Challenges of Applied Risk Management

1
Challenges of Applied Risk Management

• by
• Ralph L. Keeney
• Fuqua School of Business
• Duke University

2
What is Applied Risk Management

• Applied Risk Management is making decisions about
  risks. It involves
• Identifying and analyzing risks
• Identifying and evaluating alternatives to deal
  with these risks
• Choosing alternatives to manage risks
• Communicating about the decision.
• ARM uses models and analyses

3
Objectives of the Presentation

• Raise a few general issues about analysis
• Outline applied risk management
• Suggest important aspects of the art of applying
  risk management
• Stimulate some thought (hopefully)
• Have some fun!

4
Why Bother to Make Decisions?

• Your decisions are the only way that you can
  purposefully influence anything.
• Anything includes the quality of your life, your
  family, your organization, your country, or your
  world is through the decisions you make.
• Everything else just happens.

5
True or False

• It is reasonably likely that a moderate to
  large hurricane will cause significant damage to
  the Mid-Atlantic region sometime in the
  foreseeable future.

6
Please Define the Following

• reasonably likely
• moderate to large hurricane
• significant damage
• the Mid-Atlantic region
• the foreseeable future.

7
True or False

• It is reasonably likely that there will be a
  moderate to large earthquake in the San Francisco
  Bay Area in the near future.

8
Please Define the Following

• Reasonably likely
• Moderate to large earthquake
• San Francisco Bay Area
• Near future

9
Why Build a Model and Do Analysis?

• The real problem is too complex to understand all
  of its parts and their relationships.
• A model is a simplification of the problem.
• The model must address the complexities of the
  problem in order to provide insights about them.
• You can analyze the model to produce insights
  relevant to the real problem.

10
Complexities of Risk Decisions

• Multitude of impacts (multiple objectives)
• Some intangible impacts (hard to measure)
• Requires multiple disciplines
• Long-term impacts
• Uncertainties about impacts
• These are factual uncertainties
• Professional judgments required
• Several Stakeholders
• Values are crucial
• Aggregation of impacts is necessary
• Policy (value) judgments required
• Some vexing value tradeoffs
• Unclear decision process and decision-makers
• Need to justify decisions

11
Decision Analysis Addresses the Complexities
Inherent in ARM

• A formalization of common sense applied to
  decision problems.
• A philosophy, articulated by logical axioms, and
  the techniques and procedures, based upon those
  axioms, for analyzing complexities inherent in
  decision problems.
• Decision analysis is prescriptive vs. descriptive.

12
Keys to Effective Decision Making

• Work on the right decision problem
• Specify your objectives
• Create imaginative alternatives
• Understand the consequences
• Grapple with your tradeoffs
• Clarify your uncertainties
• Account for your risk tolerance
• Consider linked decisions

13
Elements of Decisions

• Problem
• Objectives
• Alternatives
• Consequences
• Tradeoffs
• Uncertainty
• Risk Tolerance
• Linked Decisions

Iteration among the elements is crucial in
analysis
14
Problem

• Given a decision problem then
• Initial problem statement (frame)
• Reframed with objectives and alternatives
• What is the risk
• Who is/are the decision-maker(s)
• Who are the stakeholders

15
Problem Examples

• Nuclear Repository best site or portfolio of
  three
• New Orleans Levees build a new levee, how high
  to build, how strong to build
• How to effectively communicate about a hurricane
  approaching Florida

16
Objectives

• Objectives state what you want to achieve by
  making a decision
• Define objectives using a verb and an object
• Minimize environmental degradation
• Limit loss of life
• Minimize property damage
• Objectives are elicited from individuals

17
Ex Choosing a Dissertation Topic
Suppose you have identified ten potential
dissertation topics and now have to pick one.
Write down all the objectives that matter to you
in selecting a dissertation topic.

• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______
• __________________________________________________
  ______

18
Categories to Stimulate Dissertation Objectives
We would like you to think a bit harder about the
objectives that matter when selecting a
dissertation topic. Below you will find four
categories of objectives. Many dissertation
objectives will fall into one of these four
categories. Consider each category and list any
additional objectives that matter to you that you
did not list previously.

• Academic Objectives While A Student
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• Academic Objectives After Graduating
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________

• Personal Objectives While A Student
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• Personal Objectives After Graduating
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________
• ____________________________

OTHER (any other overlooked criteria that are
relevant to deciding on dissertation topic)
AA. ____________________________ AB. _____________
_______________ AC. ____________________________
AD. ____________________________ AE. _____________
_______________ AF. ____________________________
19
Dissertation Study
Check all objectives that are important to your
choice of a dissertation topic.

