Shifting the goal posts
Who would want to be a chief financial officer with a defined benefit pension plan on his books? Sometime during summer 2006 you probably reassured your board colleagues that the International Accounting Standards Board's Phase I accounting project looked unlikely to snag your plan, only to find that within a year the goal posts had shifted.
Your first reaction is no doubt to blame the press. Those handy reassurances in the pensions press that the IASB's review of pensions accounting "does not bring into question the twin classifications of defined benefit and defined contribution accounting" doubtless helped to reinforce that impression.
In a bid to settle the debate, IPE has conducted an analysis of what individual IASB board and staff members said as they deliberated so-called ‘troublesome' plans. We set ourselves a high evidential burden. Rather than rely on individual recollections, we have quoted only official IASB sound archives, published agenda papers and the board's official journal, IASB Update. The only time we broke this rule was to quote an article by this author that relied on these sources. We have asked IASB to support any response to our analysis with documentary evidence.
The analysis was sparked by comments made by the IASB's director of research, Wayne Upton, during the December 2007 board meeting. He said: "We've always said we were going to change the measurement of pensions … The agenda proposal that the board agreed with said that we would do that. That was part of the project plan from day one. This is what we proposed in the agenda proposal, that was the decision the board reached, and this is consistent with the typed decision."
His comments were in response to the charge from board member Jim Leisenring: "The frustration people are having is that we keep saying we are not changing the measurement of pensions and we are." Leisenring added: "We've changed dramatically the population of what is in there or not." His board colleague Warren McGregor agreed.
A request for clarification to the IASB's press office has met a stone wall. Citing staff holidays, a spokesman said "on this occasion we are unable to help". A further request for clarification as IPE went to press remained unanswered.
IPE's trawl started with the observer notes accompanying IASB's December 2006 discussion on pensions accounting. These documents, prepared by IASB staff, reveal that: "The board agreed to address the accounting requirements for cash balance and similar plans in Phase I of the employee benefits project." Under Phase I, constituents can expect to see the board's much-delayed discussion paper during the first quarter of 2008. Following publication of an interim accounting standard, the board plans to mount a comprehensive review of pensions accounting with its US counterpart, the FASB, in Phase II.
Introducing this paper, pensions project manager Jenny Lee said: "The key problem that this part of the employee benefits project is trying to address is that the prescribed measurement method under IAS19 for these types of plans appears to be inappropriate. The board did look at quite a lot of the key issues that arose because of that in the education session."
The "key problem" that Lee identified in December 2006 is amplified in Agenda paper 3, which accompanied the board's 16 October 2006 education session. Detailing the project scope, this summary of the board's agreed remit is worth revisiting in detail:
"At the July 2006 meeting, the Board voted to add an employee benefits project to its active agenda. The project, to be conducted in two phases, will be a comprehensive reconsideration of the accounting required by IAS19 Employee Benefits. One of the issues to be included in the first phase of the project is the accounting for cash balance plans and the definition of defined benefit and defined contribution plans."
With the trail leading back to July 2006, a key question is how this summary suggests that the board is on the verge of grouping a UK career-average plan among the troublesome plans. It is one that IASB is in no hurry to explain.
Paragraph 2 of the agenda paper explains that cash-balance plans, or intermediate-risk plans, are "generally understood to refer to a narrower group of plans than those being considered in this aspect of the project." If any CFOs had missed the point, paragraph three of the same agenda paper spelled the message out: "A key consideration in respect of intermediate risk plans is the identification and measurement of the risk to which the entity is exposed in respect of an employee benefit promise."
Back at the October 2006 education session for board members, clues that plans such as UK career-average salary plans might feature alongside the more troublesome Belgian or Swiss intermediate-risk plans were hard to find.
With a focus on Belgian intermediate-risk plans, Deloitte's Gert de Ridder delivered a remarkably prophetic analysis of the key problem presented by the existing IAS19 discount rate and certain Belgian plan designs. "Currently, IAS19 prescribes for all defined benefit plans the projected unit credit (PUC) method," he explained on page 13 of his presentation.
By now, any CFO with a UK career-average plan on his books must have felt fairly smug, especially when De Ridder chronicled three deadly assumptions in PUC accounting that doom that methodology to conceptual inadequacy in Belgium: projection of benefits using best estimate assumptions; attribution of projected benefits to periods of service; and discounting projected benefits using a high-quality corporate bond yield. Whatever the shortcomings, if any, with PUC accounting when it meets a UK career-average plan, few would have expected this analysis to apply.
Whether he intended it or not, De Ridder nailed the key to unpicking the IASB's Phase I project: he quoted the standard-setter back at itself. Although the board will likely respond that its position is as stated in its discussion paper after completion of its due process, many in positions of influence in the EU, IPE has learned, will show an interest in those deliberations.
The August 2005 IFRIC Update, De Ridder said, argued that: "projection of assets based on an expected rate of return, different from the discount rate, leads to problems." IFRIC D9, he continued, "foresaw projection based on minimum guaranteed rate of return and recognition of an additional liability ‘just' to allow for excess returns obtained in the past". De Ridder's precision appears to exclude UK career-average plans from the project scope.
Also presenting to the board in October 2006 was Hewitt Consultants' Tim Reay. If he imagined that career-average plans were in the IASB's sights, he was not telling during a follow-up interview. "Our dialogue strayed into hybrid DB-DC plans in the UK in the shape of a retirement scheme featuring a defined contribution element for employees aged, say, below 35, and a defined benefit plan for older workers above that or some other threshold; our focus remained, however, strictly on what we understood to be problem plans."
