Risk Appetite: Linkage with Strategic Planning

Sponsored by
CAS/CIA/SOA
Joint Risk Management Section

Prepared by
Kailan Shang and Zhen Chen
March 2012

©2012 Society of Actuaries, All Rights Reserved

The opinions expressed and conclusions reached by the authors are their own and do not represent any official
position or opinion of the sponsoring organizations. The sponsoring organizations make no representation or
warranty to the accuracy of the information.
 Acknowledgements

The authors would like to thank the Project Oversight Group (POG) members for their
guidance, reviews, comments and full support throughout this project. This paper would not
have its current level of comprehensiveness and sophistication without the POG's insightful
input. The authors are grateful for funding sponsored by the Joint Risk Management Section
of the Society of Actuaries, the Casualty Actuarial Society and the Canadian Institute of
Actuaries.

“Risk Appetite: Linkage with Strategic Planning” POG members are:

• Fred Tavan, Chair
• Gaetano Geretto
• Glenn Meyers
• Howard Rosen
• Jim Reiskytl
• Mary Neumann
• Michael Kula
• Paul Anderton
• Robert Wolf
• Sivakumar Desai
• Steven Siegel
• Sylvain Fortier
• Tao Guo
• Trevor Pollitt
• Vince Granieri

The authors also would like to thank Barbara Scott for her efficient coordination of this
project.

©2012 Society of Actuaries, All Rights Reserved Page 2 of 57

 Executive Summary
With recent advances in computational capability, progressive and complex models have been
created to assist risk management. At the same time, regulators, rating agencies, and
investment markets have put more attention on risk controls of financial institutes. Those
factors, however, did not prevent financial institutions from going bankrupt or needing a
government bailout in the past few years. Most of the defaulted financial firms stated they had
good risk management policies and implementation and the resources spent on risk
management were not trivial.

The firms failed for different reasons but clearly their “good risk management” went wrong.
Without a clear understanding of the risks of their business, decision makers adopted risky
strategic plans. Advanced risk management models and stress testing may help with assessing
the quantitative impact of risks, but understanding the company’s core competence and the
risks before taking them is far more important.

It is critical to have clear answers to the following questions before making decisions:
• What is the company’s competence in the market?
• Are the decision makers familiar with the risks involved including the tail risks and
understand their potential impact?
• Is the company capable of surviving extreme events?

Risk appetite articulates the level of risk a company is prepared to accept to achieve its
strategic objectives. Risk appetite frameworks help management understand a company’s risk
profile, find an optimal balance between risk and return, and nurture a healthy risk culture in
the organization. It explains the risk tolerance of the company both qualitatively and
quantitatively. Qualitative measures specify major business strategies and business goals that
set up the direction of the business and outline favorable risks. Quantitative measures provide
concrete levels of risk tolerance and risk limits, critical in implementing effective risk
management.

Risk appetite represents the willingness and the ability to take risk. Due to the sophisticated
nature of financial institutions, it requires a lot of effort to fully understand the constraints and
the ability to assume risk. Therefore, it is a gradual cognitive process to be updated from time
to time if you enter new markets, new business and have new understanding of risks assumed.

With the help of several case studies, this paper tries to explain the role of a risk appetite
framework in effective risk management and how it can be used to make more informed
strategic decisions. The paper does not seek a golden rule for risk appetite framework and the
integration of risk appetite and strategic decision making. Assumptions validation, parameter
calibration and model risks are not the focus of this paper either.

It is hoped this paper will formalize the thinking process of assessing the risks of strategic
plans and reflecting them in the risk appetite framework. It is an integrated interactive process.

©2012 Society of Actuaries, All Rights Reserved Page 3 of 57

 Contents
1. Introduction ................................................................................................................................... 5
2. Risk Appetite, Risk Tolerance and Risk Limit Setting .................................................................. 6
2.1 Literature Review ................................................................................................................ 6
2.2 Risk Appetite Framework ................................................................................................. 10
2.3 Risk Governance and Risk Appetite ................................................................................. 16
2.4 Constraints ........................................................................................................................ 16
2.5 Risk Appetite for Each Risk Category and Risk Limit ..................................................... 18
3 Risk Appetite and Asset Allocation .............................................................................................. 25
4 Risk Appetite and New Business Budgeting ................................................................................ 29
5 Risk Appetite and Capital Allocation ........................................................................................... 34
6 Risk Appetite and Liquidity Management ................................................................................... 37
7 Risk Appetite and Performance Measurement ............................................................................. 44
8 Conclusion ................................................................................................................................... 50
9 References .................................................................................................................................... 52

©2012 Society of Actuaries, All Rights Reserved Page 4 of 57

 1. Introduction

Risk appetite has been playing a more important role in systematically determining an
institution’s level of risk tolerance and influencing its strategic planning. Similar to the risk
objective in an investment policy statement for individual investors, risk appetite is used in
different areas of strategic planning to better facilitate the analysis of risk versus return and
improve decision making. With both quantitative and qualitative analysis of the company’s
risk profile, risk appetite provides a holistic picture for investors and senior management.

This research paper explores the current practices of risk appetite framework in insurance
companies. It also discusses the reasons and the ways risk appetite should be embedded in
strategic planning, including asset allocation, new business budgeting, capital allocation,
liquidity management and performance measurement. Sections 2 to 7 cover the following:

• Section 2: Risk Appetite, Risk Tolerance and Risk Limit Setting

The principal, process and major components of risk appetite, risk tolerance and risk
limit setting are discussed. It focuses not only on the quantitative aspects such as capital
at risk (CaR), earnings at risk (EaR) and economic capital, but also on several qualitative
components such as target credit rating and reputation risk. Constraints from local
regulators, group regulators, rating agencies and investors also are critical components in
strategic planning. How those constraints are reflected and the corresponding
implications also are discussed. The full picture should at least include the following
types of risk: interest rate, credit (both credit spread risk and credit impairment), equity,
foreign exchange, concentration, asset liability mismatch, liquidity, insurance,
catastrophe and operation.

• Section 3: Risk Appetite and Asset Allocation

After risk appetite is set, it should be embedded in the investment decision. How risk
appetite can be embedded in asset allocation with regards to capital adequacy and
earning volatility are discussed in this section.

• Section 4: Risk Appetite and New Business Budgeting

The mix of new business is an important determinant in the strategic landscape of
insurance companies. What kind of business mix can better fit the requirement of risk
tolerance and also maximize the risk-adjusted return on capital (RAROC) is an important
factor in informed business decisions. In this section, both the strategic and technical
sides of the implementation will be demonstrated, with consideration of the company’s
CaR and EaR.

• Section 5: Risk Appetite and Capital Allocation

Capital allocation takes into account the link between target business mix and its impact
on risk profile. Target business mix incorporates existing business and planned new

©2012 Society of Actuaries, All Rights Reserved Page 5 of 57

 business sales. This section addresses some major considerations when aligning capital
allocation with risk appetite.

• Section 6: Risk Appetite and Liquidity Management

Liquidity has become more important, especially under stress situations such as the
financial crisis. This also is where the appropriate risk appetite plays a critical role in
protecting capital adequacy. The tolerance level of liquidity risk can be determined
considering both the capital market situation and the company’s idiosyncratic situation,
by looking backward and forward. This naturally leads to a consistent and
comprehensive framework that relies on the characteristics of asset and liability profiles.

• Section 7: Risk Appetite and Performance Measurement

Performance measurement also is a critical component in a company’s strategy as only
effective and fair performance measurement can incentivize managers to achieve the
return and risk objectives. An example in which performance measurement is linked to
risk-adjusted measures consistent with a risk appetite statement is presented.

Section 7 is followed by the conclusion.

2. Risk Appetite, Risk Tolerance and Risk Limit Setting

2.1 Literature Review

2.1.1 Risk Appetite Definition

Risk appetite, or risk preference, has long been used when discussing utility functions,
measures of relative satisfaction in economics. It is often discussed in terms of the investment
decision. In the area of risk management in the insurance community, risk appetite means a
high-level view of the risks the insurer is willing to accept in pursuit of value.

Several sources, such as regulators, consultants, academics and professional associations,

provide a definition of risk appetite. Some definitions are given below:

The total risk exposure an organization will undertake to achieve its objectives.1
The amount of risk, on a broad level, an organisation is willing to take on in pursuit of value.
Or, in other words, the total impact of risk an organization is prepared to accept in the pursuit
of its strategic objectives.2
Often taken to be the maximum amount of risk a company is willing to accept in pursuit of
its mission/objectives/plans. 3

1
Chase-Jenkins, “Risk Appetite,” 12.
2
KPMG, “Understanding and Articulating,” 3.
3
Chaplin, “Risk Appetites,” 3.

©2012 Society of Actuaries, All Rights Reserved Page 6 of 57

 Risk appetite framework has the following focuses. These appeared in articles published by
consulting firms and rating agencies.

• Protecting and creating value for the business 4 . Risk appetite describes risk with
quantitative measures and facilitates the analysis of the risk/return trade-off. This helps
senior management make informed decisions to maximize the risk-adjusted return for the
shareholder.
• Ensuring the consistency between risk appetite and risk limits 5. Both rating agencies and
investors are concerned about whether risk appetite is properly aligned with the risk limits
being set for business operations. The financial crisis clearly proved that some companies
failed to do so. Lots of government bailouts and large-scale capital raising from the
market took place to keep the companies solvent. These actions imply that the risk taken
exceeded the risk appetite for some companies.
• Integrating into business strategy and corporate culture 6. Risk appetite acts as a guideline
for risk-taking activities. Keeping risk appetite in mind for business decisions and
operations facilitates risk identification and monitoring.

We cannot find any academic papers that have in-depth discussion of risk appetite.

2.1.2 Rating Agency’s View

Rating agencies have put more focus on assessing risk management practices when evaluating
credit ratings. Other motivations such as using a more refined risk-based capital model and/or
economic capital model when assessing capital adequacy also may promote a better
understanding of the risks of the business. Those together encourage senior management to
make more efforts to enhance their understanding and management of assumed risk and
potential risk.

Standard & Poor’s clearly states that risk appetite is an essential component for an enterprise
risk management (ERM) score of strong or above: “We expect insurers with strong ERM
frameworks to have a well-defined risk appetite framework that supports the effective
selection of risks, establishing the risks that the insurer wishes to acquire, avoid, retain and/or
remove.” 7

S&P published the methodology used for assessing an insurer’s risk appetite framework in
2006. In 2010 after the financial crisis, a more refined and detailed assessment methodology
was published as S&P found that “the financial crisis exposed a number of weaknesses in
insurers' risk appetite frameworks with some insurers having been quite active in seeking

4
Deloitte & Touche LLP. “Global Risk,” 5
5
Bowser et al., “Refined Methodology,” 9.
6
Barnes, “Evaluating Risk Appetite,” 3, Deloitte & Touche LLP. “Global Risk,” 10-12, and KPMG, “Understanding
and Articulating,” 9.
7
Bowser et al., “Refined Methodology,” 3.

©2012 Society of Actuaries, All Rights Reserved Page 7 of 57

 risks that they may not have fully understood and could not manage within their specified risk
tolerances. While most insurers have set risk limits, many insurers found, in hindsight, that
these did not adequately reflect their risk appetite or were poorly understood.”8

S&P also emphasized the importance of the integration of risk appetite into a company’s
strategy and culture: “Standard & Poor's examines whether an enterprise has a clearly
articulated risk appetite process and the degree to which this process is integrated with its
strategy and culture.” 9

Clear and formal linkage between risk appetite and strategic planning is favourable in ERM
ratings/credit ratings, which is one of the motivations of many ERM projects.

2.1.3 Risk Appetite in Solvency II Framework

The European Union’s Solvency II framework directive has clearly emphasized the
integration of risk management and decision making:

That risk-management system shall be effective and well integrated into the
organisational structure and in the decision-making processes of the insurance or
reinsurance undertaking with proper consideration of the persons who effectively run the
undertaking or have other key functions.10

It links good risk management practices with meeting capital requirements to survive a
1-in-200-year event:

In order to promote good risk management and align regulatory capital requirements with
industry practices, the Solvency Capital Requirement should be determined as the
economic capital to be held by insurance and reinsurance undertakings in order to ensure
that ruin occurs no more often than once in every 200 cases or, alternatively, that those
undertakings will still be in a position, with a probability of at least 99.5%, to meet
their obligations to policy holders and beneficiaries over the following 12 months. 11

In a Committee of European Insurance and Occupational Pensions Supervisors (CEIOPS)

consultation paper, it is explained that an effective risk management system includes:

• A clearly defined and well documented risk management strategy that includes the
objectives, key principles, risk appetite and assignment of responsibilities across all
the activities of the undertaking and is consistent with the undertaking’s overall
business strategy;
• Adequate written policies that include a definition and categorisation of the risks
8
Ibid., 2.
9
Barnes, “Evaluating Risk Appetite,” 3.
10
European Union, “Directive of the European Parliament,” 158.
11
Ibid., 24. Emphasis is the authors’ own.

©2012 Society of Actuaries, All Rights Reserved Page 8 of 57

 faced by the undertaking, by type, and the levels of acceptable risk limits for each risk
type, implement the undertaking’s risk strategy, facilitate control mechanisms and
take into account the nature, scope and time horizon of the business.12

It covers all aspects of risk appetite framework: enterprise risk tolerance, and appetite for each
risk category and risk limit, although the terms used are not exactly the same.

