PRICING TORNADOES: USING CAT MODELS FOR GRANULAR RISK UNDERWRITING

Similar documents
2/27/2015. Perils. Risk. Loss PRICING TORNADOES: USING CAT MODELS FOR GRANULAR RISK UNDERWRITING WHY MODEL TORNADO RISK? FIRST, YOU HAVE TO ANSWER:

Enhanced F (EF) Scale for Tornado Damage

The AIR Severe Thunderstorm Model for the United States

ALLEY CATS: HOW EL NIÑO INFLUENCES TORNADO ALLEY AND THE THREAT OF CATASTROPHES

USING MIKE TO MODEL COASTAL CATASTROPHE RISK

THE AIR SEVERE THUNDERSTORM MODEL FOR AUSTRALIA

Regional Wind Vulnerability. Extratropical Cyclones Differ from Tropical Cyclones in Ways That Matter

The AIR Bushfire Model for Australia

Risk Assessment and Mitigation. Hurricane Checklist

CS-11 Tornado/Hail: To Model or Not To Model

A pragmatic view of rates and clustering

UNDERSTANDING RECENT TRENDS IN U.S. THUNDERSTORM LOSSES

Exposure Disaggregation: Introduction. By Alissa Le Mon

Hurricane Season 2018

The AIR Crop Hail Model for Canada

Kimberly J. Mueller Risk Management Solutions, Newark, CA. Dr. Auguste Boissonade Risk Management Solutions, Newark, CA

Enhanced Fujita Scale

CHAPTER 10. TORNADOES AND WINDSTORMS

Caribbean Tropical Cyclone Modeling

THE NEW STORM SURGE MODULE IN AIR S U.S. HURRICANE MODEL

(906) EXT JOIN US:

The Joplin Tornado: Lessons Learned from the NIST Investigation

Superstorm Sandy What Risk Managers and Underwriters Learned

Your Task: Read each slide then use the underlined red or underlined information to fill in your organizer.

3/5/2013 UNCERTAINTY IN HURRICANE FREQUENCY: HOW MODELERS APPROACH THE PROBLEM PROBLEM STATEMENT

5.4 HAZARDS PROFILES AND VULNERABILITY ASSESSMENT

Economic Losses and Extreme Tornado Events

DAMAGE SURVEY OF THE BINGER, OKLAHOMA TORNADO OF MAY 22, 1981

KCC White Paper: The 100 Year Hurricane. Could it happen this year? Are insurers prepared? KAREN CLARK & COMPANY. June 2014

The AIR Severe Thunderstorm Model for Europe

Earthquakes. & Expansive Soils

Complete Weather Intelligence for Public Safety from DTN

The Joplin Tornado: Lessons Learned from the NIST Investigation

Date: 27 April UF Research Contract Number: Project Date: Project Start: 6 November 2017 Project End: 30 June 2018

Wind Engineering Research at UIUC

National Weather Service

A bright flash that is produced due to electrical discharge and occurs during a thunderstorm.

The AIR Tropical Cyclone Model for Mexico

Steps to Reduce the Risk of Tornado Damage in Commercial Structures

AIR Tropical Cyclone Model for the Caribbean

...TORNADO CONFIRMED NEAR SANFORD...BROADWAY...HOLLY SPRINGS AND RALEIGH IN LEE AND WAKE COUNTIES COUNTY NORTH CAROLINA...

How Wind Technologies Work. WHITE PAPER 4 April 2017

Andy Bailey National Weather Service Forecast Office Pleasant Hill / Kansas City

World Meteorological Organization

The AIR Tropical Cyclone Model for India

National Weather Service

Our Challenge. It s All Wind 2/16/2017

Wind Speed Estimation Standard - International Working Group

GC Briefing. Weather Sentinel Hurricane Florence. Status at 5 PM EDT (21 UTC) Today (NHC) Discussion. September 13, 2018

[CLUB NAME] HURRICANE ACTIVATION PLAN [EXCERPT VERSION]

