Do State-Facilitated Retirement Savings Programs Have a Positive Impact on Employers Offering Plans and Worker Participation?
By Adam Bloomfield, Kyung Min Lee, Jay Philbrick, and Sita Slavov
Employer-sponsored retirement plans (ESRPs) are the largest source of private retirement savings in the U.S., although many workers lack access to such plans. While workers can establish Individual Retirement Accounts (IRAs), many do not. To increase retirement savings, states like California, Oregon, and Illinois are implementing automatic enrollment IRA (auto-IRA) programs. In these states, employers without ESRPs must either (1) adopt their own retirement plans or (2) facilitate payroll deductions that are deposited into IRAs established for workers by the state. Workers can opt out of the auto-IRA program at any time. Our recent research examines how these auto-IRA programs influence employers’ decisions to offer, and worker access to and participation in, ESRPs.
Theoretically, auto-IRA programs can increase, decrease, or leave unchanged an employer’s likelihood of offering an ESRP. Some employers might drop their ESRPs in exchange for the state program to cut costs, while others may start offering ESRPs because of the value of offering such benefits and how they align with potential changes in business norms. Our research findings suggest that employers in states that implement a state auto-IRA program are more likely to offer their own plans and workers are more likely to participate in such plans.
Our research uses individual-level data from the Current Population Survey (CPS) and firm-level data from Form 5500 (F5500) filings to examine the impact of auto-IRA rollout on firms offering ERSPs as well as worker access to and participation in them.
While most of the retirement literature considers worker saving behavior conditional on having access to a plan, this research contributes to our understanding of why employers provide fringe benefits. By focusing on an employer’s decision to provide retirement benefits, we show that saving for retirement is a function of both employer and worker decisions.
How State Programs May Affect Employer Behavior
Standard economic theory states that, in a perfectly competitive labor market, employers offer benefits when the value of the benefits to workers exceeds the cost to employers. Auto-IRA programs should not affect firms already providing ESRPs, because these policies merely change IRA enrollment defaults. Firms that drop ESRPs in this context would reduce workers’ compensation, which would have to be offset by wage increases.
However, state auto-IRA laws may alter the behavior of firms that do not offer ESRPs. Firms close to indifference between offering and not offering an ESRP may be induced to adopt one rather than use the state program. Thus, auto-IRA legislation could expand the set of firms providing ESRPs.
Considering a behavioral economics framework, auto-IRA laws could raise the prominence and awareness of retirement plans, potentially increasing their perceived value and adoption. Moreover, financial service providers might also leverage these laws to sell retirement products and services to employers.
Conversely, some firms with ESRPs might drop their relatively costly plans, opting for auto-IRA enrollment of workers. This substitution is plausible if employees are not fully rational—if they perceive the state plans as a new employer-provided benefit or equivalent, a scenario akin to public services “crowding-out” private alternatives.
Study Data and Methods
Our study uses two data sources: individual-level CPS data and employer-level F5500 data. The CPS provides detailed information about demographics, labor supply, income, education, employer characteristics, and workplace benefits. Form 5500 filings provide information about the number of participants in retirement plans, as well as about the plan’s sponsor, contributions, expenses, and assets. We extract the employer identification number (EIN) and collapse the data to the firm-year level.
Our identification strategy relies on the fact that auto-IRA programs were rolled out at different times in different states, and implementation within each state was staggered by firm size. These sources of variation provide us with a setting to estimate causal effects using several quasi-experimental econometric techniques. We estimate the impact of auto-IRA legislation on firm decisions to offer retirement benefits as well as on worker access to and participation in plans.
The results from our study indicate that auto-IRA legislation has a positive impact on the likelihood of employers offering retirement plans and individuals participating in these plans.
Using CPS data, we find that individuals in states that adopted auto-IRA legislation are 3.2 percent more likely to work for an employer who offers a retirement plan after auto-IRA policy implementation. Furthermore, there is a 7 percent increase in the probability of a worker participating in such a plan.
Using Form 5500 data, we find that the probability of a firm in an auto-IRA state offering any ESRP increases by 1.5 – 1.7 percent relative to firms in states without such programs. Worker participation in existing retirement plans rises by 3 – 5 percent.
We study the impact of state auto-IRA legislation on firm provision of ESRPs and worker access to such plans. Theoretical predictions are ambiguous about the prevailing direction of the potential effects. Using individual and firm data, our study concludes that auto-IRA policies significantly increase the likelihood of employers offering ESRPs, employee access to these plans, and participant numbers in existing ESRPs. Estimating consistent and complementary effects across two quite different data sources strengthens the robustness of these inferences.
It is important to note that determining the impact of auto-IRA legislation on overall saving or welfare is beyond the scope of this paper. In addition, this study does not identify the precise economic channel of these effects, which will be the focus of future work. However, these results are relevant for policymakers who wish to understand the impact of auto-IRA legislation on firm behavior and worker access to retirement benefits.
Adam Bloomfield is a Senior Economic Policy Advisor for the FDIC; Kyung Min Lee is an Economist for The World Bank; Jay Philbrick is a Student at Brown University; and Sita Slavov is a Professor at George Mason University and a Research Associate with NBER.
The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research, World Bank, FDIC, the U.S. Government, or the Georgetown University Center for Retirement Initiatives.
July 2023, 23-04
Bloomfield, Adam, Lee, Kyung Min, Philbrick, Jay, and Slavov, Sita, “How Do Firms Respond to State Retirement Plan Mandates? NBER Working Paper 31398, National Bureau of Economic Research, Inc., 2023.
Pew Charitable Trusts, State Automated Retirement Savings Programs Continue to Complement Private Market Plans, April 14, 2023.