Questioning the basic assumptions. (Pensions).

For years, many corporations have been using a lofty presumed return rate for their pension investments. Such returns are highly unrealistic these days, and the fallout from years of unmet targets could be catastrophic.

When Fortune magazine editor Carol Loomis sat down for an interview with Warren Buffett late last year, the conversation could have led in many directions: the economy, the stock market or the tenets of value investing. Instead, it mostly led in just one: U.S. pension fund accounting and the current pension return assumptions underlying it.

Buffett espoused in that interview (and has continued to espouse since) that pension accounting is likely to blossom into yet another scandal besmirching America's corporate boardrooms -- and is a problem potentially far larger than the accounting shenanigans caused by companies such as Enron Corp. or Adeiphia Communications.

"I invite you to ask the CFO of a company having a large defined-benefit pension fund what adjustment would need to be made to the company's earnings if its pension assumptions were lowered to 6.5 percent," Buffett stated. "And then, if you want to be mean, ask what the company's assumptions were back in 1973 when both stocks and bonds had far higher prospective returns than they do now."

According to Ryan Labs Inc., a New York-based quantitative asset management firm, the average major U.S. corporation maintains a longterm pension fund return assumption of approximately 9.25 percent a year. Many corporations set this return assumption using a five-year moving average of recent annual returns, but the Financial Accounting Standards Board makes no explicit requirement to reveal what methodology is used -- asking only that the approach be consistent over time.

Under the rules of FASB 87, which became effective in 1986, if pension returns are realized above the assumed rate, and pension liabilities remain static, companies may accrue the excess returns directly to earnings over the average life of a pension plan. If returns fall short of the assumed return, the loss does not have to be realized right away, but again, can be accrued over time.

Either way, the "smoothed" accrual typically shows up either within the Gross Income line of the balance sheet or as a part of Sales and General Administration expense. Many Wall Street analysts bemoan this practice as a potential obfuscation of true operating earnings, but the accounting community has long since signed off on either treatment.

But Buffett believes a 9.25 percent annual return assumption in today's world is just dead wrong, and to not admit that keeps corporate earnings artificially higher for a longer period of time -- something potentially deceiving and dangerous. "Companies with return on asset assumptions above 6.5 percent are not facing reality, and anyone choosing not to lower assumptions -- CEOs, auditors, actuaries all -- is risking litigation for misleading investors," Buffett argues. "And directors who don't question [this] optimism simply won't be doing their job."

By comparison, far more conservative assumptions existed in past years. In 1982, for example, corporations were using an average 6.5 percent future return assumption -- a number that represented just half of the then-current bond yield and was barely higher than then-current equity dividend yields. The current 9.25 percent assumption, on the other hand, represents almost twice the recent yield on T-bonds and eight times the dividend yield on stocks.

A Huge Issue

But how can accepted financial accounting standards cause such potential problems? And nine months after Buffet's comments, have we started to see more signs of this problem manifesting itself? At least one Baltimore-based actuary, Thomas Lowman of Bolton Offutt Donovan Inc., struggles with the issue of pension return assumptions every day. "How do I approach my clients and tell them that they should honestly be lowering their return assumptions on pension assets?" he asks. "If I push too hard on this topic, all I will succeed in doing is to lose their business."

Another expert on the subject, Robert Arnott, chief executive officer of Pasadena-based investment manager First Quadrant LP, refers to the current situation as "a potential trillion-dollar disaster." Explains Arnott: "If actual pension returns average 2-3 percent below current assumptions -- let's say near 6-7 percent annually -- and then were multiplied by the aggregate assets for the largest 1,000 public and private plans of $4.8 trillion, this difference would cost America $100-$150 billion per year in future 'revenue.' Multiply this annual [shortfall] by the average 15-year life of pension plans, and bingo, a trillion dollars falls out."

Corporate defined-benefit pension plans represent approximately a third of total pension fund assets, so continuing with Arnott's calculations, corporate America would face $33 billion to $50 billion in lower revenues per year over time. Smoothing of past gains into current losses would mitigate the impact, but incrementally, it would add up to a huge drag over time.