2.1.4 Current Practices of Insurance Companies

Some survey reports 13 about risk management practices and risk appetite produced by
consulting firms are available to gain a high-level view of how risk appetite was implemented
in financial institutions. The landscape may have improved since the surveys were done as the
recent credit crisis increased the awareness of the importance of risk management.

A survey conducted by Towers Perrin in 2009 14 indicated that although 84 percent of

respondents expected to have a documented risk appetite statement within 12 months of the
survey, only 40 percent demonstrated consistency between risk appetite and risk limits.

Deloitte’s fifth Global Risk Management Survey 15 covering 130 financial institutions in 2007
reveals a similar result. Sixteen percent of respondents did not have a statement of the firm’s
risk appetite while others had one that was either approved or informal/not approved by the
board of directors. The survey also shows that the integration between risk management and
other management initiatives was low. Only 42 percent of executives indicated that ERM
programs were well integrated with strategic planning.

Twenty-two insurance companies, including five reinsurance companies, 16 were selected to

explore risk appetite and its integration with strategic planning with information from their
2010 annual reports.

Some key findings are listed below:

• Fifteen out of 22 companies (68 percent) clearly mentioned risk appetite framework as
part of the risk management sections in their 2010 reports.
• For companies that did not mention risk appetite in their annual report, capital adequacy
was stated as its important management goal. For companies that mentioned risk appetite,
both capital adequacy and earnings volatility were important considerations in risk
12
CEIOPS, “CEIOPS’ Advice,” 17-18.
13
Chase-Jenkins, “Risk Appetite,” and Deloitte, “Global Risk Management.”
14
Chase-Jenkins, “Risk Appetite,” 5&9.
15
Deloitte, “Global Risk Management,” 12.
16
The companies selected include ACE Ltd., Admiral Group PLC, American International Group Inc. (AIG), Allianz
Group, Allstate Corp., Aviva PLC, AXA Group, Berkshire Hathaway Inc., General Reinsurance AG (Gen Re),
HDI-Gerling (2009), ING Group, Lloyd's, MetLife Inc., Manulife Financial, Munich Re, PartnerRe, Prudential PLC,
Prudential Financial Inc., Standard Life PLC, Sun Life Financial Inc., Swiss Re and Zurich Financial Services Group.

©2012 Society of Actuaries, All Rights Reserved Page 9 of 57

 management.
• Fourteen out of 15 companies that mentioned risk appetite framework stated that risk
appetite played a role in strategic planning/business decisions.
o Three companies mentioned that risk appetite was considered when deciding
asset allocation.
o Ten companies mentioned that risk appetite was considered to maximize
risk-adjusted return.
o Fifteen companies mentioned that risk appetite was considered in capital
allocation.
• All 15 companies that mentioned risk appetite framework claimed to have risk limit and
risk-monitoring processes consistent with risk appetite.

Though details about how risk appetite is integrated with strategic planning are unavailable
due to confidentiality concerns, there is a clear trend that companies with risk appetite
frameworks not only focus on protecting value but also on creating value from risk appetite.

2.2 Risk Appetite Framework

Traditional measures used by insurance companies to manage business are numerous.

Examples include:
• New business volume/market share
• Persistency rate
• Embedded value of existing and new businesses
• Appraisal value
• Operating profit

With these measures, and a conservative approach to reserving capital, sometimes the
outcome could also be too optimistic. In reality, tail events do occur. An appropriate risk
appetite framework is more concerned with the impact of tail events on the company’s
financial condition. With stochastic analysis, stress testing and correlation analysis, the
distribution and range of possible outcomes may be discovered. These are very useful tools
for senior management’s strategic planning.

Risk appetite framework

A fundamental step in a risk appetite framework is to define the organization’s willingness
and ability to take risk based on the input from the board of directors and senior executives.
Just as an individual investor must evaluate his or her risk appetite or risk objective when
managing personal portfolios and setting reasonable return requirements, an institution needs
to look at its unique situation to define its risk appetite.

Risk appetite is a high-level view of the risks the insurer is willing to assume in pursuit of
value. When insurers define the optimal level of risk, the ultimate priority is to serve
shareholders’ benefits. Before setting up the risk appetite, it is essential to have a clear picture
of the market and the company’s core competency. This will facilitate the decision on the type

©2012 Society of Actuaries, All Rights Reserved Page 10 of 57

 and magnitude of risk to be taken consistent with business strategies and market situation. At
the same time, the desired risk profile shall satisfy the constraints set by other parties such as
regulators, rating agencies, policyholders, debt holders, senior management and employees.

Some external changes have expedited the process of setting risk appetite. S&P has required a
clear statement of risk appetite as a foundation of “strong” or “excellent” ERM rating.
Solvency II also requires insurers to explicitly consider their risk appetite.

Risk appetite framework normally includes three increasingly detailed levels:

• Enterprise risk tolerance: The aggregate amount of risk the company is willing to take,
expressed in terms of
o Capital adequacy
o Earnings volatility
o Credit rating target
It represents the company’s long-term target and shall be revised only if there are
fundamental changes to the company’s financial profile, market situation and strategic
objective. Risk appetite helps prevent default by preserving capital position. This is
required by regulators, rating agencies, policyholders and debtholders. These stakeholders
show little or no interest in the upside of risk taking. On the other hand, shareholders are
interested in the upside resulting from risk taking and low earnings volatility.
• Risk appetite for each risk category: Enterprise risk tolerance needs to be allocated to
risk appetite for specific risk categories and business activities, such as selling life
insurance policies or underwriting property and casualty risks, or taking more market risk
versus credit risk. By doing this, the company’s resources, such as capital, can be
allocated to the areas in which the company feels comfortable or has competitive
advantage. When determining or updating risk appetite for different risk categories, the
platform, analytic tools and conclusions also are instrumental in identifying risk-taking
activities that have relatively high risk-adjusted return.
• Risk limit: Risk limits are the most granular level used for business operation. It
translates enterprise risk tolerance and risk appetite for each risk category into
risk-monitoring measures. The consistency between risk limit and enterprise risk
tolerance helps the company realize its risk objective and maximize risk-adjusted return.

Risk appetite includes both qualitative and quantitative measures. The following risk metrics
are normally considered in setting risk appetite given a specified time horizon:

Quantitative measure
• Capital/equity at risk: Used under local statutory/rating agencies/economic basis. It can
be stated as the maximum acceptable capital at risk would be X percent of current
available capital with a certain probability. Alternatively, with a probability of Y percent,
the company can still stand solvent.

©2012 Society of Actuaries, All Rights Reserved Page 11 of 57

 • Earnings at risk: Used under local accounting/U.S. Generally Accepted Accounting
Principles (U.S. GAAP)/International Financial Reporting Standards (IFRS) basis. It can
be stated as the maximum acceptable earning at risk would be X percent of
expected/target earning with a certain probability. Alternatively, with a probability of Y
percent, the earnings will be non-negative.
• Embedded value/market consistent embedded value (EV/MCEV): It can be stated as
the maximum acceptable decrease in EV/MCEV would be X percent of expected/target
EV/MCEV with a certain probability.

The examples in the above quantitative measures are value at risk (VaR). A more conservative
measure is tail value at risk (TVaR), aka conditional tail expectation (CTE). TVaR takes the
average of all possible values above VaR. For a variable with high skewness, TVaR is a more
conservative and appropriate choice.

Qualitative measure
• Credit ratings: Lowest desired financial strength rating or debt rating.
• Risk preferences: Certain risks that an insurer does not want to take, such as not
underwriting business in catastrophe prone regions.
• Franchise value: How much can be decreased due to adverse publicity, poor reputation
or regulatory intervention.

There are several issues that need careful consideration when implementing a risk appetite
framework. Although those limitations are kept in mind, this paper focuses on the philosophy
and approaches of embedding risk appetite in corporate strategic planning. The following
topics are critical but out of the scope of this paper:
• Assumptions setting and validation such as correlation assumption in tail events and
probability assigned to certain tail events.
• Model risk that the model does not reflect the real market dynamics due to incorrect
model specification, assumptions calibration errors and implementation errors, etc.

Enterprise risk tolerance statement

Below is an example of an enterprise risk tolerance statement. Examples of risk appetite for
each risk category will be given in section 2.5.
• Credit rating
o Financial strength rating: Achieve and maintain an AA financial strength rating
on a global basis (S&P rating) and maintain a buffer of more than 50 percent of one
year’s net income above the minimum capital requirement for the AA rating.
o Debt rating: Achieve and maintain an A rating for senior unsecured debt.
• Earnings at risk: The probability of negative IFRS/U.S. GAAP earnings for one year is
less than 5 percent.

©2012 Society of Actuaries, All Rights Reserved Page 12 of 57

 • Capital at risk: The probability of a 15 percent loss of IFRS/ U.S. GAAP equity for one
year is less than 0.5 percent. The probability of a capital deficiency under the economic
framework is 0.05 percent according to the target financial strength rating.
• The company has extremely high risk aversion to reputation risk.

Figure 2.2.1 Illustrative Quantitative Risk Appetite Statement

Measures  Earnings at Risk  Capital at Risk  Economic Capital

Confidence  99.95% (based on target

 95% 99.5% financial strength rating)
interval

Stressed metric  IFRS/GAAP earnings IFRS/GAAP equity Available economic capital

Potential 15% available capital Potential available economic
Potential negative IFRS earnings
reduction of net value during a capital < required economic
Interpretation during a ‘moderate’ stress
‘severe’ stress scenario (i.e. 1 in capital during an ‘extreme’ stress
scenario (i.e. 1 in 20 years)
200 years) scenario (i.e. 1 in 2000 years)

Risk appetite and business analysis

With risk appetite in place, the risk/return profile and the constraints of the business are
quantified, which helps prevent exceeding enterprise risk tolerance when making strategic
decisions. Below is a graph that shows the use of capital at risk in finding the optimal
risk/return trade-off.

Figure 2.2.2 Risk Return Analysis Based on Capital at Risk

Return on Risk/Return Profile

Capital
15%

Utility function with

return and risk
13%

11% BU1
Desired risk/return
BU3 BU2 tradeoff: 13%/10%
9%

Risk Appetite: Capital

7% can not lose more
Minimum required than 15% in a 1-in-
return: 6% in a 1-in- 200-year event
20-year event
5%

3% Capital
at risk
6% 8% 10% 12% 14% 16% 18% 20% 22% 24%

©2012 Society of Actuaries, All Rights Reserved Page 13 of 57

 • Capital at risk (CaR) is used to describe the risk. It is a holistic risk measure of capital
adequacy.
• Return on capital (ROC) is chosen as the return measure as it is something investors will
earn.
• The utility function can be considered as nondifferential curves, each of which represents
a certain risk-adjusted return. It represents the investor’s degree of risk aversion. In the
example, the risk-adjusted return or utility of business unit (BU) 1 is greater than those of
BU2.
• Target ROC is 10 percent.
• A minimum ROC of 6 percent represents the risk appetite for earnings volatilities: The
company does not want to see earnings drop by more than 40 percent in a 1-in-20-year
event.
• A maximum CaR of 15 percent of capital represents the risk appetite of capital adequacy:
The company does not want to lose more than 15 percent of its capital in a 1-in-200-year
event.
• An efficient frontier is built to represent the risk-return opportunities through
stochastic/scenario analysis.

The most desirable position would be 10 percent ROC with 13 percent CaR. It is within all
the constraints and maximizes the risk-adjusted return. In this example, business units 1 to 3
are operating with an expected return above the minimum requirement and CaR below the
risk tolerance. BU1 and BU3 lay on the efficient frontier while BU2 still has some room to
improve its risk/return profile. If the market is segmented and BU2 is in a market with a lower
market risk premium, the company’s capital allocation and new business plan might consider
capital reallocation from BU2 to other more efficient BUs. This is a simple example of how
risk appetite framework interacts with strategic planning.

Percentiles in risk measures

Appropriate percentages adopted in different quantitative risk measures are important to
different objectives. There are several considerations when determining the percentile:

• The higher the target credit rating, the higher the percentage used in risk measures. Some
companies refer to default experience data published by rating agencies to determine the
appropriate percentage to use, especially in capital management, or calculating CaR.

• Different risk measures may use different percentiles. For example, TVaR, aka CTE, is
more conservative than VaR. TVaR may look at a 1-in-40-year event instead of
1-in-200-year event as specified in Solvency II. EaR and CaR also may share different
percentages, as during extreme events caused by systematic risk, the investors are more
concerned about the solvency than the earnings. However, in the business decision
process under normal circumstances, the same percentages might be chosen for both EaR
and CaR.

• Tail events may change from time to time. For example, in the nonlife

©2012 Society of Actuaries, All Rights Reserved Page 14 of 57

 insurance/reinsurance industry, return periods 17 are very likely to change after certain big
events. The industry and catastrophe modellers put more load on U.S. hurricanes after
Katrina in 2005. This implies that even if the percentage does not change in each risk
measure, the severity of the tail events needs to be updated based on new information.

• Companies with different types of business usually view key tail events with different
probabilities. For example, a reinsurer specialising in the U.S. energy insurance business
is more concerned with a hurricane in the Gulf of Mexico, which may have an assigned
probability of 0.2 percent, or a 1-in-500-year event. Meanwhile, pension business writers
are much more exposed to longevity risk. A tail event of 5 percent mortality improvement
expected with a probability of 0.5 percent, or a 1-in-200-year event, could be more
appropriate for the pension industry.

• Sometimes the percentage is not important but the deterministic stress events, either
based on historical experience or hypothesized events, are critical. There are always some
assumptions to be made while assigning a probability of occurrence to certain events.
Sometimes it is hard to reach agreement on those event-related probabilities. In those
cases, designing strategies to remain solvent when those stress events happen are more
meaningful than trying to allocate a probability to those events. This is common in
catastrophe and pandemic risk management.

