IRDA Seminar, Hyderabad March 13,2006

1. IRDA Seminar, Hyderabad March 13,2006

2. + 4 hours 16 hours - From a Study of the Airline Industry….. Performance during wakeful period IRDA Seminar, Hyderabad March 13,2006

3. Airline Crashes IRDA Seminar, Hyderabad March 13,2006

4. Micro Sleep cannot be controlled ! Just after passing the Outer Marker I fell asleep and awoke at 300 feet above threshold (anonymous report from a British Captain) IRDA Seminar, Hyderabad March 13,2006

5. Misc Airport/ATC Human Weather Airplane Maintenance Error Rate: Human 1/1,000 Airplane 1/ 10,000,000 Source: FAA Human Performance Program What Causes Accidents ? IRDA Seminar, Hyderabad March 13,2006

6. Story No. 2….. Quote from an Executive, extolling the benefits of cutting quality assurance and inspection costs: ”The Company has established new global benchmarks for the generation of exceptional shareholder wealth through an aggressive and innovative program of cost cutting on its production facility. Conventional constraints have been successfully challenged and replaced with new paradigms appropriate to the globalised corporate market place. Through an integrated network of facilitated workshops, the project successfully rejected the established constricting and negative influences of prescriptive engineering, onerous quality requirements, and outdated concepts of inspection and client control. Elimination of these unnecessary straitjackets has empowered the project's suppliers and contractors to propose highly economical solutions, with the win-win bonus of enhanced profitability margins for themselves. The facility shows the shape of things to come in unregulated global market economy of the 21st Century." IRDA Seminar, Hyderabad March 13,2006

7. … the production platform IRDA Seminar, Hyderabad March 13,2006

8. the shape of things to come? 25th March 2001 IRDA Seminar, Hyderabad March 13,2006

9. Petrobras P-36 Petrobras: 20.03.2001P36 platform, Campos Basin • platform sank after suffering three explosions • Insured claim PD 500 mio US$ (only PD) • Operating difficulties not reported • Down-sized, Out-sourced, Cost-cutting operations • Flawed risk analysis • Flawed crew training • Flawed communications IRDA Seminar, Hyderabad March 13,2006

10. Story # 3Closer home…… • extract from an international risk surveyor’s 2003 report on an oil platform:-> A major release of gas from a riser…. could result ….an explosion-> Alternatively, riser failure …… could result in a jet flame that would cause severe damage • Recommendation….The risers .. are unfortunately on the outside of the jacket envelope ….. specifically require protection against boat impact… riser protection guards have been placed where boat operations are envisaged but it is recommended these be reviewed to determine if impact …. could be catastrophic IRDA Seminar, Hyderabad March 13,2006

11. Story # 3Aug 2005 • extract from loss adjuster’s first report …. it is reported that the vessel impacted with the platform and, in particular, the helideck of the vessel struck the jacket in the area of the risers in the south west corner…. …. loss reserved at $ 350 mn (PD) IRDA Seminar, Hyderabad March 13,2006

12. Rating of Property & Engineering Risks in a Detariffed Environment IRDA Seminar, Hyderabad March 13,2006

13. We will look at….. -> important aspects of risk management -> assessment of natural perils exposures -> rate determination, i.e. from risk premium to risk rate -> risk rate loadings -> some features of project insurance IRDA Seminar, Hyderabad March 13,2006

14. Risk Management, Risk Management and Risk Management IRDA Seminar, Hyderabad March 13,2006

15. The Dynamic of Incident Causation near miss IRDA Seminar, Hyderabad March 13,2006

16. The Dynamic of Incident Causation (2) IRDA Seminar, Hyderabad March 13,2006

17. Comprehensive Safety Management Safety culture “ ” Liveware Risks Interactive Software Hardware IRDA Seminar, Hyderabad March 13,2006

