Seven Surprising Findings

Back in the mid-1990s, when Michael Sherraden’s seminal book, Assets and the Poor, was beginning to make its mark, the prevailing wisdom in the poverty field was that the poor couldn’t save, so why bother? If they could save, they wouldn’t be poor, right? Thankfully, well-endowed foundations didn’t believe this, but many left-leaning academics and nonprofit leaders (sorry, I won’t name names) were actually quite dismissive of the idea. They believed that they knew best what the poor needed and what they were capable of. We certainly hoped to surprise them.

But everyone agreed that proof of concept was necessary, and thus was born the American Dream Demonstration (ADD), which tested approximately 2,400 Individual Development Accounts (IDAs) over the course of about two years in 13 sites nationwide.

When researchers reported early findings that the poor did save in IDAs, skeptics were surprised and began to take the idea seriously. Some policymakers were so inspired by this finding and so eager for new ideas to help the poor—remember, this was when welfare as we knew it came to an end—that multi-billion dollar savings proposals emerged from Capitol Hill and the White House, again surprising just about everyone. We did not secure those billions, but did manage to get the $25-million-per-year Assets for Independence Program established in 1998, a modest though important step forward. Over 10 years later, we continue to be surprised—and sometimes humbled—by what we’re learning about saving and asset-building by the poor. Here are a few of those lessons.

1. The poor save, and the poorest save more.

Households earning 200 percent of the poverty line (the maximum allowed) saved about 1 percent of their income, while those earning below 50 percent of the poverty line saved about 3 percent. In fact, not only did income not predict saving, but most other demographic features—age, gender,race, employment status and welfare receipt—did not strongly predict saving either. In other words, a broad range of low-income people saved successfully in IDAs. How was this possible? Researchers at the Center for Social Development, which Sherraden directs, believe that what really matters is who has access to a structured saving mechanism—an employer, a nonprofit, automatic payroll deductions, or easy ways to save and build assets at tax time. In fact, that’s how most wealth in this country is accumulated.

2. Low-income families want short-term, unrestricted savings, not just savings restricted to long-term, productive assets.

When 64 percent of participants in ADD made an unmatched withdrawal (i.e., they forfeited a 2-1 match just to get through some crisis), the field realized that low-income families need short and long-term savings, and that IDAs could not meet all of those needs. Since then, promoting access to unrestricted savings (at tax time, from automatic payroll deductions, through prize-linked savings, etc.) has become a key goal in the assets field. Low levels of flexible savings force many struggling families to rely on expensive, wealth-depleting alternative financial services providers (pay-day lenders, check-cashers and others). The Consumer Federation of America found that low-income families with $500 in emergency savings had better financial outcomes than moderate-income families with lower savings. Also, the Urban Institute found that households that are “liquid-asset poor” are two to three times more likely than those with liquid assets to experience “material hardship” after a job loss, health emergency, death in the family or other adverse event.

3. Matches matter less than we thought.

One of the most powerful findings in ADD was that a $1 increase in the match cap (the maximum amount of savings that would be matched) was associated with a 57-cent increase in savings. That’s a huge effect. Researchers believe that many savers turned that match cap into a savings goal, increasing their saving. In contrast, higher match rates (e.g., a 2-1 match instead of 1-1) tended to encourage people to start saving, but higher matches also resulted in less monthly savings since participants could reach their goal with less of their own savings. Lesson: Personal goals matter. If the goal is to increase saving, then higher match caps are likely to be more effective than higher match rates. However, higher match rates might result in faster and increased asset accumulation.

4. Assets matter more than we thought.

Sherraden was more right than he predicted when he posited that owning, as distinct from earning, would generate a broad range of good social outcomes. For example, William Elliot III of the University of Pittsburgh and Sondra Beverly of the Center for Social Development, in trying to understand why youth who expect to go to college do not, discovered that, remarkably, those with accounts in their names are seven times more likely to attend college than those lacking accounts. Trina Williams Shanks of the University of Michigan reported that children who grow up in homes with assets have lower rates of teen pregnancy, fewer behavioral problems, better self esteem, more confidence and a future orientation. Interestingly, the simple presence of the asset, and not the amount, seemed to matter most. Assets matter, too, for driving opportunity. New York University’s Dalton Conley found that, controlling for income, education and family background, “it is really net worth that drives opportunity for the next generation.” And Pew’s Economic Mobility Project found that financial capital, along with family structure and educational attainment, are the three strongest predictors of economic mobility in America.

5. Financial education matters, but defaults matter more.

In ADD, every hour of financial education was correlated with greater saving, but only up to 10 hours. This is an important finding, but the effects of establishing the right “defaults” in savings plans is even more powerful. In an experiment documented by Bridgitte Madrian and Dennis Shea, participation in 401(k) plans grew from 35 to 85 percent for women, 19 to 75 percent for Hispanics, and 13 to 80 percent for low-income workers when the default setting was switched to automatic enrollment (you have to opt out) instead of automatic non-enrollment (you have to opt in). Also, Richard Thaler and Shlomo Benartzi found in their “Save More Tomorrow” experiment (workers can commit to save more of future pay increases), that 78 percent elected to use it while 98 percent remained in the plan two pay raises later. These “behavioral” features are similar to the structural or institutional features mentioned earlier—that saving, like many things, is a psychological or behavioral response to an established structure.

6. Our biggest successes thus far have required no government funds.

This, too, was quite a surprise. We had assumed that we needed billions in new government funds to generate billions in savings and assets for the poor. But no funding was required when Congress clarified various 401(k) laws in the Pension Protection Act of 1996 that enabled more companies to offer opt-out policies, which is estimated to generate $44 billion in new savings every year, including new savings by low-income workers, according to the Brookings Institution. Also, taxpayers can now automatically save their refunds in three accounts(via form 8888 or “split refunds”) and can purchase savings bonds directly on their tax returns—again, at no cost to taxpayers.

Preliminary data and earlier experiments strongly suggest that low-income families are largely, even disproportionately, benefiting from both split refunds and the savings bond options: Over 60 percent of savings bond purchasers had AGIs of $50,000 or less.

The bottom line: Think big and think small. Think big, as in savings accounts at birth for all children. And think about the small, low-cost changes in regulations, tax forms, products,etc. that could result in billions of dollars in new savings, including from the poor.

7. Saving and asset building are still the right ideas—even in this economy, and even as asset values have shrunk.

The goal should not change, but how we achieve it, for whom, and when in the lifecycle should. A good example of asset building done right is Self-Help’s 50,000-plus families in the secondary market program, who repaid their loans and saw their home equity increase over the last several years. Another is the CFED-Urban Institute IDA study, which found that “[L]ow-income homeowners who participated in programs providing extensive financial education and matched savings on their down payments were two to three times less likely to lose their homes in the recent wave of foreclosure than similar families in the same communities.”

No doubt research and the savings experiences of the poor will continue to surprise and inform us. And no doubt these lessons and insights will be helpful as we design policies and programs in this tight fiscal era—a time in which all Americans will be expected to save and invest more to realize their aspirations.

This piece was originally published in the St. Louis Federal Reserve Bank’s Winter 2010-2011 ‘Bridges’ newsletter.

Ray Boshara is a Senior Research Fellow in the Asset Building Program at the New America Foundation, a Washington, D.C.-based think tank. He founded the Asset Building Program at New America, which he directed from 2002-2009.