Remarks before the National Association for Business Economics
by Bradley D. Belt
Executive Director, Pension Benefit Guaranty Corporation
"Through the Looking Glass: Adventures in Pension Land"
Good afternoon, and thank you for the opportunity to address this distinguished audience, which I am confident will appreciate the importance of the issues we are discussing today.
With more than $2 trillion in obligations, defined benefit pension plans represent one of the largest off-balance-sheet liabilities underwritten by corporate America. One out of every five private-sector employees still earns a pension benefit, and more than two-thirds of the companies in the S&P 500 still sponsor a defined benefit plan. In many cases, pension plans have grown so large that they dwarf the companies that sponsor them. Consider: With the pension assets of one of the domestic automakers, you could buy all three companies.
And yet, despite their massive size and economic importance, defined benefit pension plans have been largely hidden from view behind a nearly impenetrable thicket of often incomprehensible accounting standards, funding rules, and actuarial conventions. Indeed, when we gaze upon the pension landscape, we are struck with the peculiar sensation that much of what we were taught-about economics, about corporate finance, about accounting-no longer applies. With apologies to Lewis Carroll, we feel as though we have followed Alice down the White Rabbit's hole into a looking-glass world that is disconnected from economic reality-one where you can add risk without adding cost, where a dollar of stocks is worth more than a dollar of bonds, and where the value of assets five years ago affects your pension contributions today.
Let us begin with one of pension-land's more remarkable propositions-that employee compensation, which one generally considers to be a cost center for corporations, can be a rich source of profits. As you know, private-sector plan sponsors are required by law to set aside a certain level of assets to meet their pension obligations. But under the pension accounting rules-FAS 87-companies can book to income not the actual return on those pension assets but the expected return. Holding everything else constant, a company with $50 billion in pension assets and a ten percent assumed rate of return can claim a "pension profit" of $5 billion, even if in reality the assets decline by ten percent that year. For some companies in recent years, the pension plan has been a primary source of profits-phantom profits in the view of many, but with the power to boost share prices and executive compensation just the same.
In a series of well-researched reports, David Zion of Credit Suisse has called this the "magic" of pension accounting. Perhaps that is too generous a spin-pension accounting is beyond magic. It is alchemy. Today's financial engineers have swapped the medieval alchemist's mortar and pestle for the modern tools of smoothing and expected returns, but the end result is the same-lead miraculously transformed into gold. With pension accounting, we move beyond managed earnings to manufactured earnings, to a world where two minus two can equal five.
Imagine if this approach applied to other areas of corporate finance. That ten percent drop in sales you experienced becomes a ten percent gain, because that's what you "assumed" would happen based on past experience. That fifty percent increase in the price of raw materials never happened, because it wasn't in your projections.
You probably wouldn't buy shares in any company that kept its books this way, and I suspect that the SEC would have an issue as well. But, when it comes to pensions, we look the other way. You can be a large, insolvent annuity provider attached to a small widget maker and report profits based on phantom income from the pension plan. Caveat emptor.
I would like to make one other observation about the practice of booking assumed returns-there is no free lunch. The risk-free rate of return is currently hovering just south of five percent. Yet companies routinely book income based on the assumption that their pension assets will earn 8, 8.5 or even 9 percent.
Of course, these higher expected returns are only possible by taking on extra risk, yet that risk apparently carries no immediate cost. The extra three hundred basis points above the risk-free rate just materializes with no acknowledgement that the only way to guarantee such returns is to buy portfolio insurance whose cost, when subtracted from the assumed return, would lower your effective return to the risk-free rate.
When pension-land was dominated by actuaries and benefits professionals, it was relatively easy to get away with loose thinking about how to value pension assets and liabilities. Consider two promises to pay: one collateralized with stocks and the other collateralized with bonds. The actuarial view says the "cheaper" promise is the one backed by stocks, because stocks have higher expected returns. The financial economics view, which has been making steady inroads in the pension space, considers that approach nonsensical-a dollar of stocks is worth exactly the same as a dollar of bonds. The economic cost of a benefit promise is the same whether it is backed up by stocks, bonds, or no assets at all.
The actuarial view, which also informs FAS 87, encourages companies to meet their obligations by borrowing from employees, investing in riskier assets, and hoping the assets beat the borrowing rate. And it holds a surface appeal: If you borrow at a six percent corporate bond rate and earn eight percent on a basket of riskier assets, your cash contributions to the pension plan can be reduced accordingly. Of course, what pension plans experienced in the early part of this decade was the opposite. Companies borrowed at six percent, earned negative eight percent, and reacted with shock when the margin call of higher required contributions came due.
