INTRODUCTORY TRACK
The Introductory Track sessions will deliver a comprehensive introduction to DFA.
Familiarity with basic financial statements and concepts will be assumed, but depth and
subtleties will be limited in the interest of education. Presenters will assume no previous
exposure to DFA. The emphasis will be on "dynamics" or thinking about what could
possibly happen and probabilities associated with possible outcomes.
The Value Proposition of DFA
This session will provide various perspectives on ways in which performing DFA analysis
has provided value within an organization. Senior executives from insurance companies
and consulting firms will share their experiences in using DFA-type analyses to address a
variety of real-world problems. The discussion will highlight the ways in which DFA
helped identify solutions that weren't otherwise obvious and how suboptimal decisions
may have been made if DFA had not been performed.
Panelists:
Susan E. Witcraft, RTW, Inc.
Manuel Almagro, Guilford Specialty Group
DFA Implementation Issues
Management does not wake up one morning and say, "We will implement DFA." The
decision to implement DFA usually involves input from every corner of the company, after
careful consideration of the intended purpose and the many issues that must be addressed.
In this session, the panelists will discuss some unique and unforeseen obstacles that their
companies faced in the implementation process, and whether these obstacles were
successfully addressed. The panelists will also discuss alternative courses of action and
decisions that would have enhanced the success of the DFA implementation project.
Panelists:
Gerald S. Kirschner, Classic Solutions Risk Management, Inc.
Elizabeth R. Wiesner, Accident Fund Company
Mary E. Wills, United Services Automobile Association
Important Considerations in the Use of DFA
Panelists will share their experiences on many decisions that have to be made when
considering the use of DFA. This will include their thoughts on how they intend to apply
models in their decision making and what they look for in deciding whether to purchase a
model or build their own. The models that are available from various sources have
similarities, but have important differences as well. Economic assumptions imbedded in
models are often considered proprietary. How much due-diligence work should be
performed in evaluating these models?
Panelists will discuss how they tested various models for sensitivity to key inputs, how
they considered cost of "buying or building," and how useful they feel DFA is in helping
them make decisions.
Moderator:
Thomas E. Hettinger, MHL/Paratus
Panelists:
Kevin G. Dickson, Allstate Insurance Company
Elizabeth R. Wiesner, Accident Fund Company
A Basic Model for DFA
The content of this session is based on a paper submitted to the 1997-99 DFA Call
Paper Programs. The presenters will assume that the audience has no prior exposure to a
working DFA model. In this session, attendees will learn about a workable, publicly
available DFA model that seeks to reflect measurable risks that are most relevant to
property and liability insurers in the United States. Model parameters will be explored
with emphasis on user decision making. The final half hour is set aside for group
discussion and interactive changes to the model.
Panelists:
Richard W. Gorvett, University of Illinois
Shawna S. Ackerman, MHL/Paratus
Interest Rates: Empirical Patterns and Modeling Issues
One of the key components of a DFA model is the specification of the interest rate
process. The importance of interest rate modeling stems from the fact that many other
economic, financial, and insurance processes are correlated—either directly or
indirectly—with interest rates. This session will look at historical interest rate patterns and
will consider a variety of models that have been proposed to represent future interest rate
emergence. In addition, the relationship between interest rates and inflation will be
explored. This relationship has important implications for property and liability insurers
because changes in interest rate levels are correlated with inflation, which in turn affects
the magnitudes of future insurance claim payments.
Panelist:
Kevin C. Ahlgrim, University of Illinois
Stephen Britt, Tillinghast - Towers Perrin
STRATEGIC TRACK
The Strategic Track sessions will focus on vital uses of DFA in overall company decision
making. The emphasis will be on practical ways of "putting DFA into play," not model
specification. Prerequisite knowledge of DFA techniques should be minimal for most of
these presentations.
Assessing Balance Sheet Protection Using DFA
Three speakers will discuss how DFA can be utilized to assess the risk on a company's
balance sheet. One panelist will address process modeling as the key to better linking
asset and liability risk factors. Discussions beginning with examples of process modeling in
the economic scenario generator will expand to liability process modeling: loss ratios,
diversifiable and systematic factors, linkages with economic variables, and responsiveness
of pricing. Another panelist will present a case study that illustrates how DFA is used to
evaluate the effectiveness of alternative risk management vehicles. A third panelist will
discuss how DFA is utilized to analyze the balance sheet impact of a multiyear reinsurance
transaction, which protects against both underwriting and capital markets exposures.
