Defined benefit plan investing is in a transitional phase, evolving from an asset-only framework to one that focuses on the impact of the asset/liability relationship.
This shift is occurring as plan sponsors realize that by using asset/liability management they usually can reduce the present value of future contributions to their defined benefit pension plans by more than 20%. Also, the risks associated with the corporation's cost of contributions can be reduced by a similar magnitude.
The key to these savings is the recognition that the plan sponsor bears almost all of the downside risk, but only receives part of the upside reward.
There are only three ways a corporation is financially affected by its defined benefit plans:
corporate cash flow;
expense (accounting required under Financial Accounting Statement 87); and
borrowing cost (credit rating).
Note that all three of these are directly related to pension assets minus pension liabilities (the funded status of the plan). There is absolutely nothing that affects the corporation economically that is predicated on either the assets or the liabilities on a stand-alone basis. Proper management of the asset/liability relationship is critical.
Corporate cash flow
A defined benefit pension plan should be viewed as part of the capital structure of the corporation. The plan participants have been promised deferred wages by the corporation. This promise is met through required contributions to a pension fund as dictated by a host of laws and regulations.
The "cost" of these contributions is the true liability of the corporation. The "economic cost" of the benefit promise, from a corporate finance perspective, is the present value of the required future contributions.
The corporation also owns a call on the plan's revertible surplus. Thus, a more accurate definition of the economic cost of the plan is the present value of future contributions less the value of the call on the revertible surplus. In most cases, however, surplus cannot be reverted without a substantial penalty, so the callable amount usually will be considerably less than the potential surplus buildup.
From the corporation's perspective, the future stream of contributions looks just like a stream of payments it must make on borrowed money, e.g. debentures. Because the company cannot easily renege on its obligation without replacing it with something of equivalent value, it should be considered debt. Furthermore, this debt should be classified as unsecured debt, because, if the company became insolvent, the plan would not have a claim on the company with respect to the future contributions.
Therefore, each contribution due in the future should be discounted at the appropriate cost of borrowing for that term. This leads to the conclusion that the discount rate should be specified by the term structure of borrowing cost of the corporation.
Note that the economic liability, namely the present value of future contributions discounted at the sponsor's term structure of borrowing cost, is dependent on the assets, the liabilities and the sponsor's borrowing cost. This "economic" liability is very different from any combination of assets and liabilities calculated by the plan's actuary or the corporation's accountant or auditor.
Generally, investment managers are used to working in a framework where returns and funded status are symmetric in value, i.e. a gain of $1 or 1% is as valuable as a loss of $1 or 1% is costly. This is not true, however, with respect to the present value of contributions, which is a highly asymmetric function of funded status.
Each additional dollar added to the surplus saves the company less in terms of present value of contributions. This is easily seen when comparing the savings if a plan moves from 100% to 200% funded on an accumulated benefit obligation basis vs. a plan moving from 200% to 300% funded.
Moving from 100% to 200% funded clearly saves the company a great deal of future contributions. However, moving from 200% to 300% funded, which is the same dollar amount of increase, saves the company considerably less money. At 200% funded, the company may very well not be required to make significant (if any) future contributions.
Similarly, as the surplus diminishes (or deficit increases), every dollar lost has a greater cost to the corporation.
In addition to the asymmetry introduced by the full funding limit (contributions go to a minimum of zero but never negative), there also are asymmetries from increasing PBGC premiums, deficit reduction contributions and term structure of borrowing costs.
The accompanying table sets out an example of this based on a 30-year stochastic projection.
Note that although the funded status is changing symmetrically, the present value of contributions is affected asymmetrically as every increase of 40% in funded ratio saves less money in terms of corporate indebtedness. It is critical to manage this asymmetry in both investment and funding policy.
Expense under FAS 87
Expense - the charge against income as calculated under FAS 87 - can affect stock price. Not only do many stock analysts and investors focus on earnings reports and changes therein, but also some active and tactical managers use computer programs that select stocks on the basis of earnings ratios. Thus, there is a case for believing pension expense numbers play a role in stock price.
The actual expense calculation under FAS 87 does not, however, reflect the true costs of the plan to the corporation. The single most important flaw is that FAS 87 implicitly assumes all of the surplus in the pension plan belongs to the corporation. FAS 87 requires a reporting of income on the full earnings of the assets irrespective of how well the plan is funded. Therefore, an extremely well-funded plan will report a substantial income on the company's income statement and an asset buildup on the company's balance sheet (pre-paid expense).
