ALM processes and techniques in insurance

Transcription

1 ALM processes and techniques in insurance David Campbell 18 th November PwC

2 Asset Liability Management Matching or management? The Asset-Liability Management framework Example One: Asset risk factors for India Example Two: Optimal portfolio for guaranteed endowments Implementation issues 2

3 Trends The traditional approach was all about matching up cash flows over the contract duration. The main challenge is finding financial assets that closely approximate the long duration of actuarial liabilities. Persistent low interest rates and volatile, poorly performing equity markets have shifted the centre of attention. Managing asset-liability risk is the key concept. 3

4 Capital at Risk Break even Average Frequency Crisis Management Conditions Normal Operating Conditions Economic Capital 4

5 Capital at Risk Model uncertainties in projected cash flows on both assets and liabilities. Include interest rate, equity market, currency and all other risks on a consistent basis. In any particular situation, adapt the approach to fit the circumstances and needs. The model gives insights. It does not manage or hedge risk! 5

6 Capital at Risk The approach allows one to: Model the distributions of the Net Asset Value at future time points Identify the duration mismatch between assets and liabilities Quantify and compare different asset allocation strategies Verify or refine the current portfolio benchmark Refine product mix strategies 6

7 Risk exposures Interest Rate Interest rate Re-investment Realisation Currency Credit risks: bonds, derivatives, reinsurance, tenants 7

8 Liquidity risk Horizon financial instruments Horizon insurance products Moment of payment : lapses, renewal of contracts sometimes surplus sometimes lack of liquidity Issues Is it possible to find appropriate funding means at favorable conditions? What is the quality of the «rating» that credit institutions will grant to the company? Is it possible to realise assets? 8

9 Technical risk Life Insurance Mortality Morbidity Lapses Non-life Insurance Catastrophe Severity Frequency 9

10 Operational risk business risk Changes in the legal environment (eg. Tax issues) Competition Products Intermediaries (brokers, ) IT breakdown Fraud Elliott Spritzer 10

11 The ALM process Some basic principles The implementation of ALM requires : an active and reactive process a feed-back process environment external constraints internal constraints objectives Balance sheet Assets Liabilities Risk analysis Current position at risk ALM modelling Optimal allocation Strategic Tactical Operational feedback process 11

12 The ALM process Formalising objectives & constraints Horizon : the projection period has to be defined Run-off of the existing portfolio Introduction of future premiums From existing contracts From new production Both objectives and constraints must be clearly identified and formalised The definition of the objectives requires a precise definition of the concepts of return and risk The constraints should be expressed in a way enabling their modelling 12

13 ALM techniques some basic concepts Returns Purely financial financial return on assets (to be maximised) minimum expected return to be achieved (to be exceeded) Combination of assets and liabilities funding level : ratio assets / liabilities through a (legal or internal) measurement of the obligations RAROC 13

14 ALM techniques some basic concepts Risk Purely financial : standard deviation and related measures standard deviation of the return : measurement of the deviation to the average downside deviation : similar to the standard deviation but only considering the unfavourable cases 14

15 Example One Asset risk factors for India Identify likely fall in values from market volatility or default Practical benchmark in assessing solvency margin Crude but consistent Given the current market value and a risk- free investment rate (i.e. the investment rate generated by Government bonds of the same duration as the investment), what is the discount to current market value for which we have a certain level of statistical confidence that the market value in 12 months time will not be below such discount market value. 15

16 Example One Asset risk factors for India Equities Bombay SENSEX 10 year history Average Volatility Time Period (days) % % Best estimate discount factor Volatility 10.0% 15.0% 20.0% 25.0% 30.0% Risk Free Interest Rate Discount to market value 7.0% 9.5% 17.1% 24.3% 31.1% 37.4% 8.0% 8.5% 16.3% 23.6% 30.4% 36.8% 9.5% 7.2% 15.0% 22.4% 29.4% 35.8% 12.0% 4.8% 12.9% 20.5% 27.6% 34.2% 13.0% 3.9% 12.0% 19.7% 26.8% 33.5% 25% 16

