Risk Analysis and Risk Management in an
Uncertain World
Howard Kunreuther (kunreuther@wharton.upenn.edu)
Risk Management and Decision Processes Center
The Wharton School
University of Pennsylvania
http://opim.wharton.upenn.edu/risk
SAMSI Workshop
Duke University
Sept. 16, 2007
Radisson Hotel RTP,
Research Triangle Park, NC
Outline of Talk
Conceptual Framework
Risk Assessment
Risk Perception
Risk Management
Application to Catastrophic Risks
Natural Disasters
Role of Insurance and Mitigation Measures
2
A Conceptual Framework for
Studying Risk
Risk Assessment
& Vulnerability
Analysis
Risk Perception and
Choice
• Public Perceptions
Modeling
of Risks
Statistical
Data Building
Scenarios
• Expert/Layperson Differences
• Risk Communication
Risk Management Strategies
Public Private Partnerships
3
A Conceptual Framework for
Studying Risk
Risk Management Strategies:
Public Private Partnerships
• Information Provision
• Incentives
• Regulation
• Standards
• Compensation
• Insurance
• Liability
Evaluation of Strategies
• Impact on Society
• Impact on Interested Parties
4
Questions to Be Addressed for Undertaking Risk
Analyses of Health, Safety and Environmental Issues?
• How can one link the tools of risk assessment and our
knowledge of risk perception to develop risk management
options that are likely to be successful and implemented?
• What is the changing role of the public and private sectors in
dealing with these risks?
• How can we utilize lessons from dealing with past events in
helping to plan for the future?
5
Risk Assessment & Vulnerability
Analysis
Risk Assessment
– Encompasses estimates of both the chances of a specific set of events occurring
and their potential consequences
– Experts differ in their estimates of the risk
– Find your favorite expert to support your position
Vulnerability Analysis
– Characterize forms of physical, social, political, economic, cultural, and
psychological harms to which individuals and modern societies are susceptible
– Millions of dollars have already been spent to reduce our vulnerability
Constructing Scenarios
– What are the probabilities of specific events?
– What are the potential consequences?
6
Using Exceedance Probability (EP) Curves
to Depict Risk
7
Risk Assessment Questions to Ponder
•
What are the chances that New Orleans will have a Category 3 or
higher hurricane in the next 10 years and what will be the resulting
damage and indirect losses?
•
What is the likelihood of a severe nuclear power accident somewhere
in the United States and what would be the resulting impacts?
•
What are the chances that an airplane will crash into the Sears Tower
in the next year and how serious would the consequences be?
•
What are the chances that there will be a smallpox epidemic in the
United States in the next five years and how many people would be
affected?
8
Risk Assessment and Vulnerability Analysis
Open Issues and Questions
How accurately can experts estimate the likelihood and consequences of
disasters of hurricanes of different magnitudes and intensities?
Can one characterize the types of uncertainties that currently exist in assessing
risk, and suggest ways to improve these estimates in the future?
What are the expected costs and benefits of undertaking specific risk-reducing
measures in hurricane-prone areas, and can one rank them on the basis of
cost effectiveness?
What are the interdependencies in the system (e.g. infrastructure damage
affecting supply of electricity, water, telephone/telecommunications, and
other services to residences and businesses)?
How do these interdependencies affect the direct and indirect losses that would
result from a future natural disaster?
9
Risk Perception and Choice
Basic Concepts
Focuses on psychological and emotional factors that have been shown to have an
enormous impact on behavior.
•
Hazards where individuals have little knowledge and experience are highly dreaded
and perceived as being very risky
•
Disparity between experts and laypersons (e.g. nuclear power, storing radioactive
waste)
•
Ignoring public’s perception of risk by scientific community if it differed significantly
from their own estimates
Individuals exhibit systematic biases in processing information and making choices
• Estimating likelihood of event is influenced by salience
• “It cannot happen to me” bias before a disaster
• “It will happen to me” bias after a disaster
• Framing of information may influence choice
Individuals have difficulties learning due to biases and information processing limitations
10
Stigmatization and Social Amplification of Risk
The Alar Case
A bag of Alar
plant growth regulant
surrounded by
Red Delicious apples
11
Stigmatization and Social Amplification of Risk
The Alar Case
•
•
No scientific evidence that Alar was carcinogenic

Were based upon animal studies that were considered suspect because the
doses used had been so large as to have been acutely toxic.