18. Summing up of information IRDA Seminar, Hyderabad March 13,2006

19. Assessment Matrix for Refineries and Petrochemical Plants IRDA Seminar, Hyderabad March 13,2006

20. Control room Gas & Firedetectionsystems Isolation valves(EBV’s), pumptypes ESD, critical interlock & alarm systems, blow down Expanding the Boxes:Example ….. Process control • control systems • building protection • alarm • number, spread • emergency block valves • pump seal types, doublemechanical seal • emergency shut-down system • fired heater controls • flare system • alarm systems hard-wired IRDA Seminar, Hyderabad March 13,2006

21. Process &equipmentintegrity Operation Riskmanagement Commitmentto safety Expanding the Boxes (2):Example ….. Management • maintenance • inspection • contractors • training • procedures • permit systems • safe work practices • HAZOP studies • self-auditing • incident analysis, reco control • emergency procedures • housekeeping, order • safety policy • safety culture IRDA Seminar, Hyderabad March 13,2006

22. Business Natural & Plant layout Process Process Loss Grades: interruption Management allied perils and design hazards control prevention exposure 0 good Between unit Process & Neighbour - Fire water spacing (unit & Capacity Control room Utilities equipment hood supply 1 average tankfarms ) integrity 2 below average Material Gas & Fire Within unit Age of the movement Earthquake detection Fire brigade Operation 3 poor layout plant (feedstock/pro systems d. Supply) Extreme Construction Isolation val - Manual fire Process flow Risk weather ( incl . fire- Complexity ves ( MOV’s ) fighting configuration Management conditions proofing) Pump type systems (bottleneck) Fire and ESD, critical Fixed fire Critical equip- Drainage/Spill Commitment Flooding explosion interlock & fighting ment , Spares control to safety hazard alarm systems systems philosophy x 2 x 4 Weight factor: Risk quality: 0 - 33 48 - 66 82 - 132 34 - 47 67 - 81 in points good average poor above below average average Knock out criteria : Hardware Software Lifeware Assessment Matrix for Refineries and Petrochemical Plants IRDA Seminar, Hyderabad March 13,2006

23. Relative Inherent Risk (RIR) vs. Protect Level (PL) A Challenge for the Detariffed market:Multilocation + Multi-occupation risks in one account All locations showed in the graph above would most likely have been described as “average” risks although their RIR vs PL relationships are quite different. The full use of the RIR vs PL landscape allows for an enhanced separation between risks showing different risk profiles and hence for a better risk selection by the UW. IRDA Seminar, Hyderabad March 13,2006

24. MPL Maximum Possible LossMPL Maximum Probabale LossMPL Maximum Potential LossPML Possible Maximum LossPML Probable Maximum LossMAS Maximum Amount SubjectMML Maximum Monetary LossNML Normal Monetary LossNML Normal Maximum LossNLE Normal Loss ExpectancyLLP Large Loss PossibilityLLP Large Loss ProbabilityELLP Expected Large Loss Poss.ELLP Est’d Large Loss ProbabilityEPML Est’d Probable Maximum Loss EPML Est’d Possible Maximum LossMFL Maximum Foreseeable LossUML Ultimate Maximum LossAML Absolute Maximum LossTPL Total Probable LossTPL Total Possible LossMLE Maximum Loss ExpectancyAS Amount SubjectPS Percent SubjectVS Value SubjectLE Loss ExpectancyLE Loss EstimationEML Estimated Maximum Loss A final word….know your abbreviation IRDA Seminar, Hyderabad March 13,2006

25. Assessment of Natural Perils Exposures IRDA Seminar, Hyderabad March 13,2006

26. Features as distinct from fire losses low frequency / high severity accumulation control necessary (per peril and exposed area) company solvency is threatened if commitments exceed ability to pay losses full coverage may not be available  high deductibles may be considered necessary  insurance coverage may be prohibitively expensive, or in soft markets, exposure may be “forgotten” and insufficient premium charged Underwriting natural catastrophe perils IRDA Seminar, Hyderabad March 13,2006

