As sponsors of 401(k) plans, employers can boost plan participation by automatically enrolling new employees in plans–a practice made possible by the Pension Protection Act of 2006). They can also raise participation and contribution rates by matching a portion of each employee’s contributions.
There’s been some debate over the years about which factor—auto-enrollment or the (more expensive for the employer) match—drives participation more. A study by a team of Harvard and Yale economists in 2007 showed that most auto-enrolled participants will stay in a plan even if the employer suspends its match.
The match may be more important than previously thought, however. The results of a recent study by analysts Nadia Karamcheva and Justin Falk of the Congressional Budget Office’s Microeconomic Studies Division found that “most of the estimates from the literature substantially understate the effect of matching.”
The analysts took advantage of two natural experiments. Before 1984, the federal government offered only a defined benefit pension (without Social Security). In 1984, it began offering federal employees a defined contribution plan (the Thrift Savings Plan or TSP) with a match. It allowed people under the old CSRS system to also contribute to TSP, but with no match. This change provided an opportunity to test the impact of a match on contribution rates.
The second natural experiment took place in August 2010, when the government implemented a policy of automatic enrollment with a default contribution rate of three percent and the “G Fund” as the default investment option. (The G Fund invests in government securities. Its yield is based on the yield for Treasury notes.) This change provided a test of the impact of auto-enrollment.
A microeconomic analysis of the results of these two natural experiments showed that the match had a bigger effect. It increased contribution rates by 22 percentage points. Auto-enrollment increased it by 19 percentage points.
Looking at the long-term impact of the matching contribution, the analysts found that for those with a match, the average ratio of balance to pay was 2.5 to 1 (after an average accumulation period of 28 years). The average ratio of balance to pay for those without a match was 0.8 to 1. Looking at the impact of auto-enrollment (over an average accumulation period of five years), the average ratio of balance to pay was the same (0.4 to 1) for those who were hired just before and just after auto-enrollment was introduced.
In the 2007 Harvard-Yale study, the economists studied the behavior of participants whose plan sponsor switched from a matching contribution to a voluntary employer contribution not contingent on a worker’s contribution. They found that participation rates declined by “at most five to six percentage points” and average contribution rates fell by 0.65%.
Our “results suggest that the match has only a modest impact on opt-out rates,” wrote John Beshears, David Laibson and Brigitte Madrian of Harvard and James Choi of Yale in a 2007 paper, “The Impact of Employer Matching on Savings Plan Participation under Automatic Enrollment.” The same team also looked at data from nine different employers who all used auto-enrollment and varying match structures. It drew similar conclusions.
“We find that a one percentage point decrease in the maximum potential match as a fraction of salary is associated with a 1.8 to 3.8 percentage point decrease in plan participation at six months of eligibility,” the paper said. “We estimate that moving from a typical matching structure of 50% up to 6% of pay contributed to no match would reduce participation under automatic enrollment at six months after plan eligibility by 5 to 11 percentage points.”
The CBO and Harvard studies are quite different, so it’s impossible to say which carries more weight. CBO examined the effect of a match on workers who did not have auto-enrollment, whereas other researchers have looked at the effect of taking away the match from workers who have auto-enrollment.
© 2018 RIJ Publishing LLC. All rights reserved.