Transcript Proposed NSF Center on Climate Decision Making Carnegie Mellon

Carnegie Mellon University
Agenda -Part 2
11:25-11:30
Overview on Part 2
11:30-11:45
Insurance managers
- Daniel Hoffmann and Granger Morgan
11:45-12:00
Questions and Discussion
12:00-12:15
Forest, fisheries and ecosystem managers
in the Pacific Northwest and Western
Canada
-Tim McDaniels
12:15-12:30
Questions and Discussion
12:30-13:00
Lunch
Proposed NSF Center on Climate Decision Making
1
Carnegie Mellon University
Part 2: Studies of Decision Making
in Four Specific Contexts
We've proposed to study four specific decision contexts.
The point of this part of the project is to examine a set of
specific settings in which climate is likely to be very important,
and the decision-making implications of our limited ability to
make predictions about future climate can be worked out in
detail.
For that reason we have been very careful to select settings in
which we think climate will play an important role:
• insurance industry (potential large liability exposures)
• the high arctic (anticipated large climate changes).
• high latitude forest and fishery resources (anticipated substantial
impacts).
• power industry (probably will take the brunt to early serious controls).
Proposed NSF Center on Climate Decision Making
2
Carnegie Mellon University
Climate is only one variable
While we have been careful to select decision contexts
in which we believe that future climate is likely to be a
major factor, it is important to remember that even in
these cases climate is not the only thing that will
change over coming decades.
Indeed, often, even in these cases, climate is likely to
be of second order importance compared with other
important variables such as new technology, changing
public policy, changing economic relationships, and
social and instructional infrastructure.
Proposed NSF Center on Climate Decision Making
3
Carnegie Mellon University
Decision case studies…(Cont.)
While climate, and possible future climate policy, can
impose stress on social, economic and ecological systems,
so too can many other factors.
Just as there are limits to our ability to make predictions
about future climate, so too there are limits both to our
ability to identify likely sources of future stress, and to make
meaningful predictions about those stresses.
However, with some effort, at least some of these can be
identified, and when possible, be generalized.
Proposed NSF Center on Climate Decision Making
4
Carnegie Mellon University
Decision case studies…(Cont.)
Typically, we will not use detailed scenarios but rather will
use simpler parametric methods.
Thus, for example, if future natural gas prices look to be
critical to a specific class of decisions, in the projects in Part
2 we will not develop long detailed stories about how those
prices might be shaped by future technology and by
developments in the US, Europe, the Middle East and the
Former Soviet Union but will simply truncate the causal
chain, posit a range of possible future oil prices, and work
from there.
Later of course, outputs from Part 3 may help us refine this
treatment.
Proposed NSF Center on Climate Decision Making
5
Carnegie Mellon University
Decision case studies…(Cont.)
Once results from the climate science elicitation studies from
Part 1 become available, we will use them, along with the
other sector-specific information we have developed, to begin
to create a set of decision support tools that are appropriate
given the limitations on the predictive information that we are
likely to be able to acquire.
As results become available from the work in Part 3 of the
Center's research, these tools will be modified to incorporate
additional information and uncertainty about climate policy
and its impacts.
Different tools will be developed in different contexts
depending upon the details of the sector and the problems
they face.
Proposed NSF Center on Climate Decision Making
6
Carnegie Mellon University
Agenda -Part 2
11:25-11:30
Overview on Part 2
11:30-11:45
Insurance managers
- Daniel Hoffmann and Granger Morgan
11:45-12:00
Questions and Discussion
12:00-12:15
Forest, fisheries and ecosystem managers
in the Pacific Northwest and Western
Canada
-Tim McDaniels
12:15-12:30
Questions and Discussion
12:30-13:00
Lunch
Proposed NSF Center on Climate Decision Making
7
Carnegie Mellon University
Climate Change and Global Warming –
Risks to the Insurance Industry
Proposed NSF Center on Climate Decision Making
8
Carnegie Mellon University
Climate Change and Global Warming –
A Fact to Reckon With
1870
Rhone Glacier, Switzerland
1999
Proposed NSF Center on Climate Decision Making
9
Carnegie Mellon University
… and the same here : Grinnell Glacier and Lake (1910 to 1977)
Pretty soon we will
have to settle for
Non-Glacier
National Park!
Proposed NSF Center on Climate Decision Making
10
Carnegie Mellon University
Overview - The Insurance Industry
• Primary Insurance Carriers
– Business Area: short- and medium-term risk
– Products: Health & Life, P&C
– Backstop provided by Reinsurance Carriers
• Reinsurance Carriers
– Business Area: Backstop for primary carriers;
Long-term, catastrophic, specialty, and high exposure risks
– Products: CAT insurance, high excess, specialty products
(BI, D&O, ART, derivatives, guaranties, bonds), asset
management
Proposed NSF Center on Climate Decision Making
11
Carnegie Mellon University
Financials of the Insurance Industry
• Revenue and Assets
– Revenues from premium and investments
– Premium Revenues: US$2.2T global
– Assets (U.S. Life Insurance companies only):
• Total: US$2.8T
• Real Property Holdings: US$60B
• Equal to ~15% of total assets and reserves of major
pension funds and retirement programs
– Return on premium: mostly negative (ratio ~1.06)
– Return on Investments/Assets = Profitability
(Source: Innovest, personal communication, 2003)
Proposed NSF Center on Climate Decision Making
12
Carnegie Mellon University
The Dawning in the Insurance Industry
“The insurance business is first in line to be affected by
climate change. It is clear that global warming could
bankrupt the industry …”
Franklin Nutter,President,
Reinsurance Association of America, 2003
Why…?
- Climate Change is a global phenomenon,
- The Insurance Industry is a global player
Proposed NSF Center on Climate Decision Making
13
Carnegie Mellon University
Awareness of the Insurance Industry
• Primary Insurance Companies
– Late in awareness
• Reason: Revenue and Assets
– Awareness increases due to exposures in liability
coverage
• Reinsurance Industry
– Highly aware overall, as risk research/quantification has
shown increased exposure
– Leaders: Munich Re and Swiss Re
– Implementation of programs to
• Limit risks
• Assess business opportunities
Proposed NSF Center on Climate Decision Making
14
Carnegie Mellon University
Source: J. Holdren, KSG, Harvard University
Proposed NSF Center on Climate Decision Making
15
Carnegie Mellon University
Potential Climate Change Cost
Proposed NSF Center on Climate Decision Making
16
Carnegie Mellon University
Threats to the Insurance Industry
1.
