The Pension Protection Act at 20: Defaults Built for a Job You Keep

Consider Maria, a hypothetical worker who has spent six years with the same employer. During that time, her 401(k) contribution rate gradually increased from 3% to 8% through automatic escalation. She did not have to make a new decision each year. Her employer's retirement plan did the work for her.

Then Maria changes jobs. Her new position comes with a 10% salary increase, but her new employer's retirement plan automatically enrolls her at a 3% contribution rate.

Unless she takes action, six years of progress in increasing her retirement saving rate have effectively been reversed.

The financial consequences can be substantial. Suppose Maria earned $60,000 before changing jobs. Her new salary is $66,000. At her previous 8% contribution rate, she would contribute $5,280 annually to her 401(k).

At the new employer's 3% default, she contributes only $1,980. That is $3,300 less in annual retirement contributions, before considering employer matching contributions or investment returns.

Maria's experience illustrates a limitation in the American retirement saving system.

Automatic enrollment and escalation can help workers establish and gradually increase their saving rates. But those higher rates do not necessarily follow them when they change employers.

Twenty years ago, the Pension Protection Act, or PPA, helped transform defined-contribution, or DC, plans by encouraging automatic enrollment and escalation and establishing a framework for qualified default investment alternatives.

It was a major achievement, but an incomplete one.

On the investment side, the PPA has worked remarkably well. On the contribution side, the record is more mixed, particularly for workers who change jobs.

That raises a question for the next generation of automatic plan design: Should the account follow the worker, the saving rate, or both?

In 2005, Sarah Holden and I modeled how automatic enrollment might affect workers who were then decades from retirement. For lower-income workers reaching age 65 between 2030 and 2039, we projected a median replacement rate from 401(k) accumulations of 23% without automatic enrollment.

A 3% default contribution rate paired with a money market fund, one of the common default choices at the time, raised it to 37%; a 6% default paired with a lifecycle fund raised it to 52%.

The investment side of that comparison largely worked, especially after the Department of Labor provided a framework for diversified default options such as target-date funds and managed accounts in 2007.

By 2024, Vanguard found that 98% of automatic-enrollment plans used a target-date fund as their default investment.

Contribution defaults improved substantially as well. Sixty-one percent of plans defaulted workers at 4% or more, including 30% that defaulted at 6% or more.

Still, 39% defaulted workers at 3% or less. The higher-default case we modeled in 2005 is therefore no longer unusual.

Recent evidence, however, suggests that automatic saving policies may produce smaller long-run effects than earlier studies estimated. A study by Choi and colleagues estimated that automatic enrollment increases steady-state saving by about 0.6% of income, while default escalation adds about 0.2%.

Turnover is one reason.

Vanguard data cited by the researchers show that 29% of new hires leave within one year and more than two in five leave within two years. The researchers also estimated that, averaging equally across cohorts and firms, 42% of balances are cashed out at separation.

Escalation loses force over time as well: Across 21 plans with default escalation, 43% of participants accepted the first scheduled increase, 36% the second, and 29% the third.

Our simulations measure a different outcome.

In a 2024 Morningstar analysis of the evolution of retirement-income adequacy under ERISA with a focus on DC plans, Spencer Look and I estimated that moving remaining voluntary-enrollment plans to automatic enrollment increased the average projected wealth-to-final-salary ratio at age 65 by 16.3%, while adding auto-escalation up to a 15% cap increased it by 28.8%. The medians were much smaller: 0.5% and 8.0%.

Those estimates are not directly comparable with figures from the researchers. Their estimates measure changes in annual saving as a share of income; ours measure percentage changes in projected wealth relative to final salary.

The large difference between our mean and median results also shows that the gains in our simulations are highly uneven.

Our universal-coverage analysis reinforces the importance of sustained participation. We estimated retirement wealth increases of 28% to 49% across the base plan designs, rising to 67% to 125% among workers who sustained participation for at least 10 years.

That is a conditional comparison, not a causal estimate of continuity, but it illustrates why remaining connected to the saving system matters.

Auto portability addresses one part of that continuity problem. My 2019 EBRI analysis examined automatically moving a former employer’s account balance into a worker’s active account at a new employer as a way to reduce cashout leakage.

SECURE 2.0 later created a statutory prohibited-transaction exemption for certain automatic-portability transactions and raised the mandatory-distribution limit from $5,000 to $7,000.

It also generally requires covered 401(k) and 403(b) plans established after Dec. 29, 2022, beginning with 2025 plan years, to start automatic contributions between 3% and 10% and escalate annually until at least 10%, but no more than 15%.

These provisions address different parts of the problem. Auto portability can preserve an account balance across jobs. Section 101 of SECURE 2.0 strengthens the contribution path within a job. But neither automatically preserves the contribution rate when a worker changes jobs.

Vanguard research highlighted why this distinction matters. Among workers whose prior and new employer plans were both recordkept by Vanguard, the median job switcher received a 10% pay increase but experienced a 0.7 percentage-point decline in the retirement saving rate.

Vanguard proposed setting the worker’s new default at the higher of the new plan’s default or the worker’s prior saving rate, potentially using expanded automatic-portability data feeds.

The researchers reached a similar diagnosis from different evidence: Job transitions are a weak point in the retirement saving system. They suggest infrastructure that would allow higher contribution rates reached at one employer to carry over automatically to the next.

Their estimate that default escalation increases steady-state saving by only about 0.2% of income is an important warning against assuming that contribution-rate portability would produce large effects.

But it does not tell us what would happen if workers actually carried their previous saving rates into their next jobs. That is a different policy design and requires a different test.

What is missing is a population-level simulation of that alternative. Who would benefit? How many workers would opt out? How would balances and employer matching costs change? Would the effects differ by age, income, tenure, plan design, or frequency of job changes?

Those are questions the Morningstar Defined Contribution Outcomes Model, or DCOM, can examine.

DCOM simulates participant outcomes under alternative plan designs, allowing us to test how changes in defaults and other plan features affect retirement outcomes across different types of workers.

Evaluating alternative default architectures, including personalized default designs, is already part of our published research agenda.

One logical next step is to compare the current system with alternatives in which some or all of a worker’s previous contribution rate follows that worker to a new employer.

Such an analysis could estimate both the potential improvement in retirement outcomes and the consequences for employers and participants who respond differently to the new default.

The PPA’s legacy is not that defaults solved retirement saving. They did not. Its legacy is that defaults became central to retirement plan design.

Twenty years later, the next question is whether the default can survive the moment a worker walks out the door.