Keynote address to super bowl conference

I want to begin my remarks today with a discussion of the Pension Benefit Guaranty Corporation - what we do, why you've been reading about us in the papers, and what we're doing to address the challenges facing the pension insurance program. I will also address the subject of investment policy, both at PBGC and in the broader pension community, and attempt to draw some connections between the financial theory and the reality of managing against pension liabilities.

PBGC is really two businesses operating under one roof. We provide financial guaranty insurance covering 44 million workers and retirees in return for insurance premiums paid by plan sponsors, and we manage our asset base to provide life annuities to the beneficiaries of terminated pension plans. Unfortunately, we face both a clear shortfall in premiums relative to the risk posed by plan sponsors and a significant deficit between the assets we hold in trust and benefits we have promised to pay. The good news is that PBGC has the resources to continue paying benefits for several years; the bad news is that needed reforms grow more costly the longer we wait.

How did we get here?

The Pension Benefit Guaranty Corporation came into being following the passage of the Employee Retirement Income Security Act (ERISA) in 1974, with the purpose of providing financial guaranty insurance to the sponsors of defined benefit pension plans. Seen within the larger context of ERISA, the expectation for PBGC was that it would provide the layer of protection needed to make good on guaranteed benefits that were not fully funded by plan sponsors. The funding rules enshrined in ERISA were intended to make sure that such losses were small and could be easily shared across the sponsor population via the risk-sharing mechanism of insurance premiums.

Unfortunately, the basic "business model" and funding rules were flawed from inception.
Although PBGC began its existence with a clean balance sheet, it was inevitable that some plan sponsors would fail with less than fully funded plans, requiring PBGC to take over the assets and liabilities of the plans - and book the difference as a net loss. Over the past three decades more than 3,000 plans have been taken over by PBGC, and our balance sheet has grown dramatically. Today PBGC has over one million beneficiaries, more than $60 billion in present-valued liabilities and a $23 billion dollar negative net position. In fact, we have experienced a $30 billion swing in our net position over the past three years.

It is now clear that the original expectation for ERISA's funding rules - that they would result in plans being fully funded across time - has gone unfulfilled. Some of this has been the result of investment choices and market fluctuations (some would like to place all the blame on the "perfect storm"), but much of the problem has been the result of flaws in the structure of the funding rules themselves. At certain times over the past 30 years, Congress has toughened funding standards and raised premiums on plan sponsors in order to deal with past losses and address the risk posed by underfunded plans. We are quickly approaching such a point once again - actually, we've arrived there.

Reforming the System

The administration has been developing a reform package that will substantially improve the long-term prospects of the pension insurance system. There are four key components: improving plan funding rules to ensure that benefit promises made to employees are properly secured, implementing premiums to price insurance fairly and disincentivize risky behavior, encouraging greater transparency in the system for all stakeholders - beneficiaries, sponsors, shareholders and regulators - and provide PBGC with additional tools to better manage risks to the pension insurance fund.

Apart from the legal and moral mandate to honor promises made, well-funded pension plans are essential to the long-term viability of the pension insurance system. Plan funding rules are currently based on the concept of "current liability," which properly reflects the ongoing nature of most businesses that sponsor plans but fails to account for the significant changes to projected cash flows when a sponsor is likely to terminate a pension plan. PBGC's experience shows that plans following the "letter of the law" have often terminated in a severely underfunded state. This was the case with our largest claim to date, Bethlehem Steel, and, as we've noted publicly, it would be the case with United Airlines as well. Funding rules with clearer standards, better liability measures and fewer loopholes will go a long way towards fixing this problem.

The premium system is also overdue for reform, having remained unchanged for more than a decade despite the surge in costly plan failures. The current premium structure falls short in three ways: first, it does not adequately compensate PBGC for expected claims; second, it fails to provide an effective disincentive for risky behavior; and third, it can only be adjusted via legislation. The key will be to balance the incentive structure (i.e. higher cost for higher risk) with the need to spread risk across all plan sponsors (as with any insurance system, all the insured must pay something). However, the imbalance in the current system is reflected by the fact that UAL has paid a total of approximately $50 million in premiums while potentially presenting a claim of over $6 billion.

The pension system is critically short of transparency. Both ERISA and generally accepted accounting principles (GAAP) seem designed to obfuscate reality when it comes to pensions. Pension accounting and disclosure are inadequate to allow investors, beneficiaries and regulators to make informed judgments about the health of pension plans. This lack of useful information prevents stakeholders from effectively exercising their informed self-interest and allows risk to build in the system unobserved. Investors are increasingly aware of the impact pension underfunding and asset-liability mismatches can have on the plan sponsor, but the current state of disclosure precludes complete understanding of their exposure.

Finally, we need better tools to enable us to better manage risks and minimize losses to the pension insurance fund. This is especially true when a company enters bankruptcy where, in the case of UAL and US Airways, the companies have failed to make contributions required by law, with no adverse consequence (other than some bad press). But, we also need additional tools to help ensure that companies are appropriately funding their pension plans long before they get into financial difficulty.

