Dynamic Solvency Test Joint regional seminar in Asia, 2005 Asset Liability Management
Evolution of DST International financial reporting changed to a GAAP basis Actuarial reserves were no longer good and sufficient but appropriate Risk-based capital requirement (MCCSR) was seen as static and retrospective There was a need for a dynamic and forward looking study of a company s financial condition 1
Cash Flow Projections Can be used to examine the profile and interaction of the cash inflows and outflows from the company s in force and future new business. At the most basic level, cash flows are projected based on sample policies which give a good representation of the in force portfolio. with investment returns being calculated at a fixed rate on liabilities. Further increases in sophistication can be gained by projecting the cash-flow for each individual asset, allowing for reinvestment. and/or bringing in a stochastic model of investment returns 2
What is DST? A process whereby the business of the company is tested through cash flow projections into the future under a variety of scenarios of possible (unfavourable) experience A process of stress testing A process that allows for the testing of possible management strategies for handling adverse experience A risk management process 3
Stochastic Dynamic Solvency Testing It may be desirable to test certain factors with randomly generated scenarios derived from a known distribution Economic variables Claims (general insurance) For general insurance, this is DFA Useful for pricing Stochastic methods require greater attention paid to algorithms describing business volume, pricing, expenses, investments,.. Data storage becomes a problem Use caution in interpreting the results: Use appropriate risk measures Remember the technical competence of the reader 4
Advantage of Stochastic modelling Enables you to model path-dependency, policyholder behaviour and any type of complex option Allows calculation of market price of any option (whether a similar traded option exists or not) Can allow for management actions Theoretically easy to understand principles Most accurate solution 5
How do we do DST? Construct a computer cash-flow projection module of the company Include both assets and products Specify future experience factors Specify company operating policies Provide for revaluation of actuarial liabilities 6
Basic Risks Life Insurers Mortality New business Morbidity Expense Persistency Reinsurance Cash flow mismatch Deterioration of asset values (credit risks) Government and political action Off balance sheet 7
Basic Risks General Insurers Frequency and severity Pricing Misestimating of policy liabilities Inflation Interest rate Premium volume Expense Reinsurance Deterioration of asset values (credit risk) Government and political action Off balance sheet 8
DST-base test Company s assets and liabilities as at January 1 of the calendar year in which the DCAT is being conducted are projected forward for a 5-year period. The projections should be consistent with the company s business plans and use the actuary s best estimate assumptions without margins for interest rates, mortality, morbidity, expenses, inflation, etc. If the actuary feels the business plan is unrealistic (for example, the level of future new business sales is not achievable) he may modify the base scenario. In that case, the actuary must disclose the changes and the rationale for making them. 9
DST-additional test For life insurers, the actuary needs to consider threats to the company s capital adequacy under plausible adverse scenarios. In addition Consider integrated scenarios whereby a high probability adverse scenario is combined with a less probable Consider the consistency of each adverse scenario being tested Be not limited to these scenarios. If there is an awareness of a situation that could expose the company to future risk, the actuary is expected to test that scenario in addition and quantify the exposure. 10
Adverse scenario Plausible adverse scenarios are include but are not limited to, the 10 listed in section e) below. 1. Mortality deterioration; 2. Morbidity deterioration; 3. Persistency deterioration; 4. Asset/liability cash flow mismatch (C- risk); 5. Deterioration in asset values (asset default C-1 risk); 6. Increasing/decreasing/level new business sales relative to the business plan over the projection period; 7. Expense deterioration; 8. Insolvency or other unexpected termination of major ceded reinsurance agreements; 9. Government and/or political action 10. Off balance sheet exposures. 11
Stochastic Modelling Issues Investment Model All market-consistent models are the same aren t they? Vendor transparency Supplied calibrations Calibrated to companies risks? Option price replication? Deflator or risk neutral? Convergence Economies? Risk free rate? Source: Weapons of Mass Computation Survey, 2004. 12
Stochastic Modelling Issues Modelling Cashflows Choosing the level of detail to put into the model defining the usage of the model Ensuring risks, options and guarantees are captured especially path dependency and policyholder behaviour Assumptions? Number of model-points to use and correct bandings especially for stochastic work Managing the amount of information received from models 13
Stochastic Modelling Issues Software Choice All software products are the same aren t they? Vendor transparency Supplied models Allowable complexity Speed of implementation Speed of runtimes Hardware compatibility and scalability Consultancy support Right tool for right job? Source: Weapons of Mass Computation Survey, 2004. 14
Stochastic Modelling Issues Hardware Considerations 4 elements go into determination of the run time per unit of hardware Model Complexity Number of Modelpoints Number of Simulations Speed of Software Tendency to have as much complexity as software allows Usually mechanical process As few as possible whilst still allowing convergence and calibration Broadly similar for leading packages (exception of old-style Prophet) 15
Asset Liability and solvency regulation survey result Summary of Asset-Liability and Solvency Regulation Resilience testing Cash Flow projection DST Stochastic A-L Testing Financial Condition Australia Canada X Denmark X X X X France X X X Germany X X Singapore X X X South Africa Note: above is the survey result at year 2000 16
DST potential implementation issues in China Should we start with a stochastic or deterministic model? A sophisticated model, or using a simplified approach in trying period? Data limitation on DST analysis? What data we need to accumulated during trying period? What is the main risk factors to be considered in insurance companies in China? Prudence in assumptions Software support? Gap analysis for current solvency system? What are the main outputs need to be considered? 17
Case Study Sample company A Life company: Traditional life portfolio of three basic products, endowment, annuity and whole life at Dec 2003; The whole life business is a lapse supported products with new premium in 2003 is 15% of the total premium; More than 50% of the asset is invested in the government bonds, and remaining is the equity and cash deposit. Business plan assumes rapid growth 18
Case Study Business summary of company A: Whole Life Endowment Annuity Total Total premium in 2001 5,000 30,000 20,000 55,000 Total premium in 2002 9,000 35,000 25,000 69,000 Total premium in 2003 15,000 40,000 28,000 83,000 Reserve @ Dec, 2003 16,000 90,000 60,000 166,000 Net risk amount @ Dec, 2003 160,000 9,000 6,000 175,000 Asset Government bond 100,000 Equity 23,000 Cash 60,000 Total Asset @ Dec, 2003 183,000 19
Case Study What should we consider? What is the solvency requirement according to Chinese regulation? Is the company A solvency according to CIRC requirement? What additional information might we need for DST? What are the main risks for this company in terms of the solvency? Any other potential practical issues in your company is DST applies at this year end? 20
Thank you. 21