Shane Whelan, L527

1
Asset-Liability Management for Actuaries

• Shane Whelan, L527

2
Preface

• This course, given jointly with Mr Gareth Colgan,
  FIA.
• It is based on the Core Applications - Concepts
  (CA1) course of the Faculty Institute of
  Actuaries
• It is not an exemption subject.
• Integrates all areas of actuarial practice with
  the concept of the actuarial control cycle.
• Gives an introduction to traditional areas of
  actuarial expertice.

3
Chapter 1

• The Actuarial Control Cycle

4
Actuarial Control Cycle
Specifying the Problem
Developing a Solution
Monitoring the Experience
5
Actuarial Control Cycle
Specifying the Problem
Developing a Solution
Monitoring the Experience
6
Actuarial Control Cycle

• Actuarial Control Cycle is a schema based on a
  scientific approach to problem solving
• first, specify the problem,
• second design and implement a solution,
• third monitor the effectiveness of the solution
• But it does not end here
• Typically, revision is necessary hence the loop
  or cycle.
• The Actuarial Control Cycle must of course be
  considered in the context of the economic and
  commercial environment (e.g. legislation,
  taxation, business ethics).
• Also, as professionals, actuaries must adhere to
  requirements of professionalism.

7
Specifying the problem

• This stage of the control cycle will give an
  assessment of the risks faced and how they can be
  handled. It also provides an analysis of the
  options for the design of plans to provide
  various benefits and the prices of the plans that
  will result. It also considers the strategic
  courses of action that could be used to handle
  the particular risks in question.

8
Developing a Solution

• An examination of the major (actuarial) models
  currently in use and how they may be adjusted for
  the particular problem to be solved.
• Selection of the most appropriate model to use
  for the problem, or construction of a new model.
• Consideration and selection of the parameters to
  be used in the model.
• The parameters used in a model are critical and
  it is necessary to have a good understanding of
  their sensitivities.
• Interpretation of the results of the modelling
  process.
• Consideration of the implications of the model
  results on the overall problem. Consideration of
  the implications of the results for all
  stakeholders.
• Determining a proposed solution to the problem.
• Consideration of alternative solutions and their
  effects on the problem.
• Formalising a proposal.

9
Monitoring the experience

• It is critical that the models are updated to
  reflect current experience.
• This stage deals with the monitoring of
  experience and its feedback into the problem
  specification and solution development stages of
  the control cycle, i.e. updating the
  investigation.
• An important part of this monitoring will be the
  identification of the causes of any systemic
  departure from the targeted outcome from the
  model and a consideration as to whether such
  departures are likely to recur.
• Hence identify new problem and around the cycle
  again.

10
Applications/Examples of Actuarial Control Cycle
0

• Problem It is the 19th century and a mutual
  friendly society approaches you to design a
  simple scheme to provide for funeral expenses of
  members in consideration for a weekly premium of
  P.
• Formulate your proposal and put in place a plan
  to monitor the emerging experience.

11
Question

• Is the Actuarial Control Cycle not just the
  scientific method applied in the actuarial domain
  model, test, use results to build better model?

12
Critique of the Actuarial Control Cycle

• The Actuarial Control Cycle, while it is useful
  as an educational framework, would seem to be the
  epitome of an abstraction which has been emptied
  of content.
• Bellis, C. S. (2000), Professions in Society,
  British Actuarial Journal, 6, II, 317-344. Quote
  is from Section 10.3.17.
• In Section 10 the author reveals the true nature
  of the emperors new clothes the Actuarial
  Control Cycle is not actuarial at allThe
  Actuarial Control Cycle, so valued around the
  world, is a content-free zone.
• Hardy, M.R. (2000) Discussion of above paper, in
  BAJ, 6, II, p. 353.