The AIR Tropical Cyclone Model for the Carribean

The Weather Information Value Chain

GC Briefing. Weather Sentinel Tropical Storm Michael. Status at 8 AM EDT (12 UTC) Today (NHC) Discussion. October 11, 2018

Lee County, Alabama 2015 Forecast Report Population, Housing and Commercial Demand

The NOAA/NWS Warning Decision Training Branch (WDTB): On-line Training Resources for Emergency Managers & Intro to Dual-Polarization Radar

AN ENGINEERING ANALYSIS OF THE GRAND ISLAND, NE TORNADOES

Air Masses, Fronts and Weather Systems

4.1 Hazard Identification: Natural Hazards

A HURRICANE IS COMING. Presented by Atiba Upchurch Broward Emergency Management Division

Weathering the Storm. Steve Drews, Impact Forecasting & Paul Eaton, Aon Benfield Analytics Analytics Insights Conference July 22-24

Real-Time Loss Estimates for Severe Thunderstorm Damage: The Event of April 27-28, 2002

ASCE/SEI Tornado Wind Speed Estimation (TWSE) Standard: Background and Introduction. Marc Levitan, NIST James LaDue, NWS

Summary of 2006 NW Pacific Typhoon Season and Verification of Authors Seasonal Forecasts

CURRENT AND FUTURE TROPICAL CYCLONE RISK IN THE SOUTH PACIFIC

Emergency Preparedness Questions

Severe Weather: Tornadoes and Hurricanes

Pacific Catastrophe Risk Assessment And Financing Initiative

THE FUTURE OF U.S. WEATHER CATASTROPHES

THE IMPACT OF WEATHER

Zoning Ordinance. Requirements Used Districts R1 R2 R2.6 R3 R4 R5 B1 B2 B3 B4 M1 M2 RH Density Low Low Med Med High High Gen Local CBD Local LT Hvy Ag

How Detailed Weather Data and Geospatial Tools Can Be Used to Increase Net Income

GeoAnalytics A key component for insurance industry disruption

TORNADO RATING IN EUROPE WITH THE EF-SCALE

IN VEHICLES: Do not try to outrun a tornado. Abandon your vehicle and hide in a nearby ditch or depression and cover your head.

THE CHANGING CLIMATE OF CATASTROPHE RISK UNDERWRITING

Impact Based Warning Tornado - Infrastructure

Insurer-Funded Non-Profit Promotes Resilient Construction Research & Communications. Tornado Proof Near Ultimate Protection Tornado Resistant

The Politics of Tornadoes The Impacts of Sudden Storm Events on People

Integrated Electricity Demand and Price Forecasting

New Directions in Catastrophe Risk Models

5.4.6 SEVERE STORM. This section provides a profile and vulnerability assessment for the severe storm hazards.

Weather Middle School Teacher Instructions and Activity

Pacific Catastrophe Risk Assessment And Financing Initiative

While all thunderstorms are dangerous, the National Weather Service (NWS) defines a severe thunderstorm as one that:

10. Severe Local Storms (Thunderstorms)

GIS Capability Maturity Assessment: How is Your Organization Doing?

16B.6 IMPLEMENTATION AND APPLICATION OF THE EF-SCALE IN CANADA. Meteorological Service of Canada, EC, Winnipeg, MB 3

NOAA Surface Weather Program

Best Practices Natural Disasters Action Plan

AN ENGINEERING ASSESSMENT OF STRUCTURAL DAMAGE IN THE ALTUS, OK TORNADO: MAY

Erin Mack Ashley, PhD, LEED AP Diana Castro, PE

Pacific Catastrophe Risk Assessment And Financing Initiative

KENTUCKY HAZARD MITIGATION PLAN RISK ASSESSMENT

Weather Alerts for Engagement. Build stronger relationships with customers through meaningful interactions

The Wind Hazard: Messaging the Wind Threat & Corresponding Potential Impacts

Miami-Dade County Overview

Tornadoes. tornado: a violently rotating column of air

Baldwin County, Alabama

Tornadoes. Be able to define what a tornado is. Be able to list several facts about tornadoes.

