Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
2013 Convention 31 Oct & 1 Nov
The business case for economic capital
modelling for life insurers
Matthew Brinckmann
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
What happens if you get this wrong?
• Exceptionally bad luck?
• The big issues were:
− Too much breadth
− Insufficient depth
5
2013 Convention 31 Oct & 1 Nov
Agenda
1. Introduction
2. Real benefits
3. What to get right
4. New solutions for a new world
5. A European detour
6. Conclusions
6
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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…
2013 Convention 31 Oct & 1 Nov
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…
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
Agenda
1. Introduction
2. Real benefits
3. What to get right
4. New solutions for a new world
5. A European detour
6. Conclusions
11
2013 Convention 31 Oct & 1 Nov
Keep it simple
12
2013 Convention 31 Oct & 1 Nov
Keep it simple
13
“I think you should be more explicit here in step two”
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
Some things to get right
15
2013 Convention 31 Oct & 1 Nov
Agenda
1. Introduction
2. Real benefits
3. What to get right
4. New solutions for a new world
5. A European detour
6. Conclusions
16
2013 Convention 31 Oct & 1 Nov
Purpose
Identify
(and rank)
risks
Implementation
Risk metrics Modelling approach
Modelling constraints
A process with key decisions required
17
Validate the model
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
Actuaries in animation?
24
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
Agenda
1. Introduction
2. Real benefits
3. What to get right
4. New solutions for a new world
5. A European detour
6. Conclusions
26
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
Internal models European contrast
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
2013 Convention 31 Oct & 1 Nov
Internal models European contrast
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
2013 Convention 31 Oct & 1 Nov
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
2013 Convention 31 Oct & 1 Nov
Agenda
1. Introduction
2. Real benefits
3. What to get right
4. New solutions for a new world
5. A European detour
6. Conclusions
31
2013 Convention 31 Oct & 1 Nov
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