Parametric Cover for Tropical Cyclones in Texas 6 April 2009 Texas - - PowerPoint PPT Presentation

6 April 2009

Parametric Cover for Tropical Cyclones in Texas

Slide 2

Texas – Disaster Relief and Contingency Account – Why?

• Over 150 hurricanes, tropical storms, and tropical depressions

have affected Texas since 1851.

• Most significant events create over- dependency on FEMA for

immediate and continued response.

• Texas would like to augment FEMA post event funding, reducing

dependency in immediate aftermath of storm.

• Texas is seeking a concept that will enable them to pre-fund

significant emergency response costs. Governor Perry calls for a “Disaster Relief and Contingency Account”

2009 State of the State Address

Slide 3

Parametric / Modeled Loss Solution

• Product based on parameters that trigger release of funds when

thresholds are met. Texas pre-selects trigger levels to reflect greatest concerns.

• Population compounded by severity of wind are significant drivers
• f emergency response costs. Parameters therefore include

impacted population and pre-defined wind speed.

• Populations are determined by zip code or county, and each location

is designated as a “calculation unit”. Values of “calculation units” affected by wind are totaled to determine whether population trigger is met

• Triggering event releases 100% of preset funds for Texas to deploy

where and when most needed. Calculation Units

Slide 4

Swiss Re Proposal Modeled

Population “vulnerability” (determination of population affected by

wind) can be: – binary: single wind speed affecting population determines all or nothing payout; or, – variable: impacted population set to different wind speeds with variable credit given for populations with wind speeds below highest threshold

Model develops probability of loss based on historical storm

patterns, current population, and vulnerability

Population factors, wind speeds, calculation units, algorithm and

model are locked and stored (escrowed)

As an option several tranche levels at different trigger settings is possible and will reduce basis risk

Trigger – population at each affected calculation unit

Two different wind fields (speed and extent) Red dots indicate calculation units Calculation units within affected area are calculated to determine impacted

• population. Green
• nly would be

binary, green and white would indicate variable

Slide 6

Swiss Re Proposal: Determination of payout amount

• Payout is determined by model
• Post event storm details from National Hurricane Center are entered

into escrowed model which will develop wind footprint across calculation units.

• The model parameters and storm data will be used by the model to

determine population “impacted”.

• If combination of

– wind speed, and – population trigger cover, 100% of funds are released to Texas

As an option several tranche levels at different trigger settings is possible and will reduce basis risk

Slide 7

Parametric Hurricane Cover Trigger

Summary: How does it work? Calculation Steps

• 1. Calculate wind speed at each

population grid point (for e.g. county, zip, etc.) using information from the National Hurricane Center (NHC) and Swiss Re’s wind field model

• 2. Determine percentage of affected

population at each grid point using a pre-defined vulnerability function

• 3. Calculate total affected population by

summing up affected population from various population grid points

• 4. Determine total payout based on total

affected population, trigger, and payout Design

NHC Track and parameters Hurricane Ike (2008) – Wind footprint

Slide 8

Sample Scenarios based on Loss Curves (binary trigger)

A) Texas selects a Binary trigger: – Wind Speed Trigger: 100% at 96 mph – Population trigger: 400,000 – Cost: $8* MM for $100MM in limits – Payout Scenario: 96 MPH Wind occurs in an areas with a population totals of 400,000, Texas receives $100MM.

• Model allows tailored selection of wind speed and population. As

population trigger increases, price decreases. As wind speed trigger increases, price decreases. The opposite is equally true. As variable

• ptions increase, price increases.

Reference Wind speed in mph:

Cat 1: 74-95 Cat 2: 96-110 Cat 3: 111-130 Cat 4: 131- 155 Cat 5: > 156

* There will be a minimum premium based on limits

Slide 9

Sample Scenarios based on Loss Curves (variable trigger)

B) Texas selects variable trigger: – Wind speed trigger: 10% below 74 mph, 50% at 74 mph, 100% at 111 mph – Population trigger: 375,000 – Cost: $16.67MM for $100 MM in cover

• Payout scenario: The storm is very wide, with the strongest winds at

the center, but not in an area with a population that would trigger the coverage alone. Winds blow at 115 mph in an area with a population of 300,000, at 75 mph within an area of 150,000 and 60 mph at an area of 100,000.

• The total population affected is calculated as: 100% 0f 300,000,

50% of 150,000, 10% of 100,000, or 385,000. Texas receives a payout of $100MM

• Please note: in this example there are no partial payments,

therefore the total population affected, regardless of whether calculated with a variable or binary method, must exceed the affected population trigger. Reference Wind speed in mph:

Cat 1: 74-95 Cat 2: 96-110 Cat 3: 111-130 Cat 4: 131- 155 Cat 5: > 156

Slide 10

Trigger Index Calculation

Key Historical examples with current day population figures

1900 Galveston 1915 No Name 1932 No Name 1983 Alicia Pop affected: 2.5M Pop affected: 0.9M 1919 No Name 2008 Ike Pop affected: Pop affected: 2.2M Pop affected: 1.1M Pop affected: 0.65M

Previous storm tracks

• verlaid with current

populations show the present-day potential impact of similar storms

Slide 11

Next Steps

• 1. Texas provides population data.
• 2. Texas selects data resolution (county or zip), limits;,

wind speed, and chooses either binary (yes/no) or variable vulnerability Design.

• 3. Based on model provided loss frequency curve Texas

will set the parametric impacted population target.

• 4. Provide Swiss Re with prior relief costs per unit of

population will help in developing appropriate trigger.

• 5. All numbers are preliminary and subject to verification.

Trigger will be further refined using additional information (for e.g. estimates of pop. affected in past events, etc.) Basis for continued dialogue

* This represents a basic parametric structure. The previous slides are an indicative proposal of a potential solution. A meaningful dialogue will be necessary to structure a solution that addresses Texas’ unique needs.

Contact Information Americas

Public Sector Client Manager Swiss Re America Holding Corp. 101 Constitution Ave. NW Suite 700 Washington, DC 20001 Tel (202) 742-4623 Fax (202) 742-4615 E-mail Alex_Kaplan@swissre.com

Alex Kaplan