In practice, how do pension plans determine their risk appetite?

Question

While I understand DB pension plans tend to use an ALM and surplus management framework to determine their asset allocation and risk/return objectives, I am wondering how in practice they determine the maximum amount of risk they are willing to take on.

I imagine it would be an extremely difficult quantitative problem given that there are many interrelated and difficult to predict variables: the nature of the liabilities, the probability and path dependency of a funding shortfall or benefit cuts, the confidence level of achieving a certain return etc. In practice, what approaches to risk tolerance are used in practice to model these variables, and what has been successful?

Answer 1

I think there are many approaches to setting risk tolerance/appetite limits. Here are some examples that you may find interesting (from the most clearest):

• Risk Management Policy of NZ Superannuation (or universal retirement income):

The Investment Risk Appetite of the Fund is:

1. Relative (1 year) - Fund return 3 standard deviations from Reference Portfolio (> -6.5% or > 8.5%);
2. Absolute (1 year) - Loss greater than 25% of Fund value;
3. Absolute (since inception) - Return >3% below risk free rate;

Also read their Investment Beliefs.

• PSP Investments:

This assumption of the Government’s risk appetite, referred to as the “Risk Appetite Assumption”, is determined by reviewing, through assetliability studies, the pension funding risk resulting from simply investing in a passive portfolio (aka Policy Portfolio) that would replicate market indices of public debt and equity markets. This passive portfolio is designed with the lowest possible investment risk consistent with the Return Objective.

And going through their 2015 Annual Report I found this:

Active management activities form the second pillar of PSP Investments’ approach. These activities are implemented within an active risk limit and the risk appetite to generate additional returns over the Policy Portfolio.

So my guess here is that they are creating their in-house index of Policy Portfolio and setting tracking error limits for their active funds.

You may be able to find some other approaches, but many pension funds seem to be less clear about their methodology. For example, see this presentation to get some idea about NBIM's approach. They also use Benchmark Portfolio and I think have tracking error limits along with other metrics.