So why do so many companies operate their pension plans the same way? Clearly, much of it is driven by the pension accounting rules. The ability to book nine percent on $10 billion of assets makes it powerfully tempting to invest in assets risky enough to "justify" nine percent. But I also think a herd mentality predominates in pension-land. Everyone else is doing it exactly the same way and, besides, your performance as a pension manager is judged not by how well you defease the plan's liabilities, but by whether you beat your peers on the asset side.
That explains why the asset allocations of pension plans are remarkably similar despite vast differences in the liability profile. For example, some people think that a "young" pension plan with five active workers for every retiree can afford to invest along the traditional lines of sixty percent equity, thirty-five percent fixed income, and five percent cash. But what we observe in pension-land is that "mature" plans with five retirees for every active employee invest exactly the same way. The cash flows and inflation sensitivity of these two pension plans could not be more different. What is the same is the desire to chase returns in heretofore almost blissful ignorance of the liabilities.
Let me be clear: the pension alchemy we see practiced every day is permitted, even encouraged, by financial accounting standards, ERISA funding rules, and actuarial conventions. But it is disconnected from the economic reality in which you must operate, and it obfuscates what every worker, retiree, investor, and creditor has a right to see.
Those who take a different view will argue that we need not worry about these stakeholders of the pension system. In this rose-colored view of the world, pension plans are seen only as long-term obligations, with pension gains and losses amortized over time. Corporate income statements will catch up with economic reality eventually. Yes, that may happen. At some point pension reporting and economic reality might even agree with each other, but only by coincidence, and never for long.
And thus we come to another figment of imagination in pension-land-smoothing. "Smoothing" is a seductive marketing word. It conveys the sense that we are sparing investors from the rude jolt they would receive if pension losses were reported at full value and saving companies from the terrible burden of repairing pension deficits as quickly as they were created.
In the accounting context, smoothing allows companies to show pension losses to investors in small slivers over time rather than all at once. This helps make a company's reported earnings look smoother as well, which is to say, more divorced from economic reality. But if we have learned anything from recent economic history, it is that attempting to manage reported earnings leads to trouble. Going back a few years further, would we have avoided the need for an S&L bailout if we had allowed thrifts to smooth interest-rate spikes over a several year period? Would the economic reality of their asset and liability mismatch have been any different? In the pension context, it should be a wake-up call when the deputy chief accountant of the SEC derides smoothing for its potential to render financial statements "meaningless."
But as problematic as smoothing may be in the pension accounting context, in some ways it is even worse in the pension funding context.
Under the pension funding rules contained in ERISA and the Internal Revenue Code, a company can skip needed contributions to its pension plan on the grounds that "smoothed" assets and liabilities make the plan look well-funded. When followed by a corporate bankruptcy, this policy of ignoring economic reality and failing to make needed contributions can lead to devastating losses of retirement income for long-serving employees.
On the asset side, the funding rules allow companies to use values smoothed over five years. The only constraint is that the market value of the assets cannot be more than twenty percent different than the so-called "actuarial" value of assets. In practice, this means a pension plan with $1.2 billion in liabilities and $1.2 billion in "actuarial" assets may not be fully funded but rather $200 million short of what's needed to pay promised benefits. If I tried to pay my bills with the "actuarial" value of my bank account, I'd be bouncing checks left and right-which, unfortunately, is what some companies are doing with their pension plans.
If anything, the situation is even more perverse on the liability side. Companies are permitted to calculate the present value of their pension liability using the four-year average of a corporate bond index. It should go without saying that interest rates from four years ago have absolutely nothing to do with the value of the pension liability today (or tomorrow). This is akin to driving down the highway at a high rate of speed looking only in the rear-view mirror.
Still, I can understand why plan sponsors want the flexibility afforded by smoothing the discount rate. It is a fact of life that pension liabilities are extremely sensitive to movements in interest rates. If the discount rate drops by one hundred basis points, that can easily drive up liabilities by ten percent or more. Better to "smooth in" that rate drop slowly over time to avoid unpleasant hiccups in the plan's funded status. Of course hiding the volatility doesn't mean it isn't there.
Without these (and other) smoothing mechanisms, the argument is made that companies won't be able to "predict" their pension contributions and won't be able to budget accordingly. This is a particularly fascinating line of reasoning. How can a CFO of an airline possibly function without being able to "predict" future oil prices? Or the CFO of an auto manufacturer with respect to steel prices? Or the CFO of a multinational enterprise that has to deal with currency fluctuations? Or, perhaps most similarly, a bank or insurance company CFO whose business is especially sensitive to changes in interest rates?
