How to Increase 401(K) Saving

In their decisions about 401(k) saving, employees often "follow the path of least resistance." This is the conclusion of James Choi, David Laibson, Brigitte Madrian and Andrew Metrick in a new NBER study entitled "Defined Contri-bution Pensions: Plan Rules, Participant Decisions, and the Path of Least Resistance" (NBER Working Paper 8655). Employees generally do whatever takes the least effort -- generally doing nothing -- a phenomenon these investigators call "passive decision-making." For better or worse, 401(k) plan administrators can manipulate the path of least resistance -- through automatic enrollment rules, default contribution rates, and other plan design features -- to powerfully influence the savings and investment choices of their employees. The findings also suggest the importance of 401(k) design decisions in insuring that employees save enough for their retirement.

The study looks at the 401(k) investment decisions of about 200,000 employees at seven U.S. companies, most of which have made one or more changes to their 401(k) plans over time. The use of data from multiple companies over a period of years enables the investigators to compare saving decisions across companies with different plan design features, and to analyze how saving behavior changes in response to plan design changes.

One aspect of plan design explored in the study is whether or not there is automatic enrollment of new employees. The typical 401(k) plan requires an active election on the part of employees to initiate participation. A few plans have implemented automatic enrollment. Note that automatic enrollment does not require the participation of employees -- it just makes participation the default decision. As illustrated in figure 1, automatic enrollment dramatically increases the percentage of workers who participate in the plans. Fewer employees enroll if they need to "do something" to participate, as compared with programs where they have to "do something" not to participate. The path of least resistance matters a lot.

Two related aspects of 401(k) plan design are the default contribution rate and the default investment allocation associated with automatic enrollment. If new employees "do nothing," how much of their salary will be contributed automatically to the 401(k) plan, and how will it be invested (among stock funds, bond funds, company stock, and other investment options)? While employees always have the option of changing their contributions, these default provisions also have a very significant influence on what people do. A large proportion of automatically enrolled new employees accept both the default contribution rate and the default investment allocation associated with the plan. Figure 2 shows the percentage of new enrollees who chose to invest at the default contribution rate before and after it was implemented at three of the companies in the study. It is important to note that the defaults could raise or lower 401(k) contributions, depending on where the default rate is set. At these companies, the default was set quite low (two to three percent of salary), and so the contribution rates of new enrollees became lower after the defaults were implemented, even as many more new employees participated in the plans.

The asset allocation defaults are also important as they determine whether people's 401(k) money will be invested in assets that are likely to appreciate a lot over the long-term (such as stock mutual funds), or assets with more modest (though more stable in the short-term) returns, or assets with unusual risk (such as company stock). In this case, too, employees tend to follow the default asset allocations associated with their plan.

A fourth aspect of plan design explored in the study is whether or not automatic distributions of 401(k) assets are made to employees who leave their jobs. Yet again, the path of least resistance dominates. Employees with cash balances be-low $5000 are often given automatic cash distributions when leaving their jobs; most of these employees consume this money, rather than rolling it into an IRA or another retirement plan. By contrast, employees with cash balances above $5000 rarely receive automatic distributions; most of these employees leave the money just where it is -- accumulating in their 401(k) account.

A fifth aspect of plan design is the rate at which employers match employee contributions, and the match threshold (the maximum amount of contributions that the employer matches). An employer match is found to increase 401(k) participation, and the match threshold is an important focal point in the selection of employee contribution rates. Increasing the match threshold can raise contribution rates, particularly among those with lower prior saving rates. This makes sense, since there are real and immediate losses to the employee from contributing below the match threshold.

A final interesting aspect of the study is the role of financial education in the workplace, such as retirement planning seminars. The effect of these programs in altering saving intentions is found to be significant. However, their effect on actual saving is modest at best -- and far smaller than the plan design characteristics already noted. Even after workplace education, and despite the best of intentions, "the path of least resistance" is still to do nothing. This contrasts with a program at one of the companies in the study that implemented a pre-commitment program in conjunction with workplace education. This program pre-commits employees to allocate a portion of future raises toward higher 401(k) contributions. With the pre-commitment in place, the good intentions of the employee became the default, and increases in saving did occur.

It is clear from these findings that companies can do quite a lot more to promote retirement saving than just implement a 401(k) plan, and counting on employees to take it from there. Plan design decisions like automatic enrollment, higher default contribution rates, default investment allocations with higher expected long-run returns, and pre-commitment mechanisms to in-crease future contributions can have a huge influence.

The research was funded by the National Institute on Aging, the MacArthur Foundation, the Sloan Foundation, and a National Science Foundation Graduate Research Fellowship. It was summarized by Richard Woodbury.