Backtesting
All the quantitative risk measures in a risk appetite framework are very sensitive to economic
assumptions and are subject to model risk. A difficulty normally encountered when using
those risk measures is validation. A 1-in-200-year event is rare and may not be experienced
during a lifetime. Therefore, it is hard to demonstrate the credibility of the CaR or EaR
against reality. However, without a proper process to test those measures against experience,
those quantitative measures are likely to be misunderstood, misused or even manipulated.
This would totally be against the objective of risk appetite framework. A possible solution to
overcome this is to backtest the measures against reality at a less severe confidence level. For
example, use 80 percent CaR and test it against the experience in the past five years.
Alternatively, use 99.5 percent CaR and test it against N × 1-in-10-year event; N can be based
on the relationship between a 1-in-10-year event and a 1-in-200-year event implied by
historical data or forward-looking assumptions.

People may doubt backtesting as forward-looking views evolve quickly, especially in a highly
uncertain world with more and more emerging risks. This is true and also a major limitation
of backtesting. On the other hand, history repeats. Backtesting helps us learn from history and
adjust our view of the future. In the past decade, examples of extreme events affecting
companies' risk appetite and strategic decision are numerous.

• After recent earthquakes (2010 in Chile, Haiti and New Zealand, and 2011 in Japan),
RMS Inc., a market leader in catastrophe (CAT) modelling, backtested the catastrophe
17
A 200-year return period means a 1-in-200-year event.

©2012 Society of Actuaries, All Rights Reserved Page 15 of 57

 model and released a new version with updated natural catastrophe model parameters.
This will have an impact on the whole industry regarding the capital requirement for CAT
business and may reduce the number of market players.

• The variable annuity (VA) market in Japan felt the squeeze after the 2008 financial crisis.
There had been tremendous growth since 2002. The market share of VA business
decreased significantly due to the increasing awareness of the material downside risks for
insurance companies.

• After the events of Sept. 11, 2001, many global insurers and reinsurers invested much
more capital in aviation businesses. The hard market lasted almost five years.

From the examples above, it is obvious that backtesting of key assumptions could improve
corporate strategic decisions such as efficient capital allocation to business lines, the timing of
exit plans and rate adjustment.

2.3 Risk Governance and Risk Appetite

A clear risk appetite statement sets the guideline from the top for risk management. Normally
it needs to be approved by the board and included in company’s risk policies. Setting up a risk
appetite framework is not a trivial task. It usually requires the following steps:

1) Bottom-up analysis of the company's current risk profile.

2) Interviews with the board of directors regarding the level of risk tolerance.
3) Alignment of risk appetite with the company's goal and strategy.
4) Formalization of the risk appetite statement with approval from the board of directors.
5) Establishment of risk policies, risk limit and risk-monitoring processes consistent with
risk appetite.
6) Design and implementation of the risk-mitigation plan to be consistent with risk appetite.
7) Communication with local senior management for their buy in.

With risk appetite and risk limits in place, all the risk management activities have the same
ultimate goal: ensuring risk-taking activities are consistent with risk appetite. A risk limit
system provides measures for monitoring the risk position and not exceeding the level of risk
tolerance.

Risk appetite needs to be revisited when there is a large change in business strategy, market
situation or financial condition. Normally there is a yearly review with the board of directors
to determine if any changes need to be made.

2.4 Constraints

With a clearly defined risk appetite, constraints are known when making strategic decisions. It

©2012 Society of Actuaries, All Rights Reserved Page 16 of 57

 shows the line that cannot be crossed to remain consistent with risk tolerance.

A global company may face the following constraints:

• Local statutory requirements: capital adequacy
• Group solvency requirements: capital adequacy
• Rating agency capital requirements
• Economic capital sufficiency
• IFRS/U.S. GAAP earning volatility
• Embedded value earning volatility

Figure 2.4.1 below shows the interaction between the constraints defined in risk appetite and
strategic planning.

Figure 2.4.1 Risk Appetite and Strategic Planning

ROC > 10%

Increase/decrease
equity allocation
IFRS 90% EaR <
Credit rating
Budgeted earning Increase
ERM A+ or higher
investment

Current risk profile

Hedge rho of
VA1 business

Group solvency: Add MVA3 to

2
CAR > 160% Increase P&C4 pass through
market share investment Liquidity
risk
risk limit

Local solvency:
2
CAR > 150%

Notes:
1
VA: Variable annuity
2
CAR: Capital adequacy ratio, normally calculated as available capital/required capital
3
MVA: Market value adjustment
4
P&C: Property and casualty insurance

The green rectangle is a whole set of possible strategic plans. The blue hexagon illustrates the
space where the strategic plan will keep the company within the risk tolerance limit. Its
borders, the red lines, represent the components of the risk appetite framework. It could be
enterprise risk tolerance or risk limit. The white pool in the middle is an example of the
current risk profile. Starting from here, we can implement different strategic plans to reach the
maximum risk-adjusted return or other strategic goals in the blue hexagon. As an example,

©2012 Society of Actuaries, All Rights Reserved Page 17 of 57

 there are five strategies considered to improve the risk profile and create value:

1) Increase P&C market share: Business expansion requires capital expenditures, which will
reduce the capital adequacy ratio if there is no capital-raising activity planned.
2) Add MVA to pass through investment risk: Adjusting product features to transfer risk to
policyholders is a risk-mitigation technique. In this case, liquidity risk caused by mass
lapse behaviour due to either credit rating downgrade or disintermediation risk is reduced.
3) Increase ERM investment: Enhancing risk management policies and implementation such
as the risk appetite framework will improve the rating of ERM and benefit the business. It
may also help achieve a credit rating upgrade, which also reduces the borrowing cost.
4) Increase/decrease equity allocation: Riskier asset investment helps improve the earnings
expectation and grow the capital in the long term. However, at the same time, more risks
are taken such as high probability of capital deficiency and high earnings volatility.
5) Hedge rho of VA business: A hedging program reduces earnings volatility and stabilizes
capital position at a cost of taking a long position in derivative markets or dynamic
hedging implementation. In this case, hedging the interest rate risk of VA business
reduces CaR and EaR.

2.5 Risk Appetite for Each Risk Category and Risk Limit

Enterprise risk tolerance can be further broken down into a risk preference for each risk
category. Risk preference is normally stated quantitatively (VaR limit) and qualitatively (do
not bear certain types of risk). To efficiently and clearly communicate risk appetite to
business units and different operation functions, risk appetite needs to be translated into risk
limits in a quantitative way. This ensures the stated risk appetite is followed. Modelling is a
key tool used to create reasonable and credible linkage between overall risk appetite and risk
limits. Major risk categories are discussed below to give an example of how risk preference
can be translated into risk limits. Another important piece is diversification benefits because
of imperfect correlation among risk factors, business lines and business units. Risk tolerance
has a material impact on setting risk appetite for individual risk categories and limits as well.

While a lot of focus has been put on the quantitative aspects, companies are much more likely
to fail due to taking unidentified or unheeded risks. This is important to keep in mind before
setting up the risk appetite for each risk category. In the recent financial crisis, we saw big
players in the insurance industry get huge government bailouts to remain solvent. The cause
was not life or P&C insurance but insurance business on financial products. The lesson
learned was that a true and comprehensive picture of risks being taken and the exposure to
each risk category is far more important than quantitative risk measures. Those risk measures
are very sensitive to assumptions and as a result are exposed to a high level of model risks.
The list below includes most of the risk categories on which the insurance industry has been
focusing. Though it is desirable to have an inclusive list, new risks are emerging and need to
be identified and analyzed.

©2012 Society of Actuaries, All Rights Reserved Page 18 of 57

 2.5.1 Market Risk
Market risk appetite reflects constraints caused by risk-taking activities in interest rate risk,
equity risk, foreign exchange risk, alternative investments and volatilities of the financial
market. The appetite of market risk could be stated as:
• The company could not lose more than 10 percent of IFRS equity in a 1-in-200-year
event.
• The company will not participate in any strategy that bets on the direction and magnitude
of the foreign exchange rate movement (currency hedging) or change in equity volatility
(no butterfly 18 or straddle 19 investment strategies).

Appetite for market risk should be mapped to asset allocation limits, foreign exchange (FX)
limits, fixed income securities duration limits and/or asset liability mismatch (ALM) limits.
At this stage, strategic decisions have already been made to maximize the risk-adjusted return.
Therefore, the allocation of each risk type is already known. A case study of the linkage
between risk appetite and asset allocation is presented in section 3. Below is a description of
market risk limit setting.

To determine the risk limits, the first step is to allocate VaR for market risk to each sub risk
type: FX, interest, equity, etc. A simple example is shown below.

18
“A butterfly is a limited risk, nondirectional options strategy that is designed to have a large probability of
earning a small limited profit when the future volatility of the underlying is expected to be different from the
implied volatility.” Wikipedia, s.v. “butterfly (options),” accessed January 31, 2012,
http://en.wikipedia.org/wiki/Butterfly_%28options%29.
19
“A straddle is an investment strategy involving the purchase or sale of particular option derivatives that allows
the holder to profit based on how much the price of the underlying security moves, regardless of the direction
of price movement.” Wikipedia, s.v. “straddle,” accessed January 31, 2012,
http://en.wikipedia.org/wiki/Straddle.

©2012 Society of Actuaries, All Rights Reserved Page 19 of 57

 Figure 2.5.1.1 Market Risk Budgeting

Interest Rate

Allocation
Market Risk VaR Risk
- no 30
Additive
diversification
Market Risk VaR Limit
additive limit
- with
Equity Risk
diversification
50 15
30 FX Risk
5

Consider the company has the following interest rate risk profile:

Figure 2.5.1.2 Interest Rate Risk Profile

Value Duration Duration Mismatch
Asset 200 10
15
Liability 160 25
1 in 200 year event: Interest rate drop 2%
∆ asset 40 =200*10*2%
∆ liability 80 =160*25*2%
∆ surplus -40

The current risk profile for interest risk exceeds risk appetite for interest rate by $10 million. 20
A duration mismatch limit consistent with the risk appetite for the interest rate needs to be
worked out.

Based on the current balance sheet, the duration mismatch needs to be reduced to 12.5 years
so the loss for a 1-in-200-year event is $30 million. The increase of asset value under a
1-in-200-year event needs to be $50 million so that the change in surplus is equal to $30
million. Divide $50 million by asset value $200 million and then by the 2 percent interest rate
change. The desired asset duration is now equal to 12.5 years. The duration mismatch limit is
duration of liability – duration of asset (25 years – 12.5 years = 12.5 years).

In reality, duration itself, even with convexity, might not do a good job as there are a lot of
management actions to reduce risk exposure when things go wrong. On the other hand, the
policyholder may behave in a way that makes things worse. Calculating a limit without
considering those factors could be misleading. Therefore, dynamic modeling is a key and
difficult component in deriving a reasonable limit. Dynamic management actions include
dynamic investment strategy, dynamic policyholder dividend setting, dynamic credit interest
rate setting and dynamic premium resetting. Dynamic policyholder behaviors include
20
All dollar figures are USD unless otherwise noted.

©2012 Society of Actuaries, All Rights Reserved Page 20 of 57

 dynamic lapse and dynamic premium payment.

2.5.2 Credit Risk

Credit risk consists of downgrade and default risks related to investment and reinsurance
activities. The risk appetite statement for credit risk is in a similar format to that of market
risk. An example could be:
a. The company could not lose more than 5 percent of IFRS equity in a 1-in-200-year event.
b. The company will not invest in any bonds that have a credit rating below BBB.

Appetite for credit risk is mapped into credit limit such as obligor limit or reinsurance
counterparty credit rating limit.

2.5.3 Insurance Risk

Insurance risk, aka underwriting risk, includes mortality, morbidity, lapse and expense risks. It
will impact both the value of new business (VoNB) and the reserve of in-force business.
Appetite for insurance risk defines the risk tolerance of those risk types to limit potential
losses. An example could be:
• The company cannot lose more than 5 percent of IFRS equity in a 1-in-200-year event
due to the insurance risk’s impact on reserves.
• The company cannot lose more than 50 percent of VoNB in a 1-in-200-year event due to
the insurance risk’s impact on pricing.

Translating appetite for insurance risk into risk limit is not straightforward. To be more
effective and easier to monitor, an existing experience monitoring framework needs to be
relied on. A possible choice is the A/E ratio 21 and/or loss ratio. A stress test under a
1-in-200-year event needs to be done to get the A/E ratio limit. It is calculated as the expected
payment under the stress test divided by the expected payment with a best estimate
assumption. The regular experience study results can be monitored against the limit to protect
from insurance risk. For expense risk, a limit in terms of amount might be a better choice for
risk-monitoring purposes. However, for a company with diversified products featured even
within each product line, this approach may create practical issues regarding the required
efforts and timing for experience monitoring at the product level. In addition, for new
products or new product features, it is often challenging to set assumptions and a risk limit.
From this perspective, it may be more valuable as a tool to identify risks and adverse trends
than to set a limit.

A risk limit and risk-monitoring report for insurance risk may look like the following.

21
A/E ratio: Actual payment/expected payment. For example, the payment could be the death benefit for
mortality risk or the surrender benefit for lapse risk. It is used to track experience of insurance assumptions.

