Effective LDI: Getting to 100% and Staying

Michael J. Bazdarich
Product Specialist/Economist

Executive Summary

  • Achieving and sustaining fully funded status is the goal of any DB pension plan.
  • Aggressive risk-asset allocations are unlikely to achieve this goal unless the plan sponsor is willing either to substantially over-fund initially or else to make substantial, periodic cash contributions in excess of service costs.
  • Rather than waiting and hoping for an aggressive allocation to succeed, an alternative pension strategy is to fully fund and derisk now, reaping possible tax savings as well as the benefits of a more stable pension plan.
  • While such a strategy lacks the “sizzle” of an aggressive allocation, it is more consistent with both real-world experience and financial economic principles.

Introduction

Every defined-benefit (DB) pension plan aspires to fulfill its obligations to beneficiaries. In order to do that, it needs to achieve and sustain 100% funded status now or at some point in the future. Suffice it to say that achieving or surpassing 100% funded status is an ultimate goal of every pension plan.

However, the DB plans for most large corporations currently have funded status levels in the high 80s or lower. We argue here that such plans are unlikely to achieve and sustain full funding from the investment strategies they are currently pursuing unless they receive large cash infusions. Yet the very rationale for these strategies is for the plan sponsor to avoid such contributions. By “current investment strategies,” we mean aggressive allocations dominated by investments in equities and alternatives. While such allocations did briefly achieve full-funded status in the late-1990s, the trends in funded status for the bulk of corporate (and public) DB plans have been downward ever since, despite decent stock market performance since then.

Many plans continue to anticipate that the next move upward in bond yields and equity prices will achieve full funding, at which point they can derisk and get off the risk-asset rollercoaster. The experience of the last 17 years suggests otherwise. Defined-benefit plans generally declined to derisk in 1999, when funded status levels were around 130%, and they declined to derisk in 2006, when funded status levels were around 100%. Presently, we are nearly 7 years into a bull market, with very high equity market valuations, and plans are still heavily allocated to risk assets.

It is not impossible that equities will rally further from here, nor is it impossible that bond yields will rise substantially. However, the relevant question is how well the wait-and-hope strategy has worked so far. An alternative strategy is for sponsors to contribute enough to bring their plans to 100% funded status today and then move immediately to a low-risk allocation.

Most well-managed US corporations are flush with cash. Why not allocate some of that cash to pensions? The plan has already promised to pay the benefits; why not fund those promises? Such investments will likely boost corporate asset returns, since bond yields far exceed current near-zero money fund returns, and they are most likely to reduce corporate tax liabilities, since pension contributions are generally tax deductible up to a 130% funded status.

The objections to such a move are well known. Bond yields are too low. Equity investments reduce the cost of the pension to its sponsor. Lowering the expected return on DB plan assets—via derisking—will reduce corporate earnings. All these objections can be rebutted on the facts. All of them run contrary to financial economics, which commonly finds that 100% funding in long bonds is the only responsible, shareholder-friendly allocation for a DB plan. Just as importantly, these objections appear to be at odds with historical experience.

We believe a sensible and prudent path for corporate DB plans is to achieve full funding immediately via cash contributions and then sustain that 100% funded status via a low-risk long duration asset allocation. We’ll argue here that such a strategy does not incur the pain levels commonly associated with it and that, in any case, it will prove less painful than a high-risk strategy that fails.

Only Large Cash Contributions Have Prevented Downtrends in DB Funded Status

Exhibit 1 shows Federal Reserve (Fed) data on aggregate corporate DB plans’ funded status back to 1951 and aggregate funded status back to 1997 for the 24 current members of the Dow 30 that sponsor a DB plan.1 Prior to the 1970s, funding of corporate DB plans is best described as haphazard. After the passage of federal ERISA legislation in 1974, corporate sponsors began to take pension obligations more seriously,2 and funded status accounting protocols evolved to current practice. Ironically enough, as the green line in Exhibit 1 shows, it was only when DB plans began to be better funded that asset allocations shifted out of fixed-income and (mostly) into equities, where they remain. Whatever the reason for the shift out of fixed-income, funded status fluctuations have since become more pronounced.

