

Date

Hong Kong

China

25,000

3,000

7,500

22,500

2,750

7,000

20,000

2,500

6,500

SEE Composite Index Level

8,000

Hang Seng Index Level

TWSE Index Level

400

17,500

2,250

6,000

15,000

5,500

12,500

1,750

5,000

10,000

1,500

Date

Date

2,000

Date

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

43

towerswatson.com

EX ER C

Exercise 1: Defining extreme events

IS E



That was an example of an extreme event



Discuss with your neighbours and define 5 more 

Document on the sheet provided

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

44

towerswatson.com

Other events that could lead to losses A meteorite lands on China

Cure for cancer

Improvement in longevity

Mass increase in lapse rates

Interest rates fall to zero / rise to 10%

Nuclear Armageddon

Increase in expense

SARS

Global warming

Earthquake

The Plague kills 50% of population

Increase in credit spreads

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

45

towerswatson.com

What is economic capital?

Economic capital is the money you hold to protect you against the impact of adverse events or combinations of adverse events happening simultaneously

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

46

towerswatson.com

What is economic capital?

Economic capital is the money you hold to protect you against the impact of adverse events or combinations of adverse events happening simultaneously

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

47

towerswatson.com

How much capital is enough … do I protect myself against all events? “… for insurers to consider how this liability might change in reasonably realistic adverse scenarios …” — Individual Capital Assessment, UK “… plausible adverse scenarios …” — Dynamic Capital Adequacy Testing, Canada “… on a basis that is more conservative than best estimate and that considers scenarios of adverse experience ...” — Capital Adequacy Standard, Australia

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

48

towerswatson.com

Events can be too extreme ... outside a “risk appetite” A meteorite lands on China

Cure for cancer

Improvement in longevity

Mass increase in lapse rates

Interest rates fall to zero / rise to 10%

Nuclear Armageddon

Increase in expense

SARS

Global warming

Earthquake

The Plague kills 50% of population

Increase in credit spreads

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

49

towerswatson.com

Another complication is that events don’t come along in isolation … 

For example, equity markets fall and interest rates reduce at the same time … FTSE 100 Index Level 7,000

FTSE 100 Index Level

6,000

5,000

4,000

3,000

Date



… and, as a result, you might also experience a spike in lapse rates 25% 20% 15% 10% 5% 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Month



We need to allow for these interactions

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

50

towerswatson.com

Implementation – from simple methods to best practice Two possible approaches

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

51

towerswatson.com

Economic capital is the answer to specific questions Question

“Technical terms”

1. What risks am I exposed to?

1.

Risk universe, risk factors

2. How might the risks I am exposed to vary?

2.

Time horizon, risk model, marginal risk distribution, dependency structure, correlation, copulas

3. Given the evolution of risks, how will this affect my business performance?

3.

Balance sheet response to risk, loss function

4. What level of impact am I concerned about?

4.

Risk measure, confidence interval, risk appetite, VaR, Tail-VaR, risk of ruin, required economic capital

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

52

towerswatson.com

Two general approaches have emerged to measure economic capital

Liability runoff approach The level of total initial assets, less some measure of reserves for liabilities, required to pay all future policyholder benefits at the chosen confidence level

One-year mark-to-market approach The level of assets, in addition to the market value of liabilities, needed to cover a fall in the market value of net assets over a one-year time horizon at the chosen confidence level

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

53

towerswatson.com

Each of the two main approaches has pros and cons Risk Management and Risk-Based Performance Considerations Liability Runoff

One-year Market-Consistent

+ Measures risk over period risk is held, so better for long-term decision making

+ Facilitates link between risk quantification and risk management

+ Longer-term decision making not distorted by volatility of economic assumptions over short term, i.e. stable capital requirements

+ Short-term volatility to economic assumptions may be very relevant when assessing risk management options

+ More accurately captures risks that emerge over time

+ Risk quantification and risk management linked to performance management over the typical annual performance reporting cycle

– Does not respond to short-term market fluctuations, which might impair risk management decisions

– Can lead to volatile capital requirements from one to the next period

– Can result in a timing mismatch, with shortterm performance being compared with risk and capital based on a longer term