Panelists:
Donald F. Mango, American Reinsurance
Joan Lamm-Tennant, General Reinsurance Corporation
Daniel Isaac, Swiss Re Investors
Using DFA To Optimize the Value of Reinsurance
The first speaker will explore the effects caused by a shift from proportional to
nonproportional reinsurance. To determine which reinsurance program is better suited for
an insurance portfolio, one must study and model not only the liabilities but also the assets
and the business strategies. This will be illustrated by applying a DFA model to an
example company. The second speaker will present a practical approach that uses DFA
to evaluate alternative reinsurance programs, by assessing the trade-offs between risk,
capital, and value. DFA modeling is used to measure the ability of each component of the
reinsurance program to reduce the volatility of financial results and to protect capital. With
the specification of a suitable objective function and constraints, the optimum reinsurance
program can be found. Alternative objective functions for the optimization, including one
based on maximizing economic value, will be considered.
Panelists:
Markus Stricker, Swiss Reinsurance
Nathan Schwartz, E.W. Blanch, Inc.
Using Catastrophe Modeling in DFA
Companies subject to significant catastrophe losses are continually evaluating their
appetite and tolerance for risk. A sound process of linking catastrophe modeling results to
financial output is key to understanding the implications of risk management strategies.
Therefore, dynamic financial models are playing an increasingly important role in
evaluating catastrophe management strategies.
The first part of this session will focus on the differences between catastrophe models
used in the insurance industry and how to compensate for those differences. The second
part of this session will focus on how to use modeled catastrophes in a dynamic financial
model, simulation techniques, and issues in modeling catastrophe reinsurance structures.
Moderator:
Elizabeth E.L. Hansen, Guy Carpenter
Instrat
Panelist:
Laura A. Esboldt, E.W. Blanch Company
Securitized Risks: Engineering and Empirical Analysis with DFA Concepts
During the last several years, securitized products have become an alternative source of
capital for the insurance industry. The first part of this session will focus on the basics of insurance linked securities including
typical terminology and pricing factors. The effect on the cedant's balance sheet as well as
pros and cons versus traditional reinsurance will be discussed. The second part of this
session will present a case study of how a medium-size primary insurer used DFA
concepts to evaluate a catastrophe reinsurance program with both private and public
components, the performance of which relies heavily on public risk securitization. The
implications for reinsurance buyers as well as the bond markets and the borrowers will be
discussed.
Panelists:
Paul Puleo, Lehman Brothers Inc.
John W. Rollins, Florida Farm Bureau Insurance
Insurance Pricing and Performance Measurement: Return on Invested Capital
and Economic Value Added
DFA lets insurers better manage the returns on capital employed to support insurance
operations. Panelists will describe some of the procedures now being used by a major
property and casualty insurer. These procedures will include determining the capital used
to support the various segments of the company's business, pricing the insurance contracts
for an appropriate return on that capital, and rewarding management performance based
on the economic value added by the insurance operations.
Panelists:
Douglas M. Hodes, Liberty Mutual Group
Sholom Feldblum, Liberty Mutual Group
Neeza Thandi, Liberty Mutual Group
Anya Sri-Skanda-Rajah, Liberty Mutual Group
Strategies for Ranking DFA Results
DFA is often used to provide quantitative insights into strategic business questions. These
questions include:
- Which reinsurance program is better?
- Which asset mix is preferred?
- What is the impact of changing product mix?
- What are the ramifications of changing pricing strategy?
To assist in making these decisions, the various alternatives need to be prioritized. This
presentation will provide several ideas for ranking strategies in the context of a case study,
focusing on the choice of investment strategy. The approaches suggested can be applied
to decisions regarding any type of strategy, including those that focus on more than one
aspect of an insurance company's operations, such as a set of strategies for mitigating risk
through combinations of changes in the mix of business and reinsurance programs.
Panelist:
Susan E. Witcraft, RTW, Inc.
