Headline pension deficits make for great headline copy. The formula is simple enough. Take the market value of a pension plan’s asset portfolio - if any - and deduct the scheme’s liabilities. With any luck the net pension asset will turn out to be a number less than zero - a pretty safe bet in recent years. All you have to do now is to sprinkle around a few words like “black hole” and in no time at all you’ll have a “pensions crisis”.

If only the real world were that simple. It is dangerous to assume that this headline accounting figure, argues Heinz-Joachim Neubürger, reveals more information about the funding status of a pension plan than it in fact does: the balance sheet figure is nothing more than a present value. Neubürger was until April 2006 the chief financial officer of German multinational Siemens. Widely regarded as ‘more than just a CFO’, he oversaw the transfer of Siemens’ on-balance sheet pension scheme asset base to a Contractual Trust Arrangement on Siemen’s move from German to US GAAP reporting. Latterly, private equity outfit Kohlberg Kravis Roberts has hired him to grow its German and Asian business.

“The first comment you can make,” he says, “is that the accounting number is just one number. It contains no information content on how the plan liability develops over time. Nor does it reveal how, therefore, cash might be required either from its plan assets, or more critically if there are no plan assets, from on-going operating cashflow.” This shift into the more complex accounting territory of future cashflows - and both their analysis and predictability - is an unwelcome message to those journalists and activists who have assumed that they can extrapolate more than they really can from the headline balance sheet position.

When it comes to a headline deficit figure, big does not always equal bad. A failing business model is just one example of how small could just as bad, or ever worse. “The seriousness of a deficit,” Eric Steedman, a UK-based senior international consultant with Watson Wyatt, points out, “will depend on the company’s ability to rebuild that deficit. It is of less concern in a company that is strong and has plenty of time to repair it than in a weak company with a mature workforce. Again, this is a useful key indicator but it isn’t the whole story.”

And the development of plan liabilities over time are an important indicator of the demographics of the plan beneficiaries, adds Neubürger. “Take the demographics of a European steel manufacturer. The number of pensioners and new retirees might well peak over the next few years, after which the liability and cashflow requirement will start to decline substantially. If you have a very young plan, on the other hand, a business could have 20 or more years to fund any current underfunding, albeit that the liabilities are bound to increase annually. “It is exactly that liability development over time and the related cash call that is so critical to understand. Let’s assume as an extreme case where you are operating near your maximum borrowing capacity and you realize that the cash call of your pension plan will go up sharply - purely due to the demographics of your workforce. In such an extreme case, you may have to shift money away from your operating business towards meeting pension liabilities. This in turn explains why the CFO of a multinational corporation has to trespass on what was until now human resources’ territory.”

The current accounting framework, he explains, fails to differentiate between a business whose demographics mean that its workforce is at or nearing the peak of its retirement profile and a younger business with time on its hands to address a defined-benefit funding shortfall in their annual accounts. “It is also a very critical distinction compared with a situation where you can say ‘I have a very young workforce, their peak retirement phase will not start for the next 15 years, and so I have much more discretion about when I chose to fund that requirement.’ It is this distinction that you are not seeing in the current accounting framework.”

The lay political debate around defined-benefit pension plan deficits in some minds fixes on the size of an IAS19 deficit relative to a notion often as nebulous the “size of the plan sponsor”. For a blue-collar steelworker, it can be a daunting prospect to read in the press that your employer’s IAS19 deficit is equal to, perhaps, half its book equity. But how useful is this metric in reality? Is it any more useful than the old standby commonly resorted to by the UK media of measuring geographical areas relative to “the size of Wales”?

“My initial reaction to this ratio is that it relies on two stock measures”, responds Crispin Southgate, of Pentangle Pensions Consulting. More recently, IASB appointed him to their pensions working group advisory body. “It compares a point measure of the pension ‘debt’ with the value of the company, but it doesn’t tell you anything about the capability of the company to pay off that debt with cash. The comparison of two stock measures is not completely devoid of flow information; market cap will take some account of the cash-generating ability of the business, but what the ratio does not do is to compare free cash flow with the overall IAS19 deficit.”

“If you take the gross pension liability, without considering the plan assets, the comparison that you make is crazy,” says Neubürger.

“The ratio to market cap makes sense, if at all, relative only to the underfunding. Even then, I doubt whether the comparison makes that much sense - it gives a clue - but what

is much more relevant is net underfunding in proportion to the average free cashflow over the last few years. What the business has got to do,” says Neubürger, “is to fund this underfunding at some point in time. If the business has had a negative free cashflow then they have a big pension funding and generally financing problem at hand.”

