Can they fix it?
Stephen Bouvier sets out the problems facing the IFRIC interpretation committee on the treatment of single-A corporate bonds in discounting pension liabilities
Can a single-A corporate bond yield qualify as high quality for the purposes of discounting a pension liability under IAS19? This is the question that confronts the IFRS interpretations committee as the result of a request for guidance from the German national standard setter, the German Accounting Standards Board (GASB).
In short, the GASB wants the committee to clarify the term ‘high quality corporate bonds’ and, in doing so, determine whether entities can fall back on the so-called ‘six-A’ – AAA, AA and A – discount-rate proposition.
The IAS19 guidance on discount-rate determination is at paragraphs 83-86 of the 2011 version of the standard. According to paragraph 83: “The rate used to discount post-employment benefit obligations (both funded and unfunded) shall be determined by reference to market yields at the end of the reporting period on high quality corporate bonds.”
The paragraph continues that where there is no such deep market, “the market yields (at the end of the reporting period) on government bonds shall be used.” In addition, it advises preparers that: “The currency and term of the corporate bonds or government bonds shall be consistent with the currency and estimated term of the post-employment benefit obligations.”
Alongside those constraints, IAS19 acknowledges, “there may be no deep market in bonds with a sufficiently long maturity to match the estimated maturity of all the benefit payments.” In addition, paragraph 86 requires that: “In such cases, an entity uses current market rates of the appropriate term to discount shorter-term payments, and estimates the discount rate for longer maturities by extrapolating current market rates along the yield curve.”
The GASB agrees that paragraph 83 of the standard requires entities to discount pension liabilities by reference to market yields at the end of the reporting period on high quality corporate bonds. But, the standard setter points out, IAS19 does not specify which corporate bonds are high quality.
Accordingly, although, according to prevailing opinion, listed corporate bonds qualify as high quality if they either AAA- or AA-rated, current market conditions and disruption, has led to a fall in the number of corporate bonds rated with a AAA or AA rating. As a result, a single trade has the potential to influence market yield, distort the observable market rate, and in turn distort the discount rate.
Aon Hewitt principal Simon Robinson broadly agrees the thrust of the German analysis.
“The issue that confronts practitioners is that there is a lack of long-dated corporate bonds,” he says. This in turn raises the question: how do you derive a discount rate? In my experience, one of the reasons for a wider range of discount rates is because there are few long-dated bonds.
“I think that the falling population of high-quality corporate bonds is becoming a bigger issue. Back in 2008 there were about 250 GBP-denominated corporate bonds that you would classify as AA. That has now come down to about 70. So you have a much smaller dataset and that results in a much less credible answer because of the risk of a sampling error.”
And, Robinson agrees, arriving at what he calls a “sensible yield curve” to match the discount rate to the duration of the liability is a challenge. “Even before you start, you hit the problem that a typical UK plan has a duration of 20 years or more, whereas the durations on the corporate bonds that you will be looking at are substantially shorter. They face a similar issue in Europe.
“If you look at the longest iboxx bond index (the over-15-year index), you will see that it typically has bond durations of 12–13 years. And, you have to remember that there is not a lot of data in that index. In other words, your IAS19 discount rate might have as its starting point both a lack of available data and a duration mismatch between the scheme liability and the basket of corporate debt that it is based on. These factors have a huge effect on the discount rate.”
Monthly variations in the available yields also mean that the discount rate will hinge in part on the market yields available for sampling at a company’s annual reporting date.
Robinson notes: “As with any basket, if you have 20 bonds that you consider are long duration, the yield at an individual bond carries relatively little weight. If there are only four long-dated bonds, however, the credit rating of that individual bond will have a disproportionate effect.
“It is important to bear in mind that an individual bond might only relate to an issuance of, say, £100m [€123m]. Frankly, that is a poor choice for discounting a liability of, say, £5bn worth of liabilities. And the fewer bonds you rely on — and this is also an argument against using a single bond — any credit downgrade would have a massive impact and destroy comparability.” So, the starting point for arriving at a discount rate is an initial universe of all high-quality corporate bonds – perhaps with weeding-out criteria for outliers. This starting point is dictated by both the IAS19 requirement to make an unbiased assessment of the discount rate, and also by auditor concerns about cherry picking.
“So once you have arrived at a basket of yields,” Robinson explains, “you have to make a subjective assessment about where the yield curve will go to match the duration of the pension liability.” And it is at the point at which the duration of the pension liability exceeds the duration of the corporate bond yield, that what he calls “subjective extrapolation” leads to greater uncertainty over the assumptions that you are making.
As for where the issue is most pressing, Eric Steedman, a consultant actuary with Towers Watson, suggests it is no accident that the request for guidance emanates from the euro-zone. “Most of the focus has been on the euro-zone, although Canada is a market which also faces challenges from the shape of its bond market. The course of the debate has been different there, however,” he says.
“In the euro-zone, you have a deep market at the short and medium durations. As you come out further along the yield curve, the number of bonds starts to fall off. IAS19 specifies that you need to extrapolate – paragraph 86 of the new version.
“So there has been some dispersion in different models in the euro-zone – practitioners have extrapolated in different ways and made different judgements around the data that they have available. A major issue is the very large fall in yields in the last year. Euro-zone discount rates are down by as much as 200bps, depending on the period you are looking at and the length of the liability.”
And Clive Fortes, a partner and head of corporate consulting with Hymans Robertson, says the UK market is not entirely free of tensions. “Since the start of the credit crisis, we’ve seen in the UK a rapid reduction in the number of AA-rated corporate bonds with a large number of financials being downgraded and a wide range of yields on the remaining AA-rated corporate bonds,” he says.
“At the end of October 2012, long dated AA-rated corporate bonds yielded between 3.2% to 5.2% per annum. And this pattern is repeated across the euro-zone. The shortage of benchmark AA-rated yields to be used to discount pension scheme liabilities is even more marked at very long durations suitable to the maturity of pension scheme liabilities.
“In spite of this, UK companies have retained a relatively narrow spread of discount rates for IAS19 purposes. As at 31 December, 68% of FTSE350 companies used a discount rate for IAS19 purposes that was within 10bp of the AA over-15 year yield. That’s the herding instinct in play. Most companies want to be ‘in the pack’ and, in the UK at least, actuaries and auditors have helped to keep the pack together.”
Steedman suggests the committee will struggle to come up with a fix. “I don’t think it can be completely ‘fixed’, unless they exercise a degree of prescription that would be out of character for the way that IFRS usually works. It would be like issuing application guidance.
“So, unless standard setters get a lot more prescriptive, I think we will simply have to accept that when you have data that is thinning out, the scope for putting different interpretations on that data grows larger if the data grows more dispersed.”
So, although the committee plans to revisit the issue in January, it might be the case, as staff in effect concluded, this is simply not the committee’s problem. “We think that paragraphs 83–86 of IAS19 provide sufficient guidance for determining the discount rate.”
Entities, they conclude, “should apply their judgement in determining a rate that complies with this guidance, taking into account the characteristics of their capital markets.”