Apart from needed changes in the statutory framework, within PBGC we are working on developing our ability to measure risk and react appropriately, particularly with respect to the contingent exposure from underfunded plans with risky plan sponsors - totaling almost $100 billion. Because so much of this exposure is concentrated in a relatively small number of firms, even the best statistical modeling techniques result in fairly wide distributions of expected outcomes. As a result, we are intent on developing a greater capability to use real-time market information, such as the credit derivatives market, to assess the risk of individual plan sponsors and react more quickly to changes in our exposure.

We have also begun reshaping our investment program. The goal is to better reflect our position as an insurer and annuity provider, primarily by focusing on liability hedging using actively managed and enhanced index fixed income strategies to outperform a liability benchmark over time. This policy shift has been developed in response to the substantial balance sheet volatility of the past few years and a recognition that, as an insurer to defined benefit plans, PBGC's exposure to the markets extends beyond our own portfolio.

Looking at the Pension System

It goes without saying that the past few years have been a period of extraordinary volatility for the defined benefit pension system. PBGC has experienced this rollercoaster firsthand, both within our portfolio of assets and liabilities and through the increasing frequency and severity of plan terminations. While our experience is somewhat unique, certain observations on the pension system at large are worth discussing.

Many of the problems in the defined benefit system over the past few years are the result of a broad failure to understand the impact of a defined benefit pension plan on the corporate income statement and balance sheet. Our accounting regime has been partly to blame. GAAP accounting does not consolidate the pension plan on the balance sheet, and many investors have long been accustomed to seeing the pension plan's impact only through an annual entry on the income statement. This is like looking through the wrong end of the telescope - massive liabilities that will accrue over many decades are ignored in favor of the tiny piece of the total that rolls across the income statement each year. It has been almost impossible to gain a real understanding of a pension plan's net position on an accrual, or present valued, basis.

For many years there was little reason to care. Strong investment returns through most of the 1980s and 1990s allowed pension plans to become a source of net income rather than a use of it. In addition, because the gains on equity-heavy portfolios consistently outpaced the growth in liabilities resulting from declining interest rates, sponsors were happily ignorant of the asset-liability mismatch risk they were running. In this area particularly, our accounting regime has encouraged riskier behavior by allowing plan sponsors to book expected returns on assets immediately while amortizing real volatility over time. It is only natural that, in such a regime, plan sponsors would operate with a bias towards volatile assets with higher long-term expected returns.

With the popping of the equity bubble and the drop in interest rates post-9/11, the illusion of stability vanished. As it was, investors were shocked to discover that pension plans could result in a charge to corporate income. Had they been able to see a more accurate picture of pension plans' net position on an accrual basis, the picture would have included a very large gap on many balance sheets as well. Analysts have begun sharpening their pencils and digging into the pension accounting in search of a clearer picture. ("Turning the telescope around," if you will.) New disclosures under Financial Accounting Standard 132 are a step in the right direction, but the best analysts will tell you that we still have not reached true transparency.

It is not just the analysts who are changing their approach. Plan sponsors are beginning to reassess their approach to pension management in a way that integrates the assets and liabilities. In this regard, the traditional separation of the asset management business (which has typically been handled by a Chief Investment Officer) from the benefits business (which has been the product of collective bargaining or management negotiations with workers) is being reconsidered. Defined benefit pension plans may come to be viewed as a subsidiary business that functions like an annuity underwriter, with assets and liabilities that should be carefully balanced. Although it may take years to overcome the habits of the past, such a change in perspective would enhance the viability of the pension insurance system.

Looking Ahead

One practical approach that may have value is the use of liability benchmarks to assess the success or failure of pension investment policies. After all, what good does it do a plan sponsor if its investment managers outperformed the S&P 500 by 300 basis points when the S&P 500 underperformed liabilities by twenty percentage points? (Answer: not much.) Even in the opposite case when the mismatch works to a firm's advantage, risk is risk. The risk and reward of different asset allocations should be viewed not in absolute terms (or "asset-only") but, rather, relative to liabilities (or "asset-liability"). At the individual manager level, standard market benchmarks may be appropriate for judging performance but, at the overall plan level, a liability benchmark would seem to be more appropriate.

Ultimately, the proper yardstick for success in managing a defined benefit pension fund is maintaining a well-funded plan over time. Such a plan should be able to pay expected benefits without drawing on the resources of the plan sponsor. The desire to invest in assets with higher long-term expected returns must be balanced against the volatility of those assets and their correlation with liabilities. Simply assuming that returns in any given year, or period of years, will look like their historical average is to ignore the potential impact of asset volatility on the plan and its sponsor. Similarly, failing to account for the correlation of the assets with liabilities is to ignore the impact of mismatch risk. Investors have received a very real lesson about these risks over the past few years - I hope they remember it down the road.

Time does not allow us to delve into these issues with appropriate depth. And, there certainly are other issues that warrant consideration, such as one of the key themes from this conference, the search for alpha. Now, I'm a big believer in alpha. Markets aren't perfect, and discerning and nimble investors can identify and capture market inefficiencies to their profit. But, the notion that a couple of trillion dollars of private and public sector pension money wants to chase alpha gives me great pause. While some may realize above-market returns for short periods, collectively you are the market. Inevitably, you will spend hundreds of millions of dollars and incur substantial friction and transaction costs for what ultimately is a zero sum game. But, that is a topic for another day.

I thank you for inviting me to share these thoughts with you, and I would be pleased to take any questions you may have.