January 2008
Volume 9, Issue 1

Director’s Corner

The long-anticipated retirement of the Baby Boom generation has begun: on October 15, 2007, the Nation’s first baby boomer, Kathleen Casey-Kirschling, filed on-line for her Social Security Retirement benefits. Over the next two decades, nearly 80 million Americans will become eligible for Social Security benefits. Combined with relatively low fertility and increasing longevity, this demographic wave will certainly challenge Social Security. Guaranteeing the system's success into the future remains a task for policy makers to address.

Of the many potentially important variables influencing Social Security’s finances, one that is something of a wild card is the labor force participation of older Americans. Over the years, MRRC has funded a variety of research projects on this topic. This issue’s featured researcher, Nicole Maestas, examines several different aspects of the issue--for example, factors affecting returns to work after retirement. Should older Americans decide to increase their labor force participation substantially, the magnitude of Social Security’s impending solvency shortfalls could be reduced, and the economy's concerns about possible looming labor shortages might ease.

John Laitner, Director MRRC

FYI: File Online for Social Security Retirement Benefits

Did you know that you can file for your Social Security retirement benefits online? The Social Security Administration has developed a very user-friendly and safe way to do it. Social Security’s online services are convenient and secure.
In The Federal American Customer Satisfaction Index (ACSI) Scorecard, Social Security’s online benefit application has the highest score among all federal websites followed closely by Social Security’s application for extra help with Medicare drug costs.

When the Nation’s first Baby Boomer, Kathy Casey-Kirschling, filed for her reirement benefits last October, she did so using Social Security’s on-line application process. “Social Security provides economic protection to millions of Americans and their families,” Commissioner Astrue said. “And we remain committed to providing the public with the highest quality service possible. Boomers have always been a generation of trendsetters. Kathy is leading the way by doing business with Social Security online.”

To apply for Social Security reitrement benefits on-line, go to www.socialsecurity.gov and follow the link to “apply for benefits.”

back to the top.

Regular Features


Key Findings

Researcher Q & A

Inside This Issue

Research Briefs


MRRC Researchers Receive Prestigious Awards

Annamaria Lusardi and Olivia S. Mitchell were awarded the 2007 Fidelity Research Institute Pyramid Prize for their work on advancing understanding of the importance of financial literacy and planning in helping Americans reach their financial goals. The prize is presented to authors of published applied research that the Institute believes best helps address the goal of improving lifelong financial well-being for Americans. The award-winning paper “Baby Boomer retirement security: The roles of planning, financial literacy, and housing wealth” first appeared as MRRC working paper WP 2006-114 and was subsequently published in the Journal of Monetary Economics, Volume 54, Issue 1, pps. 205-224.

MRRC researchers John Karl Scholz and Ananth Seshadri, along with Madison Surachai Khitatrakun of the Urban Institute, were awarded the 2007 TIAA-CREF Paul A. Samuelson Award for outstanding scholarly writing on issues related to lifelong financial security. Named in honor of the Nobel Laureate and former CREF trustee, this award is given each year in recognition of an outstanding research publication containing ideas that the public and private sectors can use to maintain and improve Americans’ financial well being. Their award-winning paper, “Are Americans Saving Optimally for Retirement?,” was published in the Journal of Political Economy, August 2006.

Visit the MRRC website www.mrrc.isr.umich.edu for these researchers' MRRC papers

Key Findings

Key Findings reflect main points of MRRC Working Papers. Visit our website for more.

Are 401(k) Saving Rates Changing? Cohort/Period Evidence from the HRS by Irena Dushi and Marjorie Honig WP 2007-160

  • Participation in 401(k) pensions is nearly 50 percent greater among workers born 1948-1953 (early Baby Boom) compared to workers born 1931-41.
  • This large growth in participation over a short period may reflect differences in saving preferences between the two cohorts, changes over this time in the retirement savings environment, or the joint effect of both.

