April 2002
Volume 3, Issue 1

Director’s Corner

In December, 2001, the President’s Commission to Strengthen Social Security (CSSS) released its final report presenting three models for modifying the present Social Security program. A revised version was released in March of this year that includes the Social Security Administration’s estimates of the financial effects of the three proposed models. In the report, the Commission recommends that there be a period of dialogue lasting at least one year before lawmakers take action. In keeping with our mandate to serve the public interest by providing a scientific foundation for policy, the Michigan Retirement Research Center (MRRC) is soliciting and selecting research proposals that address aspects of the models and alternatives. Indeed, much of our past and current work is very relevant to reform debate. We are continuing our efforts to provide quick turn around and to disseminate research findings by MRRC researchers expediently through our website, newsletter, working papers, and briefs. Another important route for dissemination this year will be the annual Retirement Research Consortium (RRC) conference to be held this May in Washington, D.C. The conference will provide a public forum for the presentation of research. The theme of the conference is Directions for Social Security Reform. Thus, we hope to be a significant part of the dialogue proposed by the President’s Commission.

Regular Features
Issue in Brief


Inside This Issue
Sandell Grants Awarded

Researcher Workshop

RAND Data Available

Issue in Brief
Imperfect Knowledge, Retirement and Saving
By Alan L. Gustman & Thomas Steinmeier

Executive Summary
The standard economic model of retirement and saving suggests that people are well-informed and forward-looking in deciding on their labor market and saving behavior. A key consideration is the standard of living they will enjoy once they have retired, and how it relates to their standard of living when at work. This Issue in Brief describes an exploratory investigation that suggests people are not very well informed about their pensions and social security, but that their retirement and saving behavior is not very much affected by whether they have been overly optimistic or pessimistic. Even as people approach retirement age, they do not correct their saving or revise their retirement to better attain their desired consumption in retirement. Among those who have overestimated their benefits, rather than save more or retire later, they are content to accept whatever consumption their saving will permit in retirement, leaving their behavior as planned using misinformation.
We use data from the University of Michigan Health and Retirement Study (HRS) linked with Social Security records and employer-provided information about pensions to develop direct measures of the level of knowledge and misinformation about Social Security and pensions. We begin by examining the general level of knowledge about retirement benefits and see whether this is different for various demographic groups. We relate these measures of knowledge to activities undertaken to plan for retirement. Next, we examine the effects of knowledge and retirement planning activities on retirement outcomes, including on retirement plans, retirement behavior, revisions in retirement as one approaches retirement age, and the estimated relation between Social Security and pension incentives and retirement. Lastly, we conduct preliminary analyses to see what relationship knowledge has to wealth accumulated for retirement.

The Data
The Health and Retirement Study (HRS) is a longitudinal, nationally representative study of older Americans. This NIA supported survey began in 1992 with an initial cohort of 12,652 individuals from 7,702 households, with at least one household member born from 1931 to 1941. Social security earnings histories were linked for 9,472 respondents, or about 75 percent of the respondents to the survey.
Detailed descriptions of employer pension plans were obtained for two-thirds of respondents with plans on their current jobs and for those with a plan on their last jobs if not currently working, and for about a third of the pensions from jobs held before the current or last job. It is important to note that, for a number of reasons, measurement of pension values is less precise than the measurement of Social Security benefits.
From the HRS survey, we obtain the following measures of respondents’ knowledge of Social Security: Do you expect to receive Social Security benefits in the future? At what age? How much will the benefits be in today’s dollars? Similar questions are asked about pensions, along with additional questions about the specific features of the respondents’ pension plans. Retirement planning activities include things like talking about retirement with one’s spouse or friends or relatives, thinking about retirement, and attending employer-sponsored retirement meeting.

