Expanding workplace retirement plan access: Evidence on incentives and some potential paths forward

 Introduction

In recent decades, policymakers have tried to boost both worker access to and participation in employer-sponsored retirement plans (ESRPs).¹ At the federal level, the Pension Protection Act, SECURE, and SECURE 2.0 Acts took significant steps to incentivize ESRP provision by firms and participation by workers. In this brief, we evaluate the impact of alternative policies designed to increase workers’ access to ESRPs.

Economic theory suggests that firms substitute ESRPs and other benefits for cash wages when their value to workers exceeds their cost to the firm. Policies that increase ESRP access may therefore reduce cash compensation, resulting in both benefits and costs to workers. ESRP access is best understood as an equilibrium compensation outcome shaped by firms’ net costs and benefits, worker valuations, and firm- and worker-side frictions. Within this framework, we review evidence on how alternative policies may alter this equilibrium by affecting firms’ net costs, worker valuations, or the frictions that distort whether plans form.

Why do employers offer ESRPs?

We begin by reviewing the trends in ESRP offer rates by firms, access rates for employees, and participation rates by employees. We then introduce a conceptual framework to understand the observed rates of firm ESRP offers and employee ESRP access. We argue that firm offering, worker access, and worker participation are jointly determined: firms are more likely to offer ESRPs when they expect sufficient worker valuation and participation, and workers who value ESRPs may sort toward firms that offer them. We then discuss frictions that may distort these choices.

How should we think about ESRP offering, coverage, and participation?

To describe trends in workplace retirement coverage, we distinguish between three concepts. We define a firm as offering an ESRP if it makes a plan available to at least some workers, either by allowing employee contributions or by making employer contributions on their behalf. Second, a worker has access to (or is covered by) an ESRP if the worker is eligible to contribute to the plan or receives employer contributions. Firm offering and worker access are not the same, as groups of workers may be excluded from a firm’s ESRP based on part-time work status, union membership, or other factors. Finally, a worker participates in an ESRP during a timeframe if either employee or employer contributions are made to the worker’s account. In March 2025, the BLS National Compensation Survey (NCS) reports that 72% of private-industry workers (81% of full-time private-industry workers) had access to an ESRP, and 53% (62%) participated.² These estimates capture whether workers had access to, or participated in, at least one employer-sponsored retirement plan (defined benefit or defined contribution).

These aggregate rates mask substantial heterogeneity by industry, occupation, and establishment size. Access rises sharply with size: among workers at establishments with fewer than 50 workers, 55% had access and 38% participated, compared to 90% access and 76% participation at establishments with 500+ employees. Across industries, access is 96% in finance and insurance and 38% in leisure and hospitality; participation is 15%, respectively. Across occupations, access ranges from 90% (management, business, and financial) to 47% (service). Participation ranges from 79% to 25% across these same occupations. In some cases, access is relatively high, but participation is much lower; for example, within the retail trade industry. These figures suggest that aggregate participation rates reflect both plan availability and take-up conditional on access.

What role do ESRPs play in compensation packages?

As discussed by Rosen (1974), Woodbury (1983), and Summers (1989), employers compensate workers through a mix of cash wages and non-wage benefits like health insurance or ESRPs. Workers may value benefits more than their cost to the firm. For example, employers and workers may benefit from preferential tax treatment that fringe benefits receive, while employees may benefit from the convenience of saving through payroll deduction for ESRP contributions or risk pooling for health insurance. When the value of a benefit to workers exceeds its cost to firms, firms can substitute the benefit for cash wages in a way that increases profit without making workers worse off. Incentivizing ESRP provision, such as through tax subsidies, lowers its cost to the firm, increasing the likelihood that firms and workers can arrive at mutually beneficial compensation packages that substitute ESRPs for cash wages.

In this framework, ESRPs may create private value for firms and workers through several channels. They are a form of tax-advantaged compensation, a saving vehicle that some workers may value above its cost to the firm, and a recruiting or retention tool for workers who value retirement benefits. These channels may make ESRPs privately valuable. However, they do not imply that ESRP access is always socially valuable or beneficial for every worker. Tax advantages have fiscal costs, retention may or may not improve efficiency, and workers differ in how much they value illiquid retirement saving relative to cash wages.