Liquidity Risks
2.
Investment Portfolio Risks
Proposed NSF Center on Climate Decision Making
17
Carnegie Mellon University
Exposure of the Insurance Industry
1.
Property and Casualty (P&C) Insurance
Physical Damage to Property due to increased Frequency
and/or Severity relating to:
• Flooding – as a result of increased precipitation, rise in
sea level and change of weather patterns;
• Storms – as a result of change in ocean
currents/weather patterns
• Compounding loss effect – as a result of increasing
population, infrastructure density, increase in property
value, and event characteristics
Proposed NSF Center on Climate Decision Making
18
Carnegie Mellon University
Exposure of the Insurance Industry
in billion US$, 2001 prices. Source: Swiss Re sigma
60
50
40
2001 11 Sept
Upward trend expected to continue:
• higher insurance penetration
1992 Hurricane Andrew
• growing values
• value concentration in coastal areas
• changing hazard cycles and trends,
e.g. natural & man-made climate change
1999 Storms Lothar/
Martin
1994 Northridge
Earthquake
Total Estimated
Loss: US$38 to
50B, incl. third party
liability
30
20
10
0
1970
1975
1980
1985
1990
1995
Natural catastrophes
Man-made catastrophes
11 September loss
(property and business interruption)
11 September loss
(liability and life)
2000
Worldwide economic losses due to natural disasters appear to be doubling every 10 years and
next decade will reach US$150B
19
Proposed NSF
Center
onFinancial
Climate
Decision
Making
» Source
UNEP
Initiatives
Climate
Working Group Report 2002
Carnegie Mellon University
The Reinsurers Speak …
Relating to P&C exposure, Munich Re reported that natural
disasters caused US$55B in damage in 2002, primarily related
to weather-induced property damages across France, Austria,
Poland and Italy.
Percentage Distribution Worldwide
Source: http://www.munichre.com/pdf/natcat_natural_catastrophes_2002_e.pdf
Proposed NSF Center on Climate Decision Making
20
Carnegie Mellon University
Exposure of the Insurance Industry
2. Health and Life Insurance
Increased risk to human health as a result of
weather/climate patterns:
– Thermal Stress
– Natural Disasters
– Vector-borne Diseases (see next slide)
– Mortality Rates
Proposed NSF Center on Climate Decision Making
21
Carnegie Mellon University
Proposed NSF Center on Climate Decision Making
22
Carnegie Mellon University
Exposure of the Insurance Industry
3. Other Exposure
As a result of change in frequency and severity of events
due to climate change, unpredictability of loss exposure
relating to the following exposures:
• Business Interruption
• Agro/crop loss
• Existing CAT coverage
• Weather derivatives
• Project finance
• Directors & Officers (D&O)
• Errors & Omission (E&O)
• Technology relating to carbon mitigation and associated
technologies
Proposed NSF Center on Climate Decision Making
23
Carnegie Mellon University
In Summary
“It is estimated that US$2.7 trillion of the $10 trillion U.S.
economy is susceptible to weather-related loss of revenue,
meaning that an enormous number of companies have "off
balance sheet" risks related to climate”
John Dutton, Dean Emeritus
Penn State's College of Earth and Mineral Sciences
Of this amount…
the exposure for the global insurance industry is
approximately US$800 billion to US$1 trillion, a
significant Liquidity Risk!
Proposed NSF Center on Climate Decision Making
24
Carnegie Mellon University
Can there be any other potential
threats to the insurance industry…?
You betcha!
Proposed NSF Center on Climate Decision Making
25
Carnegie Mellon University
Yes, … and they might be big!
Do you recall Dean Dutton’s statement?
“It is estimated that US$2.7 trillion of the $10 trillion U.S.
economy is susceptible to weather-related loss of revenue,
meaning that an
enormous number of companies have "off balance
sheet" risks related to climate”
Proposed NSF Center on Climate Decision Making
26
Carnegie Mellon University
Threats to the Insurance Industry
1.
Liquidity Risks
2.
Investment Portfolio Risks
Proposed NSF Center on Climate Decision Making
27
Carnegie Mellon University
The Insurance Industry…
…is a major investor as well as a Third Party
Administrator.
• U.S. life insurance companies ALONE
– Have assets in excess of US$ 2.8T
– Account for 14% of the total assets/reserves of major
pension funds and retirement programs
– Have fiduciary responsibilities as a Third Party
Administrator for approx. US$300B in assets under
management
• … and because the above is for U.S. life insurance
companies only, that is only the tip of the iceberg
Proposed NSF Center on Climate Decision Making
28
Carnegie Mellon University
Exposure of the Insurance Industry
4. Value of Assets
Unless, rigorously evaluated from a prospective threat
emanating from emission of greenhouse gases (GHG)
resulting in climate change, the value of the assets could
be impaired due to:
• Direct but Hidden (Off-Balance Sheet) Carbon Liabilities affecting
the market value of the assets’ securities
– As a result of potential disclosure requirements/regulation of GHG,
the carbon exposure of GHG-emitting companies could be as high as
35%, resulting in a financial risk of up to 10% of current market value
(Source: Innovest, personal communication, 2003)
– Extra cost associated with climate change: The water industry could
face additional cost of $47B by 2050 and $1T by 2070
(Source: J.T. Houghton, Climate Change 2001, Oxford University Press)
Proposed NSF Center on Climate Decision Making
29
Carnegie Mellon University
Exposure of the Insurance Industry
4. Value of Assets – Part II
• Indirect Effects of Climate Change, affecting the market value
of the assets’ securities
– As a result of more complex climate variations and its
effects, increase of cost to doing business (COGS), resulting
in increase of Assets’ liabilities.
• Ex 1: Fishing Industry – As a result of current changes, increase
COGS to new fishing areas/potential loss of fishing at all.
