Reforming the U.S. 401(k) System

In his two most recent columns, my colleague John Rekenthaler outlined why he thinks the U.S. 401(k) system has failed and how to replace it with a national, universal retirement system.

In sum, he argues that a solely employer-based approach has made our system inequitable because many people, through no fault of their own, do not have access to a defined-contribution retirement plan, like a 401(k), at work. Without a workplace plan, these unlucky workers have few options to meaningfully prepare for retirement. Further, John discusses the inadequacy of the system for those that do not save enough or let their assets leak out when they change jobs or cannot pay back a 401(k) loan.

As John noted, the policy research team at Morningstar, which I head up, has spent time evaluating many of the less radical overhauls that people have proposed, such as newly legal Pooled Employer Plans and experimental state-run automatic IRAs--both of which we think have potential to improve our retirement system.

To add to John’s piece, it’s important to understand that any comprehensive overhaul of the retirement system must go beyond reforming defined-contribution plans. Any reforms aimed solely at the defined-contribution system would still represent a piecemeal approach, leaving other pillars of our system further misaligned. Let’s start with the third rail of American politics--Social Security--and then move on to our still-large pension system and our muddled retirement income policies.

A Progressive, Nearly Universal System Has Masked Flaws in Our DC System but Should Be Adjusted as Part of Overall Reform One reason that Americans tend to have pretty good income replacement in retirement, despite the problems with the 401(k) system, is that Social Security covers around 96% of workers and is highly progressive. (Most of the noncovered are government employees with state-provided defined-benefit plans.) The first pillar of the American retirement system replaces as much as 90% of income for low-income workers. The system does what it is supposed to, at least right now: The Social Security Administration reports that just 7.3% of people collecting Social Security benefits are in poverty.

Concurrent adjustments to Social Security will be critical if policymakers rethink our defined-contribution system. Combined contributions to Social Security (already 12.4% of salary split between workers and employers) and the new defined-contribution system would likely jump to more than 20% of salary, to achieve reasonable retirement income for higher-income people for whom Social Security replaces less. Regardless of whether workers or employers technically pay for these contributions, this would represent a large increase in resources devoted to retirement savings and raise other questions about the design of Social Security.

Depending on the level of contributions, do the benefits need to be as progressive going forward for those that spend their working lives in the new system? Does it make more sense to means-test for Social Security since a mandatory system reduces the moral hazard people won’t save, knowing they have a minimum benefit? On the other hand, would people take imprudent investment risks figuring a means test would act as an income floor? There are lots of ways to address these questions but leaving them unanswered would create more problems rather than fewer.

For that matter, any overhaul of the system should address Social Security’s fiscal imbalance. Unlike other challenges we face, Social Security’s looming insolvency (it will only have enough money to pay for 75% of benefits in the next decade-and-a-half according to recent projections) is a solvable math problem. It does not involve a lot of guesswork, or untested solutions: Revenue must be enough to pay benefits. Still, the solution will probably involve a further increase in the payroll tax, an increase in the minimum retirement age (which is a de facto benefits cut since most people claim early), and probably a reduction in cost-of-living adjustments or some other subtle benefits cut for future retirees.

The point is that a change in contributions (and expected income) from a defined-contribution overhaul should be done in concert with reforms to Social Security, so that these reforms increase people’s retirement security without forcing them to shoulder risks they are ill-prepared to bear.

The Defined-Benefit Pension System Is Not Dead and Needs to be Accounted For If we had a universal defined-contribution system, what would we do with the retirement plans of the vast majority of municipal, state, and federal workers who still have traditional pensions? Further, while most new hires in the private sector do not have pensions, some 36 million people will receive at least some guaranteed retirement income from a traditional pension, according to Department of Labor data.

The problem with these pension plans, particularly in the public sector, is that they are poorly funded and represent a large liability on future taxpayers. However, traditional pensions may make sense for police officers, firefighters, or others who have a hard stop to their working lives, and to whom society owes a stable retirement income. (The same could be said for many dangerous but important careers in the private sector.)

Private-sector plans are typically the best funded (except, heartbreakingly, for some of the collectively bargained plans for dangerous jobs such as coal miners), but they too will require more contributions to shore them up.

Any overhaul of the defined-contribution system will need to consider whether to incorporate defined-benefit plans or whether to crowd them out and manage a transition to a universal system. As we debate overhauling the 401(k) we need to decide whether we want to keep these pension plans, for whom, and we need a plan to help people transition from one system to another while maintaining solid prospects for a secure retirement.

Helping Convert Assets to Retirement Income May Require a Regulatory Overhaul An overhaul of the defined-contribution system would not be complete without adding some ways for people to convert their income into lifetime income. At a minimum, people getting to retirement with a lump sum of money need at least some way to insure against the risk that they live an unusually long time.

Other countries, like Chile, have integrated guaranteed income into their retirement systems, by making it easy for participants to select such products. The challenge for the U.S. is that our insurance system is largely regulated at the state level, leading to regulatory fragmentation. As a result, we have no national policy on selecting lifetime income options alongside traditional investments, and no one encourages retirees to balance allocations between these different product types.

In the final analysis, the point of a retirement savings system is for workers to have income in retirement. If policymakers adjust Social Security and 401(k) contributions and coverage, they need to also figure out some pathways for ensuring retirees do not run out of money when they retire. As an alternative, policymakers could also use Social Security as the annuity provider of last resort (which is essentially what people who delay benefits are using it for), but this could add strain to an already stressed system.

It’s hard to argue that if policymakers were designing a system from scratch, they would alight on the employer-linked system we have ended up with. Back when traditional pension plans were common, linking traditional defined-benefit pensions to employers used to make sense, as it gave employers a powerful human capital management tool: They could decide when to encourage retirement, and they could reduce turnover by giving workers a powerful incentive to stay and build up benefits.

It makes less sense to tie a defined-contribution system to a company’s benefits--there’s limited human capital management to be achieved. That’s the reason countries that have built their retirement systems from the ground up tend to largely disconnect the operations of retirement plans from employers.

All that said, reforming the defined contribution system in isolation will still mean a piecemeal system with parts that add up to less than the sum of their parts. We should prioritize comprehensive reforms across all facets of the retirement system. There are ways to include employer-sponsored plans while achieving other goals.

The United Kingdom, for example, has a mandate and a public option for workers whose companies do not offer plans. Not surprisingly, they also reformed their universal basic income system at the same time they took on adding auto-enrollment to their defined-contribution system. All these parts connect and must be moved together.