Summary of Major Findings
Knowledge of Social Security and Pensions
  • Only half of the respondents who expect Social Security benefits are able to say what they expect their benefit to be. Regarding pension, 41% say they don’t know what their pensions are worth.
  • Comparing what people expect to receive with what they actually receive, we find that 27 percent of respondents estimate their Social Security benefits within 25% of their actual benefit amount. Of respondents who make a guess that is less accurate, 14% are too pessimistic and underestimate their benefits, and 10% are too optimistic and overestimate their benefits. The numbers are a little worse for pensions with 16% accurate within 25%, and of those who are less accurate, 25% understate their likely benefits and 17% overstate them.
  • Demographic comparisons reveal that women do a poorer job of estimating their Social Security benefits than men; whites do better than Hispanics and blacks; married people are better informed as are those with more schooling. Respondents in the oldest cohorts (closer to retiring) do better than younger, and those in the lowest income and wealth deciles do worse than those in the highest. A similar pattern is revealed for knowledge of pensions.
  • Those for whom Social Security represents a larger proportion of their total wealth--are more dependent on it--are less likely to know what their benefits will be than those for whom it is a smaller proportion. The opposite is true for pensions: those who are more dependent on pensions are likely to know more about them.
  • Retirement planning activities seem to increase knowledge for both types of benefits, although only modestly. One exception is that respondents who report having asked the Social Security Administration for a calculation of their benefits are much more likely to say they know what those benefits will be.
The Relation of Knowledge of Social Security Benefits and Pensions to Retirement Outcomes
  • Overall, there is only modest support for the expectation that those who have engaged in retirement planning activities are likely to retire earlier.
  • Individuals who overestimate their Social Security and pension benefits are likely to retire later than planned, but the effects are very weak.
  • Respondents who said they did not know what their benefits would be were less likely to be retiring in the near future.
  • Adding measures of knowledge of Social Security and pension benefits to standard retirement models has little impact on the predicted effects of Social Security and pension variables on retirement.
The Relation of Knowledge of Social Security Benefits and Pensions to Other Wealth
  • Measures of knowledge are only weakly related to measures of non-Social Security and non-pension wealth.
  • Among the retirement planning measures, talking with one’s spouse about retirement is associated with slightly higher wealth. Overall, retirement planning has modest but statistically significant effects on non-Social Security and non-pension wealth.
  • Adding measures of Social Security and pension knowledge to models of saving has little impact on the estimated effects of Social Security and pension variables on saving.

As seen in this study, there is a great deal of misinformation about Social Security and pension benefits and rules. Those who are most dependent on Social Security benefits are least well informed about them. We find the opposite for pensions: those who are most dependent on pensions are most well informed about their pension benefits. Women and minorities are less well-informed about both types of retirement benefits.
Overall, there is only a modest relationship between measures of information and retirement planning activities--those who plan are somewhat better informed than those who do not. However, we find that having requested a report from the SSA is strongly related to knowledge of one’s Social Security benefits, suggesting that providing information on request is a helpful policy. However, people who request reports are likely to have more information in the first place. Future work should attempt to evaluate the SSA’s current policy of providing the report of personal earnings record and projected retirement benefits to all recipients. Lastly, we find a modest relationship of knowledge and planning activities to planned and actual retirement and to non-Social Security and non-pension wealth, and little effect of including knowledge variables in retirement and saving equations.
Although this investigation yielded some suggestive results, the puzzle is far from solved. In economic research, a common assumption is that people are well-informed and forward-looking in their behavior. Yet we find many who are not well-informed about their future retirement benefits, and they seem to behave no differently from those who are well informed with regard to saving and retirement. Moreover, despite being misinformed, they do not adjust their behavior as they approach retirement age and become aware of the true situation.

Alan L. Gustman is a Loren Berry Professor of Economics at Dartmouth College, Department of Economics, Hanover, New Hampshire 03755 (alan.l.gustman@dartmouth.edu)

Thomas L. Steinmeier is a Professor of Economics, Texas Tech University, Department of Economics, Lubbock, Texas 79409 (Thomas.Steinmeier@TTU.edu)

back to the top.

Sandell Grants Awarded
We are pleased to announce that the MRRC has awarded the Sandell grant to two recipients: Marie-Eve Lachance and David McCarthy. Marie-Eve Lachance is a Ph.D. candidate in the Department of Insurance and Risk Management at the University of Pennsylvania’s Wharton School. Her fields of specialization are pensions and finance. Specific areas of interest include the design, valuation, and investments of pension plans. Recently, many social security systems and public sector pension plans have converted their pension design from defined benefit to defined contribution. Her on-going work, along with Olivia S. Mitchell, studies the implications of these conversions with regard to transferring investment risk to individuals. Part of this research is centered on the current proposal to include Personal Retirement Accounts in the US social security system. This study evaluates the financial implications of adding a minimum guarantee to these accounts as a way of mitigating investment risk. Other work includes the evaluation of a different approach taken by the Florida Retirement System in a similar context. In that case, employees were granted an option to return to the original defined benefit system after transferring to the newly created defined contribution plan. Before attending graduate school, Marie-Eve worked for three years as a pension consultant in Canada. She received her Bachelor’s degree in Actuarial Science from Laval University (Canada) in 1996 and the professional designations of Fellow of the Society of Actuaries (FSA) and Chartered Financial Analyst (CFA).
David G. McCarthy is a Ph.D. candidate in the Department of Insurance and Risk Management at the Wharton School of the University of Pennsylvania. His areas of interest include annuities markets, portfolio choice and the institutional aspects of retirement saving. Recent work with Olivia Mitchell examines the impact of adverse selection in annuities and insurance markets, and the role of housing in the retirement portfolio. Ongoing work includes studying the role of wage-indexed claims in retirement portfolio choice, further empirical studies of adverse selection in annuities markets, and an empirical examination of pricing of PCS catastrophe derivatives. Before attending graduate school, he was an actuarial trainee at the Liberty Life Group, a South African life insurance company. David received his Bachelor’s degree from the University of the Witwatersrand in Johannesburg, South Africa in 1995, and qualified as a Fellow of the Faculty of Actuaries in Edinburgh, UK in 1999.

