Enterprise Risk Management (ERM) A risk-based approach to ... - Scor
Enterprise Risk Management (ERM) A risk-based approach to ... - Scor
Enterprise Risk Management (ERM) A risk-based approach to ... - Scor
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Focus<br />
November 2010<br />
<strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong> (<strong>ERM</strong>)<br />
A <strong>risk</strong>-<strong>based</strong> <strong>approach</strong><br />
<strong>to</strong> the management<br />
of a (re)insurance company
ENTERPRISE RISK MANAGEMENT
Summary<br />
Preface 3<br />
1 Implementation of an <strong>ERM</strong> framework 4<br />
2 Adapting solvency regulations <strong>to</strong> times of crisis 16<br />
3 Reinsurance optimization in the context<br />
of capital management 24<br />
4 Capital assessment beyond s<strong>to</strong>chastic modeling 36<br />
5 Integrating <strong>risk</strong> modeling in<strong>to</strong> the organization 44<br />
6 Investment and <strong>Risk</strong> <strong>Management</strong> strategies<br />
in a changing regula<strong>to</strong>ry framework 52<br />
7 <strong>ERM</strong> and Economic Capital Models: The A.M. Best view 64<br />
8 Macro-economic standpoint 72<br />
9 How <strong>to</strong> deal with the impact of inflation<br />
on pricing and reserving 80<br />
10 The financial markets view on the <strong>Risk</strong>/Reward strategies<br />
in the (re)insurance industry 90<br />
<strong>ERM</strong>: a system <strong>to</strong> ensure the optimal security and efficiency 98<br />
Speakers’ biographies 101
2 - November 2010 - SCOR<br />
The views and statements expressed<br />
in this publication are the sole<br />
responsibility of the authors.
Preface<br />
In the current s<strong>to</strong>chastic, financial and economic environment,<br />
which is characterized by shocks and instability, and in view of<br />
impending regula<strong>to</strong>ry changes such as Solvency II and its equivalents<br />
under consideration outside the EU, <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong><br />
or <strong>ERM</strong> has become a <strong>to</strong>p priority for the (re)insurance industry.<br />
Following SCOR’s first <strong>ERM</strong> seminar last year, which presented the<br />
main concepts involved in <strong>ERM</strong> and how they could be a driving<br />
force for our industry, the Group held a second seminar in June.<br />
The themes discussed this time around <strong>to</strong>ok matters a step further<br />
than last year, focusing on how <strong>to</strong> implement an <strong>ERM</strong> framework,<br />
on integrating <strong>risk</strong> and economic capital management in<strong>to</strong><br />
a company, and on how <strong>ERM</strong> can contribute <strong>to</strong> the decision making<br />
process, for example in terms of reinsurance optimization.<br />
The seminar also looked at why we need <strong>to</strong> adapt solvency<br />
regulation <strong>to</strong> times of crisis and at possible investment and <strong>risk</strong><br />
management strategies in a changing regula<strong>to</strong>ry framework.<br />
More generally, the June seminar revisited the macroeconomic<br />
environment and the best way for <strong>risk</strong> takers <strong>to</strong> cope with it<br />
from an <strong>ERM</strong> standpoint, and presented the financial market view<br />
of the <strong>risk</strong> reward strategies currently being applied in the<br />
(re)insurance sec<strong>to</strong>r.<br />
SCOR Global P&C hopes that this publication, which is <strong>based</strong> on<br />
the presentations delivered and the discussions held during the<br />
June seminar, will help <strong>to</strong> promote <strong>ERM</strong> culture in our industry.<br />
HEDI HACHICHA<br />
Head of Strategy and Development<br />
SCOR Global P&C
1<br />
IMPLEMENTATION<br />
OF AN <strong>ERM</strong><br />
FRAMEWORK<br />
This article presents the concept<br />
and objectives of <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong><br />
(<strong>ERM</strong>). Often <strong>ERM</strong> is presented in abstract terms,<br />
with many definitions and few concrete applications.<br />
However, this article is intended <strong>to</strong> illustrate<br />
the meaning of <strong>ERM</strong> through SCOR’s day <strong>to</strong> day<br />
experience in this field and <strong>to</strong> describe how <strong>Risk</strong><br />
<strong>Management</strong> works in practice.<br />
Fig. 1: Building blocks of <strong>ERM</strong><br />
<strong>Risk</strong> and<br />
Economic<br />
Capital<br />
Modeling<br />
Standard and Poor’s way of describing <strong>ERM</strong> has been<br />
herein adopted. The corresponding graphical presentation<br />
is a Greek temple as shown in Figure 1. The base<br />
represents <strong>Risk</strong> Culture. Built on this foundation are<br />
3 specific pillars:<br />
• <strong>Risk</strong> and Economic Capital Modeling <strong>to</strong> deal with<br />
present and upcoming <strong>risk</strong>s;<br />
4 - November 2010 - SCOR<br />
MICHEL DACOROGNA<br />
Deputy Chief <strong>Risk</strong> Officer, SCOR<br />
WAYNE RATCLIFFE<br />
Direc<strong>to</strong>r Group <strong>Risk</strong> <strong>Management</strong>, SCOR<br />
Strategic <strong>Risk</strong> <strong>Management</strong><br />
Emerging<br />
<strong>Risk</strong><br />
<strong>Management</strong><br />
<strong>Risk</strong> <strong>Management</strong> Culture<br />
Before going in<strong>to</strong> details, it is important <strong>to</strong> note<br />
that <strong>risk</strong> constitutes the core business of insurers<br />
and reinsurers and gives them the opportunity <strong>to</strong><br />
provide insured people with protection against<br />
<strong>risk</strong>s <strong>to</strong> which they are exposed. Clearly, <strong>risk</strong> in this<br />
context should not be considered just as a danger<br />
but also as an opportunity for insurers and reinsurers<br />
<strong>to</strong> provide a valuable service and receive a<br />
fair reward for taking such <strong>risk</strong>.<br />
<strong>Risk</strong><br />
Control<br />
Processes<br />
• Emerging <strong>Risk</strong> <strong>Management</strong> <strong>to</strong> deal with future <strong>risk</strong>s;<br />
• <strong>Risk</strong> Control Processes <strong>to</strong> control and support the<br />
management of these <strong>risk</strong>s.<br />
The implementation of these 3 elements supports<br />
the roof of the temple, a process referred <strong>to</strong> by<br />
Standard & Poor’s as Strategic <strong>Risk</strong> <strong>Management</strong>.<br />
This is the ultimate aim of <strong>ERM</strong>.
I. <strong>Risk</strong> Culture<br />
What does <strong>Risk</strong> Culture mean for a (re)insurer? In fact,<br />
<strong>Risk</strong> Culture forms the basis of a solid <strong>risk</strong> management<br />
policy within the company, as illustrated in the<br />
Greek temple.<br />
The foundation of <strong>Risk</strong> Culture is strong internal<br />
<strong>risk</strong>-<strong>based</strong> governance. At SCOR this governance is<br />
overseen by a Board <strong>Risk</strong> Committee which reports<br />
<strong>to</strong> the Board of Direc<strong>to</strong>rs. The main responsibilities of<br />
this committee are:<br />
• Ensuring that the company has an effective <strong>ERM</strong><br />
framework in place;<br />
• Proposing an appropriate <strong>risk</strong> appetite framework <strong>to</strong><br />
the Board and ensuring this is clearly communicated<br />
<strong>to</strong> and unders<strong>to</strong>od by all stakeholders, in particular<br />
by staff;<br />
• Moni<strong>to</strong>ring and reporting on the Group’s <strong>risk</strong> profile<br />
<strong>to</strong> the Board;<br />
• Moni<strong>to</strong>ring and reporting critical <strong>risk</strong> issues <strong>to</strong> the<br />
Board.<br />
<strong>Risk</strong> Culture benefits from the appointment of a Chief<br />
<strong>Risk</strong> Officer (CRO) who is a member of the company’s<br />
Executive <strong>Management</strong>. He/she is responsible for the<br />
management of the above areas and is expected <strong>to</strong><br />
provide regular updates <strong>to</strong> the company’s Executive<br />
<strong>Management</strong> (weekly at SCOR) and the Board <strong>Risk</strong><br />
Committee (quarterly at SCOR).<br />
At SCOR, the day-<strong>to</strong>-day management of these<br />
areas is dealt with by the Group <strong>Risk</strong> <strong>Management</strong><br />
(GRM) department which reports <strong>to</strong> the Group CRO.<br />
The operating divisions (SCOR Global P&C and SCOR<br />
Global Life) also have their own <strong>Risk</strong> <strong>Management</strong><br />
organizations, headed by a Division CRO who has<br />
a dotted line reporting <strong>to</strong> the Group CRO. Both<br />
organizations work closely with GRM.<br />
From a governance point of view it is also imperative<br />
that a clear separation of roles between <strong>risk</strong> decision<br />
takers and <strong>risk</strong> managers exists. In particular the <strong>risk</strong><br />
takers must be accountable for their business decisions.<br />
The various levels of decision making should also be<br />
<strong>risk</strong>-<strong>based</strong>, e.g. critical <strong>risk</strong>s should be owned and<br />
managed by members of Executive <strong>Management</strong>.<br />
At SCOR, various <strong>risk</strong>-related committees, at or below<br />
the Group Executive <strong>Management</strong> level, provide<br />
formalized decision making forums which enable the<br />
views of <strong>risk</strong> decision takers and <strong>risk</strong> managers <strong>to</strong> be<br />
taken in<strong>to</strong> account. For example the Group Asset<br />
Liability <strong>Management</strong> (ALM) Committee is in charge of<br />
capital allocation (<strong>to</strong> assets and liabilities) and strategic<br />
asset allocation. The Group Investment Committee is<br />
responsible for tactical asset allocation and ensures that<br />
the investment guidelines are respected.<br />
SCOR - November 2010 - 5
Fig. 2: Process <strong>to</strong> derive <strong>Risk</strong> Preference and <strong>Risk</strong> Profile<br />
COMEX/BoD decision:<br />
• <strong>Risk</strong>-Based Capital (RBC)<br />
for asset <strong>risk</strong> is less than<br />
25% of <strong>to</strong>tal RBC<br />
• S&P capital for asset <strong>risk</strong><br />
is less than 15% of S&P<br />
Net Total Adjusted Capital<br />
Investment/ALM<br />
Committee decision:<br />
• Investment guidelines<br />
• Strategic asset allocation<br />
• Limit allocation<br />
Finally <strong>Risk</strong> Culture is reinforced by:<br />
Asset <strong>Risk</strong><br />
Preference<br />
• A remuneration system which incorporates incentives/disincentives<br />
for management and staff <strong>to</strong><br />
optimize <strong>risk</strong> and returns. The formula for SCOR’s<br />
staff bonuses incorporates a significant element in<br />
respect of individual performance which is <strong>based</strong> on<br />
objective evaluation criteria including a part which<br />
rewards individual contributions <strong>to</strong> effective <strong>risk</strong><br />
management;<br />
Definition of<br />
<strong>Risk</strong> Appetite<br />
Total Capital<br />
Allocated<br />
RAC<br />
Assets<br />
Fig. 3: <strong>ERM</strong> affects the entire organization<br />
<strong>Risk</strong> moni<strong>to</strong>ring and reporting<br />
6 - November 2010 - SCOR<br />
RAC<br />
Liabilities<br />
Board of Direc<strong>to</strong>rs<br />
Liability <strong>Risk</strong><br />
Preference<br />
<strong>Risk</strong> reporting Governance and direction<br />
<strong>Risk</strong> reporting<br />
Performance moni<strong>to</strong>ring<br />
<strong>Risk</strong> identification & assessment<br />
Mitigation plans + responsibilities<br />
Performance reporting<br />
COMEX<br />
<strong>Risk</strong> <strong>Management</strong><br />
COMEX decision:<br />
• Peak limit decision<br />
• Peak limit allocation<br />
• Product selection (NPI)<br />
• LoB/geographic<br />
growth areas<br />
• Underwriting<br />
guidelines<br />
• <strong>Risk</strong>-<strong>based</strong>, Group-wide policies and guidelines in areas<br />
such as <strong>ERM</strong>, reserving, underwriting, accounting,<br />
asset management, human capital management,<br />
compliance, internal audit, etc.;<br />
• <strong>Risk</strong>-<strong>based</strong> internal control standards (including<br />
exposure limits) at the process level.<br />
Policies & Strategy<br />
Prioritization and threshold setting<br />
Appropriateness of <strong>risk</strong> mitigation plans<br />
Decisions, guidance and ”sponsorship“<br />
Process and guidelines improvement<br />
<strong>Risk</strong> / <strong>to</strong>lerances adjustment / change<br />
<strong>Risk</strong> plan execution<br />
Operational Processes Operational Processes<br />
Mission and vision, <strong>risk</strong> policy and appetite
II. <strong>Risk</strong> and Economic Capital<br />
Modeling<br />
The first pillar of <strong>ERM</strong> is the <strong>Risk</strong> and Economic Capital<br />
Modeling of the company. In order <strong>to</strong> be efficient, the<br />
model has <strong>to</strong> include certain characteristics:<br />
• Consistency (of the model), which addresses all<br />
types of <strong>risk</strong> such as underwriting <strong>risk</strong>, market <strong>risk</strong><br />
and credit <strong>risk</strong>. These three <strong>risk</strong> areas are considered<br />
<strong>to</strong> be the most significant for SCOR.<br />
• Quality of the data, which helps <strong>to</strong> resolve<br />
potential deficiencies through regular validation<br />
and processes.<br />
• Appropriateness of assumptions, ensured e.g. by<br />
stress tests.<br />
• Capacity for identifying the main <strong>risk</strong> drivers of<br />
the company.<br />
• Quality assurance of <strong>Risk</strong> Adjusted Capital. The RAC<br />
modeling process and the connection <strong>to</strong> <strong>ERM</strong> and<br />
planning processes i.e. the internal model should be<br />
embedded in the business processes of the company,<br />
thus reinforcing <strong>Risk</strong> Culture.<br />
• Accurate programming and solid change<br />
management go without saying but they are very<br />
important for securing the current status and for<br />
any future changes.<br />
Assets<br />
Investments<br />
<br />
Cash & Short-term<br />
investments<br />
Fixed Income<br />
Equities<br />
Real Estate<br />
Alternative<br />
Investments<br />
Cash flow<br />
Accounting<br />
Economic<br />
Indica<strong>to</strong>r<br />
Economy<br />
Equity indices<br />
GDP<br />
Yield curves<br />
Forex<br />
SCOR includes all relevant <strong>risk</strong>s in its Asset and Liability<br />
<strong>Management</strong> <strong>approach</strong>. Figure 4 illustrates the issue<br />
of consistency. On the balance sheet of an insurance<br />
or reinsurance company, there are assets on one side<br />
and liabilities on the other. Cash flows and accounting<br />
connect assets with liabilities. Broadly speaking, an<br />
insurance or reinsurance company receives a premium<br />
that it invests. When the company has <strong>to</strong> pay claims, it<br />
sells some shares or assets that it has in its possession.<br />
This is the way most of the internal models look at the<br />
dependency between assets and liabilities. Nevertheless,<br />
the Asset and Liability <strong>Management</strong> <strong>approach</strong> is more<br />
complicated in reality as it is impacted by the economic<br />
situation. In fact, all phenomena affecting the economy<br />
have an influence on the value of both assets and<br />
liabilities.<br />
Here are two cases illustrating the question of<br />
dependency:<br />
The first example refers <strong>to</strong> a life insurance company<br />
that is committed <strong>to</strong> paying liabilities in ten years, at an<br />
interest rate of 5%. If the economic climate deteriorates<br />
(for instance, if a crisis occurs), the market value of its<br />
liability will increase because the interest rate will have<br />
decreased. This is because a debt payable in ten years<br />
will appear bigger when the interest rate is low. On<br />
the other hand, the behavior of the assets backing the<br />
expected liability will depend on the nature of those<br />
assets. Usually a large proportion of assets will be<br />
invested in bonds, and if the bond portfolio is chosen<br />
<strong>to</strong> have exactly the same duration as the liabilities<br />
Fig. 4: SCOR integrates all models in its ALM <strong>approach</strong> (consistent model)<br />
Liabilities<br />
Lines of Business (LoB)<br />
LoB1<br />
LoB2<br />
LoB9<br />
SCOR - November 2010 - 7
then the value of the bond portfolio will also increase<br />
by the same amount as the liabilities. This so-called<br />
‘perfectly matched’ portfolio is <strong>based</strong>, however, on<br />
the assumption that the expected future liabilities<br />
can be estimated with precision. In practice this is not<br />
always possible – if the liability cash flows turn out<br />
<strong>to</strong> be far greater than originally estimated then there<br />
are no available assets backing the shortfall. But it is<br />
precisely this shortfall that determines the amount of<br />
required capital. If interest rates fall then the value of<br />
this shortfall will increase, thus increasing the amount<br />
of required capital for the company.<br />
The second example focuses on natural catastrophes.<br />
Not all natural catastrophes have the same influence<br />
on the economy. An earthquake in Tokyo is expected <strong>to</strong><br />
impact the economy in many ways. Many studies have<br />
demonstrated that such a quake is likely <strong>to</strong> influence<br />
interest rates, the s<strong>to</strong>ck market, and the Japanese Yen’s<br />
foreign exchange rate.<br />
SCOR’s model deals with such big catastrophes. They<br />
are very important in terms of calculating the capital for<br />
e.g. 1/100 year events. It is thus essential <strong>to</strong> integrate<br />
all the impacts of this type of event in<strong>to</strong> SCOR’s model<br />
in order <strong>to</strong> come up with a realistic estimate of the<br />
required capital. This is the type of information that<br />
determines whether or not the model is consistent.<br />
Named events / stress scenarios*<br />
• Analysis of clearly defined and described events<br />
• Events can have happened in the past<br />
or may be possible in the future<br />
• Events can consist of a single <strong>risk</strong> fac<strong>to</strong>r<br />
or of a combination of several <strong>risk</strong> fac<strong>to</strong>rs<br />
• Definition of direct and indirect impact,<br />
management actions, and contingency plans<br />
Figure 5 presents the pros and cons of two modeling<br />
<strong>approach</strong>es: both s<strong>to</strong>chastic models and extreme scenarios<br />
are an important part of modeling. But since both<br />
<strong>approach</strong>es have their limitations, it is better <strong>to</strong> use both<br />
<strong>approach</strong>es <strong>to</strong> complement each other. The s<strong>to</strong>chastic<br />
models allow (re)insurance companies <strong>to</strong> explore many<br />
possible outcomes. Nonetheless, it is complicated <strong>to</strong><br />
calibrate the interactions between different phenomena<br />
and <strong>to</strong> deduce the <strong>risk</strong>s involved. The extreme scenarios<br />
explore only a few scenarios, so the range of <strong>risk</strong>s is very<br />
constrained. Nevertheless, this enables companies <strong>to</strong><br />
focus on the dependency and <strong>to</strong> build it in<strong>to</strong> the various<br />
scenarios. From a catastrophe such as an earthquake in<br />
Tokyo, it is possible <strong>to</strong> determine the consequences on<br />
the s<strong>to</strong>ck market and the interest rate, depending on<br />
the parameters chosen. By comparing the results of both<br />
<strong>approach</strong>es, it is possible <strong>to</strong> verify the plausibility of the<br />
results. For instance, if the probability given by the internal<br />
model of the outcome of a scenario is consistent with<br />
the examined scenario, this makes it more likely that the<br />
s<strong>to</strong>chastic model includes the same kind of cases.<br />
Stress tests combining both <strong>approach</strong>es are difficult <strong>to</strong><br />
develop; nonetheless they are very useful because they<br />
reveal <strong>risk</strong>s that are not detectable in the balance sheet.<br />
Thus, the best <strong>approach</strong> for these economic models is<br />
<strong>to</strong> combine s<strong>to</strong>chastic models and extreme scenarios.<br />
Incidentally, this is what the Swiss Solvency test requires<br />
from insurance companies.<br />
.<br />
Fig. 5: Stress Testing: Two forms of worst case analyses should be part of the model<br />
8 - November 2010 - SCOR<br />
Examples:<br />
- Financial distress<br />
- Severe adverse development in reserves<br />
- Tokyo earthquake<br />
- Retrocessionaires‘ default<br />
* Based on Lloyd’s, RDS, scenario catalog by the Swiss Solvency Test<br />
and SCOR specific scenarios.<br />
Extreme tail scenarios<br />
• Detailed analysis of the various <strong>risk</strong> drivers<br />
of the worst 5% scenarios from the economic<br />
scenario genera<strong>to</strong>r<br />
• Heavy-tailed extrapolation of distributions<br />
deliver truly extreme scenarios<br />
• Scenarios consist of a combination<br />
of Profit & Loss (P&L) and balance sheet impacts<br />
<strong>Risk</strong> driver examples:<br />
- Aviation<br />
- Credit & Surety<br />
- Marine<br />
- Foreign exchange rates<br />
- Interest rates
III. Emerging <strong>Risk</strong> <strong>Management</strong><br />
The second pillar of the <strong>ERM</strong> concept is Emerging <strong>Risk</strong><br />
<strong>Management</strong>. It is important <strong>to</strong> note that, ten years<br />
from now, nobody would have been in a position <strong>to</strong><br />
foresee the emerging <strong>risk</strong>s that have actually emerged.<br />
The main objective of Emerging <strong>Risk</strong> <strong>Management</strong> is<br />
not so much <strong>to</strong> predict what the future <strong>risk</strong>s are as <strong>to</strong><br />
be well-prepared for their emergence. This objective<br />
could be summarized in three key words:<br />
• Identify potential emerging <strong>risk</strong>s by using internal and<br />
external sources and central information gathering;<br />
• Assess the relevance of emerging <strong>risk</strong>s by identifying<br />
affected areas, estimating economic impact and correlation<br />
with other <strong>risk</strong>s;<br />
• Mitigate emerging <strong>risk</strong>s by hedging/retrocession<br />
strategies, setting exposure limits, changing terms<br />
and conditions and securing access <strong>to</strong> liquidity<br />
(contingent capital, securitization).<br />
Therefore, insurance and reinsurance companies need<br />
<strong>to</strong> develop methods <strong>to</strong> anticipate the potential <strong>risk</strong>s that<br />
could emerge, and above all <strong>to</strong> be prepared <strong>to</strong> mitigate<br />
the <strong>risk</strong>s in question. In the past, <strong>Risk</strong> <strong>Management</strong><br />
has led insurers and reinsurers <strong>to</strong> introduce certain<br />
preventive measures. Many safety regulations for<br />
industry and traffic have been inspired by insurers.<br />
SCOR combines a moni<strong>to</strong>ring system with observers<br />
(“ECHO”) and <strong>to</strong>ols for collecting information from<br />
numerous sources (e.g. “SCORWatch”) <strong>to</strong> track emerging<br />
<strong>risk</strong>s. If a new trend or phenomenon is detected, it<br />
will be analyzed <strong>to</strong> determine if it could be an emerging<br />
<strong>risk</strong>. In the case of an emerging <strong>risk</strong>, an assessment and<br />
mitigation process is triggered.<br />
Within SCOR’s working group over 80 <strong>risk</strong>s have been<br />
identified so far as possible emerging <strong>risk</strong>s.<br />
Sample emerging <strong>risk</strong>s<br />
identified at SCOR<br />
• Global pandemic<br />
• IT / Network <strong>risk</strong>s<br />
• Climate change<br />
• Legal / Regula<strong>to</strong>ry shift<br />
• Nuclear/biological/chemical terrorism<br />
• Technology: nanotechnologies, genetically<br />
modified organisms, chemicals<br />
• Product Liability<br />
• New chronic diseases<br />
• Mega projects<br />
• Socio-economic breakdown<br />
What are the options and the practical measures<br />
that (re)insurance companies could take when facing<br />
emerging <strong>risk</strong>s? Here is a sample list:<br />
• Update business continuity plans<br />
• Explore possible reinsurance options <strong>to</strong> mitigate<br />
this <strong>risk</strong><br />
• Include pre-defined clauses in insurance<br />
contracts<br />
• Strengthen underwriting guidelines<br />
• Ensure the existence of back-up systems<br />
• Constantly update the Natural Catastrophe<br />
models<br />
• Examine the possibility of opting out of certain<br />
covers/products/features where the <strong>risk</strong>s are<br />
becoming <strong>to</strong>o high<br />
• Assess the <strong>risk</strong> of parameter uncertainty for longtail<br />
business, review dependencies calculation<br />
• Encourage geographical diversification<br />
All of these measures form part of Emerging <strong>Risk</strong><br />
<strong>Management</strong>.<br />
SCOR - November 2010 - 9
IV. <strong>Risk</strong> Control Processes<br />
The third pillar of the Greek temple is the <strong>Risk</strong> Control<br />
Process. This involves updating and moni<strong>to</strong>ring the processes<br />
that have already been implemented. The <strong>risk</strong><br />
control process is centered on three key words:<br />
• Identify<br />
• Assess<br />
• Respond<br />
As mentioned previously, a part of the <strong>Risk</strong> Culture<br />
is <strong>to</strong> establish <strong>risk</strong>-<strong>based</strong> internal control standards<br />
(including exposure limits). One of the objectives of the<br />
control process is <strong>to</strong> make sure that these standards<br />
are followed by all staff and by the company as a<br />
whole. Every <strong>risk</strong> is measured regularly and then the<br />
current exposure is compared <strong>to</strong> the limits. If need be,<br />
actions are recommended. SCOR also delivers regular<br />
<strong>risk</strong> reporting <strong>to</strong> the major decision makers, Chief<br />
Underwriters, the Executive Committee and the Board<br />
<strong>Risk</strong> Committee.<br />
Certain <strong>risk</strong>s must be controlled and measured very<br />
frequently, particularly investment <strong>risk</strong>. For instance,<br />
a bank computes the Value-at-<strong>Risk</strong> daily and weekly.<br />
Fig. 6: Loss Event <strong>Management</strong><br />
Generic Headline Loss Process<br />
Fostering transparency of numbers and process<br />
Newsworthy<br />
event occurs<br />
High level<br />
estimate of loss<br />
coordinated by<br />
Claims<br />
Claims trigger the estimation process<br />
<strong>based</strong> on significant expected loss<br />
Proposed guideline: > 1.5m<br />
It is different for an insurance company. Indeed, frequency<br />
control has <strong>to</strong> be adapted <strong>to</strong> the time horizon<br />
of the <strong>risk</strong> in question.<br />
One of the processes involved in <strong>risk</strong> control is Loss Event<br />
<strong>Management</strong>. When a major loss occurs, the question<br />
is how the company will manage the consequences of<br />
the incident. Figure 6 gives an illustration of Loss Event<br />
<strong>Management</strong>.<br />
Taking the recent Chilean earthquake, as an example,<br />
the first thing that the insurance company needs <strong>to</strong><br />
do is <strong>to</strong> evaluate its potential exposure in Chile. If the<br />
first loss estimate is higher than a certain threshold,<br />
then the process of event management is initiated<br />
over one month. During this time, the company<br />
refines its estimation, and checks the underwriting<br />
of the contracts. So it moni<strong>to</strong>rs and reports <strong>to</strong> the<br />
management the <strong>to</strong>tal of the claims involved. Different<br />
parts of the organization will be involved in this process,<br />
e.g. natural disaster specialists, actuaries, underwriters<br />
and claims management units. For a month, a weekly<br />
estimation of the potential losses is computed and<br />
claims are reported contract by contract. After this initial<br />
phase, the moni<strong>to</strong>ring enters a more standard portfolio<br />
review process for as long as is needed <strong>to</strong> assess the full<br />
extent of the loss.<br />
Evaluate Initiate<br />
Moni<strong>to</strong>r<br />
Loss estimation process<br />
is triggered with UW.<br />
Weekly estimation + reporting<br />
of that specific loss event by contract<br />
Event moves in<strong>to</strong> standard portfolio<br />
review process for Headline Large Losses<br />
Online reporting and updating of all HLL events.<br />
Monthly reporting <strong>to</strong> management as a component<br />
part of claims reporting process<br />
48 hours 1 st month Ongoing<br />
Clearly defined interfaces between claims, underwriting, technical accounting and actuarial<br />
Figure 7 gives an example of growth on a company’s<br />
netting plan.<br />
This chart illustrates how a (re)insurance company<br />
manages its exposure through the setting of limits and<br />
the use of reinsurance. The company plans a certain<br />
level of exposure growth, and then it implements a<br />
reinsurance policy so that the peak exposure is slightly<br />
smoothed out, the portfolio remains balanced and<br />
10 - November 2010 - SCOR<br />
Headline loss<br />
is not actively<br />
moni<strong>to</strong>red<br />
the diversification of the entire company is preserved.<br />
In certain regions the company was not able, for<br />
business reasons, <strong>to</strong> reduce its exposure and ended up<br />
with exposure that was <strong>to</strong>o high; nevertheless, it was<br />
able <strong>to</strong> compensate it with hedging. Thus, the dark blue<br />
portfolio was finally more diversified than the light blue.<br />
This is an integral part of <strong>risk</strong> control and processes.
Fig. 7: Limit setting and enforcement<br />
Gross <strong>to</strong> Net Impact - Illustrative example<br />
Gross and Net distributions (including RI premiums): Single-event claims, 250-year return period<br />
Europe<br />
Winds<strong>to</strong>rm<br />
Turkey<br />
Earthquake<br />
Switzerland<br />
Earthquake<br />
Gross Net of Retrocession / RIPs<br />
Japan<br />
Earthquake<br />
Reinsurance companies negotiate their Treaty P&C contracts<br />
mainly over two months in the year (November<br />
and December for Europe, February and March for<br />
Japan). During this period, it is highly important <strong>to</strong><br />
moni<strong>to</strong>r the exposures that underwriters propose <strong>to</strong><br />
their cus<strong>to</strong>mers. The curve in Figure 8 shows a sample<br />
contract exposure.<br />
The company determines its authorized capacity. It is<br />
possible <strong>to</strong> propose more than the authorized capacity<br />
because not all contract proposals will be taken up by<br />
Fig. 8: <strong>Risk</strong> Moni<strong>to</strong>ring/Limit Control<br />
Renewal Process: <strong>Risk</strong> View<br />
Limit Setting<br />
(Net Appetite)<br />
Constant<br />
Feedback<br />
Loop<br />
<strong>Risk</strong> Control<br />
(Outwards options,<br />
additional limits,<br />
scale-down)<br />
<strong>Risk</strong> Reporting<br />
(Renewal Report)<br />
Japan<br />
Typhoon<br />
06.11<br />
USA<br />
Earthquake<br />
USA<br />
Hurricane<br />
Caribbean<br />
HU<br />
cus<strong>to</strong>mers. There is, however, a point where the proposed<br />
capacities, even though they are not all accepted,<br />
will run a high <strong>risk</strong> of ending up above the set limits.<br />
This is the point where the Chief <strong>Risk</strong> Officer must intervene.<br />
In this example, the Executive Committee decided,<br />
following intervention by the CRO, <strong>to</strong> authorize more<br />
capacity <strong>to</strong> underwriters. At the end of the renewal<br />
period the bound contracts had an accumulated exposure<br />
lower than the initially authorised capacity, but the<br />
<strong>risk</strong> was <strong>to</strong>o high for the CRO not <strong>to</strong> intervene. Once<br />
the contracts are bound, it is <strong>to</strong>o late.<br />
Renewal Process: Moni<strong>to</strong>ring<br />
14.11<br />
Illustrative example:<br />
Chief <strong>Risk</strong> Officer Intervention<br />
Estimated<br />
Exposure<br />
24.11<br />
01.12<br />
08.12<br />
Authorized<br />
capacity<br />
15.12<br />
20.12<br />
27.12<br />
03.01<br />
SCOR - November 2010 - 11
V. Strategic <strong>Risk</strong> <strong>Management</strong><br />
Strategic <strong>Risk</strong> <strong>Management</strong> constitutes the “roof”<br />
of <strong>ERM</strong>. Strategic <strong>Risk</strong> <strong>Management</strong> ensures that the<br />
company’s return on capital (upside) objectives and<br />
<strong>risk</strong> (downside) constraints are reconciled. It can be<br />
decomposed in<strong>to</strong> three parts:<br />
• Definition of a <strong>risk</strong> appetite framework (which is an<br />
integral part of the company’s strategic goals);<br />
• Development of various business strategies that<br />
enable the company <strong>to</strong> satisfy the objectives and<br />
constraints set out in the <strong>risk</strong> appetite framework;<br />
• Constant management of the <strong>risk</strong> profile <strong>to</strong> ensure it<br />
remains aligned with the <strong>risk</strong> appetite framework<br />
RISK APPETITE FRAMEWORK<br />
A company’s <strong>risk</strong> appetite framework should be defined<br />
by the company’s Board of Direc<strong>to</strong>rs or equivalent<br />
supervisory body. The <strong>risk</strong> appetite framework should<br />
be systematically reviewed when a new Strategic Plan<br />
is developed. At SCOR, the review entails discussions<br />
in the Executive <strong>Management</strong> Committee and in<br />
the Board <strong>Risk</strong> Committee of the key implications of<br />
potential changes <strong>to</strong> the <strong>risk</strong> appetite framework.<br />
Recommendations are then made <strong>to</strong> the Board.<br />
In exceptional circumstances the Board may vary the<br />
amount and the composition of <strong>risk</strong> which the Group<br />
is prepared <strong>to</strong> take.<br />
The <strong>risk</strong> appetite framework encompasses three<br />
concepts: <strong>risk</strong> appetite, <strong>risk</strong> preferences and <strong>risk</strong><br />
<strong>to</strong>lerances.<br />
<strong>Risk</strong> appetite<br />
Fig. 9: Framework <strong>to</strong> Strategic <strong>Risk</strong> <strong>Management</strong><br />
12 - November 2010 - SCOR<br />
Allocation between<br />
- Lines of Business (LoB)<br />
- Perils<br />
- Markets<br />
- Regions<br />
- Contract types<br />
- Clients<br />
Retro strategy<br />
<strong>Risk</strong> appetite defines the quantity of <strong>risk</strong> which a<br />
company wishes <strong>to</strong> accept <strong>to</strong> achieve a desired level<br />
of profitability. This will determine where the company<br />
wishes <strong>to</strong> position itself on the <strong>risk</strong>-return spectrum<br />
(between extremely <strong>risk</strong> averse i.e. low <strong>risk</strong>-low return<br />
and an extreme <strong>risk</strong> taker i.e. high <strong>risk</strong>-high return).<br />
SCOR uses a target retained <strong>risk</strong> profile (probability<br />
distribution of economic profits and losses) and target<br />
expected profitability <strong>to</strong> provide a complete definition<br />
of its <strong>risk</strong> appetite. A comparison of SCOR’s actual and<br />
target retained <strong>risk</strong> profile and profitability is regularly<br />
reported <strong>to</strong> the Board via the Board <strong>Risk</strong> Committee.<br />
<strong>Risk</strong> preferences<br />
<strong>Risk</strong> preferences are qualitative descriptions of the <strong>risk</strong>s<br />
which the Group is willing <strong>to</strong> accept i.e. the type of<br />
<strong>risk</strong> the company wants <strong>to</strong> take:<br />
• Mo<strong>to</strong>r or Property on the P&C side?<br />
• Mortality, Accident & Health or Disability on the Life<br />
side?<br />
• Low uncertainty or high uncertainty business?<br />
i.e. working layers or remote catastrophe layers?<br />
• Geographical areas? E.g. Europe or Asia?<br />
SCOR wishes <strong>to</strong> cover a wide range of reinsurance <strong>risk</strong>s<br />
from many geographical areas. SCOR has no desire <strong>to</strong><br />
take operational, legal, regula<strong>to</strong>ry, tax or reputation<br />
<strong>risk</strong>s (but this does not mean that SCOR is immune <strong>to</strong><br />
these <strong>risk</strong>s).<br />
This choice will determine the <strong>risk</strong>s <strong>to</strong> be included in<br />
SCOR’s underwriting guidelines.<br />
<strong>Risk</strong> appetite, <strong>risk</strong> preference, <strong>risk</strong> profile, <strong>risk</strong> limits<br />
Strategic liability<br />
<strong>risk</strong>-return-management<br />
Strategic asset and ALM<br />
<strong>risk</strong>-return-management<br />
- Strategic asset allocation<br />
- Hedging strategies<br />
- Duration (mis-)matching<br />
- Currency (mis-)matching
<strong>Risk</strong> <strong>to</strong>lerances<br />
<strong>Risk</strong> <strong>to</strong>lerances define the limits required by the company’s<br />
stakeholders (clients, shareholders, regula<strong>to</strong>rs, etc).<br />
SCOR defines two types of <strong>risk</strong> <strong>to</strong>lerances as follows:<br />
• The amount of retained exposure for each Line of<br />
Business (LoB) and asset class is limited <strong>to</strong> ensure<br />
that the contribution of each LoB or asset class <strong>to</strong> the<br />
95% TVaR of the annual Group economic profit/loss<br />
probability distribution does not exceed 5% (7.5%<br />
DEVELOPMENT OF BUSINESS STRATEGIES<br />
THAT SATISFY THE OBJECTIVES<br />
AND CONSTRAINTS SET OUT IN THE RISK<br />
APPETITE FRAMEWORK<br />
(i.e. <strong>risk</strong>-return optimization)<br />
Situations where optimized business strategies are<br />
developed include:<br />
• Definition of the strategic plan, which may include a<br />
possible revision of the <strong>risk</strong> appetite framework.<br />
In September 2010 SCOR launched its new three-year<br />
Strategic Plan “Strong Momentum” which replaced the<br />
previous plan “Dynamic Lift”. The ”Strong Momentum“<br />
plan has three targets over the cycle:<br />
- Increase moderately (relative <strong>to</strong> the previous<br />
“Dynamic Lift” strategic plan) the target <strong>risk</strong><br />
profile in the middle of the profit/loss probability<br />
distribution with no increase on tail <strong>risk</strong>s (after<br />
hedging).<br />
for CAT business) of the Group’s available economic<br />
capital. These limits are intended <strong>to</strong> avoid concentration<br />
of <strong>risk</strong> in specific LoBs or asset classes and as such<br />
help <strong>to</strong> maintain a well diversified portfolio.<br />
• The amount of retained exposure of the Group (from<br />
one or several LoBs or asset classes) <strong>to</strong> very extreme<br />
scenarios (with a probability of 1 in 250 years) is limited<br />
<strong>to</strong> a loss of 15% of the Group’s available economic<br />
capital. These limits are designed <strong>to</strong> limit the impact of<br />
extreme scenarios on the shareholders’ capital.<br />
Fig. 10: Strategic <strong>Risk</strong> <strong>Management</strong> means basing business decisions<br />
on <strong>Risk</strong>/Return analysis<br />
Analysis<br />
of the<br />
renewal<br />
results<br />
Modeling<br />
of current<br />
portfolio<br />
- <strong>Risk</strong> appetite<br />
- Limit definitions<br />
for exposures<br />
- Pricing parameters<br />
- New product reviews<br />
Timeline<br />
January June November<br />
ALM<br />
Business planning<br />
Strategic<br />
targets<br />
(capital<br />
allocation)<br />
Business<br />
planning<br />
Modeling<br />
of the<br />
planned<br />
portfolio<br />
(<strong>risk</strong> budget)<br />
- Cat limits<br />
- Retro strategy<br />
- Capital consumption<br />
Strategic<br />
asset<br />
allocation<br />
(investments)<br />
Renewal<br />
pricing<br />
against<br />
the planned<br />
portfolio<br />
- Alignment of the asset<br />
portfolio <strong>to</strong> the plan<br />
- Strategic Asset Allocation<br />
- Allocation of extracapacity<br />
- Increase Return on Shareholders’ Equity (RoE) <strong>to</strong><br />
1,000 bps above <strong>risk</strong>-free rate over the cycle.<br />
- Reach an “AA” level of security(1)<br />
for clients<br />
by 2013.<br />
• Definition of business units’ annual underwriting<br />
plans<br />
Capital management is firmly embedded in SCOR’s<br />
business planning processes. Figure 10 shows how the<br />
process of modeling <strong>risk</strong>s is incorporated in SCOR’s<br />
P&C business unit’s planning process. After the P&C<br />
renewal period the results are analyzed and the portfolio<br />
is modeled. From this, business targets, new<br />
pricing parameters and products are defined for next<br />
year, and all this information is brought <strong>to</strong>gether in the<br />
P&C business plan. Capital allocations by LoB are also<br />
defined for next year. This allows SCOR <strong>to</strong> measure<br />
the performance of a LoB and <strong>to</strong> reduce or increase<br />
the exposure according <strong>to</strong> the potential results of that<br />
particular LoB. The business plan also includes Natural<br />
(1) This reflects the level of security provided by SCOR according <strong>to</strong><br />
the S&P scale; however it does not reflect any Rating Agencies’<br />
opinions of the Group.<br />
SCOR - November 2010 - 13
Catastrophe limits, a retrocession strategy and a <strong>to</strong>tal<br />
capital consumption concerning the next year. If there<br />
is any extra capital, a decision is made about what <strong>to</strong><br />
do with it. The projected portfolio is then modeled,<br />
enabling SCOR <strong>to</strong> define asset allocations (within the<br />
boundaries of the strategic asset allocation) for the<br />
next year. After this, the renewal period begins.<br />
• Strategic asset allocation<br />
Strategic asset allocation determines the optimal<br />
asset portfolio mix that will protect the liabilities. The<br />
strategic asset allocation should be reflected in the<br />
company’s capital model and used <strong>to</strong> define limits<br />
and <strong>risk</strong> budgets.<br />
• Individual business decisions e.g. pricing, terms<br />
and conditions, product design<br />
• Mergers and acquisitions<br />
Taken <strong>to</strong>gether, the above actions lead <strong>to</strong> portfolio<br />
optimization at the company level which, in turn,<br />
facilitates the development of a <strong>risk</strong> management<br />
culture within the company.<br />
MANAGEMENT OF THE RISK PROFILE<br />
Using various <strong>risk</strong> moni<strong>to</strong>ring and response mechanisms<br />
within its <strong>ERM</strong> Framework, SCOR’s Executive<br />
<strong>Management</strong> and the Board <strong>Risk</strong> Committee periodically<br />
review the Group’s actual <strong>risk</strong> profile and are<br />
systematically informed of any deviation from the target<br />
<strong>risk</strong> profile. In particular actual exposures are moni<strong>to</strong>red<br />
against <strong>risk</strong> exposure limits for different Lines of<br />
Business, derived from the Group’s <strong>risk</strong> <strong>to</strong>lerances.<br />
14 - November 2010 - SCOR<br />
Variations in the Group’s retained <strong>risk</strong> profile may<br />
occur due <strong>to</strong> a change in the underlying asset and<br />
liability portfolios and changes in the level of <strong>risk</strong>iness<br />
of certain parts of the asset and liabilities, e.g. due <strong>to</strong><br />
rapid or slowly emerging changes in the economic,<br />
financial, social, legal and regula<strong>to</strong>ry environments.<br />
<strong>Risk</strong> moni<strong>to</strong>ring and response mechanisms enable<br />
companies <strong>to</strong>:<br />
• Take mitigating actions <strong>to</strong> ensure that the company’s<br />
retained exposure <strong>to</strong> specific <strong>risk</strong>s remains<br />
within the defined exposure limits<br />
The first part of SCOR’s Capital Shield Policy – active<br />
hedging of peak exposures through retrocession and<br />
Insurance linked Securities (ILS) – is a key mitigation<br />
mechanism. The level of retrocession is selected <strong>to</strong><br />
ensure that the Group’s retained <strong>risk</strong> profile respects<br />
the Group’s <strong>risk</strong> <strong>to</strong>lerances.<br />
• Take capital management actions <strong>to</strong> ensure that<br />
the Group’s capital base remains aligned <strong>to</strong> the<br />
<strong>risk</strong> appetite framework<br />
The second part of SCOR’s Capital Shield Policy –<br />
holding of Buffer Capital in addition <strong>to</strong> the required<br />
solvency capital – is a key capital management<br />
mechanism. The level of Buffer Capital is chosen <strong>to</strong><br />
be consistent with SCOR’s <strong>risk</strong> appetite framework<br />
and strategic targets and enables SCOR <strong>to</strong> absorb<br />
a significant amount of inherent volatility in annual<br />
results, without having <strong>to</strong> turn <strong>to</strong> the market <strong>to</strong>o<br />
frequently <strong>to</strong> raise capital.
