Evolving auto features to improve retirement adequacy

Automatic enrollment has transformed retirement plan participation and dramatically expanded access to workplace savings. Data from our recordkeeping platform show that in 2025, plans with auto-enrollment achieved participation rates of 85%, compared with 46% in plans without it.¹ But after more than two decades of adoption, the conversation is shifting from plan participation to one of adequacy—whether participants are saving enough and whether those savings are durable over time.

Recent research highlights several emerging considerations that impact the durability of auto-enrolled savings: higher default contributions may be associated with increased household debt (Beshears et al. 2024² and Choukhmane and Palmer 2025)³; some of the gains from auto-enrollment may be eroded by leakage through loans and early withdrawals (Choi et al. 2024)⁴; and as discussed later, mid‑ and late‑career job switchers may be defaulted into lower contribution rates that fall short of their savings needs.

These insights present an opportunity to evolve automatic plan features. Auto-enrollment has proven to be a highly effective plan design tool. A next generation of “Auto Features 2.0” could build on its foundation by leveraging technology to better align default strategies with a broader set of participant characteristics, such as age, and support innovation in implementing legislative enhancements like pension‑linked emergency savings accounts (PLESAs). Advancing this evolution will require coordination across industry stakeholders and policymakers to expand what is possible, while also sustaining the long‑term gains of current automatic features.

The power of automatic designs

Automatic plan features have evolved from a behavioral nudge to a foundational design framework for defined contribution plans, shifting key decisions—such as whether to enroll, how much to contribute, or how to invest—into default settings. Auto‑enrollment significantly increased participation, particularly among historically underserved groups such as lower‑income workers and younger employees, who have traditionally exhibited lower participation rates (Copeland 2025),⁵ but experience has shown that higher participation alone does not ensure adequate retirement savings.

Contribution rates, escalation mechanisms, and other design features also play an important role in determining long‑term outcomes. This has driven industry recognition that higher default contribution rates are necessary to improve ultimate retirement outcomes. In response, many plan sponsors have begun adopting higher initial defaults alongside automatic escalation features—with minimal impact on enrollment rates (Figure 1).

Many defaulted participants actively adjust their contribution levels. In plans with low default rates, a substantial share increase contributions. Even at higher default contribution rates, participants are more likely to increase than decrease their contributions.

Further, tools such as auto‑increase and periodic auto‑reenrollment have taken on increased importance. While auto‑increase ensures that saving rates rise over time, auto‑reenrollment provides a mechanism to bring previously opted‑out participants or those not initially captured by automatic enrollment back into the plan and realign them with current defaults.

Notably, data from our recordkeeping system show that 79% of re‑enrolled participants remain in the plan. Yet only 14% of plans currently use this feature, reflecting, in part, plan sponsor concerns around potential negative employee reactions. The evidence of low opt‑out rates, however, highlights a meaningful opportunity to expand adoption.

Automatic features have also played a central role in improved asset allocation through qualified default investment alternatives (QDIAs). Our recordkeeping data show that nearly all plans (99%) now use target date strategies as their QDIA option, and 65% of participants invest their entire account balance in a target date portfolio. Participant behavior shows strong adherence to QDIAs, especially among younger participants who are more likely to be automatically enrolled and remain in these allocations over time.

How outcomes evolve over time

Examining how participant saving behavior changes with tenure reveals important differences in how saving outcomes are impacted under different plan designs. In the first three years of participation, contribution rates in auto‑enrollment plans without auto‑increase tend to lag those in voluntary enrollment plans and in plans that also use auto-increase (escalation), regardless of opt-in or opt-out implementation (Figure 2).

Over time, automatic contribution increases play a critical role in sustaining savings momentum, particularly for participants who are less likely to actively manage their contributions. As shown in Figure 2, plans with auto‑increase keep participants on pace with voluntary enrollees—who are typically more engaged and tend to contribute at higher rates—while plans without it fall behind. This divergence underscores the importance of escalation as a core component of effective plan design.

Gaps in a one‑size‑fits-all default system

While effective, current automatic plan features also reveal structural limitations. Participants with non‑linear career paths, including those who change jobs frequently or move in and out of the workforce, may be repeatedly defaulted into contribution rates that are misaligned with their age and retirement needs.

Research shows that participants tend to remain at default contribution rates for extended periods (Choi et al. 2001⁶ and Madrian and Shea 2001).⁷ Each job transition can therefore reset saving behavior and anchor individuals at lower contribution levels.

This dynamic is reflected in our recordkeeping data. Within auto‑enrollment plans, older participants with shorter tenure exhibit lower contribution rates than similarly aged peers with longer tenure (Figure 3). For example, among participants over age 50 in plans with opt‑out auto‑increase (AE+AI Opt-Out), those with 10+ years of tenure contributed an average of 12%, compared with 8% for those with 1–3 years of tenure.