The Impact of Private Participation on Disability Costs by Estelle James, Alejandra Cox Edwards, and Augusto Iglesias Palua WP 2007-161

  • The public disability insurance system in Chile differs from the traditional one in that it is largely pre-funded and the disability assessment procedure includes participation by private pension funds and insurance companies, who finance the benefit and have an interest in controlling costs.
  • We hypothesize that these procedures and incentives will keep system costs low, by cutting the incidence of successful disability claims.
  • Disability hazard rates are only 20-35% as high in the new system as in the old. Mortality rates among disabled pensioners suggest that the new system has accurately targeted those with more severe medical problems.

Financial Literacy and Stock Market Participation by Maarten van Rooij, Annamaria Lusardi, and Rob Alessie WP 2007-162

  • A basic understanding of stocks and bonds (concepts such as risk diversification) is quite limited among respondents of a Dutch national household survey.
  • Those who have low financial literacy are significantly less likely to invest in stocks.

Extra Help: Take-up of the Social Security Administration’s Low-Income Subsidy Program for Part D of Medicare in 2006 by Helen Levy and David Weir WP 2007-163

  • Take-up of Part D is very high with fewer than 10% of seniors left without coverage.
  • Those with higher use of prescription drugs or worse self-reported health in 2004 were more likely to sign up for Part D.
  • Take-up of the low-income subsidy is quite low. Many respondents report they had not heard about the extra help.

Future Beneficiary Expectations of the Returns to Delayed Social Security Benefit Claiming and Choice Behavior by Angela Hung and Jeff Dominitz WP 2007-164

  • Among workers who expect to stop working full-time before age 62, over 60% say they plan to delay claiming Social Security benefits until after they turn 63.
  • Among those who expect to stop full-time work between ages 62 and 70, only about 25% expect to delay claiming beyond their retirement age.
  • When presented with a hypothetical choice to delay claiming or not, proababilities of delayed claiming are high relative to previous findings.

back to the top.

Research Briefs

Estimating the Health Effects of Retirement

by John Bound and Timothy Waidman


As the large birth cohorts of the baby boom approach eligibility age for Social Security old-age benefits, concerns for the fiscal stability of the program have prompted proposed policies to encourage aging workers to postpone retirement. If successful, they would both increase revenues into the system and decrease benefit payments. Beyond fiscal implications, however, delayed retirement may lead to other outcomes that make the social welfare impact ambiguous. One outcome that has not been well-studied is the potential health impact of delayed retirement, and one can imagine retirement being either a positive or negative influence on health. For example, for an individual in a physically demanding job that constitutes an ongoing stressor to health, retiring could be good for health, and postponing retirement could impose an added burden on individuals as well as potentially increased costs for private and public financing of health care. Alternatively, one could imagine scenarios in which retirement leads to the worsening of a person’s health. For example, mental activity has been associated with reduced risk of dementia, and the physical activity of work may serve to reduce risks of obesity and diabetes.

Other researchers who have examined the effect of health on retirement have typically compared the health of individuals before and after they retire. The estimates produced by this method suggest that retirement has large negative effects on both physical and mental health. However, if deteriorating health leads individuals to retire, this approach will tend to seriously exaggerate the negative effects of retirement on health.

As an alternative, we attempt to isolate retirement decisions that are exogenous to health and use them to estimate the magnitude of any direct effect of retirement on health. The idea is to compare health indicators immediately before normal retirement age. In the United Kingdom, a significant portion of retirement income for large numbers of individuals is available at age 65 for men and 60 for women. This feature provides a strong incentive for retirement at those ages, and data on labor force attachment show drops at those ages. On the other hand, there is nothing unique about those ages that influences health. By contrast, in the U.S. there are two (one at 62 and one at 65) rather than one spike in retirements, and retirement affects health insurance status.

Using the second wave of the English Longitudinal Study of Aging, conducted in 2004 and 2005, we measure differences in a variety of types of health measures for men and women both before and after their respective normal retirement ages. These include clinical measures of heart disease risk, physical performance measures of balance and mobility limitations as well as self-reported indicators related to those measures, and general measures of health. We also examine age patterns of mortality in England, for several sets of causes.