However, if the shareholders tried to access this "asset," they would realize the excise tax and other costs involved in reverting the surplus to the corporation would diminish its value significantly. For this reason, it is important to control the volatility of expense, but not focus on absolute levels beyond a short period.
Borrowing cost (credit rating)
The PBGC has a claim against corporate assets in the event that an underfunded plan is terminated. Credit agencies and borrowers have grown increasingly aware and are concerned about this issue. For corporations that face large unfunded pension liabilities relative to the size of the company, credit rating and the impact on borrowing cost can be a major concern.
The impact on borrowing cost is another example of a financial impact that is highly asymmetric with changes in funded status. Moving from 120% to 170% funded will not improve credit ratings or reduce borrowing cost. However, moving from 120% to 70% funded will probably gain a considerable amount of unwanted attention from the PBGC, credit agencies, unsecured creditors and bond holders alike, and could have a costly impact above and beyond the increase in contribution requirements.
While the expense and credit rating are important considerations, the core economic cost is the present value of contributions less the value of the call on the revertible surplus. It is possible to reduce the present value of contributions with appropriate investment and funding policy.
Key sources of savings
A traditionally managed plan will experience five key sources of savings when switching to minimizing the present value of future contributions.
Better asset/liability matching. Designing the assets to move in tandem with the liabilities saves companies money. The reason for this savings lies in the asymmetric payoff pattern to the corporation described in the previous section.
Assume there is a plan that is 140% funded (i.e. assets are 140% of the ABO). Let us further assume that 140% is the full funding limit for that plan. In other words, as long as the plan is at least 140% funded, no contributions are required or tax deductible.
Now, suppose an asset was available that would perform in tandem with the liabilities so as to maintain the ratio at 140% regardless of what occurs with the capital markets. Let us call this hypothetical asset the "liability asset." If the plan is fully invested in this "liability asset," no contributions would be required this year or in any future year. Therefore, the present value of future contributions would be zero (ignoring PBGC premiums).
Suppose further there was an alternative asset class called the "risky asset" that had an expected return of 400 basis points a year above the "liability asset." Unfortunately, this 400 basis point premium also has a 15% standard deviation. If the plan fully invested in the "risky asset" in lieu of the "liability asset," the expected funded ratio would be well in excess of 200% after about 30 years.
However, many possible scenarios also would result in funded ratios substantially below 140%. Furthermore, even on those paths that generate large funded ratios at the end of the period, there may be some points in time where the funded ratio drops below 140%. If these points coincide with the actuarial funding valuation, the company will be called upon to make contributions. Thus, the present value of contributions will be substantially higher despite the higher expected returns. This increase in cost can only be recouped if the surplus at the end of the period can be recaptured by the company at a significant value.
In practice, it is impossible to find a "liability asset." It is however, possible to cause the assets and liabilities to move more in tandem than traditional management achieves. The "lowest risk" portfolio, as measured by present value of future contributions, is usually a combination of equities and long duration bonds. The actual mix for this "lowest risk" portfolio can vary substantially among plans and has ranged, in my experience, from a low of 20% equities to a high of 60% equities. The optimal bond duration is specific to each plan and the equity exposure selected. It usually exceeds 10 years and is much higher than any of the commonly used bond indexes.
Extending the duration of the fixed-income portfolio. One of the key ways of synchronizing assets and liabilities is to extend the duration of the fixed-income portion of the portfolio. Liabilities, like long duration bonds, are extremely interest sensitive. Long duration bonds not only reduce the riskiness of the asset/liability gap, but also provide a higher expected return. This higher expected return results in part from the move up the yield curve and in part from the move to a more convex portfolio (i.e., the portfolio gains more in value when interest rates drop than it loses when interest rates rise by an equally probable amount).
The amount of equity exposure. The equity exposure a plan takes is critical to the cost. The traditional 60% equities, 40% fixed income is very often not optimal and sometimes is not even on the efficient frontier. As indicated previously, the "lowest risk" portfolio varies from 20% equities to 60% equities; however, the "lowest risk" portfolio is usually suboptimal. The optimal amount of equity exposure also varies dramatically from plan to plan and from company to company. We have studied plans where the optimal equity exposure is no more than 30% and others where it lies closer to 90%.
There are five key factors affecting the optimal amount of equity exposure.
The "noise" in the liabilities. The efficacy of fixed income as a "liability matching asset" is directly related to the extent that liabilities look very similar to bonds. When the liabilities look a great deal like bonds, a close match can virtually eliminate risk. When the liabilities look less like bonds, a residual risk remains even with the "best" matched portfolio. In this latter case, the addition of equity increases returns more sharply than it increases risk.