No evidence from epidemiological studies showing Alar to be a human
carcinogen.
The reaction by the public to Alar illustrates social amplification of risk.
Media amplified the risk and effectively stigmatized the product.
Millions of consumers stopped buying apples and apple products after CBS
news story on “60 Minutes” stated Alar could cause cancer.
Convincing my wife that children from birth to 3 could drink apple juice
with Alar
12
Risk Perception:
People Don’t Think Probabilistically
• “It will happen to me” or “ it will not happen to me”
• Other factors influence protective decisions---worry;
peace of mind
• Affect and emotion play a key role in decision-making
• Availability bias: likelihood of an event is estimated
by the ease with which a person can visualize it.
13
Explain this behavior:
• I bought my first set of battery cables only after my car wouldn’t start
and it had to be towed. The towing charge was twice as much as the
cost of the battery cables.
• Most homeowners in California purchase earthquake insurance
only after they experienced a quake. When asked whether the
probability of a future event was more likely than before the disaster
most people responded, “Less likely”.
• Until seat belt laws were instituted in the United States, most drivers
refused to wear them. When asked why they did not, a typical
response was, “I won’t have an accident.” This response is
consistent with the well-document finding that 90% percent of all
drivers feel they are better than the average driver.
14
Risk Perception and Choice
Open Issues and Questions
•
What role do perceived likelihoods and resulting consequences play in how
people view a particular risk that they may face, such as a severe
hurricane?
•
How important are emotional factors such as fear, dread and anxiety in how
people perceive these risks and learn over time?
•
What role do social networks and social norms play in influencing risk
perception, choice and learning with respect to low probability events?
•
How can one best communicate information to those at risk from natural
disasters so they are aware of what actions they can take prior to and after
a disaster?
•
What is the role of past experience and the media in influencing risk
perception and choice?
15
Risk Management Strategies:
Basic Concepts
Policy options for reducing losses and aiding recovery
process
–
–
–
–
Economic incentives
Insurance
Well-enforced regulations and standards (e.g. building codes)
Disaster assistance
Relevant roles of public and private sectors in
implementing hazard management strategy
Criteria for evaluating alternative strategies
– Efficiency---allocating resources to maximize social welfare
– Equity---concern with fairness and distribution of resources
16
Risk Management Strategies
Using Economic Incentives
• Premiums reductions on insurance policies for those who
undertake loss reduction measures.
•
Overcoming myopia---long-term loans on property coupled with
insurance premium reductions
Risk Communication (Framing Information Makes a Difference)
• People are more willing to wear seatbelts if they are told they have
a .33 chance of an accident over a 50-year lifetime of driving rather
than a .00001 chance each trip.
•
People are more willing to buy insurance if they are told that there
is a 1 in 5 chance of a 100 year flood occurring in the next 25 years
than a 1 in 100 chance of it occurring next year.
17
Risk Management Strategies
Present probabilities using concrete comparisons:
Risk of an automobile accident
People need to see these decisions in the contexts of risks that they
understand.
Public Sector Involvement
Well-enforced standards (e.g. building codes)
Federal reinsurance protection
Disaster assistance
18
Risk Management Strategies
Open Issues and Questions
•
What type of economic incentives would encourage property owners to
mitigate the risks of a disaster and purchase insurance prior to a disaster?
•
How does the prospect of federal aid to victims of a disaster affect
protective decisions by individuals prior to a disaster?
•
What are the appropriate roles of standards and regulations in reducing
losses from large scale disasters and the impact of the event following the
disaster?
•
What types of financial backstops should be provided by the public sector at
the state and federal levels for dealing with catastrophic losses following
future disasters?
•
How can one link different policy tools such as information provision,
economic incentives, insurance, third party inspections, regulations and
standards to achieve the desired objectives of a hazard management
strategy?