One solution that our discussion explored at length was deconstruction of risk. Risk, Reay explained, falls primarily into four categories: asset value risk, interest rate risk, mortality risk, and what he calls "salary risk". In an uncanny anticipation of where IASB now finds itself in delimiting DB plans, we wrote that he "warns against treating any category of risk as mandating an outcome".
By the standards of the certainty and stridency of Upton's December 2007 rebuttal, IASB's senior project manager Anne McGeachin appears to have been uncertain of her own project's scope - the recording of the December 2006 meeting from 15 minutes onward makes for useful listening.
"Can I just check what you want in the scope of this aspect of the project," began McGeachin. "Say you had a benefit that was simply 10% of salary and then a guaranteed return on that of 4%. Are you saying that you want to do something different with that than the traditional defined-benefit accounting, because that is not something that I had expected? That benefit is not something on its own that I would have expected that you wanted to deal with at this stage."
She continued: "I certainly think it's defined benefit. I also think the IAS19 methodology for defined-benefit isn't so bad for that. You project forward at 4% and you discount back at what might be regarded as a slightly dodgy discount rate. The things that people complain about with IAS19, the DB methodology, are where you say you have a contribution of 10% going in every year but it's guaranteed that you're going to get the return on, say, the FTSE100. The problem there is that you project forward at the expected return on the FTSE100 and discount back at a corporate bond rate."
Whatever, if any, her confusion was, appears to have lifted by 5 June 2007 when she took part in the employee benefits working group. At paragraph 7 of agenda paper 1A of a background briefing on project scope, staff explained that: "To limit the scope of Phase 1, the board decided the issues to be addressed should meet the following criteria: … a. the issue causes current problems in post-employment benefit accounting,…b. there are alternative solutions to the problem that do not fundamentally change the techniques currently used to measure post-employment benefit obligations for the types of plans contemplated when IAS19 was written."
Given IASB's sudden reticence, an overly generous reading would suggest that there is scope for disagreement over the words "when IAS 19 was written". Equally these words do not appear in the board's published agenda decision in the July 2006 IASB Update. But even applying the staff's June 2007 test of "when IAS19 was written", it is questionable whether the plan described by McGeachin in December 2006 was contemplated at that point.
Like De Ridder, McGeachin was clear where the problem lies: "If you just had a benefit that was simply a 4% guarantee on contributions paid in, the defined-benefit methodology in IAS19 would cope fine. You would project forward your lump sum at 4% and you would discount it back, and you wouldn't get the problems that we've got." She explained that: "The problem that arises ... is the optionality between an asset-based pot and a fixed return of 4%. No-one knows how to account for that ‘higher of' benefit under IAS19."
Far from confused was then IASB member Tricia O'Malley who ranks as an IAS19 specialist. Her response to McGeachin's analysis makes interesting listening for German plan sponsors offering DB plans with, perhaps, a reinsurance feature. At some point, the sponsors of these plans - effectively DC plans - will need to justify their accounting treatment in a submission on the IASB's due-process discussion paper.
"But the whole problem is that you are calling it a DB plan instead of the fact that it's a defined contribution with a guarantee on it. If we just accounted for what it is, instead of calling it DB when it's a defined-contribution plan, then we wouldn't have the problem," she said.
One IASB member who did think that this class of plan was up for grabs was Mary Barth: "I guess that's where I got confused too, Anne, because I thought we would deal with that here and I thought that's one of the reasons you folks were looking at the optionality in guarantees and trying to split between asset-based and benefit-based plans. Maybe there's miscommunication." Maybe.
McGeachin double checked. Jenny Lee reminded the board that a 10%-of-salary plan was not a defined contribution plan under IAS19. The solution, summed up IASB chairman, Sir David Tweedie, was to revisit IAS19's definitions:
"We obviously don't agree with the definitions that we have in 19 anyway. The first thing is that Anne and Jenny need to come back with definitions that they feel are clear. Then I think it would probably help if we had these schemes set out so that we know what it is that we are talking about, because I think that we are probably talking past each other at the moment."
By now the project had started to drift a long way from what many believe was its original project scope. And no-one at the IASB is in any hurry to explain the drift.
IASB phase I pensions project timeline
• July: IASB launches phase I of a two-phase project to tackle pensions accounting, in particular ‘troublesome'
• October: Hewitt Consultants' Tim Reay and Deloitte's Gerd de Ridder present an education session to the board on intermediate-risk plans
• December: A discussion of the IFRIC D9, derivative and so-called deconstruction approach to pension obligation measurement ends with a lack of clear consensus as IASB agrees to tackle plan definitions
• February: Staff present preliminary plan definitions, labelling intermediate risk plans asset-based for the first time
• March: Some board members express a clear preference for presenting changes in a post-employment benefit
obligation in profit or loss
• May: Introducing the notion of a ‘defined-return plan', IASB reveals that "some career average plans ... would be
classified as defined-return promises using the new classifications"
• June: IASB holds first meeting of its employee benefits working group
• July: With defined-return plans still labeled defined return, IASB continues its discussion of measurement issues
• September: Board suggests measuring defined-return promises at fair value but notes it has yet to develop its
understanding of fair value. It confirms that a defined-benefit-promise is any promise that is not defined return
• November: IPE quotes IASB member Mary Barth from 19 September 2007 saying: "I think we have to be honest with the world about what we're doing as opposed to pretending that we are not changing things when we are"
• December: All change as defined-return plans switch moniker to contribution-based, a new plan classification that rolls up existing defined-contribution plans
• The much-delayed discussion paper is delayed until sometime during the first quarter of 2008.