• Ex 2: Agriculture/Food – As a result of temperature and
precipitation changes, increasing unpredictability of crop yield,
resulting in loss of market share
• Ex 3: Basic high energy consuming Industry (Steel, Chemicals) – As
a result of increased energy cost, emanating from the power
companies, COGS will increase
=> Increased COGS will result in Loss in Market Value of
Assets
Proposed NSF Center on Climate Decision Making
30
Carnegie Mellon University
Exposure of the Insurance Industry
5. Value of Insurance Industry’s Securities
The largest and ultimate threat to the Insurance industry and,
thus in the value of its own securities is based on a timing
issue.
• In fact the timing issue relating to climate change, if not
properly prepared for by the insurance industry at large,
may become its death knell.
• The ultimate threat is the compounding effect of a
CONVERGENCE
of liquidity (underwriting) exposure
AND
of investment portfolio exposure
Proposed NSF Center on Climate Decision Making
31
Carnegie Mellon University
Who is Aware of this?
… Insurance Analysts
As a result of larger exposures of the industry in 2001,
insurance analyst at Lehman Brothers lowered earning
estimates to account for “higher-than normal level of
catastrophes”
(FT.com, April 27, 2001)
Can you imagine the reaction when Convergence starts to
affect the Industry?
Proposed NSF Center on Climate Decision Making
32
Carnegie Mellon University
… and Increasingly the Insurance Industry
“The insurance business is first in line to
be affected by climate change. It is clear
that global warming could bankrupt the
industry …”
Franklin Nutter, President
Reinsurance Association of America
Proposed NSF Center on Climate Decision Making
33
Carnegie Mellon University
Plans for Research on Insurance
Building on the initial results of the other work of the Center we
will prepare a background paper in which we will develop a
preliminary taxonomy of the climate-related risks and
opportunities that confront the insurance industry, and suggest a
preliminary set of decision analytic and other tools that could
help the industry to better understand and think about these
issues.
In refining this paper we will be assisted by several industry
experts including Richard (Rich) Soja and Peter Thompson in
Chubb's global property underwriting department in Warren NJ,
and Mike Ewbank an energy underwriting specialist in Chubb's
Chicago IL office, Howard Kunruther at Wharton, and a number
of Daniel Hoffmann's professional contacts.
Proposed NSF Center on Climate Decision Making
34
Carnegie Mellon University
First Expert Workshop
We will then convene a small invitational workshop with
participants drawn from leading insurance, reinsurance,
capital investment and risk assessment firms.
We will use the workshop to revise and refine the taxonomy
and define an appropriate set of analytical needs. The result
will form the basis of the research agenda for years two and
three, during which, in collaboration with several experts
from the industry, we will undertake a program of systematic
analysis and tool development. With assistance from Paul
Fischbeck, this may include work that makes use of real
options.
Proposed NSF Center on Climate Decision Making
35
Carnegie Mellon University
Insurance…(Cont.)
While much of this work may involve the application of
existing analytical methods, given the high, and likely
irreducible, levels of some of the relevant uncertainties,
it seems probable that there will also be a need for the
development of new tools and methods (such as those
employed in our previous work on mixed levels of
uncertainty and bounding analysis).
However, the specifics should be driven by the needs
identified by key actors in the industry.
Proposed NSF Center on Climate Decision Making
36
Carnegie Mellon University
Examples of Possible Analysis
How adequate are current efforts to rate the climate-related
vulnerabilities for investments by major industrial sector?
What could be done to improve such measures?
How soon, and to what extent, will it be possible to know the
contribution that climate change makes to weather related
losses? (How big would they have to be to be detectable?
How does this compare with what we can hope to know?)
What are specific insurance/investment risks in the arctic
(shipping in NW passage; structures on permafrost); to NW
timber holdings; to power company's asset values?
Proposed NSF Center on Climate Decision Making
37
Carnegie Mellon University
Examples
…(Cont.)
While many in the
industry believe that
recent escalating weather
related losses are driven
by climate change, many
of the climate data don't
support this conclusion.
We need to look at:
- adequacy of risk
assessment tools;
- the way those
assessments are being
used;
- risk portfolios.
Losses from catastrophic weather events 1950-2000
Atlantic hurricane frequency 1948-2001
Source: IPCC (above); NOAA (below).
Proposed NSF Center on Climate Decision Making
38
Carnegie Mellon University
Second Expert Workshop
At the end of year three, with preliminary results in hand,
we will convene a second workshop at which we will
expose our work to critical review by experts from across
the industry, and seek advice on how it should be revised,
redirected, and extended.
We will communicate our results and seek input and
involvement from the expert and lay communities
concerned with insurance/investment matters via:
• The two workshops described.
• Professional and popular publication.
• Briefings to relevant government and private-sector
decision makers.
Proposed NSF Center on Climate Decision Making
39
Carnegie Mellon University
Agenda…(Cont.)
11:30-11:45
* * * * * Part 2 * * * * *
Insurance managers
- Daniel Hoffmann and Granger Morgan
11:45-12:00
Questions and Discussion
12:00-12:15
Forest, fisheries and ecosystem managers in
the Pacific Northwest and Western Canada
-Tim McDaniels
12:15-12:30
Questions and Discussion
12:30-13:00
Lunch
13:00-13:15
Arctic-region decision makers
- Hadi Dowlatabadi
13:15-14:00
Questions and Discussion
Proposed NSF Center on Climate Decision Making
40
Carnegie Mellon University
Agenda -Part 2
11:25-11:30
Overview on Part 2
11:30-11:45
Insurance managers
- Daniel Hoffmann and Granger Morgan
11:45-12:00
Questions and Discussion
12:00-12:15
Forest, fisheries and ecosystem managers
in the Pacific Northwest and Western
Canada
-Tim McDaniels
12:15-12:30
Questions and Discussion
12:30-13:00
Lunch
Proposed NSF Center on Climate Decision Making
41
Carnegie Mellon University
Basic Issue
How to manage ecosystem harvests (forestry, fisheries), and
the dual objective of maintaining rich ecosystems and
biodiversity, given irreducible uncertainties about climate,
and a host of other uncertainties regarding ecological
systems, resource productivity, values, markets, and many
other important influences?
