How prepared are Americans for retirement: Written testimony to the Special Committee on Aging

Chairman Collins, Ranking Member McCaskill, and members of the Special Committee, thank you very much for the invitation to appear before you today. I am delighted to be here. My name is Michal Grinstein-Weiss. I am an associate professor of social work at Washington University in St. Louis and the associate director of the Center for Social Development. Let me state that the views expressed here today are my own and not necessarily the views of Washington University in St. Louis or the Center for Social Development.

My work focuses on developing knowledge to inform public policy on innovative strategies that promote social and financial opportunities for all, especially for low-income households. Specifically, I am engaged in creating and evaluating large-scale experiments on methods to help low- and moderate-income families mitigate financial hardship, create emergency savings, enter the financial mainstream, and accumulate long- term assets. My work in general and my testimony today focus mostly on low- and moderate-income households, as they typically struggle to accumulate assets face many barriers to saving.

My testimony is in three parts. I first discuss the problem of low retirement savings and, secondly, I summarize relevant theories of saving that help to illustrate areas for improvement. In the final portion, I offer some specific policy recommendations based on my own work and that of my colleagues. In offering these recommendations, I seek to enhance the financial security of Americans in retirement.


Nearly half (45%) of Americans have no retirement savings, and those least likely to own a retirement account are lower-income households. Although nearly 90% of households in the highest income quartile own retirement accounts, only about a quarter of households in the lowest income quartile have such accounts. Overall, the average working household has little to nothing saved for retirement. The median retirement-account balance is only $3,000 for working-age households and only $12,000 for households approaching retirement. In two-thirds of working households with earners between ages 55 and 64 years, at least one earner has saved less than one year’s income for retirement. Such savings fall far below what they will need to maintain their present standard of living.

As you know, many low- and moderate-income households rely entirely on Social Security for postretirement income. Currently, Social Security and Supplemental Security Income account for more than 90% of income for retirees in the bottom 25% of the income distribution. Social Security payments account for  approximately 70% of the annual income of retirees in the middle 50%. This dependence on Social Security is alarming and problematic, as the Social Security program was never meant to serve as the main source of retirement income and the program is not sustainable under these conditions. Experts are therefore correct in identifying retirement savings as a central challenge for generations currently in the labor force.

Relevant Theories on Saving

The original life-cycle theory of saving holds that both household income and savings increase as members of the household age. However, some evidence suggests that household spending and income grow together, so the household saving rate remains fairly constant. Particularly during periods of stagnant wages and high unemployment, households may lack the resources needed to accrue adequate emergency savings. In my research, I draw from newer viewpoints including the institutional model for saving behavior, and insights from behavioral economists to understand how to increase savings behavior among low- and moderate- income people.

Departing from the life-cycle theory of saving, my colleagues Michael Sherraden and Sondra Beverly pioneered the institutional theory of saving, which suggests that institutional factors greatly affect an individual’s ability to save. These factors include access, information, incentives, facilitation, expectations, restrictions and security.9 The institutional model also suggests that all people can save if given access to the right systems—the correct plumbing. Access to and knowledge about savings accounts, employer-based retirement plans, and 529 accounts are several examples of these wealth building products that are less likely to be accessed by low-income Americans.

Behavioral economists have further expanded our understanding of saving behavior by showing how people are predictably irrational in their saving practices and habits. People tend to continue whatever habits, accounts, and transactions they have in place, even if those are not optimal. People also tend to adopt default choices if offered several savings options, to flee from decisions that appear difficult or complex, and to procrastinate instead of taking action. These behavioral and cognitive biases contribute to the status quo of low saving rates. However, I and other researchers have begun to translate these insights into practical strategies and program designs that can effectively improve saving behaviors. For example, our research in  the Refund to Savings Initiative has shown that it is possible to increase the number of filers who save part of their tax refund and the amount set aside: by suggesting that filers should save a specific percentage of the refund—for example, 75%—we increased the number of savers and the amount saved.

Barriers to Saving for Low- and Moderate-Income Americans

As I have indicated, American households face both individual and institutional barriers that deter saving. These barriers are often multiplied for lower-income households. However, demographic characteristics, perceptions, attitudes toward saving, deficits in financial education, and other individual features are difficult targets for policy intervention. Institutional factors offer more promising opportunities for reform, and institutional changes can have far-reaching impact. I wish to discuss several institutional factors and their effects on household saving behavior.

Indeed, research has repeatedly shown that households at all income levels not only want to save but can and do accrue savings when provided with the right institutional support. Unfortunately, many existing institutional barriers preclude or diminish the chances of low- and moderate-income households in building savings.

