Muddled Thinking About Defined Benefit Pension Plans – Part 1

Private Sector Pensions: The Strikes at Boeing and U.S. Automakers, IBM's Reopened DB Plan, and Cash Balance Plans

Photo by Sven Piper on Unsplash

Everybody Seems to Want a Defined Benefit Pension

On November 4, Boeing machinists approved a proposal to end a seven-week strike after two failed votes, the latest on October 23. The strike is estimated to have cost Boeing over $5.5 billion. The new contract has retirement provisions identical to those in the proposal that was rejected on October 23, including an unconditional employer contribution of 4% to the 401(k) defined contribution (DC) plan, along with a 100% match of the first 8% of pay that employees contribute to the plan - a total potential employer contribution of 12% of pay. Other components of pay were increased somewhat from the October 23 proposal.

Why did the union reject the October 23 proposal? Many articles attributed it to the failure to bring back the traditional defined benefit (DB) pension that was frozen in 2014.  

Bringing back DB pensions was also an unattained goal of the UAW in its strike negotiations last November with the U.S. “big 3” automakers (GM, Ford, Stellantis). And IBM did replace its 401(k) contributions with credits to a cash balance plan, which, legally, is a defined benefit plan, starting in 2024.

Are DB pensions making a comeback?

Why Do Unions (and Others) Like DB Pensions So Much?

One possibility is that people want the peace of mind of not having to think about how much to save, how to invest their retirement pot of money, and how best to draw it down once they retire. They like knowing they have a guaranteed income stream for life when they stop working: “’The 401(k) program is gambling on our retirement,’ said Kat Kinckiner, a union steward who has worked at Boeing for close to 15 years.”  

Another possibility, not mutually exclusive from the first, is that people think there is something magical about DB that makes it worth more; if you put $1,000 in a DB plan, it is somehow worth more than the same $1,000 contributed to a DC plan. While it sounds ridiculous when stated like that (and it is), the argument has been advanced by public pension advocates.

Boeing union members and officers recognize that pensions are just one component of total compensation, which means there is some amount of other compensation that would be acceptable in lieu of the traditional pension. Ms. Kinckiner went on to say, “To take away the pension and not compensate us enough to cover it? It’s just another takeaway, and there’s nothing unreasonable about wanting it back.” Jon Holden, the president of the union local that represents most of the striking employees, made the same point: “If they're not willing to give it [the pension], we've got to get something that replaces it…So it does come down to wages. It does come down to the 401(k) plan.”

But how much of other types of compensation would be enough?

DB Pension Plans: Volatile Cost and Little Value for Young Participants

There are two important implications of pension math:

1. Other forms of compensation being equal, sponsoring a traditional DB plan means older employees earn more than younger employees; and

2. The annual cost of traditional DB pensions, and thus total compensation cost, is sensitive to market interest rates. Lower market interest rates means higher cost, and vice versa.

Regarding #1: The value of $1,000 per year guaranteed for life at age 62 (for example) is worth much more at age 60 than at age 40. As of this writing, the Treasury yield curve is relatively flat, with yields around 4.25%. This means that the cost to fully fund and hedge a $1,000 per year pension accrual payable at age 62 – ignoring PBGC premiums, administrative costs, etc. – varies by age roughly as follows (based on a standard blue collar mortality table and 4.25% assumed interest):

For someone currently 60 years old, a pension of any given amount starting at age 62 costs over twice what the same age-62 pension costs at age 40 and over four times what it costs at age 25. Other benefits cost more for older employees than younger employees, one example being medical insurance, so this may be fine.

But employers and unions should understand the implications. Under the new contract, Boeing will contribute up to 12% of pay into members’ 401(k) accounts. If Boeing had instead given the union a budget of 12% of pay to spend on retirement benefits in any way they chose, would the union have decided to skew value delivery this much toward older union members? Would younger union members support that?