Proposed NSF Center on Climate Decision Making
42
Carnegie Mellon University
Context
• Forestry: how we harvest the eco-productivity of nonagricultural land ecosystems
• Fisheries: how we harvest aquatic ecosystems
• Given the scale of harvest systems, these economic
harvest flows potentially conflict with ecosystem and
biodiversity preservation
• a constant tradeoff for managers, affected publics,
NGOs, concerns about eco-service flows
• Forest land is more privately owned in WA and Oregon,
nearly all public in BC
Proposed NSF Center on Climate Decision Making
43
Carnegie Mellon University
More context
• These harvest systems are always subject to massive
uncertainties:
–scientific, social, economic and institutional
• Management has been, in technical terms, as if
uncertainties were minimal ; linked to short term political
and economic objectives
• Growing involvement of civil society advocating stronger
preservation orientation, growing emphasis on new
institutional structures to address conflicts
Proposed NSF Center on Climate Decision Making
44
Carnegie Mellon University
Recognizing Irreducible Uncertainties
• Acknowledging IU about climate means we face
issues of uncertainty about biodiversity
preservation, and continuity of economic flows in
these systems much more directly and likely with
much greater potential loss over next century or so.
• How to design, compare, build broader technical
and societal support for management alternatives
given this context?
Proposed NSF Center on Climate Decision Making
45
Carnegie Mellon University
Feedbacks and interrelated effects
• Climate (specifically winter temperature) affects pine
beetle infestations
• Infestations kill trees over huge areas, change land
cover, create massive fire hazards, but still allow
harvest
• Will change species mix, age classes
• Accelerates and then reduces harvests
• Creates massive ecosystem change in parks
Proposed NSF Center on Climate Decision Making
46
Carnegie Mellon University
Interrelated effects (cont.)
• Glacial runoff will decline (to zero?) in coming decades
• Effects on habitat at mid-high elevations will be fast and
huge
• Reduced Sp/Su flows in major rivers
• Columbia River example: changes in storage requirements,
fish flows, flood control, fish production could all be
substantial
• These are already enormously contentious, complex
decision processes, subject to heavy constraints
Proposed NSF Center on Climate Decision Making
47
Carnegie Mellon University
Decision Analysis Challenges
• Value tradeoffs: biodiversity objectives, economic
objectives, flexibility, learning
• Creating alternatives: characterizing complexity; robust,
adaptive approaches; societal learning, across whole
domains and regions
• Decision processes that involve civil society groups,
managers and technical specialists
Proposed NSF Center on Climate Decision Making
48
Carnegie Mellon University
Our basic approach
• Model archetypes of management decisions in each
domain
– Influence diagrams, consequence tables, expert
elicitations, some DA modeling
• Define robust strategies that characterize different
fundamental approaches and illustrate consequences
• Foster a greater emphasis on learning and adaptation as a
generic response to uncertainty
• Engage managers, experts and civil society groups in
comparison, discussion of strategies
Proposed NSF Center on Climate Decision Making
49
Carnegie Mellon University
Information Needed From Part 1
• Limits on confidence regarding what will be known about
the rate of and extent of climate change over next 100
years for PNW and BC
• Implications for extreme weather events, average
temperature, rainfall
• This will be input to mental model characterizations of
experts regarding the implications of these uncertainties for
particular kinds of resource management decisions
Proposed NSF Center on Climate Decision Making
50
Carnegie Mellon University
Values as crucial input
• We will use value-structuring methods to clarify what
matters for important resource management decisions, in
the views of a range of interested parties
• We can use the values to develop new, more widely
supported strategies
• Values define information needed for assessment,
evaluation
Proposed NSF Center on Climate Decision Making
51
Carnegie Mellon University
Information needed to foster
more adaptive resource
management decisions
• For a given kind of decision (say, forestry response to pest
infestations)
• When, where, how are decisions made now?
• Time scale of information collection, feedback, revising
decisions, for key variables
• Role of analysis, discourse in current processes
• Incentives, penalties for adaptive approaches
• Institutional advantages and obstacles
• Fostering better decision processes
Proposed NSF Center on Climate Decision Making
52
Carnegie Mellon University
Adaptive Management and Decision
Processes
• We have worked on ways to improve stakeholder decision
processes for AM
– McDaniels and Gregory, Learning as an objective in
structured risk management decision processes, ES&T,
forthcoming.
• We are conducting a major project to design an AM plan
for salmon aquaculture in BC, involving all stakeholders
Proposed NSF Center on Climate Decision Making
53
Carnegie Mellon University
Layers of Decisions
• Resource management decisions are often viewed on a site
basis (for salmon fisheries, on a “opening” basis)
• The relationship between the narrowest level of decisions
and broader decisions (say, at the area or regional level) is
only starting to be explored in terms of approaches to
management and regulation (McDaniels and Dowlatabadi,
2004)
• These layers of management decisions will be an important
issue in design of adaptive strategies
Proposed NSF Center on Climate Decision Making
54
Carnegie Mellon University
Property Rights and First Nations
• In Canada, First Nations (native people) have been granted
quasi property rights to be consulted about and share in
benefits of resource harvests. In US, rights are more
limited in some contexts, greater in others
• We can directly address issues of native involvement in
fisheries decisions and IU due to our work with the
BCAFC
Proposed NSF Center on Climate Decision Making
55
Carnegie Mellon University
Involving Managers, NGOs and
Communities
• We will establish an advisory group from Washington and
BC for this specific component
• Its purposes will be:
– Advise on issues of scope and emphasis
– Help provide access to technical experts for mental
models work, understanding key decisions
– Provide advice on key tradeoffs and strategy design
from various perspectives
– Provide contacts for mechanisms to communicate our
findings to interested parties
Proposed NSF Center on Climate Decision Making
56
Carnegie Mellon University
Synthesis Workshops
• We plan to hold decision synthesis workshops in the final
years of the project involving a wide range of stakeholders
• Their purpose will be to characterize the results of our
work, in terms of its design, process, findings, and the
management strategies, potential consequences and
tradeoffs
• We will seek preferences for alternatives, feedback on the
issues and management practices involved and the chance
to communicate broadly
Proposed NSF Center on Climate Decision Making
57
Carnegie Mellon University
Agenda…(Cont.)
