Research Findings in the Economics of Aging
edited by David A. Wise
University of Chicago Press, 2010
Cloth: 978-0-226-90306-4 | Electronic: 978-0-226-90308-8
DOI: 10.7208/chicago/9780226903088.001.0001


The baby boom generation's entry into old age has led to an unprecedented increase in the elderly population. The social and economic effects of this shift are significant, and in Research Findings in the Economics of Aging, a group of leading researchers takes an eclectic view of the subject. Among the broad topics discussed are work and retirement behavior, disability, and their relationship to the structure of retirement and disability policies.

While choices about  when to retire are made by individuals, these decisions are influenced by a set of incentives, including retirement benefits and health care, and this volume includes cross-national analyses of the effects of such programs on these decisions. Furthermore, the volume also offers in-depth analysis of the effects of retirement plans, employer contributions, and housing prices on retirement. It explores well-established relationships among economic circumstances, health, and mortality, as well as the effects of poverty and lower levels of economic development on health and life satisfaction. By combining micro and macro evidence, this volume continues a tradition of expanding the research agenda on the economics of aging.


David A. Wise is John F. Stambaugh Professor of Political Economy at the Kennedy School of Government at Harvard University and area director for Aging and Health Studies at the National Bureau of Economic Research.



- David A. Wise
DOI: 10.7208/chicago/9780226903088.003.0001
[disability rate, central policy, aging, health, retirement]
This chapter provides an overview of the studies contained in the volume, relying to a significant extent on the authors' own language to summarize their findings. The chapters in the volume consider work and retirement behavior, work disability, and their relationship to the structure of retirement and disability policies. These issues are central to discussions of population aging and its impact, because the ages at which people leave the labor force, whether through retirement or disability, define the effective dependency ratio in the population and, by consequence, the associated economic strain that we collectively confront. In many past studies of disability trends, a consistent finding is the strong correlation between education and functional ability. Those with more education are less likely to develop functional limitations and appear to cope more effectively with functional limitations when they do develop. The authors of this book note that nearly half of elderly people with less than a high school degree report some difficulty caring for themselves; whereas only about a quarter of college graduates report that they are disabled. The lower disability rate among those who are better educated results in substantial differences in health and medical spending, in employment and earnings, and in many aspects of functional independence. Understanding why education is related to disability and whether changes in education have contributed to disability declines is thus a central policy concern. (pages 1 - 14)
This chapter is available at:
    University Press Scholarship Online

I. Disability, Work, and Retirement

- John B. Shoven
DOI: 10.7208/chicago/9780226903088.003.0002
[age, labor force, mortality, life expectancy]
This chapter deals with several innovative approaches to thinking about age. The traditional measure of age is a measure of years-since-birth. The chapter suggests that for some purposes, age measures that are more closely associated with health and longevity might be more appropriate. For example, a simple alternative to years since birth would be a measure of age based on mortality risk. Groups whose mortality risk is high would be considered old, those with low mortality risk would be classified as young and those with the same mortality risk would be considered the same age. Another approach would be to measure age from the other end of life, based on remaining life expectancy (RLE). Those with a short RLE would be considered elderly and those with a long RLE would be considered young. One advantage of the RLE approach over the mortality risk approach is that it is measured in years, units that are widely understood. (pages 17 - 36)
This chapter is available at:
    University Press Scholarship Online

- Axel Börsch-Supan
DOI: 10.7208/chicago/9780226903088.003.0003
[demography, health, disability insurance, social safety, retirement program]
This chapter points to a tradeoff in the provision of disability insurance. Understanding the trade-off between its role as a health dependent social safety net and its use as a broader early retirement program is important for the design of a modern social security system, particularly at a time of financial strain on income support systems in most countries. Three explanations commonly offered for the large variation in disability insurance spending across countries: demographics, health, and institutions. First, while all European countries are aging, the specific age demographics of the population vary considerably from one country to another. A second potential cause for the cross-national variation is international variation in health and disability status, beyond just the difference in demographic composition. Third, public pension and disability support systems exert large incentive effects, which, according to each country's legislation and policies, significantly increase or decrease the take-up of benefits. The results of the cross-country study indicate that demographic and health-related differences explain very little of the cross-national variation in disability enrollment rates. (pages 37 - 62)
This chapter is available at:
    University Press Scholarship Online