In contrast to those who find negative health consequences of retirement, when we concentrate on exogenous retirement incentives, we find no negative health effects, and some evidence that retirement may actually improve balance, mobility and heart disease risk, at least in men. While there are problems with the interpretation of self-reports, these also tend to show improvements associated with retirement. Mortality, however, does not show a consistent pattern that can be well-explained by retirement.


It is natural to speculate about whether our estimates have implications for the advisability of changing early or normal retirement age in the U.S. Our estimates would seem to suggest that delaying retirement would tend to postpone the positive health effects of retirement. However, this does not mean that recent shifts that have been introduced to encourage the delay in retirement are a mistake. Instinctually, as economists, we tend to favor policies that do not distort behavior. In the context of retirement policy, this means policies that neither encourage nor discourage early retirement. Recent changes in Social Security rules such as the increase in the delayed retirement credit reduced the disincentives built into the program for delayed retirement. As long as individuals are aware of the health consequences of retirement, as we suspect they are, there is no reason to shy away from such policies. By the same token, our results suggest that policies based on the paternalistic notion that work is good for individuals is mistaken.

This brief is based on MRRC working paper WP 2007-168


Managing the Risk of Life

by Adeline Delavande and Robert J. Willis


An individual eligible to receive Social Security (SS) benefits may first claim benefits at age 62 or may delay claiming benefits until a later age. Life Cycle Models predict that single individuals who expect to be long-lived will delay applying for benefits because they perceive the increase in SS benefits they will receive from claiming later will be financially beneficial. The claiming decision for married individuals is more complicated because spouses are entitled to spouse’s and survivor’s benefits and can adopt a joint strategy to maximize the total lifetime amount the couple receives from SS.

Earlier work (Delavande, Perry and Willis 2006 and Hurd, Smith and Zissimopoulos 2004) found a small effect of survival expectations on claiming decisions. However, those studies overlook two important aspects of the decision-making process: (1) Individuals will delay claiming when beneficial only if they are aware that delaying claiming is associated with larger monthly SS benefits in the future, and (2) Married individuals may take into account the survival of both spouses and the benefits of both spouses when deciding to claim to maximize the lifetime SS benefits received by the couple, rather than by the individual.

We analyze the role of individual’s and spouse’s survival expectations and knowledge about SS rules on the decision to claim SS benefits. Our innovation is to take into account the incentives provided by SS to married respondents, as well as individual’s knowledge about SS rules. We take advantage of unique data collected in the Health and Retirement Study (HRS) 2004 that provides information on whether individuals are aware that claiming early reduces benefits: Respondents are asked to estimate their expected benefits if they were to claim at age 62 and at their normal retirement age (NRA). Our empirical analysis evaluates the effect of survival expectations and knowledge on the expected claiming age (ECA) reported by respondents less than 62 years old in the 2004 HRS.


Knowledge about whether claiming early reduces future SS benefits

Forty-six percent of respondents report not knowing what their SS benefits would be conditional on various claiming ages, and 14 percent are not aware that their SS benefits might be reduced if they claim before their NRA.

There are important gender, age, education, and ethnicity differences in knowledge (and perception of knowledge) about the effect of claiming age on the level of benefits. A large proportion of women report that they do not know their expected SS benefits might be in the future (51 percent compared to 39 percent of men). Respondents with more education are more likely to report their expected benefits and to correctly know the effect of claiming age on benefits than respondents with lower education. Black and Hispanic respondents are more likely to report that they do not know what their benefits would be than their counterparts. In addition, the proportion of black respondents reporting an incorrect effect of claiming age is higher than that of non-Blacks (18 percent versus 13 percent).

Determinants of expected claiming age for single individuals

We find that single men and women base their claiming decisions on their survival probabilities.