– Fails to provide information about emergence of risk over time

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

54

towerswatson.com

Both approaches pose certain implementation challenges Implementation Considerations One-year Market-Consistent

Liability Runoff + Relatively standard, i.e. consider the real world projection of liabilities until they run off

+ Generally easier to calibrate EC to target confidence levels over one year

– Inclusion of interim solvency assessments can add to the model complexity

+ All risks measured over the same time horizon, so consistent aggregation (and demonstration of diversification benefits) easier to achieve

– Can pose challenges in aggregation across different lines of business with different durations

– Market-consistent liability valuations can lead to stochastic-on-stochastic calculations; more simplistic approaches (e.g., stress testing) are less accurate and provide less insight into overall capital distribution; replicating portfolio and other advanced techniques may provide solutions

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

55

towerswatson.com

Both methods can be used, but … Europe The one-year mark-to-market approach has emerged as the standard Asia Pacific The trend is towards a one-year mark-to-market approach

North America A consensus is still forming

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

56

towerswatson.com

We will adopt a 1-year mark-to-market approach Market Value Assets

Available Economic Capital (AEC)

AEC RM BEL

Probability of outcome

Sim 1 Expected Market Value Assets

AEC RM BEL

0.5% probability

Sim 500,000

t=0

What risks am I exposed to? How might the risks I am exposed to vary?

Market Value Assets

AEC RM BEL

0.05% probability

t=1

Given the evolution of risks, how will this affect my business performance?

What level of impact am I concerned about?

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

57

towerswatson.com

Implementation – from simple methods to best practice Getting started

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

58

towerswatson.com

What risks am I exposed to? Total EC

Basic EC

Non-Life

Premium Reserve

Market

Currency

Health

SLT

Spread

Mortality

Interest rate

Longevity

Property

Expenses

Lapse

CAT

Operational

Counterparty

Life

Intangible

Non SLT

Lapse

Mortality

Premium Reserve

Expenses

Longevity

Disability

CAT

Lapse CAT

Equity

Disability

Concentration © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

59

towerswatson.com

The measure of risk needs to be defined Distribution of “profit and loss” Probability of outcome

Shaded area 0.5% of distribution 99.5% value at risk

Mean value of distribution

Profit / loss

99.5% tail-value at risk

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

60

towerswatson.com

How do I choose between VaR and Tail-VaR? Value at Risk  Is the risk measure used in Solvency II and Basel II and most rating agency models  



Tail-Value at Risk  Used by supervisors in USA, Canada and Switzerland 

Easier to implement than t-VaR



Defines the trigger point of insolvency Has a technical flaw – not a coherent risk measure – although this is not often an issue practical



Harder to implement than VaR Measures the average cost of insolvency Does not have the technical flaw faced by VaR – is a coherent risk measure

The relative simplicity of VaR has led to its emergence as the standard measure © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

61

towerswatson.com

What confidence interval shall I consider? Distribution of “profit and loss” Probability of outcome

TAIL PROBABILITIES AAA 1bp

AA 3bps

A 7bps

BBB 50bps

ECBBB = VaR 99.5%

Mean value of distribution

Profit / loss

ECA = VaR 99.93% ECAA = VaR 99.97% ECAAA = VaR 99.99% © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

62

towerswatson.com

More than one event can occur simultaneously 100% correlated

-100% correlated

Positively correlated

Negatively correlated

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

63

towerswatson.com

The approach defines capital in terms of the economic balance sheet

Free capital

Required capital

Available economic capital

Risk margin Market value of assets Market consistent value of liabilities

Time value of guarantees

Certainty equivalent liability

Discounted best estimate of liabilities

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

64

towerswatson.com

The approach poses a computational challenge: stochastic-on-stochastic modelling Available Economic Capital (AEC) Market Value Assets

AEC RM

Scenario 1

Expected Market Value Assets

AEC

Probability of outcome

BEL

RM BEL

Scenario 100,000

Market Value Assets

AEC RM BEL

T=0

Real world

T=1

Marketconsistent © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

65

towerswatson.com

Many life insurers implement the 1-year market-consistent approach by directly measuring the balance sheet at the chosen confidence level Normal Conditions