Conclusion<br />
As a conclusion, it can be asserted that <strong>ERM</strong> is nothing<br />
less than sound insurance practice:<br />
• It allows (re)insurance companies <strong>to</strong> steer the<br />
business;<br />
• It enables them <strong>to</strong> define the value drivers of<br />
insurance;<br />
• It allows them <strong>to</strong> measure the performance of the<br />
business;<br />
• It encompasses the whole organization and its<br />
processes;<br />
• It makes the company more transparent <strong>to</strong> all<br />
stakeholders.<br />
Fig. 11: Strategic Asset Allocation (SAA) <strong>based</strong> on efficient frontier<br />
<strong>Risk</strong> versus Return (efficient frontier)<br />
Expected return<br />
Scenarios of equity allocations<br />
0% equity allocation<br />
Optimum equity allocation<br />
Downside <strong>risk</strong> (<strong>based</strong> on expected shortfall)<br />
HOW IS THE STRATEGIC ASSET ALLOCATION<br />
DET<strong>ERM</strong>INED?<br />
Figure 11 shows a typical <strong>risk</strong>/reward curve, the x-axis<br />
representing the downside <strong>risk</strong> <strong>based</strong> on the expected<br />
shortfall, while the y-axis corresponds <strong>to</strong> the expected<br />
return of the strategy. The curve is computed by<br />
varying the proportion of equity in the investment<br />
portfolio. Both the <strong>risk</strong> and the reward are computed<br />
using the full asset and liability portfolio. The red<br />
point on the graph has 0% of equity. This means<br />
the company has only invested in fixed income.<br />
From 5%, it can build what is called the efficient<br />
frontier, which is basically the point at which it could<br />
obtain the maximum return for the minimum <strong>risk</strong>.<br />
This curve represents the set of optimal strategies<br />
This is why SCOR firmly believes that <strong>ERM</strong> does not<br />
simply represent a passing trend but is a concrete way<br />
for an insurance company <strong>to</strong> become more professional<br />
and more competitive. As with everything in life, there<br />
is a price <strong>to</strong> pay: <strong>ERM</strong> requires a long-term commitment<br />
<strong>to</strong> excellence from the entire organization.<br />
The image used in this article <strong>to</strong> understand the structure<br />
of the <strong>ERM</strong> concept is a Greek temple, therefore we will<br />
let a Famous Greek philosopher, Aris<strong>to</strong>tle, conclude:<br />
“We are what we repeatedly do. Excellence, therefore,<br />
is not an act, but a habit.”<br />
The investment strategy is <strong>based</strong> on:<br />
• <strong>Risk</strong>/return considerations<br />
for the entire shareholder’s equity<br />
(including liability <strong>risk</strong>)<br />
• And <strong>risk</strong> aversion as defined<br />
by <strong>to</strong>p management<br />
depending on the amount of equity in the investment<br />
portfolio. To determine the optimal asset allocation<br />
of the company, it is necessary <strong>to</strong> determine a target<br />
return. With this return it is easy <strong>to</strong> construct the<br />
corresponding straight line. The point where this<br />
line <strong>to</strong>uches the curve is the point of the optimal<br />
asset allocation in light of the company’s profile.<br />
However, optimal asset allocation is not sufficient for<br />
a complete strategic asset allocation, because given<br />
the fluctuations of financial markets it is not possible<br />
<strong>to</strong> follow it constantly. It is needed <strong>to</strong> determine<br />
bands around which the allocation can move. For<br />
this, the company also needs <strong>to</strong> have a maximum<br />
and a minimum so as <strong>to</strong> limit the <strong>risk</strong>s and <strong>to</strong> allow<br />
for tactical asset allocation within those boundaries.<br />
SCOR - November 2010 - 15
2<br />
ADAPTING SOLVENCY<br />
REGULATIONS TO<br />
TIMES OF CRISIS MICHEL DACOROGNA<br />
Deputy Chief <strong>Risk</strong> Officer, SCOR<br />
With the advent of the economic<br />
and financial crisis, the question of solvency is<br />
increasingly under discussion. This article addresses<br />
the current financial crisis and its consequences<br />
on solvency requirements and takes a fresh look<br />
at crises and their characteristics. It ultimately<br />
presents concrete proposals <strong>to</strong> make the regula<strong>to</strong>ry<br />
system more flexible in terms of responding <strong>to</strong><br />
future crises.<br />
I. Financial crises form part<br />
of the economic system<br />
The current developments in the financial industry<br />
are among the worst economic crises in his<strong>to</strong>ry, as<br />
described in the diagram below. This shows a large<br />
sample created by a his<strong>to</strong>rian on the movements of the<br />
New York s<strong>to</strong>ck market from the 1800s <strong>to</strong> 2008:<br />
Fig. 12: The current developments in the financial industry line up amongst<br />
the worst economic crises in his<strong>to</strong>ry<br />
S&P index annual returns from 1800 <strong>to</strong> 2008<br />
What is next?<br />
We are here!<br />
1965 9.1%<br />
2004 9.0%<br />
1959 8.5%<br />
1886 8.5%<br />
1845 8.1%<br />
1968 7.7%<br />
1830 7.4%<br />
1921 7.4%<br />
1871 7.3%<br />
1993 7.1%<br />
1802 6.8%<br />
1872 6.8%<br />
1899 6.5%<br />
1864 6.4%<br />
1878 6.1%<br />
1821 6.1%<br />
1926 5.7%<br />
1870 5.6%<br />
1840 5.5%<br />
1824 5.1%<br />
1832 4.8%<br />
1820 4.7%<br />
1806 4.6%<br />
1992 4.5% 1885 19.8%<br />
1807 4.4% 1852 19.6%<br />
1981 -9.7% 1856 4.4% 1999 19.5%<br />
1803 -9.6% 1866 3.6% 1943 19.4%<br />
1877 -9.4% 2007 3.5% 1976 19.1%<br />
1914 -8.6% 1889 3.5% 1898 18.9%<br />
1883 -8.5% 1812 3.5% 1963 18.9%<br />
1865 -8.5% 1916 3.4% 1924 18.8%<br />
1811 -7.5% 1906 3.1% 1850 18.7%<br />
1884 -18.8% 1819 -7.5% 1835 3.1% 1880 18.7%<br />
1903 -18.4% 1953 -6.6% 2005 3.0% 1891 17.6%<br />
1842 -18.1% 1887 -6.6% 1881 3.0% 1983 17.3%<br />
1876 -17.9% 1990 -6.6% 1831 3.0% 1951 16.5%<br />
1941 -17.9% 1934 -6.0% 1912 3.0% 1918 16.4%<br />
1973 -17.4% 1825 -5.8% 1874 2.8% 1901 15.7%<br />
1814 -16.7% 1939 -5.4% 1815 2.7% 1905 15.6%<br />
1940 -15.3% 1822 -4.8% 1956 2.6% 1972 15.6%<br />
1932 -15.2% 1805 -4.4% 1987 2.0% 1844 15.5%<br />
1846 -14.5% 1804 -4.3% 1892 1.8% 1982 14.8%<br />
1957 -14.3% 1875 -4.1% 1869 1.7% 1986 14.6%<br />
1913 -14.3% 1816 -3.8% 1838 1.6% 1858 14.3% 1915 29.0%<br />
1890 -13.5% 1848 -3.6% 1867 1.6% 1900 14.1% 1936 27.8%<br />
1841 -13.3% 1818 -3.2% 1855 1.5% 1909 14.1% 1989 27.3%<br />
1966 -13.1% 1851 -3.2% 1984 1.4% 1860 14.0% 1998 26.7%<br />
2001 -13.0% 1827 -3.1% 1902 1.3% 1919 14.0% 2003 26.4%<br />
1873 -12.7% 1960 -3.0% 1847 1.2% 1944 13.8% 1955 26.4%<br />
1853 -12.7% 1882 -2.9% 1813 1.1% 2006 13.6% 1985 26.3%<br />
1839 -12.3% 1894 -2.5% 1978 1.1% 1834 13.0% 1991 26.3%<br />
1910 -12.1% 1888 -2.5% 1809 1.1% 1964 13.0% 1980 25.8%<br />
1929 -11.9% 1896 -2.3% 1911 0.7% 1897 12.6% 1904 25.6% 1958 38.1%<br />
1946 -11.9% 1810 -2.1% 1895 0.5% 1942 12.4% 1938 25.3% 1863 38.0%<br />
1962 -11.8% 1861 -1.8% 1970 0.1% 1988 12.4% 2009 23.5% 1928 37.9%<br />
1937 -38.6% 1836 -11.7% 1994 -1.5% 1808 0.0% 1979 12.3% 1925 23.3% 1908 37.4%<br />
2008 -38.5% 1974 -29.7% 1977 -11.5% 1923 -1.5% 1823 0.0% 1817 11.9% 1961 23.1% 1995 34.1% 1933 46.6%<br />
1907 -33.2% 1930 -28.5% 1837 -11.5% 1829 -1.1% 1826 0.0% 1952 11.8% 1950 21.8% 1975 31.5% 1954 45.0%<br />
1857 -31.0% 1920 -24.5% 1969 -11.4% 1833 -0.9% 1828 0.0% 1971 10.8% 1922 20.9% 1997 31.0% 1843 45.0%<br />
1917 -30.6% 2002 -23.4% 1859 -10.7% 1948 -0.7% 1849 0.0% 1868 10.8% 1996 20.3% 1927 30.9% 1879 43%<br />
1931 -47% 1854 -30.2% 1893 -20% 2000 -10.1% 1801 -0.1% 1947 0.0% 1949 10.3% 1967 20.1% 1945 30.7% 1935 41.5ç 1862 55.4%<br />
To -50% To -40% To -30% To -20% To -10% 0-10% 10-20% 20-30% 30-40% 40-50% 50-60%<br />
Source: globalfinancialdata.com<br />
16 - November 2010 - SCOR<br />
2009 showed<br />
a +24% return
The cells (Figure 12) contain the yearly returns of the<br />
s<strong>to</strong>ck market from -50% <strong>to</strong> +50% <strong>to</strong> 60%, and are<br />
ordered by size. The 2008 crisis, at -38.5%, is the third<br />
largest yearly negative return in 210 years. The year<br />
1937 comes second, differing from 2008 by only 0.1%.<br />
The worst year is 1931 at -47%.<br />
The internet bubble that burst in 2000 caused a<br />
decrease of 10% (-10%) at the end of the year. The<br />
s<strong>to</strong>ck market returns of 2000-2002 provoked a decrease<br />
of 23% (-23%), while the crisis of 2008 generated<br />
a drop of 38.5% (-38.5%). In 2009, s<strong>to</strong>ck market<br />
returns recorded an increase of 24%; nonetheless the<br />
same phenomenon occurred for the other crises. This<br />
temporary recovery should not be taken as the end of<br />
the turbulence created by the financial crisis.<br />
The financial crisis has caused previously unimaginable<br />
wealth losses: In May 2008, the hundred largest banks<br />
in the world lost USD 384 billion. Among them were<br />
Citigroup: USD 43 billion, UBS: USD 39 billion and Merrill<br />
Lynch: USD 37 billion. The Gross Domestic Product (GDP)<br />
of all OECD countries dropped by 2% in the last quarter<br />
of 2008 and by 2.1% in the first quarter of 2009.<br />
AIG lost USD 100 billion during the last quarter of 2008.<br />
This corresponds <strong>to</strong> USD 300,000 per minute, 24 hours<br />
a day, 7 days a week! These are amounts that defy the<br />
imagination.<br />
Crises are a part of human his<strong>to</strong>ry, whether they are<br />
man-made or natural. His<strong>to</strong>ry shows that mankind<br />
has been exposed <strong>to</strong> many natural catastrophes and<br />
man-made crises, looking back over just two centuries:<br />
• Since 1825 the volcano Vesuvius has had 9 periods<br />
of eruption (1) . During the twentieth century volcanic<br />
eruptions caused fewer than 1,000 deaths per year.<br />
However, two events caused more than half the deaths<br />
recorded in the last century: 29,000 killed for the 1902<br />
eruption of Mont Pelée (on the island of Martinique)<br />
and 23,000 killed for the 1985 eruption of Nevado del<br />
Ruiz (in Colombia) (2) . More recently, the 2004 Asian<br />
Tsunami claimed more than 170,000 lives (3) .<br />
• Over this period, the S&P 500 index has lost 9 times<br />
more than 20% in a year. When looking at those<br />
facts, one is almost tempted <strong>to</strong> think that crisis is<br />
more the norm than the exception.<br />
Many crises have already been recorded; it is very likely<br />
that there are many more <strong>to</strong> come. Even if it is impossible<br />
<strong>to</strong> foresee such catastrophes, the idea is <strong>to</strong> be prepared<br />
and <strong>to</strong> adapt the financial system <strong>to</strong> face up <strong>to</strong> crises with<br />
as little damage as possible. This means that companies<br />
and regula<strong>to</strong>rs have <strong>to</strong> change their general mindset.<br />
The focus should be much more on how companies can<br />
survive these crises unharmed than on trying <strong>to</strong> predict or<br />
avoid crises. Avoiding crises is probably an unachievable<br />
goal, so let us see how <strong>to</strong> survive them instead.<br />
WHAT IS BEHIND THE CURRENT SEVERE<br />
CRISIS? WHAT ARE ITS CHARACTERISTICS?<br />
Existing measures have shown their limits in terms of<br />
preventing severe crises from recurring and preventing<br />
<strong>risk</strong> from asphyxiating the development of the financial<br />
system through excessive capital requirements and<br />
deleveraging. It is thus very important <strong>to</strong> analyze the<br />
current crisis and <strong>to</strong> come up with new ideas for how<br />
<strong>to</strong> handle such exceptional situations.<br />
Three characteristics come out of any financial crisis:<br />
• A crisis, almost by definition, comes as a surprise<br />
because it reveals links between <strong>risk</strong>s that were<br />
previously neglected;<br />
• The <strong>risk</strong> appetite of market ac<strong>to</strong>rs is irresponsibly<br />
high, and accompanied by high financial leverage;<br />
• There is an excessive concentration of aggregate <strong>risk</strong><br />
in highly leveraged financial institutions.<br />
(1) Source Volcano Discovery.<br />
(2) Environmental Hazards: Assessing <strong>Risk</strong> and Reducing Disaster.<br />
(3) World Bank.<br />
SCOR - November 2010 - 17
The surprises that have the potential <strong>to</strong> trigger severe<br />
financial crises are not simply bad realizations within<br />
a known probabilistic environment. Rather, they are<br />
changes in the environment itself. It is this “rare event”<br />
feature that holds the key as it has the potential <strong>to</strong><br />
trigger a sharp rise in perceived uncertainty and flight<br />
<strong>to</strong> quality. An example of such a surprise <strong>to</strong>ok place<br />
when the Reserve Primary Fund, a leading money market<br />
fund, did not reimburse at par after Lehman declared<br />
bankruptcy (breaking the buck). The authority thought<br />
that because Lehman did not have a very high Credit<br />
Default Swap (CDS) exposure and the CDS exposure<br />
of the Reserve Primary Fund was so high, its failure<br />
would not catch the market by surprise. Unfortunately,<br />
the Primary Fund had invested 1.2% of its assets in<br />
Lehman’s debt.<br />
Immediately after Lehman filed for bankruptcy, the<br />
fund suffered a massive run, with over USD 30 billion<br />
of redemption requests (about half of its <strong>to</strong>tal assets),<br />
before it s<strong>to</strong>pped accepting redemption requests the<br />
following day. Money market funds had been considered<br />
extremely safe, and had indeed benefited from the<br />
flight <strong>to</strong> quality the previous year, growing by about<br />
USD 850 billion (34%) since mid-2007. The drop in the<br />
Reserve Primary Fund’s NAV caused inves<strong>to</strong>rs <strong>to</strong> question<br />
the safety of the entire industry and the first under-par<br />
redemption from a Money Market fund since their<br />
creation in 1968. That week, there were USD 169 billion<br />
in redemptions from <strong>to</strong>tal investments of USD 3.4 trillion<br />
(5%), as well as a major shift from prime funds <strong>to</strong>wards<br />
funds investing exclusively in government debt.<br />
In retrospect, the consequences of Lehman’s demise<br />
on the Primary Fund could have been predicted: public<br />
filings showed large investments in Lehman as early<br />
as November 2007. Anyone who <strong>to</strong>ok the trouble <strong>to</strong><br />
connect the dots could, in principle, have foreseen<br />
what might happen. However, money market funds<br />
had a track record of stability that had always made it<br />
unnecessary <strong>to</strong> inspect their holdings. The realization<br />
that there might be further losses in previously<br />
unexamined places led inves<strong>to</strong>rs <strong>to</strong> intensify their flight<br />
<strong>to</strong> quality. The main failure was in not understanding<br />
that relatively “small effects” could generate huge<br />
impacts and create confusion for the entire system.<br />
Surprises quickly trigger a chain of unexpected events<br />
from the panic they engender. In times of crisis small<br />
effects can generate huge impacts.<br />
18 - November 2010 - SCOR<br />
The second important trigger for a severe crisis is the<br />
highly leveraged and interconnected sec<strong>to</strong>r of the<br />
economy, generally the financial sec<strong>to</strong>r, being exposed<br />
(directly or indirectly) <strong>to</strong> a surprise of the kind discussed<br />
earlier. Aggregate <strong>risk</strong>s are those exposed <strong>to</strong> aggregate<br />
shocks impacting the entire economy. Investment in<br />
structured products has exposed financial institutions<br />
<strong>to</strong> more aggregate <strong>risk</strong> and surprises than in the past.<br />
In the current crisis, banks were holding mostly AAAtranches<br />
of a large variety of new ABS (Asset Backed<br />
Securities) (85% of assets held in securitized form).<br />
These tranches rely on protection from junior tranches<br />
and from the law of large numbers in order <strong>to</strong> reduce<br />
the <strong>risk</strong> of default enough <strong>to</strong> achieve an AAA-rating.<br />
The law of large numbers implies that the loss on a pool<br />
with a sufficient number of underlying assets, as was the<br />
case with most ABS, can only occur when an aggregate<br />
shock takes place. Furthermore, the higher up a given<br />
structure is situated, the larger the aggregated shock<br />
must be for it <strong>to</strong> pierce the protection offered by the junior<br />
tranches. Losses large enough <strong>to</strong> affect the AAA-tranche<br />
only occur in situations of severe aggregate shocks,<br />
but this is exactly what large surprises do! Therefore<br />
holdings of AAA-tranches of structured products<br />
exposed financial institutions <strong>to</strong> more systemic <strong>risk</strong>s<br />
than their rating, when misinterpreted, would suggest,<br />
and certainly more than similarly rated “single name”<br />
corporate bonds. Corporate bonds are still affected by<br />
macroeconomic conditions but idiosyncratic fac<strong>to</strong>rs play<br />
a larger role. Downgraded structured financial securities<br />
lost on average between 5 and 6 notches in the period<br />
2007/2008. By comparison, during the great corporate<br />
bond downgrade of 2001/2002 (30% of corporate<br />
bonds were downgraded) the average notch-loss<br />
was 1.8. The systemic consequence of this <strong>risk</strong> was<br />
that highly leveraged institutions were bearing more<br />
aggregate <strong>risk</strong>s than would have been thought from<br />
simply observing the average ratings of their assets.<br />
Having the financial sec<strong>to</strong>r of the economy holding such<br />
<strong>risk</strong> with respect <strong>to</strong> aggregate surprise proved <strong>to</strong> be a<br />
recipe for disaster.<br />
Surprise and aggregate <strong>risk</strong>s are major causes for<br />
financial crises. Neither element is likely <strong>to</strong> disappear<br />
soon. One can thus expect that the economy, despite<br />
all attempts <strong>to</strong> protect it, will suffer shocks in the future.<br />
The correct <strong>risk</strong> management policy is <strong>to</strong> prepare the<br />
organization for the occurrence of such shocks and <strong>to</strong><br />
make sure that the company can survive them. Rather<br />
than trying <strong>to</strong> predict the next crisis or <strong>to</strong> avoid it,<br />
(re)insurance companies should concentrate on making<br />
sure that their <strong>risk</strong> management model integrates the<br />
occurrence of crisis with a reasonable probability. Only<br />
in this way will organizations be prepared <strong>to</strong> face major<br />
disruptions in the financial markets.
II. Consequences of financial<br />
crises on insurance capital<br />
requirements<br />
SCOR was actually very prudent during the crisis, building<br />
a huge cash position. Retrospectively, this proved <strong>to</strong> be<br />
the right thing <strong>to</strong> do, especially with a good model.<br />
However, this policy drew questions from the regula<strong>to</strong>rs<br />
because it departed from the usual ALM strategies. It is<br />
thus useful <strong>to</strong> look at the impact of the crisis on SCOR’s<br />
internal model <strong>to</strong> understand why the company acted<br />
like this and <strong>to</strong> see what a good alternative could be in<br />
terms of regulation during times of crisis.<br />
THE MAIN CONSEQUENCES OF THE CRISIS<br />
ON SCOR’S INTERNAL MODEL<br />
Good <strong>risk</strong> models would show an increased <strong>risk</strong> of the<br />
situation and thus come up with higher <strong>risk</strong> adjusted<br />
capital than in quieter times. This is natural because<br />
the model reflects inherent uncertainty, which naturally<br />
increases during turbulent times. Requiring companies<br />
<strong>to</strong> keep the same level of security as before the crisis<br />
would require a significant increases in capital. The<br />
following example examines the consequences of the<br />
crisis on SCOR’s internal model.<br />
Some consequences are directly linked <strong>to</strong> the fall of<br />
interest rates:<br />
Fig. 13: <strong>Risk</strong>-Free Rate development<br />
His<strong>to</strong>rical 10Y <strong>Risk</strong>-Free Rates<br />
in %<br />
10<br />
8<br />
6<br />
4<br />
2<br />
Aug. 87<br />
Aug. 89 Aug. 91 Aug. 93<br />
10Y US T-bond<br />
• The life model required more capital for mortality<br />
due <strong>to</strong> the fall in interest rates. This translated <strong>to</strong><br />
more capital for the whole SCOR portfolio. It is easy<br />
<strong>to</strong> understand that a liability payable in ten years<br />
appears bigger if the interest rate is low.<br />
• At the same time, SCOR’s available capital experienced<br />
a marked decrease. Using the same yield curve as<br />
before the crisis would have resulted in an increase in<br />
the available capital. Due <strong>to</strong> the fall in the discounting<br />
benefit of P&C reserves, the available capital looks<br />
smaller, even though aside from the interest rate<br />
nothing has changed in the portfolio.<br />
Other consequences are linked <strong>to</strong> the increased volatility<br />
of financial markets and increased credit <strong>risk</strong>:<br />
• During the crisis the volatility of s<strong>to</strong>ck returns more<br />
than doubled.<br />
• Credit spread skyrocketed for reinsurance <strong>to</strong> more<br />
than 2,000 basis points.<br />
Those cumulated effects significantly reduce the<br />
solvency ratio of companies. SCOR has seen a significant<br />
drop in its solvency ratio even though there were no<br />
significant changes in its liability portfolio and its asset<br />
portfolio was significantly de<strong>risk</strong>ed. Luckily SCOR was<br />
well capitalized; otherwise the company would have<br />
had <strong>to</strong> go <strong>to</strong> the financial markets <strong>to</strong> get more cash<br />
at a time of liquidity squeeze, thus accentuating the<br />
lack of liquidity.<br />
Aug. 95 Aug. 97 Aug. 99 Aug. 01 Aug. 03 Aug. 05 Aug. 07 Aug. 09<br />
10Y German Bund<br />
10Y French OAT<br />
10Y US T-bond: 3.68x<br />
10Y French OAT: 3.502x<br />
10Y German Bund: 3.26x<br />
Source: Reuters as of 22 Jan. 2010.<br />
SCOR - November 2010 - 19
OTHER ECONOMIC INDICATORS<br />
SUCH AS INFLATION MAY BE TAKEN<br />
INTO CONSIDERATION<br />
Figure 14, which comes from an Economic Scenario<br />
Genera<strong>to</strong>r during the crisis (last quarter of 2008), shows<br />
the importance of such an advanced indica<strong>to</strong>r. This is<br />
an input in SCOR’s model. It refers <strong>to</strong> the distribution of<br />
inflation. The blue line corresponds <strong>to</strong> the distribution<br />
of the his<strong>to</strong>rical inflation index. The brown one is the<br />
forecast distribution. The chart indicates the change<br />
in the distribution. Considering that the interest rate<br />
and the inflation will decrease, the distribution has<br />
shifted. Nonetheless, the most important thing for<br />
the capital is the widening of the distribution. SCOR’s<br />
model forecast a higher uncertainty on the inflation.<br />
The higher uncertainty on inflation required a capital<br />
increase because of the uncertainty of the outcome.<br />
Fig. 14: The ESG reflects the current<br />
uncertainty on inflation<br />
Empirical density function of 1 year returns of CPI: USD<br />
15<br />
10<br />
5<br />
0<br />
-0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5<br />
His<strong>to</strong>rical data: annualized quarterly returns<br />
Simulated data: yearly returns<br />
Data:<br />
• Number of his<strong>to</strong>rical data: 83 (relatively small)<br />
• Number of simulations: 60,000<br />
The empirical distribution of the simulated inflation is:<br />
- Out of phase because the current interest rates are lower,<br />
- Wider because the current volatility is bigger (GARCH effect).<br />
20 - November 2010 - SCOR<br />
APPLYING SOLVENCY RULES AFTER<br />
A CRISIS IS COUNTERPRODUCTIVE<br />
When a crisis occurs, regula<strong>to</strong>rs insist on respecting<br />
solvency requirements, but:<br />
• The economic situation is deteriorating, which could<br />
force companies <strong>to</strong> declare insolvency for claims they<br />
would have <strong>to</strong> pay far in the future;<br />
• By requiring a capital increase from insurance and<br />
reinsurance companies, regula<strong>to</strong>rs force companies<br />
<strong>to</strong> de-leverage their balance sheet, which would fuel<br />
the crisis.<br />
This type of action would immobilize a huge amount<br />
of additional capital, which increases the cost of<br />
the protections provided by insurers and reinsurers.<br />
Furthermore, it would dry out the capital available<br />
for the rest of the economy, further weakening nonfinancial<br />
companies, which would, therefore, reduce the<br />
quality of the asset portfolio of insurers and reinsurers,<br />
reinforcing the vicious circle.<br />
REGULATORS AND COMPANIES COME UP<br />
WITH AD HOC CHANGES IN THE RULES<br />
Several solutions were proposed <strong>to</strong> escape from the<br />
dilemma of pro-cyclicality:<br />
• Use of a swap rate <strong>to</strong> create liquidity premium, which<br />
allows companies <strong>to</strong> account for liquidity <strong>risk</strong>s;<br />
• Give companies negative capital <strong>to</strong> compensate for<br />
the capital increase;<br />
• Suspend the solvency rules <strong>based</strong> on <strong>risk</strong> models <strong>to</strong><br />
enable companies <strong>to</strong> survive.<br />
Such proposals could cast doubts on the validity and<br />
the usefulness of modeling the <strong>risk</strong>s of the company. In<br />
fact, such measures contradict the <strong>risk</strong>-<strong>based</strong> regulation<br />
principles. On the one hand, such tricks are dangerous.<br />
On the other hand, the system has <strong>to</strong> develop a certain<br />
flexibility. It is worth noting that the parameters of<br />
models are very sensitive. So if companies start changing<br />
the parameters, they can obtain any result they want,<br />
and this is not what they should be doing. They should<br />
not act as if there were no crisis, because then their<br />
model will defeat its purpose, which is <strong>to</strong> look at the<br />
<strong>risk</strong>. On the contrary, they have <strong>to</strong> accept that there is<br />
more <strong>risk</strong>. They should accept this and they should live<br />
through it. After all, if companies have capital, it is <strong>to</strong> be<br />
used in times of crisis.
III. Giving the regula<strong>to</strong>ry system<br />
sufficient flexibility<br />
<strong>to</strong> respond <strong>to</strong> the next crisis<br />
The system must have a certain amount of flexibility.<br />
In such confusing situations simple rules are necessary,<br />
simple rules that do not cast doubts on the outcome<br />
of the models.<br />
RECOGNIZING THE RISKINESS<br />
OF THE SITUATION<br />
The regula<strong>to</strong>rs’ answers <strong>to</strong> the crisis are <strong>to</strong> propose an<br />
indiscriminate increase of the target capital for insurance<br />
companies. Such requirements, if they go through, would<br />
have severe social consequences by drying up even more<br />
market liquidity and increasing the price of insurance.<br />
• What is required in such a context is simply the<br />
recognition that the situation is <strong>risk</strong>ier.<br />
• Capital requirements should be adaptive and should<br />
change according <strong>to</strong> the economic situation.<br />
Moreover, empirical studies have shown that a measure<br />
like Value-at-<strong>Risk</strong> ex-post (1) is lower during crises<br />
because the market corrects itself and the probability of<br />
rebounds increases after a big downward movement.<br />
CHANGING THE RISK MEASURE THRESHOLD<br />
Solvency II requires that the capital be computed using<br />
the Value-at-<strong>Risk</strong> at a threshold of 99.5%. This threshold<br />
has been chosen arbitrarily <strong>to</strong> be in the tail of the<br />
distribution. There is no objective reason <strong>to</strong> take 99.5%<br />
instead of 99% or 98%. This threshold does not need<br />
<strong>to</strong> be constant all the time.<br />
120%<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
28.02.1873<br />
Yearly volatility (monthly data)<br />
120%<br />
100%<br />
28.02.1877<br />
80%<br />
60%<br />
40%<br />
20%<br />
28.02.1881<br />
Return and volatility of US s<strong>to</strong>ck market (1870-2009)<br />
0%<br />
-150% -100% -50% 0% 50%<br />
YoY return<br />
28.02.1885<br />
28.02.1889<br />
Source: IMF (2009).<br />
28.02.1893<br />
28.02.1897<br />
28.02.1901<br />
28.02.1905<br />
26.02.1909<br />
28.02.1913<br />
28.02.1917<br />
100%<br />
28.02.1921<br />
150%<br />
27.02.1925<br />
28.02.1929<br />
28.02.1933<br />
26.02.1937<br />
28.02.1941<br />
• Moving from 99.5% <strong>to</strong> 99% is a simple and transparent<br />
way <strong>to</strong> recognize the reality of the economic<br />
situation: nobody can be so safe any more when the<br />
whole world is in turmoil.<br />
• Furthermore, this change will mainly compensate<br />
the increase in cost of capital due <strong>to</strong> the model’s<br />
recognition of the higher <strong>risk</strong>s.<br />
In SCOR’s model, it would represent a decrease of roughly<br />
10% of the required capital, compensating the increase<br />
due <strong>to</strong> higher market volatility and lower interest rates.<br />
The supervisory authorities should set a threshold<br />
sufficiently remote <strong>to</strong> inspire confidence in the system<br />
by all stakeholders, within a predefined range, in order<br />
<strong>to</strong> react quickly if an indica<strong>to</strong>r shows an economic<br />
degradation close <strong>to</strong> a pre-crisis.<br />
WHICH INDICATOR WOULD ENABLE<br />
AN OBJECTIVE ASSESSMENT<br />
OF THE CRISIS SITUATION?<br />
The volatility of s<strong>to</strong>ck indices is a relevant indica<strong>to</strong>r <strong>to</strong><br />
detect a crisis situation. Market volatility affects the <strong>risk</strong>s<br />
generated by the Economic Scenario Genera<strong>to</strong>r (ESG) and<br />
thus ends up requiring higher capital from the model.<br />
Figure 15 presents the yearly volatility of the s<strong>to</strong>ck index<br />
since 1870. The volatility is more correlated with negative<br />
return than positive return. When a crisis occurs,<br />
the volatility increases faster than in periods of calm<br />
and the market reaches twice the average volatility of<br />
roughly 40%. This situation has only happened twice,<br />
in 1929 (120%) and in 2008 (80%).<br />
(1) VaR ex-post means the Value-at-<strong>Risk</strong> measured at a certain point<br />
in the past.<br />
Fig. 15: Market volatility has reached such values only twice in the past 140 years<br />
Yearly volatility of US s<strong>to</strong>ck market since 1870 (monthly data)<br />
28.02.1945<br />
28.02.1949<br />
27.02.1953<br />
28.02.1957<br />
28.02.1961<br />
26.02.1965<br />
28.02.1969<br />
28.02.1973<br />
28.02.1977<br />
27.02.1981<br />
28.02.1985<br />
28.02.1989<br />
26.02.1993<br />
28.02.1997<br />
28.02.2001<br />
28.02.2005<br />
27.02.2009<br />
SCOR - November 2010 - 21
A SIMPLE RULE TO MANAGE<br />
A CRISIS SITUATION<br />
This article proposes a simple rule <strong>to</strong> manage a crisis<br />
situation. Regula<strong>to</strong>rs should declare that companies<br />
would be allowed <strong>to</strong> use a Value-at-<strong>Risk</strong> (VaR) at 99%<br />
for the next year if the s<strong>to</strong>ck index yearly volatility<br />
reaches twice its his<strong>to</strong>rical average measured over a<br />
very long period.<br />
A year later, if the volatility is below this index, the<br />
regula<strong>to</strong>rs would then res<strong>to</strong>re the 99.5% threshold<br />
and ask companies <strong>to</strong> refurbish their capital <strong>to</strong> comply<br />
with it.<br />
Such a rule would allow insurance and reinsurance<br />
companies <strong>to</strong> use part of their capital <strong>to</strong> face up <strong>to</strong><br />
the <strong>to</strong>ugh economic situation without running the <strong>risk</strong><br />
of becoming insolvent for liabilities they would have <strong>to</strong><br />
pay in the distant future.<br />
Such a flexible system combines three advantages:<br />
1. It solves the dilemma of pro-cyclicality;<br />
2. It reduces the need <strong>to</strong> lock up useless extra capital;<br />
3. It enables companies <strong>to</strong> analyze a crisis situation with<br />
transparency and objectivity.<br />
22 - November 2010 - SCOR<br />
Conclusion<br />
Crises are unforeseeable and reveal links that were<br />
previously underestimated. Reinsurance and insurance<br />
companies undoubtedly need <strong>to</strong> learn from all crises in<br />
order <strong>to</strong> reduce the <strong>risk</strong> of the same causes producing<br />
the same effects.<br />
It is essential <strong>to</strong> adapt the solvency regulations <strong>to</strong> the<br />
occurrence of crises and their seriousness with simple<br />
adaptive rules, such as modifying the threshold of the<br />
<strong>risk</strong> measure. Indeed, this is a simple way <strong>to</strong> res<strong>to</strong>re<br />
confidence in the system because it recognizes the<br />
objective situation and gives companies time <strong>to</strong> adapt<br />
and refurbish their solvency ratio.<br />
It is worth utilizing an economic indica<strong>to</strong>r such as<br />
the extreme volatility of financial markets because it<br />
avoids blaming any stakeholders for the decision and<br />
highlights the objective situation.<br />
Being prepared for the next crisis means integrating<br />
in<strong>to</strong> the models a reasonable probability that a crisis will<br />
happen and developing dynamic strategies <strong>to</strong> survive it.<br />
The Romans used <strong>to</strong> say: “si vis pacem para bellum” (if you<br />
want peace prepare for war). One will paraphrase them by<br />
saying: “if you want <strong>to</strong> survive a crisis prepare for it”.
