The evolution of terrorism insurance and what the future may hold June 2014 Agenda • Historical Background • Why is supported terrorism insurance required? • How do other countries deal with this? • How is terrorism different to other event perils? • Current UK terrorism landscape. Risk manager perspective? • How can Pool Re respond to future challenges? • How is terrorism market developing? • Questions Historical Perspective • Prior to 1993, terrorism as an insurance concept was virtually nonexistent. • Most property policies in OECD territories were silent on subject and not an underwritten peril. • Exceptions to this were Spain, where Consorcio existed since 1941, but covers broad range of ‘extraordinary events’ not just terrorism and S Africa, where SASRIA existed since 1979 with cover more akin to political violence. • Also terrorism covered in some state compensation mechanisms such as Israel and Northern Ireland. • Many global examples of Catastrophe pools or vehicles that ensured cover for such things as hurricanes, earthquakes and floods available. • Pool Re was first dedicated terrorism pool and UK were the first to be forced into a partnership with government. IRA Effect • IRA UK mainland bombing campaign culminating in Baltic Exchange car bomb in April 1992. £600m insured damage in today’s money. • Reinsurers withdrew cover for UK terrorism wef 31/12/92 which meant UK insurers would have been unable to include cover. • After intense lobbying by industry groups (AIRMIC,ABI, City of London etc) government agreed to enter into dialogue. • Insurance industry established Pool Re wef 1/1/93 in partnership with Government. Intended to be temporary arrangement. • In April 1993 Bishopsgate bomb exploded; £800m+. £260m paid out at time by Pool Re. • In many ways this created precedent for what was to happen around the world post 9/11. 9/11 • Attacks marked advent of new form of terrorism • Unprecedented loss dimension; not previously envisaged. • Scale of insured losses, ($31bn all classes) largest ever insurance event (at the time but since surpassed by Katrina) • Provoked complete re-assessment of terrorism loss scenario’s • This produced maximum loss scenarios that were previously unthinkable. • Reinforced notion that global capacity is large but finite. • Cat insurance requires capital to pay the potentially high claims amount an extreme event can deliver. 9/11 choked off a significant proportion of this capital flow for terrorism. • Global re-insurers reaction mirrored their approach 9 years earlier in UK. Cover excluded. Why is support needed • IRA and Al-Qaeda demonstrated the vulnerability of modern cities. • Density of buildings means insurers issue is not that of single risks but of the cumulative effect of many risks. • More recent development of cities like London have increased exposure density significantly. • Explosive devices positioned in urban areas can cause significant loss. Unlike natural catastrophes they can be positioned deliberately. • Global reinsurers are reluctant to offer significant capacity where magnitude of loss is too high and/or too uncertain. • This can be seen from the way they have capped exposure to California earthquake, Florida windstorm etc. • Billions of exposure in many cities around world and reinsurers not prepared to offer sufficient aggregated capacity to satisfy need. Your perspective • Interested as to where terrorism features on your risk assessments? • Do you consider your own vulnerability to an attack? • Would you also evaluate your vulnerability following an attack in the vicinity? • Is loss of infrastructure due to terrorism a consideration? • Where does Cyber terrorism feature in current risk assessments? 10 most costly Terrorist Attacks 1990 to 2012 Date Country Event Insured property loss ($ millions) Fatalities 11/9/2001 USA New York/Washington 23,870 2982 24/4/1993 UK Bishopsgate bomb 1,152 1 15/6/1996 UK Manchester bomb 946 0 10/4/1992 UK Baltic Exchange bomb 852 3 26/2/1993 USA WTC bomb 794 6 24/7/2001 Sri Lanka Colombo Airport 507 20 9/2/1996 UK South Quay bomb 329 2 19/4/1995 USA Oklahoma bomb 185 166 11/4/1992 UK Staples corner bomb 122 0 26/11/2008 India Source; Guy Carpenter Tensions Building, Dec 2012 Mumbai attacks/shootings 107 172 Supported Terrorism • Withdrawal of global terrorism capacity hit those territories with significant asset concentrations hardest. • This led to pools/schemes being set up in many developed countries. Most of these had some form of government involvement • Some form of government involvement required to offset the lack of available reinsurance cover and ensure insured’s can buy cover. • Nature of government involvement varies. • This was the birth of the concept of Supported terrorism • 12 years on from 9/11 these mechanisms endure though debate as to future underway in US. • Unlikely Pool Re structure would be same if it was formed today. Just look at debate around Flood Re Pool Re •Mutual Reinsurer established 1993 by UK insurers when reinsurance cover withdrawn in UK. •No capital as limited by guarantee from HMT •Owned by its members who are insurers or Lloyds syndicates • Fund of £5bn built since inception which would pay claims in first instance. Once exhausted draw on retrocession from HMT. • Scope is unlimited by exposure and wide cover ie inc CBRN. Excludes war and cyber. Prescribed R/I rates charged to members. • Relationship with Govt via formal retrocession agreement for which premium paid. • Act of Terrorism defined by 1993 Act. Government certification required, or binding tribunal. • Eligibility – commercial policyholders, property and BI only TRIA - USA • Financial backstop created post 9/11 • Includes commercial property/BI but also inc WC & Liability • No overt premium charge for backstop but drawdown would be recouped. • Support is capped • CBRN only included if underlying property includes • Act of terrorism certified by Secretary of State. No appeal process. • Expires Dec 31 2014. • Not compulsory (apart from WC), but insurers must offer and insured accept/refuse. USA - TRIA France – Gareat • Compulsory scheme aimed at risks with exposures > E20m (separate R/I arrangement available for ‘small’ risks) • Compulsory for FFSA members to provide terrorism where fire cover purchased • Cover available to full sum insured and includes CBRN. • Unlimited state guarantee (via CCR stop loss) on top of 5 layers of aggregate cover from reinsurance market totalling E1.9bn and insurer retention E400m. • Pool does not build reserve, surplus rebated. • ‘Act of Terrorism’ defined by French criminal code. • Premiums are based on % of underlying property premium. Pools- Different approaches • Some are traditional insurance mechanisms • Others are insurance but with a guarantee/backstop at the top • Some governments take premiums for the coverage whereas others only require payment upon drawdown of funds. • Some pools buy reinsurance in the market. • Market retention is usually retained within individual insurers but some schemes build up reserves which act as the retention. • Some examples of state compensation mechanism where claimants indemnified direct by government. • Some provide CBRN, many do not. Why is state support still needed? • These arrangements came about because of market failure and the unavailability of sufficient cover. • Governments stepped in to ensure assets were protected and the wheels of commerce could continue turning. • Most schemes envisage Government withdrawal if market normalised and sufficient aggregate R/I cover was available. • Just look at the current debate in USA around TRIA. • Post 9/11 wasn’t just about the magnitude of that event, but about the broader characteristics of terrorism risk it highlighted. • That suggests terrorism continues to be a very different risk to other catastrophe perils. Why is this? Why is terrorism different? Why is terrorism different? To arrive at an objectively based premium for property insurance, an underwriter must; • Understand the capital needed to underpin the business. • Decide how much re-insurance he/she must buy to soften the effect of large/frequent losses. • Understand the expected claims pattern from the business to be written. This can only be done if the probability and severity of future losses can be modelled to an acceptable tolerance. Future loss patterns are based on historical data....but what happens if historical data is limited and/or outdated? Why is terrorism different? • We have seen the similarity with natural catastrophe risks in that both are exposed to low probability high consequence events • NatCat models are mature and probabilistic loss simulations are informed by a century of weather and loss data. • Terrorism models share low probability characteristics but data is scarce and thus frequency is more chaotic. • Manifestation of weather perils are commonly understood, as indeed is earthquake. Moreover weather ‘happens’ every day. • Terrorism is dynamic and historically manifests in different ways • Concept of dynamic uncertainty (Michel-Kerjan) Why is terrorism different? Manifestation The effect of wind strength and direction on buildings can be modelled. The effect of water can be modelled as can the effect of earthquakes of varying magnitude Terrorism is more difficult as terrorists aim to maximise damage and can target weakpoints. Many variables such as type of attack, explosive, number, situation and building type. PML analysis complex WTC – cover bought for 1 tower as perception 2 could never be lost. Why is terrorism different? • Correlation – Both NatCat and terrorism share characteristic that they can suffer spatially correlated losses and also the simultaneous occurrence of many losses from single event. • Terrorist losses can also be correlated across many more classes of cover as evidenced by 9/11. eg Life, PA, WC, marine, aviation • Correlation hinders risk diversification and thus affects capital. • NatCat has always been a rating factor whereas terrorism was not so nothing to base a premium upon • Provision of CBRN (chemical, biological, radioactive, nuclear) cover. • Nature of risk changes in ways NatCat can’t. Why is terrorism different? • Human factor – terrorists can adapt their tactics to maximise damage, recognise weakpoints in