During the bull market of the late '90's, a plan sponsor with a pension fund covering union employees had ridden the indices upward, amassing assets with a traditional 60% equity and 40% fixed income allocation. Then came 2002, when the S&P dropped 22% for the year. The sponsor's plan's assets plummeted to $450 million, down from $650 million, and its funding ratio of assets over liabilities dropped to 80%, down from 125%. The reversal was a rude awakening for the plan sponsor.
"When the market dropped and our assets dropped by about 25%, it was a sick feeling and a harrowing experience," said one trustee of the plan. "The market went down and we went with all the other pension funds."
The loss would be next to impossible to make up through contributions. The pension fund is a nest egg for roughly 900 employees and retirees of the plan sponsor who belong to a national union. The locals face declining membership. With technological advances, fewer employees are needed to run the technical equipment. Because contribution amounts are based on the number of members, this meant that payments into the pension fund were decreasing over the long run. Because of the relatively small size of the active population, annual contributions were less than 1% of overall assets in the plan. In 2002, the population was aging fast: Over the next 15 to 20 years, the bulk of the members would retire. The market decline in 2002 showed the trustees in painful detail how the fund was exposed to the risks inherent in the equity markets. It meant the fund would eventually be required to replenish the fund with contributions, as stipulated by pension funding rules at the time.
The plan sponsor turned to Milliman and a pension fund investment consultant to find out what were the primary risks to the plan and set a new course for minimizing them. Milliman used FutureCost, a powerful modeling tool that projects the assets and liabilities of a pension plan over the next 10 years or longer.
For each year, Milliman used FutureCost to produce full actuarial valuation funding results, including plan contributions and plan funded status. The model can be run with one or more scenarios for future liability discount rates, inflation, and investment returns. It also can be used to generate a full set of stochastic results that provide the sponsor with the full range of possible outcomes for each measure, and their estimated probability. Using data from three years prior to the 2002 market decline, Milliman used FutureCost to backtest how contributions would have been affected by the drop. The predictions matched reality, giving the plan sponsor confidence in the model and a better sense of its risk.
For the coming years, Milliman then used FutureCost to show what contributions would be, from allocations ranging from 0% equity and 100% fixed income, to 100% equity and 0% fixed income. (In these scenarios, the fixed component was assumed to be in bonds of duration that matched the liabilities of the plan.) After careful review, the plan sponsor elected to implement a liability-driven investment (LDI) strategy. The bulk of assets were moved into long-duration bonds because their performance more closely matched the overall liabilities of the plan (i.e. these assets tended to increase in value at the same time liabilities were increasing). The ultimate goal was to have equity only account for roughly 20% to 30% of assets. The plan sponsor started slowly unwinding equity positions in 2004, and continued reducing the allocation over the next few years. By mid-2007, it had reached its lower target of equity exposure. Yet, it also had the good fortune to catch a good part of the recovery of the equity market mid-decade, allowing it to reach fully funded status by 2006. Maintaining and adjusting the LDI strategy going forward would help insulate the fund in the next market downturn.
The switch to the new strategy would prove timely. With the stock market collapse in 2008, the plan sponsor had inoculated itself against big drops not only in plan assets, but also increases in funding requirements. While the S&P 500 dropped 37% for the year, the plan's assets declined by only 5% and the funded level fell by less than 2% at a time when other plans' funding levels assets were falling fast.
"We were reading in the newspapers how the other funds are dropping by 25%, just like we had a few years earlier," the trustee says. "But this time we held up because the LDI portion of our portfolio kept our assets in line."
FutureCost helped show early on that the potential for interest rate fluctuations was among the plan sponsor's biggest risks. That would become all the more important under the Pension Protection Act of 2006, which stipulated plan sponsors starting in 2008 had to use current market bond rates for its interest rate assumption, and no longer use a rate based on the expected rate of return on plan assets. A falling rate would mean greater liabilities, while a rising rate would mean reduced liabilities. Plans whose ratio of assets to liabilities fell below 80% would be required to boost contributions, essentially "marking to market" its funding deficit.
That wouldn’t be possible for the plan sponsor, given its aging and dwindling member base. The primary risk the plan faced was not that assets would deviate from their benchmarks but that the funded status would decrease. FutureCost gave the plan sponsor a way to look into the future to help it make the right decisions with respect to making sure liabilities were well matched for the long run, while leaving some room for modest market exposure in an equity component.