MEASURING THE THING ITSELF
read Nikesh Patel’s engaging article Superflat with interest. Nikesh laments the focus on gilt yields and argues instead for a ‘time consistent flat discount rate’ to value pension schemes. It is always welcome to see discussion on alternatives to the current valuation methodology – marking £tr’s of pension liabilities to a tiny (by comparison) index linked gilt market has unintended consequences which Nikesh articulates well.
To my mind asset prices remain the right place to start. We need to be smarter though about how we choose them and have been working with pension schemes and insurers to design portfolios that deliver matching income with a higher yielding asset universe. See Solving The Pension Conundrum.
Pension schemes that approach funding from this perspective should be well placed to meet the member benefit promises…
Which brings me to what really caught my eye in Nikesh’s article: ‘measuring the thing itself’.
As an industry we have tended to focus on primary and secondary funding targets, accounting forecasts, company contributions etc as suitable measures to assess pension design and funding strategy. From a Trustee perspective, in particular, aren’t we forgetting something? Or someone? Lot of people!
Nikesh redresses this balance by looking at the expected shortfall in ruin. Or, if I can paraphrase, the size of broken promises.
In a similar vein, we include net benefit size and benefit shortfall in the Financial Canvas ALM toolkit. These measures focus on the amount of benefits that ultimately get paid, or not, to the members of the pension scheme. It seems obvious but somehow traditional measures don’t!
Net benefit size is the amount of benefits actually paid to members, net of their contributions.
These measures address a similar concept to expected shortfall in ruin albeit that ruin is now defined to be failure of the sponsor, triggering a buyout, rather than the scheme running out of assets. The size of benefits paid is then equal to the benefits paid out before the buyout plus whatever benefits can be secured at buyout. It can be expressed either in absolute terms or relative to the size of the current benefit promises.
Benefit shortfall is the amount of benefits promised but never paid
The calculations can be done stochastically to generate a distribution of outcomes – which is helpful to compare the outcome of different strategies – or deterministically based on scenarios of interest. Crucially it forces us to consider the impact on the sponsor company and the pension scheme together.
If we think about different questions that Trustees face (and which tPR is increasingly pressing them to resolve):
Should we stop or reduce benefit accrual? Close the scheme? Increase member contributions?
What is our integrated funding strategy? How do we define our recovery plan and set the investment strategy? Should we use a fixed discount rate, perhaps(?!)
What is the impact of a partial buy-in, buy-out, a contingent asset, …?
These choices should all be compared by measuring the thing itself
Does this increase the net benefits paid to our members?
Are we reducing the likelihood or magnitude of broken promises?
Who are the winners and losers under different strategies?
This creates a member-centric way of comparing a package of actions, explicitly incorporating the investment strategy, benefit design and the company covenant (including the interactions between the company and the scheme under different scenarios).
All sound familiar? If you’re helping Trustees develop their IRM isn’t this the place to start?
We believe these metrics add new insight to the funding decision process and we’d be delighted to hear from anybody interested in using these techniques or who would like to understand the benefit security of their members…