The Relationship Between Automatic Enrollment and DC Plan Contributions: Evidence from a National Survey of Older Workers

WP#2015-14

Abstract

Automatic enrollment has been widely embraced for raising employee participation in 401(k) plans. However, the empirical evidence is based on data with limitations that, up until now, have prevented researchers from extrapolating the effects of automatic enrollment to the broader population of workers. This paper reexamines the determinants of 401(k) participation and contributions in the presence of automatic enrollment using nationally representative data from the Health and Retirement Study (HRS) for 2006 through 2012. The results confirm previous findings that automatic enrollment is associated with a higher proportion of workers included in DC plans; however, automatically enrolled workers are less likely to contribute to their DC plans than voluntarily enrolled workers. Auto enrollment is also associated with lower employee contribution amounts and rates. However, the employers of auto-enrolled workers are more likely to contribute to their employees’ accounts than are the employers of voluntarily enrolled workers. Additionally, employer contribution amounts and rates are higher among workers who are automatically enrolled. Even so, the combined effect is that the retirement accounts of automatically enrolled older workers receive, on average, $900 less in combined annual contributions and have contribution rates that are 1.6 percentage points lower than those of voluntarily enrolled workers.

The paper found that:

  • Automatic enrollment is associated with a higher probability of being included in a DC plan.
  • On average, workers who are automatically enrolled in a DC plan tend to be less likely to contribute positive amounts than those who opt in.
  • However, the employers of automatically enrolled workers are more likely to make contributions and to contribute, on average, higher amounts and a higher percentage of their employees’ earnings.
  • The correlation between automatic enrollment and combined (employer and employee) contribution amounts and contribution rates, however, is still negative, despite controlling for a range of factors.

 
The policy implications of the findings are:

  • Auto enrollment could do a better job of boosting overall contribution levels among participants.
  • Possible ways to achieve this might be by offering a more generous employer match and by using auto escalation.
  • More research and better data are needed to assess the potential impact on retirement plan contributions of implementing automatic enrollment features in DC plans on a national scale.