Measuring Risk/Reward Trade-offs
Traditional financial planning and asset management models do not have DFA's
capabilities to reflect the risk-vs.-reward trade-offs between alternative business growth
and asset allocation options. These trade-offs include the risks embedded in varying
premium growth plans among lines of business and/or the company's changing its asset
allocations between bonds and equities. With DFA, the analyst can simulate the likelihood
that a given plan will produce an outcome within an acceptable range and the probability
that selected measures of success will be achieved. For example, with DFA the potential
trade-off between a consistent NWP-to-surplus target versus maximizing growth in the
company's future economic value, can be tested under varying asset portfolio structures.
In this session the panelists will discuss measures of success and risk-vs.-reward that they
have applied and demonstrate DFA approaches and tools for testing alternative strategies
against these measures.
Panelist:
Stephen M. Sonlin, Swiss Re Investors
DFA, Financial/Strategic Planning, and Risk Management
Planning requires the evaluation of options available to the company for growth and
profitability. Risk management seeks to minimize the potential that unexpected adverse
events could produce unacceptable outcomes. This panel will discuss how DFA tools,
such as those described in the previous session, "Measuring Risk/Reward Trade-Offs,"
are being applied to evaluate the relative risks among strategic options. Such evaluations
are in the context of testing the potential for producing unacceptable outcomes relative to
the company's measures of success. The session will touch upon how companies use this
information to draw conclusions, or to seek risk management solutions to minimize the
risks intrinsic to an otherwise attractive plan.
Moderator:
Stephen M. Sonlin, Swiss Re Investors
PERSPECTIVES TRACK
The Perspectives Track sessions will attempt to provide some different outlooks and
perspectives on DFA. Regulators, rating agency representatives, and actuaries from other
countries will provide some ideas that are a little different from those of North American
property and casualty actuaries working directly or indirectly for insurance companies.
Prerequisite knowledge of DFA techniques should be minimal for most of these
presentations.
How DFA May Affect Your Company's Rating
DFA is becoming an increasingly important risk management and capital optimization tool
for insurance companies in evaluating decisions, such as finding the optimal reinsurance
program and determining financial planning. Asset portfolio and strategic planning
decisions can also be enhanced by DFA analyses. Given that a company sees value in
using DFA internally to evaluate corporate initiatives and capital adequacy, a key question
is the degree to which rating agencies are interested in DFA for their external valuation
purposes. Can internal DFA studies be used by the company to garner an upgrade or
stave off a downgrade? Can DFA models be used in meetings with rating agencies to
support strategic or tactical changes in a company's business plan, or to support how the
company evaluates and manages risk? Our panelists will address these and related
questions. The panelists will also discuss the role that DFA and similar quantitative models
play in their rating process.
Panelists:
William M. Wilt, Moody's Investors Service
Mark Puccia, Standard & Poor's
Martin Sheffield, Ward Financial Group
Regulatory Applications of DFA
Insurance regulators have long used stresstesting and similar analyses to gauge the
solvency risk associated with various adverse underwriting and operational outcomes for
individual insurers. As DFA capabilities have developed, such analyses have been utilized
more frequently. Additionally, the scope of such analyses has expanded. In this session,
the panelists will discuss specific instances in which DFA-type analyses have been utilized
in identifying and monitoring distressed insurers. Additionally, the panelists will discuss the
process of reviewing the assumptions and results of insurer-submitted DFA analyses.
Panelists:
Victoria S. Lusk, Colorado Division of Insurance
Chester J. Szczepanski, Pennsylvania Insurance Department
Property and Casualty RAROC: Practical Risk Measurement for ERM
One of the cornerstones of an ERM program is an effective means of measuring all risks
across an enterprise and linking those risk measures to strategic decisions. Over the past
ten years, banks have developed an approach for measuring risk in terms of economic
(required) capital and for measuring profitability in terms of risk-adjusted return on capital,
or RAROC. This session presents property and casualty RAROC, an adaptation of the
banking approach to the property and casualty insurance industry.