Hurricane Irma. Willis Re Hurricane Damage Survey Report. Executive Summary

Transcription:

PRICING TORNADOES: USING CAT MODELS FOR GRANULAR RISK UNDERWRITING Kevin Van Leer, Product Manager Americas Climate National Tornado Summit Oklahoma City, OK February 24, 2015

PURPOSE OF RMS Tool for the Insurance Industry to help with: Underwriting Portfolio Management Risk Transfer 2

WHY MODEL TORNADO RISK? 3

AAL Hail Tornado Straight-Line Wind 20% OVERALL SEVERE CONVECTIVE STORM RISK Perils Tornado Hail Straight-Line Wind Lightning 20% 60% 100-Year RP Hail Tornado Straight-Line Wind 20% 35% Risk Loss High Risk to: Aggregate Covers Automobile Lines Large Single Location Risks SFD, MFD, Mobile Home, Commercial, etc. 1/3 of U.S. peril Average Annual Loss (AAL) 92% chance of a $1BN event in the U.S. every year Two $7BN events in 2011 Tornadoes drives tail return period (RP) losses; Hail drives AAL 45% 4

HOW DOES RMS MODEL TORNADO RISK? 5

FIRST, YOU HAVE TO ANSWER: WHAT IS THE PROBABILITY OF A TORNADO IMPACTING OKLAHOMA CITY, OK? 6

HISTORICAL CHALLENGES - OBSERVATIONS Cons Point estimates Human observations Few night-time observations Reporting methods differ between NWS offices Tornado Observations Pros Relatively long database Detailed location data Assimilated from multiple sources 7

HISTORICAL CHALLENGES - CLAIMS Cons Point estimates Human observations Reporting methods differ between insurers Typically only a wind claim Tornado Claims Pros Location data Damage information 8

TORNADO RISK MODELING CHALLENGES Limited Historical Record Small Footprints Challenges Few in-situ observations Intensity estimate requires damage 9

GENERATING AN EVENT CATALOGUE Numerical Modeling Detailed Claims Analysis Results Statistical Modeling of Weather Obs. WHY A HYBRID APPROACH? Fills in the gaps associated with incomplete claims and historical data records Allows RMS to model the spatial distribution of events more accurately Advanced Hybrid Methodology Identifies areas of emerging risk Implicitly captures the behavior of event clustering Events and Sub-perils 10

WHAT IS THE PROBABILITY OF AN EF-3 TORNADO IMPACTING OKLAHOMA CITY, OK? 11

WHAT IS THE PROBABILITY OF AN EF-3 TORNADO IMPACTING OKLAHOMA CITY, OK? How is this defined? 12

HOW DO WE DEFINE TORNADO HAZARD? Hurricanes vs Tornadoes NSSL Meteorologist: That is a category 5 hurricane. Designated using satellite, aircraft recon, radar, and other observations to determine peak 1-min sustained wind speed at 10 meters Wind speed maps directly to the Saffir- Simpson Hurricane Wind Scale Meteorologist: That is a tornado, but I can t say how strong it is until there is damage. Wind speed is estimated using damage from the tornado, not direct observations Damage maps to ranges of wind speed Wind speeds map to the Enhanced Fujita Scale 13

EF-SCALE Originally developed as the Fujita scale in 1971 Enhanced Fujita (EF) Scale adopted in 2007 The F and EF-scale ratings were kept nearly the same for historical record consistency 28 Damage Indicators (U.S. Version) Damage Indicators Small barns, farm outbuildings One- or two-family residences Single-wide mobile home (MHSW) Double-wide mobile home Apt, condo, townhouse (3 stories or less) Motel Masonry apt. or motel Small retail bldg. (fast food) Small professional (doctor office, branch bank) Strip mall Large shopping mall Large, isolated ("big box") retail bldg. Automobile showroom Automotive service building School - 1-story elementary (interior or exterior halls) School - jr. or sr. high school Low-rise (1-4 story) bldg. Mid-rise (5-20 story) bldg. High-rise (over 20 stories) Institutional bldg. (hospital, govt. or university) Metal building system Service station canopy Warehouse (tilt-up walls or heavy timber) Transmission line tower Free-standing tower Free standing pole (light, flag, luminary) Tree - hardwood Tree - softwood Abbreviation SBO FR12 MHSW MHDW ACT M MAM SRB SPB SM LSM LIRB ASR ASB ES JHSH LRB MRB HRB IB MBS SSC WHB TLT FST FSP TH TS 14

