lloyd’s realistic disaster scenarios 14 th october 2004
TRANSCRIPT
Lloyd’sRealistic Disaster Scenarios
14th October 2004
Background to Lloyd’s
In existence since 1688
World's leading insurance market
66 syndicates
Over 160 brokers
Capacity of approximately £15.0 billion
Profile of the Lloyd’s Market
Source: Lloyd’s
45 Managing Agents
66 Syndicates
£15.0bn Capacity
AM Best: A-S&P: A
45 Managing Agents
66 Syndicates
£15.0bn Capacity
AM Best: A-S&P: A
45 Managing Agents
66 Syndicates
£15.0bn Capacity
AM Best: A-S&P: A
CLIENTS
BUSINESS FLOW
Reinsureds
Personal
Commercial
168Lloyd’sBrokers
Service Companies
4Members’
Agents
MEMBERS
53 Corporate
Groups
2,048IndividualsMembers(Unlimited
liability)
637 Conversions
(Limited liability)
CAPITAL PROVISION
Lloyd’s Franchisor
A
Profile of Lloyd’s Capacity & Managing Agents
Capacity Movement
66
164
4568
02468
10121416
1997 1998 1999 2000 2001 2002 2003 2004
£ Bn
0
50
100
150
200
Market Capacity No. of syndicates No. of Managing Agent
Lloyd’s Realistic Disaster Scenarios(RDSs)
Lloyd’s RDSs were introduced in 1995
Requiring syndicates to carry out disaster planning
Based on a range of hypothetical scenarios
Scenarios mainly based on historical disasters
“Failure of North Sea Complex” – from Piper Alpha
“Marine Collision” – from Exxon Valdez
“Japanese Earthquake” – from 1923 Great Kanto Event
RDS Overhaul Project
Two year RDS Overhaul Project
Improve consistency of risk assessment
Raise Loss Modelling standards in the market
Collaborative Process
Market Experts Group
Lloyd’s Market Association – Marine RDS
AIR, EQECAT and RMS Florida Windstorm
California Earthquake
Overhaul to be Completed in April 2005
Will take account of feedback received in 2004
Why Overhaul the RDSs?
To provide a more reliable and accurate measurement of risk, based on a more comprehensive and consistent approach
To measure the market’s and syndicates’ exposures to a series of major catastrophes RDS assumptions and practices varied widely
Significant variations observed in “Aggregate” to “Gross Loss” relationships
Wide variations in methodologies
To provide a better understanding of Lloyd’s aggregate risk profile For catastrophes
As an input to capital setting and allocation decisions
As an input in managing Central Fund exposures
RDS Guidance and Instruction Issued in April 2004
Replaces previous “Regulatory Bulletins” with a comprehensive guide to the assessment of catastrophe exposures
Key sections include Definitions
Calculation Principles
Recommended Best Practice
“Event Subset” Submissions
Scenarios and Events
Available on Risk Management Website www.lloyds.com/riskmanagement
REALISTIC DISASTER SCENARIOS APRIL 2004
Guidance and Instructions
SINGLE RISK SCENARIOS
1 USA Windstorm
2 Marine Event
3 North Sea – Loss of Major Complex
4 Aviation Collision
5 Major Risk Loss
6 Satellite Risks
7 Liability Risks
8 Political Risks
9 Alternative RDS: A
10 Alternative RDS: B
AGGREGATING SCENARIOS
11 Second Event
12 Florida Windstorm
13 California Earthquake
14 New Madrid Earthquake
15 European Windstorm
16 Japanese Earthquake
17 Terrorism
Scenarios and Events
Marine RDS Syndicates must return two scenarios
Collision in Prince William sound
Major Cruise Vessel Incident
Increased guidance for both scenarios
Specification of losses to the International Group of P&I Associations Reinsurance Programme.