Stressed Conditions

AEC AEC MV Assets

RM

MV Assets

BEL

RM BEL

Required EC AEC

Normal

AEC

Stressed

Referred to as the “stress test correlation matrix” approach © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

66

towerswatson.com

Stress test correlation matrix approach 

Aggregate capital =

 i



 j ij C i C j

Where Ci is the capital required for the ith risk and ij is the correlation coefficient between ith and jth risk

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

67

towerswatson.com

Example: Suppose the following impacts of stresses are measured Risk factor

Stressed value 616

Equity

7

Interest rate Mortality

775

Lapse

639

Total before diversification

2,037

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

68

towerswatson.com

Example: Assume the following correlation matrix

Equity

Interest Rate

Mortality

Lapse

Equity

1

0.75

0.00

0.25

Interest Rate

0.75

1

0.00

0.50

Mortality

0.00

0.00

1

0.00

Lapse

0.25

0.50

0.00

1

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

69

towerswatson.com

Example: After diversification Risk factor

Stressed value 616

Equity

7

Interest rate Mortality

775

Lapse

639

Total before diversification

2,037 (698)

Diversification Total after diversification

1,339

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

70

towerswatson.com

Implementation – from simple methods to best practice Advanced methods

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

71

towerswatson.com

The stress test correlation matrix approach has proved useful, but firms have started to understand its limitations Area

Stress test correlation  assumption

Issue

Risks

Normally distributed

Most risks are fat‐tailed

Dependency

Correlation only captures linear  dependency

Risk dependency structures vary when moving from benign to  extreme conditions – “markets crash together”

Balance sheet  response (1)

Assumes linear response to variations  Balance sheets behave in a non‐linear way – e.g. embedded  in risk  financial options and guarantees vary non‐linearly with  underlying market exposures

Balance sheet  response (2)

Assumes separable response to  variations in risk

Response to risks is not separable – e.g. annuity response to  combined mortality and interest rate stress is not the sum of  the two individual stresses

Fungibility of own  funds and tax

Adjusted for manually outside the  model if at all

Legal or regulatory restrictions prevent losses in some legal  entities being supported by assets from other legal entities

Capital allocation

Not directly possible, requires ad hoc  methods

Embedded and active risk management requires understanding  of risk, capital consumption and value creation at the level  decisions are made 

Full distribution /  other percentiles

Not possible without re‐run of  balance sheet at target percentile

Capital at the 99.5th percentile is only one data point – best  practice risk management requires understanding of impact of  decisions across the risk and outcome spectrum

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

72

towerswatson.com

Best practice firms are revisiting the general methodology and addressing the stochastic-on-stochastic modelling challenge Available Economic Capital (AEC) Market Value Assets

AEC RM

Scenario 1

Expected Market Value Assets

AEC

Probability of outcome

BEL

RM BEL

Scenario 100,000

Market Value Assets

AEC RM BEL

T=0

Real world

T=1

Marketconsistent © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

73

towerswatson.com

Three approaches to modelling the 1 year market consistent approach have emerged

Stress test correlation matrix approach

1 year VaR

Delivery certainty

Simulation modelling of market risk, simplified modelling of insurance risk

Simulation modelling of all risks

Information gained

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

74

towerswatson.com

This complexity has been introduced to address hidden limitations with the traditional “stress test correlation matrix” approach Area

Stress test correlation  assumption

General simulation methodology

Risks

Normally distributed

Allows risks to be modelled using any marginal distribution

Dependency

Correlation only captures linear  dependency

Works with complex dependency structures – e.g. copula  functions

Balance sheet  response (1)

Assumes linear response to variations  Use of replicating models allows for full non‐linearity of the  in risk  balance sheet to be reflected in the calculations

Balance sheet  response (2)

Assumes separable response to  variations in risk

Use of replicating models allows for full non‐separability of risk  to be reflected in the calculations

Fungibility of own  funds and tax

Adjusted for manually outside the  model if at all

Legal or regulatory restrictions can be modelled directly and  reflected as balance sheet are produced for each risk outcome