The speakers will explain how the banking approach was adapted for the property and
casualty industry, and how this approach takes advantage of the "bruised knuckle"
experience of the banks. The panel will also discuss a case study using the property and
casualty RAROC framework, in conjunction with an insurance company's own data and
"best of breed" actuarial analytics, to arrive at an effective and practical capital allocation
tool for the enterprise.
Moderator:
Alex Krutov, Navigation Advisors
Panelists:
Paul J. Brehm, St. Paul Companies, Inc.
Peter K. Nakada, ERisk
Value-at-Risk Applications in the Insurance Industry
The panel will discuss how Value-at-Risk (VaR) applications measure financial strength
and performance in the insurance industry. Effective risk measurement and capital
management is critical in today's crowded and converging environment. However, the
ability to properly identify risk has become increasingly complex. The first part of this
session will focus on enhancements to A.M. Best's rating process designed to assess
emerging risks, operating performance, and capital adequacy in the rapidly changing
environment of the insurance industry. This session will introduce and demonstrate a new
enterprise risk management tool that has been developed.
The second part of the session will deal with how VaR works as well as how it is used to
measure financial performance, allocate capital, assess risk-adjusted returns (RAROC),
and make restructuring and strategic decisions. The panel will also discuss several cases
that were assessed using VaR technology.
Moderator:
Alex Krutov, Navigation Advisors
Panelists:
Michael L. Albanese, A.M. Best Company
Tim Freestone, Seabury Insurance
TACTICAL TRACK
Tactical Track sessions will focus on detailed, technical specifics of DFA modeling.
Familiarity with DFA modeling fundamentals will be assumed. The emphasis will be on
model specification and sophisticated outcome metrics.
Coherent Risk Measures
To date, DFA research has been conducted relating to parameterization, modeling
processes, and strategic applications. However, somewhere in the DFA process between
the generation of simulations and decision making, there are vital considerations to be
made in terms of what metrics to be used to measure the risk of the firm under study.
Some widely used measures, such as standard deviation and value-at-risk, can be shown
to have properties that are not always consistent with intuitive thinking.
This session draws on points made in an important paper by several international
professors of finance, and translates their points to a DFA context. The presenters will
highlight some basic examples of risk measures that are described by this paper as being
"coherent," in contrast to many widely used measures that fail such a definition. The
panelists will point out why these considerations should be made prior to designing and
applying DFA models. Finally, we will propose a few coherent measures to use in DFA
applications and discuss the merits of their use.
Moderator:
Charles C. Emma, MHL/Paratus
Panelists:
Glenn G. Meyers, Insurance Services Office, Inc.
Andrzej Czernuszewicz, EMB Zysk Ltd./Paratus Consulting
Putting the Power of Modern Applied Stochastics into DFA
From the viewpoint of the applied scientist, DFA is a platform that integrates a variety of
methods and techniques from financial and insurance mathematics, statistics, and
quantitative risk management. This session will give an overview of the points of contact of
DFA with current research in these areas. It will also explore opportunities that some
selected topics offer for the further development of DFA: multivariate stochastic models,
their scaling properties, and the use of high-frequency data for their calibration,
dependence concepts, alternative risk measures, and multiperiod portfolio management.
Moderator:
Alex Krutov, Navigation Advisors
Panelist:
Peter Blum, Zurich Reinsurance
The Cost of Capital for Property and Casualty Insurers
In this session, Professors Doherty and Phillips will discuss recent advances in academic
research on issues related to the cost of capital for insurers.
Professor Doherty will present a theoretical and conceptual discussion of two approaches
in the recent literature that suggests how insurers and other firms set hurdle rates of return
for capital allocation of capital budgeting. One set of models, such as RAROC (Risk
Adjusted Return on Capital), argues that the discount rate should be lowered to reflect
total risk. The competing approach based on models such as the Capital Asset Pricing
Model argues only risk that is undiversifiable in the capital market should be priced. Each
model by itself is incomplete. A dual risk model based on early work of Doherty and later
work of Froot and Stein will be discussed.
Professor Phillips will discuss the objectives and some preliminary results of a research
project sponsored by the CAS Committee on the Theory of Risk, "The Risk Premium
Project." The presentation will focus on the part of the project designed to overcome one
of the more difficult issues related to applying modern capital budgeting techniques for
pricing insurance risks: the general lack of reliable information on the differences in the
systematic risk across the various lines of insurance. The empirical methodology and the
preliminary results of the by-line cost of capital estimates will be presented.
