Volume 3, Issue 2
What to do about Social Security remains one of our country’s most pressing policy questions. As I assume the directorship of the Michigan Retirement Research Center (MRRC), I am confident that we have assembled under its aegis a group of individuals who are preeminently qualified to address this important question and the many related issues. The annual Retirement Research Consortium (RRC) conference, organized this year by the MRRC, and held in Washington, D.C. last May, gave testimony to this assertion. Attendees to the conference were able to witness the MRRC philosophy in action: MRRC wants to help create and strengthen the scientific foundation for policy decisions. In the spring of 2002, the Social Security Administration issued a request to the retirement research centers to place a greater emphasis on research addressing questions of individual retirement accounts for Social Security, and we were able to respond quickly. In addition to several papers that explored different aspects of individual accounts, this year’s conference included a special panel discussion presenting a variety of scholarly views on this timely issue.
On a somewhat more personal note, I want to thank my immediate predecessor, Tom Juster, for his aid and counsel through the transition. MRRC was lucky, indeed, to have his participation, and I look forward to his continued contributions. I also want to acknowledge Lee Lillard’s significant achievements as the first Director of the MRRC. In addition to his impressive academic contributions, Lee set up an organizational structure for MRRC that makes sense and works well. One of my goals is to ensure organizational continuity. A second is to maintain not only the quality of our research projects but also their priority within MRRC. MRRC has weighted research heavily and we will continue to do so -- working as well to solicit research that is timely for current policy initiatives. A third goal is to expand MRRC’s research to include international expertise.
I am glad to be at the helm of this extraordinary organization when reform of social insurance programs is occurring on a global scale. It is an exciting time to be part of the dialogue.
Issue in Brief
Modeling the Macroeconomic Implications of Social Security Reform
By John Laitner
Two models dominate the economics literature on why people save (and hence on why they accumulate wealth). In one, the life–cycle model, people save when they are young in order to have funds to support their retirement. In the other, the dynastic model, households save to build estates to pass to their descendants. From the standpoint of public policy analysis, the two models can have quite different predictions. It is also the case that research shows that either model by itself has shortcomings in describing actual savings behavior. In this Issues in Brief, I summarize how one might combine the two basic models, what advantages the combined framework might have in matching empirical evidence, and what implications for public policy — specifically Social Security reform — the combined model might have.
The life–cycle model is manifestly attractive. It emphasizes a basic pattern of life: a household’s earnings tend to rise during youth and middle age, but they often disappear in old age, with retirement. A household with low earnings might be comfortable retiring on Social Security benefits alone. To maintain reasonably level annual consumption spending, most households, however, must save in youth and middle age to build private wealth, which they can spend in their retirement.
A closer look raises some doubts about the life-cycle model as the sole explanation for savings behavior. Although the model predicts that households will seek to spend their net worth down to zero before death, economists have never been sure empirical evidence supports such complete depletion. Similarly, there has been longstanding suspicion that the life–cycle model cannot explain the vast total amount of wealth that we see in the U.S. economy (for example, in the Federal Reserve Bank’s Flow of Funds data). Furthermore, the distribution of wealth holdings in the U.S. is very unequal. For instance, the Federal Reserve Bank’s Survey of Consumer Finances 1995 shows the wealthiest 1% of U.S. households own 30% or more of U.S. private wealth, and the top 5% own over half. This is much more unequal than the distribution of earnings, where the share of the top 1% is about 11%. Simulations of life–cycle models generally fail to generate nearly as much concentration of wealth as these data indicate.
According to the dynastic model, households may save to build estates to pass to their descendants. In the simplest version of that framework, the so–called representative agent model, all households are alike and all leave bequests. Due to the framework’s analytic tractability, economists employ it widely. Another advantage of it is that it generally has little difficulty explaining the total amount of private net worth that we see. Nevertheless, in terms of matching survey data, the representative agent model is unsatisfactory. For one thing, only about half of U.S. families report that they receive an inheritance at some point in their lives, and most reported inheritances are quite small. The long time span of the Panel Study of Income Dynamics (PSID) data set makes it convenient for studying intergenerational linkages. Yet, recent PSID–based work on the correlation of the consumption of parents and their grown children, on inter vivos transfer behavior, and on inheritances, fails to support the simple dynastic model, or provides at most mixed support.
I propose a model combining life–cycle saving and estate building. Key elements of the combined system are that every household does life–cycle saving; the earnings levels of different households are different; all households care about both their own lifetime well-being and that of their descendants; and, liquidity constraints exclude negative bequests and negative net worth at any age. The model has a simple aggregate production function; labor supply is inelastic; earning ability differences among households are exogenous; and, the economy is closed.