These patterns suggest that newer employees are defaulted to a lower saving rate, highlighting the potential benefits of increasing the default savings rates with age.

A potential trade‑off: Expanding access vs. increasing financial stress?

Automatic features can bring into the retirement system a broader population of workers, many with varying levels of financial resilience. Participants enrolled automatically may include individuals who had not previously prioritized retirement saving because they face tighter household budgets (Choukhmane and Palmer 2025).⁸ As participation expands, retirement plan savings increasingly intersect with participants’ broader financial realities.

Related research suggests that participants often rely on retirement accounts during periods of financial strain. Figure 4 highlights a notable difference between plan types: loan usage was more than 30% higher in auto-enrollment plans than in voluntary enrollment plans (21% versus 16%), even though average outstanding loan balances were generally similar across age groups. Meanwhile, average loan sizes increase with age—approximately $4,700 for participants in their 20s, $8,700 for those in their 30s, and $11,500 for those in their 40s—versus roughly $10,000 overall.

At the same time, findings from our recent annual survey of a nationally representative sample of 401(k) plan participants indicate that while retirement saving remains a top priority, it is not the only one (Figure 5). Together, these trends point to an inherent tension where the same mechanisms that increase participation may also expose underlying financial fragility.

Improving retirement outcomes may, therefore, depend not only on increasing participation and contribution rates through enhanced plan design or default features, but also on how effectively plans can help accommodate short‑term financial needs through expanded access to emergency savings accounts.

Strengthening outcomes through emergency savings

Results from a recent voluntary survey of participants on our recordkeeping platform revealed that participants who have access to emergency savings tend to have higher contribution rates, higher account balances, and a lower likelihood of taking plan loans (Figure 6).

Emerging plan features and policy developments increasingly aim to address this gap. Options such as PLESAs and other workplace emergency savings mechanisms are designed to provide participants with a modest, liquid reserve that can absorb near‑term financial shocks without disrupting retirement assets. Our early experience as one of the first recordkeepers to implement a PLESA solution points to promising participant engagement. Initial anecdotal evidence suggests enrollment volumes have exceeded expectations, aided by an active plan sponsor rollout and participant education.

Aligning these features with existing plan design may enhance their effectiveness. For example, pairing automatic enrollment in retirement plans with automatic enrollment in a PLESA could extend the benefits of behavioral defaults, encourage participation, and reduce leakage.

Looking ahead, the next phase of retirement plan design will depend on how effectively plans adapt to participants’ evolving and increasingly complex financial lives. The following considerations, some of which align with proposed regulations, highlight targeted opportunities to improve both adequacy and financial resilience.

Plan design considerations

• Revisit default contribution rates—employers should reassess whether default contribution rates remain appropriate for their workforce. Data from our recordkeeping system show that many plans still default participants at relatively low contribution rates, with roughly one‑third set at around 3%.
• Make auto‑increase the default—establishing automatic contribution increases as an opt‑out feature can help ensure more consistent savings growth over time. Currently, 49% of plans with auto‑increase on our recordkeeping platform offer auto‑increase on an opt‑in basis. Participant utilization drops to just 10% under opt-in designs, compared with 60% when implemented as opt‑out, indicating room to further standardize opt‑out design. Expanding this approach could increase participation in escalation and support higher long‑term savings rates.
• Adopt automatic reenrollment⁹—reenrollment provides a mechanism to bring disengaged participants back into the plan and realign them with current defaults. Periodic use, for example, on an annual basis, can help ensure that earlier opt‑out decisions or outdated elections do not limit long‑term outcomes.
• Emergency savings integration—pairing retirement plan enrollment with automatic emergency savings accounts can help participants manage short‑term financial shocks without tapping retirement assets. Extending defaults across both savings vehicles may improve participation and reduce leakage.

Policy considerations

• Age‑based auto‑enrollment rates—static default rates may underserve participants who have mid‑ or late‑career job switches or have interrupted careers. Future legislative or regulatory changes that would more broadly allow age-based contribution structures could better match savings needs, for example, x% up to age 30, y% between ages 30–50, and z% for age 50+.
• Increase the PLESA savings cap¹⁰—current emergency savings limits may not reflect the scale of real‑world financial needs, given that the average loan amount is $10,000. Increasing the cap could enhance the effectiveness of these programs and reduce reliance on loans and hardship withdrawals.

 

Appendix

T. Rowe Price Global Retirement Savers Study (GRSS): The 2025 GRSS surveyed 7,010 adults age 18+, representative of the population of workers (on age, gender, and region) contributing to a defined contribution (DC) or similar account based workplace retirement plan in each country.  The surveys were conducted in Australia, Canada, Japan, UK and the US. 1,000 (or slightly more) adults per market were surveyed, except in the U.S., where 3,001 adults were surveyed.