SCOR - November 2010 - 23
3<br />
REINSURANCE OPTIMIZATION<br />
IN THE CONTEXT OF CAPITAL<br />
MANAGEMENT EVA SCHLÄPFER DE MONTMOLLIN<br />
Senior <strong>Risk</strong> Consultant, SCOR<br />
MAGDALENA KLAPPER-RYBICKA<br />
<strong>Risk</strong> Consultant, SCOR<br />
This article outlines the different<br />
motivations an insurance company may have for<br />
purchasing reinsurance. It then introduces the<br />
capital management view, as a driver for structuring<br />
a reinsurance program. From this perspective,<br />
specific reinsurance valuation criteria are investigated<br />
in detail using an example portfolio.<br />
I. Various criteria for the<br />
valuation of a reinsurance<br />
program<br />
Insurance companies may have different reasons for<br />
buying reinsurance:<br />
• To enhance profit and portfolio management, <strong>to</strong><br />
decrease the volatility of the portfolio, <strong>to</strong> increase<br />
diversification benefits, <strong>to</strong> protect profitability.<br />
• To support the company’s capital management strategy<br />
in order <strong>to</strong> achieve the solvency capital requested<br />
by the regula<strong>to</strong>rs, <strong>to</strong> satisfy the capital requirements<br />
of the rating agencies, or simply <strong>to</strong> provide better<br />
capacity.<br />
• The reinsurance purchase may also be cost driven:<br />
insurers buy reinsurance if expected recoveries exceed<br />
the reinsurance premium.<br />
• Various other reasons, such as: <strong>to</strong> continue the<br />
previous year’s reinsurance program; <strong>to</strong> benefit from<br />
the reinsurers’ know-how or services (e.g. product<br />
development, medical underwriting, etc.).<br />
24 - November 2010 - SCOR<br />
The purchase of reinsurance, therefore, may also be<br />
considered as a <strong>risk</strong> management <strong>to</strong>ol that enables<br />
insurers <strong>to</strong> lower their capital requirements in the short<br />
term. Insurers also have other means at their disposal<br />
<strong>to</strong> mitigate their overall business <strong>risk</strong>, such as: changing<br />
their investment strategy, raising capital and changing<br />
their underwriting policy. However, these methods take<br />
longer <strong>to</strong> implement.<br />
Figure 16: <strong>Risk</strong> <strong>Management</strong><br />
for Insurance Companies<br />
<strong>Risk</strong> mitigation instruments<br />
Short-term<br />
Long-term<br />
Buy reinsurance<br />
Change investment strategy<br />
Raise capital<br />
Change underwriting policy<br />
Different possibilities for mitigating overall business <strong>risk</strong>s<br />
according <strong>to</strong> the time needed for their application and effect.<br />
The various motivations for buying reinsurance determine<br />
the criteria <strong>to</strong> be applied <strong>to</strong> the evaluation and<br />
comparison of different reinsurance programs.<br />
Depending on the motivation involved, the best<br />
reinsurance program could be any of the following:<br />
• The program offering the best profit protection,<br />
the lowest volatility, the best diversification benefit or<br />
the highest profit for insurer and reinsurer, taking the<br />
cost of capital in<strong>to</strong> account (all of which corresponds
<strong>to</strong> the portfolio management and profit optimization<br />
perspective).<br />
• The program with the lowest required capital resulting<br />
from the company’s internal model, the lowest Solvency<br />
II standard model capital requirement or the lowest<br />
rating agency capital (if the capital management view<br />
is taken).<br />
• The program with the lowest cost of reinsurance -<br />
premium minus recovery (cost driven policy).<br />
Each criterion defined for the evaluation of a reinsurance<br />
program (independent of its nature) should definitely<br />
incorporate the costs of the reinsurance program, in<br />
order <strong>to</strong> produce a meaningful and realistic result. In a<br />
further step this article will focus on reinsurance valuation<br />
as a part of capital management practices.<br />
II. Valuation of a reinsurance<br />
program in the context<br />
of capital management<br />
In order <strong>to</strong> look at reinsurance valuation from a capital<br />
management perspective, this section will start with<br />
some general definitions.<br />
Fig. 17: The capital of an insurance company<br />
Economically adjusted capital:<br />
Available Capital<br />
Discount in loss<br />
reserves<br />
taxes<br />
Unrealized<br />
capital gains Latent<br />
Capital as reported<br />
in financial statement<br />
An insurance company receives capital from its<br />
shareholders and utilizes that capital <strong>to</strong> guarantee<br />
that it will pay claims – up <strong>to</strong> a certain probability. The<br />
primary focus of capital is, then, not <strong>to</strong> provide finance,<br />
but <strong>to</strong> absorb the <strong>risk</strong>s undertaken. Capital can thus be<br />
seen as a “commodity” used <strong>to</strong> produce the company’s<br />
business.<br />
It is necessary <strong>to</strong> distinguish between two definitions<br />
of capital: available capital and <strong>risk</strong>-<strong>based</strong> capital (or<br />
required capital).<br />
Available capital or economic capital is the capital a<br />
company has at its disposal (equity, long-term debts,<br />
unrealized gains, discount in loss reserves, latent taxes,<br />
etc). Conversely, the <strong>risk</strong>-<strong>based</strong> capital or required capital<br />
is the minimum amount of capital that the insurer<br />
needs <strong>to</strong> cover the <strong>risk</strong> taken.<br />
The difference between the two is called the “signaling<br />
capital” and corresponds <strong>to</strong> the economic resources<br />
available <strong>to</strong> develop more business or take on more<br />
<strong>risk</strong>s. As illustrated in Figure 17, the amount of reinsurance<br />
involved has an impact on the balance between<br />
required and signaling capital. It also has a slight impact<br />
on the available capital, due <strong>to</strong> the cost of reinsurance.<br />
This can be seen on the left-hand side of the graph.<br />
Capital required given management’s<br />
<strong>risk</strong> appetite framework: Required Capital<br />
Signaling Capital<br />
Required Capital<br />
for underwriting <strong>risk</strong><br />
Required Capital<br />
for investment <strong>risk</strong>s<br />
Required Capital<br />
for other <strong>risk</strong>s<br />
More reinsurance<br />
Less reinsurance<br />
SCOR - November 2010 - 25
The amount of required capital is determined by the<br />
insurance company’s <strong>risk</strong> appetite, in other words the<br />
amount of capital the company is willing <strong>to</strong> <strong>risk</strong> in a<br />
certain time horizon. Other elements are the proper<br />
definition of required capital and a capital allocation<br />
policy. These are prerequisites for the optimization of<br />
the portfolio’s profitability, because they build the foundations<br />
on which <strong>to</strong> measure the true performance of<br />
the business. Hence they have <strong>to</strong> be incorporated at the<br />
core of business processes.<br />
Further discussion of <strong>risk</strong> allocation would be outside<br />
the scope of this article, however the definition of<br />
required capital will be addressed in more detail.<br />
There are many different definitions of required capital.<br />
Clear definitions should ensure consistent use of terminology<br />
within the organization. The definition of required<br />
capital may be developed <strong>based</strong> on various mathematical<br />
concepts and should address the <strong>risk</strong> aversion of the company<br />
in question. SCOR has its own definition of required<br />
capital as used in its internal model.<br />
Let us begin with standalone <strong>risk</strong> capital. The asset and<br />
liability portfolio builds a set of <strong>risk</strong> classes: {Xi}, e.g.<br />
bonds, equities, Nat Cat, Mo<strong>to</strong>r, etc. The standalone<br />
(or undiversified) <strong>risk</strong> capital of a class is the difference<br />
between two values on the probability distribution of<br />
annual economic profit. These two values are:<br />
• The expected shortfall at 1% quantile of the distribution<br />
• The expected value of the distribution<br />
E(X i | X i < 1% percentile) – E(X i )<br />
This definition is one of several possible definitions.<br />
It expresses the shareholders’ point of view, since it<br />
measures how far the company deviates from expected<br />
profit (see Figure 18). By way of comparison, the Swiss<br />
Solvency Test definition takes zero profit as its reference<br />
point. Thus, regula<strong>to</strong>rs represent the clients’ view and<br />
therefore seek <strong>to</strong> know whether or not a company can<br />
pay its obligations.<br />
Solvency II provides another definition of required capital.<br />
The directive applies the Value-at-<strong>Risk</strong> at the 0.5%<br />
quantile. This <strong>risk</strong> measure corresponds <strong>to</strong> a point on<br />
profit distribution (i.e. not the expected shortfall) and<br />
therefore does not incorporate the <strong>risk</strong> profile below<br />
the chosen threshold. One advantage of this is that it is<br />
easier <strong>to</strong> compute analytically, with no need <strong>to</strong> apply a<br />
Monte Carlo simulation.<br />
SCOR has <strong>based</strong> its definition of required capital on the<br />
expected shortfall at the 1% quantile. It can be demonstrated<br />
empirically that required capital measured as<br />
Value-at-<strong>Risk</strong> at 0.5% is close <strong>to</strong> required capital defined<br />
with expected shortfall at the 1% quantile, so that the<br />
<strong>risk</strong> capital as measured by SCOR is in line with the<br />
Solvency II regula<strong>to</strong>ry view. The <strong>risk</strong> capital definition<br />
used by SCOR is coherent and its resulting mathematical<br />
features (additivity) allow the required capital <strong>to</strong> be allocated<br />
<strong>to</strong> the single <strong>risk</strong> classes. This leads <strong>to</strong> the definition<br />
of diversified capital.<br />
Figure 19 provides a closer look at the tail end of the<br />
profit distribution already presented in Figure 18. Since<br />
the profit distribution is the result of a Monte Carlo<br />
simulation, it is possible <strong>to</strong> analyze the single worst-case<br />
scenarios, and in particular the contributions <strong>to</strong> the<br />
scenarios of the various different <strong>risk</strong> classes.<br />
Diversified <strong>Risk</strong> Capital is defined as a contribution<br />
(measured in terms of average) of a class of <strong>risk</strong> <strong>to</strong> the<br />
expected shortfall of the company overall profitability<br />
loss distribution (Z):<br />
E(X i | Z < 1% percentile) – E(X i )<br />
Diversified Capital is the capital amount that SCOR<br />
needs <strong>to</strong> keep for a certain <strong>risk</strong> class in a given portfolio,<br />
in order <strong>to</strong> be able <strong>to</strong> pay its obligations.<br />
Fig. 18: Definition of standalone capital as the difference between<br />
mean expected value and 1% expected shortfall<br />
26 - November 2010 - SCOR<br />
Standalone <strong>Risk</strong> Capital<br />
1% ES 100% Expected value<br />
1% VaR<br />
10%<br />
Profit Distribution<br />
0%<br />
-220 -170 -120 -70 -20 30 80<br />
90%<br />
80%<br />
70%<br />
60%<br />
50%<br />
40%<br />
30%<br />
20%
Fig. 19: Diversified <strong>Risk</strong> Capital<br />
-250 -200 -150 -100 -50 0 50<br />
With this definition, the main <strong>risk</strong> contribu<strong>to</strong>rs can be<br />
identified as shown in Figure 19. The <strong>to</strong>tal required<br />
capital for the portfolio is around EUR -90 million and<br />
can be divided in<strong>to</strong> four <strong>risk</strong> classes. The highest <strong>risk</strong><br />
comes from classes 4 and 1, with diversified <strong>risk</strong> capital<br />
of EUR -44 million and EUR -26 million respectively.<br />
The contribution for the other <strong>risk</strong>s is lower.<br />
It is interesting <strong>to</strong> note that some <strong>risk</strong>s in certain scenarios<br />
show a positive contribution <strong>to</strong> the overall <strong>risk</strong>.<br />
This may occur due <strong>to</strong> the low dependency of these<br />
<strong>risk</strong>s on other <strong>risk</strong> classes.<br />
The example also demonstrates that the contribution<br />
of some <strong>risk</strong> classes may vary widely for different<br />
Diversified<br />
<strong>Risk</strong> Capital<br />
Definition of diversified <strong>risk</strong> capital as a contribution of a single <strong>risk</strong> class <strong>to</strong> the aggregated shortfall. Due <strong>to</strong> the graphical limitations,<br />
the diagram only contains around 20-30 worst-case scenarios, (not the whole 1%). The blue curve corresponds <strong>to</strong> the profit distribution.<br />
For each scenario the <strong>to</strong>tal <strong>risk</strong> may be split in<strong>to</strong> certain <strong>risk</strong> classes. The mean is therefore computed separately for each <strong>risk</strong> class.<br />
Fig. 20: <strong>Risk</strong> drivers<br />
<strong>Risk</strong> 1<br />
<strong>Risk</strong> 2<br />
<strong>Risk</strong> 3<br />
<strong>Risk</strong> 4<br />
Total<br />
scenarios. To better understand the contribution made<br />
by <strong>risk</strong> classes <strong>to</strong> scenarios of differing frequency, it is<br />
necessary <strong>to</strong> make a distinction between <strong>risk</strong> capital<br />
and <strong>risk</strong> drivers.<br />
<strong>Risk</strong> drivers are measured <strong>based</strong> on a 5% shortfall,<br />
rather than the 1% shortfall used for <strong>risk</strong> capital:<br />
E(X i | Z < 5% percentile) – E(X i )<br />
They reflect more frequent <strong>risk</strong>s, while <strong>risk</strong> capital<br />
focuses more on peak or extreme <strong>risk</strong>s. The contribution<br />
made by different <strong>risk</strong> drivers may be quite different<br />
from their contribution <strong>to</strong> the required capital, as<br />
illustrated in Figure 20.<br />
0%<br />
-220 -170 -120 -70 -20 30 80<br />
<strong>Risk</strong> 1 <strong>Risk</strong> 2<br />
<strong>Risk</strong> 3 <strong>Risk</strong> 4<br />
1% ES 5% ES 100% Mean<br />
80%<br />
60%<br />
40%<br />
20%<br />
<strong>Risk</strong> Driver<br />
<strong>Risk</strong> Capital<br />
<strong>Risk</strong> driver and <strong>risk</strong> capital reflect <strong>risk</strong>s of differing frequency. Various <strong>risk</strong> classes may contribute differently <strong>to</strong> more or less frequent <strong>risk</strong>s.<br />
For example, <strong>risk</strong> class 4 (in pastel blue) contributes much more <strong>to</strong> the peak <strong>risk</strong> than <strong>to</strong> the <strong>risk</strong> driver, while <strong>risk</strong> class 2 (in dark blue)<br />
does the opposite. The same thing can also be seen in Figure 18.<br />
-26<br />
-9<br />
-7<br />
-44<br />
-86<br />
SCOR - November 2010 - 27
Having defined the <strong>risk</strong> measures and how <strong>to</strong> allocate<br />
<strong>risk</strong> capital <strong>to</strong> particular <strong>risk</strong> classes, one can now turn<br />
<strong>to</strong> the criteria for evaluating a reinsurance program.<br />
Reinsurance is a way <strong>to</strong> reduce required capital, and<br />
the valuation criterion proposed accounts for the costs<br />
of reinsurance including the cost of capital.<br />
How is a reinsurance program evaluated in this context?<br />
On the one hand, the insurance company purchasing a<br />
reinsurance program pays a corresponding reinsurance<br />
premium and expects some recoveries and related costs.<br />
On the other hand, a reinsurance program is a way <strong>to</strong><br />
reduce capital requirements, which consequently also<br />
lowers the cost of the capital that the company needs<br />
<strong>to</strong> keep in order <strong>to</strong> pay its obligations. Therefore, this<br />
criterion compares the cost of reinsurance <strong>to</strong> the cost<br />
of capital <strong>to</strong> be kept for the portfolio.<br />
Fig. 21: Reinsurance valuation<br />
criteria used<br />
Cost of<br />
capital on<br />
required capital<br />
Cost of Reinsurance (RI):<br />
RI premium<br />
- RI Recoveries<br />
III. Example of two reinsurance<br />
structures<br />
This section will continue <strong>to</strong> look at valuation criteria<br />
using a practical example.<br />
Figure 22 shows a realistic example portfolio which is<br />
characterized by a large P&C book, mainly in Property,<br />
Casualty and Mo<strong>to</strong>r. The <strong>to</strong>tal required capital is around<br />
EUR 95 million, primarily relating <strong>to</strong> the P&C new<br />
business book. Assets are mostly invested in bonds<br />
and real estate and they do not constitute a very high<strong>risk</strong><br />
contribu<strong>to</strong>r.<br />
BEFORE APPLYING THE REINSURANCE<br />
STRUCTURE<br />
Let us briefly analyze the gross performance of the<br />
main <strong>risk</strong> contribu<strong>to</strong>r, the P&C new business book (see<br />
Figure 23). The standalone required capital for P&C<br />
new business amounts <strong>to</strong> EUR 92.9 million. Diversified<br />
with the whole portfolio, this <strong>risk</strong> class requires EUR<br />
52.4 million, giving a diversification benefit of 43.5%.<br />
The insurance company needs capital <strong>to</strong> keep this<br />
book, which involves some cost. Assuming that the<br />
cost of capital rate is 11.5%, the cost of the capital<br />
needed amounts <strong>to</strong> EUR 6.03 million. This cost should<br />
be reflected in the assessment of the P&C new business<br />
profit, which amounts <strong>to</strong> EUR 9.7 million.<br />
Fig. 22: Example portfolio – gross P&C book with main <strong>risk</strong> categories<br />
and composition of inverted assets<br />
<strong>Risk</strong> Category<br />
in EURm<br />
28 - November 2010 - SCOR<br />
Reinsurance valuation<br />
criteria optimizes capital<br />
while also taking in<strong>to</strong><br />
account the pure cost<br />
of reinsurance.<br />
Exposure Fully Diversified Capital Exposure definition<br />
Gross Non-Life New Business 210 52.4 Written Premium<br />
Gross Non-Life Reserves 240 24.0 Claims Reserves+UPR<br />
Invested Assets 83 10.1 Market Value<br />
Credit 107 5.4 Available Capital<br />
Other <strong>Risk</strong>s 107 0.9 Available Capital<br />
Operational <strong>Risk</strong> 107 1.5 Available Capital<br />
Total RBC 94.3<br />
Asset Class Market Value Fully Diversified Capital<br />
Equity 13 2.86<br />
Real Estate 40 6.06<br />
Fixed Income 29 1.18<br />
Total Assets 82 10.1
Fig. 23: P&C new business book – gross required capital,<br />
related cost of capital and profit (in €m)<br />
P&C NB<br />
Gross<br />
Standalone<br />
<strong>Risk</strong> Capital<br />
92.9<br />
P&C NB Gross Profit<br />
excluding CoC<br />
15.8<br />
P&C NB<br />
Diversification<br />
Benefit: 43.5%<br />
P&C NB Gross Profit<br />
including CoC<br />
9.7<br />
To gain a concrete view of the impact of reinsurance on<br />
diversification and capital requirements, a reinsurance<br />
program is applied <strong>to</strong> the P&C new business book with<br />
a structure identical <strong>to</strong> that shown in Figure 24.<br />
As already mentioned, this P&C new business book<br />
consists mostly of Casualty, Mo<strong>to</strong>r and Property lines;<br />
it therefore requires a specific reinsurance structure.<br />
The program is characterized primarily by a quota share<br />
contract for the Casualty and Mo<strong>to</strong>r lines, which is<br />
followed by XL cover. The Property line is protected with<br />
the 3-layer cover; the 2-layer Cat program completes the<br />
reinsurance structure.<br />
APPLYING THE REINSURANCE STRUCTURE<br />
TO THE P&C NEW BUSINESS PORTFOLIO<br />
An analysis of the net portfolio gives the following<br />
picture: net standalone required capital for the P&C<br />
P&C NB<br />
Gross<br />
Diversified<br />
<strong>Risk</strong> Capital<br />
52.4<br />
Total Gross Capital = 94.3<br />
Costs of Capital<br />
allocated <strong>to</strong> P&C NB<br />
6.03*<br />
Operational <strong>Risk</strong><br />
Other <strong>Risk</strong>s<br />
Credit<br />
Invested Assets<br />
Gross Non-Life Reserves<br />
Gross Non-Life New Business<br />
* CoC Rate = 11.5%<br />
new business book has obviously decreased due <strong>to</strong> the<br />
reinsurance structure and amounts <strong>to</strong> EUR 54.8 million.<br />
The net profit without considering the cost of capital is<br />
of course lower than the gross profit, since in this view<br />
reinsurance naturally appears as an additional cost.<br />
Within the overall portfolio, the net required capital for<br />
this <strong>risk</strong> class amounts <strong>to</strong> EUR 23.9 million, generating a<br />
better (compared <strong>to</strong> the gross portfolio) diversification<br />
effect equal <strong>to</strong> 56%. The overall portfolio is better<br />
balanced with this reinsurance structure. The cost<br />
of capital needed for the P&C new business book<br />
is now also reduced due <strong>to</strong> the lower amount of<br />
diversified capital, and amounts <strong>to</strong> EUR 2.8 million.<br />
With adjustment for cost of capital, the net profit of<br />
this book is slightly over EUR 10 million (see Figure 25<br />
for details).<br />
Fig. 24: P&C new business book and current reinsurance program (in €m)<br />
P&C Line Gross Premium RI Structure<br />
Casualty 70 QS<br />
Per <strong>Risk</strong> XL2: 60 xs 60<br />
Mo<strong>to</strong>r 50<br />
25%<br />
Per <strong>Risk</strong> XL1: 30 xs 30<br />
Property Cat 10<br />
Total Property 80<br />
Total P&C 210<br />
CAT XL2: 60 xs 40<br />
CAT XL1: 30 xs 10<br />
Per <strong>Risk</strong> XL3: 20 xs 100<br />
Per <strong>Risk</strong> XL2: 50 xs 50<br />
Per <strong>Risk</strong> XL1: 40 xs 10<br />
Portfolio<br />
characteristics:<br />
• Quota share before<br />
XL for Casualty and<br />
Mo<strong>to</strong>r combined<br />
• Property covers<br />
are XL<br />
SCOR - November 2010 - 29
Fig. 25: P&C new business book – net required capital,<br />
related cost of capital and profit (in €m)<br />
The question that arises now is whether or not this program<br />
is efficient due <strong>to</strong> the chosen valuation criteria.<br />
As mentioned, the application of the reinsurance<br />
program lowered the P&C new business book’s profit<br />
by around EUR 2.5 million (due <strong>to</strong> the costs of<br />
reinsurance). On the other hand the mitigation of the<br />
peak <strong>risk</strong>s, which resulted in a lower standalone capital<br />
as well as in improved diversification with the rest of<br />
the portfolio, lowered the cost of the capital needed<br />
<strong>to</strong> retain this EUR 3.2 million <strong>risk</strong>. Both effects should<br />
be considered <strong>to</strong>gether, as shown in Figure 26. As a<br />
consequence the profit of this book including cost<br />
Fig. 26: Impact of the reinsurance program on the profit of the P&C<br />
new business book including cost of capital (in €m)<br />
30 - November 2010 - SCOR<br />
P&C NB Net<br />
Standalone<br />
<strong>Risk</strong> Capital<br />
15.8<br />
54.8<br />
P&C NB Net Profit<br />
excluding CoC<br />
13.3<br />
13.3<br />
Gross<br />
Net<br />
P&C NB Profit<br />
excluding CoC<br />
P&C NB<br />
Diversification<br />
Benefit: 56.3%<br />
10.5<br />
P&C NB Net Profit<br />
including CoC<br />
Costs of RI<br />
2.5<br />
Total Net Capital = 71.2<br />
Costs of Capital<br />
allocated <strong>to</strong> P&C NB<br />
2.8*<br />
<<br />
P&C NB Net<br />
Diversified<br />
<strong>Risk</strong> Capital<br />
23.9<br />
of capital (the value chosen as a valuation criterion)<br />
has increased, which justifies the application of this<br />
reinsurance structure.<br />
Even though the cost of capital that the insurer<br />
can save with this reinsurance program exceeds<br />
the cost of reinsurance, the agreement is probably<br />
also advantageous for the reinsurer. Reinsurers are<br />
presumably better able <strong>to</strong> diversify their <strong>risk</strong>s than<br />
direct insurers, thus for the reinsurer, the cost of capital<br />
resulting from the transferred <strong>risk</strong> is generally lower<br />
than the cost of capital resulting from the same <strong>risk</strong><br />
for the direct insurer.<br />
Saved CoC<br />
3.2<br />
6.03<br />
Operational <strong>Risk</strong><br />
Other <strong>Risk</strong>s<br />
Credit<br />
Invested Assets<br />
Gross Non-Life Reserves<br />
Gross Non-Life New Business<br />
* CoC Rate = 11.5%<br />
2.8<br />
Gross Net<br />
P&C NB<br />
Costs of Capital
Fig. 27: Optimal reinsurance, win-win situation (in €m)<br />
Reinsurer<br />
Portfolio<br />
Impact of new <strong>risk</strong><br />
on reinsurer portfolio<br />
Costs of RI**<br />
2.5<br />
** CoC Rate = 12%<br />
P&C NB Gross<br />
Standalone<br />
<strong>Risk</strong> Capital<br />
Transferred<br />
<strong>Risk</strong><br />
P&C NB<br />
Net<br />
Standalone<br />
<strong>Risk</strong> Capital<br />
54.8<br />
<<br />
This concept is illustrated in Figure 27. Following the<br />
previous example, part of the P&C new business book<br />
is transferred <strong>to</strong> the reinsurer. On the right-hand side<br />
one can see the relief of the insurer’s diversified capital,<br />
which enables certain savings in the cost of capital. It<br />
should be kept in mind that this capital relief depends<br />
on the Insurer’s <strong>risk</strong> appetite, which determines the<br />
definition of <strong>risk</strong> capital used. It is also interesting <strong>to</strong><br />
note that the capital relief for this <strong>risk</strong> class may (and<br />
probably will) influence the capital required for other<br />
<strong>risk</strong> classes, which can also be observed in Figure 27.<br />
The stronger diversification effect of the transferred <strong>risk</strong><br />
within the reinsurer portfolio is visible on the left-hand<br />
side of the graph. All in all, the cost of capital for the<br />
insurer, even though advantageous, is greater than the<br />
cost of capital that the reinsurer needs in order <strong>to</strong> retain<br />
the transferred <strong>risk</strong> (with comparable cost of capital rates<br />
for both parties), which creates a win-win situation.<br />
APPLYING THE ALTERNATIVE REINSURANCE<br />
STRUCTURE TO THIS PORTFOLIO<br />
This reinsurance structure suffers from certain drawbacks.<br />
Cover for the Property line is limited <strong>to</strong> EUR 120 million,<br />
whereas the underlying gross model presupposes potentially<br />
higher losses. The cover is therefore “<strong>to</strong>o short”<br />
for the underlying gross portfolio. The second problem<br />
arises from the unfavourable conditions of the quota<br />
share contract for Casualty and Mo<strong>to</strong>r lines (in particular<br />
due <strong>to</strong> the rather low ceding commission). Consequently,<br />
it is clear that this program could be improved.<br />
Change of<br />
diversified capital<br />
Costs of Capital<br />
Saved with RI*<br />
3.2<br />
* CoC Rate = 11.5%<br />
Definition of <strong>Risk</strong> Capital reflects<br />
the company’s <strong>risk</strong> appetite,<br />
Costs of Capital saved with RI<br />
depend also on the chosen percentile<br />
P&C NB Net<br />
Fully Diversified<br />
<strong>Risk</strong> Capital<br />
Total Capital<br />
Net = 71.2 Gross = 94.3<br />
Reinsurer-Cedant win-win situation. Reinsurer benefits from the stronger diversification of the transferred <strong>risk</strong> within its portfolio.<br />
Fig. 28: An alternative<br />
reinsurance program for the P&C<br />
new business book<br />
P&C Line Gross<br />
Premium<br />
Adjusted RI Structure<br />
Casualty 70 Per <strong>Risk</strong> XL2: 60 xs 60<br />
Mo<strong>to</strong>r 50 Per <strong>Risk</strong> XL1: 40 xs 20<br />
Property Cat 10<br />
Total Property 80<br />
Total P&C 210<br />
CAT XL2: 60 xs 40<br />
CAT XL1: 30 xs 10<br />
Per <strong>Risk</strong> XL3: 100 xs 100<br />
Per <strong>Risk</strong> XL2: 50 xs 50<br />
Per <strong>Risk</strong> XL1: 40 xs 10<br />
In response <strong>to</strong> these drawbacks, a slightly modified alternative<br />
reinsurance structure is proposed, as shown in<br />
Figure 28. The reinsurance cover for Casualty and Mo<strong>to</strong>r<br />
lines has been restructured so that the <strong>risk</strong> mitigation<br />
impact of the quota share is (partially) compensated by<br />
the lower attachment point of the non-proportional program,<br />
i.e. EUR 20 million rather than EUR 30 million.<br />
Secondly, the limit of cover for the Property line has been<br />
increased from EUR 120 <strong>to</strong> EUR 200 million so that the<br />
extreme <strong>risk</strong>s in the portfolio are also covered.<br />
SCOR - November 2010 - 31
Fig. 29: Comparing the two structures (in €m)<br />
Evaluation of both reinsurance structures using the proposed criteria. The adjustments <strong>to</strong> the reinsurance structure lowered the cost<br />
of reinsurance and at the same time enabled further savings in the cost of capital.<br />
The impact of these changes in the reinsurance structure<br />
should be analyzed, using valuation criteria previously<br />
proposed. As expected, replacing the quota share<br />
with the extended non-proportional cover significantly<br />
lowers the cost of reinsurance. Further extension of the<br />
Property cover <strong>to</strong> include peak <strong>risk</strong>s lowers the amount<br />
of required capital for this line, and consequently also<br />
the <strong>to</strong>tal cost of capital for this portfolio (Figure 29).<br />
To conclude, the adjusted reinsurance structure is more<br />
appropriate for the considered P&C new business<br />
portfolio due <strong>to</strong> the selected valuation criteria and<br />
measures adopted for the required capital (particularly<br />
with respect <strong>to</strong> the level of <strong>risk</strong> aversion).<br />
Now, it would be interesting <strong>to</strong> look at the influence<br />
of both reinsurance structures on the solvency ratio of<br />
the company in question.<br />
Fig. 30: Solvency ratios (in €m)<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
114%<br />
94<br />
Gross<br />
151% 153% Solvency ratio<br />
71<br />
107<br />
Original Net Adjusted Net<br />
Total Required Capital Available Capital<br />
Solvency ratio for the insurer: gross and net with both investigated reinsurance structures.<br />
32 - November 2010 - SCOR<br />
Original RI Structure<br />
Adjusted RI Structure<br />
P&C NB Profit<br />
excluding CoC<br />
13.3<br />
14.3<br />
2.5<br />
Costs of RI<br />
1.4<br />
Costs of RI<br />
70<br />
3.2<br />
Saved CoC<br />
3.4<br />
Saved CoC<br />
Costs of Capital<br />
P&C NB<br />
2.8<br />
2.6<br />
As indicated in the initial portfolio characteristics, the<br />
company has available capital of EUR 107 million at<br />
its disposal, with <strong>to</strong>tal required capital amounting<br />
<strong>to</strong> around EUR 94 million. For simplification purposes,<br />
it is assumed that the reinsurance structures do not<br />
materially impact the available capital, however<br />
they reduce the <strong>to</strong>tal required capital <strong>to</strong> EUR 71<br />
and EUR 70 million respectively. The solvency ratio is<br />
computed by dividing the required capital by available<br />
capital, as shown in Figure 30.<br />
The initial gross solvency ratio of 114% appears <strong>to</strong> be<br />
rather low and significantly improves with reinsurance,<br />
rising <strong>to</strong> 151%. Further refining of the reinsurance<br />
structure improves it a little further.
Fig. 31: Impact of the reinsurance refinement on the standalone<br />
required capital by <strong>risk</strong> class (in €m)<br />
... With the original reinsurance structure ... With the adjusted reinsurance structure<br />
NB P&C Standalone Capital<br />
Property NonCat<br />
Prop. Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
DETAILED ANALYSIS OF THE IMPACT<br />
OF THE REINSURANCE STRUCTURE<br />
ADJUSTMENTS ON THE REQUIRED CAPITAL<br />
First of all, it is necessary <strong>to</strong> look at the impact of the<br />
reinsurance refinement on the standalone required<br />
capital by <strong>risk</strong> class (see Figure 31). The overall standalone<br />
required capital for the P&C new business book<br />
has moderately decreased. The capital for the Property<br />
line has decreased naturally, due <strong>to</strong> the extended cover<br />
for peak <strong>risk</strong>s.<br />
Restructuring the reinsurance for Mo<strong>to</strong>r and Casualty<br />
caused a decrease in the standalone capital for the<br />
Mo<strong>to</strong>r line and a slight increase in capital for the<br />
Casualty line. This interesting effect can be explained by<br />
the different characteristics of the underlying models.<br />
NB P&C Standalone Capital<br />
Property NonCat<br />
Prop. Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
0 10 20 30 40 50 60 0 10 20 30 40 50 60<br />
More severe losses recover better from the non-proportional<br />
structure than from the quota share. For more<br />
frequent and less severe events, the capital reduction<br />
from the non-proportional program is lower than from<br />
the quota share. The standalone capital for the Cat line<br />
remains unchanged, as expected.<br />
The reduction in diversified capital for the P&C new<br />
business book is mainly due <strong>to</strong> lower diversified capital<br />
for the Property line (see Figure 32). The restructuring<br />
of the Mo<strong>to</strong>r and Casualty reinsurance cover gives<br />
another detailed picture. The contributions made by<br />
these lines in a diversified portfolio increase slightly,<br />
especially for Mo<strong>to</strong>r. The explanation given above also<br />
applies in this case.<br />
Fig. 32: Impact of the reinsurance refinement on the diversified<br />
required capital by <strong>risk</strong> class for the new book P&C Diversified Capital (in €m)<br />
... With the original reinsurance structure ... With the adjusted reinsurance structure<br />
Property NonCat<br />
Prop. Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
Property NonCat<br />
Prop. Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
0 2 4 6 8 10 12 0 2 4 6 8 10 12<br />
SCOR - November 2010 - 33
More detailed results for the Property cover are shown<br />
in Figure 33. In the graph on the left side, the extreme<br />
<strong>risk</strong>s exceeding the original reinsurance cover are visible<br />
(although not frequent). The 1% worst net <strong>risk</strong>s after<br />
application of the adjusted reinsurance cover, presented<br />
in the graph on the right side, do not contain any losses<br />
with such a high gross loss amount.<br />
More detailed results for the Mo<strong>to</strong>r and Casualty cover are<br />
shown in Figure 34. In the original cover (the graph on the<br />
left) the quota share reduces all gross losses and the nonproportional<br />
reinsurance only impacts the peak <strong>risk</strong>s. The<br />
adjusted non-proportional structure allows the retention of<br />
more small <strong>risk</strong>s; the retention line has a sharper increase<br />
(the graph on the right). The non-proportional cover acts<br />
not only on the peak <strong>risk</strong>s but also in the middle band.<br />
Fig. 33: Detail Results: Property Non-Catastrophe, aggregate losses<br />
Aggregate Losses - Gross vs. Net / 1% worst Net Losses<br />
Net Losses (EURm)<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
Original RI<br />
0<br />
0<br />
50 100 150 200 250 300 50 100 150 200 250 300<br />
Gross Losses (EURm)<br />
Gross Losses (EURm)<br />
Net Losses (EURm)<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
Adjusted RI<br />
Refinement of the Property cover – the graphs present 1% worst aggregated net losses for both reinsurance structures. For each of these<br />
losses the corresponding gross and net loss amount is graphed in order <strong>to</strong> visualize the impact from the appropriate reinsurance program.<br />
Fig. 34: Detail Results: Mo<strong>to</strong>r, aggregate losses<br />
Aggregate Losses - Gross vs. Net<br />
Net Losses (EURm)<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
Original RI<br />
retention<br />
XL<br />
0<br />
0 20 40 60 80 100 120 140 160<br />
Gross Losses (EURm)<br />
retention<br />
Refinement of the Mo<strong>to</strong>r and Casualty cover – the graphs represent aggregated losses for both reinsurance structures, in a gross vs. net loss<br />
amount view. The impact from quota share and non-proportional reinsurance can be clearly observed. Severe gross aggregate losses are<br />
significantly reduced if they are caused by rare but severe events (triggering the layer); and are reduced less if caused by the accumulation<br />
of more frequent and less severe <strong>risk</strong>s.<br />
34 - November 2010 - SCOR<br />
Net Losses (EURm)<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
Adjusted RI<br />
XL<br />
0<br />
0 20 40 60 80 100 120 140 160<br />
Gross Losses (EURm)
IV. Partial model and fully<br />
diversified capital<br />
The main difficulty that arises when assessing valuation<br />
criteria previously mentioned lies in the need <strong>to</strong> estimate<br />
the diversified required capital. Measuring the impact<br />
of structural reinsurance changes (in particular for Non-<br />
Proportional cover) on the diversified capital is not simple and<br />
usually requires an internal model <strong>based</strong> on a Monte Carlo<br />
simulation. This is not a straightforward process because<br />
the internal model is a holistic one, a “huge machine” that<br />
needs time <strong>to</strong> input data and compute results.<br />
SCOR has recently developed a Capital Deployment Tool<br />
(CaDeT), which allows it <strong>to</strong> forecast the changes <strong>to</strong> standalone<br />
and diversified capital caused by new exposure,<br />
using the standalone and diversified capital intensity<br />
benchmark. This capital forecast, although produced by<br />
a fac<strong>to</strong>r model, accounts for the non-linear impact of the<br />
change of diversification within the considered portfolio,<br />
determined by the size of the <strong>risk</strong> classes of which it is composed<br />
and by the corresponding change in exposure.<br />
Fig. 35: Assessment of fully diversified gross capital<br />
Gross NB P&C<br />
s<strong>to</strong>chastic model<br />
Analysis<br />
Model<br />
Scenarios<br />
Data P&C NB model:<br />
• Gross Loss Models<br />
• Gross Premiums and Expenses<br />
• Dependencies<br />
• (Reinsurance structure for net view)<br />
Outputs<br />
<strong>Risk</strong><br />
Fac<strong>to</strong>rs<br />
Benchmark<br />
Gross Portfolio<br />
This new methodology has been applied <strong>to</strong> the example<br />
used in this article, as shown in Figure 35. As a starting<br />
point it is necessary <strong>to</strong> use a partial s<strong>to</strong>chastic internal<br />
model in order <strong>to</strong> gain a good picture of the diversification<br />
structure of the gross P&C new business book.<br />
The model produces gross diversified results purely<br />
within this portfolio, but it is much lighter and faster<br />
<strong>to</strong> run. These results, <strong>to</strong>gether with a number of<br />
suitable gross benchmark portfolios (reflecting<br />
diversification effects with other <strong>risk</strong> classes), were<br />
used as a starting point for the CaDeT model in order<br />
<strong>to</strong> obtain an approximated fully-diversified required<br />
capital by <strong>risk</strong> class.<br />
A similar <strong>approach</strong> has been followed in order <strong>to</strong> compute<br />
the corresponding net diversified capital Figures.<br />
For this case, however, the new partial model has been<br />
run for the P&C net business book, and fully diversified<br />
gross results were used as a benchmark for the diversification<br />
effects with other <strong>risk</strong> classes. The estimated net<br />
fully diversified capital has been utilized <strong>to</strong> compute the<br />
change in the cost of capital due <strong>to</strong> reinsurance.<br />
Non-S<strong>to</strong>chastic<br />
Capital Deployment<br />
Fac<strong>to</strong>r model:<br />
• Diversification correction<br />
• Projection of standalone capital<br />
• Aggregation of segments<br />
• Adjusted profit distribution<br />
Benchmark Data for the whole Portfolio:<br />
• Exposure<br />
• Standalone Capital<br />
• Diversified Capital<br />
• Strength of internal diversification<br />
Fully diversified<br />
Gross results<br />
Results:<br />
• Required Capital<br />
• Diversification Benefit<br />
• Sensitivity<br />
Assessment of fully diversified required capital applying CaDeT methodology. The graph illustrates the estimation of gross diversified<br />
capital. The corresponding net results were estimated analogically, applying the Net NB P&C partial model as well as gross results<br />
as a benchmark for diversification with other <strong>risk</strong>s.<br />
Conclusion<br />
The valuation of different reinsurance options depends<br />
on the motivation for buying reinsurance.<br />
When calculating the cost of reinsurance, the cost of<br />
capital savings is a way of considering the impact of<br />
reinsurance on capital. The assessment of capital relief is<br />
relative and depends on the definition of capital used.<br />
A fully diversified capital view is essential <strong>to</strong> take in<strong>to</strong><br />
account the cost of capital relief (Solvency II standard<br />
model, internal model, rating agencies).<br />
Partial s<strong>to</strong>chastic models allow the evaluation of capital<br />
impact for different reinsurance options in a more<br />
appropriate way.<br />
Fac<strong>to</strong>r models or models <strong>based</strong> on aggregate<br />
distributions (like rating agencies’ models and Solvency II<br />
standard model) are struggling <strong>to</strong> reflect the impact<br />
of non-proportional reinsurance, so the idea is <strong>to</strong> use<br />
the partial model and then <strong>to</strong> use a kind of capital<br />
assessment <strong>to</strong>ol for diversification effects.<br />
SCOR - November 2010 - 35
4<br />
CAPITAL ASSESSMENT<br />
BEYOND STOCHASTIC<br />
MODELING<br />
After an introduction <strong>to</strong> Capital<br />
Modeling at SCOR, this article addresses the motivation<br />
behind the development of a fac<strong>to</strong>r-<strong>based</strong><br />
methodology used <strong>to</strong> approximate capital assessments<br />
for different cases, as an extension of the<br />
s<strong>to</strong>chastic model. It then elaborates on the main<br />
characteristics of this methodology – namely the<br />
consideration of dependencies and diversification<br />
– through examples, highlighting some of the<br />
challenges involved.<br />
36 - November 2010 - SCOR<br />
EVA SCHLÄPFER DE MONTMOLLIN<br />
Senior <strong>Risk</strong> Consultant, SCOR<br />
MAGDALENA KLAPPER-RYBICKA<br />
<strong>Risk</strong> Consultant, SCOR<br />
I. Capital Measurement<br />
and Allocation<br />
HOW DOES SCOR COMPUTE ITS CAPITAL<br />
REQUIREMENTS?<br />
SCOR uses a sophisticated s<strong>to</strong>chastic model <strong>to</strong> calculate<br />
its <strong>risk</strong>-<strong>based</strong> capital by quantifying the combined<br />
capital impact of worst-case scenarios from various<br />
<strong>risk</strong> drivers. Assets and liabilities are modeled using<br />
simulation techniques and dependencies are applied<br />
both within and between portfolios. This ensures that<br />
dependent <strong>risk</strong>s are properly accounted for and that<br />
conservative benefits from the diversification within the<br />
portfolio are taken in<strong>to</strong> account.<br />
<strong>Risk</strong> mitigation measures such as retrocession, asset<br />
allocation and hedging are applied <strong>to</strong> ensure that the<br />
net <strong>risk</strong> profile is modeled.<br />
Figure 36 shows an example output of SCOR’s Group<br />
Internal Model.<br />
The diversification effect is the difference between the<br />
standalone capital and the diversified capital.
Fig. 36: SCOR’s capital requirements<br />
Standalone <strong>Risk</strong>-Based Capital and <strong>risk</strong> category contributions <strong>to</strong> SCOR’s <strong>to</strong>tal diversified <strong>Risk</strong>-Based Capital<br />
<strong>Risk</strong>-Based Capital 2009, SCOR Group<br />
in EURm<br />
RBC<br />
Standalone*<br />
RBC<br />
Diversified<br />
Share<br />
of RBC<br />
P&C New Business 1 595 834 26%<br />
P&C Reserves 1 646 1 249 39%<br />
Life 1 703 877 28%<br />
Assets 699 127 4%<br />
Credit <strong>Risk</strong> 280 38 1%<br />
Other <strong>Risk</strong>s (inc. FX) 321 61 2%<br />
Operational <strong>Risk</strong> 214 214 7%<br />
Total RBC 6 244 3 186<br />
* Sum of Standalone RBC<br />
Diversification Benefit 49%<br />
SCOR calculates its <strong>risk</strong> <strong>based</strong> capital by quantifying the combined capital impact of worst case scenarios from various <strong>risk</strong> drivers.<br />
Diversification between <strong>risk</strong>s is the basis of the business<br />
model for insurance and particularly reinsurance<br />
companies, for the following reasons:<br />
• <strong>Risk</strong> Diversification reduces a company’s need for <strong>risk</strong><strong>based</strong><br />
capital. This is key <strong>to</strong> insurance, reinsurance<br />
and investments.<br />
• Nevertheless, <strong>risk</strong>s are rarely completely independent:<br />
- S<strong>to</strong>ck market crashes are usually not limited <strong>to</strong><br />
one market. Recent market crashes show that<br />
global crises are interdependent.<br />
- Certain Lines of Business are affected by economic<br />
cycles, such as Liability, Credit & Surety and Life<br />
insurance. This creates an implicit dependency.<br />
- Major catastrophes can produce claims in<br />
various Lines of Business.<br />
• Dependency between <strong>risk</strong>s reduces the benefits of<br />
diversification.<br />
• The influence of dependency on the aggregated <strong>risk</strong><strong>based</strong><br />
capital is thus crucial and needs <strong>to</strong> be carefully<br />
analyzed.<br />
The processes and systems of the internal model are<br />
relatively complicated; running the model with the<br />
proper data is time consuming. The full model is very<br />
detailed but not very well suited <strong>to</strong> swift analyses. This<br />
is the main reason why SCOR has recently developed<br />
a simplified version of its internal model as a basis for<br />
dynamic decision making.<br />
This is a capital projection <strong>to</strong>ol that is less complex than<br />
the internal model, and is useful for the following:<br />
• “As if” analysis.<br />
• Strategic planning support.<br />
• A useful benchmark for the Group internal model<br />
and standard formulas.<br />
• Identification of main contribu<strong>to</strong>rs <strong>to</strong> the capital<br />
requirements.<br />
• Assessment of diversification benefit and change as<br />
a consequence of portfolio change.<br />
• Assessment of capital adequacy (solvency ratio).<br />
• Analysis of profitability.<br />
SCOR - November 2010 - 37
Fig. 37: Functionality and high-level architecture of CaDeT<br />
SCOR<br />
Group Internal Model Data<br />
Benchmark Data<br />
Data:<br />
New<br />
• Exposure<br />
exposure<br />
• Standalone Capital Adjustments<br />
• Diversified Capital<br />
• Economic Value Distribution<br />
Parameters:<br />
Strength of internal diversification<br />
Future:<br />
Mean time <strong>to</strong> payment (for discount)<br />
SCOR’S CAPITAL DEPLOYMENT TOOL (CADET)<br />
This capital projection <strong>to</strong>ol is called CaDeT. It is a fac<strong>to</strong>r<strong>based</strong><br />
model that takes output from the s<strong>to</strong>chastic<br />
Group Internal Model and is used <strong>to</strong> calculate new<br />
capital requirements under changed assumptions.<br />
A key feature of this <strong>to</strong>ol is an adjusted diversification<br />
effect for a new portfolio. This is one of the main<br />
differences compared <strong>to</strong> other fac<strong>to</strong>r-<strong>based</strong> models,<br />
such as rating agency models for example.<br />
CaDeT is a light analytics <strong>to</strong>ol that allows the user <strong>to</strong><br />
change assumptions on exposure and <strong>to</strong> approximate<br />
capital and <strong>risk</strong>. It is a major breakthrough in methodology<br />
that blends s<strong>to</strong>chastic modeling with fac<strong>to</strong>r-<strong>based</strong><br />
methods. Due <strong>to</strong> high demand, ongoing extensions are<br />
planned <strong>to</strong> the functionality presented here.<br />
CaDeT requires some benchmark portfolio data such as<br />
exposure information, standalone capital information,<br />
diversified capital information at a certain granularity<br />
and an economic value distribution for the <strong>risk</strong> classes<br />
of interest. Then the user can adjust the exposure<br />
information, for example <strong>to</strong> reflect a new business plan.<br />
The <strong>to</strong>ol extrapolates the portfolio benchmark, taking<br />
in<strong>to</strong> account certain non-linear model assumptions<br />
on diversification behavior. In this way it projects new<br />
standalone and diversified capital amounts. CaDeT also<br />
delivers new diversification benefits, updated required<br />
capital and a sensitivity analysis on the profit distribution<br />
in just a few seconds. It is a good approximation of the<br />
Group Internal Model, but since it relies on simplified<br />
methods CaDeT is obviously less accurate than the<br />
Group Internal Model.<br />
38 - November 2010 - SCOR<br />
CaDeT Result approximation<br />
Fac<strong>to</strong>r model:<br />
• Diversification correction<br />
• Projection of standalone capital<br />
• Aggregation of segments<br />
• Adjusted profit distribution<br />
Future extensions:<br />
• Retrocession optimizer<br />
• Profitability assumptions<br />
• S&P analytics<br />
Fig. 38: Example - original portfolio<br />
<strong>Risk</strong> Category<br />
in EURm<br />
• Required Capital<br />
• Diversification Benefit<br />
• Sensitivity<br />
II. Capital assessment beyond<br />
s<strong>to</strong>chastic modeling<br />
A realistic portfolio can be analyzed in order <strong>to</strong> explain<br />
the methodology in more detail and <strong>to</strong> demonstrate the<br />
impact on diversification and capital. The table below<br />
shows exposure information for the main <strong>risk</strong> categories<br />
as well as their capital contributions in EUR millions.<br />
This portfolio is a P&C book with mainly Property,<br />
Liability and Mo<strong>to</strong>r. Assets are mostly invested in bonds,<br />
real estate and some equities.<br />
Exposure* Fully diversified<br />
Capital<br />
Non-Life New Business 210 52<br />
Non-Life Reserves 240 24<br />
Invested Assets 83 10<br />
Credit 107 5<br />
Other <strong>Risk</strong>s 107 1<br />
Operational <strong>Risk</strong> 107 2<br />
Total capital 94<br />
Example portfolio – gross P&C book with main <strong>risk</strong> categories as<br />
well as contributions <strong>to</strong> the diversified capital.<br />
* Exposure means: Written premium for New business, Reserves,<br />
Market value of invested assets and Available Capital for credit,<br />
operational and other <strong>risk</strong>s.
