The SECURE Act, I lamented in December, had bipartisan support but languished in the Senate until it was finally passed as a part of a massive budget bill. And it will not surprise readers to know that I am not jumping on the bandwagon of “the SECURE Act will change everything” because I’m irritated that the much more crucial legislation-in-waiting, a fix for multiemployer pension plans, has stagnated.
And here’s (the/a) fundamental question:
Should multiemployer pension plans be required to fund at the conservative level dictated by the use of a corporate bond rate similar to what’s been required of single-employer plans since 2006, or is it acceptable to use a funding discount rate determined by the expected return on assets for the plan, similar to the method state and local public pension plans use?
This is, near as I can tell, headed towards being one of the sticking points in whatever behind-the-scenes negotiations are or will be taking place with respect to multiemployer plans, because lowering the rate increases liabilities and reduces reported funding levels, providing more secure benefits at a considerably higher cost.
And here’s the difficulty: there is no single correct answer to this question. There is no way to say that these folks over here are right and those folks are dummies, or irresponsible stewards of taxpayer money or cruel dismantlers of the multiemployer system.
Yes, Jeremy Gold insisted that proper actuarial practice requires the use of a risk-free discount rate to value liabilities. The use of an expected return-based rate places future generations at risk of covering the cost of past generations’ benefit accruals, and, what’s more, tempts plans to increase their on-paper funded status by claiming a higher expected return, either simply by claiming unwarranted optimism or, worse still, by investing in inappropriately-risky assets.
On the other hand, there’s a case to be made that, when a pension fund does indeed invest in diversified assets rather than in low-risk corporate bonds (which is, incidentally, a common practice outside the US), it is unnecessarily restrictive to limit benefits to what can be funded by the available contributions when using a conservative basis.
Now, longtime readers will recall that, a year ago, I spent some time analyzing the pension plans in Chicago. Since the main plan, for municipal employees, was 95% funded as recently as 2000, I wanted to understand what the cause was of the collapse in funded status since that point.
As it turns out, this plan has a number of commonalities with the types of multiemployer plans which are having so much trouble — in particular, both this plan and many multiemployer plans are paying out far more in benefits than is being paid into the plan in terms of contributions, or than current employees are accruing in benefits. In addition, both this plan and multi’s in general use the expected return as their valuation rate, and have seen drops in this rate over time.
Just for fun, here’s a comparison of the funded status for these two plans:
(For a refresher and sources, see “Understanding The Central States Pension Plan’s Tale of Woe,” “Actually, Central States’ Pension Plan Is Fully Funded,” and “What’s Worse Funded Than Teamsters’ Central States? Chicago’s Pensions.” Also note that the three-year gap for the Teamsters is simply a gap in the online data.)
Now, I had previously discussed the causes of the Chicago pension plans’ underfunding, and the fact that it’s a trivial statement to say that the city didn’t make the appropriate contributions – because, had the city made the “Actuarially Determined Contributions,” the lament would have been at the escalating increase in those contributions, and, even so, for multiple reasons, including the gradual rectification of underfunding in the ADC calculation method, that plan would have nonetheless been only 54% funded in 2018 had these contributions been made as scheduled.
But the municipal plan’s data which I had previously summarized is a useful source with which to ask the question: how much overfunding would have been necessary to protect against the drop in funded status in a period of weak asset returns relative to the expected return built into the valuation? This is, after all, what we’re looking at: cumulative years of returns below expectations drive a drop in the funded status, and depending on the timing of those returns, recovery from these low returns can be very difficult indeed.
Here are some hypothetical figures, assuming that the Chicago municipal plan’s liability developed as was the case historically, but that the plan contributed amounts based on the total normal cost (that is, new benefit accruals) rather than an actuary’s formula or the haphazard actual contributions.
If the plan had started this time period at 140% funded rather than 95% funded, and had continuously contributed 140% of the normal cost each year, it would have emerged from the ups and downs of this period at 70% funded.
If the plan had started at 150% funded, and contributed with a 150% overfunding target, it would have been 82% funded at year end 2017.
With a 160% overfunding level, the plan would have been 95% funded.
It would have required overfunding at a level of 164% to have been funded at 100% at the end of this period.
And as it turns out, this is as much about the argument about different discount rates as it is conservativism in funding, since the difference between liabilities using a valuation discount rate based on expected asset returns, and based on corporate bonds, is about this level anyway — that is, obliging a plan to fund its liabilities based on the more conservative rate provides this protection against low returns.
But that still doesn’t get us to answers.
Having a funding cushion isn’t the only way to deal with asset losses. If plan sponsors were ongoing financially healthy entities, these plans would be able to temporarily increase plan sponsor contributions. This requires that plan sponsors have the capacity to do so; that’s not so in the case of multiemployer plans with large numbers of “orphans,” that is, participants of bankrupt employers. And many participants would prefer to take “their share” of the ongoing plan contributions rather than leaving a significant overfunding level for the next generation of participants — even if they’d have to agree to the risk of benefit cuts as needed, in return. What’s more, while we can look at past history — asset returns for the past decade or generation or century — there is no surefire way to know a plan’s prediction of long-term asset returns will actually materialize, and whether shortfalls or surpluses are short-term or a “new normal.”
But beyond all these issues, there’s a larger practical one: plans which have made contributions up to now in good faith, based on the existing methodology, would be at risk of being hammered if, however much they are now deemed satisfactorily funded under current rules, they are required to make up newly-calculated funding deficits which might be quite substantial.
So, with apologies for the long preface to this suggestion, here’s my proposal: grandfathering.
Allow plans to continue using the existing funding methodology for existing plan accruals, as well as paying the existing PBGC premiums for the existing guarantee levels.
But for new plan accruals, offer plans a choice: more demanding funding requirements paired with higher PBGC premiums, or preservation of existing requirements with new methods of applying benefit cuts routinely and systematically for plans which drop below reasonable thresholds.
This is not going to make everyone happy. But I think it has a shot at being sufficiently acceptable to enough people to work.
What do you think? Share your thoughts at JaneTheActuary.com!