The EBRI research is the first using actual plan information on both actual auto enrollment and actual match rate information both before and after adoption of auto enrollment. The new EBRI analysis uses plan-specific data for large employers from Hewitt Associates, and finds that employers instituted more generous contribution rates after adopting automatic enrollment, and did so when measured by several different standards.
This release provides preliminary results from the research. Full results of the simulated impact of the changes on worker’s retirement wealth will be available next month in the February 2010 EBRI Issue Brief, and will be posted online at www.ebri.org
“Our recent analysis of plan-specific data shows that, at least among large 401(k) plans, plan sponsors actually increased the generosity of their contribution rates,” said Jack VanDerhei, EBRI research director and author of the analysis. He noted that EBRI has published three separate simulation studies since 2005 showing the potential importance of automatic enrollment for future retirement account accumulation, even under the assumption that employer contribution rates would not change when the 401(k) plan was modified.
“The modifications to 401(k) plans made by sponsors in response to the Pension Protection Act of 2006 will be very important for retirement income adequacy in this country. Adding these more realistic assumptions to our simulation model will allow us to more accurately demonstrate the relative improvements in retirement accounts, especially for young and low-income workers,” he added.
The EBRI results contradict an earlier publication of the Center for Retirement Research (CRR) at Boston College and by the Urban Institute, which concluded that among a sample of large 401(k) plans, match rates are lower among firms with automatic enrollment than among those without automatic enrollment after controlling for firm characteristics.
However, there were two major limitations with the CRR/Urban published analysis:
• The study was based on U.S. Department of Labor Form 5500 data that does not include specific information on 401(k) match rates. Instead, the authors constructed an estimate for the match rate as the ratio of employer-to-employee contributions for each 401(k) plan.
• They merged the Form 5500 data with information on automatic enrollment from the Pensions &Investments database of the top 1,000 pension funds, which includes a flag indicating whether plan administrators reported offering automatic enrollment in their defined contribution (401(k)-type) plans.
However, this database does not report the year that the automatic enrollment provision was adopted, so there is no way to tell from this data source how long auto-enrollment had been implemented in a planThe authors of the CRR/Urban published study present a regression analysis on this database and produced a finding that: suggests a negative relationship between automatic enrollment and match rates and is statistically significant at the firm-level. In particular, match rates are about 7 percentage points lower among firms with automatic enrollment than among those without automatic enrollment, after controlling for firm characteristics.
While the authors correctly point out that although the regressions suggest a relationship between automatic enrollment and match rates, they do not necessarily imply that auto enrollment causes lower match rates; however, this crucial qualification has been generally ignored in recent press accounts of the study.
Earlier EBRI Research
The CRR/Urban published study conflicts with previous EBRI research published in 2007,2 which surveyed Mercer Human Resource Consulting defined benefit sponsors to gauge their recent activity and planned modifications to their defined benefit (pension) and defined contribution (401(k)-type) plans. The EBRI survey also was able to determine what, if any, increases in employer contributions to defined contribution plans were provided in conjunction with reductions to their defined benefit plans.
Although the association between the adoption of automatic enrollment and employer contributions to 401(k) plans was not the focus of the EBRI study, one-third of the defined benefit sponsors surveyed indicated that they had already increased or planned to increase their employer match to a defined contribution plan, and 20.9 percent indicated that that they had already increased or planned to increase their nonmatching employer contributions to a defined contribution plan. There was some overlap between the two groups, but overall, 42.5 percent of the defined benefit sponsors surveyed indicated that they had already increased or planned to increase their employer match and/or nonmatching employer contribution to a defined contribution plan. This was particularly true of defined benefit sponsors that had either closed a defined benefit plan to new hires or frozen the plan to all members in the last two years or planned to do so in the next two years.
Moreover, the 2007 EBRI study found an extremely large correlation between the adoption of automatic enrollment for a 401(k) plan and the freezing or closing of the defined benefit plan.4 Of those defined benefit sponsors that had closed their defined benefit plans in the last two years, 80.5 percent had either already adopted or were currently considering adopting automatic enrollment features for their 401(k) plans. Of those defined benefit sponsors that had closed their defined benefit plans in the last two years, 76.1 percent had either already adopted or were currently considering adopting automatic enrollment features for their 401(k) plans.
EBRI’s New Research and Methodology
The new EBRI study analyzes in detail 225 large defined contribution plans6 that had adopted automatic enrollment 401(k) plans by 2009, but did not have them in 2005 (the last observation that was not influenced by PPA). The following information was coded for each plan:
• The default contribution rate for the automatic enrollment (AE) plan in 2009.
• The entire match rate contribution formulae for both years.
• All nonelective contributions paid to the defined contribution participants by the employer.
Whether plan sponsors were more or less generous after adopting AE was measured with three different metrics:
1. The average 2009 first-tier match rate was 87.78 percent, while the average 2005 first-tier match rate was 81.26 percent. The difference of 6.52 percentage points suggests that, to the extent that this sample is representative of the universe of large 401(k) sponsors, those sponsors adopting AE were more generous to the 401(k) participants when measured by this variable after automatic enrollment was implemented than they were before.
2. The average effective match rate8 for 2009 was 4.32 percent of compensation, but only 4.00 percent of compensation in 2005. The increase of 0.32 percentage points again suggests that large 401(k) sponsors adopting AE were more generous to the 401(k) participants when measured by this variable after the adoption of automatic enrollment than they were before3. The average total employer contribution rate9 for 2009 was 6.35 percent of compensation and 5.46 percent of compensation in 2005. The increase of 0.89 percentage points once more suggests that those large 401(k) sponsors adopting AE were more generous to the 401(k) participants when measured by this variable than before.
Influence of Defined Benefit Plan Activity
This information was then combined with the defined benefit information for the same sponsor in an attempt to analyze whether EBRI’s 2007 findings of the association between defined benefit freezing/closing and enhanced 401(k) contributions were corroborated. The average improvements for all three metrics were much higher for sponsors that had frozen/closed their defined benefit plans than for the overall average. For example, the change in the total employer contribution rate for all frozen plans was 1.64 percent of compensation versus 0.89 percent for the overall average. Employers that had closed their defined benefit plans to new employees had an even larger average improvement: 2.82 percent of compensation.
The defined benefit sponsors that had frozen or closed their plans were then split into those that had done so prior to adopting AE and those that had changed their defined benefit plans between 2005 and 2009. If the hypothesis that the 401(k) improvements were a result, at least partially, of a simultaneous quid pro quo for the decreased accruals in the defined benefit plan, one would expect that the earlier improvement would be smaller than those that took place approximately at the time of the conversion to AE. In fact, this is exactly what is found for all six comparisons. For example, the average total employer contribution improvement for firms that had frozen their plans prior to 2005 was 0.69 percent of compensation, compared with 2.45 percent for those that froze between 2005 and 2009. Similar evidence is found for those that closed their pension plans to new employees: The average improvement in total employer 401(k) contribution was only 0.56 percent of compensation for those that closed prior to 2005, but 3.34 percent for those that closed the plan between 2005 and 2009.
-Employee Benefit Research Institute