Pension Finance: Putting the Risks and Costs of Defined Benefit Plans Back Under Your Control

by M. Barton Waring, Wiley Finance (2011), Hardback. ISBN 978-1-118-10636-5.

Pension finance need not be complicated—it is in essence a classic financing problem of trading off consumption now for saving and consumption in the future. A pension plan could be seen as one long bond, one short bond and the asset fund. Assets plus the stream of future contributions have to match, in present value terms, the stream of promised future benefits. We can calculate lease payments on a car using the manufacturer’s website and enter any amount for down payment, but we cannot input our own lease rate or residual value. The plan sponsor may have some choice on timing and size of the contributions, but the liability remains the same. Despite their simplicity, pension funds in the United States are in a state of crisis, with unfunded debt, according to the author’s calculation, approaching $4 trillion. Pension plans in other countries have not fared much better. Many plans, believing that they were in surplus at the beginning of the century, declared ʻcontribution holiday’, only to see the effects of declining stock markets and interest rates turn surpluses into large deficits.

Focusing on market values as inputs to all variables, the author of this book shows how the liability and costs should be calculated from an economic standpoint and demystifies the actuarial terminology.

Accrued liability is the liability that is owed and ultimately has to be funded. The process of accruing liability is similar to making payments to amortize a debt, and in the case of a pension fund this debt would be the present value of future benefit payments to current employees. Normal costs are simply these amortizing payments. As for retirees, their benefits have already been charged as normal costs and accrued to the liability while they were employed.

The choice of method to calculate normal costs has no effect on the size of the promised benefits (except, perhaps, affecting the security of these benefits when slower methods leave a smaller portion of the present value of the future benefit payments secured). It looks simple so far, and of course it would be if it were not for the actuarial practice of using one liability measure and normal cost calculation for pension expense and another one for contributions. The inconsistency becomes even greater when a different discount rate is used for the stated liability and for the contribution calculation.

We need a discount rate to equate the value of the pension assets today with the value of liability to the retirees in the future. There is no market to trade pension obligations, so this rate cannot be directly observed, it needs to be estimated. There should be only one discount rate to calculate the present value of risk-free future benefits—it is the risk-free rate, or the rate of return on a mixture of nominal and real return bonds with matching horizons. Unfortunately, the discount rate varies from plan to plan and sponsors have an incentive to keep the discount rate assumption high to show lower accrued pension obligation. Of course, the real pension obligation is not affected by the choice of discount rate and the only way to reduce it is to cut the benefits. This will not be popular with current and future employees, while changing the discount rate assumption in the footnotes to the annual report will hardly be noticed.

This discount rate could be interpreted as the rate of interest paid to the plan participants on their deferred earnings. Naturally, sponsors would want the borrowing costs to be low so it does not make sense that they are guaranteeing their employees a higher rate of return on their deferred earnings. Yet today most sponsors use the expected return on an investment portfolio as a discount rate for their liability, making it appear lower than it really is. To make things worse, this rate is then called the ʻrequired’ rate of return, limiting the choice of investment strategies to those that can beat this target and leaving more prudent (less risky) asset mixes out. An improper discount rate is the single biggest problem in current pension practices and the author keeps coming back to remind us about it throughout the book.

The author also points out that the risk of an investment does not decrease with time and the ʻexpected’ rate of return is not something that we can expect to receive over the long term; in fact, we get the realized rate of return and the ending value of an investment is a random draw from an ever-widening (with the square root of time) distribution.

The author believes that the only amortization in use should be in accruing the normal costs into an accrued liability. There are no benefits to using amortization or smoothing anywhere else, be it the amortization of newly awarded benefits for prior service that immediately become an unfunded part of accrued liability, amortization of make-up contributions, smoothing of the liability’s volatility by keeping the discount rate constant or smoothing of the asset returns. Smoothing does not change the risk over the long term, only instead of showing the most recent volatility, it reflects the volatility of the years prior to the smoothing window. Smoothing makes sense only if markets do not follow a random walk and are mean-reverting; we know this is not the case. If smoothing is required, the author’s suggestion is to smooth only the volatility of the net surplus or deficit, not the assets or liabilities separately. Likewise, including demographic risks into the asset-liability model is only a distraction, since these risks cannot be hedged and any correlation with economic variables is spurious.

As an alternative to current asset-liability studies, the author suggests using surplus optimization and a two-fund approach with the following steps.

1.	Adopt a liability-matching asset portfolio as a policy.
2.	Set the size of the risky assets’ portfolio exposure.
3.	If desired, adopt tactical asset allocation policies for the risky assets portfolio.

However, the author believes that no matter how good an investment strategy is, it will probably not help the funds that are already in trouble. Holding more equities or exotic assets is not a solution to reducing the deficit and will only increase the risk of a plan becoming even more underfunded; only large make-up contributions and/or reductions in benefits can bring the plans back to full funding.

For those of us who are not actuaries, chapter 11 of the book provides a brief overview of various actuarial methods, perhaps only to beg the question “are they still doing this?” I can see some actuaries trying to defend the current practices, but as the author points out, there is no need to defend the logic of HP12C when calculating the time value of money and amortization payments.

If you are an investment professional, you may find the book reiterating principles that are obvious to you, but, then again, how obvious was it to the portfolio manager who provided a back-of-the-envelope estimate of the long-term return on their asset class to someone collecting these numbers ʻfor the Board’, that this estimate would come back to haunt them next year as a part of the now ʻrequired’ rate of return they are supposed to beat?

© 2013, Vadim Gracie