In a recent academic paper, Arnott writes: "In 2000 and 2001, earnings from the S&P 1500 (the S&P "Super-composite") totaled $524 billion and $252 billion, respectively, with P/E ratios of 25 and 46 times earnings. If pension return assumptions had been dropped by 3 percent, then reported earnings would have been $474 billion and $202 billion, respectively, and the yearend P/E ratios would have been 28 and 58 times earnings. Today's P/E ratio would rise from 60 to 80 times current depressed earnings. $50 billion is a significant event."

Pat McConnell, a CPA at Bear Stearns & Co. and another analyst who has studied the pension accounting issue, largely confirms Arnott's numbers.

In a detailed November 2001 report, she estimated that 2000 average operating profits for the S&P 500 would have been "2.9 percent lower than reported operating income" if pension fund income accruals were omitted. "Pension income" accruals actually represented over 15 percent of reported corporate earnings for 17 major corporate entities in 2000 (see Table 1), she adds. Heavily represented on her listing were older industrial and aerospace companies with highly unionized work forces.

Going forward, McConnell has predicted that should equity markets effectively remain flat through the balance of 2002, approximately 10 percent of S&P 500 companies will have their earnings cut by 5 percent or more due to pension fund considerations.

Pension Holidays Gone Puff

Courtesy of the 2000-02 downdraft in the equity markets, and corporate investment policies that failed to realize many of the pension fund "paper" gains of prior years, positive pension income amortizations have largely evaporated.

Consider Procter & Gamble Co. In 1992-99, annual excess gains from P&G's pension fund investments trickled to its bottom line, and its plan was $1.3 billion overfunded by the end of 1999. But that entire overfunded amount was wiped out by the 2000 equity decline, eliminating amortizations of past gains that would have nicely padded future P&G earnings.

The situation at General Motors Corp. is perhaps even worse. A close review of GM's annual reports reveals that GM earned $5.5 billion in revenue in 2000, but in the 2000-01 fiscal years, GM's $70 billion pension fund declined in value by 30 percent. This $21 billion paper loss did not fall immediately to earnings. Indeed, GM only showed $346 million in pension costs for 2000, courtesy of FAS 87 smoothing techniques.

But eventually, of course, actual market losses -- if not quickly recouped -- will need to be realized. In the last half of 2001, GM Chief Financial Officer John Devine admitted that the company's prior $1.7 billion pension paper surplus was finally gone, replaced by a $9.1 billion dollar deficit. Devine further revealed that GM would have to contribute about $2 billion to its pension plan by 2003 to avoid paying additional Pension Benefit Guaranty Corporation (PBGC) penalty premiums.

This latter entity, the PBGC, is a governmental organization that acts as the receiver and ultimate guarantor of failed corporations' pension plans. Each year, it charges an annual premium to corporations with defined benefit plans. It may also impose penalty premiums on plans deemed actuarially under-funded. Since such penalty premiums can never be recovered, most corporations try to avoid this latter path at all costs.

But presuming the markets won't come roaring back, GM's problem is not going to go away. "What Devine didn't tell people was that pension funding issues will now result in an ongoing annual charge, instead of being just a one-year exceptional item," explains one pension accounting expert. "The 2000-2001 $21 billion loss will now need to be amortized over the average life of GM's pension plan of approximately 15 years. This translates into a $1.4 billion future drag to reported earnings before GM even opens its showroom doors each and every year."

Clearly cognizant of this situation, GM issued $3 billion in convertible securities earlier this year, with much of this money earmarked to go into its pension plan. But by taking such an approach, GM has also now turned one liability on its balance sheet into two. "This is obviously not without a certain amount of danger," says Ron Ryan, President of Ryan Labs Inc. "In the end, if pension asset performance does not recover very quickly, various financially undesirable accounting measurements will occur. This includes some combination of reduced earnings, higher required pension fund contributions, higher pension insurance premiums and lower credit rating valuations."