Capital allocation

Not directly possible, requires ad hoc  methods

Enables capital to be allocated to any level modelled in the  simulation

Full distribution /  other percentiles

Not possible without re‐run of  balance sheet at target percentile

Automatically generates the full distribution forecast

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

75

towerswatson.com

Looking at each component in turn Risk factor Risk marginal marginal distributions distributions

Risk factor Risk dependency dependency structure structure

Balance sheet valuation loss functions

Calculation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

76

towerswatson.com

Some terminology…

Risk Factors Factors that contribute to either loss frequency or loss severity Loss Functions The impacts of the changing risk factors on the balance sheet

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

77

towerswatson.com

Risk factor marginal distributions

Risk factor dependency structure

Balance sheet valuation loss functions

Calculation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

78

towerswatson.com

Risks are not normally distributed

The heavy tailed distribution has a greater probability of very small and very large events The heavy tailed distribution has a lower probability of medium sized events

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

79

towerswatson.com

Distribution of extreme events Probability

Fitting heavy tailed distribution

50% Lots of data

75%

95%

98%

percentile 99.8%

Less data Some data Little data Few events © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

80

towerswatson.com

Experience shock versus allowing for long term change Experience shock

Volatility around long term trend

Long term change e.g. new disease

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

81

towerswatson.com

EX ER C IS E

Exercise 2: Fitting risk distributions

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

82

towerswatson.com

EX ER C IS

Exposed to equity: how does equity evolve?

E © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

83

towerswatson.com

EX ER C IS

Fitting a distribution…but which should we choose?

E © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

84

towerswatson.com

EX ER C IS

Assessing a risk factor goodness-of-fit

E © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

85

towerswatson.com

EX ER C

Risk factors - exercise

IS E



Which distribution do you choose? Exercise 2 : Risk factor calibration Task You are leading a team calibrating an economic capital model for a medium-sized Life company. You need to recommend a distribution to be used for the equity risk factor . To help with your decision, a colleague has prepared the following table of equity stresses. In addition, another colleague has fit four candidate distributions to historical data. Histograms and QQ -plots of these fits can be found in the appendix. Please consider the table and the appendix and make a recommen dation below. Table 1: Equity risk stresses Source

Stress

Regulator X’s regulatory capital assessment stress

-40%

Stress for equities in “killer scenario”

-35%

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

86

towerswatson.com

EX ER C

Risk factors - exercise

IS E



Which distribution do you choose?

1

Recommendation I recommend the use of the Normal

(please tick one box)

Student’s T Pearson IV Variance-Gamma

distribution for modelling equity risk in the economic capital model.

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

87

towerswatson.com

EX ER C IS

Risk factors - exercise

E © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

88

towerswatson.com

Risk factor marginal distributions

Risk factor dependency structure

Balance sheet valuation loss functions

Calculation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

89

towerswatson.com

Dependency should allow for different behaviour in the tail United Kingdom

Malaysia

6,000

800

1,200

Hong Kong

Thailand

25,000 750

1,150 5,500

700

4,500

650 1,050

20,000 1,000

950

900 4,000

17,500

600

550

500

450

15,000

850

3,500

SET Index Level

5,000

22,500

Hang Seng Index Level

Kuala Lumpur Composite Index Level

FTSE 100 Index Level

1,100

400

800

350

12,500

Date

Date

Date

10,000

Taiwan

Hong Kong

China

8,000

25,000

3,000

7,500

22,500

2,750

Date

In benign markets, equity market fluctuations will not have a big impact on lapses, but, when markets fall, there is usually a trigger point that gives rise to mass lapses … 6,500

17,500

6,000

15,000

5,500

12,500

5,000

10,000

Date

20%

15%

10%

5%

Date

0%

SEE Composite Index Level

20,000

Hang Seng Index Level

TWSE Index Level

7,000

25% 2,500

2,250

2,000

1,750

1,500

Date

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

90

towerswatson.com

Risk factor marginal distributions

Risk factor dependency structure

Balance sheet valuation loss functions

Calculation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

91

towerswatson.com

Use non-linear balance sheet models or “loss functions” 