11:30-11:45
* * * * * Part 2 * * * * *
Insurance managers
- Daniel Hoffmann and Granger Morgan
11:45-12:00
Questions and Discussion
12:00-12:15
Forest, fisheries and ecosystem managers in
the Pacific Northwest and Western Canada
-Tim McDaniels
12:15-12:30
Questions and Discussion
12:30-13:00
Lunch
13:00-13:15
Arctic-region decision makers
- Hadi Dowlatabadi
13:15-14:00
Questions and Discussion
Proposed NSF Center on Climate Decision Making
58
Carnegie Mellon University
The Arctic Region
The challenge of reconciling
rapidly evolving environmental
and social conditions with
management paradigms that
emphasize restoration of “the
natural state”
Proposed NSF Center on Climate Decision Making
59
Carnegie Mellon University
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
Proposed NSF Center on Climate Decision Making
60
Carnegie Mellon University
Overview
• NADW is a key factor in determination of atmospheric
and oceanic fluxes and sea-ice cover in the circumpolar
region.
• These have defined:
– the ecology;
– the flow and fate of pollutants in the region, and;
– The opportunities for resource exploitation.
• The local decision-makers have two classes of irreducible
uncertainties to cope with:
– The gross uncertainties in evolution of the NADW and its
impacts on the flows that shape the arctic environment.
– The higher order uncertainties in interactions that these will
precipitated within and across social and environmental
processes.
Proposed NSF Center on Climate Decision Making
61
Carnegie Mellon University
Nunavut Government objectives
• Managing environmental conditions and biodiversity
through good science and Inuit Qaujimanituqangit [i.e.
Traditional Ecological Knowledge.]
• Building Healthy Communities.
• Ensuring the wise use of resources in a manner that will
protect and enhance the environment now and for future
generations.
• Developing and supporting sustainable economies.
– Provide the support needed for people to pursue
sustainable livelihoods both in the traditional and wage
economy.
Proposed NSF Center on Climate Decision Making
62
Carnegie Mellon University
Ocean
Currents
Based on: Macdonald, R.W. and
J.M. Bewers, 1996.
Contaminants in the arctic
marine environment: priorities
for protection. ICES J. mar. Sci.
53: 537-563.
Proposed NSF Center on Climate Decision Making
63
Carnegie Mellon University
Air Mass
Flows
Based on: mean air mass
position: Li, S.M., R.W. Talbot,
L.A. Barrie, R.C. Harriss, C.I.
Davidson and J.-L. Jaffrezo, 1993.
Seasonal and geographical
variations of methane sulphanic
acid in the Arctic troposphere.
Atmos. Environ. 27A: 3011-3024.
Proposed NSF Center on Climate Decision Making
64
Carnegie Mellon University
NOx
Emissions
Based on: Benkovitz, C.M., T.M.
Schultz, J.M. Pacyna, L.
Tarrason, J. Dignon, E.C.
Voldner, P.A. Spiro, A.L.
Jernnifer and T.E. Graedel,
1995. Gridded inventories of
anthropogenic emissions of
sulfur and nitrogen. J. geophys.
Res. 101: 29239.
Proposed NSF Center on Climate Decision Making
65
Carnegie Mellon University
POPs in the
Environment
POPs are found in all
compartments of the Arctic
environment. The figure shows
how these are partitioned in the
bio-geo-chemical system and
where bioaccumulation leads to
human health.
Proposed NSF Center on Climate Decision Making
66
Carnegie Mellon University
Ice
Cover
Comparison of the
averages of Arctic sea
ice for the month of
Sept. from 1973-1976
(left) to the averages for
the month of Sept. from
1999-2002 (right).
Quick Time™a nd a
TIFF ( Unco mpre ssed ) dec ompr esso r
ar e nee ded to see this pictur e.
Source: NASA 2003.
Proposed NSF Center on Climate Decision Making
67
Carnegie Mellon University
Oil & Gas
Source: AMAP 1998. AMAP
Assessment Report: Arctic
Pollution Issues. Arctic
Monitoring and Assessment
Programme (AMAP.
Proposed NSF Center on Climate Decision Making
68
Carnegie Mellon University
Ethnic
Profiles
Source: AMAP 1998. AMAP
Assessment Report: Arctic
Pollution Issues.
Proposed NSF Center on Climate Decision Making
69
Carnegie Mellon University
Issues
• Environmental change
– Climate change, pollution flow and fate
• Economic viability
– Biological resources, mineral resources and new employment
opportunities.
• Cultural identity
– Traditional Ecological Knowledge, population movements
• Politics
– Governance, International relations
• Health
– Traditional activity patterns and diet, desk jobs and imported
foods
• Disasters and their management
– …
Proposed NSF Center on Climate Decision Making
70
Carnegie Mellon University
A Long History of Seeking to Establish
Sustainable Communities
government
services
scientific
military
oil and gas
large mining
small mining
religion
trading post
subsistence
1800
1850
1900
Yukon & Alaska
1950
Nunavut
2000
Proposed NSF Center on Climate Decision Making
71
Carnegie Mellon University
Configuration of Communities = interaction of
(drivers, local conditions, constraints) over time
Drivers
Food
Religion
Commercial trade
Non-renewable
resource extraction
Military
Local Conditions
Constraints
Access to…
Limited resources
+
Cultural discord
+
+
+
+
Distant, unstable markets
Economic dependence
Human capacity
Opportunities for
diversification
Scientific research
Proposed NSF Center on Climate Decision Making
72
Carnegie Mellon University
Example of unknowns:
Impact of climate change on drivers & constraints
Drivers
Food
Local Conditions
Access to…
Religion
Commercial trade
Non-renewable
resource extraction
Military
Constraints
Climate
Change
Limited resources
+
Cultural discord
+
+
+
+
Distant, unstable markets
Economic dependence
Human capacity
Opportunities for
diversification
Scientific research
Proposed NSF Center on Climate Decision Making
73
Carnegie Mellon University
Proposed Research
I. Applied:
• Developing indicators of the: vitality and persistence of Arctic
communities
• Assessing the importance of natural and introduced attractors in
the long-term prosperity of communities.