As predicted by theory, individuals who expect to be long-lived plan on delaying SS claiming. However, knowledge about the penalty influences single men and women differently. Women who know that claiming at 62 reduces SS benefits compared to claiming at the NRA—which is the case for 36 percent of single women— are more likely to expect to delay claiming than women who do not answer the questions about expected SS benefits and than women who do not know about the penalty for early claiming. Moreover, when we allow for differential effects of survival on knowledge, the coefficient associated with subjective survival is statistically significant only for women who know about the penalty. For men, knowledge about penalty is not statistically significant when estimating expected claiming age. Moreover, when we allow for differential effects of survival on knowledge, we find that the coefficient of subjective survival is statistically significant for all men who provided their expected SS benefits if they were to claim at age 62 and at NRA, independent of whether they knew about the penalty, but it is not significant for men who did not answer the knowledge question.

Determinants of expected claiming age for married individuals

We find that married respondents are responsive to incentives provided by the SS rules. In particular, married men base their ECA on their spouse’s survival expectations but not on their own expectations: Husbands who expect their wife to be long-lived report a higher ECA.

This is consistent with the fact that husbands want to maximize the survivor’s benefits that their wife will received, which depends on the husband’s claiming age. For married women, however, both own and spouse’s subjective survival has a statistically significant influence on the ECA, and the coefficients are positive and of similar magnitude. Thus, everything else equal, a married woman whose husband has higher subjective survival expects to claim later, which is consistent with the fact that she expects to rely on spouse’s benefits (or own), rather than survivor’s benefits, for a long period. This spouse’s (or own) benefits are reduced when a woman claims early. We also find that women married to an older husband expect to claim earlier. This result is consistent with the fact that those women expect to rely on survivor’s benefits for a longer period, which is not affected by their claiming age. Those women probably plan to enjoy reduced spousal or own benefits for a short period before receiving survivor’s benefits. If we allow for a differential effect of survival by knowledge, we find that the coefficient of own subjective survival is statistically significant only for women who know about the penalty incurred by claiming early. Married men who do not know that claiming early reduces SS benefits expect to claim earlier.


There is substantial heterogeneity in the level of knowledge about SS rules according to demographic characteristics. The lack of knowledge among underprivileged groups, such as respondents without high-school degrees or Blacks, is of concern for the well-being of those groups in retirement, because they might rely heavily on SS benefits in old age.

Our results suggest that individuals are responsive to the incentives provided by the SS rules. Single individuals take into account their survival expectations when deciding when to claim. Married individuals seem to make claiming plans to maximize the total lifetime amount the couple will receive from SS. This suggests that it is important to consider the incentives provided to couples—not just individuals—when evaluating the impact of various SS reforms. Women’s claiming decisions are importantly influenced by their level of knowledge. Given the fact that women are less knowledgeable than men, targeting education programs toward women is likely to have an impact on their claiming decisions.

This brief is based on MRRC working paper WP 2007-167

back to the top.

Researcher Q & A

Nicole Maestas discusses her MRRC-supported research.

Q: Talk about your interest in labor supply behavior. 

In graduate school, I was interested in questions of poverty. Working on an NIA training grant with Ron Lee at Berkeley, I discovered that it was at least as interesting to me to look at questions related to poverty in the elderly as it was to examine childhood poverty. In fact, some of the issues were even more interesting to me as an economist because of this complex retirement decision. The decisions people make and the life events that occur leading up to and in retirement are very important for elderly poverty and well-being.

In modeling retirement, there are several important concerns including the individual’s own preferences and constraints such as the pension environment. There has been a lot of work attempting to characterize pensions and their impact on behavior, but less attention has been paid to the role of preferences in the retirement decision. Even though economic incentives clearly are important, there is a large unexplained component of behavior that I have always found very interesting.

The Health and Retirement Study (HRS) has a wide range of variables measuring different psychological constructs that may allow us to identify, in economic models, aspects of preferences we have only guessed at until now. In my papers for MRRC, I have made use of the HRS to explore various aspects of the retirement decision. In particular, I have been interested in choices about retirement timing and decisions about either continued work or return to work after retirement.