17 Example One Asset risk factors for India Bonds Objective ratings unavailable Segment according to coupon rate relative to government bond rate in year of issue Assume uniform probability of default over the term Rating Probability of default US RBC Factors Canada RBC Factors AAA 5.6% 0.3% 0.25% 6% AA 5.1% 0.3% 0.5% 6% A 5.4% 0.3% 1.0% 6% BBB 8.6% 1.0% 2.0% 10% BB 11.4% 4.0% 4.0% 10% B 9.6% 9.0% 8.0% 10% B- 16.7% 20.0% 16.0% 17% 17

18 Example One Asset risk factors for India Mortgages and Property Map basic cash flows to bonds Judgemental adjustments for risk such as tenant default 18

19 Example Two Optimal Portfolio for guaranteed endowments Liabilities : pure endowments Model point Age Term contract Math Reserve Math Reserve PS Premium MP MP Each model point corresponds to 100 contracts Total amount of mathematical reserves =

20 Example Two Optimal Portfolio for guaranteed endowments Assets On the financial markets : 2 assets available Asset class Expected return Volatility of return Bonds 5% 2.5% Equities 9% 9% Correlation factor = Current strategy : all in bonds Total amount of Assets =

21 Example Two Optimal portfolio for guaranteed endowments Objective : to ensure a return of 5.5% on the contracts Risk : expressed in function of the funding level and related indicators Funding ratio defined as : Assets / (Reserves + solvency requirements) Probability of underfunding Capital requirement 21

22 Example Two Optimal portfolio for guaranteed endowments Horizon of projection Increase in future premiums : 10 years : 3% a year Transition Mortality : usual mortality tables +stochastic deviate Surrender : natural rate (1+factor) + stochastic deviate Natural rate : 5% Factor : function of the level of bond returns 75% of the lapses are cash-outs Regulatory solvency margin = 4% mathematical reserves Returns are normally distributed 22

23 Example Markowitz efficiency Efficient frontier 10.00% 9.00% 8.00% Return 7.00% 6.00% 5.00% 4.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 8.00% 9.00% 10.00% Volatility 23

24 Example step1 :Markowitz efficiency Portfolios Equity 0.0% 12.5% 25.0% 37.5% 50.0% 62.5% 75.0% 87.5% 100.0% Bonds 100.0% 87.5% 75.0% 62.5% 50.0% 37.5% 25.0% 12.5% 0.0% Return 5.00% 5.50% 6.00% 6.50% 7.00% 7.50% 8.00% 8.50% 9.00% StDev 2.50% 2.30% 2.70% 3.50% 4.49% 5.56% 6.68% 7.83% 9.00% The current strategy (#1) is not efficient! 24

25 Example step2 :Monte-Carlo projection Number of contracts alive Number of contracts surrendered Cash-ins Premiums Return on Assets Cash-outs Lump sums paid in case of life Surrender value in case of lapse Evolution of stocks Assets Mathematical reserves 25

26 Example step2: Monte-Carlo projection Stochastic evolution of Reserves 250 sample paths

27 Example step2 : Monte-Carlo projection Evolution of Cash-Outs 250 sample paths

28 Example step3 : information processing by the «ALM engine» Huge amount of information to process Decision criteria derived from the objective & risk definitions 5%-percentile for the funding level Probability of underfunding Capital requirements (based on a VaR level of 97.50%) 28

29 Example step3 : information processing by the «ALM engine» Funding ratio % 80.00% 60.00% 40.00% 20.00% 0.00% PcDn P ro b a b ility o f U n d e rfu n d i n g 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% A v g 0.00%

30 Example step3 : information processing by the «ALM engine» Capital requirements Allocation 4 (63% bonds, 37% equities) better meets the objective More than an expected return of 5.5% is required! Allocation 1 (current strategy) is not in line with the objective 30

31 Implementation Issues Purpose Accuracy Which business Completeness What do we have already 31

32 Implementation Issues Management practice Fair value Market consistency Crisis Management Conditions Break even Normal Operating Conditions Average Capital Extreme market events Asset correlations Frequency Economic Capital Value 32

33 Implementation Issues Testing and Audit Asset and liability model point simulations should statistically match and converge to deterministic projections Portfolio asset and liability simulations should statistically match and converge to deterministic projections Economic Scenario Generator must be correctly calibrated to local economy Management decision rules should work in a wide range of deterministic scenarios 33

34 ALM processes and techniques in insurance PwC