A 30% fall is not necessarily ¾ of a 40% fall 1,200

Losses

1,000 800 600 400 200 0 0%

‐10%

‐20%

‐30%

‐40%

‐50%

‐60%

Equity returns

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

92

towerswatson.com

Use non-linear balance sheet models or “loss functions” 

A 30% fall is not necessarily ¾ of a 40% fall Linear Non-linear

1,200

Losses

1,000

10% overestimation

800 600 400 200

10% underestimation

0 0%

‐10%

‐20%

‐30%

‐40%

‐50%

‐60%

Equity returns

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

93

towerswatson.com

Replicating models – loss functions 

A loss function is an interpretation of the values that would be derived if all 100,000 balance sheet runs were performed Value of Balance Sheet Loss

Values estimated by loss function

Actual values from actuarial cashflow model runs

Risk factor value

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

94

towerswatson.com

The impact of two risks occurring at the same time is different from the sum of the individual impacts

Losses from mortality and interest rate movements Separability assumption

Actual joint stresses

Change in mortality

Change in mortality

-20% 260

-15%

-10%

-5%

0 0

-20% 0 -25 -50

100

-10%

-5%

0 0

Interest rate movement (bps)

0 -25 -50

Interest rate movement (bps)

-75

-75 360

-15%

440

-100

-100

18% underestimation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

95

towerswatson.com

Parametric loss functions can be fitted to stress test results from projection models Change in mortality

Change in interest rates

-10%

-£120m

-20%

-20%

Balance sheet movement non-linearity

-£260m +0.5%

+£40m

-1%

-£100m

-1%

-£440m

non-separability (multivariate non-linearity)

non-linearity

Balance sheet movement = -1,000 * Δmort2 + 1,100 * Δmort - 133,333 * Δi-rate2 + 8,667 * Δi-rate - 40,000 * Δmort * Δi-rate © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

96

towerswatson.com

EX ER C IS E

Exercise 3: Fitting a loss function

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

97

towerswatson.com

EX ER C

The Loss function Game!

IS E

The Loss Function Game! Play the game with your neighbour.

PLAYER A

The Loss Function Game! Play the game with your neighbour.

Each of you has to fit loss functions to three different risks. Your neighbour has the complete loss function for your risks. Unfortunately, you only have 25 hours of computer time and each run takes 5 hours, so you can only ask for 5 results.

Each of you has to fit loss functions to three different risks. Your neighbour has the complete loss function for your risks. Unfortunately, you only have 25 hours of computer time and each run takes 5 hours, so you can only ask for 5 results.

Fill in the run request form below and ask your neighbour to supply you with the results. Plot the results on the charts and sketch your estimated loss functions.

Fill in the run request form below and ask your neighbour to supply you with the results. Plot the results on the charts and sketch your estimated loss functions.

Meet your neighbour

Stress

Result

Stress

Result

…………

…………

Run 1:

…………

…………

Run A:

Run 2:

…………

…………

Run B:

…………

…………

Run 3:

…………

…………

Run C:

…………

…………

Run 4:

…………

…………

Run D:

…………

Run 5:

…………

…………

Run E:

…………

PLAYER B

Don’t cheat! ………… …………

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

98

towerswatson.com

EX ER C

The Loss function Game!

IS E

Loss with respect to spread stress

300

200

Loss function value

100

0

-100

-200

-300 -0.4

-0.2

0

0.2

0.4

Spread stress

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

99

towerswatson.com

EX ER C

The Loss Function Game! – How to play

IS E

PLAYER A

PLAYER B

Turn to Page 1

Turn to Page 2

Take 2 minutes to consider the 3 risks

Player A will ask you for some results from your graphs

Request 5 different runs from Player B. Write the stresses in the spaces on side 1 of your sheet.

Estimate the loss function results from your graph and given them to Player A

Sketch the loss functions you think best fit the 3 risks. Turn to Page 2, swap roles and repeat!

Turn to Page 1, swap roles and repeat! © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

100

towerswatson.com

EX ER C IS

The Loss Function Game!

E © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

101

towerswatson.com

Risk factor marginal distributions

Risk factor dependency structure

Balance sheet valuation loss functions

Calculation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

102

towerswatson.com

A simulation approach models capital by business unit and risk factor Capital by Business Unit

Capital by Risk Factor

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

103

towerswatson.com

Using this modelling approach, insurers can understand risk drivers and monitor against a risk appetite Split of Economic Capital

Counterparty risk Operational risk

Insurance risk

Diversification effect Equity Real Estate

ALM -Risk

Interest Rate

Business Unit Economic Capital

Risk Category Interest rates

BU1

BU2

BU3

BU4

BU5

Other

Total

535

397

170

296

985

465

1,322 1,322

98

448

320

393

71

50

689 689

Other market risk

242

341

382

625

697

658

1,277 1,277

Insurance

591

664

284

258

755

293

1,262 1,262

Operational

131

92

78

827

554

171

1,026 1,026

12

57

8

42

18

25

849

963

603

Equity

Counterparty Total

1,177 1,177

1,529 1,529

876 876

79 2,550 2,550

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

104

towerswatson.com

Implementation – historic trends

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

105

towerswatson.com

Objectives 

Results from Towers Watson’s global ERM survey



How economic capital methodology has changed over time…



…and how it varies currently

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

106

towerswatson.com

The Towers Watson global ERM survey 





Biennial survey of enterprise risk management (ERM) in the insurance sector Largest and longest standing survey for the insurance industry 

First survey run in 2000



465 participants globally in 2010, 99 from Asia Pacific

Covers multiple areas of risk management and measurement, including 

Performance, priorities and challenges



Governance and organisation



Risk appetite, limits and reporting



Economic capital methodology



Solvency II (for European headquartered participants)

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

107

towerswatson.com

The proportion of respondents calculating economic capital globally has increased from 57% in 2008 to 65% in 2010

2010

Global

65%

2008

25%

57%

2006

28%

65%

2010

10%

15%

19%

59%

16%

32%

9%

Asia Pacific ex Japan 2008

53%

Currently calculate

24%

Plan to calculate

24%

Do not calculate and no plans

Towers Watson global ERM survey 2006, 2008 and 2010. Does your organisation calculate economic capital? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

108

towerswatson.com

Globally, one year approaches are most favoured… … alternative risk assessment periods are most often used by P&C insurers

Global

2010

68%

2008

68%

2006

15%

13%

56%

2010

14%

12%

14%

14%

3%

5%

18%

81%

14%

5%

Asia Pacific ex Japan 2008

60%

1 year

2 to 5 years

10%

Run-off

25%

5%

Other

Towers Watson global ERM survey 2006, 2008 and 2010. Over what period do you assess risk in calculating economic capital? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

109

towerswatson.com

Companies that use a fixed time period to calculate economic capital most commonly use a market-consistent terminal balance sheet Global

Asia Pacific ex Japan

Market-consistent balance sheet (mark to market or fair value)

57% 56%

24%

17%

Regulatory balance sheet

60% 60%

GAAP or IFRS balance sheet

23%

GAAP or IFRS balance sheet

Market-consistent balance sheet (mark to market or fair value)

14% 13%

26%

Regulatory balance sheet

27%

16%

Other

3%

Other

0% 0%

4%

2010

2008

Towers Watson global ERM survey 2008 and 2010. If you use a fixed time period to calculate economic capital, what basis do you use for the terminal balance sheet? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

110

towerswatson.com

The proportion of insurers using the VaR measure is higher in Asia Pacific (ex Japan) than globally Global

Asia Pacific ex Japan 72%

Value at risk (VaR) or risk of ruin (RoR)

67%

84%

Value at risk (VaR) or risk of ruin (RoR)

70%

39%

Tail value at risk (TVaR) or conditional tail expectation (CTE)

Economic cost of ruin (ECoR)

Tail value at risk (TVaR) or conditional tail expectation (CTE)

20% 21% 37%

4% Economic cost of ruin (ECoR)

5% 11%

11% 20%

3% 0%

4% Other

Other

6% 14%

2010

2008

3% 10%

2006

Towers Watson global ERM survey 2006, 2008 and 2010. What primary measure of risk tolerance do you use to calculate economic capital? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