Fig. 39: Example – original exposure and capital<br />
Detailed view on P&C new business<br />
Premium<br />
Standalone Cap<br />
Property<br />
Non Cat<br />
Property<br />
Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
10<br />
0 20 40 60 80 0 20 40 60 80 0 20 40<br />
Total contribution <strong>to</strong> the fully diversified capital of P&C<br />
new business is EUR 52 million. Property Non-Cat is the<br />
largest Line of Business and consumes the most capital<br />
50<br />
70<br />
80<br />
29<br />
43<br />
53<br />
75<br />
Diversified Capital<br />
10<br />
10<br />
9<br />
23<br />
69%<br />
66%<br />
81%<br />
79%<br />
P&C New Business<br />
Book Capital: 52<br />
Example portfolio – Detailed view of the P&C new business book with premium volumes, standalone capital and contributions <strong>to</strong> the fully diversified<br />
capital. The diversification benefits per Line of Business in the full diversification is represented in orange. All amounts are in EUR millions.<br />
WHAT HAPPENS IF EXPOSURE CHANGES?<br />
Fig. 40: Standalone Capital (in €m)<br />
Example for Mo<strong>to</strong>r – doubling the premium<br />
Mo<strong>to</strong>r (New Business)<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
~106% standalone capital intensity<br />
50<br />
Written Premium<br />
100<br />
2x<br />
53.4<br />
Standalone Capital<br />
in a standalone and a diversified view. The diversification<br />
benefit for smaller Lines of Business is larger, but<br />
Property is also well diversified.<br />
In soft market<br />
you can grow only<br />
by adding business<br />
with higher <strong>risk</strong><br />
~106% standalone capital intensity<br />
160<br />
140<br />
160<br />
Mo<strong>to</strong>r Example, what can happen <strong>to</strong> the capital if the premium doubles. Three examples are presented showing the effects<br />
of diversification for different types of growth.<br />
3x<br />
Grow if diversity<br />
(healthy market),<br />
maximum of internal<br />
diversification:<br />
sqrt(2) = 1.41<br />
1.41x<br />
grow by adding<br />
weakly dependent<br />
business<br />
1.8x<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
100<br />
100<br />
100<br />
~75% standalone<br />
capital intensity<br />
75<br />
~100% standalone<br />
capital intensity<br />
100<br />
Diversification Benefit<br />
SCOR - November 2010 - 39
In the Mo<strong>to</strong>r Line of Business, the written premium is<br />
EUR 50 million and the standalone capital is EUR 53 million,<br />
which means that the capital intensity is around 100%.<br />
If the premium volume doubles, the capital should<br />
increase but it is not immediately clear by how much. The<br />
increase in capital actually depends on how the portfolio<br />
can be expanded. This can be achieved by increasing<br />
the share of business or writing new business, or a<br />
combination of both. For example in a soft market, the<br />
only way you can grow is through writing business that<br />
is worse than the existing book from a <strong>risk</strong> perspective.<br />
Then the capital intensity increases and the required<br />
capital subsequently increases over-proportionally. In the<br />
best-case scenario – i.e. in a healthy market – it is possible<br />
<strong>to</strong> expand so that the portfolio’s internal diversification<br />
improves significantly, increasing capital by a fac<strong>to</strong>r of 1,41<br />
(square root of two). If the portfolio were expanded with<br />
weakly dependent business, the capital would increase<br />
by a fac<strong>to</strong>r of around 1.8.<br />
Now, it is possible <strong>to</strong> explore the capital projections of<br />
the methodology if the portfolio changes. Two cases<br />
will be explored, one better-balanced portfolio and<br />
one extremely unbalanced portfolio, keeping the <strong>to</strong>tal<br />
premium volume constant.<br />
The first, balanced, portfolio has a new premium<br />
volume for Property Non-Cat of EUR 56 million<br />
(previously EUR 80 million), EUR 5 million for Property<br />
Cat (previously EUR 10 million) and EUR 99 million for<br />
Casualty (previously EUR 70 million). The Mo<strong>to</strong>r premium<br />
volume does not change. The <strong>to</strong>tal premium volume<br />
stays at EUR 210 million.<br />
In the second, extremely unbalanced portfolio,<br />
EUR 180 million of the new business premium is written<br />
in Mo<strong>to</strong>r, this is more than 80% of the <strong>to</strong>tal premium.<br />
The rest is written in the smaller lines. The <strong>to</strong>tal premium<br />
volume stays at EUR 210 million.<br />
40 - November 2010 - SCOR<br />
TOTAL CAPITAL REQUIREMENTS<br />
AND PER RISK FACTORS<br />
Fig. 41: Capital requirements<br />
per <strong>risk</strong> fac<strong>to</strong>rs (in €m)<br />
Capital<br />
in €<br />
200<br />
150<br />
100<br />
50<br />
0<br />
Main <strong>Risk</strong>s<br />
Other<br />
Operational<br />
Credit<br />
Capital Deployment Overview<br />
94m<br />
52m<br />
89m<br />
46m<br />
Gross Gross Better<br />
Balanced<br />
170m<br />
127m<br />
Gross<br />
Unbalanced<br />
Invested Assets<br />
P&C Reserves<br />
P&C New Business<br />
Total capital (in orange) and main <strong>risk</strong> fac<strong>to</strong>rs <strong>to</strong> the <strong>to</strong>tal<br />
required capital for the original example portfolio as well as<br />
the two modified portfolios. In black we show the<br />
contributions of the P&C new business capital.<br />
The <strong>to</strong>tal capital in the better-balanced portfolio<br />
drops <strong>to</strong> EUR 89 million, mainly due <strong>to</strong> the lower<br />
contribution of P&C new business. The extremely<br />
unbalanced portfolio has a much higher <strong>to</strong>tal capital<br />
of EUR 170 million, again mainly due <strong>to</strong> the much<br />
bigger impact of the P&C diversified capital. Capital<br />
consumption for the other <strong>risk</strong> fac<strong>to</strong>rs only changes<br />
marginally due <strong>to</strong> the change in diversification.<br />
The impact of the different Lines of Business on the<br />
capital contributions for the balanced and extremely<br />
unbalanced portfolio should be analyzed more<br />
precisely.
Fig. 42: Example – balanced<br />
Increasing Premium volume for Casualty, decreasing for Property<br />
Property<br />
Non Cat<br />
Property<br />
Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
Premium<br />
5<br />
-5<br />
50<br />
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40<br />
P&C New Business<br />
Book Capital: 46<br />
As shown in Figure 42, the two Property lines diversify<br />
better in the more balanced portfolio. The diversification<br />
benefit increases from 69% and 66% for Property<br />
Non-Cat and Property Cat <strong>to</strong> 77% and 79% respectively.<br />
As expected, the diversification benefit for<br />
Casualty decreases from 79% <strong>to</strong> 68%, which is a natural<br />
consequence of a significant growth of this line.<br />
Below is a similar graph for the extremely unbalanced<br />
portfolio.<br />
56<br />
+29<br />
-24<br />
99<br />
Standalone Capital<br />
14<br />
50<br />
53<br />
66<br />
Diversified Capital<br />
Better balanced portfolio, P&C new business book with premium volumes, standalone capital and contributions <strong>to</strong> the fully diversified<br />
capital. The diversification benefits per Line of Business in full diversification are in orange. All amounts are in EUR millions.<br />
Fig. 43: Example – very unbalanced<br />
Basically a mono-line company<br />
Total<br />
Property<br />
Property<br />
Cat<br />
Mo<strong>to</strong>r<br />
Casualty<br />
Premium<br />
2<br />
14<br />
14<br />
180<br />
Standalone Capital<br />
11<br />
9<br />
9<br />
3<br />
11<br />
11<br />
21<br />
77%<br />
79%<br />
80%<br />
68%<br />
The standalone capital and the contribution of Mo<strong>to</strong>r<br />
<strong>to</strong> the <strong>to</strong>tal diversified capital increase significantly. The<br />
diversification benefit for Mo<strong>to</strong>r drops from 80.6% <strong>to</strong><br />
38.7%. Diversification benefits on all lines except Mo<strong>to</strong>r<br />
are very large and the small lines are allocated very little<br />
capital. This effect can often be observed when there is<br />
a dominating <strong>risk</strong> category – “the winner takes all”.<br />
0 50 100 150 200 0 50 100 150 200 0 50 100 150<br />
208<br />
Diversified Capital<br />
0<br />
1<br />
0<br />
P&C New Business<br />
Book Capital: 129<br />
Diversification Benefit<br />
96%<br />
88%<br />
39%<br />
127<br />
Extremely unbalanced P&C new business book with premium volumes, standalone capital and contributions <strong>to</strong> the fully diversified capital.<br />
The diversification benefits per Line of Business in full diversification are in orange. All amounts are in EUR millions.<br />
96%<br />
Diversification Benefit<br />
SCOR - November 2010 - 41
SOLVENCY RATIOS<br />
Finally, it is also interesting <strong>to</strong> observe the behavior of<br />
the solvency ratio between the original portfolio, the<br />
balanced portfolio and the unbalanced portfolio. The<br />
solvency ratio is defined as the quotient of the available<br />
capital of EUR 107 million and the required capital.<br />
Underwriting new business in a more balanced way<br />
improves the solvency ratio from 113% <strong>to</strong> 120%. The<br />
solvency ratio for the unbalanced case is below 100%,<br />
which means that this company would need <strong>to</strong> urgently<br />
raise capital or lower its required capital by de-<strong>risk</strong>ing the<br />
portfolio. Of course, in reality, there are limitations as <strong>to</strong><br />
how fast significant portfolio changes can be achieved<br />
due <strong>to</strong> changes in the underwriting policy. Reinsuring<br />
part of the portfolio could be a faster solution in terms<br />
of moving <strong>to</strong> a better-balanced portfolio. This example<br />
shows in particular that having a well-balanced portfolio<br />
is important from a capital perspective.<br />
Fig. 44: Solvency ratios<br />
of the three example portfolios<br />
Div. Capital<br />
in €m<br />
200<br />
150<br />
100<br />
50<br />
0<br />
42 - November 2010 - SCOR<br />
Required Capital<br />
vs. Available Capital<br />
113% 120%<br />
Gross Gross Better<br />
Balanced<br />
63%<br />
Solvency ratio<br />
Gross<br />
Unbalanced Available<br />
Capital<br />
III. Modeling of Diversification<br />
and other Challenges<br />
HOW CAPITAL CHANGES WITH EXPOSURE<br />
This section will now show in more detail how capital<br />
changes due <strong>to</strong> changes in exposure.<br />
Fig. 45: Exposure against capital<br />
intensity and against capital amounts<br />
(in %)<br />
200%<br />
180%<br />
160%<br />
140%<br />
120%<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
0<br />
(€m)<br />
1,000<br />
900<br />
800<br />
700<br />
600<br />
500<br />
400<br />
300<br />
200<br />
100<br />
0<br />
0<br />
1,000 2,000 3,000 4,000<br />
Exposure<br />
(€m)<br />
Standalone Capital Intensity<br />
Diversified Capital Intensity<br />
200 400 600 800<br />
Exposure<br />
LoB standalone Capital<br />
Diversified Capital 2009<br />
1,000<br />
(€m)
In Figure 45, exposure is shown on both X axes. The<br />
Y axis on the first graph shows capital intensity, while<br />
the Y axis on the second graph shows capital amounts.<br />
This model requires a number of parameters that need<br />
<strong>to</strong> be adapted and calibrated for different Lines of<br />
Business and markets.<br />
The behavior of the standalone capital intensity will be<br />
explained more in detail; diversified capital intensities<br />
behave similarly. In the second graph you can see<br />
that the capital amounts increase with the exposure.<br />
The development of standalone capital with growing<br />
exposure is modeled, assuming that a small portfolio<br />
can be expanded with improving internal diversification.<br />
This is where the capital intensities decrease<br />
with growing exposure. But there is a saturation<br />
point where diversified growth is no longer possible.<br />
The portfolio can only grow with worsening internal<br />
diversification, or because only <strong>risk</strong>ier new business can<br />
be underwritten. From this saturation point the capital<br />
intensities increase once again.<br />
The parameterization of the curve depends on the<br />
Lines of Business, the portfolio and the market, and<br />
should be chosen in such a way as <strong>to</strong> reflect realistic<br />
behavior.<br />
CHALLENGES AND FUTURE DEVELOPMENT<br />
It is worth mentioning that there are many challenges<br />
and areas of future development surrounding this<br />
methodology. For example, the modeling of duration<br />
matching and interest rate sensitivity needs <strong>to</strong> be<br />
improved. If the asset allocation is changed significantly,<br />
this will change the duration of the assets, which may<br />
be better or worse matched <strong>to</strong> the duration of the<br />
liabilities. This should have an impact on the capital.<br />
Furthermore, the methodology needs some parameters<br />
– for example <strong>to</strong> describe the change in standalone<br />
capital when exposure changes, as discussed in<br />
the previous section. A realistic choice reflecting the<br />
portfolio and market conditions with appropriate<br />
parameters is challenging.<br />
To summarize, the methodology as it stands <strong>to</strong>day is<br />
used for the following:<br />
• Approximate assessment of capital requirements,<br />
taking changing diversification in<strong>to</strong> account.<br />
• Identification of main contribu<strong>to</strong>rs <strong>to</strong> the capital<br />
requirements.<br />
• Assessment of capital adequacy (solvency ratio).<br />
• Approximate adjustments <strong>to</strong> the capital with changing<br />
exposure.<br />
Further planned developments include:<br />
• Profitability analysis of changing portfolio.<br />
• Impact of different reinsurance structures on<br />
capital.<br />
• Better modeling of interest rate sensitivities.<br />
This methodology can also be used <strong>to</strong> make the leap<br />
from a partial s<strong>to</strong>chastic model analysing the impact of<br />
a reinsurance structure <strong>to</strong> fully diversified capital.<br />
Conclusion<br />
For quick approximations of capital requirements,<br />
a simple proxy of a complex model, in our case the<br />
internal model, is needed. The methodology presented<br />
here shows one way of building such a proxy tailored<br />
<strong>to</strong> SCOR’s needs. Since diversification is an essential<br />
concept, this methodology is focused on more realistic<br />
modeling than Solvency II’s quantitative impact studies<br />
or the rating agencies’ models.<br />
Obviously, this methodology cannot replace a complex<br />
internal model, but it is a very useful extension since the<br />
calibration is <strong>based</strong> on the underlying complex model<br />
for fast and simple analyses. This article has shown a<br />
way of calibrating this model with benchmark data<br />
<strong>based</strong> on the Group Internal Model. Using other benchmark<br />
data allows an extension in order <strong>to</strong> address other<br />
issues or test potential business changes with the same<br />
methodology.<br />
SCOR - November 2010 - 43
5<br />
INTEGRATING RISK<br />
MODELING INTO<br />
THE ORGANIZATION JANICE COWLEY<br />
Program Direc<strong>to</strong>r of CoCPIT, SCOR<br />
CoCPIT (Cost of Capital, Pricing,<br />
IT) was a two-year program of change in terms of<br />
the cost of capital, pricing and IT, and was aimed<br />
at further integrating the internal model in<strong>to</strong> the<br />
organization. The program concluded in the<br />
summer of 2010. At its conclusion SCOR was able<br />
<strong>to</strong> demonstrate the Group’s strong involvement in<br />
modeling and the connectivity of the results <strong>to</strong> its<br />
day-<strong>to</strong>-day activities.<br />
This article provides simple and straightforward<br />
examples on how SCOR has:<br />
• Cemented the process of modeling and governance<br />
within the organization;<br />
• Integrated the use of the Group’s internal model in<strong>to</strong><br />
the organization.<br />
44 - November 2010 - SCOR<br />
I. Cementing the process<br />
of modeling and governance<br />
within the organization<br />
The CoCPIT <strong>approach</strong> is <strong>based</strong> on three components:<br />
• Data Ownership and Audit ability are increasingly<br />
important due <strong>to</strong> the upcoming Solvency II regulations.<br />
They enable <strong>to</strong> have the data that feed the<br />
models within the organization. The systems that have<br />
been developed are built on three components:<br />
- Owners sign off on their data;<br />
- Au<strong>to</strong>mation and solid design provides a robust<br />
audit trail;<br />
- The development of a centralized data reposi<strong>to</strong>ry<br />
is used as the basis for capital modeling<br />
and will be used in the future for Solvency II and<br />
standard models.<br />
• A robust governance structure constitutes the second<br />
component of CoCPIT, comprising:<br />
- A defined and transparent governance<br />
structure;<br />
- The clear establishment of roles and responsibilities;<br />
- The involvement of the Executive Team and<br />
Board in the process.<br />
• Use of expertise within the organization <strong>to</strong> make<br />
sure that:<br />
- Processes are designed and their results validated<br />
in partnership with the organization;<br />
- Processes are flexible in order <strong>to</strong> benefit from<br />
expert judgment and knowledge. It is essential<br />
<strong>to</strong> know the limitations of the models;<br />
- Transparency is ensured concerning limitations<br />
and assumptions.
A CLEAR GOVERNANCE FRAMEWORK UNDERPINS THE MODELING PROCESS<br />
Fig. 46: Modeling of <strong>risk</strong> and capital<br />
The involvement of the organization in modeling, validation, challenge and review<br />
is of paramount importance <strong>to</strong> SCOR<br />
Current<br />
Portfolio<br />
Plans<br />
Reserves<br />
Enter & Sign<br />
off data<br />
Retrocession<br />
& Hedges Economic<br />
Scenarios<br />
Model Life,<br />
P&C, Assets<br />
The process is structured so that the starting point is the<br />
provision of data and sign off before execution of the<br />
modeling of standalone and globally diversified results.<br />
The process is overlaid with the ALM Governance<br />
framework, which ensures the necessary technical and<br />
management reviews of results.<br />
Model Group<br />
Capital<br />
<strong>Risk</strong>-Based Capital<br />
& capital<br />
intensity<br />
Strategic Asset Allocation<br />
& duration matching<br />
<strong>Risk</strong> drivers<br />
& strategies<br />
Capital allocation<br />
Process<br />
Expert Review<br />
Groups<br />
ALM Technical<br />
Committee<br />
ALM Committee<br />
Board of Direc<strong>to</strong>rs<br />
Info <strong>to</strong> regula<strong>to</strong>rs<br />
& Rating Agencies<br />
SCOR - November 2010 - 45
Fig. 47: A clear governance framework underpins modeling process<br />
The pyramid above illustrates the review and approval<br />
process at SCOR.<br />
Expert reviews are designed <strong>to</strong> ensure that results are<br />
reviewed at a division level by those familiar with the<br />
<strong>risk</strong>s in that division. The review groups are composed<br />
of experts in that area. These division experts have a<br />
responsibility <strong>to</strong> ensure the technical and modeling<br />
accuracy of the results on a standalone level. The scope<br />
of the role includes ensuring the transparency of any<br />
limitations or assumptions within the model.<br />
The global technical review is performed by the<br />
ALM Technical Committee, which also reviews the<br />
consolidated results of the organization. Transparency<br />
on assumptions and limitations is provided and<br />
the impact of any changes <strong>to</strong> methodology is made<br />
explicit. The ALM Technical Committee comprises<br />
technical experts, including representatives from Group<br />
<strong>Risk</strong> <strong>Management</strong> division, Chief Reserving Actuaries,<br />
Pricing Actuaries, Head of Natural Catastrophes, Global<br />
Head of Retrocession, and so on.<br />
The ALM Committee, composed of the Group’s executive<br />
team, completes the management review of<br />
the results. The ALM Committee relies on the ALM<br />
46 - November 2010 - SCOR<br />
Board of Direc<strong>to</strong>rs via <strong>Risk</strong> Committee<br />
Endorses ALM results<br />
Approves Strategic Asset Allocation ”SAA”<br />
Endorses capital and buffer<br />
ALM Committee<br />
Optimal capital deployment <strong>to</strong> achieve <strong>risk</strong>/return<br />
Proposes <strong>risk</strong> <strong>based</strong> capital, buffers & SAA<br />
Sets <strong>risk</strong> mitigation strategies & manages <strong>to</strong> defined <strong>risk</strong> <strong>to</strong>lerances<br />
Affirms the appropriate capital allocation per portfolio for pricing<br />
ALM Technical Committee<br />
Alerts ALM Committee <strong>to</strong> any limitations of the model<br />
Assures accuracy and completeness of models + results<br />
Proposes the capital allocation per portfolio for pricing<br />
Assets, Life & P&C Technical Review Groups<br />
Assure the accuracy and completeness of standalone results<br />
Alerts the ALM Technical Committee <strong>to</strong> limitations, areas of<br />
judgement and decisions taken <strong>to</strong> assure accurate results<br />
Technical Committee for the accuracy of the results and<br />
<strong>to</strong> understand the drivers behind change, limitations<br />
and areas where judgment is required. As part of this<br />
process, the ALM Committee often focuses on areas<br />
requiring management action, for instance areas<br />
where <strong>risk</strong>s are arising, where diversification is not<br />
optimized or where there are proposals worth pursuing,<br />
for example in relation <strong>to</strong> strategic allocation or <strong>risk</strong><br />
mitigation. These proposals are then put <strong>to</strong> the Board<br />
of Direc<strong>to</strong>rs.<br />
Finally the Board of Direc<strong>to</strong>rs, via the <strong>Risk</strong> Committee,<br />
endorses the results of the modeling process, sets the<br />
strategic asset allocation and ultimately agrees on <strong>risk</strong><br />
<strong>to</strong>lerance levels.<br />
TRANSPARENCY AND INVOLVEMENT<br />
ARE AT THE CORE OF SCOR’S APPROACH<br />
SCOR’s <strong>approach</strong> is <strong>based</strong> on three main features:<br />
• Advanced Models<br />
Governance<br />
<strong>Management</strong>/<br />
COMEX<br />
Technical<br />
Expert<br />
These constitute a basis for decision. It is important <strong>to</strong><br />
acknowledge that decision making needs <strong>to</strong> be <strong>based</strong><br />
around the results, and processes must provide for that<br />
flexibility.
• Assumptions and Limitations<br />
It is crucial in terms of gaining the trust and acceptance<br />
of the organization <strong>to</strong> identify and communicate the<br />
limitations of the internal model. Assumptions are<br />
transparent, so their validity can be unders<strong>to</strong>od and<br />
challenged where necessary. Limitations are made clear<br />
so that informed decisions can be made. Judgement is<br />
applied as necessary.<br />
• Involvement and disclosure<br />
Involvement, and ensuring that a constructive dialogue<br />
takes place, are crucial <strong>to</strong> the success. The ability <strong>to</strong><br />
analyze results is provided, thereby optimizing the use<br />
of the knowledge and expertise of the organization.<br />
These highlights illustrate how SCOR has cemented the<br />
process of internal modeling within the organization.<br />
•<br />
•<br />
•<br />
<strong>Risk</strong>/return decisions<br />
Financial &<br />
Economic Results<br />
Performance<br />
Measurement<br />
Business<br />
Actions<br />
<strong>Risk</strong> Tolerances & Limits<br />
<strong>Risk</strong><br />
Budgets<br />
II. Integrating the use<br />
of the Group internal model<br />
in<strong>to</strong> the organization<br />
In this section are selected examples of how the results<br />
of the internal model are used at SCOR.<br />
THE INTERNAL MODEL IS A CORNERSTONE<br />
OF <strong>ERM</strong> AT SCOR<br />
Figure 48 shows how capital and <strong>risk</strong> are integrated in<strong>to</strong><br />
the organization, from strategy setting and business<br />
planning through <strong>to</strong> capital allocation, <strong>risk</strong> budgets,<br />
underwriting execution and performance evaluation.<br />
Strategy & <strong>Risk</strong> Appetite<br />
Capital and <strong>risk</strong> <strong>to</strong>lerance are key considerations when<br />
setting and affirming the strategy.<br />
• When considering the strategic plan, the implications<br />
on capital are confirmed and it is ensured that the<br />
Group does not exceed its <strong>risk</strong> <strong>to</strong>lerance.<br />
• When the results of the Group internal model are<br />
discussed, the Board <strong>Risk</strong> Committee ensures that<br />
SCOR’s target <strong>risk</strong> profile is respected and the<br />
expected return and volatility are unders<strong>to</strong>od.<br />
• The capital buffer is set in line with the capital shield<br />
strategy <strong>to</strong> avoid the need <strong>to</strong> regularly <strong>approach</strong> the<br />
shareholders for additional funding.<br />
Fig. 48: Integrating the internal model in<strong>to</strong> the organization’s processes<br />
The internal model is a corners<strong>to</strong>ne of <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong> at SCOR<br />
•<br />
•<br />
Strategy &<br />
<strong>Risk</strong> Appetite<br />
Available capital & <strong>Risk</strong> Tolerance<br />
Target capital & <strong>Risk</strong> Profile<br />
Portfolio<br />
Planning &<br />
Optimization<br />
Capital<br />
Allocation<br />
•<br />
•<br />
•<br />
•<br />
•<br />
•<br />
•<br />
•<br />
Capital & <strong>Risk</strong> Awareness<br />
Trade off analysis<br />
Diversification benefit<br />
Sensitivity of result<br />
CaDeT – SCOR’s <strong>to</strong>ol<br />
<strong>to</strong> support strategic planning<br />
Capital diversification benefits<br />
Principles of capital allocation<br />
Proactive operational <strong>risk</strong> allocation<br />
SCOR - November 2010 - 47
Fig. 49: Appropriate executive level governance over <strong>risk</strong> <strong>to</strong>lerance and adherence<br />
<strong>Risk</strong> Tolerance is set by the Board of Direc<strong>to</strong>rs<br />
SCOR’s Board and Executive <strong>Management</strong> team regularly review the Group’s <strong>Risk</strong> Profile and Available<br />
Capital <strong>to</strong> ensure that they remain aligned with the Group’s <strong>Risk</strong> Appetite framework<br />
Change in Economic Value in EUR bn<br />
At a strategic and Board level, an understanding of the<br />
<strong>risk</strong> profile is important. The corporate level <strong>to</strong>lerances<br />
are taken and combined with operational level limits,<br />
so that an end <strong>to</strong> end or <strong>to</strong>p <strong>to</strong> bot<strong>to</strong>m <strong>approach</strong> <strong>to</strong> <strong>risk</strong><br />
<strong>to</strong>lerance and appetite within the company is get.<br />
Portfolio Planning & Optimization<br />
The portfolio planning and optimization is carried out<br />
on the basis of:<br />
• Capital & <strong>Risk</strong> Awareness<br />
• Trade off analysis<br />
• Diversification benefit<br />
• Sensitivity of result<br />
• CaDeT – SCOR’s <strong>to</strong>ol <strong>to</strong> support strategic planning<br />
48 - November 2010 - SCOR<br />
Net income Expected<br />
growth in<br />
economic value<br />
+ Buffer Capital Required Capital Available Capital<br />
Return Period in Years<br />
CaDeT is an abbreviation for Capital Deployment Tool.<br />
CaDeT provides real time information on the impact<br />
of changed exposure (or business plans) on capital<br />
requirements and diversification benefit/optimization.<br />
The <strong>to</strong>ol is the organization’s response <strong>to</strong> the need for<br />
real time information on capital and <strong>risk</strong>.<br />
Figure 50 demonstrates its internal usefulness. The<br />
numbers used in this example are just illustrative. It is<br />
possible <strong>to</strong> alter exposure, and the standalone capital<br />
and diversified capital will respond instantly. A key<br />
feature is the consideration of the changed exposure<br />
and diversification benefit across the entire portfolio.<br />
In addition <strong>to</strong> presenting capital and <strong>risk</strong> numbers, this<br />
example shows the probable volatility of results.<br />
For SCOR, CaDeT has represented a big step forward<br />
in terms of integrating modeling results in<strong>to</strong> the organization.<br />
CaDeT has proven <strong>to</strong> be extremely beneficial<br />
for strategic decision making in terms of portfolio<br />
composition and looking at potential business<br />
acquisitions. It is also used <strong>to</strong> cement the understanding<br />
of capital and <strong>risk</strong> across the organization, because it<br />
is very simple <strong>to</strong> use and easily accessible.
Fig. 50: CaDeT example (Simplified example)<br />
<strong>Risk</strong> Fac<strong>to</strong>r Exposure Standalone Capital Diversified Capital Diversification <strong>Risk</strong><br />
Moni<strong>to</strong>r<br />
Current Change New Current New Change Current New Change 2009 Plan<br />
P&C New Business 1,500 50% 2,250 800 1,100 38% 415 700 69% 48% 36% <br />
P&C Reserves 5,000 5,000 800 800 0% 625 600 -4% 22% 25% <br />
Life 1,300 1,300 850 850 0% 450 415 -8% 47% 51% <br />
Investments 5,000 5,000 350 350 0% 65 55 -15% 81% 84% <br />
Other 2,000 2,100 150 150 0% 30 25 -17% 80% 83% <br />
Total 1,585 1,795 13%<br />
Changes in exposure<br />
(different granularities available)<br />
Available & Required Capital - Internal Model<br />
Available Capital<br />
2,000<br />
Current<br />
2,100<br />
New<br />
100<br />
THE INTERNAL MODEL IS THE BASIS<br />
FOR CAPITAL ALLOCATION<br />
Current<br />
Under CoCPIT, SCOR sought <strong>to</strong> cement an over-arching<br />
global framework around capital allocation. To achieve<br />
this, SCOR established a global multi-division, multifunctional<br />
team <strong>to</strong> look at this <strong>to</strong>pic.<br />
The key principles <strong>to</strong> follow in Capital Allocation<br />
were defined at the outset and can be summarized<br />
as follows:<br />
• Align with the Group Internal Model: the objective<br />
was <strong>to</strong> make sure that capital allocation was<br />
connected <strong>to</strong> our Group internal model results.<br />
• Allocate Gross of Retrocession: a second key principle<br />
was <strong>to</strong> allocate capital gross of retrocession – whilst<br />
SCOR allows for retrocession benefit in the pricing of<br />
treaties, the capital allocated and basis for measuring<br />
return, is gross.<br />
Required Capital<br />
Diversified (net) capital<br />
1,615<br />
1,825<br />
New<br />
210<br />
Are assessed real-time<br />
in terms of <strong>risk</strong> and capital<br />
Solvency Ratio<br />
= Available/Required<br />
124%<br />
115%<br />
2009 Plan<br />
-9%<br />
Solvency moni<strong>to</strong>r:<br />
• Capital allocated reflects <strong>risk</strong>: the general principle<br />
regarding the comparative quantum of capital<br />
allocated was that the greater the <strong>risk</strong>, the greater<br />
the dependence or the greater the development<br />
time, then the greater the amount of capital allocated<br />
<strong>to</strong> a portfolio or a treaty.<br />
• Unit of measure: the idea was <strong>to</strong> try introducing a<br />
common unit of measure for capital allocation – by<br />
using capital intensities, ensuring that the exposure<br />
measure is appropriate <strong>to</strong> the underlying business.<br />
• Moni<strong>to</strong>r and manage: it was agreed upfront that<br />
there must be a moni<strong>to</strong>ring process <strong>to</strong> review the<br />
actual capital allocated within pricing vs. expectation,<br />
so that deviations are known and unders<strong>to</strong>od and if<br />
necessary action is taken.<br />
SCOR - November 2010 - 49
Fig. 51: In pricing, allocation of capital follows strategic principles<br />
Measurement of <strong>risk</strong> and allocation of diversification benefit<br />
RISK TOLERANCES AT THE OPERATIONAL<br />
LEVEL ENSURE CONNECTIVITY OF THE<br />
STRATEGY WITH DAY-TO-DAY OPERATIONS<br />
In day-<strong>to</strong>-day decision making, capital and return on<br />
that capital is a key fac<strong>to</strong>r that SCOR examines in terms<br />
of when the business is actually written.<br />
<strong>Risk</strong> budgets<br />
<strong>Risk</strong> <strong>to</strong>lerance is set on the basis of available capital and<br />
is implemented at all levels of the company:<br />
• At Group level, <strong>to</strong> ensure that the company is not<br />
over-exposed <strong>to</strong> a single event combining all Lines<br />
of Business and asset classes;<br />
• At Lines of Business level, <strong>to</strong> avoid concentration<br />
of <strong>risk</strong> in specific Lines of Business or asset classes<br />
and hence ensure that the diversification benefits<br />
are optimized.<br />
Fig. 52: <strong>Risk</strong> / Return decisions at the treaty level are assessed<br />
against capital allocated<br />
The capital allocated <strong>to</strong> the <strong>risk</strong>s has a cost associated: the shareholder expects a return, the bondholder an interest.<br />
50 - November 2010 - SCOR<br />
<strong>Risk</strong> Capital<br />
Expenses<br />
Terms & Conditions<br />
Pure Losses<br />
<strong>Risk</strong> Capital<br />
Expenses<br />
Terms & Conditions<br />
Pure Losses<br />
Economic<br />
profit<br />
Cost<br />
of capital<br />
Expenses<br />
Expected<br />
Loss<br />
Economic<br />
profit<br />
Cost<br />
of capital<br />
Expenses<br />
Expected<br />
Loss<br />
It is worth noting that <strong>risk</strong> mitigation measures such as<br />
retrocession, and other hedging are applied <strong>to</strong> ensure<br />
that <strong>risk</strong> <strong>to</strong>lerance levels are respected.<br />
Business actions and performance<br />
measurement<br />
A key fac<strong>to</strong>r in the decision whether or not <strong>to</strong> underwrite<br />
a contract is the return on the allocated capital.<br />
Thus, at the treaty level the results of the internal model<br />
are used as a basis for decisions.<br />
As an example:<br />
Capital allocated<br />
The capital allocated is <strong>based</strong><br />
on the <strong>risk</strong> of each and every treaty<br />
Diversification benefit from<br />
the internal model is leveraged<br />
The charge <strong>to</strong> the contract<br />
is the cost of that capital<br />
The amount of capital allocated<br />
is moni<strong>to</strong>red<br />
• If 100 in capital is allocated and a profit of 15 is<br />
expected <strong>to</strong> be made, then the expected return on<br />
<strong>risk</strong>-adjusted capital is 15%.<br />
• If, however, this were a more <strong>risk</strong>y contract, potentially<br />
with a greater dependence or a longer tail, then it<br />
may have a capital allocated of 150. Consequentially<br />
the expected return would be 10%.<br />
Return on capital<br />
The capital allocated is <strong>based</strong><br />
on the <strong>risk</strong> of each and every treaty<br />
The performance / return assessment<br />
is <strong>based</strong> on that capital<br />
<strong>Risk</strong> Capital: 100, Profit: 15<br />
Return on capital: 15%<br />
<strong>Risk</strong> Capital 150, Profit: 15<br />
Return on capital: 10%
Conclusion<br />
The results of SCOR’s internal model are important<br />
<strong>to</strong> the company’s organization. The CoCPIT team has<br />
therefore invested a lot of time <strong>to</strong> ensure that the<br />
Group internal model is connected <strong>to</strong> day-<strong>to</strong>-day decision<br />
making within the organization.<br />
The involvement of the management from the outset<br />
is of paramount importance in terms of integrating<br />
capital and <strong>risk</strong> management in<strong>to</strong> the organization.<br />
Robust governance with executive involvement ensures<br />
the appropriate management of <strong>risk</strong>.<br />
The internal model provides important information <strong>to</strong><br />
assess whether the Group’s <strong>risk</strong> profile is in line with the<br />
Group <strong>risk</strong> appetite framework. Nonetheless, it does<br />
not replace the need for decisions.<br />
Finally, building the trust of stakeholders, both internally<br />
and externally, is immensely important. CoCPIT<br />
team has therefore ensured that it is very transparent<br />
around its assumptions and also its limitations, so that<br />
people know when <strong>to</strong> rely on the results and in what<br />
areas they might have <strong>to</strong> make judgment decisions.<br />
SCOR - November 2010 - 51
6<br />
INVESTMENT AND RISK<br />
MANAGEMENT STRATEGIES<br />
IN A CHANGING REGULATORY<br />
FRAMEWORK JULIEN S. HALFON<br />
Head of Coverage<br />
P-Solve <strong>Risk</strong> <strong>Management</strong> Solutions<br />
The Solvency II Directive is now<br />
almost fully drafted and its content is relatively<br />
well defined. Notwithstanding, large-scale changes<br />
in regulation, even when they are known, can<br />
bring unforeseen consequences – positive or negative<br />
– often because of external fac<strong>to</strong>rs such as<br />
market moves or changes in behaviors. The pension<br />
markets – especially in the UK and the Netherlands<br />
– have undergone similar changes over the last<br />
decade. The outcomes of these changes were not<br />
always what they were expected <strong>to</strong> be.<br />
52 - November 2010 - SCOR<br />
The main question that arises in this article is: what<br />
happens <strong>to</strong> investment and <strong>risk</strong> management strategies<br />
when the regula<strong>to</strong>ry framework changes? To answer<br />
that question, this article proposes <strong>to</strong>:<br />
• Review the changes that have taken place in terms of<br />
pension regulation in the UK and the Netherlands;<br />
• Understand the implications and challenges faced by<br />
UK and Dutch pension funds;<br />
• Analyze changes <strong>to</strong> investment and <strong>risk</strong> management<br />
strategies: How they have evolved and how pension<br />
funds have responded <strong>to</strong> regula<strong>to</strong>ry changes.<br />
The following are the views of Julien S. Halfon and do not<br />
necessarily represent the views of P-Solve Investments<br />
Limited.