Workers may value ESRPs relative to cost-equivalent cash wages for several reasons. First, workers may value the ability to defer taxes on ESRP contributions and receive preferential tax treatment on investment returns, though the value of these tax advantages varies across individuals. Second, workers may appreciate the convenience of saving for retirement through payroll deduction. Additionally, some “behavioral”³ workers who are aware that they tend to undersave may value features that make saving easier or more automatic, including auto-enrollment, auto-escalation, and the commitment value created by less liquid retirement accounts (Thaler and Benartzi 2004). While all (IRAs), those accounts have lower contribution limits and generally lack the payroll-deduction convenience and plan-design features available through ESRPs. Workers may also value employer contributions because those contributions receive favorable tax treatment relative to cash wages.⁴ At the same time, these same tax and plan features may reduce liquidity and flexibility: early withdrawals may be costly, employer matches may vest gradually, and workers facing liquidity constraints or high-interest debt may prefer cash wages to additional retirement contributions even when tax treatment is favorable. Finally, forward-looking workers who are not currently interested in saving for retirement may value access to an ESRP because it gives them the option to make contributions in the future.

Offering an ESRP has costs. First, firms bear the nominal financial costs and administrative burdens of establishing and maintaining the ESRP. Chen (2023, 2024) reports publicly advertised small-employer 401(k) options, as of late 2023, with annual base administration fees of roughly $950–$1,800 and per-participant fees of about $72–$96 per year. For a firm with 10 participants, these provider examples imply annual administration costs of roughly $1,900–$2,500. Chen also reports another provider charging a $500 setup fee and $1,200 in annual administration fees and requiring employer matching contributions. These advertised fee figures exclude employer contributions, fees such as fund expense ratios, and the value of compliance time. Beyond these administrative costs, firms must comply with nondiscrimination rules intended to ensure that ESRPs are not primarily used by highly compensated workers. Compliance requires time and effort. Firms may also find they need to provide employer contributions (above and beyond enabling workers to contribute their own wages) to pass or avoid nondiscrimination tests.⁵ These employer contributions add to the cost of offering an ESRP if not fully offset by decreases in employee wages.

This framework predicts that a firm is more likely to offer an ESRP to a group of employees when those employees, in aggregate, value the ESRP more than its cost to the firm. In this scenario, a firm can substitute an ESRP for cash wages in a way that makes the firm and eligible employees better off on net. To the extent that ESRP provision involves fixed costs, the framework predicts that larger employers will be more likely to offer ESRPs.

In making their offer decisions, firms will also consider that offering an ESRP or other benefits may change the composition of their employee pool (see e.g., Ouimet and Tate 2023). Firms cannot arbitrarily exclude individual employees from an ESRP in exchange for higher wages. Groups of employees may be excluded, although there are strict limits on how these groups may be defined (see, e.g., Freelove 2019). Thus, offering an ESRP to a group may make it easier to recruit or retain workers who value the ESRP, but harder to recruit or retain workers who prefer the cash wages. Firms may also consider whether workers who prefer ESRPs to cash wages differ in firm-relevant characteristics, such as patience, expected tenure, job attachment, or productivity. A related implication is that when a fixed group of workers gains access to an ESRP, there is redistribution within the group from those who prefer cash wages to those who prefer the ESRP.

Empirical work suggests that other employer-provided benefits like health insurance and workers’ compensation are partially offset by lower cash wages (e.g., Baicker and Chandra 2006; Olson 2002; Gruber and Krueger 1991). There is less evidence on retirement benefits and labor markets. Defined benefit pension incentives have been shown to influence labor supply (e.g., Stock and Wise 1990; Ni and Podgursky 2016). However, evidence on the extent to which ESRPs are offset by lower wages is limited (especially relative to the health insurance literature). Simon and Kaestner (2004) show that , which is consistent with limited (or no) substitution away from these benefits when wages rise exogenously. Relatedly, provide evidence that higher employer contributions to DC retirement plans increase firms’ success in filling posted jobs, with larger effects in high-income occupations and older-workforce occupations. This evidence is consistent with meaningful worker valuation of retirement benefits, This evidence base is limited and still emerging, which is an important caveat to our analysis. Future research should examine the relationship between ESRP provision and other dimensions of compensation.

