Charlton “Eli” Freeman, and William Skimmyhorn
Review of Finance, Volume 29, Issue 5, September 2025, Pages 1587–1618, https://doi.org/10.1093/rof/rfaf027
Effectively saving for retirement requires workers to coordinate savings across different accounts (e.g., Social Security, IRAs, defined contribution plans) and to select appropriate contribution rates, tasks for which they may not be prepared. Fortunately, a notable body of behavioral research establishes that programs like automatic enrollment (Madrian and Shea 2001), which maximizes participation rates, and active choice (Carroll et al. 2009), which maximize contribution rates, may help. Given the tradeoffs among programs, research on “optimal” enrollment strategies is challenging and has been primarily theoretical. The notable exception to this rule is the Save More Tomorrow program (hereafter SMART, Benartzi and Thaler 2004), which asks individuals to commit to saving more for retirement in the future. SMART has enjoyed unparalleled success, and yet we know relatively little about which aspects of the program drive its results.
Motivated by the success of precommitment programs, we study a new set of light-touch and low-cost framing nudges on the short-run savings outcomes of current Thrift Savings Plan (TSP, the federal government’s 401k equivalent) participants and non-participants. The Department of Defense (DOD) conducted a large multi-arm field experiment that encouraged servicemembers scheduled for upcoming tenure-based pay raises to save more in the Thrift Savings Plan.
We find that emails (i.e., motivational deadlines, salary raise information or both) increase participation and contribution rates among previous non-participants, with suggestive evidence of effects among previous contributors. The effect magnitudes are small (around 1 percentage point on the extensive margin and 0.05 percentage points on the contribution rate margin), but often economically significant (i.e., 60-100% for non-participants in the treatment vs. control group), comparable to previous light-touch interventions (Benartzi et al. 2017) and similarly cost-effective. Our research contributes to the retirement savings literature by demonstrating that the effects of high-touch precommitment programs can be replicated using lighter-touch methods and decision framing, by evaluating the importance of “tomorrow” element of “Save More Tomorrow” programs, and by providing suggestive evidence on the role of loss aversion in precommitment decisions.