I. Regula<strong>to</strong>ry changes<br />
Interestingly enough, the drivers behind all these<br />
changes have little <strong>to</strong> do with pensions. If anything,<br />
regulation has essentially been driven by specific<br />
events such as the collapse of the Enron and Maxwell<br />
Communication Corporation pension funds (which<br />
drove a lot of UK and US pension funds <strong>to</strong> change<br />
their investment framework) and by the attempt <strong>to</strong><br />
create a single financial solvency framework (notably<br />
in the Netherlands) between Life insurance companies<br />
and pension funds.<br />
Early on, market specialists realized that pension funds<br />
behaved like captive Life insurance companies and had<br />
<strong>to</strong> be regulated in a similar way. Therefore, changes in<br />
regula<strong>to</strong>ry frameworks have mostly, but not only, been<br />
driven by financial and solvency considerations rather<br />
than human resources considerations.<br />
Across the various markets, pension regula<strong>to</strong>ry and<br />
accounting changes in the UK and the Netherlands<br />
(and also <strong>to</strong> some extent in Switzerland and the US)<br />
were aimed essentially at:<br />
• Applying fair value <strong>to</strong> both sides of the balance sheet;<br />
• Maintaining relatively high regula<strong>to</strong>ry solvency and<br />
funding levels;<br />
• Increasing corporate sponsor responsibility; and<br />
• Increasing governance and education.<br />
These changes <strong>to</strong>uched upon many aspects of pension<br />
funds’ financial management. It is worth noting<br />
that the European Solvency II Directive will also require<br />
similar changes.<br />
APPLYING FAIR VALUE TO ASSETS<br />
AND LIABILITIES<br />
Pension assets have traditionally been (and remain)<br />
invested in liquid financial securities such as bonds,<br />
equities and properties, because these assets are:<br />
• Easy <strong>to</strong> moni<strong>to</strong>r;<br />
• Easy <strong>to</strong> periodically value; and<br />
• Easy <strong>to</strong> liquidate when required.<br />
Defined benefit pension liabilities are discounted using<br />
market-observable instruments rather than the fixed<br />
discount rates that were used for years:<br />
• Government bond yields or swap curves – for both<br />
interest rates and inflation – for solvency and funding<br />
purposes; and<br />
• AA corporate bond yields and breakeven inflation for<br />
accounting purposes.<br />
This has created a more consistent framework for both<br />
sides of the balance sheet.<br />
FUNDING LEVEL: THE CASE OF THE UK<br />
The changes in the UK funding framework responded <strong>to</strong><br />
the need for pension funds <strong>to</strong> be fully funded or close <strong>to</strong><br />
full funding at any point in time. If a pension fund is in<br />
deficit there is a requirement <strong>to</strong> get back <strong>to</strong> full funding<br />
within a relatively short period of time. Funding levels<br />
have become scheme-specific and depend on:<br />
• The size and nature of the liabilities;<br />
• The size of the deficit; and<br />
• The investment strategy.<br />
Given the existing funding deficits, most funds aim first<br />
for deficit reduction and try <strong>to</strong> achieve full funding in<br />
the medium <strong>to</strong> long term.<br />
SCOR - November 2010 - 53
What have been the main changes <strong>to</strong> the regulations?<br />
• Valuation assumptions now take in<strong>to</strong> account the<br />
<strong>risk</strong>iness of asset allocation and the probability of<br />
achieving return objectives;<br />
• Pension scheme trustees maintain a strict fund<br />
governance between valuation dates; and<br />
• Additional contributions are requested from the<br />
sponsor and recovery periods were initially set at<br />
seven years (because of the Credit Crisis, this has<br />
been extended <strong>to</strong> ten <strong>to</strong> fifteen years).<br />
UK funding requirements have become more<br />
conservative.<br />
SOLVENCY LEVEL: THE CASE<br />
OF THE NETHERLANDS<br />
The same phenomenon occurred in the Netherlands,<br />
with a somewhat different outcome.<br />
In the New Financial Assessment Framework (nFTK)<br />
produced by the Dutch National Bank, pension fund<br />
solvency measures have also been made schemespecific<br />
and dependent on:<br />
• The size and nature of the liabilities;<br />
• The size of the deficit; and<br />
• The investment strategy.<br />
The objectives were <strong>to</strong> maintain high solvency levels<br />
and reduce the volatility of the net position by creating<br />
an asset buffer that protects solvency against sudden<br />
drops. In order <strong>to</strong> define the buffer, three solvency<br />
tests (minimum, solvency and continuity) have been<br />
established using two different methods.<br />
This <strong>approach</strong> is driven by the concept of economic<br />
shortfall <strong>risk</strong>. Broadly speaking, if a pension fund has a<br />
<strong>risk</strong>y asset allocation, it should reserve more (i.e. set up<br />
a higher buffer) for solvency purposes as the probability<br />
<strong>to</strong> become insolvent increases. The specific constraint<br />
is that a pension fund should always remain funded<br />
over a twelve-month period on a 97.5 percentile<br />
probability basis.<br />
In practice, Dutch pension funds need about 25% -<br />
30% more assets than nominal liabilities <strong>to</strong> meet this<br />
requirement. The minimum funding requirement is<br />
105% on a nominal basis (i.e. without inflation) before<br />
triggering a reaction from the regula<strong>to</strong>r.<br />
Solvency requirements have also become more<br />
conservative for Dutch pension funds.<br />
54 - November 2010 - SCOR<br />
CORPORATE SPONSOR RESPONSIBILITY<br />
Corporate sponsors have been made more responsible<br />
from legal and financial standpoints in both countries.<br />
Their responsibilities and duties have been expanded:<br />
• Walking away from pension obligations has become<br />
very difficult;<br />
• Regula<strong>to</strong>rs in the UK and the Netherlands tend <strong>to</strong><br />
frown upon hazardous investment strategies and<br />
penalize them; and<br />
• Additional contribution levels and recovery periods<br />
are set on the basis of the size of the deficit and the<br />
<strong>risk</strong>iness of the sponsor and its capacity <strong>to</strong> pay.<br />
Changes in regulation and accounting standards have<br />
significantly affected the sponsoring companies.<br />
Corporate pension obligations are also becoming more<br />
transparent and are being scrutinized internally and<br />
externally. For example, in 2003 Standard & Poor’s<br />
produced a list that they called the “dirty dozen”,<br />
which consisted of around ten European companies<br />
with pension problems that could negatively impact<br />
their credit rating.<br />
GOVERNANCE AND EDUCATION<br />
Increased governance is the last corners<strong>to</strong>ne of the<br />
regula<strong>to</strong>ry changes. Pension funds have a duty <strong>to</strong>:<br />
• Draft investment principles and implementation<br />
guidelines;<br />
• Select investment and <strong>risk</strong> management advisors;<br />
• Moni<strong>to</strong>r asset performance and the evolution of<br />
liabilities.<br />
Filing with bodies such as the Dutch National Bank or the<br />
(UK) Pension Regula<strong>to</strong>r has also been made compulsory.<br />
Pension fund management bodies have therefore<br />
significantly invested in outside advice and education.<br />
They collaborate with sec<strong>to</strong>r associations, consulting<br />
firms, asset managers and investment banks offering<br />
advisory services and education programs <strong>to</strong> trustees<br />
and investment committees.
II. Implications & Challenges<br />
OVERALL IMPLICATIONS<br />
• The new regulations are moving <strong>to</strong>wards one<br />
common solvency framework for all financial<br />
institutions, as is already the case in the Netherlands<br />
and in some Scandinavian countries.<br />
• The new regulations turned pensions in<strong>to</strong> a<br />
problem scrutinized by equity and credit research<br />
analysts and by inves<strong>to</strong>rs. Insurance companies and<br />
pension funds suddenly appeared on the radars of<br />
rating agencies, equity analysts and credit research<br />
analysts.<br />
• They also turned pensions in<strong>to</strong> asset-liability and<br />
<strong>risk</strong> management problems, with long durations<br />
and potentially indexed liabilities. Before that,<br />
liabilities were fixed and pension funds had <strong>to</strong><br />
manage assets in a way that was consistent with a<br />
solvent situation.<br />
• Regula<strong>to</strong>rs require pension funds <strong>to</strong> be or <strong>to</strong><br />
remain well funded and solvent. The new regulations<br />
make it quite onerous <strong>to</strong> take <strong>risk</strong>s in the<br />
Netherlands or <strong>to</strong> count on corporate sponsors in<br />
the UK.<br />
These four points demonstrate that changes in regulation<br />
have also changed the nature of the problem. From a<br />
human resources problem it has become a financial<br />
institution problem with investment, <strong>risk</strong> management,<br />
corporate finance and credit rating implications.<br />
CHALLENGES<br />
Fig. 53: Applying fair value <strong>to</strong> Assets and Liabilities<br />
UK Liability Valuation Drivers<br />
8.0%<br />
7.0%<br />
6.0%<br />
5.0%<br />
4.0%<br />
3.0%<br />
2.0%<br />
Dec-99<br />
Jun-00 Dec-00 Jun-01<br />
UK 30-yr Gilt Yields<br />
Source: Bloomberg.<br />
Dec-01 Jun-02 Dec-02 Jun-03 Dec-03 Jun-04 Dec-04 Jun-05 Dec-05<br />
UK 30-yr BE Inflation<br />
The market observable indica<strong>to</strong>rs now used <strong>to</strong> discount<br />
liabilities are volatile. Market volatility now<br />
affects both sides of the pension fund’s balance<br />
sheet as well as those of the corporate sponsors.<br />
As shown in Figure 53, corporate bond yields moved<br />
from 5% <strong>to</strong> 7% between December 2005 and<br />
December 2008. Assuming a liability duration (average<br />
liability cash flow maturity) of about 20 years, a 2%<br />
move in interest corresponds <strong>to</strong> a 40% change in the<br />
value of liabilities.<br />
Likewise, during the same period, inflation moved from<br />
3% <strong>to</strong> almost 4%, and dropped by 1 percentage point<br />
in 2007 and 2008. These changes are also in the range<br />
of 20% <strong>to</strong> 30% of the <strong>to</strong>tal amount of liabilities.<br />
Sterling AA Credit Spread over 15yr<br />
Jun-06 Dec-06 Jun-07 Dec-07 Jun-08 Dec-08 Jun-09 Dec-09<br />
SCOR - November 2010 - 55
By moving in<strong>to</strong> a market-consistent framework for<br />
the valuation of assets and liabilities, the regulations<br />
brought large amounts of volatility <strong>to</strong> the pension funds<br />
and <strong>to</strong> the sponsors. This volatility is unrewarded as it<br />
applies <strong>to</strong> liabilities.<br />
The sponsors have become legally and financially<br />
responsible. They have <strong>to</strong> make their contributions<br />
according <strong>to</strong> the deficit and their financial capacity<br />
<strong>to</strong> help fund pension deficits.<br />
Figure 54 presents the additional pension fund<br />
contributions of 340 large public and private institutions<br />
in the UK. In the 90s these represented between<br />
GBP 500 million and 1 billion per quarter on average.<br />
Fig. 54: UK Corporate Sponsor Responsibility<br />
UK Sponsors’ Additional Contributions<br />
£ bn<br />
5.0<br />
4.0<br />
3.0<br />
2.0<br />
1.0<br />
0.0<br />
1992<br />
1993<br />
1994<br />
Special Contributions<br />
56 - November 2010 - SCOR<br />
1995<br />
1996<br />
1997<br />
1998<br />
1999<br />
4 Quarter Moving Average<br />
340 public and private funds (including some DC contributions).<br />
Source: PPF & ONS.<br />
2000<br />
When the problem arose in 2001/2002, contributions<br />
started <strong>to</strong> increase significantly. When the new<br />
regulations came in<strong>to</strong> force, the contribution levels<br />
increased <strong>to</strong> almost GBP 5 billion per quarter.<br />
In the mid 2000s sponsors could afford <strong>to</strong> pay high<br />
levels of contribution. However, since 2008, companies<br />
have had less cash flow, so they have not been able <strong>to</strong><br />
contribute as much. Contributions decreased <strong>to</strong> GBP<br />
1.5 billion. The assumption that corporate sponsors<br />
would be able <strong>to</strong> help did not prove <strong>to</strong> be accurate.<br />
Funding levels have improved but assets need<br />
<strong>to</strong> increase and/or liabilities <strong>to</strong> decrease in order<br />
<strong>to</strong> return <strong>to</strong> full funding.<br />
2001<br />
2002<br />
2003<br />
2004<br />
2005<br />
2006<br />
2007<br />
2008<br />
2009
Fig. 55: UK defined-benefit pension Assets & Liabilities<br />
UK pension Assets & Liabilities<br />
Assets & <strong>Risk</strong>-Free Liabilities (£ m)<br />
1,500,000<br />
1,250,000<br />
1,000,000<br />
750,000<br />
500,000<br />
250,000<br />
0<br />
2001<br />
Corporate Schemes Corporate Scheme Liabilities<br />
Source: OECD, PPF, Watson Wyatt & P-Solve Estimates.<br />
2002 2003 2004 2005 2006 2007 2008 2009<br />
Figure 55 shows that corporate pension plans<br />
roughly hold about GBP 1.2 trn of liabilities for about<br />
GBP 1.0 trn of assets. The actuarial funding level is<br />
estimated at 85.9% (almost the highest in the last<br />
10 years). Interestingly enough, the shock in 2008<br />
was not as marked as people thought because at the<br />
same time as bond yields went up, inflation went down.<br />
This provided relief on the liability side.<br />
Dutch pension funds’ solvency ratios have decreased<br />
and appear <strong>to</strong> remain volatile.<br />
Solvency levels fell sharply as a result of the credit<br />
crisis and remain volatile. If anything, since the new<br />
Dutch regulations came in<strong>to</strong> place, because of market<br />
issues, the funding level became even more volatile.<br />
Corporate Scheme Funding Level<br />
Fig. 56: Dutch defined-benefit pension Assets & Liabilities<br />
Dutch pension Assets & Liabilities<br />
Assets & <strong>Risk</strong>-Free Liabilities (€ m)<br />
1,000,000<br />
750,000<br />
500,000<br />
250,000<br />
0<br />
2001<br />
Assets Estimated Liabilities (Technical Provisions)<br />
Source: DNB & Dutch Insurers Association.<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
As indicated in Figure 56, the funding level in 2007<br />
was almost 150%. At the end of December 2009,<br />
Dutch pension funds roughly held about EUR 610 bn<br />
of liabilities for about EUR 666 trn of assets. The overall<br />
solvency ratio went down <strong>to</strong> below <strong>to</strong> 97% or 98%,<br />
then went back up <strong>to</strong> 109% (with a lot of pension<br />
funds still below the limit of 105%). Complex solvency<br />
models did not take extreme events in<strong>to</strong> account.<br />
These examples above show that, by themselves,<br />
regula<strong>to</strong>ry changes did not bring all of the outcomes<br />
expected. They did not bring full funding or more solvent<br />
pension systems, in the UK or in the Netherlands. Nor<br />
did they bring a reduction in funding level volatilities.<br />
2002 2003 2004 2005 2006 2007 2008 2009<br />
Estimated funding ratio (%)<br />
150%<br />
100%<br />
50%<br />
0%<br />
Funding Level<br />
Funding Level<br />
SCOR - November 2010 - 57
III. Changes in Investment and<br />
<strong>Risk</strong> <strong>Management</strong> Strategies<br />
What actually happened is, in many ways, far more<br />
interesting. To respond <strong>to</strong> the challenges of the<br />
changing regula<strong>to</strong>ry framework and environment, the<br />
investment and <strong>risk</strong> management strategies of defined<br />
benefit pension funds, as well as their benefit policies,<br />
have evolved dramatically over the last decade.<br />
Pension fund managers, trustees, investment committees,<br />
pension fund boards and corporate sponsors in<br />
both countries looked at the problem and unders<strong>to</strong>od<br />
that they had <strong>to</strong> change the way in which they thought,<br />
managed their <strong>risk</strong>s and invested their funds. They also<br />
modified their benefit and contribution structures.<br />
Liability and <strong>Risk</strong> <strong>Management</strong> had become crucial<br />
words by the second half of this decade.<br />
APPROACH TO CHANGES IN INVESTMENT<br />
AND RISK MANAGEMENT STRATEGIES<br />
The traditional de-<strong>risk</strong>ing <strong>approach</strong> for pension funds<br />
(and their corporate sponsors) consisted in selling all<br />
or part of their liabilities <strong>to</strong> insurance companies (bulk<br />
annuity buyout). However, because of their prohibitive<br />
cost, “pension liability buyout” was out of the reach of<br />
most pension funds, which preferred <strong>to</strong> decompose the<br />
entire <strong>approach</strong>, starting with a <strong>risk</strong> budgeting exercise<br />
where <strong>risk</strong>s would be identified and qualified.<br />
Figure 57 describes the various alternative stages <strong>to</strong> an<br />
annuity buyout.<br />
The process involves:<br />
• Reducing liabilities by closing down the defined<br />
benefit pension fund, limiting indexation or freezing<br />
the benefits.<br />
• Diversifying investments by moving out of listed<br />
local equities in<strong>to</strong> new asset classes.<br />
• Hedging unrewarded liability <strong>risk</strong>s such as interest<br />
rates and inflation and more recently longevity.<br />
The amount of <strong>risk</strong> reduced is similar <strong>to</strong> that of an<br />
annuity buyout, but the process is far less expensive.<br />
Fig. 57: Approach <strong>to</strong> changes in investment and <strong>Risk</strong> <strong>Management</strong> strategies<br />
Multiple instruments are used <strong>to</strong> achieve diversification and de-<strong>risk</strong>ing, starting with <strong>risk</strong> budgeting<br />
Increasing <strong>Risk</strong><br />
58 - November 2010 - SCOR<br />
Current Exposure<br />
<strong>Risk</strong> Budgeting<br />
Liability Reduction<br />
Diversification<br />
Interest Rate Hedging<br />
Inflation Hedging<br />
Longevity Hedging<br />
Insurance premium<br />
Increasing Cost<br />
Annuity buyout
RISK BUDGETING<br />
Pension funds and corporate sponsors are now<br />
developing pension <strong>risk</strong> budgets <strong>based</strong> on capital<br />
market and corporate finance models, in order <strong>to</strong> assess<br />
the implications of different investment and <strong>risk</strong> management<br />
strategies in terms of:<br />
• <strong>Risk</strong>-return profiles;<br />
• Solvency and funding;<br />
• Contributions; and<br />
• Corporate sponsor metrics (net debt, net income or<br />
free cash flow).<br />
<strong>Risk</strong> budgets tend <strong>to</strong> include:<br />
• <strong>Risk</strong> <strong>to</strong>lerance limits for both the pension fund and<br />
the sponsoring company;<br />
• Alternative investment and <strong>risk</strong> management strategies<br />
<strong>to</strong> be applied at varying levels of solvency<br />
through the use of trigger points; and<br />
• Differentiation in the treatment of rewarded and<br />
unrewarded <strong>risk</strong>s.<br />
Fig. 58: Liability reduction<br />
Membership by member type - UK<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
Actives<br />
Mar-06<br />
Source: PPF and P-Solve Estimates.<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
Deferred<br />
Approaches <strong>to</strong> investment and <strong>risk</strong> management strategies<br />
are increasingly influenced by capital market and<br />
investment banking activities.<br />
LIABILITY REDUCTION<br />
Corporate sponsors started closing down their defined<br />
benefit funds <strong>to</strong> new members in the late 1990s, but<br />
accelerated in the late 2000s. They also changed or<br />
reduced the indexation policy. More recently, they also<br />
started freezing defined benefit pension funds. Closure<br />
<strong>to</strong> new members and an increase in deferred members<br />
has reduced the pension liabilities of many pension<br />
funds. Defined contribution plans have been set up as<br />
an alternative.<br />
Figure 58 shows the impact of closing down and<br />
freezing plans on the membership profile in the UK<br />
and the Netherlands.<br />
33% 33% 36% 36% 37%<br />
41% 42%<br />
Mar-07 Mar-08 Mar-09<br />
Mar-10 (E)<br />
Retirees<br />
42% 43% 44%<br />
26% 25% 22% 21% 19%<br />
Membership by member type - NL<br />
Actives<br />
Dec-05<br />
Source: DNB and P-Solve Estimates.<br />
14% 15% 15% 15% 15%<br />
49% 50% 51% 52% 54%<br />
37% 35% 34% 33% 31%<br />
Deferred<br />
Dec-06 Dec-07 Dec-08<br />
Dec-10 (E)<br />
Retirees<br />
SCOR - November 2010 - 59
The number of UK active members fell from 26% in<br />
March 2006 <strong>to</strong> 19% in March 2010. The number of<br />
deferred members rose from 41% in March 2006 <strong>to</strong> 44%<br />
in March 2010. The number of Dutch active members<br />
fell from 37% in December 2005 <strong>to</strong> 31% in December<br />
2009. The number of deferred members rose from 49%<br />
in December 2005 <strong>to</strong> 54% in December 2009.<br />
In five years there has been a 20% reduction in active<br />
members for both countries.<br />
Fig. 59: Diversification - The case of the UK<br />
Pension Funds’ Asset Allocation<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
Dec-01<br />
60 - November 2010 - SCOR<br />
Alt.<br />
7%<br />
Bonds<br />
26%<br />
Equities<br />
67%<br />
Dec-03<br />
UK Equities<br />
Overseas Equities<br />
Alt.<br />
5%<br />
Bonds<br />
31%<br />
Equities<br />
64%<br />
Alt.<br />
11%<br />
Bonds<br />
28%<br />
Equities<br />
61%<br />
Gilts<br />
Index-Linked Gilts<br />
Corporate Bonds<br />
DIVERSIFICATION<br />
The case of the UK<br />
There has been a shift in UK pension fund asset<br />
allocations. Most of the pension funds were massively<br />
invested in equities (70%) at the beginning of the<br />
decade. By the end of the decade, however, equities<br />
represent around 53% and the UK equities home bias<br />
has been reduced (less than 25% of <strong>to</strong>tal assets). Bond<br />
allocation increased significantly with allocations <strong>to</strong><br />
index-linked and corporate bonds. However, the real<br />
change affected alternatives, which rose from less than<br />
5% <strong>to</strong> 16%. Pension funds invested in private equity,<br />
hedge funds, commodities, infrastructure and so on.<br />
Alt.<br />
11%<br />
Bonds<br />
29%<br />
Equities<br />
60%<br />
Alt.<br />
13%<br />
Bonds<br />
33%<br />
Equities<br />
54%<br />
Mar-06 Mar-07 Mar-08 Mar-09<br />
Mar-10 (E)<br />
Cash & Deposits<br />
Property<br />
Other Investments<br />
Alt.<br />
16%<br />
Bonds<br />
37%<br />
Equities<br />
47%<br />
Alt.<br />
15%<br />
Bonds<br />
32%<br />
Equities<br />
53%
Fig. 60: Diversification - The case of the Netherlands<br />
Pension Funds’ Asset Allocation<br />
100%<br />
80%<br />
60%<br />
40%<br />
20%<br />
0%<br />
Dec-01<br />
Alt.<br />
9%<br />
Bonds<br />
43%<br />
Source: DNB & OECD.<br />
Equities<br />
48%<br />
Dec-03<br />
Alt.<br />
11%<br />
Bonds<br />
43%<br />
Equities<br />
46%<br />
Euro Listed Equities<br />
Non-Euro Listed Equities<br />
Unlisted and Other Equities<br />
The case of the Netherlands<br />
Alt.<br />
8%<br />
Bonds<br />
43%<br />
Equities<br />
49%<br />
Dec-05 Dec-06 Dec-07 Dec-08<br />
Dec-09<br />
Eurozone Bonds<br />
Non-Euro Bonds<br />
Loans<br />
At the same time, Dutch pension funds started <strong>to</strong><br />
change their asset allocation, drastically reducing Eurolisted<br />
equities in traditional markets <strong>to</strong> between 10%<br />
and 15% of the <strong>to</strong>tal. The remaining equity investments<br />
are now in private equity, hedge funds, or are directly<br />
invested in companies or infrastructure funds. Dutch<br />
pension funds have also been investing in derivatives<br />
such property derivatives.<br />
Fig. 61: Liability hedging<br />
Asset Matches liability<br />
structure?<br />
Mortality derivatives<br />
Bonds<br />
Swaps<br />
Alt.<br />
8%<br />
Bonds<br />
36%<br />
Equities<br />
56%<br />
Alt.<br />
11%<br />
Bonds<br />
38%<br />
Equities<br />
51%<br />
Cash & Deposits<br />
Property<br />
Other Investments<br />
LIABILITY HEDGING<br />
Alt.<br />
17%<br />
Bonds<br />
41%<br />
Equities<br />
42%<br />
Alt.<br />
11%<br />
Bonds<br />
27%<br />
Equities<br />
62%<br />
<strong>Risk</strong> management techniques and strategies developed<br />
on the capital markets are now regularly used by<br />
pension funds. Liability hedging generally means<br />
the use of bonds, interest rates and swaps <strong>to</strong> reduce<br />
unrewarded (liability) <strong>risk</strong> by matching assets <strong>to</strong><br />
liabilities (i.e. liability-driven investment strategies).<br />
These instruments have been utilized by many pension<br />
funds in the Netherlands and in the UK. The choice<br />
between the instruments was dependent on:<br />
• liability structure; and<br />
• cost versus <strong>risk</strong> reduction benefits.<br />
Cost-effective? Currently used?<br />
SCOR - November 2010 - 61
LIABILITY-DRIVEN INVESTMENTS<br />
Fig. 62: Liability-driven investments<br />
LDI penetration in the UK<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
17%<br />
31-Mar-06<br />
20%<br />
Source: PPF & P-Solve estimates.<br />
23%<br />
26%<br />
30%<br />
31-Mar-07 31-Mar-08 31-Mar-09 31-Mar-10 (E)<br />
Liability-driven investment strategies are being adopted<br />
by an ever-increasing number of pension funds. These<br />
strategies are <strong>based</strong> on using either long-term bonds or<br />
derivatives that will genuinely match the liabilities.<br />
Liability-driven investment penetration is estimated at<br />
30% in the UK (PPF & P-Solve) and at 70% in the<br />
Netherlands (Aegon Global Pensions). According <strong>to</strong> the<br />
same source, pension funds in the two countries are<br />
well over 50% hedged.<br />
62 - November 2010 - SCOR<br />
LIABILITY BUYOUT MARKET<br />
Fig. 63: Liability buyout market<br />
Cumulative pension Liability buyouts in the UK<br />
Liabilities (GBP bn)<br />
20<br />
15<br />
10<br />
5<br />
0<br />
Source: LCP.<br />
2006<br />
2007 2008 2009 2010 - QI<br />
As may be expected, although the UK pension liability<br />
buyout market is growing, it has remained a limited and<br />
marginal phenomenon (about 1.5% of <strong>to</strong>tal corporate<br />
liabilities). According <strong>to</strong> the Dutch Insurers’ Association,<br />
pension contracts managed by insurance companies in<br />
the Netherlands represent up <strong>to</strong> 12% of <strong>to</strong>tal defined<br />
benefit pension liabilities.<br />
FIDUCIARY MANAGEMENT<br />
Finally, an increasing number of firms are offering<br />
fiduciary management services in the UK. A fiduciary<br />
manager is appointed by the pension fund <strong>to</strong> manage<br />
pension assets on its behalf, with varying degrees of<br />
freedom:<br />
• Traditional global / UK asset managers (BlackRock,<br />
SEI, Allianz Global Inves<strong>to</strong>rs);<br />
• Dutch pension aggrega<strong>to</strong>rs or <strong>risk</strong> management<br />
companies (Mn Services, Cardano, etc.); and<br />
• Investment consultants (Mercers, Hewitt, Towers<br />
Watson, P-Solve Asset Solutions).<br />
Although their <strong>approach</strong>es vary, they tend <strong>to</strong> focus<br />
on the execution of unconstrained asset allocations<br />
and the rapid implementation of investment and <strong>risk</strong><br />
management decisions. They aim <strong>to</strong>:<br />
• Help pension funds access more complex products<br />
and strategies;<br />
• Help attain a high level of diversification;<br />
• Undertake rotation of asset classes; and<br />
• Provide a higher level of professionalism.<br />
In the Netherlands, according <strong>to</strong> Bureau Bosch, 77%<br />
of assets are managed by local and foreign fiduciary<br />
managers. Often, they also manage interest rate<br />
hedges.
Fig. 64: Dutch fiduciary management<br />
Assets and Hedges<br />
(€ m)<br />
750<br />
500<br />
250<br />
Conclusion<br />
-<br />
2001<br />
Dutch Fiduciary Managers<br />
Source: Bureau Bosch.<br />
WHAT LESSONS HAVE COMPANIES<br />
LEARNED?<br />
This article has demonstrated that regula<strong>to</strong>ry<br />
changes do not always bring the expected outcomes.<br />
Nevertheless, they trigger changes in investment and<br />
<strong>risk</strong> management behavior, which take relatively little<br />
time <strong>to</strong> take shape. When pension funds are faced<br />
with such profound changes as <strong>risk</strong>-<strong>based</strong> capital<br />
and fair value for liabilities, plus governance, plus <strong>risk</strong><br />
management, plus transparency, there is no other<br />
2002 2003 2004 2005 2006 2007<br />
2008<br />
Foreign Fiduciary Managers<br />
choice but <strong>to</strong> change themselves. In the UK and the<br />
Netherlands these changes are already in place. If the<br />
same pattern is applied <strong>to</strong> the insurance industry with<br />
Solvency II, the first major changes will be perceptible<br />
in 2017 at the latest.<br />
Education and governance are unquestionably key<br />
in terms of comprehending (and benefitting from)<br />
regula<strong>to</strong>ry changes. Pension funds have made many<br />
efforts <strong>to</strong> change their behavior and <strong>to</strong> provide more<br />
professional advice, sometimes subcontracting their<br />
assets <strong>to</strong> someone else. Another fundamental change<br />
has been embedding their <strong>risk</strong> management <strong>approach</strong>es<br />
in<strong>to</strong> investment strategies.<br />
SCOR - November 2010 - 63
7<br />
<strong>ERM</strong> AND ECONOMIC<br />
CAPITAL MODELS:<br />
THE A.M. BEST VIEW MILES TROTTER<br />
General Manager,<br />
Analytics Non-Life, A.M. Best<br />
This article proposes <strong>to</strong> explore<br />
<strong>ERM</strong> and economic capital from the rating agency<br />
view. In view of the current challenges, how do<br />
<strong>ERM</strong> and models fit in<strong>to</strong> the A.M. Best rating<br />
process?<br />
This article begins with an overview of A.M. Best<br />
rating process, then it establishes the general background<br />
of <strong>ERM</strong> in the (re)insurance industry; last but<br />
not least it poses a certain number of questions concerning<br />
<strong>ERM</strong> and models in the rating process and<br />
their ongoing improvements.<br />
Responding <strong>to</strong> the insurance industry trends and<br />
practices, A.M. Best has been accompanying and<br />
supporting (re)insurance companies since 1899.<br />
64 - November 2010 - SCOR<br />
I. The rating process<br />
Fig. 65: The rating process<br />
This above graph summarizes the rating process.<br />
Market Profile<br />
Operating Performance<br />
Balance Sheet Strength<br />
The A.M. Best rating process involves 3 components:<br />
• The balance sheet strength which is the bedrock<br />
for any rating;<br />
• The operating performance which supports the<br />
balance sheet strength through retained earnings;<br />
• The market profile which is the more subjective part<br />
of the rating, access <strong>to</strong> business is notably assessed<br />
in this part of the process. For instance, if a company<br />
has dedicated access <strong>to</strong> certain business and other<br />
companies cannot compete with it, this will protect<br />
its earnings and support its financial strength.
A.M. Best rating has six secure financial strength ratings<br />
which range from B+, the lowest and the first secure<br />
rating, <strong>to</strong> A++, the highest rating. The higher the rating<br />
level is, the more important the marketing profile is. The<br />
balance sheet strength and the operating performance<br />
are givens. The profile is the component which enables<br />
<strong>to</strong> distinguish between higher rating levels.<br />
THE BEST CAPITAL ADEQUACY RATIO (BCAR)<br />
A.M. Best has developed a proprietary capital model,<br />
the Best Capital Adequacy Ratio (BCAR). This ratio is<br />
used for the analysis of the balance sheet strength.<br />
This is a simple deterministic or fac<strong>to</strong>r-<strong>based</strong> model,<br />
i.e. a market surveillance type model that can be<br />
Fig. 66: A.M. Best’s Capital Adequacy Ratio (BCAR)<br />
+ (B1) Fixed-Income Securities<br />
+ (B2) Equity Securities<br />
+ (B3) Interest Rate<br />
+ (B4) Credit<br />
+ (B5) Loss Reserves<br />
+ (B6) Net Written Premium<br />
+ (B7) Off Balance Sheet <strong>Risk</strong>s<br />
– Covariance<br />
AM Best’s Capital Adequacy Ratio (BCAR)<br />
rapidly populated with reasonably limited data. This<br />
benchmark contributes usefully <strong>to</strong> the rating process.<br />
The BCAR score expresses a percentage. As depicted in<br />
the tabular layout, the <strong>to</strong>tal adjusted capital is divided<br />
by the net required capital. The <strong>to</strong>tal adjusted capital<br />
corresponds <strong>to</strong> the available capital used <strong>to</strong> pay claims<br />
and the <strong>to</strong>tal required capital is obtained by applying<br />
fac<strong>to</strong>rs <strong>to</strong> the different elements on the left of the<br />
below table.<br />
The score is obtained from the model and a one-<strong>to</strong>one<br />
relationship between <strong>to</strong>tal adjusted capital and net<br />
required capital is needed <strong>to</strong> achieve a secure rating<br />
with A.M. Best, which is B+ on its scale.<br />
Net required capital Total adjusted capital<br />
+ Shareholders funds<br />
+ Positive adjustments. Typically:<br />
• Discounting of reserves<br />
• Difference between market<br />
and book values<br />
• Value of In-Force Business (VIF)<br />
• Economic Reserves e.g. RfBs (1)<br />
• Hybrid Equity<br />
– Negative adjustments:<br />
• Deduction of one cat PML<br />
• DAC<br />
BCAR score = <strong>to</strong>tal adjusted capital / net required capital<br />
(1) Reserven für Beitragsrückerstattung - technical refund reserve under German life policies.<br />
SCOR - November 2010 - 65
II. <strong>Risk</strong> <strong>Management</strong><br />
<strong>Risk</strong> <strong>Management</strong> <strong>to</strong>ols and practices have advanced<br />
significantly in recent years. <strong>ERM</strong> has been the industry<br />
response <strong>to</strong> growing exposure <strong>to</strong> volatility in company<br />
earnings and capital. Reinsurance and insurance companies<br />
face emerging <strong>risk</strong>s such as terrorist attacks,<br />
natural catastrophes such as hurricanes (four hurricane<br />
events in six weeks in 2004, three events in 2005) and<br />
more recently the credit crisis whose naughty child,<br />
the economic crisis, continues. The industry has been<br />
heading <strong>to</strong>wards a higher <strong>risk</strong> profile. The (re)insurance<br />
companies have taken steps <strong>to</strong> manage their exposure,<br />
mitigate <strong>risk</strong>s and preserve policyholder security.<br />
On the Non-Life side, catastrophe modeling is an<br />
example of this.<br />
ENTERPRISE RISK MANAGEMENT<br />
FRAMEWORK<br />
<strong>ERM</strong> has been a major addition <strong>to</strong> insurers’ vocabulary.<br />
Its foundations are rooted in traditional <strong>Risk</strong><br />
<strong>Management</strong> practices and controls as shown in the<br />
pyramid. It constitutes a natural extension of these<br />
practices and only the enterprise view where the<br />
(re)insurers now identify, quantify and measure the<br />
different <strong>risk</strong>s that expose the company on a holistic<br />
basis, is something new.<br />
<strong>ERM</strong> is <strong>based</strong> on three main areas:<br />
• <strong>Risk</strong> Culture: the company has <strong>to</strong> establish a <strong>risk</strong><br />
culture that embeds <strong>risk</strong> awareness and accountability<br />
in daily operations;<br />
Fig. 67: <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong><br />
framework<br />
Senior<br />
<strong>Management</strong><br />
<strong>ERM</strong><br />
and ICM*<br />
Capital<br />
<strong>Management</strong><br />
Traditional <strong>Risk</strong> <strong>Management</strong><br />
Practices and Controls<br />
* Establish <strong>Risk</strong>-aware Culture, including proper alignment<br />
of management incentives<br />
* Implement Improved <strong>Risk</strong> Identification and <strong>Management</strong><br />
* Develop Sophisticated <strong>Risk</strong> Measurement Tools, e.g. ICM<br />
The pyramid illustrates the connection between <strong>ERM</strong><br />
and (re)insurer’s traditional practices and controls.<br />
66 - November 2010 - SCOR<br />
• <strong>Risk</strong> Identification and <strong>Management</strong>: the ability <strong>to</strong><br />
identify key <strong>risk</strong>s across an organization and set uniform<br />
controls and procedures <strong>to</strong> manage their impacts;<br />
• <strong>Risk</strong> Measurement: consists of using sophisticated<br />
<strong>to</strong>ols and data collection <strong>to</strong> quantify <strong>risk</strong>, including<br />
the impact of correlation, economic conditions and<br />
extreme events. The results are then reported <strong>to</strong> senior<br />
management.<br />
<strong>ERM</strong> is very much about the aggregate view for the<br />
organization as a whole.<br />
Hopefully the hard lessons of the recent past will be<br />
learnt. Many governments have taken unprecedented<br />
steps <strong>to</strong> bail out, or in some cases not bail out, some of<br />
the largest institutions in the world. The current economic<br />
situation is very volatile and now the focus is shifting <strong>to</strong><br />
sovereign debt issues and the possibility of a double dip<br />
recession. In any case, there is lag in foreclosures and it<br />
is likely that there will be another spike in foreclosures in<br />
2010/2011.<br />
Prior <strong>to</strong> 2008, many institutions felt that their <strong>ERM</strong> capital<br />
modeling processes were really good at getting them<br />
ahead of the <strong>risk</strong> curve for a more efficient use of capital.<br />
However, 2008 has definitely come as a lesson <strong>to</strong> the<br />
effect that tail events are not just scenarios <strong>to</strong> be modeled<br />
and statistically analyzed; they actually happen.<br />
DOES THAT MEAN THAT <strong>ERM</strong> IS A FAILURE?<br />
<strong>ERM</strong> at some institutions definitely failed, but this<br />
does not mean that <strong>ERM</strong> itself is a failure. The crisis<br />
impact on (re) insurers and other financial institutions<br />
demonstrates that there were gaps in their <strong>ERM</strong><br />
programs, notably in the understanding of individual<br />
<strong>risk</strong>s and <strong>risk</strong> correlation and dependencies. Insurers<br />
were used <strong>to</strong> having a high level of liquidity, a strong<br />
cash position. The crisis taught companies that cash<br />
management and liquidity analysis must be integrated<br />
in<strong>to</strong> the <strong>Risk</strong> <strong>Management</strong> framework, including the<br />
s<strong>to</strong>chastic analysis and modeling. And it is necessary<br />
<strong>to</strong> keep on challenging the <strong>risk</strong> inven<strong>to</strong>ry and looking<br />
for emerging <strong>risk</strong>s.<br />
In the (re)insurance industry, some companies had some<br />
difficulties. The key points on the lessons they learned<br />
from the crisis could be summarized as follows:<br />
• Regularly recalibrate stress test assumptions and<br />
confirm their overall robustness;<br />
• Update assumptions concerning the dependency of<br />
securitized assets <strong>to</strong> reflect market conditions;<br />
• Ensure that the dependencies between the equity and<br />
the corporate credit <strong>risk</strong>s are adequately modeled;<br />
• Enhance the financial acumen of the reviewing team<br />
<strong>to</strong> reflect the importance of judgment in reviewing<br />
model output.
Fig. 68: Why did insurance companies fail?<br />
It is worth listening <strong>to</strong> the experiences of institutions<br />
during the crisis. And one of the critical lessons <strong>to</strong> learn<br />
from the crisis is that the fundamentals matter. Specialists<br />
recommend back <strong>to</strong> basics: credit analysis, internal<br />
underwriting control, sound financial management,<br />
common sense and discipline. Most companies<br />
emphasize that modeling is important, but they have<br />
<strong>to</strong> use their judgment. <strong>ERM</strong> is a continuous process,<br />
a mindset, not simply a task <strong>to</strong> be performed.<br />
WHY DID INSURANCE COMPANIES FAIL?<br />
As shown in the pie chart above, the leading reason<br />
for failure is a combination of deficient loss reserves<br />
and inadequate pricing, followed by excessive growth.<br />
Should a company conclude from these elements that<br />
it must focus on reserving prudently, charging the right<br />
prices and may be not pursuing for <strong>to</strong>o much growth?<br />
If <strong>ERM</strong> teaches anything, it is that the answer <strong>to</strong> that<br />
question is definitely no.<br />
<strong>Risk</strong> <strong>Management</strong> should cover the whole organization<br />
and any or all of these <strong>risk</strong>s should be managed by a<br />
company if they are relevant <strong>to</strong> the company. Any of<br />
these reasons for failure could be attributed <strong>to</strong> a failure<br />
of <strong>Risk</strong> <strong>Management</strong>.<br />
It is worth noting that in the rating process <strong>Risk</strong><br />
<strong>Management</strong> constitutes the common thread which<br />
links balance sheet strength, operating performance<br />
and business profile. <strong>Risk</strong> <strong>Management</strong> is used in the<br />
company’s strategic decision making process <strong>to</strong> define<br />
its profile but also in its financial management practices<br />
which dictate the sustainability of its performance, and<br />
ultimately its exposure <strong>to</strong> capital volatility.<br />
A failure in <strong>Risk</strong> <strong>Management</strong><br />
Deficient loss reserves/Inadequate Pricing<br />
Rapid growth<br />
Alleged fraud<br />
Catastrophe losses<br />
Affiliate impairment<br />
Investment problems<br />
Miscellaneous<br />
Significant change in business<br />
Reinsurance failure<br />
Source: A.M. Best: 1969-2008 Impairment Review,<br />
Special Report, April 6, 2009.<br />
Deficient loss reserves, inadequate pricing, and rapid growth are the leading triggers. Investment and catastrophe losses play<br />
a much smaller role. This pie chart represents the impairment of the US Property and Casualty industry from 1969 <strong>to</strong> 2008.<br />
If a company has sound <strong>Risk</strong> <strong>Management</strong> and executes<br />
its strategy effectively, it will preserve and build its<br />
capital strength and perform successfully over the long<br />
term and is likely <strong>to</strong> have a strong rating.<br />
That is why A.M. Best does not produce separate<br />
ratings of companies for <strong>ERM</strong>. <strong>Risk</strong> <strong>Management</strong> is the<br />
element which supports the whole process. All these<br />
elements form a part of A.M. Best rating.<br />
Fig. 69: Financial Strength Ratings<br />
An opinion as <strong>to</strong> an insurer’s financial strength<br />
and ability <strong>to</strong> meet its ongoing obligations<br />
<strong>to</strong> policyholders<br />
Balance<br />
Sheet<br />
Strength<br />
Insurance Company<br />
Financial Strength<br />
Operating<br />
Performance<br />
<strong>Risk</strong><br />
<strong>Management</strong><br />
Best’s<br />
Rating<br />
Business<br />
Profile<br />
SCOR - November 2010 - 67
Fig. 70: Traditional RM: silo <strong>approach</strong><br />
Underwriting<br />
<strong>Risk</strong><br />
TRADITIONAL RISK MANAGEMENT<br />
As already mentioned, one has <strong>to</strong> explore traditional<br />
<strong>Risk</strong> <strong>Management</strong> in order <strong>to</strong> comprehend the recent<br />
developments of <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong> in the<br />
(re)insurance industry. Traditional <strong>Risk</strong> <strong>Management</strong><br />
refers <strong>to</strong> the policies and procedures that companies<br />
have set <strong>to</strong> identify, quantify and measure each of their<br />
<strong>risk</strong>s separately or individually. The main categories of<br />
<strong>risk</strong> for (re)insurers are:<br />
• Credit <strong>risk</strong>;<br />
• Market <strong>risk</strong>;<br />
• Underwriting <strong>risk</strong>;<br />
• Operational <strong>risk</strong>;<br />
• Strategic <strong>risk</strong>.<br />
Market<br />
<strong>Risk</strong><br />
The leading feature of traditional management is the<br />
separate assessment of <strong>risk</strong>s which is often referred<br />
<strong>to</strong> as a silo <strong>approach</strong>. It differs from <strong>Enterprise</strong> <strong>Risk</strong><br />
<strong>Management</strong>, which utilizes a holistic process.<br />
A.M. Best does not deem <strong>ERM</strong> <strong>to</strong> be the sole answer for<br />
(re)insurers. Each company may pursue its own <strong>approach</strong>.<br />
Thus, the traditional <strong>approach</strong> could be modified, using<br />
<strong>ERM</strong> concepts, <strong>to</strong> be more effective. A company has <strong>to</strong><br />
understand the correlation between <strong>risk</strong>s. If a company<br />
modifies its traditional <strong>approach</strong> in this way, it can still<br />
be considered strong by A.M. Best.<br />
Credit<br />
<strong>Risk</strong><br />
Traditional RM is typically implemented using a “silo” <strong>approach</strong> with little or no formal<br />
interaction, communication, or alignment among <strong>risk</strong> managers.<br />
68 - November 2010 - SCOR<br />
Operational<br />
<strong>Risk</strong><br />
Strategic<br />
<strong>Risk</strong><br />
III. <strong>ERM</strong> in the rating process<br />
A.M. Best seeks <strong>to</strong> make an open-minded assessment<br />
of a company’s operating practices. The onus is on the<br />
company <strong>to</strong> show that its <strong>Risk</strong> <strong>Management</strong> process is<br />
tailored <strong>to</strong> its individual business needs, is flexible and<br />
adaptive and works.<br />
The crisis has proved that Chief <strong>Risk</strong> Officers, <strong>risk</strong> committees,<br />
corporate dashboards, corporate <strong>risk</strong> appetites<br />
and s<strong>to</strong>chastic capital models are certainly all valid parts<br />
of a robust <strong>ERM</strong> program, but strong fundamental <strong>risk</strong><br />
management in a sustainable business model comes in<br />
different shapes and sizes.<br />
Furthermore, the role of rating agencies such as<br />
A.M. Best is by no means <strong>to</strong> prescribe <strong>to</strong> companies<br />
what <strong>ERM</strong> should be or should not be. A.M. Best’s<br />
portfolio of rated companies is <strong>to</strong>o diverse <strong>to</strong> impose<br />
such a measure.
