Current Accounting Rules Understate Pension Problem

As bad as the underfunding of state pension plans appears, the economic reality could be much worse, says a vocal faction of actuaries and financial economists.

The reasoning goes like this: States can no more neglect to pay their pension obligations than they can pass on paying interest to bondholders. If it’s a near-certainty that the benefits will have to be paid, then they should be valued with a commensurate conservative discount rate, such as the yields on U.S. Treasury bonds. It’s this sort of thinking that led the Financial Accounting Standards Board and Congress in 2006 to require corporate defined benefit sponsors to value their pension obligations using yields on high-grade corporate bonds.

Instead, governmental accounting standards have allowed pension plans to value their liabilities at the assumed rates of return on their investment portfolios, which range from 7 to 8.5 percent; the average is 8 percent, although some states have started to scale back the rate a bit. The logic is that because it’s the rate the assets are going to earn, or at least are expected to earn, it’s appropriate to value the liabilities the same way.

“Although public pension funds, along with most other investors, have experienced subpar returns over the past decade, median public pension fund returns over longer periods exceed the assumed rates used by most plans,” asserts a March 2010 report by the National Association of State Retirement Administrators. The 20-year median return through 2009 was 8.1 percent, and the 25-year median was 9.3 percent.

问题是,平均不够好。To build assets sufficiently to satisfy the liability, the portfolio has to earn at least the discount rate every year. “We often falsely assume that the differences between money- and time-weighted returns are negligible,” says Norman Ehrentreich, principal of Ehrentreich LDI Consulting & Research in Minneapolis. “Yet large cash flows, high return volatility and low funding levels can drive a big wedge between these two rates. Funding costs are path-dependent on the specific return sequence, and time-weighted returns are inappropriate to use to make any statements about past funding costs.” The problem is solved, he says, by incorporating more-conservative discount rates.

Professors Joshua Rauh of Northwestern University and Robert Novy-Marx of the University of Rochester have developed estimates of the difference between funding measured at actuarial value and at market discount rates. The numbers are staggering: Recognizing future service and wage increases, they estimate liabilities at $3.6 trillion and $5.2 trillion using municipal bond and Treasury yield curves, respectively, as of June 2009. Compared with $1.8 trillion in pension fund assets, they say, the baseline level of unfunded liabilities is therefore about $3 trillion under Treasury rates. The Pew Center on the States places the total unfunded liability, albeit measured in June 2008, early in the financial crisis, at about $500 billion.

The situation is unlikely to change soon, however, as the Governmental Accounting Standards Board recently has proposed revisions to pension accounting without broad changes to valuation methods.

The dispute over discount rates is not just an academic amusement. Proper valuation is essential to the current efforts to get pension plans on an affordable long-term trajectory. If the discount rates used to value pension obligations are too high, the liability is understated by 50 to 60 percent, according to pension scholar M. Barton Waring. “And that colors everything else,” he says. “Benefit promises look less expensive than they really are, because they are based on the liabilities.”

Adds Waring: “If we keep up the fiction, funding will become worse and worse until plans go bankrupt. The only path to saving them is to build the emperor a new set of clothes, and that includes the painful and unwelcome task of renegotiating benefit levels.”