Are Auto Features Falling Short of Expectations?

While automatic enrollment and automatic escalation have been touted as effective strategies to get more people into retirement plans and save at higher rates, a new paper published by the National Bureau of Economic Research found that these features may not be as impactful on generating long-term wealth as previously estimated.

David Laibson, a professor of economics at Harvard University and one of the authors of the paper, spoke at the Defined Contribution Institutional Investment Association Academic Forum on Wednesday in New York City about how high employee turnover rates, as well as a high rate of 401(k) leakage upon job separation and low acceptance of auto-escalation defaults, diminish some the effectiveness of automatic features.

The paper, “Smaller Than We Thought? The Effect of Automatic Savings Policies,” examined nine firms that, some time between 2005 and 2011, introduced either auto-enrollment, default auto-escalation or auto-enrollment and auto-escalation simultaneously. Because automatic policies applied only to employees hired from a certain date onward, the study compared 62,430 employees hired in the year after the policy introductions to 55,937 employees hired in the year before.

“Our view is that the auto features that we have now in the U.S. system are a net positive for our workers,” Laibson said. “It’s just that they’re not producing long-run wealth formation.”

The ‘Stickiness’ Factor

When employees move from job to job, Laibson explains, auto-escalation does not have the same “bite” it would if a person stayed at the same employer. For example, an employee may be automatically enrolled into a plan at a 3% contribution rate and eventually, due to auto-escalation, increase that contribution to 6%. But if they switch employers, there is a high likelihood that they will once again be defaulted into a new plan at 3%, so that previous auto-escalation has little impact.

Auto-escalation is also a feature that many participants are opting out of, and Laibson said it lacks “stickiness.”

Using data from Alight Inc., the researchers found that about 43% of employees defaulted into auto-escalation are sticking to the solution for the first year they are escalated. But at the second year of escalation, only 36% are sticking with it. Only 29% stick at it for the third year.

“I would emphasize, even if we increase the stickiness, we’re still not going to see the effects of auto-escalation in the status quo system because of the high turnovers, because people are ramping up and then going back to 3%, ” Laibson said. “So the whole idea that you can stay in and ramp up to 10% and stay there, or ramp up to 15% and stay there, [is] just not the experience for the bottom third of the workers in the U.S.”

Difference in Outcomes Not That Drastic

When studying employees with one year of tenure and using a “naïve” assumption that 85% of participants utilize auto-escalation, Laibson showed that those with automatic enrollment would experience a 2.2-percentage-point increase in their household savings rate. However, this increase shrinks to 1.84 percentage points when looking at data for employees with five years of tenure—likely because people are not actually sticking with auto-enrollment. The increase then shrinks to 1.48 percentage points after incorporating data about people who fail to fully vest in their employer contributions.

Because people are moving jobs every one to two years, Laibson said many are leaving employer plans before they are fully vested, thus not receiving the full employer match.

After incorporating data about withdrawals and other forms of 401(k) leakage after people separate from their employer, the increase in savings rate shrinks again to 0.63 percentage points. According to the paper, 42% of 401(k) balances are cashed out upon departure from a job, and many people are not rolling over the funds into an individual retirement account or a new employer’s plan, but instead are taking it as a cash distribution, especially if they have lower account balances.

Laibson noted that auto-enrollment tends to be most beneficial for low-income workers, as it helps them to save in the first place, but at the same time, setbacks like 401(k) leakage and non-vesting are impacting these workers the most.

Potential Improvements

One audience member asked Laibson if higher initial default rates would help solve the issue of employees reverting back to their lower, original rate when they switch employers. Laibson said increasing default rates to the 6% to 10% range would be a step in the right direction. However, he noted that a 10% default rate is not right for people at the bottom of the income scale, and employers must consider that depending on how it is structured, the employer match might be more expensive with higher defaults.

Another audience member asked if implementing age-based default rates would be an effective strategy. Laibson countered that income-based default rates would be better, arguing that younger workers sometimes have more of an ability to save than older workers who are preparing to have a family and may not have as much expendable income to contribute to retirement.

At the policy level, Laibson said a solution would be to have some sort of aggregated database that would show employers what employees were contributing at their previous job, in order for them to nudge or encourage new employees to maintain or raise their contribution rate.

“We should think about ways of knitting together our employment landscape so that what happened at the previous employer can help inform what’s going to happen at the next employer,” Laibson said.

Lastly, he said a key issue is what employees do with their retirement savings when they separate from a job. He said employers should be doing more to communicate and educate workers about keeping their funds in the retirement savings system when leaving a job.