Transcript Document

EXTREME EVENTS
AND
(RE)INSURANCE
Richard J. Murnane
RPI/BIOS, 16 Jjune 2011
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
(Hurricane Fabian
Researcher Colloquium on
over
Bermuda.
Sept.
5, 2003)
Extreme
Weather
Phenomena
under Climate Change
Washington Post, June 15, 2011
Overview
• Introductory comments
• (Re)insurer’s view of science
• Climate change relative to business change
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Why Am I Here?
• Senior research scientist at BIOS
and program manager for the Risk
Prediction Initiative, a sciencebusiness partnership at BIOS
• Chief scientist with OpenRisk LLC, a
catastrophe risk business platform
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
What Is Risk?
• Risk = f(climate, exposure, vulnerability)
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Top 40 For Victims (1970-2010)
Swiss Re Sigma, 1/2011
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Top 40 For Victims (1970-2010)
Swiss Re Sigma, 1/2011
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
What Is Risk?
• Risk = f(climate, exposure, vulnerability)
• Focus on risk of losing $$$
• Probability of loss:
– Annual Average Loss (AAL)
– Return period loss, e.g., 100 year event
• Variance around the probability of loss
– Correlation of occurrence and intensity
– Clustering
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
2008 Non-life Premium Volume
US and Europe premium: ~$1.4 trillion
Global ~$1.8 trillion
Swiss Re, Sigma 3/2009
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Top 40 Property Cat Losses 1970-2010
In 2010 dollars:
Total ~$350 billion
2010 losses ~$43 billion
Swiss Re Sigma, 1/2011
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
How Do (Re)Insurers:
• Assess their risk?
– Using catastrophe risk models that provide
the “technical” price
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Generic Risk Model
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Existing Risk Models
• Public models (not open source!)
– HAZUS-MH
– Florida Public Hurricane Model
• Proprietary models
– AIR, ARA, EQECAT, RMS
– “In-house” models
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
How Do (Re)Insurers:
• Assess their risk?
– Using catastrophe risk models that provide
the “technical” price
• Price their (re)insurance?
– Model results, in part, but also market
price, investment expectations, business
considerations, …
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Underwriting Vs. Investment Returns
Aggregate of US, Canada, France, Germany, UK, and Japan
Percent Change
15
10
5
0
-5
-10
-15
1998
1999
2000
2001
2002
Underwriting result
Current investment income
Capital gain/loss
Other income/charges
Swiss Re, Sigma 2/2005
2003
2004*
Operating result
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Guy Carpenter’s Global
Property Rate On Line Index
400
0
1990
Guy Carpenter, 2011
1995
2000
Year
Katrina
100
9/11
Andrew
200
2005
Great recession?
300
2010
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Overview
• Introductory comments
• (Re)insurer’s view of science
• Climate change relative to business change
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Sea Surface Temperature
Power Dissipation Index
Changes In Hurricane Power?
“… future warming may lead
to an upward trend in tropical
cyclone destructive potential,
and – taking into account an
increasing coastal population
– a substantial increase in
hurricane-related losses in the
twenty-first century.”
Year
K. Emanuel, Nature, 2005.
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Or, No Change?
“Subjective measurements and variable
procedures make existing tropical cyclone
databases insufficiently reliable to detect
trends in the frequency of extreme
cyclones.”
Landsea et al., Science, 2006.
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Future Unfavorable Conditions?
“… the increase of [vertical wind shear] has been historically associated with
diminished hurricane activity and intensity. A suite of state-of-the-art global
climate model[s] project… [s]ubstantial increases in tropical Atlantic and East
Pacific shear …”
Vecchi and Soden, GRL, 2007.
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Upward Trend In Strongest Storms?
“We find significant upward trends
for wind speed quantiles above the
70th percentile…”
Elsner et al., Nature, 2008
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
State Of Knowledge
IPCC, 2007
CCSP, 2008
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Overview
• Introductory comments
• (Re)insurer’s view of science
• Climate change relative to business change
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Potential Impacts
• All other things being equal, losses will
increase with:
–
–
–
–
–
Sea level rise
More frequent events
More intense events
Wetter events (i.e., more floods)
Etc…..
• But, to what extent, and over what time scale,
can we say with certainty that these changes
will occur?
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
“Real World” Impacts
• Regulatory and ratings agencies
• New cat models
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
AMO And Hurricane Landfalls
Goldenberg et al., 2001
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Florida Hurricane Commission On
Loss Projection Methodology
• FHCLPM created during the 1995
Legislative Session
• Models used for rate filing in the state
must be certified by FHCLPM
• To date only models based on
climatology approved
• Models not approved by commission
used for reinsurance transactions
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
New Models
• New RMS hurricane model
–
–
–
–
Updated construction and roof types
Higher inland wind speeds
Heightened building vulnerability
Increased losses due to storm surge
• Change in losses
– Increased insured loss results range from 20 to 100
percent
– Some loss estimates in Texas have doubled, losses for
Middle Atlantic states also increased significantly
– Smaller increases in Florida
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Market Response
• Ratings:
– S&P issued a negative watch on 17 cat
bonds due to the revised model
– A.M. Best expects companies to incorporate
model revisions as soon as practical.
– Those companies that have started to use the
new version apparently are quick to move as
they see little or no impact on their rating or
capital requirements
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change
Closing Comments
• (Re)insurers time horizon is very short: quarters
and years, not decades and centuries
• To my knowledge:
– Clustering not incorporated in models
– Correlation in frequency and intensity not incorporated
in models
• Under-appreciated problem is how to combine, in
an optimal manner, results from multiple model
Researcher Colloquium on
Extreme Weather Phenomena under Climate Change