Actuarial firm Milliman USA has conducted a study showing that the 20 largest U.S. defined-benefit pension plans reported "revenues" of $6.89 billion in 2000 despite suffering actual fair-market losses of $19.2 billion -- a $26 billion difference fully attributable to FAS 87 accrual accounting.

Slow Fuse with Big Ramifications

But does the average equity investor realize that such a time bomb of negative pension fund accruals is already lurking? "At a smoothed 9.25 percent return assumption on assets, we have yet to see most corporations take pension-related earnings hits," says Lowman, the Baltimore actuary.

One of Lowman's clients is, in fact, the Pension Benefits Guaranty Corp. In his duties for the PBGC, he regularly delves behind reported corporate earnings to assess the impact of lower return assumptions. "There is no doubt in my mind that a growing drag to corporate earnings looms out there, as do increasing claims on the PBGC," says Lowman.

But when will this all hit? "I'm not sure," Lowman adds. "It's all very company-specific, but I think you will be hearing a great deal more about this issue as we move toward 2004-2005." The PBGC recently announced that unfunded pension liabilities for private companies surged to $111 billion at the end of 2001 from $26 billion in 2000.

All of these issues are concerning enough by themselves. Unfortunately, they only cover the asset side of the pension fund equation. Adding in the liability side of the equation only worsens the situation. The huge Baby Boom generation is moving toward retirement -- with more retiring people, hiking corporate costs -- but the liability calculation for these future pensioners has nothing to do with equity market performance.

Instead, under ERISA rules, corporations must calculate their pension liabilities by taking the expected number of employees that will qualify for a defined benefit plan and multiplying this number by the years that each is expected to be alive in retirement, and then, in turn, multiplying by the dollar benefit each will be due per year. This actuarial dollar result is then discounted back to a present value over the average life of the plan. ERISA mandates that a discount rate be used approximating that of highgrade debt similar to plan liabilities.

And therein lies another problem. The ERISA wording is vague enough that the actual discount rate used by corporations tends to vary. Some corporations use discount rates as low as 5 percent; others have set their discount rates as high as 7.5 percent.

Moreover, given the simple mathematics of present discounting, the absolute level of calculated pension liabilities rises when interest rates fall. And interest rates have, of course, been trending lower for a number of years. In some respects, this has proven a double whammy to many corporate pension managers, who tend to be more focused on the asset side of the equation than on the liability side.

Ryan Labs has constructed Table 2 (page 37) showing that because of such liability considerations -- and as surprising at it might at first sound -- the overall tactic of leveraged equity investing by pension funds has not been as successful in the past as most would expect. Even with the preponderance of pension plans investing heavily in equities through the 1990s, pension managers have barely kept pace with their liabilities.

A model portfolio begun in January 1989 and allocated 60 percent to equities, 30 percent to bonds, 5 percent to international equities and 5 percent to cash stood on a compounded basis at 54 basis points below calculated liabilities (the final Asset -- Liabilities column in Table 2 summed across years) through May 2002.

A few financial managers have realized that equities simply may not give them the returns they want in a risk-adjusted manner. U.K. retail giant Boots took the unprecedented step in 2001 of liquidating its entire 2.3 billion-pound equity portfolio and replacing it with a static portfolio of long-term zero-coupon government bonds and inflation-protected securities.

Several noted financial economists, including the late Fisher Black, have shared Boots' view that more pension plans need to replace equity assets with fixed-income assets to reduce the potential risks to corporations, beneficiaries and ultimately the PBGC and U.S. taypayer.

The Boots decision may have also found its roots in a 1997 white paper presented to the Institute of Actuaries by Jon Exley, an actuarial consultant for William Mercer & Co. In that paper, Exley refers to the "folly of relying on equity returns as a substitute for [proper pension fund] risk management."