©2012 Society of Actuaries, All Rights Reserved Page 21 of 57

 Figure 2.5.3.1 Illustrative Insurance Risk Monitoring Report
Risk Metric for insurance risk Risk Limit Current position
A&H* Loss ratio deviation from pricing** +5% +7% A&H Loss ratio
A/E mortality rate for life 115% 108% and lapse rate
A/E mortality rate for annuity 90% 94% exceed risk limit
A/E Lapse rate for non-lapse supportive 130% 135% and need
A/E Lapse rate for lapse supportive 90% 95% mitigation plan
Expense overrun $2.5 million $2.36 million

*A&H: Accident and health products

** pricing target ratio may float from 35-90%, depending on mix of business

Another consideration of insurance risk is the possible offsetting impact of life insurance
business and annuity business on mortality risk. It is a natural hedging and should be
considered if any reflection is appropriate when setting the limit. Other hedging instruments
in the asset portfolio such as longevity bonds should be taken into account as well.

2.5.4 Catastrophe Risk

A company’s exposure to catastrophe risk has attracted more attention due to climate change,
pandemics and a growing number of natural disasters. Risk exposure could devour a
company’s available liquidity quickly if such a catastrophe occurs and the company runs the
risk of bankruptcy without sufficient liquidity. An example of appetite for catastrophe risk
could be:
• The company cannot lose more than 20 percent of IFRS equity in a 1-in-200-year event.
• The company cannot lose more than 20 percent of IFRS equity if the 1918 Spanish flu
pandemic happens again.

Risk limit regarding catastrophes can be easily aligned with risk appetite via net amount at
risk (NAaR). NAaR is the sum assured minus the reserve. It is the additional amount that has
to be paid in excess of what has been reserved. Monitoring real-time NAaR against the limit
prevents excessive risk exposure.

2.5.5 Liquidity Risk

In a catastrophic event, insurance companies face high liquidity requirements, mass policy
lapses due to disintermediation risk and credit rating downgrade. Lessons of bankruptcy due
to insufficient liquidity are numerous. Appetite for liquidity risk could be:
a. The company maintains liquidity for a 1-in-200-year event over a time horizon of three
months.
b. The company maintains liquidity at the confidence level of 99.5 percent while the
liquidity cost to meet cash payments at the confidence level of 99.5 percent (1 in 200
years) is less than 25 percent of capital.

©2012 Society of Actuaries, All Rights Reserved Page 22 of 57

 Appetite for liquidity risk would constrain investment choices as it would indicate the level of
liquid asset investment. It is a major constraint and the company may have a lower risk
tolerance of it than that of other risk categories due to the devastating impact. This point will
be discussed in section 6 with an example.

2.5.6 Concentration Risk

Concentration risk relates to the risks caused by a dominating position in a risk category,
product line, region, distribution channel or asset class. Appetite for concentration risk could
be:
• The company has a natural diversification of geographic concentrations of insurance and
political risks.
• The company has a range of product offerings including both life and annuity products
and thereby reduces exposure to concentrations of mortality or longevity risk.

Appetite for concentration risk is an important element in strategic planning, especially when
developing business plans. A business development strategy that causes concentration in a
certain risk category or business line will put the company in danger because of the lack of
diversification. A good risk limit or risk-monitoring system can track the real experience to
identify significant deviation from appetite for concentration risk.

2.5.7 Terrorism Risk

Terrorism causes deaths, injuries, business interruption and adverse market movement. It
hurts the company’s financial strength and business. An example of appetite for terrorism risk
could be:
• The company cannot lose more than 20 percent of IFRS equity in a terrorism event.
• The company has a contingency plan in place for continuing business operations in the
event of terrorism.

Risk limit related to terrorism risk focuses on the concentration of policyholders’ locations.
For example, the company may stop underwriting life insurance coverage for additional lives
working in the same building once the total sum assured on the lives in the building reaches
$100 million. Without sufficient reinsurance coverage to cap the maximum possible claim
amount, location concentration can have a disastrous impact in a terrorism event.

2.5.8 Operational Risk

Appetite for operational risk normally is qualitative, such as “the company disallows any
behaviors that will damage our reputation.” Appetite for operational risk may not be
translated to quantitative risk limit but is ensured by company policy and/or internal control.
Brand management and antimoney laundering and compliance policies are examples of
reputation and legal risk management.

©2012 Society of Actuaries, All Rights Reserved Page 23 of 57

 2.5.9 Diversification Benefit
Insurance companies manage to provide protection from insured risk at a reasonable price due
to the pooling of individual exposures. Diversification among risk categories, products and
geographic regions reduces the aggregate level of risk. The diversification benefit is a critical
factor in determining the current risk profile with a bottom-up approach and setting the risk
limit from the top down. It is also the most important factor in economic capital management
when considering both the high sensitivity to the assumption of dependency and its magnitude
in reduction of required economic capital. Therefore, prudence is highly needed when
quantifying the diversification benefit and using it as an argument in business decisions.
Increased correlation in a tail event is a good example. Normally, there is a geographic
diversification benefit in capital management that takes into consideration the nonperfect
correlation among countries and regions. In some tail events, such as the financial crisis, the
correlation between countries tends to increase quickly due to the contagion factors. Low
interest rates, high credit risk and widespread panic among investors could happen in all
countries in a region at the same time. The high diversification benefits considered in business
decisions in normal situations were gone immediately and the once huge amount of excess
capital disappeared. Therefore, making a reasonable diversification assumption and keeping it
from manipulation is critical in risk management and business decision.

Several approaches to quantifying diversification are adopted in a risk appetite framework.

Those methods incorporate a higher correlation of tail events in different ways.
• Correlation matrix approach: This approach uses a correlation matrix to aggregate the
value distributions or certain percentiles of different risk factors.
• Copula approach: Instead of assuming a linear correlation as used in the correlation
matrix approach, the copula approach determines joint distribution based on marginal
distributions and a copula function. It allows better modelling of tail events where much
higher correlations were seen than in normal circumstances.
• Structural scenario approach: This approach uses stochastic global scenarios that include
all the risk factors. The correlations between risk factors are incorporated in the stochastic
scenarios. Asset and liability portfolios are assessed under those scenarios to obtain value
distribution.

Different approaches are appropriate for different purposes. When determining the
diversification benefit among different risk categories such as market risk and insurance risk,
and among geographic regions, correlation matrix and copula approaches are often used.
Sometimes different correlation matrices are used at different percentiles to reflect the higher
correlation in tail events. This, somehow, compensates for the disadvantage of assuming
linear correlation in the correlation matrix approach. When calculating the diversification
benefit among risk factors within each risk category, such as interest, equity, foreign exchange
and volatility under market risk category, the structural scenario approach is more appropriate
as it can build the causal relationship into scenario generation. For example, interest rate,
equity return and currency value are highly dependent on a macroeconomic environment.

©2012 Society of Actuaries, All Rights Reserved Page 24 of 57

 Building in the relationship of those risks in a macroeconomic environment in global
scenarios provides more insights about the underlying causes of market movement and,
therefore, facilitates effective management actions.

3 Risk Appetite and Asset Allocation

Strategic asset allocation (SAA) is used to determine a long-term policy portfolio reflecting
the desired systematic risk exposure. Tactical asset allocation (TAA), on the other hand,
specifies the allowable deviation from SAA to take advantage of short-term market
opportunities. A numerical example is given below.

Figure 3.1 Asset Allocation Plan

Asset class SAA TAA
Bond 50% +/–7% (43%~57%)
Equity 30% +/–10% (20%~40%)
Cash 20% +/–3% (17%~23%)

Before risk appetite frameworks became part of corporate governance, insurance companies
decided asset allocation based on the return objective and risk consideration for each business
line and surplus, separately. Segmented investment portfolios supporting different product
lines lead to multiple return objectives.

Return objective
• Minimum return: statutory rate set by actuarial assumptions to fund statutory reserve.
Assets chosen are based on the duration matching strategy.
• Enhanced margin: competitive return earned to fund a well-defined liability and a
reasonable profit.
• A surplus account usually has a riskier asset allocation to achieve higher returns.

Risk considerations
• Valuation concerns: adverse market movement that lead to reduced surplus so risk
exposure is limited to a safe capital adequacy level.
• Cash flow volatility: low tolerance of income loss.
• Reinvestment risk: duration and/or convexity match between asset and liability portfolios.
• Credit risk: need to reduce credit risk through a diversified portfolio.
• Disintermediation risk: periods of large cash outflows when a high interest rate leads to
more policy loan lapses, which decrease the duration of liability.
• Regulatory and legal constraints about eligible investment.

When determining asset allocation, those considerations can help realize return objectives
while protecting from the identified risks. However, without a clearly defined risk appetite, or
the analysis framework supporting risk appetite, there is no holistic thought of the relationship
of risk and return or the correlation of all the identified risks. For example, reinvestment risk

©2012 Society of Actuaries, All Rights Reserved Page 25 of 57

 and disintermediation risk have negative correlations. Low interest rate implies high
reinvestment risk and low disintermediation risk. With risk appetite in the picture, strategic
asset allocation will play a better role in ensuring the realization of both risk and return
objective. After the high level objective is determined, appropriate and different return
objectives can be set for individual segments to achieve the overall goal.

Case study
A simple example below shows the difference in the asset allocation decision processes and
results.

Figure 3.2 Illustrative Balance Sheet and Earnings

Balance sheet: (USD million) Earnings: (USD million)

Surplus
40
Expected
Asset
Earning
Liability
200
4
160

Two available asset classes:

Expected return Risk — Volatility
Bond 5% 8%
Equity 9% 10%
Risk free rate: 3%
Correlation between bond return and equity return: 20%
Required return: 8.5%

A typical asset allocation process without risk appetite would construct the efficient frontier
and capital market line (CML). 22 In Figure 3.3, we can see that CML touches on the efficient
frontier when the expected return is 8 percent. Assume the borrowing cost is the risk-free rate.
To achieve an 8.5 percent return objective, the most efficient way is to borrow 20 percent
risk-free assets and invest the entire available fund in the tangent risky portfolio (25 percent in
bond and 75 percent in equity). However, in reality, there are considerations that make it
deviate from the theoretical optimal solution. The borrowing cost could be higher and there
might be high cash flow volatility as the dividend income is not very stable compared to bond
coupon payments and redemptions. In this analysis, it is apparent the risk considerations are
not combined in an efficient way. Different risk considerations might drag the desired asset
allocation in the opposite direction and there is no effective way to find out the optimal or
22
CML is the tangent line of the efficient frontier and it crosses the risk-free rate at zero risk.

©2012 Society of Actuaries, All Rights Reserved Page 26 of 57

 suboptimal position to manage everything in a balanced way.

Figure 3.3 Asset Allocation without Risk Appetite

Efficient Frontier
14%
12% CML
10%
Return

8%
6%
4%
2%
0%
0% 5% 10% 15% 20% 25%
Risk - Volatility

With a mandatory risk appetite in place, strategic asset allocation can be used to consider risk
holistically. Figure 3.4 shows three additional comprehensive risk measure candidates highly
related to risk appetite.

Figure 3.4 Asset Allocation with Risk Appetite

Efficient Frontier
14%

12%
10%
10%
Expected Return

9%
8% 8%
8%

6%

4%

2%

0%
0% 50% 100% 150%
Volatility as % of asset value 95% VaR as % of asset value
95% CaR as % of available capital 90% EaR as % of expected earning

95% VaR 95% CaR 90% EaR Risk - Vol

Notes:
Volatility is presented as percent of asset value.
VaR is presented as percent of asset value.

©2012 Society of Actuaries, All Rights Reserved Page 27 of 57

 CaR is presented as percent of current available capital.
EaR is presented as percent of expected earnings.

With those risk measures, the tangent risky portfolio also changes.

Figure 3.5 Tangent Risky Portfolio Under Different Risk Measures

Risk measures Expected return Bond Equity
Volatility (traditional one) 8% 25% 75%
95% VaR 9% 0% 100%
95% CaR 10% -25% 23 125%
90% EaR 8% 25% 75%

• VaR at 95 percent: It incorporates interest rate, credit, equity and FX risks. Using a risk
measure with a confidence level consistent with the risk appetite statement is more
appropriate than using volatility as a 1-in-20-year event. This is of more concern than
utilizing one standard deviation.

• CaR at 95 percent: It is a quantitative component in the risk appetite statement. It also

implies the asset allocations that are not allowable. For example, if the risk appetite is to
not lose more than 50 percent of capital in a 1-in-20-year event, then all the efficient
portfolios with expected returns higher than 11 percent are not appropriate as 95 percent
CaR of those portfolios exceed 50 percent loss of capital.

• EaR at 90 percent: Like CaR, it is one of the key risk measures in risk appetite. Figure 3.4
shows that portfolios with expected returns higher than 11 percent would have 90 percent
earnings at risks greater than 100 percent expected earnings. In other words, negative
earnings are expected to happen in a 1-in-10-year event.

Risk appetite can help refine risk measures in strategic asset allocation and also help a
company consider asset allocation from a holistic perspective.

Then can those efficient frontiers be constructed? With the calculation engine that can support
risk appetite analysis, it is highly probable the efficient frontiers can be derived by running the
engine with a set of asset allocation plans.

Tactic asset allocation normally determines the allowable range according to the volatility of
asset classes. The higher the volatility, the wider the range. The logic would be that a volatile
asset class also presents high alpha, which an experienced portfolio manager may take
advantage of. Risk appetite may add extra constraints as we need to test if the upper and
lower bounds of TAA can trigger the situation that CaR and/or EaR exceed risk tolerance. If

23
Negative means short selling, which is probably forbidden for insurance companies. Therefore, in this case,
relying solely on 95 percent CaR cannot give a feasible and optimal asset allocation and the suboptimal allocation
plan should be used.