• Helping local authorities design and implement adaptive
management strategies that permit more rapid learning and
response across Arctic communities. E.g.:
 Support for traditional economy
 Support for infrastructure development and wage economy
…
II. Theoretical:
• Characterizing use by dates for knowledge (modern and
traditional).
• Developing an algorithm for calculation of high-order interactions
without enumeration.
Proposed NSF Center on Climate Decision Making
74
Carnegie Mellon University
Outreach and community involvement
We will seek involvement from the expert and lay
communities concerned with arctic development via:
• Local research institutions:
 Nunavut Research Institute
 Canada Climate Impact Adaptations Research Network (C-CAIRN)
North
 Canadian Polar Commission
 DewLine to SeaLane project (MCRI proposal with Arctic Institute of
North America).
• and partner communities:
 Pangnirtung
 Cambridge Bay
 Bathurst Inlet
…
Proposed NSF Center on Climate Decision Making
75
Carnegie Mellon University
Agenda - Part 2…(Cont.)
13:00-13:15
Arctic-region decision makers
- Hadi Dowlatabadi
13:15-14:00
Questions and Discussion
14:00-14:20
Electric utility managers facing capital
investment decisions about generation
and 3P versus 4P
- Paul Fischbeck and Jay Apt
14:20-14:35
Questions and Discussion
14:35-15:00
Break/Executive Session
Proposed NSF Center on Climate Decision Making
76
Carnegie Mellon University
Agenda - Part 2…(Cont.)
13:00-13:15
Arctic-region decision makers
- Hadi Dowlatabadi
13:15-14:00
Questions and Discussion
14:00-14:20
Electric utility managers facing capital
investment decisions about generation
and 3P versus 4P
- Paul Fischbeck and Jay Apt
14:20-14:35
Questions and Discussion
14:35-15:00
Break/Executive Session
Proposed NSF Center on Climate Decision Making
77
Carnegie Mellon University
Why Electric Power?
• $250 billion annual sales
– Larger than telecom, computers, s/w, autos
– $3 trillion physical assets
– 4,700 generation units
– Over 3,000 Utilities in the US
• Enormous economic leverage
– August blackout: $6 billion
• Enormous uncertainty in billion dollar decisions from
incorporation of externalities
Proposed NSF Center on Climate Decision Making
78
Carnegie Mellon University
Power Plants
Beaver
Valley
Power Plants
.
> 3000 MW
2,000-3,000 MW
1,000-2,000 MW
450-1,000 MW
< 450 MW
Voltage
DC
765kV
Proposed NSF Center on Climate Decision Making
79
Carnegie Mellon University
Power Generation Investments
• Large capital investments (~$750 M for 1 unit)
– Once committed, often expensive to modify
– Long lifetime: 50 years +
– Very large proportion of electricity cost (~60%)
• Critical factors
– Systems of different plants
– Environmental regulations
– Power market structure
– Uncertain financing
– Technology breakthroughs
US Net Generation
15 new plants per year
Proposed NSF Center on Climate Decision Making
80
Carnegie Mellon University
Indirect Influence of Climate Change
Changes in
climate
Perception of
change
Regulations
Emissions control
& plant technology
Demand for
power
Valuing
externalities
Cost of power
Power market
structure
Capital investment
in power generation
Proposed NSF Center on Climate Decision Making
81
Carnegie Mellon University
Amplifiers of Uncertainty
• Perception of climate change
– Influenced by scientific understanding
– Complex combinations of stakeholders
– International, national, and regional pressures
– Conflicting goals
• “Precautionary behavior and future generations” or “Job
creation”
• Regulations
– Political solutions to environmental problems
– Large uncertainties in predicting future regulations
– Not necessarily stable
• Criteria for upgrading coal power plants without adding
emission controls changes with administration
Proposed NSF Center on Climate Decision Making
82
Carnegie Mellon University
emissions rate
(lb./mmBtu)
Regulatory Uncertainty:
Timeline of Power Plant Emission Regulation
State and Local
smoke control
laws, Federal
research 1970
CAAA
1977
CAAA
1990
CAAA
NOX
Acid Rain
Group 2
NSPS
OTC NOX Budget
Hg
BACT/LAER
Acid Rain
RACT
A. R. Group
1
SO2
NOX SIP CALL
BACT/LAER
NSPS
Typical
uncontrolled
emissions rates
1965
C
O
2
1970
NSPS
MACT
1975
1980
1985
1990
1995
2000
2005
2010
Proposed NSF Center on Climate Decision Making
83
Carnegie Mellon University
The Future is Not Clear
40
30
20
Baseline (2002)
Sweeney (2007)
Jeffords (2007)
Waxman (2007)
Bush (2018)
Kerry 2008?
10
0
SO2 (M tons)
NOx (M tons)
Hg (tons)
Note: EPA may have authority and intent to impose Jeffords-like limits, but
may have to use rigid command-and-control policies to do so.
Sources: EPA, White House
84
Proposed NSF Center on Climate Decision Making
Carnegie Mellon University
Traditional 3P Emissions
Control Technologies
•
•
•
SO2
– Fuel switch (away from high sulfur coal)
– Flue gas desulfurization (several different kinds)
NOX
– Combustion: LNB, OFA, Lean Burn
– Post-combustion: SNCR, SCR
Mercury (Hg)
– Fuel switching (coal) and coal cleaning limited
– Traditional technologies have uncertain effects
• FF, FGD, and SCR are effective at removing various forms of Hg
– Hg-specific control technologies
• Sorbent injection and capture.
• FGD enhancements (SCR+FGD ?)
– Final disposal of Hg-containing ash or sorbent
Carnegie Mellon’s Center for Energy and Environmental Studies (CEES) has developed
relevant performance and cost models (ICEM).
Proposed NSF Center on Climate Decision Making
85
Carnegie Mellon University
Utility Managers and CO2
• Granger Morgan asked two audiences of utility executives,
“How many do not believe the US will have significant
carbon regulation by 2020?” Only one hand went up at the
2002 EPRI workshop and four at a 250-person meeting at
Alliant Energy.
– EPRI’s Board then charged EPRI with developing a new
strategic plan related to climate change.
• The CEOs of Cinergy, Excelon, and Alliant all are on record
as believing carbon caps are inevitable. They do not know
how to approach investment with this uncertainty.