19
Katrina – As of Sunday, August 28, 2005
20
Katrina’s Path
21
Nature of the Problem
A radical change in the scale and rhythm of catastrophes
Natural disasters have caused severe insured losses in
recent years
– Hurricane Katrina: $40-$55 billion
– Hurricane Andrew: $22 billion (2005 dollars)
Victims complain about receiving substantially less than
the actual costs to repair or rebuild their damaged
structures from Katrina
Federal government is committed to providing liberal
disaster assistance to aid the victims of Katrina and
rebuild the Gulf Coast
22
Who are the Interested Parties?
Insurers provide coverage against damage from disasters and can offer financial
incentives to encourage investment in mitigation measures
Reinsurers and capital markets provide coverage against some of the losses from largescale disasters
Residents and businesses located in hazard-prone areas subject to large-scale losses
from natural disasters.
Federal government (including the President and Congress) provides disaster assistance
and operates the National Flood Insurance Program
State governments (including Governors) regulate insurance rates, create State funds
and/or insurers (e.g. Citizens) are in charge of developing and enforcing building
codes and some land use policies
Local governments (county, city level), are responsible for issuing building permits and
other activities related to regional development
Mortgage lenders and related financial institutions can create financial incentives for
homeowners and businesses to invest in mitigation measures and to purchase and
23
maintain appropriate insurance coverage.
Worldwide Evolution of Catastrophe Insured Losses, 1970-2005
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
19
70
19
71
19
72
19
73
19
74
19
75
19
76
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
0
Human-caused catastrophes
Natural catastrophes
9/11/2001 attacks
(Property and business interruption (BI); in U.S.$ billon indexed to 2005)
Sources: data from Swiss Re and Insurance Information Institute
24
The 20 Most Costly Catastrophe Insurance Losses, 1970-2005
(18 of them occurred between 1990 and 2005; 10 of them occurred in the last 5 years)
Rank
U.S.$ billion
(indexed to 2005)
Event
Victims
(Dead or missing)
Year
Area of primary damage
1
45
Hurricane Katrina
1,326
2005
USA, Gulf of Mexico et al
2
35
9/11 Attacks
3,025
2001
USA
3
22.3
Hurricane Andrew
43
1992
USA, Bahamas
4
18.5
Northridge Quake
61
1994
USA
5
11.7
Hurricane Ivan
124
2004
USA, Caribbean et al
6
10.3
Hurricane Wilma
35
2005
USA, Gulf of Mexico et al
7
8.3
Hurricane Charley
24
2004
USA, Caribbean et all
8
8.1
Typhoon Mireille
51
1991
Japan
9
6.9
Winterstorm Daria
95
1990
France, UK et al
10
6.8
Winterstorm Lothar
110
1999
France, Switzerland et al
11
6.6
Hurricane Hugo
71
1989
Puerto Rico, USA et al
12
5.2
Hurricane Frances
38
2004
USA, Bahamas
13
5.2
Storms and floods
22
1987
France, UK et al
14
5.0
Hurricane Rita
34
2005
USA, Gulf of Mexico et al
15
4.8
Winterstorm Vivian
64
1990
Western/Central Europe
16
4.7
Typhoon Bart
26
1999
Japan
17
4.2
Hurricane Georges
600
1998
USA, Caribbean
18
4.1
Hurricane Jeanne
3,034
2004
USA, Caribbean et al
19
3.7
Typhon Songda
45
2004
Japan, South Korea
20
3.5
Tropical Storm Alison
41
2001
USA
Sources: Wharton Risk Center with data from Swiss Re, Insurance Information Institute and press releases
25
What’s Happening?
The Question of Attribution
Higher degree of urbanization
Huge increase in the value at risk
Population of Florida
- 2.8 million inhabitants in 1950 - 6.8 million in 1970 - 13 million in 1990
- 19.3 million population in 2010 (590% increase since 1950) (U.S. Bureau of the Census)
- A direct hit on Miami by Hurricane Andrew that would have cost $60bn in 1992 and
$120bn in damage in 2004 because the market value has doubled and is still increasing
Weather patterns
- Among the top 19 natural disasters that occurred in the past 35 years,
more than 80% were weather-related events
- Changes in climate conditions and/or return to a high hurricane cycle?