- Arie Kapteyn, James P. Smith, Arthur van Soest, James Banks
DOI: 10.7208/chicago/9780226903088.003.0004
[labor market, Panel Study of Income Dynamics, disability, insurance]
This chapter analyzes variations in work disability and enrollment in disability insurance in Europe and the United States. Data from the Panel Study of Income Dynamics (PSID) and the European Community Household Panel (ECHP) is looked upon. An additional aspect of this chapter is a simulation analysis that applies U.S. institutional variables to the population characteristics of other countries to see how self-reported work disability would change in the other countries in the study. These institutional differences include program eligibility rules, workplace accommodation of older or sick workers, and generosity of benefits. By simulating work disability using U.S. parameters (i.e., U.S. institutions and norms), but applied to European countries, there is often a considerable reduction in self-reported disability rates. (pages 63 - 98)
This chapter is available at:
    University Press Scholarship Online

II. Education and Disability

- David M. Cutler, Adriana Lleras-Muney
DOI: 10.7208/chicago/9780226903088.003.0005
[old age, education, health behavior, smoking, obesity]
There are many theories about the link between education and disability, ranging from childhood conditions that affect education and disability, occupational differences in the working years, differential health. Three factors have particularly strong influences on the education gradient in disability. The first is health behaviors. Better, educated people are significantly less likely to smoke than are less educated people; they are also less obese. About one-third of the education gradient in disability is found to be associated with differential health behaviors. Another third is explained by differences in lifetime occupation. People are broken down by hard work. Finally, differential rates of medical conditions explain another fifth of the education gradient in disability. Stroke, heart disease, and chronic conditions such as diabetes and arthritis are highly related to disability. Less educated people are more likely to have suffered from these conditions, partly as a result of their greater propensity to smoke and to be obese. (pages 101 - 120)
This chapter is available at:
    University Press Scholarship Online

- David M. Cutler, Edward L. Glaeser
DOI: 10.7208/chicago/9780226903088.003.0006
[smoking, health, policy, interpersonal]
This chapter discusses whether people are more likely to smoke when smokers surround them. Several reasons are suggested that peers might matter for health-related behaviors. In many cases, health-related behaviors are more fun to do when others are doing them too (e.g., drinking). Peers are a source of information about health (the benefits of a mammogram) or what is acceptable in society (the approbation accorded smokers). These interpersonal complementarities can have enormous social impact. Peer effects magnify the impact of policy interventions. The existence of social interactions implies that a policy intervention has both a direct effect on the impacted individual and an indirect effect, as that person's behavior affects those around them. These indirect effects create a social multiplier where the predicted impact of interventions will be greater when the interventions are imposed in larger geographic areas. (pages 123 - 141)
This chapter is available at:
    University Press Scholarship Online

- Steven J. Atlas, Jonathan Skinner
DOI: 10.7208/chicago/9780226903088.003.0007
[education, pain, college, neurology, health]
This chapter begins by documenting the dramatic differences across educational groups in the prevalence of pain. The chapter finds significant differences across educational groups, with rates of people aged 55 to 59 troubled by pain ranging from 26 percent for women with a college education to 55 percent for those without a high school diploma. The prevalence of pain declined with age. One might think that these differences result from those with lower education being more likely to have worked in manual jobs, or to experience other types of health impairments. The strong association between education and pain in the survey data and the clinical data are supportive of the view that educational attainment has an independent association with the neurological mediators of pain, or for social or even economic factors that may be associated with the perception of pain. (pages 145 - 166)
This chapter is available at:
    University Press Scholarship Online

