Room Document 7. Paris, June, 2011 OECD Headquarters, 2 rue André Pascal, Paris

Transcription

1 Room Document 7 High-Level Roundtable on the Financial Management of Earthquakes Paris, June, 2011 OECD Headquarters, 2 rue André Pascal, Paris CATASTROPHE RISK LIABILITIES TO THE GOVERNMENT OF CHILE Robert Muir-Wood Risk Management Solutions

2 1 June 7 th 2011 Robert Muir-Wood PhD Chief Research Officer Insurance building on what worked well after the 2010 Feb 27 th EQ Technical risk pricing market enabled with Cat models Focus on bringing ex ante solutions to the uninsured 1) Poorer homeowners 2) Low value commercial 3) Government Facilitating new strong motion index based risk transfer Strategies for reducing the costs of earthquakes in Chile Measuring national financial and social risk costs Actions focused around three risk hot spots Tsunami Legacy buildings Building Codes to reduce damage 2

3 March 3 rd Mw 8.1 San Antonio/Valparaiso 177 killed 121,000 houses damaged or destroyed 372,000 homeless 1.1m tsunami $1.2Bn total economic 7.2% of loss insured 5520 claims $85M insured loss 2010 Feb 27 th Mw 8.8 Concepcion/Maule 562 killed x3 370,000 houses damaged or destroyed x3 800,000 homeless X2.5 4m (10m max) tsunami $25Bn total economic X21 c30% of loss insured X successful claims X34 $8.5Bn insured loss X100 3 The performance of the local insurance and global reinsurance systems after the 2010 Feb 27 th earthquake proved a significant success Two key initiatives: PPP concession schemes and the mortgage market have driven the expansion of earthquake insurance in Chile now x2 that of California for homeowners However... the government subsidies delivered after the 2010 earthquake mean that the market in Chile is now on two diverging paths Ex-ante private insurance path 24 % of homeowners + high value commercial Ex-post subsidy path 60% of poorer homeowners with damaged properties as well as hospitals, schools, government buildings etc Meanwhile the uninsured majority small business owners and 16% of middle income uninsured homeowners (those without mortgages) received nothing Going forward Government and insurance sector should focus on providing ex-ante solutions for the three uninsured sectors in Chile Government property Small businesses Home-owners without mortgages 4

4 3 Single dominant Cat peril - EQ Reasonable understanding of seismotectonic controls of earthquake size and location Good documentation of past major earthquakes and their impacts Detailed insurance claims data available from 2010 EQ for calibrating vulnerabilities Has every potential to become a highly technical market for earthquake insurance Should aim to be Solvency II compliant For insurers consistency in the same models employed for setting rates and managing capital Focus on 200yr RP Full understanding and appreciation of how data and models should be employed by re/insurers 5 Generate Stoch. Events Assess Earthquake Motion Apply Exposure Calculate Damage Quantify Financial Loss Where? How big?what probability? (Updated after 2010 to reflect that subduction zone close to Concepcion is now unloaded) How strong are the ground motions at each location? Where is the exposure? What is its value? What types of buildings? How much are the buildings damaged? How is business interrupted? What is the financial cost?

5 Subduction interface events (eq 1960, 2010) Subduction intraslab events (eg 1939) Crustal earthquakes Gaps 2010 Maule Event ruptures compiled by Dr. Diana Comte, Universidad de Chile 8

6 5 USGS active fault maps for South America Chile dataset incomplete Augmented with local fault studies Includes active folds Slip rates are poorly constrained and generally modeled as very low Characteristic magnitudes based on fault length San Ramon Fault The great earthquake of 13 May 1647 was the most destructive shock to have affected Santiago in history. City of 300 houses, most of which collapsed. About 20 per cent of the population (600 people) were killed in the earthquake. Local seismic source seems probable possibly on San Ramón Fault or other unknown local crustal fault sources 10

7 6 Detailed geotechnical data Comuna level exposure and inventory data Differentiates structure response by: Ground motion characteristics (magnitude, frequency, soil effects) Structural characteristics (construction material, building height) In Chile, key concern is non-ductile reinforced concrete Stiff buildings with limited ductility Calibrated from 2010 claims data Stiff Soil Spectral Response Soft Soil Spectral Response Short Buildings Tall Buildings