What are the implications for ESRP offering and access?

This conceptual framework implies that firm offering, worker access, and worker participation are jointly determined. Firms do not decide to offer ESRPs in isolation from worker demand; they offer them when they expect a group of workers to value ESRP access more than its cost, presumably because enough members of the group expect to participate at some point. At very small firms, which are often the target of efforts to expand ESRP access, the preferences and anticipated participation of even one or two employees may have significant influence on a firm’s offer decision, and hence on other workers’ access.

This framework also implies that any misperceptions or limited attention regarding ESRP availability, costs, or benefits, among employers or workers, can influence ESRP provision. ESRP access may also be affected by technological or regulatory changes that alter provision costs, improve the availability or salience of plan options, change how information about costs and benefits diffuses among employers and workers, or reduce communication frictions between the two sides of the labor market.

How can policymakers incentivize ESRP access?

There are several reasons why policymakers may want to induce firms to offer or give workers access to ESRPs. First, firms may overestimate the cost of offering ESRPs, or they may be deterred by administrative burdens, paperwork, inertia, or limited awareness (Chen 2023; Scott and Olson 2024). Relatedly, nondiscrimination tests may make ESRPs costly or limit employers’ ability to target access to workers who value them. Second, workers may misperceive the value of ESRP access or participation, in either direction, because of present bias, inattention, limited financial literacy, or uncertainty about future public benefits or tax policy. If firms anticipate low worker valuation or participation, they may choose not to offer plans. Third, saving decisions may have fiscal spillovers: additional private saving may reduce future reliance on public programs, but tax-preferred retirement saving also has budgetary costs and distributional tradeoffs. Given these tradeoffs, as well as the compensation and worker-heterogeneity considerations discussed above, we do not take a stand on the normative question of whether ESRP access should increase. Instead, we present evidence on the positive question of how alternative policies affect each channel underlying ESRP access.

What policies could lower firms’ actual or perceived costs of ESRP provision?

Most workers at large firms have access to ESRPs, though variation in offers exists across much of the size distribution. Recent efforts to incentivize ESRP provision have therefore focused on smaller firms. For example, since the early 2000s, the federal government has offered tax credits to small firms —for up to three years —to offset the administrative cost of establishing and operating an ESRP. s. Policymakers have also taken steps to facilitate pooled plans —e.g., multiple employer plans (MEPs) and pooled employer plans (PEPs)— in which costs and administrative burdens are distributed across many employers. In addition, many state governments have adopted “auto-IRA” policies requiring firms to either offer an ESRP or initiate automatic employee contributions to IRAs established for workers by the state. Analyzing these policies can shed light on how firms assess and respond to the costs and benefits of providing ESRPs.

Box 1 summarizes the SECURE Act (2019) and SECURE 2.0 (2022) provisions most directly aimed at increasing employer provision of workplace retirement plans. In the remainder of this section, we summarize what the available evidence, including our recent work, suggests about how employers respond to these tools.

Theoretically, tax credits lower firms’ costs of offering ESRPs. They may also alleviate liquidity constraints to the extent that costs are front-loaded. These lower costs should necessitate smaller offsetting wage reductions, making it more likely that equilibrium compensation packages will include ESRP access. In equilibrium, employers and employees will split the tax credit. However, in contrast to the theoretical prediction, our recent work (Bloomfield et al. 2025a) suggests that most small firms that establish ESRPs do not even claim the credit. Even after the more generous credit went into effect in 2023, the take-up rate among apparently eligible firms in 2024. Since some of these ESRP-adopting firms would have established a plan regardless of the tax credit, the low take-up rate places an upper bound on the number of ESRPs induced by the tax credit.