Exhibit 1
Aggregate Funded Status of Private DB Plans and Aggregate Allocation to Fixed-Income
effective-ldi-getting-to-100-and-staying-2015-12
Source: Federal Reserve Board, Corporate Annual Reports, Data through 4Q14

The massive, protracted bull stock market of the 1980s and 1990s helped push funded status levels from the 85% range in 1979 to as much as 130% at the stock market peak in 1999.3 However, the trend in funded status levels ever since then has been clearly and substantially downward. Each stock market peak has been associated with lower peak funded status levels than previously.

Now, consider the funded status behavior of individual Dow 30 companies (with names removed), as shown in Exhibit 2.4 Funded status levels of Dow 30 plans generally follow the same trends as do the Dow 30 and private pensions in aggregate: ever-lower levels of funded status at each stock market peak.

There are five exceptions, each of which are designated by green shading in Exhibit 2. The one characteristic distinguishing these plans from the others is that all five have made substantial cash contributions well in excess of service costs. Ironically, the asset allocations of these five are generally the same as the others, dominated by investments in equities and other risk assets. It is cash contributions that drove the sharp improvements in funded status exhibited by these five plans. The plans with downtrends in funded status failed to make such contributions.

Exhibit 2
Funded Status Levels' Historical Performance for Dow 30 Companies (Percent of PBO Liabilities)
effective-ldi-getting-to-100-and-staying-2015-12
Sources: Annual reports and 10-K filings of current Dow 30 companies from 1997 through 2014, compiled by Western Asset. All charts are plotted over 1995-2014 and 2008 shaded gray

Sound Pension Management in Theory and in Practice

This experience provides a direct challenge to the principle that aggressive DB asset allocations can reduce sponsors’ pension costs. Over the last 17 years of experience, aggressive asset allocations have required substantial cash infusions to avoid deteriorating funded status. These asset allocations have not been able on their own to sustain funded status levels.

What’s more, this historical experience supports the results we reported in a recent white paper (Bazdarich 2014b). That analysis found that in a series of Monte Carlo experiments, even adequately funded plans invested in risky assets face a greater-than-50% chance of bankruptcy, the reason being that the need to make benefit payments during market downturns so depletes plan assets that asset levels are not able to fully recover even when risk-asset markets have fully recovered.

“Adequately funded plans” in that analysis meant funding levels high enough that if asset returns held constant at their long-run average, there would be sufficient plan assets to successfully fulfill all benefit payments. In the real-world, where asset returns constantly fluctuate around their means, that same plan was most likely to exhaust assets well before the last benefit payment was made. Strategies that work under “actuarial” conditions of constant returns generally fail in the real world of fluctuating returns.

In order to materially reduce the chances of bankruptcy, a sponsor would have to materially overfund the plan—or else make regular contributions. And keep in mind that the analysis in the paper concerned plans where benefit cash flows were fixed: hard-frozen plans. In order for non-frozen plans to avoid a high risk of bankruptcy, not only would initial funding have to be much more than “adequate,” but also additional ongoing contributions would have to be made to fund service costs.

Notice that such “over-funding” is exactly what DB plans hope to avoid when they embark on a risky asset allocation. The whole point of risky DB asset allocations is to “reduce costs” by funding each dollar of DB liabilities with less than a dollar of assets. Our analysis finds that such strategies are unlikely to be successful even if the plan has accurately assessed average rates of asset returns, as the inevitable combination of (temporarily) below-average returns and continued benefit payments wreaks havoc.

This finding from extensive Monte Carlo studies is verified by the real-world experience of corporate DB plans, as shown in Exhibits 1 and 2. It also jibes with financial economic theory. Writers such as Jack Treynor (1977, 1978), Merton Miller (1978), Fischer Black (1979), Irwin Tepper (1982), John Exley (1997) and others have criticized the “something for nothing” practice whereby plan sponsors “fund” each $100 worth of pension liabilities with something less than $100 of high-return, high-risk assets.

