Defined Benefit Pension Plan: 'Life Support' Over 'Full Recovery'

Speaking solely as a remote, disinterested observer and corporate benefits professional… I have been thinking about IBM’s recent change to stop employer contributions to the 401(k) plan in favor of restarting accruals in the defined benefit (DB) pension plan’s cash balance formula.

Lots of commentators in the benefits press suggest this might trigger a renewal of DB plans. I don’t think so. But I am jaded by my 31+ years in plan sponsor roles at four different Fortune 500 employers where each offered a DB pension.

Myself, I took my DB plans through a series of changes, including but not limited to:

• Inheriting previously approved changes effective 12/31/85 featuring improvements such as:

Subsidized early retirement,
Guaranteed 3% per year COLA, and
“Spiking” of covered compensation at separation, and

• As predicted by our plan actuaries immediately in August 1985, before the change, we then spent the next 25 years scaling back the DB plan, by:

Changing the 3-year Final Average Pay (FAP) formula to a 5-year FAP formula,
Reducing the participation service maximum from 40 to 35 years, while concurrently reducing the formula’s excess factor from .75 to .5 per year of participation service,
Freezing and wearing away:
- The early retirement subsidy, and
- The post-retirement COLA, and
Ultimately, wearing away the FAP formula, applying a cash balance formula for all hired after 2001, and later, applying the cash balance formula for all future participation service regardless of hire date.

Why all the changes? Why the reduction in accruals and curtailment of favorable features?

Why? No longer a pension for “retirees”!

Until the Tax Reform Act of 1986, most DB pensions vested at 10 years of service. TRA ’86 changed that to 5 years. Then, after the Pension Protection Act of 2006, vesting for plans with hybrid or cash balance formulas was reduced to 3 years.

Because median tenure of American workers has been less than 5 years for the past seven decades, before TRA ’86, and even later, until PPA ’06, most workers never vested in their DB plan. After PPA ’06, DB plans no longer primarily benefitted older, long-service workers who stopped work and commenced payout. Instead, our plan primarily benefitted term vested workers who separated many years before retirement.

In fact, when I took a longitudinal look at our population in the late 1990s, I found that less than 1 in 20 individuals who we hired would remain with the firm until retirement—arbitrarily defined as separating after reaching age 62, completing 25 or more years of service, full stop of employment, and immediate commencement of pension benefits as lifetime income.

Why? Underappreciated by workers, relative to costs!

Back in the 1980s and 1990s, the budget for DB plan expense varied between 5%-7% of payroll. However, surveys, such as one we conducted in 1986, showed that:

• ~20% of respondents (mostly short service, younger workers, including many who were accruing a benefit) didn’t know we had a DB pension, and

• < 50% of respondents had a favorable perception of our DB pension (most did not expect to retire from our organization).

When we shared those survey results with executive leadership, I remember an irritated Chief Legal Officer who quipped, “Well, maybe we should survey retirees, instead.”

Why? Mark to Market (MTM), cost variability!

Predictability is prized! Every August, I would be called upon to assist the Office of Finance in predicting benefits expenses to incorporate in the budget for the coming year. Unfortunately, changes in interest rates and valuations of equity investments could reduce the predictability and year-to-year comparability of corporate earnings. Variability might also prompt changes in pension asset allocations.

Why? Not an “attraction, retention, engagement” lever!

Perhaps most importantly, from a Human Resources perspective, not only did the DB pension offer little in immediate value, but it was also negatively impacted by hyperbolic discounting.

When we conducted engagement studies coupled with conjoint analysis to tease out participant preferences, worker assessments of the value of the DB pension suffered. Unbelievably, the change that was most likely to prompt improved associate engagement was the award of an additional three days of paid time off!

Here is a long ago chart I regularly incorporated in Human Resources/Benefits presentations:

As experienced pension professionals know, a non-contributory, final average pay formula DB pension plan with early retirement subsidies and automatic post-retirement COLA lines up solidly (and solely) on the left—a benefit that is a promise, delivered in the future, and importantly, contingent on the employer maintaining the plan, continuing the formula, remaining in business and of course, contingent on a worker’s continued employment.

That is especially true where the plan features a final average pay formula and early retirement subsidies that only apply if separation occurs after attaining a qualifying age.

Who benefits from the change at IBM?

The change had some positives for IBM, perhaps not so much for IBM workers.

It leverages the DB plan surplus. The plan was frozen at year-end 2007. This might allow for a contribution “holiday” for an extended period, potentially 14 or more years:

• The 5% of pay credits will be less for most workers compared to employer contributions to the 401k plan,

• For the first three years, the investment credits will be 6%; thereafter, investment credits will be based on 10-year Treasury returns, which for the past 20 years, have averaged less than 3%, which, in turn, may extend the contribution holiday depending on the spread between actual returns and the investment credits.