111

towerswatson.com

More than 50% of the respondents to the 2010 survey use simple correlation aggregation approaches Global

Asia Pacific ex Japan 54%

Correlation matrix

55% 56%

Simulation-based aggregation and structural models

None

32%

57%

Correlation matrix

50%

Simulation-based aggregation and structural models

32% 26%

9% 6%

22% 30%

None

11% 10%

Other

11% 10%

10%

Other

5% 7% 7%

2010

2008

2006

Towers Watson global ERM survey 2008 and 2010. What methodology do you use for aggregating risk at the corporate level? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

112

towerswatson.com

In summary

Standards for economic capital calculation have emerged over time… …but approaches still vary by region and by type and size of company

There is no single “right” answer!

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

113

towerswatson.com

Implementation – how you can get there

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

114

towerswatson.com

Objectives 

An example of a practical, phased approach to implementation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

115

towerswatson.com

Hypothetically … what are we aiming for? Key design decisions1 Time horizon

1 year

Measure of risk

Value at risk (VaR)

Risks included

Market, insurance, operational

Level of confidence to target

Complete distribution

1. Key design decisions for time horizon and measure of risk for hypothetical requirements based on most common responses made by Asia Pacific companies to relevant questions in Towers Watson global ERM survey 2010.

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

116

towerswatson.com

…therefore we need… 

Stochastic asset liability models to calculate the market consistent balance sheet



Risk models for all underlying risks



Data to support calibration and parameterisation of risk models





Modelling and aggregation techniques that allow for interactions across and between risk categories with non-linearity and non-separability Models that run quickly enough to produce results at all points on the distribution

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

117

towerswatson.com

…but we have… 



Deterministic liability cashflow models 

Project liabilities on best estimate assumptions



Investment return assumptions do not vary by period



Single risk discount rate

Assets values determined according to current GAAP requirements 

 



But (usually) easily convertible to market value

A simple asset model capable of re-valuing assets under stressed conditions The knowledge that the three key risks underlying our business are interest rate risk, lapse risk and mortality risk A reasonable amount of data (and the relevant skills) that allows us to develop (allowing for judgement where relevant) 

Probability distributions underlying these three risks



A simple correlation matrix of dependencies

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

118

towerswatson.com

What might a sensible development path look like?

Getting from what we have now… …to what we need… …to support our aims An example approach

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

119

towerswatson.com

Using a phased implementation approach – Step 1

Step 1: Certainty equivalent value of liabilities

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

120

towerswatson.com

Using a phased implementation approach – Step 2

Step 1: Certainty equivalent value of liabilities

Step 2: Market value of assets and liabilities

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

121

towerswatson.com

Using a phased implementation approach – Step 3

Step 1: Certainty equivalent value of liabilities

Step 2: Market value of assets and liabilities

Step 3: Extend risks to cover other insurance and market risks

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

122

towerswatson.com

Using a phased implementation approach – Step 4

Step 1: Certainty equivalent value of liabilities

Step 2: Market value of assets and liabilities

Step 3: Extend risks to cover other insurance and market risks

Step 4: Allow for non-separability on an approximate basis

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

123

towerswatson.com

Using a phased implementation approach – Step 5

Step 1: Certainty equivalent value of liabilities

Step 2: Market value of assets and liabilities

Step 5: Extend risks to cover operational risks

Step 3: Extend risks to cover other insurance and market risks

Step 4: Allow for non-separability on an approximate basis

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

124

towerswatson.com

Using a phased implementation approach – Step 6

Step 1: Certainty equivalent value of liabilities

Step 6: Employ advanced modelling techniques

Step 2: Market value of assets and liabilities

Step 5: Extend risks to cover operational risks

Step 3: Extend risks to cover other insurance and market risks

Step 4: Allow for non-separability on an approximate basis

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

125

towerswatson.com

Implementation – operational issues

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

126

towerswatson.com

Objectives 

Model and process governance arrangements



Achieving buy-in



Implementation constraints

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

127

towerswatson.com

Implementation – operational issues Operationalising the system

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

128

towerswatson.com

Making the system operational Define overall strategy  Provide guidance on use of economic capital  Ensure on-going appropriateness of methodology, design etc 

