Asset Liability Modelling (ALM) Approaches, Techniques, Trends In the Pension Practice

Transcription

1 Asset Liability Modelling (ALM) Approaches, Techniques, Trends In the Pension Practice Chris Brisebois, FSA, FCIA, CFA CIA Investment Seminar

2 Agenda 2 Background ALM in a pension fund context Modeling approaches Assumptions Current modeling standards Closing thoughts

3 Understanding defined benefit pension risk 3 DB pension plans can be exposed to many risks, both in respect of assets and liabilities: Asset Risk Factor Liability Risk Factor Interest rate risk Asset liability mismatch risk Currency risk Manager risk Credit risk Equity market risk Inflation risk Cash flow/liquidity risk Longevity risk Demographic risk Regulatory risk Exposure to above risks can vary, depending on the Plan and investment strategy

4 A look back at traditional investing 4 As traditional thinking goes: Pension plans, due to their long time horizon, can afford to ride out short term volatility of equity markets to realize excess returns over the long term Source: CIA Economic Statistics [4]

5 But what about the last 10 years? 5 Equities have faltered, while bonds have outperformed High bond returns driven by declining interest rates Liabilities and costs have increased significantly Source: Eckler [5]

6 Some key considerations going forward 6 Low yield environment Liability matching asset classes have low expected returns, based on current rates*: Universe Bond yield: 2.3% Long bond yield: 3.3% Impact on liabilities Increases costs Negative impact on solvency/wind up/accounting liabilities Reduces return expectations for the Fund * Source: PC Bond, at August 31, 2012

7 Some key considerations going forward (cont d) 7 Volatility of equity markets Heightened by 2008 credit crisis High correlation of equity markets in periods of distress Less tolerance for short term underperformance From all stakeholders Impact of regulatory changes Pension plan sponsors are looking to the actuary for help!

8 ALM for pension plans 8 A forecast of the financial continuum facing the Plan from both an asset and liability perspective Actuarial valuation with added dimensions of: Time Asset and/or Liability Mix Goal: Identify key risks facing the pension plan Translate economic uncertainty into some estimate of the future dollar consequences to the pension plan (or some other key financial objective) Helps to: Identify ways to control risk and maximize opportunity through investment/benefits policy See the best and worst the future may hold Better understand the dynamics of the Plan financials (stress testing)

9 Interaction between assets and liabilities 9 Wages Pension Bonuses Pension Liabilities Inflation Interest Rates Actuarial Discount Rate PV of Liabilities Funding (Surplus or Deficits) Lower Market Bond Returns Value of Assets Equity Returns Asset Values

10 Modeling approaches 10 Deterministic Develop one or more possible scenarios for how pension assets and liabilities will unfold over the next few years Helps identify key sensitivities, but does not address probability that event will occur Useful for highlighting how bad a bad scenario can be Stochastic Computer models to generate 1000 s of possible scenarios each with its own likelihood or probability of occurring Scenarios are then accumulated into a distribution of results to assess the expected, best case and worst case scenarios the future will hold

11 Typical stochastic modeling process 11 Develop economic/investment scenarios based on the capital market assumptions Run 1000 s of scenarios of inflation, interest rates, and rates of return for all of the asset classes over the projection period Liabilities Run full actuarial valuations for every year for every economic scenario Use model to analyze costs, risks, etc.

12 Stochastic modeling: sample output 12 Projected Funded % in Year 10 Orange shows 5th to 95th Percentile Range Blue shows 25th to 75th Percentile Range Projected Funded Percentage 150% 125% 100% 75% 50% 144% 144% 143% 146% 139% 139% 136% 137% 130% 130% 132% 131% 131% 110% 111% 111% 113% 114% 114% 115% 107% 109% 109% 108% 110% 108% 92% 94% 94% 94% 95% 94% 97% 97% 97% 95% 92% 94% 93% 80% 80% 80% 80% 81% 83% 83% 83% 82% 82% 80% 82% 81% 66% 65% 65% 65% 64% 69% 67% 68% 66% 69% 66% 65% 67% 25% Current 2A 2B 2C 2D 2E 2F 2G 2H 2I 2J 2K 2L Asset Mix

13 Economic models: a key tool 13 Project key financial variables: Asset class returns Bond yields Inflation, Consistency between economic variables is critical Inflation and bond yields, bond yields and bond returns, etc. As the world gets more complex economic model have become more sophisticated Has led to departure from traditional mean/variance approaches

14 Economic models: recent advances 14 Ability to segment risks (e.g. interest risk, credit risk, equity risk, etc.) Common when analyzing Liability Driven Investing (LDI) strategies Fat tails Better modeling of derivatives Value at Risk techniques Regime switching Etc. As models get more sophisticated, specialized expertise is required

15 Importance of assumptions 15 Economic variables typically require the following assumptions: Expected rate of return Expected variability of return over time Expected correlation Other model specific assumptions Liability assumptions: Starting point? Usually last filed actuarial valuation Open or closed group projections? Decrements New entrant profile if necessary Annual valuations? Going concern, solvency and accounting assumptions during projection period? Minimum required contributions? Future changes to benefits?

16 Assumption setting and validation 16 A variety of approaches to setting assumptions: Current valuation assumptions, demographics, Historical data (look at different periods, frequencies, etc.) Consensus forecasts Current pricing Etc. Validating assumptions Reasonable for the purpose Consistency of model output with assumptions Appropriateness of tails Stress testing Other considerations: Consistent risk/reward relationship Awareness of any biases Common sense always applies!

17 Canadian standards for ALM 17 ALM is not required to be done by actuaries No ALM specific standards apply to actuaries within the CIA s SoP Greater scrutiny of ALM techniques coming? E.g. use of acceptable models in NB Risk Shared Plans

18 Standards of Practice for ALM in other jurisdictions 18 Examples: US: ASOP#38 (Actuarial Standards Board) for P&C actuaries UK: Board of Actuarial Standards TAS M Key elements: Confirm reliability of model, suitability for purpose, validation/testing, appropriateness of assumptions, etc. Duties of actuary (e.g. understand the model, etc.) Appropriate disclosures (e.g. assumptions, reliance on experts if applicable, explanations, limitations of model, etc.) Do not tell the actuary which models to use

19 Closing thoughts 19 Pension plans are much more focused on managing asset and liability risks facing their plans Asset/liability modelling tools have made significant advances in recent years to help identify and manage these risks Actuaries should be aware of how these tools can be used to help their clients

20 Questions? 20