Direct consequences
- More intense weather-related events coupled with increased value at risk cost more,
much more
What will 2007 Bring?
26
Insured Coastal Exposure
as a % of Statewide Insured Exposure in 2004
da
Flori
ticut
c
e
n
Con York
New ine
Ma
setts
u
h
c
a
Mas s uisiana
Lo sey
Jer
New are
w
Dela nd
sla
I
e
d
Rho rolina
S. Ca exas
T
NH
ppi
i
s
s
i
s
Mis ma
a
Alab ia
in
Virg
NC
g ia
Geor
0%
Source: AIR Worldwide
79.3%
63.1%
60.9%
57.9%
54.2%
37.9%
33.6%
33.2%
28.0%
25.6%
25.6%
23.3%
13.5%
12.0%
11.4%
8.9%
5.9%
10%
20%
30%
40%
50%
60%
70%
80%
27
Total Value of Insured Coastal Exposure in 2004 (in $billion)
$1,937bn
$1,902bn
da
Flori
York
w
e
N
s
Texa
setts
u
h
c
a
Mas s Jersey
New ticut
ec
Conn siana
Loui ina
rol
a
C
.
S
inia
g
r
i
V
e
Main
a
rolin
a
C
h
Nort labama
A
g ia
Geor e
war
a
l
e
D
hire
s
p
m
i
Ha
New ississipp
M
land
s
I
e
Rhod
$740bn
$663bn
$506bn
$405bn
$209bn
$149bn
$130bn
$117bn
$105bn
$76bn
$73bn
$46bn
$46bn
$45bn
$44bn
0
Source: AIR Worldwide
250
500
750
1,000 1,250 1,500 1,750 2,000 2,250
28
What is at Stake and Goals of
Study
What is at stake?
– Affordability of living in risky areas
– Who ultimately bears the costs and receives the
benefits of such decisions
Develop a strategy document to help inform the
current policy debate
– Role that the private and public sectors can play in
reducing future disaster losses
– Financing the recovery process through insurance and
other means.
29
Two Principles Guiding Study
Principle 1: Risk-based Premiums
Insurance premiums should be based on risk to provide
signals to individuals as to the hazards they face and to
encourage them to engage in cost-effective mitigation
measures to reduce their vulnerability to catastrophes.
Principle 2: Dealing with Equity and Affordability Issues
Any special treatment given to lower income residents in
hazard-prone areas who cannot afford the cost of living in
those locations should come from general public funding
and not through insurance premium subsidies.
30
Phase II of Project: 2007
Examine and evaluate alternative programs for providing
protection against financial losses and reducing risks from
natural catastrophes
Program 1: Maintaining Status Quo
Program 2: Involvement of the Public Sector
Program 3: Free Market
Develop a set of case studies based on different states or
metropolitan areas/regions to examine how well alternative
insurance programs perform in encouraging mitigation and
providing funds for recovery.
Florida
New York
South Carolina
Texas
31
Summary
The Facts: Totally new era of “large-scale risks”; huge and still
growing concentration of value in high risk areas; indication
of more devastating disasters in the future
The Reality: Major issues at stake for insurance and mitigation
programs that require us to link risk assessment with risk
perception and risk management
Research and policy questions:
What are the options for risk sharing among different
interested parties following catastrophes?
What economic incentives can encourage investment in
mitigation?
32
Questions for Discussion
Should all property owners be required to have protection against
catastrophic losses?
– Mandatory insurance coverage
– Tax on all property based on actuarial risk
What role should regulation play?
– Challenges in states permitting risk-based insurance rates
– Reexamination of accounting standards for capital market instruments
What are the liability issues?
– Paying for water damage even though homeowners insurance doesn’t
cover it
– Lawsuits filed in Mississippi and Louisiana following Katrina claiming that
homeowners policies should provide protection against water damage
What type of special treatment should be given lower income families
residing in hazard-prone areas?
– Types of subsidies that should be given to this group
– Determining proportion of homes eligible for special treatment
33