Capital Sensitivity What matters in solvency risk capital assessments

Transcription

1 Capital Sensitivity What matters in solvency risk capital assessments Dr. Chris Ordowich Senior Consultant, LifeRisks Investigating Future Mortality: Blending Medical & Actuarial Science for Life & Longevity Risk Management RMS LifeRisks Seminar May 24, 2012

2 Solvency II and Mortality-Driven Risks Pillar I: Quantitative Requirements Life Underwriting Qx Life Catastrophe Longevity Pillar I: Quantitative Requirements Pillar II: Risk Management Practices and Supervisory Review Pillar III: Disclosure Mortality Lapse Expense Revision Life underwriting risk is the second most material module for life undertakings behind market risk. (EIOPA, 2011) 2

3 Solvency II Modules for Mortality-Driven Risks Deaths per 100, Life Catastrophe 400 Mortality 200 Longevity

4 Capital Requirements for Mortality-Driven Risks Solvency II offers two options for calculating the Solvency Capital Requirement (SCR) Standard formula specified in QI5 Technical Specifications Internal model approved by local regulators Risk Metric Life Catastrophe Longevity Mortality Standard Formula One-year shock of 1.5 per mille Permanent 20% rate decrease Permanent 15% rate increase Standard formula is a one-size-fits-all approach Internal models allow insurers to calculate capital requirements appropriate for their risk profile Solvency II specifies risk threshold of 99.5 th loss percentile 4

5 Scenario-Based Models Statistical models limited for estimating mortality-driven risks Few historical observations of significant mortality shocks Changes in drivers of mortality trends Scenario-based models supplement historical data with scientific knowledge Ability to link specific loss levels with scenarios Sensitivity analyses maintaining scenario specificity Exceedance Probability 2.00% 1.75% 1.50% 1.25% 1.00% 0.75% 0.50% 0.25% Exceedance Probability Curve Scenario A: 5-ton Bomb in City of London Scenario B: Highly Transmissible, Low-Virulence, Rapid Vaccine 0.00% ,000 1,500 2,000 2,500 3,000 Financial Loss (Millions) Scenario C: Medium Transmissibility, Highly Virulent, Slow Vaccine, Cytokine Storm 5

6 Sample Co. Hypothetical insurance company with three books of business within the UK Book of Business Lives Amount at Risk Whole Life 2,000, billion face amount Group Life 1,000, billion face amount Simple Annuity 100,000 1 billion annualized payments Representative exposure profiles smoothed for illustrative purposes Working age life policies Retired annuity policies 6

7 Life Catastrophe Risk Mortality risk that stems from extreme or irregular events While rare, have potential to cause significant numbers of fatalities Sources of life catastrophe risk modeled by RMS include: infectious disease, terrorism, natural hazards, and accidents Infectious disease the largest risk due to its large geographic footprint Terrorism can be significant risk for highly-concentrated geographic portfolios (group life) Pandemic Influenza Emerging Infectious Diseases Terrorism Natural Hazards Other Causes 7

8 RMS Infectious Disease Model H5N1 Reassortment Transmissibility: Virulence: Age Profile: Vaccine: Pharmaceuticals: Gov t Response: Ro= % Deaths per Case Cytokine Storm 6 months to production, 75% efficacy Antioboitics ineffective, treatable with Tamiflu Airport screening, school closures Infectious Disease Scenarios Thousands of scenarios capturing the full spectrum of infectious disease risk Influenza and emerging infectious diseases Virus mutations characterized by transmissibility and virulence Response: vaccines, pharmaceuticals, school closures, etc. Epidemiological Model Fatality rate calculated from Susceptible, Infected, Recovered (SIR) Models Differential impact by demographic group linked to virus characteristics Accounts for combinations of countermeasures Probabilistic Loss Result Virus mutation is a random process Antigenic shift/drift Probabilistic model linking likelihood to virus severity Calibrated to historical record 8

9 Sample Co. Life Catastrophe Results 2.00% 1.75% LifeCatastrophe Standard Formula 1.50% Exc ceedance Probability 1.25% 1.00% 0.75% 0.50% 0.25% 0.00% Whole Life Terrorism Infectious Disease Group Life Terrorism Other Sources All Risks Combined Loss Rate (per 1000) 0.5%Exceedance Probability Sample Co. Capital Requirements Standard Formula Modeled Loss 270 million 450 million 180 million Savings Infectious Disease Terrorism Other 9

