Sunday, August 23, 2015
Orphan pensions and their relatives
Let's first note that there is a "political" problem when people doing the same work get different benefits, especially if one group gets a tangibly lesser benefit than another. This problem arises even if the two groups are in some sense in the same overall plan but one group has a less advantageous formula. There is a lesser degree of solidarity that results which undermines support for the system.
A strictly "orphan" plan is one in which the plan is totally closed to new hires and thus has obligations to a dwindling proportion of active employees and retirees. If there is any unfunded liability in such a plan, there is no new generation available to provide additional funding. As the Legislative Analyst's Office (LAO) has pointed out, in such a case the trustees of the plan are under increasing pressure to be sure that there is enough money in the plan to pay off its liabilities. So there is both a political issue and a financing issue. There is no indefinite horizon in which to make up for any unanticipated declines in asset values. The only way to be sure all obligations are met is to invest the plan funds in relatively riskless - and thus low return - assets. But as returns go down, the cost of meeting plan obligations goes up. And there is no one new to cover the cost.
Even if a plan isn't strictly an "orphan" in the sense that there are groups of employees, some under less advantageous formulas than others but all in the same umbrella plan, there will be tensions if it appears that less advantaged groups are being "taxed" to make up for any unfunded liability that might be attributed to more advantaged groups.
Finally, let's note that creating a hybrid plan - which is what is likely to emerge from the Committee of Two deal and task force - creates orphan-type problems even if the task force somehow creates a defined contribution (DC) supplement to the now-capped defined benefit (DB) plan the deal envisions so that the sum of the DB and DC plan are by some measure equivalent to the former DB-alone plan. The essence of a DC plan is that the employer puts cash into an account which the recipient then invests with some discretion. Some investments will turn out better than others. There is an iron law that, by definition, 50% are always below the median!! So even a carefully sculpted DC supplement, that for some "typical" retiree makes up for the Committee of Two cap, there will be after-the-fact situations for individual retirees whose investment choices didn't work out well that will lead them to a sense that they didn't get what was promised. You could in theory eliminate all investment choice so that everyone gets the same guaranteed return - a plan that is technically called a "cash balance" (CB) pension. But you then have to guarantee a rate of return that precisely makes up for the cap.* The problem is that a CB plan inherently has the potential to have an unfunded liability since you are guaranteeing a return in an uncertain world. What if your plan assets turn out to earn less than the guarantee? And would the governor/Dept. of Finance even accept a cash balance plan as meeting the Committee of Two criterion? Or would there be insistence on a DC plan? No one knows.
*We are skimming over yet another problem. What you get from a traditional DB plan depends on the age at which you retire, your years of service, and your final earnings. There is no simple linear formula under DB. DB plans tend to favor long service and older workers disproportionately. A CB plan is inherently a more linear (really log linear) program. (If the guaranteed return is, say, 5% per annum, every year a dollar in your CB account turns into $1.05 the following year.) Thus, while you might make a DB+CB plan equivalent to a DB-alone plan for an employee with a certain, specified age-plus-service combination, you would be hard pressed to come up with a DB+CB combo plan that was equivalent for every retiree. To put the matter in Shakespearean terms, in the end it's either DB, or not DB. And the Committee of Two has decreed that it's not DB.