The business case for economic capital modelling for life ......The business case for economic capital modelling for life insurers ... (EC) modelling for life insurers: ... which include

2013 Convention 31 Oct & 1 Nov

The business case for economic capital

modelling for life insurers

Matthew Brinckmann

[email protected]

mailto:[email protected]


Agenda

1. Introduction

2. Real benefits

3. What to get right

4. New solutions for a new world

5. A European detour

6. Conclusions

2


A business case

• Business case = cost benefit analysis

• This presentation will explore the business case for

economic capital (EC) modelling for life insurers:

− Benefits?

− Costs and complexities?

− How do you achieve the right balance?

• Based on practical experience with economic

capital models over a number of years.

3


What happens if you get this wrong?

Source: Difficult risks and Capital Models paper

Firm Stated 1 year VaR

(YE 2007)

Frequency Loss during

2008

Extrapolated

frequency

(Normal)

AIG $14.5bn-$19.5bn 1 in 2,000

years $99.3bn 1 in 4 x 1062 years

Fortis €17.6bn 1 in 3,333

years €28bn

1 in 40,000,000

years

4


What happens if you get this wrong?

• Exceptionally bad luck?

• The big issues were:

− Too much breadth

− Insufficient depth

5


Agenda

1. Introduction

2. Real benefits




6. Conclusions

6


Why do insurers calculate EC?

7

In order to give the great security the trade of

insurance provides, is necessary that the

insurers should have a large capital

- Adam Smith


Real business benefits

8

“With the continuing development of the group’s economic capital modelling capability, the

economic capital requirement is increasingly being used in the group’s risk and capital

management processes.”

Liberty Life annual report 2012

“Use of EC models helps inform strategy and support decision making to maximise return on

shareholder capital while protecting policyholders”

Aviva EC presentation 2011

“How to optimise value creation?

3 levers to monitor and improve value creation:

• Maximize absolute value

• Optimize capital requirement constraints

• Manage volatility over time

Our principal internal tool to act on these three dimensions:

• Return on economic capital”

Axa investor day, November 2010

Wh

at

co

mp

an

ies

ha

ve

to

sa

y…


Real business benefits

9

“Economic capital – why?:

• Enables the development of consistent risk metrics across financial services conglomerates

which include banking, life and non-life insurance and asset management.

• Provides a transparent view of risks across an organisation and allows for correlations between and across risks.

• Enables effective testing of business initiatives from a risk management perspective.

• Needs to feed into incentive arrangements so that risk based decisions are promoted. “

Old Mutual, analyst roundtable – EC and other metrics, June 2013

“SCOR strengthens its solvency governance by creating a sophisticated dynamic solvency

target based on a gradual escalation process:

• The solvency target for the new strategic plan complements the existing threshold capital

with an escalation process depending on the level of available capital

• The optimal capital range enables the Group to achieve maximum profitability and satisfy

the level of solvency which SCOR targets to offer its clients

Optimal Dynamics reflects SCOR’s search for the sweet spot between solvency, profitability

and growth, given its risk appetite”

Scor optimal dynamics presentation, September 2013

Wh

at

co

mp

an

ies

ha

ve

to

sa

y…


Own risk and solvency assessment

Even if you haven’t had an EC model for internal purposes, the ORSA

probably now requires you to have one.

10

Source: FSB position paper 34 (v7) extracts

It is therefore likely that most insurers will use internal models to calculate economic

capital for ORSA purposes.

Consider all material current and

foreseeable risks

Assess the adequacy and quality of EC

over a longer time horizon than used to

determine regulatory capital

The results of the ORSA should be used

to inform and improve business

decisions, business strategy and the ERM

framework

Other risks relating to insurance groups

Other risks not considered adequately

covered by SF

Adequacy of current and future internal

(economic) capital resources given its

risk appetite and tolerance levels

OR

SA


Agenda

1. Introduction

2. Real benefits




6. Conclusions

11


Keep it simple

12


Keep it simple

13

“I think you should be more explicit here in step two”


