Who Uses the Roth 401(k)?

Since 2006, plan sponsors have had the opportunity to add a Roth 401(k) option to defined contribution pension plans. Takeup among plans has been fairly rapid, with half of 401(k) plans offering a Roth option by 2011. Yet usage of the Roth 401(k) option by participants has lagged behind, with fewer than one in five 401(k) plan participants making Roth contributions in that year. In Who Uses the Roth 401(k) and How Do They Use It? (NBER Working Paper 19193), researchers John Beshears, James Choi, David Laibson, and Brigitte Madrian explore the usage of the Roth 401(k).

Pension plan participants may have the option of making three different types of 401(k) contributions: Roth, before-tax, and after-tax. Before-tax contributions are deductible from current-year income, but the principal, interest, and capital gains are taxed at the ordinary income tax rate upon withdrawal. By contrast, Roth contributions are not deductible, but the principal, interest, and capital gains may be withdrawn tax-free if the withdrawal is "qualified" (occurs after age 59.5 and the account has been held for five years). After-tax contributions are also not deductible; at withdrawal, interest and capital gains (but not principal) are taxed at ordinary rates.

Due to these differing tax treatments, making contributions to a Roth 401(k) is a better financial deal than making before-tax contributions for workers who expect to face a higher tax rate in the future (after age 59.5) than at the current time. This might be the case for a worker who is young and expects future earnings growth, has temporarily low income this year (for example, due to a spell of unemployment), or has large deductions that reduce current tax liability (for example, a large mortgage or many children under age 18). If tax rates were to rise over time, additional groups of workers might also experience higher tax rates in retirement than at the current time.

As the authors note, other factors can affect the relative appeal of the Roth. The Roth is more attractive if workers anticipate a possible withdrawal before age 59.5, since Roth principal is exempt from the 10 percent early withdrawal penalty that applies to before-tax contributions. Moreover, workers who are constrained by the limits on total 401(k) contributions may find it advantageous to make Roth contributions because a dollar of Roth balances buys more retirement consumption than a dollar of before-tax balances. However, if the employer offers a match, the employee can earn more match dollars by making a before-tax contribution.

For their empirical analysis, the authors use administrative data for twelve large companies that introduced a Roth option to their 401(k) plans between 2006 and 2010.

The authors first report that 8.6 percent of participants in their sample use the Roth in the first year after this option was introduced. This figure rises to 11.5 percent excluding companies that automatically enroll employees with default contributions that allocate nothing to the Roth. While the fraction of employee balances held in the Roth is unsurprisingly low after only one year of contributions, contribution flows into the Roth are also modest one year after the Roth's introduction, at 5.4 percent of flows for the full sample and 8.5 percent of flows for workers without auto enrollment.

However, conditional on being used, Roth contributions average nearly two-thirds of the employee's contribution. A majority of Roth users are actively engaging in tax diversification by simultaneously making contributions to both the Roth and another 401(k) account. Among this group, very few follow a naive "50-50" rule of putting half of their contributions in each type of 401(k).

Since many past studies have shown that workers are passive in their retirement savings decisions, the low usage of the Roth may reflect a slow response to its introduction rather than an active preference for making before-tax contributions. To explore this, the authors examine how Roth participation differs between 401(k) participants who were hired before and after the Roth introduction at their firm. Workers hired after the Roth became a plan option are likely to be more aware of it and are able to choose the Roth at the time they are making their initial pension plan decision, so the cost of enrolling in the Roth is lower.

Indeed, the authors find that for workers hired after the Roth's introduction, 19.0 percent have positive Roth balances and 14.3 percent of employee contributions are flowing to the Roth, versus 7.9 percent with a positive balance and 4.7 percent of employee contribution flows for workers hired prior to the Roth's introduction.

Finally, the authors explore the demographic covariates of Roth usage within the 401(k) participant population. They find that Roth contributors are younger and more likely to be male. Salary has a negative relationship with Roth contributions for workers hired after the Roth's introduction, as expected, but a positive relationship for those hired beforehand. The authors note that this unexpected finding may be due to higher-income workers being less financially passive in general, making them more likely to update their 401(k) contribution elections when the Roth is introduced. In addition, higher-income workers have greater financial literacy and thus are more likely to be familiar with the Roth.

The authors conclude by noting that Roth 401(k) usage is relatively uncommon among workers in their sample of firms, although participation is more than twice as high among workers who were hired after the Roth introduction. As they write, "because of passivity or inattention, 401(k) participants do not react quickly to the Roth option when it is introduced after they have already joined the 401(k) plan."

The authors acknowledge financial support from the National Institute on Aging (grants R01-AG021650 and P01AG005842) and the Social Security Administration (grant FLR09010202-02 through RAND's Financial Literacy Center and grant #5 RRC08098400-04-00 to the NBER as part of the SSA Retirement Research Consortium). At least one author has disclosed a financial relationship of potential relevance for this research; further information is available at http://www.nber.org/papers/w19193.ack.