Fig. 71: Catastrophe exposure in BCAR<br />
THE CATASTROPHE EXPOSURE<br />
Standard BCAR Two-Event Cat Stress Test<br />
EVENT – 1/100 Wind or 1/250 Earthquake 1 st EVENT – 1/100 Wind or 1/250 Earthquake<br />
ADJUSTMENT<br />
– Subtract after-tax net PML from Surplus<br />
As part of the rating assessment, A.M. Best uses its<br />
own model i.e. the BCAR model. For catastrophe <strong>risk</strong><br />
management, BCAR-<strong>based</strong> tests are run. A 1/100 years<br />
winds<strong>to</strong>rm after-tax net Probable Maximum Loss (PML)<br />
or 1/250 years earthquake after-tax net PML, whichever<br />
is the largest, is deducted from the surplus, the capital<br />
being therefore always viewed net of the largest<br />
PML. But following the events of 2005, a two-event<br />
catastrophe stress test was introduced <strong>to</strong> the model.<br />
As shown in Figure 71, in addition <strong>to</strong> the deduction of<br />
the largest after-tax net PML from the surplus for the<br />
first event, 40% of the pre-tax net PML is added <strong>to</strong> loss<br />
reserves, and 40% of the reinsurance recoverables are<br />
added <strong>to</strong> the credit <strong>risk</strong>. For the second event, there is<br />
a further deduction from surplus of the largest after-tax<br />
net PML.<br />
The second event for earthquake drops down <strong>to</strong> a 1 in<br />
100 event reflecting the likelihood that a second earthquake<br />
in the same place will be less severe. The idea is<br />
obviously <strong>to</strong> test capital for multiple events. The normal<br />
<strong>to</strong>lerance would be a 30-point drop in the BCAR from<br />
the minimum guideline. If a company’s BCAR decreases<br />
under this threshold, then there would be questions <strong>to</strong><br />
ask the company.<br />
ADJUSTMENT<br />
– Subtract after-tax net PML from Surplus<br />
+ Add 40% of pre-tax net PML<br />
<strong>to</strong> Loss Reserves<br />
+ Add 40% of reinsurance recoverables<br />
<strong>to</strong> credit <strong>risk</strong><br />
2 nd EVENT – 1/100 Wind or 1/100 Earthquake<br />
ADJUSTMENT<br />
– Reduce surplus by after-tax net PML<br />
Here are some of the questions that (re)insurers are<br />
likely <strong>to</strong> face from their A.M. Best analysts:<br />
• Describe the involvement of the Board of Senior <strong>Management</strong><br />
within your <strong>risk</strong> management framework;<br />
• Has your organization established and communicated<br />
any <strong>risk</strong> management objectives <strong>to</strong> your employees,<br />
and other stakeholders?<br />
• How does your organization encourage good <strong>risk</strong><strong>based</strong><br />
decision making?<br />
• What is your organization’s process for identifying<br />
and cataloging key <strong>risk</strong>s?<br />
• What <strong>to</strong>ols does your organization use <strong>to</strong> determine<br />
required capital?<br />
• How do you fac<strong>to</strong>r in correlation/dependency of<br />
individual <strong>risk</strong>s?<br />
• How do you incorporate operational <strong>risk</strong> and strategic<br />
<strong>risk</strong> in your evaluation of required capital?<br />
• How are emerging <strong>risk</strong>s identified and evaluated?<br />
• What are the <strong>to</strong>p <strong>risk</strong>s exposures your organization<br />
faces, and how are these <strong>risk</strong>s managed?<br />
The questions focus on the usual areas: corporate<br />
culture, <strong>risk</strong> controls, <strong>risk</strong> models, particularly capital<br />
models, emerging <strong>risk</strong>s. With reference <strong>to</strong> the last<br />
question shown in the list above, analysts will always<br />
ask for examples. If the example quoted is a mistake,<br />
the company will have <strong>to</strong> explain what it learnt from<br />
this mistake. Did it lead <strong>to</strong> any changes in their <strong>risk</strong><br />
management process?<br />
SCOR - November 2010 - 69
Fig. 72: <strong>Risk</strong> <strong>Management</strong> and BCAR<br />
A.M. Best’s traditional <strong>approach</strong><br />
BCAR<br />
HOW TO DET<strong>ERM</strong>INE REQUIRED CAPITAL?<br />
As for the determination of required capital, it is <strong>based</strong><br />
on an interactive dialogue comparing:<br />
• Company’s view of capital (e.g. ICM);<br />
• BCAR score and his<strong>to</strong>rical trends;<br />
• Holding company issues;<br />
• Other models or metrics.<br />
Weak <strong>Risk</strong><br />
<strong>Management</strong><br />
Strong <strong>Risk</strong><br />
<strong>Management</strong><br />
BCAR Guideline<br />
Low Exposure <strong>to</strong> earnings and<br />
Capital volatility<br />
High<br />
A.M. Best does not import results from companies’<br />
internal models in<strong>to</strong> its model but this does not mean<br />
A.M. Best ignores company views or analysis. These<br />
form an essential part of the decision making process<br />
for the rating.<br />
In many instances, companies reaching the same<br />
BCAR position may have different ratings <strong>based</strong> on the<br />
integration of other considerations, including financial<br />
management practices and operating elements which<br />
dictate performance sustainability and capital volatility.<br />
<strong>Risk</strong> <strong>Management</strong> is a key fac<strong>to</strong>r <strong>to</strong> determine the<br />
amount of capital required for a given rating.<br />
Companies with strong <strong>Risk</strong> <strong>Management</strong> and relatively<br />
low exposure <strong>to</strong> earnings and capital volatility, will have<br />
capital at the BCAR guideline level. For companies with<br />
weak <strong>Risk</strong> <strong>Management</strong>, it is difficult <strong>to</strong> maintain the<br />
fixed minimum of required capital. Therefore, they need<br />
<strong>to</strong> obtain a higher BCAR score, even if they maintain<br />
a low exposure <strong>to</strong> volatility. As exposure increases, the<br />
BCAR requirement also increases at a faster rate.<br />
Fig. 73: <strong>Risk</strong> <strong>Management</strong> and BCAR<br />
A.M. Best’s current <strong>approach</strong><br />
BCAR<br />
What’s different…<br />
A.M. Best will consider<br />
allowing companies with<br />
STRONG <strong>Risk</strong> <strong>Management</strong><br />
<strong>to</strong> maintain lower BCAR<br />
levels relative <strong>to</strong><br />
the rating guideline<br />
Weak <strong>Risk</strong><br />
<strong>Management</strong><br />
Strong <strong>Risk</strong><br />
<strong>Management</strong><br />
BCAR Guideline<br />
Low Exposure <strong>to</strong> earnings and<br />
Capital volatility<br />
High<br />
The A.M. Best traditional view is depicted in the first chart above (Fig. 72). The vertical axis corresponds <strong>to</strong> the BCAR score, and the<br />
horizontal axis represents exposure <strong>to</strong> earnings and capital volatility. The second chart shows its current view.<br />
70 - November 2010 - SCOR<br />
The A.M. Best view has evolved in recognition of<br />
increased sophistication in <strong>Risk</strong> <strong>Management</strong> practices<br />
in the industry. Lower BCAR scores relative <strong>to</strong> the<br />
guideline are considered for companies with a strong<br />
<strong>Risk</strong> <strong>Management</strong>. In doing this A.M. Best follows the<br />
improvements of the industry and adapts its rating<br />
process.<br />
THE INTERNAL CAPITAL MODELS (ICMs)<br />
APPROACH<br />
The A.M. Best <strong>approach</strong> <strong>to</strong> ICMs is similar <strong>to</strong> <strong>ERM</strong> in<br />
that it is handled as part of the rating process and not<br />
separately, although a supplementary review may be<br />
appropriate. This is a review, not an audit, with the<br />
objective of understanding the company’s modeling<br />
and comparing this with the BCAR.<br />
The BCAR is a deterministic model, built from his<strong>to</strong>ric<br />
analysis of industry level data. So for net premiums<br />
written, A.M. Best has his<strong>to</strong>ric data by class of business<br />
and arrives at a fac<strong>to</strong>r using that. With the company’s<br />
model, it will be <strong>based</strong> on its own <strong>risk</strong> data and maybe<br />
s<strong>to</strong>chastic or may have s<strong>to</strong>chastic elements <strong>to</strong> it. So the<br />
two cannot be mixed. Rather than trying <strong>to</strong> drop the<br />
company’s analysis in<strong>to</strong> the BCAR model, A.M. Best<br />
simply tries <strong>to</strong> understand differences and why they are<br />
occurring. This sort of comparison and the conclusions<br />
that arise from it are very much a day-<strong>to</strong>-day part of<br />
the rating process.
To gain comfort in the integrity of the ICM process<br />
and results, A.M. Best seeks information on the following<br />
items:<br />
• Granularity of the model and data integrity;<br />
• Time horizons;<br />
• <strong>Risk</strong> metric (VaR… TvaR… CTE);<br />
• Assumptions and scenario testing;<br />
• Timeliness, availability and applicability of data;<br />
• Linkages <strong>to</strong> financial reporting and strategic<br />
planning;<br />
• Operational, strategic and emerging <strong>risk</strong>s;<br />
• Disclosure (internal and external);<br />
• Internal and external “audit” findings;<br />
• Next steps in ICM development.<br />
A company may have a model that looks very good, but<br />
if it takes six months for the company <strong>to</strong> run it, then it is<br />
hardly going <strong>to</strong> be fully embedded within the company’s<br />
decision making. A.M. Best also wants <strong>to</strong> know what<br />
the company does about the more esoteric or difficult<br />
<strong>to</strong> measure areas of <strong>risk</strong>, whether operational, strategic<br />
or emerging <strong>risk</strong>s and how widely disseminated the<br />
results are. Does the company see its model as the be<br />
all and end all of modeling for the company, or are<br />
there future developments expected?<br />
SOLVENCY II: THE FUTURE FOR RISK<br />
MANAGEMENT IN EUROPE<br />
Solvency II introduces <strong>risk</strong>-<strong>based</strong> regulation, the implications<br />
of which are summarized as follows:<br />
• This directive will lead <strong>to</strong> an increased focus on <strong>Risk</strong><br />
<strong>Management</strong> and <strong>to</strong> an improvement of standards<br />
in Europe;<br />
• More companies are expected <strong>to</strong> use Internal Capital<br />
Models because the level of required capital is very<br />
likely going <strong>to</strong> increase. Therefore, they will have <strong>to</strong><br />
seek approval and this will include demonstrating<br />
that the ICM is embedded in the organization. The<br />
time frame for obtaining approval is a tight one;<br />
• In the longer term, more and more companies<br />
will be using internal capital models that will result<br />
in strengthening <strong>risk</strong> management standards.<br />
The ORSA (Own <strong>Risk</strong> and Solvency Assessment)<br />
means that companies will have <strong>to</strong> incorporate <strong>risk</strong><br />
in<strong>to</strong> their strategic management.<br />
Conclusion<br />
A.M. Best will continue <strong>to</strong> look at <strong>ERM</strong> and ICM models<br />
within the existing rating process, which is an interactive<br />
one.<br />
Companies have <strong>to</strong> be practical in their <strong>approach</strong>, do<br />
what makes sense <strong>to</strong> them and advise A.M. Best why<br />
it works for them.<br />
SCOR - November 2010 - 71
8<br />
MACRO-ECONOMIC<br />
STANDPOINT MICHÈLE LACROIX<br />
Chief Investment Officer, SCOR<br />
This article presents the financial<br />
and economic difficulties caused by the current<br />
environment and the position adopted by SCOR<br />
in this context. The first section analyzes the crisis<br />
in all its phases; the second before examining<br />
SCOR’s view on the financial and economic issues<br />
involved in the second section. The last section<br />
highlights how the <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong><br />
process is implemented in<strong>to</strong> SCOR’s asset management<br />
decisions.<br />
I. The global economy is facing<br />
a critical transition phase<br />
To understand the current situation, it is necessary <strong>to</strong><br />
explore the pre-crisis period, which will be referred <strong>to</strong><br />
as the leveraging period. Saving rates decreased constantly<br />
during this period, which was itself split in<strong>to</strong><br />
3 phases:<br />
• Financing US economic growth: up until 2001, non-US<br />
residents financed the real economy, investing in<br />
corporate bonds and helping <strong>to</strong> implement real<br />
growth.<br />
72 - November 2010 - SCOR<br />
• Financing US public debt: from 2001 <strong>to</strong> 2005, nonresidents<br />
invested in the US <strong>to</strong> finance the budget<br />
deficit and the war in Iraq. This period corresponds <strong>to</strong><br />
what Alan Greenspan used <strong>to</strong> call a “conundrum”,<br />
i.e. a period with no inflation, low his<strong>to</strong>rical rates and<br />
strong incentives <strong>to</strong> borrow money.<br />
• Financing households: at this time, and as has often<br />
been the case in US his<strong>to</strong>ry, savings rates were <strong>to</strong>o<br />
low <strong>to</strong> finance consumer spending. In this context<br />
Alan Greenspan favored securitization, which<br />
enabled financial institutions <strong>to</strong> transfer their <strong>risk</strong>s <strong>to</strong><br />
other economic agents with fewer legal constraints.<br />
Securitization became the easiest way <strong>to</strong> finance<br />
local loans with non-US residents’ investments, which<br />
explains why the entire world was exposed <strong>to</strong> the US<br />
mortgage market and why the bursting of the US real<br />
estate bubble had global consequences.<br />
As shown in Figure 74, over the last three years the<br />
debt burden has simply moved from the private <strong>to</strong> the<br />
public sphere, staying at the same level and actually<br />
postponing global deleveraging of the economy.
Fig. 74: The deleveraging years<br />
Markets lost confidence<br />
in financial institutions<br />
➜ the first dip<br />
The end of the real estate rally clearly resulted in<br />
the collapse of the system: the deleveraging of US<br />
households came first, but the level of debt could not<br />
easily be lowered given the low level of inflation.<br />
The second step, i.e. the deleveraging of financial<br />
institutions, was more visible: the 2008 liquidity crisis<br />
showed how lack of confidence is critical in the financial<br />
markets and can create panic. Markets concerns were<br />
addressed by outstanding bailout programs, which<br />
resulted in the transfer of private debt <strong>to</strong> public debt;<br />
nevertheless the burden of the debt has not decreased,<br />
so the deleveraging process is far from over.<br />
Small increase in saving rates<br />
Subprime bubble Financial crisis Sovereign debt crisis<br />
➜ End of real estate rally<br />
➜ Deleveraging of consumers<br />
➜ Household have tried <strong>to</strong><br />
reduce their level of debt<br />
but low inflation does not<br />
help such a move<br />
➜ Lehman’s and AIG’s bankruptcy<br />
and market turmoil<br />
➜ Deleveraging of financial institutions<br />
through balance sheet clean up<br />
➜ Markets concerns were calmed<br />
down thanks <strong>to</strong> outstanding<br />
bail-out programs<br />
➜ Transfer from private debt<br />
<strong>to</strong> public debt<br />
➜ Tightening spreads on corporate<br />
market vs higher spreads on<br />
government bonds debt<br />
Markets confidence seemed<br />
<strong>to</strong> be res<strong>to</strong>red<br />
➜ the 2009 rally<br />
➜ The recovery seemed <strong>to</strong> be<br />
on tracks, but…<br />
➜ As long as the markets trusted<br />
the government, spreads were<br />
under control<br />
➜ Markets lost confidence in the<br />
sustainability of current public<br />
deficits in Europe<br />
➜ Local vs Paneuropean issuer<br />
➜ Lack of coordination<br />
➜ Time for a global answer<br />
Markets lost confidence<br />
in sustainability of public debt<br />
➜ the second dip?<br />
Debt burden flew from private <strong>to</strong> public but the deleveraging is not over<br />
The financial markets unders<strong>to</strong>od what was happening<br />
very quickly because, at the time of the bailout<br />
programs, corporate spreads on the markets were<br />
tightening, meaning that the markets once again trusted<br />
the corporate sec<strong>to</strong>r <strong>to</strong> cope with their debt burden.<br />
At the same time, government spreads were widening<br />
because everyone was beginning <strong>to</strong> understand that<br />
they had <strong>to</strong> pay for all this debt.<br />
Eventually the financial markets lost confidence in the<br />
public sec<strong>to</strong>r, wondering how such a level of debt could<br />
be sustained.<br />
SCOR - November 2010 - 73
Fig. 75: The global economy is facing a critical transition phase<br />
2007-2008: phase I<br />
Financial & economic<br />
collapse<br />
➜ Global recession<br />
➜ Liquidity crisis<br />
➜ Low interest rates combined<br />
with exploding credit spreads<br />
➜ Financial market disruptions<br />
➜ Fiscal and social deficits<br />
➜ Emergency interventions<br />
of Governments<br />
II. SCOR’s view on the financial<br />
and economic environment<br />
Figure 75 shows SCOR’s view on the economic and<br />
financial environment, split in<strong>to</strong> 3 major phases.<br />
As already mentioned, the liquidity and financial crisis<br />
in 2007-2008 led <strong>to</strong> the global economic depression<br />
and <strong>to</strong> distressed financial markets: governments then<br />
decided <strong>to</strong> create his<strong>to</strong>rically huge amounts of money,<br />
which sooner or later will have <strong>to</strong> be dealt with.<br />
The years 2009-2011 may be referred <strong>to</strong> as the transition<br />
phase. This phase encompasses upcoming <strong>risk</strong>s<br />
and throws up a number of questions: will interest rates<br />
decrease, remain stable or increase? Will the sovereign<br />
crisis extend <strong>to</strong> other countries after Greece, Spain and<br />
Portugal? When will the high volatility of exchange<br />
rates s<strong>to</strong>p? Will social unrest develop if governments<br />
change in countries such as Greece? Will new regulations<br />
support the financial markets or not?<br />
These are major issues that need <strong>to</strong> be addressed. For<br />
its part, SCOR anticipates a return of inflation in the<br />
coming years, an increase in interest rates and a potential,<br />
limited increase in GDP, probably with currency<br />
adjustments. Investments will be chosen accordingly.<br />
Given that the economic and financial environment<br />
is so unstable, what choices are there for inves<strong>to</strong>rs?<br />
It is worth noting that the financial markets rely on<br />
expectations: the issue is not whether or not there will<br />
be a double-digit inflation rate, but how long it will<br />
last. This is why the equity market is so bumpy in these<br />
uncertain times.<br />
74 - November 2010 - SCOR<br />
2009-2011: phase II<br />
Key uncertainties<br />
in a s<strong>to</strong>chastic world<br />
➜ Shape of recovery in front of<br />
global recession (L, V, W or √)<br />
➜ Uncertainty on monetary policies<br />
➜ Uncertainty on interest rates<br />
developments<br />
➜ Instability of exchange rates<br />
➜ Sovereign debt crisis<br />
➜ Tax and regula<strong>to</strong>ry debates<br />
~ 2012 onwards: phase III<br />
Scenario<br />
of normalization<br />
➜ Stabilization of economic activity?<br />
With increased inflation?<br />
➜ Normalization of monetary<br />
policies?<br />
➜ Increase of nominal / real interest<br />
rates? Decrease of credit spreads?<br />
➜ Stabilization of equity markets?<br />
➜ Control of public deficits?<br />
➜ Stabilization of regula<strong>to</strong>ry<br />
reforms?<br />
III. How are <strong>Enterprise</strong> <strong>Risk</strong><br />
<strong>Management</strong> processes<br />
implemented in<strong>to</strong> SCOR’s<br />
asset management decisions?<br />
To come back <strong>to</strong> <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong>, how do<br />
you deal with uncertainty? SCOR has tried <strong>to</strong> identify<br />
its main <strong>risk</strong>s, and <strong>to</strong> decide whether the largest <strong>risk</strong><br />
lies in equity, the return of inflation, the steepening<br />
of the yield curve or the effects of volatility on SCOR’s<br />
profitability. Once this has been done, a trade-off has<br />
<strong>to</strong> be made between protecting the value of assets and<br />
capturing the highest profitability: a certain amount<br />
of expected return must be relinquished in order <strong>to</strong><br />
protect the balance sheet against the most significant<br />
<strong>risk</strong>s. In times of high uncertainty, when it is difficult <strong>to</strong><br />
anticipate upcoming developments, the safest course is<br />
<strong>to</strong> prepare for the worst-case scenario so that you can<br />
very quickly adapt <strong>to</strong> new disruptions.<br />
The following four case studies highlight SCOR’s<br />
handling of uncertainty within its investment policy.
CASE STUDY N°1: POSITION THE PORTFOLIO IN CASE OF INFLATION<br />
Fig. 76: Case study no.1: position the portfolio in case of inflation<br />
General concern: inflation is more damaging<br />
for reinsurers’ balance sheet than deflation<br />
➜ Accept the trade off between hedging<br />
and profitability<br />
➜ Inflation in developed countries may be postponed<br />
<strong>to</strong> 2012<br />
➜ Inflationary stimuli are in place<br />
• Unprecedented liquidity injection in the financial system will inevitably lead <strong>to</strong> surge in inflation<br />
and interest rates<br />
• Full sterilization of excess monetary creation likely <strong>to</strong> be impossible<br />
➜ What is currently missing in the market for further inflationary developments:<br />
• Shock in demand for credit more than supply<br />
• Deleveraging is not over yet<br />
• Growth expectations are weaker than a couple of months ago and commodities are directly impacted<br />
Inflation is more damaging for an insurer’s balance<br />
sheet than deflation. Reinsurers lose a lot of money in<br />
an inflation scenario, so they need <strong>to</strong> be prepared for<br />
inflation <strong>to</strong> rise. Even in the case of deflation, reinsurers<br />
should be positioned <strong>to</strong> tackle the return of inflation<br />
if it is identified as a major <strong>risk</strong>. Once again, there is a<br />
trade-off involved. This was the case in 2008, when the<br />
financial markets were geared more <strong>to</strong>wards deflation<br />
than inflation. At that time SCOR decided <strong>to</strong> build an<br />
inflation-linked bond pocket in order <strong>to</strong> be prepared<br />
for a possible return of inflation. This bond pocket was<br />
increased up <strong>to</strong> EUR 1 billion in the first half of 2009,<br />
when inflation expectations were close <strong>to</strong> zero. SCOR<br />
bought securities with negative inflation breakevens,<br />
which means deflation expectations. The Group was<br />
absolutely out of step with the rest of the market, but<br />
it felt that it had <strong>to</strong> be protected against this <strong>risk</strong>, which<br />
is a major one for reinsurance balance sheets. Of course<br />
it is far less attractive <strong>to</strong> invest in linkers than in nominal<br />
bonds, but everything has a cost.<br />
SCOR also decided that, in order <strong>to</strong> be prepared for<br />
the return of inflation, it had <strong>to</strong> stay very liquid in its<br />
portfolio. However, being liquid does just not mean<br />
investing purely in cash. SCOR decided <strong>to</strong> implement<br />
SCOR’s answers<br />
Direct answer<br />
➜ The inflation linked bonds pocket has been<br />
increased up <strong>to</strong> EUR 1 bn in the first half of 2009<br />
when inflation expectations were close <strong>to</strong> zero<br />
➜ The carry is less attractive than that<br />
of nominal bonds<br />
➜ The portfolio is liquid <strong>to</strong> take advantage<br />
of any rise in interest rates<br />
Alternative answers<br />
➜ Diversify the portfolio <strong>to</strong> other asset classes that will<br />
benefit from inflation come-back (emerging markets,<br />
high yield, equities, real estate, commodities…)<br />
what is called the “rollover strategy”, which enables <strong>to</strong><br />
receive cash from redemptions on a regular basis. Each<br />
quarter, the Group gets money back from redemptions,<br />
which can then be reinvested <strong>to</strong> prepare for a rise in<br />
interest rates.<br />
It now became clear that inflation expectations could<br />
be postponed for some time, whereas SCOR had<br />
initially expected sooner a return of inflation. This was<br />
the chance <strong>to</strong> diversify the portfolio in<strong>to</strong> asset classes<br />
that could benefit from inflation, such as commodities,<br />
real estate, equity, emerging markets, high yield markets<br />
and so on.<br />
As indicated at the bot<strong>to</strong>m of Figure 76, the inflationary<br />
stimuli are all in place. What is currently lacking<br />
in the market for this inflation <strong>to</strong> become a reality<br />
is that there is no demand for credits in the United<br />
States or Europe. This is why there is still no inflation,<br />
despite the high level of liquidity that has been pumped<br />
in<strong>to</strong> the market over the past few years. So it is not<br />
a question of supply, but of demand. Nobody knows<br />
when it will happen, but as soon as demand returns,<br />
whether from industry, private companies or households,<br />
inflation could return <strong>to</strong> the markets.<br />
SCOR - November 2010 - 75
CASE STUDY N°2: STAY LIQUID<br />
Fig. 77: Case study no.2: stay liquid in our portfolio<br />
General concern:<br />
in case of markets collapse, cash is king<br />
➜ Accept the trade off between hedging<br />
and profitability<br />
➜ The panic has been pushed away thanks<br />
<strong>to</strong> outstanding bail-out programs<br />
➜ Current market are under pressure<br />
In the event of market collapse, cash held by companies<br />
is important. This was SCOR’s experience in 2007,<br />
2008 and the beginning of 2009, and it accepted<br />
the trade-off between hedging and profitability. As<br />
already mentioned, between early 2007 and the end<br />
of Q1 2009, SCOR accumulated up <strong>to</strong> EUR 4.6 billion in<br />
cash, which was an outstanding amount in relation <strong>to</strong><br />
our balance sheet, representing more than 40% of the<br />
Group’s invested assets. So 40% of the invested assets<br />
were invested in an asset class with a close <strong>to</strong> zero<br />
return, which was highly costly for our investments,<br />
but this was the price of protecting the Group’s<br />
assets. There is, therefore, always a trade-off between<br />
the desire for profitability, which entails a <strong>risk</strong>, and the<br />
desire <strong>to</strong> stay as secure as possible, which will lead <strong>to</strong><br />
less attractive remuneration.<br />
The Group has successfully protected the value of its<br />
balance sheets and its assets and has come through<br />
the financial crisis having suffered as little damage as<br />
possible. If you look at the net asset value of SCOR before<br />
and after the crisis, you will see that it has remained<br />
stable despite dividend distribution. This policy of “stay<br />
in cash” worked very well during the crisis.<br />
The panic has now subsided thanks <strong>to</strong> outstanding<br />
bailout programs. Since the counterparty <strong>risk</strong> involved<br />
was now easier <strong>to</strong> anticipate, SCOR began <strong>to</strong> look for<br />
76 - November 2010 - SCOR<br />
SCOR’s answers<br />
Direct answer<br />
➜ SCOR accumulated cash early 2007<br />
up <strong>to</strong> EUR 4.6 bn in Q1 2009<br />
➜ The return of cash was close <strong>to</strong> zero<br />
and 40% of invested assets were in cash<br />
➜ SCOR successfully protected the value of<br />
its assets through the financial crisis (Cf NAV)<br />
Alternative answers<br />
➜ Since Q2 2009, strong focus on liquidity while<br />
seizing short term market opportunities:<br />
➜ Current liquidity is EUR 1.5 bn at end of Q1 2010<br />
➜ But roll-over strategy in place with EUR 4.4 bn<br />
of cash flows from the invested portfolio<br />
over the next 24 months<br />
➜ Short duration<br />
➜ SCOR progressively came back on selective credit,<br />
equities and other asset classes<br />
➜ Equities because of double dip <strong>risk</strong>s<br />
➜ Currencies due <strong>to</strong> euro weakness<br />
➜ Rates due <strong>to</strong> sovereign debt concerns<br />
➜ Emerging markets because of <strong>risk</strong>s of contagion…<br />
alternative ways <strong>to</strong> be able <strong>to</strong> face liquidity needs. SCOR<br />
lowered the amount of liquidity from EUR 4.6 billion<br />
at the end of Q1 2009 <strong>to</strong> EUR 1.5 billion at the end of<br />
Q1 2010, which means that over a period of one year,<br />
it placed EUR 3 billion of invested assets in more attractive<br />
asset classes. It did this very selectively, ensuring<br />
that the duration of our portfolio remained very short.<br />
It selected its high quality corporate bonds very carefully<br />
and had strong concerns about sovereign <strong>risk</strong>s.<br />
The Group implemented this rollover strategy with a<br />
strong focus on the amount <strong>to</strong> be redeemed or received<br />
as coupons over the next 24 months. Instead of holding<br />
cash as it was done during the turmoil, SCOR is now<br />
in a position where it receives around EUR 4.5 billion<br />
cash that can be reinvested in the financial markets at<br />
the next available opportunity.<br />
The current situation is also a concern, with markets<br />
under pressure. Interest rates are stressed, equity values<br />
are stressed, foreign exchange rates are stressed, and<br />
something could explode in the coming months. The<br />
question is, what will burn out? Are companies going<br />
<strong>to</strong> face a contagious situation where one collapsed asset<br />
class infects the others, or will this be contained within a<br />
country, an area, or a specific asset class? Once again no<br />
one has the answer, so SCOR has decided <strong>to</strong> remain liquid<br />
and <strong>to</strong> continue <strong>to</strong> arbitrage in favour of liquid assets.
CASE STUDY N°3: REACT ACTIVELY TO THE SOVEREIGN CRISIS<br />
Fig. 78: Case study no.3: react actively <strong>to</strong> the sovereign crisis<br />
An early concern on sovereign <strong>risk</strong>s<br />
➜ As extensively discussed during the July 2009 Inves<strong>to</strong>rs’ Day,<br />
strong focus on sovereign <strong>risk</strong>s<br />
SCOR regularly scorecards all its investments <strong>to</strong> preserve<br />
its financial strength<br />
4.00<br />
3.50<br />
3.00<br />
2.50<br />
2.00<br />
1.50<br />
1.00<br />
Inves<strong>to</strong>rs’ Day – July 2009<br />
Watch List<br />
Sovereign <strong>risk</strong> scorecard: example of the November 2008 study<br />
South Korea<br />
Reduce<br />
Hungary<br />
Mexico<br />
0.50<br />
0.00<br />
Austria<br />
Sweden<br />
Switz<br />
-0.50<br />
Fin<br />
Hold Ger<br />
Watch List<br />
Norv<br />
-1.00<br />
Neth<br />
Greece<br />
Ireland<br />
Poland<br />
Italy Cz<br />
NZ<br />
SLV<br />
DK Port<br />
UK<br />
USA<br />
Can Esp Belg<br />
Jap Fr Lux<br />
Australia<br />
Singapore<br />
-1.50 -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50<br />
Macro-economic scoring<br />
Brazil<br />
Russia<br />
This case study is about reacting actively <strong>to</strong> the sovereign<br />
crisis. In Spring 2009, SCOR was already concerned<br />
about European public debt. In fact, SCOR started <strong>to</strong><br />
become concerned at the end of 2008, at the time of<br />
the bailout program decisions. As you probably know,<br />
Denis Kessler is an economist, so his first reaction was<br />
<strong>to</strong> say, “It’s not possible <strong>to</strong> sustain such public debt.<br />
Keep an eye on sovereign names and on sovereign<br />
individual exposure <strong>to</strong> financial markets.” In view of<br />
this, the Group decided <strong>to</strong> create a kind of scorecard<br />
of all the big countries in which it was supposed <strong>to</strong><br />
invest, and used this as a focal point at our quarterly<br />
Group Investment Committee meetings. SCOR had<br />
already decided at the end of 2008 <strong>to</strong> reduce its<br />
exposure <strong>to</strong> Spain and Italy, at a time when it no longer<br />
had any exposure <strong>to</strong> Greece. What you have <strong>to</strong> do in<br />
such difficult times is <strong>to</strong> be highly responsive. If you no<br />
longer have confidence in a country, you leave it. If the<br />
situation subsequently eases and you think that new<br />
elements should be taken in<strong>to</strong> account, then you have<br />
<strong>to</strong> reassess the situation and you can reinvest from that<br />
point on. This is not a problem. With clients, it’s a very<br />
different s<strong>to</strong>ry. When you want <strong>to</strong> run off an activity,<br />
you have <strong>to</strong> wait until the last client‘s policy has ended.<br />
When you deal with the financial markets, you can buy<br />
securities one day and sell them the next.<br />
South Africa<br />
Gvt CDS scoring COR regularly scorecards all it<br />
Source: SCOR calculation<br />
India<br />
China<br />
Iceland<br />
Turkey<br />
Financial<br />
Change/<br />
Default <strong>risk</strong><br />
Forbidden<br />
(High default<br />
<strong>risk</strong>)<br />
Exit<br />
(Significant<br />
default <strong>risk</strong>)<br />
No major<br />
change<br />
Hungary<br />
Greece<br />
Poland<br />
Czeck Rep.<br />
Moni<strong>to</strong>r Italy<br />
(Intermediate Korea<br />
default <strong>risk</strong>) Portugal<br />
Hold (Hold<br />
default <strong>risk</strong>)<br />
Canada<br />
France<br />
Germany<br />
Japan<br />
Switzerland<br />
United States<br />
Rating<br />
downgrade<br />
<strong>risk</strong><br />
Brazil<br />
Iceland<br />
Russia<br />
Belgium<br />
Denmark<br />
Ireland<br />
New Zealand<br />
Spain<br />
CDS<br />
spread <strong>risk</strong><br />
India<br />
Turkey<br />
Mexico<br />
Slovakia<br />
Combination<br />
on the two<br />
<strong>risk</strong>s<br />
China<br />
South Africa<br />
Australia<br />
United<br />
Kingdom<br />
Singapore<br />
SCOR’s answers<br />
Significant decrease in global exposure<br />
<strong>to</strong> euro peripheric countries…<br />
87<br />
288<br />
169<br />
248<br />
0 8 11<br />
5 3<br />
17<br />
Portugal Ireland Italy Greece Spain<br />
Total 31/12/2009 Total 19/05/2010<br />
… as well as exposure on government bonds<br />
& assimilated of euro peripheric countries<br />
105<br />
55<br />
0 0 3 7<br />
29<br />
0 0 0<br />
Portugal Ireland Italy Greece Spain<br />
Govies & assimilated 31/12/2009<br />
Govies & assimilated 19/05/2010<br />
SCOR was already addressing its concerns about the<br />
sovereign crisis in early 2009 and also did so at our<br />
Inves<strong>to</strong>rs’ Day last year. As you can see from the <strong>to</strong>p<br />
right-hand side of Figure 78, it still had some exposure at<br />
the end of 2009 <strong>to</strong> Italy, Spain and Portugal. By mid May,<br />
as shown in the bot<strong>to</strong>m right-hand side above graph,<br />
SCOR had sold all the Portuguese, Irish and Greek debt<br />
and had considerably reduced exposure <strong>to</strong> Italy and Spain<br />
because we no longer had confidence in these <strong>risk</strong>s.<br />
SCOR’s current concern is France. Should it continue <strong>to</strong><br />
hold a larger amount of French vs. German bonds? Is the<br />
current weight of French government bonds sustainable?<br />
What is going <strong>to</strong> happen? This does not mean that the<br />
Group does not trust the French government. It simply<br />
means that there is a <strong>risk</strong> and that, in terms of <strong>ERM</strong>,<br />
it has <strong>to</strong> assess that <strong>risk</strong>, make a decision and put it<br />
in<strong>to</strong> practice. So if it is decided that the bonds are<br />
safe, they will be kept. If it is decided that they are not<br />
sure whether or not they will be safe over the coming<br />
years, SCOR has <strong>to</strong> act accordingly. The trade-off is as<br />
follows: you sell French government bonds, you invest<br />
in German government bonds and you lose 50 basis<br />
points. Each time you arbitrage from <strong>risk</strong>y <strong>to</strong> non-<strong>risk</strong>y,<br />
you lose some kind of remuneration in one way or<br />
another, but you have <strong>to</strong> try <strong>to</strong> protect your assets.<br />
SCOR - November 2010 - 77
CASE STUDY N°4: DIVERSIFY THE PORTFOLIO<br />
Fig. 79: Case study no.4: diversify the portfolio<br />
General concern: high stress on financial markets<br />
makes it difficult <strong>to</strong> predict any evolution<br />
➜ Get ready for the worst<br />
➜ Implement strict investment controls<br />
process / stress test<br />
➜ Stay reactive<br />
The last case study addresses portfolio diversification. In<br />
order <strong>to</strong> be prepared for the worst-case scenario, SCOR<br />
has implemented a strict investment control process<br />
as well as stress tests. The way <strong>to</strong> cope with a stressed<br />
market is <strong>to</strong> enlarge the scope of your investments <strong>to</strong><br />
ensure that you capture the maximum profitability, <strong>to</strong><br />
be highly selective on each asset class, and <strong>to</strong> moni<strong>to</strong>r<br />
every position closely and be ready <strong>to</strong> change your<br />
mind. Sometimes you are right, sometimes you are<br />
wrong. Maybe you should have kept this asset class<br />
because it eventually turned out <strong>to</strong> be safer and more<br />
attractive than you expected, but you have protected<br />
your assets and that was your main concern at the<br />
time of the decision. Once again, there is always a<br />
trade-off between profitability and sheltering your<br />
assets in safer harbours.<br />
IV. Positioning the investment<br />
portfolio <strong>to</strong> absorb large shocks<br />
while seizing short-term<br />
opportunities<br />
By way of illustration and <strong>to</strong> conclude this article,<br />
Figure 80 represents the profile of SCOR’s portfolio at<br />
Q1 2009 and at Q1 2010. The <strong>to</strong>tal invested assets<br />
are roughly the same. SCOR has placed a large part of<br />
its cash and investments in fixed income. At Q1 2009,<br />
40% of its portfolio was invested in cash and shortterm<br />
investments, which means investments of less than<br />
three months and government bonds with a maturity<br />
of up <strong>to</strong> one year. One year later, only 12% is in cash<br />
and short-term investments, and the fixed income has<br />
almost doubled. This reflects the inflation program that<br />
the Group started in the second quarter of 2009, when<br />
it thought that this crisis situation was over.<br />
78 - November 2010 - SCOR<br />
SCOR’s answers<br />
➜ Enlarge the scope of investment <strong>to</strong> be sure<br />
<strong>to</strong> capture the maximum profitability<br />
➜ Be highly selective on each asset class<br />
➜ Moni<strong>to</strong>r closely every position and be ready<br />
<strong>to</strong> change your mind<br />
➜ This remains the best way <strong>to</strong> capture profitability<br />
➜ Bet on various countries / asset classes / currencies enable <strong>to</strong> take opportunities<br />
and <strong>to</strong> be able <strong>to</strong> exit quickly one of the strategy in case of bad developments<br />
The Group lengthened the duration of the T-bills that<br />
were invested in cash in<strong>to</strong> medium-term government<br />
bonds classified under fixed income, in order <strong>to</strong><br />
benefit from the steepening of the yield curve with<br />
an additional 100 basis points. It was decided <strong>to</strong> move<br />
EUR 600 million of cash investments in<strong>to</strong> corporate<br />
bonds in order <strong>to</strong> benefit from still very attractive<br />
spreads over government bonds. It was at the time of<br />
this long rally on the corporate market last year, when<br />
the market began once again <strong>to</strong> envisage economic<br />
growth, that the Group entered in<strong>to</strong> this double dip<br />
<strong>risk</strong> scenario. Finally, EUR 150 million were invested<br />
in equities <strong>to</strong> take advantage of good opportunities<br />
on the markets. The rebalancing of these portfolios<br />
is still in process. The Group wants <strong>to</strong> adjust the level<br />
of cash <strong>to</strong> between EUR 1 and EUR 1.5 billion, and<br />
<strong>to</strong> stay highly selective on the fixed income and the<br />
equity markets. As already mentioned several times,<br />
this situation could change since SCOR is constantly<br />
adapting <strong>to</strong> new developments.<br />
What is not visible in these charts is the rebalancing<br />
SCOR has conducted on alternative investments – in<br />
Q1 2009 it had already requested the redemption of all<br />
the hedged funds it had in the portfolio, but this process<br />
takes time. These redemptions were initially requested<br />
at the end of 2008, but one year later it still had<br />
a lot of them in the portfolio because the redemption<br />
process was still underway. Incidentally, this gives a<br />
good indication of what an illiquid market can look<br />
like. At the end of 2009, the Group had EUR 100 million<br />
less in alternative investments, which were placed in<br />
Q1 2010 in assets classified as alternative investments,<br />
but which have nothing <strong>to</strong> do with hedge funds. The<br />
company also classifies commodities as alternative<br />
investments. It has notably invested a little in gold, raw<br />
materials, agricultural products and so on. These are<br />
the kinds of investment you select in times of inflation<br />
return <strong>risk</strong>, because this class will benefit from inflation.