targets & purposefully adapt their strategy. • Substitution – good security at 1 risk merely means terrorists choose a weaker target, they don’t give up • Symmetry of non-information (OECD). Governments have most information on the risk but cannot share much of it. • Government influence on risk via foreign policy and approach to counter-terrorism • All of this means that terrorism risk is highly uncertain and of a nature and magnitude that cannot be predicted to any acceptable tolerance. • This is why the amount of capital deployed by re-insurers and insurers for terrorism is restricted in comparison to that deployed for NatCat. Why is terrorism different? • However, NatCat does sometimes need a different approach too • Florida Hurricane Cat Fund - $17bn capacity • California Earthquake Authority - $10bn capacity • National Flood Insurance program – paid > $38bn since 1978 • CatNat (France) • Many created to ensure risks that might otherwise be uninsurable can get cover eg high risk coastal homes. • In Florida industry models informed risk awareness of coastal areas. Could not price to reflect higher expectation of losses • Government involvement in many (direct or indirect; financial or sponsorship) Current Terrorism Landscape • IRA adapted tactics many times before ceasefire. • Al Qaeda had different strategies then fragmented. • Perception might be currently that UK threat reduced. • New MI5 chief warned recently of continuing threat and concluded that the threat no worse but that it was ‘more diffuse, more complicated and more unpredictable’ • Extremist groups still aspire to attack western countries. • 330 UK terrorist convictions from 2001 to Mar 2013. • Average of 42 arrests per year between 2001 and 2013 and total of 2465 in total since 9/11 (*) • Returning fighters from Syria an issue for UK (and others) (* - Report of the independent reviewer on the operation of the Terrorism Act 2000 & part 1 of the Terrorism Act 2006) Current Terrorism Landscape • What do you think of the current terrorism threat? • How does this inform your insurance purchases.....do you review based on threat criteria or perception or affordability? • If your building is on top of a hill, why would you still buy insurance for flood and not terrorism? • Do you consider CBRN a realistic loss scenario? • Do you get enough information about terrorism risk and insurance to inform your thinking? What might the future hold? Nature of the threat • Dynamic but won’t go away • Terrorist strategy and tactics will continue to evolve • Debate around what constitutes acceptable level of surveillance. • CBRN – intent v capability • EMP • Lack of an event doesn’t equate to lack of terrorist activity • No-one anticipated nature or magnitude of 9/11 • Possible future attacks would be equally unexpected. • Also possible future attacks wont be conventional bomb/impact type. What might the future hold? • Perception of risk in absence of events. This drives demand. • Changing risk profile – does this affect relevance of scheme? • Re-emergence of market capacity. Can Pool Re facilitate this? • Should shape of scheme change? • Can the scheme be more flexible and recognise requirements of different buying groups? Eg credit for higher deductibles • Government requirements might change • Modelling techniques improving • 20 years on the risk is different and more opaque than ever. Alternative Capacity Capital/Capacity • ‘Flood of capital’ into insurance industry • Displacement of traditional capacity by Cat bonds/ ILS/ pension funds • NatCat is a natural fit for this new capacity ie obvious triggers, mature modelling. • Will this money find its way into terrorism vehicles? • Even if it doesn’t it displaces traditional capacity who then might choose to then write some terrorism • Many pools/ vehicles already buy and rates on line on downward trend. • Remains to be seen if widespread aggregate capacity would choose to leave the shelter of the pools. Reinsurers see pools as preferred vehicle. • Regulators/ rating agencies already looking at gross terrorism exposure in USA should TRIA not renew Alternative Capacity • Alternative capacity remains carefully controlled and opportunistic. • Tends to be deployed via schemes, pools or niches ie where control of aggregations can be easily exerted. • CBRN either excluded or infrequently offered with low sub-limit • That said available capacity is growing as events disappear into history. • NatCat rates on downward spiral and loss activity remains volatile. • Terrorism looks increasingly attractive provided exposure to aggregated loss can be capped and controlled. Politics • TRIA expires 31/12/14. US debate highly polarised. • New pools still starting around world. • 23 pools/arrangements according to Guy Carpenter (12/2012) • Fiscal squeeze means Governments conscious of exposing taxpayers but governments still recognise insurance mechanism occasionally requires assistance. • Given nature of terrorism and close link with Government, it would seem risk fits as one requiring assistance in one form or another. • Shape and structure of pools/schemes may change. Questions? Any Questions?