Plan sponsors finally have statistical proof that automatic enrollment can narrow the persistent racial gap in retirement savings. Unfortunately, other stats from the same research show that auto enrollment can widen the gap.
In a new study of seven large clients by Vanguard Group, those that automatically put new hires into a 401(k) plan increased the participation rate by more than 60 percent among black employees — to 94 percent from 57 percent — and by more than 40 percent among Latino employees, to 95 percent from 67 percent. As a result, the two groups statistically caught up with the rates of white and Asian employees.
“This is another reason for introducing automatic enrollment if you haven’t offered it already,” says Stephen Utkus, head of Vanguard’s Center for Retirement Research and a co-author of the study.
The catch is that the companies allocated an average of just 3.2 percent of employees’ paychecks to the 401(k)s, a typical amount in these automatic programs but far less than retirement experts say is needed. After enrollment, black and Latino employees pretty much stayed at the default level. However, white and Asian employees increased their allocations, widening the racial disparity again.
“Automatic enrollment is a good step,” says Pamela Hess, director of retirement research at consulting firm Aon Hewitt. “But it’s not going to bridge the gap completely.”
“With automatic enrollment, differences in participation and savings are much reduced,” says Annamaria Lusardi, a professor of accounting and economics at the George Washington University School of Business. “But it might not eventually be the best way to help minorities.”
Thus, according to Lusardi, employers will have to take more steps, perhaps coupled with education starting in elementary school.
For years surveys by Aon Hewitt and others have shown that blacks and Latinos participate in 401(k)s at lower rates and save significantly less than whites and Asians in their plans. This holds up even after controlling for factors like age, income and job tenure. For instance, in the Vanguard study the participation rates for whites and Asians without auto enrollment were 73 percent and 90 percent, respectively — way above the 57 percent and 67 percent of blacks and Latinos.
Making matters worse, the latter two groups were more likely to reduce their nest eggs through withdrawals and loans, according to a survey of 3 million employees at 57 large companies that Hewitt Associates (as it was then named) and Ariel Investments conducted in 2007.
Finance and business experts say the explanations are complex, a mixture of socioeconomics and culture.
Hal Hershfield, an assistant professor of marketing at New York University’s Leonard N. Stern School of Business, is researching whether “the sense of power that one has in daily life can actually affect spending and saving behavior.” Another theory holds that if people see few members of their ethnic group living to old age — perhaps because they can’t afford medical care — it may be hard for them to envision a need to save for retirement.
Among Latinos, says Aon Hewitt’s Hess, supporting relatives often takes priority over saving for retirement.
Both Hess and Vanguard’s Utkus suggest that black and Latino employees may have less faith in financial markets. Related to that, employees from poor backgrounds, a group that includes large numbers of black and Latino workers, are less likely to learn financial basics from family, friends and employers, says George Washington professor Lusardi.
When Boston College’s Center for Retirement Research analyzed the Federal Reserve’s triennial Survey of Consumer Finances, which polled 4,500 U.S. households for 2001, 2004 and 2007, it found that factors like a college degree, home ownership and nonpension wealth were the most important determinants of 401(k) savings.
Although both blacks and Latinos generally lag the other two ethnic groups, Latinos tend to do slightly better than blacks in some categories. Their auto enrollment jump in the Vanguard survey was less dramatic because they started from a higher level of participation, and they were much less likely to make withdrawals and loans in the Ariel/Hewitt survey.
One possible reason, Hess suggests, is that black households may have fewer breadwinners, while Latinos may live with extended families. NYU’s Hershfield cites the social psychology concept of “stereotype threat.” He explains, “When people are aware of the stereotypes that exist about their group, they get anxious and flustered when they’re making decisions in the domain that stereotype applies to.” And African-Americans are more likely than Hispanics or other groups to be stereotyped as being bad at math and finance, Hershfield says.
Ariel/Hewitt and Vanguard are now respectively conducting and planning follow-up studies to delve further into the explanations.
If sponsors understood why some ethnic groups were more likely than others to join 401(k) plans, could they do something about it?
The old standby of workplace-based, targeted education could be one way to change behavior. “You’re not going to say, ‘We only want African-Americans at an educational meeting at noon,’?” Utkus points out. But an employer can schedule a meeting just for people who haven’t enrolled in the 401(k), a group that is likely to encompass a large percentage of African-Americans.
Most important is for schools to teach financial literacy every year, starting “as early as possible,” just like they teach the three Rs, says George Washington professor Lusardi.
And what about automatic enrollment? The Vanguard survey would seem to show that it works. Auto enrollment brought all four groups to statistically equal participation levels: 98 percent for Asians, 95 percent each for Latinos and whites, and 94 percent for African-Americans.
True, the Vanguard sample was small and self-selecting, and the companies had particularly “well-designed and generous retirement programs,” as the study puts it. Still, Utkus and co-author Cynthia Pagliaro say their results are valid because these companies have diverse workforces, with blue, pink and white collars, and also support auto enrollment as at least a partial solution.
The real problem is the low default savings rate. Plan sponsors typically auto-enroll people at about 3 percent, and history shows that inertia will keep most participants at whatever level they’re put at. However, experts recommend saving at least enough to max out on the company match — usually 6 percent of pay — or even as much as 15 percent.
One solution is automatic escalation: methodically increasing the contribution rate by, say, 1 percent annually. “But if companies set the default too high,” warns Hershfield, “people really feel it in their paycheck,” and then they will overcome their inertia to bolt the plan.
Except, apparently, whites and Asians, who boosted their auto defaults to an average 4.4 percent and 5.6 percent, respectively, in the Vanguard report.
