The study, based on the experience of three large firms, also suggests that trust is critical in reducing the “opt-out” rate in automatic enrollment plans – and that both financial literacy and trust appear to have more sizeable marginal effects on that behavior than does income. The researchers say that plan sponsors who are experiencing high opt-out rates should consider ways to reduce employee mistrust of financial institutions – particularly among lower-paid workers.
The findings reinforce a paper released earlier this year by the Center for Retirement Research at Boston College that found financial literacy and trust have more sizeable marginal effects on 401(k) savings behavior in voluntary and automatic enrollment 401(k) plans than income (see Trust a Critical Factor in 401(k) Participation).
The researchers noted that higher financial literacy among workers is associated with higher voluntary 401(k) participation rates or lower quit-rates in automatic enrollment plans. In examining the motivations behind participant behavior it was noted that lower income workers do tend to have lower plan participation rates. They cited the work of Munnell, Sundén and Taylor (2001/2002) which, in this regard, proffered several reasons for this finding:
- Low income households are more likely to be financially constrained;
- They face lower (or negative) tax rates and derive little (or no) tax benefit from 401(k) saving; and
- They need to save less because of the progressive benefits structure of Social Security.
Additional research cited (Mitchell, Utkus and Yang, 2005) noted that plan features such as employer matches and access to funds through loans also appear to make 401(k) plans more attractive to employees and therefore are associated with higher participation rates. However, that same research found that the incentive associated with matching contributions “…explains only a small portion of aggregate plan participation, with employee own-preferences for savings being more important in determining 401(k) savings behavior.’
Another aspect that encourages participation is plan knowledge, according to the researchers. Choi, Laibson and Madrian (2005) found that just 21% of participants contributing below their plan’s match threshold knew what that rate was – half the 41% of those contributing above the match threshold. Furthermore, a study by Chan and Stevens (2006) found that knowledgeable participants were five times more responsive to plan features than the average individual. Finally, Bernheim and Garrett (1996) find that individuals that use employer-provided informational materials are significantly more likely to participate in their plan than those who do not receive them or simply do not use them. Studies by Clark and Shieber (1998) and Nyce (2005) found that increasing the quality of communication increased participation rates in plans.
The Vanguard research grouped “prior literature” on the topic of savings behavior into three broad categories (which, it notes, are not mutually exclusive):
- Hypothesis 1: Neoclassical Agents Theory: Under this theory, both quitters and non-joiners are forward-looking, rational agents and their decision not to participate is rationally explained by things like income-related budgetary limitations, their own preferences for tax-deferred saving, or by the belief that Social Security will provide enough retirement income.
- Hypothesis 2: Rational Information Gap Theory: Under this theory, quitters and non-joiners are held back by lack of information on basic savings and plan features and the transaction costs needed to accumulate that information. In other words, they can be reached by the messages, but they’re being held back.
- Hypothesis 3: Behavioral Theory: Under this theory, quitters and non-joiners are impeded by psychological biases such as procrastination and mistrust that interfere with a purely rational assessment of their savings choices.
The report notes that it is possible that in 401(k) plans one can be dealing with both rational and irrational behavior consistent with all three theories. “In fact,’ they say, “it is difficult to determine from prior research whether the decision to participate in a 401(k) plan is rationally motivated or a function of behavioral biases.’
Contrary to hypothesis one, it does not appear that non-participants think Social Security will be enough when they retire. Indeed, 94% of the non-participants in the automatic enrollment plans and 85% of non-participants in the voluntary plan disagreed or strongly disagreed with this statement. Nor are workers misled being too young to start saving. Ninety percent of the automatic enrollment, non-participants and 91% of the voluntary, non-participants disagreed or strongly disagreed with this statement – while 92% of the automatic enrollment, non-participants and 86% of the voluntary, non-participants disagreed or strongly disagreed with the comment that retirement is too far away to worry about saving for it now.
However, the answers by participants in the study suggest that one reason some people may not be participating is because they are financially constrained. Fifty-one percent of automatic enrollment, non-participants and 37% of voluntary, non-participants agreed or strongly agreed with this statement. As for relying on a spouse’s 401(k) plan to save for retirement, nearly half (49%) automatic enrollment, non-participants and 38% of voluntary, non-participants agreed or strongly agreed that this was the case.
The Vanguard study was based on an analysis of administrative data and survey data from three 401(k) savings plans that have similar plan features, and were offered by plan sponsors who operated in similar industries. Plan A relied on voluntary enrollment, while Plans B and C have an automatic enrollment feature. However, all three plans offer employer matches, catch-up contributions and the standard immediate vesting of employees’ contributions, though the employer match and the vesting schedules for these contributions vary by plan.
The automatic enrollment plans, Plans B and C, have higher new-hire participation rates, 85% and 78% respectively, than the voluntary plan, Plan A, at 63%. However, the researchers were struck by the much higher income of employees in Plan A. Mean income for Plan A employees is approximately $70,000, compared with an approximate mean income of $38,000 for Plan B and $34,000 for Plan C. “This disparity in income makes the higher new-hire participation rates under automatic enrollment for Plans B and C even more striking,” the researchers said.
Interestingly enough, more than a third (35%) of the non-participants in the voluntary enrollment plan incorrectly thought they were participating – though that perception could have arisen from the fact that employees in the plan receive a 4% employer contribution annually whether they make elective employee contributions or not.