For Pension Funds, a Smarter Application of the 'Risk-Free Discount Rate'
Financial economics should apply to labor relations, not accounting.
For two years, a group of actuaries known as "financial economists" have been urging the governmental accounting community to change the way public pension funds value pension liabilities and calculate employer costs. They argue that public employees receive a risk-free retirement benefit that is guaranteed to them regardless of the investment performance of the trust fund, so the calculation of the liabilities should be based on a risk-free interest rate. Traditionally, public pension funds use the expected rate of return on the investment portfolio to calculate these numbers, which results in lower actuarial liabilities and employer contributions. So far, the financial economists have not persuaded the Governmental Accounting Standards Board (GASB) or the public pension establishment of their logic.
As I have written before and testified before the GASB, I see little value in this idea as a basis for governmental accounting. If pension funds are consistently invested in diversified — albeit riskier — portfolios over decades, the odds are that the traditional approach to valuing liabilities will be more accurate than the risk-free rate. I have also supported the GASB view that liabilities should be discounted by the employers' borrowing rate (or a standardized taxable muni bond index) when the benefit plan's assets are insufficient to pay the future liabilities.
But my latest work on pension reform and retirement plan redesign has convinced me that the financial economists have a valid point that should not be lost in the national effort to redesign these retirement plans on sustainable terms. Instead of focusing on the employer's side of the balance sheet and payroll stub, we should apply their logic to the calculation of employee contributions to these retirement plans.
Public employees in pension plans and defined benefit OPEB plans enjoy a guaranteed benefit that is risk-free to them. The price that we ascribe to them for their fair share of these costs is too cheap. We use the expected investment returns of the risky portfolio that trustees approve in their efforts to optimize long-term returns on public capital. That's a mistake — it's heads, they win; tails, taxpayers lose.
There is a better approach.
Employees should pay half. First, the optimal public policy is a 50-50 match of employer and employee contributions to retirement plans. When employees bear half the cost of a retirement benefit, they have equal skin in the game. There is no free lunch. Experience has shown that employees are far more responsible about retirement benefits when they pay half the cost of those benefits. From a labor-market perspective, there are very few employers outside of government that match employee contributions with equal shares. The "attract and retain" argument for healthy retirement benefits that we often hear is just a smokescreen when employees pay less than half the costs. In today's competitive labor markets, they really should be paying two-thirds of the costs of all retirement benefits to achieve parity with the private sector which more typically matches only 50 percent of employee contributions. However, that takes the argument to unnecessary extremes.
The employees' share of costs should be based on what is called "normal cost" — the current actuarial cost of this year's earned benefit. Employees don't pay toward any unfunded liabilities which result from investment underperformance or actuarial experience that are beyond their control. In most instances, employers must properly bear those risks as the inherent cost of providing a defined benefit plan. (I've previously written about one exception involving retroactive benefits increases, which was the forerunner to my new solution below.)
So who pays the unfunded liabilities? When investment performance fails to meet actuarial expectations, guess who pays the costs? Well, you can be sure that it isn't the employees. They get what I have previously called the "public pension fund straddle option" in which employees typically get a "call option" with bigger benefits when investments exceed the actuarial assumption, and a "put option" to force their employer to pay the losses when prices head south. Until now, I had considered this an inescapable fact of life for public employers. But now, I have found a way to level the playing field to protect American taxpayers: Require employees to pay their share using the full price for the real cost of their risk-free benefit, not an artificial cost.
Calculating the employee share of a risk-free benefit. When public employers set the employee contribution rates, whether by statute, ordinance or collective bargaining, the employee's half of the costs should be calculated by using a "risk free" discount rate for their share. They are receiving a guaranteed annuity (or even better, an escalating retiree health care benefit) regardless of what happens to the retirement fund's investments. So why aren't we discounting their (50 percent) share of the plan's cost using a risk-free rate?
Financial economists have struggled to agree on what constitutes a risk-free rate. Some would advocate a U.S. treasury bond. That brings objections from critics who point to the volatile nature of government bond yields, especially during periods like the recent Great Recession when yields fell to 2 percent as investors fled riskier assets for the safety of Treasury bonds. Employer costs could double or triple overnight in a financial panic, for no good reason.
My solution is to blend and average the rates on relevant U.S. Treasury bonds and taxable municipal bonds over the past five years. This blended index will be sufficiently stable to avoid wild swings in contribution rates and has a conceptual validity based on the employer's cost of borrowing and the risk-free nature of government bonds. (Notice the alignment with GASB's thinking.) In today's market, that would require a discount rate around 5 percent for new hires with 30 years of service ahead of them, and 4 percent for incumbent employees with shorter remaining career lives (averaging 12 to 15 years).
The fair cost to employees. Using a lower, risk-free discount rate for the employees will fairly assess their share of the guaranteed benefits they will receive. Needless to say, this will increase required employee contributions. Employees and unions won't like this, but their choice should be clear: If they want stock market returns on their contributions, they should invest instead in a defined contribution plan where they take the investment risks that go along with those higher investment performance expectations. No longer should employees be able to have their cake and eat it too, which is what the traditional public pension plans have unwittingly and naively allowed them to enjoy.
Two groups of employees are best suited for this new plan design strategy: new employees and recipients of retroactive benefits increases. The new employees should be given an option to contribute instead to a defined contribution plan, so there is no hardship to anybody who wants to earn a higher rate of return and take the commensurate risks. For incumbent employees, the most logical place to introduce this concept is for those employees who were awarded or bargained for non-contributory retroactive pension increases. Such employees have collected windfall benefits for which they paid little or nothing, on the premise that investment results would pay for their fuel-injected benefits increases. Now that it's obvious that political claims of "free money from the financial heavens" were illusory, it is only fair that those employees pay the risk-free rate for the guaranteed benefits they received.
No change for employers. This approach does not change the rules of the game for employers. If they continue to fund their pension plans using the expected return on risky investments, that is their choice. Employers bear all the risks of those portfolio decisions in the form of unfunded liabilities that must then be amortized if the markets fail to produce the expected returns. The employers (and thus the taxpayers) who bear the investment risk should enjoy all the benefit of the riskier portfolio's expected returns, not the employees who take no risk and get returns from risky investments. If investment portfolios produce a substantial investment surplus, the portfolio can then be "immunized" by investing in bonds to eliminate the equity risks and assure taxpayers that their obligations have been fulfilled. At that point, the employer and employee discount rates will converge to the risk-free rate.
If GASB has its way and the discount rate is blended to include the employer's borrowing cost index, that issue should reside between the accountants and the employers. But at least the employees' money will be contributed at an appropriate rate, and trustees can then decide whether they want to accept higher risks when they invest those funds.
Guidance for elected policymakers. When elected officials advocate pension and retirement plan reform, their first salvos should be to require employees to pay half. Then pension policymakers should use relevant government bond yields as the discounting rate in calculating the employees' contributions. It will take a while for members of the public retirement community to get their heads around this concept — they have instinctively resisted the financial economists' arguments as the basis for pension fund accounting. But those who think this through will eventually realize that a federalist blend of federal and taxable municipal government bond discount rates can be used prudently to gain a better design of and fund all defined benefit retirement plans that pay a guaranteed benefit.
The financial economists may not feel completely vindicated, but the world would be a better place if their analysis of the risk-free character of the pension promise results in proper pricing.
Join the Discussion
After you comment, click Post. You can enter an anonymous Display Name or connect to a social profile.