Some things to get right

Focus on the

key risks

Understand

the influence

of expert

judgement

14

George Box: All models are wrong, but some are useful

Manage the

model risk Model must

be used


Some things to get right

15


Agenda

1. Introduction

2. Real benefits




6. Conclusions

16


Purpose

Identify

(and rank)

risks

Implementation

Risk metrics Modelling approach

Modelling constraints

A process with key decisions required

17

Validate the model


Typical Economic Capital definition

Paradigm is often 1 year value at risk

18

99.95%

99.50% 99.50%

99.95% 99.93%

99.20%

99.30%

99.40%

99.50%

99.60%

99.70%

99.80%

99.90%

100.00%

Aviva AXA RSA ING Old Mutual

Confidence level

*

* Aviva aim to cover policyholder liabilities following a 1 in 200 event, followed by a further 1 in 10

event


Why not just use a regulatory or

ratings agency model?

• Different purposes

• Non-economic elements

• Don’t allow for all risks

• What if you operate across regulatory regimes?

• Until you calculate EC, you cannot tell if it will be

more or less onerous than regulatory capital.

19


Why not just follow the standard

formula approach?

• Adjust standard formula stresses and correlations?

• Established and widespread approach

• Simple to apply

• Sensible starting point

• Makes a number of strong assumptions, so:

• Potentially material weaknesses, so:

– be careful of misleading risk management

information.

20


Improving the approach

• Adjust the correlation matrix results using scenario

testing

• Full multivariate simulation

• Multivariate simulation using balance sheet proxies

• What additional risk management information does

a more sophisticated model provide?

21


A simple concept

22

Asset & liability re-valuation

Individual risk driver distributions

Joint risk driver

distribution

Copula

Aggregated loss distribution

99.5th percentile

50th percentile

Capital requirement

Base Scenario 1

Scenario 2


New (and not so new) solutions exist

Possible solutions are around proxy

methods, e.g. formula fitting, replicating

portfolios

The key remaining challenge is

really a practical one around

asset and liability re-valuation

23

Tried and tested solutions and

approaches exist, for example around

distribution fitting, dependency structures,

capital attribution approaches etc

There are a number of technical

challenges


Actuaries in animation?

24


Formula fitting

-1 000

0

1 000

2 000

3 000

4 000

5 000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

BEL

Run number

Actual vs Fitted BEL

Fitted BEL

Actual BEL

25

-40 000

-35 000

-30 000

-25 000

-20 000

-15 000

-10 000

-5 000

-

5 000

10 000

1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341

BEL

Run number

Actual vs Fitted BEL (Out of sample)

Fitted BEL

Actual BEL


Agenda

1. Introduction

2. Real benefits




6. Conclusions

26


Internal models European contrast

• Life insurers in South Africa typically currently not

seeking regulatory approved internal models

• 2012 EY European survey suggested half of

respondents are developing internal models

• Significant focus on regulatory compliance in

Europe, although costly Solvency II programmes

have been trimmed

• Survey suggests that insurers with value-based

management expect increase in return on

economic capital.

27



Source: EY survey

0% 10% 20% 30% 40% 50% 60% 70% 80%

Market risks

Underwriting risks/life & health

Underwriting risks/ non-life

Credit risks

Operational risks

Other risks

Aggregation

Standard formula compared to risk situation

Don't know About right Too high Too low

28



Source: EY survey

1%

5%

17%

37%

26%

14%

0%

5%

10%

15%

20%

25%

30%

35%

40%

Increase > 20% Increase 10%-20% Increase/Decrease <

10%

Decrease 10%-20% Decrease 20%-30% Decrease >30%

Change in SCR due to internal model

29


Solvency II and S&P – the first kiss may be some time…

Convergence of capital frameworks is badly needed

Source: Scor presentation ‘Scor focuses on optimising capital’

30


Agenda

1. Introduction

2. Real benefits




6. Conclusions

31


Conclusions

• A business case is usually argued in terms of a

cost benefit analysis

• It is not good enough to just have an EC model

• You need to know what to get right to ensure it

can be used effectively.

S&P May 2012

32

Documents

The business case for economic capital modelling for life ......The business case for economic capital modelling for life insurers ... (EC) modelling for life insurers: ... which include