Most programs intended to help Americans build savings are targeted for, and are more likely to benefit, upper-income households. Asset-building programs that operate through the tax code—such as 401(k) retirement plans, 529 college savings accounts, and the home mortgage interest deduction—primarily benefit households in the upper half of the income distribution. In fact, at least 90% of the asset-building tax benefits go to the top half of the income distribution, with 30% of those benefits going to the top 10%. In 2013, the lowest income quartile of Americans received only 0.7% of the tax subsidies for employer based retirement savings plans, whereas the top income quintile received 68.4% of those subsidies. The benefits of homeownership are similarly disproportionate: owning a home offers fewer wealth gains to low- and moderate-income households than to higher-income ones; moreover, the risk of losing assets invested in a home is greater for low- and moderate-income households because a greater share of their wealth is tied up in their homes. Without institutional supports, such as workplace-based programs, many households succumb to “saving procrastination,” leaving roughly 37% of all U.S. workers with less than $1,000 in retirement savings. Saving through workplace retirement plans and secured savings like homeownership also offer higher income households the benefit of streamlined or automated savings methods. Households without access to these systems must navigate an increasingly complex financial world to choose among many financial products.

Another problem is that small-dollar, short-term savings accounts—the sort that households use for emergency savings—are not a profit center for banks. To make these accounts financially viable, many banks require minimum balances and impose fees, but those features make the accounts unattractive to low- and middle-income households. Meager savings can be eaten away by monthly account maintenance fees. At present, transaction accounts (e.g., checking and regular savings accounts) carry interest rates that are too low to act as effective incentives for potential savers. In some cases, bank fees and penalties are high enough that being unbanked may be economically preferable to those with low balances.

Many low- and moderate-income households intend to save but experience high levels of financial shocks  that keep them from saving. Low levels of liquid savings and lack of access to affordable credit render lower- income households especially vulnerable. However, a lack of emergency savings is a not a problem restricted to low-income households. Americans across the income distribution are often unprepared to deal with negative financial shocks. When those without general savings face sudden need, they often raid dedicated forms of savings such as retirement accounts (incurring tax and interest penalties when they can least afford them) and thereby further diminish their ability to prepare for and respond to future events. Household attitudes toward liquidity and risk may also limit the demand for emergency savings. Studies indicate that households feel pressure to keep a big share of their assets in highly liquid savings products. On its face, this perceived pressure would seem to encourage high levels of emergency savings, which are usually held in liquid products, and low levels of illiquid long-term savings. However, holding savings in accounts that are highly liquid may make them too easy to spend and may lead to depletion of funds needed for future emergencies.

Unless they have access to products that encourage saving but also allow them to liquidate savings for unforeseen financial shocks, low- and middle-income households are less likely to save.


Each of these barriers contributed to the declining saving rates of American households. Several insights from researchers and practitioners offer areas for reform and improvement of our current approaches to saving. Today, I offer four recommendations that can enable more Americans to save and to prepare better for retirement: 1. Automate good financial choices  2. Maximize “golden moments” to create more saving opportunities  3. Support emergency saving  4. Facilitate saving opportunities early in life.

Automate good financial choices

In an increasingly complex financial landscape, Americans face a multitude of difficult financial decisions, including how and how much to save for retirement. I recommend that policymakers make it easier for Americans to overcome this challenge by designing automated savings opportunities into policies and programs. One way to do this is to make enrollment into savings accounts the default choice so that those who do not wish to enroll must act to opt out. Behavioral economics research and several large-scale programs demonstrate that this change to automatic enrollment can be a powerful tool for increasing the number of Americans saving for retirement and other purposes. For example, in the seminal work on defaults and organ donation, researchers found an average participation rate of 99% in countries with an opt-out organ-donation policy but that participation rates were below 28% in countries with an opt-in policy.
This same strategy—making no-action-required enrollment the default choice—has also been applied successfully to saving opportunities. For example, the SEED for Oklahoma Kids (SEED OK) experiment found that it was possible to achieve nearly universal program enrollment if a 529 college savings account were automatically opened for the children of treatment participants and $1,000 deposited into the account; in contrast, control participants had to open their own account for their child, and only 2% did so. The Alfond College Challenge has similar features. Influenced by SEED OK, the Challenge has provided Maine newborns with 529 accounts and an accompanying $500 deposit since 2008, but program take-up increased from 40% to 100% when the program switched from an opt-in model to an opt-out model in 2014. Automatic enrollment significantly increases program participation and thus opportunity to benefit from the program.