Financial sophistication and awareness may explain part of this failure to claim the credit. For example, eligible firm owners with more education are more likely to claim the credit (though even the vast majority of highly educated owners do not claim it). Moreover, firms are more likely to claim the credit if they work with a tax preparer who has already claimed the credit on behalf of another client. This finding suggests that information frictions may be substantial, . However, these factors account for only a small fraction of incomplete take-up. For example, among firms that claim the credit, the probability of claiming it again in the following year drops sharply and is typically below 50% (Bloomfield et al. 2025a). This pattern suggests that administrative or informational frictions may play a role in low take-up even among firms that have previously claimed the credit.

In other recent work (Bloomfield et al. 2025b, 2025c), we have found that state auto-IRA policies cause many affected firms to adopt new ESRPs rather than utilizing the auto-IRA program. This response is difficult to reconcile with a simple cost-minimization model in which firms fully understand the costs and benefits of all available options. The conceptual framework above predicts that most firms not already offering ESRPs (presumably because the costs exceeded the benefits) will enroll their workers in the state auto-IRA programs (which likely have only a small administrative cost to the firm).

Interestingly, we also find that firms induced to start ESRPs by state auto-IRA policies are less likely to claim the ESRP tax credit than firms that would have started an ESRP regardless (though the take-up rate is low in both groups). About 3% of these “complier” firms claimed the credit at any point from 2020 to 2022, compared with about 8% of “always-taker” firms. Such a pattern could emerge if actual (or perceived) paperwork and procedural frictions influence both auto-IRA policy-induced ESRP formation and the low take-up rate of the tax credit (Bloomfield et al. 2025a, b). If that explanation is correct, then reducing administrative burdens could increase the effectiveness of tax credits and other policies that incentivize ESRP formation.

Overall, the observed firm behavior is difficult to reconcile with a simple model in which firms are fully informed and continuously optimize over available retirement-plan options. These results lead us to consider firm-side “behavioral” factors such as inertia, limited attention, low financial literacy, and marketing. For example, a firm’s owners may initially decide that the benefits of offering an ESRP do not exceed the costs. Inertia may cause them not to revisit that decision even as the firm grows and the cost-benefit analysis shifts. The owners may only discover this shift when state policies force them to make an active choice between an ESRP and the auto-IRA program. In addition, it is possible that ESRP administrators have used the adoption of auto-IRA policies to market their services to small employers. Furthermore, some small firm owners may have low financial literacy, which could further impede their capacity to do the complex financial calculations around fringe benefit offers.

To our knowledge, however, there is no direct evidence on the role of paperwork burdens, inertia, owner financial literacy, or marketing on small firms’ ESRP offerings. We encourage researchers to examine these issues.

Theoretically, nondiscrimination rules also raise the cost of offering an ESRP. These rules directly create an administrative burden and blunt the ability for employers to target only employees who value ESRPs. Firms may also need to make employer contributions to pass or avoid nondiscrimination tests. We are not aware of direct evidence on the impact of nondiscrimination rules on ESRP offer decisions. Recent work by Ouimet and Tate (2023) suggests that, within firms, the distribution of cash wages is much more unequal than the distribution of non-cash benefits like ESRPs. On the other hand, across firms, the distribution of non-cash benefits is much more unequal than the distribution of cash wages. They argue that nondiscrimination rules may explain this finding. These rules constrain firms that choose to offer benefits to extend access to a broad range of employees. On the other hand, they do not require firms to offer benefits. In equilibrium, therefore, some firms will opt not to offer benefits (to any employees), while others will offer benefits to both low-paid and high-paid employees.

While relaxing nondiscrimination rules may incentivize ESRP formation, these rules are intended to ensure that the tax advantages of ESRPs do not only accrue to the most highly compensated employees. Policymakers may therefore be reluctant to relax them to induce broader access to ESRPs if such a change would also cause participation to become more concentrated among highly compensated employees.

What policies could increase worker valuation of ESRP access?

