A recent paper from the Center for Retirement Research at Boston College was titled “401(k) Plans in 2010: An Update from the SCF.” The SCF1 (perhaps better known to non-researchers as the Survey of Consumer Finances) is, as its name suggests, a survey of consumer households “to provide detailed information on the finances of U.S. families.” It’s conducted every three years by the Federal Reserve, and is eagerly awaited and widely used—from analysis at the Federal Reserve and other branches of government to scholarly work at the major economic research centers. The 2010 version was published in June.
The CRR analysis incorporated some of the SCF data (ownership of a retirement
plan account, participation, median 401(k)/IRA account balances, asset
allocations within those accounts, and distribution/loan patterns). The report
then brings in data from other sources on features such as automatic enrollment,
hardship withdrawals, and IRAs to complete its assessment, summarized on its
website as “Progress in the 401(k) system stalled in the wake of the economic
The summary went
on to note that “despite an increase in auto-enrollment, the percent of
employees not participating ticked up,” “401(k) contributions slipped,
while leakages through cash outs, loans, and hardship withdrawals
increased”—and that, “…the typical household approaching retirement had
only $120,000 in 401(k)/IRA holdings in 2010, about the same as in 2007.”
Setting aside for a moment the question of what “typical” is, a logical research
question arises: Are these statements a researcher’s extrapolation, or based
upon hard data, and thus “facts”?
What We Know
EBRI research has previously noted that, while the financial crisis of 2008
had a significant impact on retirement savings balances, as recently as just a
month ago, more than 94 percent of the consistent participants in the EBRI/ICI
401(k) database2 are estimated to have balances higher than they did
at the pre-recession market peak of October 9, 2007 (see “Returns
Engagement”). According to a number of industry surveys, participation
rates have remained relatively consistent, despite the soft economy and
tumultuous market environment, and EBRI research finds comparable trends in loan
activity (see “401(k)
Plan Asset Allocation, Account Balances, and Loan Activity In
2010”). Also, at year-end 2010, the data show that 401(k) loan balances
outstanding declined slightly from those in the past few years.
What else do we know?
We know that the number of future years that workers are eligible to
participate in a defined contribution plan makes a tremendous difference in
their at-risk ratings (See “Opportunity
We know that automatic enrollment, where deployed, has a significant positive
impact on retirement readiness (see “The
Impact of Automatic Enrollment in 401(k) Plans on Future Retirement
Accumulations: A Simulation Study Based on Plan Design Modifications of Large
Plan Sponsors”; “EBRI:
Auto-Enrollment Trend Boosts Retirement Readiness Ratings”).
We know that “averages” are easy to understand, and relatively easy to
calculate—but they don’t always provide an accurate3 portrayal of the
real world (see “Above
We also know that it’s important to understand the source and composition of
data—particularly when using self-reported results, drawing from multiple
sources, and creating a composite perspective.
Nevin E. Adams, JD
(1) The 2010 Survey of Consumer Finances is available here.
(2) As of December 31, 2010, the EBRI/ICI database included
statistical information on about 23.4 million 401(k) plan participants, in
nearly 65,000 employer-sponsored 401(k) plans, representing $1.414 trillion in
assets. The 2010 EBRI/ICI database covered 46 percent of the universe of 401(k)
plan participants, and 47 percent of 401(k) plan assets.
(3) When analyzing the change in participant account balances over
time, it is important to have a consistent sample. Comparing average account
balances across different year-end snapshots can lead to false conclusions. For
example, the addition of a large number of new plans (arguably a good event) to
the database would tend to pull down the average account balance, which could
then be mistakenly described as an indication that balances are declining, but
actually would tell us nothing about consistently participating