In terms of matching empirical evidence, the hybrid model offers potential advantages over either basic model taken separately. Wealth accumulation will tend to be less equal than in the pure life–cycle model. As in the life cycle model, every household saves during the first part of its life in anticipation of retirement. However, estates are a different matter. Since earning abilities within family lines regress over generations toward the mean, low earners expect that their descendants will have favorable earnings relative to their own. Therefore, they will not attempt to build an estate. High earners face the opposite circumstance: parents with high earnings must expect that their descendants will likely do worse; hence, they have incentives to build estates to share their good luck. In the end, higher earners save for two reasons—to support their retirement, and to leave an estate. Lower earners save for only one reason—to support their retirement. It is also true that the hybrid does not predict that all parents will leave bequests. On the contrary, only parents with high earnings and/or large inheritances will want or be able to do so.
My work calibrates the combined model. I set the distribution of earning abilities from the Survey of Consumer Finances 1995. Several other parameters are quite standard. Two are less so — namely, a parameter determining the degree of household risk aversion, and a parameter determining the importance which parents place on their grown children’s lifetime well–being relative to their own. I jointly calibrate the last two parameters so that the model’s equilibrium has as much private wealth (relative to gross domestic product) as the U.S. economy, and that the model generates a distribution of bequests yielding realistic federal estate tax revenues. In the best calibrations, households are quite tolerant of risk, and parents weight their grown children’s well-being about 80% as high as their own.
Summary of Findings
A simulation based on the best calibration matches several aspects of the U.S. economy surprisingly well. In the simulation, about half of all households in each birth cohort choose to leave bequests. And the simulated distribution of wealth is very unequal. For example, the richest 1% of households in every cross section own about 25% of total private net worth.
Turning to analysis of public policy, the pure life–cycle model suggests that reform instituting funded private accounts for Social Security in place of part of the current unfunded Social Security system could substantially increase the economy’s wealth accumulation in the long run. The idea is that taxes reduce household resources for saving, and benefits reduce households’ need to save for their retirement. The dynastic model has the opposite implication: since the present value of any system’s taxes and benefits exactly balance for each endless dynasty, neither a funded nor an unfunded system affects the representative agent’s dynastic well-being or consumption. Private intergenerational transfers counterbalance public programs, and the economy’s privately supported capital stock remains always the same.
My work shows the hybrid model can exhibit either reaction to policy, or indeed any reaction between the two extremes. The wide range of potential outcomes makes precise calibration of the model’s parameters very important. The best calibrations at this point yield a combined model with policy implications resembling the representative agent dynastic formulation rather than the pure life–cycle case. This is true despite the fact that in the simulation life–cycle saving alone finances about 70% of total private net worth.
Further research is needed. For example, future research will incorporate lifetime earnings uncertainty for households. Perhaps lifetime precautionary saving will further reduce the importance of estate building. Even if present results hold, other analysis suggests that reform might play a role in controlling wealth inequality. Reductions in inequality might be a desirable—though otherwise unexpected—concomitant of funding Social Security or reducing the national debt.
In conclusion, a relatively straightforward combination of existing economic models seems to offer the potential of explaining a number of empirical features of the U.S. economy that have heretofore been somewhat opaque. Precise calibrations of the parameters of the combined model are important since the potential range of policy implications is broad. At this point, analysis suggests that outcomes from Social Security reform such as greater equality, flexibility, and sustainability may be at least as important as potential long–run increases in national saving.
John Laitner is the Director of the Michigan Retirement Research Center and a Professor of Economics at the University of Michigan; Ann Arbor, Michigan 48106 (email@example.com).
back to the top.
For Your Information
How Much Do You Need to Retire?
Figuring out what you need to save for retirement is easier than ever. The American Savings Education Council (ASEC) offers savings tips and planning advice to help citizens save and prepare for a secure retirement. ASEC has developed what they call the Ballpark Estimate, which is an easy-to-use, one-page worksheet that helps you quickly identify approximately how much you need to save to fund a comfortable retirement. The Ballpark Estimate takes complicated issues like projected Social Security benefits and earnings assumptions on savings, and turns them into language and mathematics that are easy to understand. This worksheet is available for interactive use on-line or in printable copy at http://www.asec.org/ballpark/
The Social Security Administration (SSA) will be teaming up with ASEC and with State Farm Insurance Companies in the spring of 2003 to launch national campaign called “Save for Your Future” to increase the public’s awareness about the importance of saving. As noted by JoAnne Barnhardt, Commissioner of the SSA "Social Security is the foundation of a secure retirement…and is meant to be only part of a three-legged financial stool, along with pensions and personal savings. Unfortunately, only half of today's retirees have a private pension. And too few Americans save as much as they should."1
Other SSA initiatives that have been described in this column to educate workers about the importance of financial planning include SSA’s annual mailing of a Social Security Statement to all workers age 25 and older as well as the online Benefits Planner (www.ssa.gov/planners/calculators.htm) where people can calculate their estimated Social Security retirement benefit.