Their points, among others, are that:

  1. such a leveraged bet on the stock market accomplishes nothing for shareholders that they can’t accomplish on their own;
  2. such a leveraged bet is tax-inefficient, since pension asset returns are not taxed, while stocks are already tax-favored, so that a pension plan buying stocks is like a non-taxed investor buying muni bonds; and
  3. while the plan sponsor will be stuck for any underperformance of the risk assets, they will likely not benefit from any over-performance of them, since there are formidable hurdles to withdrawing excess funds from a plan.

These points all imply that underfunding via risk assets is imprudent in theory. Our results indicate that it is also unlikely to succeed in the real world.

Why Not Boost Contributions and Derisk…Now?

The alternative to risky asset allocations is to derisk, but that requires, of course, a move to full funding. Long bonds can closely hedge the movements in DB liability returns and so sustain funded status, but cash contributions are necessary for most plans to first achieve full or nearly full funding. Every DB plan understands this, and most intend to derisk at some point in the future.

Given the financial economic principles and real-world experience detailed previously and given the advanced age of the current bull market, the question becomes: why not derisk now? Yes, the plan sponsor will have to pony up cash to achieve full funding, but it is likely that the plan will have to make such contributions eventually anyway. Many corporations are flush with cash at present, and investing that cash in their pension plans would generate a number of advantages.

Tax-savings are the most immediate such advantage. As we mentioned, cash contributions to a DB plan can be tax deductible up to the point of 130% funding.5 We hear constantly about US corporations “stashing” offshore earnings abroad to shield them from domestic taxation. Channeling those funds into pension contributions could allow the corporate sponsor to repatriate the earnings on a nontaxable basis, funding a current corporate liability now that will need to be paid down the road.

Many corporate CFOs spurn derisking strategies because they fear they will hurt earnings, since the expected return for bonds is less than that for stocks, and expected returns for DB assets are a direct boost to earnings under current FASB protocols. However, this aversion ignores two facts. First, it is total return on assets that enters earnings, not the rate of return, and boosting total plan assets works to offset any lowering of the expected rate of return in determining total asset returns. Second, if the plan sponsor has ample cash, those funds are likely earning near-zero (taxable) returns today in a money fund, while contributing them to the DB plan would allow them to earn the higher (non-taxed) return on bonds. (And even if bond yields were to rise, any paper losses on the bonds within plan assets will be offset by lower valuations—gains—on DB liabilities.)

Even if the plan sponsor has to issue debt to fund the contributions, the interest paid on that debt will be tax-deductible, while the returns earned on the invested contributions are not taxed. And such a debt issue will not hurt even pre-tax earnings, since the interest earned on the bonds in plan assets offsets the interest costs of any such debt issue.

Of course, plan sponsors should be aware that full funding and derisking cannot fully eliminate DB plan risk. As we have detailed in past white papers,6 DB liabilities are not perfectly hedgeable, and any long-duration LDI allocation will experience some tracking error against liabilities. However, this risk will be orders of magnitude lower than what accrues from a risk-asset allocation. A 60/40 or 70/30 equities/bonds allocation produces 12% to 15% per year volatility in funded status. In contrast, even a simple long duration allocation can reduce funded status volatility down below 4% per year, and properly customized allocations might further reduce it toward the 2%-per-year level.

Conclusion

Full funding and derisking of corporate DB plans is commonly considered a costly procedure, but is it? And could it be any more painful than experiencing an ongoing decline in funded status? Making cash contributions to fully fund the plan does use up corporate cash, but those assets will remain on the sponsor’s balance sheet, while possibly boosting pre-tax earnings and almost certainly boosting after-tax earnings.

Under-funding a DB plan and then investing plan assets in high-return, high-risk sectors is thought to provide a low-cost “funding” of pension promises. However, financial economic analysis finds this premise to be faulty. What’s more, the real-world experience recounted here bears less than favorably on such strategies.

We argue that the best way to achieve and sustain 100% funding is to adequately fund the plan and then invest plan assets in low-risk long-duration strategies. This strategy doesn’t have the sizzle of stocks, but it does offer a better chance of success.

References

Bazdarich, Michael J. 2014a. Don’t Sweat the Small Stuff.

------------------. 2014b. Why All Defined-Benefit Plans Are Short-Term Investors.

Black, Fischer. 1980. The Tax Consequences of Long-Run Pension Policy, Financial Analysts Journal, pp. 21–28.

