Briefing: Pensions Accounting, Back to basics
The International Accounting Standards Board has revealed plans to issue a due process document on the future of pensions accounting. Stephen Bouvier asks where this might lead.
What should a pensions accounting standard deliver? That is the question that the International Accounting Standards Board must answer as it embarks on its plan to issue a blue-skies research paper on the future of employee benefits accounting.
The current methodology for calculating pension liabilities is found in International Accounting Standard 19 (IAS 19), Employee Benefits. Although this might lack definitional certainty, it manages to plug the balance sheet with a form of best estimate – albeit one that skims variables such as risk and measurement uncertainty.
Eric Steedman, a London-based consultant actuary with Towers Watson explains: “If you look at the existing accounting, it is focused on measuring the expected cash flow. There is a range of possible outcomes that do not automatically form a perfect bell curve.
“The use of a corporate bond discount rate – usually – is probably best seen as a pragmatic approach in that it lacks a strongly articulated theoretical basis but it is nonetheless something that, despite throwing up challenges, is relatively simple to apply.”
That accounting is in need of a fix is clear from the number of requests for guidance submitted to the IASB’s interpretive committee since 2011 when the board published a series of amendments to IAS 19. This predictable state of affairs – to the chagrin of many IASB constituents – was apparent from 2009, when the IASB scaled back its work on pensions (see box).
Despite the IASB’s efforts to make the pensions footnote more relevant by requiring sponsors to disclose the risks inherent in their provision, communication remains a stumbling block, according to Hymans Robertson IAS 19 specialist Jon Hatchett.
“We have many pages of notes to the accounts on pensions, but still investors feel they are poorly served, “ he says. “This tells me that there are tensions with the status quo. IAS19(r) has added more detail on the asset side, but it is unclear that it has given investors better insight into scheme risks.”
An ongoing struggle to fix pensions accounting
The IASB’s efforts to tackle pensions accounting over the past decade have met with failure. Although IAS19 methodology works well for traditional defined benefit and defined contribution plans, the emergence of so-called intermediate-risk plans has stretched the standard to breaking point. The IASB’s interpretations committee tried to break the impasse with its IFRIC D9 initiative.
That ground to a halt in 2006 when it launched its ill-fated project to develop a new approach to accounting for the new designs. The effort led in 2008 to a widely panned discussion paper, which considered a full fair-value approach for some, but not all pension promises.
Consequently, the board made quick fixes to IAS19, among them the removal of deferral and amortisation mechanisms and the introduction of the net-interest approach. In May 2012, the IASB’s interpretations committee decided to revisit the IFRIC D9 project before calling time on it in May 2014.
So what might a fundamental review of pensions accounting look into? “If you examine what benefits are being valued,” says Steedman, “there are some interesting questions around benefits that are not yet vested and amounts that you might expect to pay, such as the effects of future salary increases in a final salary or career average plan.”
He continues: “Around discounting, there are some questions about what it should capture. Time value of money, yes, but should it also capture an element of credit risk or an illiquidity premium?” Whichever way the standard-setter goes, preparers need a solution that is practical, easy to operationalise, and which offers comparability.”
According to Steedman, a further consideration is whether the board ought to evaluate the difference between settled and non-settled measures. “There is a real difference between benefits that have been settled, including DC plans, and benefits that have not,” he notes. “An interesting question would be whether the IASB feels that difference should be reflected in the accounting or not.”
And what of fair value? The IASB’s 2008 discussion paper failed because it drew a series of arbitrary lines that forced fair value on some pension promises, but not on others with similar economics.
This time, the board has learnt the lessons and is talking in terms of a single approach for all plan designs. “The IASB knows, and a lot of pensions professionals know, that an obvious solution to the challenge of accounting for hybrid pension is to value all pension obligations in the same way,” says Martin Lowes, a partner at Aon Hewitt.
“This would be a major change from now where defined benefit and defined contribution promises are approached differently, with a standard level of credit risk built in for all defined benefit promises,” he says.
“Invariably, that single measurement model is fair-value accounting. So, for example, rather than having a discount rate that is set in an arbitrary way, you adopt a fair-value approach and use something that is appropriate for the degree of risk in the plan.”
So, although fair value would be one way of dealing with the problems the standard setter has run up against when looking at contribution-based promises, the three professionals who spoke to IPE agree that any such move would have consequences.
“Using fair value, however it was defined, has the potential to remove awkward boundaries and certain incentives,” says Hatchett. “But it would likely add £100bn (€68.8bn) of pensions debt to FTSE 350 balance sheets. It wouldn’t change the underlying position on which those schemes are funded or the risks inherent in them, but it would make them look much worse at a stroke. That doesn’t feel right to me. How would that help investors?”
More ominously, Hatchett adds: “You would expect to see political resistance with numbers like that. I imagine one or two large companies would be close to balance sheet insolvency.”
Lowes says that, however good the idea might sound in practice, it would make pensions accounting more onerous. “The annual calculations will be much more challenging. It will increase liabilities for funded risks where the credit risk is limited,” he says.
But irrespective of any eventual pensions model, what should it capture? “Investors would value an understanding of the level of risk being taken with the strategy for funding the scheme,” says Hatchett. “For example, identifying the present value of the current deficit contributions payable and what it might be if, say, a one-in-six downside event happens. That would provide greater transparency around risk.
“There would be no need for organisations to set any additional funds aside, so it wouldn’t be ‘Solvency II for pensions’ through the back door”. It would provide useful information for investors and reward those companies that control pension risk.”
But taking the long view, it is hard to ignore the fear that this effort, like those before it, will crash and burn. As Steedman notes on discount rates: “I’m yet to be convinced that once they go through all of these issues that there is going to be something that emerges that is clearly better than the existing corporate bond rate, for all its imperfections and flaws.”