Proposed NSF Center on Climate Decision Making
86
Carnegie Mellon University
CO2 “Control” Technologies
• Conservation, efficiency and renewables
– Traditional emission controls reduce efficiency
• Low-carbon fossil fuels (e.g., natural gas)
– What price? Sufficient? Imports?
• Carbon ‘sinks’
– Sufficient? How permanent?
• Carbon capture and sequestration (CCS)
– What price? How acceptable?
• Geo-engineering
– How feasible? How acceptable?
Proposed NSF Center on Climate Decision Making
87
Carnegie Mellon University
The Big Questions About 3P/4P
•
What are effects of timing, level, and forms of regulation?
•
How to balance costs and benefits?
•
How to induce sufficient technological innovation to meet the goals at lowest
social dislocation?
•
How to keep the existing, coal-fired power plants available to produce lowcost power?
•
What sort of power plants (using what sort of fuels) to invest in for the future?
Bottomline: Uncertainty about future regulations can have
measurable costs for power generation owners/operators
Proposed NSF Center on Climate Decision Making
88
Carnegie Mellon University
Risk
• Risk is the set of triplets: R = {(si, pi, xi)}
– si What can happen?
– pi How likely is it to happen?
– xi If it does happen, what are the consequences?
• For short-term decisions, all three can be assessed with
confidence.
• For long-term decisions, confidence that the set of
scenarios is exhaustive disappears.
• But, it is still possible to make predictions about the
distribution of consequences.
– Limiting factors caused by physical/economic properties
– Long-term averaging because of mean-regressing processes
Proposed NSF Center on Climate Decision Making
89
Carnegie Mellon University
Scenarios in Power Generation Modeling
• Scenarios are the typical way that future uncertainties
are modeled out to 25 or 50 years
– A set of deterministic forecasts (5-10) is created to
span the variable space
– Solutions for each scenario are determined
– Robustness claims are inferred
• Limitations
– Are the scenarios a representative set?
– Are the scenarios equally likely?
– How much uncertainty is there within each
scenario?
– How can they be used to support decisions?
Proposed NSF Center on Climate Decision Making
90
Carnegie Mellon University
Industry Typically Makes
Strategic Decisions via Scenarios
Specify a few futures, with
deterministic values for fuel, NOx
price, interest rate, etc.
With full recognition of the
underlying uncertainties, the
decision surface can be
understood.
Proposed NSF Center on Climate Decision Making
91
Carnegie Mellon University
Two Research Questions
1.
2.
How will different climate change estimates and their
associated uncertainties influence perceptions and in
turn, regulations?
Given the uncertainty of future regulations, what is the
impact on decision making for power generation
assets?
Proposed NSF Center on Climate Decision Making
92
Carnegie Mellon University
Evolving Regulations and Standards
• Database of approximately 250 regulations over 40 years
dealing with transportation fuels
• Research conducted in the Center for the Study and
Improvement of Regulation (CSIR) by David Stikkers
• Evolution of standards over time
Within a regulation (between proposed to final)
Between regulations (challenged, revised, updated/follow-on)
Initiating/motivating events
More
Influence of stakeholders
Proposed
Impact of uncertainty
Regulation 2
• Preliminary stringency results
–
–
–
–
Reduced during regulation making
Increased between regulations
Varies with amount of uncertainty
Decreases lead to challenges
Stringency
–
–
–
–
–
Proposed
Final
Regulation 1
Final
Less
Time
Proposed NSF Center on Climate Decision Making
93
Carnegie Mellon University
Important Questions
• As climate change predictions evolve, what happens to the
ensuing regulations?
• What combination of “evidence” would lead to precautionary
regulations?
– Reject guaranteed short-term gains to prevent unlikely long-term
losses
– Related to other ongoing studies
• What type of climate change predictions would cause a
tightening/relaxation of regulations?
– Leaded gasoline (it’s worst than previous thought)
– MTBE (requirement, ultimate need for, clean-up)
• Can the same set of predictions lead to very different
regulations?
– Influence of other factors (election results)
• How is this uncertainty modeled?
• How are investment decisions made given this uncertainty
Proposed NSF Center on Climate Decision Making
94
Carnegie Mellon University
Expert Elicitation Protocol
• Experts
– Utility executives
– Regulators
• Conditional on the results from Part 1 (given a climate
change forecast), quantify the distributions of the resulting
regulations
– Timing (immediate to delayed)
– Stringency (none to high)
– Technology (performance-based to prescriptive)
• Protocol based on Morgan and Keith
– Capturing uncertainties
– Use of scenarios to expand thinking
• Relying on significant existing contacts in CEIC and
CSIR
Proposed NSF Center on Climate Decision Making
95
Carnegie Mellon University
Valuing Generation Assets
Given Uncertainty
• Different levels of analysis are possible
– Deterministic: Suppose you know everything
– Game theory: Multiple decision makers
– Monte Carlo: Adding uncertainties
– Portfolio: Assets cannot be valued in isolation
– Real options: Determining the value of creating future
decisions
• Each provides some level of insight, but without a
system-level framing that includes uncertainty, analyses
can lead to valuation errors
Proposed NSF Center on Climate Decision Making
96
Carnegie Mellon University
Deterministic Analysis:
Optimal Configuration by Scenario
• Scenarios well defined (allowance and fuel costs, caps)
• Lists of possible control configurations for each plant
• Find the configuration that works best with a scenario
Proposed NSF Center on Climate Decision Making
97
Carnegie Mellon University
Dominant Strategy Regret Table
Going with the dominant optimal configuration
for each plant will provide relatively good results
in 4 out of 7 scenarios
Proposed NSF Center on Climate Decision Making
98
Carnegie Mellon University
Appreciating the Importance of
Uncertainty through Sensitivity
One unit under one scenario
CO2 Emission Price $ / Ton
$
$
$
$
$
$
$
$
$
$
$
$
4.00
8.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
2.00
As Is
As Is
As Is
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
$
2.40
As Is
As Is
As Is
As Is
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
$
2.80
As Is
As Is
As Is
As Is
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
$
3.20
As Is
As Is
As Is
As Is
As Is
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass
Biomass Price $ / mmBTU
$ 3.60 $ 4.00 $ 4.40
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
Biomass As Is
As Is
Biomass Biomass As Is
Biomass Biomass Biomass
Biomass Biomass Biomass
Biomass Biomass Biomass
Biomass Biomass Biomass
$
4.80
As Is
As Is
As Is
As Is
As Is
As Is
As Is
Biomass
Biomass
Biomass
Biomass
$
5.20 $ 5.60 $ 6.00
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
As Is
Biomass As Is
As Is
Biomass Biomass Biomass
Biomass Biomass Biomass
Practical outreach: this convinced the utility that adding
uncertainty to scenario analysis was both do-able and important!