10 Sample Co. Life Catastrophe Scenarios 1 Canada Square 25 Canada Square 10 Cabot Square 25 Bank Street Terrorism 5-Ton bomb at Canary Wharf Ryder rental truck 4,500 UK population deaths Sample Co. loss of 60 million (90% group life) Infectious Disease 1918 Spanish Flu recurrence Primary cause of death: secondary bacterial infections 60,000 UK population deaths Sample Co. loss of 180 million 10

11 Life Catastrophe Sensitivity at 99.5 th Percentile Loss Rate (per ) Modelled Results Can Vary Significantly Life Catastrophe Standard Formula 0.0 Infectious Disease Terrorism Other Combined 11

12 RMS Longevity Risk Model RMS models sources of mortality improvement (vitagions) Lifestyle Health environment Regenerative Medicine Retardation of Aging Medical Intervention Each vitagion characterized by: V max Mean Trend Trend Volatility Path Volatility 12

13 Sample Co. Longevity Risk Longevity Model produces distribution of change from median liability for Sample Co. s annuity book 99.5 th percentile of this distribution is a 3% increase in annualized liability Standard formula based on 20% decrease to best estimate mortality rates Results in a 5% increase in annualized liability Modeled result is 60% of standard formula Exceedance Probability 100% 80% 60% 40% 20% 0% -4% -2% 0% 2% 4% Percentage Change in Annualized Liability Risk Metric Standard Formula Modeled Capital Requirement 575 million 345 million Life Expectancy at age 65 CMI 2011 Long-Term Rate % 3.5% 13

14 Sample Co. Longevity Risk Drivers Many different longevity pathways to a 3% increase in liability for Sample Co. s annuity book Different trajectories of smoking rates, obesity, adoption of preventative healthcare, and/or medical discoveries, etc. Single-factor example: cancer treatments developed more rapidly than expected Results can vary significantly based on characteristics of annuity book Demographics, health, socioeconomic characteristics, etc. RMS Modelled 99.5 th Percentile Change in Annualized Liability as % of Standard Formula 25% 50% 75% 100% 14

15 Sample Co. Mortality Risk Mortality risk defined as situation where mortality improvement slows compared to expected Opposite side of distribution from longevity 99.5 th percentile of Whole Life liability distribution = 580 million Mortality standard formula calculation is based on best estimate projection Standard formula is 15% increase in best estimate mortality Results in mortality capital requirement of 620 million Modeled capital requirement is about 95% of standard formula Risk Metric Standard Formula Modeled Mortality Capital Requirement 620 million 580 million 15

16 Sample Co. Mortality Risk Drivers Many different mortality pathways leading to the modeled mortality capital requirement Factors include increases in smoking, obesity rates, and/or stagnation of healthcare funding Example: improvements in healthcare outcome metrics slow to below half the rate of the past 20 years Results can vary significantly based on characteristics of life books Demographics, underlying health, lapse rates, etc. RMS Modelled 99.5 th Percentile Change in Annualized Liability as % of Standard Formula 25% 50% 75% 100% 16

17 Sample Co. Mortality-Related Capital Summary Legend Standard Formula Modelled Deaths per 100, Life Catastrophe (60% of Standard Formula) 450m 270 0m 400 Mortality (95% of Standard Formula) 620m 580m 200 Longevity (60% of Standard Formula) m 345m 17

18 Integrating Solvency II Internal Models In order to use internal models for Solvency II, models must be validated, documented, and internalized Validation facilitated by ability to explore events at key points of loss distribution RMS Models supplemented by detailed methodology documentation Ability to tailor modeled risk estimates to a company s book of business make results more actionable More likely to be integrated into internal risk management processes 18

19 Scenario-Based Mortality Model Applications Pricing/ Underwriting Reinsurance $900 Million 0.1% EP $800 Million 0.3% EP RMS Scenario-Based Models 19

20 Conclusions Solvency II internal models impose a new discipline on the risk management processes of a life insurance company Using internal models ensures that the exact amount of capital is held, appropriate for the portfolio In many cases this will be less capital than the standard formula Evidence-based models enable detailed validation and sensitivity analysis to ensure that risk assessments are sound Scenario-based models provide intuitive narrative for understanding the metrics of risks 20

21 RMS LifeRisks Solvency II Whitepaper Solvency II Whitepaper now available on rms.com/liferisks 21