8 7 $3.00 $2.50 $2.00 $1.50 $1.00 $0.50 $

9 8 At least 3 or 4 commercial Catastrophe models already exist for Chile Used by all reinsurers writing business in Chile but less so by insurers Should be used for opening up the insurance market For setting variable risk costs by location For differentiating risk costs by property type, age, number of storeys etc For setting solvency requirements for insurers Capital requirements (for 200 yr RP losses?) For pricing reinsurance treaties New risk transfer products: eg. Designing and pricing a microinsurance system for poorer home owners Creating, pricing and structuring reinsurance needs for new products aimed at providing coverage for a Government pool Designing new strong motion index triggers for securitizations 15 Schools Premiums to be funded by Government Micro- Govt. ins residual Schools pool Health Pool Premium c $50M per year 50% microinsurance Govt school pool Govt health pool Govt Premium c $60M per year Premium c $100M per year 16

10 9 Business owner deductible Self funded deductible Uninsured no subsidy Gap between Insurance & subsidies Above the subsidy Low damage no subsidy Business owner no insurance Numbers classified as in Poverty rose nationally from 16.4% to 19.4% in the months after the earthquake 17 Current Future 24% Ex-ante insurance 16% The Gap - No insurance no subsidy Ex-ante microinsurance 60% Ex-post subsidy 18

11 10 California strong motion network 900 stations California Strong Motion Instrumentation Program run by U.S. National Center for Engineering Strong Motion Data Japan strong motion network 1000 stations K-Net run by Japanese National Research Institute for Earth Science and Disaster Prevention (NIED) Science and Technology Agency 19 At time of Feb 27 th 2010 EQ there were about 30 strong motion instruments in central Chile Data from the networks is still not widely available Opportunity to develop a new network Around the principal concentrations of population Run by a reputable independent scientific organization Guarantee of network maintenance Telemetered to central locations To be used for disaster management And available for index based risk transfer 20

12 11 Chile is in the front-line of earthquake impacts worldwide Since 1972 has instituted a world-class building code with a focus on reducing earthquake fatalities Feb 27th 2010 earthquake demonstrates the success of this building code However major earthquakes continue to have a significant economic costs Expanded and improved risk transfer can only spread the costs of earthquakes and ensure ex-ante funding is in place The next wave of earthquake impact mitigation should focus on reducing the national economic impacts Chile has the opportunity to be a leader in this work Insured Property Government Individuals Insured Property Government Individuals Insured Property Government Expanded insurance and risk transfer By measuring risk and developing policies to reduce it Individuals 22

13 12 Expanded hazard models: For each stochastic subduction zone earthquake full tsunami modeling and high resolution outputs of flood depths and velocities Comprehensive national building specific exposure data Detailed high resolution (building level) national exposure data sets of values, types, storeys, ages etc Comprehensive national human population exposure data Detailed high resolution population exposure data including locations by time of day and day of the week Evacuation modeling Expanded casualty modeling Ability to model casualty impacts of building damage and collapse 23 Developed at a 1-KM URG linked to LandScan data. Comuna level exposure values distributed to URG cells based on population LOB TIV COM $ 468b IND $ 275b SFD $ 282b MFD $ 91b Total $ 1,117b 24

14 13 High resolution map of risk cost by location for both earthquake and tsunami Risk costs for different building types, ages, etc Segmentation of risk contributions by age, type, occupancy, height etc in each individual city specific cities Loss EP for the whole of Chile Casualty Exceedance probability curve for Chile Estimation of economic impacts of earthquake including interruption to ports, industrial production, oil and gas etc Map showing casualty risk cost per location to identify the most dangerous locations in Chile Arrival at alternative strategies of risk reduction and demonstration of the impact on national and local risk cost and casualty EPs. Measure benefit v costs 25 Distribution of Reinforced Concrete by building count Distribution of Unreinforced Masonry by building count 26

15 Exceedance Probability 14 Exceedance Probability M w 9.0 offshore subduction event M w 6.8 crustal thrust event San Ramon Fault 27 7% Aggregate Exceedance Probability AEP - Normalized by Nat'l GDP 7% 6% Santiago - As Is 6% Santiago - As Is 5% 4% Santiago - RM San Ramon - Mw6.8 Subduction - Mw9.0 5% 4% Santiago - RM San Ramon - Mw6.8 Subduction - Mw9.0 3% 3% 2% 2% 1% 1% 0% Loss Amount (US$ billions) 0% 0.0% 0.5% 1.0% 1.5% 2.0% Loss as Percentage of National GDP Replace all Unreinforced Masonry buildings with Reinforced Masonry 28