Marine collision in Prince William Sound
US$ 50m
US$ 550m
Oil Pollution
Owned Entries
US$ 1,050m
US$ 1,550m
US$ 2,050m
US$ 50m
US$ 550m
Owned Entries
US$ 1,050m
US$ 1,550m
US$ 2,050m
US$ 437.5m
US$ 187.5m
Co
ins
ura
nce
25
%
Co
ins
ura
nce
25
%
Co
ins
ura
nce
25
%
Co
ins
ura
nce
25
%
Layer 1
Layer 2 Layer 2
Layer 3
Layer 1
Layer 4
Protection and Indemnity
Layer 1
Layer 2 Layer 2
Layer 3
Layer 1
Layer 4
Protection and Indemnity
POOL RETENTION POOL RETENTION
Tanker Owner Cruise Vessel
Oil Pollution
US$ 600m
California EarthquakeEXPOSURE DISTRIBUTION EVENT FOOTPRINTS
Catastrophe Loss Modelling Software – Florida Windstorm & California Earthquake
Three agencies - AIR, EQECAT & RMS
Working with the three catastrophe modelling agencies has enabled us to:
Specify new events tightly, by industry loss level
US$70bn and $54bn respectively Based on Commercial (including Industrial) and Residential losses
Specify 2 events of similar magnitude
Better coverage of risk (e.g. LA and San Francisco)
Derive Average Damage Factors
Feasible events cannot destroy all properties
Florida Windstorm and California Earthquake
County level “Footprints” “Aggregate” to be based on county exposures in “Footprint”
County level damage factors Derived from discussions with AIR, EQECAT and RMS
Prompt for additional lines of business to include Marine, Specie/Fine Art, Personal Accident, Aviation, Liability
Exclusion of classes viewed as “impossible” to model Contingent Business Interruption Aviation Hull in Earthquake
RMS “Roll-up” exercise
Pilot study with five managing agents
Combining Event Loss Table output To validate choice of new RDS events
To assess feasibility of combining syndicate results
Analysis split by broad business categories Direct & Facultative
Treaty
Binder
RMS “Roll-up” exercise – California EQ
RMS Industry Loss
Llo
yd
's S
am
ple
Lo
ss
D&F
D&F Benchmark Events
RMS “Roll-up” exercise
RMS Industry Loss
Llo
yd
's S
am
ple
Lo
ss
TREATY
Treaty Benchmark Events
RMS “Roll-up” exercise
RMS Industry Loss
Llo
yd
's S
am
ple
Lo
ss
BINDERS
Binder Benchmark Events
Terrorism RDS
Terrorism RDS
2-tonne conventional bomb attack on Empire State Building
Damage zones and levels specified in scenario
TRIA and non-TRIA events
Specification of deaths, injuries and BI
Fire-following, Number of Events and CBRN Guidance
Focus on data resolution
Address-level best for urban accumulations
Guidance for proportions in 10001, 10016 and 10018
Terrorism RDS
Data Resolution
Source: Munich Re. Topics Geo 2003
RDS Results Playback
Visits
Each Managing Agent visited by two or more Risk Executives
General questions asked concerning: Data capture and resolution of data
Identifying relevant policies
Loss calculation and expected values
The return process & quality assurance
Modelling methods
Review of one RDS scenario from start to finish
RDS Results Playback
Packs
Every syndicate receives a pack prepared solely for their own use.
Contains syndicate specific analysis on: Aggregate, Gross, Net and Final Net losses
Losses in relation to capacity and hence Franchise Performance Guidelines
Ranking within the market
Includes analysis on modelling types at both market and syndicate level for each scenario.
Graphical representation of Losses, Average Damage Factors (ADFs) and Ranking.