III. Economic Circumstances and Health

- Abhijit V. Banerjee, Esther Duflo
DOI: 10.7208/chicago/9780226903088.003.0008
[age, death, dollar, low income, poor]
This chapter considers this relationship among those with extremely low incomes around the world. The research compares the consumption patterns and mortality outcomes of the very poor (living on less than one dollar per day), the poor (less than two dollars per day) and the somewhat less poor (two to four dollars per day, or six to ten dollars per day). The chapter also emphasizes the implications of relative poverty for mortality. Based on multiple pieces of evidence, the results all point in the same direction: the poor, and particularly the very poor, have a lower chance of survival than those who are somewhat better off. The proxy measure of longevity used in the study is the probability that an adult's mother and father are alive. The mother of someone who is not poor is more likely to be alive than the mother of someone who is poor. Using panel data for Indonesia and Vietnam, it is seen that older adults are more likely to have died five years later if they are poor. (pages 169 - 210)
This chapter is available at:
    University Press Scholarship Online

- Anne Case
DOI: 10.7208/chicago/9780226903088.003.0009
[health, old age, environment, nutrition]
This chapter explores the relationship between economic circumstances and health, focusing on health and environment in early childhood and its long-term effect on health and functioning in old age. There are two markers of health and environment in early life to assess their impact on health in later life. The first documents the extent to which height, as a measure of early life health and nutrition, is associated with more outcomes that are favorable in old age. The second looks at the conditions that might have affected a mother's nutrition while pregnant specifically, the success of corn crop production while she was pregnant, are predictive of health in later life, and whether this marker of mother's nutrition can explain the association between height and health outcomes in old age. The study finds that height is protective of health. Height appears to become more protective against hypertension, activities of daily living ADLs, and loss of fine motor skills at the oldest ages, when there is a higher risk of poor health. Height and corn production both have large and significant effects on health in old age. (pages 211 - 228)
This chapter is available at:
    University Press Scholarship Online

- Angus Deaton
DOI: 10.7208/chicago/9780226903088.003.0010
[health, age, gender, ethnicity, education]
This chapter looks at the effects of income and age on self-reported well-being in more than a hundred countries. It addresses in particular self-reports of life satisfaction, health, and disability, how these measures change with age, and how the effects of age differ across countries according to their level of development and their region of the world. The analysis is based on the Gallup World Poll, which collected data from samples of people in each of 132 countries during 2006. With few exceptions, the samples are nationally representative of people aged fifteen and older. Because the survey used the same questionnaire in all countries, it provides an opportunity to make cross-country comparisons while, at the same time, providing enough data to permit within-country disaggregation; for example, by age, gender, ethnicity, or education. (pages 235 - 268)
This chapter is available at:
    University Press Scholarship Online

IV. Retirement Saving

- James M. Poterba, Steven F. Venti, David A. Wise
DOI: 10.7208/chicago/9780226903088.003.0011
[retirement, lifetime earnings, 401(k) plans, social security, assets]
This chapter looks at how Social Security, 401(k) participation, and other asset accumulation fit together for households with different lifetime earnings and different Social Security wealth accumulations. Two broad categories of wealth are focused upon: dedicated retirement assets, which are made up of Social Security wealth, accrued benefits in traditional pension plans, 401(k) savings, and IRAs and Keogh plans; and undedicated assets, including nonretirement financial savings and housing equity. 401(k) participation varies substantially by income, broader measures of retirement assets show a “retirement replacement rate” (inclusive of both Social Security and retirement saving) and a “total saving rate” (including both dedicated and undedicated assets) that varies only moderately by lifetime earnings and by Social Security wealth. Combining projections of 401(k) assets with estimated Social Security wealth, the study finds that the combined rate of growth is surprisingly similar across earnings deciles, and translates to at least a doubling of retirement resources in most earnings and Social Security wealth deciles. The growth rate is lower in the bottom two deciles of lifetime earnings. These various results are indicative of a very dramatic shift in the landscape of financial resources available to retirees in the future. (pages 271 - 310)
This chapter is available at:
    University Press Scholarship Online