In my first paper for MRRC (WP 2004-085), I attempted to describe the patterns of retirement and work after retirement observed in the HRS. In the simplest life-cycle model, people retire when the cost of working an additional period just equals the benefit. The benefit includes income as well as pension accruals. The cost is seen as the value of foregone leisure. People enter retirement once and for all, completely withdrawing from the labor force. On the contrary, only half of respondents in the HRS follow this pattern. The other half pass through partial retirement on the way to full retirement and/or retire and then go back to work, or “unretire.” I wanted to think about what we would need to add to this very simple life-cycle framework in order to generate behavior like this. For example, is it just uncertainty (for example, about asset returns) or is something else going on?

I found that most people who “unretire” had planned to do so before retiring. Going back to work doesn’t seem to be something driven by unexpectedly low retirement income. For the most part, unretirement is qualitatively similar to partial retirement. However, those in worse health are more likely to partially retire rather than unretire, possibly reflecting a need for continuous health insurance coverage during retirement. Second, those who said they did not plan to work in retirement were more likely to unretire than partially retire, which suggests that unretirement remains an option for the small subset of individuals who find themselves with inadequate retirement income.

Q: What is a discouraged worker?

The findings from the uretirement paper suggest that there are many people who apparently desire to remain in or return to the labor force. There were also a large percentage who said they expected to work during retirement but who, in fact, did not. In a second paper for MRRC (WP 2006-133), Xiaoyan Li and I sought to understand more about job search outcomes for older workers. This is especially important to understand more about given calls for policies that explicitly encourage the elderly to work.

So we were interested to what extent the elderly can readily find suitable jobs. In fact, we find that only half of older searchers find jobs. After examining several potential explanations (such as varying search intensity, intervening health problems, expecting too high a wage), we find that about 13 percent of older job searchers become discouraged, which means that they are willing to work at a reasonable wage but are unable to find a job.
Q: Describe your work on job burn out and how it potentially affects labor supply decisions.

Job burnout is a psychological response to chronic work stress. People who are burned out feel exhausted, cynical, and ineffective--essentially, they’re sick of their jobs. In WP 2007-166, Xiaoyan Li and I speculated that burnout might be important for retirement behavior. We hypothesized that since burnout is related to work, you should see it rise prior to retirement and fall afterward. Also, you might expect that people who are experiencing burnout would come to favor leisure more highly. Likewise then, you might see the value of leisure rise prior to retirement but fall after retirement. As the value of leisure falls over time and the person recovers from burnout, working may begin to appear attractive again.

Our findings support this hypothesized pattern. We are really excited by these findings because they are strong and robust. Burnout is a strong predictor of retirement (both complete retirement and unretirement) even after controlling for health. Interestingly, burnout is much lower among those who partially retire, and remains steady over time.

Q: There is a lot of interest in whether or not older Americans can be attracted back to the work force. What factors from your research do you think affect that probability?

It’s a good news and bad news story, I think. The first thing is desire, and I think we see evidence that there is a good deal of interest in some form of working past normal retirement age. We find half of people in the HRS follow a nontraditional retirement path. They either partially retire on the way to full retirement, or they go back to work after retiring. That seems like a pretty good sign of interest in working.

I do think that retirement is an opportunity for many people to take a break and make a plan to do something different. The catch is that for many people something different includes working part-time. It’s not clear how willing employers are to go for part-time. It may be that, as employers become more and more in need of workers as the numbers of retirees increases, they might be more willing to offer part-time and flex-time arrangements for older workers. Certainly some employers already see some benefit in this, and I think more employers will with time. So that’s the good news.

The second thing is health, and here the news is not good. Health really distinguishes who is going to go back to work and who is not. In the burnout paper, we show very clearly that you can have two equally burned out people, but the person with other health problems will not be likely to go back to work. They are much more likely to retire completely than the person who doesn’t have health problems.