Retirement savings programs can also benefit from automatic features, which can be applied in enrollment processes, savings contribution percentages, and employer matches. Several new programs and legislation provide promising examples of how automation and defaults can be used to increase take-up rates and facilitate saving. The Illinois Secure Choice Savings Program, signed into law in January, creates automatic enrollment into a Roth IRA with a 3% default contribution rate for private-sector workers in small businesses with at least 25 employees, thereby providing a retirement plan for workers not already covered by their employers. At the federal level, Sen. Collins, Sen. Nelson, and Sen. McCaskill introduced the Retirement Security Act of 2015, which raises the cap on automatic employer-based retirement plans and has the  potential to make retirement savings plans accessible to more people. A similar version of this type of program- an automatic enrollment IRA funded through automatic direct deposit from paychecks- has been proposed for the last decade by researchers and policymakers across the nation. President Obama called for the creation of such accounts in his 2015 State of the Union address, and policymakers should pass legislation that offers every American access to an automatic IRA. Defaulting to automation will ensure that Americans, particularly low-income Americans, have a simple pathway to boost retirement savings.

Maximize “golden moments” to create more saving opportunities

Policymakers should focus on developing saving interventions that target what I like to call “golden moments.” These are moments (1) when existing touch points between government and citizens can be leveraged to reach a large number of people at little cost, and (2) when it is easy to overcome natural biases and barriers that inhibit saving. In the context of saving, these golden moments happen when people are thinking about their finances and are likely to make prudent decisions about their future: during tax time, while signing up their children for public school, or when people are enrolling in a new job.

Tax time is one of the best examples of a golden moment. It is a nearly universal touch point between the government and U.S. households. With 75% of tax filers receiving a tax refund, tax time is an attractive opportunity for saving interventions. For many low- and moderate-income households, the federal income tax refund is the largest lump sum payment received during the year. After a household receives its refund, it is possibly in its best balance sheet position for the entire year and perhaps more open to saving than at any other point. Tax time is also a prime opportunity to interact with households around a saving message, because it is one time of year when most households are thinking about the finances and their current financial situation. The introduction of the 8888 form enables tax filers to split their refund into both a savings and a checking account.

When people consider how they will act in the future, they want to act proactively and rationally—to sacrifice and to think long-term. But when the moment comes to act, people often fail to follow through with the original intention. This may be especially true for retirement saving because it is such a long-term goal. The tax filing moment may be an ideal time to introduce the concept of precommitment to those due a refund. The taxpayer knows what the refund will be but lacks access to that money and may be persuaded to commit some of it to savings.

I launched the Refund to Savings Initiative, a collaboration of the Center for Social Development, Duke University, and Intuit, with the goal of testing large-scale tax-time saving concepts with low- and moderate- income households. Working with the makers of the TurboTax Free File Edition to reach more than a million tax filers over the past two tax seasons, Refund to Savings has found that simple interventions delivered at tax time can influence the saving-related decisions and actions of taxpayers. In the 2013 tax season, 11% of our participants said that they would be interested in opening a new retirement account at tax time. In our current experiment, we are working with the Treasury Department to test ways to implement the myRA at the tax-time golden moment and to increase savings for our low- and moderate-income participants.

Another golden moment for saving interventions revolves around leveraging touch points in the public education system. Setting good saving habits at this point can produce effects that last a lifetime. Child Development Accounts (CDAs), which are savings and investment accounts for long-term developmental goals, have been implemented in several contexts. Many CDA programs are built onto existing 529 college saving plans. For example, the Nevada College Kick Start Program automatically deposits $50 into a 529 account for every public school kindergartner in the state and matches further contributions into the accounts by parents.

A third golden moment for saving is the commonly shared experience of starting a new job. When starting a new position, employees are often asked to make retirement savings decisions that have a major impact on their retirement preparedness later in life. Because people have a bias toward the status quo, those early decisions about retirement contributions are critically important. Much research has focused on increasing contributions to employer-sponsored 401(k) plans;however, several policy proposals have sought to take advantage of the employment channel even in the absence of a retirement plan. That is, the hiring process can be a golden moment to encourage new employees lacking bank accounts to open a new bank account and to set up direct deposit.

Given all of these examples, I recommend that policymakers focus attention on developing saving interventions that capitalize on existing golden moments. Policymakers should promote efforts to take advantage of the tax time moment by opening and/or depositing to a savings account. They should use the Free File Alliance to achieve a large-scale impact.

Support emergency savings

Unexpected financial shocks can affect a household’s financial stability as well as its ability to save for long- term needs like retirement. Policymakers should therefore concentrate on developing policies that encourage households to build emergency savings. Recent research, including some of my own, has revealed that many Americans are not financially prepared to handle unexpected major expenses or short-term income volatility. Without liquid assets to cover these financial shocks, households may rely on expensive and potentially harmful strategies, including skipping payment on bills, taking out payday loans, using other alternative financial services, and liquidating retirement savings. These practices can lead to situations of escalating debt and cause households to develop an array of other financial problems that may keep them from saving. In particular, households may deal with financial emergencies by delaying needed medical care or forgoing necessary nutrition. Therefore, a lack of emergency funds can hurt U.S. households in both the short- and long-term. Developing policies and tools to accumulate emergency savings will help families gather the resources to weather financial shocks. These tools will enable financial stability, so that saving for the future is possible.