16 15 First action should be to prevent any new development in tsunami zones Both insofar as this will save lives and property And in particular for redevelopment in 2010 tsunami zones Tsunami hazard zoning at 500yr RP? For other regions of Chile, begin the process to have a long term strategy to move people out of tsunami zones Planning high hazard zones to become parks or other recreational spaces Will be a tough policy to implement requiring intensive local education, consultation and long term political will 29 Valuable industrial facilities are located in relatively flat areas of Quintero bay. The most important are: The Quintero Bay LNG terminal, situated 155km northwest of Santiago, required an investment of $1.2 billion and will have an overall annual supply capacity of 2.5 million tons of LNG. In Aug 2010 peak output matched the maximum amount of gas central Chile used to import from Argentina through two cross-border pipelines. Maximum capacity of 10 million cubic meters a day = 2,300 megawatts. Las Ventanas thermoelectric station, which generates around kw/hr (1996) satisfying the requirements of V Región and part of the metropolitan area of Santiago; Las Ventanas Refinery, with an annual capacity of tonnes of 99.97% pure copper and Oxiquim S.A., company which provide the Chilean and Latin American markets with chemical products and related services. 30

17 Strict zoning laws about not building in the tsunami zone have been developed in the past in Chile (as followed 1751 earthquake when the city of Concepcion was relocated inland) However rebuilding then allowed at Penco after 1842 (which may have made sense at the time as 1835 earthquake had just happened) Where facilities have to be located in the tsunami zone they should be made tsunami resistant (to new tsunami building code) 32

18 17 Currently uninsurable (as demanded by reinsurers) However could be insured if appropriately technically rated Government/People should decide - what parts of the legacy building stock should be protected in which case it should be strengthened through low-cost loans/subsidies And what buildings should be replaced with new earthquake resistant buildings 33 The city of Valparaiso sits in the front line of major subduction zone earthquakes and megatsunamis low-lying Almandral barrios built on reclaimed land Lack of investment in the past 40 years means that the building stock of the city is vulnerable to collapse and fires in earthquakes The 1985 earthquake was far smaller than the earthquakes of 1906 or 1822 which in turn were far smaller than the M EQ The city could have high casualties in a major EQ 34

19 18 3,882 deaths and 20,000 injured. There were several destructive fires in the low lying El Almendral Much of Barrio El Almendral was flattened, along with a proportion of Barrio Puerto Cerros Alegre y Concepcion, the richest, most upmarket and mainly foreign neighbourhoods were barely affected as the hills are bedrock. 35 Equivalent to the M9 earthquake tsunami in Sendai March 11 th 2011 In Valparaiso 10m tsunami flooded the whole of Almendral and modern Avenida Argentina up to the foot of Santos Ossa Waves reported as 30m at resort town of Zapallar (50 miles north) where 1600 homes are in the tsunami zone No record of what happened at Viña del Mar or in Reñaca because no-one lived there at the time 36

20 19 In 1 st edition Recommended Lateral Force Requirements and Commentary of SEAOC 1999 it is said: This code is intended so as structures achieve a multi-tiered performance capability: Resist minor earthquake shaking without damage Resist moderate earthquake shaking without structural damage but with some damage to non structural features Resist major levels of earthquake shaking with both structural and non structural danger but without endangerment of the lives of the occupants. The purpose (as stated in Section 102 of the Uniform Building Code) is "to provide minimum standards to safeguard life or limb, health, property and public welfare." This stipulation of "minimum standards" is in keeping with constitutional curbs on police powers. Building codes can find justification only in the responsibility of government to protect public health and safety and major property damage. It is important that code requirements accomplish this and nothing more. Discussions on designs in excess of the code responsibility should be left to the owner on the advice of his structural engineer. MODERN EARTHQUAKE CODES - History & Development Frank E. McClure Chile should pioneer the enhancement of building codes to focus beyond life safety at reduced damage costs (without extra building costs) Buildings with demonstrable increased resilience should merit reductions in insurance premiums Focus on hospitals and schools new build and retrofit policies designed to ensure continuing functioning and no significant damage to these facilities 38

21 20 Take policy actions that expand ex-ante risk management for key uninsured sectors (Through a PPP?) create a microinsurance system for the poor Explore creation of new insured rental sector - move beyond subsidies Explore earthquake catastrophe pool in Government sector Expand private insurance penetration Strong motion network designed to support index risk transfer Mandate the use of catastrophe models for insurance risk management Perform Catastrophe Model analysis of overall national Earthquake risk and drive a strategy of national risk reduction Enforce tsunami hazard zonation and relocate those at highest risk Legacy buildings strengthen or replace Valparaiso city needs special attention Explore development of building codes designed to reduce damage impacts 39