Fig. 80: Positioning investment portfolio in order <strong>to</strong> absorb large shocks<br />
while seizing short-term opportunities<br />
Q1 2009<br />
Total invested assets (1) : EUR 11,360 million<br />
Cash & Short-term<br />
investments<br />
40%<br />
Impairments<br />
-5.4% +3.7%<br />
Real Estate<br />
3%<br />
Fixed Income<br />
48%<br />
Government & Assimilated<br />
25%<br />
Return on<br />
invested assets<br />
before impairments<br />
Equities<br />
6% Alternative<br />
Investments (2)<br />
3%<br />
Q1 2010<br />
Total invested assets (1) : EUR 12,656 million<br />
Cash & Short-term<br />
investments<br />
12%<br />
Real Estate<br />
3%<br />
Equities<br />
6% Alternative<br />
Investments (2)<br />
3%<br />
Structured products<br />
5% Structured products<br />
8%<br />
Corporate bonds<br />
12%<br />
Covered & Agency<br />
6%<br />
Return on<br />
invested assets<br />
after impairments<br />
Government &<br />
Assimilated<br />
38%<br />
Impairments<br />
Fixed Income<br />
76%<br />
Return on<br />
invested assets<br />
before impairments<br />
Covered & Agency<br />
7%<br />
Corporate bonds<br />
23%<br />
Return on<br />
invested assets<br />
after impairments<br />
-1.6% -0.5% +4.4% +3.9%<br />
(1) Excluding funds withheld.<br />
(2) Including hedge funds, infrastructure funds, private equity, commodities and non-listed equities.<br />
In other words, in times of uncertainty you have <strong>to</strong> be<br />
very careful in what you do, but you do not actually<br />
have <strong>to</strong> ban anything. Of course alternative investments<br />
are considered <strong>risk</strong>y, but some selective assets within<br />
this classification could help <strong>to</strong> prevent the <strong>risk</strong>s you<br />
are facing.<br />
The Group also tries <strong>to</strong> avoid investing in solutions<br />
that it does not fully understand. When it is offered<br />
new investments/solutions with which the Group is not<br />
familiar, it learns more about the asset class in question<br />
<strong>to</strong> ensure that it has ticked all the right boxes and that<br />
it knows exactly why this particular asset class is good<br />
for its balance sheet. SCOR makes sure that it knows<br />
why it should use this asset class over another one, so<br />
that it can face all the <strong>risk</strong>s involved and be sure that it<br />
has made a good choice. Of course, this does not mean<br />
that at the end of the day the Group will have predicted<br />
what is going <strong>to</strong> happen with the investment.<br />
Once again, if there is no inflation in 2012 or 2013,<br />
or even up <strong>to</strong> 2020, it will have been costly <strong>to</strong> build<br />
this inflation-linked bond pocket, but the Group will be<br />
happy with it because the trade-off is infinitely preferable<br />
<strong>to</strong> being exposed <strong>to</strong> inflation with no protection.<br />
This is how SCOR has implemented <strong>ERM</strong> in<strong>to</strong> its<br />
investment strategy.<br />
SCOR - November 2010 - 79
9<br />
HOW TO DEAL WITH THE IMPACT<br />
OF INFLATION ON PRICING<br />
AND RESERVING TONY NEGHAIWI<br />
Chief Pricing Actuary, SCOR Global P&C<br />
ERIC DAL MORO<br />
Chief Reserving Actuary, SCOR Switzerland<br />
This article addresses inflation, its<br />
origins and its impact on long-tail business lines.<br />
Based on SCOR’s experience, it shows how inflation<br />
could be integrated in<strong>to</strong> pricing and reserving.<br />
I. Liability reserves<br />
and inflation<br />
For Liability business lines, inflation sources can broadly<br />
be classified in<strong>to</strong> the following groups:<br />
Fig. 81: Sources of inflation for liability reserves<br />
Medical Price Index vs. All-item Price Index<br />
150<br />
140<br />
130<br />
120<br />
110<br />
100<br />
90<br />
80<br />
1996*<br />
80 - November 2010 - SCOR<br />
• The Consumer Price Index (CPI), defined by the United<br />
States Bureau of Labor Statistics as “a measure of the<br />
average change over time in the prices paid by urban<br />
consumers for a market basket of consumer goods<br />
and services.”<br />
• Changes in regulation or legal interpretations in any<br />
jurisdiction. An illustration of this is the regula<strong>to</strong>ry<br />
change in 1996 that completely modified reserve<br />
levels in French Non-Proportional Mo<strong>to</strong>r treaties.<br />
• Social changes: in fast-growing countries like China,<br />
the rehabilitation of victims is becoming increasingly<br />
expensive due <strong>to</strong> rapidly improving living standards.<br />
• Medical cost inflation: Figure 81 compares European<br />
medical cost inflation with the CPI in Europe. The<br />
medical cost rate is double that of the CPI.<br />
1997* 1998* 1999 2000 2001 2002 2003 2004 2005<br />
Medical price index All-item price index *Estimates.<br />
Source: Comité Européen des Assurances.
• Emerging <strong>risk</strong> inflation can be a source of inflation<br />
but is much more difficult <strong>to</strong> analyze (e.g. the possible<br />
impact of nanotechnology on future Liability claims).<br />
The precautionary principle can also be considered<br />
a source of inflation. This principle is being applied<br />
more and more and may therefore drive an increasing<br />
number of claims. This is illustrated by the recent<br />
eruption of the volcano in Iceland, for which some<br />
passengers would certainly have liked <strong>to</strong> file Liability<br />
claims.<br />
All of the above inflation sources may become drivers<br />
of superimposed inflation. Superimposed inflation is<br />
not an easy concept <strong>to</strong> grasp. In 2007, the Institute of<br />
Actuaries of Australia asked its members <strong>to</strong> come up<br />
with their best possible definitions for it. Two of these<br />
are set out below:<br />
• “The increase in claims costs that exceeds inflation, is<br />
unrelated <strong>to</strong> increases in claim frequency and ignores<br />
higher claims costs emanating from legislation changes<br />
and the like”.<br />
• “A measure of the increase in claim values in<br />
excess of the rate of wage inflation. Alternatively,<br />
a term dreamt up by actuaries <strong>to</strong> explain away their<br />
miscalculations!”<br />
Nowadays, insurers are facing these unpredictable<br />
fac<strong>to</strong>rs and the question they are asking is: when<br />
will superimposed inflation start and how can it be<br />
s<strong>to</strong>pped?<br />
The same survey by the Institute of Actuaries of Australia<br />
posed the question of how <strong>to</strong> s<strong>to</strong>p superimposed<br />
inflation. Proposed solutions included:<br />
• Changing benefits in insurance policies.<br />
• Excluding/limiting the involvement of lawyers in<br />
the Liability system. Although this answer appears<br />
a good solution at first, it should be mentioned<br />
that most Liability claims are handled by insurers.<br />
Liability claims are only occasionally presented <strong>to</strong> a<br />
court, especially when the indemnities involved are<br />
small. In the U.S., however, the <strong>to</strong>rt system, and class<br />
action suits in particular, allow courts <strong>to</strong> award large<br />
indemnity payouts. This can drive superimposed<br />
inflation and could be impacted by the limitation of<br />
lawyer involvement.<br />
• Changing the way in which claims are managed.<br />
In order <strong>to</strong> detect which claims are likely <strong>to</strong> lead <strong>to</strong><br />
difficulties or fraud, it is helpful <strong>to</strong> implement a<br />
predictive model <strong>based</strong> on the claims data that is<br />
usually available when the claim is first entered in<strong>to</strong><br />
the IT system. On the basis of this model, the more<br />
complex claims can be handled by senior adjusters<br />
and the more simple cases by junior adjusters. This<br />
allocation of claims is a very important fac<strong>to</strong>r for<br />
insurers because for Liability, especially, moral hazard<br />
is a major driver. An optimal claims management<br />
system is therefore important <strong>to</strong> prevent claims<br />
inflation arising from undetected fraudulent claims.<br />
SCOR - November 2010 - 81
Fig. 82: Sources of inflation for liability reserves<br />
Example of the impact of inflation on Non-Proportional Liability<br />
Cover<br />
10 EURm<br />
Priority<br />
1 EURm<br />
Reinsurers are partially covered<br />
against inflation for the<br />
Non-Proportional business<br />
through indexation clauses.<br />
Figure 82 illustrates the potential impact of inflation<br />
on a Non-Proportional contract. For a treaty of<br />
EUR 10 m xs 1 m, one imagines that a claim reaches<br />
EUR 1.1 m. The reinsurance cost is EUR 0.1 m.<br />
With 10% inflation, the claim will reach EUR 1.2 m<br />
and the cost <strong>to</strong> the reinsurer will be 100% higher.<br />
So, even though stabilization clauses in Non-Proportional<br />
contracts exist in many markets, inflation can have a<br />
dramatic effect for reinsurers.<br />
NON-PROPORTIONAL MOTOR BUSINESS<br />
IN FRANCE<br />
Treaty 10 EURm xs 1 EURm<br />
Claim<br />
1.1 EURm<br />
Cost<br />
reinsurance<br />
0.1 EURm<br />
10% inflation<br />
French Mo<strong>to</strong>r Non-Proportional treaties provide an<br />
important example for examining legal changes and<br />
Claim<br />
1.2 EURm<br />
Cost<br />
reinsurance<br />
0.2 EURm<br />
Cost<br />
reinsurance<br />
+100%<br />
their impact on loss development triangles. In 1996,<br />
a legal change was introduced <strong>to</strong> the calculation of<br />
pension indemnities payable <strong>to</strong> bodily injury victims.<br />
This led <strong>to</strong>:<br />
• The introduction of more conservative mortality tables;<br />
• The application of more conservative discount rates.<br />
Insurers had five years <strong>to</strong> implement the new regulations,<br />
with a retroactive effect on open claims.<br />
Below is a triangle of chain-ladder coefficients for<br />
cumulative incurred losses on French Mo<strong>to</strong>r Non-<br />
Proportional lines.<br />
Fig. 83: Mo<strong>to</strong>r Non-Proportional in France<br />
Triangle of incurred chain-ladder coefficients for Mo<strong>to</strong>r Non-Proportional in France<br />
UWY N N+1 N+2 N+3 N+4 N+5 N+6 N+7 N+8 N+9 N+10 N+11 N+12 N+13 N+14 N+15 N+16 N+17 N+18<br />
1980 0.13 14.53 1.74 1.05 1.02 1.11 1.01 0.97 1.03 1.04 0.99 1.00 1.18 1.01 1.05 1.28 0.80 1.03<br />
1981 0.00 4.97 0.99 1.04 1.21 1.16 0.86 1.04 1.02 0.92 1.03 0.94 0.97 0.98 1.26 0.93 1.10 0.97<br />
1982 0.78 0.91 1.07 1.18 1.11 1.08 1.02 0.97 1.08 1.06 1.04 1.00 1.05 1.40 0.71 1.02 1.06 1.02<br />
1983 1.07 1.14 1.03 1.01 0.95 0.94 0.99 1.02 1.04 1.02 1.05 1.06 1.43 0.70 1.05 1.11 0.86<br />
1984 1.39 1.14 1.13 1.16 1.03 1.00 0.98 1.02 0.96 0.98 1.05 1.30 0.92 1.06 1.01 1.01<br />
1985 1.54 1.29 1.09 0.94 1.03 1.06 0.86 1.00 1.00 1.01 1.28 0.95 1.03 0.95 1.04<br />
1986 1.37 0.91 1.03 1.05 0.97 1.02 1.02 1.01 1.09 1.29 0.88 1.08 1.02 0.89<br />
1987 1.37 1.01 1.02 1.02 1.08 0.94 1.10 1.00 1.29 1.09 1.10 0.97 0.97<br />
1988 1.55 0.96 0.99 0.94 1.11 0.99 1.01 1.32 0.97 1.07 1.04 0.97<br />
1989 1.16 0.92 1.14 0.97 0.99 1.06 1.39 1.01 1.22 0.99 0.98<br />
1990 NA 1.34 1.11 1.01 1.11 1.04 1.43 0.91 1.16 0.99 0.97<br />
1991 1.41 0.87 1.19 1.01 1.34 1.02 1.08 0.98 1.00<br />
1992 1.29 1.13 1.10 1.48 0.94 1.10 0.94 1.11<br />
1993 1.21 1.03 1.39 0.91 1.05 0.98 0.99<br />
1994 1.15 1.59 1.17 1.13 0.98 0.96<br />
1995 4.39 1.54 1.16 0.96 1.04<br />
1996 1.40 1.28 1.20 0.96<br />
1997 1.35 1.12 1.01<br />
1998 0.96 1.13<br />
1999 1.10<br />
The impact of the change of regulation can be clearly seen on the chain-ladder coefficients.<br />
2000<br />
How <strong>to</strong> deal with such dis<strong>to</strong>rtions?<br />
82 - November 2010 - SCOR
In the triangle (Figure 83), the impact of the regula<strong>to</strong>ry<br />
change is visible in the diagonals of 1996, 1997 and<br />
1998. For 1996, instead of having lower chain-ladder<br />
coefficients (between 1.2 and 1.3), the coefficient<br />
reaches 4.3. This high coefficient is due <strong>to</strong> cedants<br />
adapting their bodily injury reserves <strong>to</strong> the new<br />
regulation and thus reporting higher incurred values.<br />
This is typical of a retroactive change in regulation:<br />
a huge effect is visible in the years following initial<br />
implementation.<br />
Following this example, the way how SCOR deals with<br />
such dis<strong>to</strong>rtions will be presented.<br />
TREATMENT OF INFLATION IN RESERVING<br />
SCOR has two main methods for dealing with reserving<br />
inflation:<br />
• Stress testing: this consists of incorporating a loading<br />
for future unpredictable inflation in<strong>to</strong> ultimate loss<br />
ratios, <strong>based</strong> on pricing information for the most<br />
recent underwriting years. Putting up reserves in this<br />
way can be costly and, as a result, requires a strong<br />
balance sheet.<br />
• Correction of triangles for embedded inflation: the<br />
correction of triangles for inflation consists of removing<br />
the inflation effects embedded in the triangles.<br />
This correction may involve removing chain-ladder<br />
coefficient diagonals affected by legal interpretation<br />
changes (e.g. French Mo<strong>to</strong>r Non-Proportional).<br />
It may also involve correcting the whole triangle<br />
for some base inflation (e.g. TME rate in France).<br />
The methods set out below provide two ways in which<br />
<strong>to</strong> au<strong>to</strong>matically estimate this embedded inflation:<br />
- Separation method of Verbeek-Taylor<br />
(G. Taylor 1986),<br />
- Generalized Linear Model (GLM - England &<br />
Verrall 2002).<br />
The modified triangle is then projected and the final ultimate<br />
loss is calculated, fac<strong>to</strong>ring in future inflation. Future<br />
inflation estimates can be <strong>based</strong> on exposure and frequency<br />
determinants (if possible). In any case, the future<br />
claims inflation assumption remains judgmental.<br />
Stress tests<br />
In order <strong>to</strong> fac<strong>to</strong>r inflation in<strong>to</strong> the reserves, the ultimate<br />
loss ratio used for reserving purposes is split in<strong>to</strong><br />
three layers:<br />
• The first layer consists of the expected claims level<br />
with no inflation assumptions;<br />
• The second layer consists of the additional loss ratio<br />
points stemming from the original price inflation<br />
assumption. This information comes from pricing<br />
actuaries;<br />
• The third layer represents the loss ratio points<br />
additional <strong>to</strong> the second layer, stemming from<br />
stressed inflation.<br />
Figure 84 illustrates the three layers concept:<br />
The first layer, claims without inflation, represents an<br />
80% loss ratio.<br />
The second layer fac<strong>to</strong>rs in 5% inflation over a claim<br />
duration of 2 years: this would represent a further 8%<br />
on the loss ratio.<br />
Finally, the third layer fac<strong>to</strong>rs in an additional 2%<br />
inflation, which leads <strong>to</strong> an additional 3%. This is the<br />
cost of inflation unpredictability.<br />
Fig. 84: Treatment of inflation in reserving<br />
Stress-test – Applies <strong>to</strong> recent Underwriting Years, <strong>based</strong> on pricing information<br />
1 st layer<br />
2 nd layer<br />
3 rd layer<br />
Pricing Ultimate<br />
Loss Ratio<br />
without inflation<br />
Δ Ultimate Loss Ratio<br />
when original price with 5% inflation<br />
Δ Ultimate Loss Ratio<br />
for higher inflation (7%)<br />
CoC Expenses 10%<br />
Claims without<br />
inflation<br />
80%<br />
8%<br />
3%<br />
Reserving ULR 101%<br />
Note on the calculation below:<br />
We assume a claim duration of 2 years.<br />
= 80% [(1+5%) 2 –1]<br />
= 80% [(1+7%) 2 – (1+5%) 2 ]<br />
SCOR - November 2010 - 83
Triangle correction for embedded inflation<br />
The correction of triangles for embedded inflation can<br />
be conducted cell by cell on applying the following<br />
fac<strong>to</strong>r: 1<br />
(1+x) i+j<br />
where:<br />
• x is the average estimated rate of inflation in the past,<br />
• i relates <strong>to</strong> underwriting year,<br />
• j relates <strong>to</strong> development years.<br />
Details about this method can be found in England<br />
and Verrall (2002).<br />
Once the triangles have been corrected by the fac<strong>to</strong>rs<br />
above, a standard projection can be made <strong>to</strong> estimate<br />
uninflated IBNRs (e.g. using the chain-ladder method).<br />
The challenge is then <strong>to</strong> estimate the reinflated IBNRs<br />
taking future inflation fac<strong>to</strong>rs in<strong>to</strong> account.<br />
Usually, future inflation is determined by future exposure<br />
and future frequency. However, unlike Property<br />
insurance, the determinants of Liability exposures and<br />
frequency are not straightforward <strong>to</strong> define.<br />
Fig. 85: Treatment of inflation in reserving<br />
Worker’s Compensation loss ratio vs. unemployment<br />
140%<br />
120%<br />
100%<br />
80%<br />
60%<br />
40%<br />
1979<br />
84 - November 2010 - SCOR<br />
Workers comp loss ratio (LHS) Unemployment rate (RHS)<br />
Source: J.P. Morgan, APRA, ISC.<br />
As examples of Liability determinants for inflation, in 2009<br />
the Insurance Services Office in the US proposed the following<br />
measures for some Liability business lines:<br />
• Contrac<strong>to</strong>rs’ Liability exposure: the average hourly<br />
earnings of construction workers would seem <strong>to</strong> be<br />
correlated <strong>to</strong> inflation.<br />
• Manufacturers’ Product Liability exposure: future<br />
sales would be a good determinant even though<br />
they remain uncertain.<br />
• Frequency: unemployment rates and interest rates<br />
may be used. In Australia, for instance, the unemployment<br />
rate and the ultimate loss ratio of Workers’<br />
Compensation insurance seem <strong>to</strong> be correlated, as<br />
shown in Figure 85. Therefore, using the predicted<br />
unemployment rate, it should be possible <strong>to</strong> estimate<br />
the future ultimate loss ratio, which would integrate<br />
some elements of superimposed inflation.<br />
To conclude, in the context of superimposed inflation<br />
reappearing and this future inflation being different<br />
<strong>to</strong> past inflation, insurers already have many <strong>to</strong>ols <strong>to</strong><br />
deal with this new environment, as described above.<br />
However, it should be noted that the impact of inflation<br />
on reserves can be huge; fortunately, modeling<br />
solutions exist <strong>to</strong> deal with it.<br />
1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011<br />
12%<br />
10%<br />
8%<br />
6%<br />
4%<br />
2%<br />
0%
II. Pricing and inflation<br />
Projecting the cost of ultimate loss, and identifying and<br />
dealing with its drivers, is a common objective in both<br />
pricing and reserving. It also constitutes perhaps the<br />
most critical issue not just for these actuarial functions<br />
but also for the health and viability of the insurance<br />
and reinsurance industry as a whole.<br />
Fig. 86: Should Pricing/Reserving be concerned? Absolutely<br />
Reasons for US P/C Insurer Impairments, 1969-2008<br />
7%<br />
8%<br />
9%<br />
8%<br />
4% 4%<br />
8%<br />
14%<br />
38%<br />
As highlighted by Figure 86, between 1969 and 2008,<br />
38% of US P&C defaults were attributed <strong>to</strong> deficient<br />
loss reserves and inadequate pricing, by far the largest<br />
cause of insurer defaults. One may also argue that<br />
there is a strong link between pricing/reserving and<br />
some of the other causes on the chart, such as rapid<br />
growth. Unsurprisingly, frequent defaults usually follow<br />
turbulent times.<br />
Deficient loss reserves/Inadequate Pricing<br />
Rapid growth<br />
Alleged fraud<br />
Catastrophe losses<br />
Affiliate impairment<br />
Investment problems<br />
Miscellaneous<br />
Significant change in business<br />
Reinsurance failure<br />
Source: A.M. Best, 1969-2008 Impairment Review, Special Report, April 6, 2008.<br />
Deficient Loss Reserves and Inadequate Pricing are the leading cause of insurer impairments, underscoring the importance of discipline.<br />
Investment Catastrophe Losses play a much smaller role.<br />
SCOR - November 2010 - 85
Fig. 87: Some his<strong>to</strong>rical perspective on inflation<br />
Inflation is not double-digit now as it was in 1974-81<br />
Annual<br />
Inflation Rates<br />
14%<br />
12%<br />
10%<br />
8%<br />
6%<br />
4%<br />
2%<br />
0%<br />
-2%<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
1969 8<br />
1970<br />
1971<br />
1972<br />
1973<br />
1974<br />
1975<br />
1976<br />
1977<br />
1978<br />
1979<br />
1980<br />
1981<br />
1982<br />
1983<br />
1984<br />
1985<br />
1986<br />
1987<br />
1988<br />
1989<br />
1990<br />
1991<br />
1992<br />
1993<br />
1994<br />
1995<br />
1996<br />
1997<br />
1998<br />
1999<br />
2000<br />
2001<br />
2002<br />
2003<br />
2004<br />
2005<br />
2006<br />
2007<br />
2008<br />
2009<br />
2010F<br />
Source: US Department of Labor, Bureau of Labor Statistics; Blue Chip Economic Indica<strong>to</strong>rs (4/2010 issue) (2010 forecast).<br />
If you started work in the insurance industry 27 years ago, you never experienced annual inflation as high as 6%.<br />
Fig. 88: Some his<strong>to</strong>rical perspective on Non-Life companies’ impairment<br />
P&C Insurer Impairments, 1969-2009F (forecast)<br />
15<br />
12<br />
7<br />
11<br />
9<br />
34<br />
9<br />
13<br />
12<br />
19<br />
9<br />
16<br />
14<br />
13<br />
36<br />
49<br />
31<br />
34<br />
50<br />
48<br />
Source: Insurance Information Institute.<br />
The number of imparments varies significantly over the P&C Insurance Cycle, with peaks occuring well in<strong>to</strong> hard markets.<br />
Companies are currently facing economically turbulent<br />
times, where high inflation is one potential scenario as<br />
the world economy comes out of the crisis. This would<br />
have serious consequences on (re)insurance industry,<br />
especially on long-tail lines such as Liability. Even more<br />
serious is the fact that most of companies in the industry<br />
have not worked in a period of high inflation.<br />
86 - November 2010 - SCOR<br />
55<br />
60<br />
58<br />
41<br />
29<br />
16<br />
12<br />
31<br />
18<br />
19<br />
5 of the 11 are Florida companies<br />
(1 of these 5 is a title insurer)<br />
49<br />
50<br />
47<br />
35<br />
18<br />
14<br />
15<br />
5<br />
7<br />
1970<br />
1971<br />
1972<br />
1973<br />
1974<br />
1975<br />
1976<br />
1977<br />
1978<br />
1979<br />
1980<br />
1981<br />
1982<br />
1983<br />
1984<br />
1985<br />
1986<br />
1987<br />
1988<br />
1989<br />
1990<br />
1991<br />
1992<br />
1993<br />
1994<br />
1995<br />
1996<br />
1997<br />
1998<br />
1999<br />
2000<br />
2001<br />
2002<br />
2003<br />
2004<br />
2005<br />
2006<br />
2007<br />
2008<br />
2009F<br />
11<br />
However, “economic inflation” in the form of the CPI is<br />
not the only driver of loss costs. In pricing, one ultimate<br />
loss framework is composed of frequency (number<br />
of claims), and severity (the cost per claim). This is an<br />
effective and simple framework, and helps <strong>to</strong> identify<br />
loss cost drivers. Inflation is mainly linked as a driver of<br />
severity, but it is only one of many. To illustrate this with<br />
real statistics, let us take a look at Liability in the UK.
EXAMPLE FROM THE UK MARKET<br />
Figure 89 comparing the average severity of a UK Liability<br />
portfolio <strong>to</strong> the CPI and wage inflation index clearly<br />
shows that the trend in claims costs far outstrips the<br />
“economic” inflation indices: superimposed inflation<br />
is real. In the UK Liability market, many past studies<br />
indicated that the overall loss trend – the average loss per<br />
claim – had grown by around 10% per year in a period<br />
where the underlying Consumer Price Index inflation<br />
was in the range of 2-3%. It is worth noting that this<br />
superimposed inflation does not necessarily need <strong>to</strong><br />
correlate closely <strong>to</strong> the Consumer Price Index.<br />
Fig. 89: UK Example – Long-term relationship<br />
Severity Indices vs. CPI<br />
Index (1990 = 1.00)<br />
6.00<br />
5.00<br />
4.00<br />
3.00<br />
2.00<br />
1.00<br />
Median<br />
ACPC<br />
Flat increase<br />
2-yr av median<br />
4-yr av median<br />
CPI<br />
wage<br />
Economic Inflation is a broad basket of items, some of<br />
which are more relevant <strong>to</strong> loss inflation than others,<br />
such as health care costs, wage costs, and repair costs.<br />
Furthermore, there are many other fac<strong>to</strong>rs affecting<br />
claim costs that are not directly related <strong>to</strong> “economic”<br />
inflation. To name just a few: changes in laws and<br />
regulations (such as the admission or non-admission<br />
of pain and suffering); benefit schemes (such as the<br />
Ogden tables in the UK); social attitudes (e.g. tendency<br />
<strong>to</strong> file claims). These additional drivers of loss inflation<br />
are not necessarily correlated <strong>to</strong> economic inflation.<br />
Figure 90 and Figure 91 correlate the one-year loss<br />
trends <strong>to</strong> the CPI and wage inflation index.<br />
0.00<br />
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009<br />
Year<br />
Source: SCOR data.<br />
Severity Trends have far outstripped general inflation; year-on-year severity changes <strong>to</strong>o erratic <strong>to</strong> provide reliable trends; long-term trends<br />
more stable, but require evaluation of past sources of superimposed inflation and future expectations.<br />
SCOR - November 2010 - 87
Fig. 90: UK Ex. – Short-term relationship<br />
Severity comparison <strong>to</strong> wage inflation<br />
2-yr median<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
0%<br />
-10%<br />
1% 2% 3% 4% 5% 6% 7%<br />
-20%<br />
Source: SCOR data. Wage Inflation Source: SCOR data.<br />
Clearly, it is hard <strong>to</strong> find this direct correlation, and in<br />
fact the straightforward correlation <strong>to</strong> the wage index<br />
misleadingly implies a negative correlation, which<br />
practically, does not make much sense. This is not just<br />
due <strong>to</strong> the other drivers of the severity trend, but also<br />
<strong>to</strong> the random nature of claims severity, and the timing<br />
of claims. There are many “dates” affecting claims<br />
severity other than the date of policy inception: date<br />
of accident, date of settlement, date of payment, etc.<br />
When picturing the relationship between the severity<br />
index and wage inflation in the longer term, some trends<br />
become visible. As shown in Figure 92, loss inflation and<br />
wage inflation have indicated a positive correlation from<br />
1998 <strong>to</strong> 2009. Nonetheless, one can clearly see that<br />
fac<strong>to</strong>rs other than inflation affect claims severity. There<br />
was a stronger relationship between the 4-year severity<br />
index and wage inflation, but there are also many other<br />
Fig. 92: UK Ex. – Long-term relationship<br />
4-yr median loss severity index against wage<br />
inflation index<br />
Loss Inflation Index<br />
4.00<br />
3.50<br />
3.00<br />
2.50<br />
2.00<br />
1.50<br />
1.00<br />
1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40<br />
Source: SCOR data.<br />
Wage Inflation Index<br />
88 - November 2010 - SCOR<br />
From 1998 <strong>to</strong> 2009, Loss and Wage<br />
Inflation have tracked and severely,<br />
but same trend was not apparent<br />
in period of four previous years<br />
2-yr median<br />
Fig. 91: UK Ex. – Short-term relationship<br />
Severity comparison <strong>to</strong> CPI<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
0%<br />
-10%<br />
1% 1% 2% 2% 3% 3% 4% 4% 5%<br />
-20%<br />
elements <strong>to</strong> deal with. The general trend is shown in the<br />
data, but the relationship is weak.<br />
The purpose herein is not <strong>to</strong> try <strong>to</strong> minimize the effect<br />
of inflation, but <strong>to</strong> highlight the fact that there are other<br />
fac<strong>to</strong>rs involved, and most importantly that a quick<br />
look at the statistics (especially when looking at short<br />
periods) may misleadingly imply the non-existence of<br />
such a relationship.<br />
MANAGING THE TREATMENT<br />
OF INFLATION IN PRICING<br />
CPI<br />
Loss inflation has more of an effect on long-tail lines<br />
such as Liability. In this example, if the inflation estimate<br />
is changed by 1%, the effect on the loss cost would<br />
reach 8% undiscounted and 7% discounted.<br />
Loss Inflation 4-yr median<br />
Fig. 93: UK Ex. – Long-term relationship<br />
4-yr median loss severity <strong>to</strong> wage inflation<br />
20%<br />
18%<br />
16%<br />
14%<br />
12%<br />
10%<br />
8%<br />
6%<br />
4%<br />
2%<br />
0%<br />
2.0% 2.5% 3.0% 3.5% 4.0% 4.5% 5.0%<br />
Source: SCOR data.<br />
Wage Inflation
Fig. 94: Effect of Loss Inflation and Discounting on Pricing<br />
Cumulative Payments - base 100<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
1 3 5 7 9 11 13 15 17 19 21 23 25<br />
Year of Maturity<br />
10% Inflation 10% Inflation & 4% Discount 11% Inflation & 5% Discount<br />
11% Inflation 11% Inflation & 4% Discount<br />
Source: SCOR data.<br />
Known inflation can be priced in and has a more<br />
leveraged effect on long-tail Lines of Business, but<br />
this leveraging effect is reduced after discounting for<br />
investment income. When and if inflation and interest<br />
rates move in tandem, price is not materially affected,<br />
but that is a big IF.<br />
On an excess of loss basis, however, the effect is more<br />
dramatic. The leveraging effect makes loss inflation<br />
(economic and superimposed) much more dangerous.<br />
There are two effects <strong>to</strong> consider here. First, for losses<br />
already above a certain threshold, the increase caused<br />
by inflation will affect the part above the threshold;<br />
for example, if a claim is valued at EUR 110 with a<br />
threshold of EUR 100 (making excess of loss EUR 100),<br />
the increase of 1% from EUR 110 <strong>to</strong> EUR 111 will affect<br />
the excess layer by 10%! From 10 <strong>to</strong> 11. Furthermore,<br />
one should consider losses that were formerly below<br />
the threshold but now exceed it due <strong>to</strong> inflation.<br />
OTHER POTENTIAL DANGERS<br />
FROM THE ECONOMIC CRISIS<br />
One should also consider other elements that do not<br />
really constitute inflation per se but rather are related<br />
<strong>to</strong> economic scenarios. Some of these may affect<br />
frequency (or the number of claims) rather than severity.<br />
Although these are currently purely speculative, one<br />
cannot dismiss them and should keep an eye on them in<br />
order <strong>to</strong> protect the viability of (re)insurance industry.<br />
Governments are accumulating more and more debt and<br />
there will come a point when they need <strong>to</strong> deal with it.<br />
Pattern Effect: 1% inflation Index Effect<br />
10% Inflation - Undiscounted 100 0%<br />
10% Inflation - Discounted at 4% 74 -26%<br />
11% Inflation 108 8%<br />
11% Inflation - Discounted at 4% 79 7%<br />
11% Inflation - Discounted at 5% 74 0%<br />
Inflation is one possible solution in this regard. Pension<br />
or unemployment reforms are another; however, while<br />
they may not lead directly <strong>to</strong> increased losses, the people<br />
affected may be inclined <strong>to</strong> find more reasons <strong>to</strong> file<br />
claims. Another possible solution is <strong>to</strong> cut the budgets<br />
of certain government departments, thereby increasing<br />
pressure <strong>to</strong> find alternative sources of revenue. Insurance<br />
is one potentially tempting source, with the possible<br />
option of allocating more cost burdens (hospitalization<br />
for example) <strong>to</strong> the insurance sec<strong>to</strong>r.<br />
Furthermore, high unemployment rates and low or<br />
reduced income could tempt people <strong>to</strong> turn <strong>to</strong> insurance<br />
as an alternative source of revenue. The increased filing<br />
of injury claims, whether legitimate or fraudulent, and<br />
the increase in arson and theft, are all examples of such<br />
dangers. Economic pressures, which could negatively<br />
impact safety and maintenance standards and therefore<br />
cause more accidents, are another example.<br />
The most important things <strong>to</strong> remember when pricing<br />
are <strong>to</strong> stay on the alert and <strong>to</strong> take whatever<br />
materializes in the projections in<strong>to</strong> consideration.<br />
There are, however, other ways <strong>to</strong> manage these<br />
<strong>risk</strong>s, such as:<br />
• Holding more capital <strong>to</strong> cover the increase;<br />
• Using more inflation-indexed investments;<br />
• Integrating indexation clauses;<br />
• Improving the quality of data used for reserving<br />
and pricing;<br />
• Involving regula<strong>to</strong>rs in the discussion;<br />
• Using Asset Liability <strong>Management</strong>, a key technique<br />
for analyzing the combined effect on assets and<br />
liabilities.<br />
SCOR - November 2010 - 89
10<br />
THE FINANCIAL MARKETS VIEW<br />
ON THE RISK/REWARD STRATEGIES<br />
IN THE (RE)INSURANCE<br />
INDUSTRY MARCO CIRCELLI<br />
Head of Group Corporate Finance & Financial Communications, SCOR<br />
This article presents the view<br />
of the financial markets (analysts and inves<strong>to</strong>rs)<br />
with regard <strong>to</strong> the <strong>risk</strong>/reward strategies of<br />
(re)insurance companies.<br />
I. The inves<strong>to</strong>r’s reassessment<br />
of a reinsurer’s key strategic<br />
choices during the financial crisis<br />
Reinsurance companies are profiled on the basis of the<br />
key choices made around their strategic axes. These<br />
strategic axes enable us <strong>to</strong> determine the <strong>risk</strong>/reward<br />
profile of a (re)insurance company.<br />
As shown in Figure 95, which is taken from SCOR’s<br />
strategic plan, there are four key strategic corners<strong>to</strong>nes:<br />
Portfolio <strong>Management</strong>, Capital <strong>Management</strong>,<br />
Franchise <strong>Management</strong> and Cycle and Growth<br />
<strong>Management</strong>. Within these four strategic corners<strong>to</strong>nes<br />
there are four variables, <strong>to</strong> wit, volume, solvency,<br />
<strong>risk</strong>-<strong>based</strong> capital and profit. These variables have an<br />
impact on the Group as a whole, in terms of franchise,<br />
profitability, <strong>risk</strong> diversification and financial rating.<br />
Along these strategic axes, (re)insurance companies<br />
have pursued three key business models since the<br />
WTC in 2001. These three business models are called<br />
90 - November 2010 - SCOR<br />
business models A, B, and C by default. They are <strong>based</strong><br />
on the strategic corners<strong>to</strong>nes mentioned previously:<br />
• Business model A focuses more on traditional reinsurance<br />
business, and on high business diversification.<br />
It follows a cautious asset management <strong>approach</strong><br />
and a restrained capital management strategy, with<br />
a consistent dividend policy, low financial leverage,<br />
and capital allocation concentrated on business <strong>risk</strong>s<br />
with a focus on technical profitability. Concerning<br />
cycle and growth management, this business model<br />
pursues a long-term <strong>approach</strong> geared <strong>to</strong>wards liquidity<br />
and operating cash flow.<br />
• Business model B follows a more opportunity-<br />
driven business <strong>approach</strong> <strong>based</strong> on rather limited<br />
diversification (monoliners, Cat players). With regard<br />
<strong>to</strong> capital management, the objective of this business<br />
model is <strong>to</strong> maximize returns through high, but<br />
volatile, dividend distribution and share buy-backs.<br />
A large amount of capital is allocated <strong>to</strong> Cat eventdriven<br />
business.<br />
• Business model C is similar <strong>to</strong> business model A.<br />
However, business model C focuses on new business<br />
lines, including innovative business lines such as banking<br />
business and Credit Default Swap (CDS) business.<br />
A large part of capital is allocated <strong>to</strong> an aggressive<br />
asset management policy. The capital management of<br />
business model C is geared <strong>to</strong>wards the maximization<br />
of returns in the short term through share buy back<br />
and high leverage. Its financial strategy has a limited<br />
focus on liquidity and operating cash flow.