Measurements relative to market capitalisation crop up with surprising frequency; and not just at the pens of journalists. Speaking to the BBC at the end of 2006, British Airways’ finance chief Keith Williams was quoted as describing an £800m (€1,187m) pound injection into the airline’s pension plan as “very significant … relative to the company’s market capitalisation”.

An employer’s credit rating was identified by the Association of British Insurers in their March 2007 Research Paper 3 ‘Understanding Companies’ pension deficits’ as a key determinant. “The deficit figure,” argue the ABI’s researchers, “has limited significance without proper understanding of the financial strength of the sponsor and its risk-bearing capability.” In viewing the overall level of risk faced by a company, continued the report, “it is necessary to look at deficits relative to key corporate financial statistics rather than in isolation.”

In an endorsement of Neubürger’s view that it is unrealistic to look at a funding status without taking account of available plan assets, the ABI argues that a “value-at-risk analysis based on the assets of both scheme and sponsor may provide a more realistic picture of the overall risk faced by the pensions fund. In this way the strength of the sponsoring company is taken into consideration as well.”

Eric Steedman also urges caution: “A ratio of the overall pension obligation to market capitalisation is part of the picture - we compute similar ratios at Watson Wyatt. Any single ratio will contain information, but I don’t think that you can sum up the entire picture with just one ratio because the global view is much more complicated than that.”

Disclosure has always figured as a key component of pensions accounting standard setting. Indeed, the political compromise to allow smoothing - the amortising of actuarial gains and losses over time - built into FASB Statement No. 87 was predicated on the notion that the pay-back came in the shape of sufficient footnote disclosure to allow users of accounts - albeit rather sophisticated ones - to unpick the accounting and arrive at a non-smoothed position.

“Disclosure”, argues Steedman, “needs to be taken in the context of the whole approach to pensions accounting. I think the choice for standard setters is between trying to go for something that is relatively simple and not particularly customised to the situation of a particular company or plan, but is easily, if superficially, comparable between companies, or to aim at a more tailored approach that is perhaps more refined, but which requires more disclosure so that analysts can assess the reasonableness of the judgements made.

“Personally, I see quite a divide at the moment between companies that are just following the requirements and those that really want to communicate more information about their plans,” Steedman continues. “In the UK, the ASB’s best practice reporting statement on FRS17 has encouraged companies to say more, as has IAS19. But it is still early days, and I wouldn’t say that there is a uniform picture. There is a trend towards greater disclosure, but within that trend some companies are disclosing only a little bit more, and some a lot more.”

Citing his time with Siemens, Neubürger argues that he blazed a trail on transparency: “Clearly, US GAAP led the definition of pension disclosure and IFRS more or less adopted it. Everyone knows though that pension accounting and current disclosure requirements need to be reviewed and very likely be improved. The joint project between FASB and IASB has been defined but will take quite some time to produce, say, an exposure draft, since the matter is so complex.

Suffice it to say that at Siemens we always thrived to meet voluntarily additional information requirements of investors long before accounting standards had been modified, documenting thereby Siemens’ willingness to be very transparent.”

But do these claims withstand scrutiny? As IPE’s comparison of Siemens’ pension plan disclosures with Carrefour - see page24 - demonstrates, disclosure levels across Europe’s pension plan sponsors vary widely, especially in terms of decision-useful information. Even allowing for differences in the GAAP reporting base, IPE’s comparison of pension disclosure data on pages 29 and 30 highlights an urgent need for the IASB to address pensions accounting disclosures.

But what if it were possible for an analyst or investor to take up-to-the-minute accounting data and model a range of cashflow scenarios and their impact on future cashflows. This future is called eXtensible Business Reporting Language (XBRL).

XBRL at its most basic is a mark-up language. Superficially it resembles the hypertext mark-up language used to encode web pages. The theory underpinning XBRL is that businesses will tag financial data for the purposes of financial reporting and regulatory filing. The clear implication is that XBRL will free users of financial statements, such as analysts, from the single time-point estimate constraint of a balance sheet when they assess the funding status of a pension plan, or any other aspect of a business.

Thomas Seidenstein, the IASB’s director of operations, recently told the IASCF Trustees, the IASB oversight body, at their April 2007 meeting, that XBRL will “transform financial reporting”. XBRL, in short, is the passport to nothing less than an accountancy wonderland.

Not so wonderful for business says Neubürger. “From a preparer’s perspective, XBRL is unlikely to improve anything,” he argues. “It seems more a service to investors, albeit one imposed by regulators on preparers who most likely will have to bear the largest part of the cost associated with this service. I do not recall that regulators ever asked preparers for their views on this matter.

