Q&A with Michael Sherraden

Michael Sherraden, founding director of the Center for Social Development (CSD) at the Washington University in St. Louis, shares his thoughts about the benefits of asset-building policies and programs, such as universal savings accounts for social protection and development.

• What is asset-based policy?
  
• What are the benefits of asset-based programs?
  
• How have asset-based programs been implemented in industrialized countries?
  
• Could these savings schemes be replicated in developing countries? If so, what challenges may these programs encounter elsewhere?
  
• What key design principles should policymakers around the world take into consideration when developing asset-building programs?

1- What is asset-based policy?

Today there is increasing questioning of income as the sole definition of poverty and well-being. The phrase “asset-based policy” was introduced in response to predominantly income-and-consumption policies of industrialized countries. Asset-based policy is focused on accumulation and investment for long-term development. It is a complement to—not a replacement of—income-based policy.

Of course, the term “assets” has many potential meanings. These include financial wealth, tangible property, human capital, social capital, political participation and influence, cultural capital, and natural resources. While all of these meanings have value, I focus on meanings of assets that have the most direct relevance for social policy: financial and tangible wealth for purposes of household social and economic development.

Return to Top^

2- What are the benefits of asset-based programs?

Assets may have multiple positive effects. In addition to deferred consumption, some possible effects of asset holding are to: (1) improve household stability, (2) create orientation toward the future, (3) stimulate enhancement of assets, (4) enable focus and specialization, (5) provide a foundation for risk taking, (6) increase personal efficacy, (7) increase social connectedness and influence, (8) increase political participation, and (9) enhance the well-being of offspring.

For example, a study of assets, expectations, and educational performance in the U.S. found that low-income, single mothers’ assets are positively associated with children’s educational attainment. These results occur in part through expectations of the mother: Assets are associated with higher educational expectations, which are in turn associated with higher educational attainment. This implies that assets may change thinking, which in turn may change behavioral outcomes.

Another study of the “asset effect” in the U.K. found that holding assets at age 23 is associated with later positive outcomes such as better labor market experience, marriages, health, health behaviors, and political interest. These researchers also find that the presence of an asset appears to matter more than the monetary value of the asset. One implication of this may be that, if the presence of a housing asset is what matters, then policy should perhaps encourage home ownership as early in adult life as possible.

Turning to examples from applied research on IDAs, we turn to some of the results of the American Dream Demonstration (ADD). ADD was the first major demonstration of IDAs. It took place at 14 IDA programs around the United States from 1997 through 2001, with research continuing through 2005. One of the findings in ADD is that income was only weakly associated with saving outcomes, i.e., the poorest participants saved about as much as those who were not as poor, and saved a higher proportion of their income (Schreiner et al., 2002). This finding suggests that saving by the very poor should not be dismissed in public policy.

Return to Top^

3- How have asset-based programs been implemented in industrialized countries?

In the United States and many other countries, asset-based policies operate mostly through the tax system, i.e., the public transfers occur via tax benefits (either tax deferments or tax exemptions). The poor, who have little or no tax liability, receive few or no benefits. For example, of the $300 billion annually in US tax expenditures for assets (homes, investments, retirement accounts), over 90 percent goes to households with incomes over $50,000 per year. At the same time, asset limits in means-tested transfer policies discourage saving by the poor. In effect, the United States and many other countries have a dual policy, consisting of asset building subsidies for the non-poor, and asset building disincentives for the poor. This dual policy is unfair and counterproductive.

These considerations led to a proposal for Individual Development Accounts (IDAs), which were proposed as (1) universal savings accounts, (2) started as early as birth, (3) with savings matched for the poor, up to a cap, and (4) multiple sources of matching deposits, (5) accompanied by financial education, and (6) the savings are to be used for investments in homes, education, business capitalization, or other development purposes. Over 40 US states have now adopted some type of IDA policy. However, most IDA programs in the United States are very small, still in a “demonstration” phase. Perhaps more importantly, saving and asset accumulation by the poor, which was seldom discussed 15 years ago, is today a mainstream idea in the United States.

In the U.K., a serious discussion of asset-based policy began in 2000. In April 2001, Prime Minister Tony Blair proposed a Child Trust Fund for all children in the United Kingdom. Beginning in April 2005, each newborn child is given an account with an initial deposit of at least 250 pounds, and children in the bottom third of family income will receive 500 pounds. This is the first universal and progressive asset-based policy on the planet.

Of course, a CSA is not ultimately about children. After several generations of children born with a CSA, everyone could have an account throughout life. Thinking about this in terms of institutions and behavioral economics, there is a great deal of current research and discussion about “defaults”, that is, putting people in a saving plan (or some aspect of it) unless they make a choice not to participate (sometimes called “opt out” or “automatic”). We might think of a universal CSA as the ultimate “default”—every child would automatically be born with a birth certificate and an asset-building account.