Regarding #2, in a traditional DB pension plan, there is no avoiding the sensitivity of annual compensation costs to market interest rates. All else equal, in lower-interest-rate market regimes, the value of current annual DB pension accruals, and thus total compensation, is higher, and vice versa. Although higher current market interest rates make DB plans relatively more attractive than a few years ago, the sensitivity of annual compensation costs to market interest rates is a financial risk that makes sponsoring DB plans less attractive relative to DC plans.

But Criticisms of Traditional DB Plans are Often Exaggerated

Some criticisms of traditional DB plans are oversimplifications. One example is that traditional DB plans aren’t “portable.” Conventional DB plans can - and many do - allow participants to elect a lump sum option on termination of employment. Those lump sums, like those received from DC plans, can be rolled over into other employers’ DC plans or IRAs.

Another overgeneralization is that traditional DB plans cost more than DC plans. Relative cost depends on the respective level of benefits and market interest rates. The potential 12% of pay available in the new Boeing DC plan would buy a reasonable DB plan.

Finally, it is often said that people who work many jobs during their careers are better off with DC. The math illustrated above demonstrates how the ideal career path to maximize retirement wealth is to work for employers with DC plans early in one’s career and with those that sponsor DB plans (if there are any left¹) toward the end. The flip side is that an employer switching from DB to DC (or to cash balance) can hurt mid-career employees who received little value from the DB plan in earlier years and will miss out on the valuable DB accruals they were about to start receiving. This is why, during many such transitions, mid-career and older employees have been grandfathered into the traditional DB formula.

But whatever the tropes, the two implications of traditional DB pension math described above are significant and should be part of the DB-vs.-DC conversation. 

IBM Doesn’t Signal an Incipient DB Renaissance

IBM just reopened its DB plan. Does that signal a resurgence in DB plans? Are “hybrids” like IBM’s cash balance plan the right compromise between DB and DC?

Until January 2024, IBM was making 401(k) contributions of 1% of pay unconditionally plus a 100% match of employee contributions of up to 5% of pay – much less than the 4% / 8% respective figures in the new Boeing contract.

Starting this year, IBM’s 401(k) contributions are being discontinued. They have been replaced by a one-time 1% increase in pay plus a 5% “contribution” (really a notional pay credit) to a cash balance account in IBM’s existing DB plan.

DB plans with notional accounts that look like DC accounts are called cash balance plans. From a participant’s point of view, the plan seems like a 401(k) without the ability to select investments but with an option to purchase an income annuity at retirement. In IBM’s case, the notional account earns interest at 6% for three years, the 10-year Treasury rate with a 3% floor for 7 years (through 2033), and the 10-year Treasury rate with no floor thereafter.

Because the plan is technically, legally, a defined benefit plan under ERISA and the Internal Revenue Code, many point to IBM’s move as the possible beginning of a DB renaissance. In the context of the Boeing strike, some retirement experts have been quoted as saying that a cash balance plan would have been a reasonable compromise, a way to “meet in the middle”.

But that’s arguing for cutting the baby in half:

• In IBM’s case, workers who were not contributing enough to the 401(k) to receive the 5% match, presumably not that many, will be better off with the cash balance plan because they will now get the 5% credit and interest thereon. For most workers, however, a cash balance plan is worse than a 401(k) plan with the same “contribution.” In a 401(k) plan, participants can (almost universally) invest consistently with their own risk preferences, unlike a cash balance account, the returns on which are one-size-fits-all.
  

• Although the balance at retirement in a cash balance plan must, by law, be convertible to a lifetime annuity, the amount of potential lifetime income for a full-career employee in a cash balance plan with the same “contributions” as an alternative 401(k) plan is likely to be less than what could have resulted had the employee purchased an income annuity with their 401(k) balance at retirement.
  

• IBM is surely doing this switcheroo, it thinks, to save money. It has a large surplus in its DB plan, which still exists to pay benefits earned in the past, that it can only withdraw by terminating the plan and paying a confiscatory excise tax on the overfunding. By replacing its 5% 401(k) match with a cash balance credit in its DB plan, IBM realizes significant ongoing cash savings and utilizes the full value of the surplus.
  