Ah, say the inhabitants of pension-land, but our obligations are "long term." These benefits are going to be paid out over decades, so there's no need to value the liability based on what interest rates are doing today.
Nonsense. I want to know the market value of my house today even if I have a thirty-year mortgage and plan to live in it for another thirty years-it affects my net worth and my ability to borrow. Moreover, there's always the chance that I may have to sell my house earlier than I expected.
Similarly, workers and retirees need to know the funded status of the pension plan today even if the benefits are going to be paid out over thirty years. Not only should it affect their planning for retirement, but there's always the possibility that their company may go bankrupt and turn its pension plan over to the PBGC. I can assure you: At that point a liability calculation based on interest rates from the year 2002 is utterly meaningless and misleading. Yes, most pension obligations are long term. But, there have been more than 160,000 standard terminations of fully funded plans over the past thirty years. There have been 3,600 terminations of underfunded pension plans. Ask the participants in these plans whether these are necessarily long-term obligations.
This is but a glimpse of the often fantastical world of pension funding and accounting. If it were simply the product of an imaginative storyteller, we might be suitably amused. If the effects were benign, we might not care. But neither is true-the interaction of pension finance, pension policy, and the decisions made by pension managers and the PBGC affects real people in the real world. It affects workers and retirees of American corporations, who can lose billions of dollars in promised benefits when plans terminate because of legal limits on the PBGC guarantee.
It affects corporations saddled with massive unfunded legacy burdens from an earlier era. To provide just one example: When it went bankrupt, the once-mighty Bethlehem Steel had about $1 billion in business debt and $7 billion in unfunded pension and retiree medical obligations.
It affects healthy companies with well-funded pension plans, which are growing weary of paying the pension bills of bankrupt steel and airline companies through ever-higher PBGC premiums.
It affects investors who cannot penetrate the pension veil to perceive a company's true value and end up paying more than they should.
It affects taxpayers who may one day be called upon to bail out the PBGC when corporate America says enough is enough.
It affects the capital markets, which may not be operating as efficiently as they should due to the distortions of pension accounting and the subsidies created by the PBGC guarantee.
But there is reason for hope. Whereas the broad implications of the pension system were heretofore not well understood, I believe we are poised to see improvements both on the funding front and the accounting front.
As you know, both the Senate and House have passed their respective versions of pension reform. Just last week the conference committee was formed to iron out differences between the two versions, and the chairman of the committee, Senator Mike Enzi of Wyoming, said he hopes a final bill will pass by April 7.
The congressional bills share some important common elements with the proposal the Bush Administration unveiled in January of last year. They increase funding targets for plan sponsors to one hundred percent of accrued liabilities, limit the period over which funding shortfalls can be amortized to no more than seven years, and require the use of a modified yield curve to more accurately measure pension liabilities.
That said, the Congressional bills need to be strengthened and improved to better protect the benefits earned by workers and retirees. The Bush Administration set the standard for the debate when it called for the elimination of smoothing, a funding target equal to one hundred percent of accrued liabilities that are accurately measured, full disclosure of the funded status of pension plans to participants and the public, and a more rational premium structure that raises enough revenue to bring the system back into balance.
We look forward to working with the conferees to produce a bill that strengthens the funding rules, increases transparency, and puts the insurance program on a sustainable solvency path. We're not there yet. If the end product of the conference committee is a bill that fails to improve on current law, the President's senior advisers have said they will recommend a veto.
Meanwhile, there are encouraging signs that pension accounting may also be in line for much-needed overhaul. The Financial Accounting Standards Board has announced a project to take a fresh look at accounting for all postretirement benefits, including pensions. FASB Chairman Robert Herz has been forceful in calling for a new regime that will account for pensions accurately and honestly.
While we need to carefully balance often competing interests, we also must recognize that the status quo results in the misallocation of capital, does a poor job of protecting the benefits earned by employees, does a poor job of protecting responsible companies from shouldering the costs of companies that terminate underfunded pension plans, and exposes taxpayers to the risk of having to bail out the federal pension insurance program. And the status quo is not going to save the defined benefit system. Changes are needed now-otherwise the costs will be higher and the decisions that much harder down the road. The promise to provide deferred compensation to workers is a straightforward proposition that should be accounted for and funded in a clear and financially sound manner. It is time to pierce the veil and allow the stakeholders of the pension system to understand these liabilities using the same rules that govern other corporate obligations. Even Alice didn't remain in Wonderland forever. Eventually she woke up.
Thank you again for inviting me to speak, and I would be happy to take your questions.