13
Other Control Cycles

• What other control cycles are there?
• Especially those applied in a business context.
• Two-generic types
• Mathematically-assisted decision-making
• Mathematically-inspired decision-making
• The following few slides are adapted from those
  of Dr Maciej Klimek based on a seminar he gave at
  Dublin Institute for Advanced Studies, December
  2004.

14
Mathematically assisted decision making
The decision process uses results of computations
only as arguments
Required goals
Separation of tasks!!!
15
Mathematically inspired decision making
Highly quantitative and logical approach to
problem solving, with no spurious assumptions
Mathematical-like reasoning is an integral part
of the decision process!
16
Real discrete dynamical system
17
Characteristics of decision making in the spirit
of dynamical systems

• Modelling of reality as a collection of time
  dependent processes
• That is, modelled as a stochastic process with an
  associated filtration
• Awareness of non-linearity/sensitivity to initial
  conditions
• Adaptability regarded as one of the highest
  values
• Must update model and actions with filtration.
• Reliance on iterative decision loops whenever
  possible
• Embracing change as opportunity for betterment.
  

18
John R. Boyd (1927-1997) Strategist and fighter
pilot
What is the aim or purpose of strategy?  To
improve our ability to shape and adapt to
unfolding circumstances, so that we (as
individuals or as groups or as a culture or as a
nation-state) can survive on our own
terms. Boyd The most important thing in life
is to be free to do things. There are only two
ways to insure that freedom you can be rich or
you can you reduce your needs to zero. I will
never be rich, so I have chosen to crank down my
desires. The bureaucracy cannot take anything
from me, because there is nothing to take. Boyd
quoted in Spinney, F.C. (1997) Genghis John,
Proceedings of the U. S. Naval Institute, July,
pp 42-47.
Boyd in 1953
Time is a weapon
For more on Boyd, including all his lectures and
notes on strategy, see www..belisarius.com/boyd.h
tm
19
Boyds Loop OODA Observation Orientation
Decision - Action
20
Real discrete dynamical system (OODA)
21
Other Decision Loops

• The Six-Sigma Loop (Japanese roots, developed by
  General Electrics and Motorola)
• Define-Measure-Analyze-Improve-Control
• The six-sigma methodology is closely related to
  the Japanese methodologies Lean Thinking and TQM
  (Total Quality Management).
• The common sense PRIME loop
• Planning-Rehersal-Implementation-Maintainance-Evol
  ution
• The Lean Thinking 5S management rule
• SEIRI clear organization without redundant
  elements
• SEITON order, everything has its place
• SEISO absolute cleanliness to avoid errors
• SEIKETSU clarity the result of the first 3S
• SHITSUKE discipline

22
Applications/Examples of Actuarial Control Cycle
1

• Problem An established pension fund with
  significant assets wants to design and implement
  a suitable investment strategy.

23
Actuarial Control Cycle
Specifying the Problem
Developing a Solution
Monitoring the Experience
24
Actuarial Control Cycle
Analyse the Risk Assess Clients
Situation Consider high-level options
Specifying the Problem
Developing a Solution
Monitoring the Experience
25
Actuarial Control Cycle
Analyse the Risk Assess Clients
Situation Consider high-level options
Specifying the Problem
Monitoring the Experience
Developing a Solution
Built model and discuss assumptions Calculate and
intepret results Determine solution proposed and
main alternatives identified Formal
proposal/report
26
Actuarial Control Cycle
Analyse the Risk Assess Clients
Situation Consider high-level options
Specifying the Problem
Compare with model Review model Review business
context Identify scope for improvement
Monitoring the Experience
Developing a Solution
Built model and discuss assumptions Calculate and
intepret results Determine solution proposed and
main alternatives identified Formal
proposal/report
27
Investment Advisory Process
Investmentobjectives
Strategicassetallocation
Monitoring Evaluation
Fund managerstructure
Fundmanager selection
28
Completes Chapter 1

• The Actuarial Control Cycle

29
Asset-Liability Management for Actuaries

• Shane Whelan, L527