Panelists:
Richard D. Phillips, Georgia State University
Neil Doherty, The Wharton School
Actuarial Worlds Colliding
You know the DFA actuarial world: complex, removed, abstract. Then there is the pricing
actuarial world: stable, dependable, established. What happens when the underwriting
modeling aspects of DFA and innovations in pricing analysis come together? You've got
worlds colliding! Come watch.
Panelists:
Brett M. Nunes, MHL/Paratus
Robert J. Walling, MHL/Paratus
Retention Modeling
Techniques for calculating actuarially indicated rates and assessing competitive position
have become extremely refined. However, assessing the likely customer reaction to a rate
change and the resulting impact on policyholder retention and profitability, has received
little attention. Part of the reason for this is that modeling customer behavior has been
viewed as too complex. As a result, companies have historically used a number of ad hoc
rules in order to assess customer reaction, but these are not always grounded in fact.
However, with advances in computing power and modeling techniques, more rigorous
assessments of customer reaction to a rate change and the resulting impact on profitability
can be made. This panel will discuss retention modeling from two different perspectives.
First, the authors of a recent discussion paper, "Ratemaking for Maximum Profitability,"
will present a framework for retention modeling and how it explains certain elements of
current rate structures. Second, an actual case study involving private passenger
automobile insurance will be presented. As part of the case study, the practical issues
surrounding the development and parameterization of retention models will be presented.
Moderator:
Thomas P. Conway, Ernst & Young LLP
Panelists:
Charles H. Boucek, Ernst & Young LLP
Lee M. Bowron, Independent Consultant
Donald E. Manis, Alfa Mutual Insurance Company
Multidimensional Analysis Profitability and Service Provider Analysis
With new competitors on the horizon, antiquated technologies, slow-moving cultures, and
industry uncertainty, insurance companies now compete in a complex, global, and rapidly
changing post-Glass Steagle environment. In order to survive in this new competitive
landscape, insurance companies need to know their detailed profitability across
permutations and combinations of all key business drivers.
By using extremely multidimensional analyses, multiple operational areas within a property
and casualty insurance company are given the ability to generate very sophisticated,
detailed analyses. By utilizing these analyses, the user is able to spot detailed trends early,
allowing them to act quickly and specifically to achieve better pricing, improved
profitability, and better marketing.
This presentation is intended to provide an introduction on how to utilize extremely
multidimensional analysis to increase profitability by showing real industry examples and
advanced techniques.
Panelist:
Rick Baff, Pinpoint Solutions
Call Papers
The CAS Committee on Dynamic Financial Analysis (DFAC) has been soliciting
responses to a call for papers on the topic of "Dynamic Financial Analysis, A Case
Study." In this call paper program, participants have been presented with a specific
actuarial situation, including a company description and financial statements, and have
been asked to write a paper describing their approach and solution to the current
situation. By giving all participants a common starting point, DFAC has attempted to (1)
encourage creative problem solving by participants using DFA, (2) demonstrate the range
of DFA approaches and models to the CAS membership, and (3) illustrate how
appropriate capital levels can be determined using DFA. Selected authors of accepted
papers are being invited to present their work at the 2001 Special Interest Seminar on
Dynamic Financial Analysis.
In contrast to prior DFA call paper programs, this call has focused on applying DFA
approaches to a given situation in order to illustrate how appropriate capital levels can be
determined. Each participant has been expected to determine appropriate capital levels
for the given insurance company, based on standards used by ratings agencies, regulators,
or financial markets. The capital standard that is adopted for this study is left to the
participant but should be defended as to the appropriateness of its use. It is expected that
each paper will include the following:
- Description of measures of risk and reward used in evaluation;
- Description of strategies considered;
- Description of the model used;
- Description of analytical process; and,
- Interpretation of model results/evaluation of strategies.
Seminar attendees interested in the specific situation, company description, and financial statements can find them on the CAS Web Site at
www.casact.org/research/dfa/dfainsco.htm.
Return to Main Page of Brochure
Photo Credit: Greater Boston Convention and Visitors Bureau