RDS ResultsModelling Type Analysis
Modelling type - Market
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
TRIA Event
Non-TRIA Event
Liability
Tampa Bay
Miami
Specific J apanese Earthquake
Specific New Madrid Earthquake
San Francisco
Los Angeles
Specific European Windstorm
Alternative RDS A
North Sea Complex
Alternative RDS B
USA Windstorm
Second Event
Major Risk Loss
Sunk/Damaged Cruise Liner
Marine Collision in P rince William Sound
Percentage of Gross Loss by Modelling type
Licensed Softw are(AIR/EQECAT/RMS) Modelled Internally Maximum Line Market Share Premium Derived
Four-Stormed is Forearmed
14 October 2004
How Hurricanes are Formed Thunderstorms drift over warm ocean
waters. Warm air from the storm and ocean
surface combine and begin rising, creating low pressure.
Trade winds blowing in opposing directions cause the storm to start spinning.
The rising warm air causes pressure to decrease at higher altitudes.
Air rises faster and faster to fill this low pressure, in turn drawing more warm air off the sea and sucking cooler, drier air downwards.
As the storm moves over the ocean it picks up more warm, moist air. Wind speeds start to increase as more air is sucked into the low pressure centre.
It can take hours or several days for a depression to grow into a fully-formed hurricane.
Lloyd’s 2004 Realistic Disaster ScenariosDistribution of Florida Property Values
Hurricane Charley – The facts
Hurricane Charley made landfall at Sanibel Island, Florida, late on the afternoon of 13 August 2004.
Maximum sustained winds were near 145 mph after upgrading to a Category 4 storm, causing severe damage and flooding in coastal areas.
Hurricane Charley had been expected to make landfall near Tampa, 120 miles to the north, but shifted course at midday and brushed low-lying islands.
Charley was responsible for the death of 27 people in Florida.
RDS 2004 - Pinellas Hurricane Event Footprint
Estimation of track of Hurricane Charley
Hurricane Frances – The facts
Hurricane Frances made landfall at Sewall’s Point in Martin County, Florida, at 05:00 GMT on Sunday 5th September.
Maximum sustained winds were near 105 mph and Frances dropped more than 33cm of rain along Florida’s east-central coast.
Insured losses will be limited because much of the damage has been caused by flooding, which is covered by the National Flood Insurance Program.
Frances was responsible for the death of 23 people in Florida.
RDS 2004Miami-Dade Hurricane Event Footprint
Estimation of track of Hurricane Frances
Tracks of Hurricane Charley & Hurricane Frances
Source: Guy Carp.com.
Cat-i Reports Hurricane Frances Update- Report 10,
September 7, 2004
Hurricane Ivan – The facts
Hurricane Ivan made landfall just west of Gulf Shores in Baldwin County, Alabama at 07:00 GMT on Thursday 16th September.
Maximum sustained winds were near 130 mph when it came ashore and subsequently moved inland and weakened.
Ivan was responsible for the death of at least 50 people in the United States in addition to over 70 people in the Caribbean.
Tracks of Hurricanes Charley, Frances & Ivan
Source: Guy Carp.com.
Cat-i Reports Hurricane Ivan Update- Report 13,
September 17, 2004
Hurricane Jeanne – The facts
Hurricane Jeanne made landfall near Hutchinson Island just east of Stuart in Martin County, Florida at 04:00 GMT on Sunday 26th September.
Maximum sustained winds were near 120 mph (Cat 3) when it came ashore and subsequently moved inland and weakened.
Jeanne was responsible for the death of at least 3,000 people in Haiti mainly due to flooding and 8 people in Florida.
Track of Hurricane Jeanne
Source: Guy Carp.com.
Cat-i Reports Hurricane Jeanne Update- Report 7,
September 27, 2004
Tracks of the Four Hurricanes
Source: Guy Carp.com.
Cat-i Reports Hurricane Jeanne Update- Report 7,
September 27, 2004
Industry Loss Estimates
RMS EQECAT AIR
Charley USD 6-8bn USD 5-10bn USD 6-10bn
Frances USD 3-6bn USD 2-5bn USD 5.6-8.7bn
Ivan USD 3-6bn USD 4-10bn USD 3-6bn
Jeanne USD 4-8bn USD 6-14bn USD 5-9bn