back to the top.

Researcher Workshop
In March of this year, a group of MRRC researchers met in Santa Monica, California at RAND for a researcher workshop. The day was structured to allow maximal time for discussion and exchange. Researchers were asked to make very brief presentations about current and planned work with time in between for discussion led by session chairs. Tom Juster (Acting Director, MRRC) chaired a session on the Macro Effects of Reform in which John Laitner (University of Michigan), Hugo Benítez.Silva (State University of New York- Stony Brook) and Shripad Tuljapurkar (Stanford University) each presented. John Laitner chaired a session on Retirement Incentives in which Alan Gustman (Dartmouth College), Julie Zissimopoulos (RAND), and Jennifer Ward-Batts (Claremont McKenna College) presented. Robert Schoeni (University of Michigan) and Richard Burkhauser (Cornell University) presented in a session on the Distributional Effects of Reform that was chaired by Alan Gustman. Elizabeth Powers (University of Illinois, Champaign) and Steve Haider (RAND) each spoke about their work that is focused on Interactions Among Social Insurance Programs in a session that was chaired by Richard Burkhauser. To end the day, John Laitner chaired a session on New Data Sources. Robert Willis (University of Michigan) and Robert Schoeni talked about data that are coming on-line from the Health and Retirement Survey and the Panel Study of Income Dynamics, respectively. Other MRRC researchers who attended but did not make presentations included: Beth Asch (RAND), Michael Hurd (RAND), David Loughran (RAND), John Phillips Social Security Administration), Kent Smetters (Wharton College), Tom Steinmeier (Texas Tech University), David Weir (University of Michigan). Also in attendance were Arie Kapteyn (RAND) and MRRC staff members Amanda Sonnega and Becky Bahlibi.

Bob Willis, left, and John Laitner at the workshop.

back to the top.

For Your Information
How to Request Your Social Security Statement
Your Social Security Statement is a record of your earnings history and an estimate of how much you and your employer paid in Social Security taxes. It also provides you with estimates of benefits that you (and your family) may be eligible for now and in the future.
The Social Security Administration now has a service that allows you to request this statement either on-line or by mail. If you have internet access, you can visit their website and make the request on-line at www.ssa.gov/onlineservices. You can also use the SSA website to obtain a form that you can mail to them to request your benefit statement. To obtain this form go to www.ssa.gov/online/ssa-7004.html. Or you can call them at 1-800-772-1213 or “TTY” 1-800-325-0778.
Regardless of how you request the statement, you will receive a response to your request by U.S. mail in 2-4 weeks.
To request your Social Security Statement, you will need:
  • Your name as shown on your Social Security Card
  • Your Social Security Number
  • Your date of birth
  • Your place of birth
  • Your mother's maiden name- last name only (to help identify you)
You will obtain a better estimate of your benefits if you can also provide your last year’s earnings and an estimate of your current and future earnings as well as the age at which you plan to stop work.

Source: Social Security Administration Web Site, 2000

back to the top.

RAND HRS Data Now Available
The RAND HRS Data file is a cleaned and easy-to-use version of data from five waves of the original 1992 Health and Retirement Study cohort. Derived variables covering a broad, though not complete, range of measures have been constructed and named consistently across waves. It incorporates data from 1992, 1994, and 1996 Final Release versions of HRS data and from 1998 and 2000 Early Release versions. Work to incorporate the final 1998 data is in progress. There are no data for members of the AHEAD study nor for the new cohorts added to HRS in 1998. You can access these data by visiting their website at http://www.umich.edu/%7Ehrswww/docs/rand/index.html or through the MRRC website. This work is supported by the Social Security Administration and the National Institute on Aging.

back to the top.

University of Michigan
Institute for Social Research
426 Thompson Street
3026 ISR
Ann Arbor, MI 48106-1248

Phone: 734-615-0422
Fax: 734-615-2180
Email: mailto:mrrcumich@umich.edu