1 Excerpted from SSA News Release available at http://www.ssa.gov
back to the top.
John Laitner Named Director of MRRC
John Laitner received his Ph.D. from Harvard University. He is a Senior Research Scientist at the Institute for Social Research and a Professor of Economics at the University of Michigan. His research falls primarily in the area of economic theory, in particular, factors influencing long-run growth and the distribution of wealth. Among his recent publications are the following: "Random Earnings Differences, Lifetime Liquidity Constraints, and Altruistic Intergenerational Transfers," Journal of Economic Theory, December 1992; "Quantitative Evaluations of Efficient Tax Policies," Oxford Economic Papers, July 1995; "New Evidence on Altruism: A Study of TIAA--CREF Retirees" (with T. Juster), American Economic Review, September 1996; "Intergenerational and Interhousehold Economic Links," chapter 5 of Handbook of Population and Family Economics, North--Holland, 1997; "Earnings Within Education Groups and Overall Productivity Growth," Journal of Political Economy, August, 2000; "Bequest Motives: A Comparison of Sweden and the United States" (with H. Ohlsson), Journal of Public Economics, January, 2001; and "Social Security Reform and National Wealth," Scandinavian Journal of Economics, 2000. John has been associated with the MRRC since the second year and has worked with Tom Juster in the past six months to arrange the annual RRC conference and the MRRC researcher workshop.
back to the top.
Symposium on Risk Transfers and Retirement Income Security Held in April
MRRC researchers Olivia Mitchell and Kent Smetters recently co-organized a symposium on "Risk Transfers and Retirement Income Security.” The conference, which was supported in part by the Social Security Administration through the MRRC, was held at the Wharton School in Philadelphia, PA, on April 22-23, 2002 and explored the types of risk that employees and retirees bear under retirement plans, and how these risks might be better protected against--areas where pension plans are sometimes critiqued.
Participants addressed issues such as whether innovative financial responses can be designed to meet and better manage these risks. Also, given recent stock market volatility, pension experts have heard calls for guaranteed products to supplement investment options in the defined contribution environment. Participants drew lessons from United States private and public pension plans, as well as retirement systems in other countries that have already implemented a wide range of protective measures. The goal was to find better ways to manage risk associated with retirement income products. In addition, policymakers learned of new research in the area of risk protection mechanisms that might be provided to retirement account savers. The symposium was co-sponsored by the Wharton School’s Pension Research Council and Financial Institutions Center.
back to the top.
4th Annual Retirement Research Consortium Conference held in Washington, DC
The fourth annual conference of the Social Security Retirement Research Consortium was held at the National Press Club in Washington, D.C. on May 30 and 31, 2002. The two-day conference, entitled “Directions for Social Security Reform,” was planned and arranged this year by the Michigan Retirement Research Center (MRRC) in conjunction with the Center for Retirement Research at Boston College (CRR) and the Social Security Administration (SSA). To open the conference, greetings were offered by John Laitner, Director of the MRRC and Alicia Munnell, Director of the CRR. Introductory remarks were made James B. Lockhart III, Deputy Commissioner of the Social Security Administration. The keynote address was delivered by Olivia S. Mitchell who is a member of the MRRC Executive Committee and on faculty at the Wharton School and NBER. Mitchell served as a member of the President’s Commission to Strengthen Social Security. Her address provided a summary of the Commission’s findings with an opportunity for conference attendees to pose questions to her.
A special feature of this year’s conference was a panel discussion on Social Security personal retirement accounts lead by MRRC researcher and Executive Committee member Alan Gustman. The panel represented a range of views on the topic at hand and included Kent Smetters of the University of Pennsylvania, Gene Steuerle of the Urban Institute, Sylvester Schieber of Watson-Wyatt Worldwide, and John Shoven of Stanford University. Over the two days, researchers from each of the two consortium centers presented papers on current projects, all of which are funded by SSA. The topics covered in the six paper sessions included economic issues in Social Security reform, distributional effects of Social Security and perspectives on reforming Social Security, interactions among social insurance programs, health issues in retirement, and experiences from the private sector. The final session of the second day featured nine Sandell Awardees who presented their research findings.
back to the top.
MRRC booth at the American Statistical Association Meetings in New York City
The MRRC conference booth will appear at the ASA Meetings in New York City, August 11-15, 2002. MRRC staff will be available to respond to inquiries from conference attendees. In addition, the booth will have materials including MRRC working papers, Issues in Brief, a list of research topics, current and previous Center newsletters, and MRRC brochures. The booth will appear next at the Gerontological Society of America 2002 Annual Meeting November 22-26, 2002 in Boston, MA.
back to the top.
University of Michigan
Institute for Social Research
426 Thompson Street
Ann Arbor, MI 48106-1248