Proposed NSF Center on Climate Decision Making
99
Carnegie Mellon University
Adding the Uncertainty
• Missing key pieces of knowledge.
– Which pollutants will be controlled?
– When will the controls be required?
– At what level will the limits be set?
– What type of regulatory instruments will be used?
• Plant owners cannot wait for all uncertainty to be resolved
before making investment decisions.
• A strategy that is optimal under one regulatory scenario could
be very expensive under others.
– Lack of regulatory knowledge can be expensive.
– Improving knowledge about future regulations can have
economic savings for the plant operator and the industry.
• How robust are certain choices?
– May be being wrong doesn’t cost that much.
Proposed NSF Center on Climate Decision Making
100
Carnegie Mellon University
Multi-period Decision Model
• Scenarios and belief in them evolve over time.
– Even as some uncertainty is resolved, new questions
arise
• It takes time to install a new technology.
• Belief about future regulations evolves in different ways
– Informed/Uninformed/Mistaken
• Decision maker has to decide when to take action.
• Decision is dependent on:
– Belief in the likelihood of the various scenarios
– Scenario-specific values (reductions, allowance costs,
fuel prices)
– Configuration parameters (capital costs, heat rate, O&M)
• If the decision maker waits until “correct” scenario is
revealed, then benefits are delayed.
Proposed NSF Center on Climate Decision Making
101
Carnegie Mellon University
Optimal Configurations
Given a final “correct” scenario and a knowledge evolution,
what configuration will minimize expected NPV costs?
Knowledge evolution makes a difference.
With less informed forecasts of the future,
unnecessary equipment is installed and/or timing is off.
Proposed NSF Center on Climate Decision Making
102
Carnegie Mellon University
Portfolio Analysis
• The value of generation assets vary based on what’s happening
in the market
– Changes in fuel prices/demand
– Specifics of regulations
– Location on the grid
• The value of some plants would change in very similar ways
while others would not
– A sulfur regulation would negatively affect the value of a
coal plant and positively affect the value of a gas plant
• The fact that plant values are not perfectly correlated allows
investors to reduce their risk by using techniques from
economic portfolio theory
• We have developed techniques for incorporating the complexity
of the power grid and distributions of future key factors to
display power generation assets in a “risk-return” space.
Proposed NSF Center on Climate Decision Making
103
Carnegie Mellon University
DMUU and Portfolios
• By combining assets,
investments on the
“efficient frontier” can be
found
• This can be done at
generation-technology
scale or an individual
plant scale
• The value of adding an
asset varies based on
what is in the portfolio
Proposed NSF Center on Climate Decision Making
104
Carnegie Mellon University
• If, because of climate/regulatory
uncertainty, we don’t know precisely
the expected future return and standard
deviation of an asset, then it follows
that the efficient frontier must actually
be a distribution of frontiers and there
is, similarly, a distribution of optimal
portfolios
• What is the impact on portfolio value?
Return
Effects of Climate/regulatory Uncertainty
Variance
Proposed NSF Center on Climate Decision Making
105
Carnegie Mellon University
The “Environmental Frontier”
Portfolio Optimization and 3P+CO2
• Until now, the criteria have been exclusively finance-oriented
• However, portfolios can be constructed to satisfy any number of
criteria
• The Regulator’s Perspective: Emissions in the objective function
– Minimize emissions such that risk and return are within acceptable
parameters
• The Firm’s Perspective: Emissions in the constraints
– Maximize Sharpe ratio such that emissions do not exceed a certain
level
• The Consumer’s Perspective
– Minimize total expenditures such that reserve margin levels are
sufficient to prevent service interruptions
Proposed NSF Center on Climate Decision Making
106
Carnegie Mellon University
Trading Credits and Expanding the Efficient
Frontier
• In the basic model, all assets are power plants
• Suppose new assets – emissions credits – are introduced
• The introduction of these assets into the feasible set Paretoimproves market participants
• In our model, this can be seen directly by noting that the
efficient environmental frontier expands “northwest”
– More efficient combinations of assets are possible
– On a regional level, the dollar size of the Pareto gain
can be quantified
Proposed NSF Center on Climate Decision Making
107
Carnegie Mellon University
Initial Center Tasks
• Develop a protocol for assessing from experts the impact of climate
uncertainty on future regulations
– Will be directly tied to results from Part 1
– Will allow investigation of the effects of improved understanding
(reduce uncertainty) of climate uncertainty
• Develop portfolio-level models that will permit an economic analysis of
climate-induced regulatory uncertainty.
– Include probability distribution functions for generation plant
parameters, economic dispatch, and non-deterministic frontier
• With this framework
– Quantify the cost of regulatory uncertainty and regulatory
predictability
– Evaluate risk-mitigation programs
• Use feedback from real-world decision makers to develop decision
support tools.
Proposed NSF Center on Climate Decision Making
108
Carnegie Mellon University
Outreach
• We will communicate our results and seek input and
involvement from the electric power community via:
– Annual presentations to the EPRI RAC.
– Annual reports to the CEIC and CSIR advisory boards.
– Using CEIC and CSIR contacts, provide detailed briefings and
research collaborations with individual utilities.
– Periodic briefings to PUC commissioners via the EPRI Advisory
Board.
– Presentations to NGOs (e.g. CECA), political leaders, FERC.
• Underlying structure of the research is directly relevant for
other large-scale industries that would be affected by
climate-related regulations (e.g., petro-chemical and
automotive manufacturers).
– Using CSIR contacts, provide briefing to other industries
Proposed NSF Center on Climate Decision Making
109