“Secondly,” Neubürger says, “the more information you put out, the more you run the risk of misinforming people, because people are also interested in being guided towards what is important. So you always have a trade off between more information and more pertinent information. You will never be able to accommodate all transparency and informational desires.”

Nor do the levels of pension-plan disclosure sometimes demanded by investors and analysts exist in a vacuum. Alongside the demand on CFOs for more and better information - right across the board, and not just in the pensions arena - than currently required under existing accounting rules, CFOs also face calls for ever increasing levels of disclosure about their business’ exposure to litigation risk, or its hedging activities. These are all areas that might be equally critical to a business’ ongoing viability, maybe more so, than the cost of underwriting pensions and employee benefits for a workforce. So it is not just about pensions.

“While I see the utility of providing information by XBRL from the analysts’ and investors’ perspective, I suspect that companies will have a much more difficult time on their roadshows as a result,” Neubürger adds. “For instance, they might meet their investors who have taken XBRL data provided by the company and drawn up completely different financial statements based on the way they as investors look at companies and based on the way their spreadsheet models work.

“And suddenly a CEO or CFO may have to comment on his company’s audited data after it has been put into a format that he is not familiar with. I think it will complicate communication with the financial markets for companies - but we will see. I have always pursued a very transparent reporting policy and I am not backtracking on that, but one needs to define the boundaries. I am not prescribing the limits, but I think it is important to have that debate.”

However XBRL develops, the amount of information that businesses currently disclose about their defined-benefit obligations varies widely. So whereas one company in the IPE’s survey on pension funding might make extensive disclosures, even above and beyond regulatory or GAAP requirements, another might decide to publish more minimal data.Explaining this disparity, Heinz-Joachim Neubürger responds: “There are some companies that do not want to provide information because they simply don’t want to share it. There are other companies who come back - perhaps with 21 pages of footnote disclosure - and make a valid judgement between adequate and too much information. It is a different quality of non-disclosure, if you will, over those companies that say they don’t want to provide that information period.”

In terms of how he sees the debate on informational transparency developing, he predicts that “companies will face growing calls to provide data, and you eventually end up with anyone who shouts loud enough having the right to more and more information,” he explains. “So you run into financial reporting being no-longer defined by what creditors and shareholders require but what the public at large might like to see. I believe that it is therefore appropriate to re-run the debate about the purpose of financial reporting, and who we are reporting to.” If that debate takes place, it is hard to see why a CFO would want to debate pension obligation disclosure issues as a special case in isolation from other broader informational demands on the business.

One of the aims of converging towards a single set of globally acknowledged and endorsed accounting standards such as IFRS is to improve comparability across businesses reporting under the same GAAP - a sort of informational level playing field. IPE’s compilation of pension funding data reveals a less clear picture.

“One thing that does stand out,” says Eric Steedman, “is the influence

of the book reserve system in Germany. Here you are might find that a company’s pension plan assets are significantly lower than the corresponding liability because that business is to a large extent book reserving their pension obligations. I don’t think one can regard that as straightforward underfunding because they have chosen to finance in a different way.

He also warns against jumping to overly hasty conclusions: “It is possible to draw comparisons on the relative strength of funding across these schemes at a very superficial level, but you do have to go deeper than the headline figure to make any more meaningful comparisons. The German book reserve system, which you can also find to a lesser extent in Sweden, doesn’t totally rule out comparability, but it means that the comparison is at best superficial.”

Many but by no means all of the German companies featured in IPE’s tables on pages 29 and 30, particularly those with a US listing on the NYSE, have established a contractual trust arrangement (CTA). Does this development among German businesses hinder comparability? “Yes, definitely,” says Heinz-Joachim Neuburger, “but it is not the intention. It is a deliberate decision by companies whether to fund or not to fund their pension obligation. So Siemens has dedicated pension assets, whereas other German companies might not have. As long as such companies feel they can meet the obligation on an annual cashflow basis, that’s fine. While I wouldn’t want to suggest a specific course, my preference is clearly on funding liabilities with dedicated pension assets.”

Comparisons within a single country are also fraught. Whereas an unfunded German defined-benefit scheme has no direct exposure to equity-market volatility, since it holds no assets, it does impose a significant interest rate risk on the sponsor. The present value of the obligations will rise - and book net worth will fall via IAS19 - if long-term interest rates fall.