Financial emergencies and income fluctuations are often unexpected, but they are a consistent fact of life for most people. This may be especially true for low- and moderate income households. To illustrate this, my research asked a sample of roughly 8,000 low- and moderate-income individuals whether they recently experienced a major financial emergency, which in this case was defined as an unexpected medical emergency, major auto repair, period of unemployment, or legal expenses. We found that exactly two thirds of surveyed households experienced at least one of those events in the past six months, and over a quarter experienced several shocks. Households that experienced emergencies were more likely to report material hardship, such  as skipping medical care, and to have used alternative financial services.

To help households deal with emergencies, we need to develop flexible saving products that give easy for liquidity when needed. Ensuring that households have access to traditional savings accounts is clearly an important step in this direction. Although marketed as a retirement product, Treasury’s new myRA program  is an example of a flexible savings product that may provide historically underserved groups with a vehicle for dealing with unexpected expenses. In additional to other attractive program features, such as direct deposit, no fees, and tax advantages, myRA allows for unrestricted withdrawals of contributions and, thus, may serve as a saving vehicle for financial emergencies. In the short term, the accounts are likely to earn higher interest rates than those offered with traditional savings accounts. Efforts to market this type of product to low- and moderate-income households should emphasize its potential to serve as a repository of emergency savings and retirement savings.

Facilitate saving opportunities early in life

Both aging and saving for old age are lifelong processes. Thus, saving should start as early as possible. My colleagues at Center for Social Development have spent the last quarter of a century building the evidence for why we should start saving early and identifying the best parameters to encourage lifelong saving.

Overall, there is reason to believe that children who grow up in families with assets are better off than those who grow up in otherwise similar families without them. We have learned from the SEED for Oklahoma Kids experiment that the CDA has positive effects on mothers’ attitudes about the future, mothers’ mental health, and children’s early development. The effect is similar in size to at least one estimate of the effect of the Head Start Program on early social-emotional development. Having a CDA and assets for college may also matter more than the saving behavior of parents. Some evidence suggests that early positive effects such as these may improve longer-term outcomes, perhaps especially for disadvantaged kids.

Studies also show that having an account in one’s name as a child is positively associated with financial outcomes later in life. It may be because they have already entered the financial mainstream and are able to continue engaging with financial products and institutions as they grow older. Young adults who grew up with a savings account in their name are two times more likely to own checking accounts, two times more likely to own credit cards, three times more likely to own certificates of deposit, and four times more likely to own stocks. This pattern holds true across high- and low-income families. By starting early, children develop good financial habits and a mind set for saving. Good financial habits are crucial for healthy financial lives.

My colleagues at the Center for Social Development lastly recommend that policy should focus on creating long-term asset-building programs. These programs should start early—ideally at birth—enroll participants automatically, build on existing infrastructure, be universal (covering every child), and be progressive (with higher subsidies for low income families). They should restrict the use of accumulated savings for such purposes such as education, retirement, or other asset-building mechanisms. Countries around the world have invested in their children through Child Development Account programs. Singapore, Canada, the  United Kingdom, and Korea serve as examples. The United States has also taken steps to adopt child accounts by building on the 529 savings platform. The state of Maine now automatically enrolls every newborn in the state’s 529 college savings plan and seeds the child’s savings with a $500 grant through its Harold Alfond College Challenge. In Nevada, the state treasurer’s office opens and seeds a 529 college savings account with $50 for every public school kindergartner in the state. Similarly, Congress introduced the ASPIRE (America Saving for Personal Investment, Retirement, and Education) Act in 2010 and 2013 but did not pass the legislation, which would have established a universal system of Lifetime Savings Accounts  for every newborn child.


Whether it is for retirement, emergency expenses, postsecondary education, or that an asset purchase, nearly half of American households are not saving adequately, and the saving rates are significantly worse for low- and moderate-income households than for the full population. Further, Social Security cannot serve as the primary source of retirement income, as it would be insufficient to meet household expenses. Despite the current low savings rates in the United States, evidence from various studies shows that people of all income levels can and will save if institutional supports are available and accessible. Therefore, it is imperative that policymakers strive to increase the savings rate of Americans by enacting legislation that aims to (1) automate good financial choices, (2) maximize “golden moments” to create more saving opportunities, (3) foster preparation for financial shocks, and (4) facilitate saving opportunities early in life. In laying out these four policy recommendations, I have cited examples that can build upon existing infrastructure while promoting saving and streamlining the savings process at minimal cost to the federal government.

Download full testimony with references.