27. For example, earthquake -> 1. What exactly is covered? (shock only? fire following? BI?) 2. Is premium sufficient? 3. Are deductibles high enough to eliminate numerous (costly!) smaller losses? 4. Is there a sublimit or co-insurance for the cover? 5. Is building construction shock resistant? 6. What is the seismicity of the area? Are coastal regions exposed to “tsunami”? 7. Is accumulation strictly controlled? Underwriting natural catastrophe perils (2) IRDA Seminar, Hyderabad March 13,2006

28. Example.. Flood & risk topography IRDA Seminar, Hyderabad March 13,2006

29. Basics of Cat risk assessment…. a reinsurer’s approach • Basic approach: quantification of (1) hazard, (2) insured value distribution, (3) damageability, and (4) effects of cover conditions • India-specific approach: above four model components can be quantified for India (e.g. Indian EQ activity, Indian EQ-insured value distribution, damageability by EQ of Indian risks, EQ cover conditions applied in India) Hazard Vulnerability Value Distribution InsuranceConditions IRDA Seminar, Hyderabad March 13,2006

30. Example… Vulnerability, Storm IRDA Seminar, Hyderabad March 13,2006

31. Example… Damageability -> for differentoccupancies IRDA Seminar, Hyderabad March 13,2006

32. Cat perils: models needed • Low frequency, many years with low losses or even no losses, but one year with very big, disastrous loss • Burning costs analysis (as applied for e.g. fire peril or mortality) is not applicable • Cat models simulate the EQ or W/F loss history over a very long time period (e.g. 10’000 years). • By summing up all the losses at the end of this time period, and dividing this sum by 10’000 years, the average yearly loss for EQ or W/F can be derived. Likewise, event losses are sorted according to their size and the 500-, 100-year, etc loss levels can be defined. • annual expected loss can be drilled down to single risk level => move from a zonal (EQ) or country (cyclone) rate to location specific rates for natural perils exposures. IRDA Seminar, Hyderabad March 13,2006

33. sample IRDA Seminar, Hyderabad March 13,2006

34. How about Engineering Risks? Engg-specific features for project risks -> TSI exposed towards the end of the project …. for eg. assume an avg exposure of 50% of TSI? -> lower vulnerability as compared to property risks …. the more vulnerable fittings and installations are added towards the end of the project For Annual Engg risks…. - depends on coverage - assume lower exposure… say 80% if adequate info is available IRDA Seminar, Hyderabad March 13,2006

35. Structural changes which affect “cat” exposure growth of world population more concentration of people / values in exposed areas increasing use of exposed areas (coasts, flood plains) increasing standard of living  introduction of complex and fragile technology  changing building standards (enforcement)  broadening scope of cover and deteriorating conditions  global warming? Underwriting natural catastrophe perils – a dynamic scenario IRDA Seminar, Hyderabad March 13,2006

36. Global Warming? signs of trouble global surface temperature increased by 0.6°C during 20th century another increase of 1.4 - 5.8 °C projected for 1990 - 2100 probably without precedent for the last 10,000 years evidence that most warming during the last 50 years is attributable to human activities  man-made climate change will persist for centuries source: Summary for policy makers (www.ipcc.ch) Underwriting natural catastrophe perils – future uncertain? IRDA Seminar, Hyderabad March 13,2006

37. Global Warming? potential consequences for property (re-)insurers more extreme precipitation events (flash floods) likely more heat waves likely increase in tropical storms in some areas likely possible increase in storm severity and frequency  potential increase in loss figures also affected by demographic and socio-economic change, as well as by technological advances  in spite of the uncertainty involved, climate issues must be taken seriously! Underwriting natural catastrophe perils – future uncertain? (2) IRDA Seminar, Hyderabad March 13,2006

