The pension threat to credit quality

Pension deficits

In the past 12 months corporate analysts, lenders and investors have had much to contend with, operating in an environment where complex and often highly subjective accounting conventions have allowed for, at best, financial obfuscation and, at worst, outright fraud.

The experience, along with the accompanying downgrades and defaults, has heightened the degree of sensitivity to any area of financial reporting where the permissible discretion within the accounting regulations allows for potential massaging of the numbers.

The latest round of accounting quality headaches goes far beyond the realm of emerging TMT businesses, where we saw too many transactions fail to accurately represent the value of ‘clicks’ businesses using ‘bricks’ accounting methodologies. Hot topics such as the appropriate way to account for employee compensation via stock options have underlined that these accounting issues can affect even the most basic elements of operating fundamentals across every industry.

But while the subject of accounting for employee stock options gained considerable prominence earlier in the year and is under review by the Financial Accounting Standards Board (FASB), another area of accounting that is more widely used and that currently harbours potentially more negative shocks to the credit community has, until recently, been flying under the radar screen. And that is the subject of accounting for pension liabilities.

Whether it allows for misrepresentation or merely misinterpretation of credit quality, balance sheet strength, borrowing needs or intermediate liquidity fundamentals, pension accounting is perhaps belatedly being recognised as the next major threat to holders of securities. And the risk seems most pronounced for some of the mature cyclical manufacturers based in the US, where private defined benefit plans have been a union standard over many cycles.

General Motors, Ford, and DaimlerChrysler (DCX) are prime examples. The auto manufacturers, which are among the largest issuers in the dollar- and euro-denominated debt markets, are heavily exposed to these liability streams, and the erosion of pension asset values and unfavourable balance between retirees and current workers as the companies continue to downsize operations are starting to make for both a major financial threat and a competitive disadvantage.

As we drill down past the more spectacular pension numbers coming out of the auto companies, we are also seeing sharply rising risks for most mature, heavily unionised industries that offer generous retiree benefits. The capital goods, machinery and transportation sectors are most at risk.

The third-quarter earnings reporting season has started to bring out additional colour on the nature of the balance sheet effects and potential funding demands of an expanding unfunded pension liability for many benchmark corporate names. However, companies have generally remained very cagey on the topic with respect to the direction of the discretionary plan assumptions and what they will mean for reported income statement and balance sheet numbers.

We have seen early estimates out of many companies on the year-end charge to equity that will be required under the accounting rules, and for some we have seen disclosure of interim shortfalls of plan assets versus the projected benefit obligation (PBO), but at this point the information is still inadequate and we expect many ugly surprises in the next round of yearly 10-K filings.

The primary risk associated with the hits to book equity will be the need for some issuers to seek waivers on bank line covenants. In addition, bank lenders themselves may start to adjust their own risk models to account for the growing pension threat and seek additional non-public disclosure from borrowers on pension plan status. This could, in turn, lead to revisions of credit terms and even credit availability as bank lines mature.

Pensions by the book

For US companies, the accounting methodology surrounding pensions is necessarily complex. It is governed by Statement of Financial Accounting Standard No. 87 (SFAS 87) Employer’s Accounting for Pensions, which was issued by the FASB in December 1985. The rules require that plan assets be segregated in a pension fund – ie, kept off-balance sheet – and that they be applied to no purpose other than to pay plan benefits. SFAS 87 provides two possible methods for the determination of the value of plan assets.

Using a ‘fair value’ approach, all pension assets must be carried at an observed market value and illiquid assets that are not regularly traded, such as real estate or venture capital investments, must be valued by current appraisals. The second ‘market-related asset value’ amortises actual plan gains and losses over specified periods. What is troubling in the disclosure guidelines is that companies are not required to disclose what method of valuation is being employed.

However, it is not so much the valuation of assets which gives rise to the potential lack of clarity regarding the status of the pension plan. The limitations inherent in SFAS 87 arise from the fact that the formula used to calculate pension benefits relies upon many factors that are unknown and for which assumptions must be made. These include, but are not restricted to, the level of future compensation increases, years of service, the age of employees at retirement and, most importantly, the future investment returns that will be generated on the assets in the plan.

It is to be expected that these assumptions will at times diverge from current levels given the cyclicality of markets, but because of their long-term nature, relatively minor changes in any of these assumptions can lead to huge revisions in the PBO. Therefore, there is legitimate concern about the allowable subjectivity in arriving at the assumptions used and the impact they can have on the perception of actual obligations.

This concern is being exacerbated by the level of exposure that most pension funds have to the equity market and the sharp drop in equity values that has taken place over the last three years. In a natural reaction to the 1990s bull market in equities, the average percentage of pension fund balances invested in the stock market rose throughout the decade.

In 1990, 38% of aggregate pension fund assets were invested in the equity market (based on all companies in the Standard & Poor’s Compustat database). This level was 50% by the end of 1999. The bias toward equity investment created a sizeable exposure to the bear market that began in 2000 and therefore it is not surprising that the overall value of assets in pension plans fell by 11% during 2001.

The bull run in equities contributed to a degree of comfort surrounding the issue of funding pension obligations in two ways. First, the market’s rise greatly increased the value of plan assets limiting the amount of cash contributions that companies were required to make (which had a secondary effect of further benefiting earnings by reducing expenses). Second, it prompted an increase in the expected rate of return, the crucial assumption about future investment returns that dictates whether or not a plan’s current assets are deemed to be adequate to meet its future obligations.

The compound annual growth rate of aggregate reported plan balances between 1997 and 2001 was 9.09%. Against this background, the 8.66% average expected rate of return used by plans in 2001 does not seem egregious.