Return to Top^

4- Could these savings schemes be replicated in developing countries? If so, what challenges may these programs encounter elsewhere?

Of course, focusing on productive assets is not at all a new idea in economic development. Much of “development” work is about building tangible assets such as a small businesses, land (including clarification of land title), homes, equipment, or livestock.

Also, savings and development is a very well-established field. Perhaps the question above would be better posed to ask whether savings schemes in developing nations can be replicated in developed ones.

During the twentieth century, there was a nearly complete separation of discussions of “social welfare policies” in developed countries from “development policies” in developing countries. Today, this separation appears to be breaking down. The capabilities discussion introduced by Sen is providing a framework, albeit an abstract one, for linking “social welfare” and “development”. Asset building is part of this linkage, and offers concrete referents. For example, CSAs may have appeal in developing countries, and for international aid.

One of my former graduate students, Fred Ssewamala, now on the faculty at Columbia University, is testing CSAs with HIV/AIDS orphans in Uganda. His strategy is to enable the children to save enough, combined with matching funds, to pay for secondary school (four years of secondary school costs about US$600). These young people will be vastly better off economically and socially if they complete secondary school. If they do not, they are at high risk of becoming HIV/AIDS victims themselves (Ssewamala & Curley, 2005). Ssewamala now has a grant from the National Institutes of Health (NIH) to extend this research. This is another step toward Ssewamala’s goal of providing an account to all orphan children, and eventually to all children, in Uganda.

“Development” strategies to date have not focused on universal individual accounts. However, with the spread of information technology, there will be opportunities to do so. To extend the example above, if every child has an account, these could be targets for international aid that goes directly to children—avoiding administrative costs, mismanagement, and corruption. This eventually could link policies across developing and developed countries.

Challenges are many. These include access to a financial services infrastructure, financial markets that can hold and protect capital, governments that are capable and honest, and macroeconomic policies that provide stability—especially price stability to protect against asset depletion due to inflation. Many countries to do offer these conditions. In these cases, it may be preferable to concentrate on shorter-term strategies (e.g., savings for secondary education described above), and/or assets other than financial assets.

In the larger picture, information technology will some day make access to financial services, sound financial markets, and stable currencies available to everyone. This will be a “public good” of huge importance for development worldwide.

Return to Top^

5- What key design principles should policymakers around the world take into consideration when developing asset-building programs?

In CSD’s research on IDAs, we have identified the following institutional factors that may affect saving and asset accumulation: (1) access, (2) expectations, (3) information, (4) incentives, (5) facilitation, (6) restrictions, and (7) security. These constructs appear to be useful in explaining saving outcomes, and they have direct relevance for policy.

For example, we find in research on IDAs that, controlling for many other factors, each dollar increase in the monthly saving target (expectation) is associated with a 40-to-50-cent increase in average saving—a huge effect. We find that financial education (information) up to about 10 hours is associated with increased saving performance, but after 10 hours there appears to be no effect. Because financial education is expensive, this is important to know. We find that increasing the saving match (incentive) keeps people saving in the IDA program, but among the “savers” (i.e., those who saved non-trivial amounts in IDAs), increased match does not increase amounts saved. We also find that direct deposit (facilitation) keeps people saving but does not increase amounts saved by “savers”.

The goal of asset-based policy should be inclusion. By inclusion, I mean that policy should: (1) bring everyone into asset-based policy, (2) make asset-based policy life-long and flexible, (3) provide at least equal public subsidies for the poor in dollar terms; and (4) achieve adequate levels of asset accumulation, given the purposes of the policy.

If saving and asset building are to be inclusive, the policy must be in the form of a savings plan. Examples in the US include 401(k) or 403(b) plans, the Federal Thrift Savings Plan, and College Savings (529) plans. Such plans are in fact how most Americans are able to save.

Savings plans (contractual savings) have important features that lend themselves to inclusion. These features are: centralized and efficient accounting, outreach and education, a limited number of low-cost investment options, low initial and on-going deposit requirements, automatic deposits, and opportunities to establish other practices and “default settings” that increase saving performance. These include automatic enrollment, savings match, match cap (amount of savings that can be matched), a default low-cost fund, automatic increases in savings deposits with pay raises. During the payout period, it may be desirable for a required minimum annuitization for income protection.

As a closing thought, the advantages of asset building may be that: (1) it is simple and clear, (2) it is easy to communicate, (3) it has widespread appeal and acceptance, (4) it is flexible and adaptable, (5) it can be both a large policy and a local strategy, and (6) main outcomes are relatively easy to measure.

Return to Top^

Note: These remarks borrow in part from Sherraden’s testimony before the US Senate, Committee on Finance, Subcommittee on Social Security, in April 2005.