Business owner

Development team

Production team

Implement the strategy  Develop supporting models and reporting  Validate and document system 

Perform calculations  Analyse and report results 

Use results to manage the business  Input into methodology and design, challenge the approach  Use results to monitor risk taking activity 

Users

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

129

towerswatson.com

Senior management support is crucial

Economic capital is used as a tool to help management make better decisions

Senior management need to be actively involved in defining requirements

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

130

towerswatson.com

Reliable results are essential if they are to be used to manage the business Testing / Validation

Initial and on-going  Feed results back into the methodology/model development cycle  Not just around the mean – test results/validity of model across the distribution 

Business requirements – underlying principles of model, purpose of model, limitations  Detailed specification – could be used to re-build a consistent model  Testing and validation – testing completed and reasonableness of results  User manual  Guide for senior management 

Documentation / Explanation

Accurate, complete and appropriate underlying data  Internal / external loss data, market data, in-force policy data  Expert judgement

Data / Parameterisation



Systems and processes



Stable, secure, controlled and auditable system  Streamlined, systemised and automated processes  Reproducible results

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

131

towerswatson.com

A continuing cycle of improvement Known issues / testing and validation

Best practice methodology and approaches Model change programme Business uses

New / improved data sources © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

132

towerswatson.com

Implementation – operational issues Achieving buy-in

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

133

towerswatson.com

Use top-down and bottom-up communication to secure internal buy-in 

Top-down communication from senior management plus tailored training programmes 



Rationale for the decision to implement economic capital



The overall economic capital methodology



Short-term and long-term implementation plans



Modelling requirements



Using the results in decision making



The impact on processes

Bottom-up communication gives business managers the opportunity to react and provide feedback on top-down communication



(Information + Experience) x Perception = Behavioural Change © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

134

towerswatson.com

How much and when to disclose?

Upside of announcing strong enterprise risk management and economic capital implementation

versus Downside of making disclosures too early in the process

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

135

towerswatson.com

Implementation – operational issues Implementation constraints

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

136

towerswatson.com

What are the biggest challenges companies face? What are your biggest challenges in implementing or using economic capital? 49%

Availability of resources

56% 44% 44%

Approach / methodology

44%

Data (e.g. lack of, integrity)

39% 35%

Technology (e.g. run times)

30% 33%

Management buy-in

44% 21%

Relevancy (if regulation N/A) Other

31% 5% 3%

Global

Asia Pacific (ex Japan)

Towers Watson global ERM survey 2010 Q.34. (Tick all that apply.) © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

137

towerswatson.com

Do you have the right skill set? 

New concepts, new methodologies and new tools 

Significant one-off effort… — Initial development requirements – methodology, model, validation, documentation — New processes — Training



…plus on-going needs — On-going maintenance – methodology, model, validation, documentation,

processes — On-going training requirements — Production and analysis of results — Use in decision making 

Import or grow the required skills? © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

138

towerswatson.com

Do you have the systems you need?  







What do you have now…

…and what do you want from your system?

Are current systems adaptable to meet new and on-going requirements? Is there an opportunity to radically change the way actuarial reporting and processing systems are operated? Internally or externally developed software solutions?

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

139

towerswatson.com

What’s your timeframe and budget for implementation? 



Time and financial constraints are significant factors in implementation…

…and expected time and cost can depend on many factors



Size, diversity and complexity of your business



Nature of financial models already in place



New hardware and software requirements



Complexity of the proposed methodology being implemented



Phasing of implementation



Availability of data



Use of internal and external resources © 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

140

towerswatson.com

In summary…

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

141

towerswatson.com

A successful economic implementation…

…is a journey …is not just about the numbers …can lead to better informed decisions… …and aid value creation

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

142

towerswatson.com

Disclaimer This presentation has been prepared for general information purposes only and does not constitute professional advice. The information, opinions and illustrations contained here-in are derived from various sources and have not necessarily been independently verified. Views expressed are those of the presenters, and do not necessarily represent the views of any particular company.

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

143

towerswatson.com