• ? Is a topic acceptable to faculty for my
  doctorate
• ? Is of interest to a faculty member that I want
  to serve as my advisor
• ? Leads to multiple publishable papers
• ? Is of interest to me/maintains my interest
• ? Can be scoped/is tractable
• ? Uses methods generalizable to other domains
• ? Is interdisciplinary in nature/combines
  different areas
• ? Utilizes my academic strengths
• ? Does not require lots of data gathering
• ? Does not require a lot of data analysis
• ? Provides opportunity to improve my writing
  skills
• ? Addresses issues involving collaboration
  between public and private sectors
• ? Provides opportunity for sufficient
  quantitative analysis
• ? Allows for personal time during dissertation
  years
• ? Is of interest to the research/academic
  community
• ? Is enjoyable to do
• ? Helps me develop myself academically

• ? Helps me build a coherent future research
  program
• ? Helps me balance my career and personal life
• ? Causes me to learn skills that will be
  applicable to future research
• ? Prepares me as an independent researcher
• ? Improves ability to write research proposals
  for funding
• ? Puts me in control of the dissertation process
  (e.g., content, timing)
• ? Is innovative/pursues a new idea/novel
• ? Is insightful/has results that werent obvious
  prior to my work
• ? Is relevant to real-world applications/is
  implementable
• ? Will help people/organizations make better
  decisions
• ? Addresses problems that are important
• ? Influences the work of others
• ? Stimulates discussion with colleagues
• ? Leads to potentially fundable future research
• Provides basis for further research
• ? Opens new areas of research after the
  dissertation
• ? Results in an interesting job talk paper
• ? Provides opportunities to work with top scholars

20
Dissertation Study Results

• self-generated objectives

• 7.1 average unaided, 4.5 additional with
  categories

relevant objectives
21.3 average
recognized objectives
9.3 average
importance
self-generated average 7.7 recognized objectives
average 7.2 bogus objective average 2.1 baseline
objective average 4.2
21
Identifying Objectives

• From decision makers and stakeholders
• Get a complete set of objectives
• Distinguish between fundamental objectives and
  means objectives (max wind speed over land vs.
  fatalities)

22
Identifying Objectives Requires Work

• Individuals miss many objectives
• Experimental results
• Real-world decisions
• Devices to stimulate thinking
• Wish list
• Generic (health and safety, environmental,
  social, economic)
• Use alternatives
• Ask why?
• Involve multiple individuals

23
Alternatives

• Need to create alternatives
• Creative alternatives are needed
• Use objectives to stimulate thinking about
  alternatives (NYFD study)
• Iterate through decision analysis steps
• Delete poorer alternatives
• Embellish/change others

24
Exercise
Problem Effective Communication about an
Approaching Hurricane

• What are the objectives?
• What are some alternatives?

25
Consequences

• Need to measure objectives
• Natural scales
• Constructed scales
• Proxy measures
• Use models, data, and information to describe
  consequences
• Most time, effort, and cost of analysis involves
  building the model and collecting the data
• Consequences are uncertain

26
Ghost Dance Fault

• What is the probability that the Ghost Dance
  fault is active?
• What is the definition of active?
• What is the Ghost Dance Fault?

27
Implications for Applications

• Unambiguously clarify terms
• Quantify uncertainties
• Quantify consequences (or fully and completely
  describe them)
• Develop clearly defined measurement scales

28
Get the Data First

• Usually a bad idea
• The objectives, alternatives, and tradeoffs
  indicate what data and information is desired
• Much of the data collected first may turn out
  to be irrelevant
• Ex Auburn Dam Reservoir Induced Seismicity

29
Tradeoffs

• Minimizing the risk is not the objective
• Construct an objective function
• There are multiple objectives and these must be
  balanced using value tradeoffs (i.e. even
  swaps)
• Risk tolerance is also relevant
• Value judgments are required to do this

30
Tradeoffs - What is More Important

• In cleaning up hazardous waste sites, rank the
  following in order of importance (1 is most
  important)

___ Economic costs of the clean-up ___ Human
illness caused by the hazard ___ Damage caused
to the natural environment (i.e. flora and
fauna)
31
Decisions Based on Values
In cleaning up hazardous waste sites, check the
alternative that you prefer ___ Alternative A
which costs 1 billion and 20 people nationwide
subsequently get very ill for 1 week
each. ___ Alternative B which costs 2 billion
and 10 people nationwide subsequently get
very ill for 1 week each.

• Most common error which objective is most
  important

32
Uncertainty

• Uncertainty is the lack of complete knowledge of
  what is or what might occur
• Probability quantifies uncertainty (coin flip)
• Because of uncertainty, you wont know what
  consequences you will get until after deciding
• What are some decisions without uncertainties?

33
Risk Profile Introduction

• What are the key uncertainties?
• What are the possible outcomes of these
  uncertainties?
• What are the probabilities of occurrence of each
  possible outcome?
• What are the consequences of each outcome?

34
A Perspective on Analysis

• People make decisions
• The model is not the real world, so answers for
  the model are not answers to the problem
• No analysis ever makes a decision
• Analysis, done well, can provide insight to make
  good (or better) decisions
• Good analysis includes sensitivity analyses
• Analysis and insights provide a basis for
  communicating

35
On Objective Analysis

• There is no objective analysis
• The foundation of any analysis is based on
  subjective judgments
• Problem
• Objectives
• Alternatives
• Data and Information Sources
• Foundation is common sense
• Analysis can be systematic, honest, justifiable,
  consistent, understandable

36
Summary Points

• The problem drives the analysis
• The objectives and alternatives define the
  analysis
• Include all that is important to the problem
• The standard for what you should do and how you
  should do it is common sense