CS-3: Catastrophe Modeling Update Managing Zonal Risk Casualty - - PowerPoint PPT Presentation

cs 3 catastrophe modeling update managing zonal risk
SMART_READER_LITE
LIVE PREVIEW

CS-3: Catastrophe Modeling Update Managing Zonal Risk Casualty - - PowerPoint PPT Presentation

CS-3: Catastrophe Modeling Update Managing Zonal Risk Casualty Actuaries in Reinsurance George Davis, FCAS, MAAA May 6-7, 2010 New York, NY


slide-1
SLIDE 1

CS-3: Catastrophe Modeling Update Managing Zonal Risk

  • Casualty Actuaries in Reinsurance

George Davis, FCAS, MAAA May 6-7, 2010 New York, NY

slide-2
SLIDE 2

Introduction

  • The highest wind speed and often the most intense

damage from hurricanes occurs at the coast

  • The footprint of storm damage often extends far inland
  • There is significant exposure and thus loss potential from

inland exposure

  • Some companies focus heavily on managing exposure in

coastal counties, not paying sufficient attention to inland risk

2

slide-3
SLIDE 3

About 17 Percent of U.S. Industry Exposure is in Atlantic and Gulf Coastal Counties

Estimated 2007 Insured Value of Coastal Properties by State ($B)

  • 3
slide-4
SLIDE 4

The Proportion of Coastal Exposure is Similar for Residential and Commercial Lines of Business

  • 4
slide-5
SLIDE 5

Georgia: Significant Inland Risk

  • 5
slide-6
SLIDE 6

2004 and 2005 Claims Data Clearly Show Inland Loss Potential

  • 6
slide-7
SLIDE 7

AIR Hurricane Model Realistically Captures Inland Residential and Commercial Losses

Residential Ground Up Loss Cost Commercial Ground Up Loss Cost

  • 7
slide-8
SLIDE 8

About 5 to 10 Percent of All Hurricanes Go Through Re-intensification

  • 8
slide-9
SLIDE 9

Lessons from Ike: Requirements for a Robust Model

Track durations extended based on HURDAT Inland states are modeled

  • Rmax evolves along

the inland track

9

slide-10
SLIDE 10

Though Exposure is High in Coastal Areas, Significant Accumulation of Loss Can Occur Inland

Ike Opal Katrina Tier A1 19.5% 76.7% 32.7% Tier B2 55.2% 2.8% 52.9%

3

Estimated Ground Up Industry Loss

  • Tier C3

12.1% 17.3% 11.6% Tier D4 13.1% 3.2% 2.7% Tier E5 n/a n/a n/a

1 Zip codes whose zip code centroids are within 1 mile of the effective coastline 2 Counties which touch the effective coastline, excluding Tier A areas 3 Coastal states, excluding Tier A and B areas 4 Inland HU states, Excluding Tier A, B and C areas (includes MO, IL and IN) 5 Remainder of country

10

slide-11
SLIDE 11

Importance of Knowing Policy Conditions – Storm Surge

  • High-Risk Zones

Storm Surge Risk using AIR US Hurricane Model

11

slide-12
SLIDE 12

Modeled Loss Estimates are Highly Sensitive to Proper Coding of Policy Conditions

300 40.00 45.00 3 3.5

  • !

"#$$% !%&'%

Unknown percentage of policies covering storm surge Data coded to indicate which policies cover storm surge High percentage of policies covering storm surge

  • Source: Hurricane Ike: Performance of the AIR Model

50 100 150 200 250

Properly Coded high Losses

Company A

0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00

Properly Coded high Losses

Company B

0.5 1 1.5 2 2.5 3

Properly Coded high Losses

Company C AIR Real-time Ike Scenario (Low Portfolio Loss) AIR Real-time Ike Scenario (High Portfolio Loss) Actual Portfolio Loss Experience 12

slide-13
SLIDE 13

Summary

  • There is significant inland hurricane risk
  • There is significant loss potential from storm surge
  • Exposure aggregation should extend beyond coastal

counties

  • Using Catastrophe models with quality data provides a
  • Using Catastrophe models with quality data provides a

more robust view of risk and is the best approach to understanding catastrophe risk

13