“Interestingly, if a German company were to hedge that interest rate risk on its unfunded scheme, by receiving fixed on long dated swaps and paying LIBOR,” says Crispin Southgate, “investors might have some chance of seeing that they had done so through disclosures in the accounts. But if the trustees of a funded scheme did the same thing, thus reducing the sponsor’s and the scheme’s exposure to declining interest rates, current accounting standards would not require that fact to be disclosed. If schemes use swaps in LDI strategies to reduce interest rate and inflation exposures - what new balanced approaches call the unrewarded risks - investors would not know unless the sponsor chose to tell them.”

One number that Neubürger would like users to see alongside an IAS19 deficit figure is the national pension regulator’s assessment of the individual group pension scheme that come together in the IAS 19 calculation. “While one needs to look at the consolidated number, one also needs to know what the underfunding and then the cashflow requirement is from the national regulatory perspective.

“I believe that audit committees, for instance, should want to see these different numbers. If you backtrack to the national regulators who supervise the national pension obligations in each of the countries where a consolidated group operates, they might come up with a completely different number to the one shown on the balance sheet, because they measure the liabilities and the matching asset base for their individual national purposes in different ways.”

His comments in fact highlight the tension that springs up when the process of consolidating a company’s financial information takes an item such as a pension obligation. A consolidated pension obligation might in fact comprise many individual liabilities spread across the various jurisdictions in which the group operates. The process of group consolidation, he argues, infers a false notion of comparability: “They are clearly not because local laws governing group companies define the ultimate obligation,” he explains.

“Let’s assume for consolidation purposes that you use an interest rate of 5%”, Neubürger continues. “Although that is aggregated based on different national discount rates across different currencies, if the national regulator prescribes perhaps as a matter of conservatism a much lower discount rate, you obviously get a different liability in the case of an underfunding.

“And if you have a politically motivated discount rate, such as in the US where the government prescribes a much higher rate for, as an example, PBGC-premium calculation purposes than current market rates, you obviously have a different funding situation.” At least on a short-term basis, he adds, “until reality catches up”.

So were pension fund trustees to take up Neuburger’s challenge of adding together the individual regulatory measures for each of the consolidated jurisdictions, they would probably discover that their total does not equal the consolidated entity’s IAS19 net funding position.

The consensus across the consultants and the preparer is clear: there are no shortcuts to analysing a defined-benefit pension plan’s funding status.

Siemens and Carrefour: a tale of two disclosures

Siemens’ pension disclosures in its September 2006 financial report run to some 10 full pages in the notes alone. They include a further three pages on post-retirement healthcare benefits. The accounting basis is US GAAP and Siemens continues to amortise actuarial gains and losses in the September 2006 accounts. So the raw funded status of the pension plans - on a basis very close to that of IAS19 - is a number that must be found in the notes: it isn’t on the balance sheet.

In accordance with FASB Statement No. 87, Siemens must explain the differences between: (1) the smoothed number recognised so far on the balance sheet; (2) the more volatile funded status of projected benefits and (3) the minimum liability for accrued benefits.

“If that sounds like a dog’s dinner of a confusing mess,” says Crispin Southgate, “then it is US GAAP that is mainly to blame.” FASB Statement No. 158, a partial overhaul of SFAS87 to eliminate smoothing on the balance sheet, has changed this and would apply to Siemens’ future disclosures, but for the fact that it intends to adopt IFRS for its next round of annual reporting.

“Albeit hobbled by the old US GAAP framework, Siemens’ pension disclosures include reasonable attempts to explain what is going on, throughout the annual report,” Mr Southgate. The word ‘pension’, he observes, occurs 230 times in the document.

“Analysts looking at Siemens can find information on the asset allocation of the funds, a sensitivity analysis for key assumptions, and cashflow information in a table of recent and projected payments from the plans.

“By contrast, those looking at accounts will find very little to work with in a page or so of disclosures - the word ‘pension’ appears seven times in their annual report - and little attempt to explain.” His verdict? “Carrefour’s is a compliance approach that appears uninterested in communicating information - and that’s partly because IAS19 lets them do so.”

IPE invited Carrefour to explain its disclosure policy. In particular, we were curious to know why chief financial officer believes “it is necessary to devote a substantial part of the information disclosed to defining the terms defined-benefit and defined-contribution”. Carrefour’s response provides damning evidence to support Southgate’s charge of a compliance-based approach - it also supports his criticism of IAS 19’s disclosure regime.

A Carrefour spokeswoman told IPE that “[t]he annual report is compliant with IFRS legislation. It has been auditing and approved by auditors.”

She added that: “The group is not exposed in markets such as Germany, the United States or the United Kingdom, which explains the fact that pensions are a very small account.” But if you ever wanted to know the difference between a defined-benefit and a defined-contribution plan, Carrefour’s annual report is a great place to start.