Retiree liabilities are far more accurately represented by bonds than are active liabilities. Salary increases in active liabilities create "noise" (volatility in the relationship between bonds and liabilities). The greater the "noise," the greater the optimal equity exposure. Thus the proportion of retirees to actives is an important factor.
It is often possible for companies to reduce the noise level in the liabilities with careful benefit design. For instance, designing lump sums payable on the basis of market rates of interest allows them to be matched with fixed-income instruments. Generally, "noise" costs money, and it is advantageous to reduce the noise level to the extent possible.
Growth in the plan work force. A declining work force usually means retiree liabilities are growing faster than active liabilities and thus more fixed income becomes optimal. Conversely, a growing company will need more equity exposure in its plans.
The "weight" attached to surplus value. As stated previously, surplus in a pension plan has value beyond the reduction in future contributions. It may be recaptured by the company at a substantial penalty or used indirectly to reduce costs elsewhere in the company. The actual value is highly dependent on legislation and the balance of power in the struggle over who owns the pension surplus.
Because the appropriate value of surplus is unclear, it is beneficial to examine what the optimal allocation is for a range of possible surplus weights. Generally, the larger the weight, the greater the value of the surplus generated by extraordinary equity returns, and the larger the optimal equity exposure.
The funded status of the plan. In general, poorly funded plans and well-funded plans lead to high equity exposures.
For poorly funded plans, high equity exposures are necessary because the asymmetry discussed earlier virtually disappears, causing all of the upside to be valuable. For well-funded plans, there is a large cushion protecting the plan from future contribution requirements on the downside. Thus, even though the company cannot receive full value for the upside, the large cushion limits the downside risk to the extent that it makes the upside of a large equity exposure an attractive trade-off.
Plans that are neither poorly nor well funded usually will find that unless the remaining factors discussed below dominate, high equity exposures are economically unattractive.
Term structure of the company's borrowing cost. The higher the company's borrowing cost and the more steeply it rises with the term of borrowing, the more the future is discounted relative to the present. Thus equities, which save money in the long term (higher expected rate of return), but cost money in the short term (high volatility of funded status) become less attractive relative to "liability matching" fixed income.
All else being equal, higher borrowing costs result in lower optimal equity exposure.
Rebalancing rules. The fourth source of savings comes from selecting an appropriate rule for rebalancing the portfolio's optimal asset allocation as the capital markets and other events change the five factors mentioned earlier.
It is important to rebalance the portfolio to the mix that is then optimal considering the changes that have occurred. One needs to determine the sensitivity of the optimal asset allocation to each of the previously described factors influencing equity exposure. From this sensitivity analysis, one can create a rule to rebalance the portfolio. An example of such a rebalancing rule may be to add an additional 5% to the equity exposure for each 10% increase in funded ratio.
Funding policy including fund methodology and assumptions. It has long been a tenet of the actuaries and financial economists that the pace of funding is not financially relevant. The cost of the plan is the benefits paid, and funding merely transfers money from one pocket of the company to another. Once it is recognized that the corporation does not have easy access to the surplus in a pension plan, this tenet breaks down.
The present value of future contributions can be highly affected by the pace of funding. In fact, it would be advantageous for companies to add voluntary contributions so long as the cost of such additional contributions (the contribution minus the tax deduction) is less than the present value of the future contributions saved if such contributions were made.
It is often possible to reduce the present value of future contributions with appropriate changes in both long-term funding policy and annual voluntary contributions.
The economic cost of a defined benefit pension plan is the present value of future contributions. The term structure of the sponsor's borrowing cost is the appropriate discount rate. The traditional management of defined benefit pension plans (mainly the setting of investment policy in an asset-only world) results in a present value of contributions that is significantly higher than it need be.
To realize the savings, it is necessary for companies to integrate investment and funding policies. Investment policy needs to focus sharply on the nature and structure of the liabilities and needs to explore the extent to which assets can be designed to match liabilities.
In order to find the solutions that work, it is necessary to use an asset/liability/plan sponsor model that projects yield curves, equity returns, general wage increases, returns on bond portfolios of varying duration, and actuarial liabilities, etc. It also must be capable of determining contribution requirements under various scenarios, alternate funding methods and assumptions, and discounting those at the company's cost of borrowing.
In order to achieve the available savings, some significant changes may be necessary in asset allocation and funding policy. These include a significant lengthening of the duration of the fixed-income portfolio, changing the equity exposure (this could be up or down), adopting an appropriate rebalancing rule and, perhaps making changes in funding policy.