©2012 Society of Actuaries, All Rights Reserved Page 28 of 57

 that happens, the TAA band needs to be narrowed until it is in line with risk appetite. From
this new perspective, risk appetite helps ensure that alpha-seeking behaviors (TAA) will not
put capital/earning position at high risk.

4 Risk Appetite and New Business Budgeting

Insurance companies normally prepare new business budgets of certain return or value
measures each year. On the other hand, clients, shareholders, employees and regulators are
interested in understanding the amount of risk the company will take in the future. Some
commonly used return or value measures are traditional embedded value of new business for
life insurance and combined ratio for non-life insurers. 24

These return or value measures do not fully and accurately consider the level of risk being
taken. Therefore, other measures need to be used to explain the impact of the new business on
the future risk profile. New business budgets and risk appetite have to find an effective way to
communicate with each other so that the risk tolerance will not be broken.

One candidate of the measures to be considered is the risk-adjusted return on capital

(RAROC). In the insurance industry, RAROC takes into consideration the expected profit and
risk over the entire life of a policy. It allows for a consistent comparison of activities across
different types of risks and businesses.

Other candidates, such as market consistent embedded value (MCEV), adjust the approach of
value measurement used by traditional embedded values of new business. MCEV takes into
account the cost of nonhedgeable risks, the cost of options and guarantees offered in the
insurance contracts, and the frictional cost of capital explicitly.

In the following case study, RAROC is used as the key measure considered in business
planning. There are several versions of RAROC definitions in the real world. It is specified
for the case study as follows.

where PV stands for present value

PV (underwriting profit) = PV (premium) – PV (claims) – PV (acquisition costs) –
PV (other expenses)
Required capital is the required economic capital
t: effective corporate tax rate

24

©2012 Society of Actuaries, All Rights Reserved Page 29 of 57

 Case study

Reinsurance company ABC started writing worldwide property and casualty business and
several lines of products in 2011. Senior management is preparing a 2012 new business plan,
which includes:
• New business mix projection for 2012
• Total premium target of 2012 for each business unit

Normally, there is another step in preparing the new business mix for each business unit that
is consistent with the new business mix at the total company level and the allocation plan
among business units. For simplicity, it is assumed that each business unit shares the same
new business mix as that of the total company.

Step 1: New business mix projection for 2012

Assume the financial results in statutory reporting, as of Dec. 31, 2011, are as follows.

Figure 4.1 2011 Reinsurer ABC Business Mix (USD million)

Premium Underwriting Profit
Year 2011 25
income profit margin
Auto liability 10 0.9 9.0%
Specialty liability 15 1.5 10.0%
CAT (catastrophe) 50 7.5 15.0%
Homeowners/farm owners 25 0.8 3.2%
Company 100 10.7 10.7%

Reinsurer ABC has a relatively high exposure to catastrophe risk. Half of its business covers
natural catastrophe, which is also the main contributor to the underwriting profit in terms of
amount. Therefore, it seems to make sense that premium income planned for CAT should at
least be kept at the 2011 level.

However, the profit margin does not reflect the risks the company has taken. Figure 4.2 lists
the required capital and RAROC for each business line. As a startup reinsurance company, the
available capital of reinsurer ABC is $250 million at the end of 2011. For simplicity, it is
assumed 50 percent of the reinsurance contracts will expire on Dec. 31, 2011, which has a
required capital of $72 million. At the beginning of 2012, out of $250 million available capital,
$72 million is used to support existing business and $178 million can be used to support new
business activities.

25
Profit margin = Underwriting profit/premium

©2012 Society of Actuaries, All Rights Reserved Page 30 of 57

 Figure 4.2 Return and Required Capital as of Dec. 31, 2011 (USD million)
[PV (underwriting
Required PV (required profit) + PV (invest.
Year 2011 Premium 26 RAROC
capital capital) income on capital)] ×
(1-t)
27
Auto liability 10 5 10 1.8 18%
Specialty
15 10 20 3 15%
liability
CAT 50 125 250 15 6%
Homeowners/
25 20 40 1.6 4%
farm owners
Company 100 144 289 28 21.4 7.4%

Reinsurer ABC’s risk appetite statement is as follows:

1. Economic capital adequacy: The company has sufficient economic capital with a
probability of 99.95 percent (available economic capital is no less than the required
economic capital in a 1-in-2,000-year event).
2. Economic earning volatility: Reinsurer ABC has a long-term target 10 percent RAROC
(hurdle rate) over the cycle. The company does not want to see drops in earnings by more
than 40 percent in a 1-in-20-year event. In other words, the company should earn at least
6 percent RAROC with a probability of 95 percent.

At the end of 2011, required capital ($144 million) is less than available capital ($250 million),
which implies the company's capital position is consistent with risk appetite. Reinsurer ABC
earns 7.4 percent in 2011, which is much less than the target of 10 percent but above the limit
of 6 percent. This is something to be improved upon.

Before determining the new business mix in 2012, another important decision to be made is
whether the premium rate needs to be adjusted. Premium adjustment will change the view of
profitability and capital requirement. In this case study, it is assumed the company is
operating close to the average level of RAROC in the industry for each individual business
line. The activities with RAROC in excess of the hurdle rate create shareholder value, while
those below the hurdle rate deplete shareholder value.

Comparing figures 4.1 and 4.2, another interesting finding is that the line exposed to CAT has
the second lowest RAROC while its statutory profit margin is the highest. This line produces

26
Required capital here is the 99.95th percentile of loss distribution, the mean loss under the economic
framework. The correlation among all the product lines is assumed to be 50 percent for illustration purposes.
27
The four business lines (auto liability, specialty liability, commercial multi-peril and homeowners/farm owners)
are chosen to be the same as the product lines studied in Nakada, et al., “P&C RAROC: A Catalyst,” 15. The
industry average RAROC information (Exhibit 14, Page 15) available in the study was referred to in the case study.
28
The company level PV(Required Capital) takes into account diversification benefit. Therefore the total amount
is lower than the sum of line figures.

©2012 Society of Actuaries, All Rights Reserved Page 31 of 57

 a much higher return most of the time given that catastrophe has low frequency but high
severity. However, when a catastrophe event happens, the loss is significant.

As a new company, the board of directors wants to maintain a minimum 20 percent premium
growth rate. The 2012 new business mix plan needs to address the following:
• The economic capital adequacy requirement needs to be met. Capital available for new
business is $178 million.
• The overall target RAROC of 10 percent needs to be met.
• Long-term client relationships need to be maintained; therefore, reducing undesired lines
of business needs to be gradual.
• The new business projection should also consider the phase of cycles for different lines of
business (hard market 29 or soft market 30).
• Other constraints mentioned in section 2 such as appetite for catastrophe risk and
concentration risk need to be assessed.

A relative shift from CAT and homeowners/farm owners to auto and specialty liability
business is planned to increase expected RAROC and efficiency in the use of capital. It is
predicted that 2012 will be a soft market so there is no rate increase and a profitability
increase is expected. A 45 percent increase in premium income is targeted.

Figure 4.3 2012 New Business Plan (USD million)

[PV (underwriting
Required PV (required profit) + PV (invest.
2012 Premium RAROC
capital capital) income on capital)] ×
(1-t)
Auto liability 50 25.0 50.0 9.0 18%
Specialty
30 20.0 40.0 6.0 15%
liability
CAT 45 112.5 225.0 13.5 6%
Homeowners/
20 16.0 32.0 1.3 4%
farm owners
31
Company 145 148.4 296.8 29.8 10.0%

Expected RAROC is 10 percent for 2012 new business and the required capital is $148.4
million, which is less than the available capital for new business.

29
A hard market comes after a catastrophic event when the price of insurance coverage and demand increases.
30
A soft market happens several years after a big catastrophe, when the event has been gradually forgotten and
competition increases. The price of insurance coverage goes down and there is less demand in the market,
thus less margin.
31
See footnote 30.

©2012 Society of Actuaries, All Rights Reserved Page 32 of 57

 Step 2: Total premium target of 2012 for each business unit

Figure 4.4 2011 RAROC for Each Business Unit

Risk Adjusted Return on Capital (%)

A 12

B 11

C 10

D 6.8

E 3.5

0 5 10 15

Value Reduction Value Creation

Business Required EC29 RAROC (%) Excess Return (%)

Unit ($MM) (RAROC - hurdle rate)
A 30 12.0 2.0
B 40 11.0 1.0
C 20 10.0 0.0
D 45 6.8 -3.2
E 50 3.5 -6.5
Company 144 7.4 -2.6 32

Business units A and B add value to the firm. Their RAROCs of 2011 business exceed the
prespecified hurdle rate of 10 percent. On the other hand, BUs D and E reduce shareholder
value. BU C earns the rate at the average level in the market.

It is implied that from the perspective of capital efficiency, more capital needs to be allocated
to BUs A and B relative to BUs D and E to support more business expansion.

32
Required capital here is 99.95th percentile of loss distribution, the mean loss under the economic framework.
The correlation among business units is assumed to be 50 percent for illustration purposes.

©2012 Society of Actuaries, All Rights Reserved Page 33 of 57

 5 Risk Appetite and Capital Allocation

Capital available to a firm is not inexhaustible. Generally, capital should be employed to

generate enough return for shareholders on a sustainable basis. The level of required return is
determined by the level of risk taken. Otherwise, the share price of the firm will underperform
that of competitors.

Companies need to allocate capital to different risk types, business lines and business units.
Below is an example for a top-down capital allocation structure.

Figure 5.1 Capital Allocation Structure

Group Operation

BU 1 BU 2 BU 3

… Class 1 Class 2 Class 3 …

Wind
WIND Earthquake
EQ Flood
FLOOD

The company has three business units and each business unit can write coverage for damages
caused by wind, earthquake and flood.

In some cases, companies allocate capital based on a statutory framework or rating agencies'
capital models. The appropriateness of those approaches highly depends on the regulatory
rules or rating agencies' rules.

• Under a statutory framework, regulatory rules build conservatism into calculating

reserves and specify the amount of required capital needed to remain solvent. Although a
certain degree of the provision for adverse deviations (PAD) of reserving assumptions is
used, it is hard to get a clear idea of how well the PAD can protect from the risks being
taken. Regulatory required capital frameworks vary significantly across the world. Such
frameworks compare businesses from different regions using the same criteria. Most
statutory regulations adopt factor-based models where the factors are prescribed by the
regulator. However, the lack of required granularity in those models may lead to an
inaccurate estimation of the risk exposure. Those may be improved significantly after the
implementation of the Solvency II framework in Europe and the adoption of more
principal-based reserve methods.

• Rating agencies’ models are very helpful in understanding the requirement to maintain the

©2012 Society of Actuaries, All Rights Reserved Page 34 of 57

 target credit rating. However, like statutory frameworks, the model cannot cater to all the
idiosyncratic characteristics of the business.

• Both statutory and rating agencies’ capital models are prescribed and cannot reflect the
view of the board of directors and senior management.

Economic capital framework was chosen by many insurance companies for efficient capital
allocation after realizing the disadvantages of using the statutory capital model or rating
agencies’ capital model. It provides a platform to compare and select business opportunities
from different business lines, risk types and geographic regions. However, relying on
economic capital framework alone is not practical as the company still faces the constraints of
regulators and rating agencies.

Risk appetite takes into account the specifics of the business, the investors’ risk tolerance and
all the constraints. It provides the guideline for capital allocation with the board’s risk
tolerance (CaR, economic capital adequacy and EaR). The explicit statement of the
probability of loss and associated maximum loss amount facilitates capital allocation
decisions.

Capital allocation is highly correlated with new business plans. A new business plan deviating
from current capital allocation actually changes the future capital allocation. The example of a
new business mix in section 4 also applies to capital allocation except that only new business
is considered. It explains how to make decisions to meet the RAROC target (an economic
measure) and have economic capital remain sufficient. The following case study, based on the
one in section 4, highlights a situation where a rating agency’s capital requirement constrains
fast growth.

Case study

Reinsurer ABC has a target financial strength rating of AA. The rating agency’s capital model
shows that a 45 percent premium income growth in 2012 will make required capital at the AA
rating greater than the available capital. Considering that the company needs to hold a buffer
of 50 percent annual net income on top of the required capital at the AA rating, the maximum
allowable premium income growth in 2012 is 30 percent. A revised new business plan
including capital allocation according to required capital is given below.

©2012 Society of Actuaries, All Rights Reserved Page 35 of 57

 Figure 5.2 Revised 2012 New Business Plan (USD million)
[PV (underwriting
Required PV (required profit) + PV (invest.
2012 Premium RAROC
capital capital) income on capital)] ×
(1-t)
Auto liability 44.8 22.4 44.8 8.1 18%
Specialty
26.9 17.9 35.9 5.4 15%
liability
CAT
40.3 100.9 201.7 12.1 6%
(catastrophe)
Homeowners/
17.9 14.3 28.7 1.1 4%
farm owners
33
Company 130 133.0 266.1 26.7 10.0%

Some of the considerations embedded in risk appetite framework are important to capital
allocation too. They are necessary for aligning capital allocation with risk appetite.

Available capital versus required capital

If the total available capital is much greater than the total required capital, the company can
reduce the available capital (dividend payment to shareholders or share buyback) and/or take
up more business opportunities (increase the size and/or risk of the business).

If the total available capital is lower than the total required capital, the company has to take
actions to increase available capital (raise additional capital from shareholders or the market,
perhaps by issuing preferred stocks) and/or reduce the required capital (reduce the size or risk
of the business).

