Can Low-Cost Interventions Affect Retirement Saving Behavior?
The shift towards defined contribution pension plans has placed much of the responsibility and risks associated with saving for retirement on individuals. Yet a growing body of literature raises questions about how well equipped individuals are to make optimal savings decisions. For example, past studies have documented relatively low levels of financial literacy and the large impact that default rules have on saving decisions.
Employers have a number of potential levers at their disposal to affect individuals' retirement saving decisions, including the match rate; policy makers can change tax incentives or even impose a mandate. Communication interventions represent a simple, low-cost alternative to these mechanisms. But do such interventions affect saving behavior?
This question is explored in two new studies by NBER researchers. In What Will My Account Really Be Worth? An Experiment on Exponential Growth Bias and Retirement Saving (NBER Working Paper 17927) authors Gopi Shah Goda, Colleen Flaherty Manchester, and Aaron Sojourner conduct a large-scale field experiment designed to inform subjects about how a stream of retirement contributions would accumulate into an account balance at retirement and income in retirement and examine whether this information influences saving decisions.
Understanding how current contributions translate into retirement savings balances and retirement income requires knowledge of exponential growth and is a key element in optimal retirement planning. However, many individuals systematically underestimate the returns to saving that accrue from compound growth.
To conduct their experiment, the authors randomly assign employees of the University of Minnesota system to one of four groups. Three of the four groups received a brochure designed to encourage individuals to reflect on whether they are on target to reach their retirement goals, while the control group did not receive the mailing. Among the three groups, one received the basic brochure, while the other two groups also received a customized projection of how hypothetical additional contributions would translate into additional assets at retirement or into both assets and retirement income. Along with the brochure, individuals also received materials to assist them through the process of changing their contribution rate.
The authors find that providing income projections along with general plan materials resulted in an increased probability of changing contribution levels relative to the control group over a six-month period. Individuals in this group increased their annual contributions by $85 more than the control group on average, though since relatively few people changed their contribution, the magnitude of the increase among those who made a change was much larger, approximately $1,150 per year. The findings also suggest that providing planning materials without projections may have had a positive effect on contributions.
The study included a follow-up survey of subjects, which provided corroborative evidence that the intervention influenced saving decisions. The income group reported being better informed about retirement planning, more likely to have figured out how much to save, and more certain about their expected retirement income, as well as having higher overall financial satisfaction. The authors also find that the intervention was less effective for those with a higher discount rate, liquidity constraints, or procrastination tendencies.
The authors conclude "this study provides proof of concept for a policy that requires no additional mandate on individuals or subsidy for saving. Providing retirement income projections--an extremely low-cost intervention -- can actually affect individuals' saving behavior." They caution "the effects manifested were not large on average and were found in only a small share of the sample; thus, this policy is not likely to lead to a saving revolution. However, among those who made changes, effects were substantial and suggest that similar policies may help some individuals move closer to their retirement goals."
The study Small Cues Change Savings Choices (NBER Working Paper 17843) by authors James Choi, Emily Haisley, Jennifer Kurkoski, and Cade Massey also makes use of a field experiment to learn more about retirement saving decisions. The focus of this study is on learning whether inserting cues designed to activate phenomena identified in the psychology and behavioral economics literature into retirement saving materials would affect saving behavior.
To explore this, the authors randomly assigned employees of a large technology company to receive one of several versions of an email reminding them about the opportunity to change their 401(k) savings plan contribution before the end of the year. Control subjects received a message specifying their own contributions to date and reminding them to take advantage of the company match. Treatment subjects received the same message with an additional one- to two-sentence cue related to anchoring, goal setting, or a saving threshold. The added anchoring cue text read: "For example, you could increase your contribution rate by 1% of your income and get more of the match money for which you're eligible. (1% is just an example, and shouldn't be interpreted as advice on what the right contribution increase is for you)." The authors also randomized the numerical value used in the cue across subjects.
Turning to the findings, the authors report that during the first four months after the email, a low anchor (1 percent) decreased contribution rates relative to a control email with no anchoring cues by as much as 1.4 percent of income within a pay period, while high anchors (3, 10, or 20 percent) had no effect. In the longer run, the 1 percent anchor continued to depress contribution rates by up to 1.2 percent of income within a pay period, while the higher anchors increased contribution rates by up to 1.9 percent of income within a pay period. For the goal setting cue, a higher ($11,000) savings goal raised contribution rates significantly-by 2.2 percent of income at ten weeks after the email was sent-while a lower ($7,000) savings goal had little effect. In addition, highlighting high savings thresholds, such as the $16,500 annual limit on contributions or the 60 percent maximum contribution rate, led to larger increases in saving than mentioning lower thresholds.
The authors note that due to the constraints of their field setting, they "cannot establish beyond all doubt that the psychological mechanisms that motivated our cues are in fact responsible for our treatment effects." Nonetheless, if their treatment effects are driven by employees drawing inferences about their optimal contribution rates from the cues, "employees at this firm must have extremely diffuse prior beliefs about how much they should be saving in their 401(k)." They conclude "our paper's central message is, irrespective of the exact psychological channels through which they operate, small cues of the types we have tested have large effects on savings choices. Organizations and policymakers should be cognizant of these facts when designing their communications."
Shah Goda et. al. acknowledge funding from the Social Security Administration through a grant to the Financial Literacy Center, as well as from TIAA-CREF and from the University of Minnesota Carlson School of Management. Choi et. al. acknowledge funding from the National Institute on Aging (grant R01-AG-021650). At least one co-author has disclosed a financial relationship of potential relevance for this research; further information is available at http://www.nber.org/papers/w17843.ack.