  Also, IBM may be able to recognize less than 5% of pay as a compensation cost for corporate net earnings calculations, even as it credits cash balance accounts by 5% of pay. The “service cost” component of GAAP pension expense, the amount recognized as a compensation cost, is calculated by projecting the 5% of pay “contribution” forward to retirement at IBM’s assumed cash balance interest crediting rate and discounting the projected amount back to the present based on a corporate Aa bond yield curve. I would guess that this crediting-discounting whipsaw will result in a service cost of somewhat less than 5% of pay, but there are some moving parts like account interest crediting floors and the shape of the Aa discount yield curve, that make it difficult to say for sure. However, I expect this accounting was a consideration in IBM’s decision.
  

• Unlike traditional pension plans, the liability - i.e., the cumulative value of past service costs - of cash balance plans with Treasury-based interest crediting is challenging to hedge, which makes these cash balance plans riskier to employers than meets the eye.
  

• While cash balance plans do not have the same value delivery skew (by age) as traditional DB plans, and the annual cost is not as sensitive to market interest rates, DC plans are similar to cash balance plans in these regards.
  

• Sponsoring DB plans entails costs that DC plans do not, such as PBGC premiums and actuarial valuations.

The smart bet: IBM reverts to 401(k) when the DB surplus runs out.

The irony of all the DB-is-coming-back hoopla that resulted from IBM’s DB plan reopening is that when IBM transitioned to a cash balance plan in 1999, it was considered by many to be deceitful, and it sparked lawsuits and acrimony, leading to the creation of an employee group that was grandfathered into the prior DB formula. Ellen Schultz, a reporter for the Wall Street Journal, cited the trend (at the time) of employers transitioning from traditional DB plans to cash balance plans as one element of what she called a “Retirement Heist.”

The incriminating IBM pay credit in 1999 was 5%, and the interest credit was the 1-year Treasury rate plus 1%, all pretty similar to the new 2024 plan. What was considered plunder before and lamented as indicating the demise of DB pensions is now viewed by some as the hopeful sign of a DB comeback.

So, Are We Witnessing An Incipient DB Comeback in the Private Sector?

My bet is no. Traditional DB pension plans will likely remain highly unattractive to plan sponsors and many of their employees, and cash balance plans offer few advantages relative to ordinary DC plans.

As much as the ongoing strike was costing Boeing before the recent settlement – an estimated $1 billion per month plus the risk of their credit rating being downgraded to junk – Boeing was steadfast in refusing to entertain the notion of reinstating its traditional DB pension plan.

We may see other employers do what IBM is doing, replacing DC contributions with cash balance accruals in existing overfunded DB plans, at least temporarily. However, once the surpluses are gone, the DC plan will look better again. In any case, employees would almost always be better off with the DC.

Most new DB plans will likely be adopted by successful professionals (doctors, lawyers, etc.) and small business owners looking to maximize tax deferrals.

If DB makes a comeback more broadly, the most likely scenario is that it would be in a more esoteric form, like market-return cash balance plans, where account balances earn or lose in parallel with plan assets, or variable defined benefit plans, where the promised annuity varies with plan asset performance over time, both of which shift investment risk to employees, like in DC plans. If annuities are paid in these types of DB plans instead of lump sums, participants would benefit from longevity pooling and receive lifetime income, though the amount will not be guaranteed.

But it’s difficult to see the advantage of these esoteric types of DB plans over pure DC, where participants can tailor their investment risks to their own preferences over time and where lifetime income can be provided by insured income annuities purchased at retirement with DC plan balances, or, to dream big, by longevity pooling through a tontine.

Tontines are too deep a topic to discuss at length here. Still, if the lifetime income aspect is what makes DB plans attractive to unions, maybe it’s possible, even currently, for unions to set up a tontine for its members when they retire. Retirees could fund retirement income guaranteed to last for life (though the amount may vary over time) with rollover IRAs from DC plan balances. No DB plan needed.

1. DB plans are predominant in state and local plans, so transitioning to a government job at the end of one’s career could work.