It raises concerns, however, as the expected rate is virtually unchanged from that used in 2000 despite the major reversal in the equity market’s fortunes. And while the estimate is a long-term one that should exhibit stability year after year, one can reasonably expect that current assumptions should reflect the severity of the three-year bear market in equities, the prevailing exceptionally low yield environment and the generally more limited expectations of post-bubble investment returns.

If investors can cast a critical eye on the average level of expected return, then any company that is currently using an even higher rate deservedly comes under the spotlight. The case for this is even stronger if the company’s pension plan is currently underfunded. The bad news for the credit markets is that not only are there many companies (including a collection of industry bellwethers) where this is the case, the numbers involved are quite spectacular.

Little wonder that the most recent round of earnings calls were peppered with questions about the need to make cash contributions to address funding shortfalls and requests for details on exactly what calculations are producing the pension obligations that are being reported.

Adjusting balance sheets to assess default risk

The issue has given rise to a increase in pension-adjusted analysis as the market seeks ways to cut to the chase and get to grips with the matter on a company-specific basis. As companies are only required to detail their pension plan assumptions in their annual 10-K filing, bondholders and lenders are scrambling to make their own judgments about the sustainability of current plan assumptions and the hit to the balance sheet, with its potential flow on effect, that could result from any ‘adjustments’.

Risk scoring models such as CreditSights’ BondScore provide an excellent means of scenario-testing the impact that would result if these legally enforceable obligations were to be accounted for in the same manner as other debt.

To return to our example of the auto companies, during its third-quarter earnings call, General Motors (GM) provided an array of scenarios on its pension funding numbers (highlighting once again why the market is coming to regard company projections in this area with such caution) that put the plan’s shortfall at a minimum of $23 billion.

While the BondScore model as at November 15, 2002 rated GM as a BBB2 credit with a credit risk estimate (which measures the probability of default in the next year) of 0.2817%, if the $23 billion shortfall were to be included in the total debt numbers, the rating would drop to BBB3 with a substantially higher credit risk estimate of 0.7331%.

As we highlighted earlier this is not a credit impairment that is borne by GM alone given that it arises from the auto industry’s cyclicality and its ageing and highly unionised workforce. The BondScore model currently rates Ford in the same universe as GM, as a BBB2 credit with a credit risk estimate of 0.1442%. And while Ford’s unfunded pension liability is approximately one quarter that of GM’s, it has a consistently negative effect on the company’s credit quality. If the assumed $8 billion of Ford’s pension shortfall is added to the company’s total debt the BondScore credit risk estimate would rise to 0.3234%, which would also drop Ford into the realm of a BBB3 rating.

Currently, DCX enjoys the strongest BondScore rating of the big three auto manufacturers: BBB1 with a credit risk estimate of 0.0986%. Adding the company’s estimated $5.5 billion pension funding shortfall drops the ranking to BBB2 and lifts the risk estimate to 0.1634%.

Such analysis is merely the first step that can be undertaken to explore the potential for pension obligations to exert dramatic changes on the perception of comparative credit quality. This scenario-testing has taken company-provided information on pension obligations and current shortfalls at face value. And while it is not possible to completely recalculate a company’s PBO given the information that it is required to disclose regarding its assumptions, it is not difficult to identify where those assumptions are overly optimistic compared with industry averages.

Higher than average expected rates of return, discount rates well above prevailing market yield levels and low forecasts for compensation growth relative to industry norms are all signs that a company may be pushing the envelope with regard to its calculation of the PBO and thereby presenting a somewhat rosier picture on the subject that its competitors. The use of risk scoring models allows such factors to be handicapped by adjusting the assumed level of a plan’s funded shortfall to that which would bring it into line with industry-wide assumptions.

While the process is by no means an exact science, it does mean that investors should not be caught unawares when companies that have clung too long to optimistic assumptions are finally forced to make adjustments and adopt inputs that better reflect prevailing conditions.

This sensitivity testing by no means predicts likely rating actions as the agencies have provided little clarity on the subject of their treatment of pension and other contingent liabilities. But it does enable investors not only to target those companies where the dollar amount of the funded pension shortfall is significant or the assumptions that are driving calculations are out of alignment with competitors but also to measure the credit impact so that an accurate comparative assessment of credit quality may be made.

Considering pension obligations to be less onerous than any other debt commitment a company has incurred is an indulgent attitude that was fostered by a bull equity market that did much to desensitise us to this ongoing, enforceable and substantial employee expense. It is not an approach suited to the current environment where judging the amount of appropriate compensation for a given level in a company’s capital structure vis-à-vis other lenders is a much more complex and nebulous business.

Those who accept company data on pension plans at face value without doing their own forensic research and sensitivity-testing may easily find themselves having incurred a risk/reward profile very different to that envisaged, without the reporting company ever having done any less than it is required to do under current accounting rules.

The subject of pensions accounting poses as many risks to prevailing perceptions of company credit quality as any that have captured the headlines so far in 2002 (off-balance sheet financing, overreliance on short-term funding, stock options accounting). Yet the upside for a market that is still reeling from the fallout of a year of unprecedented erosion of value for bondholders and lenders alike is that this should not be an issue that catches investors off guard.

A substantial amount of information about pension plans exists, despite the fact that it is typically buried in the fine print. Tools for using that information to more accurately assess credit strength are available for those willing to use them. And the degree of attention that the subject is receiving is growing daily. Ultimately, this will give companies the choice of voluntarily providing a greater level of transparency in the broad area of reporting pension plan calculations or risk being harshly assessed by investors and lenders who, in the wake of the latest year, are inclined to have suspicious minds.

Louise Purtle is US corporate strategist at CreditSights, a US independent credit research provider