The role of diversification in capital allocation

Diversification benefit happens in many places. For a given product, there is diversification
between different risk categories it is exposed to, such as between insurance risk and financial
risk. Within a business unit, there may be diversification among different products, such as
diversification between life insurance products and annuity products regarding mortality risk
and longevity risk. There also exists diversification across different types of business at a
higher level, such as diversification benefits between general insurance business and financial
service business.

For capital allocation, available economic capital allocated to business units should be based
on the required economic capital after diversification benefits. Business units providing
significant diversification benefits to the group are preferred for extra capital allocation due to
the relatively lower marginal cost of capital.

33
See footnote 30.

©2012 Society of Actuaries, All Rights Reserved Page 36 of 57

 Capital fungibility

Capital is assumed to be freely mobile among different legal entities in the examples. If one
entity has insufficient capital, the capital can be moved from other legal entities with
sufficient excess capital. In the real world, it is not always true and subject to regulators’
approval. Entities accumulate their retained earnings to build up the capital. The easier and
normal way to transfer capital is paying the parent company in the form of dividend. If capital
is not easy to transfer between entities, the diversification benefit is reduced.

6 Risk Appetite and Liquidity Management

Liquidity is “the ability to fund increases in assets and meet obligations as they come due.”34
Liquidity risk can be described as “the measure of probability that a company’s cash
resources will be insufficient to meet current or future cash needs.” 35 When a liquidity event
occurs, the loss will happen via transaction cost, interest payments on borrowings and market
impact of fire sale. Bankruptcy costs may also occur when mass lapses take place together
with a credit rating downgrade.

Appetite for liquidity risk defines the liquidity risk tolerance and, therefore, appropriate cash
levels held in the company’s asset allocation. Liquidity risk management without a clearly
defined and stated guideline usually leads to oversimplified rules and/or an overconservative
strategy if a prudent policy is adopted. Liquidity management policies without risk appetite
consideration could include:

• Cash balance is no less than the maximum weekly cash payment in the past three months.
• Cash balance is no less than Y times the maximum daily cash payment in the past month.
• Liquid assets cannot be less than 50 percent of the total asset balance.

While these kinds of liquidity policies protect the companies to a certain extent, they come at
the cost of earning a lower yield on those liquid assets. However, the underlying risks may not
be identified. The risks are caused by both the liability structure and the exogenous market
changes.

Appetite for liquidity risk is often stated as:

• The company maintains liquidity in a 1-in-200-year event over a time horizon of three
months.
• The company maintains liquidity at the confidence level of 95 percent while the liquidity
cost to meet cash payments at the confidence level of 99.5 percent (1 in 200 years) is less
than 25 percent of capital.

34
BCBS, “Sound Practices,” 7.
35
CRO Forum, “Insurance Risk Management,” 12.

©2012 Society of Actuaries, All Rights Reserved Page 37 of 57

 Those quantitative statements, of course, need a bottom-up analysis to figure out the major
liquidity requirements, the risks that will increase the liquidity requirement dramatically, the
market impact on available liquidity and the way to position the company.

With appetite for liquidity risk in place to protect companies from severe events like a credit
crisis or credit rating downgrade, it will also constrain strategic planning in the following
ways:
• When planning for new business sales, the impact on required liquidity would have to be
quantified so it can be verified that the overall liquidity position remains consistent with
risk appetite. This also includes identifying new factors related to the new business that
might largely change the liquidity requirement.
• Strategic capital management will also be impacted as decision makers need to keep
liquidity requirements in mind when allocating capital to different risks, business units
and business lines.
• Strategic asset allocation should consider the liquidity of assets and the likely liquidity
cost if immediate liquidation is required under stress scenarios.

Case study
In this section, we will provide an example of how the appetite for liquidity risk can be
established and linked with strategic planning. Note that the bottom-up liquidity risk analysis
is a critical tool and beneficial to both risk appetite setting/updating and strategic planning.

Company ABC needs to maintain a liquidity level to meet payment requirements for a 1-in
-200-year event for a continuing period of three months.

Available liquidity
The company’s assets are classified into three tiers with respect to liquidity risk management:
• Tier 1: highly liquid, like cash and government bonds
• Tier 2: liquid, like bond coupons and redemption, equity dividends, rental income, etc.
• Tier 3: not liquid. Selling those assets or cash flows might lead to big market impact and
significant liquidity cost.
Available liquidity = 98% Tier 1 assets + 80% Tier 2 assets.

Here the haircuts on Tier 1 (2 percent) and Tier 2 (20 percent) assets are based on capital
market condition.

Figure 6.1 shows ABC company’s available liquidity. Out of a $200 million asset value, there
is $105 million available liquidity.

Figure 6.1 Available Liquidity

©2012 Society of Actuaries, All Rights Reserved Page 38 of 57

 ABC Company Available Liquidity
200

Asset Value (USD Million) 150 95

100

14
50 1 105
80
49 56
0 0
Tier 1 Tier 2 Tier 3 Total
Available Liquidity Non-Liquid

Required liquidity
Required liquidity measures the required cash payment for a certain period (three months) in
a severe event (1-in-200-year event). It is composed of, but not limited to, the following
components:
• Credit rating downgrade impact: the additional cash payment requirement due to a credit
rating downgrade caused by either overall market chaos or idiosyncratic reason. Insurance
companies normally experience an unexpected mass lapse if downgraded. It can be
estimated based on past credit crisis experience. If the company's own experience is not
available, industry experience is a good source to estimate the impact.
• Normal operational cash flow volatility. Historical company data would be the most
appropriate sources to use for the analysis. For example, weekly historical data of net
cash flow (benefit outgo + expense – premium income for an insurance company) can be
used to fit to a distribution. The 99.5th percentile can be calculated either from historical
data directly or from fitted distribution. Distribution fitting may better capture tail events
when the history of available experience data is not long enough. 36 The historical data
should be adjusted to exclude the period when a credit rating downgrade occurred and/or
when there was a sharp increase in the interest rate as those two factors are assessed
separately from cash flow volatility.

Figures 6.2 and 6.3 provide an example of historical weekly data and some fitted
distributions. To use some fat-tailed distribution types such as lognormal, the historical
data are shifted to the right to be all positive, so that lognormal distribution and other
fat-tailed distribution types can be used. The fitted data are then shifted back to the left.

36
For example, 99.5th percentile needs at least 200 data points to have an estimate based on the worst one. If
there are only 100 data points, distribution fitting would be more appropriate.

©2012 Society of Actuaries, All Rights Reserved Page 39 of 57

 Figure 6.2 Probability Density Function Figure 6.3 Cumulative Density Function
— Shifted weekly net cash flow data — Shifted weekly net cash flow data
Mean
170

99.5 percentile
Historical: 560
Fitted (lognormal): 540

shifted value of weekly net cash flow shifted value of weekly net cash flow
(USD ‘000) (USD ‘000)

Subtracting the mean $170,000 from the 99.5th percentile ($540,000 based on fitted
lognormal distribution), the 1-in-200-year event will require $370,000 extra liquidity on
top of the expected value.

Weekly data analysis needs to be extended as a three-month period is being considered.

Two approaches might be used to get a reasonable estimation.
o Conservative approach: 3 × 4.345 × weekly results = $4.82 million. Here 4.345 is the
average number of weeks per month.
o Assume it follows the Wiener process (aka, Brownian motion):

• Catastrophe risk (CAT risk): Stress test the business portfolio using some extreme events
that occurred before, like the 2011 Japan earthquake and the 1918 Spanish flu pandemic.
Calculate the required benefit payment under those stress scenarios.

• Funding commitments such as private equity investment or real estate investment.

• Interest rate risk: Assess the impact on both the interest rate risk sensitive products and
asset values when there is a significant increase in the new money rate. Normally,
increased lapse activity and a drop in asset value will be seen. This can create additional
liquidity requirements and realized losses due to asset liquidation at a depressed price.

• Adverse mortality and morbidity experience. Sometimes, actual claim experience may
deviate a lot from pricing assumption, especially for new markets, new products and new
features added to existing products. Actual claim experience is an important component
of required liquidity for start-up companies that do not have much credible experience.

©2012 Society of Actuaries, All Rights Reserved Page 40 of 57

 • Correlation among the above factors. The diversification benefit is a very important
assumption in determining the appropriate required liquidity level. Being too conservative
may hinder business due to the cost of maintaining a high liquidity level, such as low
investment return. Being too risky may lead to a position exceeding the risk tolerance.
Detailed research based on market data and reference to industry standard and peers is
essential to determine the appropriate correlation assumption. Industry standards have
many sources such as the CFO forum, 37 CRO Forum and Solvency II.

In this example, we use the following correlation matrix to calculate the diversification
benefit.

Figure 6.4 Illustrative Correlation Matrix

Normal Net Funding
Downgrade Catastrophe Interest Insurance
Correlation Matrix Cash Flow Commitmen
Risk Risk Rate Risk Risk
Volatility t
Downgrade Risk 1 0.4 0.4 1 0.5 0.4
Normal Net Cash
0.4 1 0.2 1 0.4 0.2
Flow Volatility
Catastrophe Risk 0.4 0.2 1 1 0.4 0.2
Funding
1 1 1 1 1 1
Commitment
Interest Rate Risk 0.5 0.4 0.4 1 1 0.4
Insurance Risk 0.4 0.2 0.2 1 0.4 1

Below is a brief explanation of the correlations.

o Interest rate risk and downgrade risk have a medium correlation of 50 percent. When
a downgrade happens in a credit crisis, bond yields will normally increase as the
credit spread widens. A medium correlation is set to be consistent with CRO Forum
recommendations, such as for portfolios with short durations (which is the more
common situation), and a correlation of 0.5 seems to be appropriate; whereas for a
portfolio long in duration, a correlation of 0 would be appropriate (conservative
assumption). 38
o Insurance risk has a medium correlation with financial risks (market and credit) of 40
percent. It is at the high end of the range suggested by the CRO Forum.

37
The European Insurance CFO Forum (CFO Forum) is a high-level discussion group formed and attended by the
chief financial officers of major European listed, and some nonlisted, insurance companies. Its aim is to influence
the development of financial reporting, value-based reporting and related regulatory developments for insurance
enterprises on behalf of its members, who represent a significant part of the European insurance industry.
http://www.cfoforum.nl/index.html.
38
CRO Forum, “Calibration Recommendation,” 6.

©2012 Society of Actuaries, All Rights Reserved Page 41 of 57

 Figure 6.5 Recommended Correlation Matrix by CRO Forum

Source: The Chief Risk Officer Forum, “A Framework,” 18 and 19.

Note: High correlation: 0.75~1; medium correlation: 0.40~0.75; low correlation: 0.0~0.40

o Funding commitment is an add-on item. These are scheduled events and very
predictable.
o Correlation within insurance risk is as low as 20 percent. In the fifth quantitative
impact study on Solvency II, 39 the following correlation matrix is specified for a life
underwriting risk module. A general 20 percent assumption is a reasonable value
compared to the one used in QIS5.

Figure 6.6 Correlation Matrix in QIS5

Source: CEIOPS, “QIS5 Technical Specifications,” 148.

39
CEIOPS, “QIS5 Technical Specifications,” 148.

©2012 Society of Actuaries, All Rights Reserved Page 42 of 57

 • Budget net cash flow (NCF) within the time horizon. This is an offset of required
liquidity for a mature company unless there is significant expansion that overwhelms the
contribution from in-force business.

Here two scenarios of required liquidity that will help illustrate how liquidity requirements
could influence strategic planning are presented.

Figure 6.7 Scenario 1: Required Liquidity

120

2.5
100 15.0 22.0
2.0
10.0
80
USD Million

60 60.0

40 77.3

4.8
20
25.0

0
Downgrade Risk Normal NCF Vol CAT Risk Funding Interest Rate Insurance Risk Diversification Budget NCF Required
Commitment Risk Liquidity

Required liquidity ($77 million) < available liquidity ($105 million). The current liquidity
level meets the requirement of liquidity risk appetite with a margin of $28 million. As the
buffer is not trivial, switching some liquid low-yield assets to less liquid high-yield assets
gradually to enhance investment income can benefit both policy- and shareholders.

Figure 6.8 Scenario 2: Required Liquidity

Scenario 2 has required liquidity before budget NCF at 50 percent more than that of scenario
1. Required liquidity ($116 million) > available liquidity ($105 million). Under this scenario,
the current available liquidity is below the level implied by the risk tolerance. Downgrade,

©2012 Society of Actuaries, All Rights Reserved Page 43 of 57

 interest rate and CAT risks are the candidates to reduce liquidity risk exposure. In strategic
planning, we might want to:

• Reduce CAT risk by decreasing the underwriting of CAT coverage and/or have it
reinsured to reduce exposure.
• Reduce downgrade risk and interest rate risk. Those risks will increase chances and
impact of mass lapses. In the new business planning, policies with market value
adjustment features are desired.
• Adjust strategic asset allocation to move assets with lower liquidity to assets with higher
liquidity gradually.

7 Risk Appetite and Performance Measurement

A healthy risk culture is the key to successfully implementing and gaining from a risk appetite
framework. A risk management mandate without the understanding and buy in of
management will not create value for the company. Including appropriate key performance
indicators (KPI) regarding risk appetite in managers’ performance scorecards would
encourage people to think in terms of both return and risk when making business decisions.
Three types of performance measures can serve this purpose.

