Richard Patterson is an Assistant Professor of Economics at Brigham Young University. His research interests include public economics, applied behavioral economics, and the economics of education. William Skimmyhorn is an Assistant Professor of Economics and Finance at the Raymond A. Mason School of Business at William & Mary. His research interests include household financial decision-making, human capital acquisition, and the economics of national security.
Many American households have no retirement savings, and a majority have too little saved to fund their expected needs. One potential reason is the widespread replacement of defined benefit (DB) pensions by defined contribution (DC) plans in the U.S. and elsewhere. As a result, workers have become increasingly responsible for their own retirement planning, generating too little savings overall and inequality for some.
Employers Promote Savings in Varied Ways
While foregoing costly DB plans, employers have turned to several strategies to promote their employees’ retirement savings, including providing more and better information (think of income projections during retirement for a given level of savings), simplifying employee choices (think of a simplified list of mutual funds and contribution rates), leveraging social norms (think of providing information on peers’ financial choices), establishing active-choice frameworks (requiring individuals to make a retirement savings decision instead of avoiding it), and automatically enrolling employees in plans (changing the default to an opt-out system).
Research has demonstrated positive impacts of all these strategies on retirement savings participation and savings rates, with increasing effects based on the intensity of the program intervention. So for example, simplification is helpful, but it appears to have a smaller overall effect than automatic enrollment.
What Works Best?
Gauging which approaches works best is challenging for two reasons. First, much of the past research compared approaches that occurred in different populations and environments, and at different times. For instance, it is difficult to compare a choice simplification intervention aimed young employees at a technology firm in 2016, to a social norm intervention aimed at older employees at a finance firm in 1997. Such comparisons cannot clearly determine whether differences in program effectiveness are real or simply due to differences in context. Equally important, very few studies have incorporated measures of costs into their analyses, so they cannot assess the cost-effectiveness of alternative strategies. One nice exception includes this research on the cost-effectiveness of behavioral “nudges” in a variety of settings.
Our current work draws on several natural and designed experiments affecting hundreds of thousands of people who work for the U.S. Army, to compare what worked best to boost savings in the Thrift Savings Plan (TSP). The TSP can be thought of as the military’s 401k plan (without matching) during our study period (2015-18). Specifically, we compared the effectiveness and cost of the several real-world options, including emails providing information, action steps, and target contribution rates; active choice enrollment; and automatic enrollment.
We did find meaningful effects of all these strategies on program participation, contribution rates, and cumulative contributions through six months of follow up. We also find that the sizes of these effects increased with the intensity of the program. For instance, providing contribution rate suggestions in an email increased participation in the TSP by about 9% compared to the control group. Requiring that employees make an active choice about participating in the TSP (actively opting in or out) increased participation by about 104%. And automatically enrolling employees into the TSP at hire boosted participation by 208%. Similar patterns resulted for contribution rates and cumulative contributions to the program. These results largely confirm existing estimates, providing useful scientific replication in a more controlled setting.
We also have new findings also based on a detailed cost-effectiveness analysis (following a method used in related research). For each treatment, we estimated the cost of the program per new enrollee, as well as the cost per new dollar of savings generated. For small, medium, and large firms, we found that active choice proved to be the most cost-effective enrollment program. That treatment cost about $11 per new enrollee, and $0.01 per new dollar saved. For very large firms (like the Department of Defense), automatic enrollment proved to be more cost effective. Specifically it cost about $0.02 per new enrollee and $0.001 per new dollar saved.
Making the Most of Employer-Provided Retirement Savings Programs
The American retirement savings landscape has changed, and employer-provided programs now facilitate millions of employees to save for retirement. Our research can inform plan providers and policymakers about which types of enrollment programs are both most effective and cost- effective. For most U.S. firms, however, active choice is the most cost-effective approach. Related research suggests that active choice may be especially preferred in settings where procrastination is likely and/or when employees have a wide range of savings preferences. Nonetheless, larger firms with greater resources could pursue automatic enrollment, to maximize participation and savings.
Views of our Guest Bloggers are theirs alone, and not of the Pension Research Council, the Wharton School, or the University of Pennsylvania.