There are a couple of different issues that are important to consider here. First, you might have someone whose financial plan for retirement included working. However, when they come to retirement age, they find that they’re not able to because of poor health. A second issue that we uncovered in the discouraged workers paper is that there are many older people looking for jobs, but we don’t observe them actually transitioning to jobs. Many of them have chronic health problems, but are still willing and able to work. We don’t know for sure, but it may be that they have a hard time finding employers who are willing to accommodate their health issues whether it be via health insurance, hours, performance criteria, or physical accommodations. I don’t think we have a good sense yet of what these problems are and whether and how they can be addressed.

Q: What direction do you see this work taking?

I see us moving in at least two directions. The first is going into greater depth considering the psychology of retirement. The burnout paper was a first attempt to do that, and we will certainly continue to pursue those questions in greater depth. I think our burnout paper showed that these psychological constructs matter a lot, something economists now appreciate more than ever. The other direction for us is beginning to unpack the black box of health status and understand more about different types of health trajectories that people follow from their 50s through their 70s. For example, what if based on their inventory of health conditions and risk factors at age 45 people could forecast their future potential labor supply into their 70s? This could help them figure out to what extent they can rely on future potential earnings to help fund retirement or whether they need to save more now while they can. This would be very important information for retirement planning.


Burnout and the Retirement Decision by Nicole Maestas and Xiaoyan Li WP 2007-166, October 2007.

Discouraged Workers: Job Search Outcomes of Older Workers by Nicole Maestas and Xiaoyan Li WP 2006-133, Ocotber 2006.

Back to Work: Expectations and Realizations of Work After Retirement by Nicole Maestas WP 2004-085, July 2004.

Nicole Maestas (Ph.D., Economics, UC Berkeley) is an Economist at RAND. Her research interests include applied microeconomics, the economics of aging, and health economics. She has developed and estimated models of the retirement and annuitization behavior of married couples, and is currently working on a comparative study of retirement behavior by married couples in the U.S. and U.K. In past and current MRRC-funded work, she is studying labor force transitions of older workers, with emphasis on work after retirement, job search and labor demand, and self-employment, and, most recently, psychological burnout. In other work, she has studied the effect of medical expenditure risk on the portfolio choices of older Americans, pricing in the Medigap insurance market, and the effect of the Medicare program on disparities in health, health care utilization, and treatment intensity.

back to the top.

Panel Study of Income Dynamics Conference:

Call for Papers on Pensions, Retirement Planning, and Financial Well-being Over the Life Course

The Panel Study of Income Dynamics (PSID), with support from the National Institute on Aging, announces a call for papers that use PSID data to conduct research in the area of pensions, retirement planning, and financial well-being over the life course. Papers will be presented at a two-day conference on life course financial well-being in November 2008 in Ann Arbor, Michigan.

The PSID has included extensive information on pensions, wealth, consumption/expenditures, health, and other relevant domains in various years since the survey began in 1968. All of these domains are measured in the last several waves.

The goals of this conference are to bring together scholars in the area of life course financial well-being to generate additional scientific and policy-relevant findings, facilitate future collaborations including NIH proposal submissions, and provide feedback to PSID about possible needs for content changes and/or future data collections.

How to Apply:

To apply, Please send a 1-3 page abstract describing your study including PSID data that will be used. Also include your CV.


Papers will be chosen based on thematic relevance and significance to the conference.


All persons are eligible.


Conference expenses, including transportation and lodging, will be provided to one author per paper.


Please send abstracts to Bob Schoeni in care of Patty Hall (pathall@umich.edu) by March 17, 2008. A review committee will evaluate the abstracts, and its decisions will be announced shortly after the submission deadline. Authors of accepted papers will be required to submit their paper by October 15, 2008 for circulation prior to the conference, and present their paper at the conference.

For further information please contact Bob Schoeni (bschoeni@umich.edu) or Patty Hall (pathall@umich.edu). For more information on the PSID, please visit the website: www.psidonline.org

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