• The 1% salary increase (replacing the 1% employer 401k contribution) will likely reduce organization expense, over time, because:

• The pay increase is probably a transition, limited to current workers, and

• It is not clear how this salary change impacts future adjustments/merit budgets,

• Less, or offset by:

The increased PBGC premiums, $101 per year per participant (2024) for each new participant in the plan (those hired since 2008), and
Increased employment taxes on the 1% salary increase.

The length of the contribution holiday will depend in part on IBM’s investment strategy for the DB plan assets. Most frozen DB plans consistently reduce investment risk as funded status improves. IBM may have used such a strategy during periods when pension accruals were frozen. However, this change might prompt IBM to “re-risk” investments—increasing the allocation to return-seeking assets to potentially reduce future contributions.

In terms of the organization’s interest in “re-invigorating” the DB pension plan, I think it noteworthy to reconfirm that in 2022, IBM engineered a significant Pension Risk Transfer. Coupled with that action and a guesstimate of the substantial period without contributions, the long-term future of the DB pension is unpredictable.

From a worker’s perspective, current and future, I do not believe the changes are an improvement in retirement benefits for most workers^[i], because of:

• The impact on employer contributions, even adjusted for the salary change is not likely to be positive for most current workers, and almost all future hires,

• The impact on investment earnings,

• The potential for reduced retirement savings where 401k “opt out” rates increase – even if IBM maintains automatic enrollment and escalation features,

• Elimination of the employer matching contribution which reduces the incentive to defer compensation to the 401k plan,

• The reduced access to tax-favored plan loans (for anyone whose vested 401k account balance is less than $100,000).

• All of the above negatives are likely more than offset for a small minority of workers, the 3% who were not contributing to the 401(k) plan.

With respect to access to guaranteed lifetime income, my understanding is that IBM embraced the Heuler platform decades ago. Further, some male retirees may be able to achieve a better outcome via the Heuler platform when compared to the SLA and 50% J&S options mandated for the DB plan.

IBM median tenure is ~8 years (vs other American workers, consistently < 5 years for the past 7 decades). So, more than half, maybe 75+% of all IBM employees have never participated in the DB plan. The median age is ~38. So, even if DB accruals continue indefinitely, they will only apply to a portion of IBM service.

Will IBM’s action trigger a ‘rebirth‘ of DB Pension Plans?

Define “rebirth.” How about a return to 1985—when 90% of the Fortune 100 had DB pensions (89 Traditional, 1 Hybrid/Cash Balance)^[ii] and where many featured non-contributory, final average pay formulas, with early retirement subsidies and post-retirement COLAs. No one is predicting that.

Interestingly, it was only 14 years ago that we had a “rebirth” of the 401(k)—after observers like Time Magazine declared it all but dead.^[iii]

Anyway, I think it much more likely that frozen DB plans with pension surplus will continue to engage in de-risking before they will restart accruals. IBM wasn’t the first organization to reopen their frozen DB. We may see a few organizations with frozen DB plans with a funding surplus follow IBM’s lead. But no one is predicting employers who don’t currently offer a DB plan will start one based on IBMs actions, let alone return to 1985.

• IBM 401(k) Replacement Elicits Questions from Retirement Leaders

• IBM Replacing 401(k) Match with 5% ‘Retirement Benefit Account’ Contribution

^[i] N. Zuss, Analysts Agree: IBM Pension Thaw Benefits Both Participants and Plan Sponsor: IBM’s decision to unfreeze its pension is a “strong endorsement” of plan sponsors operating parallel defined contribution and defined benefit retirement plans. 11/20/23, Plansponsor.com, Accessed 1/13/24 at: https://www.plansponsor.com/lysts-agree-ibm-pension-thaw-benefits-both-participants-and-plan-sponsor/

^[ii] WTW (then towers Watson) More Fortune 100 Companies Offering Account-Based Retirement Plans to New Salaried Employees, Towers Watson Analysis Finds Shift Away From Traditional Pension Plans Continues, 5/26/10, Accessed 1/6/24 at: https://www.prnewswire.com/news-releases/more-fortune-100-companies-offering-account-based-retirement-plans-to-new-salaried-employees-towers-watson-analysis-finds-shift-away-from-traditional-pension-plans-continues-94915504.html

^[iii] J. Towarnicky, Is it Time Yet? 11/17/19. “… The author asserted that America would be better off if we “retired” the 401(k). He favored re-introducing defined benefit pension plans. Some continue to favor such actions. …”

Accessed 1/7/24 at: https://www.psca.org/news/blog/it-time-yet