- John Beshears, James J. Choi, David Laibson, Brigitte C. Madrian
DOI: 10.7208/chicago/9780226903088.003.0012
[automatic enrollment, 401(k) plans, employer, company]
This chapter discusses the continuing series of studies on the structural features of 401(k) plans. Companies have used a variety of approaches to encourage participation in employer sponsored savings plans. The vast majority of large firms now offer the most common approach, the provision of an employer matching contribution. Even with a match, however, savings plan participation rates are often surprisingly low, and their effect on participation is found to be relatively small. Automatic enrollment is an alternative mechanism for increasing savings plan participation. All of the companies in which automatic enrollment has been studied to date have also offered an employer matching contribution. The study disentangles the effects of matching and automatic enrollment in two ways. The first is to study a large firm with automatic enrollment that replaced its employer match with a noncontingent employer contribution to the plan, thereby eliminating the incentive that was provided by the match. The second approach is to pool the participation data from nine firms, all with automatic enrollment, but with varying matching provisions. (pages 311 - 336)
This chapter is available at:
    University Press Scholarship Online

- James Banks, Richard Blundell, Zoë Oldfield, James P. Smith
DOI: 10.7208/chicago/9780226903088.003.0013
[housing, price volatility, residential mobility, household]
This chapter discusses several types of housing transitions made at older ages in Britain and the United States. The extent of residential mobility and the downsizing across multiple dimensions, including housing size (number of rooms), housing value, and ownership (as compared with renting) is looked upon. This chapter also looks at some of the determinants of mobility, with a particular focus on house price volatility, but also the role of major life events, such as retirement or widowhood. A particular contribution of the chapter is its analysis of a longer time horizon, in which there is more likely to be evidence of downsizing if it exists in the data. It is seen that after looking over a number of dimensions and over a number of transition intervals, downsizing is an important part of life for older households in both countries, but particularly in the United States. (pages 337 - 384)
This chapter is available at:
    University Press Scholarship Online

V. Medicare

- Jay Bhattacharya, Alan M. Garber, Thomas MaCurdy
DOI: 10.7208/chicago/9780226903088.003.0014
[Medicare, expenditure, balanced budget act, BBA]
This chapter investigates how the Balance Budget Act (BBA) affected Medicare expenditure trends. In the immediate post-BBA period, the most important effects were on inpatient services and home health services, where the intention was unambiguously to reduce Medicare payments. Other features of BBA, such as the introduction of prospective payment for some outpatient services and the creation of Medicare plus Choice plans, involved new payment mechanisms with the prospect for long-term expenditure control. The study assesses whether expenditures grew more or less rapidly for high-expenditure Medicare beneficiaries (as compared with people who used few Medicare-covered services) in the period following BBA's implementation. The study finds that after 1997, the growth in expenditures among the highest-cost users of Medicare-reimbursed care was less than growth among lower-cost users. Thus, the overall dispersion in expenditures fell over time. These findings suggest that the main effects of the BBA were realized as intended. (pages 387 - 412)
This chapter is available at:
    University Press Scholarship Online

- Florian Heiss, Daniel McFadden, Joachim Winter
DOI: 10.7208/chicago/9780226903088.003.0015
[Medicare, drug, survey data, Internet]
This chapter analyzes an Internet-based survey of individuals who became eligible for prescription drug coverage through Medicare Part D. The chapter focuses on the decisions made in the initial enrollment period, and the influence on plan choice of previous prescription drug use, health risks, health-related expectations, and subjective factors. It draws on all three waves of survey data. The study finds generally that seniors' choices respond in predictable ways to the incentives provided by their own health status and the plan options available to them. In some circumstances, however, consumers selected inexpensive plans even though more expensive and comprehensive alternatives were actuarially favorable. The model developed in the study also suggests that given the subsidies to the program, as well as the penalties for late enrollment, not enrolling immediately in a plan in 2006 would have been the optimal choice for just 2.5 percent of the sample. Despite these caveats, however, the proportion of individuals who appear not to have made an optimal choice is relatively small. (pages 413 - 481)
This chapter is available at:
    University Press Scholarship Online


Author Index

Subject Index