Workers’ valuation of ESRP access is an input into employer offer decisions. This valuation in turn depends on factors such as marginal tax rates (and therefore the value of tax-preferred saving), expected retirement needs, liquidity constraints, and the convenience of payroll deduction and employer contributions. For example, younger workers who have other immediate spending needs (e.g., rent or student loan payments), expect income growth over the life cycle, or anticipate lower asset market returns may not value opportunities to save for retirement. In addition, lower-income workers who anticipate a high Social Security replacement rate and are more likely to have binding budget constraints may place a low value on private retirement saving. If such workers suffer from inertia, cognitive constraints, or a lack of financial literacy, some may be made worse off by ESRPs with an automatic enrollment feature (Scott et al. 2022). Policymakers could potentially increase these workers’ valuation of workplace saving plans by making it easier to use such saving for non-retirement purposes (such as emergencies or education costs).⁶ Alternatively, they could expand employer flexibility (provided by SECURE 2.0) to make matching contributions for non-retirement expenses like student loan payments (Horneff, Maurer, and Mitchell 2024) or savings toward a future home, business, or educational experience.

Policymakers could also provide clarity about the value of private retirement saving by resolving policy uncertainty. For workers, the appropriate amount of private retirement saving— and the value of ESRP access to facilitate that saving— depends crucially on anticipated Social Security and Medicare benefits, as well as tax rates. However, the federal government faces a long-term financial shortfall, and there is significant policy uncertainty about how this shortfall will be resolved. For example, the Social Security and Hospital Insurance trust funds are projected to be depleted in less than a decade (Social Security Trustees 2026; Medicare Trustees 2026). A politically feasible Social Security reform plan will likely involve reducing benefits for higher earners (in addition to tax increases), causing the affected workers to rely more heavily on private saving. Workers who know how they will be impacted by reform can realistically assess the value of ESRP access.

There are broadly two distinct rationales for policies that seek to increase retirement savings through ESRPs. First, individuals may lack the foresight, cognitive skills, or self-control to save optimally for retirement. Relatively “sophisticated” workers who recognize these tendencies may value ESRP access—especially with automatic features— as a commitment device; others may lack such self-awareness (Laibson, Repetto, and Tobacman 1998). Second, even when individuals are optimizing privately, saving decisions can have externalities. For example, workers may rationally anticipate relying on family support during retirement or on public or private transfers available to those with low assets level (e.g., Lindbeck and Weibull 1988; Deryugina and Kirwan 2018). Furthermore, some workers may be reluctant to build up assets that cause them to lose eligibility for means-tested programs such as Medicaid (e.g., Hubbard et al. 1995; Gruber and Yelowitz 1999). Finally, income-based federal and state taxes penalize future consumption relative to current consumption, which distorts savings incentives. While workers may accurately assess the private value of retirement saving and social insurance, they may not fully internalize the impact on the federal budget. Thus, the private value of retirement saving may differ from its social value.

These behavioral and fiscal rationales imply different benchmarks for evaluating policy tools. If the concern is primarily behavioral, low-fiscal-cost nudges such as default enrollment with easy opt-out can increase saving among workers who would otherwise under-save, while limiting the risk of pushing others to save more than they privately prefer. If the concern is primarily fiscal spillovers, increasing saving for some workers can be socially beneficial even when some individuals privately prefer lower saving. Monetary incentives can help align private decisions with the broader social benefits of additional saving, but they have budgetary costs and may deliver windfalls to firms or workers who would have opted to offer a plan or save anyway. Whether such incentives are cost-effective depends on how much induced saving is net new, how much that new saving reduces future public outlays, and how well the policy is targeted. Evidence on these questions is mixed (e.g., Poterba et al. 1996; Engen et al. 1996; Chetty et al. 2014).