38. Natcat Pool ? Underwriting natural catastrophe perils – future uncertain? (3) IRDA Seminar, Hyderabad March 13,2006

39. Setting the risk rate IRDA Seminar, Hyderabad March 13,2006

40. Premium must payclaims loss adjusting costs  reserves (for losses with low frequency and the “unforeseen”) a contribution to the Insurer’s overheads and capital costs reasonable profit (hopefully) adjustments fluctuation, data quality inflation underlying changes – portfolio, law, etc… Developing appropriate rating starts with detailed loss statistics. ExpectedLoss Cost IRDA Seminar, Hyderabad March 13,2006

41. Expected loss costs:components The expected loss cost must encompass the annual average of ALL losses (frequent and infrequent) that may incur as a result of the contractual obligations. Types of losses to be considered: • Small and frequent losses: basic loss ratio • Large and infrequent individual losses: large loss loading • Large and infrequent event losses affecting many risks: Cat loading All components together form the expected loss costs IRDA Seminar, Hyderabad March 13,2006

42. Assessment methods for estimating expected loss costs Two main assessment methods: • Experience rating: calculation of expected loss based on the loss experience of the past and its projection in the future. • Exposure rating: calculation of expected loss based on risk exposure (e.g. sums insured, number of risks, risk size distribution) IRDA Seminar, Hyderabad March 13,2006

43. Portfolio analysis Portfolio / Rating Analysis (whole property portfolio, material damage/business interruption, occupancy/etc) A company’s analysis of its own figures will give the most relevant information. Some national market associations provide members with market (or tariff) info. IRDA Seminar, Hyderabad March 13,2006

44. A credibility approach to rating A credibility coefficient would account for-> premium volumes-> heteogeneity of portfolio-> fluctuation of loss ratios from year to year Example – Metals Risk Category, Fire Credibility Rates IRDA Seminar, Hyderabad March 13,2006

45. Purpose encourage prudence by the insured (in addition to the duty of due precaution) reduce insurers’ claims burden  reduce insurers’ administrative costs Discounting for deductibles may not be justified at all on a poor accountshould be based on a company’s own statistics  a company should be able to calculate (by occupancy, protection) what portion of its loss burden would be removed by which deductibles and discount its rates accordingly only makes sense if the original gross rates are adequate!! Deductibles IRDA Seminar, Hyderabad March 13,2006

46. adequate loss statistics may not be available rating/deductibles based on theoretical models  information on historical return periods may be thin  will depend on the size of the portfolio is often subject to undue market pressure, especially during a period without catastrophic losses (memory is short) Setting natural catastrophe deductibles IRDA Seminar, Hyderabad March 13,2006

47. Individual risk adjustments (+/-) for: QualityConstructionProtectionFire LoadRisk Mgt. ConditionsMPLDeductibleLoss Limit NatCatQuality,AgeConditions CoverageFLEXAPolitical risksMisc. risksEndorsementsMach. b/d BIFixed CostsNet ProfitDeductibleCBI IRDA Seminar, Hyderabad March 13,2006

48. Contingent Business Interruption A challenge to underwriting management because of • expanded BI exposure • underwriting is “once removed” from the risk • potential lack of transparency • insufficient information • possible concerns of the insured about confidentiality • another type of contingency planning analysis • potential accumulation exposures with other business • pricing difficulties IRDA Seminar, Hyderabad March 13,2006

49. SUPPLIERS CUSTOMERS COMMUNICATION SERVICE PROVIDERS INSURED UTILITIES / ENERGY SUPPLIERS 1st TIER 1st TIER Contingent Business Interruption - continued CBI is “by nature” a non-transparent product! IRDA Seminar, Hyderabad March 13,2006

50. Contingent Business Interruption - continued Interdependencies aviation industry Steel and aluminium industry Transport Foundries Chemicals Fuselage Engine Wings Gearing Coatings Electronics Assembly Components IRDA Seminar, Hyderabad March 13,2006 Engineering industry Textile industry Ceramics (breaks)