• The gap between current risk profile and risk tolerance. This is to make sure the business
is safe compared to its risk tolerance. It is reflected at different levels.
o At the top level, the current risk profile is compared with enterprise risk tolerance to
identify major gaps. Group level senior management should be responsible for it.
o At the risk category level, exposure to each risk category is checked against risk
appetite for a specific risk. Risk committees should be responsible. For example, the
credit risk management committee needs to evaluate the credit risk exposure. The
asset liability management committee needs to be responsible for any significant
deviation from the appetite for interest rate risk.
o At risk limit, or business execution, level, risk management committees and
risk-taking departments are responsible for the monitoring. For example, market risk
management and the chief investment office’s performance scores will be deducted if
the equity VaR exceeds the limit.
• Risk-adjusted return: for example, actual risk-adjusted return on capital (RAROC) vs.
expected RAROC.
• Risk-adjusted value: for example, actual economic value added vs. expected economic
value added.

A few examples regarding how to set risk limits and perform risk monitoring against them
were discussed in section 2.5. In the following case study, we will focus on risk-adjusted
return and risk-adjusted value. The most difficult part of embedding these types of
performance measurements in the management’s scorecard is how the appropriate
performance target can be determined and how the deviation of actual results from the target

©2012 Society of Actuaries, All Rights Reserved Page 44 of 57

 can be explained. Management cannot control the system risk and the financial market
movement. Therefore, it is not reasonable to penalize them for this. A consistent, reasonable
and well-accepted performance measurement framework would have to address these issues.

In recent years, banks have increasingly adopted economic profit measures such as RAROC
and economic value added (EVA). It encourages senior management to take opportunity cost
of capital into consideration and maximize shareholder’s value given their risk appetite.

Different functions in a company are held responsible for different risk types. Investment is
responsible for achieving higher returns than expected. The asset liability management (ALM)
committee is responsible for minimizing the gap between asset and liability portfolios.
Business management is held responsible for new business growth and gain and loss from
nonfinancial risk. In the banking industry, the risk-adjusted value measure EVA is calculated
as: EVA = earnings – opportunity cost × capital allocated.

Life insurance companies usually require a longer period than banks before gaining profits.
Ideally, the earnings measure is correlated with a value measure used in the insurance industry.
MCEV is a good candidate for measuring economic value. According to the definition by
CFO Forum, “MCEV represents the present value of shareholders’ interests in the earnings
distributable from assets allocated to the covered business after sufficient allowance for the
aggregate risks in the covered business. The allowance for risk should be calibrated to match
the market price for risk where reliably observable.” 40 There are two approaches to calculate
MCEV: balance sheet approach and earnings approach. 41

40
CRO Forum, “Market Consistent,” 3.
41
More details about the calculation of MCEV and definition of its components can be found at CRO Forum,
“Market Consistent.”

©2012 Society of Actuaries, All Rights Reserved Page 45 of 57

 Figure 7.1 Two Approaches of MCEV Calculation
Balance Sheet Approach Earnings Approach
Free Surplus
MCEV Required Capital MCEV
VIF
FrCoC FrCoC

MVM CRNHR PVFP

MVA

MVL TVOG
CEL
CEL-TVOG

Tax Liability Tax Liability

Notes:
MVA: market value of asset
FrCoC: frictional cost of capital. It reflects the taxation and investment costs on the assets backing
42
required capital.
MVM: market value margin
CEL: current estimate of liability
MVL: market value of liability. It is equal to CEL + Tax Liability + FrCoC + MVM.
VIF: value of in-force business. It is equal to PVFP – TVFOG – FrCoC – CRNHR.
CRNHR: cost of residual nonhedgeable risks. It includes the impact of nonhedgeable, nonfinancial risks
and nonhedgeable financial risks and should be presented as an equivalent average cost of capital
43
charge.
TVOG, aka TVFOG: time value of financial option and guarantees
PVFP: present value of future profit of in-force business

Under an MCEV framework, the opportunity cost of allocated capital has been considered as
cost of residual nonhedgeable risks (CRNHR) and frictional cost of capital (FrCoC) due to
investment and double taxation costs. However, CRNHR and FrCoC take into consideration
all future years’ cost of using capital. In performance measurement, the actual earning/cost
and expected earning/cost are compared over the measurement period. We can define EVA =
MCEV earning – cost of capital for life insurance companies. MCEV earning refers to the
change of MCEV from the beginning to the end of the measurement period.

Investment
The task is to build two benchmarks so that we can attribute EVA to different functions:
investment, ALM and business management. In section 3, the concept of SAA/TAA and its
relationship with risk appetite were discussed. Based on SAA, investment departments can
manage an asset portfolio actively as long as the resulting portfolio does not stray outside the
allowable range specified by TAA. The investment department targets extra returns by
positioning the portfolio based on market movement expectations. For example, when the
interest rate is expected to be going down, the investment department may decrease the
42
CRO Forum, “Market Consistent,” 5.
43
Ibid., 5 and 6.

©2012 Society of Actuaries, All Rights Reserved Page 46 of 57

 duration of liability and increase the duration of assets. The investment department may also
pick up yields by investing in lower credit rating bonds. Translating into the change of MCEV,
it is obvious that an active investment department will change the present value of future
benefit and cost of capital 44 as it also has an impact on required capital. EVA calculation for
the investment department could be: EVAinv = extra investment income over SAA – ∆ cost of
capital.

Business management
The business management function is supposed to focus on new business and nonfinancial
risk such as insurance and operation risks. Therefore, a good benchmark should have the
following properties:
1. It mimics the liability characteristics as much as possible so if there is a change in the
financial market, it will change in the same way as the liability portfolio.
2. The replication is valid for a wide range of market situations.
3. Its value is easy to track.

Replicating the portfolio of liability seems to be a good choice. Sometimes, it is also called a
risk-minimizing portfolio. It uses the available liquid assets in the market to replicate the
value and sensitivities of liability. Under risk appetite framework, a replicating portfolio is
more complicated due to the additional risk-adjusted measures. It needs to replicate the cash
flows, economic value (MVL), sensitivities, and the earnings, value and capital requirement
under statutory and rating agency frameworks. This is to ensure that for the major quantitative
measures used in risk appetite, the replicating portfolio will not deviate from liability too
much.

With the replicating portfolio, EVA contributed by in-force business for business management
can be calculated as expected return on the liability replicating portfolio + experience
gain/loss due to nonfinancial factors such as mortality, expense, morbidity and lapse — cost
of capital. Cost of capital is calculated as the cost of capital rate times the required capital
consistent with risk appetite. In other words, we take the maximum statutory-required capital,
rating agency-required capital at the target credit rating and economic capital as required
capital. EVA contributed by new business can be calculated as MCEV of new business at time
zero. The following formula summarizes EVA for business management function:

EVAbus = MCEV of new business + expected return on replicating portfolio + experience G/L
– cost of capital

ALM
ALM performance measurement is straightforward now. The return on SAA over the return
on a replicating portfolio of liability is attributed to asset liability management. The difference
is caused by the mismatch between the asset and liability portfolios.

EVAALM = return on SAA – return on replicating portfolio

44
Cost of capital includes both FrCoC and CRNHR incurred due to the cash flow in a certain period.

©2012 Society of Actuaries, All Rights Reserved Page 47 of 57

 If the movement from MVA to MVL is decomposed, the responsibility of different functions
can be summarized as below.

Figure 7.2 MCEV Decomposition

Alpha -
Investment
Mismatch -
ALM MCEV
Non-financial
MVA risk - business
SAA management
Replicating
Portfolio
MVL

In the following case study, an example of how to evaluate investment function, asset liability
management function and business management functions is given.

Case study
Life insurance company ABC has been calculating and testing the use of risk-adjusted values
to measure the performance of different functions. As the company has a large exposure to
market and insurance risks, the focus is on evaluating the investment department, ALM
committee and business management. Let’s illustrate the EVA calculation with a numeric
example.

Figure 7.3 Asset Portfolio (USD million) — Consistent with Strategic Asset Allocation
Asset Class MVA Duration Expected Return
Short-term bond 100 5 3%
Long-term bond 100 20 5%
Total 200 12.5 4%

MVL is $160 million with a duration of 15 years. It is fully replicated with the fixed income
portfolio with a duration of 15 years and value of $160 million. The expected return of the
replicating portfolio is 4 percent.

The investment department has an expectation of the bond yield curve flattening. The
medium-term interest rate is expected to increase and the long-term interest rate is expected to
decrease. Within the allowable range specified by tactical asset allocation, a $10 million
short-term bond is sold for long-term bond investment. During the measurement period, the
five-year interest rate increases by 1 percent and 20-year interest rate decreases by 1 percent.
The 15-year interest rate decreases by (1/3) percent. The expected change in asset value of the
SAA portfolio is calculated as:

©2012 Society of Actuaries, All Rights Reserved Page 48 of 57

 MVi: market value of asset class i
ri: interest rate of asset class i
Di: duration of asset classs i

The actual asset value after active management using the same formula above is:

After interest rate change,

MV1 = 90 – 90 • 1% • 5 = 85.5
MV2 = 110 + 110 • 1% • 20 = 132
MVtotal = MV1 + MV2 = 217.5

Asset duration changes to:

Wi: weight of asset class i

Wi = MV1/MVtotal = 85.5/217.5 = 39.3%
W2 = MV2/MVtotal = 132/217.5 = 60.7%

This active management behavior benefits company ABC in two ways:

• Positive investment income over SAA = 17.5 – 15 = 2.5.
• Decrease in cost of capital. Asset duration increases from 12.5 to 14.1. The duration
mismatch is reduced from 2.5 years to 0.9 year. It will reduce the required capital to
support asset liability mismatch risk. Assume the cost of capital rate is 4 percent and the
reduction in required capital is $20 million. The decrease in cost of capital is $0.8 million
(4% • 20).

EVAinv = extra investment income over SAA – ∆ cost of capital = 2.5 – (–0.8) = 3.3

As the 15-year interest rate decreases by (1/3) percent and MVL is 160, change in MVL is
calculated as (MVL)(–change in interest rate)(duration of liability) = (160)((1/3)%)(15) = $8
million. Therefore, EVA for ALM function is calculated as:

EVAALM = return on SAA – return on replicating portfolio = 15 – 8 = $7 million

Further assume the following experience:

• During the measurement period, the market consistent embedded value of new business
underwritten is $10 million.
• The experience of nonfinancial risk has no deviation from the best-estimate assumption.

©2012 Society of Actuaries, All Rights Reserved Page 49 of 57

 As calculated above, return on the replicating portfolio due to the shock of the interest rate is
$8 million. However, it is 100 percent offset by the increase in MVL and should not be
included in EVA for the business management function. Expected return on the replicating
portfolio is the expected 15-year interest rate (4 percent) times MVL. The maximum of
statutory-required capital, rating agency-required capital and required economic capital at the
desired probability is assumed to be $40 million. Cost of capital is (cost of capital
rate)(required capital) = (4%)(40) = $1.6 million. EVA for business operations is calculated
as:

EVAbus = MCEV of new business + expected return on replicating portfolio + experience G/L
– cost of capital = 10 + (4%) × (160) + 0–1.6 = $14.8 million

Another important step in performance measurement is to set the appropriate EVA target. This
should be consistent with strategic planning. After the strategic and business plans are decided,
the expected EVA can be set as the target. Realized EVA is compared with expected EVA to
evaluate the performance of different functions.

At this point, a risk-adjusted value framework is established for performance measurement.

However, this is the simplest part. The most difficult part of introducing a risk-adjusted
measure to the performance scorecard is communication with the management team and
getting their buy in and agreement on all the assumptions used in the determination of the
EVA target. Interactive workshops with senior management from business units and core
departments are necessary to help them understand EVA and the assumptions and calculation
methods used. Without fully understanding how it is calculated, it is hard to persuade people
to accept linking it with their compensation. With full understanding and buy in, management
is better able to make business decisions under a risk-adjusted return/value framework. The
communication is a long and lasting process. Gradually selling the idea to the management
team will have a long-term and beneficial impact on business management and risk culture.

8 Conclusion

Regulators and rating agencies have been emphasizing the integration of risk appetite and
business strategy for quite a few years. A disconnect between the company’s risk appetite and
business strategy was not uncommon, as has been seen since the financial crisis in 2008.
Strategic decision making is very critical in managing and developing business. Awareness of
the risks embedded in these planned business activities will improve the process as
demonstrated in the paper. A risk appetite framework provides a holistic picture of the
company’s willingness and ability to take risk. It can improve strategic planning from two
perspectives:

• It embeds the risk perspective in decision making and therefore may reduce the chance of
making bad and risky strategic decisions that the company, the regulator or the market are

©2012 Society of Actuaries, All Rights Reserved Page 50 of 57

 not ready for.
• Its analysis of the risk/reward tradeoff and quantification of their impact on capital
position and earnings may make it easier for senior management to make better decisions.

Risk appetite influences business decisions almost everywhere. It also changes the way
business is managed. More focus is put on the overall impact of these decisions on the
company’s solvency, earnings and risk-adjusted value. This is clearly demonstrated in the case
studies. The examples in the case studies have a focus on quantitative analysis. When
incorporating risk appetite framework in making strategic plans, qualitative perspective and
quantitative perspective are equally important. Though quantitative analysis provides more
detailed information, model risk and assumption setting are always a challenge.

©2012 Society of Actuaries, All Rights Reserved Page 51 of 57

 9 References

ACE Ltd. “Twenty-five Years of Insuring Progress, 2010 Annual Report.” (February 25,
2011).

Admiral Group PLC. “Annual Report 2010.” (2010):

http://www.admiralgroup.co.uk/pdf/annualreports/2010.pdf.