Exley, John, S.J.B. Mehta, and A.D. Smith. 1997. The Financial Theory of Defined Benefit Pension Schemes, Presented to the Institute of Actuaries, BAJ 3, pp. 835–966.

Miller, Merton. 1977. Debt and Taxes, Journal of Finance, vol. 32, pp.261–273.

Tepper, Irwin. 1981. Taxation and Corporate Pension Policy, Journal of Finance, vol. 36, pp. 1–13.

Treynor, Jack. 1977. The Principles of Corporate Pension Finance, Journal of Finance, vol. 32, pp. 627–638.

------------------. 1978. Pension Claims and Corporate Assets, Financial Analysts Journal, May, pp. 21–28.

  1. Aggregate private-sector DB data are taken from the Federal Reserve Board’s Financial Accounts of the US, table L.118b. Fixed-income assets’ shares are also taken from this source. In its estimates of liability valuation, The Fed does not follow the standard GAAP protocol of using mark-to-market AA yields. Instead, it uses its own estimates of Treasury and corporate bond yields, subject to a floor of 6%. As bond yields have been well below 6% for several years, its discount rates for liability valuation have held steady ever since then. This is why the Fed’s aggregate measure of funded status does not show declines in recent years comparable to those exhibited by the Dow 30. We show historical Fed data because the 6% discount rate floor was not a major source of distortion in the 1980s and 1990s, and we wanted to provide a longer take on funded status history than can be extracted from the 10-K information offered by individual companies. (Only since the late-1990s have FASB protocols dictated publication of funded status levels in 10-K and annual report filings.)
  2. This remark is largely based on comments in Treynor (1977), but the data shown in the chart are clearly consistent with it.
  3. For Dow 30 companies, average funded status in 1999 was 132%. Fed data show a 118% funded status levels for 1999. As explained in endnote 1, the Fed estimates do not fully allow for the effects of changing bond yields on DB liability valuations.
  4. While Dow 30 member American Express sponsors a DB plan, it ceased reporting funded status levels in its 10-Ks after 2012, so only 23 Dow 30 plans are shown in Exhibit 2.

  5. For corporate income tax purposes, such funding levels are calculated on an ABO basis, using yield curves published by the IRS to discount DB benefit payments flows. Still, that 130% “Treasury basis” funded status translates into FASB-compliant funded status levels far above those for most corporations currently.
  6. See Bazdarich (2014a), available on our website, as well as the earlier papers cited there.
Past results are not indicative of future investment results. Investments are not guaranteed and you may lose money. This publication is for informational purposes only and reflects the current opinions of Western Asset Management. Information contained herein is believed to be accurate, but cannot be guaranteed. Opinions represented are not intended as an offer or solicitation with respect to the purchase or sale of any security and are subject to change without notice. Statements in this material should not be considered investment advice. Employees and/or clients of Western Asset Management may have a position in the securities mentioned. This publication has been prepared without taking into account your objectives, financial situation or needs. Before acting on this information, you should consider its appropriateness having regard to your objectives, financial situation or needs. It is your responsibility to be aware of and observe the applicable laws and regulations of your country of residence.
Western Asset Management Company Distribuidora de Títulos e Valores Mobiliários Limitada is authorised and regulated by Comissão de Valores Mobiliários and Banco Central do Brasil. Western Asset Management Company Pty Ltd ABN 41 117 767 923 is the holder of the Australian Financial Services Licence 303160. Western Asset Management Company Pte. Ltd. Co. Reg. No. 200007692R is a holder of a Capital Markets Services Licence for fund management and regulated by the Monetary Authority of Singapore. Western Asset Management Company Ltd is a registered financial instruments dealer whose business is investment advisory or agency business, investment management, and Type II Financial Instruments Dealing business with the registration number KLFB (FID) No. 427, and members of JIAA (membership number 011-01319) and JITA. Western Asset Management Company Limited ("WAMCL") is authorised and regulated by the Financial Conduct Authority ("FCA"). In the UK this communication is a financial promotion solely intended for professional clients as defined in the FCA Handbook and has been approved by WAMCL. Potential investors in emerging markets should be aware that investment in these markets can involve a higher degree of risk.