A Practical Approach To Considering The Possible Impact Of Asset Allocation On Funded Status For The

1
A Practical Approach To Considering The Possible
Impact Of Asset Allocation On Funded Status For
The Regulation Of Pension Plans

• Doug Andrews April
  2006

2
Purpose

• Describe construction of Minimum Risk Portfolio
  (MRP)
• Consider impact of asset allocation on future
  funded position when investments deviate from MRP
• Propose an approach to be used by regulators

3
ONT Supervisory Regime

• Going-concern and solvency actuarial valuations
  required at least every 3 years
• PBGF provides limited protection in the event of
  sponsor and plan insolvency
• Annual PBGF fees based on Ont membership and the
  excess of solvency liabilities over solvency
  assets
• FSCO is testing a revised monitoring model and
  form and plans to implement this program in
  spring 2006

4
Duration and MRP

• Ideally all cash flows would be matched with
  respect to amount and timing, but this is
  impractical
• Duration could be easily incorporated in
  valuation cycle
• Solvency basis should be used for regulatory
  purposes
• Dual duration can be used to recognize
  inflation
• Plans in surplus can select Govt bonds to match
  duration

5
Problems In Constructing MRP

• There is not a unique way to construct the MRP
  resulting in different expected returns
• In the absence of CF matching, yield curve
  changes will disturb the match
• Matching very long liabilities may be
  difficult/impossible within risk guidelines
• Many plans have more liabilities than assets

6
The Stochastic Forecast

• Purpose estimate likely change in funded
  position due to asset mix
• Regulators would specify annually the probability
  that the return on various asset classes will be
  less than on a portfolio of Govt bonds
• Sponsors/administrators would calculate the
  probability of any funding shortfall
• The results could be used in setting PBGF fees
  and determining the need for scrutiny

7
Examples

• Durn 8 portfolio with 60 equity/40 bonds has
  48.9 probability of shortfall
• Durn 12 portfolio with 50 equity/50 bonds that
  is only 90 funded has 52.4 probability of
  shortfall

8
Some Problems With Approach

• Requires regulators to publish annually a table
  of probabilities of expected shortfalls
• The weighted average of the prob of shortfall is
  not equal to the prob that the weighted average
  will fall short
• The calculation does not recognize the potential
  severity of the shortfall

9
Some Problems (Continued)

• The calculation does not recognize the potential
  benefits of diversification
• Method cannot be applied without adjustment when
  liabilities exceed assets
• In many circumstances, there is a greater
  likelihood that forecast returns will fall short
  over a 1 year period than over longer periods

10
Responding To These Difficulties

• It is suggested that regulators seek input from
  interested researchers and professionals
• The approach is simple enough to be applied by
  all plans which would not likely be the case for
  more accurate methods
• The method may subject more plans to scrutiny
  which is probably not a drawback for regulatory
  purposes

11
Responding (Continued)

• Although severity is important for calculating
  PBGF fees, this method is an improvement over the
  current calculation
• Assuming that any excess of liabilities over
  assets is invested in universe bonds is a
  practical solution
• Supervisory cycles are typically 1 year

12
Directions For Future Consideration - Norton

• A calculation like MCCSR used by OSFI for
  insurers might be a better basis to assess risk
• I agree and suggest this as an area for research
  by the CIA
• It is surprising that funding snapshots show
  little consideration for asset quality, potential
  volatility due to asset mix, and financial
  strength of sponsor

13
Directions For Future Consideration - Guerin

• Regulator would provide a set of 1000 scenarios
  of 1 year returns taking into account expected
  returns, volatilities and correlations. This
  would be available online to sponsors/administrato
  rs to do calculations.
• This would be more accurate and could be applied
  by larger plans but would be too complicated for
  most plans
• It is unlikely that regulators would wish to
  maintain a system to be used by plans

14
Directions For Future Consideration - Williams

• To address severity, create a table of point
  scores or weights that take into account both the
  probability and potential severity of
  underperformance
• Points systems can be subject to criticism of
  being arbitrary and lacking a sound theoretical
  basis. In this respect the system used by OSFI
  for insurers may be a good example to follow.

15
Features Of This Approach

• Simple and relatively inexpensive for plans to
  implement
• Introduces a way of measuring the impact that
  asset allocation may have on funded position
• Provides a stochastic modeling approach to
  enhance the quality of supervision
• It will likely increase the understanding of many
  sponsors/administrators of the risks inherent in
  the current asset mix