American International Group Inc. (AIG). “2010 Annual Report.” (2010):

www.aigcorporate.com/investors/AIG_2010_Annual_Report.pdf.

Allianz Group. “Annual Report 2010.” (March 18, 2011):

https://www.allianz.com/static-resources/en/investor_relations/reports_and_financial_data/an
nual_report/ar2010/v_1310727324000/ar2010_group.pdf.

Allstate Corp. “Notice of 2011 Annual Meeting, Proxy Statement and 2010 Annual Report.”
(2010):
http://www.allstate.com/Allstate/content/refresh-attachments/annual-reports/Allstate_Summar
y_Annual_Report_2010.pdf.

Aviva PLC. “Putting you in the Picture, Annual Report and Accounts 2010.” (2010):
http://www.aviva.com/library/pdfs/reports/2010/aviva_annual_report_2010.pdf.

AXA Group. “Registration Document — Annual Report 2010.” (2010):

http://www.axa.com/lib/en/uploads/anr/group/2010/AXA_Reference_Document_2010c.pdf.

Barfield, Richard. “Risk appetite — How Hungry are You?” The Journal: Special Risk
Management Edition (2007):
https://www.pwc.com/en_GX/gx/banking-capital-markets/pdf/risk_appetite.pdf.

Barnes, Richard. “Criteria | Insurance | General: Evaluating Risk Appetite: A Fundamental

Process of Enterprise Risk Management.” Standard and Poor’s Global Credit Portal
RatingsDirect (October 31, 2006):
http://www.standardandpoors.com/ratings/articles/en/us/?articleType=PDF&assetID=124519
3944624.

Basel Committee on Banking Supervision (BCBS). “Sound Practices for Managing Liquidity
in Banking Organizations.” (February 2000):
http://www.bis.org/publ/bcbs69.pdf.

Bennet, Caroline, and Kelly Cusick. “Risk Appetite: Practical Issues for the Global Financial
Services Industry.” Presentation to the Institute of Actuaries of Australia (September 2007).

©2012 Society of Actuaries, All Rights Reserved Page 52 of 57

 Bergeron, David. “Risk Appetite: Embedding for Value Creation.” Presentation for Oliver
Wyman. (March 15, 2010).

Berkshire Hathaway Inc. “2010 Annual Report.” (2010):

http://www.berkshirehathaway.com/2010ar/2010ar.pdf.

Bodoff, Neil M. “Capital Allocation by Percentile Layer.” Variance 3, no. 1 (2009): 13-30,
http://www.variancejournal.org/issues/03-01/13.pdf.

Bowser, Marcus, Howard L. Rosen, and Laura Santorini. “Criteria | Insurance | General:
Refined Methodology for Assessing an Insurer's Risk Appetite.” Standard and Poor’s Global
Credit Portal RatingsDirect (March 30, 2010):
http://www.standardandpoors.com/ratings/articles/en/us/?articleType=PDF&assetID=124520
9863982.

Carey, Mark. “Determining Risk Appetite.” (updated May 5, 2005):

http://www.continuitycentral.com/feature0170.htm.

Centre for Risk and Insurance Studies. “Risk Appetite: What’s it All About?” CRIS Line 17
(February 2009): www.nottingham.ac.uk/business/cris/crisline_newsletter_no17.pdf.

Chaplin, Mark. “Risk Appetites for Insurers.” Watson Wyatt Worldwide presentation to the
Actuarial Society of Greater New York (May 16, 2007).
Chase-Jenkins, Linda, and Ian Farr. “Risk Appetite: A Boundary for Decisions.” Emphasis
(January 2008):
http://www.towersperrin.com/tp/showdctmdoc.jsp?url=master_brand_2/global/News/Emphasi
s/2008/01/emp_q1_art6.htm&country=global.

Chase-Jenkins, Linda, Ian Farr, and Joe Lebbens. “Risk Appetite: The Foundation of
Enterprise Risk Management.” Towers Watson (2010).

Chase-Jenkins, Linda, Eric Gesick, and Joe Lebens. “Risk Appetite.” Towers Perrin webcast
(July 14, 2009): http://www.slideshare.net/Towers_Perrin/risk-appetite.

The Chief Risk Officer Forum. “A Framework for Incorporating Diversification in the
Solvency Assessment of Insurers.” (June 10, 2005).

Committee of European Insurance and Occupational Pensions Supervisors (CEIOPS).

“CEIOPS’ Advice for Level 2 Implementing Measures on Solvency II: System of
Governance.” (draft) Consultation Paper No. 33 (March 26, 2009):
https://eiopa.europa.eu/fileadmin/tx_dam/files/consultations/consultationpapers/CP33/CEIOP
S-CP-33-09-Draft-L2-Advice-on-Governance.pdf.

———. “QIS5 Technical Specifications.” (July 5, 2010):

©2012 Society of Actuaries, All Rights Reserved Page 53 of 57

 https://eiopa.europa.eu/fileadmin/tx_dam/files/consultations/QIS/QIS5/QIS5-technical_specif
ications_20100706.pdf.

CRO Forum. “Calibration Recommendation for the Correlations in the Solvency II Standard
Formula.” (December 10, 2009).

———. “Insurance Risk Management Response to the Financial Crisis.” (April 2009):
http://www.thecroforum.org/assets/files/publications/croforum-insuranceriskmanagementresp
onsetothefinancialcrisis20090423(2).pdf.

———. “Market Consistent Embedded Value Principles.” (October 2009).

Deloitte & Touche LLP. “Global Risk Management Survey: Fifth Edition — Accelerating
Risk Management Practices.” (2007):
http://www.deloitte.com/assets/Dcom-Albania/Local%20Assets/Documents/al_fsi_global_ris
k_mgmt_survey2007.pdf.

European Union. “Directive of the European Parliament and of the Council on the Taking-up
and Pursuit of the Business of Insurance and Reinsurance (Solvency II) (Recast).” (November
25, 2009).

Ernst & Young. “Risk Appetite: The Strategic Balancing Act.” 2010.

Froot, Kenneth A., Brian S. Murphy, Aaron B. Stem, and Stephen E. Usher. “The Emerging
Asset Class: Insurance Risk.” special report from Guy Carpenter and Co. Inc. (July 1995).

Gai, Prasanna, and Nicholas Vause. “Risk Appetite: Concept and Measurement.” Financial
Stability Review 17 (December 2004):
http://www.bankofengland.co.uk/publications/fsr/2004/fsr17art12.pdf.

General Reinsurance AG. (Gen Re). “2010 Annual Review.” (2010):

http://www.genre.com/sharedfile/pdf/GRAG_AR2010-en.pdf.

HDI-Gerling Industrie Versicherung AG. (HDI-Gerling). “2009 Annual Review.” (2009):

http://www.talanx.com/docs/Publikationen/2010/gb_hg_i_2009_en.pdf.

ING Group. “Shaping Our Future, 2010 Annual Review.” (2010):

http://www.ing.com/web/file?uuid=0165217a-76a0-4613-b95f-1c0aa7e5c751&owner=b03bc
017-e0db-4b5d-abbf-003b12934429&contentid=8984.

Jones, Rob, David Ingram, Mark Puccia, and Polina Chernyak. “Criteria | Insurance |
Property/Casualty: Nonlife Insurance Risk Control Criteria and Their Role in Enterprise Risk
Management.” Standard and Poor’s Global Credit Portal RatingsDirect (October 21, 2007):
http://www.standardandpoors.com/prot/ratings/articles/en/us/?articleType=HTML&assetID=1

©2012 Society of Actuaries, All Rights Reserved Page 54 of 57

 245319391734.

Jones, Rob, Mark Puccia, and Laura Santorini. “Criteria | Insurance | General: Strategic Risk
Management: The Upside of ERM.” Standard and Poor’s Global Credit Portal RatingsDirect
(July 27, 2006):
http://www.standardandpoors.com/prot/ratings/articles/en/us/?articleType=HTML&assetID=1
245319349570.

Kimball, Ralph C. “Economic Profit and Performance Measurement in Banking.” New

England Economic Review (July/August 1998):
http://www.bostonfed.org/economic/neer/neer1998/neer498c.pdf.

KPMG Europe LLP. “Understanding and Articulating Risk Appetite.” (2010):

http://www.kpmg.com/UK/en/IssuesAndInsights/ArticlesPublications/Documents/PDF/Advis
ory/193222_Understanding_Risk.pdf.

Leitch, Matthew. “Clear Thinking and ‘Risk Appetite.’” International Risk Management
Institute Inc. (IRMI.com) (April 2007):
http://www.irmi.com/expert/articles/2007/leitch04.aspx.

———. “Making Sense of Risk Appetite, Tolerance, and Acceptance.” 2nd ed. (August 9,
2010): http://www.internalcontrolsdesign.co.uk/appetite/index.shtml.

Lloyd's. “Lloyd's Annual Report 2010.” (2010):

http://www.lloyds.com/Lloyds/Investor-Relations/Financial-performance/Annual-Reports/201
0-Annual-Report.

MetLife, Inc. “Annual Report.” (2010):

http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9NDE5MDc1fENoaWxkSUQ
9NDMyMjQzfFR5cGU9MQ==&t=1.

Manulife Financial. “For Your Future, 2010 Annual Report.” (2010):

http://www.manulife.com/public/files/202/1/mfc_annualreport2010.pdf.

Munich Re Group. “Thinking Further, Annual Report 2010.” (2010):

http://www.munichre.com/publications/302-06773_en.pdf.

Nakada, Peter, Hemant Shah, H. Ugur Koyluoglu, and Olivier Collignon. “P&C RAROC: A
Catalyst for Improved Capital Management in the Property and Casualty Insurance Industry.”
The Journal of Risk Finance 1, no. 1 (1999).

PartnerRe. “Continuity and Adaptability: Striking the Right Balance, Annual Report Summary
2010.” (2010): http://www.partnerre.com/ar10/files/PartnerRE_AR2010.pdf.

©2012 Society of Actuaries, All Rights Reserved Page 55 of 57

 Prudential PLC. “Long-term Thinking, Annual Report 2010.” (2010):
http://www.investis.com/prudential-plc/storage/ar2010.pdf.

Prudential Financial Inc. “2010 Annual Report.” (2010):

www3.prudential.com/annualreport/report2011/annual/.

Rosen, Howard L., and Mark Puccia. “Criteria | Insurance | General: Methodology: Assessing
Management's Commitment to and Execution of Enterprise Risk Management Processes.”
Standard and Poor’s Global Credit Portal RatingsDirect (December 17, 2009):
http://www.standardandpoors.com/prot/ratings/articles/en/us/?articleType=HTML&assetID=1
245319411135.

Santori, Laura, Keith Bevan, and Christopher Myers. “Criteria | Insurance | General:
Summary of Standard & Poor's Enterprise Risk Management Evaluation Process for
Insurers.” Standard and Poor’s Global Credit Portal RatingsDirect (November 26, 2007):
http://www.standardandpoors.com/prot/ratings/articles/en/us/?articleType=HTML&assetID=1
245319388290.

Santorini, Laura, Mark Puccia, and Rob Jones. “Criteria | Insurance | General: Refining the
Focus of Insurer Enterprise Risk Management Criteria.” Standard and Poor’s Global Credit
Portal RatingsDirect (June 2, 2006):
http://www.standardandpoors.com/ratings/articles/en/us/?articleType=PDF&assetID=124519
3985484.

Segal, Sim. “Defining Risk Appetite.” ERM Symposium presentation. (March 29, 2007).

Senior Supervisors Group. “Observations on Developments in Risk Appetite Frameworks and

IT Infrastructure.” SSG report (December 23, 2010):
www.cmra.com/docs/2011_SSG_Risk_Report.pdf.

Standard Life PLC. “Annual Report and Accounts 2010.” (2010):

http://ir2.flife.de/data/standardlife/igb_html/index.php?bericht_id=1000001&lang=ENG.

Stoddard, Shellie, Karin Clemens, David Ingram, and Michael Vine. “Criteria | Insurance |
General: Extending the Insurance ERM Criteria to the Health Insurance Sector.” Standard and
Poor’s Global Credit Portal RatingsDirect (November 8, 2006):
http://www.standardandpoors.com/prot/ratings/articles/en/us/?articleType=HTML&assetID=1
245319334773.

Sun Life Financial Inc. “Working to Build a Brighter Future, Annual Report 2010.” (2010):
http://cdn.sunlife.com/static/global/files/Annual%20reports/2010/Full_report_Eng.pdf.

Swiss Re. “Innovations for Success: Swiss Re in 2010, 2010 Annual Report.” (2010):
http://www.swissre.com/annualreport/.

©2012 Society of Actuaries, All Rights Reserved Page 56 of 57

 Thirlwell, John. “Risk Appetite.” PowerPoint presentation (October 11, 2007):
www.johnthirlwell.co.uk/risk_appetite07.pdf.

Zurich Financial Services Group. “Annual Report 2010, Shaping Zurich for Tomorrow.”
(2010):
http://www.zurich.com/internet/main/SiteCollectionDocuments/financial-reports/annual-repor
t-2010-en.pdf.

Mildenhall, Stephen. "Correlation and Aggregate Loss Distributions with an Emphasis on the
Iman- Conover Method," part one of “The Report of the Research Working Party on
Correlations and Dependencies Among All Risk Sources.” Casualty Actuarial Society Forum
(Winter 2006):
http://www.casact.org/pubs/forum/06wforum/06w105.pdf

©2012 Society of Actuaries, All Rights Reserved Page 57 of 57