These rationales are often cited for either mandating or incentivizing retirement saving among workers (e.g., Kotlikoff 1987) or ESRP formation by employers. Workers who can conveniently save for retirement through payroll deduction may be incentivized to participate. Moreover, ESRPs often feature employer contributions, auto-enrollment, and auto-escalation. These features can help firms comply with nondiscrimination rules. There are also tax credits available for auto-enrollment and employer matching contributions— plan design features that have been a focus of recent federal legislation. Moreover, SECURE 2.0 requires most new 401(k) and 403(b) plans (subject to exemptions) to include automatic enrollment beginning in plan years after 2024. Auto-enrollment, auto-escalation, and employer matching all tend to boost plan participation and contributions (; Thaler and Benartzi 2004; Even and Macpherson 2005), although the impact on long-term wealth accumulation is less clear due to delayed opt-outs, cash-outs, and potential crowd-out of other forms of saving (Choukhmane 2025; Beshears et al. 2022, 2024; Derby, Mackie, and Mortenson 2023; Choi et al. 2024).

Policymakers should keep two things in mind as they consider measures that incentivize or mandate ESRP offering or access. First, incentivizing employers to offer ESRPs is at best an incomplete approach for policymakers seeking universal ESRP access. Employers can currently restrict access to employees who fall within certain tenure, union membership, full-time work, or occupational groups. Therefore, even a requirement that all employers offer an ESRP to at least some employees could still leave many workers without access.⁷ Second, incentivizing employers to extend access to all their workers has tradeoffs because some employee groups may value cash wages more highly than ESRP access. These employees may be worse off if expanded ESRP access is offset by lower cash wages and plan designs that do not match their needs (e.g., illiquid saving). Finally, extending “universal access” to non-employers (e.g., the self-employed, sole proprietors, and many partnerships) will likely require pooling arrangements or technological advances that reduce fixed plan costs.

Recommendations

In this brief, we have argued that ESRP offers and access depend on firms’ net costs and benefits, worker valuations, and frictions on both sides of the labor market. Policies to incentivize ESRP access may reduce actual or perceived firm costs, improve worker decisionmaking, change worker valuation, or target fiscal spillovers. These channels have different welfare implications, so policy design should focus on the specific barrier being addressed.

On the firm side, our research suggests that smaller firms’ choices may be driven more by paperwork burdens, lack of awareness, or inertia than monetary costs. This sensitivity may limit the impact of policies like tax credits, which offset monetary costs but require firms to know about, understand, and claim them. Evidence from state auto-IRA policies is consistent with the idea that requiring firms to make an active choice can induce some employers to revisit whether offering an ESRP is worthwhile. If firms lack information about the expected costs and benefits of ESRPs, or if administrative frictions are binding, automation may matter as much as, or more than, the generosity of fiscal incentives.

Policymakers can reduce firms’ actual or perceived costs of ESRP provision by:

• Simplifying claiming of the ESRP tax credit, for example by embedding the credit calculation in standard business tax filing workflows (without requiring a separate form).
• Providing automatic reminders for tax credit eligibility in the second and third years. These reminders could include prompts in tax software or IRS communications to firms that claimed the credit once.
• Standardizing a “tax credit statement” that plan providers and payroll companies give to new-plan sponsors with the information needed to claim the credit (plan start date, eligible expenses, and employer contributions).
• Pairing monetary incentives with low-cost compliance assistance (for example, a standardized onboarding checklist and templates for small employers).
• Considering whether nondiscrimination rules discourage ESRP formation.

Policymakers could improve worker decisionmaking and target incentives by:

• Increasing flexibility in ESRP design in settings where workers may value current liquidity more than additional illiquid retirement saving.
• Addressing Social Security, Medicare, and broader fiscal imbalances so workers can better assess the role of private retirement saving.
• Studying how opt-out design affects liquidity-constrained workers.
• Targeting fiscal incentives where they are most likely to generate net new saving rather than windfalls.

Technological improvements, pooling arrangements, and administrative automation that reduce fixed costs or compliance burdens are therefore comparatively straightforward policy tools: they can expand the set of mutually valuable compensation packages without requiring strong assumptions about whether workers are generally saving too little or too much. We emphasize, however, the need for more research on “behavioral” factors that may influence firm ESRP offers, such as inertia, financial literacy, and marketing. More research is also needed on how ESRPs influence employment selection and equilibrium compensation packages.