DAMAGE INDICATORS Contain degrees of damage (DOD) that correspond to damage descriptions and wind speeds ranges This information provides a basis for developing a vulnerability curve 15

WHAT IS THE PROBABILITY OF AN EF-3 TORNADO IMPACTING OKLAHOMA CITY, OK? This is relatively simple to answer, but provides very little value! 16

WHAT ARE THE ACTUAL APPLICATIONS OF THE EF-SCALE AND THE HISTORICAL RECORD? Risk Assessment Scientific Research Mitigation Design Media Building Codes 17

WHAT IS THE PROBABILITY OF A TORNADO CAUSING 30% DAMAGE TO A 2-STORY WOOD FRAME SINGLE FAMILY DWELLING IN OKLAHOMA CITY, OK? This is much more valuable! 18

HOW VULNERABLE IS THAT BUILDING? EF-Scale Damage Curve Generate Damage Ratios Calibration RMS Vulnerability Curve Evaluate building characteristics and values Determine % of building value damaged (i.e. roof, siding Academic Research Wind Tunnels Damage Surveys Papers Sanity Checks 19

Many factors influence vulnerability: Construction Occupancy Location Mitigation (e.g. roof anchors, foundation connections) HOW DO DIFFERENT TYPES OF BUILDINGS PERFORM? Damage Ratio Hospital (high-rise) Hospital (low-rise) Masonry Wood Mobile Home (no tie-downs) Many lines of business are at or near 100% damage at EF3 intensity F0 F1 F2 F3 F4 F5 20

RISK MANAGEMENT APPLICATIONS Underwriting Determining Risk Drivers CAT models provide the ability to differentiate between exposures Answers the questions for both tornado hazard and vulnerability Mitigation Cost-Benefit Analysis Building Code Analysis Risk Transfer Example numbers from a specific exposure geographic and attribute combination 21

HOW CAN WE IMPROVE? 22

IT STARTS WITH THE EF-SCALE! The scale is overly precise Wind speeds do not always increase with increasing DOD Overlapping DODs What about construction quality guidance? 23

WHAT ARE THE NEXT STEPS? Can we round the wind speeds? HURDAT rounds to the nearest 5 knots Combined DODs? Cumulative DODs? Construction Quality Guidance? 24

WHAT ARE THE NEXT STEPS? Can we round the wind speeds? HURDAT rounds to the nearest 5 knots Combined DODs? Cumulative DODs? Construction Quality Guidance? Mitigation Community mitigation can shrink the damage path How do we account for building code variation? 25

WHAT ARE THE NEXT STEPS? Can we round the wind speeds? HURDAT rounds to the nearest 5 knots Combined DODs? Cumulative DODs? Construction Quality Guidance? Mitigation Community mitigation can shrink the damage path How do we account for building code variation? Where do we go from here? ASCE Committee for Wind Speed Estimation in Tornadoes Work to improve the scale to better match its applications Email: kevin.vanleer@rms.com 26

ABOUT RMS RMS is the world s leading provider of products, services, and expertise for the quantification and management of catastrophe risk. More than 400 leading insurers, reinsurers, trading companies, and other financial institutions rely on RMS models to quantify, manage, and transfer risk. As an established provider of risk modeling to companies across all market segments, RMS provides solutions that can be trusted as reliable benchmarks for strategic pricing, risk management, and risk transfer decisions. 2015 Risk Management Solutions, Inc. RMS and the RMS logo are registered trademarks of Risk Management Solutions,Inc. All other trademarks are property of their respective owners.