 Most pension plans rely on two sources of funds to meet future obligations:
- Assets that the plan has right now
- Contributions that will be made over future years
Deterministic ALM (stands for asset-liability modeling), answers the question from a plan sponsor:
If I follow this funding policy and investment policy for making contributions, how will my assets and liabilities develop over future years?
In other words, the ALM package extends the concept of comparing assets to liabilities. Instead of looking only at the comparison as of today, ALM looks at the comparison over an extended period such as 30 years.
ALM dynamically models the functioning of the pension plan into the future. For example, the following things happen every year:
- Members accrue benefits according to plan specifications
- Contributions come in according to the funding policy
- Benefit payments go out
- Assets earn returns
- Actuarial valuations are performed
ALM allows the plan sponsor to test different funding policies and investment policies for making contributions. Such testing would be done in order to select the policies that are most likely to meet the goals of the sponsor.
ALM allows sensitivity testing to be performed, whereby the plan sponsor can see how the asset and liability picture changes if, for example, assets earn a higher return than over recent history.
STOCHASTIC ALM
Now Lynchval is enhancing ALM by implementing stochastic projections. The term stochastic refers to projecting asset performance under a large number of interest rate and equity return scenarios. Each scenario represents a random selection of future returns. But random does not mean purely random. Each scenario is an intrinsic whole, where for example, we allow for the fact that high shortterm interest rates have historically correlated with poor equity market returns. As another example, we allow for the fact that the yield curve has completely inverted only about 1% of the time over the last 50 years. A great many such consistencies are built into the projection software, so that when a user runs 1000 scenarios, the result is a highly scientific view of likely results. Output can be analyzed in terms of best case, worst case and standard deviations from the mean or by percentiles.
The stochastic enhancement answers the following type of question from a plan sponsor: If I follow this funding policy and investment strategy, what are the chances that plan assets will be sufficient to meet plan liabilities over the next 30 years?
Both the traditional valuation software and the ALM enhancement were motivated by a desire to help the pension plan community. The same is true of the more recent stochastic development. At Lynchval we are always striving to empower our clients.
|