Modern Portfolio Theory: Culprit & Potential Savior

But Black and Exley's opinions notwithstanding, modern portfolio theory clearly had a hand in creating current pension allocation problems and pressures. In the 1980s, most people presumed that equities always advance more than bonds over the long term. And since pension liabilities are by nature long-term as well, an obvious question arose: Why not allocate a large portion of pension assets into the equity market?

The adoption of FAS 87 then gave the pension industry a means by which equity returns could be smoothed over time -- a seemingly win-win situation, at least until the equity markets nose-dived.

Now, modern portfolio theory is being put through a severe test, but it also offers some even newer potential solutions. Most plan sponsors are currently casting about for ways to boost ongoing returns.

While some companies may try to follow Boots' all-fixed income example, the supply of long-dated, highly rated fixed-income paper is not large, and the supply of inflation-protected notes is even smaller. Partly because of this, many sponsors are gravitating towards increased allocations to other alternative forms of investing -- including hedge funds. Modern portfolio theory currently lauds the added "alpha" (added return for a given level of risk) that such "alternative" investing can create.

But overall, according to Ryan, "There are no easy answers to this situation. It is potentially a decade-long problem." Ryan believes that "corporations need to go back and reexamine the economic behavior of pension liabilities -- a behavior few people have ever taken the time to truly model correctly."

Ryan, like Buffett, clearly understands this issue. But does the average Wall Street analyst, CFO, corporate treasurer or individual investor understand the scope of America's defined benefit pension problems? The answer to this latter query likely remains: no -- at least not yet.

Table 1

Pension Income as a percentage of 2000 Operating Income


U.S. Steel              440%
McDermott Intl          267
Allegheny Technologies   49
NCR Corp                 46
Northrop Grumman         42
NICOR Inc                29
PACTIV Corp              28
Unisys Corp              25
Lucent Technologies      24
Peoples Energy Corp      24
Westvaco Corp            20
Lockheed Martin Corp     19
Consolidated Edison      16
Verizon Communications   16
Unumprovident Corp       16
Potlatch Corp            15
Tektronix Inc            15

Source: Bear, Stearns & Co.
Table 2

Asset/Liability Watch

Index                 Weight    '89     '90    '91     '92     '93

RL CASH                   5%   9.34    8.73   7.42    4.12    3.51
LB AGGREGATE             30%  14.53    8.96  16.00    7.40    9.75
S&P 500                  60%  31.68   -3.15  30.45    7.64   10.07
MSCI EAFE                 5%  10.80  -23.32  12.48  -11.85   32.95

ASSETS                  100%  24.31    0.16  24.13    6.44   10.79
RL LIABILITIES          100%  25.40    3.23  19.26    7.87   22.46


Assets - Liabilities          -1.09   -3.07   4.87   -1.43  -11.67

Index                    '94     '95    '96    '97    '98     '99

RL CASH                 3.94    7.11   5.59   5.72   5.48    4.24
LB AGGREGATE           -2.92   18.47   3.63   9.65   8.69   -0.82
S&P 500                 1.29   37.57  22.93  33.34  28.55   21.03
MSCI EAFE               8.06   11.56   6.37   2.08  20.24   27.32

ASSETS                  0.55   28.67  15.21  22.98  21.37   13.69
RL LIABILITIES        -12.60   41.16  -3.70  19.63  16.23  -12.70


Assets - Liabilities   13.15  -12.49  18.91   3.35   5.14   26.39

Index                    '00     '01    '02

RL CASH                 6.49    4.97   0.71
LB AGGREGATE           11.63    8.44   2.90
S&P 500                -9.09  -11.86  -6.50
MSCI EAFE             -13.87  -21.11   2.67

ASSETS                 -2.50   -5.40  -2.84
RL LIABILITIES         25.96    3.08   1.67

                                              Total
Assets - Liabilities  -28.46   -8.48  -4.51   -0.54

Source: Ryan Labs, Inc., as updated through May 2002

Barclay T. Leib is the President of Sand Spring Advisors LLC, a New Jersey-based financial consulting and alternative asset management company. He writes regularly on asset management and may be reached at [e-mail].