Fig. 95: The profile of a reinsurance company is a matter of key choices<br />
around its strategic axes<br />
Source: SCOR.<br />
IN THE (RE)INSURANCE INDUSTRY THERE<br />
WAS A CLEAR VALUE GAP BETWEEN<br />
THE BUSINESS MODELS<br />
Figure 96 compares share price <strong>to</strong> book value per share<br />
(P/BV) versus the return on equity (ROE) of the three<br />
business models. One allocated real companies <strong>to</strong> each<br />
business model, <strong>based</strong> on their <strong>risk</strong> <strong>approach</strong>, business<br />
diversification, asset strategy, etc.<br />
As demonstrated in the graph, in 2006, before the<br />
financial crisis, business models B and C were clearly<br />
better valued by the financial markets than business<br />
model A, with P/BV at a premium.<br />
Business models A and C had similar ROEs but business<br />
model C was valued higher in terms of P/BV, as many inves<strong>to</strong>rs<br />
liked the innovative nature of its business strategy.<br />
Of all the business models, business model B had a much<br />
higher ROE and P/BV than the others, as monoliners<br />
were producing much higher returns, especially in years<br />
of low catastrophe events.<br />
Fig. 96: In the reinsurance industry there was a clear value gap<br />
between the business models - Pre-crisis multiples (2006)<br />
P/BV 2006<br />
1.45<br />
1.4<br />
1.35<br />
Portfolio<br />
<strong>Management</strong><br />
Franchise<br />
<strong>Management</strong><br />
Model A<br />
Higher<br />
diversification<br />
Model C<br />
Value gap<br />
1.3<br />
ROE 2006<br />
13% 15% 17% 19% 21% 23% 25%<br />
RISK BASED CAPITAL<br />
(economic capital)<br />
PROFIT<br />
Model B In 2006, before the financial crisis,<br />
business models B and C were clearly<br />
better valued by the market than<br />
business model A, with a higher P/BV<br />
at a premium.<br />
Business models A and C had similar<br />
ROEs but the business model C<br />
was higher valued in terms of P/BV.<br />
From all the business models, business<br />
model B had a much higher ROE<br />
and P/BV than the others.<br />
Source: Factset as of 03/06/2010.<br />
Capital<br />
<strong>Management</strong><br />
VOLUME SOLVENCY<br />
(available capital)<br />
Lower<br />
franchise<br />
Higher<br />
franchise<br />
Lower<br />
diversification<br />
<strong>Risk</strong> diversification<br />
Franchise<br />
Rating<br />
Profitability<br />
Higher<br />
ROE<br />
Lower<br />
rating<br />
A+<br />
Higher<br />
rating<br />
Lower<br />
ROE<br />
900 bps<br />
Cycle & Growth<br />
<strong>Management</strong><br />
SCOR - November 2010 - 91
WHAT HAPPENED DURING THE FINANCIAL<br />
MARKET CRISIS?<br />
In fact, the financial crisis has led <strong>to</strong> a general reassessment<br />
of these business models, contradicting the<br />
strategic axes of reinsurance companies and impacting<br />
the four key strategic corners<strong>to</strong>nes previously quoted:<br />
• The portfolios of some reinsurance companies have<br />
suffered from a lack of diversification, especially some<br />
credit reinsurance companies and other monoliners.<br />
• Their franchise management has faced difficulties<br />
due <strong>to</strong> financial activities that have led <strong>to</strong> major issues<br />
(e.g. CDS exposure results in significant write-downs).<br />
• Aggressive asset management has led <strong>to</strong> substantial<br />
asset write-downs and impairments.<br />
• In addition, over-active capital management <strong>to</strong>ols<br />
have led <strong>to</strong> stretched capital bases, low financial<br />
flexibility and questionable share buy-back programs.<br />
Shortly before the crisis, certain companies started <strong>to</strong><br />
buy back their own shares when their share price was<br />
high, and leveraged the companies by issuing debt.<br />
• Cash flow management has been neglected, which<br />
has resulted in liquidity stress and has forced sales of<br />
assets and reputational problems.<br />
These fac<strong>to</strong>rs have led <strong>to</strong> a general reassessment of the<br />
business models, particularly business model C. The<br />
latter appears <strong>to</strong> be less ‘suitable’ than business models<br />
A and B in a period when the financial environment<br />
is challenging.<br />
INVESTORS HAVE STARTED TO REWARD<br />
THE SECTOR FOR RISK MANAGEMENT…<br />
Inves<strong>to</strong>rs have started <strong>to</strong> reward the sec<strong>to</strong>r for <strong>Risk</strong><br />
<strong>Management</strong>. They have mainly distinguished between<br />
three key areas:<br />
• The investment margin is, on the Life side, a major<br />
driver for profitability; transparency is very important.<br />
It enables inves<strong>to</strong>rs <strong>to</strong> focus on guarantees and the<br />
quality of hedging, and <strong>to</strong> track economic assumption<br />
changes. On the P&C side, the market can be suspicious<br />
of outsized investment returns. The key is therefore<br />
a balanced and clearly communicated investment<br />
strategy.<br />
• The <strong>risk</strong> margin is less of a driver but still constitutes a<br />
focus in terms of quality and track record; it is a useful<br />
<strong>to</strong>ol for distinguishing between a volatile and a stable<br />
investment, especially on the Life side. Inves<strong>to</strong>rs look<br />
for operating variances and assumption track records<br />
in the development of embedded value, and for<br />
sensitivity <strong>to</strong> mortality and morbidity scenarios.<br />
• The expense margin focuses more on the P&C side<br />
by annually reassessing the track record in terms of<br />
combined ratios, reserved track records, and sensitivity<br />
<strong>to</strong> large loss scenarios.<br />
Fig.97: Resulting in the repricing of reinsurers’ business models<br />
Pre-crisis multiples (2006) Post-crisis multiples (2009)<br />
P/BV 2006<br />
1.5<br />
1.4<br />
1.3<br />
1.2<br />
1.1<br />
1.0<br />
0.9<br />
0.8<br />
92 - November 2010 - SCOR<br />
Model C<br />
Model A<br />
Model B<br />
0.7<br />
ROE 2006<br />
0.6<br />
0% 5% 10% 15% 20% 25%<br />
Source: Factset as of 03/06/2010.<br />
P/BV 2009<br />
1.5<br />
1.4<br />
1.3<br />
1.2<br />
1.1<br />
1.0<br />
0.9<br />
0.8<br />
0.7<br />
Model C<br />
Model A<br />
ROE 2009<br />
0.6<br />
0% 5% 10% 15%<br />
Model B<br />
20%
… RESULTING IN THE RE-PRICING OF REINSURERS’ BUSINESS MODELS<br />
Fig. 98: Differences in earnings expectations are also smaller<br />
Earnings Yield (1)<br />
40%<br />
30%<br />
20%<br />
10%<br />
Model A:<br />
Long-term average: 12.8%<br />
Post-crisis average: 14.4%<br />
Model B:<br />
Long-term average: 14.6%<br />
Post-crisis average: 15.6%<br />
Model C:<br />
Long-term average: 15.6%<br />
Post-crisis average: 18.1%<br />
Lehman Brothers'<br />
bankruptcy<br />
Re-pricing<br />
Model A: 14.4%<br />
Model B: 14.8%<br />
Model C: 14.0%<br />
0%<br />
Jan-06 May-06 Sep-06 Jan-07 May-07 Sep-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 May-10<br />
(1) Next 24 Months Earnings/Price calculated at 1/average. (Price/Next 24 Months Earnings).<br />
Source: Factset as of 09/06/2010.<br />
According <strong>to</strong> Figure 97, which has appeared already,<br />
business model A was previously valued lower than<br />
business models B and C. Whereas the reassessment<br />
and re-pricing of the insurance industry in 2009 led <strong>to</strong><br />
business model A being valued the highest, even though<br />
it delivers lower returns than business models B and C.<br />
Business model C has clearly been under stress over the<br />
last few years, and has become the least valued. Earning<br />
expectations are not as compelling as before.<br />
Let us compare the earning expectations of the three<br />
business models at the beginning of 2006 with those<br />
at the beginning of 2010 using the graph below. This<br />
graph shows the share price versus the next 24 months’<br />
earnings expectations (earning yields).<br />
Fig. 99: The best option is <strong>to</strong> consistently deliver over time<br />
It appears that in 2006 the earning yield, or earnings<br />
expectation, for business model A was lower, and<br />
that the yields for business models B and C were<br />
much higher. Before Lehman Brothers’ bankruptcy,<br />
the earning yields of all three business models were<br />
pretty stable. However, when the financial market<br />
crisis started, business model C’s earnings expectation<br />
drastically decreased and became very volatile. At the<br />
same time business models A and B were, compared<br />
<strong>to</strong> business model C, less affected by volatility.<br />
Given the return on equity ratios and the average for<br />
each business model, the best option for inves<strong>to</strong>rs is <strong>to</strong><br />
consistently deliver over time. Although business model<br />
A did not generate the most attractive ROE over the last<br />
five years, it turns out that it had the most consistent<br />
ROE in the standard deviation over the last five years.<br />
Company’s ROEs 2005 2006 2007 2008 2009 Average Standard<br />
Deviation<br />
Business Model A 5.7% 17.0% 14.7% 5.0% 15.1% 11.5% 5.7%<br />
Business Model B 1.8% 22.8% 20.5% 4.5% 18.8% 13.7% 9.8%<br />
Business Model C 1.9% 17.9% 19.0% -1.9% 9.2% 9.2% 9.3%<br />
Average 3.1% 19.2% 18.1% 2.5% 14.4% 11.5% 8.1%<br />
Source: Company’s data.<br />
SCOR - November 2010 - 93
II. Inves<strong>to</strong>rs’ <strong>approach</strong> <strong>to</strong> (re)insurance company valuation<br />
Fig. 100: “The valuation life cycle” – how inves<strong>to</strong>rs look at (re)insurance companies<br />
Methodology Pros Cons<br />
Appraisal value<br />
(EV (1) + multiple of life<br />
NBV (2) )<br />
HOW DOES AN INVESTOR EVALUATE<br />
A COMPANY DURING THE FINANCIAL<br />
MARKET CYCLE?<br />
As shown in Figure 100, which is taken from a Goldman<br />
Sachs presentation, inves<strong>to</strong>rs distinguish between three classic<br />
markets: the bull, the mid-cycle and the bear market.<br />
• During a bull market, inves<strong>to</strong>rs focus on values such as<br />
embedded value and future profit, which characterize<br />
inves<strong>to</strong>rs’ confidence and optimistic expectations. This<br />
means that inves<strong>to</strong>rs allocate capital <strong>to</strong> Life and Non-<br />
Life. Based on the computed ROE and capital cost of<br />
each division, they round the allocated capital up for<br />
the valuation. This <strong>approach</strong> puts a higher valuation<br />
on companies than they are actually worth.<br />
• Conversely in bear markets, inves<strong>to</strong>rs focus on tangible<br />
net asset values or equity excluding intangibles, or on<br />
classical Solvency I ratios. This is a period characterized<br />
by a fall in investment prices and pessimism, so people<br />
pay more attention <strong>to</strong> hard facts than estimates.<br />
• As for the mid-cycle market, which is marked by a<br />
period in which inves<strong>to</strong>rs are more uncertain, they<br />
compare companies according <strong>to</strong> the ROE and the<br />
price <strong>to</strong> book value.<br />
“THE VALUATION LIFE CYCLE” –<br />
HOW INVESTOR LOOK AT (RE)INSURANCE<br />
COMPANIES<br />
• Simplicity<br />
• Unreflective of returns<br />
• Varying quality of assumptions<br />
• Arbitrary new business multiples<br />
Sum of the parts • Useful for composite businesses<br />
• Allows differential valuation of excess capital<br />
RoE (3) /RoEV (4) vs. P/B (5) /<br />
P/EV (6)<br />
• Good overall comparison<br />
• Links <strong>to</strong> BV growth, which is observable<br />
P/E (7) (7<br />
• Simplicity<br />
• Good for “earnings businesses”<br />
1x P/TNAV (8) P/TN<br />
• Focus on pure tangible assets<br />
for shareholders<br />
Solvency ency I<br />
• A more accurate measure of company<br />
stress<br />
• In line with local regula<strong>to</strong>rs<br />
• Reasonably simple <strong>to</strong> moni<strong>to</strong>r<br />
Figure 101 confirms what we have just been discussing.<br />
In order <strong>to</strong> comprehend inves<strong>to</strong>r behavior, Merrill Lynch<br />
• Easy <strong>to</strong> misallocate capital between<br />
divisions<br />
• Need for x-cycle adjustments<br />
• Neither EV or BV(9)<br />
are fully comparable<br />
• Cyclicality and volatility of stated earnings<br />
• Dis<strong>to</strong>rted for fast growing life businesses<br />
• Growing inconsistencies in IFRS<br />
• Not a consistent measure<br />
• Liabilities not at fair value<br />
• Strip out DAC(10)<br />
? VOBA (11) ?<br />
• No allowance for returns<br />
• Not an economic measure<br />
• Ignores rating agencies<br />
• Ignores other triggers, e.g. leverage<br />
• Inconsistency of calculation<br />
Source: Goldman Sachs, Oc<strong>to</strong>ber 2009, “Strategic relevance of changes in key insurance metrics”.<br />
(1) EV: Embedded Value. (2) NBV: Net Business Value. (3) RoE: Return on Equity. (4) RoEV: Return on Embedded Value.<br />
(5) P/B: Price <strong>to</strong> Book. (6) P/EV: Price <strong>to</strong> Embedded Value. (7) P/E: Price Earnings ratio. (8) P/TNAV: Price <strong>to</strong> Tangible Net Asset Value.<br />
(9) BV: Book Value. (10) DAC: Deferred Acquisition Costs. (11) VOBA: Value of Business Acquired.<br />
94 - November 2010 - SCOR<br />
Bull Market<br />
Mid<br />
Cycle<br />
Bear Market<br />
led a study in which the bank asked several insurance<br />
and reinsurance inves<strong>to</strong>rs a series of questions, most<br />
notably: what is the most important valuation metric<br />
that you utilize <strong>to</strong>day?<br />
The bank first asked them this question in 2008, during<br />
the crisis, and then once again in May 2010. In 2008,<br />
inves<strong>to</strong>rs were focused mainly on tangible book value,<br />
whereas in 2010, they concentrated on <strong>risk</strong> earnings<br />
and free cash flow yield. The development over the<br />
last twelve months in the graph on the right shows<br />
that attention <strong>to</strong> free cash flow yield has significantly<br />
increased over the last twelve months in comparison<br />
with the tangible book value. Overall, the focus from<br />
inves<strong>to</strong>rs is more on earnings and free cash flow, which<br />
suggests a mid-cycle market. Whereas the Merrill Lynch<br />
study shows that in 2008 the focus was on tangible<br />
book value, which presupposes a bear market.<br />
PRICE TO ADJUSTED TANGIBLE BOOK VALUE<br />
IS THE MOST ADEQUATE VALUATION<br />
METRIC FOR (RE)INSURANCE COMPANIES<br />
Inves<strong>to</strong>rs utilize three ratios in order <strong>to</strong> evaluate<br />
a (re)insurance company:<br />
• The price <strong>to</strong> book value is widely used by analysts<br />
across the reinsurance industry and enables them<br />
<strong>to</strong> compare a s<strong>to</strong>ck market value <strong>to</strong> its book value.<br />
It is computed by dividing the share price by<br />
the equity.
Fig. 101: “Valuation life cycle” confirmed by most recent inves<strong>to</strong>r survey<br />
What is your most important valuation metric <strong>to</strong>day?<br />
Survey 2008<br />
13%<br />
14%<br />
Survey 2010<br />
16%<br />
19%<br />
11% 1%<br />
18%<br />
30%<br />
19%<br />
43%<br />
16%<br />
Tangible book value<br />
Embedded value<br />
IFRS earnings<br />
IFRS book value<br />
Free cash flow yield<br />
Dividend yield<br />
• The price <strong>to</strong> tangible book value aims <strong>to</strong> give a<br />
better valuation as it focuses on ‘utilized’ capital<br />
base and corresponds <strong>to</strong> the price of each share if<br />
the company liquidated all of its assets. This ratio is<br />
calculated by dividing the share price by the equity<br />
minus the intangibles.<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
Tangible<br />
book value<br />
Almost half of the inves<strong>to</strong>rs still look at IFRS<br />
earnings and tangible book value (P/TBV)<br />
– mid valuation life cycle?<br />
Free cash flow yield is more important,<br />
as it might give an early warning signal <strong>to</strong><br />
mismanagement resulting in major issues<br />
Embedded<br />
value<br />
IFRS<br />
earnings<br />
IFRS<br />
book value<br />
Free cash<br />
flow yield<br />
Dividend<br />
yield<br />
Jun-09 Sep-09 Dec-09 May-10<br />
Source: BofA Merrill Lynch Global Research.<br />
• The price <strong>to</strong> adjusted tangible book value is the book<br />
value after having adjusted assets and liabilities <strong>to</strong><br />
market value. The share price is divided by the equity<br />
minus the intangibles, which is added the value of<br />
business acquired.<br />
Fig. 102: Price <strong>to</strong> adjusted tangible book value the most adequate valuation<br />
metric for (re)insurance companies<br />
1 Price <strong>to</strong> Book Value (P/BV)<br />
Widely used by analysts across the reinsurance business P/B =<br />
2 Price <strong>to</strong> Tangible Book Value (P/TBV)<br />
Aims <strong>to</strong> give a better valuation as it focuses<br />
on “utilized” capital base<br />
3 Price <strong>to</strong> Adjusted Tangible<br />
Book Value (P/Adj. TBV)<br />
Removes the effect of goodwill<br />
on acquired life portfolios for<br />
recent acquisitive players<br />
P/Adj. TBV =<br />
Market Cap<br />
Shareholders’ Equity<br />
Market Cap<br />
P/TBV =<br />
Shareholders’ Equity – Intangibles (1)<br />
Market Cap<br />
Shareholders’ Equity – Intangibles (1) + VOBA (2)<br />
(1) Includes Deferred Acquisition Costs (DAC) and VOBA. (2) Value of Business Acquired (VOBA).<br />
Development over the last 12 months<br />
SCOR - November 2010 - 95
III. Inves<strong>to</strong>rs’ view of the sec<strong>to</strong>r<br />
remains uncertain despite<br />
their reassessment<br />
WHAT ARE THE ORIGINS<br />
OF THIS UNCERTAINTY?<br />
Many analysts’ industry reports have been published<br />
recently, however the range of <strong>to</strong>pics is very wide and<br />
opinions vary:<br />
• Global financial crisis effects<br />
• Solvency II<br />
• Capital <strong>Management</strong><br />
• Inflation <strong>risk</strong> and scenarios<br />
• Dividend expectation<br />
• Pricing / Underwriting discipline<br />
• Preference for Life / Non-Life?<br />
• Reserve position<br />
This clearly shows that there are still many uncertainties<br />
emanating from various sources, notably the regula<strong>to</strong>rs.<br />
The most prominent concern is unquestionably the<br />
European directive Solvency II, which will have a<br />
significant impact on (re)insurers’ activities.<br />
UNCERTAINTY CAN BE SEEN IN THE HIGHER<br />
AMOUNT OF “NEUTRAL” VIEWS<br />
ON THE SECTOR, AND IN AN INCREASED<br />
PREFERENCE FOR LIFE<br />
The Merrill Lynch survey reveals that the inves<strong>to</strong>rs<br />
questioned on their stance and preferences vis-à-vis the<br />
insurance industry placed more trust in the (re)insurance<br />
industry during the crisis in 2008 than they did in 2010.<br />
Moreover, the percentage of neutral ratings increases in<br />
2010, which indicates that inves<strong>to</strong>rs have less confidence<br />
in this industry than before. During the crisis, there was<br />
a preference for Non-Life, whereas in 2010 this trend<br />
reverses with an increased preference for Life, which<br />
means that inves<strong>to</strong>rs prefer <strong>to</strong> pursue a more stable business<br />
line. This is reflected in the following question: What<br />
will the most important driver of the sec<strong>to</strong>r be over the<br />
coming six <strong>to</strong> twelve months? It is interesting <strong>to</strong> note that<br />
regula<strong>to</strong>ry development was not a concern before. In May<br />
2010, it became one of the key concerns for inves<strong>to</strong>rs.<br />
PARADOXICALLY, INVESTORS ARE ALREADY<br />
ASKING FOR INCREASED RISK APPETITE<br />
AND MORE ACTIVE CAPITAL MANAGEMENT<br />
Indeed, 30% of the inves<strong>to</strong>rs questioned would like <strong>to</strong><br />
“raise the <strong>risk</strong> curve with regard <strong>to</strong> invested assets” over<br />
the next few months. 26% of them want <strong>to</strong> maintain<br />
a fortress balance sheet in view of the continuing<br />
uncertainty over Solvency II. And 21% of them would<br />
like <strong>to</strong> increase the <strong>risk</strong> appetite in terms of invested<br />
assets and dividends, which means being more offensive<br />
in terms of capital management.<br />
Fig. 103: Uncertainty can be seen in higher amount of “neutral” views<br />
on the sec<strong>to</strong>r and an increased preference on Life<br />
38%<br />
21%<br />
96 - November 2010 - SCOR<br />
Survey 2008 Survey 2010<br />
11%<br />
What is your current<br />
sec<strong>to</strong>r stance?<br />
5%<br />
13%<br />
6%<br />
6%<br />
32%<br />
73%<br />
Source: BofA Merrill Lynch Global Research.<br />
Very overweight<br />
Somewhat overweight<br />
Neutral<br />
Somewhat underweight<br />
Very underweight<br />
Within the sec<strong>to</strong>r, do you<br />
have a preference for life<br />
or non-life s<strong>to</strong>cks?<br />
Life<br />
Non-Life<br />
No preference<br />
51%<br />
26%<br />
21%<br />
8%<br />
5%<br />
31%<br />
53%
Fig. 104: Regula<strong>to</strong>ry developments are rising inves<strong>to</strong>rs’ attention<br />
What will be the most important driver of the sec<strong>to</strong>r over the coming 6-12 months?<br />
19%<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
8% 3% 0%<br />
22%<br />
Earnings<br />
estimate<br />
revisions<br />
GENERALLY, WHAT ARE THE KEY<br />
TAKE-AWAYS?<br />
• Prior <strong>to</strong> the financial crisis, analysts and inves<strong>to</strong>rs<br />
largely favoured business models focused on a limited<br />
number of business lines, or financial products with<br />
more aggressive behavior <strong>to</strong>wards capital and asset<br />
management. There was also a clear valuation gap<br />
between business models A, B, C and the traditional<br />
reinsurance business model.<br />
26% Earnings estimates revisions<br />
Valuation re-rating (up or down)<br />
Regula<strong>to</strong>ry developments (e.g. Solvency II)<br />
Book value development<br />
M&A<br />
Capital action (right issues, buy-backs, positive/negative dividend surprises)<br />
Disclosure changes (presumably enhancements)<br />
22%<br />
Valuation<br />
re-rating<br />
Regula<strong>to</strong>ry<br />
development<br />
Jun-09 Sep-09 Dec-09 May-10<br />
BV<br />
development<br />
M&A<br />
Capital<br />
action<br />
Disclosure<br />
changes<br />
Source: BofA Merrill Lynch Global Research.<br />
Fig. 105: At the same time inves<strong>to</strong>rs are already asking for increased <strong>risk</strong> appetite<br />
and more active capital management<br />
Given the recovery of balance sheets, do you believe that insurers should be starting <strong>to</strong>:<br />
13%<br />
21%<br />
10%<br />
26%<br />
30%<br />
Move up the <strong>risk</strong> curve, with regard <strong>to</strong> invested assets<br />
Maintain a fortress balance sheet given continuing Solvency II uncertainty<br />
Increase dividends/buy back s<strong>to</strong>ck<br />
Preserve capital – balance sheets remain stretched<br />
Consider strategic M&A, even if this results in capital issuance<br />
Source: BofA Merrill Lynch Global Research.<br />
• In fact, the crisis has led inves<strong>to</strong>rs <strong>to</strong> reassess and<br />
re-price the sec<strong>to</strong>r, and <strong>to</strong> start rewarding companies<br />
for good <strong>risk</strong> management by using different methodologies<br />
according <strong>to</strong> the “Valuation Life Cycle”.<br />
• Inves<strong>to</strong>rs continue <strong>to</strong> focus on book value (including<br />
Goodwill), paying less attention <strong>to</strong> VOBA.<br />
• The financial markets are still uncertain about the reinsurance<br />
sec<strong>to</strong>r, which is also driven by increasing inves<strong>to</strong>r<br />
attention <strong>to</strong> regula<strong>to</strong>ry developments (Solvency II).<br />
• Paradoxically, inves<strong>to</strong>rs are already asking for increased<br />
<strong>risk</strong> appetite and more active capital management.<br />
SCOR - November 2010 - 97
Conclusion<br />
<strong>ERM</strong>: A SYSTEM TO ENSURE<br />
THE OPTIMAL SECURITY<br />
AND EFFICIENCY VICTOR PEIGNET<br />
Chief Executive Officer, SCOR Global P&C<br />
GILLES MEYER<br />
Chief Executive Officer, SCOR Global Life<br />
<strong>ERM</strong> is a fact of day-<strong>to</strong>-day life<br />
at SCOR: KYR (“Know Your <strong>Risk</strong>s”) is as important<br />
as KYC (“Know Your Clients”) within the Group,<br />
and the combination of the two is what drives its<br />
operations.<br />
<strong>ERM</strong> stands for <strong>Enterprise</strong> <strong>Risk</strong> <strong>Management</strong>, although<br />
it could be argued that this acronym is short of one key<br />
word <strong>to</strong> fully represent the concept of <strong>ERM</strong> and what<br />
it means for the reality of company operations: this<br />
missing word is SYSTEM.<br />
For a global (re)insurance company like SCOR, being<br />
<strong>ERM</strong>(s) compliant means:<br />
• Operating with a single comprehensive Information<br />
System,<br />
• Being driven by capital allocation and return on<br />
capital, <strong>based</strong> on its own internal model, and<br />
• Having a strong <strong>Risk</strong> Culture embedded in its<br />
operations.<br />
98 - November 2010 - SCOR<br />
The ultimate objective of an <strong>ERM</strong> system is <strong>to</strong> ensure the<br />
optimal security and efficiency of a company’s decision<br />
making processes and operations, without affecting its<br />
entrepreneurial spirit or innovative mindset, <strong>based</strong> on<br />
the best possible knowledge of <strong>risk</strong>s and management<br />
techniques.<br />
By 2005, SCOR had managed <strong>to</strong> recover from its 2002-03<br />
crisis and had put <strong>ERM</strong> at the center of its recovery<br />
process. Although the Group had certainly regained<br />
the organizational structure of an entrepreneurial<br />
and ambitious company, having returned <strong>to</strong> good<br />
health thanks <strong>to</strong> a strong franchise, experienced<br />
and dedicated human resources, safe and efficient<br />
procedures and processes and a corporate culture,<br />
it had not yet reached the stage where it had a strong or<br />
fully operational <strong>ERM</strong> system, or a <strong>Risk</strong> Culture served<br />
by that system.<br />
Between 2005 and 2010, SCOR implemented and<br />
completed a 5-year project <strong>to</strong> introduce such a system,<br />
which is now fully deployed at all levels throughout<br />
the Group. This major initiative has had deep,<br />
structuring effects on the way in which SCOR operates,<br />
plans its business, dynamically moni<strong>to</strong>rs and controls<br />
its exposures and makes its decisions.<br />
Basically, this system has vastly improved both the<br />
quality of data and the way in which that data is<br />
treated, from how it is entered and warehoused, both<br />
before and after processing, <strong>to</strong> how its traceability and<br />
integrity are ensured.
In an integrated system like this, all the data centers (or<br />
the global data center), and all the software programs<br />
are linked and interconnected through au<strong>to</strong>matic<br />
interfaces, transfers and downloads. Double entries are<br />
therefore minimized while au<strong>to</strong>matic checks, balances<br />
and reconciliations are maximized.<br />
This single system constitutes the corners<strong>to</strong>ne of<br />
SCOR’s <strong>ERM</strong> policy. Without it, internal modeling would<br />
have <strong>to</strong> be conducted piecemeal, which would reduce<br />
efficiency and put the <strong>risk</strong> inherent <strong>to</strong> discontinuities<br />
in<strong>to</strong> the modeling process.<br />
PRACTICALLY SPEAKING, WHAT DOES <strong>ERM</strong><br />
MEAN FOR SCOR NOW?<br />
From an operational perspective, it means that treaties are<br />
underwritten using a single pricing <strong>to</strong>ol, which is centrally<br />
controlled and parametered <strong>based</strong> on local analysis by<br />
Line of Business and by market. It also means that SCOR’s<br />
“deal teams” for given accounts or portfolios (underwriters,<br />
actuaries and modelers), along with the management<br />
team, have access <strong>to</strong> the dynamic follow up of:<br />
• The expected performance of priced and booked<br />
business, expressed either as ultimate underwriting<br />
ratios or as RoRAC (“Return on <strong>Risk</strong> Adjusted<br />
Capital”), which must exceed hurdle rates and be as<br />
close as possible <strong>to</strong> the Group’s target ROE;<br />
• The tail Value-at-<strong>Risk</strong> and the Value-at-<strong>Risk</strong> of this<br />
business, which must comply with pre-defined <strong>risk</strong><br />
appetites;<br />
• The capital that has been used, compared with the<br />
allocations at any given time, and<br />
• Au<strong>to</strong>matic year on year “as if” re-built price<br />
comparisons.<br />
This means that, during renewal seasons, the build-up<br />
of the following year’s portfolio can be moni<strong>to</strong>red in<br />
real time against the plan and the budget and that, as<br />
renewals progress, adjustments can be made <strong>to</strong> the<br />
initial capital allocations by shifting capital <strong>to</strong> markets<br />
and Lines of Business with better than expected<br />
performances. With these management and reporting<br />
<strong>to</strong>ols at their disposal, SCOR’s teams can behave like<br />
portfolio managers rather than deal makers, thereby<br />
optimizing their acceptances <strong>to</strong> meet portfolio target.<br />
WHAT MAY BE THE “DRAWBACKS”<br />
OF SUCH AN <strong>ERM</strong> SYSTEM?<br />
The first drawback is that there are <strong>risk</strong>s attached<br />
<strong>to</strong> such a user-friendly underwriting system in terms<br />
of human behavior. These <strong>risk</strong>s all involve the<br />
underwriters moving away from the traditional<br />
underwriting process and relying <strong>to</strong>o heavily on the<br />
models <strong>to</strong> make their choices. From this point of view,<br />
it is equally important for underwriters:<br />
• To have a good understanding of the modeling process<br />
in order <strong>to</strong> achieve a greater commitment <strong>to</strong> and<br />
identification with it, and <strong>to</strong> become familiar with the<br />
<strong>to</strong>ols involved, which enables them <strong>to</strong> work closely<br />
with specialized actuaries and modelers dedicated<br />
<strong>to</strong> their markets and Lines of Business within the<br />
“deal teams”,<br />
• To apply their underwriting expertise <strong>to</strong> their own<br />
assessment of the results of the models and <strong>to</strong><br />
challenge these through plausibility checks and<br />
sensitivity analysis, and<br />
• To pay as much attention <strong>to</strong> trends as <strong>to</strong> final results,<br />
whilst respecting the Group’s values of consistency<br />
and continuity in their business relationships.<br />
SCOR - November 2010 - 99
The second drawback is probably that, being <strong>based</strong><br />
on such a complex matrix of correlations, the internal<br />
model is so heavy <strong>to</strong> run that it challenges the<br />
calculation power of even the latest generation of<br />
machines. For the time being, therefore, it is practically<br />
impossible <strong>to</strong> run as many operational iterations<br />
as requested, whether for sensitivity analysis and optimization<br />
purposes at the planning and budget stage, or<br />
for unexpected large contracts with potentially material<br />
impacts over the course of a renewal season.<br />
In order <strong>to</strong> make the use of the internal model more<br />
dynamic, companies need <strong>to</strong>:<br />
• Upscale their in-house calculation capabilities,<br />
although these are limited <strong>to</strong> the power of the<br />
machines available on the market, or<br />
• Use the internal model for a limited number of detailed<br />
calculations per year, dedicated <strong>to</strong> final planning,<br />
budgeting and regula<strong>to</strong>ry areas, and have a proxy<br />
model for fast runs and quick decision making.<br />
Our Group has followed the second route and developed<br />
a simplified model (CaDeT). This model can be used<br />
throughout the organization for checks such as “what<br />
if” analysis, the assessment of diversification benefits<br />
and the implications of portfolio changes.<br />
Another advantage of having a simplified model like<br />
this is that its broad, daily use by the teams will help<br />
<strong>to</strong> maintain and further reinforce the Group’s <strong>ERM</strong><br />
culture.<br />
Since the implementation of its <strong>ERM</strong> framework, the<br />
SCOR Group has gained considerable experience and<br />
expertise. Eight areas contributed <strong>to</strong> this successful<br />
implementation:<br />
• Diversification that reduces the need for <strong>risk</strong>adjusted<br />
capital is the strategy followed by our<br />
Group. This strategy is <strong>based</strong> on SCOR’s twin business<br />
engines – Life and Non-Life – with the same weight<br />
allocated <strong>to</strong> both engines, which constitute the first<br />
level of diversification. It is worth noting that the<br />
same diversification techniques are applied within<br />
the two business engines.<br />
• The <strong>Risk</strong> Control system is a key component that<br />
SCOR has strengthened. SCOR Global Life and SCOR<br />
Global P&C are each the result of the recent mergers<br />
of several companies with different <strong>approach</strong>es <strong>to</strong> <strong>risk</strong><br />
management. In view of this, the <strong>ERM</strong> framework<br />
has improved over the last three years with the<br />
development of one single <strong>approach</strong> for both Life<br />
and Non-Life, which has been embedded within the<br />
company. For example, SCOR Global Life and SCOR<br />
Global P&C have each put in<strong>to</strong> practice a clear and<br />
defined underwriting and pricing process, having<br />
undergone a process of harmonization in<strong>to</strong> one single<br />
100 - November 2010 - SCOR<br />
system as opposed <strong>to</strong> the multiple pricing systems<br />
inherited when the entities were merged. Today,<br />
this enables SCOR <strong>to</strong> look at <strong>risk</strong>s with consistency<br />
throughout the entire organization.<br />
• Stricter adherence <strong>to</strong> profitability targets forms<br />
part of this stringent <strong>risk</strong> control system. Any deviation<br />
is referred using a defined escalation process.<br />
• Capital Shield Strategy: buying retrocession protec-<br />
tion enables us <strong>to</strong> continue <strong>to</strong> reduce volatility even<br />
further. As an example on the Life side, the purchase<br />
of mortality swaps is also a part of this capital shield<br />
strategy. SCOR Global Life has decided <strong>to</strong> reduce its<br />
gross exposure by buying two swaps that are fully<br />
collateralized.<br />
• Internal Capital Model, which is completely embed-<br />
ded in<strong>to</strong> the pricing model. This means that SCOR<br />
could prove <strong>to</strong> any stakeholder, such as the rating<br />
agencies for example, that any kind of pricing given<br />
by SCOR Global Life or SCOR Global P&C is consistent<br />
with the internal capital model.<br />
• Reserve adequacy: the reserves are checked on a<br />
regular basis, both locally and centrally, and are also<br />
reviewed by a third party consultant.<br />
• <strong>Risk</strong> enquiry, which is led by the SCOR Global Life<br />
and SCOR Global P&C Chief <strong>Risk</strong> Officers. <strong>Risk</strong> enquiry<br />
means that any type of <strong>risk</strong> that could be material <strong>to</strong><br />
the Company is tracked on a regular basis. The <strong>risk</strong>s<br />
are split in<strong>to</strong> three categories – A, B and C – depending<br />
on their potential economic impact. This system<br />
involves a large number of people throughout the<br />
organization, which is <strong>to</strong> its advantage.<br />
• Solvency II: SCOR is preparing for the upcoming<br />
Solvency II regulations in Europe. One of the key elements<br />
of Solvency II is data integrity, which forms part<br />
of the Internal Capital Model. Compared <strong>to</strong> our peers<br />
in the reinsurance market, SCOR has a key competitive<br />
advantage: a single Group-wide IT system.<br />
All of the above demonstrates how the <strong>ERM</strong> process<br />
really is at the core of our business.
SPEAKERS’ BIOGRAPHIES<br />
Guest speakers<br />
JULIEN HALFON<br />
Head of Coverage, P-Solve <strong>Risk</strong> <strong>Management</strong> Solutions<br />
Julien Halfon is the Head of Coverage for P-Solve <strong>Risk</strong> <strong>Management</strong> Solutions. He advises European<br />
corporations on pension investment and <strong>risk</strong> management issues and European and American asset<br />
managers on marketing strategies for the European institutional inves<strong>to</strong>r markets.<br />
He joined P-Solve in 2008 from the investment banking division of Lazard & Co, where he led<br />
Lazard’s European corporate pension advisory initiative. Prior <strong>to</strong> joining Lazard, Julien was a Senior<br />
Investment Consultant at Hewitt Associates where he was part of the corporate advisory team and<br />
co-ran the infrastructure manager research group. He was also a Senior Associate in the Pension<br />
& Insurance Strategy Group of Goldman Sachs International in London. Julien holds a Masters<br />
degree in International Finance and Economics from the Dauphine University of Paris, an MBA from<br />
the Whar<strong>to</strong>n School of the University of Pennsylvania and an M.A. from the School of Advanced<br />
International Studies of Johns Hopkins University.<br />
MILES TROTTER<br />
General Manager, Analytics Non-Life, A.M. Best<br />
Miles Trotter is general manager at A.M. Best’s London office. He is responsible for leading A.M. Best’s<br />
non-life analytical teams for Continental Europe and the United Kingdom, including Lloyd’s and<br />
London market companies.<br />
Miles has extensive experience in the London market. He began his career as a graduate trainee<br />
at Lloyd’s broker, Bain Dawes plc. Later he specialised in financial institutions insurance as a direc<strong>to</strong>r<br />
of Holmes Johnson Lessiter Ltd, a subsidiary of Alexander Howden Ltd. In 1989, he moved <strong>to</strong><br />
underwriting at Syndicate 1095, managed by Welling<strong>to</strong>n Underwriting Agencies Ltd.<br />
After earning a master’s degree at the London Business School, Miles returned <strong>to</strong> Welling<strong>to</strong>n <strong>to</strong> help<br />
start a new joint venture company, SBW Insurance Research Ltd. Prior <strong>to</strong> joining A.M. Best in 2001,<br />
Miles spent seven years developing this company’s research and analysis products.<br />
SCOR - November 2010 - 101
102 - November 2010 - SCOR<br />
SCOR speakers<br />
GILLES MEYER<br />
Chief Executive Officer, SCOR Global Life<br />
Gilles Meyer is graduate of a French business school, he holds an MBA from the GSBA of Zürich.<br />
Gilles Meyer began his career as an underwriter at Swiss Re before taking over the optional department<br />
of the company La Baloise in Basel. After 23 years of experience in optional and treaty reinsurance,<br />
Gilles Meyer was CEO of Alea Europe from 1999 <strong>to</strong> 2006, in charge of both the Property & Casualty<br />
and the Life reinsurance and from 2005 <strong>to</strong> 2006, he was the Direc<strong>to</strong>r of the Underwriting group of<br />
Alea. He joined the SCOR Group in January 2006 and directed the German-speaking markets of <strong>based</strong><br />
in Hanover, Basel and Winterthur.<br />
End of 2006, he <strong>to</strong>ok the position of Direc<strong>to</strong>r of the Business Unit 1 of SCOR Global Life. Since<br />
February 2008, he is CEO of SCOR Global Life SE.<br />
VICTOR PEIGNET<br />
Chief Executive Officer, SCOR Global P&C<br />
Vic<strong>to</strong>r Peignet, a Marine Engineer and graduate of the École Nationale Supérieure des Techniques<br />
Avancées (ENSTA), joined the Facultative Department of SCOR in 1984 from the offshore oil sec<strong>to</strong>r.<br />
From 1984 <strong>to</strong> 2001, he held various positions in the underwriting of Energy and Marine Transport<br />
<strong>risk</strong>s at SCOR, first as an underwriter and then as Branch Direc<strong>to</strong>r. He has led the Group’s Business<br />
Solutions (facultative) division since it was created in 2000, as both Deputy Chief Executive Officer<br />
and then as Chief Executive Officer since April 2004. On 5 July 2005, Vic<strong>to</strong>r Peignet was appointed<br />
manager of all Property & Casualty Reinsurance operations at SCOR Global P&C SE. He is currently<br />
Chief Executive Officer of SCOR Global P&C SE.<br />
MARCO CIRCELLI<br />
Head of Group Corporate Finance & Financial Communications, SCOR<br />
Marco Circelli is Head of Group Corporate Finance & Financial Communications at SCOR Group.<br />
He joined Converium’s Inves<strong>to</strong>r Relations in 2003. At the beginning of 2005 he <strong>to</strong>ok over the<br />
responsibility for the Business and Market Intelligence department where he was involved in the<br />
strategic development of the Company. Since August 2006, Marco Circelli was Head of Inves<strong>to</strong>r<br />
Relations at Converium. Prior <strong>to</strong> Converium Marco had worked in <strong>Management</strong> Accounting for<br />
Unilever and in various functions at UBS and Credit Suisse. He brings along more than 12 years of<br />
experience in Finance (of which almost 7 years in reinsurance), and holds a business degree from the<br />
University of Applied Science of Central Switzerland, and a MBA in Financial Services and Insurance<br />
from University of St. Gallen (Switzerland).<br />
JANICE COWLEY<br />
Program Direc<strong>to</strong>r of CoCPIT, SCOR<br />
Janice Cowley is Program Direc<strong>to</strong>r of CoCPIT – a global program focused on capital and <strong>risk</strong> management.<br />
Reporting <strong>to</strong> the Group CRO, she is responsible for leading large change projects for<br />
SCOR. Janice has more than 10 years experience in leading roles in both reinsurance and investment<br />
banking. Janice has a degree in Business from Vic<strong>to</strong>ria University of New Zealand and has completed<br />
post graduate studies and Harvard Business School, Bos<strong>to</strong>n.
MICHEL DACOROGNA<br />
Deputy Chief <strong>Risk</strong> Officer, SCOR<br />
Michel Dacorogna is a member of senior management of SCOR SE and heading the Group Financial<br />
Analysis and <strong>Risk</strong> <strong>Management</strong> modeling team. His main responsibilities are <strong>to</strong> develop the Asset<br />
and Liability <strong>Management</strong> models for the Group and on this basis <strong>to</strong> assess the <strong>risk</strong> <strong>based</strong> capital<br />
of the company and determine the best strategic asset allocation.<br />
The co-author of: “An Introduction <strong>to</strong> High Frequency Finance”, he has also published numerous<br />
articles in scientific journals. He is an associate edi<strong>to</strong>r of Quantitative Finance.<br />
He received his Habilitation, Ph. D. and M. Sc. In Physics from the University of Geneva in Switzerland.<br />
ERIC DAL MORO<br />
Chief Reserving Actuary, SCOR Switzerland<br />
Eric Dal Moro is an engineer of the École des Mines and graduated as an actuary from the French<br />
Institute of Actuaries. He is also a member of the Swiss Institute of Actuaries.<br />
He worked at AXA Direct Japan as Technical Direc<strong>to</strong>r (1998 -2001). Then, he joined AXA Assicurazioni<br />
in Italy as Senior Actuary (2001-2002). He also worked for Ernst &Young in Paris and Zurich as<br />
Senior Manager (2002-2007). He is currently Chief Reserving Actuary and Appointed Actuary of<br />
SCOR Switzerland AG.<br />
MAGDALENA KLAPPER-RYBICKA<br />
<strong>Risk</strong> Consultant, SCOR<br />
Magdalena Klapper-Rybicka is a <strong>Risk</strong> Consultant in the Group Financial and <strong>Risk</strong> Modeling team<br />
of SCOR. Over the past two years, Magdalena contributed significantly <strong>to</strong> the development and<br />
implementation of an integrated quantitative <strong>risk</strong> management system for the SCOR Group.<br />
Magdalena Klapper-Rybicka studied computer science at AGH University of Science and Technology<br />
in Krakow, Poland. During her time at university, she worked as a scientist and on international<br />
research projects. She holds a PhD from the Department of Computer Science with a major in<br />
artificial intelligence.<br />
MICHÈLE LACROIX<br />
Chief Investment Officer, SCOR<br />
Michèle Lacroix joined SCOR in September 2008 as Chief Investment Officer of SCOR’s newly created<br />
asset management company, SCOR Global Investments. A graduate of HEC, she spent several years<br />
on the trading floor at the Caisse Centrale des Banques Populaires, first in charge of government<br />
bond activity and then covering all fixed income and hedging strategies. After 4 years of brokerage<br />
on fixed income she moved <strong>to</strong> Legal & General Asset <strong>Management</strong> (France) as a Portfolio Manager,<br />
managing the EURO fund and heading the development of new mutual fund strategies, before<br />
becoming Chairman and Chief Executive Officer of the company.<br />
SCOR - November 2010 - 103
104 - November 2010 - SCOR<br />
TONY NEGHAIWI<br />
Chief Pricing Actuary, SCOR Global P&C<br />
Tony Neghaiwi joined Converium as Chief Actuary in 2006 and became in 2008 Chief Pricing Actuary<br />
and member of senior management of SCOR Global P&C with responsibilities of all P&C pricing<br />
and modeling operations. He has more than 23 years of experience in the insurance/reinsurance<br />
industry. He worked notably for Aetna Life & Casualty and Liberty Mutual in the United States, ZFS<br />
and XL in Switzerland. He held a Masters of Science in Applied Mathematics from University of<br />
Tennessee and completed post graduate studies in Mathematics from University of Hous<strong>to</strong>n and is<br />
a Fellow of the Casualty Actuarial Society.<br />
WAYNE RATCLIFFE<br />
Direc<strong>to</strong>r Group <strong>Risk</strong> <strong>Management</strong>, SCOR<br />
Wayne Ratcliffe is a Fellow of the Institute of Actuaries (UK) and graduate of the University of<br />
Cambridge, where he studied mathematics and operations research. He began his insurance career<br />
in 1981 in Life & Pensions (Equity & Law then Prudential), as a Product Development and Marketing<br />
Actuary. In 1994 he set up the London office of SCOR Vie before moving <strong>to</strong> the Actuarial department<br />
in Paris in 1998 as Head of Anglo-Saxon markets and subsequently as Deputy Actuarial Direc<strong>to</strong>r. After<br />
a brief spell at XL Re in 2006 he returned <strong>to</strong> SCOR in 2007 as Head of Group <strong>Risk</strong> <strong>Management</strong>.<br />
EVA SCHLÄPFER DE MONTMOLLIN<br />
Senior <strong>Risk</strong> Consultant, SCOR<br />
Eva Schläpfer de Montmollin is a Senior <strong>Risk</strong> Consultant in the Group Financial and <strong>Risk</strong> Modeling<br />
team at SCOR. For two years, she has been heavily involved in the development and implementation<br />
of an integrated quantitative <strong>risk</strong> management system at SCOR Group (CoCPIT).<br />
Prior <strong>to</strong> joining SCOR she held various actuarial roles at Zurich Financial Services where she has<br />
been leading a team of actuaries developing pricing methodologies and implementing a widely<br />
used pricing platform. She holds a PhD and a MSc in Mathematics from the University of Fribourg<br />
in Switzerland. She is a fully qualified actuary (SAV, DAV).
Paper copies limited <strong>to</strong> consultation.<br />
Distribution in electronic format only.<br />
To request an electronic copy,<br />
please email: scorglobalp&c@scor.com<br />
DEJAGLMC Imprimeur<br />
All rights reserved. No part of this publication may be reproduced<br />
in any form without the prior permission of the publisher.<br />
SCOR has made all reasonable efforts <strong>to</strong> ensure that information<br />
provided through its publications is accurate at the time of inclusion<br />
and accepts no liability for inaccuracies or omissions.
Edi<strong>to</strong>r: Strategy & Development<br />
SCOR Global P&C<br />
scorglobalp&c@scor.com<br />
No. ISSN 1638-3133<br />
November 2010<br />
SCOR<br />
1, avenue du Général De Gaulle<br />
92074 Paris-La-Défense CEDEX – France<br />
www.scor.com<br />
November 2010 – Design: – Pho<strong>to</strong> credits : Corbis, Getty Images, Inspoon, Shutters<strong>to</strong>ck, Thinks<strong>to</strong>ck.